In a study sponsored by the National Institute of Mental Health and published in JAMA, St. John's wort (Hypericum perforatum) was found to be no more effective than placebo for treating depression but to be no worse than an antidepressant comparator (Hypericum Depression Trial Study Group [HDTSG], 2002). Critics of the study attributed these results to insufficient assay sensitivity; while those critical of antidepressant pharmacotherapy found the results supported their view that medication contributes little more than side effects to depression treatment.

The study, co-sponsored by the National Center for Complementary and Alternative Medicine, followed another U.S. trial that had equated the efficacy of hypericum with that of placebo (Shelton et al., 2001). That head-to-head comparison of herb and placebo was funded in part by Pfizer Inc., manufacturer of the selective serotonin reuptake inhibitor sertraline (Zoloft) that was also used as the antidepressant comparator in the NIMH study.

It had been anticipated that the HDTSG study would resolve questions about the efficacy of hypericum shown in smaller, less well-controlled European studies and in a meta-analysis of those trials (Linde et al., 1996). It was not anticipated that the study would heighten debate about the

measure and efficacy of antidepressants approved by the U.S. Food and Drug Administration, but the equivocal results emerging from this growing debate prompted the editors of JAMA to juxtapose the study with a literature review of placebo response in depression treatment trials (Walsh et al., 2002) and an editorial on the necessity of placebo control (Kupfer and Frank, 2002a).

In the eight-week, double-blind, randomized comparison of hypericum to placebo and sertraline, 31.9% of patients responded fully to placebo, compared to 24.8% responding to sertraline and 23.9% to hypericum (HDTSG, 2002). Full response was defined as achieving a Clinical Global Impression Scale-Improvement (CGI-I) score of 1 (very much improved) or 2 (much improved) and a Hamilton Rating Scale for Depression (HAM-D-17) score of ≤8. Sertraline was ranked better than placebo on the CGI-I, but neither the SSRI nor hypericum statistically separated from placebo in the HAM-D measure.

Critics argued that the study methodology was faulty and that selecting patients with long-standing depression and a score of at least 20 (moderate to severe) on the HAM-D-17 could obfuscate beneficial effects of the herb for patients with less severe illness (Cott and Wisner, 2002; Jonas, 2002; Linde et al., 2002; Wheatley, 2002). While the acuity of illness in this studied population may leave opportunities to evaluate hypericum in patients with less severe symptoms, the similarity of placebo and drug effect could foster additional questions about antidepressant effectiveness. In particular, this resonates with questions about SSRI effectiveness for severe depression posed over a decade earlier by the Danish University Antidepressant Group (1990). In addition, these results heightened consternation with the apparent potency of placebo for depression study populations.

Accompanying the HDTSG hypericum study was a literature review that found different levels of placebo response in depression studies over a period of two decades (Walsh et al., 2002). The review evaluated 75 placebo-controlled trials of treatment for adult outpatients with major depression that were conducted between 1981 and 2000. Most of the studies involved head-to-head comparison of active agents. Taking the active medication group with the greatest response, they reported that the mean proportion of patients responding to medication was 50.1% (range=31.6% to 70.4%); compared to 29.7% (range=12.5% to 51.8%) responding to placebo.