Lack of Efficacy for Antidepressants Due to Flawed Analyses?

Megan Brooks

July 17, 2019

The alleged lack of efficacy for selective serotonin reuptake inhibitors (SSRIs) in mild to moderate depression may be due to flawed analyses, rather than pharmacology, new research suggests.

An analysis conducted by investigators at the University of Gothenburg, Sweden, does not support restricting SSRIs to patients with severe depression.

It also highlights the importance of reconsidering how depression rating scales are used to define depression severity and assess treatment response as well as interpreting group-level meta-analyses with caution, the investigators note.

The study was published online July 11 in Lancet Psychiatry.

Ongoing Controversy

"A number of previous analyses that have had marked impact on the field claimed that antidepressants are effective (if at all) only in the most severe cases, and for many debaters and clinicians this still seems to be a well-established truth," study investigator Elias Eriksson, MedDr, professor of pharmacology, told Medscape Medical News.

These previous analyses typically used a disputed measure of improvement — a decrease in the sum-score of the 17-item Hamilton Depression Rating Scale (HDRS-17) — and were based on group-level rather than patient-level data, he explained.

"We, and others, have demonstrated the inadequacy of the conventional HDRS-17 measure of response in antidepressant trials, one problem being that many of the items on this scale are rated zero already at baseline in many patients, and hence cannot possibly display any drug-induced improvement," said Eriksson.

"We hence believed that the severity vs response issue must be addressed using each individual symptom rather than the sum score of the Hamilton scale as outcome measure, which to our knowledge had not been done before," he added.

The investigators conducted an item-based, patient-level, post-hoc analysis of pooled data from 8262 adults with major depression from 28 acute-phase, placebo-controlled HDRS-based trials of three SSRIs — citalopram, paroxetine, and sertraline.

"To our knowledge, this is the largest patient-level study assessing the possible association between baseline severity of depression and response to SSRIs to date, and the first including comprehensive analyses of outcome measures other than HDRS-17 sum scores," the researchers write.

Refuting previous claims that antidepressants are ineffective for non-severe depression, the investigators found no difference between patients with the lowest baseline HDRS-17 sum score compared with those with the highest sum score with respect to SSRI-induced decrease in the 6-item HDRS subscale (HDRS-6) symptoms of depression. The HDRS-6 includes depressed mood, guilt, work and interests, psychomotor retardation, psychic anxiety, and general somatic symptoms.

However, effect sizes for non–HDRS-6 symptoms were notably lower in patients with non-severe depression than in those with severe depression, which may partly be the result of the sparsity of these symptoms at baseline in this subgroup compared with the severe depression subgroup, the researchers say.

These results suggest that the "apparently lower effect size in cases defined as 'non-severe' is largely due to the fact that these patients report fewer symptoms at baseline, and therefore cannot show any improvement in symptoms that are absent from start," said Eriksson.

"Moreover, since some of the symptoms that are often absent at baseline, such as gastrointestinal problems, are common SSRI side effects, these patients may even report some deterioration with respect to these symptoms, which may mask an antidepressant response when using the sum rating of all items as effect parameter," he noted.

He also noted that key symptoms of depression, such as depressed mood, are "almost as common and severe at baseline in cases defined as non-severe as in those defined as severe, and that the effect of SSRIs is just as good in cases defined as non-severe. The common assumption that SSRIs are useless in moderate depression thus seems entirely unfounded."

Experts Weigh In

Commenting on the study for Medscape Medical News, Irving Kirsch, PhD,  associate director of the Program in Placebo Studies at Harvard Medical School, Boston, Massachusetts, and author of the book The Emperor's New Drugs: Exploding the Antidepressant Myth, which claims there are no clinical differences in the effectiveness of antidepressants vs placebo in depression, said the study is "interesting, but there are some problems with it."

"First, the HDRS-6 that they depend on for their conclusion excludes two of the defining diagnostic criteria of depression in DSM-5: weight change and suicidal ideation.

"The argument that one should do so because mildly and moderately depressed people do not have these symptoms seems specious to me. Of course, the exclusion of core features of severe depression lessens the ability to find differences in improvement it as a function of severity. That's why they should be included, not excluded," Kirsch said.

"Second, differences associated with baseline severity are very small, even in the studies in which they are shown to be statistically significant. What the data show is that antidepressant–placebo differences are very small at all levels of depression," he added.

In a commentary published with the study, Toshi Furukawa, MD, PhD, of Kyoto University Graduate School of Medicine/School of Public Health in Japan, notes the study "opens up some fundamental questions. What is depression? How can we measure it most adequately and appropriately? And what do we mean to treat when we treat people with depression? As this field of science grows, and the more we know, the more of the unknown we have to explore."

The study was supported by Swedish Medical Research Council, AFA Insurance, the Swedish Brain Foundation, Sahlgrenska University Hospital, Bertil Hållsten's Foundation, and Söderberg's Foundation. Eriksson has previously been on advisory boards, received speaker's honoraria, or research grants from Eli Lilly, GlaxoSmithKline, H Lundbeck, and Servier.  Kirsch has disclosed no relevant financial relationships. Furukawa has received personal fees from Meiji, Mitsubishi-Tanabe, Merck Sharp & Dohme, and Pfizer; a grant from Mitsubishi-Tanabe; and has a patent pending for smartphone apps.

Lancet Psychiatry. Published online July 11, 2019. Abstract, Editorial


For more Medscape Psychiatry news, join us on Facebook and Twitter


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.