Get started
Work

Optimizing Relapse Prevention and Treatment Strategies in Childhood Nephrotic Syndrome using Target Trial Emulation and Adaptive Clinical Trial Methods

Year: 2026
Type: dissertation
Abstract: Background:Nephrotic syndrome is one of the most common childhood kidney diseases. Frequent relapses and high cumulative steroid exposure cause substantial harm. Effective strategies to prevent relapses and reduce steroid burden are needed. The objectives of this thesis were to 1) compare the effectiveness of cyclophosphamide and calcineurin inhibitors for preventing nephrotic syndrome relapses, 2) determine the optimal design of a Bayesian adaptive randomized controlled trial (RCT) comparing low-dose vs. standard-dose steroids for relapse treatment, and 3) establish feasibility of this RCT through a pilot study. Methods: Aim 1: Target trial emulation comparing cyclophosphamide vs. calcineurin inhibitor treatment in children with steroid-sensitive nephrotic syndrome. Aim 2: Simulation study of a planned Bayesian adaptive RCT. For each candidate trial design, we simulated 10,000 trials to estimate operating characteristics (power, type 1 error rate, and sample size). Aim 3: Internal pilot study to evaluate the feasibility of a planned Bayesian adaptive RCT comparing low-dose vs. standard-dose steroid treatment for nephrotic syndrome relapses. Results:Aim 1: During median 5.5-year follow-up, there was no difference between calcineurin inhibitor vs. cyclophosphamide in time-to-relapse (HR 1.25, 95%CI 0.84-1.87), subsequent relapses, steroid-sparing immunosuppression use, or kidney function. Calcineurin inhibitor treatment was associated with higher risks of hospitalization and intravenous albumin use. Aim 2: We simulated 5.4 million trials to identify three adaptive RCT designs with the lowest sample size and type 1 error rate <5%. We simulated another 1.2 million trials to identify the optimal design under varying conditions (treatment effect, recruitment rate, and prior distributions). The estimated sample size required was 71% lower than for a fixed-sample design. Aim 3: Over a 10-month period, 40 children with nephrotic syndrome relapse were enrolled in a pilot RCT. All pre-specified feasibility targets were met, and low-dose steroids had acceptable safety. These findings suggest that a full-scale Bayesian adaptive RCT is feasible, safe, and likely to reach a conclusive outcome. Conclusions:Causal inference and Bayesian adaptive RCT methods can be used to answer comparative effectiveness questions in rare childhood diseases, such as nephrotic syndrome. Further research is needed to advance evidence for steroid minimization strategies in childhood nephrotic syndrome.
Language: English
Open Access status: green
OpenAlex ID: W7203961078
Cites:
Cited by:
Related to: 10