Prognostic value of blood pressure drops during the first 24h after hospital admission for risk stratification of community-acquired pneumonia: a retrospective cohort study.

Abstract

Current risk stratification in community-acquired pneumonia (CAP) does not incorporate the dynamic nature of CAP evolution. Study aim was to evaluate the predictive value of early blood pressure (BP) drop and its consideration within the CRB-65 score. We performed a retrospective cohort study including consecutive adult hospitalized CAP patients 2013-2014 without documented treatment limitations or direct ICU admission. The CRB-65 score was calculated initially and re-calculated including any BP below the threshold (BP drop) within the first 24h (CRB-65[BP24]). The primary endpoint was need for mechanical ventilation or vasopressors (MVVS) occurring after 24h. Prognostic values were evaluated by uni- and multivariate and ROC curve analyses. 28/294 patients (9.5%) met the primary endpoint. Only 3 (11%) of them showed an initial BP of < 90mmHg systolic or ≤ 60mmHg diastolic, but 21 (75%) developed a BP drop within the first 24h. 24/178 (13%) with and only 4/116 (3%) without any low BP during the first 24h needed MVVS (p = 0.004). After multivariate analysis, the predictive value of BP drop was independent of other score parameters and biomarkers (all p < 0.01). In ROC analysis, the new CRB-65(BP24) showed a better prediction than the CRB-65 score (AUC 0.69 vs. 0.62, p = 0.04). 7/13 patients (54%) with MVVS despite an admission CRB-65 of 0 or 1 showed a BP drop. In the evaluated cohort, BP drop within the first 24h was significantly associated with more need for MVVS in CAP, and its consideration improved the prognostic value of the CRB-65 score.