Variations in the risk of gastrointestinal hemorrhage with non-steroidal anti-inflammatory drugs and localization of lesions.

The aim of this prospective study was to determine the effect of ulcerogenic drugs in patients with bleeding peptic ulcers and erosions with respect to their age and sex, ulcer history and additional risk factors such as family medical history, alcohol use, smoking, coffee consumption and stress. Of 367 patients with bleeding gastroduodenal lesions admitted during a period of 15 months, 88 (24%) had previously received ulcerogenic drugs. The most frequently taken drugs were aspirin (44.3%), piroxicam (12.3%) and ibuprofen (7.4%). Bleeding lesions were 1.4 times more frequently found in male users than in female users, and 2.1 times more often in male unusers. Males were more commonly receiving drugs than females (59.8%:40.2%), particularly those aged 34 to 54 years. Forty (45.5%) users had previously suffered from ulcer disease, 48 (54.5%) had negative history. There was no additional risk factor in 48%, whereas 58% of the users had one or more risk factors. It may be concluded with great certainty that NSAIDs caused hemorrhage in 13% of all admissions. Among users, a total of 119 different gastric and duodenal lesions were found. Gastric lesions were more common (54%) than duodenal lesions (46%) in males, while in females an inverse ratio was observed (41% of gastric and 59% of duodenal lesions). Among nonusers, gastric lesions were more frequent in females (M:F, 42%:48%), and duodenal lesions in males (M:F, 58%:52%). The number of lesions increased with age in both users and nonusers. Forty-three percent of all drug users had ulcers in the prepyloric region, 23% on the lesser curvature, and 14% at the posterior wall of the gastric corpus. Gastroduodenal erosions were seen in 11% of the males and 1% of the females. In nonusers, ulcers were found on the posterior wall of the corpus (27%), on the lesser curvature (25%) and in the prepyloric region (24%). Bleeding gastroduodenal erosions were found in 4% of the patients. Distribution of bleeding duodenal ulcers was similar in drug users and in nonusers, i.e. anterior wall (45%:44%), posterior wall (28%:27%), lower wall (16%:16%) and upper wall (3%:4%) of the duodenum. NSAIDs had no influence on the localization of duodenal ulcers. In this study, there was no death from ulcer disease with NSAID use. It may be concluded that NSAIDs are a common cause of damage to gastroduodenal mucosa. The risk of drug therapy should be balanced against the risk of the disease.