Comparison of different frailty instruments for prediction of functional decline in older hypertensive outpatients (HYPER-FRAIL pilot study 2)

Abstract

Background and aims

Few studies have evaluated frailty in older hypertensive individuals and the most appropriate tools to identify frailty in this population have yet to be identified. This study compared the performance of six frailty instruments in the prediction of 1-year functional decline in older hypertensive outpatients.

Methods

The HYPERtension and FRAILty in Older Adults (HYPER-FRAIL) longitudinal pilot study involved hypertensive participants ≥75 years from two geriatric outpatient clinics at Careggi Hospital, Florence, Italy, undergoing identification of frailty with four frailty scales (Fried Frailty Phenotype, Frailty Index [FI], Clinical Frailty Scale [CFS], Frailty Postal Score) and two physical performance tests (Short Physical Performance Battery [SPPB] and gait speed). Prediction of 1-year functional decline (i.e. a ≥ 10-point Barthel Index decrease between baseline and follow-up) was examined based on ROC curve analysis and multivariable logistic regression.

Results

Among 116 participants, 24 % reported functional decline. In the ROC curve analyses, FI (AUC=0.76), CFS (AUC=0.77), gait speed (AUC=0.73) and the SPPB (AUC=0.77) achieved the best predictive performance, with FI ≥0.21 and CFS ≥4 showing the highest sensitivity (82 %) and negative predictive value (91 %). Frailty identified with FI, CFS or physical performance tests was associated with an increased risk of 1-year functional decline, independently of baseline functional status and comorbidity burden.

Conclusions

FI, CFS and physical performance tests showed similar predictive ability for functional decline in hypertensive outpatients. The CFS and gait speed might be more suitable for clinical use and may be useful to identify non-frail individuals at lower risk of functional decline.