Quantifying lead time bias when estimating patient survival in preemptive living kidney donor transplantation

Abstract

Preemptive kidney transplantation is the preferred initial renal replacement therapy, by avoiding dialysis and reportedly maximizing patient survival. Lead time bias may account for some or all of the observed survival advantage, but the impact of this has not been quantified. Using the Australia and New Zealand Dialysis and Transplant (ANZDATA) Registry, we included adult recipients of living donor kidney transplants during 1998-2017. Patients were transplanted preemptively (n = 1435) or after receiving up to 6 months of dialysis (n = 712). We created a matched cohort using propensity scores, and accounted for lead time (dialysis and estimated predialysis) using left-truncated Cox models with the primary outcome of patient survival. The median eGFR at transplantation was 6.9 mL/min per 1.73 m2 in the non-pre-emptive, and 9.6 mL/min per 1.73 m2 in the preemptive group. In the matched cohort (n = 1398), preemptive transplantation was not associated with a survival advantage hazard ratio (HR) for preemptive vs non-pre-emptive 1.12 (95% confidence interval [CI] 0.79-1.61). Accounting for lead time moved the point estimates toward a survival disadvantage for preemptive transplantation (eg, HR assuming 4 mL/min per 1.73 m2/year eGFR decline, 1.21 [0.85, 1.73]), but in all cases the 95% CIs crossed 1. The optimal timing of preemptive living donor kidney transplantation requires further study. Preemptive kidney transplantation is the preferred initial renal replacement therapy, by avoiding dialysis and reportedly maximizing patient survival. Lead time bias may account for some or all of the observed survival advantage, but the impact of this has not been quantified. Using the Australia and New Zealand Dialysis and Transplant (ANZDATA) Registry, we included adult recipients of living donor kidney transplants during 1998-2017. Patients were transplanted preemptively (n = 1435) or after receiving up to 6 months of dialysis (n = 712). We created a matched cohort using propensity scores, and accounted for lead time (dialysis and estimated predialysis) using left-truncated Cox models with the primary outcome of patient survival. The median eGFR at transplantation was 6.9 mL/min per 1.73 m2 in the non-pre-emptive, and 9.6 mL/min per 1.73 m2 in the preemptive group. In the matched cohort (n = 1398), preemptive transplantation was not associated with a survival advantage hazard ratio (HR) for preemptive vs non-pre-emptive 1.12 (95% confidence interval [CI] 0.79-1.61). Accounting for lead time moved the point estimates toward a survival disadvantage for preemptive transplantation (eg, HR assuming 4 mL/min per 1.73 m2/year eGFR decline, 1.21 [0.85, 1.73]), but in all cases the 95% CIs crossed 1. The optimal timing of preemptive living donor kidney transplantation requires further study.