Comparison of six frailty instruments in adults with heart failure: a prospective cohort pilot study

Aims To compare the frailty prevalence and predictive performance of six frailty instruments in adults with heart failure and determine the feasibility of study methods. Methods and results Prospective cohort pilot study. Adults aged 18 years or older with a confirmed diagnosis of heart failure in Sydney, New South Wales, Australia. The Frailty Phenotype; the Survey of Health, Ageing, and Retirement in Europe Frailty Instrument (SHARE-FI); St Vincent's Frailty instrument; St Vincent's Frailty instrument plus cognition and mood; The Fatigue, Resistance, Ambulation, Illnesses, and Loss of Weight (FRAIL) scale; and the Deficit Accumulation Index (DAI) were compared. Multiple logistic regression was used to develop six frailty instrument models to evaluate the association between each frailty instrument and composite all-cause rehospitalization and mortality at 12 months. One hundred and thirty-one patients were included with a mean age of 54 [+/- 14(SD)]. Frailty prevalence ranged from 33 to 81%. All instruments except one (the FRAIL scale) appeared to signal an increased odds of rehospitalization and/or mortality, yet these results were non-significant. The six frailty instrument models displayed sensitivity between 88-92% and C-statistic values of 0.71-0.73, suggesting satisfactory discrimination. Conclusion The prevalence of frailty varied across six frailty instruments yet was in the higher range despite a 'younger' heart failure cohort. Further research is required to confirm the psychometric properties of these instruments for routine clinical use in an adequately powered and more diverse heart failure cohort.