https://orcid.org/0000-0001-7763-3533
Figures
Abstract
Introduction
Mothers of very premature infants often have difficulties expressing breastmilk, which can cause distress and potential negative impact on infant health. Clinical recommendations on breastmilk expression are extrapolated from term infants’ breastfeeding patterns. This study’s objective was to analyse the association of expressing pattern with lactation outcomes after very premature birth.
Methods
132 participants were recruited after birth between 23+0 and 31+6 weeks’ gestation. Participants recorded the milk expressed in several 24-hour periods in the three weeks after birth.
Results
Expressing frequency was positively associated with 24-hour milk yield, with an adjusted 30.5g increase per expressing session on day four (95% CI 15.7 to 45.3) and 94.4g on day 21 (95% CI 62.7 to 126.2). Expressing ≥8 times per day was associated with higher adjusted milk yield than expressing <6 times (on day four, 146.8g, 95% CI 47.4 to 246.1), but not in comparison to expressing 6–7 times (on day four, 82.1g, 95% CI –25.9 to 190.1). Participants with six months or more prior breastmilk feeding experience had a higher adjusted milk yield than others (on day four, 204.3g, 95% CI 125.2 to 283.3). Night-time (2300–0700 hours) expressing sessions were not associated with increased milk yield after adjustment for time since the prior session. On average, participants who had a longest gap between expressions of less than six hours achieved the UK target of 750g breastmilk, whereas those with a longer gap did not.
Conclusion
Expressing frequency was an important determinant of milk yield. Clinical recommendations to express ≥8 times per day were supported but for some, 6–7 times was sufficient. This was particularly likely for those with six months or more of prior breastmilk feeding experience. A need to express during the night-time hours appeared to be related to minimising the gap between expressions rather than an inherent value of night-time expression.
Citation: Levene I, Fewtrell M, Quigley MA, O’Brien F (2024) The relationship of milk expression pattern and lactation outcomes after very premature birth: A cohort study. PLoS ONE 19(7): e0307522. https://doi.org/10.1371/journal.pone.0307522
Editor: Alessia Nava, Sapienza University of Rome: Universita degli Studi di Roma La Sapienza, ITALY
Received: March 25, 2024; Accepted: July 7, 2024; Published: July 29, 2024
Copyright: © 2024 Levene et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: De-identified individual participant data (including data dictionaries) are available. A review of all requests for sharing of the study data will take place as described in National Perinatal Epidemiology Unit Standard Operating Procedures on data sharing. If agreed, any sharing of the data collected during the study must be in accordance to the Nuffield Department of Population Health and University of Oxford policies. Instructions on requesting access are provided here: https://www.npeu.ox.ac.uk/ctu/data-sharing.
Funding: IL received National Institute for Health and Social Care Research (NIHR) Clinical Doctoral Fellowship 300895 (https://www.nihr.ac.uk/) The sponsor and the funder played no role in study design, data collection and analysis, decision to publish or preparation of the manuscript.
Competing interests: I have read the journal’s policy and the authors of this manuscript have the following competing interests: MF receives an unrestricted research donation from Philips for research in infant nutrition, unrelated to this work.
Introduction
Premature infants (born at less than 37 weeks’ postmenstrual age; PMA) are at higher risk of partial breastfeeding and premature cessation of breastmilk feeding than term infants [1–3]. This is despite the fact that very premature infants (those born at less than 32 weeks’ PMA) gain particular benefit from maternal breastmilk due to reduction in the serious conditions of necrotising enterocolitis [4] and retinopathy of prematurity [5], and improvement in long term neurodevelopmental outcomes [6,7]. Improving lactation support and increasing exclusive breastmilk for premature infants are research priorities in the area of premature birth in the UK [8] and internationally [9].
World Health Organisation guidance for mothers whose infants are not able to directly breastfeed are to express (extract milk from the breasts mechanically) at least 7 times in 24 hours including at least once at night [10,11]. In the UK this recommendation (from the UNICEF UK Baby Friendly Initiative) is to express 8 to 10 times in 24 hours, with a gap of no longer than six hours at night [12]. This demanding schedule competes with other priorities, including caring for the infant/s, other children and self-care [13]. Qualitative literature emphasises the burden, pressure and mental health impact of frequent expressing [14–17] and this burden can lead to cessation [18,19].
Although the correlation of expressing frequency and milk quantity is well established [20], these precise recommendations are extrapolated from the feeding pattern of term infants [12,21]. There is a single randomised controlled trial of expressing frequency, with a very small sample size [22]. It demonstrated a higher expressed milk yield for mothers advised to express four or more time a day compared to those advised to express three or fewer times, which has little clinical relevance.
Observational studies predominantly show a correlation of expressing frequency and milk yield but most have not assessed the clinically recommended frequency of at least 7–8 expressions per day. One Australian study with 25 participants [23] challenged the recommendations, finding that although expressing more than 6 times per day was associated with higher yield than expressing 4 times per day or less, there was a similar daily milk yield for participants expressing between five and nine times per day. One other small US study analysed the gaps between expressions, reporting that there was no association of milk yield with the longest daily gap between expressions [24]. Given the recommendation to express in the night, it is also important to differentiate between the influence of a long night-time gap between expressions versus a specific need to express in the night-time hours, which no studies have attempted to do. For example, prolactin concentration is higher at night [25,26], which might give particular benefit to night-time expression.
Understanding how to optimise milk synthesis and removal in pump-dependent mothers of preterm infants has been identified as a critical unresolved question [27]. There is low concordance with expressing frequency recommendations in practice, with most studies reporting a median expressing frequency of six times per day and some as low as three [28–32]. Parents want to know more personalised and quantitative information to balance the benefits of frequent expression for lactation outcomes against the risks to physical and mental health [17,33]. For example, in our preparatory study asking parents what questions they have about milk expression, these included “How often to express is actually okay—10 times a day seemed impossible”, “Do you really have to wake up every 3 hours at night” and “how often to express and what time of day is best?” [33] specific quotations personal communication]
This study therefore uses data from an existing cohort of mothers of very premature infants within a randomised controlled trial [34] to study the association of expressing pattern and lactation outcomes. Specifically, the aims of the study are to examine the relationship of expressing frequency, gaps between expressions, circadian pattern and expressing duration with the outcome of contemporaneous expressed milk yield. The results achieve these aims and will add nuance and individualisation to clinical recommendations. Analysis of the relationship between time to first expression and lactation outcomes from this cohort has been published separately [35].
Of note, the words ‘maternal’ and ‘breastmilk’ are used throughout this article. It is acknowledged and respected that some people giving birth may not identify as female, nor choose to use the term breastmilk.
Methods
Trial design
This report uses data collected for a randomised controlled trial of a relaxation intervention. A detailed protocol has been published [34]. In brief, 132 people who had given birth to one or two infants between 23+0 and 31+6 weeks’ PMA were recruited in four tertiary and local neonatal units in the United Kingdom. Recruitment took place between 2nd August 2021 and 31st October 2022.
Three of the four neonatal units have UNICEF UK Baby Friendly Initiative level three accreditation, signifying a good level of lactation support. All units provide free hospital grade pumps both at hospital and home and have dedicated infant feeding support staff. Infant feeding support staff had varying qualifications and experience at each site–there was no requirement for specific qualifications such as board certification.
Participants were given a portable scale accurate to 0.1g (Kabalo) and filled in logs each time they expressed milk for 24 hours, on three specific timepoints (day 4, 14 and 21 after birth). The specific gravity of human milk is 1.03 so 1g and 1ml are considered near equivalent [36]. Parents were taught to record the weight of the empty milk container and of the container with milk inside, so that the pure milk weight could be calculated at the analysis stage. Research staff recorded satisfaction with the participants’ ability to use the scale after randomisation and at a day four check-in contact. The participant log template for one site is provided as S1 Fig. Single recording days were selected, over a continuous log, because of parent involvement and recommendations on reducing the burden on participants (described fully elsewhere [33]). Days 4, 14 and 21 specifically were chosen as the recording days because of previous literature examining days 4 and 14 [37,38] and a desire for a later timepoint than day 14.
Other data in the expressing phase of the trial were collected at baseline and day 4, 14 and 21 questionnaires–this included participant-reported time to first expression and participant-reported skin to skin duration. As this was an exploratory analysis of a randomised controlled trial, the number of factors included on questionnaires was limited to the most important potential covariates, to limit participant burden and reduce loss to follow up.
Infant feeding status at 36 weeks’ PMA was assessed by parent text message response and infant medical notes.
The study was approved by the Bloomsbury Research Ethics Committee, London (21/LO/0279) and registered as ISRCTN 16356650. This exploratory analysis was pre-specified. The study was powered in relation to the randomised primary outcome, as described in the published protocol. Applied to the exploratory analysis, the study had 80% power to detect a difference in mean 24-hour milk quantity of 155g between two categories of a binary variable. Power would be higher for continuous variables and lower for variables with more than two categories.
Outcomes
The outcome variables were 24-hour milk yield and expressing session milk yield. The 24-hour milk yield was defined as the sum of the milk weight expressed at each session in a 24-hour period.
To derive the number of hours from the start of the first logged session of each measurement day to the prior session (where the timing of the previous session was unknown), the final logged session of the same 24-hour period was used as a proxy. This means that an assumption was made that the 24-hour expressing pattern logged on the measurement day was the same as the previous day. For example, if a day 4 log included sessions at 0900, 1200, 1500, 1900 and 0100, the gap to prior session for the 0900 session was assumed to be 8 hours.
Statistical analysis
Analysis took place on a complete case basis. Analysis of the association between individual-level variables and 24-hour milk yield used linear regression. Analysis of the association between session-level variables and session milk yield used multilevel modelling with individual as a random effect to account for repeated measures. The regression model structures are demonstrated in S2 Fig.
Multivariable adjustment was made for potential sociodemographic and perinatal confounders, and to expressing-related covariates such as time to first expression, the type of expression used and skin to skin contact duration. These potential confounders and important covariates were identified from the literature. Baseline variables were included as candidates for inclusion in multivariable analysis if they showed association with the outcome in univariable analysis with p<0.2 and were retained in the final model if p<0.05 after stepwise removal.
Where multiple timepoints for explanatory variables were available, only measures contemporaneous with the outcome variable were included in statistical models. For example, when the outcome was milk quantity on day four after birth, a variable such as expressing frequency would only be included in the regression model when measured at day four.
Sensitivity analyses were performed to assess potential bias related to outliers and related to the assumptions made in order to assign a gap between expressing sessions to the first logged session of each day (described above). Stata 18.0 was used for analysis.
Results
Participant flow
S3 Fig demonstrates the number of participants providing data at each timepoint used in this analysis. At least one 24-hour expressing record was submitted by 108 participants. At the day four assessment, 556 expressing sessions were logged by 103 individuals; at day 14, 577 sessions by 91 individuals and at day 21, 560 sessions by 91 individuals. The majority of missing data occurred at day 14 and 21, for participants who continued in the study but did not provided expressing logs on these days. Researcher observations were that this was predominantly due to participants feeling overwhelmed with caring for their unwell infant and other demands, including expressing milk.
Participant characteristics
Table 1 shows the participant baseline characteristics. More than half had given birth to extremely premature infants (less than 28 weeks’ PMA) and 70.2% planned to exclusively breastmilk feed (85/121). Overall, 59.5% were of white ethnic background (72/121), 21.4% lived in the least deprived quintile for deprivation (28/131) and 68.3% had at least 19 years of full-time education (82/120).
Lactation characteristics
Table 2 shows the expressing pattern characteristics of the cohort for day 4, 14 and 21 at both the 24-hour level and the session level.
Few participants achieved UK recommendations on optimal expressing set by UNICEF UK Baby Friendly Initiative (BFI), but achievement of World Health Organisation (WHO) recommendations were more common (Table 3). Fewer than half of participants achieved international recommendations for expressing frequency and longest gap between sessions, on any given day.
At 36 weeks’ PMA, 64.9% (74/114) of participants were giving exclusive breastmilk to their infant/s and 86.0% were giving any breastmilk (98/114). Further description of the cohort’s lactation behaviour is provided in S1 Appendix.
Relationship of expressing pattern variables with 24-hour milk yield
Fig 1 shows the average day 21 milk yield according to expressing frequency and longest gap between expressions, with reference to the UNICEF UK BFI target of 750g. 40% of participants who expressed eight or more times per day achieved this yield target (6/15), in comparison to 37% (15/41) of those expressing 6–7 times, 21% (5/24) of those expressing 4–5 times and 10% (1/10) of those expressing less than four times per day.
Expressed milk yield on day 21 by categories of expressing frequency (A) and longest gap between expressions (B). 750g is the UNICEF UK Baby Friendly Initiative recommendation for target expressed milk yield by day 10–14 (dashed line). Box plots show median, interquartile range, data range and outliers.
In univariable analysis, expressing frequency and longest gap between expressions were each significantly associated with 24-hour milk yield at day four (Table 4). Each increment of expressing frequency was associated with 29.7g higher yield (95% CI 11.8 to 47.7) and each hour of longest gap between expressions was associated with 12.7g lower yield (95% CI -19.5 to -5.9). Expressing frequency accounted for 10% of the variability in 24-hour milk yield, and longest gap for 14% (the R2 values). There was no evidence for an association of 24-hour expressing duration with 24-hour yield.
Because there was a high correlation between expressing frequency and longest gap between expressions, only one of these two variables was used in multivariable analysis (expressing frequency). Expressing frequency was chosen because the longest gap between sessions requires a higher level of derivation from the expressing log than expressing frequency, as described in the methods. The association of expressing frequency with 24-hour milk yield on day four persisted after adjustment for confounders. There was minimal change between the crude and adjusted coefficients for any of the variables included (Table 4). The multivariable model accounted for 43% of the variability in 24-hour milk yield (R2).
The same analysis was conducted for milk yield at day 14 and 21 (S1 and S2 Tables). Expressing frequency and longest gap between expressions were also associated with milk yield in unadjusted analysis at day 14 and 21. In addition, expressing duration was associated with milk yield at day 21, but this did not remain significantly associated after adjustment for expressing frequency.
As at day 4, there was minimal change in the coefficient for expressing frequency at day 14 after multivariable adjustment (unadjusted 50.3g, 95% CI 14.4 to 86.2 and adjusted 50.8g, 95% CI 17.8 to 83.9). The coefficient for expressing frequency at day 21 increased after adjustment for prior breastmilk feeding experience, from 85.7g (95% CI 51.2 to 120.3) to 94.4g (95% CI 62.7 to 126.2).
In univariable analysis, expressing frequency accounted for 8% of the variability in 24-hour milk yield at day 14 and 22% of the variability in 24-hour milk yield at day 21. The multivariable models explained 39% and 35% respectively (R2).
Assessing clinical recommendations on expressing pattern
Table 5 shows clinical recommendations for expressing frequency and longest gap between expressions at day 4 and 21. On each day, adjustment was for the variables reported in the multivariable models shown in Table 4, S1 and S2 Tables.
Expressing eight or more times per day was associated with significantly higher yield than expressing less than six times per day, but not in comparison to expressing 6–7 times per day, at day 4 and 21. At day 21, a longest gap of less than six hours was associated with higher 24-hour yield than both a longest gap of 6–7 hours and a longest gap of eight hours or more. In contrast, at day four and 14, a longest gap of less than six hours was associated with higher 24-hour yield than a longest gap of eight hours of more, but not the intermediate category of 6–7 hours.
Clinical recommendations can also be assessed by using the multivariable statistical model to predict adjusted 24-hour milk yield across a range of expressing frequencies. This is shown for day 21 in comparison to the UNICEF UK BFI 24-hour yield target of 750g (Fig 2). The model is adjusted for prior breastfeeding experience. Expressing eight times per day achieved an adjusted 24-hour yield at the target yield, but there is considerable overlap in confidence intervals in the range of 7–9 times per day.
Adjusted effect estimate and confidence intervals. 750g is the UNICEF UK Baby Friendly Initiative clinical recommendation for target expressed milk yield by day 10–14 (dashed line).
Given the very large effect estimate for the non-modifiable factor of previous breastfeeding experience on milk yield in all the multivariable models presented above, the association of 24-hour yield and expressing pattern is presented according to prior breastfeeding experience in S4 Fig. People with more than six months of prior breastfeeding experience achieved target yield with a much lower expressing frequency than others (on average, at five expressions per day).
Relationship of session-level expressing pattern variables with session milk yield
The number of hours since the prior session was significantly positively associated with session milk yield at day 14 and 21. Expressing session duration was positively associated with milk yield on all timepoints. After adjustment for the number of hours since prior session, timing of 2300 to 0700 hours (night-time period) was not associated with higher session milk yield at any timepoint (Table 6).
Sensitivity analyses
Sensitivity analyses are described in S1 Appendix. There were some changes to the results related to longest gap between expressing sessions when the assumptions described in the methods section to derive this variable were removed.
Discussion
Summary of findings
This study strongly supports the importance [23,39–41] of expressing frequency in relation to 24-hour expressed milk yield. For example, at day 21 there is an adjusted increase in 24-hour yield of 94g per expressing session. However, expressing frequency only explained 8–22% of the variability in milk yield on any one timepoint, so women expressing at the same frequency may have very different milk yields.
UK clinical recommendations to express at least eight times per day were supported because this was associated with achievement of recommended milk yield on average. However, there was wide inter-individual variability and few people expressed at this frequency. Expressing eight or more times a day was not associated with significantly higher 24-hour milk yield than expressing 6–7 times per day, which was closer to the median frequency and therefore more achievable for this cohort. However, participants expressing less than six times per day were unlikely to achieve target yields. This would be useful information for parents who feel that expressing eight times per day is unrealistic or unachievable.
We attempted to analyse the association of longest gap between expressions with 24-hour milk yield. A significant negative association was reported, with milk yield reducing by 30g for each hour increase of the longest gap between expressions on day 21. The clinical recommendation of having a gap no longer than six hours was supported. However, these findings reduced in size and significance when removing certain assumptions required to derive the gap to prior session for the first recorded session of each day. Therefore, this finding should be interpreted with some caution.
The session level analysis reported here was able for the first time to separate two factors associated with the clinical recommendation to express in the middle of the night–the circadian timing and the time since the prior expressing session. The analysis does not support an inherent value to the night-time period, once the time to prior session is accounted for. Although parents would have to express in the night period to keep the longest expressing gap under the recommended six hours, this finding supports flexibility between individuals in exactly when to express, which could take into account their sleep pattern. This is important as sleep interruption is cited as a reason to stop expressing [18].
We have demonstrated a novel exploratory finding that participants with longer prior breastmilk feeding experience (in this study defined as more than six months) express on average less frequently to achieve the same yield as those who are primiparous or have shorter duration of prior breastmilk feeding. This could allow clinicians to individualise expressing recommendations. However, this was not a pre-specified hypothesis and is based on 26 participants with more than six months of breastmilk feeding experience. Therefore this should be interpreted cautiously and requires replication.
Setting the findings in context
This cohort showed similar expressing frequency to many other cohorts in the literature [23,29,30,32,42], although higher frequency than some [31,43]. Expressed milk yield was broadly similar to previous cohorts [39,40,42,43] although notably lower than two small studies that looked in depth at expressing dynamics [23,44]. This may suggest that our study recruited a broader range of participants, not just the most motivated.
The sample is not representative of very premature infants in general in the UK–it was predominantly recruited from Southern English units, which tend to have higher breastmilk prevalence at discharge home [45], and three out of four neonatal units are level three BFI accredited, suggesting a better than average level of lactation support. This likely explains why the prevalence of any breastmilk at 36 weeks’ PMA in this report (86%) was much higher than the national prevalence reported at discharge home in the National Neonatal Audit Programme (60% in 2020) [46]. Although this may reduce generalisability, it is useful for analysis of the relationship of expressing behaviour with breastfeeding outcomes to study a cohort with high intention to breastmilk feed.
The positive association of expressing frequency and milk yield we have found is consistent with most other studies [23,31,32,39,41,47–51]. One previous study showed a smaller adjusted increase in 24-hour yield of 25ml per expressing session (at average day 14) [50]. Only one study is known to have reported on the longest gap between expressions previously, finding no association with milk yield at two or four weeks after birth [24]. Mean longest gap was shorter in that report (6.5 hours) than seen here (7.5 to 10 hours depending on the timepoint).
In the NICU context, two previous studies with mothers of very preterm infants have shown a similar magnitude of increase in milk yield with longer previous breastfeeding experience—higher milk yield by 222ml/day [52] and 22–25% greater milk yield at week one and four [53]. One study with mixed gestation NICU mothers showed a median difference in duration of exclusive breastfeeding of one month for those with any previous breastfeeding experience [54]. In contrast, one study showed no correlation of previous breastfeeding experience with ability to express at least 500g per day by week 5 [37].
It is possible that at least part of this finding is due to confounding, whereby people with longer length of breastmilk feeding are more motivated, have higher self-efficacy [55] and may put in place unmeasured supportive behaviours to achieve higher milk yield. However, there could also be an element of physiological difference in the breast tissue of someone who has experience prolonged previous lactation. In support of this idea, one study [56] has shown that milk production per unit of breast tissue is much higher after six months of lactation (for example 4g/ml at nine months compared to 2g/ml in the first six months).
It is well established in mice that mammary epithelial cells show epigenetic changes after the first pregnancy/lactation cycle that causes earlier activation of milk-related genes for subsequent pregnancy/lactation cycles [57,58]. In multiparous rats, there is increased responsiveness to prolactin and therefore a lower circulating prolactin concentration [59]. A similar phenomenon has been seen in two human studies. One showed lower circulating prolactin in the first four days after birth in multiparous women, but higher infant milk intake [60]. The second showed increased milk yield at day seven after the woman’s second birth in comparison to their first birth, despite shorter total duration of feeding time [61]. In addition, increased length of prior breastfeeding has been linked with greater epithelial area and increased expression of progesterone receptors in clinical breast tissue samples, even a considerable time after cessation of lactation [62]
Strengths and limitations
The strengths of this study are the broad population of included participants, in terms of sociodemographic variables as well as expressing behaviour. The inclusion of individual and session level variables across time has yielded novel findings and will help clinicians to personalise the advice they offer to families.
The most significant limitation of the study is the sample size, including some loss to follow up in expressing log submission. Because of the increased loss to follow up at day 14 and 21, the findings at day 4 likely have a lower risk of bias than those at day 14 and 21. A further limitation of the study is the reliance on participant-measured data on expressing pattern and yield, which could affect data accuracy; and the restriction of data collection to single days rather than a continuous log. Single day logs meant that some assumptions were required to derive the gap to prior session for the first expressing session of the day, which had a significant impact on the analysis conclusions related to the longest gap between expressions. The characteristics of the first expression of the day were significantly different from others so it is unclear whether the impact on the analysis was because of bias in the derivation assumption or because these expressing sessions have an important and differential influence on the outcomes. In addition, some expressing dynamics may relate better to cumulative expressing behaviours, which may not be adequately represented by single day logs.
Both elements of trial design (participant measurement and single day collection timepoints) were chosen after extensive parent involvement, to reduce interference and burden for participants at a stressful time [33]. This may have resulted in the broader representation of expressing behaviour and lactation outcomes than seen in other detailed studies [23,44] and may be considered a trade-off when studying milk expression at this stressful time.
There were several variables that could not be included in this analysis that are known to be important for expressed milk outcomes; most notably maternal factors such as chorioamnionitis [38,63], assisted reproduction [63,64] and obesity [63,65]. This could lead to residual confounding.
Conclusions
Overall, the data in this study provides much needed evidence related to current clinical recommendations on expressing patterns after very preterm birth. UK and international recommendations on expressing frequency and longest gap between expressions were supported, but there are significant limitations to average-based recommendations in a population with such wide inter-individual variability.
Families can be counselled that expressing pattern is only one determinant of milk yield, and that those with an otherwise good risk profile for lactation may be able to express at a lower frequency than the general recommendation. A change in emphasis can be suggested related to clinical advice on night-time expression, which appears to be valuable in relation to limiting the gap between expressions rather than because of its specific circadian timing.
More work is required to replicate the exploratory finding of the importance of prior breastfeeding experience and identify further variables that explain the residual variability. This would help families understand their risk factors in relation to milk yield, bringing potential for individualisation of recommendations and also might identify potential further targets for intervention.
Supporting information
S1 Fig. Template for the participant expressing log at one of the research sites.
https://doi.org/10.1371/journal.pone.0307522.s001
(PDF)
S2 Fig. Diagrammatic representation of one level and multilevel modelling used in this analysis for individual level and session level analysis of milk yield.
https://doi.org/10.1371/journal.pone.0307522.s002
(PNG)
S4 Fig.
(A): Scatter graph of expressing frequency and 24-hour milk yield on day 21 according to breastfeeding experience. (B) Modelled 24-hour milk yield on day 21. Unadjusted linear regression lines are shown in (A). (B) includes a statistical interaction between expressing frequency and breastfeeding experience.
https://doi.org/10.1371/journal.pone.0307522.s004
(JPG)
S1 Appendix. Supplementary text with associated figures.
https://doi.org/10.1371/journal.pone.0307522.s005
(DOCX)
S1 Table. Crude and adjusted coefficients for the outcome of 24-hour milk yield at day 14.
https://doi.org/10.1371/journal.pone.0307522.s006
(DOCX)
S2 Table. Crude and adjusted coefficients for the outcome of 24-hour milk yield at day 21.
https://doi.org/10.1371/journal.pone.0307522.s007
(DOCX)
Acknowledgments
The 132 participants, who gave so much of their time at a period of immense stress. The 675 “Public, Patient Involvement” participants who helped to define the questions and minimise ethical concerns. NPEU CTU team Pollyanna Hardy, Kayleigh Stanbury, Christina Cole, Jenny Bell, Andy King, David Murray, Richard Welsh, Fiona Goodgame, Mariko Nakahara, Adriana Ortiz, Sarah Chamberlain and Andy Kirk. NPEU staff Marian Knight, Claire Carson, Dagmar Hutt, Jenny Shilton-Osborne, Pam White and Julieta Estremadoyro. Principal Investigators Aniko Deierl and Clare Jennings. Site research staff Bryony Ward, Ryan Browne, Sharon Baugh, Cath Dixon, Jean Yong, Batia Gourin, Zoe McClure, Trinity Oke, Kirupalini Mariampillai, Annie Aloysius, Karen Mitchell, Jay Banerjee, Duaa Dafaalla, Karen Dockery, Nicola Booth, Michelle Evans, Jenna Hill and Mirsada Paskali. The infant feeding teams at all the recruiting hospitals, including Gillian Denton, Emily Hollingworth and Victoria Hedges. The Trial Steering Committee Pat Hoddinott, Kelly Handley, Nadia Leake, Mark Johnson and Beth McCleverty–Nadia and Beth provided public and parent perspective throughout the trial. The Data Monitoring Committee Shalini Ojha, Omar Omar and Claire Carter.
References
- 1. Ericson J, Flacking R, Hellström-Westas L, Eriksson M. Changes in the prevalence of breast feeding in preterm infants discharged from neonatal units: a register study over 10 years. BMJ Open. 2016;6: e012900. pmid:27965252
- 2. Akerström S, Asplund I, Norman M. Successful breastfeeding after discharge of preterm and sick newborn infants. Acta Paediatr. 2007;96: 1450–1454. pmid:17880414
- 3. Donath SM, Amir LH. Effect of gestation on initiation and duration of breastfeeding. Arch Dis Child Fetal Neonatal Ed. 2008;93: F448–F450. pmid:18450805
- 4. Corpeleijn WE, de Waard M, Christmann V, van Goudoever JB, Jansen-van der Weide MC, Kooi EMW, et al. Effect of Donor Milk on Severe Infections and Mortality in Very Low-Birth-Weight Infants. JAMA Pediatr. 2016;170: 654. pmid:27135598
- 5. Zhou J, Shukla V V, John D, Chen C. Human milk feeding as a protective factor for Retinopathy of Prematurity: a meta-analysis. Pediatrics. 2015;136: e1576–86. pmid:26574589
- 6. Belfort MB, Knight E, Chandarana S, Ikem E, Gould JF, Collins CT, et al. Associations of Maternal Milk Feeding With Neurodevelopmental Outcomes at 7 Years of Age in Former Preterm Infants. JAMA Netw Open. 2022;5: e2221608. pmid:35816314
- 7. Vohr BR, Poindexter BB, Dusick AM, McKinley LT, Higgins RD, Langer JC, et al. Persistent Beneficial Effects of Breast Milk Ingested in the Neonatal Intensive Care Unit on Outcomes of Extremely Low Birth Weight Infants at 30 Months of Age. Pediatrics. 2007;120: e953–e959. pmid:17908750
- 8. James Lind Alliance. Preterm Birth Top 10. 2014 [cited 29 Apr 2022]. Available: https://www.jla.nihr.ac.uk/priority-setting-partnerships/preterm-birth/top-10-priorities/.
- 9.
World Health Organization. Research implications. WHO recommendations for care of the preterm or low-birth-weight infant. Geneva; 2022. Available: https://www.ncbi.nlm.nih.gov/books/NBK586695/.
- 10. World Health Organisation & United Nations Children’s Fund (UNICEF). Protecting, promoting and supporting breastfeeding: the baby-friendly hospital initiative for small, sick and preterm newborns. 2020 Aug.
- 11. Maastrup R, Hannula L, Hansen MN, Ezeonodo A, Haiek LN. The Baby-friendly Hospital Initiative for neonatal wards. A mini review. Acta Paediatr. 2022;111: 750–755. pmid:34932843
- 12. UNICEF UK, FM E. The evidence and rationale for the UNICEF UK Baby Friendly Initiative standards. 2013. Available: https://www.unicef.org.uk/babyfriendly/about/evidence-and-rationale-for-the-baby-friendly-standards/.
- 13. Palmquist AEL, Holdren SM, Fair CD. “It was all taken away”: Lactation, embodiment, and resistance among mothers caring for their very-low-birth-weight infants in the neonatal intensive care unit. Soc Sci Med. 2020;244: 112648. pmid:31707144
- 14. Fernández Medina IM, Granero-Molina J, Fernández-Sola C, Hernández-Padilla JM, Camacho Ávila M, López Rodríguez M del M. Bonding in neonatal intensive care units: Experiences of extremely preterm infants’ mothers. Women and Birth. 2018;31: 325–330. pmid:29191725
- 15. Sisk P, Quandt S, Parson N, Tucker J. Breast milk expression and maintenance in mothers of very low birth weight infants: supports and barriers. J Hum Lact. 2010;26: 368–375. pmid:20930219
- 16. Mӧrelius E, Kling K, Haraldsson E, Alehagen S. You can’t flight, you need to fight—A qualitative study of mothers’ experiences of feeding extremely preterm infants. J Clin Nurs. 2020;29: 2420–2428. pmid:32223034
- 17. Ikonen R, Paavilainen E, Kaunonen M. Trying to live with pumping: Expressing milk for preterm or small for gestational age infants. MCN The American Journal of Maternal/Child Nursing. 2016;41: 110–115. pmid:26909725
- 18. Parker LA, Krueger C, Sullivan S, Mueller M. Demographic, Social, and Personal Factors Associated With Lactation Cessation by 6 Weeks in Mothers of Very Low Birth Weight Infants. Journal of Human Lactation. 2021;37: 511–520. pmid:33030992
- 19. Furman L, Minich N, Hack M. Correlates of Lactation in Mothers of Very Low Birth Weight Infants. Pediatrics. 2002;109: e57–e57. pmid:11927730
- 20. Bendixen MM, Iapicca LC, Parker LA. Nonpharmacologic Factors Affecting Milk Production in Pump-Dependent Mothers of Critically Ill Infants. Advances in Neonatal Care. 2023;23: 51–63. pmid:36700680
- 21. Kent JC. How breastfeeding works. J Midwifery Womens Health. 2007;52: 564–570. pmid:17983993
- 22. deCarvalho M, Anderson DM, Giangreco A, Pittard WB 3rd. Frequency of milk expression and milk production by mothers of nonnursing premature neonates. Am J Dis Child. 1985;139: 483–485. pmid:3984973
- 23. Lai CT, Rea A, Mitoulas LR, Kent JC, Simmer K, Hartmann PE, et al. Short-term rate of milk synthesis and expression interval of preterm mothers. Arch Dis Child Fetal Neonatal Ed. 2020;105: 266–269. pmid:31296697
- 24. Hopkinson JM, Schanler RJ, Garza C. Milk production by mothers of premature infants. Pediatrics. 1988;81: 815–20. Available: http://www.ncbi.nlm.nih.gov/pubmed/3368280. pmid:3368280
- 25. Diaz S, Seron-Ferre M, Cardenas H, Schiappacasse V, Brandeis A, Croxatto HB. Circadian Variation of Basal Plasma Prolactin, Prolactin Response to Suckling, and Length of Amenorrhea in Nursing Women*. J Clin Endocrinol Metab. 1989;68: 946–955. pmid:2715293
- 26. Stern JM, Reichlin S. Prolactin circadian rhythm persists throughout lactation in women. Neuroendocrinology. 1990;51: 31–37. pmid:2106085
- 27. Neville MC, Anderson SM, McManaman JL, Badger TM, Bunik M, Contractor N, et al. Lactation and Neonatal Nutrition: Defining and Refining the Critical Questions. J Mammary Gland Biol Neoplasia. 2012;17: 167–188. pmid:22752723
- 28. Becker GE, Smith HA, Cooney F, G.E B, H.A S. Methods of milk expression for lactating women. Cochrane Database of Systematic Reviews. 2016;2016: CD006170. pmid:27684560
- 29. Ikonen R, Paavilainen E, Helminen M, Kaunonen M. Preterm infants’ mothers’ initiation and frequency of breast milk expression and exclusive use of mother’s breast milk in neonatal intensive care units. J Clin Nurs. 2018;27: e551–e558. pmid:28960635
- 30. Gianni ML, Bezze EN, Sannino P, Baro M, Roggero P, Muscolo S, et al. Maternal views on facilitators of and barriers to breastfeeding preterm infants. BMC Pediatr. 2018;18: 1–7. pmid:30149811
- 31. Fewtrell MS, Kennedy K, Ahluwalia JS, Nicholl R, Lucas A, Burton P. Predictors of expressed breast milk volume in mothers expressing milk for their preterm infant. Arch Dis Child Fetal Neonatal Ed. 2016;101: F502–F506. pmid:26936878
- 32. Gomez-Juge C, Scarpelli V, Yellayi D, Cerise J, Weinberger B, Brewer M, et al. Breast Milk Production Variability Among Mothers of Preterm Infants. Breastfeeding Medicine. 2023;18: 571–578. pmid:37615564
- 33. Levene I, Alderdice F, McCleverty B, O’Brien F, Fewtrell M, Quigley MA. A report on parent involvement in planning a randomised controlled trial in neonatology and lactation–insights for current and future research. Int Breastfeed J. 2022;17: 69. pmid:36104819
- 34. Levene I, Bell JL, Cole C, Stanbury K, O’Brien F, Fewtrell M, et al. Comparing the effect of a lactation-specific relaxation and visualisation intervention versus standard care on lactation and mental health outcomes in mothers of very premature infants (the EXPRESS trial): study protocol for a multi-centre, unmasked, randomised, parallel-group trial. Trials. 2022;23: 611. pmid:35906655
- 35. Levene I, Quigley MA, Fewtrell M, O’Brien F. Does extremely early expression of colostrum after very preterm birth improve mother’s own milk quantity? A cohort study. Arch Dis Child Fetal Neonatal Ed. 2024; fetalneonatal-2023-326784. pmid:38442953
- 36. Kumar RK, Singhal A, Vaidya U, Banerjee S, Anwar F, Rao S. Optimizing nutrition in preterm low birth weight infants—consensus summary. Front Nutr. 2017;4: 20. pmid:28603716
- 37. Hill PD, Aldag JC. Milk Volume on Day 4 and Income Predictive of Lactation Adequacy at 6 Weeks of Mothers of Nonnursing Preterm Infants. J Perinat Neonatal Nurs. 2005;19: 273–282. pmid:16106236
- 38. Hoban R, Bigger H, Schoeny M, Engstrom J, Meier P, Patel AL. Milk Volume at 2 Weeks Predicts Mother’s Own Milk Feeding at Neonatal Intensive Care Unit Discharge for Very Low Birthweight Infants. Breastfeeding Medicine. 2018;13: 135–141. pmid:29377728
- 39. Morton J, Hall JY, Wong RJ, Thairu L, Benitz WE, Rhine WD. Combining hand techniques with electric pumping increases milk production in mothers of preterm infants. Journal of Perinatology. 2009;29: 757–764. pmid:19571815
- 40. Murase M, Nommsen-Rivers L, Morrow AL, Hatsuno M, Mizuno K, Taki M, et al. Predictors of low milk volume among mothers who delivered preterm. J Hum Lact. 2014;30: 425–435. pmid:25063573
- 41. Henderson JJ, Hartmann PE, Newnham JP, Simmer K. Effect of Preterm Birth and Antenatal Corticosteroid Treatment on Lactogenesis II in Women. Pediatrics. 2008;121: e92–e100. pmid:18166549
- 42. Héon M, Goulet C, Garofalo C, Nuyt AM, Levy E. An Intervention to Promote Breast Milk Production in Mothers of Preterm Infants. West J Nurs Res. 2016;38: 529–552. pmid:25395013
- 43. Bendixen MM, Weaver MT, Parker LA. Milk Volume Outcomes in Pump-Dependent Mothers of Critically Ill Infants. Advances in Neonatal Care. 2022;22: 253–260. pmid:33958519
- 44. Ru X, Huang X, Feng Q. Successful full lactation achieved by mothers of preterm infants using exclusive pumping. Front Pediatr. 2020;8: 191. pmid:32391294
- 45. National Neonatal Audit Programme. National Neonatal Audit Programme—Individual unit data. In: https://nnap.rcpch.ac.uk/spine.aspx.
- 46. Programme NNA. National Neonatal Audit Programme Annual Report on 2020 data. 2021. Available: https://www.rcpch.ac.uk/resources/national-neonatal-audit-programme-annual-report-2020.
- 47. Hill PD, Aldag JC, Chatterton RT. Effects of Pumping Style on Milk Production in Mothers of Non-Nursing Preterm Infants. Journal of Human Lactation. 1999;15: 209–216. pmid:10578798
- 48. Hill PD, Aldag JC, Chatterton RT. Initiation and frequency of pumping and milk production in mothers of non-nursing preterm infants. J Hum Lact. 2001;17: 9–13. pmid:11847858
- 49. Lau C, Hurst NM, Smith EO, Schanler RJ. Ethnic/racial diversity, maternal stress, lactation and very low birthweight infants. J Perinatol. 2007;27: 399. pmid:17592486
- 50. Lussier MM, Brownell EA, Proulx TA, Bielecki DM, Marinelli KA, Bellini SL, et al. Daily Breastmilk Volume in Mothers of Very Low Birth Weight Neonates: A Repeated-Measures Randomized Trial of Hand Expression Versus Electric Breast Pump Expression. Breastfeeding Medicine. 2015;10: 312–317. pmid:26204125
- 51. Bin-Nun A, Kassirer Y, Jarallah Y, Barg M, Mimouni FB, Hammerman C, et al. Early breast expression for very low birth infants admitted to neonatal intensive care unit: the challenges of cesarean deliveries. Journal of Maternal-Fetal and Neonatal Medicine. 2021. pmid:34459334
- 52. Bishara R, Dunn MS, Merko SE, Darling P. Volume of foremilk, hindmilk, and total milk produced by mothers of very preterm infants born at less than 28 weeks of gestation. J Hum Lact. 2009;25: 272–279. pmid:19383633
- 53. Hill PD, Aldag JC, Chatterton RT, Zinaman M. Primary and secondary mediators’ influence on milk output in lactating mothers of preterm and term infants. J Hum Lact. 2005;21: 138–150. pmid:15886340
- 54. Sokou R, Parastatidou S, Ioakeimidis G, Tavoulari E-F, Makrogianni A, Isaakidou E, et al. Breastfeeding in Neonates Admitted to an NICU: 18-Month Follow-Up. Nutrients. 2022;14: 3841. pmid:36145216
- 55. Kronborg H, Foverskov E, Væth M, Maimburg RD. The role of intention and self-efficacy on the association between breastfeeding of first and second child, a Danish cohort study. BMC Pregnancy Childbirth. 2018;18: 454. pmid:30466403
- 56. Kent JC, Mitoulas L, Cox DB, Owens RA, Hartmann PE. Breast Volume and Milk Production During Extended Lactation in Women. Exp Physiol. 1999;84: 435–447. pmid:10226183
- 57. Slepicka PF, Somasundara AVH, dos Santos CO. The molecular basis of mammary gland development and epithelial differentiation. Semin Cell Dev Biol. 2021;114: 93–112. pmid:33082117
- 58. dos Santos CO, Dolzhenko E, Hodges E, Smith AD, Hannon GJ. An Epigenetic Memory of Pregnancy in the Mouse Mammary Gland. Cell Rep. 2015;11: 1102–1109. pmid:25959817
- 59. Anderson GM, Grattan DR, van den Ancker W, Bridges RS. Reproductive Experience Increases Prolactin Responsiveness in the Medial Preoptic Area and Arcuate Nucleus of Female Rats. Endocrinology. 2006;147: 4688–4694. pmid:16825319
- 60. Zuppa AA, Tornesello A, Papacci P, Tortorolo G, Segni G, Lafuenti G, et al. Relationship between Maternal Parity, Basal Prolactin Levels and Neonatal Breast Milk Intake. Neonatology. 1988;53: 144–147. pmid:3370259
- 61. Ingram J, Woolridge M, Greenwood R. Breastfeeding: it is worth trying with the second baby. The Lancet. 2001;358: 986–987. pmid:11583756
- 62. Gabrielson M, Chiesa F, Behmer C, Rönnow K, Czene K, Hall P. Association of reproductive history with breast tissue characteristics and receptor status in the normal breast. Breast Cancer Res Treat. 2018;170: 487–497. pmid:29603032
- 63. Omarsdottir S, Adling A, Bonamy AKE, Legnevall L, Tessma MK, Vanpée M. Predictors of sustained maternal milk feeds in extremely preterm infants. J Perinatol. 2015;35: 367–372. pmid:25429384
- 64. Giannì ML, Roggero P, Amato O, Orsi A, Garbarino F, Garavaglia E, et al. Intervention for promoting breast milk use in neonatal intensive care unit: a pilot study. The Journal of Maternal-Fetal & Neonatal Medicine. 2013;27: 475–478. pmid:23796208
- 65. Hoban R, Bowker RM, Gross ME, Patel AL. Maternal production of milk for infants in the neonatal intensive care unit. Semin Perinatol. 2021;45: 151381. pmid:33423795