Contributors to frailty in critical illness: Multi-dimensional analysis of the Clinical Frailty Scale

Purpose: Frailty in critical illness is common and associated with poor outcomes, however little is known about contributing factors. We compared the Clinical Frailty Scale (CFS) with a multi-dimensional validated tool, the Edmonton Frail Scale (EFS), and investigated which health domains are affected by frailty in ICU. Materials and methods: This prospective cohort study enrolled patients aged >= 50 years admitted between February-June 2017, comparing the CFS and EFS using Spearman correlation and Kappa coefficients, assessing frailty status across health domains, and examining outcomes including mortality. Results: One hundred and sixty patients were enrolled, 33.8% were frail. Frail patients had greater in-hospital and 6-month mortality [25.9% vs. 8.5%; adjusted OR (95% CI) = 3.31 (1.17, 939), p = .024; and 40.4% vs. 17.3%; OR (95% CI) = 2.84 (1.18, 6.83), p = .020 respectively]. CFS and EFS scales were highly correlated [Spearman correlation coefficient = 0.85 (95% CI 0.81-0.88)], with high agreement [kappa coefficient = 0.78 (95% CI: 0.68-0.88)]. Frail patients had worse health status across the spectrum of frailty domains, in particular functional dependence, malnutrition, and prior hospital admissions. Conclusions: Frailty in the critically ill affects a range of health deficits, adequately measured via the CFS. (C) 2019 Elsevier Inc. All rights reserved.