Clinical pharmacology in neonates and young infants: the benefit of a population-tailored approach

“Concerted efforts to improve research tools on pharmacokinetics and pharmacodynamics will probably result in improved drug therapy in infants, and is also a potent driver to learn more of the intriguing world of growth and development of infancy.”

An infant is not just a small child

The most essential characteristics of childhood are growth and maturation. Both phenomena are most prominent during infancy, making the claim that ‘an infant is not just a small child’ as relevant as the more commonly used paradigm that ‘a child is not just a small adult’. There is already one log size difference in weight (0.5–5 kg) within the neonatal population, quite similar to the log size spectrum (5–50 kg) of childhood [1,2]. The birth weight increases by 50% in the first 6 weeks of life, and doubles in the first 4 months to be three-times higher at the end of infancy. In this same time interval, there is a fourfold increase in caloric needs. As a consequence, the first year of human life is characterized by a very dynamic biological system where growth, maturation and extensive variability are the key issues [1–3].

From a clinical pharmacology perspective, the consequence of such a dynamic setting is extensive interindividual variability throughout infancy in both the pharmacokinetics and pharmaco dynamics of xenobiotics [3,4]. Instead of median values for pharmacokinetic estimates or outcome variables, the range and its contributing covariates are crucial. Body composition and compartment sizes change during infancy, and all phase I (e.g., cytochromes) and phase II (e.g., glucuronidation) metabolic processes mature in an iso-enzyme-specific pattern, while renal function (glomerular filtration rate and tubular absorption/excretion) also displays age-dependent capacity [1,5]. The phenotypic variability in either drug disposition or effects during infancy is further affected by the contribution of other, non-ontogeny-related covariates (e.g., perinatal asphyxia with whole body cooling, comedication, genetic polymorphisms, sepsis or inflammation, and congenital renal impairment) [3,4,6–9].

“…the first year of human life is characterized by a very dynamic biological system where growth, maturation and extensive variability are the key issues.”

History provides the community with compound-specific observations to illustrate the negative impact of exposure to chloramphenicol (deficient glucuronidation capacity resulting in chloramphenicol accumulation and gray baby syndrome), benzyl alcohol (deficient alcohol dehydrogenase activity resulting in benzyl alcohol accumulation and gasping syndrome) or – much more recently – dexamethasone (specific vulnerability of neonatal cortical and subcortical nervous tissues, resulting in cerebral palsy and blunted brain growth) in neonates [1]. All these anecdotic observations can be considered as illustrations of failure to consider the specific characteristics of this vulnerable population [2,3].

“The phenotypic variability in either drug disposition or effects during infancy is further affected by the contribution of other, non-ontogeny-related covariates…”

However, optimism is a moral duty. Implementation of the pediatric regulation in the USA in 1997 has resulted in a rebirth of pediatric product development, including drugs for infants. Subsequent implementation of a similar pediatric legislation initiative in the EU (2007) and in WHO initiatives (e.g., the WHO ‘make drugs child size’ program, which is an ongoing program of the WHO and resulted in an essential medicines list for children in 2011) further stimulated all stakeholders (industry, academia, governmental and research organizations) to develop focused pediatric research activities. As an illustration, a search for ‘newborn/infant’ on the clinicaltrials.gov website in early October 2011 resulted in 1318 protocols for interventional studies, of which 899 were initiated (based on the sponsor’s location) in the USA and 376 in Europe. Although this is only a snapshot of the ongoing clinical research activities, it suggests that there is growing research activity aiming to further improve pharmacotherapy [4]. In addition to compound-specific observations, such studies also stimulated the development and validation of research tools in the field of pharmacokinetics (e.g., analytical techniques and population pharmacokinetics) and pharmacodynamics (population-specific ‘biomarkers’) adapted to newborns and infants [3,4].

Tailored pharmacokinetic studies in infancy: improved sampling strategies & data interpretation

Besides maturational aspects in absorption and distribution, maturation mainly relates to rena elimination and hepatic metabolic clearance. Renal elimination clearance in early life is low and almost completely reflects the glomerular filtration rate [1,2]. Despite the overall low drug clearance, interindividual variability is extensive and can be predicted by covariates including, for example, postmenstrual age, postnatal age, coadministration of a nonselective cyclooxygenase inhibitor, growth restriction or perinatal asphyxia [1–8]. Variation in phenotypic metabolic clearance is based on constitutional, environmental and genetic characteristics. In early life, it mainly reflects ontogeny, but other covariates (e.g., disease characteristics, genetic polymorphisms and treatment modalities) may also become relevant.

The feasibility of performing pharmacokinetic studies in neonates and infants improved with the use of adapted sampling techniques (e.g., saliva samples or dried blood spots) and more accurate quantification of metabolites in low-volume samples. Because of the capability to analyze sparse, unbalanced datasets, the burden for each individual infant can be minimized, and modeling and simulation using non-linear mixed-effect modeling became the preferred tool to develop effective and safe dosing regimens [3,4,9]. Moreover, such models can also be used to design the studies and sampling strategies to obtain maximal information with a limited burden for all individual infants included. The maturational patterns described and the extent of the impact of covariates can be used to predict in vivo time–concentration profiles for compounds that undergo similar routes of elimination.

However, following initial development, the clinical research community still fails to a certain extent to validate such models in the clinical setting. Besides internal and external validation, prospective clinical trials, which allow the evaluation of the model-based dosing regimens, are needed, not only to adjust the proposed dosing regimen, but also to convince pediatricians to use the information that has been generated using these modeling exercises [1,3,4]. Finally, we have to be aware that improved knowledge on developmental pharmacokinetics during infancy is, in the majority of drugs, only a first step to describe the impact of maturation on the concentration–effect relationship.

Tailored pharmacodynamic studies in infancy: the search for valid biomarkers

Similar to the expansion of biomarker research in adult human medicine, there is an active search for robust biomarkers to evaluate the effects and side effects of interventions in neonates and infants [10–12]. In essence, a biomarker is a characteristic or quantitative indicator that reflects either normal biological processes, pathological processes or pharmacological responses. Since infants are not simply small adults with regard to their (patho)physiology, disease expression or responses to therapy, the introduction of biomarkers in clinical pharmacology during infancy needs validation in this specific population [4,10,11].

“…improved knowledge on developmental pharmacokinetics during infancy is, in the majority of drugs, only a first step to describe the impact of maturation on the concentration–effect relationship.”

For most of the drugs or interventions with potential indications in early neonatal life, there is the perception that a safety assessment should also include long-term neurodevelopmental outcome data. Although assessment based on Bayley scales or similar does provide quantitative results, we have to be aware that early (first 12–18 months) developmental assessment is only a modest (in terms of sensitivity and specificity) predictor for late neurodevelopmental outcome. This outcome is further biased by social factors and does not seem to remain stable over time. The introduction of more advanced neuroimaging techniques or clinical assessment tools that focus more on executive functions may be promising approaches, but will first need prospective validation studies [13].

Another evolving field of perinatal research that will probably benefit from these advanced neuroimaging techniques, clinical assessment tools or biochemical biomarkers is perinatal asphyxia. Since the meta-analytic evidence in support of whole-body cooling (number needed to treat: 8–10) to maintain normal neurodevelopment, this modality became the standard of care. However, it took 20 years and 638 patients recruited in eight randomized controlled trials to prove this, while still a relevant proportion of neonates display cerebral palsy (26%) or severe neurodevelopmental disability or death (48%) following whole-body cooling [6]. There are different add-on treatments or interventions (e.g., xenon, more aggressive treatment of nonclinical convulsions, argon and melatonin) that have the potential to further improve this outcome. Small focused studies combined with the use of such early available biomarkers are first needed to ensure that both the number of cases and the study time needed for subsequent large-scale pivotal studies remains feasible [12].

“…there is an active search for robust biomarkers to evaluate the effects and side effects of interventions in neonates and infants.”

As a final illustration, we would like to refer to the recent introduction of plasma biomarkers for clinically relevant, symptomatic patent ductus arteriosus (PDA) or renal impairment/toxicity [14,15]. It is hereby striking that there is extensive variability in cutoff values between the different studies reported. To illustrate this, the cutoff values of N-terminal pro-brain natriuretic peptide to predict a clinically significant PDA varies between 1203 and 5000 pmol/l [14]. We are unsure as to whether this extensive difference in cutoff values merely reflects differences in clinical characteristics. Meticulous evaluation on the impact of sampling, sample handling or techniques used and validation of the quantification methods proposed (e.g., intraday and interday variability, stability and repeatability) is needed before PDA biomarkers can be introduced into routine clinical care [14]. At least, serum creatinine showed that the analytical technique used (i.e., Jaffe vs enzymatic) also displays population-specific differences, resulting in the inability to use a fixed value to convert creatinine values between both analytic techniques during infancy [15].

Conclusion

Since the implementation of the pediatric drug regulation in the USA and the subsequent initiatives throughout the world, there is an increase in ongoing research in the field of pediatric product development. This increase does not only result in more challenges in how to perform these studies, but also creates opportunities for population-tailored approaches in both pediatric product development and clinical research. Concerted efforts to improve research tools on pharmacokinetics and pharmacodynamics will probably result in improved drug therapy in infants, and is also a potent driver to learn more of the intriguing world of growth and development of infancy.

Biographies

John van den Anker

Karel Allegaert