Ensuring the Involvement of Children in the Evaluation of New Tuberculosis Treatment Regimens

William Burman and colleagues review the barriers to involving children in studies of new tuberculosis treatments and recommend strategies for overcoming these barriers.

G lobal tuberculosis control is threatened by dramatic increases in HIV-related tuberculosis and by the emergence of multidrug-resistant strains. Highly lethal outbreaks of extensively drugresistant tuberculosis among HIVinfected persons in South Africa [1] demonstrate the public health emergency that results when these two forces converge in the same setting. Fortunately, at this time of great need, tuberculosis drug development has been roused from its decades-long slumber. New ways of using existing drugs and the development of new drug classes hold great promise for the treatment of both drug-susceptible [2][3][4][5][6] and drug-resistant tuberculosis [4][5][6][7][8].
Children are a critical part of the global tuberculosis pandemic, with an estimated 900,000 cases and 100,000 deaths per year [9]. In high-burden settings, children make up as much as 20% of incident cases of active tuberculosis [9,10]. Furthermore, young children have an increased risk of severe, rapidly progressive forms of tuberculosis, such as disseminated disease and meningitis ( Figure 1) [11,12]. Therefore, it is imperative that children benefit from improvements made in tuberculosis treatment. However, children have only been included in one study of these new agents (a phase III trial of once-weekly rifapentine + isoniazid for latent tuberculosis) [13].

Barriers to Involving Children
Although not well articulated in the published medical literature, a number of barriers to the involvement of children have been raised in discussions of tuberculosis drug development (Box 1). Our concern is that these barriers may, once again, lead the field down the path of least resistance-the exclusion of children from tuberculosis drug development efforts.
What happens when children are not included in drug development. There is a rich history of clinical trials for tuberculosis treatment, beginning with the landmark streptomycin trial [14] and followed by a remarkable series of trials establishing that multidrug therapy could be curative, that it was possible to do so with ambulatory treatment, and that therapy could be shortened from two years to six months [15]. Children were almost completely left out of this series of clinical trials.
The result? Nearly 40 years after the development of short-course treatment in adults, there are still fundamental uncertainties about age-appropriate dosing of isoniazid, rifampicin, pyrazinamide, and ethambutol [16,17]. Children, particularly very young children, do not achieve adequate serum concentrations of these agents when given weight-based dosing based on pharmacokinetic data from adults. The uncertainties about pediatric dosing reflect the lamentable paucity of pharmacokinetic data for first-line drugs in children. Another consequence of the lack of involvement of children in the initial phase of tuberculosis drug development is that only in recent years has there been a substantial effort to manufacture child-friendly formulations of firstline tuberculosis drugs (crushable mini-pills, granules, oral suspensions). A number of controversies in the treatment of pediatric tuberculosis stem from the lack of clinical trials focused on child-specific questions (e.g., the optimal duration and dosing frequency of tuberculosis treatment, how to ensure safe use of ethambutol in children). There are even greater uncertainties about less common therapeutic questions (e.g., the treatment or prevention of drugresistant disease).
The history of tuberculosis drug development reflects the lack of involvement of children and the consequences of that omission. Concerted efforts are necessary to avoid repeating this unfortunate experience in this era of renewed interest in tuberculosis drug development.
The example of antiretroviral drug development. Antiretroviral drugs have been evaluated among children, agespecific pharmacokinetic data obtained, and child-friendly formulations developed and marketed. Throughout this process, the evaluation of antiretroviral drugs for children has not lagged far behind their development and licensure for adults. As a result, age-appropriate regimens are available in a range of formulations for children, and the rates of HIV-related morbidity and mortality decreased in concert for children [18][19][20] and adults [21]. The keys to this success were advocacy [22], earmarked funding for pediatric research, focused clinical trial designs on questions whose answers could not be extrapolated from research in adults, and incentives for the pharmaceutical industry to include children in drug development.

Overcoming the Barriers to the Involvement of Children
The first and perhaps the most important step toward involving children in tuberculosis drug development is to clearly articulate the necessity of doing so. To shine a light on the path of least resistance is to show how clearly unacceptable it is; children have the same right to benefit from research as do adults. Researchers, regulatory agencies, advocates, and government agencies and private foundations that fund drug development must insist that the development pathways for all new agents/regimens include specific plans for when and how children will be involved.
Once agreement has been reached on the necessity of including children in trials of new tuberculosis treatment regimens, the specific barriers to the involvement of children must be identified and then overcome ( Table 1). The difficulty of culture confirmation of active tuberculosis among children is well known [17]. Because a positive culture is both an enrollment criterion and the primary endpoint of phase II and III clinical trials for tuberculosis treatment, some observers have concluded that tuberculosis treatment trials cannot be done among children. This limitation is quite real-because of it, pediatric trials are not the setting for the definitive evaluation of the efficacy of a new drug or treatment regimen. However, it does not mean that treatment trials cannot be done among children.
Improved specimen collection techniques (e.g., induced sputum and string test) can provide culture confirmation in a higher percentage of pediatric patients than previously thought possible [23][24][25][26]. Furthermore, case definitions for culture-negative pediatric tuberculosis can be used in clinical trials, as criteria both for enrollment and for evaluating efficacy [27,28]. While imperfect, these case definitions can be applied by an events committee blinded to treatment assignment to ensure unbiased assessment of diagnosis and response to treatment.
Although concerns about pediatricspecific side effects have led some to argue against inclusion of children in clinical trials, these concerns fail to recognize that new drugs will be used off-label among children, in the absence of data on pediatric pharmacokinetics and tolerability, as soon as they are approved for adults. Children are indeed vulnerable participants in research because of their inability to provide fully informed consent. However, an overzealous attempt to protect some children from the possible harms of research perversely causes harm, by either denying access to treatment or through exposing children to the risks of inappropriate dosages of new medications.
Effective therapy is available for drug-susceptible tuberculosis. However, the limitations of current first-line tuberculosis treatment should be recognized. Despite the appeal of our current nomenclature of "shortcourse therapy," a six-month treatment duration leads to worrisome numbers of patients who do not complete treatment in many programmatic situations [29][30][31][32]. The side effects of current regimens are appreciable as well: high rates of bothersome side effects, such as nausea and vomiting, and substantial rates of serious adverse events, such as hepatotoxicity [33]. In spite of the availability of effective therapy for drug-susceptible tuberculosis, new agents should still be evaluated among  At what points in tuberculosis drug development should studies be undertaken in children? This crucial question requires discussion among investigators, the pharmaceutical industry, advocates, and regulatory officials. As a starting point for such discussions, we offer initial suggestions in Table 2. If children are to be involved at specific points in drug development, appropriate timelines are needed for initial work on formulations and pharmacokinetic studies among children. It is inappropriate to wait until the drug development plan for a new drug or regimen has been completed in adults before beginning its evaluation in children.
What kinds of trials should be undertaken among children? Not all phase III trials in adults need repeating in children; it is highly likely that children will respond well to a new regimen if given a drug formulation and dose that achieves pharmacokinetic parameters comparable to those among adults. Questions that are specific to children, or where answers from adults are unlikely to extrapolate to children, require separate evaluation. Key examples of studies that must be done among children are those assessing the pharmacokinetics and tolerability of new drugs [17].
It can be difficult to obtain blood specimens by venipuncture from very young and very ill children. This challenge should not preclude pharmacokinetic studies of new antituberculosis drugs among children; more efficient study designs and techniques for use of ultra-small quantities of blood can overcome this limitation. Sparse sampling schemes analyzed with Bayesian statistical methods that incorporate pharmacokinetic data from adults, in addition to knowledge about maturation of metabolic pathways in children, facilitate the design and implementation of pharmacokinetic studies in children [34].

Barrier Ways In Which Barriers Can Be Overcome
The difficulty of diagnosing active tuberculosis in children Clinical trials of new drugs/regimens can proceed using validated case definitions for active disease, confirmed by blinded events review committees, even though many of the participants will not have culture-confirmed tuberculosis Lack of pharmacokinetic data from children and the difficulty of performing pharmacokinetic sampling in young children Validation of methods for doing pharmacokinetic studies using more easily obtained, very lowvolume samples Identification of local sites that already have the capacity to perform such studies Unwarranted complacency about the effectiveness of therapy for children with drug-susceptible tuberculosis Initial evaluation of novel drug classes among children with proven or suspected drug-resistant tuberculosis Ensure that pharmacokinetics and tolerability of new drugs for drug-susceptible tuberculosis are evaluated among children while their efficacy is being investigated among adults Trial design issues (endpoints, appropriate sample sizes) Collaboration with pediatric trialists in the antiretroviral and cancer treatment fields Concerns about pediatric-specific side effects Appropriate monitoring during clinical trials involving children Concern that involving children in tuberculosis drug development will complicate the research oversight and regulatory aspects of clinical trials US National Institutes of Health) Provision of incentives by some regulatory authorities (e.g., US Food and Drug Administration and EU European Medicines Agency) to encourage the involvement of children in research in particular areas Concerns about diffusing the very limited funds for tuberculosis drug development into too many areas (dividing an already small research pie into too many pieces) Use of the involvement of children in tuberculosis drug development as a basis for more effective advocacy, and a way to increase overall funding for the field ("grow the pie") doi:10.1371/journal.pmed.0050176.t001

Clinical Trial Phase among Adults
Suggested Research Activities among Children I (Single and multiple-dose PK and tolerability among healthy adults) None IIa (Early bactericidal activity and PK-patients with tuberculosis) Initial work on possible formulations for young children IIb (Sputum culture conversion over the first 2-4 months of therapy) Initial PK studies among children with tuberculosis III (Randomized trial with tuberculosis outcomes as the primary endpoint) Randomized comparison of the new drug/regimen with PK and tolerability as primary endpoints; efficacy as a secondary endpoint Expanded access (compassionate use) protocol for children with known or suspected drug-resistant disease and poor treatment options IV (Further evaluation of a regimen shown to be effective in an initial phase III trial) Additional studies among key subgroups of children-those <3 years old, those with central nervous system involvement Validation of PK of formulations and doses chosen for clinical practice PK, pharmacokinetics doi:10.1371/journal.pmed.0050176.t002 An instructive example of optimizing methodologies for pharmacokinetic studies for children comes from malaria research. Severe malaria is predominantly a disease of young children, so it is critical that the pharmacokinetics of antimalarial drugs among infants and young children be well understood. In these studies, very small samples of blood (100 microliters) obtained by finger-or heel-stick are blotted onto filter paper, allowed to dry, stored at room temperature, and later used to determine concentrations of antimalarial drugs [34,35]. Thus, pharmacokinetic studies can be extended to infants and young children, and such samples can be obtained under field conditions where the disease is common. The extension of these techniques to the study of new antituberculosis drugs requires further research, but the challenge of pharmacokinetic sampling among young children calls for this kind of innovation.
There are many regulatory steps between the development of a study protocol and its implementation at study sites. Some have expressed concern that involving children in drug development will slow down the already lengthy timeline for study implementation by triggering more rigorous regulatory review. This concern may be well founded. However, key funding and regulatory agencies have policies and incentives to encourage evaluation of new drugs among children for conditions, like tuberculosis, that are common among that age group (the US National Institutes of Health requires specific justification if children are not included in a study). Some of those incentives, such as extended patent protection for the United States market, are unlikely to encourage tuberculosis drug development efforts for children. However, the US Food and Drug Administration has recently been authorized to take additional steps to promote the inclusion of children in drug development [36], and the European Union now requires that any new drug that could potentially be used in children have a Pediatric Implementation Plan for the drug to be licensed in adults [37]. Other groups are developing alternative incentive structures for diseases of poverty, such as tuberculosis [38].
Finally, there is the concern that including children in clinical trials will dilute the already inadequate funding for tuberculosis drug development [39], thus slowing down the pathway to licensure of new drugs. Proponents of this zero-sum argument may be willing to face difficult facts, but risk fostering the continued existence of an unacceptable situation. The expansion of antiretroviral therapy and the treatment of multidrug-resistant tuberculosis in high-burden settings are examples of two interventions which were said to be inadvisable, based on zero-sum arguments, but which have now been shown to be both feasible and critical for disease control [40][41][42]. Rather than being a detriment to funding for tuberculosis drug development, the inclusion of children may draw funding to the field.

Summary
We are on the threshold of revolutionary improvements in the treatment of tuberculosis. Within five to ten years, it is likely that highly effective three-month regimens will be available to treat both active and latent drug-susceptible tuberculosis. New drug classes that have the potential to dramatically improve the treatment of multidrug-resistant tuberculosis are entering clinical trials. Children have the same right as adults to benefit from research with these new treatments. By making a deliberate choice to avoid the path of least resistance, we can ensure that both adults and children benefit from these advances in tuberculosis treatment.