Women and girls whose grandfathers or great-grandfathers began smoking before puberty have been found to have increased body fat. The new details have been revealed in research that looked into 30-year-old Children of the 90s study in the United Kingdom.

Past studies in the field have shown exposure of males to certain chemicals before breeding can have effects on their offspring. However, there had been doubts as to whether this phenomenon is present in humans and whether any apparent effects may be more readily explained by other factors.

The study, published in the journal Scientific reports, states that if the paternal grandfather had started smoking pre-puberty, compared with later in childhood (1316 years), his granddaughters, but not grandsons, had evidence of excess fat.

“This research provides us with two important results. First, before puberty, exposure of a boy to particular substances might have an effect on generations that follow him. Second, one of the reasons why children become overweight maybe not so much to do with their current diet and exercise, rather than the lifestyle of their ancestors or the persistence of associated factors over the years,” Professor Jean Golding, lead author of the report said.

In order to investigate the effects of prepubertal exposures in humans, scientists from the University of Bristol studied the effect among 14,000 individuals. They found higher body fat in females whose paternal grandfathers or great-grandfathers had started smoking before age 13 compared to those whose ancestors started smoking later in childhood (age 13 to 16). No effects were observed in male descendants.

In earlier research from 2014, they had found that if a father started smoking regularly before reaching puberty (before 11 years of age), then his sons, but not his daughters, had more body fat than expected. The new research was an extension of the 2014 study.

Researchers said that if the dataset is confirmed in future this will be the first such human demonstration of a transgenerational effect extending across four generations. “If these associations are confirmed in other datasets, this will be one of the first human studies with data suitable to start to look at these associations and to begin to unpick the origin of potentially important cross-generation relationships. It is with great thanks to participants within the Children of the 90s study that we are able to carry out such pioneering research. There is much to explore.” Jean Golding added.