The Adaptive Platform Trial for Kidney Disease

Overview

Background: Randomized clinical trials (RCTs) are essential for evaluating intervention effects but are often challenged by regulatory and logistical burdens, high costs, and extended timelines. To address these challenges, the 'Adaptive Platform Trial in Kidney Disease' (APT-KIDNEY) will establish an investigator-initiated platform trial built on a unified regulatory, contractual, and operational framework. The platform emphasizes adaptive, cost-efficient methodology, automated data capture via linkage to electronic health records and administrative registers, and stakeholder engagement. Objectives: The primary objective of APT-KIDNEY is to establish an adaptive platform trial for evaluation of multiple interventions in patients with advanced kidney disease as defined by an estimated glomerular filtration rate \< 30 ml/min/1.73 m2 or end-stage kidney disease (ESKD) on dialysis or conservative care. Study design: APT-KIDNEY is a pragmatic, randomized, embedded, multifactorial, adaptive platform trial with interventions organized into domains, emphasizing low-intervention comparisons. Domains may be open-label or blinded and will be able to use response-adaptive randomization, adaptive stopping and arm-dropping, and adaptive enrichment to enhance efficiency and relevance where applicable. Study population: Adults (≥18 years) with advanced kidney disease defined by eGFR \< 30 mL/min/1.73 m2 for ≥3 months or ESKD on hemo- or peritoneal dialysis who are eligible for ≥1 one domain. Key exclusions include inability to provide informed consent; domain-specific exclusions may apply, but eligibility cannot be broadened beyond the core protocol. Trial outcomes: Core outcomes will be all-cause mortality, major adverse cardiovascular events (nonfatal myocardial infarction, nonfatal ischemic stroke, or cardiovascular death), and health-related quality of life (EQ-5D-5L). Abbreviated methods: APT-KIDNEY will permit domains to use frequentist and/or Bayesian methods. Primary analyses will target prespecified primary estimands and be conducted using the full analysis set. Prespecified sensitivity analyses will assess robustness to alternative strategies for intercurrent events and missing data, including per-protocol and as-treated supportive analyses. Outcomes are analyzed with generalized linear/mixed models and time-to-event methods with covariate adjustment. Frequentist analyses will be fixed-sample or group-sequential; results will be reported with 95% CIs and p-values, and Bayesian analyses will report posterior effects with 95% credible intervals and posterior probabilities. Bayesian domains will primarily use neutral, mildly skeptical priors. Multiplicity will be controlled at the domain level by a prespecified hierarchy: primary comparisons will precede secondary outcomes. Advanced adaptive domains will be evaluated by simulation to quantify operating characteristics including, power and Type I error, and the impact of outcome delays and missing data. Perspectives: APT-KIDNEY will establish an enduring, investigator-led platform for pragmatic, embedded nephrology trials, reducing start-up time and administrative burden through a shared regulatory and operational framework. Using standardized core outcomes and automated follow-up via electronic health records and national registers, it will generate faster, comparable, practice-relevant evidence across multiple interventions.