Chapter 191: Nonsteroidal Anti-Inflammatory Drugs

Nonsteroidal anti-inflammatory drugs (NSAIDs) are among the most widely used class of drugs in the United States, and all share inhibition of the cyclooxygenase enzyme as a mechanism of action. NSAIDs are effective antipyretics, analgesics, and anti-inflammatory agents. Because of their large therapeutic window, acute ingestion with overdoses rarely produces serious complications.^1,2 The morbidity from NSAIDs in acute overdose is far overshadowed by complications of NSAIDs at therapeutic doses, which include GI bleeding, drug-induced renal failure, and atherosclerotic heart disease.^3,4,5,6 Due to their increased risk for cardiovascular disease, rofecoxib and valdecoxib were withdrawn from the U.S. market in 2004 and 2005, respectively.

OVERVIEW

NSAIDs are structurally varied compounds with common therapeutic effects (Table 191-1). NSAIDs reversibly inhibit the enzyme cyclooxygenase, which is responsible for the production of prostaglandins from arachidonic acid (Figure 191-1). The anti-inflammatory effect of NSAIDs is through the inhibition of prostaglandin production, and they may also inhibit neutrophil migration via unclear mechanisms. NSAIDs are antipyretics through inhibition of prostaglandin E₂ in the hypothalamus. NSAIDs attenuate prostaglandin-mediated hyperalgesia and local pain fiber stimulus.

CYCLOOXYGENASE

Two isoforms of cyclooxygenase (abbreviated COX-1 and COX-2) vary in presence and distribution.⁷ COX-1 is present with steady level of activity and is found primarily in blood vessels, kidneys, and stomach. In contrast, COX-2 is not normally found to a significant degree in human tissue with the possible exception of the brain and kidneys, unless its production is induced by local inflammation.

Cyclooxygenase inhibitors can be categorized as nonselective, partially selective, or selective regarding their inhibition of COX enzyme isoforms (Table 191-1). Most NSAIDs nonselectively inhibit both COX-1 and COX-2. COX-1 inhibition is responsible for most of the unwanted GI side effects of NSAIDs. Drugs, such as etodolac and meloxicam, were created to inhibit preferentially COX-2, theoretically reducing the unwanted GI effects. Unfortunately, the COX-2–selective compounds do not appear to be more effective mediators of inflammation or analgesia and are still associated with GI side effects in addition to usually being more expensive.^8,9

PHARMACOKINETICS

All NSAIDs are rapidly absorbed from the GI tract, and most achieve peak serum levels within approximately 2 hours. They are highly protein bound in the plasma, have low volumes of distribution (approximately 0.2 L/kg), and cross the blood–brain barrier. NSAIDs undergo metabolism mainly in the liver via glucuronic acid conjugation or via liver enzymes before elimination in the urine or feces. Plasma half-lives of NSAIDs range from 2 hours for ibuprofen to greater than 50 hours for the long-acting agents piroxicam and phenylbutazone (Table 191-1). Creation of active metabolites may prolong the therapeutic effect beyond that of the parent compound.

Ingestion of large amounts of some drugs, such as ibuprofen and naproxen, exhibits slower absorption, taking 3 to 4 hours to achieve peak plasma levels. As greater amounts of NSAIDS are absorbed in large overdoses, greater fractions of free drug become available for toxicity in a nonlinear manner because protein binding is limited.

Topical NSAIDS

Topical NSAIDS are well absorbed through the skin and reach therapeutic levels in synovial fluid. They provide good pain relief for osteoarthritis, with very low blood levels and no toxic effects described to date.^10,11

SIGNIFICANT DRUG–DRUG INTERACTIONS

Aspirin

Daily aspirin therapy reduces the incidence of recurrent myocardial infarction and stroke via impairment of platelet aggregation. Ibuprofen administered three times per day competitively inhibits this aspirin effect on platelets via its interaction with the thromboxane pathway, undermining its cardioprotective effect. This observation may extend to other NSAIDs but not to diclofenac or the COX-2 inhibitors.¹²

Antihypertensive Agents

NSAIDs can decrease the effectiveness of some antihypertensives including angiotensin-converting enzyme inhibitors, diuretics, and α- and β-adrenergic antagonists. Inhibition of prostaglandin synthesis is believed to be a central mechanism for these effects. Lower prostaglandin levels result in alterations in the renin-angiotensin system causing decreased renal sodium clearance, water retention, and changes in vascular tone.¹³

Warfarin

Warfarin and NSAIDs have several important interactions that may increase the risk of bleeding.¹⁴ Some NSAIDs may displace warfarin from plasma proteins, but the effect is small and does not lead to increased anticoagulation. The selective COX-2 inhibitors at therapeutic doses also elevate prothrombin times, likely through decreased elimination of warfarin. Other nonselective NSAIDs have not been reported to change warfarin protein binding or elimination, but their use is not recommended in warfarin users because they inhibit platelet aggregation and can significantly increase the risk of bleeding.

Decreased Renal Clearance of Drugs

NSAIDs may also decrease renal clearance via inhibition of renal prostaglandin synthesis, decreasing their vasodilatory effect, causing a decrease in renal blood flow and glomerular filtration rate, and ultimately decreasing the rate of elimination of certain drugs. Therefore, NSAID use can lead to increased concentrations of consequential drugs such as lithium, methotrexate, and metformin. NSAIDs are a potential risk factor in the development of metformin-associated lactic acidosis, a rare complication of metformin therapy with a high mortality.¹⁵

TOXICITY AT THERAPEUTIC DOSES

NSAIDs are generally safe drugs but have a number of well-reported side effects at therapeutic doses (Table 191-2). It is generally believed that indomethacin and other long-acting agents, such as piroxicam, are responsible for a greater proportion of side effects, whereas the propionic acid agents, such as ibuprofen, are responsible for fewer side effects.

Central Nervous System

CNS adverse effects at therapeutic doses include headache, cognitive difficulties, behavioral change, and aseptic meningitis.^16,17 Acute psychosis has been reported with indomethacin and sulindac use, which is hypothesized to result from the structural similarity of these NSAIDs to serotonin.

Patients with NSAID-induced aseptic meningitis may experience symptoms of headache, fever, and neck stiffness occurring within hours of a therapeutic dose.^16,17,18 Cerebrospinal fluid analysis of these patients finds elevated WBC counts and protein levels with normal or decreased glucose levels. Symptoms resolve after NSAID use is stopped and in some cases recur with repeat NSAID challenges.^18,19 This phenomenon is most often seen in patients who have underlying autoimmune diseases, such as systemic lupus erythematosus, and is thought to be a hypersensitivity reaction. Exclusion of an infectious cause is necessary before the diagnosis of NSAID-induced aseptic meningitis can be assigned. Agents associated with aseptic meningitis include celecoxib, diclofenac, ibuprofen, ketoprofen, naproxen, rofecoxib, sulfasalazine, sulindac, and tolmetin.

Cardiovascular

Increased cardiovascular events, including myocardial infarction, have been reported with long-term use of both nonselective NSAIDs and selective COX-2 agents.²⁰ The observed risk increase is variable among the different agents, but risk is dose and duration dependent, as well as more profound in patients with established coronary artery disease.^20,21 Selective COX-2 inhibitors are associated with a doubling of the relative risk of myocardial infarction and heart failure.²² For nonselective NSAIDs, no increase in major vascular events has been observed with naproxen, whereas ibuprofen and diclofenac regimens were associated with a relative risk for major vascular events as much as 1.5 that of placebo.²² A stepped-care approach to NSAID use for managing musculoskeletal symptoms in those with known cardiovascular disease or risk factors is recommended.^22,23 Begin therapy with medications not associated with increased cardiovascular effects, such as acetaminophen, and progress carefully to nonselective NSAIDs, or for patients over 75 years of age, use a topical rather than oral NSAID.^24,25,26

Pulmonary

NSAIDs have been associated with adverse pulmonary reactions, including bronchospasm in asthmatics, hypersensitivity pneumonitis, and pulmonary edema. Patients with NSAID-related hypersensitivity pneumonitis may complain of fever, cough, and shortness of breath. Chest radiographs may show pulmonary infiltrates, and blood work may show leukocytosis and eosinophilia. Cessation of NSAID use resolves the symptoms, whereas rechallenging causes recurrence. NSAIDs implicated in hypersensitivity pneumonitis include diclofenac, diflunisal, naproxen, piroxicam, and sulindac. The mechanism of NSAID-mediated pneumonitis is also thought to be an immune-mediated hypersensitivity reaction.

NSAID-induced bronchospasm is a well-described phenomenon in patients with reactive airways disease. The spectrum of this reaction ranges from rhinitis to severe bronchospasm with laryngeal edema.²⁷ Patients with underlying reactive airways disease and nasal polyps are at greater risk for these complications. The mechanism of this hypersensitivity reaction to NSAIDs in asthmatics does not appear to be immunoglobulin E–mediated but rather may result from excessive production of leukotrienes.

Gastrointestinal

The greatest morbidity caused by therapeutic use of NSAIDs is GI.^9,28 These effects derive from inhibition of cytoprotective gastric prostaglandins resulting in injuries ranging from dyspepsia, heartburn, and nausea, to life-threatening bleeding.⁵ During long-term therapy (>3 months), endoscopic ulcers are observed in 10% to 30% of patients, but only 1% to 2% experience serious ulcer complications, such a bleeding or perforation. Patients at higher risk for GI bleeding include those with a history of ulcers, those taking high doses of NSAIDs or corticosteroids, and the elderly.^29,30,31 Reduced incidence of GI side effects has been observed with celecoxib, but not other COX-2–selective agents, such as rofecoxib.

Hepatic

Clinically apparent liver injury caused by NSAIDs is most likely idiosyncratic and rare, occurring in 1 to 10 cases per 100,000 prescriptions.^32,33 More commonly, asymptomatic elevations of serum transaminases are seen in up to 18% of users taking NSAIDs over a long period. Diclofenac and sulindac are the two NSAIDs most commonly linked to hepatotoxicity. Stopping the drug should lead to complete recovery while also avoiding other NSAIDs that may cause recurrence.

Renal

NSAIDs exert many toxic effects on the kidney.^5,6 COX inhibition decreases prostaglandin synthesis, leading to vasoconstriction of the renal vascular supply with resultant reversible prerenal azotemia. Continued use may lead to acute tubular necrosis. Interstitial nephritis may result from NSAID-associated activation of inflammatory cells. Glomerular injury may be caused by caused by proinflammatory leukotrienes stimulated by NSAIDs resulting in alteration of glomerular permeability. Chronic renal injury in the form of patchy necrosis, as well as medullary ischemia, is associated with chronic, large cumulative doses of NSAIDs, resulting from chronic inhibition of vasodilatory prostaglandins. Additionally, chronic interstitial fibrosis, hyperkalemia, and hyponatremia may be caused by NSAID use.³⁴

Hematologic

Nonselective NSAIDs inhibit platelet formation of thromboxane A₂, a potent stimulator of platelet aggregation, resulting in qualitative platelet deficiencies. The new COX-2 inhibitors have far fewer antiplatelet effects than traditional NSAIDs, which may explain why an increase in cardiac and neurovascular events in patients taking the COX-2 inhibitor rofecoxib was observed versus the nonselective COX-1 inhibitor naproxen. Most NSAIDs decrease platelet aggregation only when significant concentrations of the drug are present, and therefore, increased bleeding tendencies from NSAIDs are not widely reported.

Bone marrow suppression and aplastic anemia are rare hematologic complications of almost all NSAIDs with indomethacin, diclofenac, and phenylbutazone responsible for most reported cases. NSAID use has also resulted in agranulocytosis, hemolytic anemia, red cell aplasia, and thrombocytopenia.

Dermatologic

NSAIDs account for approximately 10% to 27% of all cutaneous drug reactions.³⁵ The agents most frequently involved in dermatologic complications include benoxaprofen, phenylbutazone, and piroxicam. Drug reactions to NSAIDs range from benign maculopapular rashes to Stevens-Johnson syndrome and toxic epidermal necrolysis. NSAIDs are among the most commonly implicated drugs in cases of toxic epidermal necrolysis, accounting for up to one third of drug-related cases. NSAIDs of all types (oral and topical) can cause photosensitivity reactions including increased sensitivity to sun exposure (phototoxic) and true photoallergic reactions (see chapter 249, "Generalized Skin Disorders"). Ketoprofen and piroxicam are the most frequently involved in photoallergic reactions. Isoxicam was removed from the U.S. market after fatal skin reactions.

Bone

Controversy surrounds the use of NSAIDs for control of pain resulting from a fractured bone. Although NSAIDs are well known to treat musculoskeletal pain effectively, bone healing is a complex cascade of events that involves prostaglandins' influence on the balance between bone formation and resorption.³⁶ With conflicting results regarding wound healing, some orthopedists recommend delayed administration of NSAIDs for 3 to 4 weeks following a fracture, using judicious doses, and possibly avoiding NSAIDs altogether in cases of spinal fusion.³⁷

Reproductive

Prostaglandins are found in high concentration in the uterus at term and have a stimulatory effect on normal labor. NSAIDs will impair uterine motility through inhibition of prostaglandin synthesis. Although NSAIDs are not believed to be teratogenic in humans, they cross the placenta in late pregnancy. One of the most significant effects of fetal exposure to NSAIDs is premature constriction of the ductus arteriosus that may result in fetal pulmonary hypertension. Other reported effects of in utero exposure to NSAIDs include fetal intracranial hemorrhage, necrotizing enterocolitis, oligohydramnios, and renal dysfunction. NSAID-induced inhibition of platelet aggregation may place the fetus and the mother at increased risk for peripartum hemorrhage. The safest recommendation is to avoid NSAID use during pregnancy, especially during the third trimester.

TOXICITY WITH ACUTE OVERDOSE

Although the vast majority of patients with acute overdoses suffer little morbidity, some fatalities occur with massive NSAID ingestions that often include the clinical features of altered mental status, metabolic acidosis, and shock.^38,39,40,41 Historically, phenylbutazone was most commonly associated with severe toxicity following an overdose, and although it was withdrawn from the U.S. market in the 1970s, it is still available from veterinary sources and in other countries. By a wide margin, ibuprofen is the most common agent currently reported in NSAID overdoses. Patients who develop symptoms following an ibuprofen overdose usually ingest more than 100 milligrams/kg, and initial symptoms predominantly include abdominal pain, nausea, and vomiting, all beginning within 4 hours of ingestion (Table 191-3). Most patients who manifest severe toxicity, including apnea, coma, and metabolic acidosis, ingest more than 400 milligrams/kg. There is limited published experience concerning acute overdose of COX-2 inhibitors, but they are assumed to have similar profiles as the nonselective NSAIDs.

Central Nervous System

CNS manifestations of acute NSAID overdose are usually minimal, although patients with significant overdose may have headache, diplopia, nystagmus, and altered mental status including coma.¹⁷ Mefenamic acid has been identified as a drug with high potential for seizures in overdose.⁴² Muscle twitching and seizures that are responsive to benzodiazepines have been reported.

Cardiovascular

Acute NSAID overdose can result in hypotension and bradydysrhythmia. NSAIDs are not known to be a primary cause of dysrhythmias, but fluid and electrolyte abnormalities provoked by NSAID toxicity may induce cardiac rhythm abnormalities. Cardiovascular dysfunction from acute NSAID overdose is responsive to conventional critical care management.

Metabolic

Electrolyte and acid-base abnormalities may occur in acute NSAID overdoses. Alterations in serum electrolytes may develop secondary to decreased prostaglandin synthesis or from NSAID-induced renal failure. Sodium and water retention may lead to volume overload in patients with preexisting cirrhosis, heart failure, or renal failure. Hyperkalemia, hypocalcemia, and hypomagnesemia have been reported in NSAID overdoses complicated by acute renal failure.

Increased anion gap metabolic acidosis has been observed with large overdoses of ibuprofen and naproxen, which may be due to these NSAIDs and their metabolites being weak acids. Concurrent lactic acid production in the setting of NSAID-induced seizures and shock may exacerbate the acidosis in some cases.

GI and Hepatic

Patients presenting after acute NSAID overdose may have abdominal pain, nausea, and vomiting, but life-threatening GI hemorrhage is not a typical finding after acute overdoses. Overdose may also result in hepatic injury, as measured by cholestasis and elevated transaminase levels. Rare cases of pancreatitis have been reported after overdoses of ibuprofen, naproxen, and indomethacin.^43,44,45

Renal

NSAID overdose rarely causes acute kidney injury but may place a stressed renal system at risk for failure. The clinical presentation may include hematuria and oliguria. Most patients with acute renal failure due to an NSAID overdose have eventual recovery of renal function, but some may need long-term dialysis.

Most patients who present following an NSAID overdose will be asymptomatic. Follow the general principles for managing the poisoned patient (see chapter 176, "General Management of Poisoned Patients"), including screening for possible co-ingestants, consultation with a toxicologist, and consideration of GI decontamination. Attempt to ascertain whether the amount of drug ingested was less than 100 milligrams/kg (unlikely to result in toxicity) or more than 400 milligrams/kg (significant risk for toxicity).

Asymptomatic NSAID ingestions usually require minimal laboratory evaluation and supportive care (Figure 191-2). GI decontamination is best accomplished with activated charcoal and may be considered depending on the perceived risk of severe ingestion and clinical condition. Gastric lavage is not useful in the majority of cases but may be considered for a patient who has ingested a massive quantity within the past hour, especially mefenamic acid or phenylbutazone. Fluid and electrolyte replacement is given as necessary. Consider screening for potentially dangerous co-ingestants by measuring the serum acetaminophen and salicylate levels and obtaining a 12-lead ECG.

Patients who are symptomatic with altered mental status, seizures, shock, respiratory distress, or cardiac dysrhythmias require aggressive resuscitation and stabilization (Figure 191-2). Establish an airway as needed, institute mechanical ventilation when necessary, treat hypotension with fluid boluses and/or vasopressors, and manage seizures initially with IV benzodiazepines. Consider gastric lavage for recent ingestions involving mefenamic acid or phenylbutazone. Whole-bowel irrigation is recommended for severe overdose with enteric-coated formulations. Multidose activated charcoal has no proven benefit in symptomatic NSAID overdoses. Hemodialysis, hemofiltration, or charcoal hemoperfusion are not effective in enhancing elimination because NSAIDs are highly protein bound. Likewise, manipulation of serum and urine pH through IV alkalinization is not beneficial in enhancing renal elimination.

Patients with symptomatic NSAID overdoses should have a serum chemistry panel, hepatic profile, CBC, coagulation profile, and serum salicylate and acetaminophen levels measured. Serum NSAID levels do not correlate with observed toxicity or outcomes, so serum levels for specific NSAIDs are not indicated.

Most patients with asymptomatic NSAID ingestions can be safely discharged after a 4-hour period of observation. Patients with a symptomatic overdose (altered mental status, abnormal vital signs, electrolyte abnormalities, or acute renal failure) should be admitted for observation and supportive care. All overdoses should be reported to the regional poison control center both for management assistance and for statistical tracking. The majority of NSAID overdoses will not cause significant sequelae. Even patients with major symptoms have a good prognosis if they overcome the initial insult.