The Use of Extrapolation to Promote Clinical Trials in Pediatric Nephrology

Introduction

Kidney disease is relatively rare in children. Unfortunately, only a limited number of therapies for most kidney diseases are approved for use in pediatric patients. This reflects the difficulty in performing well-designed, logistically feasible randomized controlled trials (RCTs) to evaluate the efficacy and safety of new treatments in rare diseases in this age group.¹ As a consequence of the 13-year gap between drug approval in adults versus children, there is frequent reliance on off-label use of agents in children without adequate dosing guidelines or safety profiles.² In this Perspective, we address extrapolation, a method to streamline the design of RCTs in pediatric patients with the same disease entity as adults by (1) maximizing the use of existing information gained about novel therapies that have been demonstrated to be effective in adults and (2) leveraging what is known about safety and pharmacokinetics from studies of the investigational product for other diseases in pediatric patients. This article does not address the use of extrapolation as an approach to optimize approved therapies.

Extrapolation: General Overview

Pediatric extrapolation is defined in the International Council on Harmonization E11(R1) guideline as “an approach to providing evidence in support of effective and safe use of drugs in the pediatric population when it can be assumed that the course of the disease and the expected response to a medicinal product would be sufficiently similar in the pediatric [target] and reference (adult or other pediatric) population.” Examples of drugs that were approved for pediatric patients based on the application of extrapolation include sacubitril/valsartan for heart failure, belimumab for SLE, and tocilizumab emergency use authorization for children 2 years or older with coronavirus disease 2019.

The extent to which extrapolation can be applied falls along a spectrum. If the disorder has a comparable natural history and response to treatment, then a greater degree of extrapolation may be warranted and the only formal investigations needed to obtain drug approval may be a limited RCT in children with a focus on safety, toxicology, pharmacodynamics, and pharmacokinetics of a pediatric formulation. This may be the case in primary glomerular disorders such as FSGS.³ By contrast, under circumstances in which the pediatric disease has no adult counterpart, such as congenital anomalies of the kidney and urinary tract, then approval may require a fully powered phase 3 RCT without any extrapolation. Enrolling selected pediatric patients simultaneously into RCTs for adults with diseases that have a similar trajectory in both age groups, such as adolescents with primary glomerular diseases, is a parallel approach to expedite drug approval for children.⁴

By reducing the scope of work required in the performance of RCTs in pediatric nephrology, extrapolation should lead to cost savings related to per-patient costs, number of participating sites, and study duration. The exact amount will vary depending on the target sample size and overall study design.

Extrapolation: Key Elements

The scope of the data that will be required to make extrapolation feasible centers on three key domains: (1) clinical characterization of disease trajectory and treatment response; (2) biomarkers to assess the effect of treatment, target engagement, or pharmacodynamic response; and (3) prediction tools based on readily available information to identify patients at higher risk of adverse outcomes.

Clinical Data

Clinical information about demographics, physical findings, medications, and laboratory results may be retrieved from a wide range of sources including electronic health records from single practices, hospitals, health systems, registries, observational cohort studies, or clinical trials. Data repositories vary in the depth of information, data fields such as height to enable calculation of eGFR, methods for clinical or laboratory measurements, timing of assessments, the duration of follow-up, and comprehensiveness of input regarding medications and the occurrence of comorbidities, complications, and adverse drug reactions. The recent Proteinuria and GFR as Clinical Trial Endpoints in Focal Segmental Glomerulosclerosis initiative demonstrated that aggregating diverse real-world data (nine datasets with 1626 eligible participants) can be meaningful, namely, development of a proteinuria threshold as a surrogate end point for application in FSGS trials.⁵ There is an urgent need to define the minimal standards that real-world and research datasets need to satisfy and to harmonize a minimal core dataset before available information can be relied on to provide an accurate portrayal of the natural history of the disorder and response to treatment. Standardization of laboratory measurements such urinary albumin concentration would ensure comparability of results across studies and promote extrapolation.

Biomarkers

Biomarkers can enhance pediatric–adult efficacy extrapolation. When the extrapolation approach supports doing so, a bridging biomarker can be used to extrapolate efficacy from adult to pediatric patients with a disease that is biologically similar between the two populations, assuming that the biomarker is on the causal pathway of disease, the product is safe and effective in adults with the disease, and a large proportion of the treatment effect on the hard outcome in the trial in adults can be explained by the treatment effect on the bridging biomarker. If these criteria are met in younger patients, then pediatric trials could be streamlined to evaluate the effects on the bridging biomarker. Biomarkers can also be used to confirm target engagement, document a pharmacodynamic response to the test therapy, and monitor safety. Novel candidate biomarkers are emerging that range from single-analyte assays to multiomic signatures.⁶ For example, urinary EGF has been demonstrated to be a strong indicator of kidney tissue integrity in adults with type 2 diabetes.⁷ Proteomic data from youth with type 2 diabetes suggest that multiprotein signatures outperformed clinical variables in predicting kidney complications.⁶ Notable challenges remain in pediatrics, including age-related changes, hormonal fluctuations, and the need for robust normative data across the lifespan and developmental stages. Computer modeling and machine learning methods are expected to facilitate the development of biomarker profiles comprising single or panels of analytes.⁸ There is a need to establish standardized protocols for sample collection, processing, and storage in shared biorepositories. Candidate biomarkers will require periodic re-evaluation to ensure that they are still valid.

Risk Prediction

Identification of patients with CKD at high risk of progression to kidney failure is an important consideration in clinical trial design and extrapolation. The degree of BP control and level of proteinuria are well-established risk factors for more rapid deterioration based on evidence from the Effect of Strict Blood Pressure Control and ACE Inhibition on the Progression of CRF in Pediatric Patients trial and the Chronic Kidney Disease in Children cohort study⁹ as well as real-world data from PEDSnet.¹⁰ Risk prediction tools that incorporate multiple factors including eGFR, albuminuria, and novel biomarkers such as urinary EGF may identify pediatric patients who are more likely to have disease progression. Such tools have been developed and validated in adult patients with CKD.¹¹ This work needs to be replicated if it is to be used to identify specific subgroups of pediatric patients who are being considered for enrollment in RCTs based on risk of disease progression and for whom extrapolation can be applied. Table 1 summarizes the challenges to the implementation of extrapolation in the design of RCTs in pediatric nephrology and some proposed solutions.

Table 1.

Implementation of extrapolation in pediatric nephrology

Key Element	Challenges	Proposed Solutions
Clinical and laboratory data	Wide range of data sources	Establish shared data repository
	Varied depth of information	Identify key mandatory data elements
	Lack of uniform methods of measurement	Define standards for timing and quality of measurements
Biomarkers	Wide variety of uses (e.g., target engagement, PD response, drug effect, safety)	Increased research to support valid use of biomarkers for specific purposes
	Paucity of validated bridging biomarkers	Evaluation of existing data from adult trials and extension to candidate populations (e.g., adolescents)
	Nonstandardized collection and use of specimens	Encourage collection of biosamples in all RCTs
	Limited availability of samples for study	Establish shared biosample repository
Risk prediction	Inadequate identification of high-risk patients to define disease trajectory	Use available datasets to refine risk prediction in pediatric CKD
	Limited enrollment of high-risk patients into RCTs	Incorporate validated risk prediction tools in all RCTs

PD, peritoneal dialysis; RCT, randomized controlled trial.

Conclusion and Future Directions

International collaborations are building resources to improve patient outcomes, with growing consensus that extrapolation offers a valid and feasible approach to facilitate clinical trials in pediatric nephrology. To promote the use of this method, an initiative jointly led by the Kidney Health Initiative and NephCure called Extrapolation to Support Clinical Trials in Pediatrics has been launched. Extrapolation to Support Clinical Trials in Pediatrics is centered on the three areas of focus in this Perspective: clinical data, biomarkers, and risk prediction. Awareness of extrapolation as an emerging standard element in the planning, design, and implementation of RCTs, one trial at a time, should be promoted among all participants involved in clinical research in pediatric nephrology, including adult experts. It will benefit all children with kidney disease and progressively narrow the historical 13-year gap between adult and pediatric drug labeling.

Supplementary Material

jasn-37-616-s001.pdf ^{(1.4MB, pdf)}

Disclosures

Disclosure forms, as provided by each author, are available with the online version of the article at http://links.lww.com/JSN/F433.

Author Contributions

Conceptualization: Petter Bjornstad, Michelle Denburg, Mark D. Lim, Louise Oni, William E. Smoyer, Joshua Tarnoff, Howard Trachtman.

Supervision: Howard Trachtman.

Writing – original draft: Howard Trachtman.

Writing – review & editing: Petter Bjornstad, Michelle Denburg, Mark D. Lim, Louise Oni, William E. Smoyer, Joshua Tarnoff.

Funding

None.