1. Timing of menarche and depressive symptoms in adolescent girls from a UK cohort Carol Joinson, Jon Heron 1, Tim Croudace 2, Glyn Lewis, Ricardo Araya ; University of Bristol; Department of Community Based Medicine, 1 Department of Social Medicine, 2 University of Cambridge Avon Longitudinal Study of Parents and Children ALSPAC

2. The current study Research grant from the ESRC to examine risk factors for depressive symptoms in adolescence. Socio-economic disadvantage, adversity, family structure, pubertal timing. Role of timing of menarche in the development of depressive symptoms in girls in early adolescence.

3. Background to the study Pre-adolescence: similar rates of depression in boys and girls. Marked increase in rates of depression in girls during adolescence. Early to mid-adolescence: around twice as many girls than boys with depression.

4. Development of depression from preadolescence to young adulthood Hankin et al. Development of depression from preadolescence to young adulthood: Emerging gender differences in a 10-year longitudinal study. Journal of Abnormal Psychology. Vol 107(1), Feb 1998, 128-140.

5. Gender difference in depression Gender difference persists throughout most of adult life; major concern to public health. Chronic and recurrent course; increased risk of future adverse outcomes. Need for systematic longitudinal research.

6. Why more girls? Hyde et al. The ABCs of depression: Integrating affective, biological, and cognitive models to explain the emergence of the gender difference in depression. Psychological Review. Vol 115(2), Apr 2008, 291-313.

7. Pubertal development and the gender difference in depression Pubertal status= level of physical maturation at a single point in time. Key predictor of rising adolescent depression / emerging gender difference. More mature pubertal status in girls is consistently related to higher rates of diagnosed depression & depressive symptoms.

8. Pubertal stage and depression Angold et al. (1998). Puberty and depression: the roles of age, pubertal status and pubertal timing. Psychol Med, 28, 51-61. More advanced pubertal status is associated with an increased risk of depression in girls

9. Pubertal stage or pubertal timing? Pubertal timing= pubertal status relative to age. Timing of puberty relative to one’s peers may be the critical factor. Higher levels of depression, depressive symptoms & psychological distress among girls who mature earlier than their peers.

10. Assessing pubertal timing Self-assessed level of pubertal development (Tanner). Self reported ratings of perceived pubertal timing relative to peers. Self-assessment of pubertal timing introduces a subjective element, which may be only weakly associated with ratings based on direct physical examinations Age at menarche.

11. Onset of menarche is an important transition point for the female rise in depression Higher levels of depressive symptoms in girls who have experienced menarche compared to those who have not. Girls who experience menarche earlier their peers are more vulnerable to developing depressive symptoms in adolescence. Onset of menarche

12. Association between time from menarche and odds of psychiatric morbidity Patton et al. (1996). Menarche and the onset of depression and anxiety in Victoria, Australia. J Epidemiol Community Health, 50, 661-666. Reference group= girls who had not yet reached menarche

13. Previous studies reporting a relationship between early menarche and depressive symptoms DesignAge (years) nScaleResults Caspi & Moffitt, 1991 Longitudinal cohort study 9-15298DISC-Ct=2.89, p<0.01 Ge et al. 19964-year panel study 12-16200SCL-90-Rt=2.93, p<0.01 Kaltiala-Heino et al. 2003 Cross sectional school survey 14-1619,000BDIOR=2.0 (1.6-2.5) Patton et al. 1996 2-stage cluster sampling 12-172525CIS-ROR=2.2 (1.4-3.5) Rierdan & Koff, 1991 Cross sectional11-13334BDIst=2.76, p <.01 Stice et al. 2001 Longitudinal school survey 11-15496K-SADSB=1.61(0.61, 2.62) t ratios based on planned contrasts of early vs late menarche OR: Early compared to late menarche B coefficient for level of depressive symptoms in early vs non-early menarche groups

14. Limitations of previous studies Cross sectional Small sample size; wide age range Retrospective self-reports of age at menarche - risk of recall bias, particularly in older samples where more time has elapsed since onset of menarche. Heterogeneous samples comprising different ethnic groups. Cultural differences in the association between pubertal timing and depressive symptoms (e.g. absence of association in African American girls). Few have adjusted for potential confounders.

15. Potential confounders of the association between early menarche and depressive symptoms Socioeconomic disadvantage Stressful life events (e.g. family breakdown) High body mass index

16. Study hypothesis Girls who experience onset of menarche at an earlier age than their peers are more likely to report high levels of depressive symptoms in early adolescence.

17. Outcome variable: depressive symptoms Self-reported depressive symptoms at 3 time-points: –10.5, 13 and 14 years Short form of Mood and Feelings Questionnaire Angold et al. (1995). Development of a short questionnaire for use in epidemiological studies of depression in children and adolescents. International Journal of Methods in Psychiatric Research, 5, 237–249. Correlates highly with other well-established depression measures e.g. CDI and DISC Discriminates depressed from non-depressed children in general population samples. Administered via computer in research clinic.

18. Depressive symptoms – Short MFQ 1.I felt miserable or unhappy 2.I didn't enjoy anything at all 3.I felt so tired I just sat around and did nothing 4.I was very restless 5.I felt I was no good any more 6.I cried a lot 7.I found it hard to think properly or concentrate 8.I hated myself 9.I was a bad person 10.I felt lonely 11.I thought nobody really loved me 12.I thought I could never been as good as other kids 13.I did everything wrong Asks respondent to report on the occurrence of depressive symptoms in the past 2 weeks. Each item rated: True (score= 2)/Sometimes (score= 1) /Not at all (score= 0). Maximum score= 26.

19. Explanatory variable: timing of menarche Age at onset of menarche available for 3426 girls in the ALSPAC cohort Regular questionnaires assessing pubertal development (from age 8 years onwards): Onset of menarche 3-level variable: Early onset<11 years 6 months16.7% Normative>=11 years 6 months & <13 years 6 months65.5% Late>= 13 years 6 months17.8%

20. Age at onset of menarche: distribution in the ALSPAC cohort N MeanSDRange Age at onset of menarche342612.471.047.6 - 15.2 years

21. Confounding variables Socioeconomic disadvantage: Maternal reports of major financial problems (0-9 years). 4-level variable: none; only when child was <5 years; only when child was 5-9 years years; when child was 0- 9 years. Family breakdown: Maternal reports of presence/absence of biological father (0-10 years). 3-level variable: father present; father left when child was >=5 years; father left when child was <5 years. Body mass index: Height and weight measured at a research clinic when the children were 9 years old. 3-level variable: (lowest BMI 20).

22. Model examined in the current study

23. Analysis of the association between timing of menarche and depressive symptoms Logistic regression Association between depressive symptoms at 14 years (binary indicator) and timing of menarche. Structural equation model Association between timing of menarche and depressive symptoms at 10.5, 13 and 14 years modelled as three correlated continuous latent factors.

24. Participants Starting sample (n= 7058) Have menarche data (n= 3426) Complete SMFQ at 3 time-points (n= 2501) Menarche + SMFQ (n= 2063) Menarche + SMFQ + confounders (n= 1957) Menarche + SMFQ at 14 yrs (n= 2331) Menarche + SMFQ + confounders at 14 yrs (n= 2130) Sample used for SEM Sample used for logistic model

25. Binary measure of depressive symptoms at 14 years: scale dichotomized with high levels of depressive symptoms>=11 (15.7% of girls). Scores on the SMFQ were highly skewed; majority of the girls reported little or no depressive symptoms. Binary indicator of depressive symptoms Logistic regression used to examine the association between the binary indicator of depressive symptoms and timing of menarche (3 level variable: early, normative, late).

26. Percentage of girls with high levels of depressive symptoms at age 14 years by timing of menarche (n= 2331) % with high levels of depressive symptoms Onset of menarche Chi square= 11.7 p= 0.003 normative

27. Onset of menarcheOdds ratio (95% confidence interval) Late (reference group) 1.00 Normative1.62 (1.15, 2.26) Early2.15 (1.43, 3.24) p= 0.0008 Odds ratios (95% CI) for high levels of depressive symptoms in girls with normative and early onset of menarche compared to late onset: unadjusted model p value for test of linear trend = 0.001

28. Adjusted models Onset of menarche Unadjusted Late1.00 Normative 1.62 (1.15, 2.26) Early2.15 (1.43, 3.24) p= 0.0008 Age at assessment 1.00 1.61 (1.15, 2.25) 2.15 (1.42, 3.23) p= 0.0009 Father absence 1.00 1.62 (1.15, 2.27) 2.13 (1.41, 3.21) p= 0.001 Financial problems 1.00 1.67 (1.18, 2.34) 2.12 (1.40, 3.21) p= 0.0008 Body mass index 1.00 1.64 (1.16, 2.32) 2.10 (1.37, 3.21) p= 0.0015

29. Depressive symptoms at three time points (10.5, 13 & 14 years) were modelled as three correlated continuous latent factors using a CFA model. Each factor was measured by 12 categorical items from the SMFQ. Dummy variables representing timing of menarche (early, normative, late) were added to the CFA model to assess differences in the levels of the latent factor of depressive symptoms at each of the 3 time-points. Structural equation model

30. Fac1Fac2 Item-1Item-12 Fac3 …… Item-1Item-12 …… Item-1Item-12 …… Time 1 (10.5 years) Time 2 (13 years) Time 3 (14 years) Absent father in childhood Financial problems in childhood BMI (age 9) SMFQ Confirmatory Factor Analysis model latent factors of depressive symptoms SMFQ items at each time-point confounders

31. Depressive symptom sum-scores at three time-points across groups defined by timing of menarche (early, normative and late) F = 0.19, p =0.823)F = 8.06, p < 0.001)F = 11.97, p < 0.001)

32. Time 1 (10.5 years) Time 2 (13 years) Time 3 (14 years) Timing of menarche early0.05 [-0.10, 0.20]0.26 [0.10, 0.42]0.23 [0.07, 0.38] normative0.00 ref late0.03 [-0.11, 0.16]0.00 [-0.14, 0.14]-0.22 [-0.36, -0.08] p= 0.795p= 0.004p< 0.001 Unadjusted estimates of differences in depressive symptoms across timing of menarche groups N=2063 with complete data on timing of menarche and depressive symptoms at all 3 timepoints. Regression coefficients represent differences in the level of the latent factor of depressive symptoms at each time-point across the 3-level timing of menarche variable (normative group=reference)

33. Interpretation of the results of the SEM At age 13 years the early-onset menarche group has higher symptom levels than the other two groups. By age 14, there is separation between all three groups in level of depressive symptoms: early menarche > normative, > late onset Findings are consistent with the crude ANOVA results. There is evidence for differences in level of depressive symptoms between the timing of menarche groups at both 13 and14 years.

34. Overall findings from both approaches Earlier age at menarche is associated with increased levels of depressive symptoms in early/mid adolescence. Early timing hypothesis predicts that early maturing girls are more vulnerable to psychological distress than their peers who develop on time (Caspi et al. 1991; Ge et al. 1996).

35. Why are early maturing girls more vulnerable to depressive symptoms? Puberty involves physical, biological, psychological and social changes The timing of these changes relative to one’s peers has implications for adolescent well-being Girls who mature early may be under-prepared for these changes. Developmental deviance perspective: maturing earlier than one’s peers may foster feelings of alienation and depression. Lack of social support of peers going through the same process.

36. Increased psychosocial stressors during puberty Early maturing girls may be faced with stressors / demands that that are inconsistent with their level of cognitive and emotional development: Conflicts with parents Increased social responsibilities Stress and disruption of peer relationships (exclusion, victimisation) Development of romantic relationships Higher emotional arousal Engaging in deviant behaviours

37. Physical changes Earlier increase in adipose tissue compared to peers may be distressing. Associated with changes in body image / heightened body dissatisfaction.

38. Biological changes Hormonal changes that characterize the transition into puberty may be associated with an increased risk for depressive symptoms. Early maturing girls may be less prepared to cope with the mood fluctuations that accompany puberty.

39. Study strengths First UK cohort study of the association between onset of menarche and depressive symptoms. Large sample size. Longitudinal design. Regular reports of onset of menarche from late childhood to adolescence. Well-established method for assessing depressive symptoms. Data on potential confounders.

40. Main limitations Attrition: Individuals included in our study sample were of somewhat higher socio-economic status than the original cohort. This has implications for the generalizability of the findings. Focus on early/mid adolescent period: It is unclear whether early menarche is associated with long-term adverse consequences for emotional development or whether the effect of early menarche on depressive symptoms peaks in early-mid adolescence and attenuates over time. Need further data on depressive symptoms from mid adolescence to early adulthood: this will allow us to examine longitudinally whether the increased risk of depressive symptoms in girls with early menarche persists.

41. Conclusions Early pubertal timing could be an important etiological factor in the emergence of depressive symptoms in girls during adolescence. Early maturing females could be targeted by school and family-based programmes aimed at early intervention and prevention.

42. Future directions Further collection of data on depressive symptoms and diagnosed depressive disorder in the ALSPAC cohort. Longer-term impact of early onset of menarche on rates of depressive symptoms. Potential mediating/moderating variables e.g. stressful life events, conflicts with peers and parents, body image, and self-concept.

43. Acknowledgements My colleagues: Jon Heron, Glyn Lewis, Ricardo Araya, Tim Croudace. The ALSPAC families. The ALSPAC team: interviewers, computer and laboratory technicians, clerical workers, research scientists, volunteers, managers, receptionists and nurses. The Economic and Social Research Council: RES-000-22- 2509. Contact details: Email: Carol.Joinson@bristol.ac.uk