Menstrual disorders in adolescence: a marker for hyperandrogenaemia and increased metabolic risks in later life? Finnish general population-based birth cohort study.

STUDY QUESTION: Are self-reported menstrual disorders associated with hyperandrogenaemia and metabolic disturbances as early as in adolescence?

SUMMARY ANSWER: Menstrual disorders at the age 16 are a good marker of hyperandrogenaemia, and an adverse lipid profile was associated with higher androgen levels.

WHAT IS KNOWN AND WHAT THIS PAPER ADDS: Hyperandrogenism per se has been suggested to be a significant metabolic risk factor in women and a cause of physical and psychological morbidity in adolescent girls. A weak positive correlation has been described between hyperandrogenaemia and obesity in adolescent girls, but the clinical consequences are still poorly understood. Hyperandrogenism and insulin resistance are also key features of polycystic ovary syndrome (PCOS), and women with PCOS are consequently at an increased risk of developing type 2 diabetes mellitus and/or metabolic syndrome, and may have increased cardiovascular morbidity. Our findings confirm that the association between menstrual disorders, hyperandrogenism, obesity and metabolic risks is already evident in adolescence.

STUDY DESIGN: This population-based, cross-sectional study used postal questionnaires to targeting 15-16-year-old girls in the Northern Finland Birth Cohort 1986 (n= 4567).

PARTICIPANTS AND SETTING: There were 3669 girls who answered the postal questionnaire and out of 3373 girls who also underwent clinical examinations and blood tests, 2448 were included in the analyses. The questionnaire included one question about the regularity and length of the menstrual cycle: 'Is your menstrual cycle (the interval from the beginning of one menstrual period to the beginning of the next period) often (more than twice a year) longer than 35 days?' The girls who answered 'yes' to this question were considered to be suffering from menstrual disorders and were classified as 'symptomatic'. The girls who answered 'no' were defined as 'non-symptomatic'.

MAIN RESULTS AND THE ROLE OF CHANCE: There were 709 (29%) girls who reported menstrual disorders (symptomatic girls) and 1739 who had regular periods (non-symptomatic girls). In the whole population and in both study groups, there were significant correlations between body mass index (BMI) (and waist-to-hip ratio), hyperandrogenaemia and metabolic parameters. Symptomatic girls exhibited significantly higher serum concentrations of testosterone (P= 0.010), lower levels of sex hormone-binding globulin (P =0.042) and higher free androgen indices [FAIs; geometric mean 3.38 (interquartile range (IQR): 2.27, 5.18) versus 3.08 (IQR: 2.15, 4.74), P= 0.002]. The two groups had comparable BMI and insulin sensitivity, and serum levels of glucose, insulin and lipids. There was a significant linear trend towards higher FAI values in the higher BMI quartiles in both symptomatic and non-symptomatic girls. In the whole population, there was a statistically significant linear decrease in high-density lipoprotein concentrations (P < 0.001) and higher triglyceride concentrations (P =0.004) in the upper FAI quartile.

IMPLICATIONS: Information regarding menstrual disorders in adolescence is a good marker of hyperandrogenaemia and may be an early risk factor for the development of PCOS in adulthood. The association between obesity, hyperandrogenism and metabolic risks is already evident in adolescence, which strengthens the importance of noting menstrual disorders at an early stage. BIAS, LIMITATIONS, GENERALIZABILITY: The cross-sectional nature of the study does not allow us to draw conclusions concerning the metabolic risks of this population in later life. The diagnosis of menstrual disorders was based on a questionnaire, suggesting a risk of information bias in reporting the symptoms. This study was not designed to diagnose PCOS, as ultrasonography was not available and there was no clinical evaluation of hyperandrogenism (i.e. hirsutism). However, we were able to take into account potential confounding factors in the analyses.

STUDY FUNDING/COMPETING INTERESTS: This work was supported by grants from the Finnish Medical Society Duodecim, the North Ostrobothnia Regional Fund, the Academy of Finland (project grants 104781, 120315, 129269, 1114194, SALVE), University Hospital Oulu, Biocenter, University of Oulu, Finland (75617), the European Commission (EURO-BLCS, Framework 5 award QLG1-CT-2000-01643) and the Medical Research Council, UK (PrevMetSyn/SALVE). None of the authors have any conflict of interest to declare.