Abstract
STUDY QUESTION: Is high adiposity in childhood associated with menstrual irregularity and polycystic ovary syndrome (PCOS) in later life? SUMMARY ANSWER: Overall, greater childhood BMI was associated with menstrual irregularity, and greater childhood BMI and waist/height ratio (WHtR) in white but not black participants were associated with PCOS in adulthood. WHAT IS KNOWN ALREADY: Increased childhood BMI has been associated with irregular menstrual cycles and PCOS symptoms in adulthood in two longitudinal population-based studies, but no study has reported on associations with childhood abdominal obesity. Few studies have investigated whether there are racial differences in the associations of adiposity with PCOS though there has been some suggestion that associations with high BMI may be stronger in white girls than in black girls. STUDY DESIGN, SIZE, DURATION: The study included 1516 participants (aged 26-41 years) from the Australian Childhood Determinants of Adult Health study (CDAH) and 1247 participants (aged 26-57 years) from the biracial USA Babies substudy of the Bogalusa Heart Study (BBS) who were aged 7-15 years at baseline. At follow-up, questions were asked about menstruation (current for CDAH or before age 40 years for BBS), ever having had a diagnosis of PCOS and symptoms of PCOS. PARTICIPANTS/MATERIALS, SETTING, METHODS: In CDAH, a single childhood visit was conducted in 1985. In BBS, multiple childhood visits occurred from 1973 to 2000 and race was reported (59% white; 41% black). In childhood, overweight and obesity were defined by international age-sex-specific standards for BMI and WHtR was considered as an indicator of abdominal obesity. Multilevel mixed-effects Poisson regression estimated relative risks (RRs) adjusting for childhood age, highest parental and own education and age at menarche. MAIN RESULTS AND THE ROLE OF CHANCE: The prevalence of childhood obesity was 1.1% in CDAH and 7.5% in BBS. At follow-up, menstrual irregularity was reported by 16.7% of CDAH and 24.5% of BBS participants. The prevalence of PCOS was 7.4% in CDAH and 8.0% in BBS participants. In CDAH, childhood obesity was associated with menstrual irregularity (RR = 2.84, 95% CI: 1.63-4.96) and PCOS (RR = 4.05, 95% CI: 1.10-14.83) in adulthood. With each 0.01 unit increase in childhood WHtR there was a 6% (95% CI: 1-11%) greater likelihood of PCOS. Overall, in BBS, childhood obesity was associated with increased risk of menstrual irregularity (RR = 1.44, 95% CI: 1.08-1.92) in adulthood. Significant interaction effects between race and childhood adiposity were detected in associations with PCOS. In BBS white participants, childhood obesity was associated with PCOS (RR = 2.93, 95% CI: 1.65-5.22) and a 0.01 unit increase in childhood WHtR was associated with an 11% (95% CI: 5-17%) greater likelihood of PCOS in adulthood. In BBS black participants, no statistically significant associations of childhood adiposity measures with PCOS were observed. LIMITATIONS, REASONS FOR CAUTION: The classification of menstrual irregularity and PCOS was based on self-report by questionnaire, which may have led to misclassification of these outcomes. However, despite the limitations of the study, the prevalence of menstrual irregularity and PCOS in the two cohorts was consistent with the literature. While the study samples at baseline were population-based, loss to follow-up means the generalizability of the findings is uncertain. WIDER IMPLICATIONS OF THE FINDINGS: Greater childhood adiposity indicates a higher risk of menstrual irregularity and PCOS in adulthood. Whether this is causal or an early indicator of underlying hormonal or metabolic disorders needs clarification. The stronger associations of adiposity with PCOS in white than black participants suggest that there are racial differences in childhood adiposity predisposing to the development of PCOS and other environmental or genetic factors are also important. STUDY FUNDING/COMPETING INTEREST(S): The CDAH study was supported by grants from the Australian National Health and Medical Research Council (grants 211316, 544923 and 1128373). The Bogalusa Heart Study is supported by US National Institutes of Health grants R01HD069587, AG16592, HL121230, HD032194 and P50HL015103. No competing interests existed.