Clinical implications of NSAID : Special populations, special considerations

RAYMOND A. ADELIZZI, DO

Nonsteroidal anti-inflammato­ The popularity of NSAIDs derives, in part, ry drugs (NSAIDs) are the centerpiece of from their record for safety and efficacy in the pharmacologic for most rheumat­ treatment of rheumatic pain and . But ic disorders and related conditions, and as these drugs do cause occasional serious adverse such are used in great numbers. These drugs effects. In the gastrointestinal (GI) tract, for exam­ are relatively safe and effective, but the phar­ ple, side effects range from mild symptoms, such macokinetics of NSAIDs can be substantially as dyspepsia and nausea, to the more serious altered in certain groups of patients, includ­ complications of gastric erosions, ulceration, and ing the elderly and patients with renal and . hepatic disease. In these patients, the risk of Serious gastric complications have been attrib­ NSAID toxicity is increased. An understanding uted to NSAIDs' inhibition of syn­ of NSAID pharmacokinetics in these groups thesis. are synthesized through­ can help physicians to adjust therapeutic out the body and have localized effects, including regimens in order to limit the potentially cytoprotective functions in the GI tract. Renal serious complications of long-term NSAID dysfunction is another common result of long­ therapy. The author discusses the age-relat­ term NSAID use; and NSAIDs can have hepatic, ed physiologic changes that may affect the cardiovascular, hematologic, central nervous sys­ various areas of drug pharmacokinetics and tem, and dermatologic effects as well. ways in which NSAIDs may interact with The prevalence of serious side effects in NSAID other drugs. Risk assessment and monitor­ users is relatively small. The Food and Drug ing methods are also discussed. Administration estimates that serious GI com­ (Key words: NSAIDs, arthritis therapy, plications occur in 2% to 4% of all patients receiv­ NSAID pharmacokinetics, drug therapy in ing long-term NSAID therapy for 1 year or longer.2 elderly) In some patient populations, however, the risk of side effects resulting from long-term NSAID use Nonsteroidal anti-inflammatory drugs (NSAIDs) increases substantially. This is particularly the are widely prescribed to control the pain associ­ case for patients older than 60 years; these patients ated with and other musculoskele­ receive approximately half of all NSAID pre­ tal disorders. According to the Arthritis Founda­ scriptions.3 tion,l approximately 17 million people with arthritis According to recent epidemiologic studies, in the United States take NSAIDs, and this num­ patients 60 years of age and older have two and a ber will undoubtedly increase as the population ages. half times the risk of hospitalization and three times the risk of death due to GI complications Dr Adelizzi is director, Arthritis Center of South Jersey, Strat­ secondary to NSAID therapy as do those younger ford, NJ, and clinical associate professor of medicine, than age 60. 3 Other special populations-most Department of Medicine, University of Medicine and Den­ notably, patients with renal or hepatic impair­ tistry of New Jersey-School of Osteopathic Medicine, Stratford, NJ. ment-also are at increased risk for NSAID Correspondence to Raymond A. Adelizzi, DO, Suite 101, toxicity. 215 E Laurel Rd, Stratford, NJ 08084-1504. The precise reasons for increased NSAID tox-

396 • JAOA • Vol 94 • No 5 · May 1994 Clinical practice • Adelizzi icity:in these groups is not entire­ ly clear. Most likely, a comb:ina­ Table 1 tion offactors conspire to :increase Physiologic Changes With Aging That May Affect patients' vulnerability to long­ Pharmacokinetics of Nonsteroidal Anti-inflammatory Drugs term NSAID use. Elderly patients, Physiologic for example, may not tolerate change Possible effect injury to the GI tract as well as younger patients, and may be less resilient when damage does Absorption Decreased gastric acid production Increased gastric pH occur. Decreased gastrointestinal mucosal cell mass Age-related changes:in drug Reduced gastrointestinal motility metabolism can affect the absorp­ tion, clearance, distribution, and of drugs. These changes Distribution Reduced total body water Decreased lean body mass are thought to playa role in the Increased body fat increased toxicity of NSAIDs in Reduced serum albumin the elderly and in patients with Reduced cardiac output renal and hepatic dysfunction. Preservation of flow to heart, brain, In addition, polypharmacy-a and muscles situation common both to elder­ ly patients and to those with other Metabolism Decreased hepatic blood flow diseases--can cause drug :inter­ Decreased hepatic mass actions affecting the toxicity of Reduced enzymatic activity NSAIDs and other drugs taken concomitantly. In this context, a Excretion Decreased renal plasma flow familiarity with NSAID phar­ Decreased glomerular filtration rate macokinetics :in patients at risk Decreased tubular function for complications helps the physi­ cian to adjust therapeutic regi­ Source: Used with permission from Buchanan WW: Implications ofNSAID therapy in elderly mens :in order to prevent or limit patients. J Rheumatology 1990;17(suppl 20):29-32. serious adverse reactions.

NSAID pharmacokinetics in the elderly playa "gastroprotective" role (prostagland:ins, for It is difficult to generalize about the effects of age example, may enhance mucosal blood flow and on the pharmacokinetics of any given NSAID or :increase mucus secretion), NSAIDs may compound class of NSAIDs, in large part because not all these changes, lead:ing to a direct toxic effect on patients age at the same rate or :in the same way. the GI tract. In addition, there have been few reliable studies ofNSAID pharmacokinetics :in the elderly or :in those Distribution with diseases that may predispose patients to Age-related changes :in body composition and pro­ NSAID toxicity. There are, however, age-related te:in b:ind:ing have more meaningful implications for physiologic changes that may affect all four areas altered NSAID pharmacokinetics in the elderly. of drug pharmacok:inetics: the absorption, distri­ The volume of distribution for NSAIDs may be bution, metabolism, and excretion of these drugs reduced with age by changes :in body composition, (Table 1). :including a decline in total body size; an :increase :in body fat stores; and declines :in total body water Absorption and lean body mass. Although age-related changes in the GI tract do occur, As the volume of distribution decreases, NSAID these are probably the pharmacok:inetic variables half-life increases. Because steady-state concen­ least likely to affect NSAID metabolism. We know, trations are achieved over a period of approxi­ however, that gastric acid production decreases mately four times the drug's half-life, a longer half­ with age, as do GI mucosal cell mass, motility, and life may result-at least theoretically-in blood flow, whereas gastric pH :increases. Because accumulation of the drug for a longer time and at NSAIDs inhibit prostaglandins, some of which a higher, more toxic level. This drug accumula-

Clinical practice • Adelizzi JAOA • Vol 94 • No 5 • May 1994 • 397 Table 2 Because NSAIDs are predomi­ Antihypertensive Drugs With Which nantly metabolized by the liver, Nonsteroidal Anti-inflammatory Drugs (NSAIDs) Interact* these age-related decreases in hepatic function might be expect­ NSAIDs Antihypertensive drug ed to increase plasma half-life and drug accumulation in the elderly, and thus the potential l3-blockers, spironolactone sodium Hydralazine hydrochloride for toxicity. However, other vari­ l3-blockers ables-including hepatic and Thiazide diuretics, l3-blockers renal disease, smoking, and alco­ Indomethacin Thiazide diuretics, furosemide, hol use-may be more impor­ l3-blockers, angiotensin-converting tant than chronologic age in affect­ inhibitors ing hepatic . Excretion Thiazide diuretics, l3-blockers The excretion of NSAIDs and (timolol maleate) their metabolites may be sub­ stantially affected by notable age-related declines in renal func­ l3-blockers Thiazide diuretics, furosemide, tion. Both glomerular filtration l3-blockers (propranolol hydrochloride) rate and creatinine clearance decline with age, as does renal tubular function. Decreased capac­ Aspirin and most Thiazide and loop diuretics, ity to eliminate NSAIDs may other NSAIDs spironolactone result in prolongation of plasma reduce effects of half-life and plasma accumulation. In addition, renal insufficiency may cause diminished clearance *Current drug inform ation based on limited sturues. Source: Houston MC: Interactions between NSAIDs and antihypertensive drugs. J Museuloskel of unstable acylglucuronide Med 1991;8(8) :31-46. metabolites of certain NSAIDs (for example, feno- profen, naproxen, , tion has not necessarily been found to be the case and ), which are then retained and cre­ in clinical practice, however. ate a "reservoir" for reformation of the parent In addition, serum albumin levels fall in the NSAID, resulting in further accumulation of the elderly and in persons with hypoalbuminemia, drug.5 such as those with active . Most NSAIDs are extensively bound to serum Other special populations proteins, and even a small decrease in serum Patients may have a higher risk for NSAID tox­ albumin can lead to higher concentrations of icity if they have certain underlying illnesses, unbound, pharmacologically active and more toxic especially renal or hepatic disease, or if they take drug. Studies of drug disposition in the elderly any of a number of drugs that interact with comparing bound and unbound drug concentrations NSAIDs. have suggested that dosages of some NSAIDs (for example, naproxen) may need to be reduced in Renal disease elderly patients.4 The prostaglandin-inhibition properties ofNSAIDs can exacerbate existing disease in patients with Metabolism underlying pathologic renal conditions. Hepatic blood flow decreases with age, as does Prostaglandins help to maintain glomerular filtration hepatic mass in relation to total body size. Hepat­ rate and renal blood flow, and also influence the ic enzyme activity, especially oxidation, may also tubular transport of ions and water. Their role in be reduced in the elderly. Evidence indicating to maintaining homeostasis is especially important what extent these factors actually affect NSAID for patients in whom renal function is compro­ metabolism is scant and somewhat conflicting. mised. (continued on page 401)

398 • JAOA • Vol 94 • No 5 • May 1994 Clinical practice • Adelizzi Complications that increase patient vulner­ channel blockers or central a-agonists) are ability to renal prostaglandin suppression include thought to be achieved independently of renal congestive heart failure, cirrhosis with ascites, prostaglandins, and NSAID interactions with nephrosis and nephritis, sodium depletion, and these drugs have not been reported.8 diuretic therapy; all cause hypovolemic states that may lead to higher NSAID plasma concentrations. Clinical implications Prostaglandins also help to modulate vascular In view of the heightened risk of NSAID toxicity tone and, in patients with an underlying renal in specific patient populations, a few key principles pathologic lesion, NSAIDs can accentuate renal should guide the physician when initiating NSAID vasoconstriction and induce reversible hemody­ therapy.· namic renal failure. Renal disease can also decrease serum albumin levels, leading to higher concen­ Assess risk trations of unbound, pharmacologically active In general, NSAIDs should be prescribed cau­ NSAIDs. tiously in those patients at greatest risk for toxi­ city (Table 3). Risk factors for GI complications Hepatic disease include advanced age, history of previous gastric Depending on the extent of liver disease, hepatic ulcer, prednisone use, cigarette smoking, and pos­ clearance may be markedly reduced, leading to sibly rheumatoid arthritis. toxic accumulations ofNSAIDs. Liver disease can To aid in evaluating the potential for renal or also affect serum protein concentrations, which, hepatic NSAID toxicity, baseline renal and hepat­ as aforesaid, can lead to potentially toxic increas­ ic profiles (complete blood cell [CBC] count, and es in free, pharmacologically active drugs.6 measurements of serum creatinine and serum aspartate aminotransferase [AST] or serum alanine Drug interactions aminotransferase [ALT] levels) should be con­ The frequency with which NSAIDs are prescribed ducted to detect underlying disease. A thorough for a multitude of complaints with other medica­ medical history should elicit information about tions (and possibly even other NSAIDs, in either other medications being taken, with special atten­ prescription or over-the-counter formulations), tion to antihypertensives and other drugs that allows for potentially serious drug interactions to may require use with caution when taken togeth­ occur. One form of drug interaction takes place er with NSAIDs. when some NSAIDs (that is, and salicylates) compete with other highly protein­ Assess need bound drugs (for example, oral , Before a therapeutic regimen is determined, the need sulfonylurea hypoglycemics, sulfonamides, and for NSAIDs should be carefully evaluated. Are phenytoin) and displace them or are displaced by there nonmedical approaches-behavior modifi­ them, temporarily increasing their pharmacolog­ cation, mechanical interventions, or exercises­ ic activity.5,7 Nonsteroidal anti-inflammatory drugs that might mitigate discomfort without the need may also impair renal elimination of certain drugs, for medication? If drug therapy is necessary, anal­ including methotrexate and , by compet­ gesics such as acetaminophen, which has fewer ing at the proximal tubules for active renal tubu­ gastric side effects, and intermittent dosing might lar secretion.5,7 be effective for some patients who have mild to In addition, NSAIDs may reduce the hypoten­ moderate pain. If NSAIDs are prescribed, it is best sive effects of some antihypertensive agents (Table to begin with the lowest effective dose and encour­ 2). This effect is a potential problem for a great age the patient to give the drug time to work and many patients as it is estimated that some 20 mil­ reach steady state concentrations (1 to 2 weeks, if lion people in the United States take both NSAIDs possible) before the dosage or choice of drug is and antihypertensives.8 The mechanisms of action reevaluated. for a number of antihypertensives (for example, diuretics, [3-blockers, and angiotensin-converting Choose NSAlDs carefully [ACE] inhibitors) depend on the release Most studies show little substantive difference in of renal prostaglandins; their effects are the ones efficacy between NSAIDs. There is little doubt, most likely to be attenuated by the prostaglandin­ however, that safety issues may vary considerably inhibition action of most NSAIDs. The effects of among individual NSAIDs, especially when con­ other antihypertensives (for example, calcium sidering individual organ systems (Table 4). Unfor-

Clinical practice • Adelizzi JAOA· Vol 94 • No 5· May 1994 • 401 Table 3 Risk Factors for Nonsteroidal Anti-inflammatory Drug (NSAID) Toxicity by Organ System

Gastrointestinal Renal Hepatic

Age (>60 years) Age Age History Congestive heart failure Liver disease (cirrhosis, Corticosteroid therapy Cirrhosis with ascites hepatitis) Cigarette smoking N ephrosis/nephritis Decreased renal function Rheumatoid arthritis? Hypotension High NSAlD dose Hypovolemia use Sodium depletion Polypharmacy Diuretic therapy

Table 4 Nonsteroidal Anti-inflammatory Drug (NSAID) Safety Profile by Organ System

Risk Gastrointestinal Renal Hepatic

Highest Aspirin Indomethacin (Clinoril) Phenylbutazone* Phenylbutazone* Diclofenac (Voltaren) Indomethacin (Indocin) Aspirin Phenylbutazone* (Nalfon) Aspirin

Lowest Nonacetylated salicylates Nonacetylated Fenamates (eg, Disalcid or Trilisate) salicylates Meclofenamate sodium Etodolac (Lodine) (eg, Disalcid, (Meclomen) Nabumetaone (Relafen) Trilisate) Diclofenac sodium (Voltaren) Sulindac (Clinoril) (Pons tel) Oxaprozin (Daypro) Etodolac (Lodine) Some propionic acids Piroxicam (Feldene) Ibuprofen (Motrin, Rufen) Naproxen (Naprosyn) Fenoprofen calcium (Nalfon) Piroxicam (Feldene)

*No longer avaHable. Sources: Derived with permission from Clive DM, Stoff JS: Renal syndromes associated with nonsteroidal anti-inflammatory drugs. tunately, few studies have attempted to evaluate etodolac, the :first member in a new class ofNSAIDs NSAID toxicity in patient populations most at risk called the pyranocarboxylic acids, indicate that for complications of long-term NSAID therapy. etodolac may be better tolerated by the stomach than Furthermore, individual patient preferences and aspirin or other commonly used NSAIDs,lO per­ responses to any given NSAID vary greatly. haps as the result of selective sparing of gastric In general, patients at risk for GI complica­ and duodenal prostaglandins.l1 tions should have prescribed NSAIDs that have Some reports have suggested that sulindac been found to have fewer GI side effects. In a may have less of an inhibitory effect on renal recent review of comparative data, ibuprofen and prostaglandins and therefore may be preferred in (a nonacetylated salicylate) were noted patients with hypertensive renal disease.B Not all for receiving consistently good GI toxicity rank­ studies with sulindac have been conclusive, how­ ings.9 Endoscopic and microbleeding studies with ever. Etodolac has been studied in patients with renal

402 • JAOA • Vol 94 • No 5 • May 1994 Clinical practice • Adelizzi impairment and renal failure, and in patients and other painful conditions. However, the stage with stable liver cirrhosis.6 The drug was not is set for potentially serious NSAID-induced tox­ found to accumulate in these patients, indicating icity when the pharmacokinetics of NSAIDs are that dosing adjustments may not be necessary substantially altered in elderly patients and in in patients with underlying renal or stable liver those with renal or hepatic impairment. Interac­ disease. tions with other commonly prescribed drugs are also of major concern in the elderly and in those Regular monitoring with multiple illnesses. Despite the favorable safety profiles of certain Information about the complexities ofNSAID NSAIDs, it is advisable to prescribe all NSAIDs pharmacokinetics in special patient populations with care in patients at high risk for toxicity and is useful for guiding prescribing decisions. Recent to monitor therapy with any NSAID conscien­ research has refined our understanding in this tiously. After obtaining baseline measurements, area relative to certain high-risk conditions and the CBC count, serum creatinine levels, and liver a few specific NSAID preparations. It is my opin­ function (AST or ALT) should be checked in elder­ ion that NSAIDs can be used in a safe and judi­ ly patients within the first 3 months of initia­ cious manner if the physician is somewhat knowl­ tion of NSAID therapy (I prefer checking at 1 edgeable about these drugs, and is especially month) and every 3 to 6 months thereafter. The knowledgeable about the populations at high risk serum creatinine level should be checked with­ for a toxic reaction. However, future inquiry must in 2 weeks for patients with known liver or renal further elucidate the differences between NSAIDs dysfunction. in terms of their mechanisms of action and toxic Toxic side effects will frequently manifest potential if NSAID prescribing is to become more themselves soon after initiation of NSAID thera­ science than art. py. Also, physicians should advise patients to report any warning signs of trouble, such as edema, References weight gain, changes in bowel habits, and vision 1. Arthritis Foundation: Arthritis Foundation Survey on NSAID or hearing abnormalities. Ulcer Prevention. Westport, Conn, Yakelovich, Skelley, White/Clan­ Gastrointestinal complications, on the con­ cy, Shulman, 1989. trary, are often asymptomatic, especially in elder­ 2. Paulus HE: FDA Arthritis Advisory Committee meeting. Arth Rheum ly patients, confounding attempts to detect such 1988;31:1450-1451. problems before serious GI bleeding and ulcera­ 3. Fries JF, Miller SR, Spitz PW, et al: Toward an epidemiology of gastropathy associated with nonsteroidal anti-inflammatory drug tion occur. 12 Endoscopy cannot be recommended use. Gastroenterology 1989;96:647-655. routinely in most patients. Stool studies for occult 4. Upton RA, Williams RL, Kelly J , et al: Naproxen pharmacokinetics blood may be done periodically, although I check in the elderly. Br J Clin Pharmacol 1984;18:207-214. for occult blood only if a follow-up CBC count 5. Brater DC: Clinical pharmacology of NSAIDs. J Clin Pharma­ shows a drop in the hemoglobin value. col 1988;28:518-523. , a that 6. Brater DC, Lasseter KC: Profile of etodolac: Pharmacokinetic acts as a gastroprotective agent is of use in patients evaluation in special populations. Clin Rheumatol 1989;8(suppl clearly at high risk for serious GI complications (for 1):25-35. example, the elderly, patients with previous pep­ 7. Furst DE: Clinically important interactions of nonsteroidal anti­ inflammatory drugs with other medications. J Rheumoiol1988;15(suppl tic ulcers, patients on steroid therapy, cigarette 17):58-62. smokers, and patients with multiple debilitating 8. Houston MC: Interactions between NSAlDs and antihyperten­ illnesses). Persistent GI symptoms of blood loss, sive drugs. J Musculoskel Med 1991;8(8):31-46. especially after rechallenge with another NSAID, 9. Fries JF: Choosing an appropriate NSAlD. Drug Ther 1991;21(10)19- merits endoscopy to rule out gastric erosions and 29. ulcerations. However, it must be noted that not all 10. Lanza FL, Arnold JD: Etodolac, a new nonsteroidal anti-inflam­ GI bleeding in patients treated with NSAIDs is matory drug: Gastrointestinal microbleeding and endoscopic stud­ caused by upper GI lesions. ies. Clin R heumatol 1989;8(suppl 1):5-15. 11. Taha AS, McLaughlin S, Holland PJ, et al: Effect on gastric and duodenal mucosal prostaglandins of repeated intake ofthera­ Comment peutic doses of naproxen and etodolac in rheumatoid arthritis. Ann By virtue of their generally excellent safety and Rheum Dis 1990;49:354-358. efficacy profile, NSAIDs occupy a large and increas­ 12. Sager DS, Bennett RM: Individualizing the riskibenefit ratio of ing role in the management of rheumatic diseases NSAlDs in older patients. Geriatrics 1992;47:24-31.

Clinical e ractice • Adelizzi JAOA • Vol 94 • No 5 • May 1994 • 403