Last month, psychiatry professor John Krystal set out to evaluate the cost-effectiveness of naltrexone, a well-known treatment drug for alcoholism. But to his amazement, he found instead that it didn’t work at all.

In the largest and longest placebo-controlled treatment study of its kind, Krystal studied 627 veteran patients being treated for alcohol dependence from 15 Department of Veterans Affairs medical centers. In addition to counseling, some patients received a placebo, while others received the drug naltrexone for either three or 12 months.

“We thought we were going to find not ‘does it work,’ but how well it works, and its economic value, but to our surprise, we found out it didn’t work at all,” said co-author Robert Rosenheck, a psychiatry professor in the Department of Epidemiology and Public Health.

Krystal said the study showed that naltrexone did not delay a relapse to alcohol use, nor did it reduce the intensity or frequency of drinking.

Naltrexone is supposed to work by reducing the rewarding effects felt after consuming alcohol.

The drug first appeared in published literature in the early 1990s, when Yale psychiatry professor Stephanie O’Malley, among others, postulated that the medication might enhance a patient’s recovery in conjunction with counseling.

The arrival of naltrexone “really provided a great deal of hope to people working with alcohol dependent patients because it was the first new medication treatment since antibuse,” Krystal said.

Antibuse, the only other FDA-approved drug used to treat alcoholism in the United States, is often ineffective because it makes patients feel ill when they drink alcohol. Krystal said people would simply not take the medication — compliance with the treatment was notoriously bad.

“The possibility that a medication like naltrexone, which was well-tolerated, might reduce drinking was very exciting,” Krystal said.

The professors’ interest in studying these veterans peaked when they discovered that naltrexone was not being prescribed to veterans who were treated for alcohol dependence in the various Department of Veterans Affairs medical centers, even though it was available to them.

Krystal brought his expertise about the treatment of alcoholism to the study, while Rosenheck focused on broad population issues in mental health and the economics of treating large populations.

The cost for treating an alcoholic with the drug is about $1,000 per year. Rosenheck wanted to determine whether naltrexone could reduce other kinds of costs in areas like health care or criminal justice. He also examined whether it improved the quality of life to merit the price tag — often at taxpayers’ expense.

“We thought that if we could conduct a study of veterans that could show how helpful naltrexone was, then this drug, which we believed to be important, would be more readily prescribed,” Krystal said.

But to the entire research team’s amazement, it had no effect on the sample population.

While Rosenheck said no single study ever decides things, he said he does feel strongly about the conclusions the study drew.

“The VA spends millions of dollars a year on this medication, and it’s a very explosive thing to tell doctors who are prescribing the medication, ‘Whoops, we did a study that shows this isn’t worth it,'” Rosenheck said.

But Krystal cautioned that this study should not discourage those who have had success with the drug from continuing to use it. He said that there are reasons to believe that naltrexone used in different ways and in different populations might still be helpful for some people.

“At a higher dosage level or in combination with other medications, naltrexone could be highly effective,” Krystal said.

Nonetheless, their study showed that naltrexone is limited in its ability to treat alcohol abuse, and further research is necessary to clarify what role, if any, it might play. Other possibilities for the drug are being studied by others at Yale and elsewhere across the United States.

Rosenheck said studies of this type — those that look at the cost-effectiveness of certain treatments — are becoming much more common in a health care world so focused on the bottom line.

“Much of the research in departments here and across the country have focused on the biological basis of a psychiatric disease and its pharmacological treatment,” Rosenheck said. “But it is increasingly more common to see a pairing like this to do a cost-effective study. It’s been a wonderful synergy between our two departments.”