RESUMO
OBJECTIVE: The CURB-65 score is a simple well validated tool for the assessment of severity in community-acquired pneumonia (CAP). The weight of each criterion in very low-mortality-rate settings is unclear. The purpose of this study was to determine the weight in such setting. METHODS: This study retrospectively reviewed 1,230 adult patients admitted for CAP from 2005 to 2009. RESULTS: The 30-day mortality rose sharply from 0%, 1.0%, 8.2% and 16.7%, respectively, for patients with CURB-65 scores of 0, 1, 2 and 3 to 100.0% for patients with the scores of 4 (x(2) = 219.494, p<0.001). Confusion had the strongest association with mortality (odds ratio, 22.148). The presence of low blood pressure was not associated with mortality. Confusion, urea >7 mmol.L(-1) and age ≥ 65 yrs showed independent relationships with mortality (Odds ratio, 11.537, 5.988 and 10.462; respectively). Urea >7 mmol.L(-1) was most strongly associated with the sequential organ failure assessment (SOFA) scores [rank correlation coefficient (r(s)), 0.352]. Confusion had the closest relationship with hospital length of stay (r(s), 0.114). Age ≥ 65 yrs had the strongest association with costs (r(s), 0.223). Conclusion The individual CURB-65 criteria were of unequal weight for predicting the 30-day mortality, SOFA scores, hospital length of stay and costs in a very low-mortality-rate setting, and a low blood pressure was not associated with mortality.