Spironolactone and gastrointestinal bleeding: a population based study.

BACKGROUND

Several medical reports showed an increased risk of stomach and duodenum bleeding associated with the use of spironolactone. The objective of this work is to examine the association between upper gastrointestinal (UGI) bleeding and exposure to spironolactone, using a population-based cohort design.

METHODS

Using the pharmaceutical prescriptions from 2000 to 2006, a cohort was identified of residents in Milan (Italy) with chronic exposure to spironolactone or other diuretic drugs. The main outcome was defined as the hospital admission for UGI bleeding or ulcers. To control for potential bias related to spironolactone use, the propensity score was estimated. Then, each patient treated with spironolactone was randomly matched with one treated with other diuretics and having the same propensity score using the caliper matching method. Proportional hazard models were fitted by computing hazard ratios (HR) and the corresponding 95% confidence intervals (95%CI).

RESULTS

A total of 53 550 unexposed and 10 564 exposed to spironolactone were identified. Overall in the study period, 174 patients (3.2%) developed UGI bleeding in the unexposed and 51 (4.8%) in the exposed group. HR for UGI bleeding for the spironolactone exposed group was 1.94 (95%CI 1.42-2.65). The sensitivity analysis based on the matched design using the propensity score showed a statistically significant twofold increase of gastrointestinal bleeding only among subjects exposed to high spironolactone dose (HR 2.50; 95%CI 1.08-5.79).

CONCLUSIONS

Results from a large population based study confirm that spironolactone increases the risk of UGI bleeding.