Abnormal Plasma Gut Hormones in Pathologic Duodenogastric Reflux and

Their Response to Surgery Paul Wilson, rRCS(Edin), Neil T. Welch, FRCS(Edin), Ronald A. Hinder, MD, Marco Anselmino, MD,

Margery K. Herrington, MS, Tom R. DeMeester, MD, Thomas E. Adrian, PhD, MRCPath, Omaha, Nebraska

Fasting and postprandial plasma levels of the gut hormones gastrin, choleeystoklnin (CCK), seeretin, glucose-dependent insulinotropie polypeptide, moti- lin, neurotensin, peptide YY (PYY), enterogluea- gon, glueagon, insulin, and pancreatic polypeptide were measured in 11 patients with alkaline gastritis associated with excessive duodenogastric reflux not related to previous gastric surgery (primary DGR), 12 primary DGR patients after pancreatico-biliary diversion ("duodenal switch" procedure), and in 10 age-matched healthy controls. Gastric emptying of a semi~ulid oatmeal was also measured in pa- tients with primary DGR and in patients after bile diversion. Fasting plasma levels of the distal gut hormone neurotensin and the pancreatic islet hor- mone insulin were significantly greater in patients with , primary DGR compared with controls. Neuro- tensin levels were normal in patients studied after bile diversion. Postprandial plasma levels, incre- mental integrated and total integrated responses for CCK, seeretin, insulin, neurotensin, PYY, and en- teroglueagon, were significantly greater in patients with primary DGR compared with controls. The majority of these responses normali-~qt after bile diversion; however, the postprandial response for insulin and enteroglucagon remained elevated. Pa- tients with primary DGR had a rapid early post- prandial phase of gastric emptying of solids, which showed a significant correlation with plasma neuro- tensin levels. Bile diversion produced a significant delay in this lag-phase of gastric emptying. These abnormalities in gut regulatory hormones appear to be adaptive changes to rapid early postprandial gastric emptying, probably related to antropylorie dysmotility, which has been implicated in the pathogenesis of this condition. Measurement of these gastrointestinal hormones may become useful in the diagnosis of primary DGR.

From the Departments of Surgery and Biomedical Sciences, Creighton University School of Medicine, Omaha, Nebraska. Supported by PHS grant #ROt DK40381-02.

Requests for reprints should be addressed to Thomas E. Adrian, PhD, MRCPath, Department of Biomedical Sciences, Creighton Uni- versity School of Medicine, 601 North 30th Street, Omaha, Nebraska 68131.

Presented at the 33rd Annual Meeting of the Society for Surgery of the Alimentary Tract, San Francisco, California, May 11-13, 1992.

D uodenogastric reflux (DGR) occurs under phys- iologic conditions [I] but, when excessive, may give

rise to a clinical syndrome characterized by epigastric pain, nausea, and bilious vomiting [2] associated with a distinct histopathologic form of Chronic gastritis [3]. DGR has been implicated in the development of gastric ulcer, gastric carcinoma, esophagitis, the postcholecys- tectomy syndrome, and Barrett's esophagus with its com- plications of ulceration, stricture, and dysplasia [4-8]. Abnormal DGR most commonly occurs after surgical procedures that reseet or bypass the pyloric sphincter, such as distal gastrectomy, gastrojejunostomy, and pylor- oplasty [9]. However, over recent years, a number of patients with foregut symptoms who have not undergone previous gastric surgery have been identified as having pathologic DGR [5,10,11]. In these patients, a nonfunc- tioning gallbladder or previous cholecystectomy may con- tribute to excessive bile reflux [10,12,13]. Attention has been focused on antropyloric dysmotility in the patho- physiology of this condition, but this has been difficult to quantify [14-16]. Bile reflux has been observed in the presence of antral hypomotility during the late phase II of the migrating motor complex (MMC) arising in the duo- denum [14]. Hence, the normal mechanism for clearing the duodenogastric refluxate may be impaired in these patients. Foregut motility and secretory functions are intimately regulated by several gastrointestinal neuro- peptides and hormones, in conjunction with the autonom- ic nervous system [17]. A number of studies have defined abnormalities in plasma levels of specific gut regulatory hormones in conditions characterized by disordered gas- trointestinal motility and after a number of gastrointesti- nal surgical procedures [I 8-23]. These include reduced motilin levels in idiopathic delayed gastric emptying and chronic idiopathic constipation [24,25], increased motilin levels in diabetic gastroparesis [26], increased neuroten- sin, enteroglucagon, and peptide YY (PYY) levels in malabsorptive states and in conditions characterized by rapid gastric emptying and intestinal transit, such as oc- curs in the postgastrectomy dumping syndrome, after intestinal resection, and after jejunoileal bypass [18-21]. After partial gastrectomy, elevated postprandial chole- cystokinin (CCK) levels have also been observed [22]. A number of these foregut and distal gut hormones are known to inhibit gastric emptying, slow small intestinal transit, increase colonic motility, and influence foregut and midgut MMC patterns [27-32].

Primary pathologic DGR may be related to abnormal foregut motility with changes in pyloric function, gastric emptying, and intestinal transit. If tiffs is the case, there

THE AMERICAN JOURNAL OF SURGERY VOLUME 165 JANUARY 1993 169

WILSON ET AL

Figure 1. The "duodenal switch" procedure (suprapapil!ary Roux- en-Y duodenojejunostomy) used for pancreatico-biliary diversion in patients with primary pathologic duodenogastric reflux. The operation preserves 5 to 7 cm of the suprapapillary duodenum in continuity with the innervated antropyloric region of the stomach. (Reproduced with permission from: Wilson P, Hinder RA. The duodenal switch procedure for duodenogastric reflux. In: Prob- lems in General Surgery, vol 10, number 4. Philadelphia: JB Lippincott.)

may be associated changes in gut regulatory hormone responses. We, therefore, measured fasting and postpran- dial plasma levels of a number of these gut regulatory hormones and gastric emptying of solids in patients with primary DGR and in patients with DGR treated by pan- creatico-biliary diversion.

PATIENTS AND METHODS The fasting and postprandial plasma levels of the gas-

trointestinal hormones gastrin, CCK, secretin, glucose- dependent insulinotropic polypeptide (GIP), motilin, neurotensin, PYY, and enteroghicagon and the pancreat- ic islet hormones glucagon, insulin, and pancreatic poly- peptide (PP) were measured in the following groups of subjects. (All patients underwent gastric emptying [oat- meal and milk] scans before and after bile diversion.)

Normal subjects: Ten healthy volunteers had no gas- trointestinal symptoms or previous disease or surgery. Mean age was 53.3 years (SD: 7.8 years; range: 42 to 66 years). There were four men and six women.

Primary DGR patients: Eleven patients had symp- toms of epigastric pain, nausea, and bilious vomiting. Patients had not undergone previous gastric surgery. Pathologic primary DGR was diagnosed on the basis of evidence of gastritis on endoscopy; the presence of a gas- tric "bile lake" on endoscopy; evidence of "alkaline gas- tritis" on histologic examination of gastric biopsies, that is, a composite alkaline reflux histologic score [3] of at least nine and/or the presence of moderate or severe foveolar hyperplasia; evidence of abnormal DGR on hep- atobiliary scintigraphy; or evidence of abnormal gastric alkaline exposure on 24-hour gastric pH monitoring (pos- itive discriminant score [33] and/or prolonged [greater than 10%] fasting exposure to pH >3). All patients were symptomatic, had evidence of gastritis on endoscopy, and had an additional two or more of the above pathologic DGR criteria. Sixty-four percent (7 of 11) of patients had 4 or more of these criteria. Mean age was 55.7 years (SD: 11.5 years; range: 39 to 72 years). There were three men and eight women. Mean duration of symptoms was 6.4 years (range: 1 to 15 years).

DGR patients after bile diversion: Twelve patients had primary pathologic DGR treated by pancreatico- biliary diversion (suprapapillary Roux-en-Y duodenoje- junostomy ["duodenal switch" operation], Figure 1). Pre- operatively, all these patients were symptomatic, had evidence of gastritis on endoscopy, and fulfilled an addi- tional two or more of the stated DGR criteria. Fifty-eight percent (7 of 12) of patients had 4 or more of these criteria. Postoperative testing was carried out at a mean of 2.0 years after surgery (range: 6 months to 6 years). Mean age was 55.4 years (SD: 12.0 years; range: 33 to 77 years). There were three men and nine women. Mean symptom duration was 6.6 years (range: 1 to 23 years).

Six patients with primary DGR underwent hormone estimations and gastric emptying scans before and after bile diversion.

Blood collection: After patients fasted overnight for 12 hours, blood samples (10 mL) were drawn from an antecubital vein via an indwelling venous catheter. Three basal fasting samples were collected at 15-minute inter- vals. Subjects then ingested a standard meal consisting of two hard-boiled eggs, two slices of toasted white bread with 10 g of butter each and 10 g of fruit jelly, and 200 mL of fresh orange juice (total calorific value: 530 kcal, protein 18 g, carbohydrate 66 g, fat 22 g). Further blood samples were drawn postprandially at 15, 30, 60, 120, and 180 minutes. Blood samples were collected into chilled tubes containing ethylene diaminetetraacetic acid (EDTA) (2 mg/mL) and aprotinin (Trasylol 400 KIU/mL, Sigma Chemical Co., St. Louis, MO). Sam- ples were centrifuged (3,000 rpm at 4~ for 10 minutes); the plasma was decanted and stored frozen at -800C for subsequent radioimmunoassay.

Radioimmunoassay technique: Plasma concentra- tions of gastrin, CCK, secretin, GIP, motilin, neuroten- sin, enteroglucagon, glucagon, insulin, and PP were mea-

170 THE AMERICAN JOURNAL OF SURGERY VOLUME 165 JANUARY 1993

GUT HORMONES IN DUODENOGASTRIC REFLUX

sured using specific and sensitive radioimmunoassays that have been previously described in detail [34]. For CCK and secretin assays, samples were obtained using reverse-phase C18 disposable cartridges (Sep-Pak, Wa- ters Corp., Milford, MA) [35]. Experience has shown greater than 80% recovery of CCK-8 and CCK-33 from human plasma. CCK was measured using a specific anti- body that shows full cross-reaction between synthetic sul- fated CCK-8 and CCK-33 but shows no significant cross- reaction with the gastrins [35]. Bolton and Hunter reagent was used to label CCK-8 to provide tracer for the CCK assay and was purified using reverse-phase high- pressure liquid chromatography. Plasma PYY concen- trations were measured using an N-terminally directed antibody, as previously described [36].

Gastric emptying scans: Patients with primary DGR and DGR patients after bile diversion underwent mea- surement of the rate of gastric emptying of a semisolid. After an overnight fast, subjects swallowed 1 #Ci of tech- netium 99m (99mTc)-sulfur colloid mixed with an oat- meal preparation (2 oz of dry instant oatmeal, 3A cup of hot water, and �89 oz of sugar) and drank 1 pint of milk. Anterior and posterior images were obtained by a gamma camera fitted with a low-energy, all-purpose collimator. The energy window was set at 140 keV (4- 20%). Imaging was performed for 60-second periods over the region of the stomach with subjects in the upright position. Imag- ing commenced immediately after the meal and was con- tinued at 15-minute intervals for the following 150 min- utes. Between imaging sessions, subjects were allowed to sit or stand away from the camera. Data were recorded on a Microdelt system computer using GASEMP (Sei- mans, Munich, Germany). All 99mTC counts were cor- rected for decay, and the geometric mean values were computed. The rate of emptying was calculated as the percentage gastric retention of isotope against time, and the Tt/2 (Tso) was the calculated time for 50% of the isotope to leave the stomach. The percentage of isotope retention at 10 minutes (T10) and the emptying rate from 10 to 120 minutes (Tlo-120) were also calculated.

Statistical analysis: Data were analyzed by the Krus- kal-Wallis test for nonparametric data and are expressed as the mean 4- standard error of the mean (SEM). The Wilcoxon signed-rank test was used for the analysis of paired data. After correction for multiple comparisons, a p value of <0.05 was considered to be significant. The incremental integrated response and the total integrated response were calculated in individual subjects, for each hormone, utilizing the trapezoidal method. In primary DGR patients, the amount of isotope emptied from the stomach (at 15, 30, 60, and 120 minutes) was compared with the levels of plasma CCK, secretin, insulin, entero- glucagon, neurotensin, and PYY at 15, 30, 60, and 120 minutes using linear regression analysis. Four primary DGR patients and two DGR patients after bile diversion had hypergastrinemia (mean basal levels: 211 4- 80 pmol/L) associated with achlorhydria. The gastrin data of these patients were excluded from the analysis.

Informed written consent was obtained from all sub- jects. The study was approved by the Institutional Review Board at Creighton University.

RESULTS The mean basal (fasting) plasma levels, incremental

integrated response (area under the postprandial curve above the mean basal level from 0 to 180 minutes, IIR18o), and the total integrated response (total area un- der the postprandial curve from 0 to 180 minutes, TIRlso) for all hormones measured in all groups are shown in Table I.

Gastrin: Basal levels, IIR180, and TIR180 for gastrin were not different in the three groups (Table I, Figure 2). There were seven primary DGR patients and 10 DGR patients after bile diversion included in the analysis, due to the exclusion of the patients with achlorhydria.

CCK: Mean basal plasma levels of CCK were similar in normal subjects, patients with primary DGR, and pa- tients with primary DGR after bile diversion (Table I). However, the postprandial response (IIR180 and TIR180) in primary DGR patients was significantly greater com- pared with normal subjects (both p <0.05, Table I, Fig- ure 2). After bile diversion, the postprandial response normalized. CCK concentrations in the hypergastrinemic patients were not significantly different from those of the patients with normal gastrin levels, confirming the lack of cross-reactivity of gastrin with the CCK antibody in the radioimmunoassay.

Seeretin: Mean basal levels of secretin were not sig- nificantly different from those of normal subjects (Table I, Figure 2). However, DGR patients after bile diversion had significantly lower basal levels compared with pri- mary DGR patients (p <0.05). Primary DGR patients had a significantly greater postprandial response (TIRls0) compared with normal subjects (p <0.05). Af- ter bile diversion, the postprandial response normalized (p <0.05).

GIP. Basal levels, IIRl80, and TIR18o for GIP were not different in the three groups (Table I, Figure 2).

blotilin- Basal levels, IIR18o, and TIR18o for motilin were not different in the three groups (Table I, Figure 2).

Neurotensin. Basal levels were elevated in primary DGR patients compared with normal subjects (p <0.05) (Table I, Figure 2). Primary DGR patients also had a significantly greater postprandial response (IIR180 and TIR180) compared with normal subjects (p <0.05 and p <0.01, respectively). Basal levels and postprandial re- sponse (TIR18o) normalized in DGR patients after bile diversion (both p <0.05).

PYY. Basal levels of PYY in primary DGR patients were not significantly different from those of normal sub- jects (Table I, Figure 2). However, DGR patients after bile diversion had significantly lower basal levels com- pared with primary DGR patients (p <0.05). Primary DGR patients had a significantly greater postprandial response (TIR180) compared with normal subjects (p <0.05), which was reduced in DGR patients after bile diversion but remained significantly greater (IIRls0) than in normal subjects (p <0.05).

Enteroglueagon: Basal levels of enteroglucagon were not significantly different in the three groups studied (Table I, Figure 2). The postprandial response (IIRlso and TIRlso) was significantly greater in primary DGR patients compared with normal subjects (both p <0.05).

THE AMERICAN JOURNAL OF SURGERY VOLUME 165 JANUARY 1993 171

WILSON ET AL

TABLE I Mean Hormone Levels in all Patients*

Hormone Group Mean Basal Mean IIR18o Mean TIR18o

Gastrin f Normals 16.0 - 2.3 2,656 • 609 5,533 - 787 Primary DGR 18.8 _+ 2.8 2,297 _+ 533 5,689 _ 520 After diversion 20.6 - 2.5 2,753 _+ 613 6,466 +_. 1,027

Cholecystokinin Normals 1.5 -+ 0.1 6331 _ 101 910 • 103 Primary DGR 1.8 -+ 0.2 1,141 _+ 1615 1,454 + 143 ~: After diversion 1.6 + 0.1 733 + 109 1,014 _+ 100

Secretin Normals 2.4 - 0.3 124 _+ 34 551 _ 53 Primary DGR 2.6 -+ 0.3w 165 _+ 42 712 _ 505 After diversion 2.0 -+ 0.2 115 • 22 483 • 34

Gastric inhibitory peptide Norrnals 57.0 --_ 7.3 8,371 _+ 1,501 18,636 • 2,626 Primary DGR 52.6 -+ 5.0 9,264 _ 2,270 18,734 _.+ 2,962 After diversion 57.7 -+ 6.3 8,741 _+ 1,422 19,124 _+ 2,315

Motilin Normals 59.1 • 15.6 88 - 1,176 10,752 _ 1,977 Primary DGR 58.2 --- 5.8 214 _+ 1,081 10,698 • 1,207 After diversion 58.8 -+ 6.6 533 _ 1,238 11,248 _+ 1,511

Neurotensin Normals 5.3 • 1.6 709 _+ 165 1,657 -+ 369 Primary DGR 11.0 -+ 2.8$ 1,810 +_ 484$ 3,749 • 556** After diversion 5.8 • 0.9 1,030 - 149 2,075 _.+ 248

Peptide~(Y Normals 13.0 • 1.0 407 • 101 2,746 • 183 Primary DGR 15.0 -+ 1.6 2,500 _+ 1,250 5,199 • 1,096ff After diversion 11.0 -+ 0.9*$ 1,858 - 640tt 3,866 -+ 754

Enteroglucagon Normals 25.4 _+ 2.5 1,432 • 385 5,996 • 477 Primary DGR 22.6 -+ 2.6 4,737 _+ 996tt 8,802 _+ 1,099 t t After diversion 20.8 • 1.8 5,044 _+ 1,353 t t 8,785 _+ 1,257

Glucagon Normals 14.3 -+ 0.6 112 • 77 2,681 • 106 Primary DGR 15.7 • 1.0w 205 _+ 178 3,029 _+ 269w After diversion 11.1 • 0.6 t t 72 + 90 2,061 _+ 119it

Insulin Normals 28.4 • 3.9 22,760 _ 5,702 27,878 _+ 6,122 Primary DGR 68.0 -+ 18.1tt 34,887 • 9,321 47,127 • 12,413 t f After diversion 52.3 -+ 11.4 36,783 _+ 10,000 t t 46,203 • 10,900 t t

Pancreatic polypeptide Normals 25.8 -+ 2.2 8,397 _ 1,739 13,041 _ 1,823 Primary DGR 32.0 • 3.1 16,790 • 3,494 22,551 _ 3,804 After diversion 28.7 • 3.3 9,629 _+ 2,485 14,789 • 2,930

Primary DGR = excessive duodenogastric reflux not related to previous gastric surgery. *Mean (_SEM) basal plasma hormone levels (pmol/L), mean (_SEM) incremental integrated response over 180 minutes (mean IIR180, pmol/L, 180 minutes), and mean

(-+ SEM) total integrated response (mean TIRlao, pmol/L, 180 minutes) in normal subjects (norma/s, n = 10), patients with primary pathologic duodenogastric reflux (primary DGR, n = 11), and patients with primary DGR after pancreatico-biliary diversion (after diversion, n = 12).

1After exclusion of achlorhydric patients (primary DGR patients: 4; DGR patients after diversion: 2). *p < 0.05 versus normals and DGR patients after diversion. w < 0.05 versus DGR patients after diversion. **p <0.01 versus normals, p <0.05 versus DGR patients after diversion. ttp <0.05 versus normals. $tp< 0.05 versus primary DGR.

This abnormal response persisted after bile diversion, IIRl80 remaining significantly elevated compared with that of normal subjects (p <0.05) (Table I).

Gheagon: Basal levels were not significantly differ- ent in primary DGR patients compared with normal sub- jects (Table I ). In DGR patients after bile diversion, there was a small but significant reduction in basal levels com- pared with primary DGR patients and normal subjects (p <0.001 and p <0.05, respectively). The postprandial re-

sponses (TIR 180) were not significantly different between the groups (Table I).

Insulin. Basal levels were elevated in primary DGR patients compared with normal subjects (p <0.05 ) (Table I, Figure 2). Primary DGR patients also had a signifi- cantly greater postprandial response (TIRls0) compared with normal subjects (p <0.05). Elevated basal levels and postprandial response persisted in DGR patients after bile diversion (Table I).

172 THE AMERICAN JOURNAL OF SURGERY VOLUME 165 JANUARY 1993

GUT HORMONES IN DUODENOGASTRIC REFLUX

Figure 2. Plasma concentrations of gastrointestinal hormones during a test meal in normal subjects (n = 10), primary duodenogastric reflux (DGR) patients (n -- 11), and DGR patients after bile diversion (n -- 12). For gas- trin, patients with achlorhydria were excluded (primary DGR, n -- 4; after diversion, n = 2); the group sizes are normal subjects (n = 10), primary DGR (n = 7), and DGR after bile diver- sion (n = 10). Data are expressed as mean 4- SEM. CCK -- cholecystoki- nin; GIP = gastric inhibitory peptide; PP -- pancreatic polypeptide; PYY = peptlde YY.

Plasma GASTRIN (pmol/I)

12/ Meal t

I ~ 1 Normais . 40 rs lon

20

0 . . . . . . . - 30 0 30 60 90 120 150 180 Basal Minutes

Plasma SECRETIN (pmol/I)

il Meal~ Primary DGR

§ \ After Diversion

0 . . . . . . . . - 30 o 30 60 90 12o 15o 180 Basal Minutes

Plasma GIP (pmol/I)

l M i a l ~ s ,5o I I / . ~ I Normals

IO0 l

50 ~ ion

o . . . . . . . . - 30 0 30 60 90 120 160 180 Basal Minutes

Plasma PP (pmol/I)

150 D G R

75

N o r m a l s A f t e r D ivers ion

o . . . . . . . . - 3 0 o 30 60 90 120 150 18o Basal Minutes

Plasma NEUROTENSIN (pmol/I)

1 I I , P r imary DGR 2o I

version

10

Norma ls

o . . . . . . . . -30 0 30 S0 90 120 150 180 Basal Minutes

Plasma CCK (pmol/I) 1ii Meall p ,maryDGR version - 30 o 30 60 90 12o 15o 18o Basal Minutes

Plasma MOTILIN (pmol/I) 90 ] Mea

I | A f t e r D ive rs ion l

�9 l 6O

30 P r i m a r y D G R

o . . . . . . . - 3 0 0 3 0 60 90 120 150 180 Basal Minutes

Plasma INSULIN (pmol/I) 4 5 0 A,r / A f t e r D ive rs ion Mea, ~./ 300 ~ G R

1 5 0

0 -30 0 30 60 90 120 160 180 Basal Minutes

Plasma PYY (pmol/I)

l i ~ 1 vers ion

15 ~ 9

\ No rma ls

o . . . . . . . . - 3 0 0 30 60 90 120 150 180 Basal Minutes

Plasma ENTEROGLUCAGON (pmol/I)

Meal I A f t e r D ivers ion I

So " ~ O 40 I G R

20 N o r m a n

- 3 0 0 30 60 90 120 150 180 Basal Minutes

PP: Basal levels were not significantly different in the three groups studied (Table I, Figure 2). The postprandi- al response in primary DGR patients was greater than in normal subjects, but this did not reach statistical signifi- canoe (p = 0.2). In DGR patients after bile diversion, the postprandial response normalized.

Gastric emptying scans: Mean gastric emptying

time, Tso, in primary DGR patients was 51.8 4- 9.5 min- utes. In DGR patients after bile diversion, mean Ts0 was significantly greater at 76.7 4- 8.9 minutes (p <0.05). Patients with primary DGR had a significantly more rapid rate of early gastric emptying during the lag-phase (0 to 10 minutes) compared with DGR patients after bile diversion (p <0.05, Figure 3). Thereafter, from 10 to 120

THE AMERICAN JOURNAL OF SURGERY VOLUME 165 JANUARY 1993 173

WILSON ET AL

% lOO

80

60

40

20

o

Isotope remaining

After Diversion

. . . . . . . Prima,ry DGR

15 30 45 60 75 90 105 120

Minutes mlomin

% 45

30

15

0 m

Isotope emptied AFTER DIVERSION

~11 PRIMARY DGR

T 10min m lo -12om in

Figure 3. Top panel. Gastric emptying of oatmeal/milk/techne- tium 99m-sulfur colloid (% isotope remaining in the stomach versus time) in patients with primary duodenogastric reflux (DC-~) (n -- 11) and in patients with ~ after bile diversion (n = 12). Data expressed as mean 4- SEM. Dotted lines = 5th and 95th percentiles for normal subjects. Bottom panel. Percentage iso- tope emptied from the stomach (mean 4- SEM) in the early postprandial period (lag phase, Tlo rain), and the later phase of gastric emptying (Tlo-12o r~n) for the above groups,

minutes, the rate of gastric emptying was not significant- ly different between these groups (Figure 3). There was no correlation between the percentage of isotope emptied from the stomach and levels of CCK, secretin, PYY, enteroglucagon, or insulin. In contrast, there was a good correlation between the percentage of isotope emptied and neurotensin levels at 15 minutes (r = 0.70, p <0.05 ), 30 minutes (r = 0.84, p <0.005), and 60 minutes (r = 0.79, p <0.005). The plasma neurotemin level was direct- ly proportional to the percentage of isotope emptied from the stomach at these postprandial times. Hence, patients with the most rapid gastric emptying had the greatest elevation in plasma neurotensin.

In the six primary DGR patients who underwent hor- mone studies before and after bile diversion, analysis of paired data confmned the above findings. Basal levels of secretin, neurotemin, PYY, and glucagon were signiti- canfly reduced after bile diversion (all p <0.05, Wilcoxon signed-rank test), whereas basal insulin levels did not

change significantly after bile diversion (p = 0.22). Post- prandial responses (IIRls0) for CCK, neurotensin, and PYY were reduced after bile diversion (p <0.05, p = 0.07, and p = 0.08, respectively). In contrast, the elevated postprandial responses for enteroglucagon and insulin persisted after bile diversion (p = 0.31 and p = 0.50, respectively).

COMMENTS Alkaline reflux gastritis is most commonly related to

DGR after gastric resectional surgery or procedures that damage or bypass the pyloric sphincter [9]. However, over recent years, a number of patients with foregut symptoms who have not undergone previous gastric sur- gery and in whom the pyloric sphincter is intact have been identified as having chronic gastritis, with characteristic histologic features of alkaline gastritis associated with excessive DGR. They can be regarded as having primary DGR [5,10,11]. Previous cholecystectomy or a nonfunc- tioning gallbladder may contribute to the increased vol- ume of refluxate in these patients, but the pathophysio- logic mechanisms in this condition have yet to be defined [10,12,13]. Abnormal antropyloric motility, or incompe- tence of the pyloric sphincter, may allow inappropriate reflux of the duodenal content into the stomach during fasting and inappropriate emptying of thyme into the duodenum during the early postprandial period. A num- ber of gastrointestinal hormones, such as CCK, motilin, neurotensin, and PYY, appear to be intimately involved in the coordination of antropyloric motility during the fasting and postprandial periods [17]. Many other hor- mones from both the foregut and distal gut are known to influence gastric emptying and secretion, intestinal MMC patterns, and transit [27-32].

In this study, patients with primary DGR had elevat- ed fasting plasma levels of neurotensin and insulin and elevated postprandial levels of CCK, secretin, neuroten- sin, PYY, and enteroglucagon. These patients also showed early rapid postprandial gastric emptying of a semisolid. CCK, neurotensin, PYY, and enteroglucagon all have potent inhibitory effects on gastric emptying and intestinal transit, and levels of all are elevated in condi- tions in which gastric emptying or intestinal transit is rapid [18-23].

CCK is released by the endocrine cells of the duodenal and jejunal mucosa in response to feeding [17]. Intraduo- denal protein and fat stimulate the release of CCK, the action of which is to stimulate contraction and emptying of the gallbladder and secretion of pancreatic enzymes. CCK is also a potent inhibitor of gastric emptying by inducing relaxation of the proximal stomach and by in- creasing pyloric sphincter tone [32]. After partial gas- trectomy (Billroth I and II), fasting plasma levels of CCK are similar to those of normal subjects [22]. However, the CCK response to oral fat in these patients is significantly greater than in normal subjects [22]. This effect may be related to the associated vagotomy that is performed dur- ing distal gastrectomy but is more probably due to rapid gastric emptying, which results in the activation of a large number of CCK-releasing cells [22].

174 THE AMERICAN JOURNAL OF SURGERY VOLUME 165 JANUARY 1993

GUT HORMONES IN DUODENOGASTRIC REFLUX

Neurotensin is released from endocrine cells in the ileal mucosa in response to small intestinal luminal fat [17,37]. Its effects include inhibition of gastric motility and emptying, reduction of gastric acid secretion, and a slowing of intestinal transit [28]. When exogenous neuro- tensin is infused in human subjects, it changes the pattern of the MMC from a fasting type to a fed type [29]. Elevated postprandial levels of neurotensin have been observed in patients after partial and total gastrectomy, after small bowel resection, after jejunoileal bypass for morbid obesity, and in the dumping syndrome [19-21]. Conditions or surgical procedures that result in excessive delivery of nutrients to the distal small intestine are asso- ciated with an elevation in postprandial levels of neuro- tensin.

PYY is released from specialized cells in the ileal and colonic mucosa in response to distal small intestinal intra- luminal nutrients. It slows jejunal and colonic motility, inhibits gastric secretion, inhibits gastric emptying, slows intestinal transit, and impairs distal propagation of the MMC [17,38]. Elevated postprandial levels have been observed in malabsorptive states including sprue, in con- ditions associated with diarrhea (especially Crohn's dis- ease), after bowel resection, and in the dumping syn- drome in which the quantity of PYY released is sufficient to significantly delay gastric emptying [18,30,39].

The hormonal responses observed in patients with pri- mary DGR appear to be adaptive changes, responding to the early rapid gastric emptying in these patients. An incompetent pyloric sphincter that allows excessive DGR during fasting may allow rapid inappropriate emptying of gastric contents into the duodenum in the early postpran- dial period, resulting in the observed hormonal responses, which attempt to improve postprandial pyloric sphincter tone and inhibit rapid gastric emptying. The observed slowing of gastric emptying after bile diversion is associ- ated with normalization of CCK, secretin, neurotensin, and PYY levels, presumably in response to reduced lumi- nal stimuli postoperatively. This slowing of gastric emp- tying may be due, in part, to the close proximity of the pyloric sphincter and the end-to-end proximal duodenoje- junostomy. The duodenal switch procedure involves tran- section of the suprapapillary duodenum with subsequent proximal duodenojejunostomy, in a Roux-en-Y hookup, leaving 5 to 7 cm of the proximal duodenum in continuity with the intact antrum and pylorus (Figure 1). The distal duodenum and proximal jejunum are, therefore, no long- er exposed to thyme and gastric secretion, and the release of CCK from these segments will subsequently diminish.

Postprandial enteroglucagon levels were also elevated in patients with primary DGR but remained high after bile diversion. A similar increase in enteroglueagon levels has been observed in conditions characterized by intesti- nal hurry and in the dumping syndrome. The physiologic role of enteroglucagon is unknown, but it may act as an enterogastrone [20]. Enteroglucagon is released from the same endocrine cell as PYY, under similar circum- stances, although there is some evidence that differential release occurs [40].

The elevation of PYY and enteroglucagon responses

in patients with primary DGR is interesting, particularly since the levels of these two peptides are not completely normalized after the duodenal switch operation. The lu- minal stimuli for release of these co-localized peptides apparently remain increased even after surgery. An insu- lin-rdeasing peptide (glucagon-related peptide 7-37 [GRP-7-37] amide) is co-produced with enteroglucagon from the glucagon gene. It is tempting to speculate that the increase of GRP-7-37 amide, which undoubtedly oc- curs in these patients, may be responsible for the in- creased insulin responses seen. The secretion of GIP, which also releases insulin in a glucose-dependent man- ner, is quite normal in the primary DGR patients.

Postprandial PP levels were also moderately elevated in primary DGR patients. PP is released from the pancre- as by both CCK and neurotensin, and recent studies with CCK-receptor blockade have demonstrated a physiologic role for CCK in the normal postprandial rise of this peptide with feeding. In this study, in patients with pri- mary DGR, the observed elevation of this hormone may be secondary to the elevated CCK and neurotensin levels.

Elevated basal neurotensin and insulin levels and ab- normal postprandial responses of CCK, secretin, neuro- tensin, PYY, enteroglucagon, and insulin are seen in pa- tients with primary DGR. Postprandial neurotensin levels correlate with the rate of gastric emptying in these patients. A reduction in the rate of gastric emptying after bile diversion is associated with a normalization of most of these hormonal abnormalities, although elevated post- prandial responses of PYY, enteroglucagon, and insulin persist. These adaptive gut regulatory hormonal changes may be related to disordered antropyloric motility and suggest that primary DGR is a specific foregut motility disorder. Measurement of gastrointestinal hormones may become useful in the diagnosis of primary DGR, a condi- tion that has hitherto been difficult to quantitate.

REFERENCES 1. Schindlbeck NE, Heinrich C, Stellaard F, Paumgartner G, Muller-Lissner SA. Healthy controls have as much bile reflux as gastric ulcer patients. Gut 1987; 28: 1577-83. 2. Ritchie WP. Alkaline reflux gastritis: an objective assessment of its diagnosis and treatment. Am J Surg 1980; 192: 288-98. 3. Dixon MF, O'Connor H J, Axon ATR, King RFJG, Johnston D. Reflux gastritis: a distinct histopathological entity? J Clin Pathol 1986; 39: 524-30. 4. Thomas WEG. The possible role of duodenogastric reflux in the pathogenesis of both gastric and duodenal ulcers. Scand J Gas- troenterol 1984; 19(Suppl 92): 151-5. 5. Gowen GF. Spontaneous enterogastric reflux gastritis and esophagitis. Ann Surg 1985; 201: 170-5. 6. Langhans P, Heger RA, Hohenstein J, Schlake W, Bunte H. Operation-sequel carcinoma of the stomach. Experimental studies of surgical techniques with or without resection. World J Surg 1981; 5: 595-605. 7. Attwood SE, DeMeester TR, Bremner CG, Barlow AP, Hinder RA. Alkaline gastroesophageal reflux: implications in the develop- ment of complications in Barrett's columnar-linod lower esophagus. Surgery 1989; 106: 764-70. 8. Stahlberg M, Jalovaara P, Laitinen S, Alavaiko M, Kairaluoma M. Duodenogastfic reflux: a cause for post-cholecystectomy com- plaints? Surg Res Comm 1987; 1: 27-35.

THE AMERICAN JOURNAL OF SURGERY VOLUME 165 JANUARY 1993 175

WILSON ET AL

9. Dewal P, King R, Johnston D. Bile acid and lysolecithin concen- trations in the stomach in patients with duodenal ulcer before oper- ation and after treatment by highly selective vagotomy, partial gastrectomy, or truncal vagotomy and drainage. Gut 1982; 23: 569-76. 10. Warshaw AL. Bile gastritis without prior gastric surgery: con- tributing role of cholecystectomy. Am J Surg 1979; 137: 527-31. 11. Kalima W. Reflux gastritis unrelated to gastric surgery. Scand J Gastroenterol 1982; 17(Suppl 79): 66-71. 12. Svennson JO, Gelin J, Svanvik J. Gallstones, cholecystectomy and duodenogastric reflux of bile acid. Scand J Gastroenterol 1986; 21: 181-7. 13. Cheadle WG, Patti V, Mackie CR, Cushieri A. Effect of gallbladder function on duodenogastric reflux. Gut 1984; 25:1138. 14. Miranda M, Deflllipi C, Valenzuela JE. Abnormalities of inter- digestive motility complex and increased duodenogastric reflux in gastric ulcer patients. Dig Dis Sci 1985; 30: 16-21. 15. Keane FB, Dimaguo EP, Malagelada JR. Duodenogastric re- flux in humans: its relationship to fasting antroduodenal motility and gastric, pancreatic and bilial secretion. Gastroenterology 1981; 81: 726-31. 16. Rces WDW, Go VLW, Malagelada JR. Simultaneous mea- surement of antroduodenal motility, gastric emptying and duodeno- gastric reflux in man. Gut 1979; 20: 963-70. 17. Thompson JC. Humoral control of gut function. Am J Surg 1991; 161: 6-18. 18. Adrian TE, Long RG, Fuessl HS, Bloom SR. Plasma peptide YY (PYY) in the dumping syndrome. Dig Dis Sci 1985; 89: 1145-8. 19. Blackburn AM, Christofides ND, Ghatei MA, et al. Elevation of plasma neurotensin in the dumping syndrome. Clin Sci 1980; 59: 237-43. 20. Besterman HS, Adrian TE, Mallinson CN, et at. Gut hormone release after intestinal resection. Gut 1982; 23: 854-61. 21. Sirinek KR, O'Dorisio TM, Howe B, McFee AJ. Neurotensin, vasoactive intestinal peptide, and Roux-en-Y gastrojejunostomy. Their role in the dumping syndrome. Arch Surg 1985; 120: 605-9. 22. Hopman WPM, Jansen JBMJ, Lamers CHHW. Plasma cho- lecystokinin response to oral fat in patients with Billroth I and Billroth II gastrectomy. Ann Surg 1984; 199: 276-80. 23. Adrian TE, Savage AP, Fuessl HS, et al. Release of peptide YY (PYY) after resection of small bowel, colon or pancreas in man. Surgery 1987; 17: 95-9. 24. Labo G, Bortolotti P, Vezzadini G, Bonora G, Bersani G. Interdigestive gastroduodenal motility and serum motilin levels in patients with idiopathic delay in gastric emptying. Gastroenterolo- gy 1986; 90: 20-6. 25. Preston DM, Adrian TE, Christofides ND, Lennard-Jones JE, Bloom SR. Positive correlation between symptoms and circulating motilin, pancreatic polypeptide and gastrin levels in functional bow- el disease. Gut 1985; 26: 1059-64. 26. Achem-Karam SR, Funakoshi A, Vinik AI, Owyang C. Plas- ma motilin concentration and interdigestive migrating motor com- plex in diabetic gastroparesis: effect of metoclopramide. Gastroen- terology 1985: 88: 492-9. 27. Tamalea M, Sarr M, VanLier Ribbink J. Gastro-intestinal motor patterns: motilin as a coordinating factor. J Surg Res 1989; 147: 325-31. 28. Keinld O, Wulschke S, Ehrlein HJ. Neurotensin slows gastric emptying by a transient inhibition of gastric and a prolonged inhibi- tion of duodenal motility. Digestion 1986; 34: 281-8. 29. Thor K, Rossell S, Rokaeus A, Kager L. (GLN4)-nenrotensin changes the motility pattern of the duodenum and proximal jeju- num from a fasting-type to a fed-type. Gastroenterology 1982; 83: 569-74. 30. Savage AP, Adrian TE, Carolan G, Chatterjee VK, Bloom SR.

Effects of peptide YY (PYY) on mouth to caecum intestinal transit time and on the rate of gastric emptying in healthy volunteers. Gut 1987; 70: 166-70. 31. Thor K. Rosell S. Neurotensin increases colonic motility. Gas- troenterology 1986; 90: 27-31. 32. Liddle RA, Gertz B J, Kanayama S, et al. Effects of a novel cholecystoldnin (CCK) receptor antagonist, MK-329, on gallblad- der contraction and gastric emptying in humans. Implications for the physiology of CCK. J Clin Invest 1989; 84: 1220-5. 33. Fuchs KH, DeMeester TR, Hinder RA, Stein H J, Barlow AP, Gupta NC. Computerized identification of pathologic duodenogas- tric reflux using 24-hour gastric pH monitoring. Ann Surg 1991; 213: 13-20. 34. Bloom SR, Long RG, editors. Radioimmunoassay of gut regu- latory peptides. Philadelphia: WB Saunders, 1982. 35. Joekel CS, Herrington MK, Vanderhoof JA, Adrian TE. Post- natal development of circulating cholecystokinin and secretin, pan- creatic growth, and exocrine function in guinea pigs. Int J Pancrea- tol; In press. 36. Adrian TE, Ferri G-L, Bacarese-Hamilton A J, Fuessl HS, Polak JM, Bloom SR. Human distribution and release of a putative new gut hormone peptide YY. Gastroenterology 1985; 89: 1070-7. 37. Fujimura M, Khalil T, Sakamoto "[, et al. Release of neuroten- sin by selective perfusion of the jejunum with oleic acid in dogs. Gastroenterology 1989; 96: 1502-5. 38. Adrian TE, Savage AP, Sagol GR, et al. Effect of PYY on gastric, pancreatic and biliary function in humans. Gastroenterolo- gy 1985; 89: 484-99. 39. Adrian TE, Savage AP, Bacarese-Hamilton A J, Wolfe K, Besterman HS, Bloom SR. Peptide YY abnormalities in gastroin- testinal diseases. Gastroenterology 1986; 90: 379-84. 40. Ballantyne GH, Longo WE, Adrian TE, et al. Deoxycholate stimulated release of peptide YY from the isolated perfused rabbit left colon. Am J Physiol 1989; 20: G715-24.

D I S C U S S I O N O. E. Akwar i (Durham, NC) : Is it possible to discuss

hormonal responses and gastric dysfunction without con- sidering intestinal t ransit? Did you measure intestinal transit parameters in these patients?

Pau l Wilson: W e did not study intestinal transit in these patients, only gastric emptying.

This rapid phase of gastric emptying, which is com- mon in other syndromes such as the postgastrectomy dumping syndrome, is associated with the elevation of many of the proximal and distal gut hormones such as neurotensin, peptide YY, and enteroglucagon. Bile diver- sion produced a normalization, especially of neurotensin, which we correl~ited quite closely with gastric emptying. In the early period of gastr ic emptying, neurotensin levels were most elevated.

John M. K e l l u m (Richmond, VA): Steve Bloom in London identified enteroglucagon as a marker for the dumping syndrome, part icular ly to a glucose challenge. W e had similar results after Roux-en-Y gastric bypass. Do your findings with enteroglucagon indicate duodeno- gastric reflux (DGR) or dumping in the fed state? More importantly, what are the indications for treating this disease surgically?

Pau l Wilson: W e found that the enteroglucagon lev- els in pr imary D G R were elevated similarly to those in the dumping syndrome, but the magni tude of the changes in these patients was somewhat lower than those found in

176 THE AMERICAN JOURNAL OF SURGERY VOLUME 165 JANUARY 1993

GUT HORMONES IN DUODENOGASTRIC REFLUX

the dumping syndrome. The switch operation did not normalize enteroglucagon levels. Intestinal transit may still be a little rapid, and this could be responsible for the persistence of elevated enteroglucagon levels. This group of patients may represent one end of a spectrum of disease that ranges from physiologic to pathologic. DGR is not an easy condition to diagnose.

William Silen (Boston, MA): What tests would ac- complish that goal?

Paul Wilson: In patients with gastritis, we performed 24-hour gastric pH monitoring. Drs. DeMeester and Fuchs developed a score that discriminates between phys- iologic and pathologic DGR.

A simplified score consists of measuring the interdi- gestive phase of the gastric pH record. Patients with over 10% of the fasting period above pH 3 have abnormal DGR.

We also use technetium HIDA (0-diisopropyl imino- diacetic acid) scanning to measure DGR. We also use a composite histology score, which takes into consideration whether the patient has hyperplasia, acute and chronic inflammatory changes, superficial mucosal capillary changes, etc. We use several objective tests.

Gerald M. Lal'~,on (Louisville, KY): You concluded that these hormonal changes are compensatory rather than causal. Have you studied the hormonal pattern in patients who have had a pyloroplasty with alkaline reflux or duodenal reflux with gastritis?

Also, what is the composition of the test meal? Paul Wilson: We studied two or three patients after

pyloroplasty, and they had changes that were very similar to those of the patients with primary DGR. They had a rapid early phase of gastric emptying with abnormal lev- els of neurotenSin,' insulin, and enteroglucagon.

The composition of the test meal for the gastric emp- tying studies was an oatmeal preparation mixed with water and labeled with sulfur colloid. Patients also drank a pint of milk. This was a semisolid gastric emptying study.

B. E. Terry (Columbia, MO): What are the patients' eating habits after the procedure?

Paul Wilson: In a review of all 42 patients undergoing the procedure at Creighton University School of Medi- cine, pain and bilious vomiting were relieved in the major- ity of patients. Over 80% have a very good result after surgery, which either abolishes the symptoms or signifi- cantly reduces them.

We do not have the data to determine whether the operation improves weight loss.

Carlos A. Pellegrini (San Francisco, CA): Is there a relationship between gastric emptying and the develop- ment of mild gastritis? We presented data on gastric emptying before and after bile diversion with a Roux-en- Y, and most of our patients have had a previous gastrec- tomy. About one third of our patients with a previous gastrectomy have clear evidence of a rapid gastric empty- ing, but two thirds with bile gastritis did not. Why would the patient with rapid gastric emptying develop bile gas- tritis? Intuitively, one would think that those patients would get rid of bile much faster, which would presum- ably protect their mucosa.

Paul Wilson: We theorize that a two-way pyloric incontinence exists in the patient with an incontinent pylorus. During fasting, DGR occurs, and during the postprandial period, there is rapid gastric emptying.

A. C. Almeida (Lisbon, Portugal): Were you able to demonstrate reversal of histologic changes after the di- version?

Paul Wilsom Over 35 of these patients were reviewed before and after diversion using the composite alkaline gastritis histologic score described by Dixon. There was no change in the histologic score, and the mean period of postoperative evaluation was about 2 years.

William Silen: What is the composition of the reflux- ate? Is there an abnormal bile acid and bile salt composi- tion in those patients with abnormal DGR?

Paul Wilson: We have not specifically looked at the composition of the refluxate in these patients.

THE AMERICAN JOURNAL OFSURGERY VOLUME 165 JANUARY 1993 177