Population-based study on association between birth weight and risk of asthma: A propensity score approach

Background: Previous studies that assessed the role of birth weight in the risk of asthma have been limited because of selection bias and covariate imbalance. Objective: To assess the association between birth weight and risk of asthma by applying a propensity score approach. Methods: The study was designed as a retrospective cohort study based on a birth cohort of children born between January 1, 1976, and December 31, 1979, in Rochester, Minnesota. The propensity scores for birth weight were formulated using 16 covariates. We matched the propensity scores for children having low birth weight (<2,500 g) within a caliper of 0.2 SD of logit function of propensity scores. We calculated the cumulative incidence of asthma between low and normal birth weight groups using the Kaplan-Meier curve. Results: There were 3933 eligible children born between January 1, 1976, and December 31, 1979. Before matching, 13 of the 193 children (6.7%) born weighing 2,500 g developed asthma, whereas 201 of the 3,740 children (5.4%) born weighing 2,500 g and above developed asthma (P=.42). There were significant covariate imbalances between comparison groups. However, after propensity score matching, covariate imbalance was significantly reduced, and children born weighing less than 2,500 g had a similar risk to matched children born with normal birth weight (8.3% vs 7.3%, P=.75). Conclusion: Birth weight is not associated with risk of asthma during the first 6 years of life. The propensity score may be a useful method for observational studies concerning asthma epidemiology.