Predicting the risk for gastrointestinal toxicity in patients taking NSAIDs: the Gastrointestinal Toxicity Survey

Nonsteroidal anti-inflammatory drugs (NSAIDs) are among the most widely used drugs in the world, but their use is often associated with adverse gastrointestinal (GI) events that may be life threatening in some patients. Development of simple questionnaires for predicting GI events in individuals taking NSAIDs may help to prevent use of these drugs in high-risk patients. The present study was undertaken to test a new questionnaire designed to identify patients at high risk for NSAID-associated GI events--the Gastrointestinal Toxicity Survey (NSAID Induced) (GITS [NI]). In this study, results for GITS (NI) were compared with those for an established questionnaire, the Stanford Calculator of Risk for Events (SCORE), in a small cohort of 400 patients. Feasible generalized least squares (FGLS) and multinomial logistic (MNL) regression were used to perform the comparison. The overall correlation between results for GITS (NI) responses and the total score for the SCORE questionnaire was 0.962 (P < .0001). The agreement between the 2 instruments with respect to their ability to predict the same risk for NSAID-induced GI events was similar for both FGLS and MNL. For 4 levels of risk, the agreement was approximately 80% between the 2 instruments. For 3 levels of risk, the agreement was approximately 90%. This study showed that results obtained with GITS (NI) are highly correlated with those from SCORE and that GITS (NI) may provide physicians with information that will help them avoid administering NSAIDs to patients who are at high risk for adverse GI reactions.