The relation between inter-twin birth weight discordance and total twin birth weight

We tested the hypothesis that the frequency of growth discordance among twins is not related to the uterine capacity for carrying twins.

We counted and compared the frequencies of birth weight discordance of more than 25% in an unlike-sexed twin cohort (n = 1244) and in a population-based twin cohort (n = 7570) across the deciles of the total twin birth weight (twin A + twin B) distribution. The birth order of the heavier twin was noted.

Similar frequencies of discordant pairs were found in both cohorts (11% and 12%, respectively; Mantel-Haenszel chi2 test: P = .131, odds ratio (OR) 0.9, 99% confidence interval (CI) 0.67, 1.11; Woolf test for heterogeneity: two-tailed P = .472). In the discordant pairs, twin A was considerably more often the heavier twin in all birth weight deciles (unlike-sexed cohort: P < 10(-8), OR 5.9, 99% CI 3.0, 11.7; population-based cohort: P < 10(-8), OR 3.1, 99% CI 2.3, 4.0), and in both cohorts (inter-cohort difference: P = .109, OR 1.4, 99% CI 0.83, 2.32). Both cohorts showed a similar nonlinear trend: given that X = decile order, discordance decreased as a function of 22.0 - 6.54 ln[X] for the unlike-sexed twins cohort and 23.0 - 8.18 ln[X] for the population-based cohort, with r values of 0.967.

The more favorable the uterine milieu for carrying twins, the smaller the likelihood of discordant twin growth. Birth order of the heavier twin appears to be an integral part of the discordance phenomenon. The similarity of the cohorts suggests that these conclusions are valid for both like and unlike-sexed twins.