Factors Associated with Breastfeeding Initiation and Continuation: A Meta-Analysis Sarah S. Cohen, PhD1 , Dominik D. Alexander, PhD1 , Nancy F. Krebs, MD2 , Bridget E. Young, PhD, CLC2 , Michael D. Cabana, MD, MPH3 , Peter Erdmann, PhD4 , Nicholas P. Hays, PhD4 , Carla P. Bezold, ScD1 , Elizabeth Levin-Sparenberg, PhD1 , Marco Turini, PhD4 , and Jose M. Saavedra, MD4 Objective To use a quantitative approach to evaluate the literature for quantity, quality, and consistency of studies of maternal and infant characteristics in association with breastfeeding initiation and continuation, and to conduct a meta-analysis to produce summary relative risks (RRs) for selected factors. Study design A systematic review using PubMed and CINAHL through March 2016 was conducted to identify relevant observational studies in developed nations, reporting a measure of risk for 1 or more of 6 quantitatively derived, high impact factors in relation to either breastfeeding initiation or continuation. One author abstracted data using a predesigned database, which was reviewed by a second independent author; data evaluation and interpretation included all co-authors. These factors were summarized using standard metaanalysis techniques. Results Six high impact factors were identified (smoking [39 papers], mode of delivery [47 papers], parity [31 papers], dyad separation [17 papers], maternal education [62 papers], and maternal breastfeeding education [32 papers]). Summary RR from random-effects models for breastfeeding initiation were highest for high vs low maternal education (RR 2.28 [95% CI 1.92-2.70]), dyad connection vs not (RR 2.01 [95% CI 1.38-2.92]), and maternal nonsmoking vs smoking (RR = 1.76 [95% CI 1.59-1.95]); results were similar for breastfeeding continuation. Conclusions Despite methodological heterogeneity across studies, relatively consistent results were observed for these perinatally identifiable factors associated with breastfeeding initiation and continuation, which may be informative in developing targeted interventions to provide education and support for successful breastfeeding in more families. (J Pediatr 2018;203:190-6). Breastfeeding and human breast milk are the normative standards for ideal feeding and nutrition for infants1 with many established short- and long-term benefits.2-19 Exclusive breastfeeding through the first 6 months of age with continued breastfeeding through 12 months is universally recommended, but despite the known health benefits, breastfeeding rates drop precipitously soon after birth. In 2013 in the US, 81% of infants were breastfed at birth, but by age 3 months, only 44% were exclusively breastfed.20 Worldwide from 2007 to 2014, approximately 36% of infants ages 0-6 months were exclusively breastfed.21 To develop interventions that help more women breastfeed, an understanding of the many factors affecting choice of feeding modality is critical. Many maternal and infant factors have been examined in relation to breastfeeding; in particular, maternal obesity has been well-established as a risk factor for poorer breastfeeding outcomes.22 For other factors, however, numerous studies with small sample sizes, diverse populations, and heterogeneity in study design and definitions of outcomes and exposures make it difficult to discern patterns that can inform successful interventions. Therefore, in this study, we performed a comprehensive literature review and quantitative meta-analysis of 6 select factors in relation to breastfeeding initiation and continuation. The factors chosen had not been examined collectively in prior studies and were suitable for meta-analysis: smoking, mode of delivery, parity, dyad separation, maternal education, and maternal breastfeeding education. Methodologic limitations prevented the examination of some potentially important factors in a meta-analytic context, such as socioeconomic status (SES), and this review is P-h P value for heterogeneity RR Relative risk SES Socioeconomic status WIC Women, infants, and children From the 1EpidStat Institute, Ann Arbor, MI; 2Department of Pediatrics—Section of Nutrition, University of Colorado School of Medicine, Denver, CO; 3Department of Pediatrics, Epidemiology, and Biostatistics, University of California, San Francisco, San Francisco, CA; and 4Nestlé Nutrition, Société des Produits Nestlé S.A., Vevey, Switzerland Funded by Nestlé Nutrition in the form of a research contract to EpidStat Institute (to S.C., D.A., C.B., E.L.). Nestlé Nutrition scientists were collaborators on the study and had input in the study design; the collection, analysis, and interpretation of data; the writing of the report; and the decision to submit the manuscript for publication (P.E., N.H., M.T., J.S.). S.C., E.L., and D.A. received research funding from Nestlé Nutrition. P.E., N.H., M.T., and J.S. are employees of Nestlé Nutrition. M.C. served as a consultant for Nestlé Nutrition and is an Editorial Board member for The Journal of Pediatrics. B.Y. and N.K. declare no conflicts of interest. 0022-3476/$ - see front matter. © 2018 The Authors. Published by Elsevier Inc. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). https://doi.org10.1016/j.jpeds.2018.08.008 ORIGINAL www.jpeds.com • THE JOURNAL OF PEDIATRICS