Comparative short-term mortality risk of different echocardiographic phenotypes in sepsis

Abstract

Cardiac dysfunction in sepsis manifests as diverse echocardiographic phenotypes; however, their relative effect on short-term mortality remains unclear. To compare the associations between different echocardiographic phenotypes and short-term mortality in sepsis using a network meta-analysis. PubMed, Cochrane Library, Embase, and Web of Science were searched from inception through February 2024 for observational cohort studies evaluating echocardiographic cardiac dysfunction phenotypes and reporting short-term mortality in sepsis, defined as all-cause mortality at in-hospital, 28-day, or 30-day time points. A frequentist network meta-analysis was performed using odds ratios (ORs) with 95% confidence intervals (CIs). The protocol was registered in PROSPERO (CRD42024583205). Forty-seven studies comprising 11,393 patients were included. Compared with no ventricular dysfunction (noVD), all phenotypes were associated with significantly increased short-term mortality: right ventricular dysfunction (RVD) demonstrated the highest risk (OR: 2.78; 95% CI: 1.85, 4.19), followed by biventricular dysfunction (OR: 2.37; 95% CI: 1.08, 5.18), left ventricular systolic dysfunction (OR: 2.33; 95% CI: 1.80, 3.01), and left ventricular diastolic dysfunction (OR: 2.08; 95% CI: 1.37, 3.13). Pairwise comparisons revealed no significant differences, confirming RVD as the highest-risk phenotype. Sensitivity analysis showed a 29.6% increase in the left ventricular systolic dysfunction effect estimate with global longitudinal strain versus left ventricular ejection fraction. Ventricular dysfunction is closely associated with adverse prognosis, and among all sepsis-induced cardiomyopathy phenotypes compared with noVD, RVD carries the highest mortality risk, highlighting the critical importance of right ventricular function assessment in sepsis prognostication.