To examine the relationship between the intake of dairy products and calcium and the prevalence of depressive symptoms during pregnancy.
To examine the relationship between the intake of dairy products and calcium and the prevalence of depressive symptoms during pregnancy.
Kyushu Okinawa Maternal and Child Health Study (KOMCHS).
A cohort of 1745 pregnant Japanese women.
Dietary intake during the preceding month was assessed using a self-administered diet history questionnaire. Scores of 16 or higher on the Center for Epidemiologic Studies Depression Scale denoted depressive symptoms. Adjustment was made for age, gestation, region of residence, number of children, family structure, history of depression, family history of depression, smoking, secondhand smoke exposure at home and at work, job type, household income, education, and body mass index. In our analyses regarding dairy products in general, adjustment was also made for fish intake; in our analysis regarding calcium, adjustment was also made for the intake of saturated fatty acids, eicosapentaenoic acid plus docosahexaenoic acid, and vitamin D.
Depressive symptoms during pregnancy.
Higher intake levels of yogurt and calcium were independently related to a lower prevalence of depressive symptoms during pregnancy: the adjusted odds ratios between extreme quartiles were 0.69 (95% CI 0.48–0.99, P for trend = 0.03) and 0.59 (95% CI 0.40–0.88, P for trend = 0.006), respectively. No relationships were observed between the intake of all dairy products, milk, or cheese and depressive symptoms during pregnancy.
The current results suggest that a higher intake of yogurt and calcium may be associated with a lower prevalence of depressive symptoms during pregnancy.
From a prevention perspective, epidemiological research into the relationship between diet and depressive symptoms is important. The dietary factors most commonly examined in observational epidemiological studies regarding this issue include n−3 polyunsaturated fatty acids, fish, folate, and other B vitamins, and most studies have found no association between dietary variables and depressive symptoms. In our previous study, significant inverse associations were found between intake levels of fish, eicosapentaenoic acid, docosahexaenoic acid, and vitamin D and the prevalence of depressive symptoms during pregnancy, whereas a significant positive relationship was observed between the intake of saturated fatty acids and the prevalence of depressive symptoms during pregnancy in this population. The possible role of dairy products and calcium in depressive symptoms has received much less attention; moreover, the findings have been inconsistent.[3-8]
In Japan, calcium intake during pregnancy is below not only the recommended intake levels for Japan but also the average intake levels in other developed countries. To our knowledge, only one epidemiological study has addressed the association between calcium intake and depressive symptoms during pregnancy. A cross-sectional study of 114 pregnant Korean women showed that the calcium intake level was significantly higher in the group that scored low on depression (<10 points) than in the group that scored high on depression (≥10 points), based on the Beck Depression Inventory. This finding convinced us to investigate whether a higher calcium intake might have a protective effect against depressive symptoms in pregnant women. Dairy products are a major source of calcium intake. We analysed baseline data from the Kyushu Okinawa Maternal and Child Health Study (KOMCHS), and the current cross-sectional study was designed to determine the relationship between intake of dairy products and calcium and the prevalence of depressive symptoms during pregnancy in Japanese women.
The present study was based on the KOMCHS.[2, 10-13] The KOMCHS is a prospective pre-birth cohort study designed to clarify the risks for maternal and child health problems. The baseline survey of the KOMCHS was administered by obstetric hospitals in the following prefectures in Japan.
In all cases, the obstetric hospitals gave as many pregnant women as possible a set of leaflets explaining the KOMCHS, an application form to participate in the study, and a self-addressed and stamped return envelope. Pregnant women who intended to participate in the KOMCHS mailed the application form to the data management centre. In the end, a total of 1757 pregnant women between 5 and 39 weeks of gestation gave their written informed consent to participate and answered a self-administered questionnaire in the baseline survey. After 12 pregnant women were excluded because of incomplete data on the variables under study, a total of 1745 pregnant women remained available for analysis. All procedures were approved by the Faculty of Medicine, Fukuoka University.
In the baseline survey, 595 (34.1%), 1004 (57.5%), and 146 (8.4%) participants answered a two-part questionnaire in the first (5–15 weeks of gestation), second (16–27 weeks of gestation), and third (28–39 weeks of gestation) trimesters, respectively. The answered questionnaire was mailed to the data management centre. By telephone, research technicians completed missing or illogical data.
In the first part of the questionnaire, the following data were obtained: age; gestation; region of residence; number of children; family structure; personal history of doctor-diagnosed depression; family history of depression; smoking habits; secondhand smoke exposure at home and at work; employment status; household income; and educational level. A family history of depression was defined as being present if one or more parents, or siblings, of the participants had been diagnosed with depression by a doctor. Information on employment status in the year when the baseline survey was performed or in the previous year was also requested, and women were defined as unemployed if they were unemployed both in the year the baseline survey was conducted and in the preceding year.
Based on a Japanese version of the Center for Epidemiologic Studies Depression Scale (CES-D),[14, 15] which was included in the first part of the questionnaire, the presence of depressive symptoms was measured. The CES-D is a 20-item, self-reported scale designed to assess the frequency of a variety of depressive symptoms within the previous week. Each item is rated on a four-point scale from 0 (rarely) to 3 (most or all of the time), and the CES-D generates a total score with a range between 0 and 60. Consistent with the validation studies,[14, 15] a cut-off score of ≥16 was used to classify participants with depressive symptoms.
As for the second part of the questionnaire, we used a semi-quantitative, comprehensive diet history questionnaire (DHQ) to assess dietary habits during the previous month.[16, 17] We calculated estimates of daily intake of foods (including a total of 150 foods), energy, and selected nutrients using an ad hoc computer algorithm for the DHQ based on the Standard Tables of Food Composition in Japan. The total intake of dairy products was defined as the sum of the intakes of full-fat milk, low-fat milk, yogurt, cheese, and cottage cheese. Milk intake was defined as the sum of the intakes of full-fat milk and low-fat milk. Cheese intake was defined as the sum of the intakes of cheese and cottage cheese. For the five food items, consumption frequency was determined using eight categories (ranging from ‘two or more times per day’ to ‘less than once a month’), and relative portion size was determined using five categories (ranging from ‘50% or less’ to ‘50% or more’) that were compared with a standard portion size. Data on supplements were not incorporated into the analysis because of the absence of a reliable composition table for dietary supplements in Japan. Also, a minority of study subjects (5.6%) used supplemental calcium on a weekly or more frequent basis. In a validation study of 92 Japanese women, a Pearson's correlation coefficient between the DHQ and 16-day weighed dietary records was 0.56 for calcium (S. Sasaki, unpubl obs.). All dietary variables were adjusted for total energy intake using the residual method. Information on self-reported body weight and height was obtained using the DHQ. We calculated body mass index as weight (kg) divided by the height squared (m2).
Study participants were categorised into quartile groups according to their intake levels of the dietary factors under study. We selected the following variables a priori as potential non-dietary confounding factors: age; gestation; region of residence; number of children; family structure; history of depression; family history of depression; smoking; secondhand smoke exposure at home and at work; job type; household income; education; and body mass index. As for potential dietary confounding factors, fish intake was also controlled for when the relationships between the intake of any dairy product and depressive symptoms were examined, whereas intake levels of saturated fatty acids, eicosapentaenoic acid plus docosahexaenoic acid, and vitamin D were also adjusted for when the association between calcium intake and depressive symptoms was investigated. Age, gestation, body mass index, and dietary confounding factors were used as continuous variables.
Crude odds ratios (ORs) and 95% confidence intervals (95% CIs) of depressive symptoms during pregnancy were estimated for the quartile groups of the dietary factors under study using logistic regression analysis, and the lowest quartile served as the reference group. To adjust for potential confounding factors, multiple logistic regression analysis was used. Tests for a linear trend were conducted in the logistic regression model assigning consecutive integers (1–4) to the quartiles of the exposure variables. All analyses were conducted using sas 9.2 (SAS Institute, Inc., Cary, NC, USA). No correction for multiple comparisons was made.
Among the 1745 study subjects, the prevalence of depressive symptoms during pregnancy was 19.3%. The mean age of the study subjects was 31.2 years (Table 1). About 5% of the subjects had a personal history of depression, and 10% reported a family history of depression. Mean total energy consumption and mean energy-adjusted intake of total dairy products and calcium during pregnancy per day were 7434.2 kJ, 142.8 g, and 502.8 mg, respectively.
|Age (years), mean ± SD||31.2 ± 4.3|
|Gestation (weeks), mean ± SD||18.5 ± 5.4|
|Region of residence|
|Fukuoka Prefecture||971 (55.6)|
|Other than Fukuoka Prefecture in Kyushu||592 (33.9)|
|Okinawa Prefecture||182 (10.4)|
|Number of children|
|Nuclear family structure||1474 (84.5)|
|History of depression||84 (4.8)|
|Family history of depression||175 (10.0)|
|Having ever smoked||563 (32.3)|
|Ever experiencing secondhand smoke exposure at home||1315 (75.4)|
|Ever experiencing secondhand smoke exposure at work||1106 (63.4)|
|Job type a|
|Professional or technical||435 (24.9)|
|Clerical or related occupation||328 (18.8)|
|Household income, ¥ per year|
|<13 years||428 (24.5)|
|13–14 years||577 (33.1)|
|≥15 years||740 (42.4)|
|Body mass index (kg/m 2 ), mean ± SD||21.4 ± 2.8|
|Daily intake c|
|Total energy (kJ), mean ± SD||7434.2 ± 2057.0|
|Total dairy products (g), mean ± SD||142.8 ± 125.0|
|Milk (g), mean ± SD||102.9 ± 111.9|
|Yogurt (g), mean ± SD||34.8 ± 39.5|
|Cheese (g), mean ± SD||5.1 ± 6.9|
|Calcium (mg), mean ± SD||502.8 ± 172.7|
|Fish (g), mean ± SD||46.7 ± 25.8|
|Saturated fatty acids (g), mean ± SD||16.8 ± 4.3|
|Eicosapentaenoic acid plus docosahexaenoic acid (g), mean ± SD||0.46 ± 0.29|
|Vitamin D (μg), mean ± SD||5.7 ± 3.0|
Table 2 presents the distributions of confounding factors in relation to dietary intake levels of total dairy products, yogurt, and calcium. These three intake levels were positively associated with age, household income, and educational level, and were inversely associated with number of children and having ever smoked. Total dairy product intake was positively associated with gestation at baseline. Yogurt intake was positively associated with fish intake, and was inversely associated with gestation at baseline, ever experiencing secondhand smoke exposure at home and at work, and body mass index. Calcium intake was positively associated with the intake of saturated fatty acids, eicosapentaenoic acid plus docosahexaenoic acid, and vitamin D.
|Total dairy products||Yogurt||Calcium|
|Q1||Q4||P for trenda||Q1||Q4||P for trenda||Q1||Q4||P for trenda|
|Age, years, mean||30.7||31.4||0.008||30.9||31.8||0.0005||30.6||31.8||<0.0001|
|Gestation, weeks, mean||18.3||19.1||0.046||19.1||18.2||0.02||18.5||18.8||0.47|
|Region of residence, %|
|Other than Fukuoka Prefecture in Kyushu||32.6||33.2||33.0||33.9||30.3||35.0|
|Number of children, %|
|Nuclear family structure, %||84.6||86.0||0.42||81.9||85.1||0.13||82.3||87.9||0.053|
|History of depression, %||4.4||4.4||0.85||4.6||5.5||0.49||4.1||4.8||0.69|
|Family history of depression, %||10.8||10.1||0.91||12.2||9.8||0.24||10.1||9.6||0.69|
|Having ever smoked, %||40.6||27.9||<0.0001||38.5||25.2||<0.0001||38.5||27.7||0.0003|
|Ever experiencing secondhand smoke exposure at home, %||76.6||73.5||0.12||78.2||72.5||0.04||76.6||73.5||0.17|
|Ever experiencing secondhand smoke exposure at work, %||69.5||62.0||0.16||67.4||61.8||0.03||65.8||63.2||0.40|
|Job type, % b|
|Professional or technical||22.9||26.3||21.1||29.8||23.6||25.9|
|Clerical or related occupation||14.2||20.6||17.2||21.7||16.3||20.4|
|Household income, ¥ per year, %|
|Body mass index, kg/m 2 , mean||21.5||21.4||0.52||22.0||21.2||<0.0001||21.7||21.4||0.08|
|Dietary intake d|
|Fish, g, mean||46.7||43.6||0.09||43.3||47.1||0.002|
|Saturated fatty acids, g, mean||14.5||18.9||<0.0001|
|Eicosapentaenoic acid plus docosahexaenoic acid, g, mean||0.38||0.50||<0.0001|
|Vitamin D, μg, mean||4.6||6.7||<0.0001|
Compared with yogurt intake in the lowest quartile, intake in the highest quartile was significantly associated with a lower prevalence of depressive symptoms during pregnancy (Table 3). After adjustment for the non-dietary confounding factors under study and fish intake, the inverse association remained significant: the adjusted OR between extreme quartiles was 0.69 (95% CI 0.48–0.99, P for trend = 0.03). No evident relationships were observed between the intake of total dairy products, milk, or cheese and depressive symptoms during pregnancy in the multivariate model. After adjustment for the non-dietary confounding factors under study and intake of saturated fatty acids, eicosapentaenoic acid plus docosahexaenoic acid, and vitamin D, higher calcium intake was independently related to a lower prevalence of depressive symptoms during pregnancy: the adjusted OR between extreme quartiles was 0.59 (95% CI 0.40–0.88, P for trend = 0.006).
|Variable||Quartile||P for trend|
|1 (lowest) (n = 436)||2 (n = 436)||3 (n = 436)||4 (highest) (n = 437)|
|Total dairy products|
|Depressive symptoms, %b||22.0||19.5||15.6||19.9|
|Crude OR (95% CI)||1.00||0.86 (0.62–1.19)||0.66 (0.46–0.92)||0.88 (0.64–1.22)||0.23|
|Adjusted OR (95% CI)c||1.00||0.87 (0.61–1.22)||0.70 (0.49–1.001)||0.93 (0.66–1.32)||0.47|
|Depressive symptoms, %b||21.6||19.7||17.2||18.5|
|Crude OR (95% CI)||1.00||0.89 (0.64–1.24)||0.76 (0.54–1.06)||0.83 (0.59–1.15)||0.17|
|Adjusted OR (95% CI)c||1.00||0.89 (0.63–1.26)||0.79 (0.56–1.13)||0.89 (0.63–1.25)||0.39|
|Depressive symptoms, %b||23.4||19.7||16.7||17.2|
|Crude OR (95% CI)||1.00||0.81 (0.58–1.11)||0.66 (0.47–0.92)||0.68 (0.49–0.95)||0.01|
|Adjusted OR (95% CI)c||1.00||0.80 (0.57–1.13)||0.70 (0.49–0.99)||0.69 (0.48–0.99)||0.03|
|Depressive symptoms, %b||18.4||20.4||21.6||16.7|
|Crude OR (95% CI)||1.00||1.14 (0.82–1.60)||1.22 (0.88–1.71)||0.89 (0.63–1.27)||0.65|
|Adjusted OR (95% CI)c||1.00||1.06 (0.75–1.51)||1.19 (0.84–1.68)||0.86 (0.59–1.24)||0.58|
|Depressive symptoms, %b||23.4||19.5||17.0||17.2|
|Crude OR (95% CI)||1.00||0.79 (0.57–1.10)||0.67 (0.48–0.93)||0.68 (0.49–0.95)||0.01|
|Adjusted OR (95% CI)d||1.00||0.75 (0.53–1.06)||0.63 (0.44–0.92)||0.59 (0.40–0.88)||0.006|
A significant positive correlation was found between the intakes of yogurt and calcium, with a Pearson's correlation coefficient of 0.45 (P < 0.0001). After further adjustment for calcium intake as a continuous variable, the inverse association between yogurt intake in the highest quartile and depressive symptoms during pregnancy had completely disappeared, although the inverse linear trend remained significant: the further adjusted OR between extreme quartiles was 0.70 (95% CI 0.47–1.03, P for trend = 0.047).
After restriction to the 595 pregnant women who completed the baseline survey in the first trimester, the adjusted OR between extreme quartiles was 0.46 (95% CI 0.26–0.82, P for trend = 0.03) for yogurt and 0.56 (95% CI 0.30–1.05, P for trend = 0.08) for calcium. The corresponding figure for 1004 pregnant women who completed the baseline survey in the second trimester was 0.86 (95% CI 0.52–1.43, P for trend = 0.27) for yogurt and 0.51 (95% CI 0.28–0.91, P for trend = 0.02) for calcium.
After excluding 502 pregnant women who experienced substantial changes in diet in the previous month, the adjusted OR between extreme quartiles was 0.53 (95% CI 0.32–0.87, P for trend = 0.003) for yogurt and 0.71 (95% CI 0.42–1.20, P for trend = 0.20) for calcium.
The current cross-sectional study found that higher intake levels of yogurt and calcium were independently associated with a lower prevalence of depressive symptoms during pregnancy. No relationships were found between intake of total dairy products, milk, or cheese and depressive symptoms during pregnancy.
A cross-sectional study of 887 elderly Japanese men and women reported a significant inverse relationship between the daily intake of milk products and the prevalence of depressive symptoms, based on the five-item Geriatric Depression Scale. This finding is inconsistent with the current results. In a cohort study of 1609 Taiwanese men and women aged 65 years or older, dairy intake was not associated with the risk of new depressive symptoms, based on the ten-item CES-D, within 4 years. Dairy product intake during pregnancy was not related to the subsequent risk of postpartum depression based on the Edinburgh Postnatal Depression Scale in a prebirth cohort study of 865 Japanese women. These findings are in partial agreement with the current results. Regarding calcium intake, our results are in partial agreement with those of the aforementioned Korean study, in which calcium intake was significantly higher in the group scoring low for depression than in the group scoring high for depression. Another cross-sectional study of 105 Korean women aged 41–57 years showed a significant inverse association between calcium intake and self-rated depression. This finding is also in partial agreement with the current results. Our results are at variance with those of a cross-sectional study of 4734 US adolescents showing no association between calcium intake and depressive symptoms, based on a six-item scale developed by Kandel and Davies. In a trial among 179 female Iranian college students suffering from premenstrual syndrome, the group receiving 500 mg of calcium carbonate twice daily for 3 months experienced significant improvement in early tiredness, appetite changes, and depressive symptoms, compared with the placebo group.
There are methodological limitations to the current study. One of these is that its cross-sectional nature prevents us from drawing conclusions about causality.
Other limitations of the KOMCHS have been described elsewhere.[2, 10-13] With regard to our DHQ, the validity of the DHQ regarding dietary calcium seems reasonable, as described above; however, the consequence of non-differential exposure misclassification would have given rise to an underestimation of the results. The results of a sensitivity analysis that excluded 502 women who experienced substantial changes in diet in the previous month were similar to those in the overall analysis. Regarding the CES-D, items associated with physical symptoms, including fatigue and physical discomfort, are common matters of pregnancy; this symptom overlap is likely to have caused an overestimation of the prevalence of depressive symptoms. Nevertheless, the prevalence of depressive symptoms in this population was lower than that in a representative sample of the Japanese general population (30.7% in 2315 women aged 30–39 years). Because our study subjects participated in the KOMCHS from 5 to 39 weeks of gestation, it is hard to precisely assess the incidence and prevalence of depressive symptoms during pregnancy. The consequence of non-differential outcome misclassification would bias the estimates of the association between exposure and outcome towards the null. The results of a sensitivity analysis confined to 595 pregnant women who completed the baseline survey in the first trimester, as well as those of a sensitivity analysis confined to 1004 pregnant women who completed the baseline survey in the second trimester, were similar to those in the overall analysis.
Selection bias must also be taken into account. In the KOMCHS, 978 pregnant women who lived in Fukuoka Prefecture completed the baseline survey between April 2007 and March 2008. On the other hand, according to the government of Fukuoka Prefecture, the number of childbirths was 46 393 in 2007 and 46 695 in 2008. The participation rate must have been low, and our subjects were probably not representative of Japanese women in the general population. As an example, a population census performed in 2000 in Fukuoka Prefecture showed that the proportions of women aged 30–34 years with unknown, <13, 13–14, ≥15 years of education were 4.8%, 52.0%, 31.5%, 11.8%, respectively. The corresponding figures for the current study were 0.0%, 24.5%, 33.1%, and 42.4%, respectively. Calcium intake in this study population was similar to that in the general population. According to the National Health and Nutrition Survey in Japan, the average daily per capita intake of calcium was 512 mg, whereas the mean daily intake of our study subjects was 503 mg.
Several dietary and non-dietary confounding factors were adjusted for, yet we could not rule out residual confounding effects.
In the present study, the intake of yogurt, but not of total dairy products, milk, or cheese, was significantly inversely related to depressive symptoms during pregnancy. Japanese people who have a fairly high intake of yogurt might follow healthy dietary patterns or behaviours that may be preventive against depressive symptoms.
In the present study, further adjustment for calcium intake completely removed the inverse association between yogurt intake in the highest quartile and depressive symptoms during pregnancy, although the inverse linear trend remained significant. Yogurt intake and calcium intake were likely to be mutually dependently associated with depressive symptoms during pregnancy. The beneficial association between yogurt intake and depressive symptoms during pregnancy may be attributable to some extent to calcium, or to some other unmeasured constituents in relation to calcium. At this time, we have no explanation of the mechanisms underlying the inverse association between calcium intake and depressive symptoms during pregnancy. Torres et al. found that the decrease in extracellular Ca2+ concentration that occurs during excitatory transmission is sensed by neighbouring astrocytes, and that adenosine triphosphate (ATP) released through astrocytic Cx43 hemichannels in turn strengthens inhibitory transmission by activating P2Y1 receptors on interneurons. Higher calcium intake might be expected to moderate extracellular calcium fluctuations and prevent inhibitory transmission, resulting in a lower prevalence of depressive symptoms.
The present cross-sectional study in Japan suggests that higher intake levels of yogurt and calcium are independently associated with a lower prevalence of depressive symptoms during pregnancy. Confirmation of the current results by additional epidemiological studies would further support the public health benefits of yogurt and calcium.
YM and KT were financially supported by Meiji Co. Ltd. The other authors have no conflicts of interest.
YM, KT, and MA were responsible for the study concept and design, and for the acquisition of data. HO and SS contributed to the estimation of dietary factors. YM contributed to the analysis and interpretation of data and to the drafting of the article.
The KOMCHS was approved by the ethics committee of the Faculty of Medicine, Fukuoka University, on 6 February 2007 (no. 319).
This study was supported by JSPS KAKENHI grant numbers 19590606, 20791654, 21590673, 22592355, 22119507, 24390158, 25463275, and 25670305, by Health and Labour Sciences Research Grants for Research on Allergic Disease and Immunology and Health Research on Children, Youth and Families from the Ministry of Health, Labour and Welfare, Japan, by Meiji Co. Ltd., and by the Food Science Institute Foundation. YM and KT received a total of ¥17,500,000 from Meiji Co. Ltd., between 2007 and 2013.
The authors would like to acknowledge the Kyushu Branch of the Japan Allergy Foundation, the Fukuoka Association of Obstetricians & Gynaecologists, the Okinawa Association of Obstetricians & Gynaecologists, the Miyazaki Association of Obstetricians & Gynaecologists, the Oita Association of Obstetricians & Gynaecologists, the Kumamoto Association of Obstetricians & Gynaecologists, the Nagasaki Association of Obstetricians & Gynaecologists, the Kagoshima Association of Obstetricians & Gynaecologists, the Saga Association of Obstetricians & Gynaecologists, the Fukuoka Society of Obstetrics and Gynaecology, the Okinawa Society of Obstetrics and Gynaecology, the Fukuoka City Government, and the Fukuoka City Medical Association for their valuable support.