Abstract
Kidney transplantation (KT) experts did not support the use of subjective unintentional weight loss to measure shrinking in the physical frailty phenotype (PFP); a clinically feasible and predictive measure of shrinking is needed. To test whether unintentional weight loss could be replaced by an assessment of sarcopenia using existing CT scans, we performed a prospective cohort study of adult KT recipients with original PFP (oPFP) measured at admission (December 2008-February 2020). We ascertained sarcopenia by calculating skeletal muscle index from available, clinically obtained CTs within 1-year pre-KT (male < 50 cm(2) /m(2) ; female < 39 cm(2) /m(2) ) and combined it with the original four components to determine new PFP (nPFP) scores. Frailty was classified by frailty score: 0: non-frail; 1-2: pre-frail; ≥3: frail. Mortality and graft loss hazard ratios (HRs) were estimated using adjusted Cox proportional hazard models. Model discrimination was quantified using Harrell's C-statistic. Among 1113 recipients, 18.6% and 17.1% were frail by oPFP and nPFP, respectively. Compared to non-frail recipients, frail patients by either PFP had higher risks of mortality (oPFP HR = 1.67, 95% CI: 1.07-2.62, C = 0.710; nPFP HR = 1.68, 95% CI: 1.06-2.66, C = 0.710) and graft loss (oPFP HR = 1.67, 95% CI: 1.17-2.40, C = 0.631; nPFP HR = 1.66, 95% CI: 1.15-2.40, C = 0.634) with similar discriminations. oPFP and nPFP are equally useful in risk prediction for KT recipients; oPFP may aid in screening patients for pre-KT interventions, while nPFP may assist in nuanced clinical decision-making.