Why Aren't More Pediatric Trials Performed?

July 1, 2002

Applied Clinical Trials

Applied Clinical Trials, Applied Clinical Trials-07-01-2002,

Successful pediatric research requires partnership between the pharmaceutical industry, investigators, ethics committees, and parents.

Children and adolescents represent about 25% of the European population. Yet there is a frightening lack of scientific data on the use of medicines in children, which has now reached the notice of national governments.1,2 Studies throughout Europe have shown that children receive medication not licensed for their use, or receive it at different doses, for different indications, or by different routes from those recommended.3 Because of the failure to recognize their needs and to perform appropriate research, children are denied access to safe and effective treatment that adults would demand as a fundamental right.

Children have become therapeutic orphans, which is ironic, because the regulations about drug licensing were in many cases precipitated by accidents in children. In the United Kingdom, for example, laws were introduced after cases of drug toxicity in the developing fetus (thalidomide) and newborn infant (chloramphenicol-induced "grey baby" syndrome).4 In the United States, the Food and Drug Administration (FDA) was established in 1938 by the U.S. Food, Drug and Cosmetic Act after more than 100 children died following administration of a new antibiotic. It was not the sulfa drug itself that caused the deaths, but the solvent, diethylene glycol, used to prepare the pediatric formulation, which caused acute renal failure due to precipitation in the kidneys.5

Licensing procedures worldwide demand that drugs for human use be effective, safe, and of high quality. But children are being treated daily with drugs for which those characteristics cannot be guaranteed. This has major implications both for their health and well-being and for drug manufacturers facing new proposals to limit their marketing of drugs untested in children. The desire to protect individual children from exposure to research risks may have the effect of harming children as a class by inhibiting research into pediatric health and disease.

So why are children not more widely used in research into drugs and other treatments? I will discuss the ethical, legal, and practical obstacles and what it will take to surmount them.

Why is research in children needed?
Children differ in many respects from adults. First, their diseases are different; there is no analogy to neonatal diseases in adults, and illnesses affect children and adults differently. For example, the mild infection of rubella in children is a major one in adults. Second, the effect of drugs can be different; phenobarbitone acts as a sedative in adults but can cause hyperactivity in children, and the association between aspirin and Reye's syndrome in children is not seen in adults. Third, children may not tolerate some treatments because of the taste or method of administration.

Despite what we know about these differences, physicians prescribe medicines whose actions are poorly understood in children. In a survey conducted in five European countries, 67% of children received drugs prescribed in an unlicensed or off-label manner, and 39% of the prescriptions were off-label.3 The terms off-label and unlicensed mean that a drug is used in situations not covered by the manufacturer's product license or summary of product characteristics. This can mean that it is used at a different dose or frequency, in different clinical indications, in different age groups, by another route, or in a formulation not approved for children. Analgesics and bronchodilators were the drugs most frequently used off-label. Because most unlicensed drug use involves crushing adult tablets to form a suspension, the stability data for the tablets then become meaningless. The scale of off-label and off-license prescribing is unknown, but a retrospective analysis of prescriptions for children under 12 years old revealed that 20% of antibiotics (perhaps the most common medication given to children) were given at off-label doses, and one survey showed that 90% of drugs used in babies and neonates were off-label.6

Off-label and off-license prescribing is not illegal. It is outside the terms of the product license, so the producing company has no product liability, but it can be done on the liability of the prescribing doctor. This is potentially dangerous because it increases the risk of adverse events7 and may be associated with lack of efficacy, but no cases of legal action being taken for harm so caused have been published.

Children should have access to medicines with scientifically justified prescribing information that matches their needs. The vast majority of such information in use at present comes by extrapolation from data obtained from adults, an inherently unsafe practice. It can be difficult for a pediatrician to tell parents that a new treatment is unproven but potentially valuable, gain consent for a study, then explain that the child has been randomized to standard therapy, and yet still maintain the parents' trust.

Regulations for children and research
Safe and effective drug treatment for children requires age-appropriate formulations and trial data obtained from children across the age range. For many drugs already marketed, this would entail significant additional research, which would be difficult to justify and to design. The International Conference on Harmonisation (ICH) has produced guidelines for studies,8 and the European Agency for the Evaluation of Medicinal Products (EMEA) has published guidance on their use.9 These documents emphasize the need to obtain knowledge of the effects of medicines in pediatric patients without compromising their well-being. Pharmaceutical companies, regulatory authorities, health professionals, and society as a whole share responsibility for this effort. As a result it is expectedcertainly by the UK authoritiesthat pediatric development will be done for all new chemical entities where use in children is likely.10 In diseases seen only in children, the full development program should now be carried out in children, and for serious or life-threatening diseases found in both adults and children, pediatric development can be started after initial proof of efficacy and safety in adults. For other diseases, trials in children should be started only after obtaining a large body of evidence from adult-only studies.

The European Clinical Trials Directive11 also encourages research in children as follows:
Children represent a vulnerable population with developmental, physiological and psychological differences from adults, which make age- and development-related research important for their benefit. Medicinal products, including vaccines, for children need to be tested scientifically before widespread use. This can only be achieved by ensuring that medicinal products which are likely to be of significant value for children are fully studied.

But will this advice become reality? The evidence from the United States is that it will not unless penalties or inducements are offered to pharmaceutical companies. In 1994, FDA simplified the type of information needed to demonstrate drug safety and efficacy in children to encourage drug manufacturers to submit pediatric data voluntarily. The response was minimal; whereas in 1994 40% of drugs likely to be used in children had child-specific labeling, by 1997 it was only 33%.12 Most manufacturers went back to saying that safety and effectiveness had not been established for children.13 In 1997 the Food and Drug Administration Modernization Act (FDAMA) allowed the FDA to request labeling information on safe pediatric use for a new drug that would be commonly used in children. It also allowed them to request pediatric testing of drugs already on the market in certain circumstances where lack of data could pose significant risks. The Final Rule of 1998 (effective December 2000) was mandatory where a drug or biological agent offered a "meaningful therapeutic benefit" or there would be "substantial use"14 in children "and the absence of labeling poses a risk." It allowed the FDA to "require" studies to be submitted; if data were not provided, the sponsor company could be held in contempt of Congress for failure to comply with the requirement. This could result in criminal penalties against top executives and very large daily fines. (See box, Can FDA Require Pediatric Studies?)

On 31 December 2001, FDAMA expired, and it was replaced by the Best Pharmaceuticals in Children Act (BPCA), signed into law by President Bush two days later. BPCA also grants six months' exclusivity on a manufacturer's marketing license if the company "fairly" responds to the FDA's written request for pediatric studies. It also adds neonates to the requirement, and provides for funding studies in drugs that are no longer covered by marketing exclusivity. The National Institute of Child Health and Development is the authorized grantor for such studies. It is too soon to know if this new rule will be more effective in gaining information for safe pediatric prescribing, but indications are that more clinical studies are being conducted.

Given that this research should be done, how can the researchers who carry out pediatric studies protect their subjects? There are issues of legal consent, ethics, risk/benefit balance, and practicalities.

Issues of consent
In most countries, children are the only people on whose behalf other individuals can give consent to medical procedures. (In other countries, adults unable to give consent can have that consent given for them by a relative or legal representative.) Consent given for a child is no different from adult consent; to be legally valid it must be freely given and informed, and in research terms implies accepting whichever treatment the child will receive in a randomized study.

While consent to treatment is usually seen to be in the child's best interest, research may not be, because research generally benefits a population rather than the individual. The British Medical Association (BMA) advises that
It is now widely accepted that from an ethical perspective research can be carried out on children, when there is no expected benefit for them individually, provided there is minimum risk, strict safeguards and no objection from either the child or parents.15

Those below the legal age of adulthood can consent to research if they can show that they understand medical advice and the pros and cons of receiving or refusing treatment. If a child of any age did not wish to receive a potentially life-saving experimental treatment, his or her refusal might well be overridden by parental consent, as the level for capacity needed could be claimed to be beyond that of a child. The guidance issued by the EMEA advises that this should be so.16 Conversely, hazardous experimentation authorized only by a child might actually be unlawful.

Therefore, in most cases of research using child subjects, parental consent is necessary. Parental consent to something that might harm the child is not valid in law, so the fact that it has been given does not, of itself, justify the research. But guidelines from the Royal College of Physicians clarify the matter as follows:
The duty of the parent to act in the best interests of the child can reasonably be interpreted as an obligation not to do anything clearly against the interests of the child.17

Most commentators advocate also discussing the study with the child and obtaining the child's assent or acquiescence. A reasoned refusal might well be taken as evidence of competence and so respected, but even children unable to reason should be given the opportunity to object. A child's ability to consent will develop as he or she learns to make increasingly difficult and serious decisions, and during the course of a long study there may be a need to reexplain to a child and obtain assent on several occasions. Ability to assent may depend not just on age; many children with chronic disease and much experience of hospitals are more mature than their years.

Issues of ethics
Children are perhaps the most vulnerable patient group, because they have fewer rights, may not be able to express their fears and needs, are easily bewildered, and may experience the most lasting benefits or harm from research. Recent guidelines on the conduct of medical research on children18 establish the following six principles of ethics:

  • Research on children is important for the benefit of all children.
  • Children have unique interests and are not small adults.
  • Research should be done on children only if the question cannot be answered by studying adults.
  • Research that will not directly benefit the child subject is not necessarily unethical or illegal.
  • All proposals should be submitted to a research ethics committee.
  • Legally valid consent must be obtained, with additional agreement from school-age children if consent is by parents.

Medical research on children is important to promote child health and minimize safety risks from the use of untested drugs, but it is ethical only if there is a real prospect of benefit to children as a whole and the research is well designed and performed. This latter point excludes studies repeating known work or with insufficient statistical power to be meaningful. But a gray area exists where the previous work has been done in adults. If a drug has been shown to be efficacious in adults, some argue that equipoise no longer exists, and that to deny children an active drug is not ethical. Balanced against that is the possibility that the immature liver and enzymes may handle drugs very differently and so give a different effect. But if there is collective and genuine uncertainty about the best treatment among informed clinicians, then I believe that it is ethical to perform the research.

Research ethics committees have a crucial role: their main function is to protect the patients, and child patients need maximal protection. Each ethics committee should include someone with a close practical knowledge and experience of sick children. Each study should ask an important question, the risks should be acceptable, and the research subject's autonomy should be respected. Protocols must be assessed for reasonableness and workability, and particularly for the monitoring of a child's wish to withdraw.

There have been suggestions that the role of the ethics committee should be strengthened. Parents could be told that research projects were under way at the hospital, and they would not be required to give consent, but would be told if their child was about to be included in a study. They would be able to refuse entry, but under certain safeguards the researchers and the ethics committee would give assent.19 This would preserve the parents' autonomy but would remove the risk of them feeling pressured at a time of stress and vulnerability.

Issues of risk and benefit
In research a few subjects may face risk to benefit many, but when we seek to protect individual children from exposure to research risks, children as a class may be harmed due to lack of research into disease and the effect of drugs.

To assess the potential benefit, one must consider the size and frequency of the problem and its severity. Is there truly a question that needs an answer, and if so, how likely is it that the answer will be obtained from a particular study? Will benefit be to one or many children? Potential harm, both physical and psychological, must be balanced against this. The latter is more difficult to assess because children vary, but researchers often underestimate the stress of short-term pain (such as blood sampling) in younger children.

Risk in clinical research can be estimated as minimal, low, or high. Minimal risk is seen in studies where questioning, observing, or measuring forms the basis of the research. There is still a need for informed consent and for sensitivity, but the child is unlikely to be physically discomforted. Low-risk studies would include those where blood samples are taken, with brief pain and perhaps a little bruising. High-risk procedures include tissue biopsy and cardiac catheterization, and are not justified for research purposes alone, but only when research is done as part of treatment. All research in children, irrespective of risk, requires recognition of the need for psychological and social support for the children taking part, because their needs may differ from those of adults.

But the risks of using drugs untested in children will always have greater potential for harm, given the figures for off-label and off-license use. When clinicians are exhorted to practice evidence-based care, the lack of evidence is at best confusing and at worst dangerous, yet neither doctors nor parents question the lack of pediatric dose information on commonly used drugs. The only area where children are well-served by drug trials is oncology, where, for example, most children with leukemia in the United Kingdom have been entered into Medical Research Council trials for many years.20

Issues of practicality
One of the difficulties in assessing benefit is the relative rarity in children of some diseases common in adults. It can be hard to find sufficient numbers of children with the same condition, and statistical significance may not be reached in a trial. One could imagine pressure being placed on children with a chronic disease to enter a series of studies, simply because the target population is small, yet much research is needed. Such pressure would be unethical and unfair to the children.

Another complication is the heterogeneity of children's disease. A childís response to illness varies with age; for example, a baby may be far more ill with a particular infection than an older child. This too can render statistical analysis difficult and perhaps inconclusive.

Surrogate markers are often used in research, and although there is controversy over their reliability, their relevance has been established only in adults. Their value is unknown for children, and because diseases run different courses in children, it may be incorrect to use the same markers as for adults. Similarly, quality-of-life questionnaires to assess the impact of a disease or treatment on various aspects of a patient's life require validation before use, yet this is done only in adults, not children.21 It is self-evident that a child will be affected differently from an adult with the same disease, yet for toddlers and babies it is impossible to assess even pain other than by careful observation by a researcher, although observer rating scales exist.22

A point often forgotten by authors of protocols is the small blood volume of children. A 600-g baby has only 50 mL of blood: those used to copious sampling from adults may inadvertently threaten the well-being of a child, because not more than 5% of the circulating volume can be taken without harm.

Parental involvement
I believe that most parents given a choice of medication would prefer one that had been tested in children, so in principle they would support the concept of children participating in research programs. But when asked to give consent for their child to take part in research, they may be worried about his or her well-being or feel pressured to decide quickly. They may have concerns about the consequences for their child's treatment if they refuse consent, so their consent would not truly be given without "coercion, undue influence or inducement, or intimidation."23

In two studies involving babies, 71% of parents who were asked for consent early agreed, but only 43% of those asked later agreed.24 The authors of this study interpreted this to mean that longer reflection produces less agreement. Another report showed problems with competence and understanding in over 70% of parents, three times as often in an emergency study as in a more routine one.25 More than 96% of these parents believed it was right for them to give consent, but 64% believed their own child would benefit, despite randomization, and over 39% foresaw no risk or distress to their baby. This makes one question their understanding of what they had been told. In two other studies, when consent was sought by nurses, 30% of parents refused, compared with only 6.5% when consent was sought by doctors, suggesting that parents feel unable to refuse a doctor's request.26

In a study of ventilation techniques in neonates at the North Staffordshire Hospital Trust,27 parental consent was sought when the infants were between two and four hours old. These babies were preterm and very sick, and some parents later claimed they were not told that their children would receive an experimental treatment. There is evidence that consent was sought in some cases,28 but one must question the capacity of parents to make rational choices at such a time, and therefore the validity of any consent given.

In another study, parents who refused consent were better educated and more likely to have professional or administrative jobs than those who consented, suggesting that the more vulnerable parents consent.29 Because the essence of informed consent is that it is freely given, without pressure or inducement, and after as much discussion as needed by those consenting, all these results are a source of concern.

Perhaps this varying level of understanding supports the suggestion that ethics committees should take a bigger role in consent, leaving the parents only the right of refusal. But ethics committees do not always challenge protocols as they should; in the North Staffordshire study, approval was given for using outcome scores that were not validated.

Room for improvement
Children are unique as a research group in that others give consent for them to take part. They have perhaps the most to gain and the most to risk, because any medical intervention could have long-lasting benefit or harm. Clinical research is an important means of promoting their well-being and health, particularly in ensuring they have effective and safe medication, and yet it is not done sufficiently often, despite the encouragement of ICH. The reasons for this are varied and include avoidance of harm, fear of litigation, and concerns about ethics. But the arguments for research in children being done responsibly are overwhelming, and there is good evidence from cancer studies that it can be done effectively and safely even in children with life-threatening illness.

First and foremost, research should not be done in children if the results can be obtained from studying adults, but this is often not possible given the physiological differences. So, because research must be done in children, special consideration should be given to it. The research must be for their benefit, not for the financial or academic gain of the researcher; it must be balanced and well designed; and it must not duplicate previous work.

Second, issues of consent must be improved. Some investigators seem to want to deny the rights of parents to consent, and evidence exists to show that consent is often not properly sought and the requirements that it be informed and voluntary are overlooked. In the Euricon study described earlier,25 44% of the doctors thought that parents should not be told all the risks to avoid alarming them. This is not surprising, because this study also showed that less than 5% of doctors had formal training in the issues and legality of informed consent.

Third are issues of practicality. Because of their size, children have small blood volumes, so certain non-therapeutic adult studies such as pharmacokinetic studies with frequent blood sampling cannot be done, but there are alternatives. For example, by taking one sample from many children, population kinetics can be derived, which fulfill the requirements of understanding the metabolism of a drug and its effects on the body. If we continue to use drugs that have not been tested on children, we will inevitably see more cases like that of propofol. Propofol is a short-acting intravenous anesthetic, not licensed in children but used in intensive care units for their sedation. Many children who were given propofol developed metabolic acidosis, and 15 died.30 Regulatory agencies issued warning letters against its use in children, and it is still not licensed for them. But perhaps this tragedy could have been avoided by testing it on children in closely monitored clinical trials.

For research in children to succeed, there must be partnership between the pharmaceutical industry (which funded 43% of randomized, controlled trials published in one pediatric journal over 15 years),31 researchers, ethics committees, and parents. It is not acceptable to put the lives of more children at risk by not having an evidence base for their use of medicines.


1. House of Commons Health Committee, The Specific Health Needs of Children and Young People (House of Commons, London, 1997).

2. Food and Drug Administration Modernization Act, Section 505A, Pediatric Exclusivity (FDA, Rockville, MD, 1997). The EU is also considering regulatory changes for the licensing of drugs to be used in children.

3. S. Conroy, I. Choonara, P. Impicciatore, et al., "Survey of Unlicensed and Off-label Drug use in Paediatric Wards in European Countries," British Medical Journal, 320: 79-82 (2000).

4. I. Choonara and J. Dunne, "Licensing of Medicines," Archives of Disease in Childhood, 78: 402-403 (1998).

5. R.R. Abood and D.B. Brushwood, Pharmacy Practices and the Law. (Aspen Publishers, Inc., Gaithersburg, Maryland, 1994).

6. H.S. Boardman, "Pediatric Drug Development: Principles and Practice," International Journal of Pharmaceutical Medicine 15: 35-47 (2001), p. 36, quoting S. Conroy.

7. S. Turner, A.J. Nunn, K. Fielding, and I. Choonara, "Adverse Drug Reactions to Unlicensed and Off-label Drugs on Paediatric Wards: A Prospective Study," Acta Paediatrica 88: 965&-968 (1999).

8. International Conference on Harmonisation of Technical Requirements for Registration of Pharmaceuticals for Human Use. ICH Topic E 11.

9. Note for Guidance on Clinical Investigation of Medicinal Products in the Paediatric Population (CPMP/ICH/2711/99).

10. H.S. Boardman, "Pediatric Drug Development: Principles and Practice," International Journal of Pharmaceutical Medicine 15: 35-47 (2001), p. 36, quoting J. Dunne of the U.K. Medicines Control Agency.

11. European Union Clinical Trials Directive 2001/20/EC, Article 3.

12. D. Murphy, "1999 Pediatric Initiatives," www.fda.gov/cder/present/dia-699/mu1-dia99/.

13. T. Nordenberg, "Pediatric Drug Studies: Protecting Pint-sized Patients," FDA Consumer Magazine 33 (3) (May-June 1999), www.fda.gov/fdac/features/1999/399_kids.html, quoting R. Roberts.

14. Substantial use is classified as more than 50,000 patients for the labeled indication.

15. British Medical Association, Consent, Rights and Choices in Health Care for Children and Young People. (BMJ Books, London, 2001), p. 202.

16. Note for Guidance on Clinical Investigation of Medicinal Products in the Paediatric Population (CPMP/ICH/2711/99), p. 11, section 2.6.3.

17. Royal College of Physicians, Research Involving Patients (Royal College of Physicians, London, 1990), section 7.43.

18. Royal College of Paediatrics and Child Health: Ethics Advisory Committee, "Guidelines for the Ethical Conduct of Medical Research Involving Children," Archives of Disease in Childhood 82: 177-182 (2000).

19. N. McIntosh, "Strengthen Ethical Committees' Role," British Medical Journal 307: 1496 (1993).

20. J.M. Chessels, C. Bailey, and S.M. Richards, "Intensification of Treatment and Survival in All Children with Lymphoblastic Leukaemia: Results of UK Medical Research Council trial UKALL X," Lancet, 345 1434-1438 (1995).

21. D.K. Pal, "Quality of Life Assessment in Children: A Review of Conceptual and Methodological Issues in Multidimensional Health Status Measures," Journal of Epidemiology and Community Health 50: 391-396 (1996).

22. Such as described in R.E. Grunau, T. Oberlander, L. Holsti, et al., "Bedside Application of the Neonatal Facial Coding System in Assessment of Premature Neonates," Pain 76: 277-286 (1998).

23. Council for International Organisations of Medical Sciences (CIOMS) and World Health Organization (WHO), International Ethical Guidelines for Biomedical Research involving Human Subjects (Geneva, 1993), p. 13.

24. M. Levene, I. Wright, and G. Griffiths, "Is Informed Consent in Neonatal Randomised Controlled Trials Ritual?" Lancet 347: 475 (1996)

25. S.A. Mason and P.J. Allmark, "Obtaining Informed Consent to Neonatal Randomised Controlled Trials: Interviews with Parents and Clinicians in the Euricon Study," Lancet 356: 2045 (2000).

26. T.H.H.G. Koh, L. Collie, D. Budge, and P. Butow, "Informed Consent in Neonatal Randomised Trials," Lancet 357: 9266 (2001).

27. M.P. Samuels, J. Raine. T. Wright, et al., "Continuous Negative Extrathoracic Pressure in Neonatal Respiratory Failure," Pediatrics 98: 1154-1160 (1996).

28. E. Hey and I. Chalmers, "Investigating Allegations of Research Misconduct: The Vital Need for Due Process," British Medical Journal 321: 752-756 (2000).

29. S.C. Harth and Y.H. Thong, "Sociodemographic and Motivational Characteristics of Parents who Volunteer their Children for Clinical Research: A Controlled Study," British Medical Journal 330: 1372-1375 (1990).

30. R.J. Bray, "Propofol Infusion Syndrome in Children," Paediatric Anaesthesia 8: 491-499 (1998).

31. H. Campbell, S. Surry, and E. Royle, "A Review of Randomised Controlled Trials Published in Archives of Disease in Childhood from 1982-1996," Archives of Disease in Childhood 79: 192-197 (1997).

Related Content:

Feature Article