In order to put a drug on the market, clinical trials are required which culminate in a large so-called phase 3 trial. This requires the recruitment of thousands of patients and compare those who receive the treatment versus a placebo. There must be a statistically significant difference between treated and placebo on 1 criterion called the main criterion, which implies that most treated patients respond to the treatment. However, in complex syndromes such as autism, this is virtually impossible due to the heterogeneity of the syndrome and clinical symptoms. In addition, it happens that so-called secondary criterion – subsets of clinical criteria – are positive but not the main criterion which is the average of numerous parameters. Thus, autism evaluation criteria such as SRS or CARs include dozens of sub-criteria evaluating agitation, emotivity, sensory problems, social interactions and so forth.
Astonishingly, often, the final phase 3 fails while large number of phase 2 trials made on a smaller number of patients succeed. Thus, a meta-analysis of published data has collected data from 9 phase 2 trials performed in France (our initial trials published in 2012 &2017) Holland, China, Egypt and Tunisia confirm that Bumetandie (burinex). There is a significant improvement in 1000 children of the core symptoms of autism in particular social communication and agitation.
In my recent paper, I discuss the possible reasons for the opposite results of phase 2 and 3. In addition to the intrinsic heterogeneity of neurodevelopmental disorders and in particular autism, there are the complex pedo-psychiatric parameters analyzed to determine the disorder and its possible improvement. The evaluation of success in phase 3 relies on a single parameter which is in fact a mean of many parameters (15-20 and more), an improvement in one of these sub-parameters that characterizes a sub-population of children with autism will be unnoticed challenging the requirements of European and Us Authorities and demanding more subtle requirement that might fit for a subpopulation of children- but this of course is less favorably accepted by the pharmaceutical industry. As a result, pharmaceutical companies more frequently nowadays will first identify responders, estimate the financial implications of this limited use of the treatment before embarking on very expensive phase 3. Finally, in any case this approach does not respond to the situation in which treatment reduces the severity of the syndrome and more importantly ignores the % of patients for whom this treatment is efficient. This article reviews this debate taking autism as an example although it is valid for many disorders.
