Dogaru et al. have provided the best systematic review and meta-analysis to date of published studies of the association between breastfeeding and childhood asthma. Despite careful analysis of the reviewed studies' designs and methodological quality features, the authors are unable to explain the enormous heterogeneity (I2 = 71%-92%) among the reported findings. This heterogeneity likely stems from the fact that “asthma” is a term used to denote a highly variable phenotype. The reasons for the protective association between breastfeeding and such a heterogeneous phenotype remain unclear, but may reflect nonblinding of feeding histories among observers who assess the outcome, as well as residual confounding, particularly by daycare attendance. The absence of a dose-response relationship based on breastfeeding duration or exclusivity also raises questions about the causal nature of the observed association. Future epidemiologic studies of asthma will require better and finer phenotyping to understand its etiology, including the potential protective effect of breastfeeding.