Introduction and Objectives The British Thoracic Society and National Institute for Health and Care Excellence recommend the CURB65 for severity assessment in community acquired pneumonia (CAP). The Third International Consensus Definitions for Sepsis and Septic Shock however, state that the qSOFA (quick Sequential Organ Failure Assessment) should be used to identify those with infection who are likely to have poor outcomes. This superseded the SIRS (Systemic Inflammatory Response Syndrome) criteria. There is therefore no clear international consensus regarding how severity of CAP should be assessed outside the critical care setting. We retrospectively evaluated the ability of the CURB65, SIRS and qSOFA scores to predict 30 day mortality in patients with CAP.

Methods Adults admitted to the Queen Elizabeth Hospital Birmingham, with CAP, between 10/2014 and 01/2016 were included. Radiology, admission clerkings and electronic patient records were reviewed to confirm pneumonia and calculate the scores. Cases were excluded if there was no radiological confirmation, or if they had hospital-acquired pneumonia.

Results 1545 patients were included in the final analysis (mean age 72, 30 day mortality 19.0%, 49.1% female). All scoring systems enabled stratification according to increasing risk of 30 day mortality. CURB65: 0%–3.5%, 1%–11.5%; 2%–18.5%; 3%–27.1%; 4%–42.7% and 5%–42.1% (p<0.001). SIRS: 0%–10.4%; 1%–13.5%; 2%–18.1%; 3%–22.5%; 4%–32.5% (p<0.001). qSOFA: 0%–11.9%; 1%–17.9%; 2%–30.1%; 3%–47.2% (p<0.001). Receiver operator characteristic curves calculated to determine the accuracy with which the scores were able to predict 30 day mortality, revealed areas under the curve of 0.69, 0.60 and 0.63 for CURB65, SIRS and qSOFA respectively. Using the established cut-offs of CURB65 ≥2, SIRS≥2 and qSOFA ≥2, sensitivities, specificities, negative and positive predictive values for prediction of 30 day mortality were calculated (Table 1). Of those that died within 30 days, qSOFA ≥2 correctly identified 40.5% as high risk, compared to 79.5% and 84.9% using the SIRS criteria and CURB65 respectively.

Conclusions All three scoring systems can stratify according to risk of 30 day mortality, though none of them are particularly accurate. qSOFA has poor sensitivity, and may underestimate severity and risk of 30 day mortality in CAP. New assessment tools to accurately identify CAP patients at increased risk of poor outcomes are urgently required.