Abstract
BACKGROUND: The prevalence of frailty, an important risk factor for short- and long-term outcomes in hospitalized adults, differs by sex. Studies in critically ill adults have also found differences in mortality and organ support rates in males and females. The objective of this study was to determine if these observed differences in mortality and organ support rates can be explained by sex and frailty alone, or if the interaction between sex and frailty is an important risk factor. METHODS: This is a retrospective multi-centre population-based cohort study of all adult patients (≥ 18 years) admitted to the seventeen intensive care units (ICUs) across Alberta, Canada, between 2016 and 2017. On admission, physicians assigned a Clinical Frailty Scale (CFS) score (1 = very fit, 9 = terminally ill) to all patients. Patients with missing CFS scores or who died within 24 h of ICU admission were excluded. Frailty was defined as CFS ≥ 5. Outcomes included all-cause hospital mortality, ICU mortality, and organ support rates. A propensity score for female sex was generated and 1:1 matching on sex was performed. Multivariable Cox regression or logistic regression, as appropriate, was performed to evaluate the association between sex, frailty, and the sex-frailty interaction term with outcomes. RESULTS: Of 15,238 patients included in the cohort, after propensity score matching 11,816 patients remained (mean [standard deviation] age 57.3 [16.9]). In the matched cohort, females had a higher prevalence of frailty than males (32% vs. 27%, respectively) and higher odds of frailty (odds ratio [95% confidence interval (CI)] 1.29 [1.20-1.40]). Though females were less likely to receive invasive mechanical ventilation (hazard ratio [95% CI] 0.78 [0.71-0.86]), the interaction between sex and frailty (i.e., males and females with and without frailty) was not associated with differences in organ support rates. Receipt of dialysis and vasoactive support, as well as hospital mortality and ICU mortality were associated with frailty but were not associated with female sex or the interaction between sex and frailty. CONCLUSIONS: Although frailty and sex were individually associated with mortality and differences in organ support in the ICU, there does not appear to be a significant interaction between sex and frailty with regards to these outcomes.