The outcome and role of drugs in patients with unexplained gastrointestinal bleeding.

OBJECTIVE

Studies on unexplained gastrointestinal bleeding (GIB) are lacking. We aimed to study the clinical outcomes of patients with unexplained GIB and to determine the incidence of obscure GIB.

MATERIAL AND METHODS

A population-based study on all patients undergoing endoscopy at the National University Hospital of Iceland in 2010. Indications, results of endoscopies and drug history were prospectively registered with a follow-up of 3 years. A national pharmaceutical database containing prescription data was utilized. Patients were categorized into unexplained overt and occult GIB and obscure GIB. Patients undergoing endoscopy and without GIB acted as controls.

RESULTS

Of 2471 patients undergoing endoscopy, 11% had unexplained GIB. Of those, 46% had unexplained overt GIB, 44% had unexplained occult GIB and 11% had obscure GIB. Multivariate analysis showed that patients with unexplained GIB and unexplained overt GIB had greater odds of NSAID use than controls, OR 1.8 (CI 1.03-3.03) and OR 2.0 (CI 1.01-3.77), respectively. Warfarin was strongly associated with all bleeder groups, OR 4-4.8. The incidence of obscure GIB was 10/100,000 inhabitants annually. Two (0.8%) patients were diagnosed with colon cancer 16 and 30 months after the index colonoscopy. Of patients with unexplained overt, unexplained occult GIB and controls, 5%, 6% and 3.5% (NS) had another overt bleeding episode, during the study period.

CONCLUSIONS

NSAIDs and warfarin seem to play an important role in unexplained GIB. The incidence of obscure GIB is low and missed cancers are very rare. The probability of a repeat bleeding episode is similar to that of controls.