Ann W. McMahon, M.D. and Gerald Dalpan, M.D., M.H.S.
Office of Pediatric Therapeutics (A.W.M.), and the Office of Surveillance and Epidemiology
(G.D.P.), Food and Drug Administration, Silver Spring, MD
Over the past 20 years, several legislative initiatives in the United States have resulted in
hundreds of clinical trials assessing the treatment effects of various drugs in children.1
Since 1998, a total of 712 drug labels have been updated with information from randomized trials
about drug use, efficacy, and safety in children. Among the most influential of the relevant
laws were the 2002 Best Pharmaceuticals for Children Act and the 2003 Pediatric Research
Equity Act, both of which require the Food and Drug Administration (FDA) to review case
reports of adverse events associated with the use of medical products studied in children.
This review process, conducted by members of the FDA Pediatric Advisory Committee, has
contributed substantially to our knowledge of medical product safety in children. But
exclusive reliance on this approach doesn’t allow the agency to gather data on the full range
of safety issues that may arise when children use medical products (defined as drugs, biologics,
or medical devices). By their nature, case reports are often unable to clearly link
adverse effects with a particular product or quantify relatively small yet clinically relevant
treatment-related effects. These limitations are especially relevant for adverse effects that
occur months or years after treatment and those related to growth and development.
Traditional clinical trials can provide more answers on the long-term effects of medical
products but in this context can be unethical, impractical, or expensive.
There are important challenges specific to assessing drug safety in children. Long-term
followup is needed to observe effects over multiple developmental stages. Drugs may
contribute to decreased growth rates during or after exposure, and exposure to medications
in utero may cause delayed effects. Children take medications less often than adults, and
measurement of rare events requires large databases. Finally, treating rare diseases is an
important part of pediatric care.
A wide range of study designs can be used to assess drug safety in both adult and pediatric
populations. Increasingly, studies can incorporate data obtained in the course of clinical care,
or real-world data, to generate real-world evidence. Sources of real-world data include
electronic health records, claims and billing data, product- and disease-based registries, and
mobile and wearable devices.
There are substantial gaps in evidence regarding the safety of many drugs in children. Realworld
data could help to fill these gaps in several ways. First, such data can provide greater
insight into the long-term effects of in utero exposure to medicines on growing children. For
example, one observational study used chart reviews and clinical assessments to demonstrate
that exposure to valproate in utero was associated with decreased IQ at 6 years of age,
whereas exposure to carbamazepine, lamotrigine, or phenytoin was not.2
Similarly, real-world data can inform our understanding of the long-term effects of exposure
to medicines early in children’s development. For example, infants in the neonatal intensive
care unit (NICU) are often treated with multiple medications and other interventions, some
of which may be continued for months or years. Long-term follow-up reports on children
cared for in the NICU show that as young adults, they may have abnormal
neurodevelopmental outcomes. But such assessments are often confounded by the effects of
other medical interventions or underlying disease pathophysiology, so attributing an
outcome to a specific drug or other intervention is difficult. Incorporating collection of realworld
data into thoughtful study designs — which may include randomization and
comparator groups — could help investigators move the field toward more definitive
Real-world data can also be useful for assessing the safety of medications used to treat
chronic conditions. As the number of children with certain chronic diseases has increased,
more children are being exposed to a range of drugs over a long period, but outcomes
associated with long-term use are difficult to study in traditional clinical trials.3
Childhood prescriptions for asthma medications increased by 14% in the United States between 2002
and 2010, for example, and long-term use of these drugs may result in decreased growth
rates.4 Carefully designed studies using both real-world data and clinical assessments can be
more efficient for identifying long-latency outcomes in children with chronic diseases.
In addition, because most children are generally healthy, many drugs are used relatively
infrequently in the pediatric population. Drug-associated adverse events that are rare in
adults might be even rarer — and thus more difficult to detect — in children. Collecting data
on large numbers of children is necessary for studying certain safety issues, and several data
sources may be required, as is often the case in using real-world data. Moreover, short-term
clinical trials have generally been used to assess the safety of drugs used for short periods in
children. Sources of real-world data that include long-term follow-up could offer an effective
way to assess longer-term safety. Determining long-latency effects using such data poses
challenges, however, since children, like adults, frequently move from one health insurance
plan to another, making longitudinal follow-up difficult. Using additional or alternative
sources of real-world data can help overcome this limitation. For example, the Swedish
Cancer Registry was able to track cancer outcomes over a 50-year period in patients who
had been treated for inflammatory bowel disease in childhood.5
Finally, rare diseases manifesting during childhood, some of which may affect fewer than
100 people worldwide, generally can’t be studied using large medical record–based
databases. Disease-based registries that include clinical details are needed to assess
treatment and disease-based outcomes, including adverse effects associated with drugs.
To read more from this FDA Update, click here.