Predictive value of gastrointestinal symptoms and patient risk factors for NSAID-associated gastrointestinal ulcers defined by endoscopy? Insights from a pooled analysis of two naproxen clinical trials

Abstract

Objective Nonsteroidal anti-inflammatory drugs (NSAIDs) are widely used to treat pain and rheumatic conditions. To facilitate patient management, we determined the predictive value of gastrointestinal (GI) symptoms and risk factors for the development of NSAID-associated GI injuries. Methods Post-hoc analysis of pooled data from naproxen treatment arms of two identical, randomized, double-blind, controlled phase 3 trials in arthritis patients at risk of GI adverse events. Endoscopic incidence of GI ulcers at baseline, and 1, 3, and 6 months was employed as a surrogate parameter for GI injury. For GI symptom analysis, Severity of Dyspepsia Assessment questionnaire was used. For GI risk factor analysis, the high risk factors: previous GI injury, concomitant selective serotonin reuptake inhibitors or corticosteroids, ulcer history, concomitant low-dose aspirin, and age >65 years were employed. Results Data of 426 naproxen patients were analyzed. Distribution of GI symptoms between patients with and without ulcer was similar; about one third of patients developing an ulcer reported no GI pain symptoms. GI symptoms experienced under naproxen treatment were thus not indicative of GI injury. The proportion of patients developing an ulcer increased with the number of risk factors present, however, about a quarter of patients without any of the analyzed risk factors still developed an ulcer. Conclusion GI symptoms and the number of risk factors are not reliable predictors of NSAID-induced GI injury to decide which patients need gastroprotection and will lead to a large group of patients with GI injuries. A preventive rather than reactive approach should be taken.