RESUMEN
Aim: to evaluate the influence of pulmonary involvement at admission in predicting mortality among patients with severe leptospirosis. Methods: reprospective cohort study from medical record registry in Dr. Sardjito Hospital, Yogyakarta from 2003 to 2007. Pulmonary involvement was defined by the presence of pulmonary infiltrate, consolidation or pleural effusion in thorax radiography. Pulmonary edema was excluded. Admission data were collected. Follow-up records were noted until patients were discharded or died. The correlation between predictors (some patient characteristics on admission) and outcome (mortality) were evaluated using univariate analysis, and then proceeded to multivariable logistic regression analysis. P < 0.05 is considered statistically significant. Results: sixty patients with severe leptospirosis as a main diagnosis were evaluated. Fifty-five subjects were eligible for analysis, male patients 37 (67.3%) and mean age 42 +/- 15 years old. Pulmonary involvement was presence in 7 patients (12.7%). In univariate analysis only the presence of meningismus and pulmonary involvement were associated with mortality (p=0.001 and 0.006 respectively). In multivariable logistic regression, pulmonary involvement was independently a strong predictor of mortality (OR 9.9 95% CI (1.17 - 84.03), p=0.035). Conclusion: pulmonary involvement at admission is a strong predictor of mortality among patients with severe leptospirosis.