In this pilot randomized trial involving a sample of nulliparous birthing people in the U.S., we found that a structured AME protocol involving hand expression 1–2 times per day and weekly reinforcement visits with a lactation consultant beginning at 37 weeks of pregnancy was feasible. There was high compliance with daily AME, and most participants visualized and collected antenatal milk. We found no evidence of fetal, infant, or maternal harm associated with AME.
The average volume of milk expressed over the course of study participation and per AME episode was similar to that reported in the DAME Trial, a large randomized trial of AME among women with gestational or preexisting diabetes.21 Among 241 DAME participants with diabetes assigned to receive AME education, the median volume of milk expressed from 36 weeks of pregnancy to birth over a median of 20 expressing episodes was 5.5mL.21 In our study, participants expressed a median of 5.8mL over 15.5 expressing episodes. In the pilot study to inform the DAME Trial, 26 women engaging in AME expressed a median of 39.6mL over 14 days and 24 expressing episodes (1.67mL per expressing episode).37 This was a higher volume than we found in our study, with 0.22-0.88mL expressed per AME episode, dependent on gestational week. Similar to the slight increase in milk volume per expression episode we observed over progressive weeks of pregnancy, Rietveld (2011) also found increases over gestational weeks 35–38 among 11 participants with diabetes engaging in AME.38
We found no evidence that AME was related to adverse safety outcomes, including NICU admissions, uterine hypercontractility, early onset of labor or younger gestational age at birth, lower infant birthweight, or higher rates of cesarean sections or vaginal births involving forceps or vacuum assist. The most common side effects or problems with AME in our sample were transient uterine contractility and increased fetal activity during or directly following AME. This aligns with the findings of the DAME trial; their study team conducted cardiotocography surveillance during expressing episodes, and they documented several occurrences of brief increased uterine activity during AME but no episodes of fetal tachsystole or uterine hyper-stimulation. They also found no differences in gestational age at birth, NICU admissions, Apgar scores, birthweight, or maternal onset of labor or delivery type between those assigned to AME and the control group. Pilot research conducted by the DAME researchers37 and a retrospective cohort study involving 94 women with diabetes, all of whom had been advised to engage in AME beginning at 36 weeks of gestation (16 who engaged in AME),39 found that AME was associated with younger gestational age at birth and increased rate of neonatal admission to special care units/NICUs. However, wide confidence intervals in both studies suggested that the observed differences may have occurred due to chance.40
Our study was not powered to detect between group differences in lactation outcomes. Thus, the trends we observed with regard to higher postpartum breastfeeding self-efficacy and reduced perception of insufficient milk among AME participants may have been due to chance and/or more positive initial attitudes toward breastfeeding and higher breastfeeding self-efficacy within the AME group. Likewise, the small sample size and high rate of obstetric complications may have obscured or blunted any protective effect of AME on exclusive or any breastfeeding.
Our previously published qualitative research from participant interviews conducted at 1–2 weeks postpartum provides some contextualization of the lactation outcomes we observed.36 Overwhelmingly, participants reported that AME enhanced their prenatal and postpartum confidence that they would be able to breastfeed and make sufficient milk. Other researchers have reported similar feedback from pregnant and lactating people who engaged in AME.16,37 However, participants in our study also reported inconsistent and sometimes unsupportive policies at the birth hospital for antenatal milk storage and provision. Several participants were discouraged from using their antenatal milk when supplementation of direct breastfeeding was advised. In addition, several AME participants experienced prolonged separation from their infant in the hospital due to birth complications and NICU admission. In these cases, the total volume of antenatal milk available was not always sufficient to avoid formula use altogether, particularly as the birth hospital at that time did not have a feeding policy in place limiting supplemental feed volumes based on physiologic need.41
Larger studies outside of the U.S. have observed small effects of AME on reducing early formula use. In the DAME Trial, participants allocated to the AME arm had an adjusted relative risk of 1.15 (95% CI: 1.02–1.28) of exclusive breastmilk feeding for the first 24 hours after birth compared to the control group. More DAME participants assigned to AME were also exclusively breastfeeding during the birth hospitalization (57% vs. 49%) and at three months postpartum (60% vs. 55%), though these differences were not significant.21 Similar to the DAME Trial, in a retrospective cohort study conducted at a public hospital in North Queensland, Australia and involving 357 women with diabetes in pregnancy, 80 of whom expressed milk in pregnancy, infants whose mothers engaged in AME were significantly less likely to receive formula during the birth hospitalization versus infants whose mothers did not engage in AME (OR 0.12, 95% CI: 0.05–0.32).24
Anecdotal observations suggest that AME may accelerate the transition to lactogenesis II after birth and contribute to more abundant initial postpartum milk volumes.36 The physiologic basis for such an effect may involve critical period modulations around the time of birth that increase the number of prolactin receptors in breast tissue.42–44 Two randomized controlled trials conducted in India involving 180 and 200 pregnant women (both nulliparous and multiparous), respectively, found those assigned to a group advised to begin expressing milk after 37 weeks of gestation experienced shorter intervals to full lactation where no “top feed” was required versus a control group.20,23 We did not observe any signal in our data indicating more rapid onset of lactogenesis II among AME participants. It is possible that the high rate of birth complications, including preeclampsia, masked any potentiating effect of AME on lactogenesis-II.45 Another consideration is that our sample consisted entirely of nulliparous individuals. On average, lactogenesis II occurs a half-day later and is more likely to be delayed in nulliparous-to-primiparous compared to multiparous individuals.46,47
Any impact of AME on lactation outcomes is likely to be influenced by contextual factors, including timing and format of education. The DAME Trial and other studies on AME provided oral and/or written instructions on AME at study outset between 36 and 37 weeks of pregancy.20,23,27 Casey et al. reported that all women in their retrospective cohort study in Australia were advised to begin AME at 34–36 weeks of pregnancy and were shown how to do AME by a midwife.24 O’Sullivan et al. found that an online instructional video of AME evaluated among 95 pregnant, Australian women was effective in increasing participants’ knowledge and confidence in performing AME.48 In regions where AME is not widely known or practiced, like the U.S., there may be a need to incorporate healthcare provider training and an evaluation of clinical/community infrastructures that may help to scaffold AME education. Our study used a “high touch” approach, where an IBCLC taught and reinforced AME in weekly face-to-face, one-on-one sessions, and we provided introductory education sessions to midwives at the recruitment site. Additional research is needed to examine the optimal format and “dose” of AME needed to achieve measurable impact on lactation outcomes of interest, including proximal outcomes of breastfeeding self-efficacy and breastfeeding satisfaction, while accounting for scalability, costs, community needs, and cultural preferences. Our study team is currently conducting a large randomized trial of remote, telelactation-delivered AME education for nulliparous people in the U.S. with pre-pregnancy body mass indices ≥ 25, who are at risk for adverse lactation outcomes. Details of the trial can be found at: https://clinicaltrials.gov/ct2/show/NCT04258709
In addition to the small sample size, there were several other limitations of this pilot trial. There was potential for bias in intervention delivery, in that the PI also functioned as the lactation consultant providing AME education. In addition, antenatal milk volumes were visually estimated by participants using a 5 mL anchor marking on the collection container; thus there was likely some degree of inaccuracy and imprecision in reported volumes. Finally, generalizability of findings is constrained by convenience sampling from a single midwifery practice, which resulted in a demographically homogeneous group of mostly white, college-educated participants.