Defining maternal depression

Maternal depression can broadly be divided based on the time of occurrence: during the pregnancy (in the “prenatal period”) or after the delivery of a child (in the “postnatal period”). A recent systematic review suggests that prevalence of postnatal CPMD (19.8%) is higher than the prevalence of prenatal CPMD (15.6%) in LMICs . Just as with depression outside the perinatal period, there is no universally accepted definition for screening or diagnosis of postnatal depression. Definitions vary in their method of assessment and the specified length of the “postnatal period.” For example, the DSM-IV defines “postpartum depression” to have a window of onset within 4 weeks of delivery, while the ICD-10 defines the window to be up to 6 weeks from delivery . Postnatal depression may encompass postpartum depression , and definitions have spanned from anywhere from 1 week to 1 year after the delivery of a child .

Prenatal depression

The available data suggest that postnatal depression is higher in LMICs than in HICs , though prevalence estimates vary substantially and range from 0% to 60% depending on the setting . In addition to varying time periods of interest and different cultural and socioeconomic contexts of motherhood across studies, this wide range of prevalence may be due to inconsistent measurement tools (e.g., measures of depressive symptoms versus a diagnostic measure) or sampling designs (e.g., clinical versus community samples) .

In a cross-cultural qualitative study in eleven countries, respondents universally acknowledged the phenomenon of morbid unhappiness after delivery and also universally recognized the social support from family partners as a possible remedy . Similarly, several quantitative studies have confirmed that social stressors or economic hardships are risk factors for postnatal depression . Intimate partner violence is another risk factor for prenatal or postnatal depression across culturally and economically diverse settings . Psychological violence alone or psychological, physical, and sexual partner violence together during pregnancy were predictive of postnatal depression (ORs 1.58 and 1.76, respectively) among women in Brazil . Women in rural Pakistan who experienced domestic violence were three times more likely to be depressed 3 months after the birth of a child . In addition, the history of depression or mood disorders during and prior to pregnancy has consistently been related to postnatal depression .

Postnatal depression

Depression during pregnancy has been less studied than postnatal depression . Much of the research in the prenatal period focuses on the third trimester, a time when the risk of depression is higher than that observed in the general population . Although our best pooled estimate of prenatal CPMDs, of which depression is the most common disorder, in LMICs is 15.6% , several studies of rural women report higher estimates (e.g., 18%, 25%, and 47% in rural areas of Bangladesh , Pakistan , and South Africa , respectively).

Many social, family, and socioeconomic factors have been linked to prenatal depression. Several studies have found that being single/unmarried at the time of pregnancy is related to prenatal depressive symptoms , in addition to having negative attitudes towards the pregnancy . Socioeconomic correlates of depressive symptomology during pregnancy included being in debt , having fewer years of schooling , employment insecurity, or participating in casual (as opposed to salaried) work . Poor family relationships and hostile home environments are also consistently linked to prenatal depression. Partner violence , lack of support from husbands or mothers-in-law , and lack of decision-making power in major decisions were each more likely to be reported by depressed women compared with nondepressed women.

Defining maternal depression

Maternal depression can broadly be divided based on the time of occurrence: during the pregnancy (in the “prenatal period”) or after the delivery of a child (in the “postnatal period”). A recent systematic review suggests that prevalence of postnatal CPMD (19.8%) is higher than the prevalence of prenatal CPMD (15.6%) in LMICs . Just as with depression outside the perinatal period, there is no universally accepted definition for screening or diagnosis of postnatal depression. Definitions vary in their method of assessment and the specified length of the “postnatal period.” For example, the DSM-IV defines “postpartum depression” to have a window of onset within 4 weeks of delivery, while the ICD-10 defines the window to be up to 6 weeks from delivery . Postnatal depression may encompass postpartum depression , and definitions have spanned from anywhere from 1 week to 1 year after the delivery of a child .

Prenatal depression

The available data suggest that postnatal depression is higher in LMICs than in HICs , though prevalence estimates vary substantially and range from 0% to 60% depending on the setting . In addition to varying time periods of interest and different cultural and socioeconomic contexts of motherhood across studies, this wide range of prevalence may be due to inconsistent measurement tools (e.g., measures of depressive symptoms versus a diagnostic measure) or sampling designs (e.g., clinical versus community samples) .

In a cross-cultural qualitative study in eleven countries, respondents universally acknowledged the phenomenon of morbid unhappiness after delivery and also universally recognized the social support from family partners as a possible remedy . Similarly, several quantitative studies have confirmed that social stressors or economic hardships are risk factors for postnatal depression . Intimate partner violence is another risk factor for prenatal or postnatal depression across culturally and economically diverse settings . Psychological violence alone or psychological, physical, and sexual partner violence together during pregnancy were predictive of postnatal depression (ORs 1.58 and 1.76, respectively) among women in Brazil . Women in rural Pakistan who experienced domestic violence were three times more likely to be depressed 3 months after the birth of a child . In addition, the history of depression or mood disorders during and prior to pregnancy has consistently been related to postnatal depression .

Postnatal depression

Depression during pregnancy has been less studied than postnatal depression . Much of the research in the prenatal period focuses on the third trimester, a time when the risk of depression is higher than that observed in the general population . Although our best pooled estimate of prenatal CPMDs, of which depression is the most common disorder, in LMICs is 15.6% , several studies of rural women report higher estimates (e.g., 18%, 25%, and 47% in rural areas of Bangladesh , Pakistan , and South Africa , respectively).

Many social, family, and socioeconomic factors have been linked to prenatal depression. Several studies have found that being single/unmarried at the time of pregnancy is related to prenatal depressive symptoms , in addition to having negative attitudes towards the pregnancy . Socioeconomic correlates of depressive symptomology during pregnancy included being in debt , having fewer years of schooling , employment insecurity, or participating in casual (as opposed to salaried) work . Poor family relationships and hostile home environments are also consistently linked to prenatal depression. Partner violence , lack of support from husbands or mothers-in-law , and lack of decision-making power in major decisions were each more likely to be reported by depressed women compared with nondepressed women.

Perinatal depression and suicidal ideation

Suicide ideation is a major concern among mothers who experience depression. Data from the Confidential Enquiry into Maternal Deaths in the United Kingdom demonstrate that suicide rates in the postnatal period may be as high as at other times in a woman’s life, challenging the notion that maternity protects against suicidal behavior . Several studies have examined suicidality, which is measured as thoughts of self-harm during depression screening, for example, a single item (item 10) on the Edinburgh Postnatal Depression Scale (EPDS) or a single item on the Patient Health Questionnaire (PHQ) .

Previous or concurrent depression, economic hardship, interpersonal violence, and lack of social support appear to play an important role in suicidal ideation during the prenatal period . Among rural Bangladeshi women 34–35 weeks into their pregnancy, 14% who were classified as depressed reported thoughts of self-harm, and in a qualitative follow-up, the authors found that suicidal thoughts were in particular associated with spousal violence . Nearly one-quarter of pregnant women receiving antenatal care in Brazil had “current suicide risk”; investigators found depressive symptomology (OR = 4.44) and years of education (OR = 0.45) to be significant correlates of current suicide risk . Another study of pregnant women in Brazil found the prevalence of suicidality to be 8%, and lack of education, low socioeconomic class, being single, and symptoms of anxiety and depression were all significantly higher in women who reported suicidality compared with those who did not . Among Peruvian women, symptoms of depression, lifetime intimate partner violence, history of childhood abuse, and difficulty in paying for basic needs were significantly higher among women who had thoughts of self-harm, which were measured using both the PHQ-9 and the EPDS .

In the postnatal period, the prevalence of suicidal variation varied from 4% in Nepal to 8% in South Africa . In South African women 3 months after delivery, experiencing food insecurity was associated with higher likelihood of probably depression (adjusted risk ratio [RR] = 1.05, p < .05) and suicidality (adjusted RR = 1.15, p < .05) .

Health-care seeking during pregnancy among depressed women

Women who are depressed during pregnancy do not often seek treatment for their depression . This may be in part due to the scarcity of mental health services in LMICs . As mental health conditions are stigmatized in many LMICs , the desire to acknowledge or seek treatment may also be culturally sensitive. There are several reports that depressed women tend to express their symptoms somatically in non-Western settings . From a clinical standpoint, physicians in these settings should be sensitized to diagnosing/detecting women who present with somatic symptoms. Migrant women from non-Western cultures have been observed to be more reluctant to discuss mental problems outside the household . Furthermore, pharmaceutical treatment during the pregnancy may be undesirable to women.

Beyond care for depression, depressed mothers may be less likely to seek prenatal care . One Taiwanese study found that in a sample of women with a depressive disorder within 2 years of their delivery, those that had seven or fewer prenatal visits were at approximately a fourfold, fivefold, and twofold higher risk for low birth weight (LBW), preterm birth (PTB), and small-for-gestational age (SGA), respectively .

Stress pathways linking maternal depression to fetal and child growth and development

Depression may be considered an expression of psychosocial stress, and research suggests several potential pathways linking chronic stress to negative perinatal outcomes. The effects of prenatal depression may result from both physiological changes in the mother (i.e., altered maternal hypo-pituitary axis (HPA) and catecholamine function, early evidence on epigenetics) that affect the developing fetus directly and stress-related health behaviors that impact the infant . In utero and postnatal period when the fetus is developing at an extraordinary rate, environmental stressors such as experience of depressive symptoms can have an especially pronounced impact . Given that women are the primary caretakers of infants and children globally , and stress associated with antenatal depression and anxiety tend to endure across the prenatal and postnatal periods , depression and anxiety may have a cumulative impact in the postnatal period when maternal–child interactions are critical for child growth and development . Although measurement of stress in pregnancy may differ and results may depend on the variation in the response to stressful events and interactions with other factors that modify it (e.g., underlying anxiety as well as support and coping mechanisms) , we summarize what is known about the impacts of maternal stress on fetal and infant growth and development.

During fetal development, growing evidence suggests that maternal stress and depression affect nutrition and maternal–fetal communication and may result in neurobehavioral dysregulation . One study showed that fetuses of depressed compared with nondepressed mothers had elevated baseline heart rates, attenuated response to induced vibroacoustic stimulation, and longer recovery time to return to baseline heart rates . Another similar study did not find differences at the baseline but showed increased heart rate reactivity in fetuses of depressed mothers compared with both nondepressed women and women with anxiety disorders when mothers were exposed to a standardized laboratory challenge paradigm . In terms of fetal activity, fetuses of women with high levels of depressive symptoms have been found to be more active between the fifth and seventh months of pregnancy . Maternal cortisol levels may in part mediate the relation between maternal prenatal depression and fetal weight; fetuses of mothers with elevated maternal cortisol levels had approximately 13 times higher odds of having fetuses with below average estimated fetal weight . Dysregulation of cortisol is additionally expected to adversely affect the mother and child through cardiometabolic (e.g., elevated cholesterol) and inflammatory (e.g., elevated cytokines) responses that may increase the risk of PTB .

The effects of exposure to cortisol in utero appear to carry over into the postnatal period for infants. Mothers with higher urinary cortisol levels at 20 weeks of gestation reported higher depressive symptoms; their fetuses had smaller head circumference, abdominal circumference, biparietal diameter, and fetal weight; and after birth, indicators of growth restriction persisted in their newborns . It has also been suggested that this neurobehavioral dysregulation developed prenatally subsequently interacts with depressed mothers’ poor interaction styles, potentially compounding growth and developmental problems over time . Among infants of depressed mothers, studies have found elevated cortisol and norepinephrine levels like their mothers and greater relative right frontal electroencephalogram symmetry , which is associated with high cortisol levels in infants . Poorer performance on the Brazelton assessment (administered within the first day after birth) has been observed in newborns of depressed mothers, on a number of assessments including irritability, less activity and interest, lower endurance and robustness, as well as lower orientation and motor scores .

Maternal stress may also have physiological effects leading to PTB, including (1) increases in corticotrophin-releasing hormone caused by stress, leading to upregulation of inflammatory cytokine release that stimulates myometrial contractions that in turn leads to preterm labor, (2) stress-related downregulation of functions of T and B lymphocytes that increase susceptibility to infections, which act as co-factors triggering PTB, and (3) stimulation of a low-grade inflammation reflected in increased levels of cytokines and C-reactive protein caused by chronic stress .

Although research on epigenetic mechanisms linking in utero maternal stress to fetal DNA methylation changes is in its infancy, the effect of depression and anxiety in pregnancy on epigenetic changes is one potential pathway through which pregnancy outcomes may be affected . Social stressors, such as intimate partner violence , and psychological stress , such as in response to natural disaster or post-traumatic stress disorder , have shown differential methylation in CpG sites. Likewise, infants of depressed mothers have shown differential methylation in CpG sites compared with unexposed infants . However, evidence is mixed, with not all studies finding such an association .

Maternal depression, pregnancy, and birth outcomes

Several recent reviews summarize the impact of stress and symptoms of depression and anxiety during pregnancy and birth outcomes. Grote et al pooled 29 research studies on the impact of depressive symptoms and depression in pregnancy and birth outcomes including PTB, LBW, and intrauterine growth restriction (IUGR) . Maternal depressive symptoms predicted between 39% and 49% higher risk of these three outcomes in studies with a categorical maternal depression/depressive symptom variable . Weaker non-significant or marginally significant effects were observed for a small number of studies with continuous depression measures for these same outcomes . Of the 29 studies, only three were conducted in LMICs, all of which tended to show relatively strong relationships. Prenatal psychiatric morbidity, comparing the top with the lower three quartiles of the General Health Questionnaire, was associated with a 3.5 times higher likelihood of LBW in India . Maternal antenatal depression was associated with 2.1 times higher risk of LBW in Pakistan . In Brazil, the associations (RRs) of antenatal maternal depressive symptoms with PTB, LBW, and IUGR were 2.3, 2.0, and 1.6, respectively .

Another recent meta-analysis conducted by Grigoriadis et al (2013) found prenatal maternal depression or depressive symptoms significantly associated with 37% higher odds of premature delivery, but showed a non-significant association with LBW, except among studies conducted on low socioeconomic groups . Only two studies included in the review were conducted in LMICs (one of the two in Hong Kong, resembling more closely a high-income economy). The study conducted in a low-income setting, in Pakistan, reported significant associations between maternal depressive symptoms with PTB, LBW, lower mean APGAR scores, and obstetric complications . Although that study lacked a sample size to control potential confounders, strong associations were reported.

Accortt et al (2015) also published a review focused on prenatal depression diagnosis or depressive symptoms and PTB or LBW . It included both observational studies and case-control studies, but did not pool results in a formal meta-analysis. The authors reported a proportion of studies with significant results for each outcome, concluding that evidence exists for a stronger relationship between preterm depressive symptoms and LBW (and birth weight) than with PTB (and gestational age). However, similar to the other reviews, a few studies were conducted in LMICs .

Another meta-analysis of prospective cohort studies by Ding et al published in 2014 focused on the same outcomes, PTB and LBW, but specifically studied maternal antenatal symptoms of anxiety and anxiety disorder as the exposure of interest [74.] Based on nine and six studies included in their meta-analyses, they reported overall significant increased risks of 1.5 and 1.8 for PTB and LBW births for mothers with prenatal anxiety, respectively. However, no studies from LMICs with PTB as an outcome were included in the meta-analysis. For the studies with LBW as an outcome, the association appeared stronger in those conducted in the two Asian countries included than in the three studies conducted in Europe .

Finally, Bussières et al in a 2015 meta-analysis of 88 studies report associations between maternal prenatal stress – e.g., including subjective stress, anxiety disorder, pregnancy-related anxiety, cortisol, life events, exposure to disaster – and infant birth weight and gestational age . The type of stress appeared to be related to the magnitude of the association, with the strongest effects observed for pregnancy-related anxiety, compared with other stress indicators. Both high-risk samples and studies conducted outside of Europe and North America also showed larger effect sizes. Only four of the 88 studies were from LMICs, indicating an apparent paucity of research from those settings .

Littleton et al 2009 conducted a review of 35 studies focusing birth outcomes and four categories of stressors: number of stressful events, hassles or minor stressors, perceived stress, and appraised stressful life events . Although associations were small in magnitude, the authors found significant associations between psychosocial stress in pregnancy and birth outcomes such as LBW and lower neonatal weight. Like Bussières et al, they noted that the strength of the associations differed depending on the type of stress; for example, minor stressors were more weakly associated with negative outcomes than major stressors . Apart from these reviews, two studies in Africa (Ethiopia and Ghana) found that antenatal depression was associated with higher risk for prolonged labor .

Maternal depression and early parenting practices

In addition to poor maternal–child interactions, aspects of parenting such as health-care seeking and parenting practices may also be affected by maternal depression . The impact of depression on specific parenting practices beyond feeding behaviors is not well described in LMIC settings, and evidence for this pathway largely comes from HIC. Reviews of maternal depression and early parenting practices in HIC identified the domains of infant feeding, sleep routines, preventive health care such as well-child visits and vaccinations, and negative coercive behaviors .

Of parenting practices, breastfeeding has been best studied in relation to depressive symptoms in LMICs. In a study of mothers in rural Pakistan, depressed mothers breastfed for a shorter duration than nondepressed mothers . The authors suggest that this association between depression and truncated breastfeeding may be mediated by perceived insufficient milk . Another study in rural Ethiopia found that depressed mothers delayed initiation of breastfeeding compared with nondepressed mothers . Similarly, in US studies it was found that depressed mothers were less likely to continue breastfeeding 2–4 months after delivery , and an international meta-analysis of mothers largely in high-income countries found that maternal depression was associated with a lower odds of breastfeeding initiation . Another study found that more symptoms of depression, anxiety, and stress in mothers were related to nonresponsive feeding in low-income US children . Although there is mixed evidence regarding whether mothers with depression engage in unsafe parenting practices (e.g., mothers with depression were no more or less likely to put their children to sleep in an unsafe position or fail to lower the water temperature during bathing), practices to encourage child development may be compromised among depressed mothers . After adjusting for demographic variables and socioeconomic factors, depressed mothers in the US were less likely to engage in any of the four examined practices related to child development: setting routines, showing their infant a book daily, talking to their infant while working, and playing with their infant daily . We are not aware of comparable studies in LMICs.

Health-care seeking for the child – both preventive and curative – is another important parenting responsibility that often falls on mothers. In the US, there is evidence of increased use of acute care and decreased vaccination in children under 3 years of depressed mothers . However, a study of children under the age of two in Brazil did not find an association with maternal depressive symptoms and the time of completion of the recommended vaccinations . Other studies showed an increased odds of a “problem visit” to the hospital among infants <5 months of depressed mothers, but no significant association with well-child visits . Although preventive care and vaccination may be particularly relevant in LMICs – where the burden of infectious disease remains high – the lack of well-developed health-care systems in LMICs may preclude the relevance of unnecessary emergency department visits/acute care. Other poor parenting practices associated with maternal depression in HICs , such as reduced odds of car seat usage, storage of medicines, and water heat regulation, may similarly be less relevant among the poor and lower-middle class in many LMIC settings. Finally, it is important to acknowledge that no population within or outside LMICs is homogenous, and there may be different concerns in different communities and cultural groups. In a multi-ethnic longitudinal study of child behavior problems in the US, maternal depression significantly and negatively impacted parental provision of emotional support and cognitive stimulation among white and Latino children; among black/African-American children, there was no relation between maternal depression and parenting practices .

Negative coercive behaviors include a broad range of practices such as hitting, using an unpleasant voice, and expressing hostility towards the child . In a review of 46 studies, Lovejoy and colleagues reported that maternal depression was strongly related to negative coercive behaviors, especially among mothers who were currently depressed , and the link between maternal depression and harsh or physical disciplining has also been confirmed in more recent studies in the US . Another study among US families reported that although there is a strong cross-sectional relation between maternal depression and negative parenting behaviors such as physical assault and psychological aggression, there was no evidence that a change in maternal depression (i.e., a new episode of depression) negatively impacted parenting behavior .

Postpartum depressive symptoms and child growth

Physical growth is considered one of the best indicators of child health status, with stunting (short stature-for-age relative to reference growth curves) reflecting more chronic nutritional and health insults, while other indicators such as wasting reflecting more acute undernutrition or disease . A meta-analytic review of studies in LMICs mostly including children up to the age of two years found maternal depression or depressive symptoms associated with approximately a 40% and 50% higher overall estimated risk of child stunting and underweight, respectively . Inconsistencies across studies may be attributable to either cultural/contextual differences or heterogeneity in the measurement of depressive symptoms and anthropometric indices, age of the child, and time to follow up. For stunting, half of the 12 studies found a significant impact of maternal depression or depressive symptoms, while the other half of the studies did not . Interestingly, all four of the studies in the review that were conducted in South Asia (India, Bangladesh, and Pakistan) found positive effects of postpartum depression/depressive symptoms on stunting . With the exception of one Nigerian study , no other study from the African continent (including Ethiopia, South Africa, and Malawi) confirmed this association with stunting . Among the studies published since that review, Nasreen et al documented a relation between antenatal but not postnatal depressive symptoms and stunting in Bangaldeshi infants at 6–8 months of age .

Regarding underweight during the first year of life, four of the six South Asian studies (conducted in India and Pakistan) in the review showed significant associations with maternal depression or depressive symptoms , although no association was observed in any of the studies from the African continent . Supporting the South Asian findings, a subsequent Bangladeshi study reported that maternal depressive symptoms at 2–3 months postpartum were associated with underweight in infants 6–8 months old . Some of the strongest effect estimates have been observed in South Asia; in Pakistan, a community-based study found approximately a threefold higher risk for underweight and stunting at 6- and 12-month follow-ups (OR range 2.8–3.5) .

The particularly strong associations between maternal depression and child growth failure in South Asia has led some authors to conclude that maternal depression particularly impacts infant growth in Asia compared with in Africa or South America . Reasons for these differences have been put forward; Harpham et al 2005 suggested that women in Asia may be disempowered and face particular social pressure, such that depressed mothers in these contexts have more difficulty in making sure a child is adequately nourished . Stewart conjectured that the “Asian enigma,” of higher rates of malnutrition on the Indian subcontinent than would be predicted from food availability, might be in part due to maternal mental health .