Evidence strengths and limitations

Evidence strengths and limitations

The methodological quality of the 8 randomised controlled trials (RCTs) included in this evidence summary was good because in most of the cases, the process of randomisation, blinding and allocation concealment was described adequately.

However, overall, the evidence base for using omega-3 fatty acid medicines off-label to treat schizophrenia is limited. The studies were small. The largest study was by Peet and Horrobin (2002) (n=122) but some of the evidence came from studies recruiting fewer participants, including key studies (Peet et al. 2001a and Peet et al. 2001b) that included just 14 or 15 people in the analysis of the omega-3 fatty acid arm of the trial.

Data reporting in some of the larger trials was assessed by the Cochrane review as being selective or incomplete. This meant that potentially valuable information was not usable or comparable between studies.

The studies were short in duration. None assessed efficacy outcomes for longer than 16 weeks. Short-term fluctuations in schizophrenia symptoms may have biased the results of these trials. The outcomes were measured in these trials using clinician-rated symptom scales (such as Positive and Negative Syndrome Scale [PANSS] or Clinical Global Impression [CGI]), and did not include important patient-orientated outcomes, such as social functioning, ability to work, patient and carer satisfaction, or family burden.

The study that recruited people with schizophrenia who had no previous treatment with antipsychotics (Peet et al. 2001b) had very significant limitations because initiation of antipsychotic drugs was permitted, and occurred, part way through the trial in both omega-3 fatty acid and placebo treatment groups. Therefore, it is unclear whether outcome differences between the groups were related to omega-3 fatty acids or coincided with the initiation of antipsychotic drugs.

The evidence from Peet et al. (2001b) and Berger et al. (2007) is more applicable to people newly diagnosed with schizophrenia experiencing a first episode, whereas the other RCTs are more relevant to people with chronic schizophrenia.