Study design

We conducted a secondary analysis of the Millennium Cohort Study (MCS) following the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guidelines. The MCS is a nationally representative prospective cohort study of mothers who gave birth between 2000 and 2002 in Great Britain and Northern Ireland. The study data design and collection have been described elsewhere in detail [25]. In summary, the MCS sample was obtained from electoral wards across the UK and was clustered geographically to achieve an adequate representation of the UK population. The baseline (MCS1) data collection occurred when mothers were 9 months postpartum. There have been six subsequent sweeps of data collection at 3 years postpartum (MCS2), 5 years postpartum (MCS3), 7 years postpartum (MCS4), 11 years postpartum (MCS5), 14 years postpartum (MCS6), and 17 years postpartum (MCS7). The MCS recruited 18,492 biological mothers in the baseline survey interviews, of which 11,088 participated in the follow-up at 14 years postpartum (60.8%). The focus of the 17-year postpartum sweep was the child and not the mother so not included in this current study. We excluded mothers if they had multiple births, pre-existing mental disorders, and missing data on mode of birth. Our final sample (10,507) included singleton mothers with complete mode of birth data who provided data on depression/severe anxiety at all relevant time points (Fig 1).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Flowchart of sample determination.

https://doi.org/10.1371/journal.pone.0327129.g001

Exposure and outcomes

Mode of birth was recategorized into (1) spontaneous VB [reference group], (2) assisted VB, (3) induced VB, (4) emergency CS, (5) planned CS, and (6) CS after induction of labor. Mode of birth was considered ‘spontaneous VB’ if mothers had a ‘normal delivery’ or ‘water birth’ and responded ‘no’ to the question ‘was the labor induced or attempted to be induced’. Mothers who had a ‘normal delivery’ or ‘water birth’ but responded ‘yes’ to the question ‘was the labor induced or attempted to be induced’ were considered ‘induced VB‘. Mothers who reported ‘assisted with forceps’, ‘assisted vacuum extraction’, ‘assisted breach’, or ‘other assisted delivery’ were considered ‘assisted VB. Mothers who responded with ‘emergency CS’ or ‘planned CS’ were recorded as such. Mothers who reported having any CS (emergency or planned) but responded ‘yes’ to the question ‘Was the labor induced or attempted to be induced’ were considered ‘CS after induction’. Answers that recorded ‘other’ or ‘irrelevant’ were considered missing. Data on depression/severe anxiety were collected as a single composite variable. Mothers were asked at every time point if they had ever received a diagnosis of depression/severe anxiety from a doctor. The primary outcome was self-reported diagnosis of depression or severe anxiety at any time point up to 14 years postpartum. We considered secondary outcomes as depression/severe anxiety diagnosis at each time point. A short description of co-variates included in the analysis is provided in S2 Appendix.

Statistical analysis

Data analyses were performed using Stata/SE v.16 (Stata Corporation, College Station, TX, U.S.A.). First, we conducted a descriptive analysis of the participant’s characteristics according to the mode of birth and assessed the distribution of the mode of birth using Chi-square tests. Following this, we estimated the association between mode of birth and cumulative depression/severe anxiety up to 14 years postpartum by presenting crude and adjusted odds ratios (ORs) with 95% confidence intervals (CIs) using multivariable logistic regression. We performed four logistic regression models. The first model presented crude unadjusted estimates. In the second model, we adjusted for sociodemographic factors (maternal age, ethnicity, and maternal BMI). In the third model, we adjusted for additional factors (area-level deprivation, maternal education, HDP, presence of a longstanding illness in mothers, and parity). The final model presented fully adjusted estimates controlling for factors in the second and third models.

We undertook a range of prespecified subgroup analyses to explore key findings. We stratified analysis by the presence of postnatal psychological distress at 9 months, and NICU admission. We considered only primiparous mothers to eliminate the possibility of potential confounding effects of prior pregnancy or previous mode of birth. We stratified mothers of children born preterm (born before 37^th weeks gestation). We stratified mothers by longstanding illnesses as pre-existing morbidity may increase the likelihood of a cesarean birth. In addition, we performed the following sensitivity analyses; First, we categorized the mode of birth into “any cesarean birth” and “any induction” to investigate if different modes of intervention influenced the results. Second, we repeated analyses for our primary outcome for mothers with complete data on mode of birth, depression/severe anxiety, and co-variates. Third, we repeated the main analysis for mothers who were followed up at 14 years postpartum only. For missing data, we coded relevant items as ‘missing’ instead of fitting the model to observations with complete data or using multiple imputations. Overall, statistical significance level was set at p < 0.05.

Ethics statement

The Centre for Longitudinal Studies obtained ethical approval from the South West Multi-Centre Research Ethics Committee (MREC), London MREC, Yorkshire MREC, and Yorkshire and The Humber – Leeds East MREC at each survey; MCS1: South West MREC (MREC/01/6/19); MCS2 and MCS3: London MREC (MREC/03/2/022, 05/MRE02/46); MCS4: Yorkshire MREC (07/MRE03/32); MCS5: Yorkshire and The Humber-Leeds East (11/YH/0203); MCS6: London MREC (13/LO/1786), MCS7: REC North East–York (REC ref: 17/NE/0341). The MCS obtained informed written consent from parent/ carers of all cohort children prior to participating in the study. For this present study the MCS has been fully anonymised and did not require additional ethics approval or consent for further analysis.

Comparison with previous literature

Our main findings are consistent with two studies that demonstrated no significant association between mode of birth and depression or anxiety [19,21]. A Taiwanese study found no association between modes of birth and depression or anxiety after adjusting for covariates, although the authors did not differentiate between emergency and planned CS [19]. Similarly, in a recent longitudinal study conducted among Danish mothers, there was no difference in depressive symptoms among mothers who only had spontaneous VB compared to mothers who had only CS 11.2 years after their last birth [21]. Other studies have observed associations [12,20]. In one study, mothers reported a higher risk of later treatment for depression/anxiety following elective CS compared to vaginal delivery at 3 years postpartum [20]. Another study suggested higher depression, anxiety disorders, and somatization symptoms among women who had instrumental births compared to women who had vaginal delivery [12]. Our study extends previous studies suggesting that unplanned delivery modes may be linked with developing depression [26,27]. A potential explanation underlying these findings is that unplanned delivery modes may be associated with a negative birth experience which could increase a woman’s risk of developing mental health problems [27]. Two studies suggested that unplanned CS may be associated with a loss of control, unmatched expectations, and trauma which may predispose a woman to developing a later mental health condition [26,28]. Other studies revealed that these modes of birth are linked with unexpected complications during childbirth with increased fear of future pregnancies and a possible elevated risk of depression in the longer term [29,30]. In the UK, unplanned birth modes may be performed with medical indications and risk factors, impacting maternal and fetal outcomes [31]. However, some risk factors for unplanned birth mode may also be medical interventions required for the mother or child’s health.

Among women who had postnatal psychological distress at 9 months, we observed an increased odds of depression/severe anxiety by 14 years postpartum. The finding is perhaps unsurprising, given the established link between postpartum mental disorders and subsequent mental health disorders diagnosis [32–34]. When the severity of postpartum mental disorders is considered, postnatal psychological distress may be linked with developing later mental health disorders. According to a study by Abdollahi et al, women who experienced postnatal depression were twice as likely to have depression 4 years after childbirth [33]. The findings in our study of an association between mode of birth and depression/severe anxiety among mothers with longstanding chronic illnesses is novel. A study by Brown et al reported that more women with chronic physical conditions experienced perinatal mental illnesses [35].

Levels of steroid hormones including progesterone, cortisol oestradiol are suggested to play a role in the development of postpartum mental disorders [36,37]. One study hypothesized that hormones such as oxytocin are deficient and dysregulated from normal labor physiology during cesarean births [38]. A disruption of the hormonal environment that supports maternal bonding after childbirth may increase the childbirth stressor and risk of mental health issues. Psychologically, it has been hypothesized that unplanned interventions during childbirth may increase the risk for probable childbirth-related posttraumatic stress disorders (PTSD) due to its links with sense of fear and loss of control [39]. These experiences have been shown to increase the risk of persistent depression or anxiety symptoms induced by the childbirth experience even years after delivery [40,41]_. However, concensus on the biological and psychological mechanisms that links mode of birth to adverse maternal mental health remain limited and warrants further examination in future research.

Strengths and limitations

The results of this study should be viewed in the context of some limitations. First, depression/severe anxiety diagnosis relies on self-reported data and may be subject to recall bias [42]. Second, although comorbid, depression and anxiety are distinct mental disorders. We used a combined diagnosis of depression and severe anxiety. As such, it was impossible to conduct separate analysis for depression and anxiety. Therefore, we were not able to examine the association between mode of birth and depression and anxiety separately. Third, we had data on depression/severe anxiety diagnosis, we had no information on the criteria for the severity of the diagnosis. Therefore, we could not evaluate the overall risks and/or benefits of the mode of birth. Fourth, despite adjusting for potential confounders, the role of residual confounding from unmeasured confounding variables such as previous mode of birth and fear of childbirth cannot be precluded. The absence of data on fear of childbirth may have resulted in biased estimates of the association between mode of birth and long-term maternal depression/severe anxiety. As such, some observed associations may be as a result of underlying psychological vulnerability rather than the mode of birth [43,29]. Lastly, selection bias may have been introduced due to loss of follow-up. Our sample included 18,246 mothers at baseline which reduced to 11,088 at 14 years postpartum (of which 10507 were included in the analysis). As participation can be influenced by several factors, mothers with severe mental illnesses may be more likely to be lost to follow up which may have skewed our findings. However, using the cumulative outcome approach to estimate depression/severe anxiety up to 14 years postpartum increased efficiency regardless of the timing of the diagnosis. These study limitations underscore the need for more rigorous investigations in primary studies and suggest the need for further research to characterize specific mental disorders in relation to the mode of birth.

This study also had several strengths. First, this study used a comprehensive, nationally representative large cohort of women who gave birth in the UK between 2000 and 2002. Second, our adjusted estimates included longstanding illnesses, complications during pregnancy or childbirth, and sociodemographic variables which had not been accounted for in previous studies. Third, our analysis excluded mothers with pre-delivery mental health diagnoses as the presence of a long-term mental health diagnosis may be confounded by previous mental disorders exposures. Fourth, the data allowed us to conduct stratified analyses for explanatory factors that could potentially modify the association such as postnatal psychological distress and longstanding illnesses. Lastly, MCS maternal self-reported mode of birth was reported to be highly reliable with an agreement of 94–98% compared to hospital records [44].