The mean birth weight in singletons was reduced by around 100 g if both parents were employed in the rubber industry, similar for boys and girls, and the risk for a SGA child was doubled when the mother was exposed during the pregnancy. This can be compared to the estimated effect of maternal smoking in our study groups, around 200 g. The smoking effect observed was similar to a previous report from Sweden during the 1980s (Ericson et al. 1991). The perhaps most striking observation was that the offspring sex ratio in female rubber employees was reversed, with fewer boys. It has been

hypothesized that mammalian (including human) sex ratios at birth are partially controlled by the hormone levels of both parents at the time of conception (James 2004). Another hypothesis is a more intensive

early Tubastatin A purchase embryonic selection among males. An altered offspring sex ratio has been observed in populations exposed to persistent organic compounds like PCBs (Karmaus et al. 2002; Mocarelli et al. 2000; Rogan et al. 1999; Rylander CX-6258 in vitro et al.1995), and pesticides (Hanke and Jurewicz 2004), albeit not consistently. Also, a reversed sex ratio has been observed after heavy methyl mercury pollution (Sakamot et al. 2001). The external reference cohort was cross-sectionally defined, in contrast to the rubber workers cohort, which explains the differences in calendar year of births. However, calendar year of birth did not affect the effect estimates, when tested as covariates in multivariate models. It should be noted that all main findings were congruent in the differing comparisons, using external reference cohort, internal reference cohort and within-family comparisons. The proportion of industrial workers being trade union

members has traditionally been very high in Sweden. It has been estimated selleck compound that around 90% of all rubber workers were union members in 2001 (Rosalie Andersson, IF-Metall, personal message). Thus, the differing principles for definitions of the cohorts P505-15 research buy cannot invalidate our findings. We had information on employment as a blue-collar rubber worker during the pregnancy and sperm maturation period. Some of the workers may have been absent from work during this period (i.e. sick leave, vacation), but we have no information on such absenteeism. The Swedish social insurance system with generous benefits for sick leave and parental leaves would tend to keep workers with pregnancy-related medical problems to stay employed. Thus, we do not find it likely that there is differential selection out of the work force during pregnancy between rubber workers and food workers that would affect our findings. The analysis of first-child pregnancies rules out differential selection out of the work force when being a mother.