#1721 Proteinuria as a clinical trial endpoint in focal segmental glomerulosclerosis: validation of PARASOL findings in RaDaR

Authors:

Abstract
Background and Aims: Focal segmental glomerulosclerosis (FSGS) is a rare condition that affects children and adults, lacks any approved therapy, and has a high risk of progression to kidney failure (KF). The PARASOL consortium was set up to establish evidence to inform adoption of feasible and clinically meaningful surrogate endpoints for FSGS clinical trials. Method: PARASOL combined 9 registries of biopsy-proven or genetic FSGS patients from North America and Europe. Eligible participants required: (1) UPCR ≥1.5 g/g at any time post-biopsy or diagnosis or up to 3 months prior (index UPCR); (2) eGFR ≥30 ml/min up to 3 months prior to index UPCR; (3) ≥1 additional UPCR and eGFR measurement 3 to 24 months post-index; and (4) no evidence of KF (kidney replacement therapy, eGFR sustained <15 ml/min) prior to the first post-index measurements. Relationships between change in eGFR and proteinuria response (across various thresholds) in the first 24 months post-index and subsequent probability of kidney failure beyond 24 months were assessed using Cox proportional hazards models, using a landmark approach with a landmark time of 24 months post-index. Validation was performed among participants fulfilling the same criteria in the UK National Registry of Rare Kidney Diseases (RaDaR). Results: PARASOL contained 1626 eligible participants, with median (IQR) follow-up time from kidney biopsy of 64.2 months (35.2, 103.1). At biopsy, median age was 19 years (IQR 7, 43), median UPCR was 3.8 g/g (IQR 2, 7.2) and median eGFR was 81 ml/min (IQR 54, 109). eGFR and the rate of eGFR decline were associated with KF hazard but were highly variable: sample size requirements for clinical trials using eGFR endpoints ranged from 500–1000 individuals per arm. Proteinuria reduction was strongly associated with reduced KF hazard. Using a UPCR responder threshold of 0.7 g/g, achieved in 406 participants, responders had an 85% lower hazard of kidney failure (HR = 0.15, 95% CI = 0.09–0.26) and a survival difference of +0.27 (95% CI 0.22–0.32) compared to non-responders, consistent in incident and prevalent subpopulations and in subgroups based on age and index proteinuria or eGFR level. Estimated sample size requirements for trials using this proteinuria endpoint was 100–200 per group. The findings were validated in RaDaR, which included 465 eligible participants. At index, RaDaR participants were older, median age 36 (14–55), with lower kidney function, median eGFR 71 ml/min (46–106), and higher proteinuria, median UPCR 5.3 g/g (2.7–8.9). Responder status based on UPCR at 24 months was similar to PARASOL. Stratifying by index UPCR into 1.5 to <3.5 g/g, 3.5 to <7.0 g/g and ≥7.0 g/g, in PARASOL 32%, 31%, and 34% achieved <0.7 g/g by 24 months compared with 25%, 42% and 29% in RaDaR (Fig. 1). At the 0.7 g/g threshold, adjusted hazard ratios and survival differences were similar to PARASOL at 0.07 (0.02–0.27) and 0.29 (0.22–0.36) respectively (Fig. 2), consistent in all subgroups analysed. Conclusion: For most therapeutic effect sizes, variability in eGFR within and between patients with FSGS means that sample size requirements for trials using eGFR or eGFR slope endpoints are unfeasibly large. In contrast, there was consistent and strong association of proteinuria reduction (a biologically plausible parameter) with higher KF-free survival. Proteinuria reduction is therefore a potential surrogate endpoint for FSGS clinical trials that is both feasible and clinically meaningful.