Breastfeeding in the 21st Century – Authors’ Reply

Victora CG, Rollins NC, Murch S, Krasevec J, Bahl R. Breastfeeding in the 21st century – Authors’ reply. Lancet. 2016 May 21;387(10033):2089-2090. doi: 10.1016/S0140-6736(16)30538-4. PubMed PMID: 27301819.

We appreciate the interest raised by our Series. Dylan Walters and colleagues rightly point out that the standard international indicator for exclusive breastfeeding (proportion of infants aged 0–5 months who are fed exclusively with breastmilk) is distinct from the proportion of infants who are exclusively breastfed until they reach 6 months of age. While we agree that new approaches are needed to estimate the latter from survey data, the former is the only internationally comparable indicator currently available for low-income and middle-income countries, with the advantages of not depending on recall nor relying on modelling. We disagree that the current indicator is flawed—it just represents a different metric.

John Wallingford challenges our estimate of lives saved through improved breastfeeding practices, mentioning that it is based on observational studies. We are unaware of any randomised studies on breastfeeding promotion and mortality, because these studies require large sample sizes and—because compliance with promotion is always imperfect—both intervention and comparison groups include a mixture of feeding modes. Our effect estimate was based on the few studies of mortality according to four categories of breastfeeding:2
exclusive, predominant, partial, or none. There are many other studies comparing breastfed and non-breastfed infants, starting in the early 1900s, all of which show increased risk of death in infants fed either formula or animal milk. Because in low-income and middle-income countries breastfeeding is more common in the poor, confounding will likely reduce the magnitude of the associations; studies with proper adjustment also document a protective effect. Whereas each individual observational study can be potentially flawed, the large effect sizes, dose–response associations with intensity of breastfeeding, and consistency of a large number of studies make the overall level of evidence strong. Wallingford’s non-systematic review of the literature cites a trial from Belarus but fails to cite the other two randomised trials in Mexico and India; all three show that breastfeeding protects against infectious diseases. The HIV study he mentions cannot be extrapolated to uninfected children. A large systematic review showed consistent effects over tens of studies regarding the protection against morbidity.

Maria Quigley and Claire Carson question our estimate of 0·5% prevalence of breastfeeding at 12 months in the UK. We did an extensive review of the medical literature, and none of the published studies reported on breastfeeding at this age. We then wrote to authors of several studies in the UK, and the only estimate (unpublished) we obtained was 0·5% from the UK Millennium Cohort; the authors of this estimate are acknowledged in the appendix of our paper.1
We double checked this low estimate with researchers involved in the UK Infant Feeding Surveys, and they found that the figure was credible. The reference to the 2010 Infant Feeding Survey appendix refers to the estimates of breastfeeding soon after birth and at 6 months; inadvertently, we omitted the information that the 12 month estimate was based on the Millennium Cohort. Based on the same cohort, Quigley and Carson now estimate that this proportion is close to 10%, but they agree that this potential inaccuracy will not affect the main conclusions of our study. We would like to add that the absence of reliable estimates on breastfeeding at 12 months is symptomatic of the low level of interest in such an important behaviour.

Marilyn Agranonik and colleagues cite unpublished results from their birth cohort on how breastfeeding could prevent obesogenic behaviours in children who are born small for age. In the literature review commissioned for this Series,6
we did not find studies reporting on such an interaction, so we look forward to seeing their results published.

Lastly, we fully agree with Mary Renfrew on the need for a transformational shift in the way high-risk newborn babies are managed in many neonatal units. Space limitations precluded us from addressing this topic in our Series.

NCR has received grants from the Bill & Melinda Gates Foundation during the conduct of the study. All other authors declare no competing interests.

Disponivel Em: <https://www.ncbi.nlm.nih.gov/>