Strengths and weakness of placebo controls in pediatrics RCTS
Among various clinical trial designs, double-blind and placebo-controlled RCTs are held, by researchers and regulatory agencies, as the “gold standard”, i.e. the most methodologically rigorous approach in order to assess the effect size of prospective new medicines. Indeed, the comparison of an experimental drug to placebo allows to disregard as a specific drug-effect any improvement in the patient’s conditions related either to her/his expectancy to receive a treatment or to the natural course of the disease thus, presumably, avoiding approval of not efficacious medicines.
Placebo-controlled RCTs are required for the approval of new drugs in many instances in which the withdrawal or delay of an active treatment is deemed not to cause major or irreversible damage. However, the principle of clinical equipoise should be satisfied, in other words there should be true uncertainty with regard to the effect of the two treatments, i.e. new drug and placebo. The comparison towards a placebo-control group may depend on the medical condition as well as on the therapeutic drug class. For instance, in trials aiming at establishing the efficacy/safety of new antimicrobial and anticancer drugs the prospective new drug should be evaluated either in add-on or in comparison to the best existing standard of care, because withholding an effective treatment could dramatically reduce the patient’s probability of regaining health. The same reasoning should apply for new molecules belonging to chemical/therapeutic drug classes for which there is already a number of drugs that have been either approved or used off label for a long time (as in the case of drugs for hypertension, diabetes or immune-suppressants). Assessing the efficacy of psychotropic drugs in children and adolescent, however, represents a real scientific and ethic challenge. For instance, major depression shows a response rate to placebo even higher than in adults and the results of previous trials show that the effect size of the existing medications cannot be easily predicted [
6]. For these reasons, the regulatory authorities require that efficacy of new prospective antidepressants be assessed in placebo-controlled RCTs. Pediatric major depression represents a condition associated with relevant co-morbidity, problems in family relationships, difficulties at school and social functioning, increased risk of recurrence and of suicide. Therefore, while the 8–10 weeks usually required to show efficacy in a placebo-controlled RCT may be deemed as a tolerable delay to start treatment in a child/adolescent with low-moderate severity depression that is drug-naïve or that has shown no response to existing medications, the comparison versus placebo would not be acceptable if randomization entails the withdrawal of a medication that, even if prescribed off-label, is of benefit to the patient.
According to the Declaration of Helsinki the use of placebo in medical research is allowed “for compelling and scientifically sound methodological reasons" [
7], a wording of not clear-cut interpretation and that has not helped to dissipate the controversy around the issue of placebo in medical research [
8].
Restrictions on the use of placebo had already been identified by Freedman [
9]. The use of placebo is acceptable when: there is no standard treatment, standard treatment is no better than placebo, standard treatment is placebo, the net therapeutic advantage of standard treatment has been called into question by new evidence or effective treatment exists but is not available due to high cost or short supply.
The use of placebo is not acceptable if established known data on the drug is not taken into account (unnecessary repetition of testing), the study deprives patients of a treatment proven efficacious- thus exposing them to the risk of aggravating a morbid condition- the study entails the interruption of a treatment that is already effective and accepted by the patient, the method of administration of the therapy is invasive or causes the patient unnecessary discomfort (hospitalization, invasive diagnostic tests, stress, etc.).
In order to minimize the risks in placebo-controlled RCTs, some rules regarding experimental studies have been determined: the exclusion of patients who have an increased risk of harm (more severe disease) in case of non-response; limitation of the period of exposure to placebo to the minimum required for scientific validity; careful monitoring of the enrolled subjects; administration of the necessary treatments in the presence of severe symptoms; indication of specific criteria for the suspension of the trial in patients showing adverse events [
10].
Hence, the use of placebo would be acceptable in anticipation of minimal damage and for reasons that are methodologically valid. However, the Council of Europe, taking a position on biomedical research involving persons unable to give their consent (including minors), has stated that “Research may be undertaken only if the results of the research have the potential to produce real or direct benefits to his/her health” (Chapter V, art. 15, ì) [
11]. Such a statement appears at odds with the design of most placebo-controlled RCTs in pediatrics because real or direct benefits cannot be predicted for patients randomized to receive placebo, unless avoiding the potential side effects of the experimental drug can be considered a real or direct benefit. The latter, though, can hardly be included among the expected results of a RCT.
Particularly when the expected effect size is small or not easily predictable from previous studies, as it often occurs i.e. with antidepressant drugs, a placebo-controlled RCT has the advantage,
versus an active drug-controlled RCT, of needing a lower number of patients per group in order to demonstrate a statistically significant superior efficacy of the active treatment [
12]. Of course, the clinical relevance of a statistically demonstrated superior efficacy versus placebo heavily depends on the primary end-points of the study, as well as on the long-term effects which may not be evaluated in RCTs of short duration. Nevertheless, there may be several advantages associated with enrolling a lower number of patients in placebo-controlled pediatric RCTs, for instance: a) the study’s feasibility may be greater, due to the limited number of pediatric patients affected by specific diseases, and particularly when there is a need to acquire data on pharmacokinetics, efficacy and safety stratified by age groups, b) results on efficacy are obtained with a lower number of particularly vulnerable patients exposed to the unpredicted adverse effects of the new drug.
On the whole, a placebo-controlled RCT would be more time- and cost-effective. A better time-effectiveness in reaching positive results would mean a more rapid approval that would be certainly favorable for the drug manufacturer, but also for patients, because of a shorter delay in the availability of efficacious new drugs. Such a benefit should be complemented by an adequate assessment of the drug’s safety that, however, may be hindered by the short duration of the trial, based on the principle that patients should be exposed to placebo for the minimum time required to demonstrate efficacy.
A placebo control group is frequently claimed as necessary and therefore justified, on the ground that there are no standard (i.e., approved) treatments for a particular disease. Such a claim should not be so readily accepted, or at least should be analyzed case by case, when reviewing pediatric RCTs, since a placebo-controlled design would be unacceptable if there is an efficacious, though off-label, treatment in place and the study protocol entails switching the patients to placebo.
Children are daily treated with off-label drugs and as recently stated by the American Academy of Pediatrics: “… the term “off-label” does not imply an improper, illegal, contraindicated or investigational use…rather it means that the evidence required by law to allow inclusion in the label has not been approved by the FDA (or other Regulatory Agencies)…in no way does a lack of labeling signify that therapy is unsupported by clinical experience or data in children” [
13].
Hence, it does not seem correct that drugs that have been used off label for years, and are deemed of sufficient efficacy/safety - though on the basis of evidence not as strong as that required for approval - to be reimbursed or paid for by National Health Systems, may not be considered as suitable comparators in pediatric RCTs that would address superiority, non-inferiority or equivalence of the prospective new drug, as applicable to the particular medical condition and drugs examined. Evidently, keeping in mind the child’s best interest, it seems that a major question to be answered concerns the definition of “standard treatment” and, thus, the minimum level of acceptable scientific evidence that, although variable for different drug classes and disease conditions, would be needed for an off label drug to be considered a proper standard treatment.
It is acknowledged, however, that with regard to international and multicenter pediatric RCTs this approach would require an effort to harmonize the heterogeneous off-label drug utilization in children of different regions (i.e. USA versus EU and among EU Nations).
Scientific rigor, placebo and the patient’s well-being
Also in the context of a clinical trial, the primary justification for a medical act is the search of the patient’s well-being. For this reason, the evaluation of the benefit/risk ratio is fundamental when ideating an experimental project.
When determining the risks and disadvantages of a medical act, reference is made to three criteria: 1. the probability that the patient may suffer harm; 2. the severity of the harm; 3. the acceptability of harm [
14]. These criteria do not appear sufficient in order to define the “minimum” damage in the case of clinical trials on minors. For this reason, it is proposed to quantify the “minimum” damage on the basis of the risk and/or discomfort that a child might encounter in everyday life and in the minor’s specific situation or during routine examinations or psychological tests [
15].
However, since it is difficult to define what risk or inconvenience can be defined “minimum” and therefore acceptable, the Swiss NEK-CNE proposes to refer to the criterion of proportionality: “Rather than the use of fixed or even quantitative standards for the measurement of risk, the NEK-CNE would welcome a situational assessment on the part of ethics committees, taking into account the context of the study and the specific characteristics of the subjects concerned. Here, ethics committees should be guided by the principle of proportionality between potential harm and benefits, which is to be understood in the sense of “reasonable” risks and burdens for the child […] A study-specific assessment of burdens makes it possible to take into consideration the child’s particular situation […] and also the type and extent of benefits anticipated (individual/social, minor/major)” [
16]. Complete safety cannot, however, be granted to a children who participates in a clinical trial and there is always the risk that the social advantage may override the individual benefit.
The assessment of the proportionality in the benefit/risk ratio is even more difficult in placebo-controlled RCTs, given that placebo doesn’t present in itself any real or direct benefit, unless we consider as such: 1. the absence of risks arising from the administration of an active comparator; 2. the presence of the benefits attributable to the psychobiological reactions triggered either by the overall therapeutic context or through the activation of endogenous pathways (i.e. in pain treatment) or through the notion of being the object of care and medical attention (
therapeutic ritual) [
17].
Even assuming that the utilization of placebo reduces potential iatrogenic effects, risks could instead derive from the lack of administration of an active drug, as shown by evidence of psychological damage in patients who are obliged to leave an experimental study due to the aggravation of their clinical condition [
18].
As well, if a placebo treatment does not directly cause permanent adverse reactions, one must also consider the damage brought to the patient by the omission of a potentially useful treatment or by the interruption of a previously initiated treatment.
Placebo-controlled RCTS and informed consent
Respect of patient's autonomy means allowing the patient to choose what is most appropriate for the improvement of her/his living conditions. This requires an adequate communication process that offers all the necessary information in order to obtain the patient’s consent, with the intent of removing any impediment in the exercise of autonomy. In order for information to be complete, comprehensible and result in the expression of conscious consent, it is necessary to satisfy specific requirements regarding the information’s quality and the patient’s understanding, freedom and decision-making capacity. Research, however, suggests that the participants of RCTs do not always adequately understand the studies [
19,
20]. In the case of clinical trials, greater attention is necessary in order to avoid the so-called “therapeutic misconception” - that is, the belief that the treatment provided has been specifically designed to meet the patient’s needs [
21]. The presence of adequate information and of the patient’s consent, however, does not make any experimentation ethically acceptable, nor does it diminish the doctor’s responsibility.
The consent forms usually contain information on the study’s design, but not regarding placebo. While risks and benefits of the new treatment are listed, indications of the risks/benefits deriving from placebo are lacking [
22,
23]. On the other hand, it could not be otherwise, in fact, a placebo lacks a specific effect, even though it may determine a placebo-effect [
24]. Given that each placebo-controlled RCTs should consider as outcomes the modification of all parameters, that is not only the physiological but also the psychological ones, the evaluation of the placebo-effect would be neither easy to obtain nor would it be reproducible. It is necessary to explain to the patient also the “nocebo” risk, partly resulting from the non-administration of the study treatment or an active comparator, when available.
Obtaining truly informed consent is even more complex when experimental studies involve children. In this case, it is even more necessary to carefully evaluate the contents and form of the informed consent [
25]. The pediatric age range, in fact, not only is characterized by some specific clinical aspects due to the development of the organism, but also by a different dynamic in the doctor-patient relationship. In this case, the decisional subject does not coincide with the patient: the physician relates only with the parents, especially if the child cannot be involved because of the young age. While an older minor (14–18 years) who has demonstrated understanding and judgment (“mature” minor), may take part in the decisions, very different is the case of a child who is less than 9 years old [
26].
The physician should, therefore, communicate with the parents and explain that: 1. the clinical trial is not a personalized form of treatment and that, instead, participation may preclude the possibility of receiving the standard or an off-label alternative treatment; 2. the subjects in the placebo group will not receive an active treatment; 3. some risks can derive from the non-administration of the standard/alternative treatment; 4. if the study will be interrupted, the child will begin immediately the appropriate treatment. And most importantly, the physician must explain exactly what a “placebo” is. This could, however, not be enough: the parents may have difficulty understanding the research protocol, or underestimate the possible risks, or be influenced in their decisions by the researchers’ requests.
Placebo and the standard of the “child’s best interest” (CBI)
In clinical pediatrics, when the minor cannot be involved in decision-making process, the “child’s best interest” (CBI) is used as a standard. First used in the legal field, to make decisions on issues regarding the welfare of a child (divorce, separation, adoption, etc.), the CBI standard is very difficult to define. Those who support the validity of the CBI standard, emphasize the positive value of the analysis of each individual case and the opportunity to remind physicians of their responsibility in decisions concerning the health and life of young patients [
27]. Those who criticize its validity, consider the CBI standard potentially self-destructive, individualistic, dangerous, vulnerable to forms of abuse and difficult to use since the child cannot be involved in the decision-making process and, therefore, cannot express any opinion on her/his own condition [
28]. Indeed, as pointed out by the American Academy of Pediatrics, the real evaluation of a patient’s quality of life can only be made by the patient himself [
29].
Pros and cons: there is a necessity for a reflection on the CBI standard not only to determine its correct content and method of application, but also to verify its validity and usefulness. In fact the limit of the CBI standard is, first of all, structural: it is the simple transposition of the principle of autonomy into a context where its exercise is not possible. To make the evaluation more objective, it would be necessary to isolate the child’s interests from those of the other subjects involved, for example: the regulatory agencies [
10], the pharmaceutical companies [
30], the physicians. Despite the external pressures, physicians cannot, however, forget their ethical, deontological and legal responsibilities towards their patients. Personal interests (i.e. publications/promotions) cannot be in conflict with the responsibility of caring for the patient: “If one has
to err, one should
err on the patients’ side, i.e., preserve their welfare over the scientific rigor of the study” [
10].