Safety Signal Detection and Evaluation in the Pediatric Population

SAFETY SIGNAL DETECTION AND EVALUATION IN THE PEDIATRIC POPULATION

The ‘pediatric population’ encompasses preterm babies to adult-sized adolescents, and many aspects of drug disposition, efficacy and safety profile differ over this age range, making extrapolation of safety data from adults to the pediatric population very problematic. Although drugs are used off-label, most have little or no pharmacokinetic data to support rational dosing in pediatrics. Even for those drugs that have had formal clinical studies in pediatric patients, the pre-market safety assessments are limited by inade-quately powered studies to evaluate safety. Therefore, postmarketing monitoring of drug safety in pediatrics largely falls on careful evaluation of spontaneous reports and when possible data from epidemiologic studies or sponsor conducted phase 4 post-approval studies.

Monitoring of postmarketing data has led to the detection of important drug adverse reactions which are unique to pediatrics. Examples include the use of ciprofloxacin in neonates and its effect on teeth,²² valproic acid and liver toxicity,²³ isotretinoin and depression/suicide,²⁴ as well as other examples cited elsewhere in the chapter. Continuous monitoring of spontaneous postmarketing reports supplemented by epidemiological data and data from phase 4 studies with a focus on pediatrics is critical to better define the risks of drugs in the pediatric population.

A pediatric safety signal may arise when the evalu-ation of spontaneously reported pediatric adverse drug events include the following findings:

1. serious and unexpected drug adverse events that are unique to pediatrics, i.e., not described in the approved product labelling;

2. serious drug adverse events that may be related to a labelled event but differ from the labelled event because of:

•  greater severity (hepatic necrosis vs. increase in liver enzymes or hepatitis in the labelling);

•  greater specificity (cerebro-vascular accidents vs. cerebral thrombo-embolism or cerebral vasculitis).

3. a new high-risk pediatric subgroup for ADRs is detected arising from off-label use for an unstudied pediatric age group or indication.

Once a potential safety signal is detected, evaluation of the signal for possible causality is challenged by the limitations of passive reporting systems. Both the numerator (underreporting of adverse events) and the denominator (lack of good national estimates of pedi-atric drug exposures) are uncertain and usual reporting rates calculated from these data can be misleading and difficult to interpret. The value of these calculations is further reduced by the lack of valid data on back-ground incidence rates against which the calculated reporting rates are compared. Consequently, it is often not possible to measure excess risk unless the reported event of concern has a hard endpoint (e.g., death, liver necrosis) and it has a low background incidence rate in the general pediatric population.