Expanding Treatment Options for Youth With Type 2 Diabetes

Current Problems and Proposed Solutions: A White Paper From the NICHD Diabetes Working Group

William V. Tamborlane; Morey W. Haymond; David Dunger; Ravi Shankar; Rose Gubitosi-Klug; Kathleen Bethin; Janina Karres; Paolo Tomasi; Ingrid Libman; Paula H. Hale; Ronald Portman; Georgeanna Klingensmith; Michael Reed; Jeffrey Blumer; George Giacoia


Diabetes Care. 2016;39(3):323-329. 

In This Article

Part 2: Innovative Approaches and Novel Study Designs

The Case for Extrapolation

Both the EMA and FDA have defined their concepts and necessary criteria to extrapolate the efficacy of medicinal products from adults to the pediatric population.[13] In general, to permit extrapolation, the disease must be sufficiently similar in both populations. In addition, similar responses to intervention and similar exposure-response relationships in the adult and pediatric populations would have to be substantiated. Although the basic pathophysiology of insulin resistance and progressive β-cell dysfunction is similar in adolescents and adults,[14,15] there appears to be a faster decline in β-cell function in youth than in adults with T2D, a suggestion that is supported by the higher-than-predicted failure rate of metformin monotherapy in the TODAY study.[2] In addition, due to the hormonal changes of puberty,[16] obese adolescents with T2D may be more insulin resistant than adults with T2D. The FDA and EMA have judged that there is insufficient evidence regarding the similarities between adolescents and adults with T2D to justify full extrapolation of efficacy from studies in adults to adolescents.

Partial extrapolation of efficacy can be used when uncertainty exists about the assumptions underlying full extrapolation. Partial extrapolation of efficacy can range from requiring a single adequate and well-controlled phase 3 study (as opposed to the two separate trials required in adults) to requiring only a PK/PD exposure-response study that shows similarities in exposure-response relationships between adult and pediatric patients. Safety data would also need to be collected at the recommended dose(s).

Virtually all of the recent pediatric T2D program agreements accept partial extrapolation of efficacy from adults by allowing a single phase 3 study in the pediatric population compared with the requirement for at least two separate studies. It remains to be determined whether the concept of partial extrapolation can be further extended to reduce the number of pediatric patients required in pediatric programs without compromising the adequacy of the pediatric efficacy assessment.

Studies of Drug PK and PD in Youth With T2D

It is standard procedure for one of the first clinical studies in the pediatric development of a drug to be a phase 1 pediatric study to assess drug PK based on the assumption that growth and developmental changes in factors influencing absorption, distribution, metabolism, and excretion will lead to changes in PK measures and parameters. To achieve drug exposure values, e.g., area under the drug concentration curve and peak drug concentration, in children similar to values associated with effectiveness and safety in adults, it is considered necessary to evaluate the drug PK over the entire pediatric age range in which the drug will be used. However, as almost all pediatric patients with T2D are ≥10 years of age, some of the complexities challenging clinical pharmacology studies in younger children are avoided.

In early studies, the PK of metformin,[17] glimepiride,[18] and pioglitazone[19] in adolescents with T2D were not different than the PK of approved doses of the drugs in adults. More recently, the doses of newer drugs such as exenatide, liraglutide, and sitagliptin that are being used in pivotal safety and efficacy trials in pediatric patients with T2D were also shown to be very similar to the doses approved for use in adults (Table 4). These data suggest that the PK properties of T2D drugs in adolescent patients are not significantly different from that in adults. Similar drug exposures in adolescents compared with adults with T2D are likely to be related to the age, pubertal development, and the marked obesity in this patient population. Whether increases in glomerular filtration rate or other physiologic factors contribute to increased rates of drug clearance in youth with T2D has not been established.

Depending on the sample size, performance of dedicated PK studies in this difficult-to-recruit population usually takes between 1 and 2.5 years to complete, as only about one subject per month successfully completes the study (Table 3). Due to the difficulties with performing a standard phase 1 study in this population, alternative approaches have been considered. One alternative includes the adoption of modeling and simulation to predict the pediatric exposure and dose with confirmation of exposure of the predicted dose obtained through sparse sampling for PK parameters embedded within the pivotal pediatric study to assess safety and efficacy. This approach is now being accepted by the FDA; it eliminates the need for a dedicated phase 1 pediatric study to assess PK, with significant savings to both time and cost.

New Study Designs of Pivotal Randomized Clinical Trials

Leaders at the EMA and FDA have recognized that innovative approaches to the design of regulatory studies that address feasibility and recruitment issues in pediatric T2D are needed.[5] An innovative approach presented by James Wason, MRC Biostatistics Unit Hub for Trials Methodology Research, Cambridge, U.K., at the EMA Workshop in 2013 would use simultaneous multi-agent/multicompany pivotal trials carried out by independent networks of leading academic pediatric diabetes treatment centers in the U.S., Europe, Australia, and elsewhere.[11]

Multi-agent clinical trials could be restricted to products from the same pharmaceutical company or involve multiple companies; similarly, they could include products within the same class or in different classes. For regulatory purposes, the study design would only test each new treatment against the control treatment, so that the safety and efficacy of each individual drug can be established. There would be no regulatory requirement to test the potential superiority (or noninferiority) of one new treatment against another new treatment.

The use of multiple experimental treatment groups versus a single, shared control group will substantially lower the total number of subjects needed to complete these studies compared with current studies that use a separate control for each experimental group. For example, a four-arm trial would reduce the number of subjects that would be required for separate trials by 33% because the number of control subjects would be cut by 67%.

Multi-agent rather than the current single-agent approach could be used in a number of different study designs that would make as many subjects as possible eligible for inclusion.

Subjects Who are Poorly Controlled (A1C >7.0% or 6.5%) on Treatment With Metformin and Insulin Alone or in Combination. These would be trials of the safety and efficacy of add-on therapy with experimental agents versus placebo. The primary efficacy outcome would be superiority in lowering of A1C versus placebo after a specified period of time.

Subjects who are Well Controlled (A1C <7.0 or 6.5%) on Treatment With Metformin Alone. Trials of Experimental Agents as Monotherapy of T2D: In these studies, subjects who are well controlled on metformin alone would be randomized to either remaining on metformin or switching to one of the experimental agents. The primary efficacy outcome would be noninferiority in the difference in A1C levels versus the metformin group after 12 months.

Trials of Early Combination Therapy in Adolescents With T2D: As shown in Fig. 1, in these clinical trials, subjects who are well controlled on metformin alone would be randomized to either metformin plus placebo or combination therapy of metformin plus an experimental agent. The concept of studying early combination therapy in patients with T2D who are well controlled on metformin alone rather than waiting for the failure of metformin monotherapy has already been established in youth in the TODAY study and in adults in the Glycemia Reduction Approaches in Diabetes: A Comparative Effectiveness Study (GRADE) study.[20] As in the TODAY study, long-term efficacy, as defined as time to treatment failure, would require longer follow-up periods (e.g., up to 3–4 years) to complete.

Figure 1.

Example of a multi-agent study design of early combination therapies of pediatric T2D modeled after the TODAY study. DPP-4, dipeptidyl peptidase 4; GLP-1, glucagon-like peptide 1; SGLT2, sodium–glucose cotransporter 2.

From a patient perspective, multi-agent trials increase the likelihood of getting randomized to an active treatment arm rather than a control arm as compared with a single-agent trial. An advantage for pharmaceutical companies could be to benefit from each other's know-how in terms of recruitment strategy and trial expertise. Companies would no longer have to compete against each other to recruit the few available patients and make their individual study workable but would collaborate for one joint study. It is important to note that, in Europe and the U.S., the extension of patent protection depends only on the successful completion of the agreed development plan rather than whether or not the plan leads to a pediatric indication.

Of course, multi-agent clinical trials involving multiple companies also present significant challenges, which would need to be overcome. Which agents would be included and how would the study costs be divided? Although a three-arm study involving almost 700 pediatric T2D patients was successfully completed by the TODAY Study Group, it remains to be determined whether such a large multi-arm, industry-sponsored study would be as successful.

Expanding the Pool of Eligible Subjects

Additional steps can be taken to increase the pool of subjects who are eligible for these studies. There is a strong rationale for including subjects up to 21 or even 25 years of age who had the onset of their disease prior to 18 years of age, as it has been extremely rare for such patients to be included in adult T2D pivotal trials. Moreover, as a result of the February 2013 EMA workshop on PIPs in T2D, the EMA has indicated that "young adults behave more like adolescents than adults".[11] It is also noteworthy that 21 years of age is considered the upper limit of age for pediatric studies of medical devices.

New Organizational Structures

New organizational structures are needed for the development and implementation of clinical trials for approval of new drugs for youth with T2D to replace the current system of companies competing against one another and even against themselves when they have more than one agent to study. This approach will involve close cooperation of all stakeholders, including pharmaceutical companies, investigators at pediatric diabetes research centers, and regulatory agencies. In the U.S., the NIH could also play a role as a facilitator.

Pharmaceutical industry partners could provide the financial support to develop a network of investigators and centers with recognized expertise in performing clinical studies in pediatric T2D in the U.S., the European Union, Australia, and elsewhere, as well as covering the costs of performing the studies. Within the European Network of Pediatric Research, a Diabetes and Endocrinology network is currently being established and the PDC in the U.S. has recently received additional support to increase the number of its centers to more than 35. Similar to some of the NIH-supported clinical trial groups, such as the Diabetes Research in Children Network (DirecNet), these consortia could use a hybrid approach to funding: a relatively modest amount of committed support might be provided at each site to help build the research team and this support will be supplemented by the number of subjects who are enrolled in active treatment trials. The longer study periods of the new study designs and the steady stream of new therapeutic agents that require investigation for the foreseeable future will ensure the financial stability of the centers.

Many of the regulatory and financial hurdles can be reduced by a more coordinated approach using an independent central coordinating center. The coordinating center will have the responsibility for the preparation of regulatory documents, share the collective network experience in responding to regulatory concerns, monitor the institutional review board approval, certify the sites and investigators, implement the standard research agreements across the network, and assist in the development of study budgets.

Ideally, the clinical center principal investigators and the coordinating center(s) for each study will share the responsibility for the development of study protocols in conjunction with input from the pharmaceutical industry, the FDA, and the EMA. As much as possible, guidelines from both of these agencies regarding the essential study elements would take into consideration the expert recommendations of study investigators and be harmonized between the two agencies. A steering committee composed of all the stakeholders will establish subcommittees that will be responsible for the monitoring the recruitment of subjects, implementation of study protocols, proposals for ancillary studies, preparation of publications and presentations, and evaluations of the performance of individual centers.

As each subject completes the randomized phase of each of these trials, they will be invited to continue to be followed in a T2D registry and clinic network made up of all of the clinical centers that participated in the study for the collection of additional postapproval safety and efficacy data.

Pipe Dream or Future Reality?

The authors of this white paper remain committed to the idea that appropriate studies should be done to provide the evidence needed to secure an indication for the use of these drugs in youth with T2D where appropriate rather than to advocate for off-label use of these agents in pediatrics. We have outlined a number of approaches to aid and improve the collection of pediatric T2D safety and efficacy data that could help to facilitate the approval of new, safe, and effective glucose-lowering agents for youth with T2D. We all believe that these ideas and other novel approaches can provide a real solution to the problems we currently face in providing the best possible care to youth with this difficult condition.