Supporting women with diabetes to breastfeed: use of antenatal breastmilk expression

Breastfeeding is recommended as the optimal method of infant feeding. The World Health Organization recommends that all babies are exclusively breastfed for the first 6 months of life and continue to breastfeed for up to 2 years of age and beyond (World Health Organization (WHO) and United Nations Children's Fund (UNICEF), 2018). The benefits associated with breastfeeding are extensive. Specific health benefits have been identified in cases of maternal diabetes (Victoria et al, 2016; Doughty and Taylor, 2021). Results from a meta-analysis and a number of systematic reviews indicate that breastfeeding offers protection to mothers against type 2 diabetes and may protect infants against diabetes in later life (Victoria et al, 2016). These benefits are compelling, given the increased risk of morbidity and mortality that exist for the woman and her developing fetus from maternal diabetes (National Institute for Health and Care Excellence (NICE), 2015; Nelson Piercy, 2020; Vause et al, 2020).

When pregnancy is complicated by diabetes, there are specific challenges to breastfeeding, including delayed lactogenesis II, neonatal hypoglycaemia and separation of mothers and infants because of admissions to neonatal intensive care (Stenhouse, 2018; Doughty and Taylor, 2021). Practices that are supportive of breastfeeding can reduce some of these challenges (Fallon and Dunne, 2015). One practice that is recommended in many clinical settings is antenatal breastmilk expression. The mother expresses and stores a quantity of colostrum in late pregnancy, which is then available for the baby in the early neonatal period. A recent scoping review highlights that evidence on the use of antenatal breastmilk expression for women with and without diabetes is growing (Foudil-Bey et al, 2021). Given the specific challenges faced by women with diabetes, it is important to review studies that examine antenatal breastmilk expression to determine how the practice supports breastfeeding.

Methods

A comprehensive search was carried out of the Cumulative Index to Nursing and Allied Health Literature (CINAHL) database, Scopus, Medline, Excerpta Medica DataBASE (EMBASE) and the Cochrane Library on 21 January 2022, to identify studies on the use of antenatal breastmilk expression for women with diabetes. The keywords for the search were antenatal, ante natal, prenatal, pre natal, pregnancy; breastmilk, breast milk, human milk, mother's milk, colostrum; express, pump, collect and diabet. The boolean operator OR was used with key terms to expand the search, AND was used to limit the search. Truncation was used to identify other versions of the same word. The search was limited to that conducted in humans, and primary research. The year of publication was limited to the last 10 years (2011–2022) to identify studies published since the literature review undertaken by Chapman et al (2013), which included studies from 1946–2011.

The combined search of all databases resulted in 310 citations. The search, which was undertaken by two authors (MG, AF), was limited manually by title and abstract, removing duplicates, and full text screening (Figure 1). Finally, eight studies (11 citations) met the inclusion criteria (Table 1). Three citations were conference or poster abstracts of included studies (East et al, 2013; Moorhead et al, 2019; Heal and Casey, 2020) and three articles did not meet the inclusion criteria. One was excluded as it described a practice initiative (Weinel and Cusack, 2019), one was a commentary (Berens, 2017) and one provided insufficient information (Egan et al, 2018).

Two feasibility studies, two cohort studies, one randomised controlled trial, two qualitative studies and one systematic review were included. The search was not restricted by research method, as a narrative review can include a broad range of literature, including systematic reviews (Coughlan and Cronin, 2020). It has also been reported that the majority of editors of core clinical journals consider systematic reviews to be original research: 71% in 2012 (Meerpohl et al, 2012); 80% in 2019 (Martinic et al, 2019). Studies were critiqued using Critical Appraisal Skills Programme (2019) tools. The critique highlighted that although results from individual studies can help locally, the collective findings are limited by the constraints imposed by the methods used to test the intervention.

Data were extracted from the eight methodologically different studies and synthesised narratively into three themes: ‘informational’, ‘emotional and motivational’ and ‘practical’ aspects of antenatal breastmilk expression. The framework was based on step 5 of the ten steps to successful breastfeeding implementation guidance, which recommends ‘providing emotional and motivational support, imparting information and teaching concrete skills to enable mothers to breastfeed successfully’ (WHO and UNICEF, 2018). This framework was used to highlight different aspects of antenatal breastmilk expression, while linking with the requirements for support, as antenatal breastmilk expression is widely recommended in practice settings.

Results

Informational aspects of antenatal breastmilk expression

Admission to neonatal care

‘Imparting information’ (WHO and UNICEF, 2018) is an important aspect of providing support to undertake antenatal expression. For the purpose of this review, the informational theme focuses on neonatal wellbeing, as women worry about their baby's wellbeing (Casey et al, 2019a; Moorhead et al, 2021). A concern was highlighted in this regard when two small studies (one feasibility, one cohort study) reported an increased admission rate to neonatal care for infants in the intervention groups (Forster et al, 2011; Soltani and Scott, 2012). However, neither study used inferential methods in their design and therefore could not establish a causal link. Randomised controlled trials on the other hand, are the gold standard for testing an intervention (Jacobsen, 2021). A relatively recent randomised controlled trial (Diabetes and Antenatal Breastmilk Expression (DAME) trial) found that for women at ‘low risk of complications’, infants were not at greater risk of admission to neonatal care following antenatal expression in comparison to standard care (Forster et al, 2017). A Cochrane systematic review (East et al, 2014) did not identify any further trials. This is an important finding, as it highlights the need for further randomised controlled trials. Future updates to this systematic review protocol might examine how the timing of the intervention, in terms of gestation, affects outcomes.

The primary outcome of the DAME trial was the percentage of neonates admitted to neonatal care (Forster et al, 2017). Women who were at low risk were randomly assigned in blocks to either the intervention (antenatal breastmilk expression) group (n=316), or to standard care (n=315), using a computerised random number generator. Women at ‘low risk’ were identified by the trial's eligibility criteria, which excluded women with specific medical, obstetric and fetal complications. Caregivers were not masked to group allocation because of the nature of the intervention. Intention to treat analysis, which reduces the risk of bias (McEvoy et al, 2017), was used, and outcome assessors were masked to group allocation. Infants in the intervention compared to the standard care group did not differ in admission to neonatal care (15% vs 14%), nor in mean gestational age at birth (38.6 vs 38.7 weeks) (Forster et al, 2017). The main reasons for admission to neonatal care were similar in the intervention and the standard care group and included hypoglycaemia (42% vs 36%), suspected infection (42% vs 41%) and respiratory distress (27% vs 23%). (Forster et al, 2017). However, as a result of the trial's eligibility criteria, the findings are not generalisable to women at high risk of complications.

The earlier study by Forster et al (2011), undertaken to determine the feasibility of conducting the later randomised controlled trial (Forster et al, 2017) found that admission to neonatal care was greater for infants in the pilot group compared to the audit group (30% vs 17%) (Forster et al, 2011). However, participants consisted of a small non-random sample of women (n=43), and the comparison group was identified from an audit conducted at the study sites. Similarly, a retrospective cohort study from England (Soltani and Scott, 2012) investigated antenatal expression among women with diabetes (n=94) who attended the local maternity services. However, the sample size of the expression group was small (16/94). The cohort study found that gestational age was lower (mean: 37.1 (standard deviation: 2.6) vs 38.2 (2.2), P=0.06) and admission to neonatal intensive care was higher for neonates in the expression compared to the non-expression group (33% vs 12%, P=0.06) (Soltani and Scott, 2012).

The rate of transfer to neonatal care was recently examined in a Norwegian feasibility study, which used observational trial methods to determine the practicality and acceptability of antenatal breastmilk expression, for women (n=28) with diabetes (Johnsen et al, 2021). No comparison group was used. It reported that following antenatal breastmilk expression, seven (25%) infants needed transfer to neonatal care because of blood sugar<1.4 mmol/l (n=4) or other reasons (n=3) (Johnsen et al, 2021). The authors reported that the study was not powered to determine rates of neonatal admission or hypoglycaemia and the higher prevalence of neonatal admission in comparison to the DAME trial (Forster et al, 2017) may have occurred as antenatal expression commenced one week later, at 37 weeks, which resulted in a lower total volume of expressed milk (Johnsen et al, 2021).

Neonatal hypoglycaemia

Neonatal hypoglycaemia has been identified as a possible complication of maternal diabetes (Stenhouse, 2018; Doughty and Taylor, 2021). The prevalence of neonatal hypoglycaemia was examined by Johnsen et al (2021) in the Norwegian feasibility study. They reported that the first blood sugar was <1.4mmol/l for 14% (n=4) of infants of mothers who had expressed during the week prior to their birth.

Three further studies compared the prevalence of neonatal hypoglycaemia in the intervention versus the usual care group (Forster et al, 2011; 2017; Casey et al, 2019b) and all reported that the prevalence was similar in both groups. Forster et al (2017) reported that hypoglycaemia occurred in less than half of all infants (42% in the intervention vs 46% standard care group). However, more infants in the intervention rather than standard care group in the DAME trial (Forster et al, 2017) received extra (top-up) breastfeeds (36% vs 27%), extra expressed breastmilk (64% vs 51%) and less infant formula (46% vs 59%) for the treatment of hypoglycaemia.

Forster et al (2011) reported the following median blood glucose measurements for the pilot versus standard care groups:

• First: 2.6 mmol/l vs 2.5 mmol/l
• Second: 3.0 mmol/l vs median 3.7 mmol/l
• Third: 2.8 mmol/l vs 3.8 mmol/l.

The primary aim of a retrospective cohort study from Australia was to compare the prevalence of neonatal hypoglycaemia among infants whose mothers did (n=80, 23%) and did not (n=223, 62%) undertake antenatal expression (Casey et al, 2019b). Mother–infant pairs were excluded from the study if the infant was born before 36 weeks gestation (Casey et al, 2019b). Admission to neonatal care was not reported. There was no significant difference in hypoglycaemia among infants born to mothers in the antenatal expression group compared to the non-expression group (40% vs 40%, P=0.89) (Casey et al, 2019b).

Emotional and motivational aspects of antenatal breastmilk expression

Emotional aspects

Emotional and motivational support is required with all aspects of breastfeeding (WHO and UNICEF, 2018). Two qualitative studies (Casey et al, 2019a; Moorhead et al, 2021) and two feasibility studies (Forster et al, 2011; Johnsen et al, 2021) explored emotions experienced by women when undertaking antenatal breastmilk expression. These studies draw attention to differences in individuals' experiences of antenatal expression.

A range of positive and negatives emotions were quantified by Forster et al (2011) from responses to an open question, during a semi-structured telephone interview at 6 weeks postnatal. Participants felt more confident and prepared for breastfeeding (30%) and felt positive about expressing for baby (30%), obtaining a supply of colostrum (26%) and learning how to express (26%) (Forster et al, 2011). On the other hand, some women felt anxious about the volume of colostrum expressed (26%), finding time to express (26%), the time required to express (35%) and experienced discomfort or pain (17%) (Forster et al, 2011). Overall, women were positive about expressing, as 95% reported they would be happy to do so again (Forster et al, 2011). Similarly, Johnsen et al (2021) assessed maternal satisfaction at 6–8 weeks postnatal and reported that 100% (n=27) of participants were satisfied, and all but one would recommend antenatal expression to others.

Two Australian studies (Casey et al, 2019a; Moorhead et al, 2021) using different qualitative research methods revealed that antenatal breastmilk expression can be both emotionally challenging and rewarding. Casey et al (2019a) conducted semi-structured interviews (n=6) and an analysis based on grounded theory and thematic analysis, while Moorhead et al (2021) used a qualitative descriptive study with a purposive sample of women (n=10) who participated in the DAME trial (Forster et al, 2017). Findings emerged around the emotions associated with antenatal expression. These included a concern for fetal, neonatal and maternal health, avoiding formula milk, providing the best for baby, disheartened by failure, dealing with challenges, preparation for breastfeeding, realistic expectations and the need for support while learning and undertaking antenatal expression (Casey et al, 2019a; Moorhead et al, 2021).

Motivational aspects

It is important that women know whether antenatal expression contributes to successful breastfeeding. Johnsen et al (2021) reported a high prevalence of breastfeeding at discharge (86%, n=24) and at 6–8 weeks postnatal (78%, n=21) among the participants who undertook antenatal expression. In addition, four studies compared breastfeeding outcomes in the expression versus the usual care group (Forster et al, 2011; 2017; Soltani and Scott, 2012; Casey et al, 2019b). Overall, the proportion of women who breastfed was higher in the intervention groups, although the difference was not always statistically significant. Results can be clinically significant even when statistically insignificant (Kim et al, 2020). Forster et al (2011) examined use of formula in the first 24 hours as the primary outcome. Fewer infants in the intervention group compared to the audit group received formula in the first 24 hours (40% vs 56%) and received formula while in hospital (63% vs 73%). In addition, more infants in the intervention group compared to the audit group received only breastmilk prior to discharge (60% vs 44%). The differences were not statistically significant (Forster et al, 2011). Similarly, the retrospective cohort study from England found that all infants in the expression group (n=16) breastfed for the first feed, compared to 86% of the non-expression group (n=78), although the difference was not statistically significant (P=0.19) (Soltani and Scott, 2012).

The DAME trial reported a small increase in exclusive breastmilk feeding for infants in the expression group in comparison to the usual care group in the first 24 hours (69% vs 60%), up to discharge (57% vs 49%) and up to 3 months postnatal (60% vs 55%) (Forster et al, 2017). In addition, the retrospective cohort study by Casey et al (2019b) reported a reduced use of formula in the antenatal expression group (P<0.001).

Three studies indicated that there were differences in women's uptake of antenatal expression. Socioeconomic factors, ethnicity and support seemed to play a role, although the reasons were not fully explored. The cohort study from England examined the uptake of antenatal expression in a sample of 94 women, of whom 35 (37%) were advised to express as part of local clinical guidelines (Soltani and Scott, 2012). Less than half (n=16/35, 46%) expressed, and more women with a higher level of education expressed (Soltani and Scott, 2012). Similarly, the retrospective cohort study from Australia reported that among a sample of 303 women with diabetes who were advised to express in the practice setting, 23% (n=80) expressed while 62% (n=223) did not (Casey et al, 2019b). Results on the remaining 15% of the sample were unknown. Fewer women who identified as Aboriginal and/or Torres Strait Islander in comparison to women who identified as Caucasian expressed (P=0.03), and more women who attended a lactation consultant expressed (P<0.001) (Casey et al, 2019b). The DAME trial also reported that among women allocated to the expression group (n=316), 19 (6%) did not express, 25 (8%) expressed for 2–5 times, 80 (25%) 6–19 times, and 134 (42%) ≥20 times (Forster et al, 2017). Expression practices among the remaining 58 (19%) participants were unknown (Forster et al, 2017).

Practical aspects of antenatal breastmilk expression

Timing of expression

It is recommended that women are taught concrete skills to support breastfeeding (WHO and UNICEF, 2018). Many components of the skill of antenatal breastmilk expression have been incorporated in interventions described in the studies reviewed. These include when to start expressing, how to hand express, how to store and transport breastmilk and what to do if concerned (Forster et al, 2011; 2017; Soltani and Scott, 2012; Casey et al, 2019b; Johnsen et al 2021). However, no study examined different interventions, including different approaches, teaching methods, frequency or timing of hand expression. The majority of studies cite 36 weeks as the time antenatal expression commenced. However, some women commenced at earlier or later gestations in the studies reviewed.

Soltani and Scott (2012) explained that women were given information about hand expression in local clinical guidelines which included, ‘the best time to perform antenatal breastmilk expression (36 weeks) and the practicalities of hand expression and storing in sterile containers’. Forster et al (2011) explained ‘each woman was taught how to hand express colostrum and to do this twice daily for approximately 10 minutes until she was admitted to hospital to have her baby’, ‘from 36 weeks of gestation’. The authors reported that guidance was also provided on storage and transport of expressed breastmilk and on precautions (Forster et al, 2011). Women ‘were encouraged to hand express twice per day for no more than 10 min until admission to hospital to give birth, unless any concerns arose that indicated that the intervention should cease’, from ‘36 weeks gestation’ in the DAME trial (Forster et al, 2017). The intervention used in the cohort study by Casey et al (2019b) was described as ‘women were shown how to correctly express colostrum, provided with syringes and labels for collection and storage of colostrum (in the freezer) and advised to transport their collected colostrum in a cooler pack to hospital’, and ‘were encouraged to begin expressing colostrum daily from 34–36 weeks gestation’. Johnsen et al (2021) explained in their feasibility study that ‘women were encouraged to stimulate and hand express colostrum three to five times daily for at least 10 minutes, starting from 37+0 week’. Further research is needed to determine the optimum gestation to commence antenatal breastmilk expression.

Volume of breastmilk

To support women with the skill of antenatal expression, it is important to provide a realistic expectation of the volume of breastmilk expressed. Three studies reported that women were concerned about being able to express sufficient colostrum (Forster et al, 2011; Casey et al, 2019a; Moorhead et al, 2021), although the volume of breastmilk expressed varied widely (Forster et al 2011; 2021). The volume was reported as a median of 39.6 ml (range 5–310 ml) over 24 episodes (range 7–56 episodes) (Forster et al, 2011), and in the DAME trial as a median of 5.5 ml (range 0–905 ml) over 20 episodes (range 1–59 episodes) (Forster et al, 2017). While parity, previous experience of breastfeeding and previous experience of antenatal expression were examined (Forster et al, 2017), future studies might also consider the impact of tandem breastfeeding on volume of milk expressed and other aspects of antenatal expression.

Discussion

The aim was to undertake a narrative literature review to examine how antenatal breastmilk expression supports women with diabetes to breastfeed. The findings were examined within a framework of informational, emotional and motivational, and practical aspects of antenatal breastmilk expression. The framework enabled a more in-depth analysis than an examination of breastfeeding rates alone would allow.

Informational support in this review focused on neonatal wellbeing and providing information to women in this regard, although information on all aspects of antenatal breastmilk expression is essential. The evidence to date consisted mainly of observational and retrospective cohort studies where a concern with neonatal admission was reported. A more recent randomised controlled trial found that for women with diabetes at low risk of complications, their infants were not at greater risk of admission to neonatal care (Forster et al, 2017). While there was no evidence of increased or reduced neonatal hypoglycaemia following antenatal expression, the trial found that infants from the intervention group received expressed milk earlier in the management of hypoglycaemia (Forster et al, 2017). This finding is promising as it suggests that antenatal breastmilk expression may contribute to managing one of the challenges to breastfeeding from maternal diabetes. Further research on the safety aspect of antenatal expression is required as the findings are not generalisable to women at high risk of complications.

A range of emotions were experienced by women when undertaking antenatal expression. There seems to be a fine line between emotions associated with challenges and emotions associated with joy. Ways to support women emotionally have not been fully explored, but studies concluded that when women are asked to undertake this skill, they require ongoing support, reassurance and guidance from healthcare professionals (Casey et al, 2019a; Moorhead et al, 2021). Understanding how women feel is important, as many women do not commence or continue with antenatal expression. Reasons for undertaking antenatal expression therefore require ongoing research, as although the proportion of women breastfeeding was greater in the groups who undertook the intervention in comparison to the groups who did not, the difference was not always statistically significant.

Antenatal expression is a skill that requires teaching and learning. Instructions given on antenatal expression were described to some extent in the studies reviewed. In some studies, instructions were based on guidelines already in place in clinical settings. Future studies might examine different interventions. This is particularly important for timing, as gestation at commencement of antenatal expression varied somewhat between studies. In addition, different approaches to the skill, different frequencies of expression, different methods of teaching and learning require research for women with diabetes.

Limitations

Research on antenatal breastmilk expression for women with diabetes is in its infancy, as many questions remain unanswered. Although eight studies were included in this review, the overall findings are limited by the non-causal methods, which were mostly used when the intervention was tested. The studies included are also limited to geographical region, as the majority originated from Australia (n=6), with one from England and one from Norway. While some studies considered socioeconomic issues, there is scope for further exploration of these issues given the association already established by Victoria et al (2016) between breastfeeding and income, within and between countries.

Conclusions

Antenatal breastmilk expression is recommended to women with diabetes in many clinical settings to reduce the challenges posed by maternal diabetes to breastfeeding. Although results are not conclusive and not generalisable, the evidence is promising and at the very least suggests that research should continue to determine how antenatal breastmilk expression supports women with diabetes to breastfeed.

Key points

• Information on all aspects of antenatal expression is required for women with diabetes and their healthcare professionals.
• Antenatal expression is not necessarily intuitive, therefore women need guidance and support to commence and to continue with expression.
• It is important for women to have realistic expectations about the volume of milk expressed.
• The safety aspects of antenatal breastmilk expression for women with diabetes require ongoing research

CPD reflective questions

• Why is antenatal breastmilk expression promoted?
• How can women be supported to express breast milk antenatally?
• What is the most influential way to educate healthcare professionals on antenatal breast milk expression?
• How can midwives prepare and educate women on the volume of colostrum expected on expression?
• What safety aspects of antenatal expressing should be considered?