Clinical and Research News
Controversy Erupts Over Study Of St. John’s Wort Efficacy
Psychiatric News
Volume 37 Number 10 page 26-48
Anchor for JumpAnchor for Jump

Subjects’ scores on the Hamilton Depression Scale indicated that hypericum, the active ingredient in St. John’s wort, and sertraline did not perform statistically better than placebo. But on the CGI Improvement scale, sertraline performed statistically significantly better than hypericum and placebo.

Even though the study was conducted using rigorous NIH protocols and standards, some are questioning not only the study’s conclusions, but also its methods and are insisting that the study’s original hypothesis was fundamentally flawed.

The report, which appeared in the April 10 Journal of the American Medical Association, was rolled out with typical JAMA fanfare as the lead article in the press packet for the issue, which press receive a week in advance. Press releases, briefings, and direct contact with principal investigator Jonathan R.T. Davidson, M.D., a professor of psychiatry and director of the Anxiety and Traumatic Stress Program at Duke University Medical Center, were offered to members of the press.

Critics of the study, most notably supporters of alternative medicine and herbal therapies, quickly fired off their own press releases. Two trade groups, the American Herbal Products Association (AHPA) and the National Nutritional Foods Association (NNFA), were especially critical of the study, saying its conclusions "misled the public."

In their joint press release, AHPA’s vice president for scientific and technical affairs, Steven Dentali, Ph.D., said, "It is inexplicable that JAMA has created such fanfare over the fact that St. John’s wort is not shown to be effective for a condition that it was never intended to treat."

Colleague Phil Harvey, Ph.D., director of science and quality assurance for NNFA, added, "This research in no way invalidates the scores of clinical studies and analyses that have clearly demonstrated that St. John’s wort is effective for mild to moderate depression."

Representatives for the Council for Responsible Nutrition further blasted the study as "misdirected and inconsequential."

Judging by the amount of coverage the study garnered in large, daily newspapers across the country, the media relations people at JAMA, as well as critics of the study, did their job getting the word out. The Washington Post ran the headline "Herbal Cure’s Worth Questioned," noting that the study’s findings on St. John’s wort and depression are reviving an old debate.

Others weren’t so kind. The Boston Globe ran with "Placebo Is Best in Depression Study," with the subhead "U.S. Funded Trial Pits Herb, Rivals." The Los Angeles Times told its readers that the study raised more questions than answers, noting "some medical experts argue that recent findings claiming the herb to be ineffective in treating severe depression are inconclusive."

At the center of the debate is a $6 million, three-year clinical trial funded jointly by three arms of NIH, the National Center for Complementary and Alternative Medicine, the National Institute of Mental Health, and the Office of Dietary Supplements.

"Many Americans use dietary supplements like St. John’s wort for depression without consulting a physician," principal investigator Davidson told Psychiatric News. "We felt there was a need to conduct a trial that could help us determine where St. John’s wort fits in the overall management of depression."

Davidson and his colleagues enrolled 340 adult outpatients who had major depression with a baseline total score on the Hamilton Depression Scale (Ham-D) of at least 20. Patients were randomly assigned to receive either St. John’s wort (Hypericum perforatum), placebo, or sertraline (Zoloft), as an active comparator, for eight weeks. Patients took doses ranging from 900 mg to 1500 mg of hypericum or between 50 mg and 100 mg of sertraline.

The trials’ two main outcome measures were the change in Ham-D total score from baseline to the eighth week and the rates of full response for each of the three study arms as determined by the Ham-D and score on the Clinical Global Impression (CGI).

The study showed that, in general, patients taking hypericum did no better than they did on placebo. In fact, they faired slightly (although not statistically) worse than those in the placebo group when comparing the change in Ham-D score from baseline to eight weeks. Patients taking sertraline did slightly (although, again not statistically) better than placebo (a drop on the Ham-D of 8.68 points for hypericum, 9.20 for placebo, and 10.53 for sertraline). However, on the CGI Improvement scale, sertraline was statistically significantly better in improving patients’ depressive symptoms than both hypericum and placebo.

Depending upon who was interpreting, these results were viewed as having entirely different meanings.

"Although research has shown that hypericum does have effects on serotonin, norepinephrine, dopamine, and the GABA system, as well as modulating sodium gradients in neurons," Davidson told Psychiatric News, "it is clearly not first-line treatment for patients in this particular diagnostic category."

Indeed, Davidson and his coauthors formally concluded in the paper that "this study fails to support the efficacy of H. perforatum in moderately severe major depression," and suggests that there is a complete "absence of trends" in the study suggestive of efficacy for the herb.

The Council for Responsible Nutrition argued that the patients selected for the study were simply too ill to benefit from a limited trial of any treatment, noting that the response seen in the sertraline group was not statistically significant on the Ham-D measurement.

Quoted in the press release for the council, Jerry Cott, Ph.D., a psychopharmacologist and former chief of the Psychopharmacology Research Program at NIMH, said, "It is difficult to understand why NIH chose to study this type of patient population."

Indeed, the patient population studied was the subject of the strongest criticism of the researchers’ criteria for including or excluding subjects from the study (see page 28).

"If [NIMH] had designed the study to focus on mild-to-moderate depression," Cott continued, "then we likely would have seen a very different result with both the standard antidepressant and St. John’s wort. The lack of response in the standard treatment group prevents any valid conclusions from this study regarding the effectiveness of St. John’s wort."


Other critics went even further in their condemnation of the study and its results.

"The real story here is that if this study is believable," said AHPA’s Dentali, "it showed that a blockbuster drug, with sales of over $2 billion and prescribed to millions of Americans for severe forms of depression, may be no more effective than placebo."

According to IMS Health Inc., an independent research company that tracks pharmaceutical products, sales of Zoloft, Pfizer’s brand of sertraline, totaled $2.14 billion in 2000. According to the Nutrition Business Journal, sales of St. John’s wort in 2000 amounted to $180 million.

According to Davidson, when measuring patients’ response as "full response" or remission, the study found that about 24 percent of patients taking St. John’s wort had a full response to treatment, compared with 25 percent for sertraline and 32 percent for placebo. These differences, like the changes in the Ham-D scores, were not statistically significant between the three groups.

"This is a classic illustration of the placebo effect confounding antidepressant trials," Davidson said, noting that active medications are not shown to be superior to placebo in as much as 35 percent of antidepressant trials. The fact that people seem to respond to placebos, in addition to concerns about giving an inert "treatment" to an individual who has major depression, has caused many to question the use of placebos, especially in drug trials involving psychiatric medications. As an alternative, some have called for new trial medications to be pitted against the "best available treatment"—a controversial position endorsed recently in a revision of the Declaration of Helsinki (Psychiatric News, November 17, 2000). The placebo debate has sparked intense dialogue and was reviewed in detail by the National Depressive and Manic Depressive Association in its recently issued Consensus Statement (see page 27).

Davidson acknowledged that if the results for the placebo in the current trial are ignored, effectively pitting St. John’s wort directly against sertraline as the "best available treatment, sertraline becomes statistically significantly better at relieving major depression as measured by both the Ham-D and the CGI-I." There would still be no statistically significant difference between the two as far as remission rates are concerned. Davidson, however, stands by the use of the placebo in the current study and agrees with the primary conclusions of the NDMDA’s consensus statement on placebo use in antidepressant trials.


The research team expected some discussion over the results of the paper on its publication. Davidson said he anticipated questions regarding the failure of sertraline to be statistically better, in light of its "real world success" in treating depression.

"We may have gotten ‘better’ results if we had, first of all, used higher doses of sertraline with this significantly depressed group of patients," he told Psychiatric News. The dose of 50 mg to 100 mg is at the lower end of the recommended therapeutic dosing range for the medication. "We simply used the lower end in an attempt to minimize adverse effects."

The trial might also have seen statistically significant results if the treatment period were longer than eight weeks, he added.

"And certainly," he concluded, "we need to look at patients with less-severe depression and systematically study St. John’s wort in that group."

An abstract of "Effect of Hypericum Perforatum (St. John’s wort) in Major Depressive Disorder" is posted on the Web at http://jama.ama-assn.org/issues/v287n14/abs/joc11936.html.


Anchor for JumpAnchor for Jump

Subjects’ scores on the Hamilton Depression Scale indicated that hypericum, the active ingredient in St. John’s wort, and sertraline did not perform statistically better than placebo. But on the CGI Improvement scale, sertraline performed statistically significantly better than hypericum and placebo.

Interactive Graphics


Citing articles are presented as examples only. In non-demo SCM6 implementation, integration with CrossRef’s "Cited By" API will populate this tab (http://www.crossref.org/citedby.html).
Related Articles