Predictors of mortality in patients with Klebsiella pneumoniae infection: a systematic review and meta-analysis

Abstract Background Klebsiella pneumoniae (KP) infection is a growing global threat because of increasing antimicrobial resistance and high mortality. We systematically estimated mortality at different endpoints and synthesized adjusted factors associated with death to inform risk assessment and management. Methods We searched PubMed, Web of Science, Embase, and the Cochrane Library for studies published from January 1991 to June 2025 (CRD42023456620). Prespecified meta-analyses included predictors reported by ≥ 3 studies, with ≥ 1 reporting a significant adjusted association. Adjusted odds ratios (ORs) and hazard ratios (HRs) were synthesized separately. Fixed- or random effects models were selected according to heterogeneity. Subgroup analyses examined study-level characteristics. Results We included 121 studies involving 21,660 patients (6235 deaths; 28.8%, 95% CI 28.2–29.4). Pooled mortality was 21.7% (95% CI 16.4–27.5) at 14/15 days, 33.3% (95% CI 29.9–36.8) at 28/30 days, 34.6% (95% CI 24.1–45.8) at 90 days, and 29.3% (95% CI 23.0–36.0) in hospital. Among studies published after 2020, 28/30-day mortality was 35.1% (95% CI 30.4–40.0), without significant differences by publication period. Separate OR and HR analyses consistently identified carbapenem-resistant KP (CRKP), greater comorbidity and disease severity, malignancy, sepsis/septic shock, and inappropriate empirical therapy as adverse prognostic factors of 28/30-day mortality, whereas appropriate targeted and combination therapies were associated with lower mortality. Risk of bias and publication period partly explained heterogeneity in selected OR estimates. Conclusions Mortality from KP infection remained substantial across 14/15-, 28/30-, and 90-day follow-up, with 28/30-day mortality exceeding 35% among studies published after 2020, although differences between publication periods were not statistically significant. Effect-measure-specific analyses identified antimicrobial resistance, comorbidity, disease severity, and treatment appropriateness as clinically relevant prognostic factors, which may inform individualized management. Systematic review registration PROSPERO CRD42023456620