Applying Life History Theory to Understand Earlier Onset of Puberty: An Adolescent Brain Cognitive Development Cohort Analysis

Abstract

Purpose:

Life history theory posits that multigenerational exposure to adversity and deprivation influences childhood growth and development, including pubertal maturation. We applied this ecological, evolutionary theory to examine the contributions of distal and proximal adversity on early puberty, a potentially important marker for population health.

Methods:

Baseline data from 5,645 girls in the adolescent brain cognitive development study were included. Early puberty was defined as midlate/post pubertal development by age 9–11 years. The contributions of multigenerational Black/Indigenous (Black, Indigenous and People of Color [BIPOC]) or Hispanic identities, intergenerational mental health, economic deprivation, personal trauma exposure and mental health, and proximal biological factors of premature birth and body mass index on early puberty were examined with hierarchical modeling.

Results:

1,225 girls (21.7%) had early puberty. BIPOC/Hispanic identity, familial adversity, economic deprivation, personal trauma, depression, and a higher body mass index contributed significantly toward early puberty. The effect of multigenerational adversity remained significant across models, but the likelihood of early puberty decreased sequentially for BIPOC and Hispanic youth as proximal adversities were added (e.g., OR decreased from 2.93 to 2.38 for BIPOC youth), supporting a synergistic effect of layered adversity on early puberty.

Discussion:

This analysis supports life history theory as a coherent framework to understand early puberty among girls. Findings suggest monitoring pubertal timing as a population health indictor, like birth weight, prematurity, or life expectancy. Addressing early puberty may require policy and social changes to mitigate the negative impact of multiple layers of adversity including racial/ethnic disadvantage, family, and individual mental health and trauma, as well as economic insecurity.

Early female puberty increases the risk for psychosocial problems [1,2], adverse mental [3] and physical health outcomes [4], and early mortality [5]. Indeed, menstrual cycle factors, including age at menarche, are considered a “vital sign” of female health status [6]. Thus, understanding individual, familial, and population-level factors that contribute to early pubertal onset is essential to improving health outcomes for females.

Exposure to threatening experiences during childhood (e.g., sexual/physical abuse or community violence) is associated with earlier onset or accelerated pubertal development among females [7–11]. The role of resource deprivation is less clear, as some data suggest delayed onset in the presence of neglect and food insecurity [9], while others support a relationship between low socioeconomic status and earlier pubertal onset [12,13]. How aspects of the familial social environment affect pubertal timing remains unclear [14]. Correlation of a mother and daughter’s ages at menarche when their social context is consistent [15] suggests a significant influence, yet the role of intergenerational adversities affecting the familial social context (e.g., parent and grandparent mental health problems) remains understudied. Lastly, earlier pubertal development among Black and Hispanic girls [12,16–18] may be a marker for multigenerational stress.

This pattern of individual, familial, and population-level predictors for early pubertal development may be explained by life history theory (LHT). LHT posits that reproduction may be prioritized over longevity as a strategy to promote species survival under duress [19]. According to LHT, early environmental cues regarding resource availability and threat influence how individuals allocate energy toward growth and development or reproduction [20]. Threatening environments may lead to earlier pubertal onset and reproduction, while resource deprivation causes the opposite effect [19,20]. Hypothesized mechanisms of LHT over multigenerational and intergenerational timeframes include genetic inheritance (i.e., selection), epigenetics (i.e., differential activations or weathering), and allostatic load (i.e., stress-induced changes to neuroendocrine and cardiovascular systems as well as immune and metabolic systems) [21,22].

In studies framed by LHT, intergenerational (i.e., extended family) status reflects a decades-long (e.g., ~40-year) influence on child health status. Multigenerational population status reflects a weightier, centuries-long influence (e.g., ~400 years of colonization for US indigenous peoples and 400 years since the first enslavement of people of African descent in the American colonies). LHT may be particularly salient for Black, Indigenous, and People of Color (BIPOC) girls, who are disproportionality exposed to current deprivation and threat [23]. Indeed, the accumulation of historic cultural deprivation and threat exposures added to current deprivation and threat may exert syndemic effects on pubertal timing for BIPOC youth [24,25].

Yet, BIPOC youth remain underrepresented in puberty research [26]. Studies of “difference” by race without consideration of historic and current contexts may have contributed toward the “othering” of earlier pubertal onset among Black compared to White youth [26], with a tendency toward pathologizing the difference. Through the lens of LHT, early pubertal onset may be reframed as an adaptive response to environmental stressors. An implication of using LHT and adaptation framing is that pubertal timing can be seen as an indicator of health status affected by proximal (individual, familial, and “40-year”) and distal (community, societal, and “400-year”) forces that could be sensitive to and reversed by current efforts to work toward resource and social equity. That is to say, like low birth weight, premature birth, or early mortality, early puberty might have utility as an indicator of population health [27]. The possibility of pubertal onset being sensitive to change can be seen in 20th-century population health research. A one- to 5-month per decade downward trend in age at pubertal onset has been observed across 50 years of National Health and Nutrition Examination Surveys, with greater acceleration among BIPOC versus White females [18]. A stabilization or reversal of this trend could serve as an indicator of success in improving equity, mitigating long-standing wrongs, and improving child and family well-being.

This secondary analysis of baseline data from the Adolescent Brain and Cognitive Development (ABCD) study put LHT to the test as a way to explain early pubertal development. We focused on girls because the pubertal stage is more reliably self-reported in relation to female puberty given the sentinel event of menarche. We hypothesized that multigenerational, intergenerational, and individual-level adversities–as well as child biological factors–would have cumulative effects on age at pubertal onset among healthy female adolescents. We further hypothesized that multigenerational adversities would be the largest contributor because it is already established that BIPOC youth have earlier pubertal onset. Affirming these hypotheses would provide support for LHT and help reframe early-onset puberty as a population-level response to adversity.

Methods and Data

Overview and sample

With permission from the National Institute of Mental Health (Data Use Agreement #19-UFA02777) we conducted a crosssectional secondary analysis of ABCD study data collected at baseline (2016–2018). ABCD is an ongoing longitudinal study of child health and development that includes participants recruited from 21 research sites across the United States. This study was deemed exempt by the University of Michigan Institutional Review Board.

The ABCD sample included 5,681 girls recruited at age 9–10 years using stratified probability sampling of schools selected based on demographic and geographical characteristics to approximate the diversity of the US adolescent population [28,29]. English-fluent children aged 9–10 years were included, while those with contraindications to neuroimaging or with medical or developmental problems that would interfere with their ability to complete study procedures were excluded [28,30]. The ABCD protocols include comprehensive annual assessments of physical and mental health, cognitive functioning, and participant behaviors. For this study, we included girls with complete data for the outcome, early puberty. This resulted in the exclusion of 36 girls, for a final analytic sample of 5,645.

Measures

Outcome: early puberty.

We defined early puberty based on participant’s baseline age and self-reported perceived pubertal development using the Pubertal Development Scale (PDS) [31]. PDS scores for girls are defined as 1-prepubertal, 2 = early, 3 = middle, still premenarche, 4 = late and post-menarche, and 5 = postpubertal [32], based on indicators of pubertal development (e.g., body hair and skin changes, scored from 1 = has not begun yet, 2 = barely begun, 3 = definitely begun, and 4 = seems complete) and onset of menarche (places the girl in category four or five depending on other indicators) [31]. PDS categories were recently validated against ABCD hormonal data (i.e., salivary estradiol, testosterone, and dehydroepiandrosterone) and were found to have a modest to excellent correlation (rho = 0.27–0.98) between child and parent reports in the ABCD study [33].

The distributions of baseline age (rounded to <9, 9, 10, and 11 years) and PDS scores (1–5) were examined using box and whisker plots, and then cross-tabulated to define the lowest quintile of self-reported pubertal stage for age in the sample. Based on these determinations, girls aged ≤ 9–10 in PDS category 4 or 5 (postmonarchal) were classified as having early puberty. Girls ≤ 9 in category three were classified as having early puberty, as the trajectory of maturation indicated they would reach menarche before age 11.

Multigenerational racial/ethnic adversities: BIPOC and Hispanic identities, i.e., “400-year” factors.

The child’s racial and ethnic identities were reported by parents in the baseline demographics survey. Participants were considered BIPOC if parents reported the child’s race as Black or Indigenous (Pacific Islander, Alaska Native, or first nations). Hispanic ethnicity was coded dichotomously (yes/no) per parent response.

Intergenerational familial adversity: two generations’ mental health and substance use problems, i.e., “40-year” factors.

Parents completed a baseline family health history assessing depression, anxiety, alcohol, or drug problems for the child’s mother, father, and up to four grandparents. We summed the total number of reported problems (possible range 0–24). As some children reside in single-parent households where the caregiving parent may not know the other parent’s personal or family health history, we made the conservative decision to replace missing data with zero.

Childhood adversity

Deprivation.

Economic deprivation was a dichotomous variable based on parent responses to a baseline demographics survey assessing if, in the past 12 months, the family had not been able to: afford food; were without telephone service; did not pay the full amount of their rent or mortgage or were evicted; had heat or electric services turned off; or were not able to go to a dentist or doctor because they could not afford it. Participants whose parents responded “yes” to any of these items were coded as having experienced economic deprivation.

Threat.

We included trauma exposure and sequelae (lifetime posttraumatic stress disorder diagnoses and depression symptoms) as proxies for threat. Trauma and mental health histories were assessed at baseline using a computerized version of the parent-completed Kiddie Schedule for Affective Disorders and Schizophrenia (K-SADS) [34]. The K-SADS assesses DSM-5 types of trauma exposures as well as the child’s lifetime (past or present) post-traumatic stress disorder (PTSD) diagnoses and presence of depression symptoms. The K-SADS has demonstrated concurrent validity compared to other validated psychiatric diagnostic tools and test-retest reliability (kappa coefficient for present and/or lifetime major depression 0.78–1.00 and for PTSD 0.60–0.67) [34].

Child biological factors.

Child’s history of premature birth was reported by parents at baseline. The ABCD body mass index (BMI) variable included some implausible values (i.e., <10 and >54), which we recoded as missing.

Statistical analysis

All analyses were conducted with Stata SE (version 17.0, College Station, TX), except where noted. Preliminary analyses and assessments of assumptions for statistical tests and modeling were conducted, and sample characteristics were described for those in the early puberty group versus all others. Chi-squared and independent samples t-tests were used to compare characteristics between the early puberty group and all other girls (not early puberty).

We used a blocked logistic regression to test LHT as an integrative framework for understanding early puberty. As diagrammed in Figure 1, the blocks included factors in the order from distal to proximal, consistent with LHT which considers the toll of adversities accumulated over generations to be a powerful influence in favor of earlier reproduction. Stage 1 examined BIPOC and Hispanic identities as the most distal predictors of early puberty, using White racial and non-Hispanic ethnic identities as comparators. At stage 2, we added a variable summing all three generational (mother, father, and grandparent) histories of mental health and substance use problems. In stage 3, we added the child’s exposure to economic deprivation. Stage 4 added the child’s history of exposure and response to threat (i.e., DSM-5 defined trauma, PTSD lifetime diagnosis, and lifetime presence of depression symptoms). Finally, stage 5 included the most proximal characteristics of premature birth history and current BMI.

Figure 1.

Depiction of theorized levels of influence assessed in the 5-stage model.

Results

Our final analytic sample included 5,645 females, 1,225 (21.7%) of whom were classified into the early puberty group. Full demographic information for the sample and study groups is summarized in Table 1. As shown, rates of exposure to maternal and paternal alcohol, depression, and anxiety problems, as well as paternal drug problems, were higher among girls with early puberty. Importantly, mothers and fathers mental health and drug problems were significantly associated with maternal and paternal grandparents’ problems, respectively (see Figure 2A–C), showing intergenerational patterns. The average number of exposures to these intergenerational familial adversities was significantly higher for girls in early puberty. Finally, significantly more girls in the early group experienced economic deprivation, reported trauma exposure, and lifetime depression compared to other girls.

Table 1.

Description of the sample

Variables	Early puberty (n = 1,225)	Not early puberty (n = 4,420)	Entire sample (n = 5,645)
Age (years)***	9.6 ± 0.47	10.0 ± 0.62	9.9 ± 0.62
400-year factors
BIPOC identity***	363 (29.6)	604 (13.7)	967 (17.1)
Hispanic identity*	277 (22.6)	871 (19.7)	1,148 (20.3)
40-year factors
Mother alcohol problem**	65 (5.3)	184 (4.2)	249 (4.4)
Mother drug problem	63 (5.1)	164 (3.7)	227 (4.0)
Mother depression*	310 (25.3)	975 (22.0)	1,285 (22.8)
Mother anxiety*	104 (8.5)	334 (7.6)	438 (7.8)
Summed mother problems** (Range 0–4)	0.44 ± 0.78	0.37 ± 0.70	0.39 ± 0.71
Father alcohol problem**	167 (13.6)	547 (12.4)	714 (12.7)
Father drug problem*	133 (10.9)	366 (8.3)	499 (8.8)
Father depression*	185 (15.1)	545 (12.3)	730 (12.9)
Father anxiety**	68 (5.6)	182 (4.1)	250 (4.4)
Summed father problems** (Range 0–4)	0.45 ± 0.82	0.37 ± 0.74	0.39 ± 0.76
Summed grandparent problems (Range 0–14)	1.09 ± 1.76	1.05 ± 1.58	1.06 ± 1.62
Summed intergenerational familial adversities* (Range 0-21)	1.98 ± 2.68	1.79 ± 2.37	1.84 ± 2.44
Child factors
Economic deprivation***	355 (29.0)	848 (19.2)	1,203 (21.3)
DSM-5 trauma exposure***	515 (42.0)	1,582 (35.8)	2,097 (37.2)
Child’s lifetime PTSD diagnoses	24 (2.0)	76 (1.7)	100 (1.8)
Child’s lifetime depression symptoms***	271 (22.1)	726 (16.4)	997 (17.7)
Child’s history of premature birth	206 (16.8)	833 (18.9)	1,039 (18.4)
Child’s BMI***	20.2 ± 4.5	18.6 ± 4.2	18.9 ± 4.3

Data presented as n (%) or mean ± standard deviation.

BIPOC = Black and Indigenous People of Color; PTSD = post-traumatic stress disorder; BMI = Body Mass Index.

^*p ≤ .05 level;

^**p ≤ .01,

^***p ≤ .001.

Figure 2.

(A–C) Boxplots of relationships of parent and grandparent sums of mental health/substance use problems. X-axis represents parents’ summed mental health/ substance use problems, ranging from 0 (no problems) to 4 (depression, anxiety, alcohol, and drug use problems).

The results of the theory-based statistical model are in Table 2. In the “400-year” multigenerational model (stage 1), both BIPOC and Hispanic identity were associated with early puberty, supporting an effect of distal adversity. Adding the “40 year” adversity variables (stage 2) significantly improved the model fit, supporting a significant effect of intergenerational familial adversity. Stage 3 (economic deprivation) and stage 4 (trauma exposure and lifetime depression) both affected early puberty. Finally, in stage 5, there was a small but significant effect of BMI, but not premature birth, on early puberty.

Table 2.

Theory-based blocked logistic regression modeling predicting early puberty

Variables	Stage 1: 400-year Racial/Ethnic adversity OR [95% CI]	Stage 2: 40-year familial adversity OR [95% CI]	Stage 3: Childhood economic deprivation OR [95% CI]	Stage 4: Childhood threat OR [95% CI]	Stage 5: Childhood biological factors OR [95% CI]
BIPOC identity	2.93 [2.50,3.43]***	2.98 [2.55, 3.49]***	2.76 [2.34, 3.25]***	2.69 [2.29, 3.17]***	2.38 [2.01, 2.82]***
Hispanic identity	1.45 [1.23,1.71]***	1.47 [1.25, 1.73]***	1.41 [1.20, 1.67]**	1.40 [1.19, 1.65]***	1.31 [1.11, 1.55]**
Summed intergenerational familial adversities		1.05 [1.02, 1.07]***	1.04 [1.01, 1.06]**	1.03 [1.00,1.06]*	1.03 [1.00, 1.06]
Deprivation			1.32 [1.13,1.55]**	1.28 [1.09, 1.50]**	1.20 [1.16,1.57]***
DSM-5 trauma exposure				1.17 [1.02, 1.35]*	1.17 [1.02, 1.41]*
Lifetime PTSD diagnoses				0.85 [0.52,1.40]	0.81 [0.49,1.32]
Lifetime depression Symptoms				1.31 [1.12,1.55]***	1.29 [1.10, 1.53]**
Child’s history of premature birth					0.86 [0.73,1.03]
Child’s BMI					1.05 [1.04,1.07]***
Nagelkerke R2	0.05	0.05	0.06	0.06	0.07
Stage chi 2 (df)	176.43 (2)	187.27 (3)	198.78 (4)	214.72 (7)	261.33 (9)
−2 log likelihood	5,501.61	5,490.77	5,479.26	5,463.32	5,416.70
Hosmer & Lemeshow	0.60	0.71	0.10	0.11	0.001

BIPOC = Black and Indigenous People of Color; PTSD = post-traumatic stress disorder; BMI = body mass index; OR = Odds Ratio; CI = confidence interval.

^*≤.05 level;

^**≤.01,

^***≤.001.

Discussion

Our blocked logistic regression analysis of baseline data from girls participating in the national ABCD Study supported the effect of multi- and intergenerational adversities as well as childhood adversities on early puberty. These findings are consistent with Life History Theory (LHT), which conceptualizes early pubertal development as a population-level response to generational adversity. Specifically, we found that multigenerational (BIPOC and Hispanic identities), intergenerational (greater cumulative parental and grandparental problems), and individual (exposure to economic deprivation, personal history of lifetime trauma exposure, and depression symptoms) adversities were significantly associated with earlier puberty.

Taken together, our findings support the applicability of LHT to the phenomenon of early puberty at the population level. Previous studies have separately shown effects of childhood trauma [9], maternal pubertal history [15], socioeconomic status [35], and race or ethnicity [17] on pubertal development for adolescent girls. Our modeling integrates these factors in a theory-driven approach and shows how multiple layers of adversity from the distal population level (race or ethnicity) to the most proximal child level (BMI) contribute in a negative synergy to early puberty in girls. For instance, the odds of early puberty for BIPOC girls lessened from 2.93 to 2.38 when all layers of adversity were added to the model, showing that a portion of the risk for BIPOC adolescents can be attributed to the synergistic effect of additional adversities and factors. More broadly, the adoption of LHT may help shift the conversation around BIPOC children’s pubertal timing away from deficits-based, overly pathologizing comparisons to White youth and toward the strengths-oriented language of adaptation to stress. Indeed, early puberty may be merely one aspect of an ongoing reconsideration of how BlPOC bodies are viewed and treated in the United States and what health and well-being look like for different communities.

Our findings add support to the notion of considering early puberty as an indicator of overall population health, similar to birth weight or life expectancy [27]. Given that our data support LHT, it may be that early pubertal development is a reproductive survival strategy occurring for females in populations or family groups living with adversity. Considering the well-known adverse health consequences of early puberty, future work is needed to bolster resilient responses and social support for girls with early puberty, identify clinical interventions for those with distress or maladaptive responses, and make necessary changes in policy to reverse or attenuate this effect. Further, it remains important to determine whether the utility of pubertal timing is generalizable across high- and low-income nations and if there is variability within social strata.

Strengths and limitations

Our findings are strengthened by our use of the ABCD database, which includes a large sample of girls from across the nation. Additionally, the theoretical framing of this analysis strengthens the interpretation of data, shedding a broader light on the adversity factors contributing to early puberty for girls than previous studies focusing on sexual abuse or other single-impact approaches. Our layered conceptualization and multi-block modeling underscore the concentrations of adversity risk, particularly among BIPOC girls, thus providing a more complete explanation for early puberty. The ability to generalize findings is, however, limited given a possible selection bias in the ABCD sample away from very poor, institutionalized, or homeless youth [29]. Our coefficients are therefore likely to underestimate the full impact of adversity on girls’ early pubertal development.

The ability to generalize our findings is met with several limitations. First, our findings cannot be extrapolated to boys since our sample included only girls. This was done since menarche is more reliably observable in a database study compared to indicators of male pubertal maturation. Further, we measured puberty using the self-reported PDS, which may be less reliable than clinician reports [36]. Next, we dichotomized puberty categories into early versus not early in relation to age, which reduced precision, introducing error in the direction of nonsignificance. A second limitation is our use of BIPOC and Hispanic identities as proxies for multigenerational adversity. Though this approach assumes a multigenerational history of living in North America, most parents of BIPOC youth in our study (820 out of 967; ~85%) reported that the child, parents, and grandparents had all been born in the United States. While more Hispanic-identifying participants reported that a member of their family was born outside the United States (825 out of 1,148; ~46%), “Hispanic” in the US context generally encompasses those with origins in Latin America—a region largely subjected to European colonization [37]. Finally, we did not consider potentially resilience-promoting factors associated with BIPOC/Hispanic identities, nor experiences of discrimination, which were not assessed at baseline of the ABCD study. We, therefore, advocate for rich measurement of risk and protective factors for people with BIPOC identities in future research.

Implications

Our findings suggest that early puberty in girls is impacted by multi- and inter-generational adversities in a negative synergy, thus, placing many at risk for poor lifespan experiences, morbidity, and early mortality. Early pubertal development represents a disadvantage that may signal systemic inequities that go beyond the health of the individual child. Health research across the 20th century suggests that the age at menarche changes at a readily discernible pace. Thus, recognition of early puberty as an indicator of population health status warrants further consideration. Given the urgency of improving population health and equity, we need a change-sensitive indicator we can use in the more near term. Age at menarche is a health status indicator we can observe sooner for birth cohorts born now who may benefit from current efforts to improve equity. We would not have to wait for data about the infant mortality or birth weight of their newborns or wait to analyze their age at death or years lost to disability.

IMPLICATIONS AND CONTRIBUTION.

Distal multigenerational and proximal, personal adversities are synergistically associated with early female puberty consistent with LHT, which posits that early puberty is an adaptation favoring earlier reproduction to promote survival. This integrative analysis contributes to the consideration of early puberty as a population health indicator.