FDA Pediatric Indications for Medical Devices: Extrapolation, Incentives and Precedents

FDA pediatric indications for medical devices: extrapolation, incentives and precedents

FDA's pediatric device policy rests on four pieces. The first is a statutory and regulatory definition of the pediatric population. The second is a disclosure requirement (section 515A) for PMA, HDE and PDP submissions. The third is a framework for extrapolating existing adult clinical data to pediatric uses in PMA, HDE and De Novo submissions. The fourth is a small set of incentives, most notably the HDE profit exemption for pediatric humanitarian use devices. This article summarizes each one and then shows how they appear in actual PMA, De Novo, 510(k) and HDE decisions.

Key takeaways

• "Pediatric" runs from birth through age 21. Under 21 CFR 814.3(s), pediatric patients are those aged 21 or younger at diagnosis or treatment, up to but not including the 22nd birthday ²⁸⁸. Because of this definition, many "adult" device indications (18 years and older) also cover ages 18 to 21, and FDA reviews address that group explicitly ¹⁴⁴¹⁴⁵.
• Extrapolation is a formal, final-guidance framework. FDA's 2016 final guidance Leveraging Existing Clinical Data for Extrapolation to Pediatric Uses of Medical Devices covers PMA, De Novo and HDE submissions, but not 510(k)s ^248261262.
• Safety and effectiveness are extrapolated separately. FDA expects safety extrapolation, and full extrapolation in particular, to be less common than effectiveness extrapolation ²⁶⁵²⁶⁶.
• Section 515A asks for information, not new studies. PMAs, PMA supplements, HDEs and PDPs must describe affected pediatric subpopulations and patient numbers from readily available information. That information alone does not support a pediatric indication ⁸⁰⁸¹⁸⁹.
• The main financial incentive is the HDE profit exemption. Humanitarian use devices (HUDs) indicated for pediatric patients may be sold for profit up to an annual distribution number (ADN) ^332335336.

How FDA defines the pediatric population

FDA's Premarket Assessment of Pediatric Medical Devices guidance (final, March 24, 2014) defines pediatric patients as individuals from birth through age 21 ³¹⁶³¹⁷. It gives these approximate subgroups ³¹⁸³¹⁹:

The guidance also flags low-birth-weight newborns (under 2.5 kg), very-low-birth-weight newborns (under 1.5 kg) and preadolescents (typically 11 to 13 years) as groups that clinical studies and labeling may need to address. It cautions that size and developmental stage can matter more than chronological age for a given device ³²⁰. The regulation follows the same structure: 21 CFR 814.100(c) names neonates, infants, children and adolescents as the pediatric subpopulations ²⁸⁹.

Premarket expectations for pediatric devices

The 2014 premarket assessment guidance sets out several expectations for sponsors seeking pediatric indications:

• Clinical evidence and subject protection. Clinical evaluation may be needed to support pediatric marketing. Those studies should follow IDE requirements and be designed to protect a vulnerable population, taking account of growth, development and differences from adults ³²¹³²².
• Design and bench testing. Sponsors should consider whether an adult device needs redesign for pediatric height, weight, anatomy, physiology, activity level, maturity, immune status and growth. Relevant testing may include biocompatibility, sterility, environmental exposure and device-specific bench or animal work ³²⁴.
• Labeling and risk. Labeling should be pediatric-specific where needed. Pediatric risks should be handled through study and device design, not by assuming that adult use translates directly ^325322308.

Section 515A pediatric use information

Section 515A applies to new HDEs, PMAs, PMA supplements and PDPs submitted on or after April 10, 2014. These submissions must describe, from readily available information, the pediatric subpopulations affected by the disease or condition and the number of affected pediatric patients ⁸⁰⁸¹. FDA's Providing Information about Pediatric Uses of Medical Devices guidance (final, May 1, 2014) recommends the following ⁷⁹:

• Put the material in a separate section titled "515A Pediatric Device Use Information" and document a reasonable search effort ⁸⁴.
• Draw on public-domain biomedical, regulatory and product-research sources. A new clinical study is not required ⁸²⁸³.
• Present a table of indication, pediatric incidence and prevalence, subpopulation and age range, applicable device or component, and data source ⁸⁵⁸⁶.
• Update the information with each PMA supplement, or state that nothing new is available and refer back to the last complete submission ⁸⁷⁸⁸.

The codified requirements are 21 CFR 814.20(b)(13) for PMAs, 814.104(b)(6) for original HDEs and 814.39(c)(2) for PMA supplements. FDA can also request the information in an amendment or as a condition in an approvable letter ^{290291292293294295}. The guidance states that 515A information on its own does not establish safety and effectiveness for a new pediatric indication ⁸⁹.

Extrapolation of adult data to pediatric uses

Scope and statutory basis

The 2016 extrapolation guidance traces its authority to the Pediatric Medical Device Safety and Improvement Act (PMDSIA), Title III of the 2007 FDA Amendments Act. PMDSIA permits adult effectiveness data to support pediatric effectiveness when the course of the disease and the device's effects are sufficiently similar. Where appropriate, it also permits extrapolation between pediatric subpopulations ³³⁷. The guidance applies to pediatric indications sought through PMA, De Novo or HDE, and not to 510(k) submissions ^261262323.

When extrapolation is appropriate

FDA asks four questions ²⁶³²⁶⁴:

1. Are the disease course and the device's effects sufficiently similar in adults and children?
2. Are the adult endpoints relevant to pediatric patients?
3. Are the existing data high-quality valid scientific evidence?
4. Taken together, do the data support a fair and responsible conclusion of reasonable assurance of safety and effectiveness, or of probable benefit for an HDE?

The decision pathway

1. Threshold questions. Does the condition occur in the intended pediatric population, and do the existing data include an endpoint relevant to the pediatric device effect? If the answer to either is no, extrapolating adult data is not appropriate ²⁶⁷.
2. Differences assessment. FDA looks for clinically meaningful differences in implantation or contact, device characteristics, pediatric patient characteristics, disease characteristics and other factors ²⁶⁸. The device effect should point in the same direction in adults and children, with a similar but not necessarily identical magnitude of benefit ²⁶⁹.
3. Full extrapolation. If no meaningful differences are expected and the adult data are good enough, adult data can replace prospective pediatric data for the endpoint ²⁷⁰²⁷¹.
4. Partial extrapolation. If differences can be accounted for and the adult data are of adequate quality, adult and pediatric data are combined statistically ²⁷⁰²⁷². If the data cannot support a sound conclusion, they should not be extrapolated ²⁷³.

Statistical borrowing

For partial extrapolation, FDA describes Bayesian methods that combine adult prior information with pediatric results. How much information is borrowed depends partly on how similar the data turn out to be ²⁷⁴²⁷⁵. Hierarchical models borrow more when study results agree. They do not assume adult and pediatric patients can simply be pooled, and observed differences limit the borrowing ²⁷⁶. Models can adjust for measured age-related covariates such as size or growth ²⁷⁷. For a single adult study, the guidance discusses discounting approaches such as power priors and their limitations, and it notes that non-Bayesian borrowing methods are also acceptable ²⁷⁸²⁷⁹.

Illustrative scenarios from the guidance

• Full extrapolation for effectiveness only. For an adult-approved pleural air-leak sealant, adult effectiveness data could replace a pediatric effectiveness study. FDA still recommended a pediatric safety study because of concerns about adhesions and possible reoperation ³⁵⁶.
• Partial extrapolation of a reference standard. For a diagnostic proposed down to age 2, the adult reference test needed too much blood to use safely in young children. Calibrated adult reference-standard data could be borrowed, with body size modeled ³⁵⁷.
• Partial extrapolation across follow-up periods. Eight-month adult data were modeled with prospectively collected adolescent data to predict a 12-month response, which reduced the adolescent sample size ³⁵⁸.

The guidance also says extrapolation need not follow statutory age-group boundaries. Extrapolating from adults to adolescents may be easier than extrapolating to younger children ³⁵⁹.

510(k) devices: the X-ray imaging example

The extrapolation guidance does not cover 510(k)s. FDA does, however, address pediatric evidence for some cleared device types through device-specific guidance. The Pediatric Information for X-ray Imaging Device Premarket Notifications guidance (final, November 28, 2017) covers devices indicated for pediatric use and general-use X-ray devices expected to see considerable pediatric use ¹¹³³⁷¹. Its recommendations include:

• Pediatric risk-mitigation features, such as size-appropriate automatic exposure control and pediatric protocols ^373374375.
• Laboratory image-quality and dose testing across the indicated range of patient sizes, using size-representative phantoms ³⁷⁶³⁷⁷.
• Pediatric-specific labeling, including the caution "Use special care when imaging patients outside the typical adult size range" ^379380381.

FDA favors laboratory testing in this guidance. It will request pediatric clinical images only if laboratory testing and extrapolation from adult images are not enough to show substantial equivalence ³⁷⁸.

Pediatric device incentives and support programs

The 2019 HDE guidance repeats that new HDE applications must include pediatric subpopulation information. FDA treats this as a disclosure requirement, not an incentive ³³⁹.

Examples of devices authorized for pediatric populations

PMA approvals with pediatric indications

Summaries of safety and effectiveness data (SSEDs) for PMAs now usually include a "Pediatric Extrapolation" section. Many state that no clinical data were leveraged, for example Symplicity Spyral, WiSE CRT and SAPIEN M3 ^410411413. Several companion diagnostics take a narrower route, extending adult data to ages 18 to 21. Examples are OncoMate MSI Dx and Signatera CDx ^4950412.

De Novo authorizations

FDA press announcements also record De Novo authorizations with pediatric indications ^{463464465467468469}:

• Monarch eTNS for ADHD, ages 7 to 12.
• IB-Stim for IBS-related functional abdominal pain, ages 11 to 18.
• CARPEDIEM for critically ill pediatric patients weighing 2.5 to 10 kg.

510(k) clearances

Cleared pediatric devices often define their population by weight or size rather than age, and they usually rely on a pediatric predicate rather than new pediatric clinical data:

• ANNE Pediatric (K221530, Sibel, November 22, 2022). Indicated for neonates and infants up to 2 years ¹⁷³²⁰⁵. Clinical testing in 137 neonates included a heart-rate comparison against a reference monitor ²⁰⁶.
• Quantum Blood and Fluid Warming System (K192325, Life Warmer, October 18, 2019). Pediatric patients older than 28 days were added without new performance testing. The submission relied on prior testing, an updated pediatric risk analysis and pediatric-specific labeling ^181224226.
• 3880 MRI Patient Monitoring System (K182900, Iradimed, December 14, 2018). Indications use FDA's infant, child and adolescent age bands ¹⁸⁴²³⁰.
• F&P Optiflow Junior 2/2+ (K222197, April 6, 2023) and Comfort Flo CubCannula (K241778, December 19, 2024). Both are neonatal and pediatric respiratory interfaces cleared against pediatric predicates ^163194209.
• Embrace Neonatal MRI System (July 2017). Cleared for neonates weighing 1 to 4.5 kg ⁴⁵⁹⁴⁶⁰.

Humanitarian device exemptions

• Melody Transcatheter Pulmonary Valve (H080002, January 25, 2010). Approved for pediatric and adult patients with dysfunctional RVOT conduits ⁴³⁶⁴³⁹.
• Flourish Pediatric Esophageal Atresia Device (2017). Approved under an HDE for infants up to one year old ⁴⁵⁷⁴⁵⁸.

Practical considerations for sponsors

• Address ages 18 to 21 explicitly. Because the pediatric definition extends through age 21, an "adult" indication starting at 18 already includes pediatric patients. Recent De Novo and PMA decisions handle this with a short extrapolation rationale stating that patients aged 18 to 21 have no additional differences or risks compared with the adults studied ^14415415749.
• Plan extrapolation before the adult pivotal study. Partial extrapolation depends on relevant shared endpoints and on data quality ²⁶⁷²⁷⁰. Adult protocols designed with pediatric borrowing in mind make later hierarchical modeling more feasible ²⁷⁶²⁷⁷.
• Handle safety separately. As the sealant example shows, a full effectiveness extrapolation can still come with a dedicated pediatric safety study ³⁵⁶²⁶⁵.
• Weigh HDE early for rare pediatric conditions. The pediatric profit exemption and ADN can make a HUD commercially viable ³³²³³⁶. Some pediatric devices first approved under HDE, such as EXCOR, later moved to PMA ⁴³⁵.
• For 510(k)s, use device-specific guidance and predicates. The extrapolation framework does not apply, so pediatric support usually comes from bench testing across patient sizes, risk analysis and pediatric labeling, as in the X-ray imaging guidance and K192325 ^376224226.