Depression and erythrocyte folate levels among women in the Southampton Women's Survey: cross-sectional and longitudinal analyses

Inskip, H., Dunn, N., Robinson, S., Oestmann, A., Barnett, J., Godfrey, K., Cooper, C. and Kendrick, T. (2007) Depression and erythrocyte folate levels among women in the Southampton Women's Survey: cross-sectional and longitudinal analyses. Proceedings of the Nutrition Society, 66 (Supp/OCA-B), p.105A. (doi:10.1017/S0029665107005952).

Abstract

Lower blood folate levels have been associated with depression in several cross-sectional surveys, but longitudinal studies are needed to assess whether depression is a consequence or a cause of folate deficiency.
The Southampton Women’s Survey (SWS) comprises a cohort of 12 500 non-pregnant women recruited from the general population between 1998 and 2002 who are being followed through subsequent pregnancies. The SWS has been shown to be broadly representative of the general population1. From March 2000 all 7210 women recruited into the study were asked to complete a GHQ12 questionnaire to assess depression and anxiety2. All women were asked to provide venous blood samples from which erythrocyte folate (RCF) was assayed using an Abbott microparticle enzyme immunoassay IMx-folate kit and an IMx analyzer. Consent was obtained from the women to access their general practitioner (GP) records to obtain evidence of incident depression over the 2-year period following the baseline interview. Complete GHQ12 data were provided by 7020 women (97%) and RCF measurements were obtained for 5051 (72%) of them. Among the 5051 women 1588 (31%) were identified as depressed at the baseline survey. Using Poisson regression modelling with RCF as a continuous variable it was found that RCF was inversely associated with the risk of depression. The prevalence ratio associated with an increase in RCF of 100 nmol/l was 0.985 (95% CI 0.974, 0.995; P=0.005); those with RCF levels <960 nmol/l were 14% more likely to be depressed than those with higher levels. The association was attenuated after adjustment for confounding factors (P=0.05) but still indicated that lower RCF levels were linked to depression (prevalence ratio 0.98 (95% CI 0.97, 1.00; P=0.05). Follow-up data were available for 3996 women whose RCF levels had been measured (79%). The 1264 women who were identified as depressed at baseline either from the GP notes or from the GHQ12 questionnaire were excluded. Of the remaining 2732 women 307 (11%) had an incident episode of depression recorded by their GP in the 2 years following baseline interview. In a Cox regression model no relationship between RCF and incident depression was identified; the unadjusted hazard ratio per 100 nmol/l increase in RCF was 0.99 (95% CI 0.96, 1.02; P=0.4) and the adjusted hazard ratio was 1.00 (95% CI 0.97, 1.03; P=0.9). The finding of an association between RCF and prevalence of depression in a cross-sectional analysis but not with incident depression during follow-up of the cohort indicates that lower RCF levels may be more a consequence than a cause of depression. However, a positive association with incident depression cannot be excluded as smaller numbers of women contributed to the follow-up than to the cross-sectional analysis. Nonetheless, the 95% CI associated with the hazard ratio indicates that even a 100 nmol/l increase in RCF would be unlikely to reduce the risk of depression by more than about 3%. Thus, the contribution of low folate status to depression appears modest, and folate supplementation to prevent depression does not seem warranted on the basis of these findings.

This record has no associated files available for download.

Contributors

Author: H. Inskip

Author: S. Robinson

Author: K. Godfrey

Author: C. Cooper

Download statistics