In our study, the association between mother's menarche and offspring obesity could remain open to potential residual confounding by the mother's BMI, which was assessed by self-reported height and weight and therefore likely to be underestimated, or by other unmeasured social or behavioural factors related to the mother's obesity. However, the mother's BMI was unrelated to offspring infant weight gain (correlation: r = 0.03) [16], and therefore the associations between mother's age at menarche and her infant's weight gain and growth are unlikely to be confounded by mother's BMI. Age at menarche in the mothers was reported by recall many years after the event. Other studies have shown very good correlations (r = 0.67–0.79) between age at menarche by recall during middle age and the original childhood data [41,42], and the validity of our data in the mothers is supported by the expected strong associations with mother's BMI, and with early menarche in their daughters.
