Association between mode of delivery and postpartum depression: A systematic review and network meta-analysis

Abstract

Background:

Postpartum depression is one of the most common postpartum diseases, which has an important impact on the interaction between mother, infant, partner and family, as well as the long-term emotional and cognitive development of infants. However, there are still great disagreements on whether the delivery mode will affect the risk of postpartum depression. The purpose of this study is to explore whether the mode of delivery will affect the risk of postpartum depression through the comprehensive network meta-analysis of elective cesarean section, emergency cesarean section, instrumental vaginal delivery and spontaneous vaginal delivery.

Methods:

We searched in three electronic databases: PubMed, EMBASE and Cochrane Library.

Results:

This paper included 43 studies with a total sample size of 1,827,456 participants. Direct meta-analysis showed that the odds ratio of postpartum depression risk was 1.33 (95% confidence interval = [1.21, 1.46]) between cesarean section and vaginal delivery. The odds ratios of high Edinburgh Postpartum Depression Scale score between cesarean section and vaginal delivery in the three postpartum periods (within 2 weeks, within half a year and over half a year) were basically the same. There was no difference between cesarean section and vaginal delivery in the risk of severe postpartum depression at the Edinburgh Postpartum Depression Scale cut-off point ⩾13 (odds ratio = 1.07; 95% confidence interval = [0.99, 1.16]). Network meta-analysis showed that the risk of postpartum depression in the pairwise comparisons emergency cesarean section vs spontaneous vaginal delivery and elective cesarean section vs spontaneous vaginal delivery was odds ratio = 1.53 (95% confidence interval = [1.22, 1.91]) and 1.47 (95% confidence interval = [1.16, 1.86]).

Conclusion:

The mode of delivery has a significant effect on the occurrence of mild postpartum depression. Women who give birth by cesarean section, especially who give birth by emergency cesarean section, are at a higher risk of mild postpartum depression. We should carefully monitor the progress of postpartum mental disorders in women who delivered by cesarean section and make it possible for women to have a quick access to mental healthcare.

Keywords

Postpartum depression cesarean section vaginal delivery EPDS network meta-analysis

Introduction

Postpartum depression (PPD) is one of the most common postpartum diseases. About 10–15% of women have PPD (O’Hara and Swain, 2009). Although the birth of a baby is a pleasant and exciting thing, PPD can be the beginning of a nightmare for women who have experienced the disease (Buultjens and Liamputtong, 2007). Typical symptoms of PPD include depression, anxiety, lack of happiness, appetite and sleep disturbances, physical agitation, fatigue, feeling worthless and excessively guilty, decreased concentration and constantly thinking about death or suicide (Chandrasekaran et al., 2018). PPD is a health issue that has a significant impact on the mother, family, partners, mother–infant interaction and the long-term emotional and cognitive development of the baby. Henderson et al. (2003) suggest in report that PPD may have significant negative effects on mother-to-child interaction, child growth and breastfeeding. In addition, the children of the mothers with PPD tend to form insecure attachments with peers and develop social difficulties. PPD differs from other forms of depression in that it occurs at a specific time in life characterized by new childbearing events that require physical and psychological adaptation (Meky et al., 2020). Women experience adaptive physical and psychological changes during pregnancy, childbirth and the puerperium, making them vulnerable to mental illness (Zanardo et al., 2018). In addition to changes in the reproductive system and the physiological changes associated with pregnancy, women also experience psychosocial changes after childbirth, transition to parenthood, re-establish relationships and strive to meet the physical and emotional needs of newborns and other family members (Goker et al., 2012).

PPD occurs because of a combination of biological and psychosocial factors, genetic factors and important life experiences. However, there is still a considerable disagreement as to whether delivery patterns affect the risk of PPD, and many studies have not found an independent relationship between them (Carter et al., 2006; Eisenach et al., 2008; Hannah et al., 2002; Nielsen Forman et al., 2000; Patel et al., 2005; Sword et al., 2011). Studies have even shown that cesarean section (CS) is a protective factor for PPD in some cases (Chaaya et al., 2002). Some studies, however, suggest that the delivery mode may be a risk factor for PPD. According to the latest research of Zanardo et al. (2018), the scores of women who gave birth by CS are significantly higher than women who gave birth through normal vaginal delivery (VD) on the Edinburgh Postpartum Depression Scale (EPDS), especially with regard to anxiety and depression symptoms.

Based on the above introduction, this study aims to conduct a comprehensive network meta-analysis of elective cesarean section (ELCS), emergency cesarean section (EMCS), instrumental vaginal delivery (IVD) and spontaneous vaginal delivery (SVD) to explore whether the delivery mode affects the risk of PPD.

Methods

Protocol and guidance

This systematic review and network meta-analysis followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines (Liberati et al., 2009).

Eligibility criteria

In this paper, cohort and cross-sectional studies were included. All studies had to cover one or more of the following four delivery modes: (1) SVD, (2) IVD (vacuum- or forceps-assisted delivery), (3) EMCS and (4) ELCS. ELCS were those with prior planning, and EMCS were those without prior planning or that took place after labor started (Patel et al., 2005). The participants who delivered by SVD were assigned to a control group. The participants who met the eligibility criteria were all age-appropriate mothers without limited parity, and their psychological status was assessed after childbirth. The outcome of the included studies was the prevalence of high EPDS score or the odds ratio (OR) of delivery methods as a risk factor for PPD.

Information sources and search strategy

We searched articles published before 20 January 2020, regardless of language or data, in three electronic databases: PubMed, EMBASE and Cochrane Library. The articles were selected by manual screening. The search strategy was: search ((Postpartum depression[Title]) OR (post-partum depression[Title]) OR (post-natal depression[Title]) OR (Postnatal depression[Title]) OR (puerperal depression[Title]) OR (postnatal psychiatric[Title]) OR (Depression, Postpartum[MeSH Terms])) AND ((Cesarean[Title/Abstract]) OR (Caesarean[Title/Abstract]) OR (Cesarean Section[MeSH Terms]) OR (forceps[Title/Abstract]) OR (vacuum extraction[Title/Abstract]) OR (assisted vaginal[Title/Abstract]) OR (Operative delivery[Title/Abstract]) OR (delivery mode[Title/Abstract]) OR (Delivery method [Title/Abstract]) OR (mode of delivery[Title/Abstract]) OR (Type of delivery[Title/Abstract]) OR (method of birth[Title/Abstract]) OR (Mode of birth[Title/Abstract])).

Study selection

The two methodologically trained independent reviewers screened titles and abstracts to determine whether they met the eligibility criteria. The reviewers read the full text and extracted relevant data after a consensus was reached. Any differences were resolved through discussion and arbitration, if necessary, by a third reviewer. The reasons for inclusion or exclusion were recorded in detail. Case reports, letters and minutes of meetings were excluded. The PRISMA flow diagram was used to summarize the study selection processes.

Data extraction

The two investigators used a predefined data extraction sheet to independently extract data from each included study, such as authors, publication year, study design, sample size, grouping and number of patients in the group, cross-sectional data, including counts and effect estimates (OR), country, depression assessment time point, checklist used for depression assessment, title and conclusion. The third investigator independently reviewed the data to ensure accuracy. If no data in digital format were available, we used the free software Plot Digitizer to estimate data from the graphs.

Definition of outcomes

The outcome of the network meta-analysis is the OR of high EPDS score (defined as EPDS score ⩾ 9) in pairwise comparisons between any two of the four delivery modes. The outcome of the direct meta-analysis is the OR of high EPDS score between the cesarean delivery and the VD mode.

EPDS was used to evaluate PPD in this study. EPDS was designed and developed by the University of Edinburgh in Scotland in 1987. It is a simple and comprehensive scale for assessing PPD (Cox et al., 1987). The EPDS score from ⩾9 to ⩽12 is defined as mild PPD in this study.

Statistical analysis

The direct meta-analysis was used to compare the risk of high EPDS score between CS and VD. The generic inverse variance was used to combine the OR of high EPDS score. The heterogeneity was assessed using Q test and I² statistic. The network meta-analysis was used to evaluate the comparative effect by combining direct and indirect evidence of the four delivery modes. First, to visualize network geometry and node connectivity, we summarized the geometry of the evidence network using network plots. We conducted a network meta-analysis of the relationship between delivery modes and PPD using a multivariate random-effects (restricted maximum likelihood estimation) meta-analysis model. To obtain treatment hierarchies, we used a parametric bootstrap procedure with 5000 resamples to compute ranking probabilities. Mean rankings as well as surface under the cumulative ranking curve (SUCRA) values were computed for each delivery mode. We checked the consistency of the network using local and global inconsistency tests. All analyses were conducted in Stata/SE, version 14.

Assessment of risk of bias in individual studies

The quality of studies was independently assessed by two authors. Any disagreement was resolved by a senior author. We applied the cross-sectional study evaluation criteria of AHRQ (the agency for healthcare research and quality) to evaluate cross-sectional studies, including 11 items, with ‘yes’, ‘no’ and ‘unclear’ answers, respectively.

Assessment of small study effects

To evaluate the presence of small study effects, we visually inspected comparison-adjusted funnel plots for each outcome. We produced funnel plots not only for direct comparison between CS and VD but also for indirect comparison between four delivery modes.

Subgroup analysis

Since the timing of PPD assessment may affect the analysis results, the women were divided into three subgroups: within 2 weeks after giving birth, within 6 months after giving birth and more than 6 months after giving birth. Moreover, since a threshold score of EPDS for the diagnosis of PPD varies between different studies, we formed a subgroup with an EPDS score ⩾13 for more detailed evaluation. Finally, considering that PPD is affected by a variety of social and clinical factors, and the crude OR may be more heterogeneous, we conducted a pooled study to provide an adjusted OR of high EPDS score between CS and VD for a more comprehensive analysis.

Results

Study selection

The initial search of three databases yielded 1072 articles. We obtained 131 articles after reading the title and abstract, excluding duplicates and irrelevant articles. After screening the full texts manually, 88 articles were excluded for reasons including review article (n = 11), not relevant to delivery mode (n = 36), not relevant to PPD (n = 6) and not available data (n = 35). Eventually, 43 studies were included in this network meta-analysis (Figure 1).

Figure 1.

Flowchart of study selection.

Study characteristics

In this network meta-analysis, 43 studies were included, comprising a total sample size of 1,827,456 participants. The 43 studies included 16 cross-sectional studies and 27 cohort studies. Among them 13 studies involved SVD, 11 studies concerned IVD, 15 studies reported EMCS and 14 studies discussed ELCS, all of them were subject to network meta-analysis. A total of 36 studies were subject to direct meta-analysis of VD and CS. The summary data of each included study are shown in Table 1, and the network plot is shown in Figure 2.

Table 1.

Characteristics of included studies.

Author	Publication year	Study design	Postpartum depression evaluation time point	Sample size	Checklist
Adams et al. (2012)	2012	Cohort study	From 30 weeks of gestation to 6 months postpartum.	55,814	SCL-8 ⩾ 2
Adewuya et al. (2005)	2005	Cross-sectional	6 weeks after delivery	876	EPDS ⩾ 9
Ahmed et al. (2012)	2012	Cross-sectional	6–8 weeks postpartum	1000	EPDS ⩾ 10
Alharbi and Abdulghani (2014)	2014	Cohort study	8–12 weeks postpartum	352	EPDS ⩾ 10
Astbury et al. (1994)	1994	Cross-sectional	8–9 months after birth	764	EPDS ⩾ 13
Barbadoro et al. (2012)	2012	Cross-sectional	–	5812	Standardized questionnaire
Bener et al. (2012)	2012	Cross-sectional	6 months after delivery	1379	EPDS ⩾ 12
Bergant et al. (1998)	1998	Cross-sectional	5 days after delivery	1250	EPDS ⩾ 10
Blom et al. (2010)	2010	Cohort study	2 months after delivery	4941	EPDS ⩾ 12
Boyce and Todd (1992)	1992	Cohort study	3 months postpartum	188	EPDS ⩾ 12.5
Chaaya et al. (2002)	2002	Cohort study	24 h and 3–5 months after delivery	396	Standardized questionnaire
Chang et al. (2015)	2015	Cohort study	3 months postpartum	351	CES-D ⩾ 16
Chen et al. (2017)	2017	Cohort study	First postpartum year	25,238	–
Cirik et al. (2016)	2016	Cohort study	2–6 months after delivery	149	EPDS ⩾ 13
Cryan (2001)	2001	Cohort study	6 and 12 weeks postpartum	377	EPDS > 12
Eckerdal et al. (2018)	2018	Cohort study	6 weeks postpartum	3888	EPDS ⩾ 12
Edwards et al. (1994)	1994	Cohort study	Questionnaire between 21 and 32 months after the delivery	196	Dunedin
Farr et al. (2014)	2014	Cross-sectional	4 months after their infant’s birth	4454	Standardized questionnaire
Garel et al. (1988)	1988	Cohort study	2 months after delivery	200	Standardized questionnaire
Ghosh and Goswami (2011)	2011	Cohort study	4th–7th day postpartum	6000	EPDS ⩾ 13
Hannah et al. (2002)	2002	Cohort study	3 months after delivery	1596	EPDS > 12
Hutton et al. (2015)	2015	Cohort study	3 months after delivery	2570	EPDS > 12
Iwata et al. (2015)	2015	Cohort study	6 months after delivery	477	EPDS ⩾ 9
Kim and Dee (2018)	2018	Cross-sectional	One living infant younger than 12 months old	223	EPDS ⩾ 10
Krause et al. (2009)	2009	Cross-sectional	6 weeks postpartum	483	EPDS ⩾ 13
Maliszewska et al. (2016)	2016	Cross-sectional	First week after delivery	546	EPDS > 12
Mathisen et al. (2013)	2013	Cross-sectional	<6 weeks	86	EPDS ⩾ 10
Meky et al. (2020)	2019	Cross-sectional	8 and 16 weeks postnatal	370	EPDS ⩾ 10
Meltzer-Brody et al. (2017)	2017	Cohort study	9–12 months postpartum	392,458	Standardized questionnaire
Nelson et al. (2013)	2013	Cohort study	4 weeks after delivery	17,663	EPDS ⩾ 13
Nicklas et al. (2013)	2013	Cohort study	4–15 weeks postpartum	71	EPDS ⩾ 9
Nielsen Forman et al. (2000)	2000	Cohort study	4 months after giving birth	5091	EPDS ⩾ 13
Patel et al. (2005)	2005	Cohort study	8 weeks postnatal	11,505	EPDS ⩾ 13
Petrosyan et al. (2011)	2011	Cohort study	3 months old children	335	EPDS ⩾ 12
Rowlands and Redshaw (2012)	2012	Cross-sectional	3 months after birth	4882	Checklist
Sadat et al. (2014)	2014	Cross-sectional	2 and 4 months postpartum	300	EPDS ⩾ 13
Sheela and Venkatesh (2016)	2016	Cohort study	4–7 days postpartum	1600	EPDS ⩾ 13
Sylven et al. (2017)	2017	Cohort study	5 days and 6 weeks postpartum	530	EPDS ⩾ 12
Taherifard et al. (2013)	2013	Cross-sectional	6th and 8th week after delivery	197	EPDS ⩾ 13
Xie et al. (2011)	2011	Cohort study	2 weeks postpartum	534	EPDS ⩾ 13
Youn et al. (2017)	2017	Cross-sectional	The first year after delivery	1,269,130	The depression diagnosis within the first year after delivery
Zanardo et al. (2018)	2018	Cohort study	Second day postpartum	2234	EPDS > 12
Zanardo et al. (2017)	2017	Cohort study	2 days after delivery	950	EPDS > 9
Total studies 43				1,827,456

EPDS: Edinburgh Postpartum Depression Scale. SCL-8: Hopkins Symptom Checklist-25; CES-D: Taiwanese version of the Center for Epidemiologic Studies Depression Scale.

Figure 2.

Network plot for all studies.

Risk of bias within studies

An 11-point scale was used to measure AHRQ, and the average total score of 43 included articles was 9.3 points. One point-lost item ‘Describe how confounding was assessed and / or controlled’ had an average 0.5 out of 1 point. Another point-lost item ‘Clarify what follow-up, if any, was expected and the percentage of patients for which incomplete data or follow-up was obtained’ had an average 0.5 out of 1 point.

Synthesis of results

The OR of high EPDS score (directly compared between VD and CS)

First, the OR of PPD between CS and VD was 1.33, 95% confidence interval (CI) = [1.21, 1.46], p-value < 0.00001, I² = 61%, indicating that the risk of high EPDS score in CS was 1.33 times that in VD. Second, the OR of PPD between CS and VD was: for ‘within 2 weeks postpartum’ subgroup, OR = 1.38, 95% CI = [1.07, 1.79], p-value = 0.01, I² = 80%, Figure 3(a); for ‘from 2 weeks to 6 months postpartum’ subgroup, OR = 1.3, 95% CI = [1.13, 1.50], p-value = 0.0003, I² = 59%, Figure 3(b); for ‘More than 6 months postpartum’ subgroup, OR = 1.31, 95% CI = [1.21, 1.46], p-value < 0.00001, I² = 61%, Figure 3(c). The above results showed that the risk of high EPDS score after CS was significantly higher than that after VD in three time periods. Third, the OR of PPD in the subgroup with EPDS score ⩾ 13 was 1.07, 95% CI = [0.99, 1.16], p-value = 0.08, I² = 46%, indicating that there was no significant difference between CS and VD in risk of high EPDS score in the subgroup with severe PPD, Figure 4(a). Finally, the adjusted OR of the risk of PPD was OR = 1.34, 95% CI = [1.15, 1.56], p-value = 0.0002, I² = 44%, indicating that the risk of PPD in CS was 1.34 times that in VD after adjusting for various influential factors related to PPD, Figure 4(b). Considering the across-studies heterogeneity may be caused by different PPD measurement checklists, we conducted a subgroup analysis involving only EPDS and found that the variety of conflated effect value was too small to change the statistical significance (Supplementary Figure S1). Furthermore, we conducted subgroup analyses of cohort studies and cross-sectional studies and found that the combined results of the two types of studies were basically the same, except for the ‘within 2 weeks postpartum’ subgroup in cohort study, where the combined results changed to statistically insignificant (Supplementary Figures S2 and S3).

Figure 3.

Forest plot of odds ratio of PPD between CS and VD: (a) within 2 weeks postpartum, (b) 2 weeks to 6 months postpartum and (c) more than 6 months postpartum.

Figure 4.

(a) Forest plot of odds ratio of PPD in the subgroup with EPDS score ⩾ 13 and (b) Forest plot of the adjusted odds ratio of the risk of PPD.

The OR of high EPDS score in four delivery modes analyzed by network meta-analysis approach

The risk of high EPDS score in EMCS was 1.53 times that in SVD, and the risk of high EPDS score in ELCS was 1.47 times that in SVD, showing significant differences. The ORs were 1.53 (95% CI = [1.22, 1.91]) and 1.47 (95% CI = [1.16, 1.86]), respectively. The risk of high EPDS score was statistically insignificant in the pairwise comparisons of other delivery modes (Figure 5(a)). Furthermore, we conducted a subgroup analysis to differentiate between the use of vacuum or forceps and found no difference between their risks of high EPDS score, as displayed in Supplementary Figure S4.

Figure 5.

(a) Forest plot of odds ratio of PPD pairwise comparisons among four delivery modes (SVD, IVD, EMCS, ELCS) and (b) Forest plot of odds ratio of PPD pairwise comparisons among four delivery modes (SVD, IVD, EMCS, ELCS) in the subgroup with EPDS score ⩾ 13.

The network meta-analysis of the risk of high EPDS score in four delivery modes with EPDS ⩾ 13 showed that the risk of EPDS score ⩾ 13 in ELCS was 1.47 times that in IVD and the OR was 1.47 (95% CI = [1.09, 1.99]). The risk of EPDS score ⩾ 13 was statistically insignificant in the pairwise comparisons of other delivery modes (Figure 5(b)).

Ranking probabilities

According to SUCRA, the risk of high EPDS score in four delivery modes is shown in Table 2. From the view of point of the high EPDS score, EMCS had the highest risk, followed by ELCS, IVD ranked third and SVD had the lowest risk.

Table 2.

Ranking probabilities.

Delivery mode	SUCRA	PrBest	Mean rank
VD	98.0	94.0	1.1
IVD	65.5	5.9	2.0
EMCS	13.4	0.0	3.6
ELCS	23.2	0.1	3.3

VD: vaginal delivery; IVD: instrumental vaginal delivery; ELCS: elective cesarean section; EMCS: emergency cesarean section; SUCRA: surface under the cumulative ranking curve.

Inconsistency test

No evidence for statistically significant inconsistency (Global inconsistency tests p = 0.200) was found.

Risk of bias across studies

Publication bias was detected in the direct meta-analytical funnel plot (Figure 6). According to the trim-and-fill method, we filled nine studies and adjusted the pooled effect value from 1.33 (95% CI = [1.21, 1.46]), p-value < 0.00001, to 1.20 (95% CI = [1.09, 1.33]), but did not produce substantial effect, indicating that the results of the direct meta-analysis were not significantly affected by publication bias (Figure 7). The network meta-analytical funnel plot was shown in Figure 8. The funnel plot was relatively symmetrical, showing no evidence of publication bias.

Figure 6.

Funnel plot of odds ratio of PPD between CS and VD.

Figure 7.

Filled funnel plot of odds ratio of PPD between CS and VD.

Figure 8.

The network meta-analytical funnel plot: A = SVD, B = IVD, C = EMCS and D = ELCS.

Discussion

This meta-analysis included 43 studies with a total sample size of 1,827,456 participants. This study systematically explored the relationship between delivery mode and PPD from two perspectives. On one hand, we measured the high EPDS score after CS and VD. On the other hand, we analyzed the high EPDS score in ELCS, EMCS, IVD and SVD by network meta-analysis.

The comparative risk of PPD between CS and VD

Our study found that the unadjusted risk of PPD in CS was 1.33 times that in VD. Baby blues usually begins a few days after childbirth and lasts for 10–14 days and can be regarded as a mild and brief postpartum disorder. Although for most women (26–85%) PPD is relatively mild and usually resolves within 2 weeks (Complojer et al., 2009), 10–15% of women experience more severe and lasting PPD (Warner et al., 1996). Since the timing of PPD assessment may affect the prevalence, we formed three subgroups for comparative analysis: 2 weeks after delivery, 2 weeks to 6 months after delivery and more than 6 months after delivery. Basically, the same results were obtained for three time periods: the risk of PPD in CS group was 1.38 times, 1.30 times and 1.31 times that in VD, respectively. In addition, due to the different threshold scores of EPDS in the diagnosis of PPD in different studies, we further analyzed PPD women with EPDS scores ⩾ 13. (A woman with a score of 13 or more is considered to be severely depressed, with scores of 10–12 being classified as marginal cases and scores of 0–9 as being without depression) (Cox et al., 1987; Eberhard-Gran et al., 2001). Our study found that there was no significant difference in the risk of severe PPD between women who delivered by CS and women who delivered by VD, and women who delivered by CS were more likely to have mild PPD than women who delivered by VD. In addition, PPD is the result of many factors. On the basis of comprehensive factors, we calculated and obtained the adjusted OR, which showed that the adjusted risk of PPD in women delivering by CS was 1.34 times that in women delivering by VD, which was consistent with the unadjusted value, fully showing that CS was an important factor of PPD. This result can be explained both physically and psychologically. Physiologically, a large amount of cortisol is released during VD. Cortisol is a stress hormone caused by a stress response. Studies have shown that the link between high cortisol levels and PPD caused by surgical stress may be related to the reduction of brain 5-hydroxytryptamine (5-HT), which makes susceptible subjects prone to depression (Dinan, 1994; Edwards et al., 1994). Second, the inflammation, associated pain and stress caused by surgery theoretically increase the susceptibility of women to depression (Eckerdal et al., 2018). Third, oxytocin also affects PPD. Oxytocin is important for uterine contraction and breastfeeding. Oxytocin levels are usually lower in women delivering by CS than in women delivering by VD. In the process of VD, the content of oxytocin in the brain of pregnant women will increase, which will help reduce stress and improve happiness, thereby reducing the possibility of postpartum mental disorders (Kim et al., 2014; Lonstein et al., 2014; Mottolese et al., 2014). Psychologically, many women have negative emotions after VD, such as inferiority, sense of failure, loss of control or disappointment. Nursing wounds, slow postpartum recovery and the effects of complications may increase the risk of PPD. In addition, women who undergo CS have a lack of confidence because they cannot give birth naturally. Women’s lack of self-confidence can be stressful when raising children. In addition, changes in physical pain and adaptation to life after childbirth will cause mental disorders and then develop into PPD (Loto et al., 2010).

The comparative risk of PPD between ELCS, EMCS, IVD and SVD

This study used a network meta-analysis to analyze the risk of PPD in ELCS, EMCS, IVD and SVD. The results showed that the risk of PPD in women giving birth by EMCS was 1.53 times that of women by SVD, and the risk of PPD in women giving birth by ELCS was 1.47 times that of women by SVD, but there was no significant difference in the risk of PPD among women giving birth by IVD and SVD. We also analyzed women with an EPDS score ⩾ 13 and found that women who gave birth through one of four delivery modes had no significant difference in the risk of severe PPD. We found that EMCS was associated with the highest risk of mild PPD, followed by ELCS and IVD, while SVD was associated with the lowest risk of mild PPD. From a physiological point of view, this difference may be related to the use of anesthetics. Some studies found that the majority of participants giving birth by EMCS used general anesthesia, while the participants giving birth by ELCS used almost spinal anesthesia (Meky et al., 2020). General anesthesia may cause depression through certain biochemical mechanisms that function at the molecular level. There is an evidence that the volatile anesthetics inhibit 5-HT uptake; that thiopentone decreases transmission at serotonergic, dopaminergic and noradrenergic synapses; and that nitrous oxide can interfere with the synthesis of S-adenosyl methionine and thereby that of monoamines. The effect of barbiturates and benzodiazepines in increasing GABAergic transmission is well known, and there is a possibility that these drugs can precipitate depression (Edwards et al., 1994). In addition, from a physiological point of view, EMCS, as well as the accompanying obstetric and neonatal complications, leads to negative experience and increases the risk of PPD. It has been reported that EMCS has more complications and a longer recovery time than SVD and ELCS (Bergholt et al., 2003; Jansen et al., 2007). Women who give birth by EMCS may experience physical fatigue due to the failure of VD, as well as the psychological burden and subjective feeling of being unable to give birth. All of these can lead to a negative delivery experience, resulting in depression (Eckerdal et al., 2016; Silverman et al., 2017).

Strengths

Our research has several advantages. First of all, our study is the largest meta-analysis of the relationship between delivery modes and PPD on basis of analysis of 1.82 million samples. Second, we analyzed not only VD and CS, but also four delivery modes through network meta-analysis. This is a comprehensive meta-analysis of the relationship between various delivery modes and PPD. Third, this study is the first meta-analysis to distinguish between severe and mild PPD and modes of delivery.

Limitations

First, most of the included studies used questionnaires to assess PPD, while EPDS is used to screen for clinical symptoms of PPD, not PPD itself. However, EPDS scores are closely related to clinical diagnosis and are often used to measure PPD. Second, PPD is related to many confounding factors. Most studies provide only crude results, and the number of studies we used to combine the adjusted effect values was slightly smaller.

Conclusion

In conclusion, although there are many factors that affect PPD, our study shows that the mode of delivery has a significant impact on the occurrence of mild PPD. The risk of mild PPD was 33% higher in women delivering by CS than in women delivering by VD, and 53% higher in women delivering by EMCS than in women delivering by SVD. Given the large number of CSs per year, the number of women who may be affected by PPD will be high. If used properly, CS is an important lifesaving operation. However, unnecessary CS will not only increase the cost of medical treatment but also increase the risk of maternal and infant health. Therefore, we suggest to take necessary measures to reduce unnecessary CS. The risks associated with CS should be informed to the maternal and relevant medical professionals. We should carefully monitor the progress of postpartum mental disorders in women who delivered by CS and make it possible for women to have a quick access to mental healthcare.

Supplemental Material

Supplemental_material – Supplemental material for Association between mode of delivery and postpartum depression: A systematic review and network meta-analysis

Supplemental material, Supplemental_material for Association between mode of delivery and postpartum depression: A systematic review and network meta-analysis by Lei Sun, Su Wang and Xi-Qian Li in Australian & New Zealand Journal of Psychiatry

Footnotes

Acknowledgements

We would like to thank Beijing Zhiyun data technology co. LTD for providing data analysis service.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship and/or publication of this article.

ORCID iD

Xi-Qian Li

Supplemental Material

Supplemental material for this article is available online.

References

Adams

Eberhard-Gran

Sandvik

, et al. (2012) Mode of delivery and postpartum emotional distress: A cohort study of 55,814 women. British Journal of Obstetrics and Gynaecology (BJOG) 119: 298–305.

Adewuya

Fatoye

Ola

, et al. (2005) Sociodemographic and obstetric risk factors for postpartum depressive symptoms in Nigerian women. Journal of Psychiatric Practice 11: 353–358.

Ahmed

Alalaf

Al-Tawil

(2012) Screening for postpartum depression using Kurdish version of Edinburgh postnatal depression scale. Archives of Gynecology and Obstetrics 285: 1249–1255.

Alharbi

Abdulghani

(2014) Risk factors associated with postpartum depression in the Saudi population. Neuropsychiatric Disease and Treatment 10: 311–316.

Astbury

Brown

Lumley

, et al. (1994) Birth events, birth experiences and social differences in postnatal depression. Australian Journal of Public Health 18: 176–184.

Barbadoro

Cotichelli

Chiatti

, et al. (2012) Socio-economic determinants and self-reported depressive symptoms during postpartum period. Women Health 52: 352–368.

Bener

Burgut

Ghuloum

, et al. (2012) A study of postpartum depression in a fast developing country: Prevalence and related factors. International Journal of Psychiatry in Medicine 43: 325–337.

Bergant

Nguyen

Moser

, et al. (1998) Prevalence of depressive disorders in early puerperium. Gynäkologisch-Geburtshilfliche Rundschau 38: 232–237.

Bergholt

Stenderup

Vedsted-Jakobsen

, et al. (2003) Intraoperative surgical complication during cesarean section: An observational study of the incidence and risk factors. Acta obstetricia et gynecologica Scandinavica 82: 251–256.

10.

Blom

Jansen

Verhulst

, et al. (2010) Perinatal complications increase the risk of postpartum depression. The Generation R Study. British Journal of Obstetrics and Gynaecology (BJOG) 117: 1390–1398.

11.

Boyce

Todd

(1992) Increased risk of postnatal depression after emergency caesarean section. Medical Journal of Australia 157: 172–174.

12.

Buultjens

Liamputtong

(2007) When giving life starts to take the life out of you: Women’s experiences of depression after childbirth. Midwifery 23: 77–91.

13.

Carter

Frampton

Mulder

(2006) Cesarean section and postpartum depression: A review of the evidence examining the link. Psychosomatic Medicine 68: 321–330.

14.

Chaaya

Campbell

El Kak

, et al. (2002) Postpartum depression: Prevalence and determinants in Lebanon. Archives of Women’s Mental Health 5: 65–72.

15.

Chandrasekaran

De Souza

Urquia

, et al. (2018) Is anemia an independent risk factor for postpartum depression in women who have a cesarean section?—A prospective observational study. BMC Pregnancy and Childbirth 18: 400.

16.

Chang

Chen

, et al. (2015) Depressive symptoms, pain, and sexual dysfunction over the first year following vaginal or cesarean delivery: A prospective longitudinal study. International Journal of Nursing Studies 52: 1433–1444.

17.

Chen

Lai

Hwang

, et al. (2017) Understanding the relationship between cesarean birth and stress, anxiety, and depression after childbirth: A nationwide cohort study. Birth 44: 369–376.

18.

Cirik

Yerebasmaz

Kotan

, et al. (2016) The impact of prenatal psychologic and obstetric parameters on postpartum depression in late-term pregnancies: A preliminary study. Taiwanese Journal of Obstetrics and Gynecology 55: 374–378.

19.

Complojer

Schweigkofler

Schwitzer

, et al. (2009) The preconditions of postpartum dysphoria. Neuropsychiatrie 23: 35–41.

20.

Cox

Holden

Sagovsky

(1987) Detection of postnatal depression. Development of the 10-item Edinburgh Postnatal Depression Scale. British Journal of Psychiatry 150: 782–786.

21.

Cryan

Keogh

Connolly

, et al. (2001) Depression among postnatal women in an urban Irish community. Irish Journal of Psychological Medicine 18(01): 5–10.

22.

Dinan

(1994) Glucocorticoids and the genesis of depressive illness. A psychobiological model. British Journal of Psychiatry 164: 365–371.

23.

Eberhard-Gran

Eskild

Tambs

, et al. (2001) Review of validation studies of the Edinburgh Postnatal Depression Scale. Acta Psychiatrica Scandinavica 104: 243–249.

24.

Eckerdal

Georgakis

Kollia

, et al. (2018) Delineating the association between mode of delivery and postpartum depression symptoms: A longitudinal study. Acta obstetricia et gynecologica Scandinavica 97: 301–311.

25.

Eckerdal

Kollia

Lofblad

, et al. (2016) Delineating the association between heavy postpartum haemorrhage and postpartum depression. PLoS ONE 11: e0144274.

26.

Edwards

Porter

Stein

(1994) A pilot study of postnatal depression following caesarean section using two retrospective self-rating instruments. Journal of Psychosomatic Research 38: 111–117.

27.

Eisenach

Pan

Smiley

, et al. (2008) Severity of acute pain after childbirth, but not type of delivery, predicts persistent pain and postpartum depression. Pain 140: 87–94.

28.

Farr

Dietz

O’Hara

, et al. (2014) Postpartum anxiety and comorbid depression in a population-based sample of women. Journal of Women’s Health (Larchmt) 23: 120–128.

29.

Garel

Lelong

Kaminski

(1988) Follow-up study of psychological consequences of caesarean childbirth. Early Human Development 16: 271–282.

30.

Ghosh

Goswami

(2011) Evaluation of post partum depression in a tertiary hospital. Journal of Obstetrics and Gynaecology of India 61: 528–530.

31.

Goker

Yanikkerem

Demet

, et al. (2012) Postpartum depression: Is mode of delivery a risk factor? ISRN Obstetrics and Gynecology 2012: 616759.

32.

Hannah

Hodnett

, et al. (2002) Outcomes at 3 months after planned cesarean vs planned vaginal delivery for breech presentation at term: The international randomized Term Breech Trial. Journal of the American Medical Association (JAMA) 287: 1822–1831.

33.

Henderson

Evans

Straton

, et al. (2003) Impact of postnatal depression on breastfeeding duration. Birth 30: 175–180.

34.

Hutton

Hannah

Ross

, et al. (2015) Maternal outcomes at 3 months after planned caesarean section versus planned vaginal birth for twin pregnancies in the Twin Birth Study: A randomised controlled trial. British Journal of Obstetrics and Gynaecology (BJOG) 122: 1653–1662.

35.

Iwata

Mori

Tsuchiya

, et al. (2015) Predicting early post-partum depressive symptoms among older primiparous Japanese mothers. Japan Journal of Nursing Science 12: 297–308.

36.

Jansen

Duvekot

Hop

, et al. (2007) New insights into fatigue and health-related quality of life after delivery. Acta obstetricia et gynecologica Scandinavica 86: 579–584.

37.

Kim

Soeken

Cromer

, et al. (2014) Oxytocin and postpartum depression: Delivering on what’s known and what’s not. Brain Research 1580: 219–232.

38.

Kim

Dee

(2018) Sociodemographic and obstetric factors related to symptoms of postpartum depression in Hispanic women in rural California. Journal of Obstetric, Gynecologic & Neonatal Nursing 47: 23–31.

39.

Krause

Ostbye

Swamy

(2009) Occurrence and correlates of postpartum depression in overweight and obese women: Results from the active mothers postpartum (AMP) study. Maternal and Child Health Journal 13: 832–838.

40.

Liberati

Altman

Tetzlaff

, et al. (2009) The PRISMA statement for reporting systematic reviews and meta-analyses of studies that evaluate healthcare interventions: Explanation and elaboration. British Medical Journal (BMJ) 339: b2700.

41.

Lonstein

Maguire

Meinlschmidt

, et al. (2014) Emotion and mood adaptations in the peripartum female: Complementary contributions of GABA and oxytocin. Journal of Neuroendocrinology 26: 649–664.

42.

Loto

Adewuya

Ajenifuja

, et al. (2010) Cesarean section in relation to self-esteem and parenting among new mothers in southwestern Nigeria. Acta obstetricia et gynecologica Scandinavica 89: 35–38.

43.

Maliszewska

Bidzan

Swiatkowska-Freund

, et al. (2016) Personality type, social support and other correlates of risk for affective disorders in early puerperium. Ginekologia Polska 87: 814–819.

44.

Mathisen

Glavin

Lien

, et al. (2013) Prevalence and risk factors for postpartum depressive symptoms in Argentina: A cross-sectional study. International Journal of Women’s Health 5: 787–793.

45.

Meky

Shaaban

Ahmed

, et al. (2020) Prevalence of postpartum depression regarding mode of delivery: A cross-sectional study. Journal of Maternal-Fetal & Neonatal Medicine 33: 3300–3307.

46.

Meltzer-Brody

Maegbaek

Medland

, et al. (2017) Obstetrical, pregnancy and socio-economic predictors for new-onset severe postpartum psychiatric disorders in primiparous women. Psychological Medicine 47: 1427–1441.

47.

Mottolese

Redoute

Costes

, et al. (2014) Switching brain serotonin with oxytocin. Proceedings of the National Academy of Sciences of the United States of America 111: 8637–8642.

48.

Nelson

Freeman

Johnson

, et al. (2013) A prospective study of postpartum depression in 17 648 parturients. Journal of Maternal-Fetal & Neonatal Medicine 26: 1155–1161.

49.

Nicklas

Miller

Zera

, et al. (2013) Factors associated with depressive symptoms in the early postpartum period among women with recent gestational diabetes mellitus. Maternal and Child Health Journal 17: 1665–1672.

50.

Nielsen Forman

Videbech

Hedegaard

, et al. (2000) Postpartum depression: Identification of women at risk. British Journal of Obstetrics and Gynaecology (BJOG) 107: 1210–1217.

51.

O’Hara

Swain

(2009) Rates and risk of postpartum depression—A meta-analysis. International Review of Psychiatry 8: 37–54.

52.

Patel

Murphy

Peters

(2005) Operative delivery and postnatal depression: A cohort study. British Medical Journal (BMJ) 330: 879.

53.

Petrosyan

Armenian

Arzoumanian

(2011) Interaction of maternal age and mode of delivery in the development of postpartum depression in Yerevan, Armenia. Journal of Affective Disorders 135: 77–81.

54.

Rowlands

Redshaw

(2012) Mode of birth and women’s psychological and physical wellbeing in the postnatal period. BMC Pregnancy and Childbirth 12: 138.

55.

Sadat

Kafaei Atrian

Masoudi Alavi

, et al. (2014) Effect of mode of delivery on postpartum depression in Iranian women. Journal of Obstetrics and Gynaecology Research 40: 172–177.

56.

Sheela

Venkatesh

(2016) Screening for postnatal depression in a tertiary care hospital. Journal of Obstetrics and Gynaecology of India 66: 72–76.

57.

Silverman

Reichenberg

Savitz

, et al. (2017) The risk factors for postpartum depression: A population-based study. Depression and Anxiety 34: 178–187.

58.

Sword

Landy

Thabane

, et al. (2011) Is mode of delivery associated with postpartum depression at 6 weeks: A prospective cohort study. British Journal of Obstetrics and Gynaecology (BJOG) 118: 966–977.

59.

Sylven

Thomopoulos

Kollia

, et al. (2017) Correlates of postpartum depression in first time mothers without previous psychiatric contact. European Psychiatry 40: 4–12.

60.

Taherifard

Delpisheh

Shirali

, et al. (2013) Socioeconomic, psychiatric and materiality determinants and risk of postpartum depression in border city of Ilam, Western Iran. Depression Research and Treatment 2013: 653471.

61.

Warner

Appleby

Whitton

, et al. (1996) Demographic and obstetric risk factors for postnatal psychiatric morbidity. British Journal of Psychiatry 168: 607–611.

62.

Xie

Lei

Wang

, et al. (2011) Cesarean section and postpartum depression in a cohort of Chinese women with a high cesarean delivery rate. Journal of Women’s Health (Larchmt) 20: 1881–1886.

63.

Youn

Lee

Han

, et al. (2017) Obstetric risk factors for depression during the postpartum period in South Korea: A nationwide study. Journal of Psychosomatic Research 102: 15–20.

64.

Zanardo

Giliberti

, et al. (2018) The role of elective and emergency cesarean delivery in maternal postpartum anhedonia, anxiety, and depression. International Journal of Gynaecology and Obstetrics 143: 374–378.

65.

Zanardo

Giliberti

Volpe

, et al. (2017) Cohort study of the depression, anxiety, and anhedonia components of the Edinburgh Postnatal Depression Scale after delivery. International Journal of Gynaecology and Obstetrics 137: 277–281.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

1.23 MB