Comparison of CURB-65 and CRB-65 as predictors of death in community-acquired pneumonia in adults admitted to an ICU

Introduction

Community-acquired pneumonia is one of the most common causes of sepsis and ICU admissions. Patients with CAP who demand critical care had mortality rates of 25 to 50%. Thereby, the assessment of the severity is essential to guide the treatment. There are several severity scores for CAP and some of the most acknowledged are CURB-65 and CRB-65. The objective of this study was to evaluate the accuracy of CURB-65 and CRB-65 as predictors of death in patients with community-acquired pneumonia.

Methods

A prospective study during 6 months was conducted with patients diagnosed with CAP admitted to the ICU of the Hospital Santa Luzia, Brasília, DF, Brazil. Patients were stratified according to CURB-65 (0 to 5) and CRB-65 (0 to 4) and their risk categorized as: low (CURB-65: 0 to 1 and CRB-65: 0), moderate (CURB-65: 2 and CRB-65: 1 to 2) and high (CURB-65: 3 to 5 and CRB-65: 3 to 4). The sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), likelihood ratio positive (LR+), and likelihood ratio negative (LR-) were calculated. Validity and reliability were assessed with the Spearman correlation coefficient. Patients with chronic kidney failure and those submitted to mechanical ventilation at the time of admission were excluded.

Results

A total of 62 patients were included. Twenty-seven with low risk, 24 with moderate risk and 11 with high risk according to CURB65 and their mortality rates were 7.4%, 8.3% and 54.5%, respectively. According to CRB-65, 11 were low risk, 44 moderate risk and seven had high risk. The mortality on CRB-65 stratification was 0%, 15.9% and 42.9% for low, moderate and high risks, respectively. When we gathered moderate and high risks, CRB-65 was more sensitive (1.00 vs. 0.80) and had better LR- (0.00 vs. 0.41), and NPV (1.00 vs. 0.92). CURB-65 had better specificity (0.48 vs. 0.21), LR+ (1.54 vs. 1.26), and PPV (0.23 vs. 0.20). The receiver operating characteristic curves of CURB-65 and CRB65 had areas of 0.758 and 0.686, respectively. The Spearman correlation coefficient was 0.612 (P = 0.00). See Figure 1.

Figure 1

abstract

Conclusion

CURB-65 and CRB-65 had a high correlation. CRB-65 was more sensitive as a predictor of death as well as a guidance for hospitalization. Moreover, CRB-65 is a more practical score since it does not use laboratorial variables.