Improvement in Long-Term Breastfeeding for Very Preterm Infants

Abstract

Introduction:

The extensive health benefits of breastfeeding preterm infants for both mother and infant have been widely reported. However, establishing and maintaining breastfeeding for very preterm (VP) infants remain challenging. The aim of this study was to examine changes in breastfeeding of VP infants over time.

Subjects and Methods:

Breastfeeding questionnaires were administered to two cohorts of parents of VP infants (<32 weeks) cared for at the tertiary perinatal or surgical neonatal unit in Western Australia. Of these, 488 infants were included in cohort 1 (C1) (births from January 1, 1990 to June 30, 1992) and 253 in cohort 2 (C2) (from January 1, 2011 to September 14, 2012).

Results:

More mothers (96.8%) initiated breastfeeding in C2 compared with those in C1 (65.6%) (p<0.001). Additionally, 41.4% of mothers in C2 breastfed for more than 6 months, relative to 25.8% in C1 (p<0.001). The benefits of breastfeeding were endorsed by more women in C2 (45.8%) compared with C1 (11.4%) (p<0.01). Reasons for stopping feeding remained largely consistent.

Conclusions:

Significant improvements were evident in the initiation and duration of breastfeeding of the VP infant over time. This improvement was associated with attitudinal shifts in mothers about the benefits of breastfeeding.

Introduction

The extensive health and developmental benefits of breastfeeding, for both mother and infant, have been widely reported in term and preterm populations.^1–4 However, establishing and maintaining breastfeeding for very preterm (VP) infants born at less than 32 weeks of gestation presents unique challenges. Previous research has identified several barriers to establishing lactation, including separation of mother and infant after birth, mother's need to express breastmilk for weeks after birth, low milk supply, high stress and noise of the neonatal units, and troubles transitioning from gastric tube feeding to direct breastfeeding due to small size and fragility of the infants.⁴ A shorter length of stay in hospital has also been associated with higher breastfeeding rates on discharge from hospital in very low birth weight infants.⁵ Despite these challenges, there has been limited research exploring the barriers to establishing and maintaining long-term breastfeeding after VP birth, with a predominant focus on initiation rates and short-term continuation of breastfeeding from discharge to 4 weeks after hospital discharge.⁶ Breastfeeding rates of VP infants on discharge from the hospital varied between 49%⁴ to 97%.⁷

Cross-cultural differences were evidenced between countries, with higher rates reported in Sweden^7,8 relative to low rates reported in the United States.⁶ The two available studies from Sweden (1996–2004)⁸ and Australia (2006/2007)⁹ that reported long-term duration of breastfeeding in preterm and sick infants indicated lower rates of successful initiation and maintenance of breastfeeding in the preterm and sick infant group relative to healthy term babies. Moreover, in the Swedish study there did not appear to be any improvement over time across two birth cohorts from 1996 and 2004, respectively, reported within the same study.⁸ Both of studies included small numbers of VP infants less than 32 weeks (n=32⁹ and n=109⁸).

From 1990 to 2011 several changes in neonatal care had occurred within the tertiary neonatal and surgical unit of Western Australia. These changes included the tertiary perinatal hospital achieving accreditation with the “Baby Friendly Hospital Initiative” in 2004, employment of specialist lactation consultants who are available 7 days a week, written guidelines for staff on supporting breastfeeding of preterm infants, availability of pasteurized donated human milk for VP infants as the preferred option to preterm formula from 2006, and establishment of collaborative research on breastmilk and breastfeeding.

The aims of this audit were twofold: (1) given the changes in neonatal care, to report and compare long-term breastfeeding duration between two large contemporary (2011–2012) and historical (1990–1992) birth cohorts of VP infants; and (2) to explore barriers (and facilitators) to establishing and continuing breastfeeding of the VP infant in terms of mothers' attitudes toward breastfeeding as well as reasons for weaning.

Subjects and Methods

Participants

The inclusion criterion was mothers of infants born at less than 32 weeks gestation. At 12 months corrected age of each VP infant, the mother completes a feeding questionnaire as part of routine follow-up care. Returned completed questionnaires were reviewed for this audit.

Participants included mothers of 536 infants from cohort 1 (C1) (births from January 1, 1990 to June 30, 1992; 30 months). These mothers were enrolled in a total birth cohort study. Cohort 2 (C2) (from January 1, 2011 to September 14, 2012; 21 months) included mothers of 518 infants. These mothers were not enrolled in a study.

All infants were cared for in Western Australia's only tertiary perinatal neonatal intensive care unit (NICU), responsible for the entire geographic region of Western Australia (total land area of 2,529,875 km²). The birth rate for Western Australia in C1 was approximately 25,000 per year,¹⁰ and the NICU had 20 intensive care cots. The birth rate for Western Australia in C2 had increased to approximately 33,000 per year,¹¹ and the number of intensive care cots had increased to 34. All VP infants requiring transfer to the surgical neonatal unit attached to the children's hospital were included in this audit. Details on feeding at discharge from hospital were obtained from the unit's databases.

Procedures

Mothers were mailed a breastfeeding questionnaire when their child was 12 months corrected age and asked to complete and return the questionnaire. No reminders were given. The questionnaire asked about the nature and duration of breastfeeding and attitudes toward breastfeeding. There were no definitions of breastfeeding given to mothers when completing the questionnaire; thus breastfeeding in this audit refers to any breastfeeding. Duration of breastfeeding was measured using a Likert-type rating scale from “Not at all” through to “longer than 6 months.” Questions are shown in Table 1.

Table 1.

Breastfeeding Duration and Attitude Questionnaire

Breastfeeding questionnaire
1. Are you still breastfeeding?
2. How did you feed your baby immediately after leaving the hospital?
a. Breast alone
b. Breast plus formula
c. Formula alone
3. If you breastfed, for how long
a. Less than 1 month
b. 1–2 months
c. 3–6 months
d. more then 6 months
4. What was your reason for stopping breastfeeding? (Up to 3 choices)
a. Inadequate milk supply
b. Baby unsettled after feeds
c. Breastfeeding too tiring
d. Painful nipples
e. Baby refused to breastfeed
f. Too time consuming
g. Maternal illness
h. Breast abscess
i. Return to work
j. Disliked breastfeeding
k. Infant illness
l. Natural weaning
m. Partner/family wanted to feed baby
n. Partner was not keen on me feeding
o. Other….
5. What was your attitude to breastfeeding this baby? (As many choices as necessary)
a. Enjoyed, but would not do it again
b. Disliked
c. Breastfed as thought better for baby
d. Family pressure to breastfeed
e. Enjoyed, would breastfeed other babies
f. Enjoyed, would have breastfed for longer if possible

During admission in the NICU, mothers from C1 were asked to express away from their infant in a separate room of the main nursing floor. C1 mothers were regularly supported by a home visiting nurse during the first postnatal year. In contrast, participating mothers from C2 were encouraged to express cot-side and were supported during their admission by lactation consultants. After discharge, metro-dwelling mothers in C2 were supported by a home visiting nurse for 4 weeks. All children from both cohorts were offered neurodevelopmental follow-up. Rural and remote-dwelling mother were linked to their local Child Health Nurse. Child Health Nurses are registered nurses with qualifications in child and family health who advise on issues of child health, parenting, and child development for children 0–4 years of age.

Neonatal data (gestation, birth weight, sex, and plurality and method of feeding at hospital discharge) were obtained from the hospital neonatal follow-up database.

Ethics approval was granted for this audit (Audit 596QK) from the Ethics Committee of the Women and Newborn Health Service, Perth, WA, Australia.

Statistical analysis

Sample characteristics were examined with descriptive and frequency statistics, and bivariate correlation was undertaken to explore significant relationships between primary outcome variables: breastfeeding duration interval reasons for ceasing breastfeeding, as well as attitudes toward breastfeeding. The two cohorts were divided for analysis of group differences using either independent-samples t test or Mann–Whitney U tests (when nonparametric assumptions were not met). Given the impact of gestational age and birth weight as a marker of maturity, group and individual differences in gestational age and birth weight within each cohort and between cohorts were explored, and where a significant relationship was found, gestational age and birth weight were statistically controlled in subsequent analysis. Statistical analysis was performed with SPSS version 20.0 software (SPSS, Inc., Chicago, IL). All statistical tests were interpreted according to a two-tailed estimate of p values and 95% confidence intervals. A value of p≤0.05 was considered statistically significant. Infants from multiple pregnancies were counted separately.

Results

Sample characteristics

Of the eligible 536 infants from C1, completed questionnaires were available for 432 (79%). Of the eligible 518 infants from C2, completed questionaries were available for 253 (49%). The difference in return rate of questionnaires between the two cohorts could reflect different recruitment procedures, with C1 mothers but not C2 mothers enrolled in a birth cohort study. Discharge information from the neonatal unit's database was retrieved to establish any with-in group differences in the rates of formula-only feeding at discharge for the C2 group, as an indication of problematic sample representation. Similar numbers of infants whose mothers returned questionnaires (27%) and whose mothers did not return questionnaires (35%) were fed with formula only at discharge from hospital (p=0.12).

Independent-samples t test or Mann–Whitney U test of median differences indicated there were no significant differences in sex or plurality between the two cohorts. A small but significant differences in median gestational age (U=58003.0; z=−3.26; p=0.001) and birth weight (U=61490.0; z=−2.03; p=0.042) were found. Gestational age and birth weight were statistically controlled in all subsequent analyses. Sample characteristics for both cohorts are presented in Table 2.

Table 2.

Demographic Sample Characteristics

	Cohort 1 (1990–1992)	Cohort 2 (2011–2012)	Mann–Whitney U (z)	p
Number	432	253
Sex (M/F %)	54/46	53/47	−0.59	NS
Gestational weeks [median (range)]	30 (23–31)	29 (22–31)	−3.26	0.001
Birth weight [median (range)]	1,315 (505–2575)	1,227.5 (430–2,390)	−2.03	0.042
Singletons (%)	73	71	−0.44	NS

F, female; M, male; NS, not significant.

Group differences

There were significantly more mothers who initiated breastfeeding in C2 (96.8%) compared with 65.6% of mothers in C1 (p<0.001). The duration of breastfeeding was longer in C2, with 41.7% of C2 mothers breastfeeding for 6 or more months relative to 25.8% in C1 (p<0.001). The frequency of breastfeeding at each duration interval for both groups is shown in Table 3.

Table 3.

Group Differences in Breastfeeding Initiation and Duration

Duration of breastfeeding	Cohort 1 (1990–1992)	Cohort 2 (2011–2012)	p
Not at all	141 (33.4)	8 (3.2)	<0.001
Less than 1 month	42 (10)	24 (9.5)	0.86
1–2 months	42 (10)	35 (13.9)	0.11
3–6 months	88 (20.8)	80 (31.7)	0.001
More than 6 months	109 (25.8)	104 (41.7)	<0.001

Data are number (%).

Similar results were seen for mothers of multiple infants, with more mothers of multiple infants breastfeeding for longer than 3 months in C2 (78.9%) compared with C1 (45.5%) (p<0.01).

There was a significant difference in mother's attitude toward breastfeeding between the two cohorts. Independent-sample t tests showed a significant group difference, with more women in C2 endorsing the statement that they breastfed as it was better for their baby (45.8%) compared with mothers in C1 (11.4%) (t₇₃₁=−10.09, p<0.01). In contrast, more women in the historical C1 group (31.9%) reported they enjoyed breastfeeding and would have breastfed longer if possible relative to women in C2 (21.3%) (t₇₃₂=4.08, p≤0.01). Similar numbers of mothers in each cohort reported they enjoyed breastfeeding and would breastfeed other babies.

Overall, the most common reasons for stopping breastfeeding were inadequate milk supply (36.3%), baby unsettled after feeds (14.4%), and natural weaning (10.8%). More women (49%) in the C2 identified inadequate milk supply as the reason to discontinue breastfeeding relative to the historical C1 mother (30.4%) (t₇₃₁=−4.08, p<0.01). Fewer women (11.1%) identified an unsettled baby as a reason to stop breastfeeding in C2 than in C1 (16.0%) (t₇₃₁=−2.51, p<0.05). More women in C2 identified return to work (7.9%) as a reason to stop breastfeeding relative to mothers in C1 (1.1%) (t₇₃₄=−4.71, p<0.01). There was a small but marginal difference between the groups for natural weaning as the explanation for breastfeeding: 11.9% in C1 relative to 8.3% in C2 (t₇₃₄=1.97, p<0.05).

Bivariate correlation

The attitudinal factors associated with duration of breastfeeding produced consistent results across the two cohorts. Significant relationships as determined by bivariate correlation analysis between attitudes toward breastfeeding and duration, statistically controlling the influence of gestational age and birth weight where appropriate, are reported for each group and reported in Table 4.

Table 4.

Attitudes Toward Breastfeeding and Its Association with Breastfeeding Duration

	Duration of breastfeeding
	r	p
Cohort 1^a
Gestational age	0.17	0.00
Birth weight	0.20	0.00
Breastfed as thought better for baby	−0.13	<0.01
Enjoyed breastfeeding and would breastfeed again	0.27	0.00
Inadequate milk supply	−0.35	0.00
Natural weaning	0.44	0.00
Partner/family wanted to feed baby	0.11	<0.05
Cohort 2
Gestational age	0.00	NS
Birth weight	−0.02	NS
Breastfed as thought better for baby	0.25	0.00
Enjoyed breastfeeding and would breastfeed again	0.32	0.00
Enjoyed and would have breastfed longer if possible	0.15	<0.05
Inadequate milk supply	−0.18	<0.01
Maternal illness	−0.22	<0.01
Natural weaning	0.22	0.00

Gestational age and birth weight were controlled in bivariate correlation analysis.

NS, not significant.

Duration of breastfeeding was significantly correlated with mothers' belief that breastfeeding was better for their baby, however, in opposing directions. For the earlier C1, the attitude of “breastfed as thought better for baby” was associated with shorter durations of breastfeeding, whereas in the contemporary C2, this was associated with longer duration. There was a significant relationship between duration of breastfeeding and enjoyment with the intention of breastfeeding in the future. Inadequate milk supply was associated with reduced duration of breastfeeding in both groups, and natural weaning was also associated with longer duration in both groups. Maternal illness was negatively associated with duration in the contemporary C2 cohort. Also, partner/family wanting to feed the baby was significantly associated with longer duration of breastfeeding in the first C1 group. There was no significant correlation between multiple pregnancy and breastfeeding duration in either cohort.

Discussion

An improvement in breastfeeding initiation and duration of breastfeeding in VP infants over time has been shown in this audit. This is the first report to suggest an improvement in long-term breastfeeding rates in VP infants. Multiple pregnancies made up 29% of this audit's VP population, and the improvement in both establishing breastfeeding and maintaining breastfeeding for longer than 6 months was also found within this subsample of infants.

Findings from the contemporary C2 group indicate that 41.7% of VP infants were breastfed for more than 6 months, which is a figure approaching that reported for term infants in Western Australia of 45.8%.¹² However, breastfeeding rates for term and preterm infants at 6 months remain below national targets.¹³ The rate of breastfeeding in our contemporary cohort compared favorably with the limited data available from published studies of breastfeeding VP infants. For example, 14% of mothers from Ohio who intended to breastfeed their VP infants were still breastfeeding at 4 months corrected age,¹⁴ whereas 41% of Swedish mothers reported breastfeeding at 6 months postnatal age.⁸

Despite the improvement over time in this report, the duration of breastfeeding rates of VP infants remains lower than for term babies in Western Australia,¹² consistent with other studies.^8,9 A previous report from our unit showed that the 32 infants born at less than 33 weeks of gestation were more likely to cease breastfeeding early.⁹ Similarly, very low-birth-weight (<1,500 g) infants have also been reported to have a shorter duration of breastfeeding than term infants over 2,500 g.¹⁵

Understanding barriers to successful establishment and maintenance of breastfeeding is an important future direction for research. The most common reasons for stopping breastfeeding across both participant cohorts in this study were natural weaning and inadequate milk supply. These findings are consistent with other results reported in the literature.^4,16 Of note is that maternal concern over inadequate milk supply is also one of the most common reasons mothers of term infants wean their babies.¹⁷ Inadequate milk supply may be a proxy for concern about inadequate consumption by the infant and is an important question for future exploration. In addition, exploration by what mothers mean by natural weaning would be helpful.

Further studies are needed to explore both reasons mothers stop breastfeeding and strategies to improve duration of breastfeeding. There is potential to improve duration of breastfeeding in VP infants as one in five mothers in the C2 group reported they would have liked to have breastfed for longer. Accurately identifying whether inadequate milk supply and/or inadequate milk consumption underpins decisions to wean infants from breastfeeding is important to targeting strategies to assist mothers address either of these concerns.

Results from this study were consistent with previous research that has reported the positive association between the mother's positive attitude toward breastfeeding with initiation and longer breastfeeding duration in healthy infants and in sick term and/or preterm infants.^12,18–20 The duration of breastfeeding was positively correlated with the mother's attitudes to breastfeeding in C2. Longer breastfeeding duration was associated with mothers reporting they breastfed as they thought it was better for the baby. Such findings may point to a cultural change at the level of mothers' exposure to breastfeeding information prepreganancy and prenatally and to support within the NICU as well as within the general community as they are discharged and in the oversight care of community nursing. Improvements in breastfeeding of VP infants in the contemporary cohort may be associated with changes in neonatal care at the perinatal hospital and the surgical neonatal unit over the interval and have contributed to a positive cultural shift in pro-breastfeeding beliefs, information, and services. These changes include:

1. The (NICU) perinatal hospital achieving accreditation in the “Baby Friendly Hospital Initiative” in 2004

2. Employment of specialist lactation consultants who are available 7 days a week

3. Introduction of written policy guidelines on breastfeeding preterm infants. The guidelines include first oral feeds by breast with mothers who have indicated their intention to breastfeed, non-nutritive sucking at the breast is encouraged beginning at 28–29 weeks, nipple shields are used to help VP infants with attachment to their mother's breast, double pumping is encouraged, and encouragement for expressing milk cot-side and increased availability of breast pumps for this purpose.

4. Availability of pasteurized donated human milk for VP infants as the preferred option to preterm formula since 2006

5. Establishment of collaborative research on breastmilk and breastfeeding

6. Improved education opportunities for nursing and medical staff on the benefits of breastfeeding

Despite the pleasing results obtained in this audit, results are to be interpreted with caution as the audit was not without limitation. The poor questionnaire return rate in the contemporary group limits the generalizability of the findings. Althugh accountability measures were undertaken to assure the representativeness of the contemporary sample (i.e., similar rates of formula feeding between participant and nonparticipants in this birth cohort), a 50% return rate introduced the potential for sampling bias. For example, women with breastfeeding failure experiences may have provided a different perspective on why women stop breastfeeding, relative to women who have had success. No definition of breastfeeding accompanied the questionnaire, so it is possible that mothers reported breastfeeding as a proxy for the VP infant receiving expressed breastmilk via a bottle. The current study was also not able to control for the impact of socioeconomic influences, which has previously been associated with lower rates of breastfeeding. Other potentially influential factors such as maternal smoking, prior breastfeeding experience, and length of stay or neonatal course were not included in the current audit study.

Nonetheless, the current audit involving two birth cohorts of VP infants, including those infants transferred for surgical care, is the largest to report on the long-term breastfeeding duration in VP infants and the first to show an improvement in the long-term duration of VP infants. It provides a foothold to extend future research in this area.

Conclusions

Improvements in breastfeeding initiation and duration in VP infants were seen over time. This improvement was associated with more mothers endorsing the benefits of breastfeeding.

Footnotes

Acknowledgments

The authors wish to acknowledge the assistance of Sandra Andersen, lactation consultant, and Dr. R. Hagan with this project.

Disclosure Statement

No competing financial interests exist.

References

Section on Breastfeeding. Breastfeeding and the use of human milk. Pediatrics, 2012; 129:e827–e841.

, Chung

, Raman

, et al. Breastfeeding and maternal and infant health outcomes in developed countries. Evid Rep Technol Assess (Full Rep), 2007;(153):1–186.

Underwood

. Human milk for the premature infant. Pediatr Clin North Am, 2013; 60:189–207.

Callen

, Pinelli

. A review of the literature examining the benefits and challenges, incidence and duration, and barriers to breastfeeding in preterm infants. Adv Neonatal Care, 2005; 5:72–88; quiz 89–92.

Kirchner

, Jeitler

, Waldhor

, et al. Long hospitalization is the most important risk factor for early weaning from breast milk in premature babies. Acta Paediatr, 2009; 98:981–984.

Ross

, Browne

. Feeding outcomes in preterm infants after discharge from the neonatal intensive care unit (NICU): A systematic review. Newborn Infant Nurs Rev, 2013; 13:87–93.

Funkquist

, Tuvemo

, Jonsson

, et al. Preterm appropriate for gestational age infants: size at birth explains subsequent growth. Acta Paediatr, 2010; 99:1828–1833.

Akerstrom

, Asplund

, Norman

. Successful breastfeeding after discharge of preterm and sick newborn infants. Acta Paediatr, 2007; 96:1450–1454.

Perrella

, Williams

, Nathan

, et al. Influences on breastfeeding outcomes for healthy term and preterm/sick infants. Breastfeed Med, 2012; 7:255–261.

10.

Castles

. 1993 Births Australia 3301.0. Australian Government Publishing Service, Canberra, 1994:Table 2, p. 5.

11.

, Zeki

, Hilder

, et al. Australia's Mothers and Babies 2011. Perinatal Statistics Series number 28. Catalog number PER 59. Australian Institute of Health and Welfare, Canberra, 2013.

12.

Scott

, Binns

, Oddy

, et al. Predictors of breastfeeding duration: Evidence from a cohort study. Pediatrics, 2006; 117:e646–e655.

13.

Australian Health Minister's Conference 2009. The National Breastfeeding Stratey 2010–2015. Australian Government Department of Health and Aging, Canberra, 2009:9-12.

14.

Furman

, Minich

, Hack

. Correlates of lactation in mothers of very low birth weight infants. Pediatrics, 2002; 109:e57.

15.

Killersreiter

, Grimmer

, Buhrer

, et al. Early cessation of breast milk feeding in very low birthweight infants. Early Hum Dev, 2001; 60:193–205.

16.

Kavanaugh

, Mead

, Meier

, et al. Getting enough: Mothers' concerns about breastfeeding a preterm infant after discharge. J Obstet Gynecol Neonatal Nurs, 1995; 24:23–32.

17.

Neifert

, Bunik

. Overcoming clinical barriers to exclusive breastfeeding. Pediatr Clin North Am, 2013; 60:115–145.

18.

Scott

, Binns

, Graham

, et al. Temporal changes in the determinants of breastfeeding initiation. Birth, 2006; 33:37–45.

19.

Scott

, Shaker

, Reid

. Parental attitudes toward breastfeeding: Their association with feeding outcome at hospital discharge. Birth, 2004; 31:125–131.

20.

Colaizy

, Saftlas

, Morriss

. Maternal intention to breast-feed and breast-feeding outcomes in term and preterm infants: Pregnancy Risk Assessment Monitoring System (PRAMS), 2000–2003. Public Health Nutr, 2012; 15:702–710.