Mode of delivery and child and adolescent psychological well-being: Evidence from Hong Kong’s “Children of 1997” birth cohort

Mode of delivery (vaginal or cesarean section) is thought to affect gut microbiota, which in turn may affect psychological well-being. As such, mode of delivery is potentially a modifiable factor for psychological well-being. Here we examined the association of mode of delivery with child and adolescent psychological well-being. We used multivariable linear regression in a population-representative Hong Kong Chinese birth cohort, “Children of 1997,” to examine the adjusted associations of mode of delivery with behavioral problems assessed from parent-reported Rutter score at ~7 (n = 6294) and ~11 years (n = 5598), self-esteem assessed from self-reported Culture-Free Self-Esteem Inventory score at ~11 years (n = 6937) and depressive symptoms assessed from self-reported Patient Health Questionnaire-9 score at ~13 years (n = 5797). Cesarean Section (CS) was associated with children born in private hospitals, boys, and firstborns, higher maternal body mass index, higher maternal age, preeclampsia, higher socioeconomic position (SEP) and maternal birth in Hong Kong. CS was unrelated to behavior, self-esteem and depressive symptoms adjusted for infant characteristics (sex, gestational age, birthweight, parity and breast feeding), maternal characteristics (mother’s age and place of birth) and SEP. In a developed non-Western setting, mode of delivery was not clearly associated with childhood or early adolescent psychological well-being.

also implying that findings were confounded 18 . Curran et al. (2015) reported null associations of mode of delivery with ADHD and ASD in the Millennium Cohort in the United Kingdom 19 . However, the association of mode of delivery with other indicators of psychological well-being, as self-esteem during childhood and adolescent depressive symptoms during early adolescence has not, to our knowledge, previously been examined. CS rates tend to be higher than recommended by the World Health Organization (WHO) and thus represent a modifiable intervention target. To clarify the role of mode of delivery in psychological well-being, we examined prospectively in a large population-representative birth cohort, Hong Kong's "Children of 1997, " the association of mode of delivery with child and adolescent psychological well-being, including measures of behavior, self-esteem, and depressive symptoms.
About a quarter of the births were CSs (26.8%), 56.6% were unassisted vaginal birth and 16.6% assisted vaginal birth (Table 1). Private hospitals had higher rates of CS (48%) than public hospitals (18%). CS was positively associated with male sex, being firstborn, no breastfeeding, higher maternal BMI, higher maternal age, preeclampsia, higher SEP (parental education, occupation and household income at birth) and a Hong Kong-born mother. Table 2 shows that mode of delivery was not associated with Rutter score or its subscores in any model. Mode of delivery was not associated with overall self-esteem, or most of its subscores, but was associated with lower social self-esteem. Mode of delivery was not associated with depressive symptoms in any model. Table 3 shows that results were generally similar considering vaginal birth as unassisted and assisted. In a fully adjusted model (Model 3), CS compared to unassisted or assisted vaginal delivery was associated with lower social self-esteem. Results from a complete case analysis were similar (see Supplementary Tables S1 and S2).

Discussion
In this large, population-representative birth cohort from the developed non-Western setting of Hong Kong, consistent with one of two previous studies 16 , mode of delivery was not associated with child or adolescent behavior. Our study adds by showing CS was also not associated with depressive symptoms or overall self-esteem in adolescence, although CS was perhaps associated with lower social self-esteem.
"Children of 1997" is a large, population-representative birth cohort from a unique setting; nevertheless, our study has limitations. First, type of delivery was obtained by parent-report (mostly mothers), as opposed to clinical records, which could introduce measurement error. However, type of delivery was obtained shortly after delivery and is quite memorable. Second, screening tools were used to assess behavioral problems, self-esteem, and depressive symptoms as opposed to clinical diagnosis. However, these questionnaires have been validated in Chinese populations [24][25][26] , and our large sample size compensates for any lack of precision. Third, difficult and stressful births can lead to hypoxia and asphyxia resulting in neonatal encephalopathy and long-term neurodevelopmental impairments 27 . We do not know whether this occurred for any specific type of delivery. Fourth, we have no information regarding analgesic use during the prenatal and perinatal periods. Analgesic exposure during the fetal or neonatal period has been found to be associated with changes to the brain in animal studies 28,29 . However, no association has been found between obstetric anesthetic techniques and behavior or long-term neurodevelopmental outcomes in humans 30 . Fifth, we did not have data differentiating whether CS was elective or emergency. Infants born via emergency CS are typically exposed to the mother's vaginal flora, as compared to elective CS, where infants are not 20 . A recent study reported that emergency CS, compared to spontaneous vaginal delivery, was associated with delayed gross motor function at 9 months and null associations across various modes of delivery with cognitive and motor development at 3 years 20 . However, there are no studies examining the longer term effects of emergency CS with psychological well-being. Finally, given the large number of associations examined, any positive associations might be false positives due to multiple comparisons and need to be confirmed in future studies.  Our study is not consistent with a link between gut microbiota, established early in life, and psychological well-being. In adults, probiotics, of the Lactobacillus or Bifidobacterium genera 31-33 , may have a role in the treatment of depression, as seen in some preliminary randomized controlled trials 31,32 . Differences may have occurred for a number of reasons. Mode of delivery may affect proportions of bacteria from the Firmicutes and Bacteroidetes families, but may only be associated with short-term differences in gut microbiota, rather than representing differences that continue into adolescence, although early life is a key phase for neurological development. Alternatively, any effects of events at birth may operate via different mechanisms, such as cortisol from a stress response [34][35][36] leading to child behavioral problems. For example, Li et al. (2011) found elective cesarean delivery associated with fewer behavioral and emotional problems compared to unassisted vaginal delivery while assisted vaginal delivery was associated with more problems 17 . However, we found no associations of assisted vaginal delivery with behavioral problems. Furthermore, Li et al. (2011) did not examine the differences between the implications for CS, thereby not allowing further examination of how cortisol may mediate the pathway between mode of delivery and psychological well-being.
Mode of delivery does not appear to be associated with the psychological well-being of children and adolescents. However, a clearer understanding of the long-term effects of mode of delivery on gut microbiota will help elucidate the impact of CS on health, delineate any possible pathways involving the gut-brain-axis and identify where modification of gut microbiota in infants could potentially be beneficial.

Methods
Ethics Statement. Since our participants are children, informed consent was obtained from the parents, next of kin, caretakers or guardians (informants) on behalf of the participants by completing the questionnaire at enrollment as approved by The University of Hong Kong Medical Faculty Ethics Committee. Ethical approval for further studies was obtained from the University of Hong Kong-Hospital Authority Hong Kong West Cluster, Joint Institutional Review Board and/or the Ethics Committee of the Department of Health, Government of the Hong Kong SAR as appropriate. All methods were performed in accordance with the relevant guidelines and regulations.
Participants. The Hong Kong "Children of 1997" birth cohort is a population representative Chinese birth cohort (n = 8,327) that covered 88.0% of all births in Hong Kong from April 1, 1997 to May 31, 1997, and has been described in detail elsewhere 37 . The study was initially established to investigate the effects of secondhand smoke exposure on infant health. Families were recruited at the first postnatal visit to one of the 49 Maternal and Child Health Centers (MCHCs) in Hong Kong, which parents of all newborns were encouraged to attend. Characteristics obtained using a self-administered questionnaire in Chinese at recruitment and subsequent routine visits included maternal characteristics (mother's birthplace and age at birth), infant characteristics (type of delivery, type of birth hospital, sex, gestational age, birth weight, parity and breastfeeding) and socioeconomic position (SEP).  Passive follow-up via record linkage was instituted in 2005 to obtain routinely collected information including bi-annual assessments of emotional and behavioral problems using the Revised Parent's Rutter Scales in Chinese (grade 2, 4 and 6 (i.e., ages 7-8, 9-10 and 11-12 years)) and self-esteem using Form A of the Culture-Free Self-Esteem Inventories for Children by Battle in Chinese (grade 4 (age 9-10 years) onwards) from the Student Health Service, Department of Health, which provides free annual check-ups for all school students. Exposures. Mode of delivery was categorized as vaginal delivery or CS. Mode of delivery was obtained from the baseline survey completed by the primary caregiver (typically the mother). Mode of delivery was recorded as normal vaginal birth without instrumentation, vaginal birth with suction or forceps or CS. We considered vaginal births together (including unassisted and assisted vaginal delivery). Given that unassisted and assisted vaginal deliveries would be expected to have the same effect on gut microbiota, we also classified mode of delivery as unassisted vaginal delivery, assisted vaginal delivery or CS to check whether any associations were specific to CS. We did not collect information on the possible indications for CS or whether the CS was primary versus repeat.

Outcomes. Emotional and behavioral problems. Emotional and behavioral problems at ~7 years (6-<9
years) and ~11 years (9-<13 years) were assessed from the Revised Parent's Rutter Scales in Chinese for parents 25,38 . The scales consist of a set of 31 items describing emotional and behavioral difficulties, with each item scored 0 for does not apply, 1 for applies somewhat or 2 for certainly applies. A total score (range 0-62) and subscores for conduct problems (5 items, range 0-10), emotional problems (5 items, range 0-10) and inattention/ hyperactivity (3 items, range 0-6) were calculated, where a higher score indicated more emotional and behavioral problems. When considering the presence of emotional and behavioral problems, a total score of 13 or more was considered as overall emotional and behavioral problems 39 . Presence of conduct problems, emotional problems or hyperactivity was defined as having the corresponding subscore above the sex-specific 97th percentile (≥5 for boys and ≥4 for girls for conduct, ≥4 for boys and ≥4 for girls for emotion, 6 for boys and ≥5 for girls for hyperactivity) for this population.  Self-esteem. Self-esteem at ~11 years (9-<13 years) was assessed from the Form A of the Culture-Free Self Esteem Inventories (SEI) in Chinese for children 40 . Assessment inventories developed in Western countries may be less valid in non-Western settings. However, the items of the SEI were chosen with concern for content that is least sensitive to change across cultures 41 . The SEI has good validation in assessing self-esteem in adolescents and has been used in Hong Kong Chinese 42 . Responses are of the forced-choice variety that the respondent checks either "yes" or "no" for each item. A total score (range 0-50) and subscores for general self-esteem (perception of self-worth in general) (20 item, range 0-20), social self-esteem (perception of quality of relationships with peer) (10 items, range 0-10), academic self-esteem (perception of ability to academic success) (10 items, range 0-10), and parent-related self-esteem (perception of status at home) (10 items, range 0-10) were calculated, where a lower score indicated lower self-esteem. Based on the findings from the same study, we excluded unreliable scores with a lie scale of 2 or less (n = 158), although a lie scale of 5 or more had been used to indicate a lack of defensiveness and a reliable self-esteem inventories 40 . Presence of low self-esteem was defined as having a total score of 19 or less 40 . Each subscale score can be classified into one of five categories ('very low, ' 'low, ' 'intermediate, ' 'high' or 'very high'). A 'very low' subscale score indicated low self-esteem in the respective subscale category (≤7 for general self-esteem and ≤2 for social, academic and parent-related self-esteem) 40 .
Depressive symptoms. Depressive symptoms at ~13 years (12-≤15 years) were assessed from the PHQ-9 43 in Chinese. The PHQ-9 scale has good sensitivity and specificity in detecting depression in youth 44 , is an effective screening tool for the risk of depression among adolescents 45 and has been validated in Chinese adolescents 46 . The scales consist of a set of 9 items describing symptoms and functional impairment, with each item scored 0 for not at all, 1 for several days, 2 for more than half the days or 3 for nearly everyday. A total PHQ-9 score (range 0-27) was calculated, where a higher score indicated more depressive symptoms. Presence of PHQ-9 Major Depressive Disorder was defined as having a total score of 11 or more mapping on the DSM-IV-TR 44 .
Statistical analysis. We used multivariable linear regression to estimate the adjusted associations of mode of delivery with Rutter, SEI and depressive symptom scores from which estimated beta-coefficients with 95% confidence intervals (CIs) are presented. We adjusted for potential confounders, which are presented in three models. Model 1 adjusted for sex, age at assessment and survey mode for depressive symptoms since for PHQ-9 was obtained by postal questionnaire, telephone interview and in-person interview. Model 2 additionally adjusted for SEP (highest parental education at birth, highest parental occupation at birth and monthly household income per head at birth), mother's birthplace and birth hospital. Model 3 additionally adjusted for birth weight, gestational age, parity, maternal age, maternal body mass index (BMI), gestational diabetes and preeclampsia. Among the 7,681 birth cohort members included, 3.6% were missing mode of delivery. Overall, 13.2% of data points were missing, mainly maternal BMI (58.8% missing) and gestational diabetes (23.6%) because these were obtained from Survey I or Survey III. We used multiple imputation to predict missing values of exposures and confounders incorporating data on the primary outcomes (Rutter, SEI and PHQ-9 scores), exposures (mode of delivery), confounders and other factors potentially associated with mode of delivery based on a flexible additive regression model with predictive mean matching 47 as multiple imputation is the gold standard in this situation 48 . We used the 'Hmisc' package in R 3.0.1 (R Development Care Team, Vienna, Austria) for the multiple imputation. We summarized the results from 20 imputed datasets into single estimated beta-coefficients with CIs adjusted for missing data uncertainty. The results were based off 20 imputations due to fraction of missing information (γ) and the preventable power falloff of less than 1%. We also performed a complete case analysis for comparison using Stata version 10 (Stata Corp, College station, Texas). Data Availability. Data are available upon request from the "Children of 1997" data access committee: apr-may97@hku.hk.