THERAPEUTIC REVIEW 588 #{149} J ClIn Pharmacol 1991;31:588-598 Nonsteroidal Anti-Inflammatory Drugs: Effects on Kidney Function Andrew Whelton, MD, FACP, F’CP, and Cindy W. Hamilton, PharmD Nonsteroidal anti-inflammatory drugs (NSAIDs) are capable of inducing a variety of renal function abnormalities, particularly in high-risk patients with decreased renal blood perfusion who depend on prostaglandin synthesis to maintain normal renal func- tion. Fluid retention is the most common NSAID-related renal complication, occurring to some degree in virtually all exposed individuals; however, clinically detectable edema occurs in less than 5% of patients and is readily reversible on discontinuation of the NSAID. Other electrolyte complications, notably hyperkalemia, are seen infrequently and occur in specific at-risk patients. The next most worrisome complication is acute deterioration of renal function, which occurs in high-risk patients and is also reversible. Nephrotic syndrome with interstitial nephritis is a rare problem of NSAID use and is reversible. Papillary necrosis is the only permanent complication of NSAIDs and is very rare. Altogether, these renal function abnormalities, with the exception of mild fluid retention, are clinically detectable in approximately 1% of exposed patients. Given the number of patients who take NSAIDs on a prescription or over-the-counter basis, the absolute number of at-risk patients is relatively large. Consequently, an appreciation for the risk factors and pathophysiology of NSAID-induced renal function abnormalities is required for optimal use of these drugs. A pproximately 1-5 of persons who are exposed to a nonsteroidal anti-inflammatory drug (NSAID) will manifest one of a variety of renal func- tion abnormalities. Although this percentage ap- pears relatively low, the number of at-risk individ- uals is enormous because of the current use profile of NSAIDs, either as prescription or over-the-counter drugs. One in seven Americans is likely to be treated with an NSAID for a chronic rheumatologic disorder. If patients who take NSAIDs for acute problems are considered, the exposure rate will be even higher. Thus, of the 50 million Americans expected to use NSAIDs intermittently or routinely this year, at least 500,000 are likely to develop some degree of renal functional abnormality. In descending order of frequency, the primary NSAID-related renal abnormalities are 1) fluid and electrolyte disturbances, 2) acute deterioration of renal function, 3) nephrotic syndrome with intersti- From the Department of Medicine (Dr. Whelton), Johns Hopkins Univer- sity School of Medicine, Baltimore, Maryland, and Virginia Beach (Dr. Hamilton), Virginia. Address for reprints: Andrew Whelton, MD, The Johns Hopkins Hospital, 1830 East Monument Street, Rm 815, Balti- more, MD 21205. tial nephritis, and 4) papillary necrosis (Table I). So- dium chloride and water retention, the most com- monly encountered renal effect of NSAID use, occurs to some degree in virtually all exposed persons but results in clinically detectable edema in less than 5% of patients. This rate is probably higher in selected at-risk patients. NSAID-induced fluid retention is typically benign, reversible on discontinuation of the NSAID, and easily managed in patients who require treatment. Other electrolyte abnormalities are also induced by NSAIDs, the most important of which is potassium retention and hyperkalemia. A high-risk group can also be identified for this electrolyte abnor- mality. From the clinical point of view, the most worri- some renal side effect of NSAIDs is hemodynami- cally mediated acute renal failure, which occurs in individuals with pre-existing reduced renal blood perfusion. Ordinarily, the kidneys of such at-risk pa- tients produce vasodilatory prostaglandins to main- tain renal perfusion and function. The inhibitory ef- fects of NSAIDs on renal prostaglandin production lead to acute, reversible renal failure in these pa- tient. Acute deterioration of renal function occurs in 0.5 to 1% of patients who take NSAIDs on a chronic basis.
Transcript
THERAPEUTIC REVIEW
588 #{149}J ClIn Pharmacol 1991;31:588-598
Nonsteroidal Anti-Inflammatory Drugs:Effects on Kidney Function
Andrew Whelton, MD, FACP, F’CP, and Cindy W. Hamilton, PharmD
Nonsteroidal anti-inflammatory drugs (NSAIDs) are capable of inducing a variety ofrenal function abnormalities, particularly in high-risk patients with decreased renalblood perfusion who depend on prostaglandin synthesis to maintain normal renal func-tion. Fluid retention is the most common NSAID-related renal complication, occurring tosome degree in virtually all exposed individuals; however, clinically detectable edemaoccurs in less than 5% of patients and is readily reversible on discontinuation of theNSAID. Other electrolyte complications, notably hyperkalemia, are seen infrequentlyand occur in specific at-risk patients. The next most worrisome complication is acutedeterioration of renal function, which occurs in high-risk patients and is also reversible.Nephrotic syndrome with interstitial nephritis is a rare problem of NSAID use and is
reversible. Papillary necrosis is the only permanent complication of NSAIDs and is veryrare. Altogether, these renal function abnormalities, with the exception of mild fluidretention, are clinically detectable in approximately 1% of exposed patients. Given the
number of patients who take NSAIDs on a prescription or over-the-counter basis, the
absolute number of at-risk patients is relatively large. Consequently, an appreciation forthe risk factors and pathophysiology of NSAID-induced renal function abnormalities isrequired for optimal use of these drugs.
A pproximately 1-5 of persons who are exposedto a nonsteroidal anti-inflammatory drug
(NSAID) will manifest one of a variety of renal func-tion abnormalities. Although this percentage ap-pears relatively low, the number of at-risk individ-uals is enormous because of the current use profile ofNSAIDs, either as prescription or over-the-counterdrugs. One in seven Americans is likely to be treatedwith an NSAID for a chronic rheumatologic disorder.If patients who take NSAIDs for acute problems areconsidered, the exposure rate will be even higher.Thus, of the 50 million Americans expected to useNSAIDs intermittently or routinely this year, at least500,000 are likely to develop some degree of renalfunctional abnormality.
In descending order of frequency, the primaryNSAID-related renal abnormalities are 1) fluid andelectrolyte disturbances, 2) acute deterioration ofrenal function, 3) nephrotic syndrome with intersti-
From the Department of Medicine (Dr. Whelton), Johns Hopkins Univer-sity School of Medicine, Baltimore, Maryland, and Virginia Beach (Dr.Hamilton), Virginia. Address for reprints: Andrew Whelton, MD, TheJohns Hopkins Hospital, 1830 East Monument Street, Rm 815, Balti-
more, MD 21205.
tial nephritis, and 4) papillary necrosis (Table I). So-dium chloride and water retention, the most com-monly encountered renal effect of NSAID use, occursto some degree in virtually all exposed persons butresults in clinically detectable edema in less than 5%of patients. This rate is probably higher in selectedat-risk patients. NSAID-induced fluid retention istypically benign, reversible on discontinuation of theNSAID, and easily managed in patients who requiretreatment. Other electrolyte abnormalities are alsoinduced by NSAIDs, the most important of which ispotassium retention and hyperkalemia. A high-riskgroup can also be identified for this electrolyte abnor-mality.
From the clinical point of view, the most worri-some renal side effect of NSAIDs is hemodynami-cally mediated acute renal failure, which occurs inindividuals with pre-existing reduced renal bloodperfusion. Ordinarily, the kidneys of such at-risk pa-tients produce vasodilatory prostaglandins to main-tain renal perfusion and function. The inhibitory ef-fects of NSAIDs on renal prostaglandin productionlead to acute, reversible renal failure in these pa-tient. Acute deterioration of renal function occurs in0.5 to 1% of patients who take NSAIDs on a chronicbasis.
NONSTEROIDAL ANTI-INFLAMMATORY DRUGS
THERAPEUTIC REVIEW 589
TABLE I
Documented Renal Effects of Nonsteroldal Anti-Inflammatory Drugs
Renal Effects*Brand Name!
Drug ClassGeneric Name Manufacturer Edema tIc ARF NS PN
Salicylates Aspirin (various) Ci CiDiflunisal Dolobid/Merck Ci CI
Propionic acids Ibuprofen Motrin/Upjohn CI CI CiNaproxen Naprosyn/Syntex Ci ClFenoprofen calcium Nalfon/Liily Ci CIKetoprofen Orudis/Wyeth-Ayerst Ci ClFiurbiprofen Ansaid/Upjohn Ci Ci Cl
Indolacetic acids Indomethacin indocin/Merck Ci Ci ClSuiindac Clinoril/Merck Ci Ci ClTolmetin Tolectin/McNeil Cl ClDiclofenac Voitaren/Cibia.Geigy Cl Cl
Pyrazolones Phenylbutazone Butazolidin/Ciba-Geigy Cl CiOxicams Piroxicam Feldene/Pfizer Cl Cl Ci
ClCiClClCiCitClCiCICICIClClCl
ClAnClAnCl
AnCItAnAnClAnClClCi
* ARF = acute renal failure; NS = interstitial nephritis and nephrotic syn t Causes interstitial nephritis without nephrotic syndrome.
drome; PN = papillary necrosis; fK hyperkalemia; Cl = reported in clinical Reported in combination with phenylbutazone.studies; An = described in studies in animals (but not in humans). (Adapted from Clive and Stoff,’ with permission.)
The nephrotic syndrome, with associated intersti-tial nephritis, is seen on rare occasions. Once again,it is reversible on discontinuation of the NSAID inquestion.
According to the respective manufacturers’ pre-scribing information, chronic administration ofnearly all NSAIDs produces papillary necrosis in lab-oratory animals; and a few clinical case reports ofpapillary necrosis can be found in the recent medicalliterature. Within the framework of our present un-derstanding of NSAID effects on the kidney, this ap-pears to be the only irreversible form of renal tox-icity.
Many of the renal abnormalities that are encoun-tered as a result of NSAID use can be attributed to theaction of these drugs on prostaglandins. Hence, abrief overview of the interactions between prosta-glandins and renal function will be presented, fol-lowed by an analysis of the pathophysiology, clinicalmanifestations, patient risk factors, and preventiveapproaches to NSAID-induced renal syndromes.
THE PROSTAGLANDIN PATHWAY
Prostaglandins are ubiquitous substances that influ-ence renal function along with a variety of otherbody systems.1’2 Conceptually, they may be consid-ered local hormones or “autocoids” because they actin a paracrine or autocrine fashion. Biologic activityis limited to the site of action by the short half-life of
prostaglandins in circulation. In addition, prostaglan-dins are not stored in tissue, but are synthesized ondemand.
Prostaglandins are derived from phospholipids bya common pathway (Figure 1). Phospholipids, ofcourse, are widely distributed in cell membranesthroughout the body. The most important precursorfor prostaglandins is arachidonic acid. Cyclooxygen-ase is the catalyst for oxygenation of arachidonicacid, which is the step that is inhibited by NSAIDs.The interaction between aspirin and cyclooxygenase(acetylation) is irreversible, whereas that with otherNSAIDs is reversible.
Arachidonic acid can also be metabolized to othermediators, depending on the cell type. For example,lipoxygenase catalyzes the production of leuko-trienes, and mixed-function oxygenases catalyze theproduction of epoxyeicosatrienoic acids. Collec-tively, these oxygenated metabolites of arachidonicacid are known as eicosanoids because of their originfrom a 20-carbon (eicosa-) polyunsaturated acid.3
Continuing along the common pathway (Figure 1),oxygenation of arachidonic acid results in produc-tion of prostaglandin G2, which is converted to prosta-glandin H2 by hydroperoxidase and loss of a free radi-cal. At this point, metabolism becomes highly spe-cific for individual cell types, although many, if notall, of the metabolites are produced in the kidney.Prostaglandin E2 is a vasodilator, which, in the kid-ney, promotes diuresis and natriuresis. Prostaglan-
Figure 1. Prostaglandin pathway (and prostanoid functions). Pg = prostaglandin; t = stimulate or increase;
WHELTON AND HAMILTON
590 #{149}J COn Pharmacol 1991;31:588-598
Phospholipids(found in cell
membranes)
1[Achidonic
acid (an“e’,cosano’ud”)
I Cyclooxygenase and 02, (inhibited by NSAIDs)
fPgG2 I
free radical
PgH2
= inhibit or decrease.2’3
din E2 also inhibits lymphocytes and other cells thatare involved in inflammation and allergic responses,which, as will be discussed later, may play a role insome NSAID-induced renal syndromes. Prostaglan-din F2,, enhances excretion of sodium chloride andwater. Prostacyclin, also known as prostaglandin 12,
has a wide variety of actions including vasodilation,renin release, and inhibition of platelet aggregation.Prostaglandin D2 is a vasodilator of peripheral resis-tance vessels but is better known for its associationwith mast cell activation and bronchoconstriction.Thromboxane A2 is the principal metabolite of pros-taglandin H2 in platelets and can act as a major vaso-constrictor within the kidney. These pharmacologi-cally active metabolites of prostaglandin H2 are col-lectively known as prostanoids.3
PROSTAGLANDIN EFFECTS ON RENAL
FUNCTION
Given the diversity of cell populations within thekidney and their various functions, the complexityof the interactions between prostaglandins and renalfunction is not unexpected. Prostaglandins are in-volved in renin release, local vascular tone, regional
circulation, sodium and water homeostasis, and po-tassium balance (Table H). The following sections de-scribe these diverse effects. Detailed overviews ofthese interactions can be found in excellent reviewsby Patrono and Dunn2 and Oates and colleagues.3
An important caveat in the following sections isthat prostaglandins are not primary mediators of ba-sal renal function in normal individuals. Prostaglan-dins typically operate in conjunction with a varietyof other mediators, which, even in the absence ofprostaglandins, can preserve homeostasis. Prosta-glandin production is increased as needed in re-sponse to stress (e.g., decreased renal blood flow orblood volume). Thus, inhibition of prostaglandinfunction by NSAIDS is more likely to cause compli-cations in at-risk patients with decreased renal bloodperfusion than in the otherwise normal subjectwhose prostaglandins are merely one of many factorscontributing to homeostasis.
Renin Release
Prostaglandins stimulate renin release, which playsan important role in the regulation of arterial blood
NONSTEROIDAL ANTI-INFLAMMATORY DRUGS
THERAPEUTIC REVIEW 591
TABLE II
Principal Renal Sites of Prostaglandin Synthesisand Major Actions
(Adapted from Pa trono and Dunn2, with permission.)
pressure, blood volume, and electrolyte balance.Prostaglandins can act independently or synergisti-cally with the $-adrenergic system.4 Although theexact prostanoid mediator is not yet known, it islikely that prostacyclin is synthesized in response toa change in arteriole pressure or chloride reabsorp-tion in the macula densa of the nephron.3
Local Vascular Tone
Prostanoids are one of several local mediators thatgovern vascular tone through their actions on norepi-nephrine release at peripheral nerve endings. Prosta-glandins E2 and D2 and, to a lesser extent, prostacy-din promote vasodilation by inhibiting norepineph-rine release. Prostaglandin E2 also antagonizes theeffects of angiotensin II, a powerful vasopressor, onthe neuroeffector junction. Conversely, prostaglan-din F2,, and thromboxane A2 are vasoconstrictors.3
Regional Circulation
Sodium and Water Homeostasis
All prostanoids are capable of acting in the renal cor-tex to regulate sodium and water homeostasis; how-ever, prostanoids are only one of many factors thatshare this function.3 Prostaglandins E2 and D2, pros-tacyclin, and, to a lesser extent, prostaglandin F2a
increase the rate of salt and water excretion. Prosta-glandin E2 inhibits sodium chloride transport in thethick ascending limb of the loop of Henle and thecollecting duct.5’6 In addition, prostaglandins antago-nize the effects of antidiuretic hormone.7’8
Prostanoids do not have a direct effect on glomeru-lar filtration rate; however, vasodilation associatedwith prostaglandin E2, prostacyclin, and prostaglan-din D2 increases renal blood flow, and, as previouslymentioned, shunts blood flow from the cortical tojuxtamedullary nephrons. The net result is en-hanced diuresis and natriuresis due to reduced med-ullary hypertonicity and increased interstitial pres-
3sure.
Potassium Balance
Prostanoids indirectly lower potassium by their ef-fects on glomerular filtration and renin.3 Aspreviously mentioned, vasodilatory prostaglandinsincrease renal blood flow. This may enhance the di-rect intratubular delivery of potassium into the dis-tal nephron for excretion. Alternatively, this mayserve to quantitatively increase sodium delivery intothe distal nephron with resultant reabsorption of so-dium in exchange for potassium, which is then ex-creted in the urine. Secondly, prostacyclin is be-lieved to promote renin release. Activation of therenin-angiotensin pathway ultimately causes aldo-sterone to stimulate potassium excretion in the distalconvoluted tubule and collecting duct. However, po-tassium balance is also regulated by a number ofother factors such as insulin and the 9-adrenergicsystem.
FLUID AND ELECTROLYTE DISTURBANCES
Prostanoids contribute to regional circulation viatheir influence on local vascular tone. Under normalconditions, prostanoids do not regulate renal perfu-sion per se. However, certain conditions such as de-creased renal blood flow are associated with the pro-duction of vasodilatory prostaglandins. Prostaglan-din E2, prostacyclin, and prostaglandin D2 shiftregional blood flow from cortical to juxtamedullarynephrons.3
Sodium and Water Retention
The most common and universal renal complica-tions of NSAIDs are sodium retention and edema.According to prescribing information accompanyingnearly all NSAIDs, edema occurs in at least 3% ofpatients. The incidence is probably higher in pa-tients who take therapeutic doses over prolonged pe-riods. The onset of fluid retention usually occursearly in the course of therapy and can be dramatic as
WHELTON AND HAMILTON
592 #{149}J COn Pharmacol 1991;31:588-598
illustrated by the 15-kg weight gain in a 70-year-old
man who took ibuprofen for only 17 days.9Occasionally, the patient may retain water in ex-
cess of sodium. Severe, reversible hyponatremia (118
tmol Na/L) occurred in a patient who took ibupro-fen for only 3 days. This patient had underlying renalimpairment (CrC1 12 mL/min).1#{176}
The multiple mechanisms by which NSAIDs in-terfere with water and sodium metabolism may ex-plain the frequency of this complication. Aspreviously mentioned, NSAIDs have the potential todisrupt diuresis and natriuresis by interfering withprostaglandin-mediated sodium chloride transport,antidiuretic hormone, and distribution of blood flowfrom cortical to juxtamedullary nephrons.13 The hy-pothesis for the pathogenesis of the nephrotic syn-drome is also operative in this situation. By shuntingarachidonic acid metabolism from prostaglandins tolipoxygenase products, NSAIDs may favor produc-tion of eicosanoid derivatives that increase capillarypermeability.1
Hyperkalemia
Hyperkalemia is an unusual complication ofNSAIDs, presumably because of the multiplicity offactors that are capable of maintaining potassium bal-ance, even in the absence of prostaglandins. Hyper-kalemia is more likely to occur in patients with pre-existing renal impairment,11’12 cardiac failure,13 dia-betes,12 or multiple myeloma’14 or in patients whoreceive potassium supplementation,15 potassium-sparing diuretics,16 or angiotensin-converting en-zyme (ACE) inhibitors. Indomethacin appears to bethe major NSAID associated with this complicationand has produced hyperkalemia in patients withoutapparent risk factors.17 Thus, indomethacin mayhave a direct effect on the cellular uptake of potas-sium,18 in addition to the known effects of NSAIDs onpotassium delivery to the distal tubule as well as onthe renin-angiotensin and aldosterone pathways.
NSAID-induced hyperkalemia often occurs in thesetting of NSAID-induced acute renal deteriorationor worsening of underlying renal impairment. How-ever, the severity of hyperkalemia can be dispropor-tionate to that of renal impairment. For example,Tan and colleagues reported a patient who wastreated with indomethacin and had a serum potas-sium of 6.2 mEq/L in spite of only mildly abnormalrenal function.1#{176} In this patient, plasma renin andaldosterone levels were suppressed and did not re-spond to furosemide or postural changes. Urinaryprostaglandin E2 was also suppressed. Discontinua-tion of indomethacin resulted in normalization of po-
tassium, prostaglandin E2, and a rebound of renin
and aldosterone.
ACUTE DETERIORATION OF RENAL FUNCTION
Role of Prostanoids in MaintainingRenal Blood Flow
Although NSAIDs do not impair glomerular filtrationin normal individuals,20’21 acute renal decompensa-tion may occur in at-risk patients with various extra-renal or renal disease processes that lead to de-creased renal perfusion (Table III). Renal prostaglan-dins play an important role in the maintenance ofhomeostasis in these patients, so drug-induceddisruption of counter-regulatory mechanisms canproduce clinically important and even severe renalfunctional deterioration.2’3
Acute renal deterioration in this setting can be at-tributed to the interruption of the delicate balancebetween hormonally mediated pressor mechanismsand prostaglandin-related vasodilatory effects (Fig-ure 2). In at-risk patients, volume contraction trig-gers pressor responses via adrenergic and renin-an-giotensin pathways. Ordinarily, vasodilatory renalprostaglandins counterbalance the vasoconstrictiveeffects of norepinephrine and angiotensin H. The ad-dition of NSAIDs increases the risk of azotemia andpossibly ischemic damage to the kidney by removingthe protective effects of vasodilatory prostaglandinsand allowing unopposed vasoconstriction.
Clinical Features of Acute Renal Failure
Initially, this NSAID-induced renal syndrome is ofmoderate severity and is characterized by increasingBUN, creatinine, potassium, and weight with de-creasing urine output. NSAID-induced acute renalfailure is usually reversible over 2 to 7 days afterdiscontinuation of therapy; however, morbid conse-quences can occur if the diagnosis is not recognizedearly. Continued NSAID therapy in the setting of de-
TABLE iii
At-Risk Patients for NSAID-inducedAcute Renal Failure
Figure 2. Mechanism by which NSAIDs disrupt the compensatoryvosodilation response of renal prostaglandins to vasoconstrictorhormones in patients with prerenal conditions. A solid line (-)
indicates stimulation; a dashed line (---) indicates inhibition,
teriorating renal function may progress rapidly tothe point wherein dialysis support is required.22 De-spite this profound level of renal functional impair-ment, the kidney will nonetheless recover severaldays to weeks after discontinuation of the NSAID.Development of this type of “total” renal failure,which is often inappropriately designated as “acutetubular necrosis,” represents the extreme end of thespectrum of hemodynamic insult rather than a sepa-rate clinical entity.
Risk Factors for Acute Renal Failure
The risk of acute renal deterioration is highest inpatients with liver disease, pre-existing renal im-pairment, cardiac failure, protracted volume contrac-tion due to diuretic therapy or intercurrent disease,or old age. NSAID-induce renal decompensation hasbeen well documented in patients with cirrhosis, par-
ticularly when ascites is present.3 Urinary excretionof prostaglandin E2, prostacyclin metabolites, andthromboxane A2 is increased in these patients.23’24An analogous situation exists in patients with under-lying congestive heart failure,25 nephrotic syn-drome,26’27 or lupus nephritis.28’29
Patients with chronic renal impairment are at in-creased risk of NSAID-induced renal failure becauseof inadequate renal prostaglandin production. Wedocumented NSAID-induced acute renal failure inpatients with asymptomatic mild, but chronic, renalfailure (serum creatinine between 1.5 and 3.0 mg/dL).3#{176}Baseline excretion of urinary prostaglandin E2and 6-keto-prostaglandin Fia was quantitativelylower in the individuals who developed NSAID-in-duced renal decompensation than in those who didnot, and ibuprofen proved to be more problematicthan either piroxicam or sulindac. On initiation ofibuprofen, urinary prostaglandin excretion fell in allpatients, but trough concentrations were quantita-tively lower in the subset of patients who experi-enced acute renal failure.
Volume contraction due to diuretic therapy or anintercurrent disease that results in dehydration rep-resents another important risk factor for the deve lop-ment of NSAID-induced acute deterioration of renalfunction.22’31-32 Elderly patients are also at increasedrisk. We estimate that age of 80 years or greater isan independent risk factor because the physiologyof ageing within the kidney results in 50% loss offunction in 50% of the population at age 80, primar-ily as a result of the progression of arteriolonephros-clerosis.
Pharmacodynamics of Acute Renal Failure
NSAID-induced acute renal decompensation is apharmacologically predictable phenomenon that oc-curs in a dose-related fashion. In our triple-crossoverstudy of 12 women with mild renal failure, ibuprofen(800 mg three times daily) was discontinued on day 8
because of worsening renal function (�1.5 mg/dL in-crease in serum creatinine) or hyperkalemia (potas-sium � 6 mEq/mL) in 3 patients. When these pa-tients were rechallenged at a 50% lower dose of ibu-profen, two patients again had evidence of acuterenal deterioration.30
Another important finding in our study was thetime of onset of acute renal decompensation.3#{176} Ibu-profen-induced renal failure occurred rapidly(within days), but piroxicam and sulindac did notcause renal deterioration during the 11-day treat-ment period. A pharmacokinetic analysis in thesepatients provides insight. Ibuprofen, which has ashort elimination half-life, reached maximum serum
WHELTON AND HAMILTON
594 #{149}.1 ClIn Pharmacol 1991;31:588-598
concentrations quickly. In contrast, piroxicam andsulindac have longer half-lives and continued to ac-cumulate throughout the treatment period. Thesefindings are consistent with basic pharmacologicprinciples and suggest that NSAIDs having shortelimination half-lives will reach steady state and ex-ert maximum pharmacologic effects before NSAIDshaving longer half-lives.
“Renal Sparing” NSAIDs -?
Although all NSAIDs have the potential to induceacute renal impairment, some quantitative differ-ences may exist. Sulindac has been hypothesized tobe renal sparing, possibly because of its unusual met-abolic pathway.33 The parent compound, sulindacsulfoxide, is an inactive prodrug that undergoes he-patic metabolism to sulindac sulfide, which is themetabolite that exerts anti-inflammatory activity.Sulindac sulfoxide is also metabolized to a muchlesser extent to an inactive metabolite, sulindac sul-fone. It has been hypothesized that, within the kid-ney, sulindac sulfide is reversibly oxidized to the in-active parent compound, sulindac sulfoxide, suchthat renal prostaglandin production would not be in-fluenced.
In clinical studies, urinary prostaglandin levelsand renal effects were unchanged in patients withnormal renal function34’35 and states of proteinuria.36However, the duration of sulindac in these studiesmay have been insufficient to appreciate the fullpharmacologic effect of sulindac. NSAID-inducedchanges may not have been detectable because ofthe presence of only very mild renal impairment orabsence of renal failure altogether in these studies.Longer courses of sulindac in patients with slightlymore severe renal impairment have been associatedwith statistically significant reductions in urinaryprostaglandins3#{176} and glomerular filtration rate.37
The ability of sulindac to inhibit prostaglandin syn-thesis and impair renal function has been confirmed
in a different high-risk group, namely patients withhepatic cirrhosis and ascites.38 We have also identi-fied the development of profound acute renal failurein high-risk patients who received sulindac for sev-eral days to weeks. Collectively, these clinical experi-ences indicate the need for cautious and timelymonitoring of high-risk patients who receiveNSAIDs.
NEPHROTIC SYNDROME WITH INTERSTITIALNEPHRITIS
NSAIDs also cause another type of renal dysfunctionthat is associated with various levels of functional
impairment and characterized by the developmentof the nephrotic syndrome with interstitial nephri-tis.1’22’39’4#{176}The clinical features, absence of risk fac-tors, and pathophysiology distinguish this from otherNSAID-induced renal syndromes and from classicdrug-induced allergic interstitial nephritis.
The features of this NSAID-induced renal syn-drome are variable. The patient may experienceedema, oliguria, and/or foamy urine.41 Systemicsigns of allergic interstitial nephritis such as fever,drug rash, peripheral eosinophilia, and eosinophi-luria are generally absent.1’22’40’41 The urine sedimentcontains microscopic hematuria and pyuria.1’41 Pro-teinuria typically is in the nephrotic range.1’39 Wehave noted that renal functional deterioration canrange from minimal to severe.
Characteristically, this form of nephrotic syn-drome consists of minimal change glomerulonephri-tis with interstitial nephritis, which is an unusualcombination of histologic findings. NSAID-inducednephrotic syndrome without interstitial disease israre but has been reported in a handful of patientswho took fenoprofen, sulindac, or diclofenac. Con-versely, interstitial disease without nephrosis hasbeen reported in a few patients, but this may, in fact,represent allergic interstitial nephritis.41
In spite of nephrotic-range proteinuria, the mostimpressive histopathologic findings involve the in-terstitium and tubules. A focal diffuse inflammatoryinfiltrate can be found around the proximal and dis-tal tubules. We reported that the infiltrate primarilyconsisted of cytotoxic T lymphocytes but also con-tained other T cells, B cells, and plasma cells.39Changes in the glomeruli were minimal and resem-bled those of minimal change glomerulonephritiswith marked epithelial-foot process fusion. Other in-vestigators have reported similar findings.1’22’41-42
The onset of NSAID-induced nephrotic syndromeis usually delayed, having a mean time of onset of 5.4months after initiation of NSAID therapy4#{176}and rang-ing from 2 weeks to 18 months.1 NSAID-induced ne-phrotic syndrome is usually reversible 1 month to Iyear after discontinuation of NSAID therapy. Duringthe recovery period, some patients may require dialy-sis. Corticosteroids have been used empirically, butit is not clear whether they hasten recovery.1’22’39 Ifproteinuria does not significantly remit within 2weeks after discontinuation of the NSAID, we recom-mend a standard, 2-month trial of corticosteroid ther-apy as would be employed in a nephrotic adult withidiopathic minimal change or membranous glomeru-lonephritis.
Risk factors are not well understood. Underlyingrenal impairment does not appear to be a risk factor.Old age has been suggested as a risk factor,22’4#{176}but
NONSTEROIDAL ANTI-INFLAMMATORY DRUGS
THERAPEUTIC REVIEW 595
this may also be a reflection of the usual candidatefor chronic NSAID therapy. The syndrome has beenmore commonly reported with fenoprofen thanother NSAIDs. Approximately two-thirds of caseshave been associated with fenoprofen. Hence, thestructure of the drug itself appears to be of majorimportance. The syndrome has been attributed,nonetheless, to virtually all NSAIDs, including thosefrom structurally distinct classes.1-22’39-40’41
The mechanism of NSAID-induced nephrotic syn-drome has not been fully characterized. The associa-tion of this syndrome with structurally distinctNSAIDs suggests a common denominator. T lympho-cytes may function as immune mediators instead ofthe humoral factors that are responsible for classicdrug-induced allergic interstitial nephritis. In keep-ing with this hypothesis, NSAID-induced prostaglan-din inhibition may play an indirect role. By inhibit-ing cyclooxygenase, NSAIDs may promote metabo-lism of arachidonic acid to non-prostaglandineicosanoids. Indeed, leukotrienes, the products ofthe interaction between lipoxygenase and arachi-donic acid, are known to recruit T lymphocytes andpromote the inflammatory process. Leukotrienesmay also contribute to proteinuria by increasing vas-cular permeability.1’40’41
PAPILLARY NECROSIS
Papillary necrosis with interstitial nephritis is awell-known complication of chronic phenacetinabuse that has been reviewed extensively else-where.43 Fortunately, the incidence of the lattercomplication has diminished considerably becauseof a better understanding of the pathophysiology andpatient education. It has been suggested that chronicaspirin alone may also induce papillary necrosis,44but it is not clear that this can actually occur. What isclinically apparent is that chronic (10 to 20 years)exposure of the kidney to high doses of analgesiccombinations such as salicylate and acetaminophen(the metabolite of phenacetin), often with the addi-tion of caffeine, can and will produce chronic, pro-gressive papillary necrosis.
The black pigmentation found within necrotic pa-pillae associated with phenacetin abuse (or phenace-tin-containing combinations) is absent in patientswho ingest aspirin alone or other NSAIDs. This blackpigmentation may represent a breakdown product ofphenacetin.43
In preclinical studies, nearly all of the NSAIDs pro-duced papillary necrosis in experimental animalmodels. Clinical toxicity is exceedingly rare but hasbeen reported for ibuprofen,45 phenylbutazone,46’47
fenoprofen,48 and mefenamic acid,49 and accordingto prescribing information, several other NSAIDs.
The typical candidate for NSAID-induced papil-lary necrosis is a middle-aged woman with a historyof ingesting over-the-counter, combination analge-sics for headache. Closer questioning may revealthat the patient takes the analgesic for the mood-al-tering effects of caffeine. Renal manifestations mayinclude loin pain, macroscopic hematuria, ureteralobstruction, and/or uremia. Urinary tract infectionand hypertension are common secondary findings.Reversibility is determined by the extent of deterio-ration and ability to discontinue NSAID therapy.43Recent reports from the FDA5#{176}of spontaneous grosshematuria associated with NSAIDs such as ibupro-fen (three cases) suggest that papillary necrosis alsooccurs with newer NSAIDs. These data suggest aminor degree of papillary damage, but chronic pro-gressive deterioration of renal function is not a fea-ture of most reports.
The mechanism of NSAID-induced papillary ne-crosis is not clear. The causative role of NSAIDs isdifficult to delineate because of the presence of con-founding factors such as underlying disease, urinarytract infection, and/or concomitant medications. Se-lected NSAIDs may exert a direct toxic effect onrenal papillae, particularly combinations of aspirinand acetaminophen, a major metabolite of phenace-tin. Both drugs are highly concentrated in the me-dulla. Aspirin depletes cellular glutathione, whichwould otherwise neutralize the acetaminophen me-tabolite, N-acetyl-benzo-quinoneimine. Without glu-tathione, this highly reactive metabolite could leadto cell death.43
Prostaglandin inhibition may also play a role.1Medullary ischemia, a possible precipitating factorin development of papillary necrosis, results fromNSAID-induced reduction in blood flow to the renalmedulla in experimental models.51’52
OTHER NSAID-INDUCED RENAL SYNDROMES
Phenylbutazone, suprofen, and benoxaprofen pro-duce unique renal syndromes that are of historic in-terest. These complications are rarely encounteredbecause phenylbutazone use has diminished be-cause of the availability of safer drugs, and suprofenand benoxaprofen have been removed from themarket.
Two mechanisms have been identified for phenyl-butazone-induced acute oligo-anuric renal failure.1Phenylbutazone is known to inhibit uric acid reab-sorption, which may cause hyperuricosuria, and ulti-mately, bilateral ureteral obstruction due to uricacid stones.53 Secondly, an idiosyncratic reaction has
WHELTON AND HAMILTON
596 #{149}J ClIn Pharmacol 1991;31:588-598
been reported that results in acute tubular injurywithout uric acid precipitation.54 Underlying renalimpairment is a risk factor for the latter reaction.Also, patients experiencing this reaction appear to bepredisposed to subsequent renal injury from otherNSAIDs. These observations suggest that prostaglan-din inhibition may play a role in the development ofthe idiosyncratic reaction.1
Suprofen-induced acute renal failure is character-ized by acute flank and/or abdominal pain, occur-ring within 12 hours after starting therapy. In a seriesof 16 patients described by Hart and colleagues,55 themean peak serum creatinine was 3.6 mg/dL (range:2-8 mg/dL) and was within normal limits at follow-up in most patients. Urinalysis revealed microhema-tuna (8/12 patients) and proteinuria (7/12 patients)but no crystals. One of our patients with suprofen-induced flank pain syndrome had birefringent crys-tals in the urine several hours after the injection ofthe drug and at a time when rehydration had alreadybeen commenced. We did not determine if thesecrystals were uric acid or drug metabolites.
The mechanism of suprofen-induced flank painand acute renal failure was never conclusively iden-tified before the drug was removed from the market.No obvious risk factors were identified in theprevious series since all patients appeared to be ingood health and took NSAIDs for acute symptomaticrelief. It has been hypothesized that the suprofenflank pain syndrome is related to acute uric acid
crystal precipitation within the nephron leading toacute urinary flow obstruction.50’55 Suprofen isknown to have uricosuric activity. The finding of hy-peruricemia (mean: 10.8 mg/dL) in four of four pa-tients suggests that this may be a risk factor.55
Benoxaprofen, an NSAID with a long half-life, wasremoved from the market in 1982, within weeksafter its introduction, because of adverse effects. It isremembered for severe hepatic toxicity that occa-sionally resulted in death; however, renal failurewas also a contributing factor. Risk factors for benox-aprofen-induced toxicity were old age and concomi-tant diuretic therapy, two factors known to increasethe risk of acute functional renal failure.
CONCLUSIONS
NSAIDs are considered safe and suitable for thetreatment of a variety of chronic and acute condi-tions. The risk of renal failure after the initiation ofany given NSAID is low; however, the number ofat-risk patients is high because of the widespread useof these drugs.
In most cases, NSAID-induced renal syndromesare a direct or indirect result of prostaglandin inhibi-tion, which has important clinical implications. Atthis time, it is not clear whether it is possible to com-pletely separate the effects of NSAIDs on systemicprostaglandins, which mediate anti-inflammation ac-tivity, from renal effects. Thus, under the right cir-
: angiotensin K-sparing diureticAcute deterioration of 4, Prostaglandin and Liver disease Stop NSAID
renal function disruption ofhemodynamicbalance
Renal diseaseHeart failureDehydrationOld age
Avoid use in high-riskpatients
Nephrotic syndrome with f Lymphocyte Fenoprofen Stop NSAIDinterstitial nephritis recruitment and
activationDialysis and (?)
steroids as neededPapillary necrosis Direct toxicity Phenacetin abuse
Aspirin-acetaminophencombination
Stop NSAIDAvoid chronic
analgesic use
NONSTEROIDAL ANTI-INFLAMMATORY DRUGS
THERAPEUTIC REVIEW 597
cumstances, virtually any NSAID can produce renal
complications. Fortunately, these complications areusually reversible if the diagnosis is recognizedpromptly and NSAID therapy is discontinued.
With an understanding of the pathophysiology in-volved, preventive clinical measures can be put intooperation. Risk factors have been identified for mostNSAID-induced renal syndromes (Table IV). It isprudent to avoid high-dose, chronic NSAID therapyin at-risk patients (Table III). Unfortunately, this isnot always possible. If NSAIDs are necessary in thesehigh-risk groups, the patients should be monitoredclosely and receive appropriate counselling. Moni-toring should begin within a week after initiation of ashort-acting NSAID (e.g., ibuprofen) and continue in-definitely for signs of syndromes having delayed on-set (e.g., nephrotic syndrome with interstitialnephritis).
In the event of NSAID-induced renal failure, theNSAID should be discontinued promptly. The pa-tient should receive supportive care as needed. After
stabilization of renal function, rechallenge with thesame dose of the offending drug or even a struc-turally unrelated NSAID is likely to reproduce theadverse effect. (Patients who have recovered from anepisode of protracted dehydration due to diuretics orintercurrent disease are an exception to this rule.)Thus, if anti-inflammatory therapy is mandatory,underlying risk factors should be identified and elim-inated, if possible. Unfortunately, this is often notpossible, as in the case of old age or chronic heart,kidney, or liver disease. These patients may requirealternative therapy using corticosteroids or othersupportive drugs such as acetaminophen or colchi-cine.
REFERENCES
1. Clive DM, Stoff JS: Renal syndromes associated with nonste-roidal anti-inflammatory drugs. N Engl I Med 1984;310:563-572.
2. Patrono C, Dunn MJ: The clinical significance of inhibition ofrenal prostaglandin synthesis, Kidney mt 1987:32:1-12.
3. Oates JA, FitzGerald GA, Branch RA, Jackson EK, Knapp HR,Roberts U 11: clinical implications of prostaglandin and throm-boxane A2 formation (2 parts). N Engl J Med 1988;319:689-698,761-767.
4. Thames MD, DiBona GF: Renal nerves modulate the secretionof renin mediated by nonneural mechanisms. Circ Res1979;44:645-652.
5. Stokes JB. Effect of prostaglandin E2 on chloride transportacross the rabbit thick ascending limb of Henle: Selective inhibi-tion of the medullary portion, J Clin Invest 1979;64:495-502,
6. Stokes JB, Kokko JP: Inhibition of sodium transport by prosta-glandin E2 across the isolated, perfused rabbit collecting tubule.Clin Invest 1977;59:1099-1104,
7. Orloff J, Handler JS, Bergstrom 5: Effect of prostaglandin (PGEI)on the permeability response of the toad bladder to vasopressin,
theophylline and adenosine 3’, 5’-monophosphate. Nature1965;205:397-398.
8. Anderson RJ, Berl T, McDonald KM. Schrier RW: Evidence foran in viva antagonism between vasopressin and prostaglandin inthe mammalian kidney. I CIin Invest 1975:56:420-426,
15. Akbarpour F, Afrasiabi A, Vaziri ND: Severe hyperkalemiacaused by indomethacin and potassium supplementation. SouthMed 11985:78:756-757.
16. Mar R, Pitlik S, Rosenfeld JB: Indomethacin- and Moduretic-induced hyperkalemia. Isr I Med Sci 1983:19:535-537.
17. Goldszer RC, Coodley EL, Rosner MI, Simons WM, SchwartzAB: Hyperkalemia associated with indomethacin. Arch In tern
Med 1981:141:802-804.
18. MacCarthy EP, Frost GW, Strokes CS: Indomethacin-inducedhyperkalaemia, Med J Aust 1979;1:550,
19. Tan SY, Shapiro R, Franca R, Stockard H, Mulrow PJ: Indo-methacin-induced prostaglandin inhibition with hyperkalemia,Ann Intern Med 1979:90:783-785.
20. Berg KJ: Acute effects of acetylsalicylic acid on renal functionin normal man, Eur I Clin Pharmacol 1977:11:117-123.
21. Donker AJ, Arisz U, Brentjens JR, van der Hem GK, HollemansHJ: The effect of indomethacin on kidney function and plasmarenin activity in man. Nephron 1976:17:288-296.
22. Blackshear JL, Napier JS, Davidman M, Stillman MT: Renalcomplications of nonsteroidal anti-inflammatory drugs: Identifi-cation and monitoring of those at risk. Semin Arthrit fiheum1985:14:163-175.
23. Zipser RD, Hoefs JC, Speckart PF, Zia PK, Horton R: Prosta-glandins: Modulators of renal function and pressor resistance inchronic liver disease. 1 Clin Endocrinol Metab 1979;48:895-900.
24. Zipser RD, Radvan GH, Kronborg IJ, Duke R, Little TE: Uri-nary thromboxane B2 and prostaglandin E2 in the hepatorenalsyndrome: Evidence for increased vasoconstrictor and decreasedvasodilator factors. Gastroenterology 1983:84:697-703.
25. Walshe JJ, Venuto RC: Acute oliguric renal failure induced byindomethacin: Possible mechanisms, Ann Intern Med 1979:91:47-
49’
26. Arisz U, Donker AJM, Brentjens JRH, van der Hem GK: Theeffect of indomethacin on proteinuria and kidney function in thenephrotic syndrome. Acta Med Scand 1976:199:121-125.
27. Kleinknecht C, Brayer M, Cubler M-C, Palcoux J-B: Irrevers-
ible renal failure after indomethacin in steroid-resistant nephro-sis. N Engl 1 Med 1980:302:691.
28. Kimberly RP, Gill JR Jr, Bowden RE, Keiser HR, Plotz PH: Ele-vated urinary prostaglandins and the effects of aspirin on renalfunction in lupus erythematosus. Ann Intern Med 1978:89:336-
34. Brater DC, Anderson 5, Baird B, Campbell WB: Effects of ibu-
profen, naproxen, and sulindac on prostaglandins in men. KidneyInt 1985;27:66-73.
35. Sedor JR, Williams SU, Chremos AN, Johnson CU, Dunn MJ:Effects of sulindac and indomethacin on renal prostaglandin syn-thesis. Clin Pharmacol Ther 1984:36:85-91.
36. Ciabattoni C, Cinotti GA, Pierucci A, et al: Effects of sulindacand ibuprofen in patients with chronic glomerular disease: Evi-
dence for the dependence of renal function on prostacyclin. NEngl J Med 1984:310:279-283.
37. Mistry CD, Late CJ, Gokal R. Currie WJ. Vandenberg M, Mal-
lick NP: Effects of sulindac on renal function and prostaglandinsynthesis in patients with moderate chronic renal insufficiency.
Clin Sci 1986:70:501-505.
38. Quintero E, Gin#{233}sP, Arroyo V, et al: Sulindac reduces the
urinary excretion of prostaglandins and impairs renal function in
patients with cirrhosis and ascites. Nephron 1986;42:298-303.
39. Bender WU, Whelton A, Beschorner WE, Darwish MO, Hall-
Craggs M. Solez K: Interstitial nephritis, proteinuria, and renal
failure caused by nonsteroidal anti-inflammatory drugs: Immuno-
logic characterization of the inflammatory infiltrate. Am I Med
1984;76:1006-101 2.
40. Abraham PA, Keane WF: Glomerular and interstitial disease
induced by nonsteroidal anti-inflammatory drugs. Am J NephrolI 984;4: 1-6.
41. Levin MU: Patterns of tubulo-interstitial damage associated
with nonsteroidal anti-inflammatory drugs. Semin Nephrol1988:8:55-61.
42. Stachura I, Jayakumar S, Bourke E: T + B lymphocyte subsets
in fenoprofen nephropathy. Am J Med 1983:75:9-1 6.
43. Kincaid-Smith P: Effects of non-narcotic analgesics on thekidney. Drugs 1986;32(Suppl. 4): 109-128.
