In the second phase of our study, these same subjects were then supplemented with MaxEPA fish oil (Seven Seas Healthcare Ltd, Hull) daily for 6 wk. Each capsule contained 171 mg of EPA and 114 mg of DHA. The scores on both PANSS and AIMS scales improved significantly (Mellor et al., 1996). Total PANSS score fell from 79 ± 18 to 66 ± 9 (p < 0.002). The AIMS score fell from 20 ± 7 to 12 ± 5 (p > 0.001). RBC membrane levels were measured by the method of Manku et al. (1983). Laboratory staff were blind to the coding system. RBC membrane levels of EPA as well as other n-3 PUFA further down the metabolic chain and also the total n-3 fatty acid content increased significantly during supplementation. Multiple regression analysis showed that changes in RBC membrane total n-3 PUFA were strongly associated with changes in total PANSS scores.
Previously a small number of studies had attempted to treat schizophrenia by PUFA supplementation. None of them showed the sort of encouraging results that our group achieved. For example, Vaddadi et al. (1986) treated 21 neuroleptic-resistant inpatients with depot neuroleptic medication and dihomo-y-linolenic acid (DGLA), placebo depot medication and DGLA, or placebo depot medication and placebo DGLA. There were no significant treatment effects on behavior or schizophrenic symptomatology. In a subsequent study, evening primrose oil was given to patients with movement disorders who were predominantly schizophrenic and who had long-term exposure to neuroleptic medication (Vaddadi et al., 1989). The study used a crossover design with 16-wk treatment periods, making the study difficult to interpret because of potential carryover effects. There was a significant treatment effect on schizophrenic symptoms but no clinically important effect on tardive dyskinesia. In another small study, evening primrose oil did not benefit either schizophrenic symptoms or tardive dyskinesia (Wolkin et al., 1986). Supplementation with n-3 in the form of linseed oil (50% a-linolenic acid) was given to five patients in an open study, with some apparent benefit (Rudin, 1981).
Encouraged by our open-pilot study, we embarked upon a double-blind, placebo-controlled trial of n-3 supplementation (Peet et al., 2001). Subjects for the study were schizophrenics diagnosed according to DSM-IV (American Psychiatric Association) criteria who were still significantly symptomatic even with medical treatment that was considered optimal by the responsible psychiatrist. All patients were on stable medication and none suffered from significant physical illness or other psychiatric disorders. Patients were taking a variety of antipsychotic drugs in both oral and depot preparations, and some were also taking anticholinergic medication to treat side effects. Patients continued on their normal medication, which the responsible consultant psychiatrist was asked to keep stable during the 3 mo of treatment. Patients were randomly allocated to treatment with either an EPA-enriched oil (Kirunal, Laxdale Limited, Stirling, UK), a DHA-enriched oil, or a corn oil placebo. Patients were rated on the PANSS scale at the beginning and end of treatment. In addition, blood was taken for the measurement of RBC PUFA levels as described earlier, at the beginning and end of treatment.
Patients were well matched at baseline for age, gender, and severity of illness, with no significant differences among the three treatment groups. Positive symptoms, general psychopathology, and total PANSS score improved most in the EPA-supplemented patients. There was no treatment effect on negative symptoms. The treatment effect for positive symptoms reached statistical significance even with this small group of patients. Using a 25% improvement criterion, EPA was significantly superior to both DHA and placebo (p < 0.05).
This was a treatment-resistant group of patients who were substantially symptomatic despite current orthodox clinical treatment. The average 25% improvement in positive schizophrenia symptoms in the EPA group was significant not only statistically but was also clinically meaningful. Furthermore, the EPA treatment was not associated with adverse clinical side effects.
As expected, the highest levels of EPA and of DHA in RBC membranes of 12 patients from each group were found in the appropriate treatment groups. This confirms that the patients as a group were taking their medication and that it was having the predicted effect on membrane PUFA levels. No correlation was found between these changes in membrane PUFA levels and clinical improvement.
The apparent differential response between EPA- and DHA-enriched oils was unexpected and requires explanation. The brain contains only small amounts of EPA relative to the large quantities of DHA (Horrobin et al., 1991). It is therefore uncertain that dietary
EPA would produce any marked or physiologically significant changes in neuronal membrane EPA levels. It is possible that EPA is working through a quite different mechanism than incorporation into membranes. One possibility is through the inhibition of PLA2. It has been shown in epidermal cells in vitro that PLA2 is inhibited by EPA but not by DHA (Finnen & Lovell, 1991). However, it is unclear whether this is the same form of PLA2 that is elevated in schizophrenia. Both EPA and DHA can induce glutathione peroxidase in human endothelial cells. Glutathione peroxidase may then protect cells and tissues from oxidative damage (Crosby et al., 1996). By these two mechanisms, EPA could be preventing both enzymatic membrane breakdown and oxidative damage. This is, however, very speculative.
This initial study has been followed up by a double-blind, placebo-controlled trial of ethyl-EPA at three dosage levels (1, 2, and 4 g) used as an adjunct to existing antipsychotic treatment (Peet et al., unpublished data). Patients were taking either typical antipsychotic drugs such as clopromazine and haloperidol, new atypicals such as olanzapine, or clozapine. The overall analysis so far has suggested that the 2 g dosage of EPA is the most effective. Patients on typical antipsychotic drugs responded poorly to the addition of ethyl-EPA to their treatment, but those patients already taking clozapine showed substantial extra benefit in clinical symptomatology relative to placebo. This is of particular interest because clozapine is currently regarded as the most effective treatment for schizophrenia, and it is used in cases which have failed to respond to other treatments. Clozapine is reserved as a last-resort treatment because of its tendency to produce agranulocytosis, which requires careful blood monitoring of patients taking the drug. Another unwanted effect is elevation of triglyceride levels. In this study, it was found that treatment with 2 or 4 g of ethyl-EPA normalized the elevated triglyceride levels. Also of interest is the effect of the different dosages of ethyl-EPA on RBC membrane fatty acid levels. There was a dose-related increase of membrane EPA levels. The 2-g dose of ethyl-EPA also resulted in elevations of DHA and AA. These changes were particularly marked in the clozapine-treated patients. In contrast, the 4-g dose of ethyl-EPA resulted in a decrease of membrane AA and DHA, presumably because of competition for incorporation into the membrane. Multiple-regression analysis suggested that increased membrane levels of AA were the most important determinant of clinical response.
Further studies have been conducted into the possibility of using EPA as a sole treatment for schizophrenia. Puri et al. (1998) have reported a single case who showed dramatic remission following treatment with EPA. This one case was investigated in great detail. It was found that during EPA treatment, previously low membrane levels of PUFA were normalized (Richardson et al., 2000). Also, progressive enlargement of the cerebral ventricles reversed following EPA treatment (Puri et al., 2000). No other treatment has been shown to have this effect in schizophrenic patients. We have now conducted a double-blind trial of EPA as a sole treatment in schizophrenia. The work was done with collaborators in Baroda, India. The protocol for this study was that subjects were schizophrenic patients who came for treatment at the outpatient clinic but who were not currently on any antipsychotic medication. About a third of the patients had never been previously treated, whereas others had received antipsychotic medication in the past. Thirty patients were treated at random with either EPA or placebo, and the treating clinicians were asked not to prescribe antipsychotic medication unless this was clinically imperative. The main outcome measure was the requirement for antipsychotic drugs. In addition, the PANSS rating scale was carried out at the beginning and end of the study.
When the code was broken after 3 mo of treatment, it was found that all patients in the placebo group required antipsychotic medication by the end of the study, whereas one-third of the patients on EPA completed the 3-mo trial period off antipsychotic medication. Avoiding the use of antipsychotic medication is clinically interesting only if patients did as well as those who received the conventional treatment together with placebo. The PANSS scores demonstrated that there was no significant difference between the patients treated with EPA or placebo at baseline, but by the end of the 3-mo treatment period, the EPA-treated patients had significantly lower PANSS scores than did the placebo group. Thus, the patients taking EPA had a better outcome even though a third of them were being managed without antipsychotic drugs by the end of the trial period (Peet et al., 2000).
Taking the evidence as a whole, it is apparent that the greatest effect of EPA on schizophrenic symptoms is seen in those patients who are either on no antipsychotic treatment or who are taking clozapine. Antipsychotic drugs such as clopromazine are known to have marked effects on some of the enzymes that affect phospholipid metabolism, and high-dose antipsychotic medication can in itself lead to depleted membrane levels of PUFA (Fischer et al., 1992). There is evidence that clozapine does not have this effect (Horrobin, 1999). Thus, the therapeutic effect of EPA may be impaired by concurrent treatment with neuroleptic drugs that damage cell membrane phospholipids.
In summary, existing drug treatments for schizophrenia are of limited efficacy and have substantial side effects. New treatment can arise only on the basis of a new hypothesis. The phospholipid hypothesis of schizophrenia provides the theoretical basis for treatment with PUFA supplementation. Previous studies using n-6 supplementation have had mixed results. We now have evidence from a double-blind, placebo-controlled trial that EPA, but not DHA, is effective in reducing the symptoms of schizophrenia. It is possible that the response to EPA is impaired by concomitant treatment with antipsychotic drugs that damage membrane phospholipids. The best treatment effects of EPA have been seen in patients who are otherwise unmedicated or who are currently taking clozapine. This remains to be explored further.
Was this article helpful?