Background: Community-acquired pneumonia (CAP) is the leading cause of hospitalization among infectious diseases, and is mainly caused by Streptococcus pneumoniae. Modifications were tested to improve the accuracy of CRB-65 as a simple but useful bedside scoring system, and to compare it with 3 established severity scoring systems (PSI, CURB-65 and CRB-65) to predict 30-day mortality in bacteraemic pneumococcal CAP.
Methods: A retrospective analysis was performed on data from 375 adult patients with bacteraemic pneumococcal pneumonia. Mortality, sensitivity, specificity, positive predictive value, negative predictive value, and area under the receiver operating characteristic (ROC) curve were calculated for stratifications of the PSI, CURB-65 and CRB-65. The prognostic accuracy after addition of underlying disease and/or a peripheral oxygen saturation (SaO₂) < 90% was evaluated (DS CRB-65).
Results: The mean age of the patients was 61.5 y, and the 30-day mortality was 9%. Coexisting conditions defined according to the pneumonia severity index (PSI) rule (malignancy, liver, cerebrovascular, and renal disease and congestive heart failure, p = 0.006) and SaO₂ < 90% (p < 0.0001) were independently associated with mortality. By adding these variables, the area under the ROC curve of CRB-65 increased from 0.77 (95% confidence interval (CI) 0.66-0.84) to 0.83 (95% CI 0.73-0.89) (p = 0.01), similar to that of PSI (0.84) and CURB-65 (0.81).
Conclusions: Modification of CRB-65 with the addition of 1 point for the presence of any underlying disease according to the PSI rule, and with 1 point if SaO₂ was < 90%, increased its prognostic accuracy in bacteraemic pneumococcal pneumonia with retained independence of laboratory data. The modified CRB-65 may have potential use in the assessment of prognosis in patients with CAP.