NSAID switching and short‐term gastrointestinal outcome rates after the withdrawal of rofecoxib

OBJECTIVE: The consequences of the rofecoxib withdrawal on upper GI toxicity are largely unknown. We sought to estimate the effect of switching from selective Cox-2 inhibitors to non-selective non-steroidal anti-inflammatory drugs (nsNSAIDs) on the incidence of upper GI adverse events following the withdrawal of rofecoxib on 30 September 2004. METHODS: We identified a cohort of 33 045 patients with arthritis and chronic use of any selective Cox-2 inhibitor during the 6 months before the withdrawal of rofecoxib in claims data from several US health plans. We calculated monthly rates of hospitalization for upper GI adverse events or upper GI endoscopy for the 6 months before and 3 months after the switching and compared the time trends in outcomes. RESULTS: In the subgroup of 15 916 patients using rofecoxib immediately before its withdrawal, 2626 (17%) switched to nsNSAIDs without co-prescribing of a gastroprotective drug, 146 (1%) with a gastroprotective drug, and 5246 (33%) switched to either celecoxib or valdecoxib. Among those switching to nsNSAID without gastroprotection, time trends of upper GI hospitalization rates and endoscopies did not significantly increase compared with those switching to celecoxib or valdecoxib (+0.3 per 1000 per month; 95%CI -3.0 to 3.5). The visit rate for peptic ulcer disease (PUD) increased more in the group switching to nsNSAIDs without gastroprotection (+5.2 visits per 1000 per month; 1.2-9.2) compared with the group switching to another coxib. CONCLUSIONS: Short-term follow-up data suggest that the sudden shift from rofecoxib to nsNSAIDs without gastroprotection did not increase the rate of hospitalization for GI complications but increased outpatient visits for PUD.