The recognition that nonsteroidal anti-inflammatory drugs (NSAIDs) increase not only gastrointestinal (GI) but cardiovascular (CV) adverse events as well has created a dilemma for practicing physicians. Clinicians selecting appropriate NSAID therapy must estimate each patient’s baseline risk for both (GI) and (CV) adverse events, and then estimate the impact of each medication (and its dose) for the individual patient. To synthesize a rational current treatment approach, we have developed a 2×2 table to guide NSAID medication choice, considering the use of concomitant aspirin as well as gastroprotective therapy. COX-2 inhibitors were an important scientific advance in pain therapy, and using them in a safe and cost-effective manner is possible when all the competing risks are carefully weighed.
“The whole imposing edifice of modern medicine is like the celebrated tower of Pisa – slightly off balance.” —Charles, Prince of Wales
The recognition that users of nonsteroidal anti-inflammatory drugs (NSAIDs) experience an increased likelihood of cardiovascular (CV) adverse events has led patients and clinicians to steer clear from their use. This astonishing recognition of an unexpected side effect, from such a commonly used medication class, may never have been discovered had novel compounds not been developed to reduce their well-documented gastrointestinal (GI) complications. NSAIDs provide important benefits by controlling pain and inflammation for many patients. The alternatives, undertreatment of pain or use of narcotic analgesics, have their own significant adverse consequences. As weighing risks and benefits of any medical intervention is an essential component of clinical decision making, it has become increasingly complex to decide in whom the use of any NSAID—with/without a gastroprotective agent or with/without concomitant aspirin—is appropriate. Given the multiplicity, and in many cases rarity, of adverse events, it is unlikely that a single “megatrial” will adequately address the numerous safety and effectiveness issues. In the meantime, careful synthesis of the literature—not complete reliance on individual studies that address one outcome—is required to inform medication selection. This step-wise approach is crucial, since appropriate NSAID choice should be driven by an individual patient’s CV and/or GI risk assessment rather than a medication’s reported adverse event rate in a study of a heterogeneous population.
Clinicians are faced with a real challenge, as they must estimate each patient’s baseline risk for both GI and CV adverse events independent of drug therapy and then multiply the impact of various risk modifiers (individual drugs and doses) to estimate the overall patient risk. For example, reducing an otherwise healthy 25-year-old patient’s risk of a GI bleed on anti-inflammatory therapy by 50% by selecting a safer drug translates to absolute risk change of less than 1/1000. Compare this to the impact in a 70-year-old with a history of peptic ulcer, in whom the risk of recurrent bleeding at 1 year on an NSAID is at least 10%—two orders of magnitude greater. The same concepts apply to the confirmed but less well-understood increased risk of adverse CV events. This review provides clinicians a framework to implement a rational approach to these competing risks by providing prescribers a simplified method of individualized medication selection.
What are the Cardiovascular risks of Nonsteroidal Anti-Inflammatory Drugs?
Both nonselective and cyclooxygenase (COX)-2–selective NSAIDs have been associated with an increased incidence of hospitalization for congestive heart failure and elevated blood pressure. Their specific effect on other serious CV events, including myocardial infarction (MI) and death, has been more controversial. Although there are theoretical reasons why COX-2 selective agents may be prothrombotic, such as decreased production of the vasodilator prostacyclin (which also inhibits platelet activation) with concomitant unimpeded COX-1–mediated production of platelet thromboxane A2, the mechanistic relationship to CV events remains incompletely understood.
The CV risk of certain NSAIDs may be aggravated due to a clinically important interaction with low-dose aspirin that has been identified for some, but not all, drugs. Catella-Lawson and colleagues investigated the potential interactions between aspirin and several different NSAIDs and found that ibuprofen could mitigate the cardioprotective effects of aspirin by interfering with aspirin’s platelet inhibition. This interaction, thought to be due to ibuprofen blocking aspirin’s access to the COX-1 enzyme, was not seen with COX-2 selective inhibitors as well as diclofenac. Platelet aggregation studies support preserved antiplatelet activity in patients taking naproxen, which may override any competition the drug may have at the platelet-binding site.
The Vioxx GI Outcomes Research Study (VIGOR) trial studied rheumatoid arthritis patients (who are at increased CV risk due to their chronic inflammatory disease) and showed a higher rate of adverse CV events (nonfatal MI, nonfatal stroke, and death) on 50 mg rofecoxib every day compared with 500 mg naproxen twice daily (0.8% vs 0.4%, P <.05). These results were largely due to differences in MI (0.4% vs 0.1%; P <.01). No low-dose aspirin was allowed in this study. Although debate centered on whether these findings were due to a cardioprotective effect of naproxen or a deleterious effect of rofecoxib, the adenomatous Polyp Prevention on Vioxx (APPROVe) trial, a randomized, placebo-controlled study that enrolled 2600 patients with a history of colorectal adenomas, confirmed an increased incidence of MI or stroke for those taking 25 mg rofecoxib daily (3.5% vs 1.9%; P <.001). The manufacturer subsequently withdrew rofecoxib from the market.
Valdecoxib was withdrawn on request of the Food and Drug and Administration in April 2005. Two trials in patients undergoing coronary artery bypass grafting showed that there was an increase in CV and thromboembolic events. A systematic review and meta-analysis of CV risk with etoricoxib therapy have also been undertaken, and, although the data are limited, there is evidence of an increased CV risk (odds ratio [OR], 1.49) as well.
Celecoxib is the only COX-2 selective inhibitor remaining on the US market. The Adenoma Prevention with Celecoxib trial included 2026 patients who were assigned to one of three groups: placebo, celecoxib 200 mg twice daily, or celecoxib 400 mg twice daily. Like APPROVe, it was designed to assess the impact of celecoxib in preventing the recurrence of polyps. The study was stopped early by the National Cancer Institute after a mean follow-up of 33 months due to a significantly higher number of celecoxib-treated patients experiencing a MI, stroke, or CV death compared with those receiving placebo (OR, 2.5 for celecoxib 400 mg daily; 95% confidence interval [CI], 1.0–7.0; P = .06, and OR, 3.4 for celecoxib 800 mg daily 95% CI, 1.4–9.3; P = .009). In the Prevention of Spontaneous Adenomatous Polyps (PreSAP) trial, no increase in CV events was found for those patients assigned 400 mg celecoxib daily compared with placebo. Although there are no data to implicate celecoxib 200 mg daily with adverse CV outcomes, we must await a large CV outcome trial in progress. A patient-level pooled analysis of adjudicated data from 7950 patients in 6 placebo-controlled trials comparing celecoxib with placebo for conditions other than arthritis was recently published. With 16,070 patient years of follow-up, the hazard ratio overall was 1.6 (95% CI, 1.1–2.3). The risk, increased with dose, was lowest for the 400 mg every day dose (hazard ratio, 1.1; 95% CI, 0.6–2.0), intermediate for the 200 mg twice a day dose (hazard ratio, 1.8; 5% CI, 1.1–3.1), and highest for the 400 mg twice a day dose (hazard ratio, 3.1; 95% CI, 1.5–6.1). Celecoxib was associated with increased risk regardless of baseline aspirin use—suggesting that the mechanism of CV toxicity is more complex than unopposed COX-2 inhibition. Patients at highest baseline risk demonstrated disproportionately greater risk of celecoxib-related adverse events, strongly emphasizing the issue of appropriate patient selection. This has been emphasized by a recent observational study of Medicare beneficiaries as well. The patient characteristics that were found to increase the risk of cardiovascular disease (CVD) events in that study included age greater than 80 years, hypertension, prior MI, prior CVD, rheumatoid arthritis, chronic renal disease, and chronic obstructive pulmonary disease.
What are the Cardiovascular risks of nonselective Nonsteroidal Anti-Inflammatory Drugs?
In a widely quoted meta-analysis of randomized, controlled trials, performed by the antiplatelet trialist collaboration, the authors concluded that COX-2 NSAIDs as a class were associated with a two-fold greater risk of MI, a 1.5-fold greater risk of vascular death, and no increased risk of stroke. However, compared with placebo, nonselective NSAIDs other than naproxen were also associated with an increased risk and may be associated with a higher risk of stroke than COX-2 NSAIDs. For patients at higher risk for CV events, switching from a COX-2 NSAID to a nonselective NSAID may not decrease treatment-related CV risk, but it is likely to increase GI risk. This meta-analysis suggests, however, that for patients with higher CV risks, naproxen may be the safest NSAID. A large systematic review of observational studies confirmed these results as well. Among nonselective drugs, diclofenac had the highest risk with a summary relative risk (RR) of 1.40 (95% CI, 1.16–1.70). The other drugs had summary RRs close to one: naproxen, 0.97 (95% CI, 0.87–1.07); piroxicam, 1.06 (95% CI, 0.70–1.59); and ibuprofen, 1.07 (95% CI, 0.97–1.18).
Although a small, single study, the Alzheimer Disease Anti-inflammatory Prevention Trial reported an excess number of unadjudicated CV events among those assigned naproxen (at over-the-counter low doses) compared with those in the placebo group. Methodological limitations limit this trial from reversing the evidence provided by meta-analysis.
The overall conclusion is that an increase in overall CV risk is not confined to COX-2 NSAIDs and may be shared by a number of nonselective NSAIDs. It is likely that dose and duration of COX-2 inhibition are important, which may explain why celecoxib at twice-daily dosing has been associated with adverse CV events but every day dosing has not. A further mechanism that may underlie differences between medications is their variable effect on endothelial function, supported by recent data, demonstrating marked intra- and inter-individual responses to COX-2 NSAIDs.
What are the Cardiovascular risks of nonselective Nonsteroidal Anti-Inflammatory Drugs?
In a widely quoted meta-analysis of randomized, controlled trials, performed by the antiplatelet trialist collaboration, the authors concluded that COX-2 NSAIDs as a class were associated with a two-fold greater risk of MI, a 1.5-fold greater risk of vascular death, and no increased risk of stroke. However, compared with placebo, nonselective NSAIDs other than naproxen were also associated with an increased risk and may be associated with a higher risk of stroke than COX-2 NSAIDs. For patients at higher risk for CV events, switching from a COX-2 NSAID to a nonselective NSAID may not decrease treatment-related CV risk, but it is likely to increase GI risk. This meta-analysis suggests, however, that for patients with higher CV risks, naproxen may be the safest NSAID. A large systematic review of observational studies confirmed these results as well. Among nonselective drugs, diclofenac had the highest risk with a summary relative risk (RR) of 1.40 (95% CI, 1.16–1.70). The other drugs had summary RRs close to one: naproxen, 0.97 (95% CI, 0.87–1.07); piroxicam, 1.06 (95% CI, 0.70–1.59); and ibuprofen, 1.07 (95% CI, 0.97–1.18).
Although a small, single study, the Alzheimer Disease Anti-inflammatory Prevention Trial reported an excess number of unadjudicated CV events among those assigned naproxen (at over-the-counter low doses) compared with those in the placebo group. Methodological limitations limit this trial from reversing the evidence provided by meta-analysis.
The overall conclusion is that an increase in overall CV risk is not confined to COX-2 NSAIDs and may be shared by a number of nonselective NSAIDs. It is likely that dose and duration of COX-2 inhibition are important, which may explain why celecoxib at twice-daily dosing has been associated with adverse CV events but every day dosing has not. A further mechanism that may underlie differences between medications is their variable effect on endothelial function, supported by recent data, demonstrating marked intra- and inter-individual responses to COX-2 NSAIDs.
Gastrointestinal adverse events associated with Nonsteroidal Anti-Inflammatory Drug therapy
The pathophysiological basis for NSAID-induced GI injury is that a relative deficiency of gastroduodenal mucosal prostaglandin initiates ulceration. This promotes disruption of the protective mucus layer, inhibition of protective bicarbonate secretion, vasoconstriction (causing local tissue hypoxia), and a topical effect, in which the NSAIDs are “trapped” within cell membranes, leading to superficial epithelial necrosis. Most of the mucosal prostaglandins are derived from COX-1 in the upper GI tract. The “COX-2 hypothesis,” that sparing COX-1 can eliminate ulcer risk, was challenged by animal studies that indicated that both COX-1 and COX-2 must be inhibited for gastric ulceration to occur. Thus, the explanation for the reduced GI toxicity of COX-2–specific inhibitors appears to be their lack of dual COX inhibition, rather than COX-1 sparing effects. This explains why in patients taking both cardioprotective aspirin (primarily a COX-1 inhibitor) and a COX-2 inhibitor, the ulcer risk of a dual COX inhibitor would be predicted. The large outcome studies support these predictions.
The serious complications of NSAIDs include symptomatic gastric and duodenal ulcers and their complications—perforation and hemorrhage—which may occur during acute or chronic therapy. Ulcer bleeding is a composite of both medication-related injury and drug-induced increased propensity for bleeding. Thus, drugs with more prolonged and complete COX-1 inhibition will not only produce more ulcers but, due to their antiplatelet action, also promote bleeding. This is particularly magnified by the use of aspirin in addition to any chronic NSAID treatment regimen.
Although NSAID use may be accompanied by dyspeptic symptoms, many patients have no antecedent symptoms, particularly the elderly, whose first presentation of NSAID damage may be GI hemorrhage or perforation. Endoscopic surveys indicate that up to 20% to 30% of regular NSAID users develop ulcers, most never becoming clinically apparent, although symptoms—dyspepsia and heartburn—may occur in up to 60% of patients taking NSAIDs. The annual incidence of NSAID-related clinical upper GI events (complicated and symptomatic ulcers) lies between 2.5% and 4.5%, and the annual incidence of serious NSAID complications (perforation, hemorrhage, and obstruction) amounts to 1% to 1.5%.
The risk of developing complications of NSAID therapy is related to a number of factors, which include the type and dose of NSAID used, whether or not aspirin is coprescribed, the use of multiple NSAIDs, the duration of NSAID use, and concomitant drug prescription, such as antiplatelet, anticoagulant, or selective serotonin reuptake inhibitor therapy. Nondrug risk factors include the patient’s age, whether there is a history of peptic ulcer, uncomplicated or complicated, or of dyspepsia, particularly if it persists on antisecretory therapy, Helicobacter pylori infection status, and the presence of concurrent diseases. For example, the RR of developing clinically significant adverse events rises to 4.4 with concomitant use of corticosteroids at doses above 10 mg/d and to 12.7 with concomitant anticoagulant prescription. There are similar substantial increases in the RR of ulcer complications for patients aged over 75 years (RR, 10.6) and those with a history of complicated peptic ulceration (RR, 12.5–15.4).
Related posts:
The Role of Proton Pump Inhibitors in the Management of Upper Gastrointestinal Bleeding
Management of Massive Peptic Ulcer Bleeding
Predicting Poor Outcome from Acute Upper Gastrointestinal Hemorrhage
Nonsteroidal Anti-Inflammatory Drugs and Lower Gastrointestinal Complications
Helicobacter pylori -Negative Nonsteroidal Anti-Inflammatory Drug-Negative Ulcer
Differences in Peptic Ulcer Between the East and the West
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
