Decreasing sex difference in birth weight.

To the Editor: Among manufactured chemicals, antiandrogenic endocrine disrupters may be especially toxic during fetal life. Indeed, a temporal increase in the prevalence of male genital malformations, including hypospadias and undescended testis, has been observed in recent population-based studies; and geographical differences suggest the role of environmental exposure to antiandrogenic endocrine disrupters. Birth weight has consistently been shown to be higher in boys than in girls. This is apparently due to androgen action: the birth weight of 46 XY individuals with complete androgen insensitivity is similar to that of girls. We hypothesized that if environmental toxicants have disrupted androgen action in male fetuses, the difference between the weight of male and female neonates should have decreased over time. We used graphical methods to examine temporal trends in mean birth weight of all singleton live births from Statistics Canada’s database for the years 1981–2003 (N 5,086,550). Because of a strong temporal increase in preterm birth and decrease in postterm birth (both of which would tend to reduce overall mean birth weight), the primary analysis was restricted to term births (37–41 completed weeks). To provide a control comparison, we analyzed the corresponding temporal trends in differences in birth weight between infants of multiparous and primiparous women. We also carried out multiple linear regression analyses with birth weight as the dependent variable; gestational age, gestational age squared, maternal age ( 20, 20 –34, 35 years), fetal sex, parity (multiparous vs. primiparous), and year were the independent variables, along with sexby-year and parity-by-year interaction terms. A negative sex-by-year interaction indicates that the male-female difference has decreased over the study period. All statistical analyses were carried out using SAS version 8.2 (SAS Institute, Cary, NC). Between 1981 and 2003, mean birth weight for boys increased from 3391 to 3507 g; the corresponding figures for girls were 3248 and 3375 g. Similarly, mean birth weight in infants of multiparous mothers increased from 3368 to 3494 g, whereas the mean for infants of primiparous mothers increased from 3261 to 3375 g. The Figure demonstrates a steady temporal decrease in the male-female difference contrasting with a slight increase in the multiparousprimiparous difference. The multiple linear regression analysis confirmed the graphical results (interaction between sex and year 0.57 g/yr [95% confidence interval 0.69 to 0.44] and between parity and year 0.17 g/yr [ 0.05 to 0.30]). Recent temporal trends in birth weight seem to differ by infant sex. Control for gestational age and for maternal age and parity in the multiple regression analysis helps ensure that temporal changes in those variables do not confound the declining sex difference in birth weight. Comparisons of temporal trends in the sex difference in birth weight between jurisdictions with variable exposure to endocrine disrupters may help to uncover the biologic mechanisms underlying our findings. Guy Van Vliet Department of Pediatrics University of Montreal, Montreal Quebec, Canada [email protected]