Intended for healthcare professionals


Testing new pharmaceutical products in children

BMJ 2003; 326 doi: (Published 11 January 2003) Cite this as: BMJ 2003;326:64

A positive step, but ethical concerns remain

  1. Alastair G Sutcliffe, senior lecturer in paediatrics
  1. Royal Free and University College Medical School, London WC1E 6BT

    Ideally all medicines that could benefit children should be licensed for this purpose. To date this has been the exception rather than the rule, resulting in children becoming therapeutic orphans sometimes with tragic consequences. Many medicines prescribed for children, whether in hospital or primary care, are either unlicensed or off label (used for an indication, age, dose, or route of administration outside the terms of the product licence). 1 2 Other problems surround the prescribing, dispensing, and administering of medicines to children.3 Many difficulties could be overcome if manufacturers were obliged to test their products on children as well as adults.

    In the United States a “carrot and stick” approach has been adopted to achieve this end.4 The carrot is the “paediatric exclusivity provision,” which grants an additional six months of patent protection or market exclusivity to companies that voluntarily test the relevant drug on children. The stick is the “paediatric rule,” a later requirement by the Food and Drug Administration that companies test their products on children under certain circumstances. These include the likelihood of (1) usage in a substantial number of children, (2) meaningful therapeutic benefits, (3) risk to children in the absence of labelling (licensing), (4) usage in different paediatric age groups. Recently this rule has been challenged successfully, principally on the grounds that the Food and Drug Administration exceeded its authority in imposing these requirements on drug companies, leaving uncertainty about how the increased testing of medicines is to be achieved.5

    Although including children in trials has benefits, technical, practical, and ethical problems remain.6 Technically it will be necessary to develop further non- or minimally invasive techniques for pharmacokinetic and pharmacodynamic assessments. Pharmacodynamic assessments are required to determine accurate dosage even in clinical trials. This is especially important in children whose metabolism of drugs and sensitivity of end organs to them vary considerably with age and stage of development. Practical difficulties arise from the small number of children with specific disorders, the need to study different ages, the formulation of drugs, and problems of recruitment.

    Ethically, the obligation to act in a child's best interests entails protecting children from both the potential risks of research and the harms produced by the use of inadequately tested drugs, as well as respecting their autonomy. To satisfy these requirements drug trials should be scientifically and socially valid, adequately powered, of favourable risk-benefit ratio, subject to independent ethical review and informed consent, and conducted to an appropriate standard.7 But inevitably, pharmaceutical studies carry risks of physical or psychological harm that may be difficult to quantify. Acceptance of some risk is necessary for therapeutic advances to occur, for example, in the treatment of acute lymphoblastic leukaemia.8 Difficulties in recruitment may mean that studies lack the statistical power to answer the questions they pose. In contrast large studies are expensive and, if commercially sponsored, may tend to favour the drug tested. Commercially funded research understandably includes drugs that are likely to have high volume sales—for example, antibiotics—or high unit costs, such as surfactant.8 Research ethics committees must assess all these factors but do need sufficient expertise in paediatrics to do so safely and effectively.

    Obtaining adequately informed voluntary consent for participation in a drug trial requires the children's assent or consent, commensurate with their understanding and experience, with or without the permission or consent of their parents.8 Imparting sufficient, comprehensible information to distressed parents and ill children is difficult and leads to questions of their competence to make decisions. Understanding of such terms as randomisation, trial, and placebo may be imperfect.9 Parents may accept drug treatment as being the only hope for their child irrespective of the risks entailed; they may feel obliged to researchers and believe that refusal to enter a trial will compromise their child's treatment.4 Doubt therefore exists whether truly informed consent is possible, especially in acute life threatening situations.

    These factors and issues raised by the Bristol and Alder Hey inquiries and the Griffiths's report have created a climate in which the testing of medicines in children poses difficulties. Overcoming them is essential if children are to benefit from the development of safe effective drugs. One possible solution would be to introduce legislation such as that in the United States. But such legislation may not in itself lead to a greater number of children being enrolled in drug trials or deal with the difficulty in obtaining sufficiently informed consent. Moreover, the exclusivity provision has been criticised as benefiting the interests of companies rather than children.4 An alternative, but long term, strategy is the greater involvement of children and their families in the planning and implementation of research projects, which should have notable educational impact.10 This approach could both increase recruitment and satisfy the criteria for informed consent.

    In the meantime, those who prescribe for children should use drugs that are licensed or accepted as offering the best possible prospect of benefit by a responsible body of medical opinion. Equally they have a duty of advocacy on behalf of children, supporting the therapeutic orphans who, like fictional character Oliver Twist, are saying: “Please sir, we want some more.”


    • Competing interests None declared.


    1. 1.
    2. 2.
    3. 3.
    4. 4.
    5. 5.
    6. 6.
    7. 7.
    8. 8.
    9. 9.
    10. 10.
    View Abstract