Cow's milk exposure and asthma in a newborn cohort: repeated ascertainment indicates reverse causation

The effect of cow's milk consumption on childhood asthma has been debated for several years. This study attempts to provide further insight into this association through the use of a longitudinal study design. Newborns from parents with atopic history were recruited from Germany, Austria, and England (n = 696). For five repeated ascertainments, information was collected on cow's milk exposure, incidence of doctor-diagnosed asthma, and confounders. Generalized estimation equations, incorporating different models (concurrent, delayed, combined, and reverse causation), were used to determine this association. No association between cow's milk consumption and childhood asthma was found for the concurrent effects model (OR = 0.81, 95% confidence interval [CI]: 0.55, 1.20). In the delayed effects model, the direction of the association varied with time of follow-up. Thus, we stratified by period, which resulted in a significant protective delayed effect at 36 months (OR = 0.18, 95% CI = 0.06, 0.49). However, reverse causation negated this finding since the presence of asthma in prior months led to a reduction in further exposure to cow's milk (OR = 0.40, 95% CI = 0.16, 0.99). Hence, cow's milk consumption does not protect against childhood asthma. The apparent protection of cow's milk against asthma may result from parents of asthmatic children avoiding cow's milk, rather than actual prophylaxis.