– Tumors of the Appendix» MUCOCELE AND PSEUDOMYXOMA
PERITONEI• Morphology.
– Peritoneum– Inflammation
» PERITONEAL INFECTION• Morphology.
» SCLEROSING RETROPERITONITIS» MESENTERIC CYSTS
– Tumors
798
Esophagus
Normal
The esophagus develops from the cranial portion of the foregut and is recognizable by the third week of gestation. The normal esophagus is a hollow, highly distensible muscular tube that extends from the epiglottis in the pharynx, at about the level of the C6 vertebra, to the gastroesophageal junction at the level of the T11 or T12 vertebra. Measuring between 10 and 11 cm in the newborn, it grows to a length of about 25 cm in the adult. For the endoscopist, the esophagus is recorded as the anatomic distance between 15 and 40 cm from the incisor teeth, with the gastroesophageal junction located at the 40-cm point. Several points of luminal narrowing can be identified along its course—proximally at the cricoid cartilage, midway in its course alongside the aortic arch and at the anterior crossing of the left main bronchus and left atrium, and distally where it pierces the diaphragm. Although the pressure in the esophageal lumen is negative compared with the atmosphere, manometric recordings of intraluminal pressures have identified two higher-pressure areas that remain relatively contracted in the resting phase. A 3-cm segment in the proximal esophagus at the level of the cricopharyngeus muscle is referred to as the upper esophageal sphincter (UES). The 2- to 4-cm segment just proximal to the anatomic gastroesophageal junction, at the level of the diaphragm, is referred to as the lower esophageal sphincter (LES). Both "sphincters" are physiologic, in that there are no anatomic landmarks that delineate these higher-pressure regions from the intervening esophageal musculature.
The wall of the esophagus consists of a mucosa, submucosa, muscularis propria, and adventitia, reflecting the general structural organization of the gastrointestinal tract. [1] The
799
mucosa has a smooth, glistening, and pink-tan surface. It has three components: a nonkeratinizing stratified squamous epithelial layer, lamina propria, and muscularis mucosa. The epithelial layer has mature squamous cells overlying basal cells. The basal cells, constituting 10% to 15 % of the mucosal thickness, are reserve cells with great proliferative potential. A small number of specialized cell types, such as melanocytes, endocrine cells, dendritic cells, and lymphocytes, are present in the deeper portion of the epithelial layer. The lamina propria is the nonepithelial portion of the mucosa, above the muscularis mucosae. It consists of areolar connective tissue and contains vascular structures and scattered leukocytes. Finger-like extensions of the lamina propria, called papillae, extend into the epithelial layer. The muscularis mucosae is a delicate layer of longitudinally oriented smooth-muscle bundles.
The submucosa consists of loose connective tissue containing blood vessels, a rich network of lymphatics, a sprinkling of leukocytes with occasional lymphoid follicles, nerve fibers (including the ganglia of the Meissner plexus), and submucosal glands. Submucosal glands connected to the lumen by squamous epithelium-lined ducts are scattered along the entire esophagus but are more concentrated in the upper and lower portions. Their mucin-containing fluid secretions help lubricate the esophagus.
As is true throughout the alimentary tract, the muscularis propria consists of an inner circular and an outer longitudinal coat of smooth muscle with an intervening, well-developed myenteric plexus (Auerbach plexus). The muscularis propria of the proximal 6 to 8 cm of the esophagus also contains striated muscle fibers from the cricopharyngeus muscle. Besides creating a unique histologic interplay of smooth muscle and skeletal muscle fibers, this feature explains why skeletal muscle disorders can cause upper esophageal dysfunction.
In sharp contrast to the rest of the gastrointestinal tract, the esophagus is mostly devoid of a serosal coat. Only small segments of the intra-abdominal esophagus are covered by serosa; the thoracic esophagus is surrounded by fascia that condenses around the esophagus to form a sheathlike structure. In the upper mediastinum, the esophagus is supported by this fascial tissue, which forms a similar sheath around adjacent structures, the great vessels and the tracheobronchial tree. This intimate anatomic proximity to important thoracic viscera is of significance in permitting the ready and widespread dissemination of infections and tumors of the esophagus into the posterior mediastinum. The rich network of mucosal and submucosal lymphatics that runs longitudinally along the esophagus further facilitates spread.
The main functions of the esophagus are to conduct food and fluids from the pharynx to the stomach, to prevent passive diffusion of substances from the food into the blood, and to prevent reflux of gastric contents into the esophagus. These functions require motor activity coordinated with swallowing, namely a wave of peristaltic contraction, relaxation of the LES in anticipation of the peristaltic wave, and closure of the LES after the swallowing reflex. The mechanisms governing this motor function are complex, involving both extrinsic and intrinsic innervation, humoral regulation, and properties of the muscle wall itself.
The control of the lower esophageal sphincter (LES) is critical to esophageal function.[2] Maintenance of sphincter tone is necessary to prevent reflux of gastric contents, which are under positive pressure relative to the esophagus. During deglutition, both active inhibition of the muscularis propria muscle fibers by inhibitory nonadrenergic/noncholinergic neurons and cessation of tonic excitation by cholinergic neurons enable the LES to relax. Many chemical agents (e.g., gastrin, acetylcholine, serotonin, prostaglandin F2α , motilin, substance P, histamine, and pancreatic polypeptide) increase LES tone, while some agents (nitric oxide, vasoactive intestinal peptide) decrease the tone. However, their precise roles in normal esophageal function remain unclear.
Lesions of the esophagus run the gamut from highly lethal cancers to the merely annoying "heartburn" that has affected many a partaker of a large, spicy meal. Esophageal varices, the result of cirrhosis and portal hypertension, are of major importance, since their rupture is frequently followed by massive hematemesis (vomiting of blood) and even death by exsanguination. Esophagitis and hiatal hernias are far more frequent and rarely threaten life. Distressing to the physician is that all disorders of the esophagus tend to produce similar symptoms, namely heartburn, dysphagia, pain, and/or hematemesis.
Heartburn (retrosternal burning pain) usually reflects regurgitation of gastric contents into the lower esophagus. Dysphagia (difficulty in swallowing) is encountered both with deranged esophageal motor function and with diseases that narrow or obstruct the lumen. Pain and hematemesis are sometimes evoked by esophageal disease, particularly by those lesions associated with inflammation or ulceration of the esophageal mucosa. The clinical diagnosis of esophageal disorders often requires specialized procedures such as esophagoscopy, radiographic barium studies, and manometry.
Congenital Anomalies
Ectopic tissue rests are not uncommon in the esophagus. The most common is ectopic gastric mucosa in the upper third of the esophagus ("inlet patch"), occurring in up to 2% of individuals. Sebaceous glands or ectopic pancreatic tissue are much less frequent. The acid secretions of the ectopic gastric mucosa or pancreatic enzymatic secretions can produce localized inflammation and discomfort.
Embryologic formation of the foregut can also give rise to congenital cysts. These are usually duplication cysts, containing double smooth muscle layers and derived from the lower esophagus in 60% of cases. Rarely, bronchial or parenchymal pulmonary tissue may arise from the upper gut and is denoted bronchogenic cyst or pulmonary sequestration, respectively. These lesions usually present as masses. Lastly, impaired formation of the diaphragm may permit herniation of abdominal viscera into the thorax. When severe, this lesion is incompatible with life, since the lungs are severely hypoplastic at the time of birth. This condition is to be distinguished from hiatal hernias, to be discussed presently.
800
ATRESIA AND FISTULAS
Although developmental defects in the esophagus are uncommon, they must be corrected early because they are incompatible with life. Because they cause immediate regurgitation when feeding is attempted, they are usually discovered soon after birth. Absence (agenesis) of the esophagus is extremely rare; much more common are atresia and fistula formation ( Fig. 17-1 ). In atresia, a segment of the esophagus is represented by only a thin, noncanalized cord, with a proximal blind pouch connected to the pharynx
and a lower pouch leading to the stomach. Atresia is most commonly located at or near the tracheal bifurcation. It rarely occurs alone, but is usually associated with a fistula connecting the lower or upper pouch with a bronchus or the trachea. Associated anomalies include congenital heart disease, neurologic disease, genitourinary disease, and other gastrointestinal malformations. Atresia sometimes is associated with the presence of a single umbilical artery.[3] Aspiration and paroxysmal suffocation from food are obvious hazards; pneumonia and severe fluid and electrolyte imbalances may also occur.
WEBS, RINGS, AND STENOSIS
Esophageal mucosal webs are uncommon ledgelike protrusions of the mucosa into the esophageal lumen. These are semicircumferential, eccentric, and most common in the upper esophagus. Well-developed webs rarely protrude more than 5 mm into the lumen, with a thickness of 2 to 4 mm. The webs consist of squamous mucosa and a vascularized submucosal core. Webs can be congenital in origin, or they may arise in association with long-standing reflux esophagitis, chronic graft-versus-host disease (GVHD), or blistering skin diseases. When an upper esophageal web is accompanied by an iron-deficiency anemia, glossitis, and cheilosis, the condition is referred to as the Paterson-Brown-Kelly or Plummer-Vinson syndrome, with an attendant risk for postcricoid esophageal carcinoma.
Esophageal rings are concentric plates of tissue protruding into the lumen of the distal esophagus. One occurring above the squamocolumnar junction of the esophagus and stomach is referred to as an A ring. One located at the squamocolumnar
Figure 17-1 Esophageal atresia and tracheoesophageal fistula. A, Blind upper and lower esophageal segments. B, Fistula between blind upper segment and trachea. C, Blind upper segment, fistula between blind lower segment and trachea. D, Blind upper segment only. E, Fistula between patent esophagus and trachea. Type C is the most common variety. (Adapted from Morson BC, and Dawson IMP, eds., Gastrointestinal Pathology. Oxford, Blackwell Scientific Publications, 1972, p. 8.)
junction of the lower esophagus is designated a Schatzki ring or a B ring. Histologically, these rings consist of mucosa, submucosa, and sometimes a hypertrophied muscularis propria. Schatzki rings may have columnar gastric epithelium on their undersurface.
Esophageal webs and rings are encountered most frequently in women over age 40 and are of uncertain etiology. Episodic dysphagia is the main symptom associated with webs and rings, usually provoked when an individual bolts solid food. Pain is infrequent.
Esophageal stenosis consists of fibrous thickening of the esophageal wall, particularly the submucosa, with atrophy of the muscularis propria. The lining epithelium is usually thin and sometimes ulcerated. Although occasionally of congenital origin, stenosis is more frequently the result of severe esophageal injury with inflammatory scarring, as from gastroesophageal reflux, radiation, scleroderma, or caustic injury. Stenosis usually develops in adulthood and becomes manifest by progressive dysphagia, at first to solid foods only but eventually to all foods, which constitutes the major symptom. In severe stenosis, virtually total obstruction may result.
Lesions Associated with Motor Dysfunction
Coordinated motor function is critical to proper function of the esophagus; gravity alone is not sufficient to move food from the pharynx to the stomach, nor to prevent reflux of gastric contents—witness the blissful suckling of the supine infant. The major entities (achalasia, hiatal hernia, diverticulum and Mallory-Weiss tear) that are caused by or induce motor dysfunction of the esophagus are diagrammed in Figure 17-2 .
ACHALASIA
Achalasia means "failure to relax." It is characterized by three major abnormalities: (1) aperistalsis, (2) partial or incomplete relaxation of the LES with swallowing, and (3) increased resting tone of the LES. The pathogenesis of primary
Figure 17-2 Major conditions associated with esophageal motor dysfunction.
achalasia is poorly understood. It is thought to involve dysfunction of inhibitory neurons containing nitric oxide and vasoactive intestinal polypeptide in the distal esophagus.[4] [5] Degenerative changes in neural innervation, either intrinsic to the esophagus or in the extraesophageal vagus nerves and the dorsal motor nucleus of the vagus, may also occur. Secondary achalasia may arise in Chagas disease, in which Trypanosoma cruzi causes destruction of the myenteric plexus of the esophagus, duodenum, colon, and ureter, with resultant dilation of these viscera. Disorders of the dorsal motor nuclei, particularly polio or surgical ablation, can cause an achalasia-like illness, as can diabetic autonomic neuropathy and infiltrative disorders such as malignancy, amyloidosis, and sarcoidosis. In most instances, however, achalasia occurs as a primary disorder of uncertain etiology.
Morphology.
In primary achalasia there is progressive dilation of the esophagus above the level of the LES. The wall of the esophagus may be of normal thickness, thicker than normal owing to hypertrophy of the muscularis, or markedly thinned by dilation. The myenteric ganglia are usually absent from the body of the esophagus, but may or may not be reduced in number in the region of the LES. The mucosal lining may be unaffected, but sometimes inflammation, ulceration, or fibrotic thickening may be evident just above the LES.
Achalasia usually becomes manifest in young adulthood, but may appear in infancy or childhood. The classic clinical symptom of achalasia is progressive dysphagia. Nocturnal regurgitation and aspiration of undigested food may occur. The most serious aspect of this condition is the hazard of developing esophageal squamous cell carcinoma, said to occur in about 5% of patients, typically at an earlier age than those without this disease. Other complications include Candida esophagitis, lower esophageal diverticula (see below), and aspiration with pneumonia or airway obstruction.
HIATAL HERNIA
Hiatal hernia is characterized by separation of the diaphragmatic crura and widening of the space between the muscular crura and the esophageal wall. Two anatomic patterns are recognized (see Fig. 17-2 ): the axial, or sliding hernia, and the nonaxial, or paraesophageal hiatal hernia. The sliding hernia constitutes 95% of cases; protrusion of the stomach above the diaphragm creates a bell-shaped dilation, bounded below by the diaphragmatic narrowing. In paraesophageal hernias, a separate portion of the stomach, usually along the greater curvature, enters the thorax through the widened foramen.
The cause of hiatal hernia is unknown. It is not clear whether it is a congenital malformation or is acquired during life. Based on barographic studies, hiatal hernias are reported
802
in 1% to 20% of adult subjects, with incidence increasing with age. However, hiatal hernias are well recognized in infants and children. Only about 9% of adults with a sliding hernia suffer from heartburn or regurgitation of gastric juices into the mouth. These symptoms are attributed to incompetence of the LES and are accentuated by positions favoring reflux (bending forward, lying supine) and obesity.
Complications of hiatal hernias are numerous. Both types may ulcerate, causing bleeding and perforation. Paraesophageal hernias can become strangulated or obstructed, and early surgical repair has been advocated. Reflux esophagitis (discussed later) is frequently seen in association with sliding hernias, but compromise of the LES with regurgitation of peptic juices into the esophagus is probably the result of, rather than the cause of, a sliding hernia. The uncommon paraesophageal hernias may be caused by previous surgery, including operations for sliding hernia.
DIVERTICULA
A diverticulum is an outpouching of the alimentary tract that contains all visceral layers; a false diverticulum denotes an outpouching of mucosa and submucosa only ( Fig. 17-2 ). True diverticula are usually discovered in later life and may develop in three regions of the esophagus:
• Zenker diverticulum (pharyngoesophageal diverticulum) immediately above the UES • Traction diverticulum near the midpoint of the esophagus • Epiphrenic diverticulum immediately above the LES.
Disordered cricopharyngeal motor dysfunction with or without gastroesophageal reflux disease (GERD) and diminished luminal size of the UES are implicated in the genesis of Zenker diverticulum. Scarring resulting from mediastinal lymphadenitis (as from tuberculosis) was presumed to be a cause of traction on the esophagus that gave rise to mid-esophageal diverticula. However, arguments have been advanced in favor of traction diverticula actually arising from motor dysfunction or being a congenital lesion. Dyscoordination of peristalsis and LES relaxation are the proposed cause of epiphrenic diverticula.
Zenker diverticula may reach several centimeters in size and can accumulate significant amounts of food. Typical symptoms include dysphagia, food regurgitation, and a mass in the neck; aspiration with resultant pneumonia is a significant risk. While midesophageal diverticula are generally asymptomatic, epiphrenic diverticula can give rise to nocturnal regurgitation of massive amounts of fluid.
LACERATIONS (MALLORY-WEISS SYNDROME)
Longitudinal tears in the esophagus at the esophagogastric junction or gastric cardia are termed Mallory-Weiss tears and are believed to be the consequence of severe retching or vomiting.[6] They are encountered most commonly in alcoholics, in whom they are attributed to episodes of excessive vomiting in the setting of an alcoholic stupor. Normally, a reflex relaxation of the musculature of the gastrointestinal tract precedes the antiperistaltic wave of contraction. During episodes of prolonged vomiting, it is speculated that this reflex relaxation fails to occur. The refluxing gastric contents suddenly overwhelm the contraction of the musculature at the gastric inlet, and massive dilation with tearing of the stretched wall ensues. Since these tears may occur in persons who have no history of vomiting or alcoholism, other mechanisms must exist; underlying hiatal hernia is a known predisposing factor.
Morphology.
The linear irregular lacerations are oriented in the axis of the esophageal lumen and are several millimeters to several centimeters in length. They are usually found astride the esophagogastric junction or in the proximal gastric mucosa ( Fig. 17-3 ). The tears may involve only the mucosa or may penetrate deeply enough to perforate the wall. The histology is not distinctive and reflects trauma accompanied by fresh hemorrhage and a nonspecific inflammatory response. Infection of the mucosal defect may lead to an inflammatory ulcer or to mediastinitis.
Esophageal lacerations account for 5% to 10% of bleeding episodes in the upper gastrointestinal tract. Most often, bleeding is not profuse and ceases without surgical intervention, although massive hematemesis may occur. Supportive therapy, such as vasoconstrictive medications and transfusions, and sometimes balloon tamponade, is usually all that is required. Healing tends to be prompt, with minimal to no residua. The rare instance of esophageal rupture is known as Boerhaave syndrome and may be a catastrophic event.
Esophageal Varices
Regardless of cause, portal hypertension, when sufficiently prolonged or severe, induces the formation of collateral bypass channels wherever the portal and caval systems communicate. The pathogenesis of portal hypertension and the
Figure 17-3 Esophageal laceration (Mallory-Weiss tears). Gross view demonstrating longitudinal lacerations extending from esophageal mucosa into stomach mucosa (arrow). (Courtesy of Dr. Richard Harruff, King County Medical Examiner's Office, Seattle, WA.)
803
locations of these bypasses are considered in Chapter 18 . Here we are concerned with the collaterals that develop in the region of the lower esophagus when portal blood flow is diverted through the coronary veins of the stomach into the plexus of esophageal subepithelial and submucosal veins, thence into the azygos veins, and eventually into the
systemic circulation. The increased pressure in the esophageal plexus produces dilated tortuous vessels called varices. Varices develop in 90% of cirrhotic patients and are most often associated with alcoholic cirrhosis. Worldwide, hepatic schistosomiasis is the second most common cause of variceal bleeding.
Morphology.
Varices appear as tortuous dilated veins lying primarily within the submucosa of the distal esophagus and proximal stomach; venous channels directly beneath the esophageal epithelium may also become massively dilated. The net effect is irregular protrusion of the overlying mucosa into the lumen, although varices are collapsed in surgical or postmortem specimens ( Fig. 17-4 A ). When the varix is unruptured, the mucosa may be normal, but often it is eroded and inflamed because of its exposed position. Variceal rupture produces massive hemorrhage into the lumen, as well as suffusion of the esophageal wall with blood. In this instance the overlying mucosa appears ulcerated and necrotic ( Fig. 17-4 B ). If rupture has occurred in the past, venous thrombosis and superimposed inflammation may be present. Varices can be detected by hepatic venogram ( Fig. 17-4 C ).
Clinical Features.
Varices usually produce no symptoms until they rupture, causing massive hematemesis. Among
Figure 17-4 Esophageal varices. A, A view of the everted esophagus and gastroesophageal junction, showing dilated submucosal veins (varices). The blue-colored varices have collapsed in this postmortem specimen. B, Low-power cross-section of a dilated submucosal varix that has ruptured through the mucosa. A small amount of thrombus is present within the point of rupture. C, Hepatic venogram after injection of dye into portal veins (PV) to show a large tortuous gastroesophageal varix (arrow) extending superiorly from the patent main portal vein. (C, courtesy of Dr. Emily Sedgwick, Brigham and Women's Hospital,
patients with advanced cirrhosis of the liver, half the deaths result from rupture of a varix. Some patients die as a direct consequence of the hemorrhage and others of the hepatic coma triggered by the hemorrhage. It must be remembered, however, that even when varices are present, they account for less than half of all episodes of hematemesis. Collectively, concomitant gastritis, esophageal laceration, or peptic ulcers are more common causes. Factors leading to rupture of a varix are unclear: silent inflammatory erosion of overlying thinned mucosa, increased tension in progressively dilated veins, and vomiting with increased vascular hydrostatic pressure are likely to play roles. Once begun, the hemorrhage rarely subsides spontaneously, and endoscopic injection of thrombotic agents ("sclerotherapy") or balloon tamponade is usually required. Forty to fifty percent of patients die in the first bleeding episode. Among those who survive, rebleeding occurs in over half within 1 year, with a similar rate of mortality for each episode.
Esophagitis
Inflammation of the esophageal mucosa is known as esophagitis. Injury to the esophageal mucosa with subsequent inflammation is common worldwide. In the United States and other Western countries, esophagitis is present in about 5% of the adult population; much higher prevalence is encountered in selected regions such as northern Iran and portions of China. Esophagitis may be caused by a variety of physical, chemical, or biologic agents. We review the common ones encountered in the clinical practice.
Reflux of gastric contents into the lower esophagus is the most important cause of esophagitis. Many causative factors are involved:[7]
• Decreased efficacy of esophageal antireflux mechanisms, particularly LES tone. Central nervous system depressants, hypothyroidism, pregnancy, systemic sclerosing disorders, alcohol or tobacco exposure, or the presence of a nasogastric tube may be contributing causes. However, in most instances no antecedent etiology is identified. • Presence of a sliding hiatal hernia • Inadequate or slowed esophageal clearance of refluxed material
• Delayed gastric emptying and increased gastric volume, contributing to the volume of refluxed material • Reduction in the reparative capacity of the esophageal mucosa by protracted exposure to gastric juices.
Any one of these influences may assume primacy in an individual case, but more than one is likely to be involved in most instances. The action of gastric juices is critical to the development of esophageal mucosal injury; in severe cases refluxed bile from the duodenum also may contribute to the mucosal disruption.
Morphology.
The anatomic changes depend on the causative agent and on the duration and severity of the exposure. Simple hyperemia ("redness") may be the only alteration. In uncomplicated reflux esophagitis, three histologic features are characteristic ( Fig. 17-5 ):
• The presence of inflammatory cells, including eosinophils, neutrophils, and excessive numbers of lymphocytes, in the squamous epithelial layer • Basal zone hyperplasia exceeding 20% of the epithelial thickness • Elongation of lamina propria papillae with capillary congestion, extending into the top third of the epithelial layer.
Infiltrates of intraepithelial eosinophils are believed to be an early histologic abnormality, since they occur even in the absence of basal zone hyperplasia. Intraepithelial neutrophils, on the other hand, are markers of more severe injury such as ulceration rather than reflux esophagitis per se.
Clinical Features.
Although largely limited to adults over age 40, reflux esophagitis is occasionally seen in infants and children. The clinical manifestations consist principally of dysphagia, heartburn, and sometimes regurgitation of a sour brash, hematemesis, or melena. The severity of symptoms is not closely related to the presence or degree of histologic esophagitis; most people experience reflux symptoms without damage to the distal esophageal mucosa, due to the short duration of the reflux. Anatomic damage appears best correlated with prolonged exposure of the lower esophagus to refluxed material. Rarely, chronic symptoms are punctuated by attacks of severe chest pain that may be mistaken for a "heart attack." The potential consequences of severe reflux esophagitis are bleeding, ulceration, development of stricture, and a tendency to develop Barrett esophagus, with its attendant risks.
Figure 17-5 Reflux esophagitis. Low-power view of the superficial portion of the mucosa. Numerous eosinophils within the squamous epithelium, elongation of the lamina propria papillae, and basal zone hyperplasia are present.
BARRETT ESOPHAGUS
Barrett esophagus is a complication of long-standing gastroesophageal reflux, occurring over time in up to 10% of patients with symptomatic gastroesophageal reflux disease (GERD). It is the single most important risk factor for esophageal adenocarcinoma. In Barrett esophagus, the distal squamous mucosa is replaced by metaplastic columnar epithelium, as a response to prolonged injury. Two criteria are required for the diagnosis of Barrett esophagus: (1) endoscopic evidence of columnar epithelial lining above the gastroesophageal junction and (2) histologic evidence of intestinal metaplasia in the biopsy specimens from the columnar epithelium.[8] Barrett esophagus is further classified as long segment (extending cephalad more than 3 cm from the manometric gastroesophageal junction) or short segment (extending less than 3 cm cephalad). Barrett esophagus patients tend to have a long history of heartburn and other reflux symptoms and appear to have more massive reflux with more and longer reflux episodes than most reflux patients. It is unknown why the columnar epithelium develops in some patients with reflux and not in others.
The pathogenesis of Barrett esophagus remains unclear, but it appears to result from an alteration in the differentiation program of stem cells of the esophageal mucosa.[9] The concept of "intestinal metaplasia" in Barrett esophagus may be not entirely correct, since true absorptive enterocytes are not observed. Rather, admixed with intestinal mucin-secreting goblet cells are columnar cells exhibiting both secretory and absorptive ultrastructural features; this is a phenotype not observed elsewhere in the alimentary tract. Nevertheless the term "intestinal metaplasia" continues to be used to denote the altered histology of the mucosa.
Morphology.
Barrett esophagus is recognized as a red, velvety mucosa located between the smooth, pale pink esophageal squamous mucosa and the lusher light brown gastric mucosa. It may exist as tongues or patches (islands) extending up from the gastroesophageal junction or as a broad irregular circumferential band displacing the squamocolumnar junction several
centimeters cephalad ( Fig. 17-6 ). A small zone of metaplastic mucosa may be present only at the esophagogastric junction (short-segment
805
Figure 17-6 Barrett esophagus. A, B, Gross view of distal esophagus (top) and proximal stomach (bottom), showing A, the normal gastroesophageal junction (arrow) and B, the granular zone of Barrett esophagus (arrow). C, Endoscopic view of Barrett esophagus showing red velvety gastrointestinal mucosa extending from the gastroesophageal orifice. Note the paler squamous esophageal mucosa.
Barrett mucosa), sometimes less than 0.5 cm in length. Microscopically, the esophageal squamous epithelium is replaced by metaplastic columnar epithelium, complete with surface epithelium and mucosal glands. The metaplastic mucosa may contain only gastric surface and glandular mucus-secreting cells, making clinical distinction from a hiatal hernia difficult. Definitive diagnosis is made when the columnar mucosa contains intestinal goblet cells ( Fig. 17-7 ).
Critical to the pathologic evaluation of patients with Barrett mucosa is the search for dysplasia, the presumed precursor of malignancy, in columnar epithelium with intestinal metaplasia. Dysplasia is recognized by the presence of cytologic and architectural abnormalities in the columnar epithelium, consisting of enlarged, crowded, and stratified hyperchromatic nuclei and loss of intervening stroma between adjacent glandular structures.[10] Dysplasia is classified as low-grade or high-grade, with the predominant distinction being a basal orientation of all nuclei in low-grade dysplasia versus nuclei consistently reaching the apex of epithelial cells in high-grade dysplasia. Approximately 50% of patients with high-grade dysplasia may already have adjacent adenocarcinoma, according to some studies;[11] therefore, persistent high-grade dysplasia demands clinical intervention.
Most of the patients with first diagnosis of Barrett esophagus are between ages 40 and 60, although children can also occasionally develop this condition. The incidence is highest among white males. In addition to the symptoms of reflux esophagitis, Barrett esophagus is clinically significant due to the secondary complications of local ulceration with bleeding and stricture. Of greatest importance is the development of adenocarcinoma, which, in patients with over 3 cm of Barrett mucosa, occurs at an estimated 30- to 40-fold increased rate over the general population. The presence of short-segment Barrett esophagus also appears to impart risk for adenocarcinoma, but at what rate is not yet known.
INFECTIOUS AND CHEMICAL ESOPHAGITIS
In addition to gastroesophageal reflux (which is, in fact, a chemical injury), esophageal inflammation may have many origins, as follows:
• Ingestion of mucosal irritants such as alcohol, corrosive acids or alkalis (in suicide attempts), excessively hot fluids (e.g., hot tea in Iran); or heavy smoking • Cytotoxic anticancer therapy, with or without superimposed infection • Infection following bacteremia or viremia; herpes simplex viruses and cytomegalovirus (CMV) are common offenders in immunosuppressed patients
806
• Fungal infection in debilitated or immunosuppressed patients or during broad-spectrum antimicrobial therapy; candidiasis by far the most common; mucormycosis and aspergillosis may occur • Uremia in the setting of renal failure.
Figure 17-7 Barrett esophagus. Microscopic view showing squamous mucosa and intestinal-type columnar epithelial cells (goblet cells) in a glandular mucosa.
Esophageal inflammation may also arise following radiation treatment and in association with systemic graft-versus-host disease (GVHD), autoimmune diseases, or the desquamative dermatologic conditions of pemphigoid and epidermolysis bullosa. On rare occasion, esophagitis occurs in the setting of Crohn disease.
Morphology.
Esophagitis of different causes have their own characteristic features; the final common pathway for all is severe acute inflammation, superficial necrosis and ulceration with the formation of granulation tissue, and eventual fibrosis.
• In candidiasis, patches of the entire esophagus become covered by adherent, gray-white pseudo-membranes teeming with densely matted fungal hyphae. • Herpesviruses typically cause punched-out ulcers; the nuclear inclusions of herpesvirus are found in a narrow rim of degenerating epithelial cells at the margin of the ulcer. CMV causes liner ulceration of the esophageal mucosa; the histologic findings of CMV-associated change with both intranuclear and cytoplasmic inclusions are found in capillary endothelium and stromal cells in the base of the ulcer. In both forms of infection, immunohistochemical staining for virus-specific antigens provides a sensitive and specific diagnostic tool if routine histology is equivocal. • Pathogenic bacteria account for 10% to 15% of cases of infectious esophagitis and exhibit bacterial invasion of the lamina propria with necrosis of the squamous epithelium. • Injury induced by chemicals (lye, acids, detergents) may produce only mild erythema and edema, sloughing of the mucosa, or outright necrosis of the entire esophageal wall. Localized esophageal ulceration may result from pharmaceutical tablets or capsules "sticking" in the esophagus. • Following irradiation of the esophagus, submucosal and mural blood vessels exhibit marked intimal proliferation with luminal narrowing. The submucosa becomes severely fibrotic, and the mucosa exhibits atrophy, with flattening of the papillae and thinning of the epithelium. • GVHD shares features with the skin manifestations (e.g., apoptosis of basal epithelial cells, separation of epithelium and lamina propria, atrophy, and fibrosis of the lamina propria with minimal inflammation).
Clinical Features.
Infections of the esophagus may occur in otherwise healthy individuals, but most often occur in the debilitated or immunosuppressed. Chemical injury in children is usually accidental, as opposed to attempted suicide, which is an adult phenomenon.
Tumors
BENIGN TUMORS
Benign tumors of the esophagus are mostly mesenchymal in origin and lie within the esophageal wall. Most common are benign tumors of smooth muscle origin, called leiomyomas. Fibromas, lipomas, hemangiomas, neurofibromas, and lymphangiomas may also arise. Mucosal polyps are usually composed of a combination of fibrous, vascular, or adipose tissue covered by an intact mucosa, known as fibrovascular polyps or pedunculated lipomas. Squamous papillomas are sessile lesions with a central core of connective tissue and a hyperplastic papilliform squamous mucosa. When the papilloma is associated with human papillomavirus (HPV) infection, the term condyloma applies. In rare instances a mesenchymal mass of inflamed granulation tissue, called an inflammatory polyp, may resemble a malignant lesion, hence its alternative name inflammatory pseudotumor.
MALIGNANT TUMORS
In the United States, carcinomas of the esophagus represent about 6% of all cancers of the gastrointestinal tract but cause a disproportionate number of cancer deaths. They remain asymptomatic during much of their development and are often discovered too late to permit cure. With rare exceptions, malignant esophageal tumors arise from the epithelial layer. In the United States, most esophageal cancers used to be of squamous cell origin, but the incidence of these tumors has declined with a steady increase of adenocarcinomas. Worldwide, squamous cell cancers constitute 90% of esophageal cancers, but in the United States squamous cell carcinoma and adenocarcinoma exhibit comparable incidence rates. Rare tumors (undifferentiated, carcinoid, malignant melanoma, lymphoma, sarcoma, and adenocarcinomas arising from the submucosal glands) are not discussed here.
Squamous Cell Carcinoma
Squamous cell carcinoma is the most common type of carcinoma in the esophagus. Most squamous cell carcinomas occur in adults over age 50. The male-to-female ratio varies, in different studies, from 2:1 to as high as 20:1. While squamous cell carcinoma of the esophagus occurs throughout the world, its incidence varies widely between countries and within regions of the same country. The regions with higher incidence are Iran, central China, South Africa, and southern Brazil, where annual incidence rates are as high as 100 per 100,000, with deaths from cancer of the esophagus constituting over 20% of all cancer deaths. Other areas of high incidence include Puerto Rico and Eastern Europe. In the United States, it affects from 2 to 8 persons per 100,000 yearly and is predominantly a disease of adult males (the male-to-female ratio is 4:1). Blacks throughout the world are at higher risk than are whites; incidence in this group in the United States is fourfold higher than for U.S. whites.
Etiology and Pathogenesis.
The marked differences in epidemiology strongly implicate dietary and environmental factors ( Table 17-1 ), with a contribution from genetic predisposition.[12] The majority of cancers in Europe and the United
TABLE 17-1 -- Factors Associated with the Development of Squamous Cell Carcinoma of the Esophagus
Dietary
Deficiency of vitamins (A, C, riboflavin, thiamine, pyridoxine)
Deficiency of trace elements (zinc, molybdenum)
Fungal contamination of foodstuffs
High content of nitrites/nitrosamines
Betel chewing
Lifestyle
Burning-hot beverages or food
Alcohol consumption
Tobacco use
Urban environment
Esophageal Disorders
Long-standing esophagitis
Achalasia
Plummer-Vinson syndrome
Genetic Predisposition
Long-standing celiac disease
Ectodermal dysplasia
Epidermolysis bullosa
Racial disposition
States are attributable to alcohol and tobacco usage. Some alcoholic drinks contain significant amounts of such carcinogens as polycyclic hydrocarbons, fuel oils, and nitrosamines, along with other mutagenic compounds. Nutritional deficiencies associated with alcoholism may contribute to the process of carcinogenesis.
Alcohol and tobacco cannot be invoked as risk factors in many high-incidence regions of the world. The presence of carcinogens, such as fungus-contaminated and nitrosamine-
containing foodstuffs in China, may play a significant role in the extraordinary high incidence of carcinoma in this region. Dietary deficiencies in vitamins and essential metals have been documented in China and South Africa. Human papillomavirus DNA is found frequently in esophageal squamous cell carcinomas from high-incidence regions, but is infrequent in cancer-bearing patients in North America. [13]
Based on the above considerations, dietary and environmental factors have been proposed to increase risk, with nutritional deficiencies acting as promoters or potentiators of the tumorigenic effects of environmental carcinogens. For example, methylating nitroso compounds in the diet and in tobacco smoke may be the reason for the broad spectrum of p53 point mutations present in over half of esophageal cancers. Other genetic alterations, such as mutations in p16INK4, and amplification of CYCLIN D1, C-MYC, and epithelial growth factor receptor (EGFR), are prevalent in these cancers as well. This is in keeping with the concept that stepwise acquisition and accumulation of genetic alterations ultimately give rise to cancer.[14] Notably rare in esophageal squamous cell carcinomas are K-RAS and adenomatous polyposis coli (APC) mutations.
Finally, the chronic esophagitis so commonly observed in persons living in areas of high incidence may itself be the result of sustained exposure to the carcinogens listed earlier. This chronic esophagitis results in an increased epithelial cell turnover, which, over a length of time in a continuously carcinogenic environment, progresses to dysplasia and eventually to carcinoma. The rate of progression along the chronic esophagitis-dysplasia-cancer sequence may well be modified or modulated by genetic or racial factors.
Morphology.
Like squamous cell carcinomas arising in other locations, those of the esophagus begin as apparent in situ lesions (intraepithelial neoplasm or carcinoma in situ). When they become overt, about 20% of these tumors are located in the upper third, 50% in the middle third, and 30% in the lower third of the esophagus. Early lesions appear as small, gray-white, plaque-like thickenings or elevations of the mucosa. In months to years, these lesions become tumorous masses and may eventually encircle the lumen. Three morphologic patterns are described: (1) protruded (60%), a polypoid exophytic lesion that protrudes into the lumen; (2) flat (15%), a diffuse, infiltrative form that tends to spread within the wall of the esophagus, causing thickening, rigidity, and narrowing of the lumen; and (3) excavated (ulcerated, 25%; Fig. 17-8 ), a necrotic cancerous ulceration that excavates deeply into surrounding structures and may erode into the respiratory tree (with resultant fistula and pneumonia) or aorta (with catastrophic exsanguination) or may permeate the mediastinum and pericardium. The fortunate patient is found at the stage of superficial esophageal carcinoma, in which the malignant lesion is confined to the epithelial layer (in situ) or is superficially invading the lamina propria or submucosa ( Fig. 17-9 ).
Most squamous cell carcinomas are moderately to well differentiated. Several histologic variants may be seen, such as verrucous squamous cell carcinoma, spindle cell carcinoma, and basaloid squamous cell carcinoma. Irrespective of their degree of differentiation, most symptomatic tumors are quite large by the time they are diagnosed
and have already invaded the wall or beyond. The rich lymphatic network in the submucosa
Figure 17-8 Large ulcerated squamous cell carcinoma of the esophagus.
808
Figure 17-9 Squamous cell carcinoma of the esophagus: low-power microscopic view showing invasion into the submucosa.
promotes extensive circumferential and longitudinal spread, and intramural tumor cell clusters may often be seen several centimeters away from the main mass. Local extension into adjacent mediastinal structures occurs early and often in this disease, possibly due to
the absence of serosa for most of the esophagus. Tumors located in the upper third of the esophagus also metastasize to cervical lymph nodes; those in the middle third to the mediastinal, paratracheal, and tracheobronchial nodes; and those in the lower third most often spread to the gastric and celiac groups of nodes.
Clinical Features.
Esophageal carcinoma is insidious in onset and produces dysphagia and obstruction gradually and late. Patients subconsciously adjust to their increasing difficulty in swallowing by progressively altering their diet from solid to liquid foods. Extreme weight loss and debilitation result from both the impaired nutrition and the effects of the tumor itself. Hemorrhage and sepsis may accompany ulceration of the tumor. Occasionally, the first alarming symptom of this neoplasm is aspiration of food via a cancerous tracheoesophageal fistula. Although the insidious growth of these neoplasms often leads to large lesions by the time a diagnosis is established, resectability rates have improved modestly (from less than half to over 80%) with the advent of endoscopic screening in patient populations at risk and accurate staging by endoscopic ultrasonography. The five-year survival rate in patients with superficial esophageal carcinoma is about 75%, compared to 25% in patients who undergo "curative" surgery for more advanced disease and 9% for all patients with esophageal squamous cell carcinoma. Local and distant recurrence following surgery is common. The presence of lymph node metastases at the time of resection significantly reduces five-year survival.
Adenocarcinoma
Adenocarcinoma of the esophagus is a malignant epithelial tumor with glandular differentiation. Because of confusion in the past with gastric cancers arising at the gastroesophageal junction, true esophageal adenocarcinomas were thought to be unusual. With increasing recognition of Barrett mucosa, it is apparent that most adenocarcinomas in the lower third of the esophagus are true esophageal cancers, rather than gastric cancers straddling the esophagogastric junction. Accordingly, adenocarcinoma now represents up to half of all esophageal cancers reported in the United States, and the incidence has been increasing in recent decades, particularly among white men. The majority of cases arise from the Barrett mucosa. In rare instances, adenocarcinoma originates from heterotopic gastric mucosa or submucosal glands.
Etiology and Pathogenesis.
The discussion of adenocarcinoma focuses on Barrett esophagus. The lifetime risk for cancer development from Barrett esophagus is approximately 10%. Tobacco exposure and obesity are risk factors, but there is no close association between alcohol ingestion and the development of adenocarcinoma of the esophagus. Helicobacter pylori infection may be a contributing factor, but there is no general agreement about this issue.
Molecular studies have suggested that the pathogenesis of adenocarcinoma from Barrett esophagus is a multistep process with a long latency period associated with many genetic changes. The development of dysplasia seems to be a critical step in this process ( Fig.
17-10 ). Barrett epithelial cells have higher proliferative activity, and dysplastic epithelial cells have lost cell-cycle control. Several growth factors, oncogenes, and tumor suppressor genes are implicated in this process.[15] Overexpression of p53 and an increased proportion of cycling cells are present in the dysplastic epithelium, presumably the result of chronic cell and DNA damage induced by gastric reflux. In high-grade dysplasia, chromosomal abnormalities, such as chromosome 4 amplification, are generally present.[16]
When the dysplastic epithelium develops into adenocarcinoma, additional genetic changes, including nuclear translocation of β-catenin and amplification of c-ERB-B2, are present. Although specific genetic abnormalities associated with the Barrett esophagus-carcinoma transition have not been identified, p53 mutations, along with tetraploidy and aneuploidy, seem to occur early. These genetic alterations may be suitable biomarkers of disease progression.
Morphology.
Adenocarcinomas arising in the setting of Barrett esophagus are usually located in the distal esophagus and may invade the adjacent gastric cardia. Initially appearing as flat or raised patches of an otherwise intact mucosa, they may develop into large nodular masses up to 5 cm in diameter or may exhibit diffusely infiltrative or deeply ulcerative features ( Fig. 17-11 A ). Microscopically, most tumors are mucin-producing glandular tumors exhibiting intestinal-type features ( Fig. 17-11 B ); less often they are made up of diffusely infiltrative signet-ring cells of a gastric type or even poorly differentiated small cell-type tumor. Multiple foci of dysplastic mucosa are frequently adjacent to the tumor, which is the basis for the recommendation of multisite biopsy when performing endoscopic screening for dysplasia and malignancy.
809
Figure 17-10 Transition from Barrett esophagus to adenocarcinoma.
Clinical Features.
Adenocarcinomas arising in Barrett esophagus chiefly occur in patients over age 40, with a median age in the sixth decade. Similar to Barrett esophagus, adenocarcinoma is more common in men than in women, and whites are affected more frequently than blacks, in
contrast to squamous cell carcinomas. As in other forms of esophageal carcinoma, patients usually present because of difficulty swallowing, progressive weight loss, bleeding, chest pain, and vomiting. Long-term symptoms of heartburn, regurgitation, and epigastric pain related to concurrent gastroesophageal reflux disease and sliding hiatal hernias are present in less than half of newly diagnosed patients.
Figure 17-11 Adenocarcinoma of the esophagus. A, Gross view of an ulcerated, exophytic mass at the gastroesophageal junction, arising from the granular mucosa of Barrett esophagus. The gray-white esophageal mucosa is on the top, and the folds of gastric mucosa are below. (A, courtesy of Dr. James Gulizia, Brigham and Women's Hospital, Boston, MA.) B, Microscopic view of malignant intestinal-type glands in adenocarcinoma arising from Barrett esophagus.
The prognosis for esophageal adenocarcinoma is as poor as that for other forms of esophageal cancer, with under 20% overall five-year survival. Identification and resection of early cancers with invasion limited to the mucosa or submucosa improves five-year survival to over 80%. Although dysplasia appears to be a requisite for the development of adenocarcinoma, patients with low-grade dysplasia may not progress to cancer over long periods of follow-up, and apparent regression may occur. Regression or ablation of Barrett esophagus has not yet been shown to eliminate the risk for adenocarcinoma.
810
Stomach
Normal
The stomach develops from the distal part of the foregut. It is a saccular organ with a volume of 1200 to 1500 mL, but a capacity of over 3000 mL. It extends from just left of the midline where it is joined to the esophagus, to just right of the midline where it connects to the duodenum. The concavity of the right, inner curve is called the lesser curvature, and the convexity of the left, outer curve is the greater curvature. An angle along the lesser curve, the incisura angularis, marks the approximate point at which the stomach narrows prior to its junction with the duodenum. The entire stomach is covered by peritoneum; an exaggerated peritoneal fold, the greater omentum, extends beyond the greater curvature to the transverse colon.
The stomach is divided into five anatomic regions ( Fig. 17-12 A ). The cardia is the narrow conical portion of the stomach immediately distal to the gastroesophageal junction. The fundus is the dome-shaped portion of the proximal stomach that extends superolateral to the gastroesophageal junction. The body, or corpus, comprises the remainder of the stomach proximal to the incisura angularis. The stomach distal to this angle is the antrum, demarcated from the duodenum by the muscular pyloric sphincter.
The gastric wall, like the rest of the gastrointestinal tract, consists of mucosa, submucosa, muscularis propria, and serosa. The interior surface of the stomach exhibits coarse rugae (meaning "folds"). These infoldings of mucosa and submucosa extend longitudinally and are most prominent in the proximal stomach, flattening out when the stomach is distended. A finer mosaic-like pattern is delineated by small furrows in the mucosa. The delicate texture of the mucosa is punctuated by millions of gastric foveolae, or pits, leading to the mucosal glands.
The normal gastric mucosa has two compartments: the superficial foveolar (meaning leaflike) compartment and the deeper glandular compartment. The foveolar compartment is relatively uniform throughout the stomach. In contrast, the glandular compartment exhibits major differences in thickness and in glandular composition in different regions of the stomach ( Fig. 17-12 B , C ). The foveolar compartment consists of surface epithelial cells (the foveolar cells) lining the entire mucosal surface as well as the gastric pits. The lush undulation of the mucosal surface and pits imparts the leaflike texture to the gastric mucosa. The tall, columnar mucin-secreting foveolar cells have basal nuclei and crowded, small, relatively clear mucin-containing granules in the supranuclear region. Deeper in the gastric pits are so-called mucous neck cells, which have a lower content of mucin granules and are thought to be the progenitors of both the surface epithelium and the cells of the gastric glands. [17] Mitoses are extremely common in this region, as the entire gastric mucosal surface is totally replaced every 2 to 6 days. The glandular compartment consists of gastric glands, which vary between anatomic regions:
• Oxyntic (also called gastric or fundic) glands are found in the fundus and body and contain parietal cells, chief cells, and scattered endocrine cells. The term oxyntic means acid-forming (derived from Greek, oxynein). • Antral or pyloric glands contain mucus-secreting cells and endocrine cells.
The main cell types in these glands are the following:
• Mucous cells populate the glands of the cardia and antral regions and secrete mucus and pepsinogen II. The mucous neck cells in the glands of the body and fundus secrete mucus as well as group I and II pepsinogens. • Parietal cells line predominantly the upper half of the oxyntic glands in the fundus and body. They are recognizable
811
by their bright eosinophilia on H & E stained preparations, which are attributable to their abundant mitochondria. The apical membrane of the parietal cell is invaginated, forming an extensive intracellular canalicular system complete with microvilli. In the resting state, vesicles lie in close approximation to the canalicular system. These vesicles contain the proton pump, a unique hydrogen-potassium-ATPase (H+ ,K+ -ATPase) that pumps hydrogen across membranes in exchange for potassium ions. Within minutes of parietal cell stimulation, the vesicles fuse with the canalicular system, thereby creating an apically directed acid-secreting membrane of enormous surface area. Parietal cells also secrete intrinsic factor, which binds luminal vitamin B12 in the duodenum and permits its absorption in the ileum. • Chief cells, concentrated more at the base of gastric glands, are responsible for the secretion of the proteolytic proenzymes pepsinogen I and II. Chief cells are notable for their basophilic cytoplasm, and ultrastructurally are classic protein-synthesizing cells, having an extensive rough endoplasmic reticulum, a prominent supranuclear Golgi apparatus, and numerous apical secretory granules. Upon stimulation of chief cells, the pepsinogens contained in the granules are released by exocytosis. The pepsinogens are activated to pepsin by the low luminal pH and inactivated above pH 6.0 upon entry into the duodenum. • Endocrine or enteroendocrine cells are scattered among the epithelial cells of gastric and antral glands. The cytoplasm of these triangular cells contains small brightly eosinophilic granules, which are concentrated on the basal aspect of the cell. These cells can act in an endocrine mode, releasing their products into the circulation, or a paracrine mode, via secretion into the local tissue. In the antral mucosa, most of the endocrine cells are the gastrin-producing cells, or G cells. In the body (gastric) mucosa, the endocrine cells produce histamine, which binds the histamine-2 (H2 ) receptor on the parietal cells to increase acid production. These cells are also referred as enterochromaffin-like (ECL) cells. Other ECL cells in the gastric mucosa include D cells (producing somatostatin) and X cells (producing endothelin). These cells play an important role in modulating acid production.
Figure 17-12 Anatomy and histology of the stomach. A, Gross anatomy. B, Microscopic view of antral mucosa. C, Microscopic view of fundic mucosa.
Gastric Mucosal Physiology
ACID SECRETION
The hallmark of gastric physiology is secretion of hydrochloric acid, divided into three phases.
• The cephalic phase, initiated by the sight, taste, smell, chewing, and swallowing of palatable food, is mediated by vagal activity. • The gastric phase involves stimulation of stretch receptors by gastric distention and is mediated by vagal impulses; it also involves gastrin release from endocrine cells, the G cells, in the antral glands. Gastrin release is promoted by luminal amino acids and peptides and possibly by vagal stimulation.
• The intestinal phase, initiated when food containing digested protein enters the proximal small intestine, involves a number of polypeptides besides gastrin.
All signals converge on the gastric parietal cell to activate the proton pump:
• Acetylcholine released from cephalic-vagal or gastric-vagal afferents stimulates the parietal cell via the muscarine-3 cholinergic receptor, resulting in an increase in cytosolic Ca2+ and subsequent activation of the proton pump. • Gastrin activates a gastrin receptor, resulting in an increase of cytosolic Ca2+ within the parietal cells. • An oxyntic gland ECL cell plays a central role: gastrin and vagal afferents induce the release of histamine from the ECL cell, thereby stimulating the H2 receptor on parietal cells. This pathway is considered to be the most important for activation of the proton pump.
Activation of some receptors on the parietal cell surface inhibit acid production. They include receptors for somatostatin, prostaglandins of the E series, and epidermal growth factor.
MUCOSAL PROTECTION
At maximal secretory rates the intraluminal concentration of hydrogen ion is 3 million times greater than that of the blood and tissues. The "mucosal barrier" protects the gastric mucosa from autodigestion and consists of:
• Mucus secretion: The thin layer of surface mucus in the stomach and duodenum exhibits a diffusion coefficient for H+ that is one quarter that of water. Acid- and pepsin-containing fluid exits the gastric glands as "jets" passing through the surface mucus layer, entering the lumen directly without contacting surface epithelial cells. • Bicarbonate secretion: Surface epithelial cells in both the stomach and duodenum secrete bicarbonate into the boundary zone of adherent mucus, creating an essentially pH-neutral microenvironment immediately adjacent to the cell surface. • The epithelial barrier: Intercellular tight junctions provide a barrier to the back-diffusion of hydrogen ions. Epithelial disruption is followed rapidly by restitution, in which existing cells migrate along the exposed basement membrane to fill in the defects and restore epithelial barrier integrity. • Mucosal blood flow: The rich mucosal blood supply provides oxygen, bicarbonate, and nutrients to epithelial cells and removes back-diffused acid. • Prostaglandin synthesis: Production of prostglandins by the mucosal cells impacts on many other components of mucosal defense. For example, prostglandins favor production of mucus and bicarbonates, and they inhibit acid secretion by parietal cells. In addition, by their vasodilatory action, prostglandins E and I improve mucosal blood flow. Drugs that block postglandin synthesis reduce this cytoprotection and thus promote gastric mucosal injury and ulceration.
When the mucosal barrier is breached, the muscularis mucosa limits injury. Superficial damage limited to the mucosa can heal within hours to days. When damage extends into the submucosa, weeks are required for complete healing. Imperfect as our understanding of these defensive mechanisms may be, they are clearly a physiologic marvel, or our gastric walls would suffer the same fate as a piece of swallowed meat.
In addition to the well-characterized barrier function and digestive function of the gastric mucosa, mucosal endocrine
812
cells also produce hormones that are involved in growth regulation. Ghrelin is a recently identified growth hormone that regulates body growth and appetite via a possible effect on the gastrointestinal-hypothalamic-pituitary axis.[18]
Pathology
Gastric lesions are frequent causes of clinical disease. In Western industrialized nations, peptic ulcers develop in up to 10% of the general population at some point during life. Chronic infection of the gastric mucosa by the bacterium H. pylori is the most common infection worldwide. Lastly, gastric cancer remains a leading cause of death in the United States, despite its decreasing incidence.
Congenital Anomalies
Heterotopic rests of normal tissue may be present in the stomach, and are usually asymptomatic. With pancreatic heterotopia, nodules of essentially normal pancreatic tissue up to 1 cm in diameter may be present in the gastric submucosa, muscle wall, or at a subserosal location. When in the pylorus, localized inflammation may lead to pyloric obstruction. With gastric heterotopia, small patches of ectopic gastric mucosa in the duodenum or in more distal sites may present as perplexing sources of bleeding, due to peptic ulceration of adjacent mucosa.
Defective closure of the diaphragm leads to weakness or partial to total absence of a region of the diaphragm, usually on the left. Resultant herniation of abdominal contents into the thorax in utero produces a diaphragmatic hernia. Usually, the stomach or a portion of it insinuates into the pouch, but occasionally small bowel and even a portion of the liver accompany it. The herniation may be asymptomatic or may engender potentially lethal respiratory problems in the newborn.
Rarely, and in keeping with the foregut origin of the stomach, a bud of pulmonary tissue complete with bronchial structures may be attached to the stomach. This pulmonary sequestrum may become infected or present as a mass lesion.
Congenital hypertrophic pyloric stenosis is encountered in infants as a disorder that affects males three to four times more often than females, occurring in 1 in 300 to 900 live births. Familial occurrence implicates a multifactorial pattern of inheritance; monozygotic twins have a high rate of concordance of the condition. Pyloric stenosis also may occur in association with Turner syndrome, trisomy 18, and esophageal atresia. Regurgitation and persistent, projectile, nonbilious vomiting usually appear in the second or third week of life. Physical examination reveals visible peristalsis and a firm, ovoid palpable mass in the region of the pylorus or distal stomach, the result of hypertrophy, and possibly hyperplasia, of the muscularis propria of the pylorus. Edema and inflammatory changes in the mucosa and submucosa may aggravate the narrowing. Surgical muscle splitting is curative.
Acquired pyloric stenosis in adults is one of the long-term risks of antral gastritis or peptic ulcers close to the pylorus. Carcinomas of the pyloric region, lymphomas, or adjacent carcinomas of the pancreas are more ominous causes. In these cases, inflammatory fibrosis or malignant infiltration narrow the pyloric channel, producing pyloric outlet obstruction. In rare instances, hypertrophic pyloric stenosis is the result of prolonged pyloric spasm.
Gastritis
This diagnosis is both overused and often missed—overused when it is applied loosely to any transient upper abdominal complaint in the absence of validating evidence, and missed because most patients with chronic gastritis are asymptomatic. Gastritis is simply defined as inflammation of the gastric mucosa. It is a histologic diagnosis. Inflammation may be predominantly acute, with neutrophilic infiltration, or chronic, with lymphocytes and/or plasma cells predominating and associated intestinal metaplasia and atrophy.[19]
ACUTE GASTRITIS
Acute gastritis is an acute mucosal inflammatory process, usually of a transient nature. The inflammation may be accompanied by hemorrhage into the mucosa and, in more severe circumstances, by sloughing of the superficial mucosa (mucosal erosion). This severe erosive form of the disease is an important cause of acute gastrointestinal bleeding.
Pathogenesis.
The pathogenesis is poorly understood, in part because the normal mechanisms for gastric mucosal protection are not entirely clear. Acute gastritis is frequently associated with:
• Heavy use of nonsteroidal anti-inflammatory drugs (NSAIDs), particularly aspirin • Excessive alcohol consumption
• Heavy smoking • Treatment with cancer chemotherapeutic drugs • Uremia • Systemic bacterial or viral infections (e.g., salmonellosis or CMV infection) • Severe stress (e.g., trauma, burns, surgery) • Ischemia and shock • Suicidal attempts, as with acids and alkali • Gastric irradiation or freezing • Mechanical trauma (e.g., nasogastric intubation) • Distal gastrectomy.
One or more of the following influences are thought to be operative in these varied settings: increased acid secretion with back-diffusion, decreased production of bicarbonate buffer, reduced blood flow, disruption of the adherent mucus layer, and direct damage to the epithelium. Not surprisingly, mucosal insults can act synergistically. Thus, ischemic injury worsens the effects of back-diffusion of hydrogen ions. Other mucosal insults have been identified, such as regurgitation of detergent bile acids and lysolecithins from the proximal duodenum, and inadequate mucosal synthesis of prostaglandins. It must be emphasized that a substantial portion of patients has idiopathic gastritis, with no associated disorders.
813
Morphology.
In the mildest form of acute gastritis, the lamina propria exhibits only moderate edema and slight vascular congestion. The surface epithelium is intact, and scattered neutrophils are present among the surface epithelial cells or within the epithelial layer and lumen of mucosal glands. The presence of neutrophils above the basement membrane (within the epithelial space) is abnormal and signifies active inflammation ("activity"). With more severe mucosal damage, erosion and hemorrhage develop. "Erosion" denotes loss of the superficial epithelium, generating a defect in the mucosa that does not cross the muscularis mucosa. It is accompanied by a robust acute inflammatory infiltrate and extrusion of a fibrin-containing purulent exudate into the lumen. Hemorrhage may occur independently, generating punctate dark spots in an otherwise hyperemic mucosa or in association with erosion. Concurrent erosion and hemorrhage is termed acute erosive hemorrhagic gastritis ( Fig. 17-13 A ). Large areas of the gastric mucosa may be denuded, but the involvement is superficial and rarely affects the entire depth of the mucosa ( Fig. 17-13 B ). These lesions are but one step removed from stress ulcers, to be described later.
Clinical Features.
Depending on the severity of the anatomic changes, acute gastritis may be entirely asymptomatic; may cause variable epigastric pain, nausea, and vomiting; or may present
with overt hemorrhage, massive hematemesis, melena, and potentially fatal blood loss. Overall, it is one of the major causes of massive hematemesis, as in alcoholics. In particular settings, the condition is quite common. As many as 25% of persons who take daily aspirin for rheumatoid arthritis develop acute gastritis, many with bleeding.
CHRONIC GASTRITIS
Chronic gastritis is defined as the presence of chronic mucosal inflammatory changes leading eventually to mucosal atrophy and intestinal metaplasia, usually in the absence of erosions. The
Figure 17-13 Acute gastritis. A, Gross view showing punctate erosions in an otherwise unremarkable mucosa; adherent blood is dark due to exposure to gastric acid. B, Low-power microscopic view of focal mucosal disruption with hemorrhage; the adjacent mucosa is normal.
epithelial changes may become dysplastic and constitute a background for the development of carcinoma. Chronic gastritis is notable for distinct causal subgroups and for patterns of histologic alterations that vary in different parts of the world. In the Western world, the prevalence of histologic changes indicative of chronic gastritis in the later decades of life is higher than 50%.
Pathogenesis.
The major etiologic associations of chronic gastritis are:
• Chronic infection by H. pylori • Immunologic (autoimmune), in association with pernicious anemia • Toxic, as with alcohol and cigarette smoking • Postsurgical, especially following antrectomy with gastroenterostomy with reflux of bilious duodenal secretions • Motor and mechanical, including obstruction, bezoars (luminal concretions), and gastric atony • Radiation • Granulomatous conditions (e.g., Crohn disease)
Helicobacter pylori Infection and Chronic Gastritis.
By far the most important etiologic association with chronic gastritis is chronic infection by the bacillus H. pylori. The link was discovered in 1983, when the bacterium was called Campylobacter pyloridis.[20] Since then, studies on H. pylori have yielded tremendous knowledge on the property of the bacteria and their role in the pathogenesis of gastric diseases. The complete genome of this bacterium has now been sequenced.[21] Effective treatment with antibiotics has revolutionized the way chronic gastritis and peptic ulcer disease are managed.[22]
In addition to chronic gastritis, this organism plays a critical role in other major gastric and duodenal diseases ( Table 17-2 ). Peptic ulcer disease is now approached as an infectious disease that can be treated by antibiotics. H. pylori is present in 90% of patients with chronic gastritis affecting the antrum. Colonization rates increase with age, reaching 50% in asymptomatic American adults over age 50. Prevalence of infection among adults in Puerto Rico exceeds 80%. In this and other
814
TABLE 17-2 -- Diseases Associated with Helicobacter pylori Infection
Disease Association
Chronic gastritis Strong causal association
Peptic ulcer disease Strong causal association
Gastric carcinoma Strong causal association
Gastric MALT lymphoma * Definitive etiologic role
* MALT, mucosa-associated lymphoid tissue
areas where infection is endemic, the organism seems to be acquired in childhood and persists for decades. The mode of transmission of H. pylori has not been well defined, although oral-oral transmission, fecal-oral transmission, and environmental spread are among the possible routes. Most infected persons also have the associated gastritis but are asymptomatic. Nevertheless, infected persons are at increased risk for the development of peptic ulcer disease and possibly gastric cancer.
H. pylori is a nonsporing, curvilinear gram-negative rod measuring approximately 3.5 × 0.5 µm. H. pylori is part of a genus of bacteria that have adapted to the ecologic niche
provided by gastric mucus, which is lethal to most bacteria. The specialized traits that allow it to flourish include:
• Motility (via flagella), allowing it to swim through viscous mucus • Elaboration of a urease, which produces ammonia and carbon dioxide from endogenous urea, thereby buffering gastric acid in the immediate vicinity of the organism • Expression of bacterial adhesins, such as BabA, which binds to the fucosylated Lewis B blood-group antigens, enhances binding to blood group O antigen bearing cells.[23] • Expression of bacterial toxins, such as cytotoxin association gene A (CagA) and vacuolating cytotoxin gene A (VacA).[24] These are discussed later under "Peptic Ulcer."
The H. pylori genome is 1.65 million base pairs and encodes approximately 1500 proteins. Extensive molecular studies suggest that the bacteria cause gastritis by stimulating production of pro-inflammatory cytokines and by directly injuring epithelial cells (discussed later).
After initial exposure to H. pylori, gastritis occurs in two patterns: a predominantly antral-type gastritis with high acid production and elevated risk for duodenal ulcer, and a pangastritis that is followed by multifocal atrophy (multifocal atrophic gastritis) with lower gastric acid secretion and higher risk for adenocarcinoma. The underlying mechanisms contributing to this difference are not completely clear, but host-microorganism interplay appears to be critical. IL-1β is a potent pro-inflammatory cytokine and a powerful gastric acid inhibitor. Patients who have higher IL-1β production in response to H. pylori infection tend to develop pangastritis, while patients who have lower IL-1β production exhibit antral-type gastritis. [25]
A number of diagnostic tests have been developed for the detection of H. pylori. Noninvasive tests include a serologic test for antibodies, fecal bacterial detection, and a urea breath test. The breath test is based on the generation of ammonia by bacterial urease. Invasive tests are based on the identification of H. pylori in gastric biopsy tissue. Detection methods in gastric tissue include visualization of the bacteria in histologic sections, bacterial culture, a rapid urease test, and bacterial DNA detection by the polymerase chain reaction.
Patients with chronic gastritis and H. pylori usually improve when treated with antibiotics. Relapses are associated with reappearance of the organism. The current treatment regimens include antibiotics and hydrogen pump inhibitors.[22] Prophylactic and therapeutic vaccine development is still in the early research stage, but it holds the promise to eradicate or at least greatly reduce the worldwide prevalence of H. pylori infection.
In addition to H. pylori, humans can also be infected by Helicobacter heilmannii, a spiral bacterium found in dogs, cats, and nonhuman primates.[26] This bacterium causes a relatively mild gastritis.
Autoimmune Gastritis.
This form of gastritis accounts for less than 10% of cases of chronic gastritis. It results from the presence of autoantibodies to components of gastric gland parietal cells, including antibodies against the acidproducing enzyme H+ ,K+ -ATPase,[27] gastrin receptor, and intrinsic factor. Gland destruction and mucosal atrophy lead to loss of acid production. In the most severe cases, production of intrinsic factor is lost, leading to pernicious anemia. This uncommon form of gastritis is seen in association with other autoimmune disorders such as Hashimoto thyroiditis, Addison disease, and type 1 diabetes. Patients with autoimmune gastritis have a significant risk for developing gastric carcinoma and endocrine tumors (carcinoid tumor).
Morphology.
Chronic gastritis may affect different regions of the stomach and exhibit varying degrees of mucosal damage.[19] Autoimmune gastritis is characterized by diffuse mucosal damage of the body-fundic mucosa, with less intense to absent antral damage, probably due to the autoantibodies against parietal cells. Gastritis in the setting of environmental etiologies (including infection by H. pylori) tends to affect antral mucosa or both antral and body-fundic mucosa (pangastritis). The mucosa is usually reddened and has a coarser texture than normal. The inflammatory infiltrate may create a mucosa with thickened rugal folds, mimicking early infiltrative lesions. Alternatively, with long-standing atrophic disease, the mucosa may become thinned and flattened. Irrespective of cause or location, the histologic changes are similar. An inflammatory infiltrate of lymphocytes and plasma cells is present within the lamina propria ( Fig. 17-14 ). "Active" inflammation is signified by the presence of neutrophils within the glandular and surface epithelial layer. Active inflammation may be prominent or absent. Lymphoid aggregates, some with germinal centers, are frequently observed within the mucosa. Several additional histologic features are characteristic:
• Regenerative Change. A proliferative response to the epithelial injury is a constant feature of chronic gastritis. In the neck region of the gastric glands mitotic figures are increased. Epithelial cells of the
815
surface mucosa, and to a lesser extent the glands, exhibit enlarged, hyperchromatic nuclei and a higher nuclear-cytoplasmic ratio. Mucus vacuoles are diminished or absent in the superficial cells. When regenerative changes are severe, particularly with ongoing active inflammation, distinguishing regenerative change from dysplasia may be difficult.
• Metaplasia. The antral, body, and fundic mucosa may become partially replaced by metaplastic columnar absorptive cells and goblet cells of intestinal morphology (intestinal metaplasia), both along the surface epithelium and in rudimentary glands. Occasionally, villus-like projections may appear. Although small intestinal features predominate, in some instances, features of colonic epithelium may be present. • Atrophy. Atrophic change is evident by marked loss in glandular structures. Atrophy is quite frequently associated with autoimmune gastritis and pangastritis caused by H. pylori. Parietal cells, in particular, may be conspicuously absent in the autoimmune form. Persisting glands frequently undergo cystic dilatation. A particular feature of atrophic gastritis of autoimmune origin or chronic gastritis treated by inhibitors of acid secretion is hyperplasia of gastrin-producing G-cells in the antral mucosa. This is attributed to the hypochlorhydria or achlorhydria arising from severe parietal cell loss. The G-cell hyperplasia is responsible for the increased gastrinemia, which stimulates hyperplasia of enterochromaffin-like cells in the gastric body. As will be discussed later, the ECL cell hyperplasia is the frequent background for gastric carcinoid tumor formation. • Dysplasia. With long-standing chronic gastritis, the epithelium develops cytologic alterations, including variation in size, shape, and orientation of epithelial cells, and nuclear enlargement and atypia. Intestinal metaplasia may precede the development of dysplasia. Dysplastic alterations may become so severe as to constitute in situ carcinoma. The development of dysplasia is thought to be a precursor lesion of gastric cancer in atrophic forms of gastritis, particularly in association with pernicious anemia (autoimmune gastritis) and H. pylori- associated chronic gastritis.
Figure 17-14 Chronic gastritis, showing partial replacement of the gastric mucosal epithelium by intestinal metaplasia (upper left) and inflammation of the lamina propria (right) containing lymphocytes and plasma cells.
In those individuals infected by H. pylori, the organism lies in the superficial mucus layer and among the microvilli of epithelial cells. The distribution of organisms can be very patchy and irregular, with areas of heavy colonization adjacent to those with no organisms. In extreme cases, the organisms carpet the luminal surfaces of surface epithelial cells, the mucous neck cells, and the epithelial cells lining the gastric pits; they do not invade the mucosa. This is most easily demonstrated with silver stains ( Fig. 17-15 ), although organisms can be seen on Giemsa- and routine H & E-stained tissue. Even in heavily colonized stomachs, the organisms are absent from areas with intestinal metaplasia. Conversely, organisms may be present in foci of pyloric metaplasia in an inflamed duodenum and in the gastric-type mucosa of Barrett esophagus.
Clinical Features.
Chronic gastritis usually causes few symptoms. Nausea, vomiting, and upper abdominal discomfort may occur. Individuals with advanced gastritis from H. pylori or other environmental causes are often hypochlorhydric, owing to parietal cell damage and atrophy of body and fundic mucosa. However, since parietal cells are never completely destroyed, these patients do not develop achlorhydria or pernicious anemia. Serum gastrin levels are usually within the normal range or only modestly elevated.
Figure 17-15 Helicobacter pylori. A Steiner silver stain demonstrates the numerous darkly stained Helicobacter organisms along the luminal surface of the gastric epithelial cells. Note that there is no tissue invasion by bacteria.
When severe parietal cell loss occurs in the setting of autoimmune gastritis, hypochlorhydria or achlorhydria and hypergastrinemia are characteristically present. Circulating autoantibodies to a diverse array of parietal cell antigens may be detected. A small subset of these patients (10%) may develop overt pernicious anemia after a period of years. The familial occurrence of pernicious anemia is well established; a high prevalence of gastric autoantibodies is also found in asymptomatic relatives of patients with pernicious anemia. The distribution suggests that the inheritance of autoimmune gastritis is autosomal dominant.
Most important is the relationship of chronic gastritis to the development of peptic ulcer and gastric carcinoma. Most patients with a peptic ulcer, whether duodenal or gastric, have H. pylori infection. H. pylori is thought to contribute to the pathogenesis of both gastric carcinoma and lymphoma. The long-term risk of gastric cancer in patients with autoimmune gastritis is 2% to 4%, which is considerably greater than that of the normal population.
SPECIAL FORMS OF GASTRITIS
Eosinophilic gastritis is an idiopathic condition that features a prominent eosinophilic infiltrate of the mucosa, muscle wall, or all layers of the stomach, usually in the antral or pyloric region. This disorder typically affects middle-aged women, and the primary symptom is abdominal pain, although swelling of the pylorus may produce gastric outlet obstruction. It may occur in association with eosinophilic enteritis and is often accompanied by a peripheral eosinophilia. Steroid therapy is usually effective.
Allergic gastroenteropathy is a disorder of children that may produce symptoms of diarrhea, vomiting, and growth failure. An infiltrate of eosinophils limited to the mucosa can usually be demonstrated in antral biopsies.
Lymphocytic gastritis is a condition in which lymphocytes densely populate the epithelial layer of the mucosal surface and gastric pits and suffuse the lamina propria. The intraepithelial lymphocytes are exclusively T lymphocytes, mostly CD8+ cells. The gastritis is generally restricted to the body of the stomach. This condition produces indistinct symptoms such as abdominal pain, anorexia, nausea, and vomiting. Although idiopathic in nature, 45% to 60% of cases are associated with celiac disease. Therefore, an immune-mediated pathogenesis is most likely.
Granulomatous gastritis. The presence of intramucosal epithelioid granulomas can usually be attributed to Crohn disease, sarcoidosis, infection (tuberculosis, histoplasmosis, anisakiasis), a systemic vasculitis, or as a reaction to foreign materials. Granulomatous gastritis is the term reserved for patients without these concurrent conditions. This idiopathic disorder is clinically benign. The predominant pathologic finding is narrowing and rigidity of the gastric antrum due to transmural granulomatous inflammation.
Graft-versus-Host Disease. Gastritis associated with GVHD can be encountered in the setting of bone marrow transplantation. Histologically, there is a relatively mild
lymphocytic infiltrate in the lamina propria and apoptosis of glandular epithelial cells, in particular the mucous neck cells.
Reactive gastropathy is a group of disorders that exhibit characteristic mucosal histologic changes ( Fig. 17-16 ) that may include: foveolar hyperplasia with loss of mucin and glandular regenerative changes, mucosal edema and dilation of mucosal capillaries, and smooth muscle fibers extending into the lamina propria between the glands. The key to the definition is the absence of active (neutrophilic) inflammation of the epithelium. Reactive gastropathy is fairly common. The etiology is related to chemical injury from cyclooxygenase inhibition (aspirin and NSAIDs) or bile reflux, and from mucosal trauma resulting from prolapse. In particular, gastric antral trauma or prolapse induce a characteristic lesion referred to as gastric antral vascular ectasia. Endoscopy shows longitudinal stripes of edematous erythematous mucosa alternating with less severely injured mucosa (watermelon stomach). Histologically, the antral mucosa exhibits reactive gastropathy and dilated capillaries containing fibrin thrombi.
Peptic Ulcer Disease
Ulcers are defined histologically as a breach in the mucosa of the alimentary tract that extends through the muscularis mucosa into the submucosa or deeper. Although they may occur anywhere in the alimentary tract, none are as prevalent as the peptic ulcers that occur in the duodenum and stomach. Acute gastric ulcers may also appear under conditions of severe systemic stress or ingestion of NSAIDs. Ulcers are to be distinguished from erosions, in which there is epithelial disruption within the mucosa but no breach of the muscularis mucosa.
PEPTIC ULCERS
Peptic ulcers are chronic, most often solitary, lesions that occur in any portion of the gastrointestinal tract exposed to the aggressive action of acid/peptic juices. Peptic ulcers are usually solitary lesions less than 4 cm in diameter, located in the following sites, in order of decreasing frequency:[28]
Figure 17-16 Reactive gastropathy. Gastric mucosa, showing hyperplasia of foveolar surface epithelial cells, glandular regenerative changes, and smooth muscle fibers extending into lamina propria.
817
• Duodenum, first portion • Stomach, usually antrum • At the gastroesophageal junction, in the setting of gastroesophageal reflux or Barrett esophagus • Within the margins of a gastrojejunostomy • In the duodenum, stomach, and/or jejunum of patients with Zollinger-Ellison syndrome • Within or adjacent to an ileal Meckel diverticulum that contains ectopic gastric mucosa.
Epidemiology.
In the United States, approximately 4 million people have peptic ulcers (duodenal and gastric), and 350,000 new cases are diagnosed each year. Around 180,000 patients are hospitalized yearly, and about 5000 people die each year as a result of peptic ulcer disease.[28] The lifetime likelihood of developing a peptic ulcer is about 10% for American males and 4% for females.
Peptic ulcers are relapsing lesions that are most often diagnosed in middle-aged to older adults, but they may first become evident in young adult life. They often appear without obvious precipitating conditions and may then, after a period of weeks to months of active disease, heal with or without therapy. Even with healing, however, the tendency to develop peptic ulcers remains, in part because of recurrent infections with H. pylori. Although it is difficult to obtain estimates of the prevalence of active disease, autopsy studies and population surveys indicate a prevalence of 6% to 14% for men and 2% to
6% for women. The male-to-female ratio for duodenal ulcers is about 3:1, and for gastric ulcers about 1.5 to 2:1. Women are most often affected at or after menopause. For
Figure 17-17 Diagram of causes of, and defense mechanisms against, peptic ulceration. Diagram of the base of a nonperforated peptic ulcer, demonstrating the layers of necrosis (N), inflammation (I), granulation tissue (G), and scar (S), moving from the luminal surface at the top to the muscle wall at the bottom.
unknown reasons, there has been a significant decrease in the prevalence of duodenal ulcers over the past decades but little change in the prevalence of gastric ulcers.
Pathogenesis.
Peptic ulcers are produced by an imbalance between gastroduodenal mucosal defense mechanisms and the damaging forces, [29] particularly gastric acid and pepsin ( Fig. 17-17 ). However, hyperacidity is not a prerequisite, as only a minority of patients with duodenal ulcers has hyperacidity, and it is even less common in those with gastric ulcers. Rather, gastric ulceration occurs when mucosal defenses fail, as when mucosal blood flow drops, gastric emptying is delayed, or epithelial restitution is impaired.
H. pylori infection is a major factor in the pathogenesis of peptic ulcer. It is present in virtually all patients with duodenal ulcers and in about 70% of those with gastric ulcers. Furthermore, antibiotic treatment of H. pylori infection promotes healing of ulcers and tends to prevent their recurrence. Hence, much interest is focused on the possible mechanisms by which this tiny spiral organism tips the balance of mucosal defenses. Some likely possibilities include:
• Although H. pylori does not invade the tissues, it induces an intense inflammatory and immune response. There is increased production of pro-inflammatory cytokines such as interleukin (IL)-1, IL-6, tumor necrosis factor (TNF), and, most notably, IL-8. This cytokine is produced by the mucosal epithelial cells, and it recruits and activates neutrophils. • Several bacterial gene products are involved in causing epithelial cell injury and induction of inflammation. H.
818
pylori secretes a urease that breaks down urea to form toxic compounds such as ammonium chloride and monochloramine. The organisms also elaborate phospholipases that damage surface epithelial cells. Bacterial proteases and phospholipases break down the glycoprotein-lipid complexes in the gastric mucus, thus weakening the first line of mucosal defense. • H. pylori enhances gastric acid secretion and impairs duodenal bicarbonate production, thus reducing luminal pH in the duodenum. This altered milieu seems to favor gastric metaplasia (the presence of gastric epithelium) in the first part of the duodenum. Such metaplastic foci provide areas for H. pylori colonization. • Several H. pylori proteins are immunogenic, and they evoke a robust immune response in the mucosa. Both activated T cells and B cells can be seen in chronic gastritis caused by H. pylori. The B lymphocytes aggregate to form follicles. The role of T and B cells in causing epithelial injury is not established, but T-cell-driven activation of B cells may be involved in the pathogenesis of gastric lymphomas. • Thrombotic occlusion of surface capillaries is promoted by a bacterial platelet-activating factor. • Other antigens (including lipopolysaccharide) recruit inflammatory cells to the mucosa. The chronically inflamed mucosa is more susceptible to acid injury. • Damage to the mucosa is thought to permit leakage of tissue nutrients into the surface microenvironment, thereby sustaining the bacillus.
With the unraveling of the H. pylori genome, the basis of the pathogenicity of this organism is beginning to be understood. [22] [30] Over 80% of patients with duodenal ulcers are infected by strains that are cytotoxin-associated antigen (CagA) positive. This antigen elicits a strong serologic response, but more importantly it is a marker for the Cag pathogenicity island, a 37 kb DNA fragment that encodes 29 genes, some of which are involved in the pro-inflammatory and tissue damaging effects of H. pylori. In keeping with this, infection with Cag positive strains is associated with greater number of organisms in the tissue, more severe epithelial damage, greater acute and chronic inflammation, higher likelihood of peptic ulceration and an increased risk for gastric cancer (discussed later). One of the important genes regulated by CagA is the vacuolating toxin (VacA); the CagA gene is essential for the expression of VacA. This toxin causes cell injury (characterized by vacuole formation) in vitro and gastric tissue damage in vivo. VacA also behaves as a passive urea transporter thereby increasing the permeability
of the epithelium to urea. As discussed above, urea is broken down into toxic intermediates by bacterial urease.
Only 10% to 20% of individuals worldwide infected with H. pylori actually develop peptic ulcer. Why most infected persons are spared and some are susceptible remains an enigma. Perhaps there are unknown interactions between H. pylori and the mucosa that occur only in some individuals. Another perplexing observation is that in patients with duodenal ulcer, the actual infection by H. pylori is limited to the stomach. Increased acid production by H. pylori infection seems to play a role. Suffice it to say that while the link between H. pylori infection and gastric and duodenal ulcers is well established, the interactions leading to ulceration remain to be defined.
Other events may act alone or in concert with H. pylori to promote peptic ulceration. Gastric hyperacidity, when present, may be strongly ulcerogenic. Hyperacidity may arise from increased parietal cell mass, increased sensitivity to secretory stimuli, increased basal acid secretory drive, or impaired inhibition of stimulatory mechanisms such as gastrin release. The classic example is Zollinger-Ellison syndrome, in which there are multiple peptic ulcerations in the stomach, duodenum, and even jejunum, owing to excess gastrin secretion by a tumor and, hence, excess gastric acid production.
Chronic use of NSAIDs suppresses mucosal prostaglandin synthesis; aspirin also is a direct irritant. Cigarette smoking impairs mucosal blood flow and healing. Alcohol has not been proved to directly cause peptic ulceration, but alcoholic cirrhosis is associated with an increased incidence of peptic ulcers. Corticosteroids in high dose and with repeated use promote ulcer formation. In some patients with duodenal ulcers, there is too-rapid gastric emptying, exposing the duodenal mucosa to an excessive acid load. Duodenal ulcer also is more frequent in patients with alcoholic cirrhosis, chronic obstructive pulmonary disease, chronic renal failure, and hyperparathyroidism. In the latter two conditions, hypercalcemia stimulates gastrin production and therefore acid secretion. Genetic influences appear to play no major role in peptic ulceration. Finally, there are compelling arguments that personality and psychological stress are important contributing factors, even though hard data on cause and effect are lacking. Indeed, we might develop ulcers by trying to fathom their cause(s).
Morphology.
At least 98% of peptic ulcers are located in the first portion of the duodenum or in the stomach, in a ratio of about 4:1. Most duodenal ulcers occur within a few centimeters of the pyloric ring. The anterior wall of the duodenum is affected more often than the posterior wall. Gastric ulcers are predominantly located along the lesser curvature, in or around the border zone between the oxyntic mucosa and the antral mucosa. Less commonly, gastric ulcers may occur on the anterior or posterior walls, or along the greater curvature. Although the great majority of individuals have a single ulcer, in 10% to 20% of patients with gastric ulceration there may be a coexistent duodenal ulcer.
Wherever they occur, chronic peptic ulcers have a fairly standard, virtually diagnostic gross appearance ( Fig. 17-18 ). Small lesions (<0.3 cm) are most likely to be shallow
erosions; those over 0.6 cm are likely to be ulcers. Although over 50% of peptic ulcers have a diameter less than 2 cm, about 10% of benign ulcers are greater than 4 cm. Since carcinomatous ulcers may be less than 4 cm in diameter and may be located anywhere in the stomach, size and location do not differentiate a benign from a malignant ulcer.
The classic peptic ulcer is a round to oval, sharply punched-out defect with relatively straight walls. The mucosal margin may overhang the base slightly, particularly on the upstream portion of the circumference. The margins are usually level with the surrounding mucosa or only slightly elevated. Heaping-up of these margins is rare in the benign ulcer but is characteristic of the malignant lesion. The depth of these ulcers varies, from superficial lesions involving only the mucosa and muscularis mucosa to
819
Figure 17-18 Peptic ulcer of the duodenum. Note that the ulcer is small (2 cm) with a sharply punched-out appearance. Unlike cancerous ulcers, the margins are not elevated. The ulcer base is clean. (Courtesy of Robin Foss, University of Florida, Gainesville, FL.)
deeply excavated ulcers having their bases on the muscularis propria. When the entire wall is penetrated, the base of the ulcer may be formed by adherent pancreas, omental fat, or liver. Free perforation into the peritoneal cavity may occur.
The base of a peptic ulcer is smooth and clean, owing to peptic digestion of any exudate that may form. At times, thrombosed or even patent blood vessels (the source of life-threatening hemorrhage) are evident in the base of the ulcer. Scarring may involve the entire thickness of the stomach; puckering of the surrounding mucosa creates mucosal folds that radiate from the crater in spokelike fashion. The gastric mucosa surrounding a gastric ulcer is somewhat edematous and reddened, owing to the almost invariable gastritis.
The histologic appearance varies from active necrosis, to chronic inflammation and scarring, to healing (see Fig. 2-26 , Chapter 2). In active ulcers with ongoing necrosis,
four zones are demonstrable: (1) the base and margins have a superficial thin layer of necrotic fibrinoid debris not visible to the naked eye; (2) beneath this layer is a zone of non-specific inflammatory infiltrate, with neutrophils predominating; (3) in the deeper layers, especially in the base of the ulcer, there is active granulation tissue infiltrated with mononuclear leukocytes; and (4) the granulation tissue rests on a more solid fibrous or collagenous scar. Vessel walls within the scarred area are typically thickened by the surrounding inflammation and are occasionally thrombosed.
Chronic gastritis is virtually universal among patients with peptic ulcer disease, occurring in 85% to 100% of patients with duodenal ulcers and in 65% with gastric ulcers. H. pylori infection is almost always demonstrable in patients with gastritis. Gastritis remains after the ulcer has healed; recurrence of the ulcer does not appear to be related to progression of the gastritis. This feature is helpful in distinguishing peptic ulcers from acute erosive gastritis or stress ulcers, since the adjacent mucosa is generally normal in the latter two conditions.
Clinical Features.
The great majority of peptic ulcers cause epigastric gnawing, burning, or aching pain. A significant minority first comes to light with complications such as iron-deficiency anemia, frank hemorrhage, or perforation. The pain tends to be worse at night and occurs usually 1 to 3 hours after meals during the day. Classically, the pain is relieved by alkalis or food, but there are many exceptions. Nausea, vomiting, bloating, belching, and significant weight loss (raising the possibility of some hidden malignancy) are additional manifestations. With penetrating ulcers, the pain is occasionally referred to the back, the left upper quadrant, or chest. This type of pain may be misinterpreted as being of cardiac origin.
Peptic ulcers are notoriously chronic, recurring lesions. They more often impair the quality of life than shorten it. When untreated, it takes an average of 15 years for healing a duodenal or gastric ulcer. With present-day therapies aimed at neutralization of gastric acid, promotion of mucus secretion, inhibition of acid secretion (H2 receptor antagonists and parietal cell H+ ,K+ -ATPase pump inhibitors), and eradication of H. pylori infection, most ulcers heal within a few weeks, and victims usually escape the surgeon's knife.
The complications of peptic ulcer disease are listed in Table 17-3 . Malignant transformation does not occur with duodenal ulcers and is extremely rare with gastric ulcers. When it occurs, it is always possible that a seemingly benign lesion was, from the outset, a deceptive ulcerative gastric carcinoma.
ACUTE GASTRIC ULCERATION
Focal, acutely developing gastric mucosal defects are a well-known complication of therapy with NSAIDs. Alternatively, they may appear following severe physiologic stress, whatever its nature—hence the term stress ulcers. Generally, there are multiple lesions located mainly in the stomach and occasionally in the duodenum. They range in depth from mere shedding of the superficial epithelium (erosion) to deeper lesions that
involve the entire mucosal thickness (ulceration). The shallow erosions are, in essence, an extension of acute erosive gastritis. The deeper lesions comprise well-defined ulcerations, but they are not precursors of chronic peptic ulcers.
Stress erosions and ulcers are most commonly encountered in patients with shock, extensive burns, sepsis, or severe trauma; in any intracranial injury that raises intracranial pressure; and following intracranial surgery. Those occurring in the proximal duodenum and associated with severe burns or trauma are called Curling ulcers. Gastric, duodenal, and esophageal ulcers arising in patients with intracranial injury, operations, or
TABLE 17-3 -- Complications of Peptic Ulcer Disease
Bleeding
• Occurs in 15% to 20% of patients
• Most frequent complication
• May be life-threatening
• Accounts for 25% of ulcer deaths
• May be the first indication of an ulcer
Perforation
• Occurs in about 5% of patients
• Accounts for two thirds of ulcer deaths
• Rarely, is the first indication of an ulcer
Obstruction from edema or scarring
• Occurs in about 2% of patients
• Most often due to pyloric channel ulcers
• May also occur with duodenal ulcers
• Causes incapacitating, crampy abdominal pain
• Rarely, may lead to total obstruction with intractable vomiting
820
tumors are designated Cushing ulcers and carry a high incidence of perforation.
The genesis of the acute mucosal defects in these varied clinical settings is poorly understood. No doubt, many factors are shared with acute gastritis, such as impaired oxygenation. NSAID-induced ulcers are related to decreased prostaglandin production from the inhibition of cyclooxygenase. In the case of lesions associated with intracranial injury, the proposed mechanism involves the direct stimulation of vagal nuclei by increased intracranial pressure, leading to hypersecretion of gastric acid, which is common in these patients. Systemic acidosis, a frequent finding in these clinical settings, may contribute to mucosal injury by lowering the intracellular pH of mucosal cells. These cells are also hypoxic as a consequence of stress-induced splanchnic vasoconstriction.
Morphology.
Acute stress ulcers are usually less than 1 cm in diameter and are circular and small. The ulcer base is frequently stained a dark brown by the acid digestion of extruded blood ( Fig. 17-19 ). Unlike chronic peptic ulcers, acute stress ulcers are found anywhere in the stomach, the gastric rugal pattern is essentially normal and the margins and base of the ulcers are not indurated. While they may occur singly, more often there are multiple stress ulcers throughout the stomach and duodenum. Microscopically, acute stress ulcers are abrupt lesions, with essentially unremarkable adjacent mucosa. Depending on the duration of the ulceration, there may be a suffusion of blood into the mucosa and submucosa and some inflammatory reaction. Conspicuously absent are scarring and thickening of blood vessels, as seen in chronic peptic ulcers. Healing with complete reepithelialization occurs after the causative factors are removed. The time required for complete healing varies from days to several weeks.
Clinical Features.
Most critically ill patients admitted to hospital intensive care units develop histologic evidence of gastric mucosal damage. Bleeding from superficial gastric erosions or ulcers sufficient to require transfusion develops in 1% to 4% of these patients. Although prophylactic H2 -receptor antagonists and proton pump inhibitors may blunt the impact of stress ulceration, the single most important determinant of
Figure 17-19 Multiple stress ulcers of the stomach, highlighted by dark digested blood on their surfaces.
clinical outcome is the ability to correct the underlying condition(s). The gastric mucosa can recover completely if the patients do not succumb to their primary disease.
Miscellaneous Conditions
Gastric dilation may arise from gastric outlet obstruction (e.g., pyloric stenosis) or from the functional atony of the stomach and intestines (paralytic ileus) that may develop in patients with generalized peritonitis. The stomach may contain as much as 10 to 15 L of fluid; on rare occasion, gastric rupture may occur. This is a calamitous event followed rapidly by shock or death if not treated immediately. Rarely, spontaneous gastric perforation may occur in the newborn, during labor and delivery, severe vomiting, or cardiopulmonary resuscitation, and following ingestion of extreme amounts of carbonated beverages.
The stomach has the dubious privilege of being the major site for formation of luminal concretions of indigestible ingested material. Phytobezoars are derived from plant material, including fibers, leaves, roots, and skins of almost any plant matter. Trichobezoars, better known as "hairballs," consist of ingested hair within a mucoid coat containing decaying foodstuff ( Fig. 17-20 ). The dysmotility following partial gastrectomy or partial gastric outlet obstruction is conducive to bezoar formation from more conventional ingested food. Bizarre bezoars have developed among partakers of illicit pharmaceuticals, glue swallowers, and children or patients with neuropsychiatric disorders, who have been known to ingest pins, nails, razor blades, coins, gloves, and even leather wallets.
HYPERTROPHIC GASTROPATHY
This designation includes a group of uncommon conditions, all characterized by giant cerebriform enlargement of the rugal folds of the gastric mucosa ( Fig. 17-21 ). The rugal enlargement is caused by hyperplasia of the mucosal epithelial cells, without inflammation. Three variants are recognized:
• Ménétrier disease, resulting from profound hyperplasia of the surface mucous cells with accompanying glandular atrophy
821
• Hypertrophic-hypersecretory gastropathy, associated with hyperplasia of the parietal and chief cells within gastric glands • Gastric gland hyperplasia secondary to excessive gastrin secretion, in the setting of a gastrinoma (Zollinger-Ellison syndrome).
All three conditions are of clinical importance for two reasons: (1) they may mimic infiltrative carcinoma or lymphoma of the stomach on endoscopic and radiographic examinations; and (2) the enormous increase in acid secretions in hypertrophic-hypersecretory gastropathy and Zollinger-Ellison syndrome places patients at risk for peptic ulceration. A pure form of parietal cell hypertrophy, without hyperacidity, may
occur in long-term takers of acid secretion inhibitors. Cessation of therapy may cause a transient rebound of excess acid secretion.
Ménétrier disease is most often encountered in males (male : female ratio 3 : 1) in the fourth to sixth decade of life but occasionally is seen in children. The etiology of this disease is unknown, but a role has been suggested for growth factor overexpression in superficial gastric epithelium. Transgenic mice with transforming growth factor-α (TGF-α) expressed in the stomach exhibit a disorder clinically and histologically similar to human Ménétrier disease. Although the disorder may be asymptomatic, it often produces epigastric discomfort, diarrhea, weight loss, and sometimes bleeding related to superficial rugal erosions. The hypertrophic change may predominantly involve the body-fundus or antrum or may affect the entire stomach. The gastric secretions contain excessive mucus and in many instances little to no hydrochloric acid due to glandular atrophy. In some patients, there may be sufficient protein loss in the gastric secretions to produce hypoalbuminemia and peripheral edema, thus constituting a form of protein-losing gastroenteropathy. Infrequently, the mucosal hyperplasia becomes metaplastic, providing a soil for the development of gastric carcinoma.
GASTRIC VARICES
Gastric varices develop in the setting of portal hypertension but less often than esophageal varices. Most gastric varices lie within 2 to 3 cm of the gastroesophageal junction, arising from longitudinally placed submucosal veins. They often appear as masslike nodular and tortuous winding elevations of the mucosa in the cardia or fundus. Due to their deep submucosal or subserosal location and the normal color of the overlying mucosa, it may be difficult to distinguish varices from enlarged rugae or even malignancy. Since they rarely occur in the absence of esophageal varices, diagnosis can usually be made without resorting to a potentially disastrous biopsy.
Tumors
As in the esophagus and intestines, tumors arising from the mucosa predominate over mesenchymal and stromal tumors. These can be classified as benign and malignant lesions.
BENIGN TUMORS
In the alimentary tract, the term polyp is applied to any nodule or mass that projects above the level of the surrounding mucosa. Use of the term is generally restricted to mass lesions arising in the mucosa, although occasionally a submucosal lipoma or leiomyoma may protrude, generating a polypoid lesion. The mucosal polyps are classified as non-neoplastic or neoplastic. Gastric polyps are uncommon.[31] Although they are usually found incidentally, dyspepsia or anemia resulting from blood loss may prompt the search for a gastrointestinal lesion.
The great majority of gastric polyps (up to 90%) are non-neoplastic and appear to be of a hyperplastic nature. These polyps are composed of a variable mixture of hyperplastic surface epithelium (foveolar epithelium) and cystically dilated glandular tissue, with a lamina propria containing increased inflammatory cells and smooth muscle ( Fig. 17-22 ). The surface epithelium may be regenerative in response to surface erosion and inflammation, but true dysplasia is not present. Most hyperplastic polyps are small and sessile and are commonly located in the antrum; some may approach several centimeters in diameter and have an apparent stalk. In about 20% to 25% of cases multiple polyps, sometimes more than twenty, are observed.
The adenoma of the stomach constitutes 5% to 10% of the polypoid lesions in the stomach. By definition, an adenoma contains proliferative dysplastic epithelium and thereby has malignant potential. Adenomatous polyps are much more common in the colon, and they are described in considerable detail in the discussion of the colon. Gastric adenomas may be sessile (without a stalk) or pedunculated (stalked). The most common location is the antrum. These lesions are usually single, and may grow up to 3 to 4 cm in size before detection ( Fig. 17-23 ). In contrast to the colon, adenomatous change in the stomach may cover a large region of flat gastric mucosa without forming a mass lesion.
Other specific types of gastric polyps are relatively uncommon. Among these are fundic gland polyps,
822
Figure 17-22 Gastric hyperplastic polyp. Low-power microscopic view of the polyp showing hyperplastic foveolar epithelium and inflammation.
hamartomatous Peutz-Jeghers polyps, and juvenile polyps. The fundic gland polyp is an innocuous cystic dilation of glands in the oxyntic mucosa. These polyps usually occur sporadically, but they can occur in the syndrome of familial adenomatous polyposis (FAP, described later). Curiously, sporadic fundic gland polyps also exhibit mutations in β-catenin with high frequency. [32] The hamartomatous polyps may occur in isolation, but gastric Peutz-Jeghers polyps are most commonly seen as part of the Peutz-Jeghers syndrome, and gastric juvenile polyps as part of the juvenile polyposis syndrome. All of these conditions will be discussed later.
The inflammatory fibroid polyp (eosinophilic granuloma) is a striking lesion in that it is a bulky submucosal growth composed of inflamed vascularized fibromuscular tissue with a prominent eosinophilic infiltrate and a tenuous mucosa stretched over the surface ( Fig. 17-24 ). These polyps may occur anywhere
Figure 17-23 Gastric adenoma. Gross photograph showing a large polyp in the stomach.
in the alimentary tract but are found most frequently in the distal stomach. As they protrude into the lumen, they may occlude the pyloric channel and present abruptly as acute gastric outlet obstruction. Their origin is unknown. Whether these are inflammatory or neoplastic lesions is still debatable.
Clinical Features.
Hyperplastic polyps are seen most frequently in the setting of chronic gastritis. They are regarded as having no malignant potential as such but are nevertheless found in about 20% of stomachs resected for carcinoma. This is attributed to the tendency of chronically inflamed gastric mucosa both to form hyperplastic polyps and to develop into malignancy.
As with the colonic counterpart, the incidence of gastric adenomas increases with age, particularly into and beyond the seventh decade of life. The male-to-female ratio is 2:1. Up to 40% of gastric adenomas contain a focus of carcinoma at the time of diagnosis, particularly the larger lesions. The risk of cancer in the adjacent gastric mucosa may be as high as 30%. Unlike colonic adenomas, which usually arise from apparently normal mucosa, the usual substratum for gastric adenomas is chronic gastritis with intestinal metaplasia. Autoimmune gastritis can also lead to gastric adenoma formation.
Otherwise innocuous hyperplastic polyps may occasionally harbor foci of adenomatous epithelium. As non-neoplastic and adenomatous polyps cannot reliably be distinguished endoscopically, histologic examination of gastric polyps is mandatory.
GASTRIC CARCINOMA
Carcinoma is the most important and the most common (90% to 95%) of malignant tumors of the stomach. Next in order of frequency are lymphomas (4%), carcinoids (3%), and mesenchymal tumors (2%), which include gastrointestinal stromal tumors, leiomyosarcoma, and schwannoma.
Epidemiology.
Gastric carcinoma is the second most common tumor in the world. Its incidence, however, varies widely, being particularly high in countries such as Japan, Chile, Costa Rica, Colombia, China, Portugal, Russia, and Bulgaria,
Figure 17-24 Inflammatory fibroid polyp; microscopic photograph showing submucosal growth of inflamed vascularized fibromuscular tissue with prominent eosinophilic infiltrate.
823
and fourfold to sixfold less common in the United States, the United Kingdom, Canada, Australia, New Zealand, France, and Sweden. It is more common in lower socioeconomic groups and exhibits a male-to-female ratio of about 2:1. In most countries, there has been a steady decline in both the incidence and the mortality of gastric cancer over the past six decades. In 1930, gastric cancer was the most common cause of cancer death in the United States. Between 1930 and 1998, the annual mortality rate in the United States dropped from about 38 to 5 per 100,000 for men, and from 28 to 3 per 100,000 for women.[33] Yet it causes 2.5% of all cancer deaths in the United States and is the leading cause of cancer death worldwide. Although five-year survival rates have improved since the advent of endoscopy in the 1960s, they remain poor (about 20% in the United States).
There are several classification systems for gastric carcinoma. The most commonly used are the Laurén and the World Health Organization (WHO) classifications. In 1965, Laurén classified gastric carcinoma into two subtypes: those exhibiting an intestinal morphology with the formation of bulky tumors composed of glandular structures and those with diffuse, infiltrative growth of poorly differentiated discohesive malignant cells. The intestinal and diffuse sub-types appear to have a different pathogenetic basis. The intestinal type predominates in high-risk areas, and develops from precursor lesions. By contrast, the incidence of the diffuse type is relatively constant, and the tumors have no identifiable precursor lesions. The intestinal type exhibits a mean age of incidence of 55 years and a male-to-female ratio of 2:1. Diffuse gastric cancer occurs in slightly younger patients (mean age, 48), with an approximately equal male-to-female ratio. Although the intestinal type was far more common, the drop in incidence of gastric cancer has occurred only for this type. As a consequence, the incidence of intestinal and diffuse cancer is now approximately the same. The WHO classification system has been in use since 1977, is relatively simple, and has gained wide acceptance. It classifies the tumor based on histologic appearance alone ( Table 17-4 ).
Pathogenesis.
The major factors thought to affect the genesis of gastric cancer are summarized in Table 17-5 . They apply more to the intestinal type, as the risk factors for diffuse gastric cancer are not as well defined.
Helicobacter pylori Infection.
Chronic infection with H. pylori generally increases the risk for developing gastric carcinoma by five- to six-fold. The bacterial infection causes chronic gastritis, followed by atrophy, intestinal metaplasia, dysplasia, and carcinoma.[34] The sequential alterations depend on both the presence of bacterial proteins and the host immune response; the latter is influenced by the host genetic background. In particular, long-standing mucosal inflammation reduces acid secretion (hypochlorhydria) and pepsin secretion. This favors bacterial growth and perpetuation of chronic inflammation, sustained mucosal epithelial cell proliferation, and hence increased risk of genomic mutation. The increased oxidative stress further promotes DNA damage. However, the vast majority of individuals infected with H. pylori will not develop cancer and not all H. pylori infections increase the risk of cancer. Therefore, other factors must be involved in tumorigenesis. The risk for tumor
development is greatly increased in patients in whom mucosal inflammation progresses to multifocal mucosal atrophy and intestinal metaplasia. Dysplasia of the gastric mucosa is the final common
TABLE 17-4 -- WHO Histologic Classification of Gastric Tumors
Epithelial Tumors
Intraepithelial neoplasia: adenoma
Adenocarcinoma *
• Papillary adenocarcinoma
• Tubular adenocarcinoma
• Mucinous adenocarcinoma
• Signet-ring cell carcinoma
• Undifferentiated carcinoma
• Adenosquamous carcinoma
Small-cell carcinoma
Carcinoid tumor
Nonepithelial Tumors
Leiomyoma
Schwannoma
Granular cell tumor
Leiomyosarcoma
Gastrointestinal stromal tumor (GIST) (gradation from benign to malignant)
Kaposi sarcoma
Others
Malignant Lymphoma
* The Laurén classification subdivides adenocarcinomas into intestinal and diffuse types.
pathway by which intestinal-type gastric cancers develop. Adenomas containing mucosal dysplasia can also become malignant.
Environmental influences may be critical in gastric carcinogenesis. [35] When families migrate from high-risk to low-risk areas (or the reverse), successive generations acquire the level of risk that prevails in the new locales. The diet
TABLE 17-5 -- Factors Associated with Increased Incidence of Gastric Carcinoma
Environmental Factors
Infection by H. pylori
• Present in most cases of intestinal-type carcinoma
Diet
• Nitrites derived from nitrates (water, preserved food)
• Smoked and salted foods, pickled vegetables, chili peppers
• Lack of fresh fruit and vegetables
Low socioeconomic status
Cigarette smoking
Host Factors
Chronic gastritis
• Hypochlorhydria: favors colonization with H. pylori
• Intestinal metaplasia is a precursor lesion
Partial gastrectomy
• Favors reflux of bilious, alkaline intestinal fluid
Gastric adenomas
• 40% harbor cancer at time of diagnosis
• 30% have adjacent cancer at time of diagnosis
Barrett esophagus
• Increased risk of gastroesophageal junction tumors
is suspected to be a primary factor, and adherence to certain culinary practices is associated with a high risk of gastric carcinoma. Lack of refrigeration; consumption of preserved, smoked, cured, and salted foods; water contamination with nitrates; and lack of fresh fruit and vegetables are common themes in high-risk areas. The consumption of dietary carcinogens, such as N-nitroso compounds and benzopyrene, appears to be particularly important. Conversely, intake of green, leafy vegetables and citrus fruits, which contain antioxidants such as ascorbate (vitamin C), alpha-tocopherol (vitamin E), and beta-carotene, is negatively correlated with gastric cancer. A specific protective role for any one of these nutrients cannot be assumed, however, since intake of fresh food may simply displace consumption of preserved foods.
So far there is no conclusive evidence linking alcohol intake and cigarette smoking to the development of gastric cancer. Despite initial concern, to date there appears to be no increased risk of stomach cancer from the use of antacid drug therapies.
Host.
Autoimmune gastritis, like H. pylori infection, increases the risk of gastric cancer, presumably due to chronic inflammation and intestinal metaplasia. It has been noted that blood group A patients have higher risk but it is not yet clear whether this is related to the binding of H. pylori to Lewis B antigen, or to other mechanisms.
Within the United States, blacks, Native Americans, and Hawaiians have a higher risk of developing gastric cancer. But since only about 8% to 10% of patients with gastric cancer have a family history of this disease, genetic factors are unlikely to be a major influence. Environmental factors mentioned above, are likely to play a major role in the higher incidence of gastric cancer among these various groups. Genetic traits play a critical role in some familial cases of gastric cancer, including gastric carcinoma occurring in the hereditary nonpolyposis colorectal cancer (HNPCC) syndrome. Recently, E-cadherin gene (CDH1) germ-line mutations have also been identified as the underlying genetic basis for another familial gastric cancer syndrome that is characterized by early occurrence of diffuse type adenocarcinoma.[36] These patients are also at risk for developing lobular breast cancer.[37]
Other Risk Factors.
Peptic ulcer disease per se does not impart increased risk for development of gastric cancer. However,
Figure 17-25 Diagram of growth patterns and spread of gastric carcinoma. In early gastric carcinoma (A), the tumor is confined to the mucosa and submucosa and may exhibit an exophytic, flat or depressed, or excavated conformation. Advanced gastric carcinoma (B) extends into the muscularis propria and beyond. Linitis plastica is an extreme form of flat or depressed advanced gastric carcinoma.
patients who have had partial gastrectomies for peptic ulcer disease have a slightly higher risk of gastric cancer in the residual gastric stump, attributed to the hypochlorhydria, bile reflux, and chronic gastritis that occur in the post-gastrectomy state. Ménétrier disease is also a risk factor for gastric carcinoma.
Multiple genetic alterations have been described in gastric cancers, mostly in studies involving intestinal-type cancers. Among these are allelic losses in various chromosomal loci, and microsatellite instability in several genes including TGFβRII, BAX and IGFRII. [36] [37] [38] Moreover, p53 mutations are present in a majority of tumors, and abnormalities in E-cadherin expression are quite frequent. Nevertheless, it has not been possible so far to define a clear sequence of events in gastric tumorigenesis. It appears that intestinal and diffuse gastric cancers may develop through different genetic pathways.
Morphology.
The location of gastric carcinomas within the stomach is as follows: pylorus and antrum, 50% to 60%; cardia, 25%; with the remainder in the body and fundus. The lesser curvature is involved in about 40% and the greater curvature in 12%. Thus, a favored location is the lesser curvature of the antropyloric region. Although less common, an ulcerative lesion on the greater curvature is more likely to be malignant.
Gastric carcinoma is classified on the basis of: (1) depth of invasion; (2) macroscopic growth pattern; and (3) histologic subtype. The morphologic feature having the greatest impact on clinical outcome is the depth of invasion. Early gastric carcinoma is defined as a lesion confined to the mucosa and submucosa, regardless of the presence or absence of perigastric lymph node metastases. Some early tumors cover large areas of the gastric mucosa (up to 10 cm in diameter) and yet show no invasion into the muscular wall. Early gastric carcinoma is not synonymous with carcinoma in situ, as
the latter is confined to the surface epithelial layer. Advanced gastric carcinoma is a neoplasm that has extended below the submucosa into the muscular wall and has perhaps spread
825
Figure 17-26 Gastric carcinoma. Gross photograph showing an ill-defined, excavated central ulcer surrounded by irregular, heaped-up borders.
more widely. All cancers presumably begin as "early" lesions, which develop over time into "advanced" lesions.
The three macroscopic growth patterns of gastric carcinoma, which may be evident at both the early and advanced stages, are: (1) exophytic, with protrusion of a tumor mass into the lumen; (2) flat or depressed, in which there is no obvious tumor mass within the mucosa; and (3) excavated, whereby a shallow or deeply erosive crater is present in the wall of the stomach ( Fig. 17-25 ). Exophytic tumors are readily identified by radiographic techniques and at endoscopy and may contain portions of an adenoma. In contrast, flat or depressed malignancy may not be apparent to even the experienced eye, except as regional effacement of the normal surface mucosal pattern. Excavated cancers may closely mimic, in size and appearance, chronic peptic ulcers. In advanced cases, cancerous craters can be identified by their heaped-up, beaded margins and shaggy, necrotic bases, as well as by the overt neoplastic tissue extending into the surrounding mucosa and wall ( Fig. 17-26 ). Uncommonly,
Figure 17-27 Gastric carcinoma. A, Intestinal type demonstrating gland formation by malignant cells, which are invading the muscular wall of the stomach. B, Diffuse type demonstrating signet-ring carcinoma cells.
a broad region of the gastric wall or the entire stomach is extensively infiltrated by malignancy, creating a rigid, thickened "leather bottle," termed linitis plastica. Metastatic carcinoma, from the breast and lung, may generate a similar picture.
The histologic subtypes of gastric cancer have been variously subclassified, but the two most important types, as noted earlier, are the intestinal type and diffuse type of the Lauren classification ( Fig. 17-27 ). The intestinal variant is composed of neoplastic intestinal glands resembling those of colonic adenocarcinoma (see Fig. 17-27 A ), which permeate the gastric wall but tend to grow along broad cohesive fronts in an "expanding" growth pattern. The neoplastic cells often contain apical mucin vacuoles, and abundant mucin may be present in gland lumens. The diffuse variant is composed of gastric-type mucous cells, which generally do not form glands, but rather permeate the mucosa and wall as scattered individual cells or small clusters in an "infiltrative" growth pattern. These cells appear to arise from the middle layer of the mucosa, and the presence of intestinal metaplasia is not a prerequisite. In this variant, mucin formation expands the malignant cells and pushes the nucleus to the periphery, creating a "signet ring" conformation (see Fig. 17-27 B ). If the signet-ring cells are more than 50% of the tumor, the tumor is classified as signet-ring cell carcinoma under the WHO classification. Regardless of cell type, the amount of mucin formation varies, and in poorly differentiated portions of the tumor it may be absent. Conversely, excessive mucin production may generate large mucinous lakes that dissect tissue planes; isolated tumor cells or glands may be difficult to identify in such areas. Infiltrative tumors often evoke a strong mural desmoplastic reaction, in which the scattered cells are embedded; the fibrosis creates local rigidity of the wall, which provides a valuable clue to the presence of an infiltrative lesion.
Whatever the classification and variant, all gastric carcinomas eventually penetrate the wall to involve the serosa and spread to regional and more distant lymph nodes. For
obscure reasons, gastric carcinomas frequently metastasize to the supraclavicular sentinel (Virchow) node as the first clinical manifestation of an
826
occult neoplasm. The tumor can also metastasize to the periumbilical region to form a subcutaneous nodule. This nodule is called a Sister Mary Joseph nodule, after the nun who noted this lesion as a marker of metastatic carcinoma. Local invasion of gastric carcinoma into the duodenum, pancreas, and retroperitoneum also is characteristic. At the time of death, widespread peritoneal seeding and metastases to the liver and lungs are common. A notable site of visceral metastasis is to one or both ovaries. Although uncommon, metastatic adenocarcinoma to the ovaries (from stomach, breast, pancreas, and even gallbladder) is so distinctive as to be called Krukenberg tumor.
Clinical Features.
Gastric carcinoma is an insidious disease that is generally asymptomatic until late in its course. The symptoms include weight loss, abdominal pain, anorexia, vomiting, altered bowel habits, and less frequently dysphagia, anemic symptoms, and hemorrhage. As these symptoms are essentially nonspecific, early detection of gastric cancer is difficult. The proportion of cancers diagnosed as early gastric carcinoma clearly depends on the intensity of the diagnostic effort to uncover asymptomatic disease. In Japan, where mass endoscopy screening programs are in place, early gastric cancer constitutes about 35% of all newly diagnosed gastric cancers. In Europe and the United States, this figure has remained at 10% to 15% over several decades.
The prognosis for gastric carcinoma depends primarily on the depth of invasion and the extent of nodal and distant metastasis at the time of diagnosis; histologic type has minimal independent prognostic significance. The prognostic value of biomarkers such as p53 mutation and c-ERB-B2 amplification remains to be determined. Clinical prognosis of gastric cancer largely depends on the depth of tumor invasion and the presence or absence of nodal or visceral metastasis. Surgical resection is still the standard treatment option, without or with adjuvant chemotherapy and radiation. The five-year survival rate of surgically treated early gastric cancer is 90% to 95%, with only a small negative increment if lymph node metastases are present. In contrast, the five-year survival rate for advanced gastric cancer remains below 15%.
LESS COMMON GASTRIC TUMORS
Gastric Lymphoma.
Gastric lymphomas represent 5% of all gastric malignancies. However, the stomach is the most common site for extranodal lymphoma (20% of such cases). Nearly all gastric lymphomas are B-cell lymphomas of mucosa-associated lymphoid tissue (MALT lymphomas). Nodal-type lymphomas that may develop in the stomach are unrelated to MALT lymphomas and similar to lymphomas originating in lymph nodes (discussed in
Chapter 14 ). While some gastric lymphomas appear to arise de novo, the majority (>80%) are associated with chronic gastritis and H. pylori infection. The role of H. pylori infection as an important etiologic factor for gastric lymphoma is supported by the elimination of about 50% of gastric lymphomas with antibiotic treatment for H. pylori. Tumors that do not regress with this type of treatment usually contain genetic abnormalities, particularly Trisomy 3 and t(11;18) translocation. This translocation brings together the API2 (apoptosis-inhibitor 2) gene on chromosome 11 with the MLT (mutated in MALT lymphoma) gene on chromosome 18. The protein encoded by the fused genes is thought to inhibit apoptosis, but its precise contribution to the development of MALT lymphoma remains to be established.
Morphology.
Gastric lymphoma commonly occurs in the mucosa or superficial submucosa. In the MALT lymphoma, a monomorphic lymphocytic infiltrate of the lamina propria surrounds gastric glands massively infiltrated with atypical lymphocytes and undergoing destruction (the "lymphoid epithelioid" lesion; Fig. 17-28 ). These gut-type lymphomas are usually CD5, CD10, and CD23 negative. In contrast, nodal-type lymphomas exhibit features characteristic of lymphomas arising de novo in lymph nodes, with frequently positive immunoreactivity for CD5, GYCLIN D1, CD10, or BCL-2. Rarely, Burkitt lymphoma, AIDS-associated lymphoma, and Hodgkin lymphoma may occur in the stomach.
Gastrointestinal Stromal Tumor.
A wide variety of mesenchymal neoplasms may arise in the stomach. Those originating in nerve sheaths are known as schwannomas. All of these tumors are rare. Much more common are gastrointestinal stromal tumors, also called GISTs. It is thought that GISTs originate from the interstitial cells of Cajal, which control gastrointestinal peristalsis. These tumors have a special phenotype in that 95% of them stain with antibodies against c-KIT, and approximately 70% stain for CD34. Despite these phenotypic similarities, GISTs show different histological patterns, and can be sub-classified into spindle and epithelioid types. Tumors that show features of enteric plexus differentiation (called gastrointestinal autonomic nerve tumors or GANTs) are often classified among GISTs. On rare occasions, gastric GISTs occur as part of a tumor syndrome, such as Carney's triad (gastric GIST, paraganglioma and pulmonary chondroma), or neurofibromatosis type 1.
Morphology.
GISTs can be solitary or multiple. The tumor can protrude into the lumen with an overlying attenuated mucosa or extrude on the serosal side of
Figure 17-28 Gastric MALT lymphoma. Note the lymphoepithelial lesions (arrows). (Courtesy of Dr. Melissa Li, University of Florida, Gainesville, FL.)
827
Figure 17-29 Gastrointestinal stromal tumor. A, Gross photograph of the tumor arising from the muscularis propria of the gastric wall. B, Microscopic view of the tumor showing spindle cell feature. C, Immunohistochemical stain showing the tumor cell c-KIT positivity.
the gastric wall ( Fig. 17-29 A ). The cut surface of the tumor is tan and usually lacks the whirling smooth muscle pattern of leiomyomas or leiomyosarcomas. It varies from slightly firm to soft, and hemorrhagic changes are common. Necrosis or cystic changes can be seen in a large tumor. Microscopically, the tumor can exhibit spindle cells ( Fig. 17-29 B ), plump "epithelioid" cells, or a mixture of both. Most of the tumors are quite cellular, and mitotic activity is variable. The majority of the tumor cells are positive for c-KIT (CD117), as demonstrated by positive immunohistochemical staining ( Fig. 17-29 C ).
Pathogenesis.
The identification of c-KIT mutations and platelet-derived growth factor receptor-α (PDGFRA) mutations in these tumors constitute significant progress in understanding the pathogenesis of GISTs.[39] [40] c-KIT is the receptor for stem cell factor, and PDGFRA is a receptor for platelet-derived growth factor (PDGF). It is known that 85% of GISTs have c-KIT mutations and 35% of GISTs with normal c-KIT contain PDGFRA mutations.[41] Both c-KIT and PDGFRA have cytoplasmic tyrosine kinases that activate similar intracellular pathways. The mutations lead to constitutive activation of the tyrosine kinase signaling pathway, promoting cell proliferation and inhibiting apoptosis. c-KIT mutations and PDGFRA mutations appear to be mutually exclusive.
Based on these pathogenic insights, a newly designed tyrosine kinase inhibitor (STI571) has been shown to be effective against this tumor.[42] Recall that this drug is used to treat chronic myeloid leukemia, also associated with abnormal tyrosine kinase activity ( Chapter 14 ). Currently, STI571 is widely used as an agent to treat GISTs; this demonstrates the application of a targeted therapeutic approach to the treatment of human malignancies.
Gastric Neuroendocrine Cell (Carcinoid) Tumors.
Most gastric carcinoid tumors originate from the ECL cells in the oxyntic mucosa. The tumor can arise in the setting of chronic atrophic gastritis or multiple endocrine neoplasia type 1 (MEN1) and Zollinger-Ellison syndrome. The underlying pathogenesis is probably related to the hypergastrinemic state, resulting in ECL cell hyperplasia, a presumed pretumorous condition. Less common is the sporadic gastric carcinoid without a hypergastrinemic state, for which the pathogenesis is not known. Gastric carcinoid tumors exhibit similar histologic features to other carcinoid tumors. The clinical course is quite variable. To date, there are no reliable pathologic markers to predict the tumor behavior.
Lipomas.
Lipomas are a benign neoplasm of adipose tissue, usually present in the submucosa.
Metastatic involvement of the stomach is unusual. The most common sources of gastric metastases are systemic lymphomas. Metastases of malignant melanoma and carcinomas tend to be multiple and may develop central ulceration. Breast and lung carcinoma may mimic diffuse gastric carcinoma by diffusely infiltrating the gastric wall to generate linitis plastica, as described earlier for primary gastric carcinoma.
828
Small and Large Intestines
Normal
Anatomy
The small intestine in the human adult is approximately 6 meters in length, and the colon (large intestine) approximately 1.5 meters. The first 25 cm of the small intestine, the duodenum, are retroperitoneal; the jejunum marks the entry of the small intestine into the peritoneal cavity, terminating where the ileum enters the colon at the ileocecal valve. The demarcation between jejunum and ileum is not clearly defined; the jejunum arbitrarily constitutes the proximal third of the intraperitoneal portion and the ileum the remainder. The colon is subdivided into the cecum and the ascending, transverse, and descending colon. The sigmoid colon begins at the pelvic brim and loops within the peritoneal cavity, becoming the rectum at about the level of the third sacral vertebra. Halfway along its 15-cm length, the rectum passes between the crura of the peroneal muscles to become extraperitoneal. The reflection of the peritoneum from the rectum over the pelvic floor creates a cul de sac known as the pouch of Douglas.
Vasculature
The arterial supply of the intestine, from the proximal jejunum to the hepatic flexure of the colon, is derived from the superior mesenteric artery. The inferior mesenteric artery feeds the remainder of the colon to the level of the rectum. Each artery progressively divides as it approaches the gut, with rich arterial interconnections via arching mesenteric arcades. Numerous
Figure 17-30 A, Normal small-bowel histology, showing mucosal villi and crypts, lined by columnar cells. B, Normal colon histology, showing flat mucosal surface and abundant vertically oriented crypts.
collaterals connect the mesenteric circulation with the celiac arterial axis proximally and the pudendal circulation distally. The lymphatic drainage essentially parallels the vascular supply but does not have the intricate patterns of arcades.
The upper rectum is supplied by the superior hemorrhoidal branch of the inferior mesenteric artery. The lower portion receives its blood supply from the hemorrhoidal branches of the internal iliac or internal pudendal artery. The venous drainage follows essentially the same distribution and is connected by an anastomotic capillary bed between the superior and inferior hemorrhoidal veins, providing a connection between the portal and systemic venous systems. Since the colon is a retroperitoneal organ in the ascending and descending portions, it derives considerable accessory arterial blood supply and lymphatic drainage from a wide area of the posterior abdominal wall.
Small Intestinal Mucosa
The most distinctive feature of the small intestine is its mucosal lining, which is studded with innumerable villi ( Fig. 17-30 A ). These extend into the lumen as finger-like projections covered by epithelial lining cells. The central core of lamina propria contains blood vessels, lymphatics, a minimal population of lymphocytes, eosinophils and mast cells, and scattered fibroblasts and vertically oriented smooth muscle cells. Between the bases of the villi are the pitlike crypts of Lieberkühn, which contain stem cells that replenish and regenerate the epithelium. The crypts extend down to the muscularis
mucosa. The muscularis mucosa is a smooth, continuous sheet, serving to anchor the configuration of villi and
829
crypts alike. In normal individuals, the villus-to-crypt height ratio is about 4 to 5:1. Within the duodenum are abundant submucosal mucous glands, termed Brunner glands. These glands secrete bicarbonate ions, glycoproteins, and pepsinogen II and are virtually indistinguishable from the pyloric mucous glands.
The surface epithelium of the villi contains three cell types. Columnar absorptive cells are recognized by the dense array of microvilli on their luminal surface (the brush border) and the underlying mat of microfilaments (the terminal web). Interspersed regularly between the absorptive cells are mucin-secreting goblet cells and a few endocrine cells, described below. Within the crypts reside stem cells, goblet cells, more abundant endocrine cells, and scattered Paneth cells. Paneth cells have apically oriented bright eosinophilic granules containing a variety of antimicrobial proteins (such as defensins), which play a role in mucosal innate immunity against bacterial infection.[43]
The villi of the small intestinal mucosa are the site for terminal digestion and absorption of foodstuffs through the action of the columnar absorptive cells. The crypts secrete ions and water, deliver immunoglobulin A (IgA) and antimicrobial peptides to the lumen, and serve as the site for cell division and renewal. The mucous cells of both crypts and villi generate an adherent mucous coat, which both protects the surface epithelium and provides an ideal local milieu for uptake of nutrients. Specific receptors for uptake of macromolecules are also present on the surface epithelial cells, such as those in the ileum for intrinsic factor-vitamin B12 complexes.
Colonic Mucosa
The small intestine accomplishes its absorptive function with a highly liquid luminal stream. The function of the colon is to reclaim luminal water and electrolytes. Unlike the mucosa of the small intestine, the colonic mucosa has no villi and is flat. The mucosa is punctuated by numerous straight tubular crypts that extend down to the muscularis mucosa ( Fig. 17-30 B ). The surface epithelium is composed of columnar absorptive cells, which have shorter and less abundant microvilli than found in the small intestine, and goblet mucous cells. The crypts contain abundant goblet cells, endocrine cells, and stem cells. Paneth cells are occasionally present at the base of crypts in the cecum and ascending colon. The intestinal mucosa, particularly in the ileum, is colonized by endogenous bacteria, particularly non-pathogenic strains of E. coli and organisms such as Proteus, Enterobacter, Serratia and Klebsiella. The components of the endogenous flora may be displaced by exogenous bacteria, such as pathogenic strains of E. coli that cause damage to the mucosa.
The regenerative capacity of the intestinal epithelium is remarkable. Cellular proliferation is confined to the crypts; differentiation and luminal migration serve to replenish superficial cells lost to senescence and surface abrasion. Within the small intestine, cells migrate out of the crypts and upward to the tips of the villi, where they are shed into the lumen. This journey normally takes between 96 and 144 hours, leading to normal renewal of the epithelial lining every 4 to 6 days. Turnover of the colonic surface epithelium takes 3 to 8 days. The rapid renewal of intestinal epithelium provides a remarkable capacity for repair but also renders the intestine particularly vulnerable to agents that interfere with cell replication, such as radiation and chemotherapy for cancer.
Endocrine Cells
A diverse population of endocrine cells is scattered among the epithelial cells lining the gastric glands, small intestinal villi, and small and large intestinal crypts. Comparable cells are present in the epithelia lining the pancreas, biliary tree, lung, thyroid, and urethra. As a population, gut endocrine cells exhibit characteristic morphologic features. In most cells, the cytoplasm contains abundant fine eosinophilic granules, which harbor secretory products. The main portion of the cell is at the base of the epithelium, and the nuclei reside on the luminal side of the cytoplasmic granules.
These cells exhibit a marked diversity of secretory peptides and distribution of cell subtypes. Secretory granules are released at the basal surface of the endocrine cell or along the basal part of its lateral surface; apical secretion (into the lumen) has not been observed. The various secretory products, some of which are also present in the mural autonomic neural plexus, act as chemical messengers and modulate normal digestive functions by a combination of endocrine, paracrine, and neurocrine mechanisms. Each endocrine cell type, therefore, exhibits a distribution tailored to meet the physiologic needs pertinent to a gut segment.
Intestinal Immune System
Humans are exposed to an enormous load of environmental antigens through the gastrointestinal tract. The surface area of the gastrointestinal tract through which ingested antigens may enter far exceeds that of the skin and pulmonary tract. The immune system
must balance tolerance of harmless ingested substances against active defense reactions to potential microbial invaders. Dysfunction of this regulatory machinery may cause smoldering chronic disease and, occasionally, life-threatening acute conditions. Throughout the small intestine and colon are nodules of lymphoid tissue, which lie either within the mucosa or span the mucosa and a portion of the submucosa.[44] The lymphoid nodules distort the surface epithelium to produce broad domes rather than villi; within the ileum confluent lymphoid tissue becomes macroscopically visible as Peyer patches. The surface epithelium over lymphoid nodules contains both columnar absorptive cells and M (membranous) cells, the latter found only in small and large intestinal lymphoid sites. M cells are able to transcytose antigenic macromolecules intact from the lumen to antigen-presenting cells under the surface epithelium. Antigen-presenting cells include macrophages and dendritic cells. Throughout the intestines, T lymphocytes are scattered within the surface epithelium, usually at the basolateral aspects of the cell. These T cells are referred to as intraepithelial lymphocytes and include cytotoxic CD8+ cells. The lamina propria contains helper T cells (CD4+), activated B cells, and plasma cells. The lamina propria plasma cells secrete dimeric IgA, IgG, and IgM, which enter into the splanchnic circulation. IgA is transcytosed directly across enterocytes or across hepatocytes for secretion into bile; both are mechanisms for delivering IgA to
830
the intestinal lumen. The intestinal lymphoid nodules, mucosal lymphocytes, and isolated lymphoid follicles in the appendix and mesenteric lymph nodes constitute the MALT, mentioned in the discussion of gastric tumors.
Neuromuscular Function
Small intestinal peristalsis, both anterograde and retrograde, mixes the food stream and promotes maximal contact of nutrients with the mucosa. Colonic peristalsis prolongs contact of the luminal contents with the mucosa. Although intestinal smooth muscle cells are capable of initiating contractions, both small and large intestinal peristalsis is mediated by intrinsic (myenteric plexus) and extrinsic (autonomic innervation) neural control. The myenteric plexus consists of two neural networks: Meissner plexus resides at the base of the submucosa, and Auerbach plexus lies between the inner circumferential and outer longitudinal muscle layers of the muscle wall; lesser neural twigs extend between smooth muscle cells and ramify within the submucosa.
Many conditions, such as infections, inflammatory diseases, motility disorders, and tumors, affect both the small and large intestines. These two organs will therefore be considered together. Collectively, disorders of the intestines account for a large portion of human disease.
Congenital Anomalies
Rare anomalies of gut formation may occur.
• Duplication of the small intestine or colon, usually in the form of saccular to long, cystic structures • Malrotation of the entire bowel, resulting from improper embryologic rotation of the gut • Omphalocele, in which the abdominal musculature fails to form, leading to birth of an infant with herniation of abdominal contents into a ventral membranous sac • Gastroschisis, in which a portion of the abdominal wall fails to form altogether, causing extrusion of the intestines.
The above lesions may be silent (malrotation) or catastrophic (gastroschisis). A far more common and innocuous lesion is heterotopia of normal pancreatic tissue but occasionally of gastric mucosa. Both heterotopias may occur anywhere in the intestine and usually are small, 1- to 2-cm nodules in the mucosa or intestinal wall.
ATRESIA AND STENOSIS
Congenital intestinal obstruction is an uncommon but dramatic lesion that may affect any level of the intestines. Duodenal atresia is most common; the jejunum and ileum are equally involved, but the colon is not involved. The obstruction may be complete (atresia) or incomplete (stenosis). Atresia may take the form of an imperforate mucosal diaphragm or a stringlike segment of bowel connecting intact proximal and distal intestine. Stenosis is less common and is due to a narrowed intestinal segment or a diaphragm with a narrow central opening. Single or multiple lesions appear to arise from developmental failure, intrauterine vascular accidents, or intussusceptions (telescoping of one intestinal segment within another) occurring after the intestine has developed. Failure of the cloacal diaphragm to rupture leads to an imperforate anus.
MECKEL DIVERTICULUM
Failure of involution of the vitelline duct, which connects the lumen of the developing gut to the yolk sac, produces a Meckel diverticulum. This solitary diverticulum lies on the antimesenteric side of the bowel, usually within 2 feet (85 cm) of the ileocecal valve ( Fig. 17-31 ). This is a true diverticulum, in that it contains all three layers of the normal bowel wall: mucosa, submucosa, and muscularis propria. Meckel diverticula may take the form of only a small pouch, or of a blind segment having a lumen greater in diameter than that of the ileum and a length of up to 6 cm. Although the mucosal lining may be that of normal small intestine, heterotopic rests of gastric mucosa or pancreatic tissue are found in about one half of these anomalies. Meckel diverticula are present in an
estimated 2% of the normal population, but most remain asymptomatic or are discovered incidentally. When peptic ulceration occurs in the small intestinal mucosa adjacent to the gastric mucosa, intestinal bleeding or symptoms resembling those of an acute appendicitis may result. Alternatively, presenting symptoms may be related to intussusception, incarceration, or perforation.
Hirschsprung disease is a congenital disorder characterized by aganglionosis of a portion of the intestinal tract. The enteric neuronal plexus develops from neural crest cells, which migrate into the bowel wall during development, mostly in a cephalad to caudad direction. Congenital megacolon, or
Figure 17-31 Meckel diverticulum. The blind pouch is located on the antimesenteric side of the small bowel.
831
Hirschsprung disease, results when the migration of neural crest cells arrests at some point before reaching the anus or when the ganglion cells undergo inappropriate premature death.[45] This produces an intestinal segment that lacks both Meissner submucosal and Auerbach myenteric plexuses. Depending on the severity of migration arrest, a variable length of the distal gut is not innervated. Loss of enteric neural coordination leads to functional obstruction and intestinal dilation proximal to the affected segment. Note that the dilated segment may contain normal ganglia; ganglia are absent or nearly so in the contracted region.
The cause of the developmental failure is unclear, but at least eight susceptibility genes have been identified. Mutations in these genes are associated with varying degrees of intestinal aganglionosis and other congenital anomalies. The phenotypic expression (penetrance) among these gene mutations varies. The long-segment and short-segment
disease (see explanation under "Morphology") appear to have different modes of inheritance.[46] [47] [48] About 50% of familial cases and approximately 15% of sporadic cases are a consequence of mutations in the RET gene ( Chapter 7 ) that inactivate the kinase activity of this receptor. A much smaller proportion of cases (3%–5%) may be caused by mutations in the endothelin/endothelin-receptor system. RET and its ligands (members of the Glial-Derived Neurotrophic Factor family) promote survival and growth of neurites, and provide direction to migrating neural crest cells. The endothelin system participates in the regulation of morphogenesis during embryonic development. Despite the identification of the involvement of these and other genes in Hirschprung disease pathogenesis, the genetic defect is unknown in a large number of cases.
Morphology.
Hirschsprung disease is characterized by the absence of ganglion cells and ganglia in the muscle wall and submucosa of the affected segment. The rectum is always affected, with involvement of more proximal colon to variable extent. Most cases involve the rectum and sigmoid only (short-segment disease), with longer segments in a fifth of cases, and rarely the entire colon (long-segment disease). Absence of mural ganglion cells is sometimes accompanied by thickening and hypertrophy of nonmyelinated nerve fibers, representing ramifications of the lumbosacral preganglionic fibers. Proximal to the aganglionic segment, the colon undergoes progressive dilation and hypertrophy, beginning with the descending colon. With time, the proximal innervated colon may become massively distended, sometimes achieving a diameter of 15 to 20 cm (megacolon). When distention outruns the hypertrophy, the colonic wall becomes markedly thinned and may rupture, usually near the cecum. Mucosal inflammation or shallow, so-called stercoral ulcers may appear. Unequivocal diagnosis of Hirschsprung disease can be made histologically by the failure to detect ganglion cells in intestinal submucosa samples stained for acetylcholinesterase.
Clinical Features.
Hirschsprung disease occurs in approximately 1 out of 5,000 live births and is present with increased frequency (3.6%) in siblings of index cases. Males predominate 4:1. Short-segment aganglionosis with megacolon is more common in males, whereas females predominate among patients with long affected segments. Ten percent of all cases of Hirschsprung disease occur in children with Down syndrome, and serious neurologic abnormalities are present in another 5%, raising the possibility that this disease is only one feature of more generalized abnormal development of the neural crest.[47] [48]
Hirschsprung disease usually manifests itself in the immediate neonatal period by failure to pass meconium, followed by obstructive constipation. In those instances when only a few centimeters of the rectum are affected, the build-up of pressure may permit occasional passage of stools or even intermittent bouts of diarrhea. Abdominal distention develops if a sufficiently large segment of colon is involved. The major threats to life in this disorder are superimposed enterocolitis with fluid and electrolyte disturbances and perforation of the colon or appendix with peritonitis.
Acquired megacolon is a condition that may occur at any age and may result from: (1) Chagas disease (see Chapter 8 ), in which the trypanosomes directly invade the bowel wall to destroy the enteric plexuses; (2) organic obstruction of the bowel, by a neoplasm or inflammatory stricture; (3) toxic megacolon complicating ulcerative colitis or Crohn disease (see later); or (4) a functional psychosomatic disorder. Save for Chagas disease, where inflammatory involvement of the ganglia is evident, the remaining forms of megacolon are not associated with any deficiency of mural ganglia.
Enterocolitis
Diarrheal diseases of the bowel make up a veritable Augean stable of entities (a messy situation cleaned by the fifth task of Hercules). Many are caused by microbiologic agents; others arise in the setting of malabsorptive disorders and idiopathic inflammatory bowel disease. Consideration should first be given to the conditions known as diarrhea and dysentery.
DIARRHEA AND DYSENTERY
A healthy adult drinks 2 L of fluid per day, to which is added 1 L of saliva, 2 L of gastric juice, 1 L of bile, 2 L of pancreatic juice, and 1 L of intestinal secretions. Of these 9 L of fluid presented to the intestine, less than 200 gm of stool are excreted per day, of which 65% to 85% is water. Jejunal absorption of water amounts to 3 to 5 L/day, ileal absorption 2 to 4 L/day. The colon normally absorbs 1 to 2 L/day, but is capable of absorbing almost 6 L/day.
A precise definition of diarrhea is elusive, given the considerable variation in normal bowel habits. An increase in stool mass, stool frequency, and/or stool fluidity are perceived as diarrhea by most patients. For many individuals, this consists of daily stool production in excess of 250 gm, containing 70% to 95% water. However, over 14 L of fluid may be lost per day in severe cases of diarrhea (i.e., the equivalent of the circulating blood volume). Diarrhea is often accompanied by pain, urgency, perianal discomfort, and incontinence. Low-volume, painful, bloody diarrhea is known as dysentery.
The major causes of diarrhea are presented in Table 17-6 . The principal mechanisms of diarrhea, one or more of which may be operative in any one patient, are as follows:
• Secretory diarrhea: Net intestinal fluid secretion leads to the output of more than 500 mL of fluid stool per day, which is isotonic with plasma and persists during fasting. • Osmotic diarrhea: Excessive osmotic forces exerted by luminal solutes lead to output of more than 500 mL of stool per day, which abates upon fasting. Stool exhibits an osmotic gap (stool osmolality exceeds plasma electrolyte concentration by ≥50 mOsm). • Exudative diseases: Mucosal destruction leads to output of purulent, bloody stools that persist on fasting; stools are frequent but may be small or large volume. • Deranged motility: Improper gut neuromuscular function may produce highly variable patterns of increased stool volume; other forms of diarrhea must be excluded. • Malabsorption: Improper absorption of gut nutrients produces voluminous, bulky stools with increased osmolarity combined with excess stool fat (steatorrhea). The diarrhea usually abates on fasting.
INFECTIOUS ENTEROCOLITIS
Intestinal diseases of microbial origin are marked principally by diarrhea and sometimes ulcerative and inflammatory changes in the small and/or large intestine. Infectious enterocolitis is a global problem of staggering proportions, causing more than 12,000 deaths per day among children in developing countries, and constituting one half of all deaths before age 5 worldwide. [49] Although far less prevalent in industrialized nations,
these infections still have attack rates of one to two illnesses per person per year, second only to the common cold in frequency. This results in an estimated 99 million acute cases of either vomiting or diarrhea per year in the United States, equivalent to 40% of the population. The infections are mainly associated with contaminated food and water.
Acute, self-limited infectious diarrhea, which is a major cause of morbidity among children, is most frequently caused by enteric viruses. In infants, infectious diarrhea may cause severe dehydration and metabolic acidosis, which may result in hospitalization in developed countries and death in developing countries. Bacterial infections, such as enterotoxigenic Escherichia coli, are also common offenders. However, many pathogens can cause diarrhea; the major offenders vary with the age, nutrition, immune status of the host, environment (living conditions, public health measures), and special predispositions, such as hospitalization, wartime dislocation, or foreign travel. In 40% to 50% of cases, the specific agent cannot be isolated.
Viral Gastroenteritis
Symptomatic human infection is caused by several distinct groups of viruses ( Table 17-7 ). Rotavirus accounts for an estimated 140 million cases and 1 million deaths worldwide per year. The target population is children age 6 to 24 months, but young infants and debilitated adults are susceptible to symptomatic infection. This virus accounts for 25% to 65% of severe diarrhea in infants and young children.[50] Rotavirus is an encapsulated virus with a segmented double-stranded RNA genome. Rotavirus selectively infects and destroys mature enterocytes in the small intestine, without infecting crypt cells. The surface epithelium of the villus is repopulated by immature secretory cells. With the loss of absorptive function and excess of secretory cells, there is net secretion of water and electrolytes, compounded by an osmotic diarrhea from incompletely absorbed nutrients. The minimal infective inoculum is approximately 10 viral particles, whereas an individual with rotavirus gastroenteritis typically sheds up to 1012 particles/mL stool. Thus, outbreaks among pediatric populations in hospitals and day-care centers are very common. The clinical syndrome has an incubation period of approximately 2 days, which is followed by vomiting and watery diarrhea for several days. Viral infection can induce protective immunity, but the protection for reinfection is often short-lived. Antirotavirus
Person-to-person, water, cold foods, raw shellfish
1–3 days/4 days
Norwalk-like viruses
Sapporo-like viruses
Enteric adenoviruses
dsDNA 80 8 Child <2 years
Person-to-person
3–10 days/7+ days
Astroviruses ssRNA 28 4 Child Person-to-person, water, raw shellfish
24–36 hours/1–4 days
Data from Goodgame RW: Viral causes of diarrhea. Gastroenterol Clin North Am 30:779,2001.
ds, double-stranded; ss, single stranded.
antibodies are present in mother's milk, so rotavirus infection is most frequent at the time of weaning.
Among the numerous types of adenovirus, the subtypes (enteric serotypes) Ad40, Ad41, and Ad31 appear to be responsible for enteric infections and are a common cause of diarrhea among infants. They can be distinguished from adenoviruses that cause respiratory disease by their failure to grow easily in culture. Adenoviruses cause a moderate gastroenteritis with diarrhea and vomiting, lasting for a week to 10 days after an incubation period of approximately 1 week. In the small intestine, adenoviral infection causes atrophy of the villi and compensatory hyperplasia of the crypts similar to rotavirus, resulting in malabsorption and fluid loss. The virus can also cause colitis. Immunohistochemical stain of nuclear inclusions facilitates the diagnosis.
Caliciviruses include two major groups: the classic Caliciviruses (Sapporo-like viruses) and the Norwalk-like viruses (small round structured viruses). Sapporo-like viral infection is rare, while Norwalk virus, the prototype of Norwalk-like viruses, is responsible for the majority of cases of nonbacterial food-borne epidemic gastroenteritis in all age groups. Norwalk-like viruses are small icosahedral viruses containing a single-stranded RNA genome. They cause epidemic gastroenteritis with diarrhea, nausea, and vomiting among children. Outbreaks occur following exposure of multiple individuals to a common source. The clinical syndrome has an incubation period of 1 to 2 days, which is followed by 12 to 60 hours of nausea, vomiting, watery diarrhea, and abdominal pain.
Astrovirus is named after its starlike appearance. It primarily affects children, (it accounts for 4% of acute gastroenteritis in young children), and has a worldwide distribution. Those infected develop anorexia, headache, and fever. Other viruses such as enterotrophic coronaviruses and toroviruses are occasionally implicated in human diarrheal disease.
Despite the high incidence of viral gastroenteritis, insights into disease pathogenesis have been slow in coming.
Morphology.
Although the enteric viruses are genetically and morphologically different from each other, the lesions they cause in the intestinal tract are similar. The small intestinal mucosa usually exhibits modestly shortened villi and infiltration of the lamina propria by lymphocytes. Vacuolization and loss of the microvillus brush border in surface epithelial cells may be evident, and the crypts become hypertrophied. Viral particles may be visualized by electron microscopy within surface epithelial cells. In infants, rotavirus can produce a flat mucosa resembling celiac sprue (discussed later).
Bacterial Enterocolitis
Diarrheal illness may be caused by numerous bacteria ( Table 17-8 ). There are several pathogenic mechanisms for bacterial enterocolitis (also termed food poisoning):
• Ingestion of preformed toxin, present in contaminated food. Major offenders are Staphylococcus aureus, Vibrio, and Clostridium perfringens. Symptoms develop within a matter of hours; explosive diarrhea and acute abdominal distress herald an illness that passes within a day or so. Ingested systemic neurotoxins, as from Clostridium botulinum, may produce rapid, and fatal, respiratory failure. • Infection by toxigenic organisms, which proliferate within the gut lumen and elaborate an enterotoxin. An incubation period of several hours to days is followed by diarrhea and dehydration if the primary pathogenic mechanism is a secretory enterotoxin, or dysentery if the primary mechanism is a cytotoxin. Traveler's diarrhea (Montezuma's revenge, turista) usually occurs following ingestion of fecally contaminated food or water; it begins abruptly and subsides within 2 to 3 days. It affects 20% to 50% of the 35 million people who travel worldwide from industrialized countries to developing countries each year.
• Infection by enteroinvasive organisms, which proliferate, invade, and destroy mucosal epithelial cells, also leading to dysentery. As with ingestion of toxigenic organisms, the incubation period is several hours to days.
The main properties of bacteria that contribute to the pathogenesis of enterocolitis are: (1) the ability to adhere to the mucosal epithelial cells and replicate, (2) the ability to elaborate enterotoxins, and (3) the capacity to invade.
Bacterial Adhesion and Replication.
In order to produce disease, ingested organisms must adhere to the mucosa; otherwise they will be swept away by the fluid stream. Adherence of enterotoxigenic organisms such as E. coli and Vibrio cholerae is mediated by plasmid-encoded adhesins. These proteins are expressed on the surface of the organism, sometimes in the form of fimbriae or pili, which are rigid or wiry surface
834
TABLE 17-8 -- Major Causes of Bacterial Enterocolitis
OrganismPathogenic Mechanism Source Clinical Features
Escherichia coli
Traveler's diarrhea, including:
• ETEC Cholera-like toxin, no invasion
Food, water Watery diarrhea
• EHEC Shiga-like toxin, no invasion
Undercooked beef products
Hemorrhagic colitis, hemolytic-uremic syndrome
• EPEC Attachment, enterocyte effacement, no invasion
Fever, pain, bloody diarrhea, following antibiotic use, nosocomial acquisition
Clostridium perfringens
Enterotoxin, no invasion
Meat, poultry, fish
Watery diarrhea, food sources, "pigbel"
Mycobacterium tuberculosis
Invasion, mural inflammatory foci with necrosis and scarring
Contaminated milk, swallowing of coughed-up organisms
Chronic abdominal pain; complications of malabsorption, stricture, perforation, fistulae, hemorrhage
ETEC, enterotoxigenic E. coli; EHEC, enterohemorrhagic E. coli; EPEC, enteropathogenic E. coli; EIEC, enteroinvasive E. coli.
projections. Adherence of enteropathogenic and enterohemorrhagic organisms, including E. coli and Shigella, is also dependent on plasmid-encoded proteins, but the nature of these proteins is not known. Adherence causes effacement of the apical enterocyte membrane, with destruction of the microvillus brush border and changes in the underlying cell cytoplasm.[35] The factors regulating bacterial replication are not well understood, particularly since pathogenic organisms must compete with the normal bacterial flora to achieve a critical population density.
Bacterial enterotoxins are polypeptides that cause diarrhea. Some enterotoxins cause intestinal secretion of fluid and electrolytes without causing tissue damage; this is accomplished by binding of the toxin to the epithelial cell membrane, entry of a portion of the toxin into the cell, and massive activation of electrolyte secretion accompanied by water. Cholera toxin, elaborated by Vibrio cholerae, is the prototype secretagogue toxin. The toxin causes increased levels of intracellular calcium, resulting in dysfunction of the fluid and electrolyte transport, as discussed below under Cholera. Strains of E. coli (enterotoxigenic E. coli) that produce heat-labile (LT) and heat-stable (ST) secretagogue toxins are the major cause of traveler's diarrhea. The LT toxin is similar to cholera toxin, and the ST toxin induces cyclic guanosine monophosphate, resulting in increased fluid excretion. Leukocytes are absent from the stool of patients with traveler's diarrhea. A second group of enterotoxins are cytotoxins, exemplified by Shiga toxin produced by Shigella dysenteriae and Shiga-like toxins produced by enterohemorrhagic E. coli (e.g., E. coli O157:H7). These toxins cause direct tissue damage through epithelial cell necrosis. Staphylococcal enterotoxins, which are major causes of food poisoning, represent yet another group of enterotoxins; are proteins that bind to the antigen receptors of large numbers of T cells and activate the lymphocytes to secrete cytokines. The cytokines stimulate intestinal motility and fluid secretion.
Bacterial Invasion.
Both enteroinvasive E. coli and Shigella possess a large virulence plasmid that confers the capacity for epithelial cell invasion, apparently by microbe-stimulated endocytosis. This is followed by intracellular proliferation, cell lysis, and cell-to-cell spread. Salmonella quickly pass through intestinal epithelial cells via transcytosis with minimal epithelial damage; entry into the lamina propria leads to a 5% to 10% incidence of bacteremia, which can sometimes cause typhoid fever, meningitis, endocarditis, and osteomyelitis (commonly in the setting of sickle cell disease). Yersinia enterocolitica penetrates the ileal mucosa and multiplies within Peyer patches and regional lymph nodes. Bacteremia is rare and usually occurs in the setting of iron-overload disease, since iron is a growth factor for Yersinia.
Bacterial cytotoxins and invasion give rise to bacillary dysentery, which generates its own unique misery for its victims: abdominal cramping and tenesmus with loose stools containing blood, pus, and mucus. Bacillary dysentery, which results in as many as 500,000 deaths among children in developing countries each year, is caused by Shigella dysenteriae, Shigella flexneri, Shigella boydii, and Shigella sonnei as well as certain O type enterotoxic E. coli. (Amebic dysentery is caused by the protozoan parasite Entamoeba histolytica, discussed later in this chapter).
Shigella Bacillary Dysentery
Shigella species are gram-negative facultative anaerobes that infect only humans. S. flexneri is the major cause of endemic bacillary dysentery in locations of poor hygiene, including large regions of the developing world and institutions in the
835
developed world. Epidemic shigellosis can occur when individuals consume uncooked foods at picnics or other events.
Pathogenesis.
Transmission is fecal-oral and is remarkable for the small number of organisms that may cause disease (10 ingested organisms cause illness in 10% of volunteers, and 500 organisms cause disease in 50% of volunteers). Shigella bacteria invade the intestinal mucosal cells but do not usually go beyond the lamina propria. Dysentery is caused when the bacteria escape the epithelial cell phagolysosome, multiply within the cytoplasm, and destroy host cells. Shiga toxin causes hemorrhagic colitis and hemolytic-uremic syndrome by damaging endothelial cells in the microvasculature of the colon and the glomeruli, respectively ( Chapter 20 ). In addition, chronic arthritis secondary to S. flexneri infection, called Reiter syndrome, may be caused by a bacterial antigen; the occurrence of this syndrome is strongly linked to HLA-B27 genotype, but the immunologic basis of this reaction is not understood. [51]
Salmonellosis and Typhoid Fever
Salmonellae are flagellated, gram-negative bacteria that cause a self-limited food-borne and water-borne gastroenteritis (S. enteritidis, S. typhimurium, and others) or a life-threatening systemic illness, typhoid fever, marked by fever and systemic symptoms (S. typhi). In the United States, Salmonella species cause approximately 500,000 reported cases of food poisoning, and many cases go unreported. Because Salmonella species other than S. typhi infect most commercially raised chickens and many cows, the major sources of Salmonella in the United States are feces-contaminated beef and chicken that are insufficiently washed and cooked. Stringent hygiene in the production plants and the home kitchen helps minimize risk of contamination. In contrast, humans are the only host of S. typhi, which is shed in the feces, urine, vomitus, and oral secretions by acutely ill persons and in the feces by chronic carriers without overt disease. Therefore, typhoid fever from S. typhi is a disease largely of developing countries, where sanitary conditions are insufficient to stop its spread. Typhoid fever is a protracted disease that is associated with bacteremia, fever, and chills during the first week; widespread mononuclear phagocyte involvement with rash, abdominal pain, and prostration in the second week; and ulceration of Peyer patches with intestinal bleeding and shock during the third week.
Pathogenesis.
Salmonella invades intestinal epithelial cells as well as tissue macrophages. Invasion of intestinal epithelial cells is controlled by invasion genes that are induced by the low oxygen tension found in the gut. These genes encode proteins involved in adhesion and in recruitment of host cytoskeletal proteins that internalize the bacterium. Similarly, intramacrophage growth is important in pathogenicity, and this seems to be mediated by bacterial genes that are induced by the acid pH within the macrophage phagolysosome. The enteric nervous system also is a critical regulator of fluid secretion in the normal gut.
Neural reflex pathways increase epithelial fluid secretion in response to enteric pathogens such as Salmonella and Clostridium difficile.[52]
Campylobacter Enterocolitis
This comma-shaped, flagellated, gram-negative organism was once classified with the vibrios. When special culture conditions permitted its isolation in the 1970s, it became apparent that Campylobacter was an important cause of enterocolitis and septicemia in humans. In the United States, Campylobacter jejuni is responsible for twice the enteric disease of Salmonella and four times that of Shigella. Most infections with Campylobacter are sporadic and are associated with ingestion of improperly cooked chicken, which may be contaminated with Campylobacter and/or Salmonella. Sporadic infections may also be associated with contact with infected dogs. Outbreaks of Campylobacter are usually associated with unpasteurized milk or contaminated water.
Pathogenesis.
Invasiveness is strain dependent. Flagella of Campylobacter, which give the organism its comma shape and motility, are necessary for the bacterium to penetrate mucus covering epithelial surfaces. Three clinical outcomes of Campylobacter infection are possible: (1) diarrhea, which is independent of bacterial invasion; (2) dysentery with blood and mucus in the stool; and (3) enteric fever when bacteria proliferate within the lamina propria and mesenteric lymph nodes. Postinfectious complications of Campylobacter infections include reactive arthritis in HLA-B27 carriers (as with Shigella infection) and Guillain-Barré syndrome, a demyelinating disease of peripheral nerves due to autoantibodies against gangliosides GM1 and GQ1b, described in Chapter 27 . Recently, C. jejuni was found to be associated with the development of immunoproliferative small intestinal disease (discussed later).
Cholera
Vibrio cholerae are comma-shaped, gram-negative bacteria that have been the cause of seven great long-lasting epidemics (pandemics) of diarrheal disease. Many of these pandemics began in the Ganges Valley of India and Bangladesh, which is never free from cholera, and then moved east. Although there are 140 serotypes of V. cholerae, until recently only the 01 serotype was associated with severe diarrhea. Beginning in 1992, a new V. cholerae serotype (0139, also known as Bengal) has been associated with severe, watery diarrhea.[53]
Pathogenesis.
The vibrios never invade the epithelium but instead remain within the lumen and secrete an enterotoxin, which is encoded by a virulence phage. Flagellar proteins involved in motility and attachment are necessary for efficient bacterial colonization, as has been described for Campylobacter. (This is in contrast to Shigella species and certain E. coli strains, which are nonmotile and yet invasive.) The Vibrio hemagglutinin, which is a metalloprotease, is important for detachment of Vibrio from epithelial cells.
The secretory diarrhea characteristic of the disease is caused by release of cholera toxin ( Fig. 17-32 ). Cholera toxin is composed of five binding peptides B and a catalytic peptide A. The B peptides, serving as a "landing pad," bind to carbohydrates on GM1 ganglioside on the surface of epithelial cells of the small intestine, enabling calveolar-mediated endosomal entry of toxin subunit A into the cell. Reverse transport of the subunit A from the endosome into the cell cytoplasm is followed by cleavage of the disulfide bond linking the two fragments of peptide A (A1 and A2). Catalytic peptide A1 is generated, leading to the following sequence:
• A1 interacts with 20-kD cytosolic proteins called ADP-ribosylation factors (ARF). • The ARF-A1 complex catalyzes ADP-ribosylation of a 49-kD G-protein (called Gsα ).[54] • Binding of NAD and GTP generates an activated Gsα , which in turn binds to and stimulates adenylate cyclase. ADP-ribosylated Gsα is permanently in an active GTP-bound
836
state, resulting in persistent activation of adenylate cyclase. • The activated adenylate cyclase generates high levels of intracellular cAMP from ATP. • Cyclic AMP stimulates secretion of chloride and bicarbonate, with associated sodium and water secretion. Chloride and sodium resorption are also inhibited.
Figure 17-32 Mechanisms of cholera toxin action.
The reabsorptive function of the colon is overwhelmed, and liters of dilute "rice water" diarrhea containing flecks of mucus—up to 14 L/day, equivalent to the circulating blood volume, causing dehydration and electrolyte imbalances. Because overall absorption in the gut remains intact, oral formulas can replace the massive sodium, chloride, bicarbonate, and fluid losses and reduce the mortality rate from 50% to less than 1%.
This entity is an acute colitis characterized by formation of an adherent layer of inflammatory cells and debris overlying sites of mucosal injury, a so-called pseudomembrane. It is usually caused by toxins of Clostridium difficile, a normal gut commensal. The two major toxins produced by C. difficile are toxin A and toxin B, which modulate cellular signaling pathways, induce cytokine production, and cause host cell apoptosis.[55] The disease occurs most often in patients without a background of chronic enteric disease, following a course of broad-spectrum antibiotic therapy. Nearly all antibacterial agents have been implicated. Presumably toxin-forming strains flourish
following alteration of the normal intestinal flora; factors favoring the initiation of toxin production are not understood. Rarely, the condition may appear in the absence of antibiotic therapy, typically after surgery or superimposed on a chronic debilitating illness. Infrequently, the small intestine is involved.
Antibiotic-associated colitis occurs primarily in adults as an acute or chronic diarrheal illness, although it has been recorded as a spontaneous infection in young adults without predisposing influences. Diagnosis is confirmed by the detection of the C. difficile cytotoxin in stool. Response to treatment is usually prompt, but relapse occurs in up to 25% of patients.
Morphology.
Given the variety of bacterial pathogens, the pathologic manifestations of enteric bacterial disease are quite variable. Dramatic, even lethal, diarrhea may occur without a significant pathologic lesion, as in cholera resulting from V. cholerae. Alternatively, characteristic histology may enable diagnosis with reasonable certainty, as with C. difficile-induced pseudomembranous colitis. Most bacterial infections exhibit a non-specific pattern of damage to the surface epithelium, decreased epithelial cell maturation and an increased mitotic rate ("regenerative change"), hyperemia and edema of the lamina propria, and variable neutrophilic infiltration into the lamina propria and epithelial layer. In the small intestine, modest villus blunting may occur; in the colon, mucosal architecture is usually preserved. With recovery, epithelial damage and neutrophilic inflammation subside, leaving the residua of regenerative change and lymphoplasmacytic infiltration of the lamina propria. Alternatively, progressive destruction of the mucosa leads to erosion, ulceration, and severe submucosal inflammation. Notable features of particular infections are summarized below:
• Shigella primarily affects the distal colon, first with hyperemia and edema and enlargement of mucosal lymphoid nodules, creating small, projecting nodules. Within 24 hours, the acute mucosal inflammation and erosion generate a patchy and then confluent purulent exudate ( Fig. 17-33 ). The mucosa then becomes soft and friable, and irregular ulcerations appear; severe infection generates large denuded tracts of mucosa. The recovery phase is marked by formation of mucosal granulation tissue and eventual regeneration of the mucosal epithelium. • Salmonella (multiple species, including S. typhimurium and S. paratyphi) primarily affects the ileum and colon, generating blunted villi, vascular congestion, and mononuclear inflammation. Peyer patch involvement produces swelling, congestion, and eventual ulceration with linear ulcers. With S. typhi, bacteremia and systemic dissemination
837
cause proliferation of phagocytes with enlargement of reticuloendothelial and lymphoid tissues throughout the body. Peyer patches in the terminal ileum become sharply delineated, plateau-like elevations up to 8 cm in diameter, with
enlargement of draining mesenteric lymph nodes. Shedding of the mucosa and swollen lymphoid tissue creates oval ulcers with their long axes along the axis of the ileum. Microscopic examination reveals macrophages containing bacteria, red blood cells, and nuclear debris. Intermingled with the phagocytes are lymphocytes and plasma cells, whereas neutrophils are present near the ulcerated surface. The spleen is enlarged, soft, and bulging, with uniformly pale red pulp, obliterated follicular markings, and prominent sinus histiocytosis and reticuloendothelial proliferation. The liver shows small, randomly scattered foci of parenchymal necrosis in which the hepatocytes are replaced by a phagocytic mononuclear aggregate, called a typhoid nodule. These distinctive nodules also occur in the bone marrow and lymph nodes. Gallbladder colonization, which may be associated with gallstones, causes a chronic carrier state. • Campylobacter jejuni and other species may involve the entire intestine from the jejunum to the anus. The small intestine exhibits a decrease in the villus-to-crypt ratio. In invasive colonic infection, the colonic mucosa appears friable and superficially eroded on proctoscopy. Histology reveals multiple superficial ulcers, mucosal inflammation, and a purulent exudate. The formation of colonic crypt abscesses and mucosal ulceration may be confused with those of ulcerative colitis (discussed later). • Yersinia enterocolitica and Y. pseudotuberculosis involve ileum, appendix, and colon. They cause mucosal hemorrhage and ulceration, bowel wall thickening, Peyer patch and mesenteric lymph node hypertrophy with necrotizing granulomas, and systemic spread with peritonitis, pharyngitis, and pericarditis. • Vibrio cholerae affects the small intestine, especially the more proximal segment. The mucosa essentially remains intact, with mucus-depleted crypts. • Clostridium perfringens and Clostridium difficile. C. perfringens infection is usually similar to V. cholerae, but with some epithelial damage; some strains produce severe necrotizing enterocolitis (NEC) with perforation ("pigbel"). C. difficile-induced pseudomembranous colitis derives its name from the plaquelike adhesion of fibrinopurulent-necrotic debris and mucus to damaged colonic mucosa ( Fig. 17-34 A )—these are not true "membranes" since the coagulum is not an epithelial layer. Pseudomembrane formation is not restricted to C. difficile-induced colitis: It also may occur following any severe mucosal injury, as in ischemic colitis, volvulus, and with necrotizing infections (staphylococci, shigella, candida, NEC). What is striking about C. difficile toxin-induced colitis is the microscopic lesion ( Fig. 17-34 B ). The surface epithelium is denuded, and the superficial lamina propria contains a dense infiltrate of neutrophils and occasional capillary fibrin thrombi. Superficially damaged crypts are distended by a mucopurulent
838
exudate, which erupts out of the crypt to form a mushrooming cloud that adheres to the damaged surface—the coalescence of this "cloud" forms the pseudomembrane.
• Enteropathogenic E. coli: At least four distinct types of pathogenic E. coli are known to cause significant diseases—the enterotoxigenic (ETEC), enterohemorrhagic (EHEC), enteroinvasive (EIEC), and enteroadherent (mainly, enteropathogenic, EPEC). In North America, the most important one is EHEC serotype E. coli O 157:H7.[56] EHEC are intestinal commensals in many animals. Humans are usually infected by contaminated meat. These bacteria produce Shiga-like toxins, which damage enterocytes and vascular endothelial cells. In addition to abdominal pain and diarrhea, some patients, particularly children, may develop life-threatening hemolytic-uremic syndrome characterized by the clinical triad of hemolytic anemia, renal failure, and thrombocytopenia.
Figure 17-33 Shigella enterocolitis. Segment of colon showing pale, granular, inflamed mucosa with patches of coagulated exudate.
Figure 17-34 Pseudomembranous colitis from C. difficile infection. A, Gross photograph showing plaques of yellow fibrin and inflammatory debris adherent to a reddened colonic mucosa. B, Low-power micrograph showing superficial erosion of the mucosa and an adherent pseudomembrane of fibrin, mucus, and inflammatory debris.
The complications of severe bacterial enterocolitis are the expected consequences of massive fluid loss or destruction of the intestinal mucosal barrier and include dehydration, sepsis, and perforation. Without quick intervention, death ensues rapidly,
particularly in the very young. Alternatively, an infection may produce extreme discomfort without being life threatening. All enteroinvasive organisms can mimic acute onset of idiopathic inflammatory bowel disease.
Yersinia and Mycobacterium tuberculosis may also present as subacute diarrheal illnesses. Tuberculosis is covered in detail in Chapter 8 .
Bacterial Overgrowth Syndrome
A key mechanism for clearing bacteria from the small intestine is the normal motility, which ensures that bacteria entering into the small intestine are propelled downstream before they can adhere to the mucosa and proliferate. Gastric hypoacidity, immunologic deficiencies, and intestinal dysmotility with intestinal stasis may enable bacteria to proliferate within the small bowel, so-called bacterial overgrowth syndrome. Surgical procedures in particular may set the stage for bacterial overgrowth. These include Billroth procedures, in which the gastric antrum is resected, thereby decreasing the time for exposure of ingested bacteria to gastric acid. Surgical creation of Roux-en-Y loops, as in the Billroth II procedure or in a pancreatoduodenectomy (Whipple) procedure, creates a blind intestinal loop that is a site for bacterial overgrowth. The bacterial populations are mixed enteric populations, without specific dominant species.
Patients usually present with chronic diarrhea, abdominal pain, malabsorption, and weight loss. The clinical diagnosis largely depends on the clinical history and demonstration of the presence of bacteria in the proximal segment of the small intestine by direct culture of an aspirate. Breath tests for volatile bacterial byproducts may be a noninvasive option for a presumptive diagnosis of bacterial overgrowth syndrome.
Parasitic Enterocolitis
Although viruses and bacteria are the predominant enteric pathogens in the United States, parasitic disease and protozoal infection collectively affect over one half of the world's population on a chronic or recurrent basis. The small intestine can harbor as many as 20 species of parasites, including nematodes (the roundworms Ascaris and Strongyloides, hookworms, pinworms), cestodes (flatworms, tapeworms), trematodes (flukes), and protozoa. Some of the parasitic infections are covered in Chapter 8 . Here we will briefly discuss the common parasitic infections of the intestinal tract.
Nematodes
Ascaris lumbricoides is the most common nematode, infecting over a billion individuals worldwide. Infection occurs by ingestion of eggs as a result of human fecal-oral contamination. The ingested ova hatch in the intestine, and larvae penetrate the intestinal mucosa. The disease associated with this parasitic infection is related to larval migration from the splanchnic circulation to the systemic circulation (jejunum-to-liver-to-lung), with formation of hepatic abscess or Ascaris pneumonitis. Larvae migrate up the trachea, are swallowed, and arrive again in the intestine to mature into adult worms. Adult worm
masses can physically obstruct the intestine or the biliary tree. Diagnosis is usually made by detection of the eggs in the feces.
Strongyloides larvae live in fecally contaminated ground soil and penetrate through unbroken skin. They migrate through the lungs, generating pulmonary infiltrates with eosinophilia, and arrive in the intestine to mature into adult worms. Unlike other intestinal worms, which require an ova or larval stage outside the human, the eggs of Strongyloides can hatch within the intestine and larvae can penetrate the mucosa, causing autoinfection. Hence, Strongyloides infection can persist in one individual for life; immunosuppressed individuals can have overwhelming autoinfection. Strongyloides incites a strong tissue eosinophilic reaction, causing eosinophilia as well.
Hookworm (Necator duodenale and Ancylostoma duodenale) infection affects an estimated 1 billion people worldwide and causes significant morbidity. The infection initiates from larva penetration through the skin. The larva develops further in the lungs and gains access to the duodenum by upward migration in the bronchial tree, followed by swallowing. The worms attach to the mucosa, suck blood, and reproduce. The small intestinal mucosa usually exhibits multiple superficial erosions, focal hemorrhage, and inflammatory infiltrates. Long-term infection causes iron deficiency anemia. Diagnosis can be made by detection of the eggs in fecal smear.
Enterobius vermicularis (pinworms) do not invade host tissue and live their entire life within the intestinal lumen. Pinworm infections occur in industrialized countries as well as developing countries; in the United States more than 60 million people have pinworms. Because these are noninvasive worms, they rarely cause serious illness. Enterobius vermicularis infection (enterobiasis) occurs in situations where fecal-oral contamination is common. Adult worms living in the intestine migrate to the anal orifice at night, where the female deposits eggs on the perirectal mucosa. As the eggs are quite
839
irritative, rectal and perineal pruritus ensue. Human-to-human contact is aided by digital manipulation of the area. Both eggs and adult pinworms remain viable external to the body, and reinfection is common. Diagnosis is easily made by applying cellophane tape to the perianal skin and examining the tape for eggs under the microscope.
Trichuris trichiura (whipworm) is less common, occurring primarily in young children. Similar to enterobius vermicularis, these worms do not penetrate the intestinal mucosa and rarely cause serious disease. Heavy infections, however, may cause bloody diarrhea and rectal prolapse. Diagnosis is established by finding the characteristic eggs in the stool.
Cestodes
The intestinal cestodes reside only within the intestinal lumen and never invade beyond the intestinal mucosa. Three species of cestodes (multisegmental flatworms, tapeworms) are Diphyllobothrium latum (fish tapeworm), Taenia solium (pork tapeworm), and Hymenolepsis nana (dwarf tapeworm). Infection occurs by ingestion of raw or undercooked meat that contains encysted larvae. Release of the larvae enables attachment to the intestinal mucosa through its head, or scolex. The worm derives its nutrients from the food stream and enlarges by formation of egg-filled proglottids (segments). Humans are generally infected by one worm only; since the worm does not penetrate the intestinal mucosa, eosinophilia does not generally occur. Nevertheless, the parasite burden can be staggering, as adult worms can grow to many meters in length. Shedding of proglottids or individual eggs produces copious fecal release of eggs. Diagnosis is established by examination of stool for the ova.
Amebiasis
Entamoeba histolytica (ameba) is a dysentery-causing protozoan parasite spread by fecal-oral transmission. This protozoan infects approximately 500 million persons in developing countries such as India, Mexico, and Colombia, resulting in approximately 40 million cases of dysentery and liver abscess.
Pathogenesis.
E. histolytica cysts, which have a chitin wall and four nuclei, are the infectious form because they are resistant to gastric acid. Ingested quadrinucleate cysts colonize the surface of colonic mucin epithelial cells. Cysts release trophozoites, the ameboid forms, which reproduce under anaerobic conditions without harming the host. Because the parasites lack mitochondria or Krebs cycle enzymes, amebae are obligate fermenters of glucose to ethanol. Metronidazole, the best drug to treat invasive infections with entamoebae (as well as other parasites such as Giardia and trichomonads), targets ferridoxin-dependent pyruvate oxidoreductase, an enzyme critical in such fermentation that is present in these organisms but is absent in humans.
Amebae cause dysentery when they attach to the colonic epithelium, as they cause epithelial cell apoptosis, invade the crypts of colonic glands, and burrow into the lamina propria. The organisms then burrow laterally to create, with the accompanying inflammation and tissue necrosis, a flask-shaped ulcer with a narrow neck and broad base. Amebic proteins that may be involved in tissue invasion include: (1) cysteine proteinases, which are able to break down proteins of the extracellular matrix; (2) a lectin on the parasite surface that binds to carbohydrates on the surface of colonic epithelial cells and red blood cells; and (3) a channel-forming protein called the amebapore, which makes holes in the plasma membrane of host cells and lyses them. The presence in stool of trophozoites containing ingested red blood cells is indicative of tissue invasion by virulent organisms.
Morphology.
Amebiasis most frequently involves the cecum and ascending colon, followed in order by the sigmoid, rectum, and appendix. In severe fullblown cases, however, the entire colon is involved. Amebae can mimic the appearance of macrophages because of their comparable size and large number of vacuoles; the parasites, however, have a smaller nucleus, which contains a large karyosome ( Fig. 17-35 ). Amebae invade through the crypt epithelium and burrow into the mucosa and submucosa, eliciting a neutrophilic reaction. They are stopped by the muscularis propria and fan out laterally to create a flask-shaped ulcer with a narrow neck and broad base. These maturing ulcers contain few host inflammatory cells and exhibit extensive liquefactive necrosis. As the lesion progresses, the overlying surface mucosa is deprived of its blood supply and sloughs. The mucosa between ulcers is often normal or mildly inflamed. On occasion, the formation of profuse circumferential granulation tissue can create colonic stricture.
In about 40% of patients with amebic dysentery, parasites penetrate splanchnic vessels and embolize to the liver to produce solitary, or less often multiple, discrete abscesses, sometimes exceeding 10 cm in diameter. Amebic liver abscesses have a scant inflammatory reaction at their margins and a shaggy fibrin lining. Because of hemorrhage into the cavities, the abscesses are sometimes filled with a chocolatecolored, odorless, pasty material. Secondary bacterial infection may make these abscesses purulent.
Figure 17-35 Entamoeba histolytica in colon. High-power view of the organisms. Note some of the organisms ingesting red blood cells.
840
Clinical Features.
Clinically, the patient may present with abdominal pain, bloody diarrhea, or weight loss. Occasionally, acute necrotizing colitis and megacolon can occur, which carry significant
mortality. Rarely, amebic abscesses reach the lung and the heart by direct extension from the liver or spread from the liver through the blood into the kidneys and brain. Such abscesses remain long after the acute intestinal illness has passed.
Giardiasis
Giardia lamblia is the most common pathogenic parasitic infection in humans.[57] It is an intestinal protozoan spread by fecally contaminated water or food. Infection may be subclinical or may cause acute or chronic diarrhea, steatorrhea, or constipation. Because Giardia cysts are not killed by chlorine, Giardia is endemic in public water supplies that are not filtered through sand and in contaminated streams accessed by campers.
Pathogenesis.
In the United States, Giardia infections are especially frequent in institutions for the mentally ill and in day-care centers. Giardia, like Entamoeba, ferments glucose, lacks mitochondria, and exists in two forms: (1) a dormant but infectious cyst spread by the fecal-oral route from person to person (as well as from beavers to persons); and (2) trophozoites that multiply in the intestinal lumen. Transition from trophozoites to cysts is induced by decreases in availability of cholesterol as Giardia moves from duodenum to jejunum. In contrast to Entamoeba, Giardia trophozoites have two nuclei rather than one, are flagellated, reside in the duodenum rather than the colon, adhere to but do not invade the intestinal epithelial cells, and so cause diarrhea rather than dysentery.
Infection can occur by ingestion of as few as 10 cysts. Giardia trophozoites adhere to sugars on intestinal epithelial cells through a parasite lectin that is activated when it is cleaved by proteases, which are plentiful in the lumen of the duodenum. Tight contact between the parasite and the epithelial cell is made by a sucker-like disc, composed of cytoplasmic tubulin and unique intermediate filaments called giardins. Although Giardia does not secrete toxin, it contains a cystein-rich surface protein that resembles diarrhea-causing toxins secreted by certain snakes. The physical presence of rapidly proliferating trophozoites and their toxic proteins damages the microvillus brush border, causing a malabsorptive state.[52]
Immunity mediated by antibodies, including secretory IgA is important in resistance to Giardia, because agammaglobulinemic individuals are severely affected by the parasite. Immunity to Giardia, however, is limited by the parasite's ability to vary its major surface proteins into antigenically distinct forms, encoded by more than 50 different genes.
Morphology.
In stool smears, G. Iamblia trophozoites are pear shaped and binucleate. Duodenal biopsy specimens are often teeming with sickle-shaped trophozoites, which are tightly bound by the concave attachment disc to the villus surface of the intestinal epithelial cells ( Fig. 17-36 ). As Giardia does not actually invade the mucosa, small intestinal morphology
Figure 17-36 Giardia lamblia. Trophozoite (arrow) of the organism immediately adjacent to the duodenal surface epithelium.
may be virtually normal. However, many patients exhibit marked blunting of the small intestinal villi with a mixed inflammatory infiltrate in the lamina propria. The brush borders of the surface absorptive epithelial cells are irregular, and sometimes there is virtual absence of villi, resembling the atrophic stage of celiac disease.
Clinical Features.
Infected patients exhibit a malabsorptive diarrhea, owing to mucosal epithelial cell injury. Functional lactase deficiency also occurs in 20% to 40% of chronically infected patients by mechanisms that are not understood. The infection can last for months or years. The symptoms may be severe in immunocompromised patients. Diagnosis is readily made by examination of stool for cysts; small intestinal biopsy or examination of a small intestinal aspirate also permits identification of the organisms. While giardiasis is responsive to oral antimicrobial therapy, recurrence is common following cessation of treatment.
Necrotizing Enterocolitis
Necrotizing enterocolitis (NEC) is an acute, necrotizing inflammation of the small and large intestines with the severe consequence of transmural necrosis of intestinal segments. While it can occur at any age, NEC is particularly devastating in the neonate.[58] It is the most common acquired gastrointestinal emergency of neonates, particularly those who are premature or of low birth weight. It may occur at any time in the first 3 months of life, but its peak incidence is around the time when infants are started on oral foods (2 to 4 days old). This condition is described in Chapter 10 .
Collagenous and Lymphocytic Colitis
Collagenous colitis is a distinctive disorder of the colon characterized by chronic watery diarrhea and patches of bandlike collagen deposits directly under the surface epithelium.
Lymphocytic colitis is characterized by chronic watery diarrhea and a prominent intraepithelial infiltrate of lymphocytes. Collagenous
841
colitis occurs primarily in middle-aged and older women; lymphocytic colitis affects males and females equally. Both sets of patients can have between 3 to 20 nonbloody, watery bowel movements per day, accompanied by cramping abdominal pain. Radiographic studies are unremarkable, and endoscopy characteristically reveals normal mucosa. The pathogenesis of both these conditions remains unclear; they do appear to be separate conditions.[59] Lymphocytic colitis shows a strong association with autoimmune diseases, including celiac sprue, thyroiditis, arthritis, and autoimmune gastritis. Both diseases are benign in nature, with neither debilitating weight loss nor malignancy as potential outcomes. While collagenous colitis and lymphocytic colitis are uncommon, they must be considered in every adult patient who presents with a noninflammatory, watery diarrhea.
MISCELLANEOUS INTESTINAL INFLAMMATORY DISORDERS
Acquired Immunodeficiency Syndrome (AIDS)
Diarrheal illness occurs in 50% of AIDS patients in North America and may approach 100% in developing countries.[60] Some patients exhibit a malabsorptive syndrome with small intestinal villus atrophy or a colitic syndrome resembling ulcerative colitis in the absence of demonstrable pathogens. Although most cases are probably due to coexistent infections in these immunocompromised patients, the concept of AIDS enteropathy, attributable to direct mucosal damage by human immunodeficiency virus infection, has been proposed but remains unproven. The spectrum of infections occurring in AIDS is discussed in Chapter 6 .
Transplantation
Diarrhea is a significant complication of bone marrow transplantation. Pretransplant conditioning may cause direct toxic injury to the small intestinal mucosa, evident as villus blunting, degeneration and flattening of crypt epithelial cells, decreased mitoses, and atypia of cell nuclei. Abrupt onset of severe watery diarrhea is a major feature of acute GVHD. A distinctive histologic lesion is focal crypt cell necrosis, in which debris from necrotic cells occupies lacunae within the epithelial layer, with minimal to absent inflammatory cell response in the lamina propria ( Fig. 17-37 ).[61] In more advanced GVHD, necrosis may become so severe as to lead to total sloughing of the mucosa. In addition to fluid and electrolyte derangements, the life-threatening complications of sepsis and intestinal hemorrhage may ensue. Alimentary tract symptoms are less evident in chronic GVHD but may include dysphagia secondary to esophageal involvement and occasionally malabsorption due to chronic intestinal injury. Small intestinal transplantation also carries with it a spectrum of complications, including rejection and CMV infection.
When one considers the vast quantities of drugs ingested by humans, with or without the blessings of the medical profession, it is remarkable that the gastrointestinal tract escapes relatively unscathed. Focal ulceration can occur when a pill adheres to the mucosa and releases all of its contents locally, as may occur in the esophagus with "dry swallows." Drug-induced
Figure 17-37 Graft-versus-host disease of the colon. Note the apoptotic cell in the crypt (arrow).
acute erosive gastritis has been mentioned earlier. The small intestine and colon are susceptible to drug-induced enterocolitis, most commonly associated with the use of NSAIDs.[62] A nonspecific pattern of small intestinal inflammation may lead to malabsorption. Colonic inflammation may produce an acute or chronic diarrheal illness; ulceration and stricture formation also occur. Drug-induced gastrointestinal injury must be considered whenever abdominal illness is encountered. Considering this possibility may, on occasion, spare patients from erroneous diagnosis of a chronic intestinal illness.
Radiation Enterocolitis
Abdominal irradiation may severely impair the normal proliferative activity of the small intestinal and colonic mucosal epithelia. Acute radiation enteritis manifests as anorexia, abdominal cramps, and a malabsorptive diarrhea, attributable to acute mucosal injury. Chronic radiation enteritis or colitis may exhibit more indolent symptoms than the acute form or may present as an inflammatory colitis. The mucosal damage may be perpetuated by radiation-induced vascular injury and may be accompanied by ischemic fibrosis and stricture.
Typhlitis was a nineteenth-century term for severe acute and chronic inflammation of the cecal and appendiceal region, which in retrospect was probably the outcome of untreated acute appendicitis. The term is now used to describe a life-threatening acute inflammatory destruction of the mucosa of the cecal region occurring in neutropenic individuals. The presumed pathogenesis is impaired mucosal immunity in combination with compromised blood flow in the cecal region.
Diversion Colitis
Diversion colitis is an inflammatory mucosal lesion occurring in segments of the colon that have been surgically isolated. Colonic enterocytes derive a significant portion of their
842
caloric supply from short-chain fatty acids present in the luminal stream. Surgical diversion of the stream, as through an ileostomy, renders the colonic mucosa susceptible to nutritional deprivation. The changes may range from very mild with increased lamina propria lymphocytes, to a severe exudative diarrheal disease that resembles ulcerative colitis. Restoration of fecal flow through the colon, or enemas containing short-chain fatty acids, permits mucosal recovery.
Solitary Rectal Ulcer Syndrome
Solitary rectal ulcer syndrome is an inflammatory condition of the rectum resulting from motor dysfunction of the anorectal musculature. Dysregulation of the anorectal sphincter, in particular impaired relaxation of the anorectal sling, may create sharp angulation of the anterior rectal shelf. Abrasion of the overlying rectal mucosa creates an oval ulcer and surrounding mucosal inflammation, frequently with the formation of an inflammatory polyp. Associated partial prolapse of the rectal mucosa is common. Patients experience a characteristic triad: rectal bleeding, mucus discharge from the anus, and superficial ulceration of the anterior rectal wall.
Malabsorption Syndromes
Malabsorption is characterized by defective absorption of fats, fat-soluble and other vitamins, proteins, carbohydrates, electrolytes and minerals, and water. The most common clinical presentation is chronic diarrhea, and the hallmark of malabsorption is steatorrhea (excessive fecal fat content). At the most basic level, malabsorption is the result of disturbance of at least one of these normal digestive functions:
1. Intraluminal digestion, in which proteins, carbohydrates and fats are broken down into assimilable forms. The process begins in the mouth with saliva, receives a major boost from gastric peptic digestion, and continues in the small intestine, assisted by the emulsive action of bile salts (see Chapter 18 ).
2. Terminal digestion, which involves the hydrolysis of carbohydrates and peptides by disaccharidases and peptidases, respectively, in the brush border of the small intestinal mucosa.
3. Transepithelial transport, in which nutrients, fluid, and electrolytes are transported across the epithelium of the small intestine for delivery to the intestinal vasculature. Absorbed fatty acids are converted to triglycerides and, with cholesterol, are assembled into chylomicrons for delivery to the intestinal lymphatic system.
The major diseases and disorders causing malabsorption are listed in Table 17-9 . This classification is most helpful for diseases in which there is a single, clear-cut abnormality. In many malabsorptive disorders a defect in one pathophysiologic process predominates, but others may contribute or may be secondary outcomes of the primary cause. Although many causes of malabsorption can be established clinically, small intestinal mucosal biopsy may be required to satisfactorily identify or exclude celiac disease.
Clinically, the malabsorption syndromes resemble each other more than they differ. The consequences of malabsorption affect many organ systems:
TABLE 17-9 -- Major Malabsorption Syndromes
Defective Intraluminal Digestion
Digestion of fats and proteins
• Pancreatic insufficiency, owing to pancreatitis or cystic fibrosis
• Zollinger-Ellison syndrome, with inactivation of pancreatic enzymes by excess gastric acid secretion
Solubilization of fat, owing to defective bile secretion
• Ileal dysfunction or resection, with decreased bile salt uptake
• Cessation of bile flow from obstruction, hepatic dysfunction
Nutrient preabsorption or modification by bacterial overgrowth
Primary Mucosal Cell Abnormalities
Defective terminal digestion
• Disaccharidase deficiency (lactose intolerance)
• Bacterial overgrowth, with brush border damage
Defective epithelial transport
• Abetalipoproteinemia
• Primary bile acid malabsorption owing to mutations in the ileal bile acid transporter
Short-gut syndrome, following extensive surgical resection
Distal ileal resection or bypass
• Alimentary tract: diarrhea, both from nutrient malabsorption and excessive intestinal secretions, flatus, abdominal pain, weight loss, and mucositis resulting from vitamin deficiencies • Hematopoietic system: anemia from iron, pyridoxine, folate, and/or vitamin B12 deficiency and bleeding from vitamin K deficiency • Musculoskeletal system: osteopenia and tetany from calcium, magnesium, and vitamin D deficiency • Endocrine system: amenorrhea, impotence, and infertility from generalized malnutrition; hyperparathyroidism from protracted calcium and vitamin D deficiency • Epidermis: purpura and petechiae from vitamin K deficiency, edema from protein deficiency, dermatitis and hyperkeratosis from deficiencies of vitamin A, zinc, essential fatty acids and niacin • Nervous system: peripheral neuropathy from vitamin A and B12 deficiencies.
The passage of abnormally bulky, frothy, greasy, yellow, or gray stools (steatorrhea) is a prominent feature of malabsorption, accompanied by weight loss, anorexia, abdominal distention, borborygmi, and muscle wasting. The malabsorptive disorders most commonly encountered in the United States are celiac disease, pancreatic insufficiency, and Crohn disease.
843
Pancreatic insufficiency, primarily from chronic pancreatitis or cystic fibrosis, is a major cause of defective intraluminal digestion. Typical features of defective intraluminal digestion are an osmotic diarrhea from undigested nutrients and steatorrhea. Excessive growth of normal bacteria within the proximal small intestine (bacterial overgrowth, discussed earlier) also impairs intraluminal digestion and can damage mucosal epithelial cells, causing impaired terminal digestion and epithelial absorption.
CELIAC DISEASE
Celiac disease (also referred to as celiac sprue, gluten-sensitive enteropathy) is a chronic disease, in which there is a characteristic mucosal lesion of the small intestine and impaired nutrient absorption, which improves on withdrawal of wheat gliadins and related grain proteins from the diet. [63] Celiac disease occurs largely in Caucasians and is rare or nonexistent among native Africans, Japanese, and Chinese. Its prevalence in the United States is somewhat difficult to define; in Europe the prevalence is in the range of 1:100 to 1:200.[64] The disease was first described more than a century ago, but its connection to gluten was not known until the 1940s, changing its clinical management.
Pathogenesis.
The fundamental disorder in celiac disease is a sensitivity to gluten, which is the alcohol-soluble, water-insoluble protein component (gliadin) of wheat and closely related grains (oat, barley, and rye). The hallmark of this disease is a T-cell mediated chronic inflammatory reaction with an autoimmune component, which most likely develops as a consequence of a loss of tolerance to gluten. Interplay between genetic predisposing factors, the host immune response, and environmental factors is central to disease pathogenesis. The small intestinal mucosa, when exposed to gluten, accumulates intraepithelial CD8+ T cells and large numbers of lamina propria CD4+ T cells, which are sensitized to gliadin. The recognized epitopes are confined to residues 57–75 of gliadin. [65]
It has been long known that family history is important in celiac disease. Almost all individuals with celiac disease share the major histocompatibility complex class II HLA-DQ2 or
Figure 17-38 Celiac disease (gluten-sensitive enteropathy). A, A peroral jejunal biopsy specimen of diseased mucosa shows diffuse severe atrophy and blunting of villi, with a chronic inflammatory infiltrate of the lamina propria. B, A normal mucosal biopsy.
HLA-DQ8 haplotype. It has been proposed that gliadin is deamidated by the enzyme transglutaminase and that deamidated gliadin peptides bind to DQ2 and DQ8. Recognition of these peptides by CD4+ T cells leads to secretion of interferon γ, which damages the intestinal wall. Although this is an attractive hypothesis, its key elements remain to be proven. It is also unclear how CD8+ T cells accumulate in the epithelium. They do not recognize gliadin, but seem to respond to stress-induced molecules on epithelial cells. The epithelial cells secrete large amounts of IL-15 that activates CD8+ T cells and increases the risk of lymphoma development.
Morphology.
By endoscopy, the small intestinal mucosa appears flat or scalloped, or may be visually normal. Biopsies demonstrate diffuse enteritis, with marked atrophy or total loss of villi. The surface epithelium shows vacuolar degeneration, loss of the microvillus brush border, and an increased number of intraepithelial lymphocytes ( Fig. 17-38 ). The crypts, on the other hand, exhibit increased mitotic activity and are elongated, hyperplastic, and tortuous, so that the overall mucosal thickness remains the same. The lamina propria has an overall increase in plasma cells, lymphocytes, macrophages, eosinophils, and mast cells. All these changes are usually more marked in the proximal small intestine than in the distal, since it is the duodenum and proximal jejunum that are exposed to the highest concentration of dietary gluten. Although these changes are characteristic of celiac disease, they can be mimicked by other diseases, most notably tropical sprue. Mucosal histology usually reverts to normal or near-normal following a period of gluten exclusion from the diet.
Clinical Features.
The symptoms of celiac disease vary tremendously from patient to patient. Symptomatic diarrhea and failure to thrive may be evident during infancy, yet adults may seek attention only in their fifth decade of life. The classic presentation includes diarrhea,
flatulence, weight loss, and fatigue. However, extraintestinal manifestations of malabsorption may overshadow the intestinal symptoms. A characteristic
844
skin blistering lesion, dermatitis herpetiformis, can occur in patients with celiac disease. Neurologic disorders are occasionally seen. Detection of circulating anti-gliadin or "anti-endomysial" antibodies strongly favors the diagnosis; antibodies against tissue transglutaminase also may be detected, as this is the autoantigen recognized by anti-endomysial antibody. Definitive diagnosis rests on (1) clinical documentation of malabsorption; (2) demonstration of the intestinal lesion by small bowel biopsy; and (3) unequivocal improvement in both symptoms and mucosal histology on gluten withdrawal from the diet. If there is doubt about the diagnosis, gluten challenge followed by rebiopsy has been advocated. Serologic tests used for screening or treatment follow-up include the detection of antibodies against tissue transglutaminase and gliadin.
Most patients with celiac disease who adhere to a gluten-free diet remain well indefinitely and ultimately die of unrelated causes. However, there is a long-term risk of malignant disease, which includes non-Hodgkin lymphoma (moderate risk), small intestinal adenocarcinoma, and esophageal squamous cell carcinoma (50- to 100-fold higher risk than the general population).
TROPICAL SPRUE (POSTINFECTIOUS SPRUE)
This condition is so named because it is a celiac-like disease that occurs almost exclusively in people living in or visiting the tropics. The distribution of the disease is curious: It is common in the Caribbean (but not in Jamaica), central and southern Africa, the Indian subcontinent and Southeast Asia, and portions of Central and South America. The disease may occur in endemic form, and epidemic outbreaks have occurred. No specific causal agent has been clearly associated with tropical sprue, but bacterial overgrowth by enterotoxigenic organisms (e.g., E. coli and Hemophilus) has been implicated.
Morphology.
Intestinal changes are extremely variable, ranging from near normal to severe diffuse enteritis. Unlike celiac sprue, injury is seen at all levels of the small intestine. Patients frequently have folate and/or vitamin B12 deficiency, leading to markedly atypical enlargement of the nuclei of epithelial cells (megaloblastic change), reminiscent of the changes seen in pernicious anemia.
Malabsorption usually becomes apparent within days or a few weeks of an acute diarrheal enteric infection in visitors to endemic locales and may persist if untreated. The mainstay of treatment is broad-spectrum antibiotics, supporting an infectious etiology. Intestinal lymphoma does not appear to be associated with this disorder.
WHIPPLE DISEASE
Whipple disease is a rare disease caused by the bacterium Tropheryma whippelii. It is a systemic condition that may involve any organ of the body, but principally affects the intestine, central nervous system, and joints. [66] The disease was first described as intestinal lipodystrophy by George Whipple in 1907. The bacterial etiology was discovered in the 1960s on the basis of ultrastructural observations. The pathogenesis of Whipple disease is still not clear. The causal organism T. whippelii is a gram-positive actinomycete, named on the basis of molecular phylogenetic analysis. The bacteria proliferate preferentially within macrophages and invoke no significant host immune reaction.
Morphology.
The hallmark of Whipple disease is a small-intestinal mucosa laden with distended macrophages in the lamina propria. The macrophages contain PAS-positive, diastase-resistant granules (which are lysosomes stuffed with partially digested microorganisms) and rod-shaped bacilli on electron microscopy ( Fig. 17-39 ). The PAS stain is not specific for this bacterium. In untreated cases, bacilli can be seen as well in neutrophils, the extracellular space of the lamina propria, and even in epithelial cells. Expansion of the villi by the dense infiltrate of macrophages imparts a shaggy gross appearance to the intestinal mucosal surface; edema of the mucosa thickens the intestinal wall. Accompanying these changes is involvement of mesenteric lymph nodes by the same process and lymphatic dilation, suggesting lymphatic obstruction. The lymphatic blockade is believed to be responsible for lipid deposition in the villi, thus the original impression of intestinal lipodystrophy. Bacilli-laden macrophages also can be found in the synovial membranes of affected joints, the brain, cardiac valves, and elsewhere. At each of these sites, inflammation is essentially absent. Functional impairment nevertheless can be considerable at each affected site.
Clinical Features.
Whipple disease is principally encountered in Caucasians in the fourth to fifth decades of life, with a strong male predominance of 10:1. Many of the published cases came from rural regions, suggesting an environmental influence. It usually presents as a form of malabsorption with diarrhea and weight loss, sometimes of years' duration. Arthropathy is often the initial presentation. Atypical presentations, with polyarthritis, obscure psychiatric complaints, cardiac abnormalities, and other symptom complexes, are common. Lymphadenopathy and hyperpigmentation are present in over half of patients. Currently, the diagnosis still rests on demonstration of small intestinal PAS-positive macrophages that contain rod-shaped organisms on electron microscopy. The culture of this bacterium in 1997 and the completion of its whole genome sequence in 2003 will most likely give rise to more specific molecular diagnosis of this disease.[67] Response to antibiotic therapy is usually prompt, although some patients have a protracted, refractory course.
The disaccharidases, of which the most important is lactase, are located in the apical cell membrane of the villous absorptive epithelial cells. Congenital lactase deficiency is a very rare condition, but acquired lactase deficiency is common, particularly among Native Americans and African Americans. Incomplete breakdown of the disaccharide lactose into its monosaccharides glucose and galactose leads to osmotic diarrhea from the unabsorbed lactose. Bacterial fermentation of
845
Figure 17-39 Whipple disease. A, Note foamy macrophages in the lamina propria. B, PAS stain showing the positive granules in the foamy macrophages. C, Electron micrograph of a lamina propria macrophage showing many bacilli within the cell (arrow) and in the extracellular space (arrowhead). Inset, Higher magnification of macrophage cytoplasm showing cross-sectional profiles of bacilli and their cell walls
(small arrows). (C, courtesy of George Kasnic and Dr. William Clapp, University of Florida, Gainesville, FL.)
the unabsorbed sugars leads to increased hydrogen production, which is readily measured in exhaled air by gas chromatography.
When caused by an inherited enzyme deficiency, malabsorption becomes evident with the initiation of milk feeding. The infants develop explosive, watery, frothy stools and abdominal distention. Malabsorption is promptly corrected when exposure to milk and milk products is terminated. In the adult, lactase insufficiency apparently develops as an acquired disorder, sometimes in association with viral and bacterial enteric infections or other disorders of the gut. Neither light nor electron microscopy has disclosed abnormalities of the mucosal cells of the bowel in either the hereditary or acquired form of the disease.
846
ABETALIPOPROTEINEMIA
Inability to synthesize apolipoprotein B is a rare inborn error of metabolism transmitted by autosomal recessive inheritance. It is characterized by a defect in the synthesis and export of lipoproteins from intestinal mucosal cells. Free fatty acids and monoglycerides that are produced by hydrolysis of dietary fat enter the absorptive epithelial cells and are re-esterified in the normal fashion but cannot be assembled into chylomicrons. As a consequence, triglycerides are stored within the cells, creating lipid vacuolation that is readily evident under the light microscope, particularly with special fat stains. Concomitantly, there is complete absence in plasma of all lipoproteins containing apolipoprotein B (chylomicrons, very-low-density lipoproteins, and low-density lipoproteins). The failure to absorb certain essential fatty acids leads to lipid membrane defects, readily evident in the characteristic acanthocytic erythrocytes (burr cells). The disease becomes manifest in infancy and is dominated by failure to thrive, diarrhea, and steatorrhea.
Idiopathic Inflammatory Bowel Disease
Idiopathic inflammatory bowel disease is a set of chronic inflammatory conditions resulting from inappropriate and persistent activation of the mucosal immune system, driven by the presence of normal intraluminal flora.[68] The two disorders known as
inflammatory bowel disease (IBD) are Crohn disease (CD) and ulcerative colitis (UC). These diseases share many common features but have distinctly different clinical manifestations. IBD is common in developed countries, with up to 1 in 200 of individuals of Northern European descent affected by these diseases. The annual incidence of IBD in the United States is approximately 3 to 10 new cases per 100,000 people.
Both CD and UC are chronic, relapsing inflammatory disorders of obscure origin. CD is an autoimmune disease that may affect any portion of the gastrointestinal tract from esophagus to anus, but most often involves the distal small intestine and colon. UC is a chronic inflammatory disease limited to the colon and rectum. Both exhibit extraintestinal inflammatory manifestations. Before considering these diseases separately, the pathogenesis of IBD is considered.
ETIOLOGY AND PATHOGENESIS
A remarkable attribute of the normal gastrointestinal tract is that the mucosal immune system is always poised to respond against ingested pathogens but is unresponsive to normal intestinal microflora.[69] In IBD, this state of homeostasis is disrupted, leading to two key pathogenic abnormalities—strong immune responses against normal flora, and defects in epithelial barrier function. The basis of these abnormalities is still not established, which is why both CD and UC are considered idiopathic diseases. However, recently, extensive investigations of animal models, and more limited analyses of lesions from patients, have led to some important conclusions about the pathogenesis of IBD.[70] It is postulated that IBD results from unregulated and exaggerated local immune responses to commensal microbes in the gut, in genetically susceptible individuals. Thus, as in many other autoimmune disorders ( Chapter 6 ), the pathogenesis of IBD involves failure of immune regulation, genetic susceptibility, and environmental triggers, specifically microbial flora. Below we summarize some salient points about each of the factors that contribute to IBD.
Genetic Susceptibility.
Fifteen percent of IBD patients have affected first-degree relatives, and the lifetime risk if either a parent or sibling is affected is 9%. Dizygotic twins have the concordance rates expected for siblings; monozygotic twins exhibit a 30% to 50% concordance rate for CD. These associations clearly indicate that genetic susceptibility plays an important role in the development of IBD. The disease is a complex multigenic trait, and is not inherited in Mendelian fashion. Many candidate genes are known to be associated with, and likely contribute to, the development of IBD. These include HLA associations; an HLA-DR1/DR1/DQw5 allelic combination has been observed in 27% of North American white patients with CD, whereas HLA-DR2 is increased in patients with UC. A gene called NOD2 (so named because the encoded protein has a nucleotide-binding oligomerization domain) has recently been shown to be associated with CD.[71] The NOD2 protein is expressed in many types of leukocytes as well as epithelial cells, and is thought to function as an intracellular receptor for microbes. Upon binding microbial components, it may trigger the NF-κB pathway; recall that NF-κB is a transcription factor that triggers the production of cytokines and other proteins involved in innate immune
defense against infectious pathogens ( Chapter 6 ). The NOD2 mutations that are associated with Crohn disease may reduce the activity of the protein, resulting in the persistence of intracellular microbes and uncontrolled, prolonged immune responses. There is, however, no direct proof in support of this hypothesis. Other gene(s) associated with CD have been localized to chromosome 5q31; although a candidate gene in this locus has not been identified, this region is rich in genes encoding several cytokines that may contribute to IBD. Finally, it is worth mentioning that in inbred mice, the knockout of many different genes, individually, leads to the common pathologic manifestation of chronic inflammation in the intestine. It is, therefore, likely that immune dysregulations arising by multiple mechanisms may contribute to IBD in humans.
Role of Intestinal Flora.
Animal studies have definitively established the importance of gut flora in IBD.[72] If gene-knockout mice that normally develop IBD are made germfree, the disease disappears. However, the hunt for a specific microbe as the underlying cause has been largely fruitless. There is also no clear evidence that reducing intestinal flora has a beneficial effect on the course of IBD in humans. Microbes could exacerbate immune reactions by providing antigens and inducing costimulators and cytokines, all of which contribute to T-cell activation ( Chapter 6 ). Defects in the barrier function of the intestinal epithelium could allow luminal flora to gain access to the mucosal lymphoid tissue, and thus trigger immune responses.
Abnormal T-Cell Responses.
It is believed that the exaggerated local immune response in IBD is a consequence of too much T-cell activation and/or too little control by regulatory T lymphocytes. Both aspects have been clearly illustrated in animal models of the disease, and the lesions in humans show clear evidence of T-cell reactions (see below).
847
Thus, susceptibility genes, intestinal flora, and defective control of immune reactions all seem to play a role in the initiation and progression of IBD. One interesting, and largely unanswered question is, are these diseases caused by "true" autoimmunity, i.e. are the immune responses directed against self-antigens in the intestinal epithelium or only against the antigens of intestinal microbes. Regardless of the specificity of the pathogenic immune response, several features of the response and its role in the disease are known.
• In both CD and UC, the prime culprits appear to be T-cells, particularly CD4+ T-cells, and the lesions are likely caused by T-cells and their products. Although antibodies against certain self-antigens, such as tropomyosin, have been detected in some patients with UC, it is not clear that these autoantibodies play a pathogenic role.
• Crohn disease appears to be the result of a chronic delayed-type hypersensitivity reaction induced by IFN-γ-producing TH 1 cells. The nature of the inflammatory infiltrate, especially the presence of granulomas, is consistent with a TH 1 response. • Although animal models suggest that ulcerative colitis is caused by excessive activation of TH 2 cells, in the human disease the signature TH 2 cytokine, IL-4, has not been found in the lesions. It may be that the lesions are caused by an atypical TH 2 response, or that there is no consistent pattern of T cell activation or dominant cytokine production.
Diagnosis of IBD.
Since the exact etiology of IBD is not known, the diagnosis of IBD and the distinction between CD and UC are dependent on clinical history, radiographic examination, laboratory findings, and pathologic examination of tissue. There is no single test upon which a diagnosis is made. Even with the best efforts, the distinction between the two diseases still cannot be made in some cases. As discussed later, pathologic appearance, both macroscopic and microscopic, plays a central role in establishing a definitive diagnosis. In recent years, considerable effort has been given to developing accurate noninvasive laboratory tests.[73] The pANCA (perinuclear antineutrophilic cytoplasmic antibody) is positive in 75% of patients with UC and in only 11% with CD.[74] Another test detects an antibody against the cell wall mannan polysaccharide of Saccharomyces cerevisiae (ASCA). This antibody appears to be elevated in CD patients. The clinical utility of these tests remains to be proven.
CROHN DISEASE
When first described by Crohn, Ginsburg, and Oppenheimer in 1932, this idiopathic disorder was thought to be limited to the terminal ileum, hence the designation terminal ileitis. Recognition that sharply delineated bowel segments might be affected, with intervening unaffected ("skip") areas, led to the alternative name regional enteritis. Predominant involvement of the colon gave rise to the term granulomatous colitis. It is now clear that any level of the alimentary tract may be involved and that there are systemic manifestations; thus, the eponymic name Crohn disease is preferred. When fully developed, Crohn disease is characterized pathologically by (1) sharply delimited and typically transmural involvement of the bowel by an inflammatory process with mucosal damage, (2) the presence of noncaseating granulomas, and (3) fissuring with formation of fistulae.
Epidemiology.
Crohn disease occurs throughout the world, but primarily in Western developed populations. Its annual incidence in the United States is around 3 per 100,000, with rates between 4 and 10 per 100,000 reported in Great Britain and Scandinavia. The incidence and prevalence of CD has been steadily rising in the United States and Northern Europe. It occurs at any age, from young childhood to advanced age, but peak ages of detection are the second and third decades of life with a minor peak in the sixth and seventh decades. Females are affected slightly more often than males. Whites appear to develop
the disease two to five times more often than do nonwhites. In the United States, CD occurs three to five times more often among Jews than among non-Jews. Smoking is a strong exogenous risk factor.
Morphology.
In CD, there is gross involvement of the small intestine alone in about 40% of cases, of small intestine and colon in 30%, and of the colon alone in about 30%. CD may involve the duodenum, stomach, esophagus, and even mouth, but these sites are distinctly uncommon. In diseased bowel segments, the serosa is granular and dull gray, and often the mesenteric fat wraps around the bowel surface (creeping fat). The mesentery of the involved segment is also thickened, edematous, and sometimes fibrotic. The intestinal wall is rubbery and thick, as a consequence of edema, inflammation, fibrosis, and hypertrophy of the muscularis propria. As a result, the lumen is almost always narrowed; in the small intestine this is evidenced on x-ray as the "string sign," a thin stream of barium passing through the diseased segment. Strictures may occur in the colon but are usually less severe. A classic feature of CD is the sharp demarcation of diseased bowel segments from adjacent uninvolved bowel. When multiple bowel segments are involved, the intervening bowel is essentially normal ("skip" lesions).
A characteristic sign of early disease is focal mucosal ulcers resembling canker sores (aphthous ulcers), edema, and loss of the normal mucosal texture. With progressive disease, mucosal ulcers coalesce into long, serpentine linear ulcers, which tend to be oriented along the axis of the bowel ( Fig. 17-40 ). As the intervening mucosa tends to be relatively spared, the mucosa acquires a coarsely textured,
Figure 17-40 Crohn disease of ileum, showing narrowing of the lumen, bowel wall thickening, serosal extension of mesenteric fat ("creeping fat"), and linear ulceration of the mucosal surface (arrowheads).
848
cobblestone appearance. Narrow fissures develop between the folds of the mucosa, often penetrating deeply through the bowel wall ( Fig. 17-41 ) and leading to bowel
adhesions and serositis. Further extension of fissures leads to fistula or sinus tract formation, either to an adherent viscus, to the outside skin, or into a blind cavity. Free perforation or localized abscesses may also develop.
The characteristic histologic features of CD are:
• Mucosal inflammation. The earliest lesion in CD appears to be focal neutrophilic infiltration into the epithelial layer, particularly overlying mucosal lymphoid aggregates. As the disease becomes more established, neutrophils infiltrate isolated crypts; when a sufficient number of neutrophils have traversed the epithelium of a crypt (both in the small and large intestines), a crypt abscess is formed, usually with ultimate destruction of the crypt. • Chronic mucosal damage. The hallmark of inflammatory bowel disease, both CD and UC, is chronic mucosal damage. Architectural distortion is manifested in the small intestine as variable villus blunting; in the colon, crypts exhibit irregularity and branching. The degree of the glandular architectural distortion in CD is usually less severe than in UC. Crypt destruction leads to progressive atrophy, particularly in the colon. The mucosa may undergo metaplasia: This may take the form of gastric antraltype glands (pyloric metaplasia) or the development of Paneth cells in the distal colon, where they are normally absent (Paneth cell metaplasia). • Ulceration. Ulceration is the usual outcome of severe active disease. Ulceration may be superficial, may undermine adjacent mucosa in a lateral fashion, or may penetrate deeply into underlying tissue layers. There is often an abrupt transition between ulcerated and adjacent normal mucosa. • Transmural inflammation affecting all layers. Chronic inflammatory cells suffuse the affected mucosa and, to a lesser extent, all underlying tissue layers. Lymphoid aggregates are usually scattered throughout the bowel wall. • Noncaseating granulomas. In about half of the cases, sarcoid-like granulomas may be present in all tissue layers, both within areas of active disease and in uninvolved regions of the bowel ( Fig. 17-42 ). Granulomas have been documented throughout the alimentary tract, from mouth to rectum, in patients with CD limited to one bowel segment. Conversely, the absence of granulomas does not preclude the diagnosis of CD. • Other mural changes. In diseased segments, the muscularis mucosa usually exhibits reduplication, thickening, and irregularity. Fibrosis of the submucosa, muscularis propria, and mucosa eventually leads to stricture formation. Less common findings are mucosal and submucosal lymphangiectasia, hypertrophy of mural nerve fibers, and localized vasculitis.
Figure 17-41 Crohn disease of the colon; a deep fissure extending into the muscle wall, a second, shallow ulcer (on the upper right), and relative preservation of the intervening mucosa. Abundant lymphocyte aggregates are present, evident as dense blue patches of cells at the interface between mucosa and submucosa.
Clinical Features.
The clinical manifestations of Crohn disease are extremely variable. They are generally more subtle than those of UC. The disease usually begins with intermittent attacks of relatively mild diarrhea, fever, and abdominal pain, spaced by asymptomatic periods lasting for weeks to many months. Often the attacks are precipitated by periods of physical or emotional stress. Although emotional influences are thought not to have any direct role in the initiation of the disease, they may contribute to flare-ups. In those with colonic involvement, occult or overt fecal blood loss may lead to
Figure 17-42 Crohn disease of the colon. A noncaseating granuloma is present in the lamina propria of an uninvolved region of colonic mucosa (arrow).
849
anemia over time, but massive bleeding is uncommon. In about one-fifth of patients the onset is more abrupt, with acute right lower quadrant pain, fever, and diarrhea sometimes suggesting acute appendicitis or an acute bowel perforation. The course of the disease includes bouts of diarrhea with fluid and electrolyte losses, weight loss, and weakness.
During this lengthy, chronic disease, complications may arise from fibrosing strictures, particularly of the terminal ileum, and fistulas to other loops of bowel, the urinary bladder, vagina, or perianal skin, or into a peritoneal abscess. Extensive involvement of the small bowel, including the terminal ileum, may cause marked loss of albumin (protein-losing enteropathy), generalized malabsorption, specific malabsorption of vitamin B12 (resulting in pernicious anemia), or malabsorption of bile salts, leading to steatorrhea.
Extraintestinal manifestations of this disease include migratory polyarthritis, sacroiliitis, ankylosing spondylitis, erythema nodosum, and clubbing of the fingertips. Hepatic primary sclerosing cholangitis (see Chapter 18 ) occurs, but the association is not as strong as in UC. Any of these manifestations can develop before onset of intestinal symptoms. Deranged systemic immunity is thought to underlie these related disorders. Uveitis, nonspecific mild hepatic pericholangitis, and renal disorders secondary to trapping of the ureters in the inflammatory process sometimes develop. Systemic amyloidosis is a rare late consequence.
There is an increased incidence of cancer of the gastrointestinal tract in patients with long-standing progressive CD, with a five- to six-fold increased risk over age-matched populations. However, the risk of cancer in CD is considerably less than in patients with chronic UC.
Ulcerative colitis is an ulceroinflammatory disease limited to the colon and affecting only the mucosa and submucosa except in the most severe cases. Unlike CD, UC extends in a continuous fashion proximally from the rectum. Well-formed granulomas are absent. Like CD, UC is a systemic disorder associated in some patients with migratory polyarthritis, sacroiliitis, ankylosing spondylitis, uveitis, hepatic involvement (pericholangitis and primary sclerosing cholangitis; Chapter 18 ), and skin lesions.
Epidemiology.
UC is global in distribution and varies in incidence relative to CD, supporting the concept that they are separate diseases. In the United States, Great Britain, and Scandinavia the incidence is about 4 to 12 per 100,000 population, which is slightly greater than CD. As with CD, the incidence of this condition has risen in recent decades. In the United States it is more common among whites than among blacks, and females are affected more often than males. The onset of disease peaks between ages 20 and 25, but the condition may arise in both younger and considerably older individuals. Nonsmoking is associated with UC; ex-smokers are at higher risk for developing UC than never-smokers.
Morphology.
Ulcerative colitis involves the rectum and extends proximally in a retrograde fashion to involve the entire colon ("pancolitis") in the more severe cases. It is a disease of continuity, and "skip" lesions such as occur in Crohn disease are not found ( Fig. 17-43 ). In 10% of patients with severe pancolitis, the distal ileum may develop mucosal inflammation ("backwash ileitis"). This is probably due to the incompetence of the iliocecal valve, resulting in reflux of the inflammatory material from the colon. In contrast to CD, the ileitis is often diffuse and limited to within 25 cm from the ileocecal valve. The appendix may be involved with both CD and UC.
In the course of colonic involvement with UC, the mucosa may exhibit slight reddening and granularity with friability and easy bleeding. With fully developed severe, active inflammation, there may be extensive and broad-based ulceration of the mucosa in the distal colon or throughout its length ( Fig. 17-44 ). Isolated islands of regenerating mucosa bulge upward to create pseudopolyps. Often the undermined edges of adjacent ulcers interconnect to create tunnels covered by tenuous mucosal bridges. As with CD, the ulcers of UC are frequently aligned along the axis of the colon, but rarely do they replicate the linear serpentine ulcers of CD. With indolent chronic disease or with healing of active disease, progressive mucosal atrophy leads to a flattened and attenuated mucosal surface ( Fig. 17-45 ). Unlike CD, mural thickening does not occur in UC, and the serosal surface is usually completely normal. Only in the most severe cases of ulcerative disease (UC, CD, and other severe inflammatory diseases) does toxic damage to the muscularis propria and neural plexus lead to complete shutdown of neuromuscular function. In this instance the colon progressively swells and becomes gangrenous (toxic megacolon) ( Fig. 17-46 ).
The mucosal alterations in UC are similar to those of colonic CD, with inflammation, chronic mucosal
Figure 17-43 Comparison of the distribution patterns of Crohn disease and ulcerative colitis, as well as the different conformations of the ulcers and wall thickenings.
850
Figure 17-44 Ulcerative colitis. Ulcerated hemorrhagic surface with knobby pseudopolyps. (Courtesy of Dr. Kim Bechard, Brigham and Women's Hospital, Boston, MA.)
damage, and ulceration ( Fig. 17-47 ). First, a diffuse, predominantly mononuclear inflammatory infiltrate in the lamina propria is almost universally present, even at the time of clinical presentation. Neutrophilic infiltration of the epithelial layer may produce collections of neutrophils in crypt lumina (crypt abscesses). These are not specific for UC and may be observed in CD or any active inflammatory colitis. Unlike CD, there are no granulomas, although rupture of crypt abscesses may incite a foreign body reaction in the lamina propria. Second, further destruction of the mucosa leads to outright ulceration, extending into the submucosa and sometimes leaving only the raw, exposed muscularis propria. Third, with remission of active disease, granulation tissue fills in the ulcer craters, followed by regeneration of the mucosal epithelium. Submucosal fibrosis and mucosal architectural disarray and atrophy remain as residua of healed disease.
Figure 17-45 Ulcerative colitis. Low-power micrograph showing marked chronic inflammation of the mucosa with atrophy of colonic glands, moderate submucosal fibrosis, and a normal muscle wall.
Figure 17-46 Toxic megacolon. Complete cessation of colon neuromuscular activity has led to massive dilatation of the colon and black-green discoloration signifying gangrene and impending rupture.
A key feature of UC is that the mucosal damage is continuous from the rectum and extending proximally. In CD, mucosal damage in the colon may be continuous but is just as likely to exhibit skip areas. It should be noted that quiescent UC, particularly treated disease in which active neutrophilic inflammation is not present, may appear virtually normal histologically. This does not preclude risk for dysplasia, as now described.
Particularly significant in ulcerative colitis is the spectrum of epithelial changes signifying dysplasia and the progression to frank carcinoma. Nuclear atypia and loss of cytoplasmic differentiation may be present in inflamed or uninflamed colonic mucosa. Epithelial dysplasia is referred to as being low-grade or high-grade; [75] cytologic features are the key to evaluating dysplasia. Distinguishing between regenerative changes and dysplasia can be very difficult and sometimes impossible. Pathologists are allowed some latitude in noting atypical changes that may not be definitive for a diagnosis of dysplasia.[76] Plaque-like dysplastic lesions, overt polypoid dysplasia (adenomas),
Figure 17-47 Ulcerative colitis. Microscopic view of the mucosa, showing diffuse active inflammation with crypt abscess and glandular architectural distortion.
851
or invasive carcinoma are the ultimate lesions arising from flat dysplasia. It should be noted that elderly patients with UC are also at risk for sporadic adenomas. Distinction between IBD-associated dysplasia and a coexistent incidental adenoma may be difficult.
Clinical Features.
Ulcerative colitis typically presents as a relapsing disorder marked by attacks of bloody mucoid diarrhea that may persist for days, weeks, or months and then subside, only to recur after an asymptomatic interval of months to years or even decades. In the fortunate patient, the first attack is the last. At the other end of the spectrum, the explosive initial attack may lead to such serious bleeding and fluid and electrolyte imbalance as to constitute a medical emergency. In most patients, bloody diarrhea containing stringy mucus, accompanied by lower abdominal pain and cramps usually relieved by defecation, is the first manifestations of the disease. In a small number of patients, constipation may appear paradoxically, due to disruption of normal peristalsis. Often the first attack is preceded by a stressful period in the patient's life. Spontaneously, or more often after appropriate therapy, these symptoms abate in the course of days to weeks. Flare-ups, when they do occur, may be precipitated by emotional or physical stress and rarely by concurrent intraluminal growth of enterotoxin-forming C. difficile. Sudden cessation of bowel function with toxic dilatation (toxic megacolon) rarely develops with severe acute attacks; perforation is a potentially lethal event.
The outlook for patients with UC depends on two factors: (1) the severity of active disease and (2) its duration. About 60% of patients have clinically mild disease. In these individuals, the bleeding and diarrhea are not severe, and systemic signs and symptoms
are absent. However, almost all patients (97%) have at least one relapse during a 10-year period, and about 30% of patients require colectomy within the first 3 years of onset due to uncontrollable disease. On rare occasion,
TABLE 17-10 -- Distinctive Features of Crohn Disease and Ulcerative Colitis *
Feature Crohn Disease - SI Crohn Disease - C Ulcerative Colitis
Macroscopic
Bowel region Ileum ± colon Colon ± ileum Colon only
Distribution Skip lesions Skip lesions Diffuse
Stricture Early Variable Late/rare
Wall appearance Thickened Thin Thin
Dilation No Yes Yes
Microscopic
Inflammation Transmural Transmural Limited in mucosa
Pseudopolyps No to slight Marked Marked
Ulcers Deep, linear Deep, linear Superficial
Lymphoid reaction Marked Marked Mild
Fibrosis Marked Moderate Mild
Serositis Marked Variable Mild to none
Granulomas Yes (50%) Yes (50%) No
Fistulae/sinuses Yes Yes No
Clinical
Fat/vitamin malabsorption
Yes Yes, if ileum No
Malignant potential Yes Yes Yes
Response to surgery Poor Fair Good
*SI, Crohn disease of the small intestine; C, Crohn disease of the colon. Features are often not all present in a single case.
the disease runs a fulminant course; unless medically or surgically controlled, this toxic form of the disease can lead to death soon after onset.
The most feared long-term complication of UC is cancer. There is a tendency for dysplasia to arise in multiple sites, and the underlying inflammatory disease may mask the symptoms and signs of carcinoma. UC is characterized by DNA damage with microsatellite instability in mucosal cells. More recently, genomic instability was detected in non-dysplastic areas of patients with UC, suggesting that these patients have DNA repair deficiency and genomic instability throughout the intestinal tract.[77] The associated carcinomas are often infiltrative without obvious exophytic masses, further underscoring the importance of early diagnosis. Historically, the risk of cancer is highest in patients with pancolitis of 10 or more years' duration, in whom it is 20- to 30-fold higher than in a control population.[78] However, recent screening programs of patients with UC now indicate that the rate of progression to dysplasia and carcinoma is in fact quite low, provided that initial examinations were negative for dysplasia. Since great cost is involved in mass screening, the debate over the cost-effectiveness of repeated colonoscopies in patients with long-term inactive disease continues; the modest improvement in patient outcome may be related to better patient care, rather than identification of dysplasia per se.
The features of CD and UC are compared in Table 17-10 .
Vascular Disorders
ISCHEMIC BOWEL DISEASE
Ischemic lesions may be restricted to the small or large intestine, or may affect both, depending on the particular vessel(s) affected. Acute occlusion of one of the three major supply trunks of the intestines—celiac, superior mesenteric, and inferior mesenteric arteries—may lead to infarction of several meters of intestine. However, insidious loss of one
Figure 17-48 Acute ischemic bowel disease. Schematic of the three levels of severity, diagrammed for the small intestine.
vessel may be without effect, due to the rich anastomotic interconnections. Lesions within the end arteries, which penetrate the gut wall, produce small, focal ischemic lesions. As depicted in Figure 17-48 , the severity of injury ranges from: (1) transmural infarction of the gut, involving all visceral layers; to (2) mural infarction of the mucosa and submucosa; to (3) mucosal infarction, if the lesion extends no deeper than the muscularis mucosae. Almost always, transmural infarction is caused by mechanical compromise of the major mesenteric blood vessels. Mucosal Mucosal or mural infarction more often results from hypoperfusion, either acute or chronic. Mesenteric venous thrombosis is a less frequent cause of vascular compromise. The predisposing conditions for ischemia are as follows:
• Arterial thrombosis: severe atherosclerosis (usually at the origin of the mesenteric vessel), systemic vasculitis, dissecting aneurysm, angiographic procedures, aortic reconstructive surgery, surgical accidents, hypercoagulable states, and oral contraceptives • Arterial embolism: cardiac vegetations, angiographic procedures, and aortic atheroembolism • Venous thrombosis: hypercoagulable states, oral contraceptives, antithrombin III deficiency, intraperitoneal sepsis, the postoperative state, invasive neoplasms (particularly hepatocellular carcinoma), cirrhosis, and abdominal trauma • Nonocclusive ischemia: cardiac failure, shock, dehydration, and vasoconstrictive drugs (e.g., digitalis, vasopressin, propranolol) • Miscellaneous: radiation injury, volvulus, stricture, amyloidosis, diabetes mellitus, and internal or external herniation.
Embolic arterial occlusion most often involves the branches of the superior mesenteric artery. The origin of the inferior mesenteric artery from the artery is more oblique, and this may contribute to the relative sparing of this arterial axis from embolism. Despite the multiplicity of possible causes, there remains a significant percentage of cases in which no well-defined basis for the vascular insufficiency can be identified. Mesenteric vascular spasm has been invoked in some cases, without definitive proof.
Ischemic injury has two phases: the initial hypoxic injury at the onset of blood supply compromise and secondary reperfusion injury at the time of blood resupply to the hypoxic tissue. Most of the intestinal injury in ischemic bowel disease is actually caused by reperfusion. The underlying pathophysiology in reperfusion injury is a complex process. Among the most important factors in this process are the generation of oxygen free radicals, neutrophil infiltration, and the production of inflammatory mediators in tissue that has insufficient metabolic reserve to detoxify injurious free radicals and other mediators ( Chapter 1 ).
The severity of vascular compromise and the time frame during which it develops are major determinants of the morphology of ischemic bowel diseases. The most severe, acute lesions are considered first.
Transmural Infarction.
Small intestinal infarction following sudden and total occlusion of mesenteric arterial blood flow may involve only a short segment, but more often involves a substantial portion. The splenic flexure of the colon is at greatest risk of ischemic injury because it is the watershed between the distribution of the superior and inferior mesenteric arteries, but any portion of the colon may be affected. With mesenteric venous occlusion, anterograde and retrograde propagation of thrombus may lead to extensive involvement of the splanchnic bed. Regardless of whether the arterial or venous side is occluded, the infarction appears hemorrhagic because of blood reflow into the damaged area. In the early stages, the infarcted bowel appears intensely congested and dusky to purple-red ( Fig. 17-49 ), with foci of subserosal and submucosal ecchymotic discoloration. With time, the wall becomes edematous, thickened, rubbery, and hemorrhagic. The lumen commonly contains sanguineous mucus or frank blood. In arterial occlusions the demarcation from normal bowel is usually sharply defined, but in venous occlusions the area of dusky cyanosis fades gradually into the adjacent normal bowel, having no clear-cut definition between viable and nonviable bowel. Histologically, there is obvious edema, interstitial hemorrhage, and sloughing necrosis of the mucosa. Normal features of the mural musculature, particularly cellular nuclei,
Figure 17-49 Infarcted small bowel, secondary to acute thrombotic occlusion of the superior mesenteric artery.
become indistinct. Within 1 to 4 days, intestinal bacteria produce outright gangrene and sometimes perforation of the bowel. There may be little inflammatory response.
Mucosal and Mural Infarction.
Mucosal and mural infarction may involve any level of the gut from the stomach to the anus. The lesions may be multifocal or continuous and widely distributed. Affected areas of the bowel may appear dark red or purple, owing to the accumulated luminal hemorrhage. However, hemorrhage and an inflammatory exudate are absent from the serosal surface. On opening the bowel, there is hemorrhagic, edematous thickening of the mucosa, which may penetrate more deeply into the submucosa and muscle wall. Superficial ulceration may be present.
In the mildest form of ischemic injury, the superficial epithelium of the colon or the tips of small intestinal villi may be necrotic or sloughed. Inflammation is absent, and there may only be mild vascular dilation. With complete mucosal necrosis, epithelial sloughing leaves behind only the acellular scaffolding of the lamina propria ( Fig. 17-50 ). When severe, there is extensive hemorrhage and necrosis of multiple tissue layers. Secondary acute and chronic inflammation is evident along the viable margins underlying and adjacent to the affected area. Bacterial superinfection and the formation of enterotoxic bacterial products may induce superimposed pseudomembranous inflammation, particularly in the colon. Thus, the mucosal changes may mimic enterocolitis of nonvascular origin.
Chronic Ischemia.
With chronic vascular insufficiency to a region of intestine, mucosal inflammation and ulceration may develop, mimicking both acute enterocolitis from other causes and idiopathic IBD. Submucosal chronic inflammation and fibrosis may lead to stricture ( Fig. 17-51 ). Although colonic strictures
Figure 17-50 Mucosal infarction of the small bowel. The mucosa is hemorrhagic, and there is no epithelial layer. The remaining layers of the bowel are intact.
Figure 17-51 Chronic ischemia of the colon, resulting in chronic mucosal damage and a stricture.
typically occur in the watershed area of the splenic flexure, both acute and chronic mucosal ischemia are notoriously segmental and patchy.
Clinical Features.
Bowel infarction is an uncommon but grave disorder that imposes a 50% to 75% death rate, largely because the window of time between onset of symptoms and perforation is small. It tends to occur in older individuals, when cardiac and vascular diseases are most prevalent. Preexistent abdominal disease also increases the risk of bowel infarction, due to adhesions and torsion. Severe abdominal pain and tenderness develop suddenly in the setting of transmural infarction, sometimes accompanied by nausea, vomiting, and bloody diarrhea or grossly melanotic stool. Patients may progress to shock and vascular collapse within hours. Peristaltic sounds diminish or disappear, and spasm creates board-like rigidity of the abdominal wall musculature. Because there are far more common causes of these physical signs, such as acute appendicitis, perforated peptic ulcer, and acute cholecystitis, the diagnosis of intestinal gangrene may be delayed or missed, with disastrous consequences.
Mucosal and mural infarction, by themselves, may not be fatal, particularly if the cause of vascular compromise is corrected. A confusing array of nonspecific abdominal complaints, combined with intermittent bloody diarrhea, may be the only indication of nonocclusive enteric ischemia. Nevertheless, bowel embarrassment may progress to more extensive infarction, and sepsis or serious blood loss may set in. Chronic ischemic colitis
may present as an insidious inflammatory disease, with intermittent episodes of bloody diarrhea interspersed with periods of healing, mimicking IBD.
Several clinical conditions that cause ischemic intestinal injury merit emphasis. Ischemic bowel or infarction may occur in the setting of severe atherosclerosis of the aorta and mesenteric vasculature. Cholesterol emboli dislodged from large vessels occlude smaller vessels downstream, leading to regions of localized compromise. Second, vasculitis affecting the mesenteric vasculature may cause ischemic injury. The commonly seen vasculitides that affect the intestine are polyarteritis
854
nodosum, Henoch-Schönlein disease, and Wegener granulomatosis. Third, amyloidosis often affects mesenteric blood vessels and may actually present as chronic intestinal ischemia.
ANGIODYSPLASIA
Angiodysplasia is a non-neoplastic intestinal lesion of vascular dilation and malformation. Tortuous dilations of submucosal and mucosal blood vessels are seen most often in the cecum or right colon, usually only after the sixth decade of life. Although the prevalence of these lesions is less than 1% in the adult population, they account for 20% of significant lower intestinal bleeding; intestinal hemorrhage may be chronic and intermittent, or acute and massive. Most angiodysplasias span the mucosa and submucosa and contain a small amount of smooth muscle, suggesting that they are ectatic nests of preexisting veins, venules, and capillaries. The vascular channels may be separated from the intestinal lumen by only the vascular wall and a layer of attenuated epithelial cells, explaining the propensity toward bleeding.
The pathogenesis of angiodysplasia remains speculative, but it is attributed to mechanical factors operative in the colonic wall, with possibly a congenital contribution. Normal distention and contraction may intermittently occlude the submucosal veins that penetrate through the muscle wall. This then leads to focal dilation and tortuosity of overlying submucosal and mucosal vessels. According to LaPlace's Law, tension in the wall of a cylinder is a function of intraluminal pressure and diameter. Because the cecum has the widest diameter of the colon, it develops the greatest wall tension, perhaps explaining the distribution of these lesions. Vascular degenerative changes related to aging may also play some role. The evidence to support a congenital cause is the association of angiodysplasia with other congenital abnormalities such as aortic stenosis and Meckel diverticulum.
HEMORRHOIDS
Hemorrhoids are variceal dilations of the anal and perianal venous plexuses. These extremely common lesions affect about 5% of the general population and develop
secondary to persistently elevated venous pressure within the hemorrhoidal plexus. The most frequent predisposing influences are constipation with straining at stool and the venous stasis of pregnancy. Except for pregnant women, they are rarely encountered in persons under age 30. More rarely, but much more importantly, hemorrhoids may reflect collateral anastomotic channels that develop as a result of portal hypertension ( Chapter 18 ).
Morphology.
The varicosities may develop in the inferior hemorrhoidal plexus and thus are located below the anorectal line (external hemorrhoids). Alternatively, they may develop from dilation of the superior hemorrhoidal plexus and produce internal hemorrhoids. Commonly, both plexuses are affected, and the varicosities are referred to as combined hemorrhoids. Histologically, these lesions consist only of thin-walled, dilated, submucosal varices that protrude beneath the anal or rectal mucosa. In their exposed, traumatized position, they tend to become thrombosed and, in the course of time, recanalized. Superficial ulceration, fissure formation, and infarction with strangulation may develop.
Diverticular Disease
A diverticulum is a blind pouch leading off the alimentary tract, lined by mucosa that communicates with the lumen of the gut. Congenital diverticula involve all three layers of the bowel wall. The prototype is the Meckel diverticulum, discussed earlier; congenital diverticula are not uncommon in the ascending colon.
Virtually all other diverticula are acquired and either lack or have an attenuated muscularis propria. Acquired diverticula may occur in the esophagus, stomach, and duodenum, but the most common site is the left side of the colon, with the majority in the sigmoid colon. Acquired duodenal diverticula occur in over 1% of adults, possibly reflecting defects from healed peptic ulcer disease. Multiple diverticula of the jejunum and ileum are rare, occurring in the setting of abnormalities in the muscle wall or myenteric plexus.
Unless otherwise specified, diverticular disease refers to acquired outpouchings of the colonic mucosa and submucosa. Colonic diverticula are rare in persons under age 30, but in Western adult populations over age 60 the prevalence approaches 50%. They generally occur multiply and are referred to as diverticulosis. They are much less frequent in nonindustrialized tropical countries and in Japan.
Morphology.
Most colonic diverticula are small, flask-like or spherical outpouchings, usually 0.5 to 1 cm in diameter and located in the sigmoid colon ( Fig. 17-52 A ). However, the descending colon or entire colon may be affected. They tend to occur alongside the taeniae coli and are elastic, compressible, and easily emptied of fecal contents. As these sacs dissect into the fat-containing peritoneal pouches on the surface of the colon
(epiploic appendices), they may be missed on casual inspection. Histologically, colonic diverticula have a thin wall composed of a flattened or atrophic mucosa, compressed submucosa, and attenuated or totally absent muscularis propria ( Fig. 17-52 B ). Hypertrophy of the circular layer of the muscularis propria in the affected bowel segment is usually seen; the taeniae coli are also unusually prominent.
Obstruction and/or perforation of diverticula leads to inflammatory changes, producing peridiverticulitis and dissecting into the immediately adjacent pericolic fat. In time, the inflammation may lead to marked fibrotic thickening in and about the colonic wall, sometimes producing narrowing sufficient to resemble a colonic cancer. Extension of diverticular infection may lead to pericolic abscesses, sinus tracts, and sometimes pelvic or generalized peritonitis.
Pathogenesis.
The morphology of colonic diverticula strongly suggests that two factors are important in their genesis: (1) focal weakness in the colonic wall and (2) increased intraluminal pressure. The colon is unique in that the longitudinal
855
Figure 17-52 Diverticulosis. A, Section through the sigmoid colon, showing multiple sac-like diverticula protruding through the muscle wall into the mesentery. The muscularis propria in between the diverticular protrusions is markedly thickened. B, Low-power photomicrograph of diverticulum of the colon, showing protrusion of mucosa and submucosa through the muscle wall. A dilated blood vessel at the base of the diverticulum was a source of bleeding; some blood clot is present within the diverticular lumen.
muscle coat is not complete, but is gathered into three equidistant bands (the taeniae coli). Where nerves and arterial vasa recta penetrate the inner circular muscle coat alongside the taeniae, focal defects in the muscle wall are created. The connective tissue sheaths accompanying these perforating vessels provide points of weakness for herniations. Exaggerated peristaltic contractions, with spasmodic sequestration of bowel segments,
are the likely cause of increased intraluminal pressure. It has been proposed that diets low in fiber reduce stool bulk, which in turn leads to increased peristaltic activity, particularly in the sigmoid colon. Exaggerated contractions sequester segments of bowel (segmentation); this deranged motility can lead to symptoms in the absence of inflammation.
Clinical Features.
Most individuals with diverticular disease remain asymptomatic throughout their lives, and the lesions are most often discovered incidentally. Only about 20% of those affected ever develop manifestations. These may include intermittent cramping or continuous lower abdominal discomfort, constipation, distention, and a sensation of never being able to completely empty the rectum. Patients sometimes experience alternating constipation and diarrhea. Occasionally there may be minimal chronic or intermittent blood loss, or rarely massive hemorrhages.
Longitudinal studies have shown that diverticula can regress early in their development or may become more numerous and prominent with time. Whether a high-fiber diet prevents such progression or protects against superimposed diverticulitis is still unclear. Diets supplemented with high fiber may provide symptomatic improvement, but the treatment may seem worse than the disease. Even when diverticulitis supervenes, it most often resolves spontaneously. Relatively few patients require surgical intervention for obstructive or inflammatory complications.
Intestinal Obstruction
Obstruction of the gastrointestinal tract may occur at any level, but the small intestine is most often involved due to its narrow lumen. The causes of small and large intestinal obstruction are presented in Table 17-11 . Tumors and infarction, although the most serious, account for only about 10% to 15% of small-bowel obstructions. Four of the entities—hernias, intestinal adhesions, intussusception, and volvulus—collectively account for 80% ( Fig. 17-53 ). The clinical manifestations of intestinal obstruction include abdominal pain and distention, vomiting, constipation, and failure to pass flatus. If the obstruction is mechanical or vascular in origin, immediate surgical intervention is usually required.
HERNIAS
A weakness or defect in the wall of the peritoneal cavity may permit protrusion of a pouch-like, serosa-lined sac of peritoneum called a hernial sac. The usual sites of such weakness are anterior at the inguinal and femoral canals, umbilicus, and in surgical scars. Rarely, retroperitoneal hernias may occur, chiefly about the ligament of Trietz. Hernias are of concern chiefly because segments of viscera frequently protrude and become trapped in them (external herniation). This is particularly true with inguinal hernias, since they tend to have narrow orifices and large sacs. The most frequent intruders are small-bowel loops, but portions of omentum or large bowel also may
TABLE 17-11 -- Major Causes of Intestinal Obstruction
Mechanical Obstruction
Adhesions
Hernias, internal or external
Volvulus
Intussusception
Tumors
Inflammatory strictures
Obstructive gallstones, fecaliths, foreign bodies
Congenital strictures; atresias
Congenital bands
Meconium in mucoviscoidosis
Imperforate anus
Pseudo-obstruction
Paralytic ileus (e.g., postoperative)
Vascular—bowel infarction
Myopathies and neuropathies (e.g., Hirschsprung)
856
Figure 17-53 Schematic depicting the four major causes of intestinal obstruction: (1) Herniation of a segment in the umbilical or inguinal regions; (2) adhesion between loops of intestine; (3) intussusception; (4) volvulus formation.
become trapped. Pressure at the neck of the pouch may impair venous drainage of the trapped viscus. The resultant stasis and edema increase the bulk of the herniated loop, leading to permanent trapping, or incarceration. With time, compromise of arterial supply and venous drainage (strangulation) leads to infarction of the trapped segment.
ADHESIONS
Surgical procedures, infection, and even endometriosis often cause localized or more general peritoneal inflammation (peritonitis). As the peritonitis heals, adhesions may develop between bowel segments and/or the abdominal wall and operative site. These fibrous bridges can create closed loops through which other viscera may slide and eventually become trapped (internal herniation). The sequence of events following herniation—obstruction and strangulation—is much the same as with external hernias. Quite rarely, fibrous adhesions arise as congenital defects. Intestinal herniation must be considered, then, even without a previous history of peritonitis or surgery.
INTUSSUSCEPTION
Intussusception occurs when one segment of the intestine, constricted by a wave of peristalsis, suddenly becomes telescoped into the immediately distal segment of bowel. Once trapped, the invaginated segment is propelled by peristalsis farther into the distal segment, pulling its mesentery along behind it. When encountered in infants and children, there is usually no underlying anatomic lesion or defect in the bowel, and the patient is otherwise healthy. Some cases of intussusception are associated with rotavirus infection, suggesting that localized intestinal inflammation may serve as a traction point for the intussusception. However, intussusception in adults signifies an intraluminal mass or tumor as the point of traction. In both settings, intestinal obstruction ensues, and trapping of mesenteric vessels leads to infarction.
VOLVULUS
Complete twisting of a loop of bowel about its mesenteric base of attachment also produces intestinal obstruction and infarction. This lesion occurs most often in large redundant loops of sigmoid, followed in frequency by the cecum, small bowel (all or portions), stomach, or (rarely) transverse colon. Recognition of this seldom-encountered lesion demands constant awareness of its possible occurrence.
Tumors of the Small and Large Intestine
Epithelial tumors of the intestines are a major cause of morbidity and mortality worldwide. The colon (including the rectum) is host to more primary neoplasms than any other organ in the body. Colorectal cancer ranks second only to bronchogenic carcinoma among the cancer killers in North America. Adenocarcinomas constitute the vast majority
of colorectal cancers and represent 70% of all malignancies arising in the gastrointestinal tract. Curiously, the small intestine is an uncommon site for benign or malignant tumors despite its great length and vast pool of dividing mucosal cells.
The classification of intestinal tumors is the same for the small intestine and colon and is summarized in Table 17-12 . Although small intestinal tumors are addressed first, the bulk of our discussion is devoted to colorectal neoplasia.
TUMORS OF THE SMALL INTESTINE
While the small bowel represents 75% of the length of the alimentary tract, its tumors account for only 3% to 6% of gastrointestinal tumors, with a slight preponderance of benign tumors. The most common benign tumors in the small intestine are adenomas and mesenchymal tumors (see later discussion on gastrointestinal stromal tumors). Lipomas,
857
TABLE 17-12 -- Tumors of the Small Intestine and Colon
Non-neoplastic (Benign) Polyps
Hyperplastic polyps
Hamartomatous polyps
• Juvenile polyps
• Peutz-Jeghers polyps
Inflammatory polyps
Lymphoid polyps
Neoplastic Epithelial Lesions
Benign
• Adenoma *
Malignant
• Adenocarcinoma *
• Carcinoid tumor
• Anal zone carcinoma
Mesenchymal Lesions
Gastrointestinal stromal tumor (GIST) (gradation from benign to malignant)
* Benign and malignant counterparts of the most common neoplasms in the intestines; virtually all lesions are in the colon.
angiomas, and rare hamartomatous mucosal lesions comprise the remainder. One of the enigmas of medicine is the rarity of malignant tumors of the small intestine—annual U.S. death rate is under 1000, representing only about 1% of gastrointestinal malignancies. Small intestinal adenocarcinomas and carcinoids have roughly equal incidence, followed in order by lymphomas and sarcomas. As the latter three exhibit a broader distribution than the small intestine, they are discussed later.
Adenomas
Adenomas account for approximately 25% of benign small intestinal tumors, with benign mesenchymal tumors (especially leiomyomas), lipomas, and neuromatous lesions following in frequency. Most adenomas occur in the region of the ampulla of Vater. The usual presentation is that of a 30- to 60-year-old patient with occult blood loss, rarely with obstruction or intussusception; some are discovered incidentally during radiographic investigation. Patients with familial polyposis coli (discussed later) are particularly prone to developing periampullary adenomas. Macroscopically, the ampulla of Vater is enlarged and exhibits a velvety surface ( Fig. 17-54 ). Microscopically, these adenomas resemble their counterparts in the colon (discussed later). Frequently, there is extension of adenomatous tissue into the ampullary orifice, rendering surgical excision difficult, short of a pancreatoduodenectomy to remove the entire ampullary region. Like its counterpart in the colon, the small intestinal adenoma is a premalignant lesion. The adenoma-carcinoma sequence has been demonstrated in small intestinal tumors.
Figure 17-54 Adenoma of the ampulla of Vater, showing exophytic tumor at the ampullary orifice.
Adenocarcinoma
The large majority of small intestinal adenocarcinomas occur in the duodenum, usually in 40- to 70-year-old patients. These tumors grow in a napkin-ring encircling pattern or as polypoid exophytic masses, in a manner similar to colonic cancers. Tumors in the duodenum, particularly those involving the ampulla of Vater, may cause obstructive jaundice early in their course. More typically, intestinal obstruction is the presenting event, with symptoms of cramping pain, nausea, vomiting, and weight loss. As in patients with adenoma, fatigue from occult blood loss may be the only sign. Rarely, the tumorous mass is a lead point for intussusception.
A major risk factor for adenocarcinoma of the small intestine is the chronic inflammation associated with CD, although most tumors are sporadic and have no identifiable predisposing condition. Other conditions with increased risk for small intestinal adenocarcinoma are: celiac disease, familial adenomatous polyposis (FAP), hereditary nonpolyposis colorectal cancer (HNPCC) syndrome, and Peutz-Jeghers syndrome. From a broader epidemiologic perspective, alcohol and tobacco consumption are considered risk factors.
At the time of diagnosis, most tumors have already penetrated the bowel wall, invaded the mesentery or other segments of the gut, spread to regional nodes, and sometimes metastasized to the liver and even more widely. Despite these problems, wide en bloc excision of these cancers yields about a 70% five-year survival rate.
TUMORS OF THE COLON AND RECTUM
Colorectal carcinoma is one of the most common malignancies of Western countries. Consideration must first be given to the panoply of non-neoplastic and neoplastic but benign tumorous lesions of the colon and rectum. These are collectively known as polyps. Polyps of the colorectal mucosa are extraordinarily common in the older adult
population. Several concepts pertaining to terminology must be emphasized ( Fig. 17-55 ):
• A polyp is a tumorous mass that protrudes into the lumen of the gut. Presumably all polyps start as small, sessile lesions without a definable stalk. In many instances, traction on the mass may create a stalked, or pedunculated polyp.
858
• Polyps may be formed as the result of abnormal mucosal maturation, inflammation, or architecture. These polyps are non-neoplastic and do not have malignant potential per se. An example is the hyperplastic polyp. • Those epithelial polyps that arise as the result of proliferation and dysplasia are termed adenomatous polyps, or adenomas. They are true neoplastic lesions and are precursors of carcinoma. • Some polypoid lesions may be caused by submucosal or mural tumors. However, as with the stomach and small intestine, unless otherwise specified the term polyp refers to lesions arising from the epithelium of the mucosa.
Figure 17-55 Diagrammatic representation of two forms of sessile polyp (hyperplastic polyp and adenoma) and of two types of adenoma (pedunculated and sessile). There is only a loose association between the
tubular architecture for pedunculated adenomas and the villous architecture for sessile adenomas.
TABLE 17-13 -- Hereditary Syndromes Involving the Gastrointestinal Tract
Syndromes Altered GenePathology in GI
Tract
Familial adenomatous polyposis (FAP)
APC Multiple adenomatous polyps
• Classic FAP
• Attenuated FAP
• Gardner syndrome
• Turcot syndrome
Peutz-Jeghers syndrome STK11 Hamartomatous polyps
Juvenile polyposis syndrome SMAD4 Juvenile polyps
BMPRIA
Hereditary nonpolyposis colorectal carcinoma
Defects in mismatch DNA repair genes
Colon cancer
Tuberous sclerosis TSC1 Inflammatory polyps
TSC2
Cowden disease PTEN Hamartomatous polyps
Non-Neoplastic Polyps
The overwhelming majority of intestinal polyps occur on a sporadic basis, particularly in the colon, and increase in frequency with age. Non-neoplastic polyps include the hyperplastic polyp, the hamartomatous polyp, the inflammatory polyp, and the lymphoid polyp. Hyperplastic polyps represent about 90% of all epithelial polyps in the large intestine. They may arise at any age but usually are discovered incidentally in the sixth and seventh decades. They are found in more than half of all persons age 60 and older. It is believed that the hyperplastic polyp results from decreased epithelial cell turnover and accumulation of mature cells on the surface. Harmatomatous polyps are malformations of the glands and the stroma. They can occur sporadically or occur in the setting of genetic syndromes ( Table 17-13 ). Inflammatory polyps, also known as pseudopolyps, represent islands of inflamed regenerating mucosa surrounded by ulceration. These are seen
primarily in patients with severe, active IBD. Lymphoid polyps are an essentially normal variant of the mucosal bumps containing intramucosal lymphoid tissue.
Morphology.
Hyperplastic Polyps.
These are small (usually <5 mm in diameter) epithelial polyps that appear as nipple-like, hemispheric, smooth, moist protrusions of the mucosa, usually positioned on the tops of mucosal folds. They may occur singly but more often are multiple, and over half are found in the rectosigmoid colon. Histologically, they are composed of well-formed glands and crypts lined by non-neoplastic epithelial cells, most of which show differentiation into mature goblet or absorptive cells. The delayed shedding of surface epithelial cells leads to infoldings of the crowded epithelial cells and fission of the crypts, creating a serrated epithelial profile and an irregular crypt architecture ( Fig. 17-56 A ). Although large hyperplastic polyps may rarely coexist with foci of adenomatous change, the usual small, hyperplastic polyp is considered to have virtually no malignant potential. However, the hyperplastic polyps occurring in the setting of the rare hyperplastic polyposis syndrome can harbor epithelial cell dysplasia (adenoma), and hence are considered at risk for carcinoma. The
859
underlying genetic basis for this syndrome is not known.
Hamartomatous Polyps.
Juvenile polyps represent focal hamartomatous malformations of the mucosal epithelium and lamina propria. For the most part they are sporadic lesions, with the vast majority occurring in children younger than age 5. Isolated hamartomatous polyps may be identified in the colon of adults; these incidental lesions are referred to as retention polyps. In both age groups, nearly 80% of the polyps occur in the rectum, but they may be scattered throughout the colon. Juvenile polyps tend to be large (1 to 3 cm in diameter), rounded, smooth or slightly lobulated lesions with stalks up to 2 cm in length; retention polyps tend to be smaller (<1 cm diameter). Histologically, lamina propria comprises the bulk of the polyp, enclosing abundant cystically dilated glands. Inflammation is common, and the surface may be congested or ulcerated. In general they occur singly and being hamartomatous lesions have no malignant potential. However, the rare autosomal dominant juvenile polyposis syndrome, in which there are multiple (50 to 100) juvenile polyps in the gastrointestinal tract, does carry a risk of adenomas and hence adenocarcinoma. Mutations in the SMAD4/DPC4 gene (which encodes a TGF-β signaling intermediate) account for some cases of juvenile polyposis syndrome.[79]
Peutz-Jeghers polyps are hamartomatous polyps that involve the mucosal epithelium, lamina propria, and muscularis mucosa. These hamartomatous lesions may also occur singly or multiply in the Peutz-Jeghers syndrome. This rare autosomal dominant
syndrome is characterized by multiple hamartomatous polyps scattered throughout the entire gastrointestinal tract and melanotic mucosal and cutaneous pigmentation around the lips, oral mucosa, face, genitalia, and palmar surfaces of the hands. Patients with this syndrome are at risk for intussusception, which is a common cause of mortality. Peutz-Jeghers polyps tend to be large and pedunculated with a firm lobulated contour. Histologically, an arborizing network of connective tissue and well-developed smooth muscle extends into the polyp and surrounds normal abundant
Figure 17-56 Non-neoplastic colonic polyps. A, Hyperplastic polyp; high-power view showing the serrated profile of the epithelial layer. B, Peutz-Jeghers polyp; low-power view showing the splaying of smooth muscle into the superficial portion of the pedunculated polyp.
glands lined by normal intestinal epithelium rich in goblet cells ( Fig. 17-56 B ). The distribution of polyps in patients is reported as follows: stomach, 25%; colon, 30%; and small bowel, 100%. While these hamartomatous polyps themselves do not have malignant potential, patients with the syndrome have an increased risk of developing carcinomas of the pancreas, breast, lung, ovary, and uterus. The well-documented and characteristic tumors include sex cord tumors of the ovary, adenoma malignum of the uterine cervix, and Sertoli cell tumors of the testis. When gastrointestinal adenocarcinoma occurs, it arises from concomitant adenomatous lesions. The underlying genetic basis for Peutz-Jeghers syndrome is the mutation of the gene STK11 (LKB1) located on chromosome 19. The gene encodes a protein with serine/threonine kinase activity.
Two other hamartomatous polyposis syndromes merit comment: Cowden syndrome and Cronkhite-Canada syndrome.
Cowden syndrome is an autosomal dominant genetic syndrome characterized by multiple hamartomas involving organs derived from all three germinal layers. The commonly involved sites are gastrointestinal tract and mucocutaneous locations. Intestinal hamartomatous polyps, facial trichilemmomas, acral keratoses, and oral papillomas are characteristic. While these hamartomas do not have malignant potential, the syndrome predisposes the patient to develop thyroid and breast cancers. The underlying genetic
abnormality is the germ line mutation of the PTEN gene located on chromosome 10 ( Chapter 7 ).
Cronkhite-Canada syndrome is a nonhereditary disorder characterized by the presence of gastrointestinal hamartomatous polyposis and ectodermal abnormalities (such as nail atrophy, skin pigmentation, and alopecia). The etiology of this disorder is currently unknown.
Adenomas
Adenomas (adenomatous polyps) are intraepithelial neoplasms that range from small, often pedunculated lesions to large neoplasms that are usually sessile. The prevalence of
860
colonic adenomas is about 20% to 30% before age 40, rising to 40% to 50% after age 60. Males and females are affected equally. There is a well-defined familial predisposition to sporadic adenomas, accounting for about a fourfold greater risk among first-degree relatives and also a fourfold greater risk of colorectal carcinoma.
Adenomatous polyps are segregated into three subtypes on the basis of the epithelial architecture:
• Tubular adenomas: tubular glands • Villous adenomas: villous projections • Tubulovillous adenoma: a mixture of the above.
There is considerable overlap among these categories, so by convention, tubular adenomas exhibit more than 75% tubular architecture, villous adenomas contain more than 50% villous architecture, and tubulovillous adenomas contain 25% to 50% villous architecture. Tubular adenomas are by far the most common; about 5% to 10% of adenomas are tubulovillous, and only 1% are villous.
All adenomatous lesions arise as the result of epithelial proliferative dysplasia, which may range from low-grade to high-grade dysplasia (carcinoma in situ). Furthermore, there is strong evidence that adenomas are a precursor lesion for invasive colorectal adenocarcinomas (discussed below).[80] The period required for an adenoma to double in size is estimated to be about 10 years. Thus, they are slow growing and must certainly have been present for many years before detection. The following concepts are pertinent:
• Most tubular adenomas are small and pedunculated; conversely, most pedunculated polyps are tubular. • Villous adenomas tend to be large and sessile, and sessile polyps usually exhibit villous features.
Figure 17-57 A, Pedunculated adenoma showing a fibrovascular stalk lined by normal colonic mucosa and a head that contains abundant dysplastic epithelial glands, hence the blue color with the H & E stain. B, A small focus of adenomatous epithelium in an otherwise normal (mucin-secreting, clear) colonic mucosa, showing how the dysplastic columnar epithelium (deeply stained) can populate a colonic crypt and create a tubular architecture.
The malignant risk with an adenomatous polyp is correlated with three interdependent features: polyp size, histologic architecture, and severity of epithelial dysplasia, as follows:
• Cancer is rare in tubular adenomas smaller than 1 cm in diameter. • The risk of cancer is high (approaching 40%) in sessile villous adenomas more than 4 cm in diameter. • Severe dysplasia, when present, is often found in villous areas.
Thus, the most worrisome lesions are villous adenomas greater than 4 cm in diameter. However, since all degrees of dysplasia (low-grade and high-grade) and even invasive adenocarcinoma may be encountered in an adenoma of any subtype, it is impossible from gross inspection of a polyp to determine its clinical significance.
It must be mentioned that not all adenomas are protuberant polyps. Some adenomas are essentially "flat" and can only be identified by histologic examination. These adenomas are referred to as flat adenoma, depressed adenoma, or microscopic adenoma.
Morphology.
Most tubular adenomas (90%) are found in the colon, but they can occur in the stomach and small intestine, especially in the vicinity of the ampulla of Vater. About half the time
they occur singly; in the remainder, two or more lesions are distributed at random. The smallest tubular adenomas are smooth-contoured and sessile; larger ones tend to be coarsely lobulated and have slender stalks ( Fig. 17-57 A ). Uncommonly, they exceed 2.5 cm in diameter. Histologically, the stalk is composed of fibromuscular tissue and prominent blood vessels (derived from the submucosa) and it is usually covered by
861
normal, non-neoplastic mucosa. However, adenomatous epithelium may extend down the stalk and into adjacent regions of the mucosa, particularly in the stomach. Whether small or large, adenomatous lesions are composed of neoplastic (dysplastic) epithelium, which lines glands as a tall, hyperchromatic, somewhat disordered epithelium that may or may not show mucin vacuoles ( Fig. 17-57 B ). In the clearly benign tubular adenoma, the branching glands are well separated by lamina propria and the degree of dysplasia is low-grade. However, high-grade dysplasia may be present and may merge with areas of overt malignant change confined to the mucosa (intramucosal carcinoma). Carcinomatous invasion into the submucosal stalk of the polyp constitutes invasive adenocarcinoma.
Villous adenomas are the larger and more ominous of the epithelial polyps. They tend to occur in older persons, most commonly in the rectum and rectosigmoid colon, but they may be located elsewhere. They generally are sessile, up to 10 cm in diameter, velvety or cauliflower-like masses projecting 1 to 3 cm above the surrounding normal mucosa. Their histology is that of frondlike villiform extensions of the mucosa ( Fig. 17-58 A ), covered by dysplastic, sometimes very disorderly columnar epithelium ( Fig. 17-58 B ). All degrees of dysplasia may be encountered. When invasive carcinoma occurs, there is no stalk as a buffer zone, and invasion is directly into the wall of the colon (submucosa or deeper).
Tubulovillous adenomas are typically intermediate between the tubular and villous lesions in terms of their frequency of having a stalk or being sessile, their size, and the general level of dysplasia found in such lesions. The risk of harboring in situ or invasive carcinoma generally correlates with the proportion of the lesion that is villous.
Clinical Features.
Colorectal tubular (and tubulovillous) adenomas may be asymptomatic, but many are discovered during evaluation of anemia or occult bleeding. Villous adenomas are much more frequently symptomatic than the other
Figure 17-58 A, Sessile adenoma with villous architecture. Each frond is lined by dysplastic epithelium. B, Portion of a villous frond with dysplastic columnar epithelium on the left and normal colonic columnar epithelium on the right.
patterns, and often are discovered because of overt rectal bleeding. Rarely, villous adenomas may hypersecrete copious amounts of mucoid material rich in protein and potassium, leading to either hypoproteinemia or hypokalemia. Notably, screening programs are intended to detect asymptomatic adenomas before they progress to malignancy.
The clinical impact of malignant change in an adenoma depends on the following:
• High-grade dysplasia (carcinoma in situ) has not yet acquired the ability to metastasize and is still a clinically benign lesion. • Because lymphatic channels are largely absent in the colonic mucosa, being present erratically only at the very base of the lamina propria, intramucosal carcinoma with lamina propria invasion only is regarded also as having little or no metastatic potential. • If the lesion penetrates through the muscularis mucosa into the submucosal space, the resultant invasive adenocarcinoma is a malignant tumor with metastatic potential. Nevertheless, endoscopic removal of a pedunculated adenoma is regarded as an adequate excision provided that three histologic conditions are met: (1) the adenocarcinoma is superficial and does not approach the margin of excision across the base of the stalk; (2) there is no vascular or lymphatic invasion; and (3) the carcinoma is not poorly differentiated. • Invasive adenocarcinoma arising in a sessile polyp cannot be adequately resected by polypectomy, and further surgery may be required.
• Regardless of whether carcinoma is present, the only adequate treatment for a pedunculated or sessile adenoma is complete resection. If adenomatous epithelium remains behind the patient still has a premalignant lesion or may even be harboring invasive carcinoma in the residual lesion.
Familial Syndromes
Familial polyposis syndromes are uncommon autosomal dominant disorders. Their importance lies in the propensity for malignant transformation and in the insights that they have provided in unraveling the molecular basis of colorectal
862
cancer. Peutz-Jeghers syndrome, described earlier, is characterized by hamartomatous polyps and a modestly increased risk of cancer, frequently in extragastrointestinal sites. Juvenile polyposis syndrome and Cowden syndrome have also been mentioned earlier. FAP exhibits innumerable adenomatous polyps and has a frequency of progression to colon adenocarcinoma approaching 100%. Hereditary nonpolyposis colorectal cancer syndrome (HNPCC or Lynch syndrome) is characterized by the development of colorectal carcinoma, endometrial carcinoma, and carcinoma of the small intestine, ureter, or renal pelvis. As we discuss later, many of the molecular events underlying these syndromes have been identified.
Familial Adenomatous Polyposis (FAP) Syndrome.
FAP is the archetype of the adenomatous polyposis syndromes. It is caused by mutations of the adenomatous polyposis coli (APC) gene on chromosome 5q21 ( Chapter 7 ).[81] The same gene mutations cause a broad spectrum of clinical manifestations. Based on the clinical presentation, FAP can be further classified as classic FAP, attenuated FAP, Gardner syndrome, and Turcot syndrome.
In the classic FAP syndrome, patients typically develop 500 to 2500 colonic adenomas that carpet the mucosal surface ( Fig. 17-59 ). Occasionally as few as 150 polyps are present; a minimum of 100 polyps is necessary for a diagnosis of classic FAP. Multiple adenomas may also be present elsewhere in the alimentary tract, including the region of the ampulla of Vater. Histologically, the vast majority of polyps are tubular adenomas; occasional polyps may have villous features. Some patients already have cancer of the colon or rectum at the time of diagnosis. Cancer-prevention measures include early detection and prophylactic colectomy in siblings and first-degree relatives at risk. In addition to colonic polyps, FAP patients can have polyps in the stomach (adenomas or fundic gland polyps) and small intestine (especially around the ampulla of Vater).
In attenuated FAP, patients tend to develop fewer polyps (average, 30), and most of the polyps are located in the proximal
Figure 17-59 Familial adenomatous polyposis in an 18-year-old woman. The mucosal surface is carpeted by innumerable polypoid adenomas.
colon. The lifetime risk of cancer development is usually around 50%.
Patients with Gardner syndrome exhibit intestinal polyps identical to those in classic FAP, combined with multiple osteomas (particularly of the mandible, skull, and long bones), epidermal cysts, and fibromatosis. Less frequent are abnormalities of dentition, such as unerupted and supernumerary teeth, and a higher frequency of duodenal and thyroid cancer.
Turcot syndrome is a rare clinical syndrome marked by the combination of adenomatous colonic polyposis and tumors of the central nervous system. Two thirds of patients with Turcot syndrome have APC gene mutations and develop brain medulloblastomas. The remaining one third have mutations in one of the genes associated with HNPCC and develop brain glioblastomas.
Hereditary Nonpolyposis Colorectal Cancer (HNPCC) Syndrome.
HNPCC is an autosomal dominant familial syndrome (extensively described by Henry Lynch, hence the alternative name of Lynch syndrome).[82] It is characterized by an increased risk of colorectal cancer and extraintestinal cancer, particularly of the endometrium. Adenomas occur in low numbers and considerably earlier than in the general adult population. However, the colonic malignancies that develop in this syndrome often are multiple and are not usually associated with pre-existing adenomas. The hallmark of HNPCC is mutations in DNA repair genes, leading to microsatellite instability, as discussed in Chapter 7 .
Most colorectal carcinomas occur sporadically in the absence of well-defined familial syndromes. Like the majority of cancers in other organs, there are conditions associated with risk of tumor development. Regardless of the inciting event, a well-described set of genetic alterations occurs that ultimately leads to colorectal malignancy. The model proposed by Fearon and Vogelstein is widely accepted as the prototypical sequence for colorectal cancer development.[83] The pathologic basis for this model is the adenoma-carcinoma sequence, which has been documented by these observations:
• Populations that have a high prevalence of adenomas have a high prevalence of colorectal cancer, and vice versa. • The distribution of adenomas within the colorectum is more or less comparable to that of colorectal cancer. • The peak incidence of adenomatous polyps antedates by some years the peak for colorectal cancer. • When invasive carcinoma is identified at an early stage, surrounding adenomatous tissue is often present • The risk of cancer is directly related to the number of adenomas, and hence the virtual certainty of cancer in patients with familial polyposis syndromes. • Programs that assiduously follow patients for the development of adenomas and remove all that are suspicious reduce the incidence of colorectal cancer.
The occurrence of colorectal carcinoma without evidence of adenomatous precursors suggests that some dysplastic lesions can degenerate into malignancy without passing through a polypoid stage.
Molecular Carcinogenesis.
Study of colorectal carcinogenesis has provided fundamental insights into the general
863
mechanisms of cancer evolution. Many of these principles were discussed in Chapter 7 . Here we will discuss concepts specifically pertinent to carcinogenesis in the colon.
It is now believed that there are two pathogenetically distinct pathways for the development of colon cancer,[84] both of which involve the stepwise accumulation of multiple mutations. However, the genes involved and the mechanisms by which the mutations accumulate are different.
The first pathway, sometimes called the APC/β-caterin pathway, is characterized by chromosomal instability that results in stepwise accumulation of mutations in a series of oncogenes and tumor suppressor genes. The molecular evolution of colon cancer along this pathway occurs through a series of morphologically identifiable stages. Initially, there is localized colon epithelial proliferation. This is followed by the formation of small adenomas that progressively enlarge, become more dysplastic, and ultimately develop
into invasive cancers. This is referred to as the adenoma-carcinoma sequence ( Fig. 17-60 ). The genetic correlates of this pathway are as follows:
Loss of Adenomatous Polyposis Coli (APC) Gene.
The APC gene has been mapped to 5q21. Its mutation is the genetic basis for FAP syndrome and fulfills the "first hit" concept advanced by Knudson in the 1970s.[85] Loss of this gene is believed to be the earliest event in the formation of adenomas. This dual-function tumor suppressor gene encodes a protein that binds to microtubule bundles and promotes cell migration and adhesion. APC also acts as a gatekeeper protein, as it regulates levels of β-catenin, an important mediator of the Wnt/β-catenin signaling pathway (see Fig. 7-38 , Chapter 7). This signaling pathway plays a critical role in the normal intestinal epithelial development. It is also involved in development of colorectal carcinomas. More than 80% of colorectal carcinomas have inactivated APC, and 50% of cancers
Figure 17-60 Schematic of the morphologic and molecular changes in the adenoma-carcinoma sequence. It is postulated that loss of one normal copy of the tumor suppressor gatekeeper gene APC occurs early. Indeed, individuals may be born with one mutant allele of APC, rendering them extremely likely to develop colon cancer. This is the "first hit," according to Knudson's hypothesis. The loss of the normal copy of the APC gene follows ("second hit"). Mutations of the oncogene K-RAS seem to occur next. Additional mutations or losses of heterozygosity inactivate the tumor suppressor gene p53 (on chromosome 17p) and SMAD2 and SMAD4 on chromosome 18q, leading finally to the emergence of carcinoma, in which additional mutations occur. It is important to note that while there seems to be a temporal sequence of changes, as shown, the accumulation of mutations, rather than their occurrence in a specific order, is more important.
without APC mutations have β-catenin mutations. β-catenin is a member of the cadherin-based cell adhesive complex, which also acts as a transcription factor if the protein is translocated to the nucleus. When it is not bound to E-cadherin and participating in cell-to-cell adhesion, a cytoplasmic degradation complex (consisting of APC, Axin, GSK-3β, and β-catenin) leads to β-catenin phosphorylation and degradation. In the setting of APC
mutations (loss of normal function), β-catenin accumulates in the cytoplasm and is translocated to the nucleus to bind to a family of transcription factors called T-cell factor or lymphoid enhancer factor (TCF or LEF) proteins. The TCF contributes a DNA-binding domain and β-catenin contributes a transactivation domain. Genes activated by the β-catenin-TCF complex are thought to include those regulating cell proliferation and apoptosis, such as c-MYC and CYCLIN D1. Hence, normal APC function promotes cell adhesion and regulates cell proliferation; absence of APC function leads to decreased cell adhesion and increased cellular proliferation.
Reported mutations in the APC gene include missense mutations and deletions, resulting in synthesis of truncated APC proteins. Mutant β-catenin loses binding affinity to GSK-3β, the kinase that phosphorylates and degrades β-catenin, in normal cells. APC mutations are present in 80% of sporadic carcinomas.
Mutation of K-RAS.
The K-RAS gene ( Chapter 7 ) is the most frequently observed activated oncogene in adenomas and colon cancers. K-RAS plays a role in intracellular signal transduction and is mutated in fewer than 10% of adenomas less than 1 cm in size, in about 50% of adenomas larger than 1 cm, and in approximately 50% of carcinomas.
Loss of SMADs.
A common allelic loss in colon cancer is on 18q21. Initially, DCC (deleted in colon cancer), was thought to be the suppressor gene involved in colorectal
864
cancer, which was located in this region. However, the role of DCC in colorectal carcinogenesis has been questioned, since mutant mice lacking both alleles of DCC show no abnormalities.[88] SMAD2 and SMAD4 ( Chapter 7 ), involved in TGF-β signaling, are located on 18q21. Lack of SMAD4 increases gastrointestinal tumorigenesis.[89]
Loss of p53.
Losses at chromosome 17p have been found in 70% to 80% of colon cancers, yet comparable losses are infrequent in adenomas. These chromosomal deletions affect the p53 gene, suggesting that mutations in p53 occur late in colon carcinogenesis. The critical role of p53 in cell-cycle regulation is discussed in Chapter 7 .
Activation of Telomerase.
Telomeres plays a role in stabilizing the chromosome. They shorten with each cell division until cell senescence develops ( Chapter 1 ). Telomerase is a ribonucleoprotein complex with telomeric reverse transcriptase (TERT) as the catalytic subunit. Telomerase activity is required to maintain telomere stability and hence cell immortality, a
prerequisite for all cancer cells ( Chapter 7 ). Most adenomas lack telomerase activity, but a majority of cancers in humans, including colorectal carcinoma, have increased telomerase activity.[90]
Although the sequence of events outlined above is common, it should be emphasized that the accumulation of mutations is more important than their occurrence in a specific order.
Microsatellite Instability Pathway.
The second pathway is characterized by genetic lesions in DNA mismatch repair genes ( Chapter 7 ). It is involved in 10% to 15% of sporadic cases and in the HNPCC syndrome. As in the APC/β-catenin schema, there is accumulation of mutations, but the involved genes are different, and, unlike in the adenoma-carcinoma sequence, there are no clearly identifiable morphologic correlates. Defective DNA repair caused by inactivation of DNA mismatch repair genes is the fundamental and the most likely initiating event in colorectal cancers that travel this road. Inherited mutations (germ-line mutations) in any of five genes that are involved in DNA repair are responsible for the familial syndrome of HNPCC. These human mismatch repair genes, hMSH2 (chromosome 2p22), hMLH1 (chromosome 3p21), MSH6 (chromosome 2p21), hPMS1 (chromosome 2q31-33), and hPMS2 (chromosome 7p22), are involved in genetic "proofreading" during DNA replication and have earned the moniker of caretaker genes ( Chapter 7 ).[86] The majority of the mutations (90%) involve MSH2 and MLH1. Mutations in the mismatch repair genes cause alteration of microsatellites, leading to microsatellite instability. Microsatellites are fragments of repeat sequences in the human genome, which contains approximately 50,000 to 100,000 microsatellites. These sequences are prone to misalignment during DNA replication. In normal cells, the misalignment is repaired by the caretaker genes. Patients with HNPCC inherit one mutant DNA repair gene ("the first hit") and one normal allele. For unclear reasons, cells in some organs (colon, stomach, endometrium) are susceptible to a second, somatic mutation ("the second hit" of the Knudson hypothesis), which inactivates the normal allele (loss of heterozygosity or LOH). With homozygous loss of mismatch repair genes, mutation rates are up to 1000 times higher than normal, and most of the HNPCC tumors show microsatellite instability. About 10% to 15% of sporadic colon cancers have mutations in similar DNA repair genes, Most microsatellite sequences are in noncoding regions of the genes, and, hence, mutations in these genes are probably harmless. However, some microsatellite sequences are located in the coding or promoter region of genes involved in regulation of cell growth. Such genes include type II TGF-β receptor and BAX. TGF-β signaling inhibits the growth of colonic epithelial cells, and the BAX gene causes apoptosis. Loss of mismatch repair leads to the accumulation of mutations in these and other growth-regulating genes, culminating in the emergence of colorectal carcinomas.
Although there is no readily identifiable adenoma-carcinoma sequence that typifies tumors arising from defects in mismatch repair, it has been noted that some of the so-called hyperplastic polyps seen on the right side of the colon display microsatellite instability and may well be precancerous.[87] Fully developed tumors that arise via the mismatch repair pathway do show some distinctive morphologic features, including
proximal colonic location, mucinous histology, and infiltration by lymphocytes. In general, these tumors have better prognosis than stage-matched tumors that arise by the APC pathway.
Colorectal Carcinoma
Virtually 98% of all cancers in the large intestine are adenocarcinomas. They represent one of the prime challenges to the medical profession, because they usually arise in polyps and produce symptoms relatively early and at a stage generally curable by resection. Yet, there are an estimated 148,300 new cases per year and about 56,600 deaths, accounting for 10% of all cancer-related deaths in the United States.[91]
Epidemiology, Etiology, and Pathogenesis.
The peak incidence for colorectal carcinoma is between ages 60 and 79. Fewer than 20% of cases occur before age 50. When colorectal carcinoma is found in a young person, pre-existing ulcerative colitis or one of the polyposis syndromes must be suspected. With lesions in the rectum, the male-to-female ratio is 1.2:1; for more proximal tumors there is no gender difference. Colorectal carcinoma has a worldwide distribution, with the highest death rates in the United States, Australia, New Zealand, and Eastern European countries. Its incidence is substantially lower—up to 10-fold—in Mexico, South America, and Africa. Environmental factors, particularly dietary practices, are implicated in these striking geographic contrasts in incidence. Japanese and Polish families that have migrated from their low-risk areas to the United States have acquired, over the course of 20 years, the rate prevailing in the new environment. Both groups, for the most part, adopted the common dietary practices of the U.S. population. Other studies implicate obesity and physical inactivity as risk factors for colon cancer.[92] [93]
The dietary factors receiving the most attention as predisposing to a higher incidence of cancer are (1) excess dietary caloric intake relative to requirements, (2) a low content of unabsorbable vegetable fiber, (3) a corresponding high content of refined carbohydrates, (4) intake of red meat, and (5) decreased intake of protective micronutrients. It is theorized that reduced fiber content leads to decreased stool bulk, increased fecal transit time in the bowel, and an altered bacterial flora of the intestine. Potentially toxic oxidative byproducts of carbohydrate degradation by bacteria are therefore present in higher concentrations in the stools and are held in contact with the colonic mucosa for longer periods of time. Moreover, high cholesterol intake in red meat enhances the
865
synthesis of bile acids by the liver, which in turn may be converted into potential carcinogens by intestinal bacteria. Refined diets also contain less of vitamins A, C, and E, which may act as oxygen-radical scavengers ( Chapter 1 ). Intriguing as these speculations may be, the putative mechanisms of dietary effects remain unproven.
Indeed, recent studies have challenged the notion that low-fat, high-fiber diets protect against recurrence of colorectal adenomas, the precursors of colon cancer.
Several epidemiological studies suggest that use of aspirin and other NSAIDs exerts a protective effect against colon cancer. In the Nurses' Health Study, women who used four to six tablets of aspirin/day for 10 years or more had a decreased incidence of colon cancer. Two recent studies have revealed that aspirin reduces the risk of recurrent adenomas in patients with previous colorectal carcinomas or adenomas.[94] The mechanism of such chemoprevention is not fully understood, but it is likely mediated by inhibition of cyclooxygenase-2 ( Chapter 2 ). This enzyme is overexpressed in neoplastic epithelium and seems to regulate angiogenesis and apoptosis. On the basis of these findings, the U.S. Food and Drug Administration has approved the use of COX-2 inhibitors as chemopreventive agents in patients with the familial adenomatous polyposis syndrome.
Morphology.
The distribution of the cancers in the colorectum is as follows: cecum/ascending colon, 22%; transverse colon, 11%; descending colon, 6%; rectosigmoid colon, 55%; and other sites, 6%. The right-sided colon cancers tend to have greater microsatellite instability. Ninety-nine per cent of carcinomas occur singly, but when multiple carcinomas are present, they are often at widely disparate sites in the colon. While most cases occur sporadically, about 1% to 3% of colorectal carcinomas occur in patients with familial syndromes (i.e., FAP or HNPCC) or IBD.
Although all colorectal carcinomas begin as in situ lesions, they evolve into different morphologic patterns. Tumors in the proximal colon tend to grow as polypoid, exophytic masses that extend along one wall of the capacious cecum and ascending colon ( Fig. 17-61 ). Obstruction is uncommon. When carcinomas in the distal colon are discovered, they tend to be annular, encircling lesions that produce so-called napkin-ring constrictions of the bowel ( Fig. 17-62 ). The margins of the napkin ring are classically heaped up, beaded, and firm, and the midregion is ulcerated. The lumen is markedly narrowed, and the proximal bowel may be distended. Both forms of neoplasm directly penetrate the bowel wall over the course of time (probably years) and may appear as subserosal and serosal white, firm masses, frequently causing puckering of the serosal surface. Uncommonly, but particularly in association with ulcerative colitis, colorectal cancers are insidiously infiltrative and difficult to identify radiographically and macroscopically. Such lesions tend to be exceedingly aggressive, and spread at an early stage in their evolution.
Unlike the gross pathology, the microscopic characteristics of right- and left-sided colonic adenocarcinomas are similar. Differentiation may range from tall, columnar cells resembling their counterparts in adenomatous lesions, which now invade the submucosa and muscularis propria ( Fig. 17-63 ), to undifferentiated,
Figure 17-61 Carcinoma of the cecum. The fungating carcinoma projects into the lumen but has not caused obstruction.
frankly anaplastic masses. Invasive tumor incites a strong desmoplastic stromal response, leading to the characteristic firm, hard consistency of most colonic carcinomas. Many tumors produce mucin, which is secreted into the gland lumina or into the interstitium of the gut wall. Because this secretion
Figure 17-62 Carcinoma of the descending colon. This circumferential tumor has heaped-up edges and an ulcerated central portion. The arrows identify separate mucosal polyps.
866
Figure 17-63 Invasive adenocarcinoma of colon, showing malignant glands infiltrating the muscle wall.
dissects through the gut wall, it aids the extension of the malignancy and worsens the prognosis.
Certain specific features should be noted. Foci of endocrine differentiation may be found in about 10% of colorectal carcinomas. Alternatively, in some cancers the cells take on a signet-ring appearance. The small-cell undifferentiated carcinoma appears to arise from endocrine cells per se and elaborates a variety of bioactive secretory products. Some cancers, particularly in the distal colon, have foci of squamous cell differentiation and are therefore referred to as adenosquamous carcinomas. In contrast, carcinomas arising in the anorectal canal constitute a distinct subgroup of tumors, dominated by squamous cell carcinoma. Tumors associated with HNPCC tend to be poorly differentiated and rich in mucin.
Clinical Features.
Colorectal cancers remain asymptomatic for years; symptoms develop insidiously and frequently have been present for months, sometimes years, before diagnosis. Cecal and right colonic cancers are most often called to clinical attention by the appearance of fatigue, weakness, and iron-deficiency anemia. These bulky lesions bleed readily and may be discovered at an early stage, provided the colon is examined thoroughly radiographically and during colonoscopy. Left-sided lesions come to attention by producing occult bleeding, changes in bowel habit, or crampy left lower quadrant discomfort. In theory, the chance for early discovery and successful removal should be greater for patients with lesions on the left side, because these patients usually have prominent disturbances in bowel function such as melena, diarrhea, and constipation. However, cancers of the rectum and sigmoid tend to be more infiltrative at the time of diagnosis than proximal lesions, and therefore have a somewhat poorer prognosis. It is a clinical maxim that iron-deficiency anemia in an older male means gastrointestinal cancer until proven otherwise. In females the situation is less clear, since menstrual losses, multiple pregnancies, or abnormal uterine bleeding may underlie such an anemia. Systemic manifestations such as weakness, malaise, and weight loss are ominous, in that they usually signify more extensive disease.
All colorectal tumors spread by direct extension into adjacent structures and by metastasis through the lymphatics and blood vessels. In order of preference, the favored sites of metastatic spread are the regional lymph nodes, liver, lungs, and bones, followed by many other sites, including the serosal membrane of the peritoneal cavity, brain, and others. In general, the disease has spread beyond the range of curative surgery in 25% to 30% of patients. Anal region carcinomas are locally invasive and metastasize to regional lymph nodes and distant sites.
The single most important prognostic indicator of colorectal carcinoma is the extent of the tumor at the time of diagnosis, the so-called stage. A staging system formerly widely
used is that described by Aster and Coller in 1954, which represents a modification of classifications proposed by Dukes and Kirklin. Currently, the system most widely used is the tumor-nodes-metastasis (TNM) classification and staging system from the American Joint Commission on Cancer ( Table 17-14 ). The criteria for pathologic staging are shown in Figure 17-64 . Regardless of the system used, survival at 1, 5, and 10 years is strongly correlated with the stage of disease at the time of surgical resection. Staging can be accurately applied only after the extent of spread is determined by surgical exploration and anatomic examination.
The overriding challenge is to discover these neoplasms when curative resection is possible, preferably when they are still adenomatous polyps. Indeed, each death from colonic cancer in the United States must be viewed as a preventable tragedy, but progress has been relatively slow in coming.
Carcinoid Tumors
The first carcinoid tumor was identified in the ileum by Lubarsch more than 100 years ago. The term carcinoid was used by Oberndorfer in 1907 because the tumor was described as a carcinoma-like lesion but with a much more indolent clinical course. Carcinoid tumor is derived from resident endocrine cells, with the gastrointestinal tract and lung as the predominant sites of occurrence.
TABLE 17-14 -- TNM Classification of Carcinoma of the Colon and Rectum
Tumor Stage Histologic Features of the Neoplasm
Tis Carcinoma in situ (high-grade dysplasia) or intramucosal carcinoma (lamina propria invasion)
T1 Tumor invades submucosa
T2 Extending into the muscularis propria but not penetrating through it
T3 Penetrating through the muscularis propria into subserosa
T4 Tumor directly invades other organs or structures
Figure 17-64 Pathologic staging of colorectal cancer. Staging is based on the depth of tumor invasion.
Mucosal endocrine cells generate bioactive compounds, particularly peptide and nonpeptide hormones, and play a major role in coordinated gut function. Although they are derived from epithelial stem cells in the mucosal crypts, they are designated endocrine cells because of their endocrine and paracrine function and their resemblance to endocrine cells elsewhere, as in the pancreas. Mucosal endocrine cells are abundant in other organs, including the lungs, but the great preponderance of carcinoid tumors arising from these cells are in the gut. A scattering of carcinoid tumors arises in the pancreas or peripancreatic tissue, lungs, biliary tree, and even liver. The peak incidence of these neoplasms is in the sixth decade, but they may appear at any age. They comprise less than 2% of colorectal malignancies but almost half of small intestinal malignant tumors.
The classification of carcinoid tumors is still controversial. The prevailing view is that carcinoid tumor may represent a well-differentiated end of the spectrum is the small cell carcinoma. Carcinoid tumors may be confined to the mucosa and submucosa or may be malignant in behavior with deep invasion and metastatic spread to regional lymph nodes and the liver. Intriguingly, there is no reliable histologic difference between seemingly benign and obviously malignant carcinoid tumors. While there are no reliable molecular markers to
Figure 17-65 Carcinoid tumor. A, Multiple protruding tumors are present at the ileocecal junction. B, The tumor cells exhibit a monotonous morphology, with a delicate intervening fibrovascular stroma. C, Electron micrograph showing dense core bodies in the cytoplasm.
predict tumor behavior, the tendency for aggressive behavior correlates with the site of origin, the depth of local penetration, the size of the tumor, and the histologic features of necrosis and mitosis. Hence, it is possible to establish a reasonable clinical assessment of these tumors. Appendiceal and rectal carcinoids infrequently metastasize, even though they may show extensive local spread. By contrast, 90% of ileal, gastric, and colonic carcinoids that have penetrated halfway through the muscle wall have spread to lymph nodes and distant sites such as the liver at the time of diagnosis. This is especially true for tumors greater than 2 cm in diameter.
Morphology.
The appendix is the most common site of gut carcinoid tumors, followed by the small intestine (primarily ileum), rectum, stomach, and colon. However, the rectal tumors may represent up to half of tumors that come to clinical attention. Those that arise in the stomach and ileum are frequently multicentric, but the remainder tend to be solitary lesions. In the appendix they appear as bulbous swellings of the tip, which frequently obliterate the lumen. Elsewhere in the gut, they appear as intramural or submucosal masses that create small, polypoid or plateau-like elevations rarely more than 3 cm in diameter ( Fig. 17-65 A ). The overlying mucosa may be intact or ulcerated, and the tumors may permeate the bowel wall to invade the mesentery. A characteristic feature is a solid, yellow-tan appearance on transection. The tumors are exceedingly firm owing to striking desmoplasia, and when these fibrosing lesions penetrate the mesentery of the small bowel they may cause angulation or kinking sufficient to cause obstruction. When present, visceral metastases are usually small, dispersed nodules and rarely achieve the size seen with the primary lesions. Notably, rectal and appendiceal carcinoids almost never metastasize.
Histologically, the neoplastic cells may form discrete islands, trabeculae, stands, glands, or undifferentiated sheets. Whatever their organization the tumor cells are monotonously similar, having a scant, pink granular cytoplasm and a round to oval stippled nucleus. In most tumors there is minimal variation in cell and nuclear size and mitoses are infrequent
or absent ( Fig. 17-65 B ). In unusual cases there may be more significant anaplasia and sometimes mucin
868
secretion within the cells and gland formations. Rarely, tumors arise resembling small-cell carcinomas of the lung ( Chapter 15 ) or contain abundant psammoma bodies similar to those seen in thyroid carcinomas ( Chapter 24 ). By electron microscopy ( Fig. 17-65 C ), the cells in most tumors contain membrane-bound secretory granules with osmophilic centers (dense-core granules) in the cytoplasm. Most carcinoids contain chromogranin A, synaptophysin, and neuron-specific enolase. Specific hormonal peptides may occasionally be identified by immunocytochemical techniques.
Clinical Features.
Gastrointestinal carcinoids only rarely produce local symptoms, which are caused by angulation or obstruction of the small intestine. Many (especially rectal and appendiceal) are asymptomatic and are found incidentally. However, the secretory products of some carcinoids may produce a variety of syndromes or endocrinopathies, depending on their anatomic site ( Chapter 7 ). Gastric, peripancreatic, and pancreatic carcinoids can release their products directly into the systemic circulation, and can produce, for example, the Zollinger-Ellison syndrome related to excess elaboration of gastrin, Cushing syndrome associated with corticotropin secretion, and hyperinsulinism. In some instances, these tumors may be smaller than 1 cm and extremely difficult to find, even during surgical exploration.
Some neoplasms are associated with a distinctive carcinoid syndrome ( Table 17-15 ). The syndrome occurs in about 1% of all patients with carcinoids and in 20% of those with widespread metastases. Uncertainties remain about the precise origin of the carcinoid syndrome, but most manifestations are thought to arise from excess elaboration of serotonin (5-hydroxytryptamine, 5-HT). Elevated levels of 5-HT and its metabolite, 5-hydroxyindoleacetic acid (5-HIAA), are present in the blood and urine of most patients with the classic syndrome. 5-HT produced by gastrointestinal carcinoid tumors is degraded to functionally inactive 5-HIAA in the liver. Thus, hepatic metastases are usually present for the development of the syndrome from gastrointestinal carcinoids, because under these conditions, a sufficient amount of substances produced by the tumors can reach the systemic circulation without metabolic degradation by the liver. Not surprisingly, hepatic
TABLE 17-15 -- Clinical Features of the Carcinoid Syndrome
• Vasomotor distubances
Cutaneous flushes and apparent cyanosis (most patients)
Couth, wheezing, dyspnea (about one third of patients)
• Hepatomegaly
Nodular liver owing to hepatic metastases (some patients)
• Systemic fibrosis (some patients)
Cardiac involvement
Pulmonic and tricuspid valve thickening and stenosis
Endocardial fibrosis, principally in the right ventricle
(Bronchial carcinoids affect the left side)
Retroperitoneal and pelvic fibrosis
Collagenous pleural and intimal aortic plaques
metastases are usually not required for the production of a carcinoid syndrome by extraintestinal carcinoids (such as those arising in the lungs or ovaries), because active substances produced by the tumors are directly released into the systemic circulation. Other secretory products of carcinoids such as histamine, bradykinin, kallikrein, and prostaglandins may also contribute to the manifestations of the carcinoid syndrome.
The overall five-year survival rate for carcinoids (excluding appendiceal) is approximately 90%. Even with small-bowel tumors with hepatic metastases, it is better than 50%. However, widespread disease will usually cause death.
GASTROINTESTINAL LYMPHOMA
Any segment of the gastrointestinal tract may be secondarily involved by systemic dissemination of non-Hodgkin lymphomas. However, up to 40% of lymphomas arise in sites other than lymph nodes, and the gut is the most common location. Conversely, about 1% to 4% of all gastrointestinal malignancies are lymphomas. By definition, primary gastrointestinal lymphomas exhibit no evidence of liver, spleen, mediastinal lymph node, or bone marrow involvement at the time of diagnosis—regional lymph node involvement may be present. Primary gastrointestinal lymphomas usually arise as sporadic neoplasms but also occur more frequently in certain patient populations: (1) Chronic gastritis caused by H. pylori, (2) chronic spruelike syndromes, (3) natives of the Mediterranean region, (4) congenital immunodeficiency states, (5) infection with human immunodeficiency virus, and (6) following organ transplantation with immunosuppression.
Intestinal tract lymphomas can be classified into B-cell and T-cell lymphomas. The B-cell lymphoma can be subdivided into MALT lymphoma, immunoproliferative small-intestinal disease (IPSID), and Burkitt lymphoma.
1. MALT lymphoma is a sporadic lymphoma, which arises from the B cells of MALT (mucosa-associated lymphoid tissue, described under gastric lymphoma). This type of lymphoma is the most common form in the Western hemisphere. The biologic features of these lymphomas are different from node-based lymphomas in that (1) many behave as focal tumors in their early stages and are amenable to surgical resection; (2) relapse may occur exclusively in the gastrointestinal tract; (3) genotypic changes are different than those observed in nodal lymphomas: the t(11;18) translocation is relatively common in MALT lymphoma; and (4) the cells are usually CD5- and CD10-negative. This type of gastrointestinal lymphoma usually affects adults, has no gender predilection, and may arise anywhere in the gut: stomach (55% to 60% of cases); small intestine (25% to 30%), proximal colon (10% to 15%), and distal colon (up to 10%). The appendix and esophagus are only rarely involved.The pathogenesis of these lymphomas is under intense scrutiny. The concept has been advanced that lymphomas of MALT origin arise in the setting of mucosal lymphoid activation and that these lymphomas are the malignant counterparts of hypermutated, postgerminal-center memory B cells. As discussed earlier, Helicobacter-associated chronic gastritis, in particular, has been proposed as a driving force for the development of gastric MALT lymphoma, the result of antigen-driven somatic mutation of
869
gastric lymphoid tissue. However, the etiologic factors for intestinal lymphoma are still unknown, although history of IBD appears to increase the risk.
2. IPSID is also referred to as Mediterranean lymphoma. It is an unusual intestinal B-cell lymphoma arising in patients with Mediterranean ancestry, having a background of chronic diffuse mucosal plasmacytosis. The plasma cells synthesize an abnormal Igα heavy chain, in which the variable portion has been deleted. A high proportion of patients have malabsorption and weight loss preceding the development of the lymphoma. The diagnosis is made most commonly in children and young adults, and both sexes appear to be affected equally. The exact etiology of this type of lymphoma is not known, although infection appears to play a role.[95]
3. The intestinal T-cell lymphoma is usually associated with a long-standing malabsorption syndrome (such as celiac disease) that may not constitute a true gluten-sensitive enteropathy. This lymphoma occurs in relatively young individuals (age 30 to 40), often following a 10- to 20-year history of symptomatic malabsorption. Alternatively, a diffuse enteropathy with malabsorption may accompany the development of a lymphoma. Intestinal T-cell
lymphoma arises most often in the proximal small bowel, and its overall prognosis is poor (reported 11% five-year survival rate).
Morphology.
Gastrointestinal lymphomas can assume a variety of gross appearances. Since all the gut lymphoid tissue is mucosal and submucosal, early lesions appear as plaque-like expansions of the mucosa and submucosa. Diffusely infiltrating lesions may produce full-thickness mural thickening, with effacement of the overlying mucosal folds and focal ulceration. Others may be polypoid, protruding into the lumen, or form large, fungating, ulcerated masses. Tumor infiltration into the muscularis propria splays the muscle fibers, gradually destroying them. Because of this feature, advanced lesions frequently cause motility problems with secondary obstruction. Large tumors sometimes perforate because of lack of stromal support; reduction in tumor bulk during chemotherapy also may lead to perforation.
In the earliest histologic lesions, atypical lymphoid cells may be seen infiltrating the mucosa, with effacement and loss of glands and massive expansion of lymphoid tissue. Extreme numbers of atypical lymphoid cells may populate the superficial or glandular epithelium (lymphoepithelial lesion). With established lymphomas, the mucosa, submucosa, and even muscle wall are replaced by a monotonous infiltrate of malignant cells, consisting of a mixture of small lymphocytes and immunoblasts in varying proportions. Lymphoid follicles are occasionally formed. Most gut lymphomas are of B-cell type (over 95%) and are evenly split between low- and high-grade tumors. The small fraction of T-cell lymphomas occurring in the intestine are commonly high-grade lesions.
Clinical Features.
With the exception of T-cell lymphomas, primary gastrointestinal lymphomas generally have a better prognosis than do those arising in other sites. Ten-year survival for patients with localized mucosal or submucosal disease approaches 85%. Early discovery is key to survival; thus, gastric lymphomas generally have a better outcome than those of the small or large bowel. In general, the depth of local invasion, size of the tumor, the histologic grade of the tumor, and extension into adjacent viscera are important determinants of prognosis.
MESENCHYMAL TUMORS
Mesenchymal tumors may occur anywhere in the alimentary tract. The nomenclature for these tumors is largely based on the tumor cell phenotypes. Lipomas show a propensity for the submucosa of the small and large intestines, and lipomatous hypertrophy may occur in the ileocecal valve. A variety of spindle-cell lesions may arise in the muscle wall of any gut segment. The great majority of these tumors are of smooth muscle origin, and hence can be termed leiomyomas and leiomyosarcomas. Gastrointestinal stromal tumors (GISTs), are now considered to be a distinctive tumor type, characterized by c-KIT immunoreactivity, as discussed earlier (see "Gastric Tumors"). The small intestine is the second most common location for this tumor, (the stomach being the most common).
Both benign and malignant versions of GIST may occur at any age and in either sex. Vascular tumors such as Kaposi sarcomas are considered elsewhere (see Chapter 11 ).
Morphology.
Lipomas are usually well-demarcated, firm nodules (almost always less than 4 cm in diameter) arising within the submucosa or muscularis propria. The overlying mucosa is stretched and attenuated. Rarely, they grow to larger size and produce hemispheric elevation of the mucosa with ulceration over the dome of the tumor. Malignant stromal tumors (primarily leiomyosarcoma) tend to produce large, bulky, intramural masses that eventually fungate and ulcerate into the lumen or project subserosally into the abdominal space. Histologically, lipomas, leiomyomas, and leiomyosarcomas resemble their counterparts encountered elsewhere ( Chapter 26 ). In the case of the stromal tumors (e.g., leiomyomas and leiomyosarcomas), large size and a high mitotic rate are correlated with an aggressive course.
Clinical Features.
Most mesenchymal tumors are asymptomatic. In the stomach, larger lesions (benign or malignant) may produce symptoms resembling those of peptic ulcer, particularly bleeding that is sometimes massive. Intestinal lesions may present with bleeding, and for the small intestine, rare obstruction or intussusception. Benign lesions are easily resectable. Surgical removal is usually possible for the malignant lesions as well, since they tend to grow as cohesive masses. Five-year survival rate for leiomyosarcoma, for example, is 50% to 60%. Metastases, however, are present in about one third of cases.
TUMORS OF THE ANAL CANAL
The anal canal is the terminal portion of the large intestine. It is divided into three zones: the upper (covered with rectal mucosa), the middle (partially covered with a transitional mucosa), and the lower (covered by stratified squamous mucosa). The tumors located in this anatomic location are designated as carcinoma of the anal canal. Patterns of differentiation
870
include a basaloid pattern, squamous cell carcinoma, and adenocarcinoma.
Anal canal carcinoma with basaloid differentiation is a tumor populated by immature proliferative cells derived from the basal layer of a stratified squamous epithelium. These tumors may occur sporadically and be uniform in their histologic features. Alternatively, basaloid differentiation may be a component of a tumor that exhibits more genuine squamous cell differentiation and/or the mucin vacuole-containing features of adenocarcinoma. All such tumors remain classified as anal canal carcinoma.
Pure squamous cell carcinomas of the anal canal are closely associated with chronic HPV infection.[96] Some rare cases are also related to immunosuppression, as encountered in renal transplantation and in AIDS patients. As with the genital tract, chronic HPV infection of the anal canal often causes precursor lesions such as condyloma acuminatum, squamous epithelium dysplasia, and carcinoma in situ.
Pure adenocarcinoma of the anal canal is often the extension of rectal adenocarcinoma. Rarely, other tumors may arise from the anal canal, notably Paget disease, small-cell carcinoma, and melanoma.
Appendix
Normal
The appendix is an underdeveloped residuum of the otherwise voluminous cecum. The adult appendix averages 6 to 7 cm in length, is partially anchored by a mesenteric extension from the adjacent ileum, and has no known function. The appendix has the same four layers as the remainder of the gut and possesses a colonic-type mucosa. A distinguishing feature of this organ is the extremely rich lymphoid tissue of the mucosa and submucosa, which in young individuals forms an entire layer of germinal follicles and lymphoid pulp. This lymphoid tissue undergoes progressive atrophy during life to the point of complete disappearance in advanced age. In the elderly the appendix, particularly the distal portion, sometimes undergoes fibrous obliteration.
Pathology
Diseases of the appendix loom large in surgical practice; appendicitis is the most common acute abdominal condition the surgeon is called on to treat. Appendicitis is one of the best-known medical entities and yet may be one of the most difficult diagnostic problems to confront the emergency physician. A differential diagnosis must include virtually every acute process that can occur within the abdominal cavity, as well as some emergent conditions affecting organs of the thorax.
Acute Appendicitis
Inflammation in the right lower quadrant was considered a nonsurgical disease of the cecum (typhlitis or perityphlitis) until Fitz recognized acute appendicitis as a distinct entity in 1886. Appendiceal inflammation is associated with obstruction in 50% to 80% of cases, usually in the form of a fecalith and, less commonly, a gallstone, tumor, or ball of worms (oxyuriasis vermicularis). Continued secretion of mucinous fluid in the
obstructed viscus presumably leads to a progressive increase in intraluminal pressure sufficient to cause eventual collapse of the draining veins. Ischemic injury then favors bacterial proliferation with additional inflammatory edema and exudation, further embarrassing the blood supply. Nevertheless, a significant minority of inflamed appendices have no demonstrable luminal obstruction, and the pathogenesis of the inflammation remains unknown.
Morphology.
At the earliest stages, only a scant neutrophilic exudate may be found throughout the mucosa, submucosa, and muscularis propria. Subserosal vessels are congested, and often there is a modest perivascular neutrophilic infiltrate. The inflammatory reaction transforms the normal glistening serosa into a dull, granular, red membrane; this transformation signifies early acute appendicitis for the operating surgeon. At a later stage, a prominent neutrophilic exudate generates a fibrinopurulent reaction over the serosa ( Fig. 17-66 ). As the inflammatory process worsens, there is abscess formation within the wall, along with ulcerations and foci of suppurative necrosis in the mucosa. This state constitutes acute suppurative appendicitis. Further appendiceal compromise leads to large areas of hemorrhagic green ulceration of the mucosa and green-black gangrenous necrosis through the wall, extending to the serosa, creating acute gangrenous appendicitis, which is quickly followed by rupture and suppurative peritonitis.
The histologic criterion for the diagnosis of acute appendicitis is neutrophilic infiltration of the muscularis propria. Usually, neutrophils and ulcerations are also present within the mucosa. Since drainage of an exudate into the appendix from alimentary tract infection may also induce a mucosal neutrophilic infiltrate, evidence of muscular wall inflammation is requisite for the diagnosis.
Figure 17-66 Acute appendicitis. The inflamed appendix shown below is red, swollen, and covered with a fibrinous exudate. For comparison, a normal appendix is shown above.
Clinical Features.
Acute appendicitis is mainly a disease of adolescents and young adults, but it may occur in any age group and affects males slightly more often than females. The lifetime risk for appendicitis is 7%. Classically, acute appendicitis produces the following manifestations, in the sequence given: (1) pain, at first periumbilical but then localizing to the right lower quadrant; (2) nausea and/or vomiting; (3) abdominal tenderness, particularly in the region of the appendix; (4) mild fever; and (5) an elevation of the peripheral white blood cell count up to 15,000 to 20,000 cell/µL. Regrettably, this classic presentation is more often not present. While pain, nausea, and vomiting usually develop, tenderness may be deceptively absent or present in atypical locations. In some cases, a retrocecal appendix may generate right flank or pelvic pain, while a malrotated colon may give rise to appendicitis in the left upper quadrant. The peripheral leukocytosis may be minimal or so high as to suggest alternative diagnoses. Nonclassic presentations are encountered more often in young children and in the very elderly, populations with a host of other plausible abdominal emergencies.
There is general agreement that highly competent surgeons make false-positive diagnoses of acute appendicitis and remove normal appendices about 20% to 25% of the time. The discomfort and risks associated with an exploratory laparotomy and discovery of "no disease" are far outweighed by the morbidity and mortality (about 2%) associated with appendiceal perforation. Besides perforation, uncommon complications of appendicitis include pyelophlebitis with thrombosis of the portal venous drainage, liver abscess, and bacteremia. In instances when the appendix is normal, most often no disease of any kind is found during abdominal exploration. Definable conditions that mimic appendicitis are mesenteric lymphadenitis, usually secondary to an enterocolitis (often unrecognized) caused by Yersinia or a virus; systemic viral infection; acute salpingitis; ectopic pregnancy; mittelschmerz (pain caused by trivial pelvic bleeding at the time of ovulation); cystic fibrosis; and Meckel diverticulitis.
True chronic inflammation of the appendix is difficult to define as a pathologic entity, although occasionally granulation tissue and fibrosis associated with acute and chronic inflammation of the appendix suggest an organizing acute appendicitis. Much more frequently, recurrent acute attacks underlie a seemingly chronic condition. Since in some individuals the appendix is a mere fibrous cord from birth, it cannot be assumed that appendiceal fibrosis is the result of a previous inflammation.
Tumors of the Appendix
The most common appendiceal tumor is the carcinoid, discussed earlier. It is usually discovered incidentally at the time of surgery or examination of a resected appendix.[97] This neoplasm most frequently involves the distal tip of the appendix, where it produces a solid bulbous swelling up to 2 to 3 cm in diameter. Although intramural and transmural
extension may be evident, nodal metastases are very infrequent, and distant spread is rare. One unique type of appendiceal carcinoid tumor is goblet cell carcinoid (adenocarcinoid). Histologically, the tumor shows a typical carcinoid pattern, but with plump mucin vacuole-containing cells. The biologic behavior of the tumor is between that of typical carcinoid and adenocarcinoma. Genetic alterations have been found in both typical carcinoid tumors and goblet cell carcinoids.[98]
Conventional adenomas or non-mucin-producing adenocarcinomas of the appendix may cause a typical neoplastic enlargement of this organ. Hyperplastic polyps may occur in this location as well. Benign and malignant mesenchymal growths resemble their counterparts in other areas.
MUCOCELE AND PSEUDOMYXOMA PERITONEI
Mucocele is the macroscopic description of a dilated appendix filled with mucin. The true pathologic nature of mucocele runs the gamut from an innocuous obstructed appendix containing inspissated mucin, to a mucin-secreting adenoma (mucinous cystadenoma) and adenocarcinoma (mucinous cystadenocarcinoma). In the last instance, invasion through the appendiceal wall with intraperitoneal seeding and spread of tumor may occur.
Morphology.
All mucinous lesions are associated with appendiceal dilatation secondary to mucinous secretions. With the simple mucocele, globular enlargement of the appendix by inspissated mucus occurs, usually the result of obstruction by a fecalith or other lesion such as an inflammatory stricture. Eventually, the distention produces sufficient atrophy of the mucin-secreting mucosal cells and the secretions stop. Rarely, a focus of mucin-secreting hyperplastic epithelium appears to be the culprit. This condition is usually asymptomatic; rarely a mucocele ruptures, spilling otherwise innocuous mucus into the peritoneal cavity.
The most common mucinous neoplasm is the benign mucinous cystadenoma, which replaces the appendiceal mucosa and is histologically identical to analogous tumors in the ovary. The luminal dilation is associated with appendiceal perforation in 20% of instances, producing localized collections of mucus attached to the serosa of the appendix or lying free within the peritoneal cavity. Histologic examination of the mucus, however, reveals no malignant cells.
872
Malignant mucinous cystadenocarcinomas are one fifth as common as cystadenomas. Macroscopically they produce mucin-filled cystic dilatation of the appendix indistinguishable from that seen with benign cystadenomas. Penetration of the appendiceal wall by invasive cells and spread beyond the appendix in the form of localized or disseminated peritoneal implants, however, is frequently present ( Fig. 17-
67 ). In its fully developed state, continued cellular proliferation and mucin secretion fills the abdomen with tenacious, semisolid mucin—pseudomyxoma peritoneii. Anaplastic adenocarcinomatous cells can be found, distinguishing this process from mucinous spillage. Instances in which pseudomyxoma peritoneii is accompanied by both appendiceal and ovarian mucinous adenocarcinomas are usually ascribed to spread of an appendiceal primary lesion.
Mucoceles are generally encountered as an incidental lesion. Mucinous cystadenomas and adenocarcinomas may present with pain, attributable to distention of the viscus. Laparotomy for presumed acute appendicitis is a typical diagnostic
Figure 17-67 Mucinous cystadenocarcinoma of the appendix, with spread into the immediate periappendiceal tissues.
setting. For lesions confined to the resected specimen (appendix or more radical excision), the outlook is excellent. Pseudomyxoma peritoneii may be held in check for years by repeated debulking procedures but in most instances eventually runs its inexorable fatal course.
Peritoneum
Inflammation
Peritonitis may result from bacterial invasion or chemical irritation. The most common causes of peritonitis are as follows.
• Sterile peritonitis from mild leakage of bile or pancreatic enzymes • Perforation or rupture of the biliary system, which evokes a highly irritating peritonitis, usually complicated by bacterial superinfection • Acute hemorrhagic pancreatitis (see Chapter 19 ), with leakage of pancreatic enzymes and digestion of adipose tissue to produce fatty acids. These in turn precipitate with calcium to produce chalky white precipitates in areas of fat digestion and necrosis. Globules of free fat may be found floating in the
peritoneal fluid, and bacterial permeation of the bowel wall leads to a frank suppurative exudate after 24 to 48 hours. • Surgical procedures. The reaction to surgically introduced foreign material such as talc is usually localized and minimal, leaving residual foreign-body type granulomas and fibrous tissue. Abrasion of serosal surfaces during abdominal surgery may lead to fibrous adhesions between visceral structures. While usually asymptomatic, these occasionally are the points of internal herniation or intestinal obstruction. • Gynecologic conditions. Endometriosis may introduce irritant blood into the peritoneal cavity, and ruptured dermoid cysts may invoke an intense peritoneal granulomatous reaction.
PERITONEAL INFECTION
Bacterial peritonitis is almost invariably secondary to extension of bacteria through the wall of a hollow viscus or to rupture of a viscus. The more common disorders leading to such bacterial disseminations are appendicitis, ruptured peptic ulcer, cholecystitis, diverticulitis, strangulation of bowel, acute salpingitis, abdominal trauma,, and peritoneal dialysis. Virtually every bacterial organism has been implicated, most commonly E. coli, alpha- and beta-hemolytic streptococci, Staphylococcus aureus, enterococci, gram-negative rods, and Clostridium perfringens. The last organism is a frequent inhabitant of the gut and contributor to peritonitis but rarely causes gas gangrene in the abdominal cavity. Gynecologic infections may introduce gonococcus and Chlamydia.
Spontaneous bacterial peritonitis may develop in the absence of an obvious source of contamination. It is an uncommon disorder seen most often in children, particularly those with the nephrotic syndrome. Among adults, 10% of cirrhotic patients with ascites develop spontaneous bacterial peritonitis during the course of their illness. The usual causal agents of the latter are E. coli and pneumococci, but the manner by
873
which they invade the peritoneal cavity is unknown, possibly blood-borne.
Morphology.
Depending on the duration of the peritonitis, the membranes show the following changes. Approximately 2 to 4 hours after initiation, there is loss of the gray, glistening quality of the peritoneal surface, and it becomes dull and lusterless. At this time, there is a small accumulation of essentially serous or slightly turbid fluid. Later the exudate becomes creamy and obviously suppurative. In some cases, it may be extremely thick and plastic and even inspissated, especially in dehydrated patients. The volume of exudates varies enormously. In many cases, it may be localized by the omentum and viscera to a small area of the abdominal cavity. In generalized peritonitis, it is important to remember that
an exudate may accumulate under and above the liver to form subhepatic and subdiaphragmatic abscesses. Collections in the lesser omental sac may likewise create residual persistent foci of infection.
The inflammatory process is typical of an acute bacterial infection anywhere and produces the characteristic neutrophilic infiltration with fibrinopurulent exudation. The reaction usually remains superficial and does not penetrate deeply into the visceral structures or abdominal wall. Tuberculous peritonitis tends to produce a plastic exudate studded with minute, pale granulomas.
These inflammatory processes can heal either spontaneously or with therapy. In the course of healing, the following may occur: (1) The exudate may be totally resolved, leaving no residual fibrosis; (2) residual, walled-off abscesses may persist, eventually to heal or serve as foci of new infection; or (3) organization of the exudate may occur, with the formation of fibrous adhesions, which may be delicate or quite dense.
SCLEROSING RETROPERITONITIS
Dense fibromatous overgrowth of the retroperitoneal tissues may sometimes develop, designated sclerosing retroperitonitis or idiopathic retroperitoneal fibrosis (also called Ormond disease). In some instances the mesentery is also involved. The fibrous overgrowth is entirely nondistinctive and, although infiltrative, does not display anaplasia. There is usually an accompanying inflammatory infiltrate of lymphocytes, plasma cells, and neutrophils, suggesting inflammatory rather than neoplastic disease. The fibrosis may encroach on the ureters to produce hydronephrosis. Alternatively, fibrous tissue may surround retroperitoneal organs and extend into the mesentery. In some ways this process is an analogue of the desmoid tumor. The cause of this curious condition is obscure; in some instances there is a history of intake of the drug methysergide, an ergot derivative used for migraine, or history of previous surgery or radiation therapy. However, most cases have no obvious cause. Similar fibrotic changes seen in other sites (mediastinal fibrosis, sclerosing cholangitis, and Riedel fibrosing thyroiditis) suggest that the disorder is autoimmune and systemic in origin, preferentially involving the retroperitoneum.
MESENTERIC CYSTS
Large to small cystic masses are sometimes found within the mesenteries in the abdominal cavity or attached to the peritoneal lining of the abdominal wall. These cysts sometimes offer difficult clinical problems because they present on palpation as abdominal masses. Many classifications have been attempted; the most widely used is based on pathogenetic origins: (1) those arising from sequestered lymphatic channels; (2) those derived from pinched-off enteric diverticula of the developing foregut and hindgut; (3) those derived from the urogenital ridge or its derivatives (i.e., the urinary tract and male and female genital tracts); (4) those derived from walled-off infections or following pancreatitis, more properly called pseudocysts; and (5) those of malignant origin, most often resulting from peritoneal involvement by intra-abdominal adenocarcinomas.
Tumors
Virtually all tumors of the peritoneum are malignant and can be divided into primary and secondary forms.
Primary tumors arising from the mesothelium of the peritoneum are extremely rare and are called mesotheliomas. These exactly duplicate mesotheliomas found in the pleura and the pericardium, but the prognosis is poor. Like the supradiaphragmatic tumors, peritoneal mesotheliomas are associated with asbestos exposure in at least 80% of cases. How inhaled asbestos induces a peritoneal neoplasm remains a mystery. It has been recently suggested that genetic factors or viral infections may play a role in the genesis of peritoneal mesothelioma. The histopathologic diagnosis of mesothelioma is not always straightforward. In many occasions, immunohistochemical stains are required to differentiate this tumor from forms of adenocarcinoma.
Desmoplastic small round cell tumor is a rare tumor arising from peritoneum. The exact histogenesis and pathogenesis of this tumor are still not known. Molecular marker studies have suggested that this tumor is in the family of small round cell tumors such as Ewing sarcoma, rhabdoid myosarcoma, and primitive neuroectodermal tumor. The characteristic genetic marker for this tumor is the reciprocal chromosome translocation t(11;22) (p13;q12) resulting in EWS-WT1 fusion.[99]
Secondary tumors of the peritoneum are, in contrast, quite common. In any form of advanced cancer, penetration to the serosal membrane or metastatic seeding (peritoneal carcinomatosis) may occur. The most common tumors producing diffuse serosal implantation are ovarian and pancreatic. Appendiceal mucinous carcinomas may produce pseudomyxoma peritoneii, as described earlier. However, any type of intra-abdominal malignancy, and occasionally tumors from extra-abdominal locations, may be implicated in peritoneal seeding.
Additional mention might be made of the very uncommon tumors that may arise from retroperitoneal tissues (i.e., fat, fibrous tissue, blood vessels, lymphatics, nerves, and the lymph nodes alongside the aorta). These native structures may give rise to benign or malignant tumors derived from any of the indigenous mesenchymal cell types, resembling their counterparts arising elsewhere in the body.
References
1. DeNardi FG, Riddell RH: The normal esophagus. Am J Surg Pathol 15:296, 1991.
2. Hornby PJ, Abrahams TP, Partosoedarso ER: Central mechanisms of lower esophageal sphincter control.
11. Jankowski JA, et al: Molecular evolution of the metaplasia-dysplasia-adenocarcinoma sequence in the esophagus. Am J Pathol 154:965, 1999.
12. Souza RF: Molecular and biologic basis of upper gastrointestinal malignancy—esophageal carcinoma. Surg Oncol Clin N Am 11:257, 2002.
13. Kok TC, et al: No evidence of known types of human papillomavirus in squamous cell cancer of the oesophagus in a low-risk area. Rotterdam Oesophageal Tumour Study Group. Eur J Cancer 33:1865, 1997.
15. Jenkins GJ, et al: Genetic pathways involved in the progression of Barrett's metaplasia to adenocarcinoma. Br J Surg 89:824, 2002.
16. Croft J, et al: Analysis of the premalignant stages of Barrett's oesophagus through to adenocarcinoma by comparative genomic hybridization. Eur J Gastroenterol Hepatol 14:1179, 2002.
17. Hanby AM, et al: The mucous neck cell in the human gastric corpus: a distinctive, functional cell lineage. J Pathol 187:331, 1999.
18. Kojima M, et al: Ghrelin is a growth-hormone-releasing acylated peptide from stomach. Nature 402:656, 1999.
19. Owen DA: Gastritis and carditis. Mod Pathol 16:325, 2003.
23. Covacci A, Rappuoli R: Helicobacter pylori: after the genomes, back to biology. J Exp Med 197:807, 2003.
24. Backert S, et al: Functional analysis of the cag pathogenicity island in Helicobacter pylori isolates from patients with gastritis, peptic ulcer, and gastric cancer. Infect Immun 72:1043, 2004.
25. Furuta T, et al: Interleukin 1β polymorphisms increase risk of hypochlorhydria and atrophic gastritis and reduce risk of duodenal ulcer recurrence in Japan. Gastroenterology 123:92, 2002.
26. Solnick JV, Schauer DB: Emergence of diverse Helicobacter species in the pathogenesis of gastric and enterohepatic diseases. Clin Microbiol Rev 14:59, 2001.
27. Toh BH, van Driel IR, Gleeson PA: Pernicious anemia. N Engl J Med 337:1441, 1997.
28. Sandler RS, et al: The burden of selected digestive diseases in the United States. Gastroenterology 122:1500, 2002.
30. Prinz C, Hafsi N, Voland P: Helicobacter pylori virulence factors and the host immune response: implications for therapeutic vaccination. Trends Microbiol 11:134, 2003.
31. Ming SC: Cellular and molecular pathology of gastric carcinoma and precursor lesions: a critical review. Gastric Cancer 1:31, 1998.
32. Abraham SC, et al: Sporadic fundic gland polyps: common gastric polyps arising through activating mutations in the β-catenin gene. Am J Pathol 158:1005, 2001.
33. Wingo PA, et al: Long-term trends in cancer mortality in the United States, 1930–1998. Cancer 97:3133, 2003.
34. Normark S, et al: Persistent infection with Helicobacter pylori and the development of gastric cancer. Adv Cancer Res 90:63, 2003.
35. Kelley JR, Duggan JM: Gastric cancer epidemiology and risk factors. J Clin Epidemiol 56:1, 2003.
36. Powell SM: Stomach cancer. In: Vogelstein B, Kinzler W (eds): The Genetic Basis of Human Cancer, 2nd Edition. New York: McGraw-Hill, 703, 2002.
37. Naumann M, Crabtree JE: Helicobacter pylori-induced epithelial cell signalling in gastric carcinogenesis. Trends Microbiol 12:29, 2004.
56. Procop GW: Gastrointestinal infections. Infect Dis Clin North Am 15:1073, 2001.
57. Katz DE, Taylor DN: Parasitic infections of the gastrointestinal tract. Gastroenterol Clin North Am 30:797, 2001.
58. Caplan MS, Jilling T: New concepts in necrotizing enterocolitis. Curr Opin Pediatr 13:111, 2001.
59. Pardi DS: Microscopic colitis. Mayo Clin Proc 78:614, 2003.
60. Cohen J, West AB, Bini EJ: Infectious diarrhea in human immunodeficiency virus. Gastroenterol Clin North Am 30:637, 2001.
61. Cruz-Correa M, et al: Endoscopic findings predict the histologic diagnosis in gastrointestinal graft-versus-host disease. Endoscopy 34:808, 2002.
62. Cipolla G, et al: Nonsteroidal anti-inflammatory drugs and inflammatory bowel disease: current perspectives. Pharmacol Res 46:1, 2002.
63. Farrell RJ, Kelly CP: Celiac sprue. N Engl J Med 346:180, 2002.
64. Catassi C, Fasano A: New developments in childhood celiac disease. Curr Gastroenterol Rep. 4:238, 2002.
65. Green PH, Jabri B: Celiac disease. Lancet 362:383, 2003.
66. Marth T, Raoult D: Whipple's disease. Lancet 361:239, 2003.
67. Bentley SD, et al: Sequencing and analysis of the genome of the Whipple's disease bacterium Tropheryma whipplei. Lancet 361:637, 2003.
68. Podolsky DK: Inflammatory bowel disease. N Engl J Med 347:417, 2002.
69. Mowat AM: Anatomical basis of tolerance and immunity to intestinal atigens. Nat Rev Immunol 3:331, 2003.
70. Bouma G, Strober W: The immunological and genetic basis of inflammatory bowel disease. Nat Rev Immunol 3:521, 2003.
71. Hugot JP, et al: Association of NOD2 leucine-rich repeat variants with susceptibility to Crohn's disease. Nature 411:599, 2001.
72. McKay DM: Intestinal inflammation and the gut microflora. Can J Gastroenterol 13:509, 1999.
875
73. Dubinsky MC, et al: Clinical utility of serodiagnostic testing in suspected pediatric inflammatory bowel disease. Am J Gastroenterol 96:758, 2001.
74. Bansi DS, Chapman RW, Fleming KA: Prevalence and diagnostic role of antineutrophil cytoplasmic antibodies in inflammatory bowel disease. Eur J Gastroenterol Hepatol 8:881, 1996.
76. McKenna BJ, Appelman HD: Dysplasia can be a pain in the gut. Pathology 34:518, 2002.
77. Chen R, et al: DNA fingerprinting abnormalities can distinguish ulcerative colitis patients with dysplasia and cancer from those who are dysplasia/cancer-free. Am J Pathol 162:665, 2003.
78. Eaden JA, Abrams KR, Mayberry JF: The risk of colorectal cancer in ulcerative colitis: a meta-analysis. Gut 48:526, 2001.
79. Miyaki M, Kuroki T: Role of Smad4 (DPC4) inactivation in human cancer. Biochem Biophys Res Commun 306:799, 2003.
80. Burgart LJ: Colorectal polyps and other precursor lesions. Need for an expanded view. Gastroenterol Clin North Am 31:959, 2002.
81. Bienz M: Apc. Curr Biol 13:R215, 2003.
82. Lynch HT, de la Chapelle A: Hereditary colorectal cancer. N Engl J Med 348:919, 2003.
84. Jass JR: Pathogenesis of colorectal cancer. Surg Clin N Am 82:891, 2002.
85. Knudson AG: Two genetic hits (more or less) to cancer. Nat Rev Cancer 1:157, 2001.
86. Kinzler KW, Vogelstein B: Colorectal tumors. In: Kinzler KW, Vogelstein B (eds): The Genetic Basis of Human Cancer, 2nd Edition. New York: McGraw-Hill 583, 2002.
87. Wynter CV, et al: Methylation patterns define two types of hyperplastic polyp associated with colorectal cancer. Gut 53:573, 2004.
88. Bader S, et al: MBD1, MBD2, and CGBP genes at chromosome 18q21 are infrequently mutated in human colon and lung cancers. Oncogene 22:3506, 2003.
89. Taketo MM, Takaku K: Gastro-intestinal tumorigenesis in Smad4 mutant mice. Cytokine Growth Factor Rev 11:147, 2000.
90. Rudolph KL, et al: Telomere dysfunction and evolution of intestinal carcinoma in mice and humans. Nat Genet 28:155, 2001.
91. Jemal A, et al: Cancer statistics, 2002. CA Cancer J Clin 52:23, 2002.
92. Wei EK, et al: Comparison of risk factors for colon and rectal cancer. Int J Cancer 108:433, 2004.
93. Lieberman DA, et al: Risk factors for advanced colonic neoplasia and hyperplastic polyps in asymptomatic individuals. JAMA 290:2959, 2003.
94. Imperiale TF: Aspirin and the prevention of colorectal cancer. N Engl J Med 348:879, 2003.
95. Parsonnet J, Isaacson PG: Bacterial infection and MALT lymphoma. N Engl J Med 350:213, 2004.
96. Matczak E: Human papillomavirus infection: an emerging problem in anal and other squamous cell cancers. Gastroenterology 120:1046, 2001.
97. Goede AC, Caplin ME, Winslet MC: Carcinoid tumour of the appendix. Br J Surg 90:1317, 2003.
98. Stancu M, et al: Genetic alterations in goblet cell carcinoids of the vermiform appendix and comparison with gastrointestinal carcinoid tumors. Mod Pathol 16:1189, 2003.
99. Reynolds PA, et al: Identification of a DNA-binding site and transcriptional target for the EWS-WT1 (+KTS) oncoprotein. Genes Dev 17:2094, 2003.
