Objective. To examine factors that predict the initiation of expressed milk feedings and the transition to direct breastfeedings among mothers of very low birth weight (VLBW) infants.
Methods. The sample consists of 361 mother-infant pairs enrolled in a follow-up study of children aged 6 to 8 years who were born weighing <1501 g in 1 of 5 hospitals between 1991–1993. Chart review at birth provided data on neonatal characteristics and demographic factors at delivery were obtained by postpartum maternal interview. Information regarding infant feeding practices was obtained at follow-up.
Results. In this study, 60% of mothers initiated expressed milk feedings for their VLBW infants. However, the duration of these feedings was brief with 52% of infants receiving 1 to 3 months or less of human milk feedings. Greater educational attainment, private insurance, and breastfeeding experience were each independently associated with the decision to provide expressed milk feedings. Only 27% of mothers reported directly breastfeeding their VLBW infants. The transition from expressed milk feedings to direct breastfeedings was positively associated with sociodemographic factors including maternal age, insurance status, and breastfeeding experience as well as the length of hospitalization, an indicator of infant health.
Conclusions. Sociodemographic factors were associated with both the decision to initiate expressed milk feedings and the transition to direct breastfeedings. However, factors relating to infant health only influenced the transition to direct breastfeedings. Intervention programs need to consider the sociodemographic factors that influence infant feeding decisions as well as specific challenges encountered by mothers of VLBW infants.
Breastfeeding is the optimal method of infant feeding with numerous nutritional and immunologic advantages.1,2 National breastfeeding goals have been established to increase the proportion of mothers initiating breastfeeding to at least 75% and to maintain at least 50% of infants breastfeeding at 6 months.3,4 In 1997, the American Academy of Pediatrics extended its recommendation of breastfeeding to include premature infants.5 Very low birth weight (VLBW) infants represent a vulnerable population at increased risk of neonatal and neurodevelopmental impairments that are far less likely to receive the benefits of breastfeedings compared with healthy term infants.6–9 The low incidence and duration of breastfeeding among this high-risk population is unfortunate as human milk feedings convey advantages specific to the premature or low birth weight infant including more rapid gastric emptying, improved fat absorption, and reduced risk of necrotizing enterocolitis.10,11
Sociodemographic and attitudinal determinants of breastfeeding initiation and duration among mothers of healthy term infants have been investigated in numerous studies.12–17 Fewer studies have focused on the challenges faced by mothers of VLBW infants.7,18,19 Most VLBW infants cannot be directly breastfed at birth, requiring mothers to first initiate milk expression; mothers delivering prematurely are more likely to experience delivery complications; and they face significant stress regarding their infant’s health and the procedures of the neonatal intensive care unit (NICU).20,21 Potential institutional barriers include inadequate or contradictory information regarding the benefits of human milk feedings for VLBW infants; lack of consistent advice and support from health care professionals regarding the initiation and maintenance of milk expression and the transition to direct breastfeedings; and difficulty securing appropriate equipment and supplies to express milk.
A greater understanding of the determinants of breastfeeding among mothers of VLBW infants may be achieved by examining the decision-making process as distinct stages. Mothers of VLBW infants must first decide whether to initiate expressed milk feedings. A second critical juncture occurs with the transition from expressed milk feedings to direct breastfeedings. A final decision relates to the duration of human milk feedings (eg, the decision to continue milk expression or direct breastfeedings). Studies of breastfeeding among VLBW infants have recruited small numbers of mothers from a single medical center, limiting the generalizability of the findings.7,18,19,22,23With a large sample of VLBW infants enrolled in a multicenter study, this study investigates the influence of sociodemographic and neonatal factors on the initiation of expressed milk feedings, the transition to direct breastfeedings, and the duration of human milk feedings. Identification of specific barriers may help focus efforts to help mothers who may be interested in providing human milk feedings for their VLBW infants.
We conducted a retrospective analysis of the determinants of human milk feedings among mothers of VLBW infants participating in a follow-up study designed to assess neuropsychological outcomes at age 6 to 8 years. The original cohort, a multicenter study of 1607 infants weighing <1501 g born between January 1991 and December 1993 in 1 of 5 hospitals in New York, New Jersey, and Massachusetts, was assembled to investigate the epidemiology of brain injuries associated with VLBW and has been described in previous articles.24–26 Data collection is ongoing with a target sample size of 796 children. A total of 444 mothers had completed questionnaires by September 2001 and were eligible for inclusion in this analysis. After selection of a single infant for pregnancies that resulted in a multiple birth, 361 mother-infant pairs remained.
At baseline, maternal demographic and delivery information was collected from postpartum maternal interviews. Infant birth and neonatal data were abstracted from obstetric and neonatal charts. At follow-up, information regarding human milk feedings was obtained by parent questionnaires. The majority of questionnaires (95%) were completed by the participant’s biological mother. Mothers reported whether their child received expressed milk feedings during hospitalization and whether their child was directly fed from the breast. The duration of human milk feedings was recorded in five categories (<1 week, 1–4 weeks, 1–3 months, 4–6 months, >6 months) and mothers specified whether they had breastfed other children. Other breastfeeding experience may have included the same birth sibling (in the case of multiple births), subsequent breastfeeding of younger siblings born after the participating child or previous breastfeeding experience of older siblings. Also during the follow-up interview, mothers completed the Peabody Picture Verbal Test as an estimate of verbal intelligence (IQ).27
Group comparisons were conducted with t tests for continuous variables and χ2 tests for categorical data. To examine duration of human milk feedings, expressed milk feedings were divided into 2 categories, less than or more than 4 weeks, and direct breastfeedings were dichotomized less than or more than 3 months for the univariate and multivariate analyses. Logistic regression was used to produce univariate odds ratios and 95% confidence intervals. Multivariate logistic regression was employed to determine which factors were independently associated with the decision to provide expressed milk feedings and the transition to direct breastfeedings. Variables that were related to the outcome (P < .1) were added in a forward stepwise fashion. The final model retained variables that, when excluded, significantly changed the χ2 statistic with the relevant degrees of freedom.
The majority of mothers of VLBW infants in this cohort initiated human milk feedings. Of the 361 mother-infant pairs, 215 women (60%) provided expressed milk feedings for their VLBW infants (Fig 1). Less than half (43%) of the women who initiated milk expression progressed to direct breastfeedings. Two infants were directly breastfed without first receiving expressed milk feedings and 2 mothers who directly breastfed their infants were uncertain whether expressed milk feedings were given during hospitalization. A total of 97 mothers (27% of cohort) provided direct breastfeedings for their VLBW infants.
Although more than half of the VLBW infants received human milk, these feedings were brief. Thirty percent of infants received human milk feedings for 1 month or less (Table 1). The duration of human milk feedings was markedly shorter among infants who received only expressed milk feedings compared with infants who transitioned to direct breastfeedings. After 4 months postpartum, expressed milk feedings were received by <10% of infants while breastfeedings continued for 72% of infants who progressed to direct breastfeedings. In total, 47 infants (22%) received either expressed milk feedings or direct breastfeedings beyond 6 months. There were 228 women who reported having children other than the study participant and 60% of these mothers reported breastfeeding at least 1 other child.
Decision to Provide Expressed Milk Feedings
Factors associated with socioeconomic advantage were positively associated with the decision to initiate expressed milk feedings (Table 2). Mothers who provided expressed milk feedings for their VLBW infants tended to be older, white, married, nonsmokers, high school graduates, and to have higher verbal IQs and private insurance. Other breastfeeding experience was strongly associated with the decision to provide expressed milk feedings. Sixty-three percent of women who breastfed other children provided expressed milk feedings for their VLBW infants compared with 22% of women who did not breastfeed other children. Factors relating to delivery or neonatal health, eg, mode of delivery, gender, multiple birth, birth weight, or gestational age, were not significantly associated with the decision to provide expressed milk. None of the infant factors were associated with the initiation of milk expression including gender, birth weight, or gestational age. In the multivariate model, maternal education, other breastfeeding experience, and presence of private insurance were each independently associated with the initiation of expressed milk feedings (Table 2).
Transition to Direct Breastfeedings
Many of the same sociodemographic characteristics associated with the decision to provide expressed milk feedings also influenced the transition from expressed milk feedings to direct breastfeedings in the univariate analysis (Table 3). Maternal age, education, verbal IQ, marital status, and presence of insurance were also positively associated with direct breastfeedings. Mothers who were nonwhite, smokers, and lacked other breastfeeding experience were less likely to transition their infants to direct breastfeedings.
Additionally, the transition to direct breastfeedings was influenced by the hospital center and the health status of the infant. A lower proportion of mothers delivering at hospital center C initiated expressed milk feedings and significantly fewer infants received direct breastfeedings. Infants who progressed to direct breastfeedings had higher birth weights, more advanced gestational ages and shorter hospital stays. Infants who received direct breastfeedings spent an average of 51 days in the hospital compared with 69 days for infants who received only expressed milk feedings.
In the multivariate model, maternal age, insurance status, breastfeeding experience, and length of hospital stay were each independently associated with the transition to direct breastfeedings. Adjusting for maternal age, insurance status, and breastfeeding experience, each additional week of hospitalization reduced the odds of an infant transitioning to direct breastfeeding by 14%.
Duration of Human Milk Feedings
There were no consistent differences in any of the sociodemographic or neonatal factors between women who discontinued milk expression within 4 weeks compared with women who continued to provide expressed milk feedings beyond the first month. Similarly, no differences were observed between the mother-infant dyads who were breastfeeding for less than or more than 3 months (data not shown). The statistical power to detect differences was limited because of small numbers in each category. Further, the categorization of the original data and the resulting cut-offs distinguishing short and long duration for expressed milk feedings and direct breastfeedings may obscure differences if present.
The majority of mothers of VLBW infants in this cohort initiated human milk feedings. A higher proportion of women initiated expressed milk feedings (60%) in this study compared with the proportion of mothers delivering at a midwestern hospital (47%)18 but comparison of breastfeeding rates across studies is problematic because of variation across geographic location and time. In a 1995 nationwide survey, 60% of mothers reported breastfeeding their term infants and 48% of mothers of LBW infants initiated breastfeeding.6 Women who reside in New England are more likely to initiate breastfeeding than women from southern states and the highest prevalence of breastfeeding is reported among women from western states.6 Studies among VLBW infants are further complicated by small sample sizes from a single institution that may include programs influencing infant feeding patterns. The VLBW infants in this study were born between 1991–1993, a period when national breastfeeding rates were increasing.6 Several factors including the Baby Friendly Hospital Initiative, which began in 199128 and efforts to promote breastfeeding through the Special Supplemental Nutrition Program for Women, Infants, and Children29 may also increase the rate of breastfeeding within our study population.
In this study, low maternal education and lack of private insurance were more common among women who did not commence expressed milk feedings. Maternal age and measures of economic advantage such as education, verbal IQ, marital status, and family income have been frequently associated with increased breastfeeding initiation among mothers of term infants.14 These findings suggest that public health efforts to increase knowledge, feasibility, and social acceptability of breastfeeding, targeting groups with low breastfeeding rates such as women with lower educational and economic backgrounds, may also improve uptake of breastfeeding among mothers of VLBW infants.
Less than half (43%) of the mothers who provided expressed milk feedings went on to directly breastfed their VLBW infants. Another study reported that 8 (20%) of 39 women who were expressing milk for their VLBW infants transitioned to direct breastfeedings and did not observe any differences between these mothers and those who continued to express milk.18 The study stipulated that breastfeedings must provide half of the infant’s total daily feeding for >1 month for the transition to be considered successful. In contrast, our study included all women who directly breastfed their VLBW infants regardless of the proportion or duration of breastfeedings. Further, both sociodemographic and neonatal factors influenced the transition to direct breastfeedings in this cohort of VLBW infants. Maternal age, private insurance, breastfeeding experience, and shorter hospital stays were positively associated with the transition to direct breastfeedings. Duration of hospitalization likely reflects the overall health of the infant and the morbidity encountered during the neonatal course. Inclusion in a multiple birth did not appear to affect either the initiation of expressed milk feedings, the transition to direct breastfeedings, or the duration of human milk feedings. Infants from multiple births in our cohort had higher birth weights, more mature gestational ages and shorter hospital stays as well as sociodemographic advantages that contributed to their unexpectedly high breastfeeding rates which did not differ from singleton births.
In examining the results of the current study, a number of methodologic issues should be considered. Information regarding the initiation and duration of breast milk feedings was reported by mothers 6 to 8 years after the birth of their VLBW infants. Maternal recall of major infant feeding events has been examined after intervals up to 15 years.30–32 Recall of breastfeeding initiation was in agreement with written records for 85% of mothers interviewed after 14 to 15 years31 but the recall of breastfeeding duration was less precise with 37% to 79% of mothers reporting accurately within 1 month.31,32 Misclassification of breastfeeding duration may have contributed to this study’s inconclusive results regarding predictors for the duration of expressed milk and direct breastfeedings. Nearly all respondents were the biological mothers of the VLBW infants and none of them reported that they were did not know whether they directly breastfeed their infant. However, 3 biological mothers were uncertain whether their infant received expressed milk feedings. Information about the proportion of expressed milk or direct breastfeedings relative to other feedings was not available but a recent study that recorded daily nutritional intake during hospitalization did not observe any effect of age at first or full enteral feed or age of first bottle or breastfeed on the proportion of infants whose mothers continued lactation beyond 40 weeks.19 Further, although the specific breastfeeding policies and practices at each hospital center were not known, the cohort was drawn from 5 different hospitals located in 2 major cities and a suburban site. Altogether, the subjects represented a more diverse population than previously reported on. VLBW infants selected for follow-up were at particular risk for neurodevelopmental disabilities and they may have experienced increased morbidity during their neonatal course compared with other VLBW infants. Given that our data indicated that infant characteristics did not influence initiation of expressed milk feeding, these factors are most influential with the transition to direct breastfeeding. As survival of extremely premature infants continues to improve, infants at high risk will constitute an increasing proportion of VLBW infants.
Mothers need to receive information describing the benefits of human milk for their VLBW infants along with the practical means to help them express breast milk until their child is able to directly breastfeed. It is difficult to sustain sufficient milk volumes for extended periods without direct breast stimulation33,34 and an electric breast pump that allows for frequent and simultaneous emptying of both breasts may help maintain long-term milk expression.35 To increase the likelihood of transition to direct breastfeedings, health professionals need to provide support for long-term pumping as well as specific advice regarding the timing and progression to direct breastfeedings. Manipulating the volume of milk in the breast through partial or complete milk expression may permit a VLBW infant to prepare for direct breastfeedings even before sucking and swallowing movements are fully coordinated.36
The decision to breastfeed is part of a broad cultural context. The sociodemographic factors identified in this study, eg, maternal education, maternal verbal IQ, and insurance status, represent indirect predictors of breastfeeding behavior. Knowledge of specific factors associated with initiation of expressed milk feedings and the transition to direct breastfeedings may facilitate identification of women who are more favorably inclined to breastfeed as well as women who may benefit from additional encouragement and support.
Mothers of VLBW infants are confronted with many challenges that influence their infant feeding decision. Many sociodemographic factors are common to mothers of term infants and can be addressed through public health initiatives aimed at the general population and through efforts to educate women during prenatal care. However, other challenges are unique to this special population of high-risk infants and these factors are most influential with respect to the transition to direct breastfeedings. The continued contact necessitated by the prolonged hospitalization can be viewed as an opportunity for health care providers to provide mothers with additional support and information while their infants remain in the NICU. A women’s breastfeeding experience with her VLBW infants may affect subsequent infant feeding decisions. A public health perspective can promote a supportive breastfeeding environment both in the NICU and in the community.
This work was supported in part by National Institutes of Health grant NS 36285-01.
- Received May 9, 2002.
- Accepted September 16, 2002.
- Address correspondence to Louise Kuhn, PhD, Sergievsky Center, PH-19, Box 16, Columbia University, 630 168th St, New York, NY 10032. E-mail:
Reprints not available.
- ↵US Department of Health and Human Services. HHS Blueprint for Action on Breastfeeding. Washington, DC: US Department of Health and Human Services; 2000
- ↵US Department of Health and Human Services. Developing Objectives for Healthy People 2010. Washington, DC: US Department of Health and Human Services; 1997
- ↵American Academy of Pediatrics, Work Group on Breastfeeding. Breastfeeding and the use of human milk. Pediatrics1997;100 :1035– 1039
- ↵Ryan AS. The resurgence of breastfeeding in the United States. Pediatrics.1997;99(4) . Available at: http://www.pediatrics.org/cgi/content/full/99/4/e12
- ↵Forste R, Weiss J, Lippincott E. The decision to breastfeed in the United States: does race matter? Pediatrics.2001;108 :291– 296
- Arora S, McJunkin C, Wehrer J, Kuhn P. Major factors influencing breastfeeding rates: mother’s perception of father’s attitude and milk supply. Pediatrics.2000;106(5) . Available at: http://www.pediatrics.org/cgi/content/full/106/5/e67
- ↵Ertem IO, Votto N, Leventhal JM. The timing and predictors of the early termination of breastfeeding. Pediatrics.2001;107 :543– 548
- ↵Furman L, Minich N, Hack M. Correlates of lactation in mothers of very low birth weight infants. Pediatrics.2002;109(4) . Available at: http://www.pediatrics.org/cgi/content/full/109/4/e57
- ↵Affonso DD, Hurst I, Mayberry LJ, Haller L, Yost K, Lynch ME. Stressors reported by mothers of hospitalized premature infants. Neonat Netw.1992;11 :63– 70
- Leviton A, Paneth N, Susser M, et al. Maternal receipt of magnesium sulfate does not seem to reduce the risk of neonatal white matter damage. Pediatrics.1997;99(4) . Available at: http://www.pediatrics.org/cgi/content/full/99/4/e2
- ↵Dunn LM, Dunn LM. Peabody Picture Vocabulary Test. 3rd ed. Circle Pines, MN: American Guidance Service; 1997
- ↵Launer LJ, Forman MR, Hindt GL. Maternal recall of infant feeding events is accurate. J Epidemiol Commun Health.1992;46 :203– 206
- ↵Tienboon P, Rutishauser IHE, Wahlqvist ML. Maternal recall of infant feeding practices after an interval of 14 to 15 years. Aust J Nutr Dietetics.1994;51 :25– 27
- ↵Meier P, Engstrom JL, Mangurten HH, Estrada E, Zimmerman B, Kopparthi R. Breastfeeding support services in the neonatal intensive care unit. J Obstet Gynecol Neonat Nurs.1992;22 :338– 346
- Copyright © 2003 by the American Academy of Pediatrics