Assessing frailty in hemodialysis patients: utility of a self-rated visual analogue scale and the role of routine parameters in clinical prediction

Abstract Background There is still no established consensus on the optimal frailty screening tool for routine clinical application in patients with chronic kidney disease (CKD). We evaluated the accuracy of a visual analogue scale (VAS) for frailty screening and identified clinical predictors of frailty based on the modified Fried criteria. Methods This cross-sectional, multicenter study included 122 adult patients on hemodialysis. Frailty status was assessed using modified Fried criteria and a 5-point self-rated VAS. The agreement between these two methods was evaluated using Cohen’s Kappa. For the frailty prediction model we applied logistic regression with backwards selection among a set of clinical parameters associated with this condition. Pre-frail and robust individuals were analyzed as non-frail. Results Median age was 68 years, 71% were male and 42.6% were considered frail according to Fried criteria. Cohen’s Kappa was 0.36 (95% CI [0.21, 0.51]), indicating a fair level of agreement between VAS and Fried criteria. The VAS correctly identified most non-frail patients (specificity 94.3%, PPV 82.6%) while it missed a significant number of frail patients (sensitivity 36.5%, NPV 67.0%). Patients at risk for frailty based on the modified Fried criteria were more likely to have a central venous hemodialysis catheter, lower diuresis, lower creatinine and albumin, a higher prevalence of cardiovascular disease and use of analgesics. A model including these factors demonstrated good predictive performance with an AUC of 0.816 [95% CI 0.74, 0.90]. Conclusion A self-rated VAS for frailty has the potential to simplify frailty screening and enhance its integration into routine nephrology care. A model based on routine parameters can support the identification of frail patients.