Understanding drug attrition is worth billions. Essentially, every Phase III trial is worth at least 1 billion dollars, taking into account the cost and net present value of the drug candidate (NPV). Most of this value is lost with failure to reach the endpoint or termination of the trial for safety, even taking into account the rare occurrence of trial rescue.
The surprisingly few audit-type reviews of recent late-stage drug development efforts to learn from past mistakes may be due to the secretive nature of the drug development process. Much of the information is private or privileged within the companies pursuing the drug. Often this data never comes to light.
Fortunately, Parasrampuria and colleagues did some heavy lifting for our benefit and distilled the failures down to two reasons: 1) avoidable mistakes, which they termed “...rigorous...decision-making and...discipline” and 2) insufficient scientific foundation, which they termed whether the science was “mature” enough or not. For recent clinical trials, the avoidable mistakes included the recent appreciation of patient stratification or focusing on the individuals best matched to the drug. Poor understanding of the pharmacokinetics and pharmacodynamics were also common avoidable mistakes as were failure to properly model or simulate the trial in advance.
Testing a drug candidate in Phase III for the most pressing unmet medical needs almost always involves going to clinical trial before the biology of the disease is fully understood. This, of course, was the source of most of the unavoidable mistakes. Accordingly, the neuroscience and oncology areas were the most frequent transgressors as these have the most complext pressing unmet medical needs. For example, these failures have been seen uniformly in clinical trials for Alzheimer’s Disease. This aspect is, of course, the hardest for which to develop best practices, except perhaps to have the discipline to avoid these diseases until the science is sufficiently established. Unfortunately, these diseases are the biggest markets and waiting for the science is the surest way to miss the boat.
While some of the unavoidable, “immature” science failures were due to lack of efficacy, most were due to “unpredictable” adverse events. Interestingly, late-stage safety events were most commonly tissue-specific, such as neuro- or hepato- toxicity. Genecentrix makes tools to profile drug off-targets in tissues, such as brain areas and the liver. To us it seems unimaginable not to check for the expression of your drug target, or its entire polypharmacologic ensemble, in these tissues, when our tools make it so easy to do so.