A large body of evidence indicates that girls are twice as likely as boys to experience depression during adolescence; a finding that is robust across clinical and epidemiological samples of young people, and across different types of assessment (Thapar et al., 2011).

Researchers have looked to the role of pubertal processes in an effort to explain this sex difference (Hamilton et al., 2014). Associations between early menarche (the onset of first menstrual bleeding) and elevated rates of depressive symptoms and clinical depression in girls have been widely reported in the literature (e.g. Joinson et al., 2013; Opoliner et al., 2014). However, it is unclear whether this effect persists into late adolescence and adulthood, and whether this relationship is causal.

Consequently, new research recently published in the British Journal of Psychiatry, led by Maija-Eliina Sequeira (2016), has used Mendelian randomisation analysis to examine the possible causal association between early onset of menarche and depressive symptoms and diagnosed depression in girls during adolescence, and whether this association is still apparent in late adolescence.

Methods

Design

Randomised controlled trials (RCTs), in which participants are randomly allocated to receive or take part in different treatment groups, are a rigorous method for examining possible cause and effect relationships between variables (Sibbald & Roland, 1998). However, an RCT would not be appropriate for use in the present study because participants cannot be randomly allocated to different conditions related to timing of menarche, as this is not something that can be experimentally manipulated. To address this issue, the authors used Mendelian randomisation analysis as a proxy for an RCT design to explore the potential cause and effect relationship between early menarche and depression in adolescents. Mendelian randomisation analysis uses a natural source of randomisation, i.e. the naturally occurring genetic variation in timing of menarche, which simulates the randomisation of participants in a RCT.

Sample

The authors drew on data from the Avon Longitudinal Study of Parents and Children (Golding et al, 2001; Fraser et al, 2013) for their analyses. Pregnant women living in the Avon area of England were recruited to take part in ALSPAC when their expected delivery date was between the 1^st April 1991 and the 31^st December 1992. The ALSPAC dataset includes data on age at menarche for 3,579 adolescent girls, and genetic data for 3,006 adolescent girls (all of White ethnicity), which formed the starting sample for the present study.

Measures

The authors derived data on age at menarche from the following sources:

• Parents in ALSPAC completed a pubertal development questionnaire approximately annually from their child’s 8^th birthday (until their 17^th birthday)
• Adolescents themselves completed a questionnaire at 12 years 10 months and 13 years 10 months, which asked whether they had begun menstruation yet and at what age.

The authors derived data on the adolescents’ levels of depression from the following sources:

• The adolescents in ALSPAC completed the Short Mood and Feelings Questionnaire (SMFQ), a widely used, standardised self-report measure (Turner et al., 2014) at age 14, 17, and 19
• The adolescents in ALSPAC also completed the Clinical Interview Schedule-Revised (CIS-R) at age 18 to assess for the presence of a diagnosis of depression, according to ICD-10 criteria (Patton et al., 1999).

Covariates

The authors controlled for a range of covariates, again drawing on the ALSPAC dataset, related to child characteristics, parent characteristics, and family socioeconomic status (SES) that could, according to the literature, potentially confound or interfere with the relationship between timing of menarche and depression. These included:

• Child body mass index (BMI)
• Maternal antenatal and postnatal depression
• Father absence
• Family social class

Results

• The highest proportion of girls with depressive symptoms at age 14 consisted of those with early timing of menarche
• The lowest proportion consisted of those with late timing of menarche
• However, by age 19 there were negligible differences between the groups in terms of their levels of depressive symptoms.

After adjusting for potential confounding variables, the authors found that:

• 1-month increase in age at menarche was associated with 1% lower odds of having depressive symptoms at 14 years, but this relationship diminished at 17 years
• Moreover, age at menarche was not significantly associated with diagnosed depression at 18 years.

The Mendelian randomisation analyses indicated that there was strong evidence for a causal relationship between earlier menarche and higher levels of depressive symptoms at 14 years:

There will be an additional 6 girls per 1,000 with depressive symptoms, for each month of earlier menarche (p. 5)

However, this relationship was not apparent at age 17 and 19 years for depressive symptoms, nor at 18 years for diagnosed depression.

Strengths and limitations

The strengths of this study include the use of the longitudinal, population-based ALSPAC dataset, which provided the authors with genetic data, timing of menarche data, and data from repeated measures of adolescent depressive symptoms. The use of Mendelian randomisation analysis also allowed the authors to explore the possible causal relationship between early menarche and elevated rates of depressive symptoms, when it was not feasible to use an RCT design to do so.

The limitations of this study include:

• The sample only included girls of White ethnicity, potentially because ethnicity could be another potential confounder of the relationship between timing of menarche and depression (e.g. Karapanou & Papadimitriou, 2010)
• As the authors acknowledge, lack of power due to sample attrition between age 14 and 19 years may have affected the detection of a causal effect of timing of menarche at the later time-points
• In a discussion of the limitations of Mendelian randomisation analysis, Davey Smith and Ebrahim (2003) have cautioned that sample sizes need to be very large in studies using this technique
• While the authors controlled for maternal antenatal and postnatal depression as a possible confounder of the relationship between timing of menarche and adolescent depression in their analyses, the authors did not control for maternal recent or current depression, which could have arguably affected the adolescents’ levels of depressive symptoms (e.g. Garber et al., 2011)

Conclusions and implications

Overall, the authors have provided strong evidence for the possible causal relationship between early menarche and elevated rates of depressive symptoms in girls at age 14. The authors suggest that a possible explanation for this finding could lie in the “inconsistency between levels of biological and cognitive maturation” that girls with early menarche may experience, “as well as feelings of being ‘different’ to one’s peers” (p. 7). Moreover, as the authors acknowledge, the fact that adolescence is a time when young people are exposed to a range of new stressors and experience increases in hormone levels could also contribute to this.

Yet, the authors also found a ‘protective effect’ of late menarche on girls’ levels of depressive symptoms at 14 and 17 years, in that these girls had lower odds of having depressive symptoms compared to girls with normative onset menarche. This seems at odds with the ‘feelings of being different to one’s peers’ explanation given above, as conceivably this would also affect girls with later onset menarche.

The authors did not find evidence that the relationship between early menarche and elevated rates of depressive symptoms continues into late adolescence, as there were no differences in levels of depressive symptoms among the girls according to timing of menarche at this time-point. Possible explanations for this suggested by the authors include that girls with later onset menarche may ‘catch up’ with girls with earlier onset menarche and experience similarly elevated levels of depressive symptoms, or that girls with earlier onset menarche and elevated levels of depressive symptoms may have ‘recovered’ by late adolescence. Disentangling these explanations could be an important avenue for future research in this area to explore.

In terms of the clinical implications of the findings of the present study, the authors highlight the importance for healthcare practitioners of being aware that early menarche can be associated with elevated depressive symptoms in mid-adolescence. It also seems important for parents and teachers to be aware of this relationship. Potentially should an adolescent present with feelings of depression, earlier onset of menarche could be explored as a possible factor contributing to this. This then points to the importance, as the authors also emphasise, of exploring the mechanisms behind this relationship, as it is perhaps difficult to see otherwise how one could intervene to mitigate this factor.

Links

Primary paper

Sequeira, M., Lewis, S. J., Bonilla, C., Davey Smith, G., & Joinson, C. (2016). Association of timing of menarche with depressive symptoms and depression in adolescence: Mendelian randomisation study. The British Journal of Psychiatry, DOI: 10.1192/bjp.bp.115.168617

Other references

Thapar, A., Collishaw, S., Pine, D. S., & Thapar, A. K. (2011). Depression in adolescence. Lancet, 379, 1056-1067. [Abstract]

Hamilton, J. L., Hamlat, E. J., Stange, J. P., Abramson, L. Y., & Alloy, L. B. (2014). Pubertal timing and vulnerabilities to depression in early adolescence: Differential pathways to depressive symptoms by sex. Journal of Adolescence, 37, 165-174.

Joinson, C., Heron, J., Araya, R., & Lewis, G. (2013). Early menarche and depressive symptoms from adolescence to young adulthood in a UK cohort. Journal of the American Academy of Child and Adolescent Psychiatry, 52, 591-598. [Abstract]

Opoliner, A., Carwile, J. L., Blacker, D., Fitzmaurice, G. M., & Austin, S. B. (2014). Early and late menarche and risk of depressive symptoms in young adulthood. Archives of Women’s Mental Health, 17, 511-518.

Sibbald, B. & Roland, M. (1998). Understanding controlled trials: Why are randomised controlled trials important? BMJ: General Practice, 316, 201.

Golding, J., Pembrey, M., Jones, R., & Team, A. S. (2001). ALSPAC – the Avon Longitudinal Study of Parents and Children. I. Study methodology. Paediatric and Perinatal Epidemiology, 15, 74-87. [Abstract]

Fraser, A., Macdonald-Wallis, C., Tilling, K., Boyd, A., Golding, J., Davey Smith, G., … Lawlor, D. A. (2013). Cohort profile: The Avon Longitudinal Study of Parents and Children: ALSPAC mothers cohort. International Journal of Epidemiology, 42, 97-110.

Turner, N., Joinson, C., Peters, T. J, Wiles, N., & Lewis, G. (2014). Validity of the short mood and feelings questionnaire in late adolescence. Psychological Assessment, 26, 752-762. [Abstract]

Patton, G., Coffey, C., Posterino, M., Carlin, J., Wolfe, R., & Bowes, G. (1999). A computerised screening instrument for adolescent depression: Population-based validation and application to a two-phase case-control study. Social Psychiatry and Psychiatric Epidemiology, 34, 166-172. [Abstract]

Karapanou, O. & Papadimitriou, A. (2010). Determinants of menarche. Reproductive Biology and Endocrinology, DOI: 10.1186/1477-7827-8-115

Davey Smith, G. & Ebrahim, S. (2003). ‘Mendelian randomization’: Can genetic epidemiology contribute to understanding environmental determinants of disease? International Journal of Epidemiology, 32, 1-22. [Abstract]

Garber, J., Ciesla, J. A., McCauley, E., Diamond, G., & Schloredt, K. A (2011). Remission of depression in parents: Links to healthy functioning in their children. Child Development, 82, 226-243.

Photo credits

• Lady Fresh CC BY 2.0
• Andres Rodriguez CC BY 2.0
• Eric E Castro CC BY 2.0
• sethdickens CC BY 2.0
• Tricia CC BY 2.0
• Hernán Piñera CC BY 2.0