Women are at particular risk of high levels of depressive symptoms during pregnancy and postnatally, with high prenatal levels being more common than high postnatal levels.1 The adverse effects of postnatal depression on the development and subsequent behavior of the child are well known.2,3 Prenatal depression may strain the marital relationship and have adverse effects on how the woman feels about the pregnancy. Depressive symptoms in late pregnancy have been associated with increased risk of postnatal depression4 as well as increased risk of operative delivery and admission of the child to a special care baby unit.5 The prevention of such depressive symptoms could have major public health advantages.
A committee of the American Psychiatric Association has recommended consuming 2 or more portions of fish per week to avoid depression among the nonpregnant adult population, with additional omega-3 supplementation in the presence of emotional disorders.6 No large studies have previously assessed whether women who avoid fish in pregnancy are more likely to develop depression than those who eat fish regularly. This question is important because the current recommendations in many countries are for pregnant women to reduce the amount of fish they eat because of fears of harm to the developing fetus resulting from contamination of fish with mercury.7 Our analysis of information collected in the early 1990s is designed to address the issue of whether women who eat fish are at lower risk of developing symptoms of depression than those who have minimal fish intake and, in particular, whether omega-3 fatty acids from fish have a protective effect.
The Avon Longitudinal Study of Parents and Children (ALSPAC) was initiated to identify features of the environment that influence the health and development of children and their parents.8 The initial population consisted of 14,541 pregnant women who resided in the Avon health authority area in southwest England and had an expected date of delivery between April 1991 and December 1992. Data were collected using a variety of strategies.9 This paper uses the information collected on diet and depressive symptoms from a maternal self-completed questionnaire at 32 weeks’ gestation. Women with outcomes of fetal or infant death, and those with multiple births were excluded.
The Edinburgh Postnatal Depression Scale (EPDS) was completed by the mother at 32 weeks’ gestation. This scale was devised to measure maternal depression after childbirth,10 but has also been validated for use during pregnancy.11 The scale was designed to exclude symptoms ascribable to somatic effects of pregnancy and childbirth (eg, weight gain, sleeplessness, and tiredness). It is a 10-item psychometric rating scale, each item having 4 responses from 0 to 3 with a minimum score of 0 and a maximum score of 30. Scores on the EPDS can be categorized by well-validated cut-off points; those scoring 13 or more are classified as having high levels of depressive symptoms. Sensitivities and specificities of 95% and 93%, respectively, have been achieved by identifying women with scores of 13 or more and comparing the result with clinical interview by using the Diagnostic and Statistical Manual of Mental Disorders III (DSM III)criteria for depression.12
Maternal food consumption was estimated at 32 weeks’ gestation using a self-reported food frequency questionnaire, which was also used to derive energy intake.13 Three questions assessed seafood consumption: “How many times nowadays do you eat (a) white fish (cod, haddock, plaice, fish fingers, etc.), (b) dark or oily fish (tuna, sardines, pilchards, mackerel, herring, kippers, trout, salmon, etc.), or (c) Shellfish (prawns, crabs, cockles, mussels etc.)?” Each response had 5 predefined categories: never or rarely, once in 2 weeks, 1–3 times per week, 4–7 times per week, and more than once a day. These were converted to weekly frequencies of consumption (portions per week) of 0, 0.5, 2, 5.5, and 10, respectively. Portion sizes and types of fish were based on typical consumption patterns in Britain at the time. Fatty acid compositions were calculated using British food composition tables.14 Seafood was virtually the only source of omega-3 in the diet in Britain at this time; only 221 (2%) of the study women were taking omega-3 supplements. The food frequency questionnaire also provides information on the frequency of consuming 12 other food groups: sausages or burgers, pies or pastries, meat, poultry, cabbage or similar vegetables, other green vegetables, salad, chips (French fries), fresh fruit, fruit juice, crisps (potato/corn chips), and biscuits (cookies).
The following factors were considered as potential confounders because of an association with depression or with fish eating or both: maternal age (<25, 25+ years); parity, the number of previous pregnancies resulting in a live birth or a late fetal death (0, 1, 2+); outcome of immediately preceding pregnancy (none, survivor, other); maternal education based on the highest educational qualification achieved (low: no more than a vocational qualification; medium: O-level or equivalent; high: A-level or higher); housing tenure (owned/mortgaged, council rented [public housing], other); crowding, persons per room in pregnancy (<1, 1+); the mothers’ life events in childhood scale, a 108-point measure concerning major events (<21, 21+); a scale of 44 recent life events (<90th, ≥90th centile); chronic stress as measured by a Family Adversity Index (<90th centile, 90th centile +); maternal smoking (no, yes); alcohol (none, any); and maternal ethnic origin (white, nonwhite).
Three categories of fish consumption were created to assess the relationship of the results to the US Advisory to limit seafood consumption during pregnancy7: (a) none, (b) consumption between 1 and 340 g/wk, and (c) consumption greater than 340 g/wk. This approximates the following average portion of fish per week: none; less than 3; 3 or more. For analyses by estimated omega-3 intake from fish, the data were grouped into none; 0.01–0.4 g; 0.4–1.5 g; more than 1.5 g week. The group with the highest intake was taken as the reference group. Because the omega-3 fatty acid content differs according to the type of seafood, there is a strong but indirect relationship between the amount of fish consumed and estimated omega-3 intakes (Table 1).
We adjusted for energy intake to allow for different individual energy requirements. The adjustment was made using energy as a linear covariate. Potential confounders were considered in cross-tabulations and used in the logistic regression if there was no clear likelihood of reverse causality. Logistic regression results were presented as odds ratios with 95% confidence intervals (CIs) and the P value for trend. (Results were also examined for departures from trend but no evidence for nonlinearity was found.) Data for unadjusted analyses were restricted to the mothers for whom data were available on all confounders. Although logistic regression was used to model a linear relationship between the log odds of depressive symptoms and omega-3 from fish, other exponential curves were explored using nonlinear least squares on the adjusted means from a logistic regression model on omega-3 intake categorized as none and 6 equal groups. STATA version 8.2 was used for all analyses.
As reported elsewhere,1 the rate of high levels of depressive symptoms was 14% at 32 weeks’ gestation. Variations in rates of depressive symptoms with demographic, social, and lifestyle factors are shown in Table 2. Rates of depressive symptoms are highest with maternal youth, living in council-owned (public) accommodation, in overcrowded circumstances, low levels of maternal education, nonwhite mothers, high levels of adverse life events in childhood, high levels of recent adverse life events, high levels of chronic stressors as measured by the Family Adversity Index, loss of the preceding pregnancy, and parity of 2 or more. Variables concerning alcohol consumption and interpregnancy interval showed no consistent trend and were omitted from further analysis. Of these remaining 11 confounders, 8 were also associated with intakes of omega-3.
We have previously shown in this population15,16 that seafood consumption varies substantially with socioeconomic and lifestyle factors including many of those listed in Table 2. Before and after adjustment for these factors, together with the estimated total energy intake (Table 3), there was a trend (albeit with reduced effect sizes after adjustment) such that mothers who ate no seafood were about 50% more likely to have high levels of depressive symptoms compared with those who consumed substantive amounts of omega-3 from seafood.
The way in which the adjusted prevalence of depressive symptoms at 32 weeks varied with omega-3 intake from seafood is modeled in the Figure replacing the categorical omega-3 variable with its continuous equivalent. Differences between the 2 models were used as a test of deviation from linearity. This test showed no evidence of nonlinearity (P = 0.1379).
To assess whether omega-3 from seafood was a marker of other foods, of a healthier lifestyle related to high levels of depressive symptoms, we repeated the logistic regression analysis including the 12 food groups listed previously.13 This resulted in a reduction of effect sizes to 1.39 (1.12–1.72), 1.22 (1.00–1.49), and 1.15 (0.98–1.36) for omega-3 levels of none, 0.1–0.4 and 0.5–1.5 g/wk, respectively, compared with more than 1.5 g/wk (P trend = 0.0029). Exclusion of mothers taking fish oil supplements from the analyses had a negligible impact on these results.
To investigate whether the results were independent of a history of earlier depressive symptoms, analyses were repeated adjusting for high levels of depressive symptoms at 18 weeks’ gestation: results at 32 weeks were resilient to this adjustment (OR for none = 1.51; 1.21–1.88; P trend = 0.0004).
Although dietary data were available only at 32 weeks’ gestation, the associations between high levels of depressive symptoms and this estimate of omega-3 intake was explored at other times: 18 weeks’ gestation and 2 or 8 months postnatally. Weak associations were seen at 18 weeks’ gestation and 8 months postnatal but not at 2 months postnatal. However as there was a significant interaction across the 4 time points (P = 0.0017), this may suggest that either low omega-3 had a varying association with depressive symptoms or may just reflect that this estimated intake at 32 weeks was a less reliable indicator of actual intakes at other times.
We have demonstrated a dose-response relationship between omega-3 intake from seafood and increased risk of the mother having high levels of depressive symptoms during the last trimester of pregnancy, thus confirming our original hypothesis. Biologic evidence supports the possibility that low fish consumption is related to postnatal depression. Until recently, fish has been the major source of omega-3 fatty acids, docosahexaenoic acid (DHA) and eicosapentaenoic acid (EPA) in the British diet. Furthermore, low fish consumption tends to be associated with low blood levels of these fatty acids. Fatty acids are normally selectively concentrated in the nervous system and are necessary for optimal neural function. They cannot be made de novo in sufficient quantities at times of high demand, such as pregnancy.
Direct measurements have shown that women with postnatal depression have relatively low levels of DHA in plasma17 and other biologic indicators of DHA deficiency.18 An ecologic study found that the risk of postnatal depression was 50-fold greater in countries with little seafood consumption compared with those with the highest consumption; the likelihood that this was related to omega-3 fatty acids was imputed from the available levels of DHA in breast milk in those countries.19 One randomized controlled trial among pregnant women with depression reported lower depression scores and higher rates of clinical response (62%) to 3.5 g/d of omega-3 fatty acids compared with placebo (27%).20 Another randomized controlled trial reported a trend toward efficacy but was described by the authors as underpowered.21 Recent nonblinded trials of eicosapentaenoic acid and DHA supplementation have reported a 50% reduction in depressive symptoms.22,23
There are 2 prior epidemiologic studies (from New Zealand and Japan) that reported no association between seafood consumption and pregnancy-related depressive symptoms.24,25 The New Zealand study included only 80 women24 and consequently lacked power. Although the Japanese study had much larger numbers (865), it used a cutpoint of 9+ (rather than 13+) on the depression scale. Furthermore, almost all the subjects had a relatively high intake of oily fish with very few having no omega-3 from seafood.25 Sontrop et al,26 studying pregnant Canadian women, found robust associations between depressive symptoms and low intakes of omega-3 from seafood in current smokers and in single women, although not in women overall.
Limitations and Strengths
The estimated omega-3 fatty acid intakes from seafood are based on food frequency questionnaires. This method has potential problems of inaccurate recall and inaccurate assumptions concerning portion sizes. However, we have shown that the reported frequency of maternal seafood consumption in this study is positively related to biochemical markers of fish intake in subpopulations of this cohort. For example, maternal prenatal levels of DHA in the red cell membranes (as a proportion of total red blood cell fatty acids) varied from about 2.3% in mothers who ate no oily fish to 2.4% for those eating them once in 2 weeks and 2.9% in those eating fish more often (P < 0.0001 trend).27,28 Mercury is found in fish (as well as elsewhere in the environment). Levels of mercury in the umbilical cord were shown to increase with the amount of fish eaten–the levels when the pregnant woman ate no fish at all being considerably lower (0.10 μg/g) than those eating 4 or more fish meals a week (0.14 μg/g).
Although fish is a main contributor to dietary long chain omega-3 fatty acids, these fatty acids are also present in small amounts in a few other foods (such as meat); in addition, they may be produced from the precursor linolenic acid found in seed oils. Omega-3–enhanced foods were not available to the British public at the time of the survey.
As with all longitudinal studies, there was attrition over time. We have shown elsewhere with these data4 that the mother's emotional status in 1 questionnaire was associated with the likelihood of returning a subsequent questionnaire. Consequently, there are likely to be proportionately fewer women with high levels of depressive symptoms answering the dietary questionnaire. We believe that this is more likely to reduce the observed relationship than contribute to it.
We have assumed that the negative associations between seafood consumption and depressive symptoms were the result of deficiency in omega-3 fatty acids, but other components of fish (such as vitamin D) may be on the causal pathway. Mechanisms for attempting to unravel the components of any causal association could benefit from further randomized controlled trials (those reported so far have been for treatment rather than prevention) or from a Mendelian randomization approach,29 assessing the relationships with genetic variants related to the metabolic pathways of constituents of fish such as the vitamin D receptor gene or the FADS genes (for omega-3).
We took into account many social and environmental factors that are associated with the risk of depression, but acknowledge that there may be others that are linked both with depression and with fish eating that may explain this association. We have examined further possibilities such as other aspects of the diet and found that the association with omega-3 from seafood persisted, albeit with reduced effect sizes.
Public Health Implications
The primary prevention of maternal prenatal depression is important because antidepressant use in pregnancy may have adverse consequences on the developing child.30 Advice on the amount of fish to eat during pregnancy varies from country to country. In the United States, the advice from 2004 has been to limit seafood consumption in pregnancy to less than 3 portions a week.7 Because we have shown here that the risk of depressive symptoms was lowest among those consuming more than 1.5 g of omega-3 from seafood per week (≥3 portions of seafood per week), it is possible that limiting intake in accordance with this advice could increase the risk of maternal depressive symptoms during pregnancy.
We thank the families who took part in this study, the midwives for their help in recruiting them, and the whole Avon Longitudinal Study of Parents and Children team, which includes interviewers, computer and laboratory technicians, clerical workers, research scientists, volunteers, managers, receptionists, and nurses.
1. Evans J, Heron J, Francomb H, Oke S, Golding J. Cohort study of depressed mood during pregnancy and after childbirth. BMJ
2. O'Connor TG, Heron J, Glover V; ALSPAC Study Team. Antenatal anxiety predicts child behavioral/emotional problems independently of postnatal depression. J Am Acad Child Adolesc Psychiatry
3. Sharp D, Hay DF, Pawlby S, Schmücker G, Allen H, Kumar R. The impact of postnatal depression on boys’ intellectual development. J Child Psych Psychiatry
4. Heron J, O'Connor TG, Evans J, Golding J, Glover V; ALSPAC Study Team. The course of anxiety and depression through pregnancy and the postpartum in a community sample. J Affect Disord
5. Chung TK, Lau TK, Yip AS, Chiu HFK, Lee DTS. Antepartum depressive symptomatology is associated with adverse obstetric and neonatal outcomes. Psychosom Med
6. Freeman MP, Hibbeln JR, Wisner KL, et al. Omega-3 fatty acids: evidence basis for treatment and future research in psychiatry. J Clin Psychiatry
7. US Department of Health and Human Services, U.S. Environmental Protection Agency. What you need to know about mercury in fish and shellfish 2004 EPA and FDA advice for: women who might become pregnant women who are pregnant nursing mothers young children. Washington, DC. EPA-823-R-04-005 March 19, 2004. Available at: http://www.cfsan.fda.gov/∼dms/admehg3.html
. Accessed March 21, 2006.
8. Golding J, Pembrey M, Jones R; ALSPAC Study Team. ALSPAC–the Avon Longitudinal Study of Parents and Children. I. Study methodology. Paediatr Perinat Epidemiol
10. Cox JL, Holden JM, Sagovsky R. Detection of postnatal depression: development of the 10 item Edinburgh Postnatal Depression Scale. Br J Psychiatry
11. Murray D, Cox J. Screening for depression during pregnancy with the Edinburgh Postnatal Depression Scale (EPDS). J Reprod Infant Psychol
12. Harris B, Huckle P, Thomas R, Johns S, Fung H. The use of rating scales to identify postnatal depression. Br J Psychiatry
13. Rogers I, Emmett P; ALSPAC Study Team. Diet during pregnancy in a population of pregnant women in South West England. Avon longitudinal study of pregnancy and childhood. Eur J Clin Nutr
14. Ministry of Agriculture Fisheries and Food and Royal Society of Chemistry. Fatty Acids Supplement to Mccance and Widdowson's the Composition of Foods
. London: Ministry of Agriculture, Fisheries, and Foods/ Royal Society of Chemistry; 1998.
15. Rogers I, Emmett P, Baker D, Golding J; ALSPAC Study Team. Financial difficulties, smoking habits, composition of the diet, and birthweight in a population of pregnant women in the South West of England. Eur J Clin Nutr
16. Daniels JL, Longnecker MP, Rowland AS, Golding J; ALSPAC Study Team. Fish intake during pregnancy and early cognitive development of offspring. Epidemiology
17. Otto SJ, Houwelingen AC, Antal M, et al. Maternal and neonatal essential fatty acid status in phospholipids: an international comparative study. Eur J Clin Nutr
18. Otto SJ, de Groot RM, Hornstra G. Increased risk of postpartum depressive symptoms is associated with slower normalization after pregnancy of the functional docosahexaenoic acid status. Prostaglandins Leukot Essent Fatty Acids
19. Hibbeln JR. Seafood consumption, the DHA composition of mothers’ milk and prevalence of postpartum depression: a cross-national analysis. J Affect Disord
20. Su KP, Huang SY, Chiu TH, et al. Omega-3 fatty acids for major depressive disorder during pregnancy: results from a randomised, double-blind placebo-controlled trial. J Clin Psychiatry
21. Rees AM, Austin MP, Parker GB. Omega-3 fatty acids as a treatment for perinatal depression: randomised double-blind placebo-controlled trial. Aust N Z J Psychiatry
22. Freeman MP, Hibbeln JR, Wisner KL, Brumbach BH, Watchman M, Gelenberg AJ. Randomized dose-ranging pilot trial of omega-3 fatty acids for postpartum depression. Acta Psychiatr Scand
23. Freeman MP, Hibbeln JR, Wisner KL, Watchman M, Gelenberg AJ. An open trial of omega-3 fatty acids for depression in pregnancy. Acta Neuropsychiatrica
24. Browne JC, Scott KM, Silvers KM. Fish consumption in pregnancy and omega-3 status after birth are not associated with postnatal depression. J Affect Disord
25. Miyake Y, Sasaki S, Yokoyama T, et al. Risk of postpartum depression in relation to dietary fish and fat intake in Japan: the Osaka Maternal and Child Health Study. Psychol Med
26. Sontrop J, Avison WR, Evers SE, Speechley KN, Campbell MK. Depressive symptoms during pregnancy in relation to fish consumption and intake of n-3 polyunsaturated fatty acids. Paediatr Perinat Epidemiol
27. Newson R; ALSPAC Study Team. Multiple test procedures and smile plots. Stata J
28. Williams C, Birch EE, Emmett PM, Northstone K; ALSPAC Study Team. Stereoacuity at age 3.5 y in children born full-term is associated with prenatal and postnatal dietary factors: a report from a population-based cohort study. Am J Clin Nutr
29. Davey Smith G, Ebrahim S. ‘Mendelian randomization’: can genetic epidemiology contribute to understanding environmental determinants of disease? Int J Epidemiol
30. Feucht C. Treatment of depression during pregnancy. US Pharm