Experimental Study Design Could Lead to Improved Drug Safety

Trend-in-trend model may predict safety concerns of new drugs.

Randomized clinical trials are typically the gold standard when determining the safety and efficacy of an investigational drug; however, clinical trials can have significant flaws.

Clinical trials do not always account for real-world situations, which can cause certain side effects to remain hidden at the time of approval. Additionally, the sample size of some clinical trials can be too small or may not include patients who are taking other drugs or who have comorbidities. First-in-class drugs may also not have a comparator drug, which can lead to inaccurate results.

A new research construct published by Epidemiology may offer improvements to drug design that may not be accounted for using current methods.

For the new trend-in-trend study design, investigators monitor changes in frequency of outcomes related to drug exposure in groups that initiate the drug at different rates. The authors believe that this method could be used to explore newly approved drugs or those with a growing or decreasing population taking it.

This new method may be welcomed, as the FDA is approving drugs faster than ever before, specifically medications granted priority review. Since the FDA is quickly approving drugs, it is likely that many will have safety issues at approval. In fact, one-fifth of drugs approved receive a black box warning, and 4% are removed from the market due to safety concerns.

In the study, the authors use the experimental method to examine the painkiller rofecoxib (Vioxx). They investigators were able to reproduce the known link between rofecoxib and heart attack using the novel approach, according to the study.

When the drug was first approved, there was a large uptake; however, the drug was quickly abandoned when safety concerns were discovered. The authors wrote that this cycle makes rofecoxib an ideal drug to explore the efficacy of the model.

The investigators also said the trend-in-trend model can be used to look into popular drugs that have yet to be examined, according to the study.

“Epidemiologic studies can get the wrong answer if there are differences between people who take the drug and people who don’t take the drug,” said senior author Sean Hennessy, PharmD, PhD. “This kind of study is immune to that bias, because it’s not comparing users to non-users, it’s looking at trends in the frequency of outcome as a function of trends in the frequency of exposure. Even when there are unmeasured factors that are different between groups and those factors affect the outcome — this study will give the correct answer.”

The authors will next look at any cardiac risks linked to testosterone supplements since they became extremely popular and then unpopular. This would allow investigators to determine the cardiovascular effects of testosterone, which are currently unknown due to a limited number of clinical trials, the study concluded.