PURPOSE: The CURB-65 score is one of the most used severity scores for mortality risk in patients with pneumonia. It has been only validated in patients with community-acquired pneumonia (CAP). However, health-care associated pneumonia (HCAP) is associated with high mortality and more resistant causative pathogens. There has not been a study that has evaluated the validity of CURB-65 score in patients with HCAP.
METHODS: We performed a retrospective study of patients admitted to our ICU with a diagnosis of pneumonia between 07/2003 and 12/2009, entered in the Project Impact database. Inclusion criteria were age > 17 years and admission through the emergency department. We recorded demographic data, the presence of components of the CURB-65 score, criteria for HCAP and mortality. We analyzed the data using Chi-square, extended Mantel Haenszel, ROC curve and Mann-Whitney U test.
RESULTS: Of 406 patients identified, 357 had complete data. 156 met criteria for HCAP. HCAP patients were older (median 66 v. 59 years), had a higher CURB-65 score (1.9 v. 2.4) (both p < 0.001), a trend toward higher APACHE II score (p = 0.074), a shorter ICU length of stay (LOS) (1.5 v. 1.8; p = 0.03), but a similar mortality (25.6% v. 26.9%; p = 0.81). As the CURB-65 score increased (from 0 to 4), mortality risk increased: 18 to 43 % (p = 0.039 for difference, 0.005 for trend). A high score (>2) was significantly associated with mortality (OR 2.2 [1.1-4.8]; p = 0.043). A high CURB-65 score was not associated with increased ICU LOS (p = 0.859). The ROC curve for CURB-65 score was greater in patients with HCAP (0.648) than with CAP (0.574).
CONCLUSION: Patients with HCAP were older and had higher CURB scores but mortality rates similar to patients with CAP.
CLINICAL IMPLICATIONS: A high CURB-65 score is associated with higher mortality in patients with HCAP, and has a better predictive value than in CAP.
DISCLOSURE: Ankur Kalra, No Financial Disclosure Information; No Product/Research Disclosure Information