Impact of frailty on the short-term outcomes of elderly intensive care unit patients

Background Frailty leads to multiple unfavourable outcomes in older adults. However, few studies have investigated correlations between frailty and its impacts on morbidity and mortality of elderly patients in intensive care units (ICUs) in Taiwan. Aims To investigate the impact of frailty on the risk of hospital and 30-day mortality and functional outcomes of elderly Taiwanese ICU patients. Study Design A prospective observational study was conducted. Patients aged 65 years or older were recruited from three medical ICUs. We defined 'frailty' according to the Clinical Frailty Scale (CFS) higher than 4 within 1 month prior to admission. The primary outcomes were hospital and 30-day mortality. The secondary outcome was CFS changes at ICU admission, hospital discharge, and 30-day follow-up. Logistic/Cox regression was used to analyse the data. Results We recruited a total of 106 patients, 57 (54%) of whom were classified as frail. The overall mortality rate was 21%. Hospital mortality and mortality within 30 days after discharge were higher in the frail patients without a significant statistical difference (hospital mortality: 17.5% vs. 12.2%, p = .626; 30-day mortality: 26.3% vs. 14.3%, p = .200). The risk of 30-day mortality for frail patients was up to 2.84 times greater than that of non-frail patients in the Cox model (hazard ratio = 2.84, 95% confidence interval [0.96, 8.38]). Both non-frail and frail patients had a worse CFS score on admission, but the CFS score of surviving non-frail patients improved significantly over the medium term. Conclusion Frailty tended to increase short-term ICU mortality risk and worsen functional outcomes in the elderly Taiwanese population. This information might guide critical medical decisions. Relevance to Clinical Practice Frailty could be included in the prognostic evaluation of either mortality risk or functional outcome. Prompt palliative care might be one last piece of holistic elder care.