Almost a third of new therapeutic drugs released onto the US market with Food and Drug Administration (FDA) approval subsequently have what are called “safety events”. Most often, these lead to warnings on packets or to doctors indicating the drug should not be taken by people with certain rare conditions or under particular circumstances. In only a tiny minority are the events serious enough to require withdrawal of the product.
Drug approval is a never-ending balancing act. Make trials too comprehensive and people are left waiting for a product that might save, or at least greatly improve, lives. Yet if you don't test widely enough, you risk another thalidomide – a drug with great benefits for many people, but devastating during pregnancy.
Dr Joseph Ross of Yale University explored one aspect of how our current efforts are working by looking at what happened after the release of 222 products granted approval by the FDA between 2001 and 2010. The findings have been published in the Journal of the American Medical Association.
How you view Ross's findings depends on whether you are a glass half-empty or glass half-full type of person. For those inclined to negativity, problems emerged with 32 percent of the therapeutics, leading to changes in the terms of their approval. Since some of these had more than one issue arise, there were a total of 123 safety events recorded. The number is likely to grow for the pharmaceuticals and biologics (approved therapeutics that are biologically derived rather than synthesized) released towards the end of the period.
More positively, in only three cases (1.4 percent) was the issue so serious or widespread that the product had to be withdrawn. Another 61 warnings were added to the products' boxes, while 59 involved safety communications to prescribers. The average gap between approval and the first post-market safety event was 4.2 years.
Drug trials can only involve so many people without becoming prohibitively expensive. Most involve fewer than 1,000 participants. Unless there are grounds to suspect the drug will be a problem for people with a rare condition, it's unlikely many people with that condition will participate in the trial. So if the drug has negative side effects for just that group, this probably won't be detected during clinical trials. Once a product is on the market however, and given to hundreds of thousands or millions of people, patterns can start to emerge in which groups experience side effects.
Consequently, it is unsurprising many boxes will need warnings added to them, or doctors instructed that the drug is contra-indicated for certain people. Ross's findings can help feed a debate on whether the current rate is acceptable or if we should be spending more to make trials larger.
