Which patients taking non-aspirin non-steroidal anti-inflammatory drugs bleed? A case-control study

To identify risk factors for gastrointestinal bleeding (GIB) among users of non-aspirin, non-steroidal anti-inflammatory drugs (NANSAIDs).

Case-control study.

A total of 120 patients aged over 60 years and using NANSAIDs were hospitalized between January 1988 and September 1992 for GIB related to erosions or ulceration of the gastroduodenal mucosa. A group of 100 general practitioners selected two controls matched for age and sex, receiving NANSAIDs and without GIB, for each patient.

The same questionnaire was used to interview patients and controls about their medical history, use of NANSAIDs and other drugs, alcohol and tobacco use, recent stress and nutritional status.

The adjusted odds ratios (OR) for the risk factors related to the pattern of NANSAID use were 3.39 [95% confidence interval (CI) 1.77-6.47] when the intake of NANSAIDs was followed by decubitus, 3.00 (95% CI 1.54-5.85) when NANSAIDs were taken before a meal, 6.05 (95% CI 2.10 17.43) with a high dose of NANSAIDs, 5.87 (95% CI 2.00-17.25) with recent NANSAID use, 3.35 (95% CI 1.47-7.64) with NANSAIDs associated with aspirin use, 3.46 (95% CI 1.15-10.36) with more than one NANSAID, and 10.70 (95% CI 1.06-108.07) when NANSAIDs were associated with corticosteroids. The patient-related risk factors and their OR were 9.94 (95% CI 3.29-24.28) for irregular food intake, 3.94 (95% CI 1.45-10.69) for previous peptic ulcer, 3.71 (95% CI 1.26-10.89) for recent weight loss, 4.44 (95% CI 1.48-13.30) for heavy alcohol abuse, 2.92 (95% CI 1.36-6.26) for recent stress and 5.26 (95% CI 1.19-23.33) for a past history of GIB.

This study identified a group at 'high risk' for GIB which would benefit from the development of a prophylactic therapy.