Researchers have developed a computer algorithm that can comb through FDA's adverse event database to identify previously unknown drug side effects, as well as adverse events that could be caused by certain drug combinations, according to a study published in the journal Science Translational Medicine, Nature reports (Ledford, Nature, 3/14).
About the Algorithm
For the study, Stanford University researchers developed an algorithm that mined information from FDA's database of patient- and physician-reported adverse events (McBride, FierceBiotechIT, 3/15).
According to the researchers, adverse event reports often contain bias because people taking the drugs could have multiple medical conditions. For example, adverse event reports might incorrectly link cholesterol-lowering drugs with heart attacks because the population taking cholesterol medication -- typically older patients -- tend to be more prone to heart attacks.
To eliminate such bias, the algorithm matched data from each drug-exposed patient to a nonexposed patient with the same illness. The method helped researchers create a database of 1,332 medications that listed potential side effects of the drugs. Researchers found an average of 329 previously unknown side effects for each drug.
Database on Drug Pairs
The researchers also used the algorithm to create a database of potential interactions between pairs of drugs (Nature, 3/14).
Using the database of drug pairs, researchers found that people who were taking a class of antidepressants known as selective serotonin reuptake inhibitors -- or SSRIs -- were significantly more likely to develop a dangerous heart condition called prolonged QT if they also were taking a class of high blood pressure medications called thiazides ("80 Beats," Discover Magazine, 3/15).
Researchers then searched through electronic health records from Stanford University Hospital to confirm the interaction between SSRIs and thiazides. They found that patients were one-and-a-half times more likely to have prolonged QT if they were taking both drugs than if they were taking either drug by itself.
Possible Uses for FDA
Russ Altman -- the study's lead author, a bioengineer at Stanford University and an adviser on FDA's Science Board -- said he plans to present the results of the study to FDA. He said FDA could use the algorithm in conjunction with its existing drug-surveillance programs to eliminate bias when searching for potential adverse events.
However, Altman acknowledged that it might be challenging for FDA to address all of the new data generated by the algorithm. He said, "We've just released a database with 10,000 or more adverse events. I do not expect the FDA to uncritically take these results and add them to every drug label" (Nature, 3/14).