AIM: To establish whether there exists a causal association between omega 3 polyunsaturated fatty acids (PUFAs) and depression

HYPOTHESIS: We hypothesise that maternal levels of omega 3 PUFAs could be a contributing factor in postnatal depression (PND)

VARIABLES: Genetic variants will be used as instrumental variables for the omega 3 PUFAs docosahexaenoic acid (DHA) and eicosapentaenoic acid (EPA). The outcome variable will be PND as measured by the Edinburgh Postnatal Depression Scale (EPDS) completed at 8 weeks postpartum.

Literature surrounding the association between PUFAs and depression reports conflicting results regarding efficacy. A number of observational studies reported in the psychiatric literature suggest that a deficiency of maternal DHA can lead to the development of major depressive disorder (MDD) during pregnancy and postpartum, whereas clinical benefits of supplementation with EPA are reported in more intervention studies (Mozurkewich et al., 2011). In an attempt to clarify the situation, meta-analyses of supplementation with omega 3 PUFAs have been performed which point to EPA, rather than DHA, being the key PUFA associated with depression (Martins et al., 2012; Freeman et al., 2006). However, these analyses may be limited due to issues including small sample sizes and heterogeneity between studies.

We plan to use Mendelian Randomisation (MR) to explore the association between maternal genotype and occurrence of postnatal depression. Genetic variants discovered through genome-wide association studies (GWAS) (Lemaitre et al., 2011; Tanaka et al., 2009) will be used as instrumental variables for the PUFAs. The use of MR should help to eliminate the issues with confounding and reverse causation to which conventional observational studies are subject.

Women completing the EPDS at 8 weeks postpartum will be classified as cases or controls depending on their score, with those scoring greater than 12 being designated as cases (Cox et al., 1987). This threshold will help to distinguish women who are likely to be suffering from a depressive illness of varying severity. Given the nutritional demands of breast feeding on the mother, we will investigate whether the effect of fatty acids differs according to the occurrence of breast-feeding. A secondary analysis will be carried out using EPDS scores at 8 months postpartum.

Participants will be excluded from the study if they experienced a stillbirth or early neo-natal death (within 27 days).