A swathe of pollutants and indicators of poverty have been linked to changes in the ratio of baby boys to girls born to millions of parents. A study of half the US population and the entire Swedish population examined more than 100 possible factors and found, for example, that mercury, chromium and aluminium pollution correlated with more boys being born, while lead pollution increased the proportion of girls. Proximity to farming also affected the sex ratio, possibly due to higher chemical exposures. Measures of deprivation, such as a high number of fast food restaurants and vacant buildings, were also linked to statistically significant changes in sex ratios, as were indicators of stress, including road deaths and the Virginia Tech shootings. However, other factors such as the season of birth, weather temperature, rates of violent crime and unemployment, did not show significant correlations. The research only demonstrates correlations between the various factors and sex ratios at birth, not cause and effect. Future work to examine the effects of chemicals on human cells or animal models in the laboratory would be needed to show causal links, the researchers said. “This is a list of suspects to investigate, and all the suspects have some credible evidence, but we’re very far from conviction,” said Andrey Rzhetsky at the University of Chicago, who led the research. The sex of babies is determined at conception, when exactly half of embryos should be girls and half boys. But hormonal factors can terminate more female embryos, or more male ones, during pregnancy, leading to skewed sex ratios. “The question is why, and there are many suggested factors, such as stress or something in the environment,” said Rzhetsky. “It makes sense that it can go up and down because the physiology of male and female embryos is not identical. They have different hormonal backgrounds.” “There are a lot of myths about sex ratio and birth, but when you dig into the research, it turns out that everything that was tested on real data was done on relatively small samples [risking spurious correlations], and some statements are not founded in observations at all,” he said. The new research, published in the journal Plos Computational Biology, is the first investigation of numerous chemical pollutants and other environmental factors using large datasets from two continents. It used data on 150 million people in the US over eight years, and data on 9 million Swedish people over 30 years. The factors that linked to significant changes, such as mercury pollution and proximity to industrial plants, saw the sex ratio shift by up to 3%. In a population of 1 million, that would mean 60,000 more girls than boys, or vice versa. Previous research on the impact on sex ratio of toxic pollutants called PCBs were inconclusive, but the new analysis found it significantly increased the number of boys. The researchers also examined two highly stressful events in the US: Hurricane Katrina in 2005 and the Virginia Tech shootings in 2007. They only found a significant change in the sex ratio 34 weeks after the Virginia Tech shootings, with more girls born. To investigate this further, Rzhetsky said, studies would need to focus on those most affected by the events, rather than whole populations. The size of the populations analysed meant the study had strong statistical power, but the scientists noted they did not have data on the sex of stillborn babies and that the US population studied all had private medical insurance, and so were not fully representative of the entire population. Gareth Nye, an expert in pregnancy at the University of Chester, UK, who was not part of the study team, said: “The thinking is that pollutants may influence the very early stage of fertilisation, leading to one sex direction over another. [But] without cellular research, these will always be associations.” “There is no doubt that pollutants play roles in health and disease and that this form of computational research has a role in helping us understand why,” he said. But he said that, while the populations studied were large, the number of US births was relatively small, with the total over 8 years less than the annual average of 3.8m births: “This doesn’t necessarily mean [the study is] of lower quality, but does indicate the conclusions may not be as certain.”