Childhood Adversity and Pubertal Development among Puerto Rican boys and girls

Abstract

Objective:

Evidence stemming largely from retrospective studies suggests that childhood adversity (CA) is associated with earlier age at menarche, a marker of pubertal timing, among girls. Little is known about associations with pubertal tempo, among boys or racial/ethnic minorities. We examined the association between CA and timing and tempo of pubertal development among boys and girls.

Methods:

The Boricua Youth Study is a longitudinal study of Puerto Rican youth residing in the San Juan metro area in Puerto Rico and the South Bronx, NY. CA was based on caretaker reports of parental loss and parental maladjustment and youth reports of child maltreatment and exposure to violence. Youth completed the Pubertal Developmental Scale (PDS) yearly for three years. In linear mixed models stratified by sex we examined the association between CA and pubertal timing and tempo, adjusting for site, socioeconomic status and age.

Results:

Among the 1949 children who were age 8 or older by wave 3, cumulative CA was associated with higher PDS scores among girls compared to girls not exposed to CA (PDS score: 2.63 (95%CI 2.55, 2.71) vs 2.48 (95%CI 2.37, 2.58). In contrast, among boys, experiencing adversities was associated with lower pubertal developmental stage or later timing (PDS: 1.77 (95%CI 1.67, 1.87) vs 1.97 (95%CI 1.85, 2.10) compared to those not exposed to adversities.

Conclusions:

Associations between CA and pubertal development may vary by sex. Understanding the etiological role of adversities on pubertal development and identifying targets for intervention is of utmost importance in ameliorating the impact of CA on child health.

Introduction

Puberty is a complex, developmental period during which boys and girls become sexually mature and experience physical, biological and psychosocial changes (1). Pubertal development is a continuous process often characterized as a series of maturational stages; these stages serve as the basis of measurement and tracking of pubertal development by researchers and clinicians.(1) Pubertal development can be described in terms of timing and tempo. Timing refers to how physically mature children are compared to their same sex and age peers.(1) It has also been conceptualized as the age of a child at a given stage in pubertal development compared to their same sex-peers. Age at menarche, a stage of pubertal development that is not equivalent to the onset of puberty is often used to characterize timing due to ease of measurement. Thus, much of the existing literature on pubertal development comes from girls, and largely relies on retrospectively reported age at menarche. Tempo, refers to the pace at which children progress through pubertal developmental stages(1) and has been infrequently studied among both boys and girls.

Age of puberty onset continues to decline in the 21^st century,(2) and there is evidence that length of time from onset to completion of puberty is becoming longer. Early or late pubertal onset is a public health concern because both have been linked to a range of unfavorable outcomes throughout the lifecourse, including those occurring in later adolescence such as depression (3), eating disorders (4) and externalizing psychopathology (5), as well as physical outcomes that extend well into adulthood including adiposity (6), cardiovascular disease (7), and cancer (8). Less is known about the effects of pubertal tempo. However, tempo is important to further study since it has also emerged as a risk factor for outcomes related to subsequent behavioral problems (5) and psychopathology (1). Recent studies have shown faster tempo to be associated with psychopathology while slower tempo has been associated with cancer. (1)

Early life experiences, particularly adverse childhood experiences which threaten the child’s bodily, familial, or social safety or security, play a significant role in pubertal development. A growing number of studies have reported a relationship between various types of childhood adversity, and timing of pubertal development (9–12), noting that adversity is associated with earlier pubertal development. For example, studies have examined associations between childhood maltreatment (different types of abuse and neglect) with early pubertal timing (13, 14). Associations between sexual abuse in childhood and pubertal timing are robust. Among children (15–17) and adults (13, 18, 19) with histories of sexual abuse, both developmental changes like breast and skin changes, and age at menarche usually occur at younger ages. The literature on the effects of physical abuse and neglect is not consistent. Physical abuse has been linked with both earlier (18, 19) and later pubertal development (20). Neglect has also been associated with both early and late pubertal development (21, 22) among both boys and girls. Other studies have also reported relationships between high cumulative childhood adversity and later timing of pubertal development. (19, 20, 22) In a prospective study, cumulative adversity was associated with later pubertal development among girls, although the study only examined age at menarche.(20) Among boys and girls in the 1958 British cohort, cumulative abuse was associated with later pubertal development, however early pubertal development was not examined, as pubertal development was only assessed at ages 11 and 16. (22)

There are few studies examining childhood adversities and pubertal tempo. Mendle et al. (2011) studied a sample of 100 girls in foster care and observed that exposure to sexual abuse in childhood was associated with earlier pubertal onset, but physical abuse was associated with faster tempo(15). Negriff et al. noted a slower tempo among boys who experienced neglect but earlier timing of pubertal development among girls who experienced sexual abuse suggesting sex differences in relation to adversity types that alter pubertal timing and tempo.(16) An additional study reported a relationship between low quality parental relationships (i.e., marital conflict, parental depression, family security) and faster pubertal tempo among boys and girls (23).

The few studies that examined racial/ethnic differences in pubertal development(24), identified variations in pubertal development among racial/ethnic minorities; namely that breast development and age at menarche may occur relatively earlier in African-American and Hispanic girls, relative to non-Hispanic White or Asian girls (25, 26). Obeidallah et al. (27) noted that Hispanic girls reached menarche at earlier ages than their White counterparts, however after adjusting for socioeconomic indicators, these age differences disappeared. Only one study has described the relationship between childhood adversity and pubertal development in a Hispanic population, Peruvians(9) living in Peru, and this may not be generalizable to other Hispanic subgroups, such as Puerto Ricans. This is an important limitation since child adversities are relatively more common in the Puerto Rican population, relative to the overall Hispanic population as well as the general US population(28). Our recent work has demonstrated that pubertal development starts earlier among Puerto Rican children compared to average US age for development.(29) In addition, Puerto Ricans are at increased risk for unfavorable mental and physical health outcomes compared to other Hispanic subgroups.(30)

There are several knowledge gaps in the literature. First, most studies have exclusively considered girls and relied on age at menarche when examining the relation between childhood adversity and pubertal development. Second, most studies have not considered effects of childhood adversity on tempo. Lastly, few studies have utilized a comprehensive approach to examine the various forms of childhood adversity (i.e., child maltreatment, household dysfunction, parental separation). We aim to address existing gaps in the literature by examining childhood adversity and the timing and tempo of pubertal development in a population-based sample of Puerto Rican boys and girls.

Methods

Study Sample

The Boricua Youth Study (BYS) is a longitudinal study of multistage probability household samples representative of Puerto Rican children in the South Bronx, New York and the Standard Metropolitan area of San Juan and Caguas, Puerto Rico. The sample has been described in detail previously.(31) Briefly, a household was eligible for the study if: (1) there was at least one child residing in the household aged 5 through 13 years (2) at least one of the child’s parents or primary caretakers residing in the household also self-identified as being of Puerto Rican background. All eligible children in each household were selected to participate, with a maximum of three children per household. In households with more than three eligible children, three were selected at random (using Kish tables). Families participated in the survey after providing consent and assent of children over age 7 years. The sample was followed annually over 3 Waves with over 80% retention rates (2001, N =2491: 2002, N=2286; 2003, N=2187). Children who had not reached age 8 by Wave 3 or who did not have pubertal development information on any wave or who reported logically inconsistent pubertal developmental information (detailed below) were excluded from the analyses leaving 1949 children (1304 families) for the current analysis. Consent forms and procedures were approved by the New York State Psychiatric Institute (NYSPI) IRB and by the IRB at the University of Puerto Rico Medical School.

Childhood Adversity

Lifetime exposure to childhood adversities was reported by caretakers and youth and assessed at Wave 1 using eleven items from validated scales (32–35), and described in detail elsewhere(36). Briefly, four main constructs of adversity were assessed, namely: 1) parental loss (death or divorce/separation), (2) youth reported child maltreatment (neglect and physical, sexual and emotional abuse), (3) parent reported parental maladjustment (intimate partner violence, antisocial personality, substance use problems, emotional problems), and (4) youth reported exposure to violence. The first three constructs encompass the ten types of adversities which are part of the original Adverse Childhood Experiences Study (37) that focused on adverse experiences in the home. We added a fourth construct to include exposure to violence assessed through youth’s report of being a victim of any of several violent acts (38). Each adversity construct was coded as present or absent and scores were then summed across all adversity types. The number of types of childhood adversities was categorized into three groups (0 adversities, 1 adversity, 2 or more adversities). Individual types of adversity were also examined.

Pubertal Development Score (PDS)

Several indicators of pubertal development were assessed through questions from the Pubertal Development Scale (PDS)(39) at the three study waves. The PDS is a measure of perceived pubertal development with good reliability and validity(40). It is widely used in investigations of perceived pubertal development and has been utilized with Latino youth. (41) The pubertal development scale (PDS) is an index of pubertal development that can approximate Tanner staging. Tanner staging is typically assessed via physician exam, it is an established staging system for breast, testis, and pubic hair, with staging raging from no development to full adult maturation.(39) Previous work has established that self-report indicators of growth are also acceptable correlates of Tanner stage.(40, 42, 43)

In boys and girls, two questions regarding secondary body development were assessed: breast growth (girls only), facial hair growth (boys only), and axillary body hair growth (for both sexes). Development on each characteristic was rated on a 4-point scale ranging from 1 (no development) to, 2, 3 and 4 (development is completed). Girls were additionally asked about age of menarche, which was scored dichotomously (1 = has not occurred or 4 = has occurred).

Covariates

Sociodemographic factors were collected based on caretaker’s report: child’s sex and age, maternal education (years) and receipt of public assistance. Socio-cultural context was defined as living in the South Bronx, NY (SBx) or Puerto Rico.

Data Analyses

Examination of PDS scores across the three waves revealed that 392 participants had logically inconsistent developmental information, that is, PDS scores going backwards instead of staying stable or increasing. This is consistent with other studies that have noted pubertal development inconsistencies.(44) Of these 392 participants, 19 had PDS scores that went backwards in two Waves (wave 1> wave 2 and wave 3) and 373 had PDS scores that went backwards at only one wave point (i.e., wave 2<wave 1 or wave 3<wave 2 or wave 3<wave 1). Participants who had logically inconsistent developmental information at 2 waves were excluded from the analyses (n=19). To address the 373 participants whose scores went backwards only once, we imputed PDS scores as follows: first, if two scores in a row were identical, that score was taken for all three waves. Second, we used a carry-backward imputation approach, in which a later PDS score was used to impute the earlier (higher) score. For example, if a participant’s original PDS profile was 3–1–3 for three waves, the imputed trajectory would be 1–1–3.

To compare pubertal timing (PDS score at Wave 1) and tempo (Wave 1–3 PDS slopes) between participants with and without child adversity (CA) at Wave 1, we examined binary Wave 1 CA and PDS scores at each wave using linear mixed effects models, adjusted for site, Wave 1 age and SES (maternal education and receipt of public assistance). An interaction between wave and CA (waves*CA) was examined. Each adversity was tested in a separate model, given the correlation across some of the adversities. To account for the joint impact of the adversities we created a cumulative adversities score.

Given the differential natural history of puberty for boys and girls all models were stratified by sex. Interactions by study site were examined but were not statistically significant.

Several sensitivity analyses were conducted, first excluding all participants with PDS scores that went backwards (n=392 were excluded). A second sensitivity analyses imputed PDS scores that went backwards using a carry-forward direction, i.e, a score of 3–1–3 would be imputed as 3–3–3. Third, analyses were conducted with an alternate grouping of CA; characterizing cumulative childhood adversity into three groups (0 adversities, 1 or 2 adversities, 3 or more adversities). Fourth, analyses were conducted restricting our sample to participants who had complete PDS scores at all three time points (N=1269). Lastly, analyses examining tempo were conducted within the restricted sample (N=1269) adjusting for PDS score at wave 1. All analyses incorporated sampling weights, and were conducted in SAS software (version 9.4, 2013, SAS, Inc., Cary, NC).

Results

Table 1 describes the study sample and shows the distribution of PDS scores for boys and girls for the three study waves. Girls had higher PDS scores than boys at each study wave. Most childhood adversities were relatively common. Physical abuse, exposure to violence, and endorsing two or more childhood adversities were more common in boys than in girls.

Table 1.

Prevalence of childhood adversity and pubertal developmental score (PDS) by sex, Boricua Youth Study, children (Mean age 10, N=1949)

	Boys	Girls
	N = 1007	N = 942
Wave 1 PDSa*	2.12±0.89	2.95±1.04
Wave 2 PDSa*	2.33±0.97	3.09±1.06
Wave 3 PDSa*	2.61±1.03	3.31±1.05
Sociodemographic factors	N (%)	N (%)
Site (NY)	446 (44.3)	427 (45.1)
Public assistance	430 (45.8)	385 (43.1)
Maternal education (Years) (M, SD)	11.78±2.92	11.86±2.71
Individual Adversitiesb	N (%)	N (%)
Loss of a parent
Death	51 (4.5)	45 (4.3)
Divorce/separation	505 (51.1)	489 (53.1)
Child maltreatment
Neglect	110 (10.6)	75 (8.7)
Physical abuse*	200 (20.7)	96 (10.3)
Sexual abuse	49 (5.0)	38 (3.9)
Emotional abuse	154 (14.9)	124 (13.0)
Parental maladjustment
Intimate partner violence	29 (3.2)	27 (2.6)
Antisocial personality	128 (13.0)	141 (15.0)
Substance use problems	165 (16.3)	143 (14.3)
Emotional problems	261 (26.0)	248 (26.2)
Environmental stressors
Exposure to violence*	242 (24.2)	132 (14.0)
Cumulative Child Adversity	N (%)	N (%)
None	181 (18.9)	215 (23.9)
One	292 (29.2)	303 (32.9)
Two or more*	526 (51.9)	419 (43.3)

^*P-value < 0.05 for sex differences adjusted for SES

^aMean±SD is reported for PDS at each wave.

^bAdversities reported at Wave 1

Cumulative Childhood adversity.

In adjusted analyses, girls experiencing two or more adversities had a higher pubertal developmental stage at Wave 1 or earlier timing compared to girls not exposed to adversities [PDS: 2.63 (95%CI 2.55, 2.71)] vs 2.48 (95%CI 2.37, 2.58)] (Table 2A). Among boys, experiencing one adversity was associated with lower pubertal developmental stage or later timing [PDS: 1.77 (95%CI 1.67, 1.87) vs 1.97 (95%CI 1.85, 2.10)] compared to those not exposed to adversities (Table 2B). Experiencing two or more adversities was also associated with lower pubertal developmental stage among boys albeit this was not statistically significant [PDS: 1.87 (95%CI 1.79, 1.94) vs 1.97 (95%CI 1.85, 2.10)]. Regarding tempo, no statistically significant associations were noted between cumulative childhood adversity and pubertal tempo among girls and boys.

Table 2A.

Childhood adversity and mean pubertal development scores (PDS) among girls, for those with and without childhood adversity (CA) in the Boricua Youth Study (N=942), linear mixed effects models

	W1 PDS score					PDS slope
	With CA		Without CA			With CA		Without CA
	Score	95% CI	Score	95% CI	p-value	Slope	95% CI	Slope	95% CI	p-value
Cumulative CAs
1 CA vs 0 CA	2.52	(2.43–2.6)	2.48	(2.37–2.58)	0.5717	0.42	(0.38–0.46)	0.37	(0.32–0.42)	0.1146
2 or more CAs vs 0 CA	2.63	(2.55–2.71)	2.48	(2.37–2.58)	0.0225	0.33	(0.29–0.36)	0.37	(0.32–0.42)	0.1984
Individual CAs
Loss of a parent
Death	2.6	(2.35–2.85)	2.55	(2.5–2.6)	0.7028	0.45	(0.32–0.57)	0.36	(0.34–0.39)	0.1969
Divorce or separation	2.58	(2.51–2.65)	2.52	(2.44–2.59)	0.2131	0.37	(0.34–0.41)	0.36	(0.33–0.4)	0.6405
Child maltreatment
Neglect	2.62	(2.43–2.81)	2.54	(2.49–2.6)	0.4479	0.41	(0.31–0.5)	0.36	(0.34–0.39)	0.3886
Physical abuse	2.58	(2.42–2.75)	2.55	(2.5–2.61)	0.7483	0.29	(0.21–0.37)	0.37	(0.35–0.4)	0.0577
Sexual abuse	2.67	(2.4–2.93)	2.55	(2.5–2.6)	0.4048	0.27	(0.14–0.4)	0.37	(0.35–0.4)	0.1374
Emotional abuse	2.66	(2.52–2.8)	2.54	(2.48–2.59)	0.1110	0.26	(0.19–0.32)	0.39	(0.36–0.41)	0.0004
Parental maladjustment
Intimate partner violence	2.54	(2.25–2.83)	2.56	(2.5–2.61)	0.9172	0.34	(0.2–0.49)	0.37	(0.34–0.39)	0.7330
Antisocial personality	2.69	(2.56–2.83)	2.53	(2.48–2.58)	0.0270	0.31	(0.25–0.38)	0.38	(0.35–0.4)	0.0717
Substance use problems	2.62	(2.49–2.76)	2.54	(2.49–2.6)	0.2848	0.38	(0.31–0.44)	0.37	(0.34–0.39)	0.7826
Emotional problems	2.62	(2.52–2.72)	2.53	(2.47–2.59)	0.1389	0.35	(0.3–0.4)	0.37	(0.35–0.4)	0.3375
Environmental stressors
Exposure to violence	2.57	(2.43–2.71)	2.55	(2.5–2.61)	0.8220	0.37	(0.3–0.43)	0.37	(0.34–0.39)	0.9385

W1= Wave 1; PDS=Pubertal Developmental Score; CA=Child Adversity

Noted in bold P < 0.05

All models adjusted for site, baseline age and SES, stratified by sex

Table 2B.

Childhood adversity and mean pubertal development scores among boys, for those with and without childhood adversity in the Boricua Youth Study (N=1007), linear mixed effects models

	W1 PDS score					PDS slope
	With CA		Without CA			With CA		Without CA
	Score	95% CI	Score	95% CI	p-value	Slope	95% CI	Slope	95% CI	p-value
Cumulative CAs
1 CA vs 0 CA	1.77	(1.67–1.87)	1.97	(1.85–2.1)	0.0142	0.4	(0.35–0.45)	0.41	(0.35–0.47)	0.7910
2 or more CAs vs 0 CA	1.87	(1.79–1.94)	1.97	(1.85–2.1)	0.1600	0.39	(0.35–0.42)	0.41	(0.35–0.47)	0.4819
Individual CAs
Loss of a parent
Death	1.82	(1.56–2.07)	1.86	(1.8–1.91)	0.7575	0.41	(0.28–0.53)	0.4	(0.37–0.42)	0.8732
Divorce or separation	1.83	(1.75–1.9)	1.88	(1.8–1.96)	0.3354	0.39	(0.35–0.43)	0.4	(0.36–0.44)	0.7314
Child maltreatment
Neglect	1.93	(1.76–2.1)	1.85	(1.79–1.91)	0.3887	0.34	(0.26–0.43)	0.4	(0.37–0.43)	0.1893
Physical abuse	1.84	(1.72–1.96)	1.86	(1.8–1.92)	0.7449	0.37	(0.31–0.43)	0.4	(0.37–0.43)	0.3556
Sexual abuse	1.91	(1.67–2.15)	1.86	(1.8–1.91)	0.6708	0.36	(0.25–0.48)	0.4	(0.37–0.43)	0.5812
Emotional abuse	1.93	(1.79–2.08)	1.85	(1.79–1.9)	0.2651	0.34	(0.27–0.41)	0.41	(0.38–0.43)	0.1055
Parental maladjustment
Intimate partner violence	2.12	(1.82–2.42)	1.84	(1.79–1.9)	0.0759	0.32	(0.17–0.46)	0.4	(0.37–0.43)	0.2692
Antisocial personality	1.77	(1.62–1.92)	1.87	(1.81–1.92)	0.2515	0.42	(0.34–0.49)	0.39	(0.36–0.42)	0.5171
Substance use problems	1.86	(1.73–2)	1.85	(1.79–1.91)	0.8737	0.46	(0.39–0.53)	0.38	(0.36–0.41)	0.0534
Emotional problems	1.78	(1.67–1.89)	1.88	(1.82–1.94)	0.1049	0.41	(0.36–0.47)	0.39	(0.36–0.42)	0.4105
Environmental stressors
Exposure to violence	1.92	(1.81–2.03)	1.84	(1.78–1.9)	0.1993	0.33	(0.28–0.39)	0.41	(0.38–0.44)	0.0126

W1= Wave 1; PDS=Pubertal Developmental Score; CA=Child Adversity

Noted in bold P < 0.05

All models adjusted for site, baseline age and SES, stratified by sex

Individual Childhood Adversity.

On pubertal timing, girls exposed to parental antisocial personality had higher initial PDS scores compared to those not exposed [PDS 2.69 (95%CI 2.56,2.83) vs 2.53 (95%CI 2.48, 2.58)]. Among boys, individual adversities were not statistically significantly associated with pubertal timing.

On pubertal tempo, girls who experienced emotional abuse [PDS slope: 0.26 (95%CI 0.19, 0.32) vs 0.39 95%CI 0.36, 0.41)] had smaller increases in PDS over time indicating a slower pubertal progression or slower tempo. Boys who experienced violence exposure had slower pubertal progression (slower tempo) compared to boys not exposed [PDS slope: 0.33 (95%CI 0.28, 0.39) vs 0.41 (95%CI 0.38, 0.44)].

Figure 1 shows the adjusted trajectories of pubertal development in boys and girls stratified by childhood adversity. We noted no statistically significant interactions by site (Puerto Rico vs Bronx, NY).

Figure 1.

Childhood adversity and mean pubertal development scores (adjusted) across study waves, among boys and girls in the Boricua Youth Study (N=1949).

Sensitivity Analyses

When excluding participants who reported any backward PDS scores (N=392) (Supplemental Table 1) or when implementing a carry-forward imputation (Supplemental Table 2), we note consistent findings in relation to magnitude and direction as initial results, albeit some results were not statistically significant. We also note consistent findings when using an alternate characterization of cumulative childhood adversity (Supplemental Table 3).

When restricting the sample to participants with PDS scores at all 3 waves (647 girls, 622 boys) associations between childhood adversity and pubertal tempo were consistent with findings from the main analysis (Supplementary Table 4). Findings from models examining tempo, adjusting for Wave 1 PDS scores were also consistent with main analyses (Supplementary Table 5).

Discussion

In this sample of Puerto Rican children, we note that experiences of childhood adversity are differentially associated with timing and tempo of pubertal development among boys and girls. Cumulative experiences of childhood adversity were associated with earlier timing of pubertal development for girls but later timing of pubertal development for boys. Among both boys and girls, individual adversities were associated with slower progression (tempo) of pubertal development.

The association between social risk factors and girl’s pubertal development has been extensively studied, generally relying on age at menarche. Several studies have noted that adversity in childhood is associated with earlier timing of pubertal development.(45) However, only a handful of studies have examined the cumulative impact of different types of childhood adversity on pubertal development. Findings have been mixed, for example in the National Child Development Study, greater numbers of childhood adversities was associated with later age at menarche.(20) In contrast, and consistent with our findings, results from the National Comorbidity Survey-Replication note that increasing number of adversities is associated with early menarche(12). Noted discrepancies may be explained by a reliance on menarche data, which does not incorporate other measures of pubertal development. We expand upon these studies to define pubertal development with respect to timing and tempo. The direction, adversities predicting earlier timing but slower tempo, and magnitude of our findings are consistent with previous studies examining individual adversities, such as child maltreatment, among girls. Expanding on prior studies, we note parental antisocial personality is associated with earlier pubertal development among girls which may reflect general household dysfunction encompassing cumulative adverse experiences(46).

Pubertal development among boys has been less studied. Two studies have noted childhood adversity to be associated with later pubertal development, among boys.(22) This is consistent with our findings which note cumulative adversity to be associated with later timing of pubertal development. The high co-occurrence of adversities in this sample suggests that examining the cumulative measure of adversity may be a more appropriate characterization of adversities in this population.

Only a handful of studies have examined associations between adversities and pubertal developmental tempo, noting mixed findings. Our findings that adversities are associated with slower tempo among boys is consistent with a longitudinal study of children referred to the department of family services and a comparison group, which noted maltreatment to be associated with slower pubertal tempo (16) among boys. In contrast, two other studies have noted child abuse and low quality parental environment to be associated with faster tempo among boys and girls.(15, 23) Discrepant findings could be due to variations in sample selection, size, age and assessment of adversities. Future work examining pubertal tempo among diverse populations is warranted.

The coupling of Hypothalamic-Pituitary-Adrenal Axis (HPA) and Hypothalamic-Gonadal (HPG) axis to modulate pubertal development among girls is one potential mechanisms by which childhood adversity can impact pubertal development. Child adversities have been shown to dampen the Hypothalamic-Pituitary-Adrenal Axis in numerous studies.(47) In turn this dampening of HPA axis which results in lower cortisol excretion has been shown to accelerate pubertal development among girls. (16) In addition, among girls childhood adversities have been shown to increase adiposity in childhood, (48) which in turn is associated with earlier pubertal development. However, given the lack of anthropometric data, we are unable to examine mediators such as body mass index (BMI) which could partly explain the relation between adversities and pubertal development. As others have noted, few studies have examined the timing and tempo of pubertal development among boys and as such conceptual models that can guide our understanding of antecedents and consequences of the timing and tempo of pubertal development have been developed among girls.(24) A differential impact of child adversity on pubertal timing and tempo among boys and girls could be due to a number of factors. First, it is plausible that the differential impact noted in our study, with boys having slower development in response to adversity, may be due to the differential development timing among boys and girls. Girls pubertal development occurs earlier than boys, thus its plausible that accelerated development in response to adversity is noted at a later age for boys compared to girls. For example, there is evidence that accelerated biological aging (shorter telomere, accelerated epigenetic aging) is greater among adult men compared to women (49, 50) suggesting accelerated aging processes may occur at developmental different periods for men and women. In our study the average age for boys and girls is similar precluding us from examining pubertal development at different time points for boys and girls. Second, the differential puberty and adversity association may be due to the differential impact of HPA axis dysregulation on pubertal changes. Attenuated cortisol in response to chronic stress has been associated with greater pubertal changes among girls, however this association has not been noted among boys.(47) Lastly, it’s also plausible that the subjective nature of pubertal development among boys compared to the more objective measure (age at menarche) of pubertal development among girls may bias associations towards slower rate of pubertal progression among boys.

The fact that the relationship between adversities and pubertal development in boys seems to be different from girls accentuates the need for further investigation focused on boys, which could corroborate our results and, most importantly, bring insights about mechanisms that may drive these associations.

Some limitations should be considered. While the Pubertal Developmental Scale is a validated measure to assess pubertal development, other secondary body characteristics included in the Tanner scale were not included as part of the assessment. Furthermore, an objective measure of pubertal development was not utilized, and we relied on children’s self-report of their pubertal development. This resulted in some children reporting inconsistent or backwards PDS scores. To address this, we used several imputation methods, which in addition to utilizing available PDS scores to impute logically consistent PDS scores over time, preserved the sample size, thus improving efficiency. In sensitivity analyses we note our findings are robust to different methods of imputation. Given that we have information on the timing for some but not all adversities we could not explore sensitive periods related to the timing of experiences.

This study also has several strengths. Given the longitudinal study design we are able to examine trajectories of pubertal development over three years of follow-up. Furthermore, we examine a comprehensive number of childhood adversities, as well as the accumulation of adversities. An additional strength is the prospective assessment of adversities using validated scales, future exploration of sensitive time periods would be an important area of research. We expand upon the literature by examining timing and tempo, which has only been examined in a handful of studies, and do so within a large sample of Puerto Rican boys and girls, a population at high risk of experiencing childhood adversities and may thus be particularly vulnerable to the impact of adversities on pubertal development.

In this large sample of Puerto Rican children, we note childhood adversities are associated with pubertal development among boys and girls. Early pubertal onset is associated with depression, substance abuse, and delinquency in adolescents, and faster tempo may be associated with internalizing and externalizing problems, and depression. Our findings underscore the importance of examining the timing of the full spectrum of puberty. Understanding whether modifiable factors mediate the association between adversities and pubertal development may inform the development of interventions that would ameliorate the impact of adversities on the health and wellbeing of children.

Abbreviations:

CA

child adversity

PDS

Pubertal Development Score