Skip navigation

Evening primrose oil and fish oil for schizophrenia

Clinical bottom line:

There is no convincing evidence from small trials to show whether supplementation with polyunsaturated fatty acids is effective in relieving the symptoms of schizophrenia, though one study showed beneficial effects with fish oil. Adverse effects were uncommon. The studies assessed a variety of outcomes.

Schizophrenia affects approximately 1 in 100 individuals; over 45 million people are affected worldwide. Some studies have shown that sufferers have low levels of certain essential fatty acids in their neural cell membranes. It has been hypothesised that boosting the level of these essential fatty acids in the body helps reduce the symptoms of schizophrenia.

Systematic review

Joy CB, Mumby-Croft R, Joy LA. Polyunsaturated fatty acid (fish or evening primrose oil) for schizophrenia (Cochrane Review). In: The Cochrane Library, Issue 3, 2000. Oxford: Update Software.

Date review completed : February 2000

Number of trials included : Four

Number of patients : Total 204

Control group : Standard treatment, placebo, other polyunsaturated fatty acid.

Main outcomes : Use of antipsychotics, adverse effects, mental state, movement disorders.

Inclusion criteria were randomised clinical trial which assessed the effectiveness of polyunsaturated fatty acid supplementation in people with schizophrenia or similar chronic mental illness.

Biological Abstracts, Embase, Medline and PsychLit were searched to 1998, Cochrane Library (Issue 4, 1999) and the Cochrane Schizophrenia Group's Register (February 2000) were also searched. Bibliographies of retrieved reports were checked and authors were contacted for further studies. Study quality was assessed using a validated three-item, five point scale. An intention-to-treat analysis was conducted. Dichotomous information was used to calculate relative risk and number-needed-to-treat (NNT), with 95% confidence intervals; weighted mean difference, with 95% confidence intervals, was calculated for continuous data.


All included studies were randomised and double blind, and scored a minimum of two out of five on the quality rating scale. Study duration was 6-12 weeks. Only one study stated the age of patients (18-65 years). Continuous data (mental state, movement disorders) was collected using the Positive and Negative Syndrome Scale (PANSS) and the Abnormal Involuntary Movement Scale (AIMS). The PANSS is made up of 30 items; each is scored from one (absent) to seven (extreme) with a low score indicating low levels of psychopathology. The AIMS has 12 items rated from zero (none) to four (severe) and describes the level of dyskinetic movement; a low score indicates low levels of dyskinetic movement.

Fish oil (omega-3 fatty acids) versus placebo

One trial (Trial A) assessed 30 patients with schizophrenia who were not taking antipyschotic medication at study entry, but were experiencing a new episode of illness. The dose of fish oil (eicosapentaenoate) was not stated. The need for and use of antipsychotic medication was assessed over 12 weeks. Significantly fewer patients needed antipsychotic medication with fish oil (9/15 patients) than with placebo (15/15 patients); the relative risk was 0.6 (95% confidence interval 0.4 to 0.9) and the NNT was -2.5 (-6.5 to -1.5). The NNT means that one in every three patients would not require antipsychotic medication with fish oil who would have done with placebo. The study also assessed mental state, using the PANSS score. Greater reductions in average total PANSS score were recorded with fish oil than with placebo; the weighted mean difference (MWD) was 12.5 (3 to 22).

Another trial (trial B) assessed supplementation of antipsychotic with eicosapentaenoate (1, 2 or 4 mg) or placebo in 28 schizophrenic patients taking neuroleptics for 12 weeks. There was no significant difference between supplementation with fish oil or placebo for the percentage change in average total PANSS score; the WMD was -11 (23 to -1).

Movement disorders:

One study (trial C) assessed evening primrose oil (1, 2 or 4 g per day) for six weeks compared with placebo in 115 patients, aged 18-65 years, with DSM-IV schizophrenia who were taking neuroleptics. There was no significant difference between fish oil compared with placebo for akathisia hypertonia/stiffness or tremor; the relative risks were 0.38 (0.01 to 18), 0.13 (0.01 to 3.0) and 0.37 (0.02 to 5.7) respectively.

Evening primrose oil (omega-6 fatty acids) supplementation versus placebo

Trial B also assessed supplementation of antipsychotic with evening primrose oil (docosahexanoic acid; dose unspecified) or placebo in 29 patients taking neuroleptics for 12 weeks. Fourteen patients received evening primrose oil. There was no significant difference in the percentage change in average total PANSS score for supplementation with evening primrose oil compared with placebo; the WMD was -1.3 (-15 to 13).

Movement disorders:

One trial (trial D) assessed gamma linoleic acid (omega-6; uncertain of dose) compared with placebo in 16 chronically ill schizophrenic patients who had had a long exposure to antipsychotic drugs and had tardive dyskinesia for at least one year. Patients continued their antipsychotic drugs. There was no significant difference between evening primrose oil and placebo (16 patients) for AIMS score after six weeks, WMD 1.3 (-2.0 to 4.6). There was no difference between study treatments in the change in AIMS score (data not provided).

Fish oil supplementation versus evening primrose oil supplementation

Trial B compared fish oil against evening primrose oil in 29 patients. There was no significant change in PANSS score with fish oil compared with evening primrose oil; the WMD was 12 (0.7 to 25).

Adverse effects

Adverse effects were uncommon. Palpitations, diarrhoea and nausea were reported with evening primrose oil in one study (Trial C). There was no significant difference in the reported incidence of adverse effects with the different study treatments compared with placebo.


Information was poorly reported in these trials. The dose of fatty acid used and the severity of schizophrenia were not stated in three of the four included studies. One study reported 37% of patients to have been lost to follow-up in each treatment arm which means the validity of the study is questionable. All of the studies were small; the largest had only 115 patients split between four treatment groups. Since the trials assessed different outcomes, the analyses were based on the results of single trials and the quantitative estimates of treatment effect cannot be robust. Significant superiority of fish oil over placebo for both change in average total PANSS score and need for antipsychotics was reported in one study, but it was not stated how great a reduction in PANSS score would make a clinically relevant difference to a patient. The NNT suggested that one in three patients given fish oil would not require antipsychotic medication who would have done with placebo (in schizophrenic patients with a new episode of illness). This suggests that fish oil is effective. The problem is that more trial data is needed from studies which assess the same outcomes, before the true effect of fish oil can be determined.