Dozens of previous trials, most conducted in Europe, have found some benefit from the herb, which has been used for thousands of years.

But Dr Richard C Shelton, a psychiatry professor at Vanderbilt University and the study's lead author now says: "We can say with confidence that it (the study) calls into serious question the effectiveness (of St John's Wort) in moderate to major depression."

"It means we know nothing now about St John's Wort," he said. "It puts into question everything that came before."

Although it is unclear whether the herb helps people with milder depression, Shelton said its use should be viewed with scepticism until better-designed trials are conducted, because of the risks of depression, including disability and suicide.

The study, which appears in a recent Journal of the American Medical Association, was funded by Pfizer Inc., which manufactures antidepressants and St John's Wort extract.

The researchers studied 200 adult outpatients diagnosed with major depression at 11 academic medical centres in the United States.

Each patient was evaluated for depression and randomly assigned to receive St. John's Wort extract or a placebo for eight weeks.

Patients were not permitted to use any anti-depressant drugs during the trial; some patients were in psychotherapy.

Researchers concluded that St John's Wort "failed to produce significant differences vs placebo" based on several previously established rating scales designed to evaluate depressed patients. Shelton said his study was limited because it did not also compare the use of St John's Wort to an anti-depressant drug. Such a study is now being conducted by the National Institutes of Health, he said.

Dr Darrel A Regier, director of the American Psychiatric Association's research division, said previous reports of St John's Wort's effectiveness might be attributed to a placebo effect ? some people will get better simply by believing it works.

Shelton's study shows that St John's Wort should not be recommended for anyone who is depressed, said Regier, who praised its "state-of-the-art methodology". He added that many St John's Wort users do not seek medical advice and are unaware of many of the herb's risks. Some studies have shown, for example, that St. John's wort can interfere with drugs used to treat HIV-infection and heart transplant patients.

Shelton criticised previous studies for being poorly designed, having too few participants or using inadequate doses of the anti-depressant medicines to which St John's Wort was compared. He said it was inconceivable that so many trials ? there have been at least 31 ? all were positive.

But a spokesperson for a trade group for nutritional supplements says the conclusions of Shelton's study are overstated.

"Nobody ever said St John's Wort was effective for major depression," said John Cardellina, vice-president for botanical science for the Council for Responsible Nutrition. "Now these guys run one trial of major depression and say it should outweigh all other trials."