Adolescent Dosing and Labeling Since the Food and Drug Administration Amendments Act of 2007

Importance  During pediatric drug development, dedicated pharmacokinetic studies are generally performed in all relevant age groups to support dose selection for subsequent efficacy trials. To our knowledge, no previous assessments regarding the need for an intensive pharmacokinetic study in adolescents have been performed.

Objectives  To compare US Food and Drug Administration (FDA)–approved adult and adolescent drug dosing and to assess the utility of allometric scaling for the prediction of drug clearance in the adolescent population.

Design  Adult and adolescent dosing and drug clearance data were obtained from FDA-approved drug labels and publicly available databases containing reviews of pediatric trials submitted to the FDA. Dosing information was compared for products with concordant indications for adolescent and adult patients. Adolescent drug clearance was predicted from adult pharmacokinetic data by using allometric scaling and compared with observed values.

Main Outcomes and Measures  Adolescent and adult dosing information and drug clearance.

Results  There were 126 unique products with pediatric studies submitted to the FDA since the FDA Amendments Act of 2007, of which 92 had at least 1 adolescent indication concordant with an adult indication. Of these 92 products, 87 (94.5%) have equivalent dosing for adults and adolescent patients. For 18 of these 92 products, a minimum weight or body surface area threshold is recommended for adolescents to receive adult dosing. Allometric scaling predicted adolescent drug clearance with an overall mean absolute percentage error of 17.0%.

Conclusions and Relevance  Approved adult and adolescent drug dosing is equivalent for 94.5% of products with an adolescent indication studied since the FDA Amendments Act of 2007. Allometric scaling may be a useful tool to avoid unnecessary dedicated pharmacokinetic studies in the adolescent population during pediatric drug development, although each development program in adolescents requires a full discussion of drug dosing with the FDA.

Pediatric drug research is critical to providing safe and effective therapies for children. For more than 20 years, the US Food and Drug Administration (FDA) has sought to provide more complete information about the use of drugs in the pediatric population. The FDA Modernization Act of 1997 addressed the need for improved information regarding drug use in children by providing incentives to the pharmaceutical industry for conducting pediatric trials. Subsequently, the Best Pharmaceuticals for Children Act (BPCA) and the Pediatric Research Equity Act (PREA) were enacted in 2002 and 2003, respectively, and have furthered the incentives and requirements for pharmaceutical products to be studied for use in pediatrics. Since the reauthorization of both of these acts under the FDA Amendments Act of 2007 (FDAAA 2007), more than 50 000 pediatric patients have been enrolled in clinical trials evaluating more than 130 drugs and biologic agents to promote the safe and efficacious use of these products in children.¹ In 2012, BPCA and PREA were reauthorized and became permanent under the FDA Safety and Innovation Act.²

Dedicated pharmacokinetic studies are generally performed during pediatric drug development to determine the pharmacokinetic parameters in different age groups to support dose selection for subsequent efficacy and/or safety trials. The complexity of performing pediatric pharmacokinetic trials during drug development has been discussed elsewhere.³ Conventionally, to account for any age-related differences in maturation, the pharmacokinetics of a drug or biologic agent are evaluated over the entire pediatric age range in which the agent is likely to be used. In all cases, the need for pediatric data must be addressed within the ethical framework of scientific necessity. Pediatric patients should be enrolled only in research that is scientifically necessary and in which the objective(s) cannot be met by enrolling a less vulnerable population, such as adult subjects who can provide informed consent.⁴

Interpolation and modeling and simulation are alternative approaches that have been used to minimize the pediatric population needed to identify an appropriate dosing regimen. Given the extensive information on drug disposition (absorption, distribution, metabolism, and excretion) often available in adults, simplified methods for scaling drug doses, such as allometry, may allow for fewer pediatric patients to be exposed to an experimental drug or biologic agent while still providing adequate pharmacokinetic data.

Allometry is a method of predicting drug clearance for one population based on data from another population, taking into consideration morphologic characteristics and body function in relation to size.⁵ Interspecies allometric scaling has been used extensively during the preclinical to clinical transition in drug development to predict pharmacokinetic parameters in humans. Allometric principles also may be applicable within groups of humans, including between adult and pediatric populations.⁶ Allometric scaling alone may be too simplistic an approach to determine dosing for young children, in whom rapid maturational changes take place in organ systems that affect drug disposition.⁷ However, in adolescence, which can be defined as ranging from 12 to 16 or 18 years of age,⁸ renal capacity and hepatic enzyme expression approach adult levels.⁹ Therefore, allometric power models may provide a useful framework for the initial estimation of drug clearance and dosage requirements in adolescent patients, which could minimize the number of pharmacokinetic trials in the pediatric population. The objective of the current study was to examine pediatric trial results submitted under FDAAA 2007 to compare adolescent and adult dosing information and drug clearance.

The study included products with pediatric studies conducted under BPCA and PREA and submitted to the FDA from September 27, 2007, through February 1, 2012. Adult and pediatric indications for each product were recorded, including indications that may have existed before FDAAA 2007. Dosing information from FDA drug labels was compared for products with concordant indications in adolescent and adult patients. Dosing for adolescents and adults was considered equivalent when dosing was identical in these 2 populations or, if dosing was based on body weight, when an adolescent patient weighing at least 50 kg would receive the adult dose. An adolescent weight of 50 kg was selected to approximate the mean weight of the lower age range of the adolescent population. The National Health and Nutrition Examination Survey, a multistage probability sample of the United States noninstitutionalized population, reports mean weights of 50.8 kg and 52.9 kg for 12-year-old boys and girls, respectively.¹⁰ For the purposes of this study, adolescents were defined as persons 12 to less than 18 years of age.

Adult and adolescent drug clearance data were obtained from FDA-approved drug labels and publicly available databases containing medical, statistical, and clinical pharmacology reviews of pediatric trials submitted to the FDA under section 505A and 505B of the Federal Food, Drug, and Cosmetic Act, as amended by the FDAAA 2007.¹¹ Intravenously and orally administered products that contained a publicly available pharmacokinetic assessment in adolescent patients were included. Biologic agents, hormonal contraceptives, locally acting drugs, and products administered by topical, inhalation, subcutaneous, or intramuscular routes were excluded in this component of the study.

Orally administered products were evaluated separately from intravenously administered products to differentiate between clearance and apparent oral clearance (CL/F, where CL represents clearance and F, oral bioavailability). Drug clearance was scaled with an allometric power function with a fixed exponent of 0.75, which has been shown to adequately describe clearance parameters over a wide range of body weights.¹² For drugs with available pharmacokinetic data in adolescent patients, adolescent clearance (CL_adolescent) was predicted from adult clearance (CL_adult) by using the following equation:

CL_adolescent = CL_adult× (BW_adolescent/BW_adult)^0.75,

where BW_adolescent and BW_adult are the body weights in adolescents and adults, respectively. For clearance prediction, the mean body weight from the adolescent population from each study was used along with a standardized adult body weight of 70 kg.

The percentage error and absolute percentage error between predicted and observed clearance values were calculated for each product.¹³ Linear regression was used to examine the relationship between predicted and observed adolescent drug clearance.

For 126 unique products with pediatric studies submitted to the FDA from trials completed under FDAAA 2007 and the reauthorization of PREA and BPCA (Figure 1), approved adolescent and adult dosing was compared for each indication, with a separate evaluation of adolescent and adult pharmacokinetic clearance.

Ninety-eight of the 126 products (77.7%) have at least 1 labeled indication in adolescent patients. Overall, these 98 products have 186 indications in adults and 127 in adolescents. Of these, 121 indications are concordant between adults and adolescents, 6 are unique to adolescents, and 65 are unique to adults. The unique adolescent indications in labeling include treatment of attention deficit hyperactivity disorder (guanfacine hydrochloride), growth failure due to inadequate secretion of normal endogenous growth hormone (somatropin), juvenile idiopathic arthritis (abatacept, adalimumab, and tocilizumab), and heterozygous familial hypercholesterolemia (rosuvastatin).

Of these 98 products with a labeled adolescent indication, 92 had at least 1 indication concordant with an adult indication for which recommended dosing was compared. Of these 92 products, 87 (94.5%) have equivalent dosing for adult and adolescent patients. For 18 of these 92 products, a minimum weight or body surface area (BSA) threshold is recommended for adolescents to receive the adult dose (Table 1). The 5 products with indications that are concordant between adolescents and adults but not with equivalent dosing are ribavirin-peginterferon alfa-2a, fluocinolone acetonide, paliperidone, candesartan, and valsartan.

Twenty-seven of the 126 products had an available pharmacokinetic assessment with clearance data for the adolescent population. Ninety-nine drugs were excluded, including 43 administered by a nonoral or nonintravenous route, 8 biologic agents, 7 hormonal contraceptives, 3 nonabsorbed orally administered drugs, and 38 products that either did not have a pharmacokinetic assessment in adolescents or did not have such data publicly available. The mean adolescent body weights in the pharmacokinetic studies ranged from 31.6 to 77.1 kg.

Of these 27 products, 8 were administered intravenously. For these 8 products, the percentage error of predicted relative to observed adolescent clearance ranged from −27.6% (caspofungin) to 32.7% (gadobutrol) (Table 2).^11,14,15 The mean absolute percentage error (MAPE) was 16.7%. Allometric scaling was most accurate with acetaminophen, which had a predicted adolescent drug clearance within 2.7% of observed clearance (observed vs predicted, 17.02 vs 16.56 L/h). There was a small and statistically insignificant difference in the clearance prediction between the intravenous drugs that are predominantly metabolized by the liver and those that are predominantly eliminated renally (MAPE, 13.6% vs 19.9% [4 drugs per group]; P > .05). A positive correlation was found between allometry-predicted and observed adolescent clearance values for intravenous products (R² = 0.97; y = 1.08x − 1.1) (Figure 2). The observed adolescent clearance averaged 88.6% of the adult clearance for the intravenously administered drugs.

For the 19 orally administered products, the percentage error of predicted relative to observed apparent clearance in adolescents ranged from −33.8% (anastrazole) to 36% (eplerenone) (Table 3).^11,14,16-19 The MAPE was 17.1%. The predicted apparent oral clearance of ariprazole was closest to the observed value (observed vs predicted, 3.44 vs 3.46 L/h). There were no statistically significant differences in the MAPE between hepatically and renally eliminated drugs (18.2% vs 15.6%; P > .05). A positive correlation was found between allometry-predicted and observed apparent adolescent clearance values for oral products (R² = 0.97; y = 1.02x − 4.2) (Figure 2). Observed apparent oral clearance values in adolescents averaged 95.1% of adult values.

A high degree of congruence was observed between approved adolescent and adult drug doses. Of the 92 products with concordant indications between adolescents and adults, only 5 had different dosing in adolescent patients. These 5 products have indications for hepatitis B (ribavirin-peginterferon alfa-2a), schizophrenia (paliperidone), atopic dermatitis (fluocinolone acetonide), and hypertension (candesartan and valsartan). The dosing for ribavirin-peginterferon alfa-2a is 180 µg/wk in adults vs 180 µg/1.73 m² × BSA/wk in pediatric patients. Most adolescent patients have a BSA of less than 1.73 m² and would receive less than the adult dose. The dosing for paliperidone stipulates an adult dose range of 3 to 12 mg/d vs 3 to 6 mg/d in pediatric patients weighing less than 51 kg and 3 to 12 mg/d in those weighing at least 51 kg. Fluocinolone acetonide, which is indicated for the treatment of atopic dermatitis, is applied topically 3 times daily for adults vs twice daily for pediatric patients. However, no specific adolescent information is provided, and pediatric dosing encompasses all pediatric patients older than 2 years. Finally, candesartan and valsartan, both of which are indicated for the treatment of hypertension in patients 6 to 16 years of age, have lower recommended starting dose ranges in pediatric patients than in adults. This is in contrast to olmesartan, another angiotensin II receptor blocker included in this review, which has equivalent dosing for adult and pediatric patients who have hypertension and weigh at least 35 kg.

Drug clearance is a key determinant of dosing requirements. In this study, allometric scaling predicted adolescent clearance with an overall MAPE of 17.0%. Moreover, actual adolescent clearance averaged 93.2% of adult values, yet adult dosing was acceptable for adolescent patients for 87 of 92 products reviewed. This discrepancy suggests that increased drug exposure due to lower clearance in the adolescent population was not associated with an unfavorable adverse event profile. However, this assumption may not be true in all cases; 28 drugs reviewed had no adolescent indication, and it is unknown whether safety was a factor for these products.

Previous assessments of pediatric pharmacokinetics have examined differences between adolescents and adults. Oncology is a therapeutic area in which several pharmacokinetic studies have been conducted in adolescent patients outside the drug development process. In a 2010 review from St Jude Children’s Research Hospital,²⁰ drug dispositional changes in pediatric patients were reviewed in association with a number of important examples of oncologic agents. Although drug disposition is different between adolescent and younger pediatric patients, no evidence was presented to suggest any difference between adults and adolescents in the disposition of oncologic agents.

The results of the present pharmacokinetic analysis are consistent with existing information on drug elimination in adolescents but did not include all aspects of drug metabolism. For example, cytochrome P450 1A2 (CYP1A2) is a phase I drug-metabolizing enzyme that could be altered during adolescence, but the current study did not contain any CYP1A2 substrates. Findings of a 1982 study²¹ suggested that theophylline, which is metabolized by CYP1A2, has a sex-based difference in clearance, with women having a 22% lower clearance rate at age 20 years. A 1986 study²² found that CYP1A2-mediated demethylation of caffeine correlates to the stages of sexual maturity in adolescent boys and girls, although no statistical differences in the caffeine breath test were found between Tanner stages IV and V in comparison with adult values. These observations may be due to the inhibition of CYP1A2 by estrogen, as suggested by findings in studies of the inhibitory effects of oral contraceptive agents on the clearance of theophylline.²³ Therefore, changes in estrogen associated with puberty could potentially cause a mild to moderate effect on CYP1A2 expression and activity. For the drugs in this review that are hepatically metabolized, metabolic pathways including the cytochromes P450 3A4, 2D6, 2C19, 2C9, and 2E1 are involved. Because no CYP1A2 substrates are included, the utility of allometric scaling for the prediction of CYP1A2-mediated drug clearance in adolescents is unknown.

Allometric scaling has clear drawbacks, as discussed by Björkman in 2006²⁴ and Johnson in 2008.²⁵ In the former study, allometric scaling failed to predict the clearance of hepatically metabolized drugs in neonates. In the latter, multiple scaling methods using body weight and BSA were used to predict doses in comparison with the British National Formulary for Children dose. Percentages of predicted doses within 50% of the British National Formulary for Children dose ranged from 27% to 97% for the various pediatric age groups, but the adolescents were the only subjects in whom allometric scaling methods predicted doses with 97% accuracy. The study by Björkman used both body weight to the 0.75 power to scale to metabolic rate and body weight to the 0.67 power to scale to BSA. As discussed elsewhere,²⁶ all scaling methods are unsatisfactory when applied linearly across the pediatric population, and a combination of approaches may be necessary to make accurate predictions, especially in the youngest children.

The current study has limitations. Clearance was the only pharmacokinetic parameter considered; however, volume of distribution is another important parameter that influences dosing requirements. Allometric scaling approaches have been applied for predicting the volume of distribution, which is usually linearly related to body weight (eg, an allometric exponent of 1.0).⁵ Next, mean data and not individual-level data were used to perform allometric scaling. Individual patient data may improve the accuracy of predictions because pediatric subjects aged 12 to 17 years may have a wide range of body weights. In addition, most adult pharmacokinetic data are from healthy volunteers, whereas all pediatric data are from patients with illnesses and conditions requiring treatment. Disease states, such as inflammatory conditions, may alter drug disposition secondarily to changes in the expression and/or activity of metabolic enzymes and drug transporters.²⁷ Furthermore, chronic disease may affect growth and puberty, and this analysis of adolescent pharmacokinetics did not factor in pubertal development. A single exponent of 0.75 was used for clearance prediction using allometric scaling. Other exponents may provide more accurate predictions.⁶ Finally, the sample of products reviewed was limited, thereby precluding the ability to make broad generalizations about all pediatric drug development programs.

In summary, a review of 126 products with at least 1 pediatric trial completed under FDAAA 2007 identified 92 products with adolescent indications concordant with adult indications. Of these 92 products, 87 (94.5%) have identical adolescent and adult dosing. In some instances, adolescent doses may be able to be derived from adult data without the need for a dedicated pharmacokinetic study. Confirmatory pharmacokinetic data may then be obtained through sparse sampling during pivotal efficacy and safety trials. However, owing to the complexities of drug development and the limitations of allometric scaling, some products may require a more extensive pharmacokinetic evaluation in the adolescent population. The approach should be chosen on a case-by-case basis. Considerations include the therapeutic index of the drug and the degree of intraindividual and interindividual variability in drug disposition observed in adults. This position is supported by a recent FDA advisory committee²⁸ and consistent with current European Medicines Agency guidelines, which state that pharmacokinetics in adolescent patients are often similar to those in adults.²⁹ Because new legislation contained in the FDA Safety and Innovation Act requires earlier submission of pediatric plans during the drug development process, modeling and simulation will continue to play a key role in optimizing the clinical pharmacology strategy for drug candidates while avoiding unnecessary pediatric trials.

Corresponding Author: Gilbert J. Burckart, PharmD, Associate Director, Office of Clinical Pharmacology, Office of Translational Sciences, Center for Drug Evaluation and Research, US Food and Drug Administration, White Oak Bldg 51, Room 3184, 10903 New Hampshire Ave, Silver Spring, MD 20903 (gilbert.burckart@fda.hhs.gov)

Accepted for Publication: February 10, 2013.

Published Online: August 5, 2013. doi:10.1001/jamapediatrics.2013.465.

Author Contributions: Drs Momper, Mulugeta, Green, Karesh, and Burckart had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: Momper, Mulugeta, Karesh, Burckart.

Acquisition of data: Momper, Mulugeta, Green.

Analysis and interpretation of data: All authors.

Drafting of the manuscript: Momper, Karesh, Burckart.

Critical revision of the manuscript for important intellectual content: All authors.

Statistical analysis: Momper.

Administrative, technical, and material support: Karesh, Sachs, Burckart.

Study supervision: Mulugeta, Sachs, Yao, Burckart.

Conflict of Interest Disclosures: None reported.

Disclaimer: The opinions expressed in this article are those of the authors and should not be interpreted to be the position of the US Food and Drug Administration.