Some people who took a new schizophrenia drug for a year got better with only a few side effects, but many dropped out of the research, the company said Thursday.

The results highlight the difficulties in treatment schizophrenia, a serious mental illness which can cause people to hear voices, feel paranoid, and withdraw from others. High dropout rates are typical in drug trials for schizophrenia.

Finding a drug that works can be a long ordeal punctuated by seizures and hospitalizations. The side effects of existing medications—weight gain, tremors, restlessness—lead some people to stop taking the medication and relapse.

There are high hopes among doctors for Cobenfy, which was approved in September, because it works in the brain differently than other schizophrenia drugs. Instead of blocking dopamine receptors, Cobenfy’s main ingredient, xanomeline, works on a different receptor that indirectly blocks the release of dopamine.

Cobenfy also contains trospium, which blocks some of the side effects. The most common are nausea, vomiting and indigestion. Unlike the weight gain seen with other schizophrenia drugs, people lost a few pounds while taking Cobenfy, made by Bristol Myers Squibb.

Dr. John Krystal of Yale University has led research on other schizophrenia drugs but was not involved in the new studies. He noted that only 10% to 20% of participants in the new studies dropped out because of side effects.

“It’s pretty good,” he said, noting that fewer or milder side effects could mean people stay on treatment longer. This could mean fewer problems associated with untreated mental illness: substance use, homelessness and unemployment.

So why did some patients continue treatment while others dropped out? Krystal said it will be important to understand more about this as doctors start prescribing the drug.

The Food and Drug Administration approved Cobenfy based on two encouraging five-week studies sponsored by the company and other safety data. The latest results, announced Thursday at the Psych Congress meeting in Boston, come from two longer studies, providing a more complete picture.

In one study, focused on critically ill patients, 78% dropped out, leaving only 35 people for the final analysis. In the other, focused on more stable people, 51 percent left the study, leaving 283 who took the drug for a year.

“It’s no higher or lower than what we typically see” in studies of schizophrenia, said Dr. Greg Mattingly of the Washington University School of Medicine in St. Louis. Louis. Mattingly is a consultant for Bristol Myers Squibb and a researcher for one of the studies.

In the most severely ill group, 69% of people had a significant improvement in their symptoms at the end of the year. In the other group, 30% saw a significant benefit.

Results of interviews with a sample of study participants, conducted by an independent research team and shared by Bristol Myers Squibb, showed the likelihood of continued treatment. After six months, 36 said they would continue to take Cobenfy after the trial if given the option; 10 said they don’t want to. Some participants said the drug reduced the voices, while others said it didn’t work for them.

The estimated annual cost of Cobenfy is $22,500, compared to $540 for a generic antipsychotic. Krystal and others worry that insurers will require people to try cheaper drugs first before covering Cobenfy. Most patients’ out-of-pocket costs will be much lower, depending on insurance and other factors.

A cheaper generic called clozapine is generally considered one of the best treatments for schizophrenia, Krystal said. It is underutilized in the US compared to other countries due to a cumbersome blood testing schedule.

The FDA started the blood tests to track the risk of severe neutropenia, a rare side effect that can be fatal. But doctors and families told the FDA that patients relapsed when their clozapine was withheld or delayed because of testing requirements.

Sally Littlefield, 29, of Alameda, Calif., said what works for her is a monthly injection of a long-acting antipsychotic drug. Littlefield, who has schizophrenia and bipolar disorder, wants to hear more about the experiences of people who have taken Cobenfy, and not just gamblers with a financial stake.

Mindy Greiling of Roseville, Minn., wants to see data on how Cobenfy compares to clozapine, which works for her 47-year-old son Jim. Weight gain was a problem for him, but since taking diabetes medication, he’s back to a normal weight, Greiling said.

Cobenfy “gets a lot of ballyhoo, like any new drug,” Greiling said. “It’s just a non-starter for me unless it turns out to be better than clozapine.”

___

The Associated Press Department of Health and Science receives support from the Howard Hughes Medical Institute’s Educational and Science Media Group. AP is solely responsible for all content.