BACKGROUND: Twins have a higher-than-expected risk of infantile hemangiomas (IHs), but the exact reasons for this association are not clear. Comparing concordant and discordant twin pairs might help elucidate these factors and yield more information about IH risk factors. METHODS: A prospective cohort study of twin pairs from 12 pediatric dermatology centers in the United States, Canada, Argentina, and Spain was conducted. Information regarding maternal pregnancy history, family history of vascular birthmarks, zygosity (if known), and pregnancy-related information was collected. Information regarding twins (N = 202 sets) included birthweight, gestational age (GA), presence or absence of IHs, numbers and subtypes of IHs, presence of other birthmarks, and other medical morbidities. RESULTS: Two hundred two sets of twins were enrolled. Concordance for IH was present in 37% of twin pairs. Concordance for IH was inversely related to gestational age (GA), present in 42% of GA of 32 weeks or less, 36% of GA of 33 to 36 weeks, and 32% of GA of 37 weeks or more. Twins of GA of 34 weeks or less were more than two and a half times as likely to be concordant as those of GA of 35 weeks or more (odds ratio (OR) = 2.66, 95% confidence interval (CI) = 1.42-4.99; p < 0.01). In discordant twins, lower birthweight conferred a high risk of IH; of the 64 sets of twins with 10% or greater difference in weight, the smaller twin had IH in 62.5% (n = 40) of cases, versus 37.5% (n = 24) of cases in which the higher-birthweight twin was affected. Zygosity was reported in 188 twin sets (93%). Of these, 78% were dizygotic and 22% monozygotic. There was no statistically significant difference in rates of concordance between monozygotic twins (43%, 18/42) and dizygotic twins (36%, 52/146) (p = 0.50). In multivariate analysis comparing monozygotic and dizygotic twins, adjusting for effects of birthweight and sex, the likelihood of concordance for monozygotic was not appreciably higher than that for dizygotic twins (OR = 1.14, 95% CI = 0.52-2.49). Female sex also influenced concordance, confirming the effects of female sex on IH risk. The female-to-male ratio was 1.7:1 in the entire cohort and 1.9:1 in those with IH. Of the 61 concordant twin sets with known sex of both twins, 41% were female/female, 43% were female/male, and 16% were male/male. CONCLUSIONS: These findings suggest that the origin of IHs is multifactorial and that predisposing factors such as birthweight, sex, and GA may interact with one another such that a threshold is reached for clinical expression.