The Effect of Postpartum Depression on Infant Feeding Practice in Sub-Saharan African Countries: Systematic Review and Meta-analysis

Background: Postpartum depression (PND) is a type of mood disorder that affects women after childbirth. Despite the detailed research related to these maternal and infant health outcomes, the relationship between postpartum depression and infant breastfeeding remains ambiguous. Hence, the aim of this study was to assess the effect of maternal postpartum depression on infant feeding practice in Sub-Saharan African Countries. Method: This systematic review and meta-analysis was reported according to the Preferred Reporting Items for Systematic Reviews and Meta-Analysis. PubMed, Google Scholar, Science Direct and Cochrane Library were systematically searched for relevant articles. STATA version 14 was used to calculate the pooled effect size with 95% condence intervals (95% CI) of postpartum depression on infant feeding practice using the DerSimonian and Laird random effects meta-analysis. The heterogeneity and publication bias was assessed by using I 2 test statistics and Egger’s test respectively. Result: Total of 1058 published and unpublished article were retrieved from different data bases. Finally four studies fullled the inclusion criteria for this systematic review and meta- analysis. This meta-analysis found that postpartum depression has no signicant effect on the infant feeding practices (OR = 0.46, 95% CI: 0.18, 1.14). Conclusion: We did not nd sucient evidence to conclude the effect of postpartum depression on infant feeding practice. Thus, the investigators strongly recommend the researchers to conducted randomised control trail studies in Sub-Sahara African countries. keywords were: Postnatal, postpartum, depression, mental disorders and infant feeding keywords


Background
Postpartum depression is a type of depression experienced by a mother following childbirth, typically arising from the combination of hormonal changes, psychological adjustment to motherhood and fatigue (1). Among highly devastating psychological disorders depression is estimated to affect 350 million people worldwide. It is a common mental health disorder which is estimated to affect 10-15% of all mothers in the postpartum period (1)(2)(3).
Studies found that in the rst three months after childbirth, 14.5% of women have an episode of minor or major depression with nearly 40% of these women having experienced symptoms during pregnancy (4,5).
Postpartum depression has a considerable burden on partners and close family members, affecting social and leisure activities and causing nancial problems within the family. Also, postpartum depression has an adverse effect on the marital relationship (6)(7)(8).
There has been signi cant research in developed countries on the risk factors for postpartum depression.
Meta-analyses of these studies have identi ed past history of psychopathology, emotional disturbance during pregnancy, di cult marital relationships, inadequate social support and stressful life events as the primary risk factors for developing postnatal depression (7,9).
Mothers with postpartum depression are more likely to have an unhealthy lifestyle, including poor diet and sleep patterns, compared to mothers free of postnatal depression (10)(11)(12). Women with postnatal depression tend to stop breastfeeding earlier than non-depressed mothers (13)(14)(15).
Despite some evidence of higher risk for depression, most low-income and ethnic minority women remain undiagnosed or untreated for postpartum depression (16). Postpartum depression is a signi cant public health concern with wide-ranging negative consequences for women and their children (17,18).
Postpartum depression is associated with impairment of the mother-infant bond which can result in longer term disruption of the emotional and cognitive development of the infant (7). Mothers with postpartum depression may be less able to interpret and respond appropriately to infant signals; they show more negative than positive emotions toward their infants and are more intrusive in their interactions with their infants (19). In comparison, a study conducted in South Africa showed no clear effects of postnatal depression on infant feeding practice, although postnatal depression at two months was found to be associated with low infant weight at 18 months (20).
A study conducted in Malawi found a signi cant association of postpartum depression with infant feeding practice, All of the studies conducted in South Africa, Malawi and Ethiopia identi ed a high prevalence of postpartum depression (34%, 26% and 33%, respectively), but failed to identify signi cant associations between postpartum depression and infant feeding practice (20)(21)(22).
In Ethiopia, postpartum depression was not associated with adverse infant feeding practice in any aspects (23). Such con icting results from primary studies, coupled with the absence of any systematic reviews focused on the effect of maternal postnatal depression on infant feeding practice in sub-Saharan African countries, indicate the need for clari cation of the effects of maternal postnatal depression on infant feeding practice. A preliminary search for systematic reviews on this topic was performed in PubMed, CINAHL, DARE and PROSPERO. No existing systematic reviews that reported on the effect of maternal postpartum depression on infant feeding practice in sub-Saharan African countries were identi ed in these databases. Therefore, this systematic review seeks out the best available evidence regarding the effect of maternal postnatal depression on infant feeding practice in sub-Saharan African countries.

Search strategy
A three-step search strategy was be used in this review. First, an initial limited search of MEDLINE/PubMed has be undertaken followed by an analysis of the text words contained in the title and abstract, and the index terms used to describe the article. A second search using all identi ed keywords and index terms was undertaken across all included databases. Third, the reference list of all identi ed reports and articles was searched for additional studies. Studies published between 2001 and 2020 were considered for inclusion in this review, because in sub-Saharan Africa, there was no publication on effects of postpartum depression on infant feeding practice in Sub-Sahara Africa before 2001. The databases searched: PubMed, Google Scholar, Science Direct and Cochrane Library. The search for unpublished: MedNar and ProQuest. Initial keywords were: Postnatal, postpartum, depression, mental disorders and infant feeding practice. These keywords were used separately and/or in combination using Boolean operators like "OR" or "AND".

Types of exposure
This review was consider studies that examine the infant feeding practice whose mother has postpartum depression compared with the infant feeding practice of mothers who do not have postpartum depression. postpartum depression is de ned as depression that starts within one month after childbirth and whose symptoms last more than two weeks. We can measure postnatal depression by using the Diagnostic and Statistical Manual of Mental disorders or Edinburgh Post-natal Depression Scale and selfreporting questionnaire-20. This review was exclude studies that have been conducted on mothers with preexisting psychological disorders.

Outcomes
This review considered studies that include the following outcome measures: infant feeding practices.

Types of studies
This review was considered both analytical and observational study designs: prospective and retrospective cohort studies, case control studies and analytical cross-sectional studies reporting association between postpartum depression and infant feeding practice.

Methodological quality
Articles selected for the inclusion was assessed by two independent reviewers for methodological quality before inclusion in the review using standardized critical appraisal instruments of Newcastle Ottawa assessment tool (38,39). Any disagreements that arise between the reviewers were resolved through discussion with a third reviewer.

Data extraction
Data extraction tool was prepared in Microsoft excel spreadsheet. The extracted data includes the speci c details about the interventions, populations, study methods and outcomes of signi cance to the review question and speci c objectives. The authors of primary studies were contacted by email in case there is incomplete information.

Data synthesis
Data was analyzed using STATA version 14. Before conducting the meta-analyses, heterogeneity was assessed using the standard I 2 and visual inspection of forest plot. Owing to the possibility of low power, if there are few studies, we used a signi cance level of P < 0.1 to protect against the possibility of falsely stating that there is no heterogeneity present. Effect sizes expressed as weighted mean differences (for continuous data) and their 95% con dence intervals was calculated. Effect sizes expressed as standardized mean differences and their 95% con dence intervals was calculated using the DerSimonian and Laird method. Moreover, odds ratio and their 95% con dence interval was be calculated.

Characteristics of the Studies
Total of 1085 studies were retrieved from different data bases. (Figure 1) It has been reported according to the preferred reporting items for systematic reviews and meta-analyses (PRISMA) guideline (40). All duplicate articles (n = 33) were removed. From the remaining 1052 articles, 951 articles were excluded because their titles and abstracts were not in line with our inclusion criteria. Then, 137 articles were retrieved for abstracts, 125 articles were excluded after screening abstracts. Lastly, total of 12 full text studies were downloaded and assessed for eligibility criteria. Among accessed full text articles, eight articles were excluded because outcome of interests was not reported. Finally four articles were used for meta-analysis. Three of the studies employed cross-sectional study designs (25)(26)(27) whereas only one study was prospective cohort (24) with study population of 4153 from included studies. The minimum sample size for the included study was 159 (24) whereas the maximum sample size for the included study was 3494 both studies were from South Africa (25). (Table 1) Table 1: Characteristics of studies included in the systematic review and meta-analysis on the effect of postpartum on infant feeding practice in Sub-Sahara Africa, 2001-2020

SN Authors
Year of publication This systematic review and meta-analysis assessed the effect of postpartum depression on the infant feeding practice. Heterogeneity was modern across studies (I 2 = 75.1%, p = 0.007) which enabled us to use a random effects model. Using this method, our meta-analysis found that postpartum depression has no signi cant effect on the infant feeding (OR = 0.46, 95% CI: 0.18, 1.14). (Figure 2)

Heterogeneity and Publication Bias of included studies
We have tried to nd the possible sources of heterogeneity using different statistical techniques.
Publication year and sample size were used as covariates by Univariate meta-regression analysis.
However, none of these variables were statistically signi cant for explaining heterogeneity ( Table 2).
The presence of publication bias was also assessed using funnel plot and Eggers' and Beggs' statistical tests at 5% signi cant level. The funnel plot shows a symmetrical distribution ( Figure 2). And the Egger and Beggs' tests showed no signi cant publication bias with p-values of 0.642 and 0.734 respectively.
Therefore, publication bias is not a problem. The in uence of a single study on the overall meta-analysis estimate was assessed by sensitivity analysis using a random effects model was performed. It revealed that no single study in uenced the overall effect of postpartum depression on the infant feeding practice.

Discussion
By the year 2020, World Health Organization projected that depression will become the second cause of global disease burden (28). One in ve women in low and middle income countries developed a postpartum depression according to reviews conducted in Low and Middle Income Countries (LMICS) and a review conducted in Africa (29,30). Evidences showed that common mental disorders including depression during the postpartum period are more prevalent compared with non-pregnancy periods (31).
A study conducted in Malawi found the effect of postpartum depression with infant feeding practice. All of the studies conducted in South Africa, Malawi and Ethiopia identi ed a high prevalence of postpartum depression (34%, 26% and 33%, respectively), but failed to identify effect of postpartum depression on infant feeding practice (22)(23)(24). This showed that there is con icting result seen for the effect of postpartum depression on infant feeding practice (17,18). Hence, this systematic review and metaanalysis is perhaps the rst of its kind to be conducted at the Sub Sahara Africa to examine the effect of postpartum depression on infant feeding practice.
This systematic review and meta-analysis revealed that postpartum depression has no signi cant effect on the infant feeding practice. This nding is in agreement with individual studies conducted in Republic of Korea (32), South Africa (24), Malawi (22), Ethiopia (23), Malaysia (33), Brazil (34), and Saud Arabia (35) whereas the nding of this study was in contrary to a qualitative systematic review (36), a study conducted in Saud Arabia (37) and a study conducted in Mexico (37). These disagreements could be result of socio demographic and socioeconomic differences between the countries. The other potential explanations for the observed differences might be the use of a different instrument to assess the effect of postpartum depression on the infant feeding practice. The difference of sample size and different study periods may be the additional causes for these variations. Because in the above studies the results were based on the individual study, but this systematic review and meta-analysis were pooled the effect of postpartum depression on the infant feeding practice based on the four studies conducted in different study area and periods. However, a study conducted in Canada showed that the effect of postpartum depression on infant feeding depends on the duration of infant feeding practices (14).

Limitations Of The Study
Maximum efforts have been made to include all published and unpublished articles from sub-Sahara African Countries. Moreover, three out of four included articles in this meta-analysis employed a crosssectional study design. Chicken egg dilemma, therefore, cannot be shown in this review. Finally, we were unable to get studies from most of Sub Sahara African countries and this affects generalizability.

Conclusions And Recommendations
In Sub-Sahara Africa country, the studies conducted to assess effect of postpartum depression on infant feeding practice in Sub Sahara African were low in number. The current systematic review and metaanalysis based on existing studies revealed that postpartum depression has no signi cant effect on infant feeding practice. Thus, the investigators strongly recommend to conduct randomized control trail in Sub-Sahara African countries.

Declarations
Ethics approval and consents to participation: Not applicable Consent for publication: Not applicable Availability of data and material This study was based on literature review of published studies in Sub-Sahara Africa. Anyone who needs to access the data can contact the author concerning the studies included in the analysis. The reference list can also be used to directly access the articles.

Competing interests
The authors declare that they have no competing interests   Funnel plot of on effect of postpartum depression on infant feeding practice in Sub Sahara Africa, 2001-2020