Article Text

Download PDFPDF

Why are certain age bands used for children in paediatric studies of medicines?
  1. Kalle Hoppu1,2,
  2. Helena Fonseca3,4
  1. 1 New Children's Hospital, Helsinki University Hospital, Helsinki, Finland
  2. 2 Department of Clinical Pharmacology, University of Helsinki, Helsinki, Finland
  3. 3 Adolescent Outpatient Clinic, Department of Paediatrics, Hospital de Santa Maria, Lisbon, Portugal
  4. 4 Paediatric Committee (PDCO), European Medicines Agency, Amsterdam, The Netherlands
  1. Correspondence to Dr Kalle Hoppu, New Children's Hospital, Helsinki University Hospital, 00029 Helsinki, Finland; kalle.hoppu{at}fimnet.fi

Abstract

Rational prescribing of medicines requires evidence from clinical trials on efficacy, safety and the dose to be prescribed, based on clinical trials. Regulatory authorities assess these data and information is included in the approved summary of product characteristics. Regulatory guidelines on clinical investigation of medicinal products in the paediatric population generally propose that studies are done in defined age groups but advise that any classification of the paediatric population into age categories is to some extent arbitrary or that the age groups are intended only as a guide. The pharmaceutical companies tend to plan their studies using age groups the regulatory guidelines suggest, to avoid problems when applying for marketing authorisation. These age bands end up in the paediatric label, and consequently into national paediatric formularies. The age bands of the most commonly used age-subsets: neonates, infant/toddlers, children and adolescents, are more historical than based on physiology or normal development of children. Particularly problematic are the age bands for neonates and adolescents. The age of 12 years separating children from adolescents, and the upper limit of the adolescents set by the definition of paediatric age in healthcare, which varies according to the region, are particularly questionable. Modern pharmacometric methods (modelling and simulation) are being increasingly used in paediatric drug development and may allow assessment of growth and/or development as continuous covariables. Maybe time has come to reconsider the rational of the currently used age bands.

  • adolescent health
  • growth
  • neonatology
  • pharmacology

Data availability statement

Data sharing not applicable as no datasets generated and/or analysed for this study. No data are available.

Statistics from Altmetric.com

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Today, in the era of evidence-based medicine, rational prescribing of medicines requires, among other information, evidence from clinical trials on efficacy and safety. Most importantly for the clinician is to be sure about the dose to be prescribed. Optimally, this information is available in the summary of product characteristics of the medicine approved by regulatory authorities. In the recent past, still such information was weak or non-existing, and the therapeutic needs of the paediatric population had to be met by off-label use, provided sufficient information for appropriate, rational and safe prescribing was available.1 2

Brief regulatory history from the perspective of paediatric medicines

In the 19th century ‘medicinal products’ could be made by anyone and sold for whatever use the originator or the ‘snake oil salesman’ claimed.3 Since then, breakthroughs in life sciences, particularly in chemistry, physiology and pharmacology, laid a foundation for the modern drug research and development and made possible modern medicines regulation.

Following the 1936 sulfanilamide elixir disaster in the USA, where 107 children died of diethylene glycol poisoning, the US Congress in 1938 amended the Food and Drug Cosmetic Act to require safety of drugs and truthful labelling of a product’s composition. Toxicity studies and approval of a new drug application were required before a drug could be promoted and marketed.4 Still, no rigorous proof of efficacy was required. Without objective data the benefit/risk ratio was poorly defined and seldom mentioned when a drug was brought to market. The Harris-Kefauver amendments to the US Food and Drug and Cosmetic Act of 1962, after the thalidomide disaster, gave FDA a major regulatory role in protecting citizens from ineffective medications. New drugs had to be demonstrated to be safe and effective based on well-controlled, scientific clinical trials before they could be marketed. Safety had to be demonstrated by having a sufficiently large safety database so that the benefit-versus risk ratio can be determined.4 5

Paediatric tragedies served as the origin of these new legislations, yet the actions and efforts that followed benefitted primarily drug development for adults. No specific paediatric regulatory requirements or guidelines existed. As a result, children were rarely included in clinical trials of medicines introduced after 1962. When paediatric prescribing information was not generated for inclusion in the label, many of the new approved medicines released since 1962 carried an ‘orphaning’ clause like ‘safety and effectiveness of …. have not been established in paediatric patients’.5 6 Excuses for not studying medicines for infants and children included that it was unethical for children to participate in clinical studies, and concern about the costs and difficulties of conducting studies in children not able to consent for themselves.5

Emergence of guidelines on development of medicines for labelling for children

The paediatric community, particularly in the USA, made efforts to counter the claims that participation of children in clinical trials was unethical and to support studies of drugs in children. The American Academy of Pediatrics (AAP) in 1974 issued a report, commissioned by FDA, ‘General Guidelines for The Evaluation of Drugs to Be Approved for Use During Pregnancy and for Treatment of Infants And Children’, but it was not published or endorsed by the FDA at that time.7 Later, in 1977, the FDA adopted the AAP report as a guidance document, a recommendation not legally binding.8 Real progress with increase of studies of medicines in the paediatric population began first with the FDA Modernisation Act of 1997 and the subsequent legislation in Europe that in 2007 introduced requirements to study some new medicines in paediatric patients.3 9

Around the same time, in Europe, the European Medicines Agency (EMA) produced a guideline ‘Note for Guidance on Clinical Investigation of Medicinal Products in Children’ in 1997.10 It formed the basis for the International Conference of Harmonisation (ICH) guideline ICH E11 ‘Guideline on Clinical Investigation of Medicinal Products in the Paediatric Population’ adopted in 1999.11–13 These legislative and other initiatives have started a real change for better, at least in the USA and Europe.14–16 However, still not all medicines on the market for adults are authorised for children and off-label prescribing in children remains high.17 18

The right dose

Before methods for clinical trials of medicines evolved and performing clinical trials in children became ethically acceptable, ‘efficacy’, ‘safety’ and the right dose of a medicine were based on experience. Experience had shown that growth and development influenced use of medicines in children and size had to be considered for dosing. Children’s ‘sensitivity’ to medicines was also known to vary compared with adults and according to children’s age. However, ‘sensitivity’ was difficult to study before blood concentration of medicines could be measured in children.

Size and age, being relatively simple to measure, were used as proxy for growth and development. ‘Experience-based’ paediatric doses extrapolated from adult doses were listed, using age bands, in textbooks (box 1), formularies or ‘dosing tables’.19 20 Another approach was to extrapolate from adult dose using rules for calculation of the paediatric dose based on child’s weight, for example, Clark’s rule or age of an infant, for example, Fried’s rule named after Dr Kalman Fried (1914–1999). Some eminent paediatricians, like Arvo Ylppö, had their own approaches (table 1). Ylppö noted that paediatric doses given in literature were generally too small.

Box 1

Example of old age bands used for paediatric dosing instructions20

Classification of neonates by gestational age

  • Post-term neonate: >42 weeks of gestational age.

  • Term neonate: 37–42 weeks of gestational age.

  • Moderate to late preterm infant: 32 to <37 weeks gestational age.

  • Very preterm infant: 28 to <32 weeks gestational age.

  • Extremely preterm infant: 24 to <28 weeks gestational age.

  • Preterm infants at the border of viability: 22 to <24 weeks gestational age.

Classification of neonates by birth weight

  • LBW infants: <2500 g birth weight.

  • VLBW infants: <1500 g birth weight.

  • ELBW infants: <1000 g birth weight.

  • Preterm infants at the border of viability: <600 g.

Table 1

Historical proposal for dosing medicines for children31

Paediatric dosing is still commonly related to body size, relying on the assumption of a linear correlation between dose and size. In fact, when using doses per kg or in relation to body surface area (BSA), one implicitly assumes that fractioning of the dose will result in comparable drug levels; that is, concentrations change in a linear fashion with weight or BSA, respectively. However, as developmental changes are mostly nonlinear, this so-called empirical dosing can lead to overdosing or underdosing, especially in specific age groups such as neonates and (extremely) low-birth weight infants, thereby increasing the risk of toxicity or reduced efficacy. Consequently, while it may make some sense to use doses corrected for size in age bands, as a proxy for differences in physiological or organ function across the various subgroups of the paediatric population, it can only be considered appropriate when based on data on drug exposure and exposure–response relationship of a drug.21

The age bands commonly used for dosing come from guidelines mapping the road to regulatory approval of medicinal product for use in children (box 2). These guidelines generally advise that any classification of the paediatric population into age categories is to some extent arbitrary or that the age groups are intended as a guide, bearing in mind that individual growth and development will vary around the norm.10 13 Newer guidelines may suggest that, sometimes, it may be more appropriate to collect data over broad age ranges and examine the effect of age as a continuous covariant. For efficacy, different endpoints may be established for paediatric patients of different ages, and the age groups might not correspond to the ‘classical’ age subsets.13 One possible example illustrating this issue is a clinical trial involving anticonvulsants where academic performance/learning outcomes can be important measures of cognitive/emotional functioning in school-aged children, while not possible to evaluate in the same way in younger ages. Therefore, it would make sense to stratify the 2–11 years subset in such studies accordingly. Notwithstanding, an important point is whether cognition is being assessed for efficacy, in which case the use of age-related specific tools is crucial (eg, as for paediatric investigation plans (PIP) in Niemann-Pick, Fragile X syndrome or metachromatic leukodystrophy), or for safety. The pharmaceutical companies, understandably, prefer to plan their studies using age-groups the regulatory guidelines suggest, to avoid problems when applying for market authorisation, and these age bands end up in the paediatric label.

Box 2

Age bands in regulatory guidelines on clinical investigation of medicinal products in children

Classification of neonates by gestational age

  • Post-term neonate: >42 weeks of gestational age.

  • Term neonate: 37–42 weeks of gestational age.

  • Moderate to late preterm infant: 32 to <37 weeks gestational age.

  • Very preterm infant: 28 to <32 weeks gestational age.

  • Extremely preterm infant: 24 to <28 weeks gestational age.

  • Preterm infants at the border of viability: 22 to <24 weeks gestational age.

Classification of neonates by birth weight

  • LBW infants: <2500 g birth weight.

  • VLBW infants: <1500 g birth weight.

  • ELBW infants: <1000 g birth weight.

  • Preterm infants at the border of viability: <600 g.

Age bands in current guidelines for paediatric clinical trials

Principally, dosing recommendations in age bands and accounting for size, usually in mg/kg, can better adapt to the nonlinear developmental changes, provided the age bands are well defined. From the perspective of clinical trials aiming to provide data on efficacy, safety and the right dose, age bands should be defined to include a physiologically and developmentally homogenous group of children, to control variability. Less variability increases precision of the results and reduces the number of subjects needed to achieve statistical significance in comparisons, an ethical imperative in paediatric studies. Moreover, in most paediatric programmes, a staggered approach has been used aiming to protect children from unexpected adverse events, starting with the oldest adolescents. Nevertheless, this approach may delay the completion of paediatric studies with a potential for off-label use in the youngest subsets, in theory the most vulnerable ones.

Neonates (neonatal period)

Different terms are used to reflect maturation or clinical metrics of a neonate, such as birth weight (feasible to obtain in all clinical settings) and developmental age (optimally based on an assessment in early pregnancy and the neonate’s examination at birth). Developmental issues of the neonate are often related to gestational age whereas birth weight based classification is often used in relation to dosing.22

Defining the neonatal age band as time from birth up to 1 month is problematic, as it does not take into consideration developmental age.7 8 When the traditional upper end, 27 days after birth, was defined, survival of preterm neonates was low. With neonatal survival at earlier stages of gestation this definition has become increasingly inappropriate to thoroughly describe the neonatal population.

Subdivision into preterm new-born infants (born at <36 or at <37 weeks of gestation) and term new-born infants (age 0–27 days), as in ICH11, is better.12 22 23 The EMA Guideline on the Investigation of Medicinal Products in the Term and Preterm Neonate defines the Neonatal period as from birth up to and including 27 days in term neonates or in preterm neonates from birth up to a postmenstrual age (PMA) of 40 weeks and 27 days.22 Other classifications have also been proposed (box 3), but it is obvious that an age band based only on postnatal age (PNA), is neither simple nor optimal anymore. In the neonate, in addition to size (eg, body weight), maturation (gestational age) at birth and PNA are important determinants of variability of drug exposure and changes in pharmacokinetic (PK) parameters.

Box 3

Classification of neonates in: EMA/CPMP/PDCO. Guideline on the Investigation of Medicinal Products in the Term and Preterm Neonate, 200722

Classification of neonates by gestational age

  • Post-term neonate: >42 weeks of gestational age.

  • Term neonate: 37–42 weeks of gestational age.

  • Moderate to late preterm infant: 32 to <37 weeks gestational age.

  • Very preterm infant: 28 to <32 weeks gestational age.

  • Extremely preterm infant: 24 to <28 weeks gestational age.

  • Preterm infants at the border of viability: 22 to <24 weeks gestational age.

Classification of neonates by birth weight

  • LBW infants: <2500 g birth weight.

  • VLBW infants: <1500 g birth weight.

  • ELBW infants: <1000 g birth weight.

  • Preterm infants at the border of viability: <600 g.

Depending on the medicinal product concerned and the disease to be treated, stratification of the trial population might be appropriate or necessary. Frequently, stratification by term gestation is needed in clinical trials, as PK and pharmacodynamic (PD) properties differ between preterm and full-term neonates. The same applies to age and PMA. For instance, stratification regarding neonatal nephrogenesis should be before and after 34 weeks of PMA. However, PK and PK/PD data should be analysed for association with size-related covariates (age, weight and so on) as continuous covariates. Stratification of some analyses may be considered, in conjunction with measures to quantify the homogeneity of a treatment effect.22

The complexities of the neonatal period for classifications, clinical trials and dosing have been discussed in more detail elsewhere.22 23 It seems obvious that in the context of current neonatology, thinking of neonates as one group with an age band defined on the basis of calendar age (PNA) does not make much sense, neither for dosing recommendations nor for clinical trials.

Infants and toddlers

The age band of this group, age 28 days up to 2 years (23 months), is a period of continuing rapid total body growth, central nervous system (CNS) maturation and immune system development. There is often considerable interindividual variability in maturation. Hepatic and renal clearance pathways continue to mature rapidly. Growth changes the proportions of different tissues in relation to weight (eg, the ratio of liver volume to unit body weight). By 1–2 years of age, clearance of many drugs on a mg/kg basis may exceed adult values.

As discussed above, with the increased survival of preterm neonates the lower end of the Infants age band has become challenging. The use of PNA alone does not necessarily provide insight into organ function at an individual patient level. For example, extremely and very preterm neonates exhibit lower glomerular filtration rates (GFR) values at birth and a slower pattern of GFR development because the complete nephrogenesis is not achieved before 34 weeks of PMA. For a baby born at 28 weeks of gestation complete nephrogenesis is reached not until around 42 days of PNA, well in the Infants age band. For those born prematurely the Infant age band would be better defined to begin after PMA of 40 weeks and 27 days (see above).22 23

Children (age 2–11 years)

Children in this age group reach several important milestones of psychomotor development. Somatic growth and development proceed at a more or less predetermined rate. Specific strategies should be addressed in protocols to ascertain any effects of the medicinal product on growth and development. Most pathways of drug clearance (hepatic and renal) are mature, with clearance often exceeding adult values. Changes in clearance of a drug may be dependent on maturation of specific metabolic pathways. Guidelines indicate that in clinical trials stratification by age within this category is often unnecessary.13 This could be interpreted as suggesting that this age band is well defined and developmentally homogenous. However, a highly variable onset of puberty may occur in this age band. Puberty occurs earlier in girls, in whom it may occur at 9 years of age or even earlier. So, this age band includes both pre- and postpubertal adolescents. Physiologically, children in this age band but having already entered puberty would fit better in the Adolescent group.

Puberty is a period of dramatic physiological as well as psychological changes, which impact on growth and development and can affect the apparent activity of enzymes metabolising medicines.24 Dose requirements for some medicinal products on a mg/kg basis may decrease dramatically. A good example is what happens with theophylline, a substrate of CYP1A2, which has an apparent sex difference in pattern of ontogenic expression during puberty. In young children, theophylline plasma clearance generally exceeds adult values; however, once entering puberty, caffeine 3-demethylation in adolescent women appears to decline to adult levels at sexual maturity rating (Tanner stage) 2 as compared to men where it occurs at stages 4/5.25 26 In clinical trials, it would be appropriate to record Tanner stages of pubertal development or obtain biological markers of puberty, and examine data for any potential influence of pubertal changes or specifically assess the effect of puberty on a medicinal product by stratifying prepubertal and postpubertal paediatric patients, corresponding to a Tanner stage below or above 2. Many conditions are well known to change with the onset of puberty, but to what extent puberty may also affect PK/PD relationship has not been thoroughly investigated, so far.

Adolescents: 12–18 years

The definition of this age band is the most non-physiological of all age bands discussed. The problems related to onset of puberty are discussed above. Physiologically, children in this age band but without onset of puberty would fit better in the previous group.

The definition of adolescence as 10–19 years of age comes from the mid-20th century, when patterns of adolescent growth and the timing of role transitions were quite different from current patterns. An expanded and more inclusive definition of adolescence as 10–24 years of age (adolescents and young adults (AYA)) aligns more closely with contemporary patterns of adolescent growth and popular understandings of this life period.27

Adolescents progress through stages of pubertal development but this progression does not follow the same timelines for every adolescent. As there are distinct rhythms of biological maturation according to early, average or late maturity patterns, two adolescents of the same chronological age can be totally different one from the other.28 Moreover, body compartments change remarkably across adolescence. For example, the amount of adipose tissue increases at a great speed in women throughout puberty with body fat increasing from 15.7% (prepubertal) to 26.7% at Tanner stage 4. The opposite happens with men whose fat mass decreases from 14.3% (prepubertal) to 11.2% at Tanner stage 2, and a significant lean body mass increase due to an increment in the circulating androgens. Apart from gender differences, due to the great variation in the timing of puberty, there are also ‘same gender’ significant variations with resulting differences in physical maturity of similar-aged ‘same gender’ adolescents. These biological specificities have to be taken into account as they have important implications for understanding PK and PD.

As regards the adolescent brain, profound changes occur in brain connections and in signalling mechanisms during adolescence. A second cycle of overproduction of synapses takes place, just before puberty, followed by an extended pruning process. The prefrontal cortex is still quite immature throughout adolescence achieving its full maturity only around the age of 25. This region is responsible for organisational ability, strategic thinking and impulse control, which may explain why adolescents have so much difficulty in understanding the potential consequences of an exploratory behaviour. This late maturation of brain functioning may affect PDs and the safety profile of CNS active drugs.

As we all are aware of, adolescent patients may have delayed access to potentially effective therapies, as the initial paediatric trials for many drugs are often conducted many years later, after the drug has been approved in adults. As a result, in severe diseases (e.g, cancers, infectious diseases), there might occur off-label use and consequently an increased difficulty in the integration of adolescent patients into future paediatric trials. In the field of oncology and aiming to enable earlier access to investigational and approved drugs for adolescent patients with cancer, FDA recommends the inclusion of adolescent patients in disease-appropriate and target-appropriate adult oncology clinical trials.29 The recommendation is that adolescent patients should be eligible for enrolment in adult oncology clinical trials at all stages of drug development if their cancer is similar in histology and biological behaviour to those found in adults.

The upper limit of this age band was set by the definition of paediatric age in healthcare, which varies according to region. Older adolescents in some countries are hospitalised together with adults and not in paediatric wards or hospitals, which may pose logistical and administrative problems for clinical trials. Additionally, using the AYA definition, national legal age is reached before the end of the period, which impacts on legal issues (e.g, informed consent). Depending on factors such as the condition, the treatment and the study design, it may be justifiable to include paediatric adolescent subpopulations in adult studies or adult adolescent subpopulations in paediatric studies. Finally, given some of the unique challenges of adolescence and young adulthood, it may be appropriate to consider studying AYA patients (whether they are to be included in adult or separate protocols) in centres knowledgeable and skilled in the care of this specific population.13

Discussion

The paediatric age bands defined by regulatory guidelines and discussed above are not well based on physiological growth and development. The regulatory environment is known to be conservative. When the EMA working group developed the Note for Guidance on Clinical Investigation of Medicinal Products in Children in 1995, which became the foundation for ICH E11, the age bands were hotly debated, particularly the breakpoint of 12 years separating children from adolescents. In search of the rationale for the 12 years it was informally discussed with eminent pioneers of paediatric clinical pharmacology, unsuccessfully. Yaffe (1923–2011), called the father of Paediatric Clinical Pharmacology, informed that he did not know the rationale for the 12 years, ‘nor could he see any reason for it at the present time’ (Yaffe; Personal communication, 1997). The breakpoint of 12 years remained, only for historical reasons, and is still there.

The revised ICH E11 (R1) guideline proposes new approaches to optimise paediatric drug development.13 Extrapolation based on information from other populations (adults or paediatric subgroups) that integrate and leverage existing knowledge, and PK and PD studies in relevant child populations may allow assessment of growth and/or development as continuous covariables. Less weight is placed on the old age bands as chronological age alone may not serve as an adequate categorical determinant to define developmental subgroups in paediatric studies.13 21 In the last decade, innovative strategies such as modelling and simulation to support the acceptability/strengthen the appropriate use of extrapolation have led to agreed paediatric drug development programmes that do not require controlled trials to establish efficacy. HIV infection, the majority of infectious diseases and oncology are areas where paediatric extrapolation has been particularly used. However, the construct of an extrapolation concept to make predictions for efficacy even between different paediatric age subsets may not be possible in diseases where there are differences in terms of neurodevelopmental stages, including biopsychocognitive development that may impact on both efficacy and safety endpoints.30

As discussed previously, chronological age alone may not always be the most appropriate categorical determinant to define developmental subgroups in paediatric studies. There are many other factors to be considered in determining appropriate paediatric subsets, such as physiological development and organ maturation, drug target ontogeny, pathophysiology and the natural history of the disease. Optimally, the distinct ‘age groups’ for clinical trials should be chosen based on what is known about the prevalence and incidence of the disease, taking into account the role of developmental growth, maturation processes and ontogeny, all of which can affect PKs, PDs and the safety profile of a drug.21 The ICH E11(R1) admits that the arbitrary division of paediatric subgroups by chronological age for some conditions may have no scientific basis and could unnecessarily delay development of medicines for children by limiting the population under study. Maybe it would be time for regulatory authorities to reconsider the use of the ‘traditional’ age bands.

Data availability statement

Data sharing not applicable as no datasets generated and/or analysed for this study. No data are available.

Ethics statements

References

Footnotes

  • Contributors KH and HF conceptualised this review. KH drafted the manuscript which both authors contributed to. Both authors approved the final draft

  • Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.

  • Disclaimer The views expressed in this article are the personal views of the author and may not be understood or quoted as being made on behalf of or reflecting the position of the European Medicines Agency or one of its committees or working parties.

  • Competing interests None declared.

  • Provenance and peer review Commissioned; externally peer reviewed.