COMPARATIVE PROGNOSTIC VALUE OF THE PNEUMONIA SEVERITY INDEX, CURB-65 AND HAEMATOLOGICAL RATIOS (NLR, MLR, PLR) FOR PREDICTING IN-HOSPITAL MORTALITY IN COMMUNITY-ACQUIRED PNEUMONIA

Abstract

Background: Rapid, reliable risk stratification is central to the management of community-acquired pneumonia (CAP). The pneumonia severity index (PSI) and the CURB-65 score are well validated, but they are time-consuming or require laboratory data that may not be immediately available. Inexpensive haematological ratios derived from a routine complete blood count, the neutrophil-to-lymphocyte ratio (NLR), monocyte-to-lymphocyte ratio (MLR) and platelet-to-lymphocyte ratio (PLR), have been proposed as surrogate prognostic markers. We compared their performance against PSI and CURB-65 for predicting in-hospital mortality.

Methods: We conducted a retrospective observational cohort study of adults admitted with CAP to POF Hospital, Wah Cantt, between 1 April and 15 July 2024. Demographic, clinical and admission complete-blood-count data were retrieved from electronic records. PSI, CURB-65 and the three ratios were calculated for each patient. Survivors and non-survivors were compared with the Mann–Whitney U test; discrimination for in-hospital mortality was assessed with the area under the receiver-operating-characteristic curve (AUC), AUCs were compared with the DeLong test, and optimal thresholds were derived using the Youden index.

Results: Of 150 admission records, 143 unique patients were analysed (median age 68 years; 62% ≥ 65 years), of whom 121 had a complete admission blood count permitting ratio calculation. In-hospital mortality was 7.0% (10/143). PSI (AUC 0.88, 95% CI 0.75–0.96) and CURB-65 (AUC 0.84, 95% CI 0.72–0.94) discriminated mortality well, and both scores together with the National Early Warning Score were markedly higher in non-survivors (all p ≤ 0.001). In contrast, NLR (AUC 0.44), MLR (AUC 0.52) and PLR (AUC 0.39) showed no discriminatory ability; both clinical scores out-performed every ratio on DeLong testing (all p ≤ 0.005), and the ratios correlated only weakly and non-significantly with PSI and CURB-65 (all |ρ| < 0.08).

Conclusions: In this single-centre cohort, PSI and CURB-65 substantially outperformed the haematological ratios, which carried no independent prognostic value for in-hospital mortality. NLR, MLR and PLR should not replace established severity scores, although their low cost may warrant evaluation as adjuncts in adequately powered, multicentre studies.