Novel Therapies, High-Risk Pediatric Research, and the Prospect of Benefit: Learning from the Ethical Disagreements

Despite recent efforts by the US Food and Drug Administration (FDA) to improve the quality and quantity of clinical research data,¹ two-thirds of drugs prescribed currently to children have not been studied for safety and efficacy in pediatric populations, and information on the efficacy and safety of drugs in children is not methodically collected and analyzed.² According to some studies, the majority of pediatric drugs prescribed for children involve unlicensed drugs or off-label prescribing across all medication categories.³ There is evidence of a greater risk of a severe adverse drug reaction occurring in association with the off-label or unlicensed use of drugs in children.⁴

Obtaining safety and efficacy information on pediatric therapies requires systematic data collection and clinical research trials. Conducting research with human beings requires the balancing of two important, and sometimes conflicting, aims: ensuring access to the potential benefits that scientific research can offer and protecting human subjects from research risks and harms. This tension is all the more salient in pediatric research. Children's cognitive, psychological, and social immaturity limits their ability to understand what is involved in a research trial and to make sound decisions about participation. Because of children's vulnerability, federal regulations mandate that institutional review boards (IRBs) apply additional protections before they can approve pediatric research; such regulations allow IRBs to approve only research that either offers the prospect of direct benefit to the individual children participating or involves minimal risk or a minor increase over minimal risk.

Both the category of research that involves minimal risk and the one involving a minor increase over minimal risk have received a significant amount of attention.^5,6,7,8 With some exceptions,^9,10 however, the category that concerns high-risk pediatric research with the prospect of direct benefit has been subjected to less scrutiny. Moreover, although there are significant disagreements over whether phase I trials can be said to offer a “prospect of direct benefit,”^{11,12,13,14,15,16,17} much of that discussion has taken place in the context of trials involving competent adults rather than children, and much of it centers on oncology trials.

We focus here on high-risk pediatric research with the prospect of direct benefit and point out some aspects that have raised significant debate. In particular, we call attention to disagreements related to two essential aspects of this type of research: (i) determining what constitutes a “prospect of direct benefit” in phase I trials that involve gene transfer technologies and (ii) assessing when in these trials the risk is justified by the anticipated benefit to the participant children. Although much of our discussion is applicable to other types of high-risk pediatric trials, as an example of the dilemma this type of research poses we use pediatric trials that involve gene transfer technologies. We focus on clinical trials for late infantile neuronal ceroid lipofuscinosis (LINCL). Exploring the ethical implications of some of these disagreements might identify resources for determining how best to deal with the ethical concerns at stake in high-risk pediatric research. We thus offer some recommendations for responding to these concerns.

Research involving children

Most people would agree that research with children is needed to improve pediatric medicine. As with any research involving human subjects, current research guidelines, in both the United States and abroad, attempt to address this need by permitting the enrollment of children in research when it offers an appropriate risk–benefit balance.¹⁸ If the research offers no prospect of direct benefit but involves a sufficiently low degree of risk, the requirement to balance risks and benefits calls attention to the need to assess the likelihood that the research will produce generalizable knowledge that is relevant to children's health or to particular diseases or conditions affecting children. However, although IRBs can approve research with adults that involves greater than minimal risk without the prospect of direct benefit, federal regulations for pediatric research, in accordance with recommendations of the National Commission for the Protection of Human Subjects of Biomedical and Behavioral Research, require that IRBs use a stricter standard.¹⁹

Pediatric research that does not offer the prospect of direct benefit is approvable by IRBs provided that the risks are sufficiently low (Table 1). Specifically, Title 45 of the Code of Federal Regulations, Part 46, Subpart D (Title 21, Part 50, Subpart D, for research involving US Food and Drug Administration (FDA) regulated products), allows pediatric research without the prospect of direct benefit when the research involves minimal risk (§46.404/50.51) and the researchers obtain the child's assent and permission from at least one parent or guardian. It also allows research with no prospect of direct benefit that involves a greater than minimal risk (§46.406/50.53) provided that (i) the risk represents a minor increase over minimal risk; (ii) the intervention or procedure presents experiences to subjects that are reasonably commensurate with those inherent in their actual or expected medical, dental, psychological, social, or educational situations; (iii) the intervention or procedure is likely to yield generalizable knowledge about the subjects' disorder or condition that is of vital importance for the understanding or amelioration of the subjects' disorder or condition; and (iv) adequate provisions are made for soliciting assent of the children and permission of their parents or guardians.

Table 1. Research involving children that the DHHS/FDA is permitted to fund or conduct.

Sometimes, however, research can involve a greater than a minor increase over minimal risk. This type of research is approvable by IRBs only when it offers the prospect of direct benefit to the children participating (§46.405/50.52). Moreover, such research can be approved provided that (i) the risk is justified by the anticipated benefit to the participant children; (ii) the relationship of the anticipated benefit to the risk is at least as favorable to the subjects as that presented by available alternatives; and, as with prior categories, (ii) adequate provisions are made for soliciting assent of the children and permission of their parents or guardians. Research that involves greater than minimal risk with no prospect of direct benefit cannot be approved by IRBs. However, if this type of research, called “research otherwise not approvable” (§46.407/50.54), presents an opportunity to understand, prevent, or alleviate serious health problems, it can be approved by the Secretary of the US Department of Health and Human Services (DHHS) or the FDA Commissioner, after receiving a recommendation from an expert committee (Figure 1).

Figure 1.

Decision tree regarding risk considerations for pediatric research. The information is based on 45 CFR 46, Subpart D/21 CFR 50, Subpart D. CFR, Code of Federal Regulations; FDA, US Food and Drug Administration; HHS, US Department of Health and Human Services.

Do phase I pediatric trials offer the prospect of direct benefit?

Early-phase gene transfer trials present a good illustration of the ethical difficulties related to the assessment of risks and potential benefits (Tables 2 and 3). These trials usually address rare, fatal or very serious diseases for which few or no other alternatives exist, are associated with significant risks (including death), and often involve highly novel approaches and surgical procedures.^20,21,22,23 For instance, LINCL, an autosomal recessive lysosomal storage disease that affects the brain and retina,²⁴ is inevitably fatal in childhood, following progressive neurological deterioration. There is currently no treatment for LINCL other than management of symptoms. Clinical trials for this disease have involved the administration of an AAV2hCLN2 vector²³ directly to the central nervous system (CNS) of children with LINCL in a neurosurgical procedure involving general anesthesia, six burr holes, catheter insertion, and infusion of the vector in a total of 12 sites over several hours.

Table 2. Benefits of gene transfer intervention in early-phase clinical trials of nonmalignant disease in the pediatric population.

Table 3. Serious drug-related adverse events in early-phase gene transfer clinical trials of nonmalignant disease in the pediatric population.

Although phase I studies like the one for LINCL play a critical role in the translation of basic research into clinical application, they also present researchers and IRBs with the significant difficulty of determining whether such trials offer the prospect of direct benefit to the participant children. Children might benefit from participating in a clinical trial because the trial involves more careful monitoring, or some children might benefit from participating in research because such participation allows them access to medical care that they would otherwise lack. Such benefits however, are not direct but collateral benefits of research participation.¹¹ Similarly, clinical trials usually involve aspirational benefits. These are benefits to society that result from conducting scientific research, such as generation of scientific knowledge and future improvements in treatment. However, as currently accepted, direct benefits to the children participating in research are those that result from receiving the particular intervention being tested.¹¹

Of course, phase I trials can involve a variety of purposes and designs, and not all of them would present the same degree of difficulty when assessing the prospect of direct benefit. Nonetheless, legitimate disagreements exist about whether phase I trials that involve novel procedures, such as gene transfer, can be said to offer the prospect of direct benefit. Some of them involve epistemological disagreements about the strength of the evidence. For instance, preclinical studies provided justification for initiating several gene transfer phase I trials involving children.^{20,23,25,26,27} In the case of LINCL, preclinical studies with rodents showed that administration of the vectors AAV2hCLN2 and AAVrh.10hCLN2 was associated with a decrease in the abnormal accumulation of storage material in CNS neurons.^23,28 However, there is significant evidence that what are clearly encouraging results encountered in preclinical studies often fail to translate into efficacy in clinical studies.^29,30 Clearly, the biology of humans shares many commonalities with rodents and nonhuman primates, but inferences from animal models to human beings are fraught with difficulties. Reasons for this disconnect between preclinical and clinical outcomes include problems with the animal models used, poor methodological quality of animal studies, and publication bias.^31,32,33 Of course, it could be the case that, for some phase I trials, the preclinical studies used as evidence to argue for a prospect of direct benefit are strong, but the epistemic difficulty of determining when such is the case might still be present.

Others, however, argue that, rather than considerations about the amount or quality of evidence for the experimental procedure, what are important for determining whether a trial offers the prospect of direct benefit are the intentions of investigators when conducting research. For these commentators, the intentions of the researchers are morally relevant, and they should be taken into account when making determinations about anticipated benefits.³⁴ Clinical research is thought to often contain a mixture of procedures. Some of those procedures (e.g., drugs, biologics, and surgical and behavioral interventions) are administered with therapeutic intent, whereas others (e.g., venipuncture for pharmacokinetic drug levels, additional imaging procedures, and genetic analysis not used in clinical practice) are used solely to answer a particular research question. Given the moral relevance of researchers' intentions when using therapeutic and nontherapeutic procedures, such interventions require different moral standards of evaluation. Therapeutic procedures in particular must meet the ethical standard of clinical equipoise.³⁵ Under this view, clinical equipoise requires that at the beginning of a trial there exists a state of honest, professional disagreement in the community of expert practitioners as to the preferred treatment. In at least some phase I trials involving gene transfer, experts can reasonably disagree about whether the experimental procedure, offered with therapeutic intent, is better than the standard of care, which, in the case of LINCL, involves only management of symptoms and comfort care.²³

Nonetheless, others maintain that researchers' intent is irrelevant to the ethics of clinical research.³⁶ Clearly, it is often difficult to determine what someone's intentions might be. More importantly, people's intentions are distinct from the consequences that follow from their actions. Researchers might intend to directly benefit subjects who participate in phase I trials—or any other type of trial—but whether subjects will benefit depends not on the researchers' intentions but on the properties of the intervention given to subjects. In the case of gene transfer for LINCL, for instance, the properties of the vector and not the intentions of the investigators are the relevant factors when evaluating the possibility of direct benefit to the participant children. Similarly, subjects might benefit from participating in research even when researchers are intending to only obtain data. One can certainly agree that researchers' intent might be a relevant factor in judging their moral characters, but many have quandaries accepting the claim that intentions are a morally relevant factor in judging whether a research trial offers the prospect of direct benefit.

Another area of disagreement that makes it difficult to assess a prospect of direct benefit in many phase I trials is related to the aims of phase I trials. Traditionally, phase I studies are conducted in order to determine the metabolic and pharmacological actions of the drug in humans, as well as the adverse effects associated with increasing doses, and, if possible, to gain early evidence regarding effectiveness.¹⁵ Because of the small number of subjects that phase I trials enroll, it is in fact difficult to assess any evidence of effectiveness. Hence, disagreements exist about whether, given the goals of phase I trials, it is appropriate to present these trials as offering the prospect of benefit.

However, others argue that, even though the primary goal of a phase I trial might be to evaluate safety, this goal is independent of whether the particular intervention being tested might in fact benefit the participants.^14,36 Here, as in the case of researchers' intentions, the purpose of a trial and its anticipated consequences are distinct categories, and although phase I trials might not be powered to measure benefit, that fact by itself does not mean that benefit cannot be a possible outcome.

When are risks justified by the anticipated direct benefits?

Disagreements about the ethical conduct of pediatric research extends beyond what constitutes a prospect of direct benefit to assessments about when the risks are justified by the anticipated benefits. The epistemic concerns mentioned above about the strength of the preclinical evidence are obviously relevant when making judgments about the balance of risks and benefits. But to the difficulty of assessing the evidence to determine what direct benefits a phase I trial might offer or whether one can reasonably say that it offers any prospect of direct benefit, we must add questions about the strength of the evidence about risks.

Clearly, no matter how good the preclinical evidence might be, given the differences between animals and humans, some risks resulting from novel procedures such as gene transfer will be unknown (Table 3). For instance, important aspects that need to be considered when assessing gene transfer risks include the choice of animal models with respect to the pharmacological activity of the gene transfer product, the immunogenicity of the viral vector, and questions about the biodistribution of the vector after administration and insertional mutagenic effects.^{21,22,37,38,39,40} Even when preclinical and other clinical studies might provide evidence of risks that may be associated with gene transfer, in some cases it might be difficult to determine probabilities for some of those risks and in others we will have no knowledge at all about the risks that might be involved.^21,22,40 Given the uncertainties with evidence pertaining to both risks and benefits of novel procedures, divergence in assessments of the balance of such risks and anticipated benefits is thus quite reasonable.

Pediatric clinical trials involving experimental procedures such as gene transfer present different ethical and risk considerations from those with trials recruiting adults. First, although normally adults can provide informed consent, as we mentioned earlier, a child—in practice and by law—does not have sufficient intellectual and emotional capacity to understand the balance of risk and benefit, and thus to make informed choices. Moreover, parental consent is likely to be clouded by complicated emotions, including fear, guilt, and even desperation. Second, questions of risk assessment are likely to be exacerbated for experimental procedures directed to children when compared with those treating adults. Although questions about therapeutic persistence and concerns about long-term safety can arise for any clinical trial, such concerns are more salient when dealing with children because of the difficulty of examining long-term effects in the relatively short-term clinical studies normally conducted. Also, in cases involving the use of viral vector–mediated gene transfer, vector immunity may preclude the use of interventions that may become available in the children's longer life span.

Another area of disagreement that affects the requirement to justify risks by attending to the anticipated benefits relates to dose-escalation studies. Some commentators argue that, in these studies, determining the boundary for optimal dosage requires a study design that exposes at least some participants to doses that are ineffectual and, at the higher dose range, doses that have an unfavorable ratio of risk to direct benefit when compared with the standard of competent care outside the study (i.e., no treatment).¹⁷ This is so because, to determine optimal dosage for testing in subsequent studies, different doses must be used. Initial doses are low and tend to be relatively safe, but in these cases it is likely that no benefit will ensue. If the initial dose is not toxic, then dosage is escalated in new participant cohorts until major safety concerns appear. If this is the case, then at least some aspects of phase I clinical trials cannot be said to involve risks that are justified by anticipated benefits. Even if the risk–benefit balance presented by study participation is favorable in the aggregate, the structure of dose-escalation studies means that it is unlikely that all the participants will receive interventions that result in an appropriate balance of risks and direct benefits. Hence, at least for some children, the balance of risks and benefits will not be favorable, even if we cannot know in advance of the trial which children those will be.¹⁷

Ethical implications and recommendations

Although there has been considerable debate over various aspects of the ethics of clinical research with children, less attention has been given to the possible ethical implications of some of the disagreements discussed here. This, however, might be a mistake, as lack of attention to what arguably are reasonable disagreements can prevent reflection on ways to deal with the ethical issues that arise when conducting high-risk pediatric research.

As noted, some of the disagreements about whether it is valid to argue that some phase I trials offer the prospect of direct benefit pertain to the strength of the evidence about risks and possible benefits. These disagreements call attention to the need to reflect on and develop appropriate standards of scientific evidence for both preclinical and clinical studies.⁴¹ Developing standards of evidence that take into account the complexity of using novel drugs and techniques in clinical trials might go a long way toward helping to reduce disagreements about whether some phase I trials can reasonably be said to offer the prospect of direct benefit.

Nonetheless, it is unlikely that better standards of evidence will end these types of disagreements. Hence, it is also important to remember that discrepancies among and between IRBs, researchers, and parents about whether a trial offers the prospect of direct benefit or whether the risks are justified by the possible benefits are not necessarily the expression of irrational beliefs on someone's part. Considerations about risks and benefits necessarily involve value judgments. Risk evaluations related to human health include central ethical assumptions, such as what counts as a serious risk, what is an acceptable level of risk of death or sickness, what are the standards to judge that unmanageable risks are not present, and what is the relevant time frame for investigating such risks. There is evidence that social affiliations such as profession, gender, and political ideologies influence what one determines to be a risk.⁴² Experts might reasonably disagree with one another. Also, the views of laypeople regarding risk often differ from those of experts.⁴³ Laypersons tend to value both the context of risk as well as its content, whereas experts usually place greater emphasis on risk end points, such as human health hazards or toxicity levels, rather than the context in which they may unfold. Also important are considerations about alternatives, the voluntary nature of the risks, concerns about equitable distribution, and trust in those in charge of imposing and managing risks.^44,45 All these factors shape what one perceives to be a risk that might or might not be thought worth taking, as well as one's assessment of the benefits of a particular course of action. Thus, accusations of irrationality when people disagree about risk assessment might be misplaced. What is needed is transparency in the values that ground people's actions so that such values can be critically evaluated. Even when this does not lead to agreements, it might be helpful in facilitating communication among the different stakeholders.

Similarly, once IRBs classify research as involving a greater than a minor increase over minimal risk, concerns about the strength of scientific evidence raise questions about the fact that federal regulations for high-risk pediatric research do not promote considerations about how strong the evidence about risks and possible benefits for an intervention might be. Hence, a phase I trial involving a very novel drug for which only preclinical studies are available, a phase I trial involving a drug for which data on human safety exist (as in phase I trials of oncology interventions), and a phase III trial in which evidence about a drug's risks and potential benefit might be more reliable could all be classified under category §46.405 (i.e., research that involves a greater than a minor increase over minimal risk with the prospect of direct benefit). But if the strength of the evidence is relevant when determining whether one can reasonably argue that a particular clinical trial offers the prospect of direct benefit, then categorizing all these very different trials under the rubric of high risk with the prospect of direct benefit might be counterproductive.⁹ The lack of attention in the federal regulations to different levels of evidence in this context might contribute to overestimating risks or underestimating benefits, simply because all high-risk trials, approvable by an IRB, have to fit the same category.^9,46 It is therefore necessary to carefully consider these matters and to propose regulations that are consistent with current scientific practice.

Indeed, current regulations for high-risk trials involving the prospect of benefit are in tension with concerns about the therapeutic misconception. Many have argued that research participants—and often researchers—make decisions about their participation in research influenced by the therapeutic misconception.^47,48,49 Significant disagreements exist about how to adequately characterize this misconception.^50,51,52 The therapeutic misconception was initially characterized as a failure of research subjects to understand that research involves practices that conflict with the traditional aims of medical care.⁴⁷ However, many also characterize the therapeutic misconception as a tendency to overestimate the clinical benefit from an experimental intervention and underestimate the risks of such intervention.^11,48

Those who interpret the therapeutic misconception in this second way argue that phase I trials for novel procedures such as gene transfer should not be reasonably thought of as offering a potential for direct benefit to subjects. Hence, when consent forms include references to the prospect of direct benefit, or are ambiguous and vague about the fact that to believe in such prospects is unreasonable, this promotes the therapeutic misconception and threatens the validity of the informed-consent process.^48,49 But, in the case of pediatric trials, this involves a conflict. Indeed, insofar as a clinical trial is classified under category §46.405/50.52, the trial has been thought to provide both a prospect of direct benefit to the individual children participating and a balance of risks and benefits. If this is so, then subjects—and parents—cannot be properly said to act under the misconception that the trial offers a reasonable prospect of benefit. Moreover, if this is so, then it is, in principle, appropriate for consent forms to indicate that these trials offer the prospect of direct benefit to participants, even when clearly this would need to be done in ways that do not promote false beliefs in participants. On the other hand, if subjects—and parents—are indeed correctly thought of as suffering from the therapeutic misconception, then at least some phase I trials are mistakenly classified as §46.405/50.52. Indeed, to the extent that it is appropriate to judge that participants in at least some pediatric phase I trials incorrectly believe that the trial offers a prospect of direct benefit and a balance of risk to direct benefit, then such trials cannot be approved under category §46.405/50.52 but would have to be considered under category §46.407/50.54. Either way, it seems that more attention needs to be given to addressing this tension in pediatric research.

Disagreements about whether phase I trials can be said to offer a balance of risks and direct benefits, given the traditional goals and design of such trials (i.e., to assess safety, dose-escalating studies), leads to requirements to seriously contemplate new trial designs. Such designs should be aimed to both assess safety and evaluate any promise of efficacy in meaningful ways and should be constructed to enhance possible benefits to participants.^53,54,55 Hence, rather than simply assume that research goals (to obtain generalizable knowledge) always trump, or are incompatible with, ethical purposes (to offer therapeutic benefit), efforts should be directed to develop and implement trial designs that try to maximize the likelihood of achieving both types of aims.

All these concerns call attention to the difficulty of deciding how to move forward in developing therapies for fatal pediatric disorders when trials involve risky experimental interventions and might require invasive procedures that are themselves risky and are associated with significant discomfort. Reasonable people would agree that, for a fatal pediatric disorder, without research there can be no cure. Still, we must ensure that vulnerable populations such as children are protected. And we must not forget the parents who, particularly in cases of inherited disorders, feel responsible for their contribution to their child's fatal disease.

In its current formulation, §46.405; 50.52 might have the negative effect of compelling investigators and IRBs to overestimate the benefits of their proposed phase I study so as to ensure that the study meets the criterion of presenting the prospect of direct benefit to the individual subjects. As we have indicated previously, claims about prospect of benefit in phase I studies for fatal pediatric conditions are grounded in preclinical data. There is significant agreement that, no matter how many experimental animal studies are carried out, risks and potential benefits to human subjects are simply unknown until evidence from human subjects is obtained. After all, that is the reason why the research is being conducted.

If this view is correct, then phase I trials for fatal pediatric disorders are unlikely to be covered by §46.405; 50.52, and they would fall into category of §46.407; 50.54, which requires that the research be approved by the Secretary of the DHHS or the FDA Commissioner after receiving a recommendation of an expert committee. Although the purpose is, understandably, to protect children and their families, we suggest that this level of escalation, with the likely associated media attention, might have the unintended effect of limiting the development of potential therapies for these fatal disorders.

Given these considerations and the previously discussed tensions in the existing guidelines, we suggest that the DHHS Secretary and FDA Commissioner empanel a group of experts in the ethics and design of such trials to reconsider §46.405; 50.52 with the recognition that (i) research is research, and a prospect of direct benefit might often fail to be a justification for proceeding with such trials, and (ii) the appropriateness of the trial design, as well as the quality of the experimental animal studies, should be a more significant focus of an evaluation to determine whether trials should move ahead.

Research with human subjects presents us with difficult ethical quandaries. When that research involves children, the ethical concerns are even more salient. Nonetheless, confronting such ethical issues and addressing them in appropriate ways is essential to the conduct of pediatric research. Without it, children will be put at risk of harm by the lack of adequate treatments that have been proven safe and effective.

Given that the well-being of children and their parents is at stake when conducting pediatric research, particularly research that involves significant risk to the children, it is not surprising that important disagreements exist about how to assess risks and benefits. In this article we have examined what some of the disagreements about how to approach pediatric research tell us about current scientific and regulatory practices. They show that there is a pressing need to develop standards of evidence that take into account the complexity of using novel drugs and techniques in clinical trials and to develop and implement trial designs that try to maximize the likelihood of reaching the goals of obtaining generalizable knowledge and offering therapeutic benefit to the participants. These disagreements also call attention to the importance of taking into account the fact that considerations about risks and benefits necessarily involve value judgments. Requiring transparency of these values so that they can be critically evaluated might substantially advance communication among stakeholders. Finally, and importantly, these disagreements show that current federal regulations are inadequate to respond to current scientific practice. Federal regulations for pediatric research should be revised to account for distinctions between trials that involve different strengths of evidential support in relation to the prospect of benefit and the balance of risks and benefits. In doing so, they will facilitate the task of researchers and IRBs, encourage more appropriate judgments about risks and possible benefits, and discourage the therapeutic misconception.