The final stage of a landmark federal study on treating depression suggests that two-thirds of sufferers eventually can be helped if they are patient enough to keep trying medications until they find one that works.

Thirteen percent of the 123 study participants who did not get better on the first three drugs they tried were helped by a fourth, researchers found.

But there is a downside to so many attempts: The more tries people made, the more likely it was that they later would relapse and slide back into depression.

"It's a sobering message when you get down to requiring three or four steps," said the chief researcher, Dr. A. John Rush of the University of Texas Southwestern Medical Center in Dallas. "It says that followup is critical" to make sure that people stay on the drugs, he said.

Results were published Wednesday in the November issue of the American Journal of Psychiatry.

About 15 million Americans each year suffer depression, the nation's top mental health problem. Drugs to treat the disorder have had only limited testing until now. The government launched the six-year, $35 million study to test a variety of these medications in "real world" settings — people seeking help at community clinics and doctor's offices.

The project started out with 3,671 adults diagnosed with major depression. Half of them had a family history of it, and more than half had suffered multiple bouts of it themselves.

All were started on Celexa, made by Forest Laboratories and one of a widely used class of drugs called SSRIs. Nearly 37 percent had their depression go into remission after this first try. The rest switched to another antidepressant or continued with Celexa and added a second treatment. This second step helped 31 percent of that group.

The third and fourth attempts brought success rates of 14 percent and 13 percent, respectively. When all results from these various groups of participants were looked at collectively, 67 percent of the total group had been helped by one or more drugs.

However, 40 percent of those who achieved remission on their first drug relapsed within a year. That rose to 55 percent of those who took two tries to succeed and 65 percent and 70 percent of those requiring three and four tries, respectively.

"More steps are still worthwhile, but you have a decreasing return," Rush said.

Dr. Thomas Insel, director of the National Institute of Mental Health, said that because there was no comparison group of people who got no treatment, it is hard to know how many of those got better on their own, with enough passage of time, or because of the drugs.

"I think the overall results are hopeful," he said. "The problem with depression is that people and their families feel hopeless. The message here is that medication can be helpful."

Why did one-third of people not get better with any drugs? Possible explanations include genetic or biological differences, and life circumstances or other medical conditions that interfered with treatment. Medications may have been started too late to do any good, and some forms of depression may not respond to drugs.

Doctors say that 14 weeks is enough time for a drug to have a chance to work. After that, people should consider trying something else or adding a second treatment.

Soon to be published are separate results on the effectiveness of counseling, which some participants got in the second stage of the study.