The Impact of Omega-3 Fatty Acids on Depressive Disorders and Suicidality: Can we Reconcile Two Studies with Seemingly Contradictory Results? (original) (raw)

. Author manuscript; available in PMC: 2013 Aug 6.

Published in final edited form as: J Clin Psychiatry. 2011 Dec;72(12):1574–1576. doi: 10.4088/JCP.11com07463

Investigation into the psychotropic efficacy of the omega-3 fatty acids (n-3 FAs) has grown tremendously in the past decade and a half. Since the publication of Stoll et al’s1 study of n-3 for treatment of bipolar disorder, many clinical trials, in adults and children, have examined the n-3s in unipolar major depressive disorder (MDD), postpartum depression, bipolar disorder, attention deficit disorder (ADD), psychotic disorders, obsessive-compulsive disorder (OCD), and borderline personality disorder2–5. Psychiatric clinical trials of n-3 FAs in mood disorders now number in the thirties, and various metaanalyses and systematic reviews have generally, but not unequivocally, suggested benefit2,6–16. These efficacy findings are consistent with what is being learned about antidepressants in general, with the growing availability of unpublished studies and databases17, but interpreting the n-3 data is complicated by several factors:

  1. Preparations of the psychiatrically relevant n-3 FAs, eicosapentaenoic acid (EPA) and docosahexaenoic acid (DHA), range from pure EPA or DHA to combinations of the two at different ratios. Most studies have used EPA alone or a combination of EPA and DHA2; few studies have used pure DHA18,19. The possible EPA:DHA ratios are varied, and thus far there is no clear advantage for any one preparation.
  2. Doses in studies have ranged from less than1000 mg/day to 10 grams/day. With regard to unipolar depression, the data generally support doses in the range of 1–2 grams per day2, though studies with higher doses have also shown benefit. Optimal doses have yet to be characterized.
  3. The EPA:DHA ratio, which varies widely in commercial preparations, may be important when considering clinical effects of n-3, but we do not know whether there is an optimal ratio for humans. Likewise, omega-6 (n-6) fatty acids, which people from industrialized countries consume in high levels, may also be relevant. Some studies that have found therapeutic windows for EPA and DHA suggest an optimal n-6/n-3 ratio for humans. For example, our study of EPA monotherapy20 found an association between lowering of the n-6/n-3 ratio and clinical improvement in subjects who received EPA, but we did not find a significant association in our smaller study of DHA19.
  4. The n-3 FAs are vulnerable to oxidation and may be affected by dietary and consumptive practices. Investigators often attempt to control for this, by excluding smokers and people who already consume high amounts of n-3 in the diet21,22. However, these methods are not 100% reliable; patients may lie, minimize, or exaggerate lifestyle factors to be admitted into studies.
  5. Diagnostic criteria vary, with some studies requiring DSM-IV MDD and others having more liberal diagnostic requirements; depressive severity may also vary; and certain comorbid conditions and concurrent medications may or may not determine exclusion.

The body of n-3 data is therefore heterogenous, which makes it difficult to compare and contrast studies in a systematic manner that provides practical information for investigators and clinicians, not to mention the general public.

JCP has recently published two excellent, meticulously crafted reports examining clinical impact of n-3 FAs. Lewis and colleagues23 presented a case-control study of suicide deaths among active duty military, in which higher serum DHA demonstrated a protective effect against suicide. Surprisingly, EPA appeared to confer no significant protective effect. Following this report, a metaanalysis by Sublette and colleagues24 examined the relative effects of EPA and DHA for MDD, and found that EPA, rather than DHA, appeared to have the main antidepressant effect. Can these seemingly contradictory findings be reconciled? Let us first contrast the two studies:

  1. Lewis and colleagues selected a specific, unambiguous outcome: suicide death. In contrast, Sublette and colleagues examined depressive improvement, an outcome more subject to patient and investigator bias.
  2. Lewis and colleagues examined a narrow segment of the population: active duty, predominantly male military personnel serving during the post-9/11 era--a population that may be especially vulnerable to psychiatric disorders. Sublette and colleagues, on the other hand, examined studies with depressed individuals of both genders (presumably with a more balanced distribution than in the Lewis sample), from many walks of life, which would make their sample more representative of the general population with depression, though not necessarily of the population as a whole, since their reviewed studies did not include healthy controls.
  3. While Lewis and colleagues focused on dietary serum levels of omega-3, Sublette and colleagues examined a wide range of n-3 supplementation regimens, including monotherapy, augmentation (including one study with psychotherapy as the baseline treatment), different doses and EPA:DHA ratios, and different treatment durations. Diagnoses included MDD of different severities, bipolar disorder, dysthymia, depressive episode, Parkinson’s, and coronary heart disease. Again, this may limit generalizability with regard to MDD per se.
  4. Both studies reported difficulty controlling for medical and psychiatric comorbidity, including substance abuse and PTSD, which may be especially relevant in a military population. Soldiers who had witnessed extreme violence, for example, had a greater suicide risk than those who had not. Although access to comorbidity data was limited in the Sublette meta-analysis, the studies reviewed typically allow some comorbid disorders (except for active substance abuse), provided that they do not constitute the primary complaint. We cannot be sure whether substance abuse or PTSD were contributing factors to suicide in the Lewis study.
  5. Lewis and colleagues broke down their serum n-3 findings into octiles, allowing for a wide range of comparisons. The authors observed that their sample had an unusually low serum DHA status compared to general populations, which may make detection of an association more difficult, and yield results less generalizable to the population at large. Sublette and colleagues designed their meta-analysis with a binary dose schedule: EPA at levels of greater than or less than 60%, relative to DHA. They provide a strong rationale for their dichotomy, based on the observation that all significant positive studies had at least 60% EPA and all studies with less than 60% EPA were negative. They could not, however, examine the impact of blood levels of omega-3, which were not obtained in many of their studies. It is not clear how blood levels may relate to levels in the brain, where the n-3s presumably carry out their psychotropic effects.

These issues make the studies difficult to compare on a head-to-head basis, and could explain any apparent differences in their findings. But are these findings contradictory? One study says that DHA is protective against suicide and EPA is not; the other says that EPA is more effective than DHA for alleviating depression. These findings are not mutually exclusive:

  1. Although the relationship between depressive disorders and suicide is well established, suicidal ideation or completed suicide should not be considered as a proxy measure for MDD. Only 15% of people with depression will commit suicide25, and at least 40% of suicides consist of people without diagnosed depression26. Suicide may be influenced by many factors, e.g. situational stressors that can be difficult to systematically assess and quantify in study samples. Therefore, conclusions about suicide and about depression cannot necessarily be comingled.
  2. Although EPA did not significantly associate with suicide in the Lewis study, there was a trend to significance (p<0.08), suggesting that EPA may in fact have a benefit, albeit more modest. Other fatty acids that are seemingly unrelated to mood, also appeared to have protective effects against suicide: dihomogamma linolenic acid (DGLA), palmitoleic acid, cis-vaccinic acid, and stearic acid. Therefore, the overall clinical impact of the different FAs and their ratios, in oral preparations and in the blood, is difficult to fully elucidate within the limitations of these studies.
  3. Serum levels of n-3 may not necessarily reflect brain levels. EPA is a precursor of DHA and may indirectly increase brain DHA or cause secondary effects, such as in the inflammatory system that could ultimately impact on brain. While Sublette and colleagues suggest that DHA could block the effects of EPA, this seems unlikely given everything else that is known about DHA’s impact on mood and suicidality.
  4. PTSD in the Lewis sample, even if not formally diagnosed, may have influenced suicidality. Also, psychiatric pathology in general, measured indirectly by a greater number of outpatient and inpatient mental health visits, may have also conferred a greater suicide risk for these soldiers. Interestingly, substance abuse visits were not associated with increased suicide risk, though it is possible that many individuals were underdiagnosed in this regard.

Considering the above, these studies should not be viewed as contradictory, and can in fact be reconciled. EPA may have benefits for the overall constellation of symptoms known as DSM-IV MDD, that exceed those of DHA. On the other hand, DHA may impact the mechanism(s) in the brain that regulate suicidal behavior, independently of MDD status. These findings should encourage investigators who study omega-3s, as well as other antidepressants, to closely scrutinize suicidal symptoms, in addition to standard efficacy analyses based on depressive symptoms as a whole.

On a final note, our group at MGH, in collaboration with Emory University, is nearing completion of a 3-arm randomized clinical trial comparing EPA, DHA, and placebo for MDD, including analysis of plasma PUFA levels as moderators and mediators of response. We hope that our findings will help to shed further light on the comparative efficacy of EPA and DHA.

Acknowledgments

This work was supported in part by grant R01 MH074085 to Dr Mischoulon from the National Institutes of Health (NIH) and National Institutes of Mental Health (NIMH).

Footnotes

Disclosures

Dr Mischoulon has received research support from Nordic Naturals, and Ganeden. He has served as a consultant to Bristol-Meyers-Squibb Company. He has received writing honoraria from Pamlab and Nordic Naturals. He has received royalties from Back Bay Scientific for PMS Escape, and royalties from Lippincott Williams & Wilkins, for textbook “Natural Medications for Psychiatric Disorders: Considering the Alternatives” (David Mischoulon and Jerrold F Rosenbaum, Eds.). He has received honoraria from Reed Medical Education (a company working as a logistics collaborator for the MGH Psychiatry Academy). The education programs conducted by the MGH Psychiatry Academy were supported through Independent Medical Education (IME) grants from pharmaceutical companies co-supporting programs along with participant tuition. Commercial entities currently supporting the MGH Psychiatry Academy are listed on the Academy’s website www.mghcme.org.

References