Updated

Studies of schizophrenia drugs are increasingly finding lesser effects because more patients are responding to drug-free placebos used for comparison, according to a new United States government study.

Drugs used to treat schizophrenia are known as antipsychotics. For some people, the drugs can suppress symptoms like hallucinations and delusional thinking, and allow them to live a more normal life.

But it's well known that antipsychotics do not work for everyone. And their side effects can be so serious -- including sedation, substantial weight gain and diabetes -- that people often stop taking them.

What's more, recent clinical trials of second-generation antipsychotics -- which emerged 20 years ago and now dominate the market -- have been finding smaller treatment effects compared with trials from the early 1990s.

"Treatment effect" refers to the difference between clinical trial patients who get the real drug and those given a placebo for comparison.

For the new study, researchers at the Food and Drug Administration (FDA) looked at 32 clinical trials that were submitted to the agency between 1991 and 2008. The trials were all part of companies' applications to the FDA for approval of a new drug for schizophrenia.

The researchers found North American trials done in more recent years turned up smaller treatment effects than older studies.

But it was not because drugs in the newer studies were less effective, said Dr. Thomas P. Laughren, head of the FDA's division of psychiatry products and one of the researchers on the study.

What changed was study patients given placebo pills started showing bigger responses.

Exactly why is unclear. One possibility, Laughren said, is that schizophrenia patients in clinical trials might be less sick than in the past -- and those people may be more likely to improve even if they are on a placebo.

But Laughren said more research is needed to figure out what's really going on.

More trials failing

The trend is concerning, according to the FDA researchers, because clinical trials with big placebo responses are more likely to fail -- meaning they don't show a statistically clear difference between the treated group and the placebo group.

The more often trials fail, the less interested drug companies will be in even trying to develop new schizophrenia drugs, Laughren said.

And new treatments are needed, he noted.

The second-generation antipsychotics now on the market include Risperdal, known generically as risperidone; Zyprexa (olanzapine); Abilify (aripiprazole) and Seroquel (quetiapine).

When they first arrived, the new antipsychotics were seen as better alternatives to older antipsychotics, such as Haldol (haloperidol) and Thorazine (chlorpromazine) -- which can cause substantial problems, like uncontrollable body movements.

But the new generation carries its own list of potential side effects, including an increased risk of heart disease. The drugs also cost far more than their predecessors.

Trials in the new FDA study included 21 done solely in North America. The rest were done in multiple regions.

When the researchers looked at the North American trials, they found patients' responses to the real study drug remained stable over time. On average, their scores on a standard symptom scale, measured from 30 to 210, started at 89 and declined by 13 points over four to eight weeks.

In contrast, placebo responses changed over the years, the researchers report in the Journal of Clinical Psychiatry.

In trials done between 1991 and 1998, placebo patients' symptom scores dipped by only two points, on average. But in trials done between 1999 and 2008, the average decline was seven points.

That meant patients on the real treatment were still doing better. And overall, 75 percent of trials done between 1999 and 2008 were deemed a success -- meaning the drugs had a statistical advantage over the placebo.

But trials done in earlier years had a higher success rate, at almost 85 percent. (The studies done outside of North America all took place in later years, so the FDA team could not make the same comparison.)

According to Laughren, the findings do not mean schizophrenia drugs are ineffective, or that new ones will be approved on flimsier evidence.

"We haven't changed the standards for approving drugs," he said.

But, Laughren said, it will be important to uncover the reasons why placebo responses are rising. The FDA researchers are going to take a closer look at data on individual patients involved in these trials, to see if there are clues.

From there, Laughren said, it might be possible to design better studies.

That could mean anything from changes to how studies measure symptoms to how long they follow patients, he noted.