Limited Formula for Breastfeeding Infants: Too Much or Just Enough?
Flaherman VJ, Narayan NR, Hartigan-O'Connor D, Cabana MD, McCulloch CE, Paul IM. The Effect of Early Limited Formula on Breastfeeding, Readmission, and Intestinal Microbiota: A Randomized Clinical Trial. J Pediatr. 2018 May;196:84-90.e1. doi: 10.1016/j.jpeds.2017.12.073. Epub 2018 Mar 14. PubMed PMID: 29550235.(1)
Miren B. Dhudasia, MBBS, MPH
Research Assistant, Division of Neonatology
Children's Hospital of Philadelphia/University of Pennsylvania
Phone: (215) 590-7513
Dustin D. Flannery, DO
Research Fellow, Division of Neonatology
Children's Hospital of Philadelphia/University of Pennsylvania
Phone: (215) 829 5248
Sagori Mukhopadhyay, MD, MMSc
Assistant Professor, Division of Neonatology
Children's Hospital of Philadelphia/University of Pennsylvania
Phone: (215) 510-7916
Breastfeeding, formula, newborn, microbiome
Among term, singleton newborns with weight loss ≥75th percentile on The Newborn Weight Tool, do those supplemented with early limited formula, when compared with those who continue exclusive breastfeeding, have longer duration of breastfeeding, fewer readmissions, and/or demonstrate no changes to intestinal microbiota, at 1 week and 1 month of age?
Design: Randomized controlled trial
Allocation: Allocation was stratified by study site and delivery mode and concealed using a password-encoded Excel spreadsheet generated by an independent statistician.
Blinding: Parents of enrolled infants, providers, and study nurses were not blinded to study group assignment. The research assistant assessing study outcomes on follow-up was blinded to the study group assigned to the participant.
Follow-up period: From birth to 1 month of age.
Setting: Two study sites located in geographically distinct regions: University of California San Francisco (UCSF) Medical Center, San Francisco, CA, and Penn State Milton S. Hershey Medical Center, Hershey, PA
Inclusion criteria: Exclusively breastfeeding, term (≥37 weeks) singleton newborns, born between January 2015 and September 2016, with weight loss ≥75th percentile on The Newborn Weight Tool, whose mothers had not yet begun copious milk production.
Exclusion criteria: 1) Birth weight less than 2500g; 2) medical indication for formula supplementation (including weight loss ≥10% birth weight) or contraindication to breastfeeding 3) infants requiring more than Level I care, not discharged with mother from hospital, or with a diagnosis of narcotic abstinence syndrome 4) infants whose mothers were <18 years old or non-English speaking.
Intervention: Supplementation with early limited formula (ELF), defined as 10 mL of extensively hydrolyzed formula fed with a syringe after each breastfeeding, and stopped at the onset of copious maternal milk production.
Control: Continued exclusive breastfeeding. To balance the time and attention given by the study nurse to the intervention arm, 15 minutes of routine discharge education was provided after enrollment to mothers assigned to control arm.
Outcomes: The pre-specified primary outcome was breastfeeding duration to be assessed at 12 months' age ((https://clinicaltrials.gov/ct2/show/NCT02313181)). The current study reports interim results including the proportion of infants still breastfeeding at 1 week and 1 month of age and other pre-specified secondary outcomes.
Pre-specified secondary outcomes:
Assessed at 1 week and 1 month: neonatal readmissions, the State Trait Anxiety Inventory (State Subscale) (STAI-SS) score, and the Edinburgh Postnatal Depression Scale (EPDS) score.
Assessed at 1 week only: the Breastfeeding Self-Efficacy Scale-Short Form (BSES-SF) score, and the Satisfaction with Maternal and Newborn Health Care Following Childbirth (SMNHC) score.
Exploratory outcomes: Intestinal microbiota assessed at 1 week and 1 month of age.
Analysis and Sample Size:
A sample size of 164 was calculated to provide the study a power of 90% (with α = 0.05) which would allow, 1) to detect a relative risk of 1.4 between the intervention and control arms with respect to the proportion of infants still breastfeeding at 6 months, 2) to detect a 5-point difference between the 2 arms with respect to the STAT-SS score (at 1 week and 1 month), and a 4-point difference between the 2 arms with respect to the SMNHC score (at 1 month).
χ2 test, Fisher exact test and Student t test were used to compare the outcomes between the 2 arms. No sub-analysis or adjusted models were developed. All analyses were intention-to-treat.
Stool samples for 15 subjects (8 from the intervention arm and 7 from the control arm) were analyzed at baseline, 1 week and 1 month of age using 16S rRNA gene sequencing. The 2 groups were compared with respect to specific taxa using the Wilcoxon rank-sum test at 1 week and 1 month, accounting for multiple comparisons.
Outcomes were determined for 152/164 (92.7%) infants, with 8 infants lost to follow-up in the intervention arm and 4 infants in the control arm (Table 1). Compared with controls, mothers assigned to ELF were significantly less likely to be married (66% vs 82%, p = 0.03) and had a shorter intended duration of breastfeeding (8.6 ± 3.4 months vs 9.9 ± 4.4 months, p = 0.049). The 2 groups compared similarly on other demographic and clinical characteristics. Mothers assigned to ELF group used ELF 5.4 ± 3.0 times per day for a median of 2 days (IQR 1-4 days, range 1-7 days). Loss to follow-up did not differ between the 2 groups.
Results from the exploratory outcome of intestinal microbiota: In the sub-sample of 15 infants, the shifts in microbial abundance through 1 month of age were comparable between the 2 arms. Specifically, ELF use did not alter the abundance of Lactobacillus or Clostridia taxa in the intervention group at 1 week or at 1 month of age.
Table 1: Main Results (Adapted from Flaherman et al).
|Outcomes (at 1 month)|||||||
(n = 74)
Cont'd exclusive breastfeeding
(n = 78)
|Breastfeeding (overall), n (%)||64 (86.5)||70 (89.7)||0.53|
Breastfeeding (by center), %
· Penn State
|Formula use in past 24 h, n (%)||30 (45.5)||25 (34.3)||0.18|
|Neonatal readmissions, n||1||5||NR*|
|STAI-SS score, mean ± SD||28.5 ± 8.6||27.3 ± 8.1||0.39|
|EPDS score, mean ± SD||4.2 ± 3.6||3.4 ± 3.1||0.13|
|BSES-SF (at 1 week), mean ± SD||52.6 ± 8.6||52.3 ± 11.5||0.87|
|Modified SMNHC score (at 1 week), mean ± SD||45.7 ± 7.9||47.7 ± 6.7||0.12|
*Not Reported. Readmission rates were compared at 1 week with no readmission in the ELF arm and 4 in the control arm (p 0.06)
The investigators conclude that the targeted use of transient low volume formula supplementation in otherwise healthy infants with pronounced weight loss does not affect rates of breastfeeding or intestinal microbiota patterns through 1 month of age, and potentially reduces the rate of neonatal readmission in the first month.
Exclusive breastfeeding during early life is associated with multiple life-long benefits.(2) In-hospital supplementation with formula during the first few days after delivery is associated with early breastfeeding cessation and decreased rates of exclusive breastfeeding following discharge.(3) There are multiple mechanisms which explain this association.(4) First, providing formula decreases the biological stimuli for breastmilk production by suppressing cues from the infant. Second, formula is more readily available and easy to administer. Third, during the first days after birth, when milk production may physiologically be low, formula administration can foster the perception of inadequacy in mothers, especially in the setting of weight loss in infant, thus negatively impacting further lactation success. Furthermore, there are concerns that formula supplementation affects the infant's developing microbiome, potentially increasing the risk of adverse health outcomes.(5) Given these concerns, among infants with early weight loss, standard management consists of close monitoring of infants along with continued lactation support for mothers. Continued weight loss in such infants, however, is often encountered and may result in increased morbidity such as electrolyte imbalances, heightened bilirubin levels, and ultimately, need for formula supplementation.
Flaherman et al hypothesized that early limited formula supplementation implemented as a standardized regimen can minimize the concerns with early unregulated supplementation, and increase the duration of breastfeeding. A major strength of the study is the carefully designed intervention that addresses each of the concerns with early formula supplementation noted above while providing a practical regimen for implementation. The limited volume of supplementation attempts to minimize the reduction of physiological stimuli from the baby to the mother via feeding cues. The use of hydrolyzed formula instead of regular formula attempts to discourage continuation of formula after discharge, as hydrolyzed formula is more difficult to procure than standard formula. Also, the use of this 'special' product like a medicine rather than as food, has the advantage of being perceived by mothers as a transient intervention, and not as evidence of failure of lactation. This last finding is supported by studies on maternal perception when supplemented by donated breast milk.(6) The later however, is a resource limited product that may not be available across all health care faculties with delivery provisions. The intervention in this study specifically targeted a population of infants who demonstrated weight loss but did not meet supplementation threshold in standard practice, and had pre-defined criteria for stopping supplementation. The investigators also conducted an exploratory comparison of infant's intestinal microflora.
The study did not find a significant difference between the intervention and control arms, with respect to the proportion of infants breastfeeding at 1 week or 1 month, or any of the secondary outcomes reported (Table 1). While the number of readmissions in the control arm were higher at 1 week (n=4) compared to the intervention arm (n=0) the difference was not statistically significant (p = 0.06). The study reports the sample size calculation for proportion of infants still breastfeeding at 6 months.
The primary end point of the study is reported as 1 year at clinicaltrials.gov, suggesting that the current report presents interim results for 1 month. The power calculation reported at clinicaltrails.gov is also for a different end-point of 6 months. These varying reported end-points create difficulty in interpreting some of the findings. This could have been avoided by explicitly stating the intention at clinicaltrials.gov to report interim results at 1 month, and consistently using this time period for the power calculation as well. To clarify the power estimate for the 1-month breastfeeding outcome, we contacted the investigators who communicated that they estimated 114 total subjects would detect a 16% difference in the proportion of infants still breastfeeding at 1 month with a power of 0.8. Thus, if the study is considered adequately powered for the stated effect size, then one can conclude that there is no significant difference between the 2 arms with respect to the breastfeeding rates and the secondary outcomes at 1 month.
The authors concluded from these results that ELF supplementation did not decrease breastfeeding rate as compared to standard management with exclusive breastfeeding and that ELF supplementation has the potential to reduce readmission rates. However, the finding of a lack of difference does not provide proof for equivalency between the interventions. Based on the design of a superiority trial, the most appropriate interpretation of the results would be that ELF supplementation did not increase the proportion of women breastfeeding at 1 month (the study hypothesis). Additionally, the direction of change for the primary outcome contrasts in different subgroups, further limiting equivalency interpretations in a generalized manner. Between the 2 study sites, the proportion of infants still breastfeeding at 1 month in the ELF group was higher in the UCSF site participants and lower in the Penn State participants, compared to control participants at each site (Table 1). Finally, the study was not powered to detect significant differences in readmission rate and a non-significant effect of low event outcomes should be interpreted with caution.
While the study attempts to provide some information for microflora similarities between the cohorts, this important exploratory aim requires further study. Recent information from other groups support the finding that transient short duration in-hospital supplementation when followed by exclusive breastfeeding has minimal effect of the newborn microbiota.(7)
Identifying the right population that may benefit from targeted supplementation, and an appropriate regimen, are important aims that can support breastfeeding families. Although the results of this study do not provide clear evidence to alter the standard of care, they can inform future study design with regards to effect sizes, alternate outcome definitions, and analytic choices.
The Journal Club is a collaboration between the American Academy of Pediatrics - Section of Neonatal Perinatal Medicine and the International Society for Evidence-Based Neonatology (EBNEO.org).
Superiority, equivalence, and non-inferiority randomized controlled trials
Randomized controlled trials are considered the gold standard study design for comparing the effects of interventions. Based on the direction of the hypothesis, there are 3 generally accepted types of trials with differing objectives:
o Superiority trial: to determine a clinically relevant difference between two interventions
o Equivalence trial: to determine whether a new intervention is neither worse nor better than another established intervention
o Non-inferiority trial: to determine whether a new intervention is not inferior to another established intervention
The differences between the 3 types of trials may seem subtle, but have important implications for power and sample size calculations, statistical methods, and interpretation of results.(8) In superiority trials, the standard analysis uses statistical tests of significance to accept or reject the null hypothesis that there is no difference between treatment arms. Flaherman et al hypothesized that early limited formula would improve duration of breastfeeding compared to standard care, suggesting a superiority design.(1) The statistical methods for equivalence and non-inferiority trials are different than for superiority trials: investigators must define a margin of non-inferiority or equivalence a priori, and if the95% confidence interval of the difference does not cross the margin, then the new intervention is deemed to be non-inferior or equivalent, respectively. The testing is two-sided for equivalence trials and one-sided in non-inferiority trials. Interpretation of results from different study designs can be confusing; it is important to remember that a negative result of a superiority trial cannot prove that the studied interventions are equivalent.(9).
1. Flaherman VJ, Narayan NR, Hartigan-O'Connor D, Cabana MD, McCulloch CE, Paul IM. The Effect of Early Limited Formula on Breastfeeding, Readmission, and Intestinal Microbiota: A Randomized Clinical Trial. J Pediatr. 2018 May;196:84,90.e1.
2. Section on Breastfeeding. Breastfeeding and the use of human milk. Pediatrics. 2012 Mar;129(3):e827-41.
3. Ip S, Chung M, Raman G, Trikalinos TA, Lau J. A summary of the Agency for Healthcare Research and Quality's evidence report on breastfeeding in developed countries. Breastfeed Med. 2009 Oct;4 Suppl 1:S17-30.
4. Chantry CJ, Dewey KG, Peerson JM, Wagner EA, Nommsen-Rivers LA. In-hospital formula use increases early breastfeeding cessation among first-time mothers intending to exclusively breastfeed. J Pediatr. 2014 Jun;164(6):1339,45.e5.
5. Madan JC, Hoen AG, Lundgren SN, Farzan SF, Cottingham KL, Morrison HG, et al. Association of Cesarean Delivery and Formula Supplementation With the Intestinal Microbiome of 6-Week-Old Infants. JAMA Pediatr. 2016 Mar 1;170(3):212-9.
6. Kair LR, Flaherman VJ. Donor Milk or Formula: A Qualitative Study of Postpartum Mothers of Healthy Newborns. J Hum Lact. 2017 Nov;33(4):710-6.
7. Forbes JD, Azad MB, Vehling L, Tun HM, Konya TB, Guttman DS, et al. Association of Exposure to Formula in the Hospital and Subsequent Infant Feeding Practices With Gut Microbiota and Risk of Overweight in the First Year of Life. JAMA Pediatr. 2018 Jul 2;172(7):e181161.
8. Hulley SB, Cummings SR, Browner WS, Grady DG, Newman TB. Designing Clinical Research. Wolters Kluwer Health; 2013.
9. Christensen E. Methodology of superiority vs. equivalence trials and non-inferiority trials. J Hepatol. 2007 May;46(5):947-54.