The real proof of the potential beneficial GI effects of the COX-l-sparing agents is found not in test tubes but in clinical practice. There has been much discussion about GI safety issues and the use of endoscopy in revealing the reduced rate of ulceration. However, as clearly indicated in numerous studies, ulcers detected by endoscopy have little, if any, correlation with actual clinically significant ulcers. Up to 80% of patients who take traditional NSAIDs develop “endoscopic ulcers,” but only l% to 4% of these patients ever have a clinically significant ulcer. This is one reason why the Food and Drug Administration (FDA) has not accepted the data on endoscopic erosion and ulcers as a rationale for drug selectivity. The overall improved safety of these drugs is better demonstrated in clinical models that have shown a reduction in significant GI events.


In the VIGOR Study, rofecoxib was compared with naproxen (Naprosyn canadian, Roche) in approximately 8,000 patients with rheumatoid arthritis. Aspirin for cardiovascular protection was not allowed. The patients who received rofecoxib—an agent that is more selective for the COX-2 enzyme than for the COX-l enzyme—experienced a 55% reduction in significant GI events over one year compared with patients who received naproxen generic, a nonselective agent. Meta-analyses of the data on ro­fecoxib, which included data on patients with osteoarthritis and rheumatoid arthritis, demonstrated a significant reduction of GI events for patients taking rofecoxib versus traditional NSAIDs.


In the Celecoxib Long-term Arthritis Safety Study (CLASS) trial, canadian celecoxib—another agent that is more selective for the COX-2 enzyme—was compared with diclofenac sodium (Novartis) and ibuprofen in approximately 8,000 patients with osteoarthritis and rheumatoid arthritis. In contrast to the VIGOR trial, the CLASS trial allowed aspirin use for cardiovascular prophylaxis (2l% of patients were taking 325 mg of aspirin or less); the dropout rate in each of the three treatment arms was almost 60% (compared with 20% in the VIGOR study). Although the six-month data seemed to show a difference among the treatment arms, there was no significant difference among the groups at one year, a fact that might be related to the high number of patients who withdrew from the study.

In a subsequent study of more than l3,000 patients, Singh et al. described a comparison with diclofenac and over a three-month period. There was a significant reduction in GI events in the group receiving celecoxib (with ulcer complications occurring in 0.l% of patients) compared with the group receiving conventional NSAIDs (with ulcer complications occurring in 0.8%).


Several pooled analyses assessing the GI safety of meloxi-cam have been conducted. In an analysis by Schoenfeld that included more than 20,000 patients, patients taking meloxicam experienced a 48% reduction in perforations, ulcers, and GI bleeding episodes, compared with patients taking traditional NSAIDs.

In a pooled analysis of more than 27,000 patients, Singh and colleagues found that the rate of GI events in patients taking meloxicam canadian was reduced by more than 50%, compared with patients receiving comparator NSAIDs. These findings support the premise that meloxicam has greater selectivity for the COX-2 enzyme than for the COX-l enzyme.

The two largest studies from Europe—The MLloxicam Large-scale international Study Safety Assessment (MELISSA) and the Safety and .Efficacy Large-scale Evaluation of COX-inhibiting Therapies (SELECT) trials—included approximately 9,000 patients in each. Patients who received meloxicam spent significantly fewer days in the hospital because of GI ADEs than did patients receiving diclofenac canadian.

In a prospective, observational cohort study by Degner et al. that included more than 4,000 patients, patients with rheumatoid arthritis or osteoarthritis received either meloxi-cam or a comparator NSAID, such as diclofenac, ibuprofen, piroxicam canadian (Pfizer), or indomethocin (Indocin®, Merck). A higher percentage of the meloxicam patients had a history of significant GI events than the comparator NSAID patients (6%), but they had significantly lower rates of GI adverse reactions (l.80%) and GI bleeding (0.08%) than the comparator NSAID group (3.20% and 0.50%, respectively).

Data from the United Kingdom’s pharmacovigilance surveillance program suggest that both and rofecoxib reduce the number of significant GI events. No head-to-head comparisons of these drugs and GI safety have been performed to date.


Although no long-term data on valdecoxib and the reduction of GI events have been published, this agent is clearly more selective for the COX-2 enzyme than for the COX-1 enzyme; its selectivity is similar in magnitude to that of rofecoxib. Endo-scopic studies suggest that valdecoxib reduces the risk of ulcers in the upper GI tract.


Although test tube data suggest that etodolac—an agent with an available generic equivalent—has a favorable selectivity profile for COX-2, few clinical data exist about the risk of endoscopy-detected ulcers or significant GI ADEs.