This study was conducted at a university hospital located in an urban area in Japan. Mothers were surveyed twice, one during antepartum medical checkups (from 20 weeks of gestation to 36 weeks of gestation) and one during postpartum medical checkups (1month after birth). The participants were enrolled in this study based on the following criteria 1) age over 20 years, 2) without present or history of psychiatric diseases, and 3) permission of attending doctor. Out of 215 women who gave written informed consent to participate, 162 women (75.3%) completed the study. As the reasons of dropping out of the study were the NICU hospitalization of neonates and the changes in the birth institution or home address. Finally, we analyzed 134 participants who completely finished the study in both of antepartum and postpartum period. This study was approved by the Ethics Committee of the University of Tokyo.

This study was a longitudinal study which investigates the mental health and the bonding to the fetus or neonate of mothers in both of antepartum and postpartum. The mental state of mothers in this study means that it was assessed by EPDS and MIBS.

1) Characteristics of Participants

Information on characteristics of participants was collected from medical records in antepartum and postpartum. Information collected in the antepartum consisted of occupational status, marital status, parity, morning sickness, worries about costs. Information collected in the postpartum consisted of gestational week at the delivery, mode of delivery, obstetrical disorders during labor, baby’s weight and sex, APGAR score, umbilical blood pH, breast feeding and troubles with breast feeding.

2) Family Function

We examined Family APGAR score two times to evaluate the degree of the social support. The Family APGAR was introduced by Smilkstein in 1978 to assess satisfaction with social support from the family in adulthood, which was composed of five items measuring perceived family support in the domains of adaptation, partnership, growth, affection, and resolve. It was translated into Japanese by Nagamine [8] . This is scored on a three-point Likert scale (0 - 2), and the total score range from 0 - 10 points. Score of less than 7 points indicate family function deficiency. The reliability and the validity of this score were tested [9] [10] .

3) Measurement of Mental Health

a) EPDS

EPDS is a self-reported questionnaire composed of 10 items scored on a four-point Likert scale (0 - 3) designed to assess antepartum and postpartum depression [11] . It does not include items related to physical symptoms because it was thought up not to be affected by the physical symptoms in the postpartum. Score of 9 points or more are considered to be suspicious in depression after delivery as Japanese definition [12] . In this study EPDS scoring are classified into two group; 9 points or more and less than 8 points.

b) MIBS

MIBS was based on a study by Kumar and Marks [13] , and it was translated into Japanese by Yamashita [14] . MIBS is a self-reported questionnaire composed of 10 items on a four-point Likert scale. A high MIBS score indicates worse of mother’s bonding forward to baby. If the score of MIBS is zero, mothers do not have any bonding disorders [13] . MIBS evaluates two factors, the first one is “lack of affection” and the other one is “anger and rejection”. “Anger and rejection” is used as a direct risk factor for child abuse [15] .

4) Statistical analyses

SPSS Statistics 23 for Windows was used to analyze the descriptive statistics. Antepartum and postpartum scores were tested using the paired t-test and Fisher’s exact test. The dependent variables were postpartum EPDS score and Bonding score, and the independent variables were each score of antepartum and participants characteristics and Family APGAR score. The level of statistical significance was set to P < 0.05.

Table 1 shows the participants characteristics. The mean age was 34 ± 4.2 years (range 24 to 43 years). Among the participants, 58.9% were employed. The percentages of

*Delivery and baby care costs.

primipara and multipara were 58.9% and 38.8% respectively. Seventy-nine percent of the participants had morning sickness, and 11% were worries about delivery and baby care costs. The mean gestational week of pregnancy was 40 ± 2.3 weeks. The rates of vaginal delivery and caesarean section were 93.2%, and 6% respectively. Among the participants, 71.6% had an abnormality, for example, labor induction during the delivery process.

The mean weight of the newborns was 6.7 ± 0.9 lb. Concerning on breastfeeding, breastfeeding only, combination of breastfeeding and bottle-feeding, were 47.7%, 40.2%, respectively. Trouble with breastfeeding was found in 25.3% of the participants had.

Family APGAR score which measured the family function in the antepartum period was 8.9 ± 1.5 points, while that in the postpartum was 8.7 ± 1.6 points (Table 1).

1) Depressive state: An antepartum depressive state (EPDS score more than 9 point) was 14.1% (19/134), while postpartum depressive state was 17.9% (24/134). The mean EPDS score in the antepartum and postpartum periods were 4 ± 3.8 points and 4.9 ± 3.8 points, respectively. There were significant differences in depressive points assessed by EPDS between the antepartum and postpartum periods (p = 0.015) (Table 2). Mothers with an antepartum depressive state were 14.7 times more likely to have a postpartum depressive state than those without an antepartum depressive state (Table 3). The factors which influenced the depressive state were “worries about costs” (p = 0.00), “postpartum family function” (p = 0.04) in the antepartum period, “worries about costs” (p = 0.02) and “postpartum family function” (p = 0.04) in the postpartum period (Table 4).

2) Bonding disorder: We judged not antepartum bonding disorder was 41.7% (56/ 134), and not postpartum bonding disorder was 28.3% (38/134). The mean ± standard deviation MIBS score was 2.4 ± 2.4 points in the antepartum period and 1.5 ± 2 points in the postpartum period. There were significant differences in bonding points assessed by MIBS between the antepartum and postpartum periods (p = 0.00) (Table 2). Mothers with an antepartum bonding disorder were 5.7 times more likely to have a postpartum bonding disorder than those without antepartum bonding disorder (Table 5). The mother who had “lack of affection” of the one factor in antepartum was 55.9% (75/134) and postpartum was 39.5% (53/134). And, the mother who had “anger and rejection” of the other factor in antepartum was 52.9% (71/134) and postpartum was 46.2% (62/134). We compared the score of postpartum with antepartum with a paired t-test. As a result,

Paired t-test.

*Fisher’s exact test p = 0.00.

*Delivery and baby care cost.

“lack of affection” significantly decreased from antepartum to postpartum (p = 0.00). In “anger and rejection” which were no significant differences between antepartum and postpartum (p = 0.3) (Table 2). The factors which influenced the bonding disorder were “morning sickness” (p = 0.03), “worries about costs” (p = 0.01) and “antepartum family function” (p = 0.00) in the antepartum period, and “antepartum family function” (p = 0.003), “breast feeding” (p = 0.04), and “postpartum family function” (p = 0.00) in the postpartum period (Table 4).

We examined the relationship between bonding disorder in antepartum and relation of the depressive state in postpartum. As a result, 23 of 96 mothers who had bonding disorder in antepartum had depressed mental state in postpartum. Mothers with an antepartum bonding disorder were 11.7 times more likely to have a postpartum depressive state than those without antepartum bonding disorder (Table 6).

*Fisher’s exact test p = 0.000.

*Fisher’s exact test p = 0.001.