Pages Menu
Categories Menu

Posted by on Feb 15, 2010 in Neuropsychiatry, Pharma

Accutane-Depression Link: Clinical Studies Not Overwhelming

Accutane-Depression Link: Clinical Studies Not Overwhelming


Given the exclusion of Doug Bremner’s plaintiff-funded PET study from Accutane litigation, specifically Palazzolo et al v Hoffman La Roche et al, the Emory psychiatrist (if he is still allowed to testify*) will need to rely on other clinical studies to convince a jury that Accutane causes depression and, in particular, suicidal depression. (For background on the trial’s decision, go here and here.)

However, on this point, results from available clinical studiesnone of which is well controlledare mixed. And some data actually suggest an improvement in depression or anxiety with Accutane treatment.

So what are the data? The following is my review, based on several PubMed searches, of published clinical studies that attempted to assess any link between Accutane (or isotretinoin) use and depression or suicide. Overall I found 5 prospective, cohort studies in which patients received Accutane or a comparator treatment (eg, oral antibiotics), 4 prospective studies without comparator treatment, and 1 large retrospective population-based study. (Note: Case reports or case series are not included in this review.)

Prospective, cohort studies (treatments not randomized)

Ng et al, 2002 (University of Melbourne): Patients with acne (N = 215) received Accutane or antibiotic/topical treatments. No treatment-related differences were noted at 6 months in mean depression or quality-of-life scores, and there was no correlation between Accutane dose and depression score. However, 5 Accutane-treated patients (3%) withdrew from the study because of “worsening of mood.” Financial support for the study is unclear.

Chia et al, 2005 (St. Louis University): Patients aged 12-19 years with moderate-severe acne (N = 132) received Accutane or conservative therapy (ie, oral or topical antibiotics or topical retinoid). Among the 101 subjects (77%) who completed the study, CES-D scores suggestive of clinically significant depression and rates of new-onset depression at 3 months were comparable in the 2 treatment groups. The authors found that improvement in acne with either treatment was associated with reduced depressive symptoms. Patients who were unavailable for follow-up did not have significantly different baseline depression scores, and dropout rates were comparable in the 2 treatment groups. The authors reported “None” to financial disclosures.

Bremner et al, 2005 (Emory University): Adults aged 18-50 years with “treatment-resistant” acne (N = 28) received Accutane or oral antibiotics. There were no significant increases in Hamilton depression scores at 4 months with either treatment and no significant between-treatment differences in these scores. Changes in “depression related to acne” were also not significantly different between the treatment groups. There was no correlation between baseline metabolism in the orbitofrontal cortex (an area of the brain implicated in depression), as determined by PET, and depression on any measure used in the study. The study was funded by Accutane plaintiffs (Liam Grant, 80%; Accutane-litigation lawyers, 20%).

Cohen et al, 2007 (University of Calgary): Patients aged 14 years or older with acne (N = 200) received Accutane or oral or topical antibiotics. At 2 months, there was no correlation between Accutane treatment and scores indicating the development of depression (CES-D or Zung Depression Status Inventory). The study was funded by the Centre for Advancement of Health, Calgary.

Kaymak et al, 2009 (University of Gazi): Patients with acne (N = 78) received Accutane or topical treatment. At 2 months, quality of life (measured by using the DLQI) was more impaired in the group receiving topical treatment. At 4 months, measures of quality of life, depression, and anxiety (the latter 2 via the BDI and HAD scale) showed improvement with Accutane treatment. Funding for the study is unclear.    

Prospective Accutane studies without comparator treatment

Rubinow et al, 1987: Patients with cystic acne (N = 72) were treated with 1 of 3 dosages of isotretinoin. The authors observed “significant reductions in anxiety.” “[M]itigation of anxiety and depression [was] most robust in those patients with the greatest dermatologic improvement.” Financial support of the study is unclear.

Strauss et al, 2001 (University of Iowa). Patients with severe recalcitrant nodular acne (N = 600) received either a micronized formulation of Accutane or standard Accutane treatment in double-blind fashion. At 20 weeks, the rates of adverse events were “generally lower” with the micronized formulationspecifically mucocutaneous events (eg, dry nose, dry eyes) and hypertriglyceridemia. Three patients taking the micronized drug and no patient taking the standard drug withdrew from the study because of depression. Psychiatric adverse events (eg, depression, anxiety, panic) were reported in 11 patients (4%) taking the micronized drug and 1 patient (0.3%) taking the standard formulation (who experienced a panic reaction). The study was funded by Roche. (N.B.To my knowledge, the micronized formulation never made it to market.)

Ferahbas et al, 2004 (University of Ericyes, Turkey): Patients with “severe recalcitrant acne” (N = 45) received Accutane for 16 weeks. Among the 23 patients (51%) who completed the final assessment, the authors noted decreases in depression and anxiety scores (the latter being statistically significant) from baseline. No patient attempted or committed suicide. The abstract (I was not able to obtain a copy of the full article) does not provide information on the subjects who withdrew from the study. Funding for the study is also unclear.

Rehn et al, 2009 (Helsinki City Health Center): Male military conscripts (N = 135) received Accutane for 12 weeks. Among those who completed follow-up (93%), the mean depression score (via the BDI) and the proportion of patients with clinically significant depressive symptoms or suicidal ideation declined with treatment (the former, significantly). The authors, however, note that “one non-depressed patient attempted suicide while intoxicated with alcohol.” It is assumed that the study was funded by the Finnish government, given the nature of the subjects.

Retrospective, population-based cohort study

Jick et al, 2000 (Boston Collaborative Drug Surveillance Program): By using the Canadian Saskatchewan Health Database and the UK General Practice Research Database, the authors analyzed computer-recorded histories from patients with acne who took Accutane (n = 7535) or oral antibiotics (n = 14,376). There was no increased risk of newly diagnosed depression or psychosis or attempted or committed suicide with Accutane exposure. This study, which was funded by Roche, has been criticized for its methodological limitations.


Ten imperfect clinical studies, assessing the relationship between Accutane treatment and depression or suicide, were discovered. Two of these studies (the 2 largest) were funded by Roche, 1 study was funded by Accutane plaintiffs, and the other 7 studies were supported by independent or unclear sources. Overall these studies do not support an association between the available formulation of Accutane and depression or suicide; although Ng et al reported that 5 Accutane-treated patients (3%) dropped out of their study because of worsening mood, and 1 small study reported a high dropout rate for unclear reasons (Farahbas et al). Some studies indicate an improvement in mood with Accutane treatment, on the basis of an improvement in acne.

The prescribing information for Accutane, which is no longer sold by Roche in the United States, contains warnings regarding the risks of depression or suicide with treatment; although “[n]o mechanism of action has been established.” On the basis of their study, Ng et al concluded that the potential risk of depression with Accutane treatment is “a rare unpredictable idiosyncratic side-effect.”

BDI = Beck Depression Inventory; CES-D = Center for Epidemiologic Studies Depression scale; DLQI = Dermatology Life Quality Index; HAD = Hospital Anxiety and Depression; PET = positron emission tomography. 

* In the recent appellate ruling, the judge wrote, “We therefore remand this case to the trial judge to consider whether Dr. Bremner should be permitted to testify as an expert on general causation, without reference to the PET study. We leave to the judge’s discretion whether to permit or require any further proceedings prior to issuing a decision on remand. We do not retain jurisdiction.”

Photo of Roaccutane, aka Accutane, from Wikipedia.

bmartin (1130 Posts)

A native East Tennessean, Barbara Martin is a formerly practicing, board-certified neurologist who received her BS (psychology, summa cum laude) and MD from Duke University before completing her postgraduate training (internship, residency, fellowship) at the Hospital of the University of Pennsylvania in Philadelphia. She has worked in academia, private practice, medical publishing, drug market research, and continuing medical education (CME). For the last 3 years, she has worked in a freelance capacity as a medical writer, analyst, and consultant. Follow Dr. Barbara Martin on and Twitter.


  1. Prospective cohort studies with comparitor groups.
    Every single time I read one of these I scream inside my head “Why won’t you just randomise them? You idiots”. It’s pretty much the same amount of work and a bajillion times more useful.
    The other option is that they’ve lied about the prospective nature of the hypothesis generation to study design thought train. Which I sincerly hope I’m wrong about.

  2. Yeah, the reason or reasons for not randomizing treatments in these comparative Accutane studies are not explained. One possible reason, as you indicate, is that these studies were, in fact, not prospective.
    Specifically Bremner et al don’t address the reason(s) for not randomizing treatments; although they conclude that a randomized study is warranted. They do explain, however, their reason for the lack of blinding: “Because of the side effects of isotretinoin (severe skin dryness) it was decided that it would not be possible to blind the subjects or the raters to treatment condition.”
    Whether this explanation is satisfactory…I’m not sure. And specifically whether blinding was at least confined to the interpreters of the PET images was not addressed by the authors. Their silence on this point leads me to believe that they weren’t.