According to a new look at past antidepressant trials, up to a fifth of patients on Cymbalta and similar medications may actually do worse than those given drug-free placebo pills.

Researchers found that patients' symptoms over the first couple months of antidepressant use separated them into "responders," who got progressively better, and "non-responders," who didn't improve with treatment but may still have suffered side effects.

However, "It's difficult to say a priori who will be in which group," Ralitza Gueorguieva, the study's lead author from the Yale University School of Medicine in New Haven, told Reuters Health.

The findings highlight the importance of identifying as soon as possible which patients will and won't respond to certain drugs, her team said.

The researchers combined data from seven studies that randomly assigned patients to receive Eli Lilly's drug Cymbalta (known generically as duloxetine), other antidepressants, or a placebo pill for two months. Those trials involved a total of about 2,500 people with major depression.

People getting the placebo tended to report small, gradual improvements in depression symptoms. On the other hand, those on Cymbalta or another antidepressant fell into one of two categories: most had steeper, steady improvements in depression symptoms, but a sizeable chunk didn't seem to get any better.

About four in five patients on all antidepressants were responders. For Cymbalta in particular, about 84 percent of patients improved and 16 percent did not.

Medication responders saw significantly bigger improvements in their depression symptoms than patients assigned to the placebo. Non-responders, however, did worse.

Differences between antidepressant responders and non-responders were seen as early as a week or two into treatment, and the researchers wrote in their Archives of General Psychiatry report that initial improvements seem to predict who will have a better outcome on Cymbalta, along with the other drugs.

"You know within the first couple weeks of starting a treatment who's the most likely to benefit because they're already starting to show improvement," said Dr. Michael Thase, a psychiatrist from the University of Pennsylvania Perelman School of Medicine who wasn't involved in the new study.

"The first few weeks are revealing, and obviously if the patient's getting worse instead of better, I would use that as a strong indicator that this particular treatment isn't likely to work," he told Reuters Health.

"I think this finding holds true for the antidepressants that are most commonly used today," he said of the gap between responders and non-responders.

Thase pointed out that side effects of antidepressants, such as stomach problems and poor sleep, could make some patients score lower on measures of depression, perhaps explaining the worse symptoms seen in non-responders compared to placebo patients.

He added that if patients don't benefit from a first antidepressant, they could still respond to a different type of drug, although the chances fall with each successive treatment attempt.

Another scientist said the latest research has been trying to pick out certain patient characteristics, genetics or specific depression and anxiety symptoms, for instance, that could help determine who will end up in the responder category, and who won't see any benefit from individual drugs.

"If you can identify people who would be potential responders to a particular medication...it would be a great, huge advantage for the field," said C. Hendricks Brown, who has studied depression treatments at the University of Miami Miller School of Medicine, but wasn't linked to the new research.

Gueorguieva agreed. "Identifying variables that are associated with response is a very important question that we haven't quite tackled," she said.

One of the study's authors is an employee of Eli Lilly and another is on the company's scientific advisory board.