Lab-Based Frailty Score Predicts Death Risk in ICU Patients With ARDS

Acute respiratory distress syndrome (ARDS) carries a devastating in-hospital mortality rate of 27–45%, yet traditional frailty assessments are often impractical at ICU admission. Researchers sought to determine whether a fully objective, laboratory-derived frailty index (FI-Lab) — calculated at the moment of ICU admission — could predict who would survive hospitalization.

Using the MIMIC-IV v3.1 database, the team retrospectively identified 859 adult ARDS patients (ICD-9/10 coded) admitted between 2008 and 2022 at Beth Israel Deaconess Medical Center. FI-Lab was constructed from a prespecified 33-item panel: 30 laboratory variables spanning hematology, electrolytes, renal and liver function, coagulation, inflammatory markers, arterial blood gases, and urinalysis, plus 3 vital signs averaged over the first ICU day. Each item was binary-coded (normal = 0; abnormal = 1) and the index expressed as the proportion of abnormal items, yielding a continuous 0–1 score. Patients missing more than 12 items were excluded.

Of the 859 patients, 332 (38.7%) died in hospital. Non-survivors were older (median 65 vs 58 years) and had significantly higher FI-Lab scores (median 0.59 vs 0.54, P < .001). All four acute severity scores — SOFA, APS III, SAPS II, and OASIS — were also markedly worse in non-survivors. To minimize confounding, 1:1 propensity score matching produced 262 well-balanced pairs. In this matched cohort and in multivariable logistic regression models adjusting for one severity score at a time, FI-Lab remained independently associated with in-hospital mortality: each 0.1-unit increase in FI-Lab conferred an odds ratio of 1.26 (95% CI 1.14–1.40; P = .003).

Kaplan–Meier survival analysis stratified by FI-Lab tertiles showed stepwise worsening of survival, with the highest tertile carrying a hazard ratio of 1.98 (95% CI 1.52–2.57; P < .001) compared to the lowest. Critically, FI-Lab added prognostic information beyond established severity indices, consistent with its conceptual role as a frailty-based vulnerability marker rather than a pure organ-dysfunction score.

The authors argue that FI-Lab is particularly valuable in ARDS because it is fully automated from existing electronic health record data, requires no patient cooperation, and can be computed instantly at admission. However, the study is retrospective, single-center, and based on a US academic hospital cohort, limiting generalizability. External validation in prospective multicenter studies is essential before FI-Lab can be incorporated into clinical decision-making.