NSAIDs traditionally are grouped by their chemical characteristics. Following the development of the selective COX-2 inhibitors, the classification into tNSAIDs, which inhibit both COX-1 and COX-2, and COX-2–selective NSAIDs, emerged. Initially, only agents designed specifically for the purpose of COX-2–selective inhibition—colloquially termed the coxibs—were attributed to the group of COX-2–selective NSAIDs. However, some older NSAIDs (e.g., diclofenac, meloxicam, nimesulide) show a degree of selectivity for COX-2 that is similar to that of the first coxib, celecoxib. Thus, these drugs might better be classified as COX-2–selective NSAIDs, although this is not, as yet, commonplace (Figure 34–1). Other classifications of NSAIDs were developed based on t1/2, such as those with a shorter (<6 hours) or longer (>10 hours) t1/2 (Figure 34–1).
Classification of NSAIDs by chemical similarity(panel A), cyclooxygenase (COX) isoform selectivity (panel B), and plasma t1/2(panel C). The COX selectivity chart is plotted from data published in Warner et al., 1999, and FitzGerald and Patrono, 2001. tNSAIDs, traditional nonsteroidal anti-inflammatory drugs.
Most NSAIDs are competitive, reversible, active site inhibitors of the COX enzymes. However, aspirin (ASA) acetylates the isozymes and inhibits them irreversibly; thus, aspirin often is distinguished from the tNSAIDs. Similarly, acetaminophen, which is antipyretic and analgesic but largely devoid of anti-inflammatory activity, also is conventionally segregated from the group, despite sharing many properties with tNSAIDs relevant to its clinical action in vivo.
Chemistry. NSAIDs are a chemically heterogeneous group of compounds, which nevertheless share certain therapeutic actions and adverse effects. The class includes derivatives of salicylic acid (e.g., aspirin, diflusinal), propionic acid (e.g., naproxen, ibuprofen, flurbiprofen, ketoprofen), acetic acid (e.g., indomethacin, etodolac, diclofenac, ketorolac), enolic acid (e.g., piroxicam, phenylbutazone), fenamic acid (e.g., mefenamic acid, meclofenamic acid), alkanones (nabumetone), and diaryl heterocyclic compounds (e.g., celecoxib, valdecoxib, rofecoxib, etoricoxib) (Figure 34–1).
The vast majority of tNSAID compounds are organic acids with relatively low pKa values (Figure 34–1). Even the non-acidic parent drug nabumetone is converted to an active acetic acid derivative in vivo. As organic acids, the compounds generally are well absorbed orally, highly bound to plasma proteins, and excreted either by glomerular filtration or by tubular secretion. They also accumulate in sites of inflammation, where the pH is lower, potentially confounding the relationship between plasma concentrations and duration of drug effect. Most COX-2–selective NSAIDs are diaryl heterocyclic compounds with a relatively bulky side group, which aligns with a large side pocket in the AA binding channel of COX-2 but hinders its optimal orientation in the smaller binding channel of COX-1 (Figure 34–2) (Smith et al., 2000). Both tNSAIDs and the COX-2–selective NSAIDs generally are hydrophobic drugs, a feature that allows them to access the hydrophobic arachidonate binding channel and results in shared pharmacokinetic characteristics. Again, aspirin and acetaminophen are exceptions to this rule.
Structural basis for cyclooxygenase-2 (COX-2)-selective inhibition. The active centers of COX-1 and COX-2 are shown crystallized with the nonselective inhibitor flurbiprofen (COX-1) (Picot et al., 1994) and the experimental COX-2 inhibitor SC-558 (Kurumbail et al., 1996). The active center of COX-2 is characterized by a larger side pocket, which can accommodate molecules with bulkier side chains than COX-1. (Courtesy of Dr. Vineet Sangar.)
Cyclooxygenase Inhibition. The principal therapeutic effects of NSAIDs derive from their ability to inhibit PG production. The first enzyme in the PG synthetic pathway is COX, also known as PG G/H synthase. This enzyme converts AA to the unstable intermediates PGG2 and PGH2 and leads to the production of the prostanoids, TxA2, and a variety of PGs (see Chapter 33).
There are two forms of COX, COX-1 and COX-2. COX-1, expressed constitutively in most cells, is the dominant (but not exclusive) source of prostanoids for housekeeping functions, such as gastric epithelial cytoprotection and hemostasis. Conversely, COX-2, induced by cytokines, shear stress, and tumor promoters, is the more important source of prostanoid formation in inflammation and perhaps in cancer (see Chapter 33). However, both enzymes contribute to the generation of autoregulatory and homeostatic prostanoids, and both can contribute to prostanoid formation in syndromes of human inflammation and pain (see "Inflammation and Pain" at the beginning of this chapter). Importantly, COX-1 is expressed as the dominant, constitutive isoform in gastric epithelial cells and is thought to be the major source of cytoprotective PG formation. Inhibition of COX-1 at this site is thought to account largely for the gastric adverse events that complicate therapy with tNSAIDs, thus providing the rationale for the development of NSAIDs specific for inhibition of COX-2 (FitzGerald and Patrono, 2001).
Aspirin and NSAIDs inhibit the COX enzymes and PG production; they do not inhibit the lipoxygenase (LOX) pathways of AA metabolism and hence do not suppress LT formation (see Chapter 33). Glucocorticoids suppress the induced expression of COX-2, and thus COX-2–mediated PG production. They also inhibit the action of PLA2, which releases AA from the cell membrane. These effects contribute to the anti-inflammatory actions of glucocorticoids (see Chapter 35).
At higher concentrations, NSAIDs also are known to reduce production of superoxide radicals, induce apoptosis, inhibit the expression of adhesion molecules, decrease NO synthase, decrease pro-inflammatory cytokines (e.g., TNF-α, IL-1), modify lymphocyte activity, and alter cellular membrane functions in vitro. However, there are differing opinions as to whether any of these actions might contribute to the anti-inflammatory activity of NSAIDs (Vane and Botting, 1998) at the concentrations attained during clinical dosing. The hypothesis that their anti-inflammatory actions in humans derive solely from COX inhibition alone has not been rejected based on current evidence.
Observational studies suggest that acetaminophen, which is a very weak anti-inflammatory agent at the typical dose of 1000 mg, is associated with a reduced incidence of GI adverse effects compared to tNSAIDs. At this dose, acetaminophen inhibits both COXs by ∼50%. The ability of acetaminophen to inhibit the enzyme is conditioned by the peroxide tone of the immediate environment (Boutaud et al., 2002). This may partly explain the poor anti-inflammatory activity of acetaminophen, because sites of inflammation usually contain increased concentrations of leukocyte-generated peroxides.
Irreversible Cyclooxygenase Inhibition by Aspirin. Aspirin covalently modifies COX-1 and COX-2, irreversibly inhibiting COX activity. This is an important distinction from all the NSAIDs because the duration of aspirin's effects is related to the turnover rate of COXs in different target tissues. The duration of effect of non-aspirin NSAIDs, which inhibit the active sites of the COX enzymes competitively, relates to the time course of drug disposition. The importance of enzyme turnover in recovery from aspirin action is most notable in platelets, which, being anucleate, have a markedly limited capacity for protein synthesis. Thus, the consequences of inhibition of platelet COX-1 (COX-2 is expressed in megakaryocytes and perhaps immature platelet forms) last for the lifetime of the platelet. Inhibition of platelet COX-1–dependent TxA2 formation therefore is cumulative with repeated doses of aspirin (at least as low as 30 mg/day) and takes ∼8-12 days—the platelet turnover time—to recover fully once therapy has been stopped. Importantly, even a partially recovered platelet pool—just a few days after the last aspirin dose—may afford recovery of sufficient hemostatic integrity for some types of elective surgery to be performed. However, such a partial platelet function also may predispose noncompliant patients to thrombotic events.
COXs are configured such that the active site is accessed by the AA substrate via a hydrophobic channel. Aspirin acetylates serine 529 of COX-1, located high up in the hydrophobic channel. Interposition of the bulky acetyl residue prevents the binding of AA to the active site of the enzyme and thus impedes the ability of the enzyme to make PGs. Aspirin acetylates a homologous serine at position 516 in COX-2. Although covalent modification of COX-2 by aspirin also blocks the COX activity of this isoform, an interesting property not shared by COX-1 is that acetylated COX-2 synthesizes 15(R)-hydroxyeicosatetraenoic acid [15(R)-HETE]. This may be metabolized, at least in vitro, by 5-LOX to yield 15-epi-lipoxin A4, which has potent anti-inflammatory properties (see Chapter 33). Repeated doses of aspirin that acutely do not completely inhibit platelet COX-1–derived TxA2 can exert a cumulative effect with complete blockade. This has been shown in randomized trials for doses as low as 30 mg/day. However, most of the clinical trials demonstrating cardioprotection from low-dose aspirin have used doses in the range of 75-81 mg/day.
The unique sensitivity of platelets to inhibition by such low doses of aspirin is related to their presystemic inhibition in the portal circulation before aspirin is deacetylated to salicylate on first pass through the liver (Pedersen and FitzGerald, 1984). In contrast to aspirin, salicylic acid has no acetylating capacity. It is a weak, reversible, competitive inhibitor of COX. Salicylic acid derivates, rather than the acid, are available for clinical use. The lack of acetylation often is used as justification to prefer trisalicylate or salsalate over aspirin in presurgical patients. High doses of salicylate inhibit the activation of NFκB in vitro, but the relevance of this property to the concentrations attained in vivo is not clear (Yin et al., 1998). Salicylic acid also may inhibit the expression of COX-2 by interfering with the binding of CCAAT/enhancer binding protein (C/EBP) β transcription factor to the COX-2 promoter (Cieslik et al., 2002). This was observed in vitro at concentrations of salicylic acid that are attained in humans.
Selective Inhibition of Cyclooxygenase-2. The therapeutic use of the tNSAIDs is limited by their poor GI tolerability. Chronic users are prone to experience GI irritation in ≤20% of cases. Following the discovery of COX-2, it was proposed that the constitutively expressed COX-1 was the predominant source of cytoprotective PGs formed by the GI epithelium. Because its expression is regulated by cytokines and mitogens, COX-2 was thought to be the dominant source of PG formation in inflammation and cancer. Thus, selective inhibitors of COX-2 were developed based on the hypothesis that they would afford efficacy similar to tNSAIDs with better GI tolerability (FitzGerald and Patrono, 2001). Six COX-2 inhibitors specifically designed for such purpose, the coxibs, were initially approved for use in the U.S. or E.U.: celecoxib, rofecoxib, valdecoxib and its prodrug parecoxib, etoricoxib, and lumiracoxib. Most coxibs have been either severely restricted in their use or withdrawn from the market in view of their adverse event profile. Celecoxib (CELEBREX) currently is the only COX-2 inhibitor licensed for use in the U.S.
The relative degree of selectivity for COX-2 inhibition is lumiracoxib = etoricoxib > valdecoxib = rofecoxib ≫ celecoxib (Figure 34–1). Although there were differences in relative hierarchies, depending on whether screens were performed using recombinantly expressed enzymes, cells, or whole-blood assays, most tNSAIDs expressed similar selectivity for inhibition of the two enzymes. Some compounds, conventionally thought of as tNSAIDs—diclofenac, meloxicam, and nimesulide (not available in the U.S.)—exhibit selectivity for COX-2 that is close to that of celecoxib in vitro (Figure 34–1) (Warner et al., 1999; FitzGerald and Patrono, 2001). Indeed, meloxicam achieved approval in some countries as a selective inhibitor of COX-2. Thus, selectivity for COX-2 should not be viewed as an absolute category; the isoform selectivity for COX-2 (just like selectivity for β1 adrenergic receptors) is a continuous rather than a discreet variable, as illustrated in Figure 34–1.
Absorption, Distribution, and Elimination
Absorption. Most NSAIDs are rapidly absorbed following oral ingestion, and peak plasma concentrations usually are reached within 2-3 hours. All COX-2–selective NSAIDs are well absorbed, but peak concentrations are achieved with lumiracoxib and etoricoxib in ∼1 hour compared to 2-4 hours with the other agents. The poor aqueous solubility of most NSAIDs often is reflected by a less than proportional increase in area under the curve (AUC) of plasma concentration–time curves, due to incomplete dissolution, when the dose is increased. Food intake may delay absorption and sometimes decreases systemic availability (i.e., fenoprofen, sulindac). Antacids, commonly prescribed to patients on NSAID therapy, variably delay, but rarely reduce, absorption. Most interaction studies performed with proton pump inhibitors suggest that relevant changes in NSAID kinetics are unlikely. Little information exists regarding the absolute oral bioavailability of many NSAIDs, as solutions suitable for intravenous administration often are not available. Some compounds (e.g., diclofenac, nabumetone) undergo first-pass or presystemic elimination. Acetaminophen is metabolized to a small extent during absorption. Aspirin begins to acetylate platelets within minutes of reaching the presystemic circulation.
Distribution. Most NSAIDs are extensively bound to plasma proteins (95-99%), usually albumin. Plasma protein binding often is concentration dependent (i.e., naproxen, ibuprofen) and saturable at high concentrations. Conditions that alter plasma protein concentration may result in an increased free drug fraction with potential toxic effects. Highly protein bound NSAIDs have the potential to displace other drugs, if they compete for the same binding sites. Most NSAIDs are distributed widely throughout the body and readily penetrate arthritic joints, yielding synovial fluid concentrations in the range of half the plasma concentration (i.e., ibuprofen, naproxen, piroxicam). Some substances yield synovial drug concentrations similar to (i.e., indomethacin), or even exceeding (i.e., tolmetin), plasma concentrations. Most NSAIDs achieve sufficient concentrations in the CNS to have a central analgesic effect. Celecoxib is particularly lipophilic, so it accumulates in fat and is readily transported into the CNS. Lumiracoxib is more acidic than other COX-2–selective NSAIDs, which may favor its accumulation at sites of inflammation. Multiple NSAIDs are marketed in formulations for topical application on inflamed or injured joints. However, direct transport of topically applied NSAIDs into inflamed tissues and joints appears to be minimal, and detectable concentrations in synovial fluid of some agents (i.e., diclofenac) following topical use are primarily attained via dermal absorption and systemic circulation.
Elimination. Plasma t1/2 varies considerably among NSAIDs. For example, ibuprofen, diclofenac, and acetaminophen have relatively rapid elimination (t1/2 of 1-4 hours), while piroxicam has a t1/2 of ∼50 hours at steady state that can increase to up to 75 hours in the elderly. Published estimates of the t1/2 of COX-2–selective NSAIDs vary (2-6 hours for lumiracoxib, 6-12 hours for celecoxib, and 20-26 hours for etoricoxib). However, peak plasma concentrations of lumiracoxib at marketed doses considerably exceed those necessary to inhibit COX-2, suggesting an extended pharmacodynamic t1/2. Hepatic biotransformation and renal excretion is the principal route of elimination of the majority of NSAIDs. Some have active metabolites (e.g., fenbufen, nabumetone, meclofenamic acid, sulindac). Elimination pathways frequently involve oxidation or hydroxylation (Table 34–1). Acetaminophen, at therapeutic doses, is oxidized only to a small fraction to form traces of the highly reactive metabolite, N-acetyl-p-benzoquinone imine (NAPQI). When overdosed (usually >10 g of acetaminophen), however, the principal metabolic pathways are saturated, and hepatotoxic NAPQI concentrations can be formed (see Chapters 4 and 6). Rarely, other NSAIDs also may be complicated by hepatotoxicity (e.g., diclofenac, lumiracoxib). Several NSAIDs or their metabolites are glucuronidated or otherwise conjugated. In some cases, such as the propionic acid derivatives naproxen and ketoprofen, the glucuronide metabolites can hydrolyze back to form the active parent drug when the metabolite is not removed efficiently due to renal insufficiency or competition for renal excretion with other drugs. This may prolong elimination of the NSAID significantly. NSAIDs usually are not removed by hemodialysis due to their extensive plasma protein binding; salicylic acid is an exemption to this rule. In general, NSAIDs are not recommended in the setting of advanced hepatic or renal disease due to their adverse pharmacodynamic effects.
Table 34-1Classification and Comparison of Nonsteroidal Analgesics ||Download (.pdf) Table 34-1 Classification and Comparison of Nonsteroidal Analgesics
|CLASS/DRUG (SUBSTITUTION) ||PHARMACOKINETICS || ||DOSINGd || ||COMMENTS ||COMPARED TO ASPIRIN |
|Aspirin (acetyl ester) || |
t1/2, toxic dose
325-650 mg every 4-6 hours
1 g every 4-6 hours
10 mg/kg every 4-6 hours
Permanent platelet COX-1 inhibition (acetylation)
Main side effects:
GI, increased bleeding time, hypersensitivity
Avoid in children with acute febrile illness
|Diflunisal (defluoro-phenyl) || |
|250-500 mg every 8-12 hours || || |
Not metabolized to salicylic acid
Competitive COX inhibitor
Excreted into breast milk
Analgesic and antiinflammatory effects 4-5 times more potent
Antipyretic effect weaker
Fewer platelet and GI side effects
|Para-aminophenol derivative |
|Acetaminophen || |
Glucuronide conjugates (60%); sulfuric acid conjugates (35%)
|10-15 mg/kg every 4 hours (maximum of 5 doses/24 hours) || || |
Weak nonspecific inhibitor at common doses
Potency may be modulated by peroxides
Overdose leads to production of toxic metabolite and liver necrosis
Analgesic and antipyretic effects equivalent
Anti-inflammatory, GI, and platelet effects less than aspirin at 1000 mg/day
|Acetic acid derivatives |
|Indomethacin (methylated indole) || |
O-demethylation (50%); unchanged (20%)
|25 mg 2-3 times/day; 75-100 mg at night || ||Side effects (3-50% of patients): frontal headache, neutropenia, thrombocytopenia; 20% discontinue therapy ||10-40× more potent; intolerance limits dose |
|Sulindac (sulfoxide prodrug) ||Peak Cp |
1-2 hours; 8 hours for sulfide metabolite; extensive enterohepatic circulation
Sulfone and conjugates (30%); sulindac and conjugates (25%)
7 hours; 18 hours for metabolite
|150-200 mg twice/day || ||20% suffer GI side effects; 10% get CNS side effects ||Efficacy comparable |
|Etodolac (pyranocar-boxylic acid) || |
|200-400 mg 3-4 times/day || ||Some COX-2 selectivity in vitro ||100 mg etodolac has similar efficacy to 650 mg of aspirin, but may be better tolerated |
|Tolmetin (heteroaryl acetate derivative) || |
Oxidized to carboxylic acid/other derivatives, then conjugated
|400-600 mg 3 times/day for children (anti-inflammatory) ||20 mg/kg/day in 3-4 divided doses || |
Food delays and decreases peak absorption
May persist longer in synovial fluid to give a biological efficacy longer than its plasma t1/2
|Efficacy similar; 25-40% develop side effects; 5-10% discontinue drug |
|Ketorolac (pyrrolizine carboxylate) || |
99% glucuronide conjugate (90%)
|<65 years: 20 mg (orally), then 10 mg every 4-6 hours (not to exceed 40 mg/24 hours); >65 years: 10 mg every 4-6 hours (not to exceed 40 mg/24 hours) || || |
Commonly given parenterally (60 mg IM followed by 30 mg every 6 hours, or 30 mg IV every 6 hours)
Available as ocular preparation (0.25%); 1 drop every 6 hours
|Potent analgesic, poor anti-inflammatory |
|Diclofenac (phenylacetate derivatives) || |
Glucuronide and sulfide (renal 65%, bile 35%)
|50 mg 3 times/day or 75 mg twice/day || || |
Available as topical gel, ophthalmic solution, and oral tablets combined with misoprostol
First-pass effect; oral bioavailability, 50%
|More potent; 20% develop side effects, 2% discontinue use, 15% develop elevated liver enzymes |
|Fenamates (N-phenyl-anthranilates) |
|Mefenamic acid || |
2-4 hours High Conjugates of 3-hydroxy and 3-carboxyl metabolites (20% recovered in feces)
|500-mg load, then 250 mg every 6 hours || || |
Isolated cases of hemolytic anemia
May have some central action
|Efficacy similar; GI side effects (25%) |
|Meclofenamate || |
Hepatic metabolism; fecal and renal excretion
|50-100 mg 4-6/day (maximum of 400 mg/day) || || ||Efficacy similar; GI side effects (25%) |
|Flufenamic acid ||Not available in the U.S. || || || || || |
|Propionic acid derivatives ||Intolerance of one does not preclude use of another propionate derivative ||Usually better tolerated |
|Ibuprofen ||Peak Cp |
Conjugates of hydroxyl and carboxyl metabolites
200-400 mg every 4-6 hours
300 mg/6-8 hours or 400-800 mg 3-4 times/day
10-15% discontinue due to adverse effects
Children's dosing Antipyretic: 5-10 mg/kg every 6 hours (max: 40 mg/kg/day)
Anti-inflammatory: 20-40 mg/kg/day in 3-4 divided doses
|Naproxen || |
99% (less in elderly)
6-demethyl and other metabolites
250 mg 4 times/day or 500 mg twice/day
|5 mg/kg twice/day || |
Peak anti-inflammatory effects may not be seen until 2-4 weeks of use
Decreased protein binding and delayed excretion increase risk of toxicity in elderly
|More potent in vitro; usually better tolerated; variably prolonged t1/2 may afford cardioprotection in some individuals |
|Fenoprofen || |
Glucuronide, 4-OH metabolite
|200 mg 4-6 times/day; 300-600 mg 3-4 times/day || || ||15% experience side effects; few discontinue use |
|Ketoprofen || |
25 mg 3-4 times/day
| ||30% develop side effects (usually GI, usually mild) |
|Flurbiprofen ||Peak Cp |
Hydroxylates and conjugates
|200-300 mg/day in 2-4 divided doses || ||Available as a 0.03% ophthalmic solution || |
|Oxaprozin || |
Oxidates and glucuronide conjugates
|600-1800 mg/day || ||Long t1/2 allows for daily administration; slow onset of action; inappropriate for fever/acute analgesia || |
|Enolic acid derivatives |
|Piroxicam || |
Hydroxylates and then conjugated
|20 mg/day || ||May inhibit activation of neutrophils, activity of proteoglycanase, collagenases ||Equipotent; perhaps better tolerated 20% develop side effects; 5% discontinue drug |
|Meloxicam ||Peak Cp |
|7.5-15 mg/day || || ||Some COX-2 selectivity, especially at lower doses |
|Nabumetone (naphthyl alkanone) || |
O-demethylation, then conjugation
|500-1000 mg 1-2 times/day || ||A prodrug, rapidly metabolized to 6-methoxy-2-naphthyl acetic acid; pharmacokinetics reflect active compound ||Shows some COX-2 selectivity (active metabolite does not) Fewer GI side effects than many NSAIDs |
|Diaryl heterocyclic NSAIDs (COX-2 selective) ||Evidence for cardiovascular adverse events ||Decrease in GI side effects and in platelet effects |
|Celecoxib [diaryl substituted pyrazone; (sulfonamide derivative)] || |
Carboxylic acid and glucuronide conjugates
|100 mg 1-2 times/day || || |
Substrate for CYP2C9; inhibitor of CYP2D6
Coadministration with inhibitors of CYP2C9 or substrates of CYP2D6 should be done with caution
|See the text for an overview of COX-2 inhibitors |
| || || || || |
All NSAIDs, including selective COX-2 inhibitors, are antipyretic, analgesic, and anti-inflammatory, with the exception of acetaminophen, which is antipyretic and analgesic but is largely devoid of anti-inflammatory activity.
Inflammation. NSAIDs find their chief clinical application as anti-inflammatory agents in the treatment of musculoskeletal disorders, such as rheumatoid arthritis and osteoarthritis. In general, NSAIDs provide mostly symptomatic relief from pain and inflammation associated with the disease and are not considered to be DMARDs. A number of NSAIDs are approved for the treatment of ankylosing spondylitis and gout. The use of NSAIDs for mild arthropathies, together with rest and physical therapy, generally is effective. When the symptoms are limited either to trouble sleeping because of pain or significant morning stiffness, a single NSAID dose given at night may suffice. Patients with more debilitating disease may not respond adequately to full therapeutic doses of NSAIDs and may require aggressive therapy with second-line agents. Substantial inter- and intraindividual differences in clinical response have been noted.
Pain. When employed as analgesics, these drugs usually are effective only against pain of low to moderate intensity, such as dental pain. Although their maximal efficacy is generally much less than the opioids, NSAIDs lack the unwanted adverse effects of opiates in the CNS, including respiratory depression and the potential for development of physical dependence. Co-administration of NSAIDs can reduce the opioid dose needed for sufficient pain control and reduce the likelihood of adverse opioid effects. NSAIDs do not change the perception of sensory modalities other than pain. They are particularly effective when inflammation has caused peripheral and/or central sensitization of pain perception. Thus, postoperative pain or pain arising from inflammation, such as arthritic pain, is controlled well by NSAIDs, whereas pain arising from the hollow viscera usually is not relieved. An exception to this is menstrual pain. The release of PGs by the endometrium during menstruation may cause severe cramps and other symptoms of primary dysmenorrhea; treatment of this condition with NSAIDs has met with considerable success (Marjoribanks et al., 2003). Not surprisingly, the selective COX-2 inhibitors such as rofecoxib and etoricoxib also are efficacious in this condition. NSAIDs are commonly used as first-line therapy to treat migraine attacks and can be combined with second-line drugs, such as the triptans (e.g., TREXIMET, which is a fixed-dose combination of naproxen and sumatriptan), or with antiemetics to aid relief of the associated nausea. NSAIDs lack efficacy in neuropathic pain.
Fever. Antipyretic therapy is reserved for patients in whom fever in itself may be deleterious and for those who experience considerable relief when fever is lowered. Little is known about the relationship between fever and the acceleration of inflammatory or immune processes; it may at times be a protective physiological mechanism. The course of the patient's illness may be obscured by the relief of symptoms and the reduction of fever by the use of antipyretic drugs. NSAIDs reduce fever in most situations, but not the circadian variation in temperature or the rise in response to exercise or increased ambient temperature. Comparative analysis of the impact of tNSAIDs and selective COX-2 inhibitors suggests that COX-2 is the dominant source of PGs that mediate the rise in temperature evoked by bacterial lipopolysaccharide (LPS) administration (McAdam et al., 1999). This is consistent with the antipyretic clinical efficacy of both subclasses of NSAIDs. It seems logical to select an NSAID with rapid onset for the management of fever associated with minor illness in adults.
Fetal Circulatory System. PGs have long been implicated in the maintenance of patency of the ductus arteriosus, and indomethacin, ibuprofen, and other tNSAIDs have been used in neonates to close the inappropriately patent ductus. Conversely, infusion of PGE2 maintains ductal patency after birth. Both COX-1 and COX-2 appear to participate in maintaining patency of the ductus arteriosus in fetal lambs (Clyman et al., 1999), while in mice COX-2 appears to play the dominant role (Loftin et al., 2002). It is not known which isoform or isoforms are involved in maintaining patency of the fetal ductus in utero in humans.
In mice, PGE2 maintains a low ductal smooth muscle cell tone via the Gs-coupled receptor EP4 that increases intracellular cyclic AMP. In the late gestational period, pulmonary expression of an enzyme that eliminates PGE2 from the bloodstream, PG 15-OH dehydrogenase (PGDH), is rapidly upregulated (Coggins et al., 2002). At birth, PGE2 blood concentrations are dramatically reduced by pulmonary PGDH and by additional exposure of circulating PGE2 to this enzyme with the increase of pulmonary blood flow. Due to reduced EP4 signaling, intracellular cyclic AMP concentrations in the ductus arteriosus drop, and vasodilating effects are outweighed by vasoconstrictor signals, which initiate remodeling. However, surprisingly little information about the human ductal physiology exists, and a role for PGDH remains to be established in humans.
Cardioprotection. Ingestion of aspirin prolongs bleeding time. For example, a single 325-mg dose of aspirin approximately doubles the mean bleeding time of normal persons for 4-7 days. This effect is due to irreversible acetylation of platelet COX and the consequent inhibition of platelet function until sufficient numbers of new, unmodified platelets are released from megakaryocytes. It is the permanent and complete suppression of platelet TxA2 formation that is thought to underlie the cardioprotective effect of aspirin. Aspirin reduces the risk of serious vascular events in high-risk patients (e.g., those with previous myocardial infarction) by 20-25%. Low-dose (<100 mg/day) aspirin, which is relatively (but not exclusively) selective for COX-1, is as effective as higher doses (e.g., 325 mg/day) but is associated with a lower risk for GI adverse events. However, low-dose aspirin is not risk free. Placebo-controlled trials reveal that aspirin increases the incidence of serious GI bleeds, reflecting suppression not just of platelet thromboxane, but also reduction of gastroepithelial PGE2 and PGI2. It also increases the incidence of intracranial bleeds. Although benefit from aspirin outweighs these risks in the case of secondary prevention of cardiovascular disease, the issue is much more nuanced in patients who have never had a serious atherothrombotic event (primary prevention); here, prevention of myocardial infarction by aspirin is numerically balanced by the serious GI bleeds it precipitates (Patrono et al., 2005).
Given their relatively short t1/2 and reversible COX inhibition, most other tNSAIDs are not thought to afford cardioprotection, a view supported by most epidemiological analyses (García Rodríguez et al., 2004). Data suggest that cardioprotection is lost when combining low-dose aspirin with ibuprofen. An exception in some individuals may be naproxen. Although there is considerable between-person variation, a small study suggests that platelet inhibition might be anticipated throughout the dosing interval in some, but not all, individuals on naproxen (Capone et al., 2005). Epidemiological evidence of cardioprotection is less impressive; it suggests an ∼10% reduction in myocardial infarction, compared to 20-25% with low-dose aspirin (Antithrombotic Trialists' Collaboration, 2002). This would fit with heterogeneity of response to naproxen. Reliance on prescription databases may have constrained the ability of this approach to address the question with precision. In the Alzheimer's Disease Anti-inflammatory Prevention Trial (ADAPT Research Group, 2008), naproxen was associated with a higher rate of cardiac events than celecoxib. Hence, naproxen should not be used as a substitute for aspirin for cardioprotection. COX-2–selective NSAIDs are devoid of antiplatelet activity, as mature platelets do not express COX-2.
Other Clinical Uses
Systemic Mastocytosis. Systemic mastocytosis is a condition in which there are excessive mast cells in the bone marrow, reticuloendothelial system, GI system, bones, and skin. In patients with systemic mastocytosis, PGD2, released from mast cells in large amounts, has been found to be the major mediator of severe episodes of flushing, vasodilation, and hypotension; this PGD2 effect is resistant to antihistamines. The addition of aspirin or ketoprofen has provided relief (Worobec, 2000). However, aspirin and tNSAIDs can cause degranulation of mast cells, so blockade with H1 and H2 histamine receptor antagonists should be established before NSAIDs are initiated.
Niacin Tolerability. Large doses of niacin (nicotinic acid) effectively lower serum cholesterol levels, reduce low-density lipoprotein, and raise high-density lipoprotein (see Chapter 31). However, niacin is tolerated poorly because it induces intense facial flushing. This flushing is mediated largely by release of PGD2 from the skin, which can be inhibited by treatment with aspirin (Jungnickel et al., 1997), and would be susceptible to inhibition of PGD synthesis, or antagonism of its DP1 receptor.
Bartter Syndrome. Bartter syndrome includes a series of rare disorders (frequency ≤1/100,000 persons) characterized by hypokalemic, hypochloremic metabolic alkalosis with normal blood pressure and hyperplasia of the juxtaglomerular apparatus. Fatigue, muscle weakness, diarrhea, and dehydration are the main symptoms. Distinct variants are caused by mutations in a Na+-K+-2Cl− co-transporter, an apical ATP-regulated K+ channel, a basolateral Cl− channel, a protein (barttin) involved in co-transporter trafficking, and the extracellular Ca2+-sensing receptor. Renal COX-2 is induced, and biosynthesis of PGE2 is increased. Treatment with indomethacin, combined with potassium repletion and spironolactone, is associated with improvement in the biochemical derangements and symptoms. Selective COX-2 inhibitors also have been used (Guay-Woodford, 1998).
Cancer Chemoprevention. Chemoprevention of cancer is an area in which the potential use of aspirin and/or NSAIDs is under active investigation. Epidemiological studies suggested that frequent use of aspirin is associated with as much as a 50% decrease in the risk of colon cancer (Kune et al., 2007). Similar observations have been made with NSAID use in this and other cancers (Harris et al., 2005). NSAIDs have been used in patients with familial adenomatous polyposis (FAP), an inherited disorder characterized by multiple adenomatous colon polyps developing during adolescence and the inevitable occurrence of colon cancer by the sixth decade.
Studies in small numbers of patients over short periods of follow-up have shown a decrease in the polyp burden with the use of sulindac, celecoxib, or rofecoxib (Steinbach et al., 2000; Cruz-Correa et al., 2002; Hallak et al., 2003). Celecoxib is approved as an adjunct to endoscopic surveillance and surgery in FAP based on superiority in a short-term, placebo-controlled trial for polyp prevention/regression. However, more recently, the Adenoma Prevention with Celecoxib (APC) Trial showed a significant reduction in the incidence of adenomatous polyps in patients with a history of colorectal adenomas at high doses of celecoxib (200 mg twice a day and 400 mg twice a day versus placebo) (Bertagnolli et al., 2009). The trial was prematurely terminated because of a 2.5 times increase in cardiovascular risk for patients taking 200 mg twice a day of celecoxib and a 3.4 times increase in risk for patients taking 400 mg twice a day (Solomon et al., 2005). Similarly, the Prevention of colorectal Sporadic Adenomatous Polyps (PreSAP) trial found a reduction of polyps at a single daily dose of 400 mg of celecoxib (Arber et al., 2006), which was offset by an increase in cardiovascular risk (Solomon et al., 2006). Finally, the APPROVe trial of 25 mg of rofecoxib showed a reduction in the incidence of adenomatous polyps (Baron et al., 2006) and an increase in cardiovascular adverse events (Bresalier et al., 2005; Baron et al., 2008) that led to the termination of the trial and to rofecoxib's withdrawal from the market. Controlled evidence is not available to determine if selective COX-2 inhibitors differ from non-COX-2–selective tNSAIDs or aspirin in the extent of adenomatous colorectal polyp reduction in patients with FAP. Likewise, it is unknown whether there is even a clinical benefit from the reduction. Increased expression of COX-2 has been reported in multiple epithelial tumors, and in some cases, the degree of expression has been related to prognosis. Deletion or inhibition of COX-2 dramatically inhibits polyp formation in mouse genetic models of polyposis coli. Although the phenotypes in these models do not completely recapitulate the human disease, deletion of COX-1 had a similar effect. Speculation as to how the two COXs might interact in tumorigenesis includes the possibility that products of COX-1 might induce expression of COX-2. However, the nature of this interaction is poorly understood, as are its therapeutic consequences.
Alzheimer's Disease. Observational studies have suggested that NSAID use, in particular ibuprofen, is associated with lower risk of developing Alzheimer's disease. However, more recent prospective studies, including a randomized, controlled clinical trial comparing celecoxib, naproxen, and placebo (ADAPT Research Group, 2008), did not find a significant reduction in Alzheimer's dementia with the use of NSAIDs.
Adverse Effects of Nsaid Therapy
Common adverse events that complicate therapy with aspirin and NSAIDs are outlined in Table 34–2. Age generally is correlated with an increased probability of developing serious adverse reactions to NSAIDs, and caution is warranted in choosing a lower starting dose for elderly patients. NSAIDs are labeled with a black box warning related to cardiovascular risks and are specifically contraindicated following coronary artery bypass graft (CABG) surgery.
Table 34-2Common and Shared Side Effects of NSAIDs ||Download (.pdf) Table 34-2 Common and Shared Side Effects of NSAIDs
|SYSTEM ||MANIFESTATIONS |
|GI || |
|Platelets || |
Inhibited platelet activationa
Propensity for bruisinga
Increased risk of hemorrhagea
|Renal || |
Salt and water retention
Edema, worsening of renal function in renal/cardiac and cirrhotic patients
Decreased effectiveness of antihypertensive medications
Decreased effectiveness of diuretic medications
Decreased urate excretion (especially with aspirin)
|Cardiovascular || |
Closure of ductus arteriosus
|CNS || |
|Uterus || |
Prolongation of gestation
Inhibition of labor
|Hypersensitivity || |
Gastrointestinal. The most common symptoms associated with these drugs are gastrointestinal, including anorexia, nausea, dyspepsia, abdominal pain, and diarrhea. These symptoms may be related to the induction of gastric or intestinal ulcers, which is estimated to occur in 15-30% of regular users. Ulceration may range from small superficial erosions to full-thickness perforation of the muscularis mucosa. There may be single or multiple ulcers, and ulceration may be uncomplicated or complicated by bleeding, perforation, or obstruction. Blood loss can be gradual, leading to anemia over time, or acute and life-threatening. The risk is further increased in those with Helicobacter pylori infection, heavy alcohol consumption, or other risk factors for mucosal injury, including the concurrent use of glucocorticoids. Although there is a perception that tNSAIDs vary considerably in their tendency to cause erosions and ulcers, this is based on overview analyses of small and heterogeneous studies, often at single doses of individual tNSAIDs.
Large-scale comparative studies of tNSAIDs have not been performed, and there is no reliable information on which to assess the comparative likelihood of GI ulceration on anti-inflammatory doses of aspirin versus tNSAIDs. Thus, most information is derived from the use of surrogate markers or from epidemiological data sets and suggests that the relative risk for serious adverse GI events is elevated about 3-fold in tNSAID users compared to non-users. Epidemiological studies suggest that combining low-dose aspirin (for cardioprotection) with other NSAIDs synergistically increases the likelihood of GI adverse events (see "Drug Interactions"). Similarly, the combination of multiple tNSAIDs and high dosage of a single tNSAIDs (including inappropriately high dosage in the elderly) has been found to raise the risk for ulcer complications by 7- and 9-fold, respectively. Age >70 years alone increases the likelihood of complications almost 6-fold. The risk of ulcer complications is increased in patients with past uncomplicated (6-fold) or complicated ulcers (13-fold), or concurrent drug therapy including warfarin (12-fold), glucocorticoids (4-fold), or selective serotonin reuptake inhibitors (SSRIs; 3-fold) (Gabriel et al., 1991; Dalton et al., 2003; García Rodríguez and Barreales Tolosa, 2007).
COX-2–selective NSAIDs were originally designed for a niche indication, to improve treatment safety for patients at high risk for GI complications requiring chronic tNSAIDs—a population of <5% of tNSAID users (Dai et al., 2005). Prescription behavior changed over time, however, and more than one-third of patients at the lowest risk for GI events received a COX-2 inhibitor in 2002. Paradoxically, the incidence of GI adverse events had been falling sharply in the population prior to the introduction of the coxibs—which were developed to reduce the risk of serious GI complication, perhaps reflecting a move away from use of high-dose aspirin as an anti-inflammatory drug strategy. All selective COX-2 inhibitors are less prone to induce endoscopically visualized gastric ulcers than equally efficacious doses of tNSAIDs (Deeks et al., 2002). A more detailed discussion on this topic can be found in the prior edition of this textbook.
Gastric damage by NSAIDs can be brought about by at least two distinct mechanisms (see Chapter 33). Inhibition of COX-1 in gastric epithelial cells depresses mucosal cytoprotective PGs, especially PGI2 and PGE2. These eicosanoids inhibit acid secretion by the stomach, enhance mucosal blood flow, and promote the secretion of cytoprotective mucus in the intestine. Inhibition of PGI2 and PGE2 synthesis may render the stomach more susceptible to damage and can occur with oral, parenteral, or transdermal administration of aspirin or NSAIDs. There is some evidence that COX-2 also contributes to constitutive formation of these PGs by human gastric epithelium; products of COX-2 certainly contribute to ulcer healing in rodents (Mizuno et al., 1997). This may partly reflect an impairment of angiogenesis by the inhibitors (Jones et al., 1999). Indeed, coincidental deletion or inhibition of both COX-1 and COX-2 seems necessary to replicate NSAID-induced gastropathy in mice, and there is some evidence for gastric pathology in the face of prolonged inhibition or deletion of COX-2 alone (Sigthorsson et al., 2002). Another mechanism by which NSAIDs or aspirin may cause ulceration is by local irritation from contact of orally administered drug with the gastric mucosa. Local irritation allows backdiffusion of acid into the gastric mucosa and induces tissue damage. However, the incidence of GI adverse events is not significantly reduced by formulations that reduce drug contact with the gastric mucosa, such as enteric coating or efferent solutions, suggesting that the contribution of direct irritation to the overall risk is minor. It also is possible that enhanced generation of LOX products (e.g., LTs) contributes to ulcerogenicity in patients treated with NSAIDs.
Co-administration of the PGE1 analog, misoprostol, or proton pump inhibitors (PPIs) in conjunction with NSAIDs can be beneficial in the prevention of duodenal and gastric ulceration (Rostom et al., 2002).
Several groups have attached NO–donating moieties to NSAIDs and to aspirin in the hope of reducing the incidence of adverse events. Benefit might be attained by abrogation of the inhibition of angiogenesis by tNSAIDs during ulcer healing, as observed in rodents (Ma et al., 2002); however, the clinical benefit of this strategy remains to be established. Similarly, LTs may accumulate in the presence of COX inhibition, and there is some evidence in rodents that combined LOX-COX inhibition may be a useful strategy.
Cardiovascular. COX-2–selective NSAIDs were developed to improve the GI safety of anti-inflammatory therapy in patients at elevated risk for GI complications. However, placebo-controlled trials with three structurally distinct compounds—celecoxib, valdecoxib (withdrawn), and rofecoxib (withdrawn)—revealed an increase in the incidence of myocardial infarction, stroke, and thrombosis (Bresalier et al., 2005; Nussmeier et al., 2005; Solomon et al., 2006). The risk appears to also extend to those of the older tNSAIDs, which are quite selective for COX-2, such as diclofenac, meloxicam, and nimesulide (Grosser et al., 2006). Regulatory agencies in the U.S, E.U., and Australia have concluded that all NSAIDs have the potential to increase the risk of heart attack and stroke.
The cardiovascular hazard is plausibly explained by the depression of COX-2-dependent prostanoids formed in vasculature and kidney (Grosser et al., 2006). Vascular PGI2 constrains the effect of prothrombotic and atherogenic stimuli, and renal PGI2 and PGE2 formed by COX-2 contribute to arterial pressure homeostasis (see Chapter 33). Genetic deletion of the PGI2 receptor, IP, in mice augments the thrombotic response to endothelial injury, accelerates experimental atherogenesis, increases vascular proliferation, and adds to the effect of hypertensive stimuli (Cheng et al., 2002; Egan et al., 2004; Kobayashi et al., 2004; Cheng et al., 2006). Genetic deletion or inhibition of COX-2 accelerates the response to thrombotic stimuli and raises blood pressure. Together, these mechanisms would be expected to alter the cardiovascular risk of humans, as COX-2 inhibition in humans depresses PGI2 synthesis (Catella-Lawson et al., 1999; McAdam et al., 1999). Indeed, a human mutation of the IP, which disrupts its signaling, is associated with increased cardiovascular risk (Arehart et al., 2008).
Patients at increased risk of cardiovascular disease or thrombosis are likely to be particularly prone to cardiovascular adverse events while on NSAIDs. This includes patients with rheumatoid arthritis, as the relative risk of myocardial infarction is increased in these patients compared to patients with osteoarthritis or no arthritis. The risk appears to be conditioned by factors influencing drug exposure, such as dose, t1/2, degree of COX-2 selectivity, potency, and treatment duration. Thus, the lowest possible dose should be prescribed for the shortest possible period. NSAIDs with selectivity for COX-2 should be reserved for patients at high risk for GI complications.
Blood Pressure, Renal, and Renovascular Adverse Events. NSAIDs and COX-2 inhibitors have been associated with renal and renovascular adverse events. NSAIDs have little effect on renal function or blood pressure in normal human subjects. However, in patients with congestive heart failure, hepatic cirrhosis, chronic kidney disease, hypovolemia, and other states of activation of the sympathoadrenal or renin–angiotensin systems, PG formation becomes crucial in model systems and in humans. NSAIDs are associated with loss of the PG-induced inhibition of both the reabsorption of Cl− and the action of antidiuretic hormone, leading to the retention of salt and water.
Experiments in mice that attribute the generation of vasodilator PGs (PGE2 and PGI2) to COX-2 raise the likelihood that the incidence of hypertensive complications (either new onset or worsened control) induced by NSAIDs in patients may correlate with the degree of inhibition of COX-2 in the kidney and the selectivity with which it is attained (Qi et al., 2002). Deletion of receptors for both PGI2 and PGE2 elevate blood pressure in mice, mechanistically integrating hypertension with a predisposition to thrombosis. Although this hypothesis has never been addressed directly, epidemiological studies suggest hypertensive complications occur more commonly in patients treated with coxibs than with tNSAIDs.
NSAIDs promote reabsorption of K+ as a result of decreased availability of Na+ at distal tubular sites and suppression of the PG-induced secretion of renin. The latter effect may account in part for the usefulness of NSAIDs in the treatment of Bartter syndrome (see "Bartter Syndrome").
Analgesic Nephropathy. Analgesic nephropathy is a condition of slowly progressive renal failure, decreased concentrating capacity of the renal tubule, and sterile pyuria. Risk factors are the chronic use of high doses of combinations of NSAIDs and frequent urinary tract infections. If recognized early, discontinuation of NSAIDs permits recovery of renal function.
Pregnancy and Lactation. In the hours before parturition, there is induction of myometrial COX-2 expression, and levels of PGE2 and PGF2α increase markedly in the myometrium during labor (Slater et al., 2002). Prolongation of gestation by NSAIDs has been demonstrated in model systems and in humans. Some NSAIDs, particularly indomethacin, have been used off-label to terminate preterm labor. However, this use is associated with closure of the ductus arteriosus and impaired fetal circulation in utero, particularly in fetuses older than 32 weeks' gestation. COX-2–selective inhibitors have been used off-label as tocolytic agents; this use has been associated with stenosis of the ductus arteriosus and oligohydramnios. Finally, the use of NSAIDs and aspirin late in pregnancy may increase the risk of postpartum hemorrhage. Therefore, pregnancy, especially close to term, is a relative contraindication to the use of all NSAIDs. In addition, their use must be weighed against potential fetal risk, even in cases of premature labor, and especially in cases of pregnancy-induced hypertension (Duley et al., 2004).
Hypersensitivity. Certain individuals display hypersensitivity to aspirin and NSAIDs, as manifested by symptoms that range from vasomotor rhinitis, generalized urticaria, and bronchial asthma to laryngeal edema, bronchoconstriction, flushing, hypotension, and shock. Aspirin intolerance is a contraindication to therapy with any other NSAID because cross-sensitivity can provoke a life-threatening reaction reminiscent of anaphylactic shock. Despite the resemblance to anaphylaxis, this reaction does not appear to be immunological in nature.
Although less common in children, this cross-sensitivity may occur in 10-25% of patients with asthma, nasal polyps, or chronic urticaria and in 1% of apparently healthy individuals. It is provoked by even low doses (<80 mg) of aspirin and apparently involves COX inhibition. Cross-sensitivity extends to other salicylates, structurally dissimilar NSAIDs, and rarely acetaminophen (see "Adverse Effects" in the "Acetaminophen" section). Treatment of aspirin hypersensitivity is similar to that of other severe hypersensitivity reactions, with support of vital organ function and administration of epinephrine. Aspirin hypersensitivity is associated with an increase in biosynthesis of LTs, perhaps reflecting diversion of AA to LOX metabolism. Indeed, results in a small number of patients suggest that blockade of 5-LOX with the drug zileuton or off-label use of the LT-receptor antagonists may ameliorate the symptoms and signs of aspirin intolerance, albeit incompletely.
Aspirin Resistance. All forms of treatment failure with aspirin have been collectively called aspirin resistance. Although this has attracted much attention, there is little information concerning the prevalence of a stable, aspirin-specific resistance or the precise mechanisms that might convey this "resistance." Genetic variants of COX-1 that co-segregate with resistance have been described, but the relation to clinical outcome is not clear.
Reye's Syndrome. Due to the possible association with Reye's syndrome, aspirin and other salicylates are contraindicated in children and young adults <20 years of age with viral illness–associated fever. Reye's syndrome, a severe and often fatal disease, is characterized by the acute onset of encephalopathy, liver dysfunction, and fatty infiltration of the liver and other viscera (Glasgow and Middleton, 2001). The etiology and pathophysiology are not clear, nor is it clear whether a causal relationship between aspirin and Reye's syndrome exists (Schror, 2007). However, the epidemiologic evidence for an association between aspirin use and Reye's syndrome seemed sufficiently compelling that labeling of aspirin and aspirin-containing medications to indicate Reye's syndrome as a risk in children was first mandated in 1986 and extended to bismuth subsalicylate in 2004. Since then, the use of aspirin in children has declined dramatically, and Reye's syndrome has almost disappeared. Acetaminophen has not been implicated in Reye's syndrome and is the drug of choice for antipyresis in children, teens, and young adults.
Concomitant NSAIDs and Low-Dose Aspirin. Many patients combine either tNSAIDs or COX-2 inhibitors with "cardioprotective," low-dose aspirin. Epidemiological studies suggest that this combination therapy increases significantly the likelihood of GI adverse events over either class of NSAID alone.
Prior occupancy of the active site of platelet COX-1 by the commonly consumed tNSAID ibuprofen impedes access of aspirin to its target Ser 529 and prevents irreversible inhibition of platelet function (Catella-Lawson et al., 2001). Epidemiological studies have provided conflicting data as to whether this adversely impacts clinical outcomes, but they generally are constrained by the use of prescription databases to examine an interaction between two drug groups commonly obtained without prescription. Evidence in support of this interaction has been observed in comparing ibuprofen-treated patients with and without aspirin in two coxib outcome studies (CLASS and TARGET), but the trials were not powered to address this question definitively. In theory, this interaction should not occur with selective COX-2 inhibitors, because mature human platelets lack COX-2. However, the GI safety advantage of NSAIDs selective for COX-2 is lost when they are combined with low-dose aspirin.
Angiotensin-converting enzyme (ACE) inhibitors act, at least partly, by preventing the breakdown of kinins that stimulate PG production. Thus, it is logical that NSAIDs might attenuate the effectiveness of ACE inhibitors by blocking the production of vasodilator and natriuretic PGs. Due to hyperkalemia, the combination of NSAIDs and ACE inhibitors also can produce marked bradycardia leading to syncope, especially in the elderly and in patients with hypertension, diabetes mellitus, or ischemic heart disease. Corticosteroids and SSRIs may increase the frequency or severity of GI complications when combined with NSAIDs. NSAIDs may augment the risk of bleeding in patients receiving warfarin both because almost all tNSAIDs suppress normal platelet function temporarily during the dosing interval and because some NSAIDs also increase warfarin levels by interfering with its metabolism; thus, concurrent administration should be avoided. Many NSAIDs are highly bound to plasma proteins and thus may displace other drugs from their binding sites. Such interactions can occur in patients given salicylates or other NSAIDs together with warfarin, sulfonylurea hypoglycemic agents, or methotrexate; the dosage of such agents may require adjustment to prevent toxicity. Patients taking lithium should be monitored because certain NSAIDs (e.g., piroxicam) can reduce the renal excretion of this drug and lead to toxicity, while others can decrease lithium levels (e.g., sulindac).
Pediatric and Geriatricindications and Problems
Therapeutic Uses in Children. Therapeutic indications for NSAID use in children include fever, mild pain, postoperative pain, and inflammatory disorders, such as juvenile arthritis and Kawasaki disease. Inflammation associated with cystic fibrosis has emerged as a potential indication for pediatric NSAID use (Konstan et al., 1995); however, concern about GI adverse effects has limited NSAID use for this indication. The choice of drugs for children is considerably restricted; only drugs that have been extensively tested in children should be used (acetaminophen, ibuprofen, and naproxen).
Kawasaki Disease. Aspirin generally is avoided in pediatric populations due to its potential association with Reye's syndrome (see "Reye's Syndrome"). However, high doses of aspirin (30-100 mg/kg/day) are used to treat children during the acute phase of Kawasaki disease, followed by low-dose antiplatelet therapy in the subacute phase. Aspirin is thought to reduce the likelihood of aneurysm formation as a consequence of the vasculitis particularly in the coronary arteries. Small randomized studies did not conclusively show whether aspirin adds benefit beyond the standard treatment of Kawasaki disease with intravenous immunoglobulin (Baumer et al., 2006).
Pharmacokinetics in Children. Despite the recognition that age-dependent differences in gastric emptying time, plasma protein binding capacity and oxidative liver metabolism affect NSAID pharmacokinetics in children, dosing recommendations frequently are based on extrapolation of pharmacokinetic data from adults. The majority of pharmacokinetic studies that were performed in children involved patients >2 years of age, which often provide insufficient data for dose selection in younger infants. For example, the pharmacokinetics of the most commonly used NSAID in children, acetaminophen, differ substantially between the neonatal period and older children or adults. The systemic bioavailability of rectal acetaminophen formulations in neonates and preterm babies is higher than in older patients. Acetaminophen clearance is reduced in preterm neonates probably due to their immature glucuronide conjugation system (sulphatation is the principal route of biotransformation at this age). Therefore, acetaminophen dosing intervals need to be extended (8-12 hours) or daily doses reduced to avoid accumulation and liver toxicity. Aspirin elimination also is delayed in neonates and young infants compared to adults bearing the risk of accumulation.
Disease also may affect NSAID disposition in children. For example, ibuprofen plasma concentrations are reduced and clearance increased (∼80%) in children with cystic fibrosis. This probably is related to the GI and hepatic pathologies associated with this disease. Aspirin's kinetics are markedly altered during the febrile phase of rheumatic fever or Kawasaki vasculitis. The reduction in serum albumin associated with these conditions causes an elevation of the free salicylate concentration, which may saturate renal excretion and result in salicylate accumulation to toxic levels. In addition to dose reduction, monitoring of the free drug may be warranted in these situations.
Pharmacokinetics in the Elderly. Physiological changes in absorption, distribution, and elimination in aging patients can be expected to affect the pharmacokinetics of most drugs, including NSAIDs. Coincident diseases may further complicate the prediction of the response to the drug. The clearance of many NSAIDs is reduced in the elderly due to changes in hepatic metabolism. Particularly NSAIDs with a long t1/2 and primarily oxidative metabolism (i.e., piroxicam, tenoxicam, celecoxib) have elevated plasma concentrations in elderly patients. For example, plasma concentrations after the same dose of celecoxib may rise up to 2-fold higher in patients >65 years of age than in patients <50 years of age (U.S. Food and Drug Administration, 2001), warranting careful dose adjustment. The capacity of plasma albumin to bind drugs is diminished in older patients and may result in higher concentrations of unbound NSAIDs. Free naproxen concentrations, e.g., are markedly increased in older patients, although total plasma concentrations essentially are unchanged. The higher susceptibility of older patients to GI complications may be due to both a reduction in gastric mucosal defense and to elevated total and/or free NSAID concentrations. Generally, it is advisable to start most NSAIDs at a low dosage in the elderly and increase the dosage only if the therapeutic efficacy is insufficient.