[The usefulness of the Baylor score in patients with upper gastrointestinal bleeding in a hospital in Lima, Peru].

Objective: To demonstrate the usefulness of the Baylor score in patients with upper gastrointestinal bleeding (UGB) due to peptic ulcer, in terms of mortality and recurrent bleeding at 30 days follow-up.

Material And Methods: This study has collected information prospectively into a registration form from medical histories. Patients included were those who came to the "Hospital Nacional Cayetano Heredia", Lima, Peru, in the period between June 2009 and May 2011, with UGB due to peptic ulcer demonstrated by endoscopy. The data was analyzed with the ROC curve (Receiver Operating Characteristic).

Results: We included 181 from 380 patients with UGB (47.63%), 74% were male, mean age 56 years old; 20 patients died during follow-up: 8 due to gastrointestinal bleeding and 12 from other causes, 10 patients had recurrence of bleeding during a 30 days follow-up. When performing the analysis of the ROC curve with the Baylor score and mortality, it was reported IC 95% value of 0.89 (0.83-0.95), with recurrence of bleeding an IC 95% value of 0.81 (0.68-0.93) and need for transfusion of over two globular packages an IC 95% value of 0.79 (0.70-0.86).

Conclusions: Baylor score is a good predictor of mortality and recurrence of bleeding during the first 30 days after a first episode of an upper gastrointestinal bleeding and an acceptable predictor of the need to transfuse more than two globular packages during hospitalization.