BARCELONA, Spain—Two large epidemiologic studies reported by Gurkirpal Singh, MD, and colleagues at the EULAR 2007 meeting show that the risk of stroke with NSAIDs in arthritis patients is a function of the individual drug's effect on blood pressure and is not related to COX-2 selectivity and that the apparent gastroprotective effect of some coxibs may be largely an artifact of the comparator NSAID chosen for each clinical trial.1,2

In the first gastroprotection study, Dr. Singh, with the department of medicine at Stanford University School of Medicine, in Palo Alto, California, and colleagues used data from the California Medicaid program—the largest in the US with 7 million participants. For the stroke risk analysis the investigators conducted a nested case-control review of 15,659 cases of stroke, including 3.1 million person-years of follow-up, in arthritis patients treated with an NSAID.

"It is intriguing that NSAIDs known to cause hypertension, such as rofecoxib, valdecoxib, and ibuprofen, increase the risk of stroke, regardless of their COX-2 selectivity. NSAIDs that do not increase blood pressure to the same extent, such as celecoxib and diclofenac, do not increase stroke risk."—Gurkirpal Singh, MD.
"Several NSAIDs increase the risk of stroke in patients with arthritis. It is intriguing that NSAIDs known to cause hypertension, such as rofecoxib, valdecoxib, and ibuprofen, increase the risk of stroke, regardless of their COX-2 selectivity. NSAIDs that do not increase blood pressure to the same extent, such as celecoxib and diclofenac, do not increase stroke risk," Dr. Singh said.

In the second gastroprotection study, an attempt to differentiate among the various safety profiles of NSAIDs included 688,424 patients contributing to 2,665,611 person-years of observation time. Dr. Singh found that 11,303 patients were hospitalized because of complicated gastric or duodenal ulcers, compared with 45,212 matched controls. Multivariate analysis showed that all of the nonselective NSAIDs were associated with increased GI risk, but that this risk was greater for naproxen than for other drugs, such as diclofenac.

"Close attention should be paid to the choice of comparator NSAID in studies purporting to show a GI safety advantage of selective COX-2 inhibitors. Since nonselective NSAIDs differ in their risk of serious GI toxicity, a clinical trial comparison using a ‘high-toxic' NSAID, such as naproxen, is more likely to lead to a statistical advantage of selective COX-2 inhibitors [as in the VIGOR trial] as opposed to a comparison using a ‘low toxic' NSAID, such as diclofenac, [as in the EDGE and MEDAL studies]," Dr. Singh concluded. "There may be individual differences between GI safety profiles of selective COX-2 inhibitors, but this will require head-to-head clinical trials to confirm."

References

1. Singh G, Mannalithara A, Mithal A, et al. When COX-2 selectivity does not matter: risk of stroke with NSAIDs in patients with arthritis. Presented at: EULAR 2007 Meeting; June 13-16, 2007; Barcelona, Spain. Abstract SAT0264.
2. Singh G, Wang H, Mithal A, et al. Comparing apples with oranges: choice of comparator non-selective NSAIDs alters the gastrointestinal safety advantage of COX-2 inhibitors in clinical trials of arthritis patients. Presented at: EULAR 2007 Meeting; June 13-16, 2007; Barcelona, Spain. Abstract SAT0263.