TABLE 1

Studies That Examined Infant-Feeding Outcomes and Depressive Symptomatology

AuthorsStudy DesignSamplePPD OutcomeInfant-Feeding OutcomeResults SummaryMethodologic Comment
Breastfeeding and increased depressive symptomatology
    Alder and Cox11 (1983)Retrospective postal questionnaire 1–2 y after childbirth89 UK women who had taken part in a prospective study regarding PPDClinical depression at 18 moBreastfeeding: “total” (exclusive breastfeeding until 12 wk) or “partial” (introduction of formula before 12 wk)Mothers who “totally” breastfed their infants for at least 12 wk had a significantly higher incidence of PPD than those who “partially” breastfedRetrospective study (risk of recall bias); small sample size
    Alder and Bancroft12 (1988)Descriptive study with assessments early in pregnancy and again 12 and 24 wk after delivery91 UK women from 1 hospitalClinical depression at 12 and 24 wkBreastfeeding: number of weeks infant received nourishment from breast milkBreastfeeding mothers were more depressed at 12 wk but not at 24 wkSmall sample size
Bottle feeding and increased depressive symptomatology
    Astbury et al19 (1994)Longitudinal study using mailed questionnaires 32–36 wk after delivery790 Australian women who had given birth in Victoria during 1 weekEPDS score > 12 at 32–36 wkBreastfeeding: unspecifiedBottle-feeding mothers were significantly more likely to have depressive symptomatologyImprecise definition of breastfeeding
    Green et al16 (2006)Longitudinal study with in-hospital interview and follow-up questionnaires via mail and/or telephone125 United Arab Emirates women from 1 hospital during a single monthEPDS (Arabic version) score > 12 at 12 and 24 wkBreastfeeding unspecifiedMothers who were bottle feeding at 12 wk had significantly higher EPDS scores that mothers who were breastfeedingImprecise definition of breastfeeding; homogeneous sample; 44% of the sample had an EPDS score of >9 at 12 wk after delivery; high attrition rate (32% loss to follow-up at 12 wk and 55% at 24 wk)
    Groer and Morgan17 (2007)Secondary analysis from a larger study using data collected during 1 home visit between 4 and 6 wk after delivery200 US women from 1 university hospitalPOMS at 4–6 wk (top decile of scores were used to classify mothers as depressed)Exclusively or nearly exclusively breastfeeding or formula feedingDepressed mothers were more likely to bottle feed than nondepressed mothers at 4–6 wkImprecise definition of breastfeeding; weak measure of PPD
    Gross et al18 (2002)Population-based study using a mailed questionnaire 8–24 wk after delivery14 609 US women from 3 states (Colorado, New York, and North Carolina)Self-reported “very depressed” between 8 and 24 wkBreastfeeding: unspecifiedWomen who were bottle feeding were significantly more likely to report being “very depressed” than mothers who were breastfeedingLarge population-based sample; retrospective study; poor measure of PPD; imprecise definition of breastfeeding
    Hannah et al13 (1992)Descriptive study using an in-person questionnaire 5 d and a mailed questionnaire 6 wk after delivery217 UK women from 1 hospital between 3 and 5 d after deliveryEPDS score > 12 at 5 d and 6 wkBreastfeeding: unspecifiedBottle feeding was significantly associated with depressive symptomatologyImprecise definition of breastfeeding; EPDS was adapted to mood over the past 5 d instead of 7 d
    Mancini et al15 (2007)Cross-sectional study using an in-person questionnaire 6 wk after delivery755 US women from a collaborative obstetrics and midwifery practice (11 obstetrician and 9 nurse-midwife practices)PDSS score ≥ 80 at 6 wkBreastfeeding: unspecifiedWomen with depressive symptomatology were significantly more likely to be bottle feedingImprecise definition of breastfeeding; the number of women who declined to participate was not tracked
    Warner et al14 (1996)Cross-sectional study using interviews 6–8 wk after delivery2375 UK women from 2 maternity unitsEPDS score > 12 at 6–8 wkBreastfeeding: unspecifiedBottle feeding was identified as 1 of 4 variables associated with depressive symptomatologyImprecise definition of breastfeeding
Breastfeeding and decreased depressive symptomatology
    Abou-Saleh et al25 (1998)Longitudinal study using clinician administered questionnaire at 8 ± 2 wk after delivery34 United Arab Emirates women from 1 hospital in DubaiEPDS score > 11 on day 7 and PSE at 8 wkBreastfeeding: unspecifiedBreastfeeding mothers had significantly lower levels of depressive symptomatology in comparison with bottle-feeding mothersSmall sample size; early assessment of maternal mood; imprecise definition of breastfeeding
    Groer21 (2005)Cross-sectional study using home-visit interviews between 4 and 6 wk after delivery183 US women from physicians' offices or the postpartum unit from 1 hospitalPOMS at 4–6 wkExclusive breastfeeding or exclusive bottle feeding on entry into studyExclusively breastfeeding mothers had significantly lower levels of depressive symptomatology in comparison with those bottle feedingThose in the breastfeeding group were more likely to be older and married and to have a higher income
    Hatton et al22 (2005)Cohort study using a mailed questionnaire 6 wk after delivery and 1 during an out-patient clinic visit 12 wk after delivery377 US women from a large multicenter studyEPDS score > 14 at 6 and 12 wkBreastfeeding: unspecifiedBreastfeeding mothers had lower levels of depressive symptomatology in comparison with bottle-feeding mothers at 6 wk but not at 12 wkLow response to survey (48%); imprecise definition of breastfeeding
    Lane et al23 (1997)Longitudinal study using in-hospital interview and mailed questionnaire 6 wk after delivery289 Irish women from 1 hospitalEPDS score > 12 at 3 d and 6 wkBreastfeeding: unspecifiedBreastfeeding mothers had significantly lower levels of depressive symptomatology in comparison with those bottle feedingImprecise definition of breastfeeding; early assessment of maternal mood
    Mezzacappa et al26 (2002)Within-subjects design assessing maternal mood before and after infant feeding (breast or bottle) with each mother24 US mothers (infants were between 4 and 28 wk old) recruited through advertisingPANAS before and after feeding methodAll mothers breastfeeding and bottle feedingBreastfeeding was associated with a decrease in depressive symptomatology prefeeding to postfeeding, whereas bottle feeding was associated with an increase in depressive symptomatologySmall sample size; no definition of breastfeeding; voluntary participation
    Tammentie et al24 (2002)Cross-sectional study using a mailed questionnaire 6–8 wk after delivery373 Finnish women from 1 maternity hospitalEPDS score > 12 at 8 wkBreastfeeding: unspecifiedBreastfeeding women had significantly lower levels of depressive symptomatology than bottle-feeding womenLow response rate (39%) to survey; imprecise definition of breastfeeding; homogeneous sample of families comprising 2 adults may limit generalizability
    Yonkers et al20 (2001)Multistage screening study using questionnaires and interview; mothers scoring above the threshold for depressive symptoms at 3 wk were given repeat questionnaires at 4 and 5 wk if they continued to score above the threshold802 US women from 4 publicly funded inner-city community maternal health clinicsEPDS score > 11 at 3 wkBreastfeeding: unspecifiedBreastfeeding mothers were less likely to have depressive symptomatology at 3 wkEarly assessment of maternal mood; imprecise definition of breastfeeding; the sample, composed of primarily socioeconomically disadvantaged mothers, may limit generalizability
Breastfeeding initiation
    Green and Murray27 (1994)Longitudinal study using mailed questionnaires antenatally during each trimester and 6 wk after delivery1272 UK women from 1 antenatal clinicEPDS score of >14 antenatally and >12 at 6 wkBreastfeeding: “not at all,” “gave up,” or “still”Depressive symptomatology antenatally decreased breastfeeding initiation and duration
    Pippins et al29 (2006)Longitudinal study using 4 telephone surveys (3 during pregnancy and 1 after delivery)1448 US women from the Project Women and Infants Starting Health (WISH) cohort study in the San Francisco Bay area (California)Short-form CES-D score > 10Breastfeeding initiated (yes/no) and breastfeeding at 1 mo (yes/no)Depressive symptomatology antenatally was not associated with breastfeeding initiation; women with persistent depressive symptomatology (symptoms at 2 time points) were more likely to discontinue breastfeeding in the first week than 4 wk after delivery; depressive symptomatology at 1 time point was not associated with breastfeeding durationImprecise definition of breastfeeding; high rates of breastfeeding initiation and duration in this sample may limit generalizability; 63% of eligible women participated in study
    Seimyr et al28 (2004)Longitudinal study using questionnaires antenatally and 8 and 52 wk after delivery434 pregnant Swedish women from 6 antenatal clinicsEPDS score > 9 antenatally and at 8 and 52 wkBreastfeeding: unspecifiedDepressive symptomatology negatively influenced breastfeeding initiation and duration ratesRaw data not reported; imprecise definition of breastfeeding
Breastfeeding duration
    Bick et al34 (1998)Descriptive study using home-based interview906 UK women returned a mail questionnaire as part of long-term postpartum health studyEPDS score > 12 at a mean of 45 wkBreastfeeding: unspecifiedDepressive symptomatology in the first 12 wk predicted breastfeeding discontinuationRetrospective survey, subject to recall bias; imprecise definition of breastfeeding
    Cooper et al36 (1993)Descriptive study using semistructured interview with 2 cohorts of mothers (Oxford or Cambridge)483 UK women; the Oxford cohort was from an antenatal clinic, and the Cambridge cohort was from a postnatal wardPSE and SPI at 8 wkBreastfeeding: unspecifiedPPD preceded breastfeeding discontinuationSample recruited from 2 different sites; imprecise definition of breastfeeding
    Dunn et al30 (2006)Secondary analysis of a telephone survey 6 wk after delivery526 Canadian women from 4 regional hospitalsEPDS score > 11 at 6 wkBreastfeeding: unspecifiedWomen with depressive symptomatology were more likely to discontinue breastfeedingSecondary analysis of a population-based study; imprecise definition of breastfeeding
    Galler et al40 (1999)Longitudinal study using interviews and questionnaires 7, 12 and 24 wk after delivery93 Barbados women from 1 maternity hospitalSDS at 7 and 24 wkBreastfeeding: unspecifiedDepressive symptomatology 7 wk after delivery was significantly correlated to infant-feeding practices at 7 and 24 wkSmall sample size; imprecise definition of breastfeeding
    Galler et al41 (2006)Longitudinal study using interviews and questionnaires 7, 12, and 24 wk after delivery226 women from Barbados from 1 maternity hospitalSDS at 7 and 24 wkBreastfeeding: unspecifiedDepressive symptomatology at 7 wk was significantly associated with infant-feeding practices at 24 wkImprecise definition of breastfeeding
    Henderson et al37 (2003)Longitudinal study using in-hospital and mailed questionnaires 8, 24, and 52 wk after delivery1745 Australian women from 2 large hospitalsEPDS score > 12 at 8, 24, and 52 wkBreastfeeding: “any,” “full,” “partial,” or “no” breastfeedingEarly discontinuation of breastfeeding was significantly associated with depressive symptomatology; a majority of the mothers reported that their depression preceded breastfeeding cessation
    Jardri et al35 (2006)Methodologic study using questionnaires between the third and fifth days after delivery and an interview at 8 wk427 French women from 1 hospital during 2 periods of 2 moStructured interview (MINI) at 8 wkBreastfeeding: unspecifiedDiscontinuing breastfeeding in the first 4 wk after delivery for nonmedical reasons was identified as 1 of 5 risk markers for PPDImprecise definition of breastfeeding; 64 participants lost to follow-up; early assessment of mood
    Misri et al38 (1997)Retrospective study using mailed questionnaires51 Canadian women from the reproductive psychiatry clinics of 2 hospitalsAll women were suffering from major depressionAll women had stopped breastfeedingDepressive symptomatology preceded breastfeeding discontinuationRetrospective study; small sample size; all women were suffering from depression and had stopped breastfeeding, which may limit generalizability
    McLearn et al31 (2006)Secondary data analysis of families that participated in the National Evaluation of Healthy Steps for Young Children using questionnaires in-hospital and via telephone 8–24 wk after delivery5565 US families from 24 pediatric practices across the countryModified 14-item CES-D score > 10 at 8–16 wkBreastfeeding: unspecifiedMothers with depressive symptomatology were less likely to continue to breastfeed to 8–16 wkImprecise definition of breastfeeding; secondary data analysis
    Papinczak and Turner33 (2000)Longitudinal study using interviews in-hospital and at home at 12 and 24 wk159 Australian women from 1 maternity hospitalUnspecified at 12 and 24 wkBreastfeeding: “exclusive” or “partial”Increased breastfeeding duration was significantly associated with decreased depressive symptomatologyMeasure of depression unspecified
    Taj and Sikander39 (2003)Cross-sectional study using interviews between 8 wk and 2 y after delivery100 Pakistani women whose child had a checkup from the outpatient department of the children's hospitalHADS (Urdu version) at 8 wk to 2 yBreastfeeding: unspecifiedDepressive symptomatology preceded breastfeeding discontinuationRetrospective study; imprecise definition of breastfeeding
    Taveras et al32 (2003)Cohort study using in-person interview 1–2 d after delivery and via telephone at 2 and 12 wk1163 US low-risk women and infants from a group model managed care organizationCES-D score > 15 at 2 wkBreastfeeding: “exclusive” (no supplementing with >12 oz/d); low “partial” or “token” breastfeeding was classified as discontinuationDepressive symptomatology at 2 wk was associated with breastfeeding discontinuation at 12 wkEarly assessment of mood; excluded mothers discharged later than 48 h, which would include mothers-infants with cesarean delivery and/or complications
Breastfeeding exclusivity
    Clifford et al43 (2006)Longitudinal study using mailed questionnaires 1 and 24 wk after delivery856 Canadian women from 2 hospitalsEPDS score > 12 at 1 and 24 wkFull breastfeeding (breast milk is sole source of nutrition) and partial breastfeeding (receipt of nonhuman milk along with breast milk)Depressive symptomatology at 1 week negatively associated with full breastfeeding at 1 week; a similar association was found at 24 wk; however, depressive symptomatology was not significant in multivariate analysisHigh attrition rate (35%); homogenous sample
    McCarter-Spaulding and Horowitz44 (2007)Descriptive study of women with depressive symptoms (EPDS score > 9) at 2–4 wk, who received 3 home visits between 4–8, 10–14, and 14–18 wk after delivery122 US women from 2 hospitals with an EPDS score of >9 compared with a random sample of New England mothersEPDS and BDI-II at 4–8, 10–14, and 14–18 wkBreastfeeding: unspecified (undefined responses were breastfeeding, bottle feeding, or both)Compared with the random sample, the level of exclusive breastfeeding was significantly lower among depressed mothers than the level of combination feedingImprecise definition of breastfeeding; at last assessment, <10% of the sample had an EPDS score of >9, resulting in a very small sample size of depressed mothers
    Thome et al42 (2006)Cross-sectional survey using questionnaires 8–12 wk after deliveryPopulation-based sample of 734 Icelandic women from the Icelandic National RegisterEPDS at 8–12 wkBreastfeeding: “exclusive” (breast milk only with the exception of vitamins or medicine)Exclusively breastfeeding mothers had lower levels of depressive symptomatology in comparison with nonexclusively breastfeeding mothersCross-sectional survey of ∼25% of total births during study period
Breastfeeding difficulties
    Chaudron et al45 (2001)Cohort study using in-home interviews in second trimester and 4 and 16 wk after delivery465 US women from the Wisconsin Maternity Leave and Health ProjectDIS or CES-D score > 15 at 4 and 16 wkBreastfeeding: unspecifiedBreastfeeding did not predict PPD; however, women who worried about breastfeeding were significantly more likely to become depressed than those who did not worryImprecise definition of breastfeeding
    Edhborg et al47 (2005)Qualitative study using a grounded-theory approach22 Swedish women from 1 hospitalEPDS score > 9 at ∼12 wkBreastfeeding: unspecifiedMothers with depressive symptomatology experienced many breastfeeding problemsSmall sample size; imprecise definition of breastfeeding
    Fergerson et al48 (2002)Cohort study using the EPDS to screen for depressive symptoms vs usual care (no screening)37 low-income US women from 1 inner-city hospitalEPDS score > 9 at first postpartum visitBreastfeeding: “exclusive breastfeeding,” “exclusive bottle feeding,” “mixed feeding” (breast and bottle), or attempted breastfeeding but discontinuedDepressive symptomatology was associated with a failed breastfeeding attemptSmall sample size; low participation rate; high-risk sample
    Forman et al49 (2000)Descriptive study using mailed questionnaires 16 wk after delivery528 Danish women from 1 hospital over a 28-mo periodEPDS score > 12 at 16 wkChart review of mothers who “never established” breastfeedingMothers who never established breastfeeding were more likely to experience depressive symptomatologyPotential selection bias resulting from high rate of nonresponders
    Tamminen46 (1988)Pilot study using questionnaires in the early postpartum period119 primiparous Finnish womenBDI in early postpartum periodBreastfeeding: unspecifiedMothers with depressive symptomatology experienced more breastfeeding difficultiesPilot study; imprecise definition of breastfeeding
Breastfeeding self-efficacy
    Dai and Dennis52 (2003)Methodologic study using translated questionnaires in-hospital and by telephone 4 and 8 wk after delivery186 Chinese women from 1 hospitalEPDS at 4 and 8 wkBreastfeeding: “exclusive” “almost exclusive,” “high,” “partial,” “token,” or “bottle feeding”Mothers with depressive symptomatology had lower levels of breastfeeding self-efficacy
    Dennis51 (2003)Longitudinal study using mailed questionnaires 1, 4, and 8 wk after delivery491 Canadian women recruited antenatally from physician or midwife offices or postnatally by public health nurses after discharge from hospitalEPDS at 1, 4, and 8 wkBreastfeeding: “exclusive” “almost exclusive,” “high,” “partial,” “token,” or “bottle feeding”Mothers with depressive symptomatology had lower levels of breastfeeding self-efficacy
Directional effect of infant-feeding outcomes
    Dennis and McQueen53 (2007)Secondary analysis of a longitudinal study using mailed questionnaires 1, 4, and 8 wk after deliveryPopulation-based sample of 594 Canadian women recruited antenatally from physician or midwife offices or postnatally by public health nurses after discharge from hospitalEPDS at 1, 4, and 8 wkBreastfeeding: “exclusive” “almost exclusive,” “high,” “partial,” “token,” or “bottle feeding”No relationship was found between infant-feeding outcomes at 1 wk after delivery and the development of depressive symptomatology; conversely, mothers with depressive symptomatology 1 wk after delivery were significantly more likely at 4 and/or 8 wk to discontinue breastfeeding, be unsatisfied with their infant-feeding method, experience significant breastfeeding problems, and report lower levels of breastfeeding self-efficacy
No relationship
    Cox et al55 (1982)Longitudinal study using interviews completed twice antenatally and twice postnatally at 10 d and 12–20 wk105 UK women attending 1 antenatal clinicSPI during pregnancy and postnatally at 10 d and 12–20 wkBreastfeeding: unspecifiedBreastfeeding did not influence maternal moodImprecise definition of breastfeeding
    Josefsson et al58 (2002)Case-control study using data from standardized medical charts132 Swedish women with depression and 264 controls without depressionEPDS score > 9 at 6–8 and 24 wkBreastfeeding: unspecifiedNo association was found between breastfeeding and depressive symptomatologyImprecise definition of breastfeeding
    Kendell et al54 (1981)Descriptive study using visual analog scales81 UK women from 1 hospitalVisual analog scales daily to 3 wkBreastfeeding: unspecifiedNo significant difference in mood scores between mothers who were breastfeeding and those who were bottle feedingValidity of visual analog scales not provided; early assessment of maternal mood; imprecise definition of breastfeeding
    Lau and Chan59 (2007)Cross-sectional study using questionnaires 2–5 d after delivery1200 Hong Kong women from 1 hospitalEPDS (Chinese version) score > 9 at 2–5 dBreastfeeding, mixed feeding, and artificial feedingEarly depressive symptomatology was not associated with infant-feeding methodEarly assessment of mood; homogenous sample
    McKee et al60 (2004)Longitudinal study using interviews in the third trimester and 2 and 12 wk after delivery174 low-income black and Hispanic women from 3 community primary health care centersBDI at 12 wkBreastfeeding: exclusive, breast and bottle feeding, or exclusive bottle feedingNo relationship was found between breastfeeding practices (partial or exclusive) and depressive symptomatology; depressive symptomatology did not predict breastfeeding continuation33% of eligible women declined participation; high-risk sample
    O'Neill et al57 (1990)Cross-sectional study using questionnaires142 Irish women (recruitment not specified)EPDS score > 12 at 6 wkBreastfeeding: unspecifiedInfant-feeding method did not predict depressive symptomatologyImprecise definition of breastfeeding
    Ramsay et al56 (2002)Longitudinal study using questionnaires at 1, 8, 24, 40, and 56 wk after delivery409 Canadian women from 1 hospitalEPDS score > 12 at 8, 24, 40, and 54 wkBreastfeeding: unspecifiedDepressive symptomatology was not predictive of infant-feeding practicesImprecise definition of breastfeeding
  • BDI indicates Beck Depression Inventory; PDSS, Postpartum Depression Screening Scale; PSE, Present State Examination; POMS, Profile of Mood; PANAS, Positive and Negative Affect Scale; SPI, Standardized Psychiatric Interview; SDS, Zung Depression Scale; CES-D, Center for Epidemiologic Studies Depression Scale; HADS, Hospital Anxiety and Depression Scale; DIS, Diagnostic Interview Schedule.