Systematic Review and Meta-Analysis of Intensive Care Unit Scoring Systems' Performance in Patients with Pre-Existing...

Several scoring systems are used to provide prognostic estimates of mortality among patients admitted to the intensive care unit (ICU) but do not account for baseline kidney function. This systematic review and meta-analysis evaluates the performance of commonly used ICU scoring systems in patients with nondialysis chronic kidney disease (CKD), end-stage kidney disease (ESKD) on maintenance dialysis, and kidney transplant recipients (KTRs).

We searched Ovid Medline, Embase, and Scopus from inception through October 10, 2024, to identify observational studies reporting the prognostic performance of ICU scoring systems for mortality in patients with pre-existing kidney disease, analyzed independently or compared with a general ICU population.

Twelve studies involving 578,027 patients met inclusion criteria, evaluating 5 ICU scoring systems (sequential organ failure assessment [SOFA], acute physiologic and chronic health evaluation [APACHE] II/III, and simplified acute physiologic score [SAPS] II/III). Most studies involved patients with ESKD on maintenance dialysis and KTRs; data on patients with nondialysis CKD were very limited. For ESKD on maintenance dialysis, the area under the receiver operating characteristic curve (AUROC) showed excellent discrimination for all ICU scoring systems. However, discrimination was consistently lower in ESKD when compared with general ICU population, and calibration metrics showed consistent overestimation of mortality risk in ESKD. Among KTRs, scoring systems displayed consistently poor discrimination, for both SAPS III and SOFA, with variable calibration metrics.

Common ICU scoring systems frequently overestimate mortality risk among patients with ESKD on dialysis and kidney transplantation. These findings highlight the need to develop or recalibrate scoring systems for patients with pre-existing kidney disease.