Impact of Maternal Anxiety on Breastfeeding Outcomes: A Systematic Review

ABSTRACT

Prenatal and postpartum anxiety may impair maternal functioning and disrupt mother–infant behaviors including breastfeeding. The objective of this narrative review is to examine the association between maternal anxiety from pregnancy to 12 mo postpartum and breastfeeding initiation, duration, and exclusivity. Using a combination of Medical Subject Headings terms and text words, relevant studies were identified through PubMed and PsycINFO. Studies that were conducted in high-income countries, assessed anxiety during gestation and/or postpartum through a standardized measure, and evaluated the impact of anxiety on any of the primary outcomes were included. Sixteen studies met the eligibility criteria although they varied greatly in methodological rigor. A negative association between postpartum anxiety and breastfeeding initiation, duration, and exclusivity was suggested. No associations were found between prenatal anxiety and breastfeeding initiation or exclusivity. Evidence is mixed regarding the association between prenatal anxiety and breastfeeding duration. All studies included in the review were of low or very low quality. Although there was consistency in the association between maternal anxiety and breastfeeding outcomes in the included studies, future studies with greater methodological rigor are needed to determine the extent of the relation between prenatal and/or postpartum anxiety and breastfeeding outcomes.

Introduction

Women have a greater risk of experiencing anxious symptomatology than men. The prevalence of anxiety in the US adult population is 18.1% and women are 60% more likely than men to experience an anxiety disorder in their lifetime (1). In addition, ∼7–10% of women in developed countries (2) and 25% of women in developing countries experience anxiety during pregnancy (3). A study that followed German women from 10–12 weeks of gestation to 16 mo postpartum suggests that ∼16% of these women experienced anxiety during this time (4). Although the relatively low prevalence of anxiety during pregnancy could suggest that anxiety is uncommon, its prevalence is similar to or greater than the prevalence of other medical conditions identified by the US CDC as being common and of importance to address during pregnancy (5), including gestational diabetes (6, 7) and hypertension (8–10). Together, these findings suggest that it is important to understand if and how anxiety during pregnancy and postpartum impacts maternal and infant health.

Pregnancy and the postpartum period are associated with many psychosocial stressors, which are known to be associated with early, negative infant health outcomes (11). In addition, becoming a new mother can be especially difficult as it involves taking on a new role and new responsibilities. Experiencing mental health needs can make this time period even more challenging. For example, there are already well-established links between depression and breastfeeding outcomes. A systematic review of 48 studies found that depression during pregnancy predicted a shorter breastfeeding duration but did not impact initiation. Specifically, postpartum depression was a predictor of breastfeeding cessation (12).

Anxiety often becomes apparent during pregnancy and will persist if not treated (13). Postpartum anxiety can impair maternal functioning, cause distress, and may interrupt mother–infant bond formation (13). Poor maternal health outcomes associated with anxiety during pregnancy include an increased likelihood of nausea and vomiting, pre-eclampsia, preterm delivery, and breastfeeding difficulties (13). Breastfeeding has important positive impacts on children's physical and cognitive health outcomes (14–16), and it may be associated with improved maternal sensitivity and maternal–child attachment (17, 18), both of which are robust predictors of children's later developmental outcomes (19). Evidence from observational studies shows that in general a mother's negative affect can impact the developing child in multiple ways (20). Mental health issues such as anxiety and depression may pose barriers to optimal breastfeeding among women who experience the symptoms of these disorders. Indeed, 2 mechanisms by which anxiety influences breastfeeding outcomes have been proposed (Figure 1). In the first pathway, anxiety diminishes maternal self-esteem, negatively impacting mother–child interactions and breastfeeding. In the second pathway, anxiety is related to maternal stress, which can interfere with oxytocin release, impacting the milk ejection reflex and having a physiologically detrimental effect on breastfeeding (21).

FIGURE 1.

Hypothesized mechanisms by which anxiety influences breastfeeding outcomes (21).

This narrative review examines the association between anxiety and breastfeeding behaviors. We reviewed published evidence on the presence of anxiety symptoms from the first trimester of pregnancy to 12 mo postpartum and its impact on breastfeeding initiation, duration, and exclusivity among women in high-income countries. Findings from this review can help inform health care providers on the need to address maternal anxiety in a meaningful way during pregnancy and in the postpartum period, and to identify research gaps to help improve breastfeeding programs and policies.

Current Status of Knowledge

There are several contemporary systematic reviews on maternal depression and breastfeeding (12). By contrast, there are no updated systematic reviews focusing on maternal anxiety and breastfeeding during both the prenatal and postpartum periods (22). This is a major gap given that maternal anxiety is relatively common and prenatal anxiety is a risk factor for postpartum maternal anxiety. Examining evidence that takes into account both prenatal and postpartum periods in the same review is important to generate a more complete picture of the impact of maternal mental health on breastfeeding behaviors. We hypothesized that there would be an inverse association between maternal anxiety and breastfeeding outcomes. Specifically, we expected that more anxiety during pregnancy and in the postpartum period would be associated with decreased initiation, duration, and exclusivity of breastfeeding.

Criteria and search methods for identification of studies

For this review, a search protocol was created using clearly defined inclusion and exclusion criteria (Table 1). Using specific subject heading terms and text terms, 2 databases (PubMed and PsycINFO) were utilized to find appropriate articles through 21 October, 2017. The search terms and strategies for each database can be found in Table 2. After the search, the results were imported into EndNote X7 (Thomson Reuters) to facilitate the screening process. Articles were excluded if they did not meet the eligibility criteria specified in Table 1. Included studies assessed anxiety during gestation and/or postpartum. There were no restrictions placed on study design but as expected, given how challenging it is to conduct anxiety reduction studies in the period surrounding birth, all of the studies were observational. Studies that were not published in English or not conducted in high-income countries, as defined by the World Bank (23), were excluded. The primary outcomes in this review were breastfeeding initiation, duration, and exclusivity.

TABLE 1.

Inclusion and exclusion criteria for review studies

Criteria	Qualifications
Types of study	No restrictions placed on study design. Studies which evaluated the impact of anxiety on breastfeeding and examined anxiety using a standardized measure were included.
Types of participants	Studies were included if the study participants did not suffer from major mental and physical health issues apart from depression and gave birth to healthy infants.
Types of setting	Studies were included if they were conducted in high-income countries, as defined by the World Bank.
Types of interventions	Studies were included if women experienced anxious symptomatology at any point during pregnancy and 12 mo postpartum.
Types of outcome measure	Studies were included if the primary outcome of interest was breastfeeding initiation, breastfeeding duration at 1 mo, 3 mo, or 6 mo, and/or breastfeeding exclusivity at 3 mo or 6 mo.

TABLE 2.

Concept terms used in electronic search

Anxiety	Breastfeeding	Pregnancy
Anxiety	Breastfeed*	Pregnancy
Anxiety disorders	Breast feeding	Pregnant
	Bottle feeding	Maternal
	Infant feeding	Prenatal
	Breastfed	Perinatal
		Postnatal

^*An asterisk after a term is the truncation symbol in PubMed.

Data collection and analysis

The titles and abstracts, identified by the search, were evaluated separately by each of the first authors (CEH and NM) for potential eligibility. These authors then met to compare their findings from the initial screen and any disagreements were resolved by consensus. The full text of articles was reviewed by each of these authors independently to make final decisions about eligibility and extract data. Relevant information, including list of authors, study year, population, study design, exposures, outcomes, results, and conclusions, was extracted independently into an evidence worksheet. Studies were then rated for their methodological quality based on prespecified criteria (see the next subsection: Assessment of Methodological Quality) and designated a quality rating of high, moderate, low, or very low. Information was eventually merged into a single consensus document after a discussion of each article.

Assessment of methodological quality

The Grading of Recommendations, Assessment, Development, and Evaluation (GRADE) approach, recommended by Cochrane, was utilized to evaluate the quality of evidence for each outcome (24). GRADE assesses methodological flaws, the presence or lack of consistency across studies, effectiveness, and generalizability. The first authors independently graded each study and consensus was reached via discussion if there were disagreements. Following GRADE, observational studies were initially graded as low quality. Studies were downgraded to very low quality if they included publication bias, heterogeneity, imprecision (i.e., wide CIs), and design limitations. Studies were upgraded if the magnitude of effect was large, a dose-response gradient was present, and/or direction of plausible bias was present. Study quality upgrading factors were considered in relation to downgrading criteria to assign the final quality grading to each study.

Results of the search

We screened the titles and abstracts of 382 articles, 32 of which were deemed to meet the preliminary criteria (see Table 1). After full-text review of these articles, 16 (21, 25–39) met the full eligibility criteria and were included in the final review (Figure 2). Articles were excluded if they were conducted in low- or middle-income countries (n = 1); featured a population that did not fit the inclusion criteria (n = 1); did not include the exposure and/or ≥1 outcome of interest (n = 7); did not use a standardized measure of anxiety (n = 3); and did not explore the association between anxiety and breastfeeding (n = 4) in the correct direction (i.e., the effect of anxiety on breastfeeding). Because there were multiple outcomes assessed in this review, the studies needed to examine ≥1 relevant outcome to be included. No randomized controlled trials or intervention studies were identified.

FIGURE 2.

Systematic review flowchart description, as recommended by the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) diagram.

Study characteristics

Study designs

Of the 16 studies included in this review, 2 were cross-sectional and the remaining 14 were prospective cohort studies. For the cohort studies, follow-up periods ranged from 1 mo (4 wk) to 18 mo. All 16 studies relied on maternal self-report for the exposure (anxiety) and a combination of maternal self-report and medical records for the outcomes (initiation, duration, and exclusivity). To measure anxiety, 12 studies (21, 25–28, 30–32, 35, 37–39) used the State-Trait Anxiety Inventory (STAI) (40) to measure state and/or trait anxiety, and 2 of these studies (25, 35) also included an additional measure of anxiety (Hamilton-Anxiety Scale and Edinburgh Postpartum Depression Scale—Anxiety subscale). State anxiety is the transitory experience of unpleasant feelings when faced with a threat, such as feeling anxious before a big test or event. Trait anxiety refers to differences between individuals in terms of their disposition or inclination to experience state anxiety when faced with a threat, for example being an “anxious” person (41). The remaining 4 studies used other validated measures of anxiety, including a pregnancy-related anxiety score (n = 2) (29, 34), the tension-anxiety subscale of the Profile of Mood States (n = 1) (33), and the Duke Health Profile (n = 1) (36).

Participants

Eight studies (25, 28–30, 33–35, 39) enrolled women during pregnancy and 8 studies enrolled women shortly after delivery (21, 26, 27, 31, 32, 36–38). Women in the included studies were adults (≥18 y of age) with no major physical or mental health issues and were recruited from clinical settings (hospitals, clinics with prenatal or postpartum services). Four of the 16 studies excluded infants who were not full-term (i.e., born before 37 weeks of gestation) (21, 25, 27, 32). The remaining 12 studies included a combination of preterm and full-term infants, as long as the infants had no significant health problems that would impact breastfeeding. All studies were conducted in high-income countries, including the United States, Canada, Germany, Australia, and Italy. The participants in 15 of the studies were primarily Caucasian, well-educated, married, and middle- or high-income, whereas 1 study (35) investigated anxiety and breastfeeding outcomes in Latina populations in the United States.

Eleven of the studies (27–30, 33–39) included a measure of depressive symptoms, and thus included women who may have been depressed in addition to being anxious. These studies were retained in the analysis because anxiety frequently co-occurs with depression during pregnancy and in the postpartum period (42). Studies that included women with serious mental health issues (e.g., active psychosis, suicidality) or other serious health conditions were not included in this review because their breastfeeding behaviors might have been affected by their pre-existing conditions and bias the results.

Quality of the studies

All of the 16 studies included were graded as having low to very low quality of evidence. One study (28) was downgraded to very low because although it measured both breastfeeding duration and exclusivity, only the results for the association between anxiety and breastfeeding duration were presented, which increases the risk of publication bias. This study was still included in the final review because it met the prespecified criteria and included results for ≥1 outcome of interest. One cross-sectional (32) and 1 prospective cohort study (35) were downgraded to very low owing to a very short follow-up period. The lack of confounder control caused the rating of 2 additional studies (36, 37) to be downgraded to very low.

Main findings

The results are displayed in Table 3 by the outcomes of interest in this review: breastfeeding initiation, breastfeeding duration, and breastfeeding exclusivity. Two of the 16 studies (25, 26) reported on the relation between anxiety and all 3 of our primary outcomes (breastfeeding initiation, duration, and exclusivity); 2 studies (36, 39) reported on initiation and duration; 1 study (31) reported on duration and exclusivity; 1 study (29) reported on initiation only; 6 studies (21, 28, 33–35, 37) reported on duration only; and 4 studies (27, 30, 32, 38) reported on exclusivity only.

TABLE 3.

Summary of studies¹

Authors	GRADE quality	Sample population	Study design	Exposures	Outcome: breastfeeding initiation	Outcome: breastfeeding duration	Outcome: breastfeeding exclusivity
Adedinsewo et al. (25)	Low	255 Canadian women	Cohort	Hamilton Anxiety Scale (18–23 and 24–26 weeks of gestation, 3, 6, and 12 mo postpartum); STAI (18–23 weeks of gestation, 3 and 6 mo postpartum)	Prenatal anxiety not associated with breastfeeding initiation	No association between prenatal anxiety and breastfeeding duration; higher state and trait anxiety at 3 mo associated with 4% and 7% decreased odds of any breastfeeding at 12 mo, respectively	No association between prenatal anxiety and breastfeeding exclusivity at 3 or 6 mo; a 1-point increase in the Hamilton Anxiety Scale score at 3 mo associated with 11% reduced odds of exclusive breastfeeding at 6 mo
Britton (26)	Low	265 American women	Cohort	STAI—State (before hospital discharge, 1 mo postpartum)	Less likely to have initiated breastfeeding after delivery if women experienced high state anxiety at discharge	Women with high state anxiety at hospitalization after delivery and 1 mo postpartum had 4.40 and 5.16 times the odds of terminating breastfeeding at 1 mo postpartum, respectively	61% and 58% reduced odds at hospitalization after delivery and 1 mo postpartum of exclusively breastfeeding at 1 mo postpartum among women with high state anxiety (STAI > 40), respectively
Clifford et al. (27)	Low	856 Canadian women	Cohort	STAI (1 wk and 6 mo postpartum)	Not measured	Not measured	High trait, but not state, anxiety at 1 wk postpartum associated with 43% greater risk of not exclusively breastfeeding at 6 mo
Cooke et al. (28)	Very low; downgraded for high risk of publication bias	365 Australian women	Cohort	STAI—Trait (28–36 weeks of gestation); STAI—State (3 mo postpartum)	Not measured	No association between prenatal anxiety and breastfeeding duration; no association between postpartum state anxiety at 3 mo and breastfeeding duration	Not measured
Fairlie et al. (29)	Low	1436 American women	Cohort	Pregnancy-related anxiety questionnaire (first trimester)	Pregnancy-related anxiety not associated with breastfeeding initiation	Not measured	Not measured
Fisher et al. (30)	Low	791 Australian women	Cohort	STAI—Trait (third trimester); STAI—State (third trimester, 4 mo postpartum)	Not measured	Not measured	Prenatal trait anxiety not associated with breastfeeding exclusivity at discharge nor at 4 mo postpartum
Flaherman et al. (31)	Low	1107 American women	Cohort	STAI—State (before hospital discharge, 2 wk postpartum)	Not measured	Higher state anxiety at 2 wk postpartum associated with lower prevalence of any breastfeeding at 2 mo	Higher trait anxiety associated with increased formula use at 2 wk and 2 mo postpartum
Gagnon et al. (32)	Very low; downgraded for short follow-up period	564 Canadian women	Cross-sectional	STAI (2 wk postpartum)	Not measured	Not measured	Higher anxiety at 2 wk postpartum associated with 61% greater risk of supplementing breast milk with formula during postdelivery hospitalization
Hammarberg et al. (33)	Low	183 Australian women	Cohort	Tension-Anxiety subscale of the Profile of Mood States (first and third trimesters, 3, 8, and 18 mo postpartum)	Not measured	High prenatal anxiety (75th quantile of anxiety scores vs. remaining scores) associated with 157% greater risk of breastfeeding for <6 wk and 124% greater risk of stopping breastfeeding before 8 mo postpartum	Not measured
Kehler et al. (34)	Low	780 Canadian women	Cohort	Symptom Questionnaire Anxiety (during pregnancy)	Not measured	High prenatal anxiety associated with 80% greater risk of early cessation of breastfeeding (breastfeeding for <6 mo) than for those with low prenatal anxiety	Not measured
Lara-Cinisomo et al. (35)	Very low; downgraded for short follow-up time and lack of adequate confounder control	34 Latina American women	Cohort	STAI (4 and 8 wk postpartum); Edinburgh Postpartum Depression Scale—Anxiety subscale (third trimester, 4 and 8 wk postpartum)	Not measured	Significantly higher prenatal anxiety scores among women who stopped breastfeeding before 4 or 8 wk postpartum than among those who did not stop; compared with women who did not, those who stopped breastfeeding at 8 wk postpartum had higher anxiety scores at 4 and 8 wk	Not measured
Papinczak and Turner (36)	Very low; downgraded for lack of adequate confounder control	159 Australian women	Cohort	Duke Health Profile (3 and 6 mo postpartum)	Anxiety scores at 3 mo postpartum not associated with breastfeeding initiation before hospital discharge	Higher anxiety at 3 mo, but not 6 mo, postpartum associated with shorter duration of any breastfeeding	Not measured
Paul et al. (37)	Very low; downgraded for lack of adequate confounder control	1123 American women	Cohort	STAI—State (before discharge, 2 wk, 2 mo, and 6 mo postpartum)	Not measured	Shorter duration of any breastfeeding in the first 6 mo postpartum among women who screened positive for anxiety at postpartum hospitalization than among those who did not screen positive; true for first-time mothers and mothers who delivered via vaginal delivery	Not measured
Tully et al. (38)	Low	105 American women	Cross-sectional	STAI—State (before discharge, 1 mo postpartum)	Not measured	Not measured	Mothers of late preterm infants with higher state anxiety during postpartum hospitalization less likely to exclusively breastfeed during hospitalization; association did not persist at 1 mo postpartum; no associations between anxiety and breastfeeding exclusivity at either time point for mothers of full-term infants
Wallwiener et al. (39)	Low	330 German women	Cohort	STAI (third trimester, 3–4 d postpartum)	Prenatal anxiety not associated with breastfeeding initiation; higher state anxiety scores at 3–4 d postpartum were associated with decreased likelihood of initiating breastfeeding during postpartum hospitalization	Prenatal anxiety not associated with breastfeeding duration at 4 mo postpartum; history of postpartum anxiety associated with ceasing breastfeeding before 4 mo postpartum	Not measured
Zanardo et al. (21)	Low	204 Italian women	Cohort	STAI (before discharge)	Not measured	Higher state, but not trait, anxiety during postpartum hospitalization associated with 7% reduced odds of continuing to breastfeed at 3 mo postpartum	Not measured

¹STAI, State-Trait Anxiety Inventory.

Breastfeeding initiation

Three studies assessed the association between prenatal anxiety and breastfeeding initiation, and 3 studies examined the association between postpartum anxiety and breastfeeding initiation. All studies measuring prenatal anxiety found no association between prenatal anxiety and breastfeeding initiation (25, 29, 39). Of the studies that measured postpartum anxiety and breastfeeding initiation, 2 studies found that state anxiety at hospital discharge postdelivery or 3–4 d postpartum was associated with decreased breastfeeding initiation (26, 39). Another study (36) found no association between postpartum anxiety and breastfeeding initiation. However, in this study anxiety was measured at 3 mo postpartum, and it is unclear if anxiety levels at this time point are as relevant as the anxiety levels postdelivery for an impact on breastfeeding initiation.

Breastfeeding duration

Six studies examined the association between prenatal anxiety and breastfeeding duration, and 9 studies examined the association between postpartum anxiety and breastfeeding duration. The studies measuring breastfeeding duration varied in their follow-up time from 4 wk to 12 mo postpartum.

One prospective cohort study found that high prenatal tension-anxiety was associated with 157% greater risk of breastfeeding for <6 wk, and 124% greater risk of stopping breastfeeding before 8 mo postpartum (33). A study that assessed anxiety as yes/no based on telephone questionnaires found that high prenatal anxiety was associated with 80% greater risk of early cessation of breastfeeding (i.e., breastfeeding for <6 mo) compared with those with low prenatal anxiety (34), and another study found that those who stopped breastfeeding before 4 or 8 wk postpartum had significantly higher prenatal state anxiety scores than those who did not stop breastfeeding at 4 or 8 wk (35). Three studies found no association between prenatal anxiety and breastfeeding duration, with follow-up times ranging from 3 to 6 mo postpartum (25, 28, 39).

Eight of the 9 studies found a significant association between postpartum anxiety and breastfeeding duration. Women with high state anxiety in the maternity ward after delivery and 1 mo postpartum had 4.40 and 5.16 times the odds of terminating breastfeeding at 1 mo postpartum compared with those with low anxiety, respectively (26). Two studies measuring state anxiety found that women with higher state anxiety measured at 2 wk postpartum (31) or 4 and 8 wk postpartum (35) had lower prevalence of any breastfeeding at 2 mo postpartum. Another study found that those who screened positive for state anxiety at postpartum hospitalization had a shorter duration of any breastfeeding in the first 6 mo postpartum than those who did not screen positive for anxiety (37). This was especially true for first-time mothers and mothers who delivered via vaginal delivery (i.e., the association was modified by parity and delivery method). Several studies measured both state and trait anxiety in the postpartum period. Higher state, but not trait, anxiety during postpartum hospitalization was associated with 7% reduced odds of continuing breastfeeding at 3 mo postpartum according to 1 study (21). However, the study with the longest follow-up time period found that higher state and trait anxiety, measured at 3 mo postpartum, were associated with 4% and 7% decreased odds of any breastfeeding at 12 mo, respectively (25). Another study found that higher overall anxiety at 3 mo, but not 6 mo, postpartum was associated with shorter duration of any breastfeeding (36). A study that measured anxiety disorders found that postpartum anxiety disorders were associated with ceasing breastfeeding before 4 mo (39). Finally, 1 study observed no association between state anxiety at 3 mo postpartum and breastfeeding duration (28).

Breastfeeding exclusivity

Two studies explored associations between prenatal anxiety and breastfeeding exclusivity, and 6 studies examined associations between postpartum anxiety and breastfeeding exclusivity. The 2 studies that examined prenatal anxiety found no association between prenatal anxiety and breastfeeding exclusivity (25, 30).

All 6 studies that examined postpartum anxiety and breastfeeding exclusivity found significant, inverse associations between them (25–27, 31, 32, 38); the follow-up times for these studies ranged from a few days postdelivery to 6 mo postpartum. One study found differences in the association between state anxiety and breastfeeding exclusivity for late preterm and full-term infants. Specifically, mothers of late preterm infants with higher state anxiety during postpartum hospitalization were less likely to exclusively breastfeed in the maternity ward although not at 1 mo postpartum, and there were no associations between anxiety and breastfeeding exclusivity for mothers of full-term infants (38). Women with high state anxiety in the maternity ward after delivery and 1 mo postpartum had 61% and 58% reduced odds of exclusively breastfeeding at 1 mo postpartum, respectively (26). Three studies found that higher postpartum trait anxiety was associated with decreased breastfeeding exclusivity. Higher postpartum trait anxiety was associated with increased formula use at 2 wk and 2 mo postpartum (31), and women with higher trait anxiety at 2 wk postpartum had 61% greater risk of supplementing breast milk with formula during their stay in the maternity ward than those with lower anxiety (32). Similarly, high trait, but not state, anxiety at 1 wk postpartum was associated with 43% greater risk of not exclusively breastfeeding at 6 mo (27). One study of general anxiety, measured using the Hamilton Anxiety Scale, observed that a 1-point increase in anxiety scores at 3 mo postpartum was associated with 11% reduced odds of exclusive breastfeeding at 6 mo (25).

Study strengths and limitations

An important strength of the body of evidence is that the great majority of studies used a validated measure of anxiety, such as the STAI for Adults. Other strengths include moderate to large sample sizes in some studies, detailed assessments of infant feeding, and data on a range of potential predictors such as BMI, education, and marriage. All of the included studies were cross-sectional and cohort studies, which are of weaker design than randomized controlled trials. Some of the studies were based on small sample sizes of <100 participants (35, 38) and lacked confounder control despite ascertaining predictors (36, 37). In some cases (36, 39), postpartum anxiety was measured after the outcome (initiation) so it is impossible to draw a connection about the direction of the relation and confounding could be possible. Furthermore, there was a short follow-up period (<8 wk) in 2 studies (32, 35). Generalizability posed a significant limitation in the majority of studies, because the populations tended to be primarily Caucasian, educated, and married women; the results of the studies therefore might not apply more broadly to minority and under-resourced populations.

Summary of results

Although all of the studies were of low or very low quality, which was expected because all were observational, the bulk of the evidence suggests an inverse relation between maternal anxiety and breastfeeding outcomes. This is especially true for postpartum anxiety and breastfeeding outcomes, because the majority of studies that assessed postpartum anxiety found that greater postpartum anxiety was associated with decreased breastfeeding initiation, shorter breastfeeding duration, and decreased breastfeeding exclusivity. The results for prenatal anxiety and breastfeeding outcomes were more ambiguous. There were no clear associations between higher prenatal anxiety and breastfeeding initiation or breastfeeding exclusivity, and there was mixed evidence on the association between high prenatal anxiety and breastfeeding duration. However, strong conclusions cannot be reached at this time owing to the poor quality and lack of generalizability of the included studies.

Discussion

Our review summarizes the findings of studies that have examined the association of prenatal and/or postpartum anxiety with breastfeeding outcomes in high-income countries. Previous reviews examined the associations between either prenatal or postpartum anxiety and breastfeeding outcomes, but these reviews have not examined both time periods together to understand if and when maternal anxiety affects breastfeeding outcomes. Higher levels of postpartum anxiety were associated with lower rates of breastfeeding initiation, shorter duration of any breastfeeding, and shorter periods of exclusive breastfeeding. Despite these consistent negative associations, no strong conclusions can be made owing to the limited methodological rigor of the included studies. Evidence was mixed regarding the association between prenatal anxiety and breastfeeding duration, and no associations were present between prenatal anxiety and breastfeeding initiation and exclusivity. The lack of associations in relation to prenatal anxiety may be due to limited statistical power in several studies (34, 36, 37, 39). The studies also have limited generalizability because the majority only included educated, married, and affluent women. The consistent negative association between postpartum anxiety and breastfeeding outcomes could be due to the effects of anxiety on self-esteem, negatively impacting mother–child interactions. Anxiety could also be related to maternal stress, which can interfere with oxytocin release, impacting the milk ejection reflex and having a physiologically detrimental effect on breastfeeding (Figure 1) (21).

Our study aligns with findings previously reported regarding depression and suboptimal breastfeeding outcomes (12). Women who were considered anxious in the studies identified in this review also may have been depressed. These studies did not examine an anxiety-by-depression interaction, but it is possible that some of the associations found could be attributed in part to the coexistence of depression among women with anxiety. Indeed, a previous systematic review found pregnancy and postpartum depression to be associated with shorter breastfeeding duration (12). Overall, the results of our study indicate that higher levels of postpartum anxiety are associated with poorer breastfeeding outcomes. Our study tentatively suggests anxiety does pose a barrier to optimal breastfeeding, but more rigorous research is needed before concrete conclusions can be reached.

Strengths and limitations of the review

The findings in this review are limited owing to generalizability considerations. The included studies were limited to those conducted in high-income countries and there was no consideration of grey literature or studies not published in English. Many of the women who participated in the studies included in this review were Caucasian, well-educated, and married. It is well established that racial and ethnic minority women, and women with lower education and income levels, have lower rates of any breastfeeding and exclusive breastfeeding, experience high rates of mental health issues, and face more barriers to accessing mental health care (43, 44). This could mean that the relation between maternal anxiety and breastfeeding behaviors is more severe than what was documented in this review. It is also important to acknowledge that even though this review focused on high-income countries, depressed women in low- and middle-income countries have also been found to be more likely to follow suboptimal breastfeeding practices (45). More research is needed to determine the relation between anxiety and breastfeeding, in the context of the additional barriers and unique stressors faced by women in low- and middle-income countries.

The findings of this review are further limited by the low to very low quality of the included studies. Recall or reporting bias may have been present in most studies owing to the extensive use of self-reported data. Several of the studies lacked statistical power, and although the majority of studies utilized the STAI, they did not consistently use established clinical thresholds to assess the presence or absence of anxiety. More rigorous, well-designed observational studies, especially prospective cohort studies and intervention studies (i.e., impact of antenatal and postnatal maternal depression reduction interventions on breastfeeding outcomes), that address the methodological limitations of existing studies could further clarify the association between anxiety and breastfeeding outcomes and should be the focus of future research.

Implications of key findings and next steps

Optimal breastfeeding is essential for an infant's healthy development. Given the limitations of observational studies, definitive conclusions regarding the effect of anxiety on breastfeeding outcomes cannot be drawn from this review. The consistency of evidence in our review calls for future studies to examine these associations with greater methodological rigor so that more definitive conclusions can be drawn. Furthermore, studies should also examine the impact of anxiety on breastfeeding among diverse and at-risk populations using standardized methods to assess anxiety and breastfeeding outcomes.

Notes

AKRV is supported by a Yale School of Public Health doctoral fellowship.

Author disclosures: CEH, NM, AKRV, and RP-E, no conflicts of interest.

CEH and NM contributed equally to this work and share first co-authorship.