NEW YORK (Reuters Health) Apr 29 – The results of a meta-analysis of cohort and case control studies of severe upper gastrointestinal (GI) complications suggest that the use of nonsteroidal anti-inflammatory drugs (NSAIDs) increases the risk of perforation, ulcers, and bleeds by a factor of approximately three, US investigators report.
Previously reported meta-analyses were marred by restrictions on type of study and language of publication, and by omission of nonpublished data, Dr. Joshua J. Ofman, of Cedars-Sinai Health System and Zynx Health Inc., in Beverly Hills, California, and associates note. Their analysis, which appears in The Journal of Rheumatology for April, attempts to overcome these restrictions.
The researchers evaluated 4881 published titles and identified randomized controlled trials involving three US Food and Drug Administration studies and 13 published reports regarding outcomes of perforation, ulcers and bleeds. They also evaluated 23 case control studies and 9 cohort studies.
Randomized control trials involved 4431 patients and yielded a pooled odds ratio of 5.36 compared with placebo. Case control studies, which contributed more than 25,000 patients, yielded a pooled odds ratio of 3.0.
When Dr. Ofman’s group stratified their analysis according to whether studies were published before or after 1990, they found no significant difference between the results of randomized control trials and case control studies. However, the risk ratio for cohort studies that appeared before 1990 was 2.0, while the risk ratio for those published in 1990 or later was 3.4.
Dr. Ofman and his associates did not address differences between selective and nonselective cyclooxygenase (COX) inhibitors. In an editorial, Dr. Chris J. Hawkey, of the University Hospital Nottingham in the UK, recommends that similar analysis be conducted for both classes of drugs and for non-GI toxicities.
“The challenge of such studies will be to assess the effect of drugs on overall health, morbidity, and mortality,” Dr. Hawkey concludes.
J Rheumatol 2002;29:804-812.