Current drug approval regulations in Europe and the United States require that the treatment group demonstrates a clear benefit compared with the control group. These laws should be updated to reflect patient heterogeneity in clinical trials, and allow for approval of drugs that show efficacy in only a subset of treated patients.
At the May 2014 Herrenhausen Symposium on Autism and Neurodevelopmental Disorders, held in Hannover, Germany, and organized by Nature Medicine and the Volkswagen Foundation, a group of leaders in the field of autism discussed hurdles blocking autism translational research and drug development and proposed practical methods to overcome these challenges. Much discussion focused on the striking patient-to-patient heterogeneity of autism spectrum disorder. This affliction has been linked to a long list of genetic and environmental factors, and it manifests in a variety of symptoms of different severities in each affected individual. This heterogeneity in etiology and clinical symptoms is likely to also extend to heterogeneity in treatment responses to experimental drugs. However, even though the problem of disease heterogeneity applies across many disorders, from cancer to autoimmunity, current regulations at the European Medicines Agency and the US Food and Drug Administration typically do not allow investigational drugs to be approved on the basis of a subpopulation of responders in a clinical trial. Instead, approval of drugs requires that the group of patients receiving the experimental treatment demonstrates a clear statistically significant benefit compared with the control group. Several Herrenhausen Symposium attendees therefore proposed that patients and doctors need to lobby to convince legislators to change the laws such that drugs that demonstrate efficacy in only a subset of patients in a clinical trial—and are safe—may still be approved.
The US Congress seems ready to listen. On 30 April, the House Energy and Commerce Committee announced an initiative called 21st Century Cures, with the goal of accelerating the pace of medical breakthroughs. The committee recognizes that the current drug approval process is a “relic of another era” and plans to spend the next few months closely examining it in collaboration with the National Institutes of Health, the Food and Drug Administration and other stakeholders by commissioning white papers and holding congressional hearings. The committee members say they want to bring regulatory policies closer to advances in scientific knowledge (http://energycommerce.house.gov/cures). One clear way to achieve this goal would be to bring an appreciation of patient heterogeneity in treatment responses into the regulations that guide the drug approval process.
For example, new methods for analyzing the outcomes of clinical trials may help reveal responses in subsets of patients. Of course, because drug companies that initiate clinical trials have a vested interest in the approval of their investigational drugs, regulatory agencies require that trials be designed and analyzed using such methods as group randomization and predetermination of trial endpoints to avoid bias. Therefore, any change to the current regulations must introduce flexibility while also avoiding potential biases. However, over the past several years, many groups have been trying to devise analytical methodologies that grapple with the issue of heterogeneity of treatment effect in clinical trials yet still prevent the introduction of bias. For example, some researchers have proposed including plans for subgroup analyses in clinical trial designs, with primary analyses that would be based on prior knowledge and secondary analyses that would be more exploratory in nature and could inform future trials in that patient population (Trials 11, 85, 2010 ). Although regulatory agencies do have clear procedures in place to provide feedback to organizations planning clinical trials that include these or other new methodologies, the extent to which these novel methodologies could increase the probability that an investigational drug would be approved remains to be seen.
Another way to tackle the problem would be to reduce patient heterogeneity in clinical trials altogether by targeting investigational drug treatments to a subset of patients known to have a common cause for their disease. However, for many diseases, including schizophrenia and autism, there is still no clear way to subclassify most patients. Delaying the inception of clinical trials targeting these diseases until subpopulations are more clearly understood would seem necessary under the current regulatory system but would not benefit patients currently struggling with these disorders.
Even in scenarios permitting selection of patients with a common disease etiology, symptom heterogeneity can still complicate clinical trials. For example, for the types of autism for which a genetic cause is known, affected individuals display such a wide variety of symptoms of differing severities that it can still be very challenging to observe clear efficacy of an investigational drug in the treated group of patients compared with controls. Last year, the biotech company Seaside Therapeutics terminated development of arbaclofen, a drug targeting the GABA receptor, for use in idiopathic autism and fragile X syndrome when it failed to show a clear beneficial response when comparing all treated patients with all controls (http://www.nytimes.com/2013/06/07/business/an-experimental-drugs-bitter-end.html). Nevertheless, around 30% of patients with fragile X did demonstrate some degree of benefit—and some showed worsening. Additional research into the physiological basis of these varied responses is necessary, yet some families whose children received the drug are still seeking to obtain the treatment—some have even petitioned Congress to this end. This situation suggests that the regulatory environment needs to change to permit drugs that fail under the current regulatory system to reach patients who may benefit. One way to achieve this would be to allow safe drugs that show only partial population responses to at first be approved on a provisional basis and to require the collection of additional safety and efficacy data for eventual full approval.
For some illnesses, such as depression, patients may need to try many different medicines before they find one that helps them. But they have many options. For diseases such as autism, approving drugs that are safe in all treated patients but effective in only a subset—even without an understanding of the molecular and cellular basis for the subpopulation-selective effect—would hasten the development of an armamentarium of safe drugs that patients with autism can serially try until they achieve some degree of benefit for their disorder. But this can only happen if patients and doctors lobby legislators to change the current regulatory environment and enhance the flexibility of the rules that govern the drug approval process.
About this article
Cite this article
Embracing patient heterogeneity. Nat Med 20, 689 (2014). https://doi.org/10.1038/nm.3632