Background: The Clinical Frailty Scale (CFS) is widely utilized for risk stratification in emergency departments (EDs); however, its predictive value across various age groups remains unclear.

Methods: In this retrospective multicenter study, we analyzed 6,310 patients in the ED aged ≥ 65 years, categorized into young-old (65-74 years, n = 2,750), middle-old (75-84 years, n = 2,400), and old-old (≥ 85 years, n = 1,160) groups. According to CFS, patients were categorized as robust (scores 1-3), pre-frail (score 4), or frail (scores 5-9). Multivariable logistic regression analyses were performed to evaluate the independent association between frailty categories and clinical outcomes (ICU admission and in-hospital mortality), adjusting for age, sex, and illness severity. Predictive performance was assessed using the area under the receiver operating characteristic curve (AUROC).

Results: The predictive value of CFS varied significantly across age groups. In the young-old group, frail status was independently associated with increased ICU admission (adjusted odds ratio [aOR] 1.49, 95% confidence interval [CI] 1.076-2.062) and in-hospital mortality (aOR 3.232, 95% CI 1.738-6.009). The middle-old group demonstrated the strongest relationship with mortality for frail patients (aOR 5.361, 95% CI 2.872-10.007), but no significant association with ICU admission after adjustment. In the old-old group, neither pre-frail nor frail status significantly predicted outcomes. AUROC analysis showed the highest discriminative capability for ICU admission in the young-old group (0.616, 95% CI 0.597-0.634) and for mortality in the middle-old group (0.730, 95% CI 0.712-0.748), with reduced predictive value observed in the old-old group.

Conclusions: The prognostic value of CFS varies significantly by age group, demonstrating the strongest performance in young-old patients and diminishing predictive value in the old-old group. These findings suggest the need for age-specific frailty assessment strategies in emergency care, with additional clinical indicators potentially necessary for risk stratification in the oldest patients. CLINICAL TRIAL NUMBER: Not applicable.