Predictive value of the clinical frailty scale for intensive care unit mortality in elderly and very elderly patients: a prospective cohort study

The Clinical Frailty Scale (CFS) is a clinician-rated tool used to assess frailty. Previous studies have shown that it predicts mortality and morbidity in older adults aged ≥ 65 years. However, its predictive value in the very elderly population (≥ 80 years) remains unclear. This study aimed to evaluate the ability of CFS to predict intensive care unit (ICU) mortality among elderly patients and to determine whether this relationship differs in the very elderly subgroup. This prospective cohort study was conducted after ethical approval. All patients aged ≥ 65 years admitted to the ICU and staying > 48 h during the study period were included when frailty assessment was feasible. CFS scores at ICU admission were recorded, and patients were categorized as survivors or non-survivors. Subgroup analyses were performed for elderly patients aged 65–79 years (Group E) and very elderly patients aged ≥ 80 years (Group VE). Multiple logistic regression analysis was applied to identify independent predictors of ICU mortality, and receiver operating characteristic (ROC) analysis was used to assess model performance. Relative risk (RR) for mortality according to age subgroups was calculated for CFS categories. A total of 173 patients were included (Group E: 96 [55.5%]). The mean CFS score was 6.2 ± 1.8 (median: Group E = 6, Group VE = 7; p = 0.066), and the mean Charlson Comorbidity Index was 8.3 ± 2.5 (p = 0.406 between groups). The mean ICU stay was 21 ± 24 days. Overall, 102 patients (59%) were discharged, while 71 (41%) died. CFS, renal replacement therapy (RRT), APACHE II, lymphocyte count, procalcitonin, and albumin levels were independent predictors of ICU mortality. CFS and RRT had the greatest contributions to the model (χ² = 16.32 and 20.36, respectively). The model demonstrated 77.5% sensitivity, 85.3% specificity, and an area under the curve (AUC) of 0.90. When considered alone, CFS yielded an AUC of 0.74 with an optimal cutoff of ≥ 7 (sensitivity 77.4%, specificity 56.8%). ICU mortality risk was 2.5-fold higher for all patients with CFS ≥ 7, 2.3-fold in Group E, and 2.8-fold in Group VE (p < 0.05). CFS independently predicts ICU mortality among elderly patients. A CFS ≥ 7 and the need for RRT identify a high-risk subgroup, emphasizing the importance of routine frailty assessment and early individualized management in this population.