Table 4.

Impact of expressing on breastfeeding outcomes

Author, year, country	Design	Location, participants, year of study and recruitment	Study aims and outcome measures	Results	Strengths/Limitations
Chapman et al. 2001 USA [45]	RCT	Hartford, Connecticut	Effects of expressing before the onset of lactation :	- No significant difference in milk transfer or breastfeeding duration between women who expressed breast milk and those who did not.	Only women who had a caesarean section
		n = 60	- on early milk transfer	- Primiparous women in pumping group breastfed for 5 months less than those in control group but this finding was not statistically significant.	Study underpowered for primiparous women
		1997–1998	- on subsequent breastfeeding duration
		Convenience sample 8–24 hours post Caesarean Section
Schwartz et al. 2002 USA [42]	Prospective cohort	Detroit, Ann Arbor and Southfield, Michigan and Omaha, Nebraska	- Determine demographic, behavioural and clinical factors associated with weaning from breast in the first 12 weeks	- Michigan women (n = 711) who expressed breast milk were 3 times more likely to wean than those who didn’t (Hazard Ratio: 3.0 95% CI 1.3,6.7)	Large study
		n = 946		- Nebraska women (n = 235) showed no association between pumping and weaning (HR: 0.6, 95% CI 0.3,1.5)	Only measured to 12 weeks
		1994–1998			Non-representative sample
		Recruitment:			Michigan group were recruited from an alternative birthing centre and were significantly more likely to be older than 30 years, have a bachelor’s degree, have 3 or more children and have had a vaginal birth
		Michigan - at birth centre orientation
		Nebraska - on maternity leave application to large company
Ortiz 2004 USA [37]	Clinical audit	Burbank, California	- Duration of breast milk feeding related to a range of employee chosen lactation support options	- 98% (452/ 462) breastfeeding initiation	Large study over 4.5 years
		n = 462		- 74% (246/332) expressed milk until infant at least 6 months	Limited differentiation between breastfeeding and expressing / breast milk feeding
		1993–1999		- 24% (81/332) expressed milk until infant at least 12 months	No information re any other infant feeding/exclusivity of breast milk feeding
		Antenatal recruitment in workplace		- Mean age of infants at maternal cessation of pumping at work 6.3 months	No consideration of options in the workplace to breastfeed at the breast
Geraghty et al. 2005 USA [1]	Retrospective cohort	Cincinnati, Ohio	Measure breast pump use	Of breast milk feeding mothers:	Large sample size
		n = 346	Identify relationships between breast pump use and:	- 10% (24/346) breastfed exclusively for a minimum of 6 months	Breastfeeding / breast milk feeding clearly differentiated
		2002	- singleton vs. multiple pregnancy	- 16% (55/346) breastfed exclusively for duration of their breast milk feeding	Periodic reports re. proportion of expressing versus breastfeeding (at 1 day, 3 days, 2 weeks and monthly until 6 months)
		Random selection Postal recruitment when infants were between 2 and 3 years old	- gestation at birth	- 77% (182/236) expressed at some time in first 6 months	Retrospective data, possible recall bias as participants were recruited 2 or more years post birth
			- breastfeeding outcomes	- 59% (140/236 ) ceased breast milk feeding by 6 months
				Of the 140 women who had ceased breast milk feeding by 6 months, at the time point just prior to exclusive formula feeding:
				- 76% (106/140) were either expressing exclusively or combining expressing with breastfeeding
				- 24% (34/140) were breastfeeding
				Early breastfeeding associated with a longer duration of breast milk feeding
Win et al. 2006 Australia [55]	Prospective cohort	Perth, Western Australia	- Investigate association between breast milk expression and breastfeeding duration	- Mothers who expressed at least once more likely to be breastfeeding at 6 months (RR: 0.71, 95% CI 0.52,0.98)	Prospective design assisting recall
		PIFS II			Ever “expressed” / “any” breastfeeding
		n = 587			? lower socio economic bias
		2002–03			No account of frequency of expressing
		Recruited in hospital at birth.
Meehan et al. 2008 USA [43]	Quasi-experimental	Los Angeles, California	- Evaluation of program to facilitate breastfeeding for low income mothers	- Electric pump loan associated with more breastfeeding at 6 months. Mothers loaned a breast pump 5.5 times more likely to than those who hadn’t received one to not have requested formula by 6 months	Limited reliability of proxy measurement to assess breast milk feeding prevalence or duration
		n = 208	- Maternal request for formula from WIC program used as proxy measurement to give indication of partial breastfeeding	(OR: 5.5, 95% CI 2.0,15.1)	No differentiation between breastfeeding and breast milk feeding
		2001
		Breast pump loan program for low income Women with Children (WIC) recipients
Fein 2008 USA [41]	Prospective cohort	National	- Examine strategies used to combine work and breastfeeding	Median duration of breast milk feeding associated with workplace practices:	Large National study
		n = 810	- Identify strategies associated with enhanced breastfeeding intensity/longer duration		Prospective design
		2005–2007		- expressing and breastfeeding (32.4 weeks) (n = 75)	Questionnaire design with 7 day recall
		from IFPS II		- breastfeed at the breast only (31.4 weeks) (n = 250)	No description of feeding method away from workplace
		Recruitment via postal questionnaire in late pregnancy		- expressing only (26.3 weeks) (n = 75)	Older, less educated, low income and women from racial/ethnic minority groups underrepresented
				- neither breastfeeding or expressing (14.3 weeks) (n = 128)
Clemons & Amir 2010 Australia [5]	Cross-sectional	State-wide, Victoria	- Prevalence of breast milk expression	- 27% (218/903) indicated that expressing had allowed them to breastfeed for longer	Large study
		n = 903	- Demographic characteristics of women who express, why and how they do it		Possible selection bias (members of ABA)
		2008	- Women’s experience of using breast pumps		Timing of questionnaire, possible recall bias
		online questionnaire
		ABA members with internet addresses
Dabritz et al. 2010 USA [56]	Retrospective cohort	Yolo County, California	- Assess relationship between maternal experience in hospital and any breastfeeding at six months	- Almost exclusive breastfeeding at 6 months associated with not using a breast pump in hospital 77% (93/121) compared to 21% (25/121) who did use a pump in hospital (OR: 0.6 95% CI 0.3,1.0)	Differentiation between breastfeeding and breast milk feeding unclear
		n = 382
		2006–07			Possible recall bias - interviews 6–9 months after birth
		Recruited in community after birth - 8 months
Hornbeak et al. 2010 Singapore [6]	Retrospective cohort	Singapore	- Record prevalence and patterns of breastfeeding in Singaporean Chinese mothers	- Breast milk feeding initiation and duration increased over time and were independently associated with higher maternal education, increased milk expression and complementary feeding	Large representative sample of Chinese Singaporean mothers
		n = 3009		Changes between 2000–01 and 2006–08:	Possible recall bias - recruitment 6–72 months after birth
		2006–2008		Infant formula feeding 31% (66/210) to 18% (118/656)	Gestational age not indicated
		Recruited mothers of 6–72 month Chinese Singaporean children through STARS		Breast milk feed initiation 69% (144/210) to 82% (538/656)
		Mailed invitation		Expressed breast milk 9% (18/210) to 18% (118/656)
				Combination feeding 26% (54/210) to 41% (269/656)
Geraghty et al. 2012 USA [29]	Prospective cohort	Cincinnati, Ohio	- Determine who expresses their milk by end of 4 weeks and how long they continue feeding	- Milk expression common in first month postpartum	Prospective design
		n = 60		- Milk expression by 4 weeks did not significantly influence duration of breast milk feeding	Clear differentiation between breastfeeding and breast milk feeding
		2004–2007			Recruitment of women who planned to breastfeed for 6 months or more
		Participants enrolled in a research human milk bank recruited at home in first week postpartum			Mothers recruited for study knew they were going to be assisted to pump and may have been more likely to be comfortable with this.
					Possible introduction of bias as weekly collection of breast milk was initiated at 1 week by research nurse using an electric breast pump