Endoscopic Ulcers with Low-Dose Aspirin and Reality Testing

David Y Graham,Francis K L Chan

DOI: https://doi.org/10.1053/j.gastro.2005.01.022

IF: 29.4

2005-01-01

Gastroenterology

Abstract:Recently Merck reported the results of a large multicenter trial comparing placebo, low-dose (enteric-coated) aspirin, low-dose aspirin plus rofecoxib, and ibuprofen among patients with osteoarthritis.1Laine L. Maller E.S. Yu C. Quan H. Simon T. Ulcer formation with low-dose enteric coated aspirin and the effect of COX-2 selective inhibition a double-blind trial.Gastroenterology. 2004; 127: 395-402Abstract Full Text Full Text PDF PubMed Scopus (177) Google Scholar The outcome measure was the presence of endoscopic ulcers within a 12-week follow-up. They report that gastroduodenal ulcers occurred in 4.2%, which is an annual rate of 16.8% (16.8 per hundred patients per year), which was similar to the prevalence of in the low-dose aspirin-alone group. The reality test for this study is whether the proportion of ulcer in the placebo group is within the range expected among asymptomatic individuals taking no ulcerogenic medications. The source of the enteric-coated aspirin is not given. Enteric coating have been shown to be toxic to the colon and it is possible that the enteric coating used was also toxic in the stomach.2van Velzen V.D. Ball L.M. Dezfulian A.R. Southgate A. Howard C.V. Comparative and experimental pathology of fibrosing colonopathy.Postgrad Med J. 1996; 72: S39-S48PubMed Google Scholar Although the inclusion of patients with H pylori infection or prior history of ulcers or ulcer complications is a definite potential bias,3Chan F.K. Graham D.Y. Review article prevention of non-steroidal anti-inflammatory drug gastrointestinal complications-review and recommendations based on risk assessment.Aliment Pharmacol Ther. 2004; 19: 1051-1061Crossref PubMed Scopus (122) Google Scholar the issue of the effect of prior upper GI events was addressed by stratification whereas the presence of H pylori infection was not.1Laine L. Maller E.S. Yu C. Quan H. Simon T. Ulcer formation with low-dose enteric coated aspirin and the effect of COX-2 selective inhibition a double-blind trial.Gastroenterology. 2004; 127: 395-402Abstract Full Text Full Text PDF PubMed Scopus (177) Google Scholar However, the authors examined whether the presence of H pylori infection or the history of prior upper GI events explained the high incidence of ulcers and they did not. Exclusion of both would have produce a cleaner study.3Chan F.K. Graham D.Y. Review article prevention of non-steroidal anti-inflammatory drug gastrointestinal complications-review and recommendations based on risk assessment.Aliment Pharmacol Ther. 2004; 19: 1051-1061Crossref PubMed Scopus (122) Google Scholar There are several possibilities that might explain the unexpectedly high prevalence of endoscopic ulcers in the placebo group. One is the unreported use of ulcerogenic medications among the patients with osteoarthritis during the study period. Another possibility is that the endoscopic determination of the presence or absence of an endoscopic ulcer has a definite false-positive rate. Overall the results could be explained by a false-positive rate of approximately 4% (ie, accuracy of endoscopy = ∼96%). Considering that there the protocol included 2 different endoscopic evaluations while on therapy, a 2% false-positive rate would have resulted in a 4% cumulative incidence of ulcers. It is known that the endoscopic grading system used for such studies overestimates the prevalence of ulcers,4Sung J.Y. Lau J.Y. Chan F.K. Graham D.Y. How often are endoscopic ulcers in NSAID trials diagnosed as actual ulcers by experienced endoscopists?.Gastroenterology. 2001; 120: A597Abstract Full Text PDF Google Scholar the question about the accuracy of a large number of endoscopists in identifying ulcers or erosions has not been addressed as a potential problem. The “reality” test is what proportion of normal not H pylori–infected individuals not taking ulcerogenic medications would be expected to develop an ulcer within any 1 year of follow-up. Based on current thinking, the number would be approximately zero and studies that show a higher rate should examine why. The universal availability of videoendoscopy makes it possible to obtain documentation (video preferred) of all lesions making confirmation by an expert panel is not only possible but critical for assessing outcomes expected to be infrequent. We propose that the conclusion that low-dose, enteric-coated aspirin causes no more damage than placebo is untenable when it is based on the presence of a high proportion of ulcers in the placebo group. Failure to pass the reality test in relation to the proportion of ulcers in the placebo-only group calls the conclusions of the entire study into question. These results also call into question the assumptions underlying the study, such as endoscopists are 100% accurate in finding and correctly categorizing suspected lesions in the stomach and duodenum. Ulcer formation with low-dose enteric-coated aspirin and the effect of COX-2 selective inhibition: A double-blind trialGastroenterologyVol. 127Issue 2Preview We assessed the risk of ulcers with low-dose aspirin and the interaction of low-dose aspirin with a COX-2 selective inhibitor in a double-blind trial that compared placebo, low-dose aspirin, rofecoxib + low-dose aspirin, and ibuprofen. Osteoarthritis patients ≥50 years of age without ulcers or erosive esophagitis at baseline endoscopy were assigned randomly to placebo, enteric-coated aspirin 81 mg/day, rofecoxib 25 mg combined with aspirin 81 mg/day, or ibuprofen 800 mg 3 times a day. Repeat endoscopies were performed at 6 and 12 weeks. Full-Text PDF ReplyGastroenterologyVol. 128Issue 3PreviewDrs. Graham and Chan comment that our double-blind, randomized trial of nearly 800 patients taking placebo or 81 mg of enteric-coated aspirin fails their reality test because of a 4% rate of ulcers in the placebo group. They suggest that the ulcer rate in the placebo group was higher than they expected-possibly due to unreported NSAID use, “false-positive” ulcers in the placebo group, or toxicity of enteric coating. However, this rate of ulcers in the placebo group is exactly in line with many recently published double-blind endoscopic trials, including one authored by Dr. Full-Text PDF