Mining drug data reveals thousands of sneaky side effects

When you’re taking several drugs at once, it’s difficult to predict all the side effects of those drug-drug interactions.

And that’s because clinical trials are conducted on small groups of people, and only common effects that are detected with enough confidence are listed on a drug’s package insert. Also, these trials don’t routinely investigate interactions since their main goal is to establish the safety and efficacy of a single drug.

This week, researchers revealed a computer algorithm that trawled through data on hundreds of thousands of drug effects and interactions from Food and Drug Administration databases – and from that, they created two new databases that may offer guidance.

“The average 70-year-old is taking seven different prescription medications,” Stanford’s Russ Altman says in a news release. “The FDA has a database for patients and physicians to report possible adverse drug events, but it’s very difficult to uncover true side effects because people vary in their medical histories, conditions and drug regimens, as well as in age, gender and environment.”

He adds: “Some researchers have gone so far as to say, ‘No one will ever get useful information out of all of this data.’”

But once a drug hits the market, things can get messy as unknown side effects pop up. And that’s where Altman’s algorithm comes in, Nature News explains.

1. Reports of adverse drug events are notoriously prone to bias. For example, cholesterol-lowering treatments are often taken by older patients, so conditions associated with aging (such as heart attack) could be wrongly linked to the drug as a side effect.
2. To reduce this bias, the team developed an algorithm to match data from drug-exposed patients to non-exposed control patients with the same condition – automatically correcting for several known sources of bias, including those linked to gender, age, and disease.
3. Then they used this method to compile a database of 1,332 drugs and possible side effects that weren’t listed on the labels for those drugs.
4. The algorithm came up with an average of 329 previously unknown adverse events for each drug – far surpassing the average of 69 side effects listed on most drug labels.

Among the discoveries: patients taking certain a combination of antidepressants and drugs to treat hypertension are almost one-and-a-half times more likely to develop abnormal heart rhythms, compared with patients taking either drug alone. (These predictions were confirmed with e-records from the Stanford University Hospital emergency room.)

“It’s a step in the direction of a complete catalogue of drug–drug interactions,” Altman says.

The study was published in Science Translational Medicine this week.

[Via Nature News, Stanford University]

Image: data-mined drug-drug interactions / Science/AAAS