The CURB-65 scoring system performs well at identifying patients with pneumonia who have a low risk of death. Whether it predicts mortality in community-acquired pneumonia (CAP) better than the 2007 Infectious Disease Society of America (IDSA)/American Thoracic Society (ATS) minor criteria in low-mortality-rate settings is not clear. The purpose of this study was to determine the hypothesis.A total of 1,230 adult inpatients admitted to our hospital from 2005 to 2009 for CAP were reviewed retrospectively.The hospital mortality was 1.3 %. Percentage mortality increased significantly with CURB-65 score and the increasing number of IDSA/ATS minor criteria present. The number of CURB-65 criteria or IDSA/ATS minor criteria present had significant increased odds ratios for mortality of 7.547 and 2.711, respectively. The sensitivities of a CURB-65 score of ≥ 3 and the presence of ≥ 3 minor criteria in predicting mortality was 25 % and 37.5 %, which increased to 75 % and 62.5 %, while the cut-off values reduced to ≥ 2 criteria, respectively. The area under the receiver operating characteristic curve for CURB-65 was greater than the corresponding area for IDSA/ATS minor criteria in predicting hospital mortality (0.915 vs. 0.805, p = 0.0091).CURB-65 score predicted hospital mortality better than IDSA/ATS minor criteria, and a CURB-65 score of ≥ 2 or the presence of ≥ 2 minor criteria might be more valuable cut-off values for "severe" CAP in a low-mortality-rate setting.
