Abstract

As economic development increases, puberty occurs at younger ages, and this could contribute to an increase in the incidence of cardiovascular diseases and hormone-related cancers. The factors that determine pubertal timing are poorly understood. The growth axis that is active during puberty is active in the first 6 months of life and interacts with the immune system. The authors examined whether prior infections, proxied by number of hospital admissions for infections at different ages, were associated with age at pubertal onset (Tanner stage II) using interval-censored regression in the Children of 1997 cohort, which is a population-representative Chinese birth cohort (n = 7,527). Mediation by growth was also examined. Girls, but not boys, who were hospitalized for infections at least twice in the first 6 months of life experienced pubertal onset about 8 months later (mean = 10.3 years, time ratio = 1.08, 95% confidence interval: 1.04, 1.12) than did those without such hospitalizations (mean = 9.6 years) after adjustment for infant characteristics and socioeconomic position (sex interaction: P = 0.02). There were no such associations for infections at 6 months to ≤8 years of age. Growth did not mediate the association. Early infectious morbidity in girls may be associated with later puberty, perhaps via suppression of the gonadotropic axis. The lowering of the number of infections in early life that accompanies economic development could be an additional factor that contributes to earlier puberty.

Early puberty is associated with hormone-related cancers (1), cardiovascular disease (2) and its associated risk factors (3), and poor psychosocial adjustment (4). Average age at puberty has decreased sharply with economic development and is still falling in long-term developed Western countries (5, 6), developing populations (7, 8), and recently developed populations (9). Reasons for the continuing fall in age of puberty are not well understood, and they could encompass many environmental exposures, including nutrition (10, 11) and endocrine disruptors (12, 13). Another possible factor that has rarely been considered recently is exposure to or experience of infections (14). With smaller families, more vaccinations, and improved living standards, the rate of exposure to infections in infancy and childhood is declining (15). Serious infectious morbidity and mortality rates in infancy and childhood have been greatly reduced with economic development (16, 17), as has the age at pubertal onset (8, 9).

Pubertal onset is later in children with prenatally acquired human immunodeficiency virus infection (18, 19) or inflammatory bowel diseases (20). In the general population, findings on prior infections and age of puberty are limited and seemingly contradictory; more infections in infants or young children (21) but not in school-aged children are associated with later puberty (22). Experience of infections is well known to down-regulate the gonadotropic axis, either by reducing sex-steroid production or possibly through regulation of sex-steroid receptors in both humans (23–25) and animals (26–28). Infections would be expected to be associated with later puberty, perhaps with the greatest effect at times when the gonadotropic axis is active, that is, during the mini-puberty of early infancy (birth to about 6 months of age) (29, 30) or puberty (31), consistent with the observations to date (21, 22). However, it is also possible that these observations are due to infections that preclude optimal infant growth (32, 33) through growth faltering (34). To clarify this question, we took advantage of data on a large population-representative birth cohort from a developed setting, Hong Kong's Children of 1997, in which infections would be expected to have a limited influence on infant growth (35). Specifically, we tested the hypothesis that infectious morbidity in infancy, but not childhood, would be associated with later puberty without affecting growth.

MATERIALS AND METHODS

Data sources

The Children of 1997 birth cohort consists of 8,327 infants born in Hong Kong in April and May of 1997, comprising 88% of all births during that period. They were recruited from all 49 governmental Maternal and Child Health Centres (MCHCs), which provide free-of-charge immunizations and health and developmental surveillance (36). A self-administered survey designed to study the effects of secondhand smoke on infant health was completed by parents at the birth of the children and when those children were 3, 9, and 18 months of age (37). Passive follow-up via record linkage was instituted in 2005 to extract data on: 1) weight and height of the children from birth to 5 years of age from the MCHC records; 2) weight, height, and pubertal status from 6 to 12 years of age from the Student Health Service, Department of Health, which provides free annual check-ups for all school students (38); 3) hospital discharge records from the Hospital Authority, which manages all public hospitals (accounting for 81.4% of beds in 2005) (39); and 4) death records from the death registry. Active follow-up via direct contact was instituted in 2007. Over 89% of the original cohort had been recontacted. As of June 30, 2010, a total of 7,933 of the original 8,327 cohort members were alive, had not withdrawn from the study, and were living in Hong Kong.

Exposures

The main exposure was number of hospital admissions (including same-day discharge and inpatient admission for at least 24 hours) for infections in public hospitals. The control exposure to examine any residual confounding by family socioeconomic position was the number of hospital admissions for accidents. Accidents, like infections, are socially patterned. With adequate control for socioeconomic position, there should be no relation between hospital episodes for accidents and age at pubertal onset. A lack of association between accidents and pubertal onset indicates less risk of residual confounding. As previously, children without any public hospital discharge records were assumed to have had no hospital admissions (40–42). The numbers of hospital admissions were considered at different ages, corresponding to different developmental stages during the first 2 main phases of growth (43), that is, mini-puberty (9 days–<6 months of age) and then 6–<24 months, 2–<5 years, and 5–8 years of age. Hospital admissions in the immediate neonatal period (≤8 days of age) were excluded. Admissions are recorded differently between the vaginally born and the surgically born with the same conditions because the vaginally born have a shorter average hospital stay for delivery than the surgically born (3.0 days vs. 7.9 days) (44). Because hospitalization data are usually not normally distributed, with a long right tail and a mode at zero (45), at each stage, the number of hospital admissions for infections was categorized as 0, 1, or ≥2 and the number for accidents was categorized as 0 or ≥1.

Consistent with our previous studies (40, 41), infections, as determined from the principal discharge code based on the International Classification of Diseases, Ninth Revision Clinical Modification (ICD-9-CM), consisted of respiratory infections (ICD–9-CM codes 33, 34.0, 381–2, 460–6, 477, 480–7, 477, and 493), gastrointestinal infections (ICD–9-CM codes 1–9, 535.0, 535.4, 535.5, 538, 558.9, and 787.91), and other infectious and parasitic diseases (ICD–9-CM codes 10–32, 34.1–139, 320–1, 370, 372.0–372.3, 390–2, 540–2, 590, 595, 599.0, 680–6, 771, 780.3, 780.6, and 787.91). Hospital admissions with ICD–9-CM codes 800–999 or E800–E999 were classified as accidents.

Mediators

As in our previous study, infant growth was considered as change in the weight z score (standard deviation) from birth to 3 months of age and from 3 to 12 months of age (46). Childhood growth was considered as height and body mass index (BMI) z score at about 7 years (based on the closest measurement available for 6–<9 years of age) (47), corresponding to prepubertal growth, relative to the 2007 World Health Organization growth references (48).

Outcome

The outcome was age at pubertal onset, as the earliest age when Tanner stage II for breasts (girls) or genitalia (boys) was recorded. Pubertal status was visually assessed (from age 6 years) by physicians according to the criteria of Marshall and Tanner (49, 50). Children with precocious puberty (i.e., signs of puberty before age 7.5 years in girls (n = 35) or 9 years in boys (n = 3)) were included. Children with infeasible sequences of pubertal stages, such as pubertal stage II before pubertal stage I, were excluded.

Statistical analysis

We used multivariable interval-censored regression (51) with a lognormal distribution to examine the adjusted associations between the exposures and age at pubertal onset, from which time ratios with 95% confidence intervals are presented. A time ratio gives the comparison of ages at pubertal onset between groups, so a time ratio greater than 1 indicates older age at pubertal onset, whereas a time ratio less than 1 indicates younger age at pubertal onset. The distribution of age at pubertal onset was estimated by censoring each child in one of 3 ways: 1) left censoring if the child had experienced pubertal onset (Tanner stage II) by the first examination date; 2) interval censoring if the child had pubertal onset between 2 examination dates (i.e., the first examination date recorded Tanner stage I and the second examination date recorded Tanner stage II); and 3) right censoring if the child had not experienced pubertal onset (Tanner stage II) by the most recent examination date.

We also checked whether the associations varied by sex, from the significance of interaction terms and the heterogeneity of effects across strata. We included the interaction terms of sex with number of hospital admissions and sex with all confounders in the same model to ensure the independence of the interaction term of interest (52).

To assess whether infant or childhood growth mediated any association, we used the 4 criteria of Baron and Kenny (53), which in this case means that 1) number of hospital admissions for infections should be associated with age at pubertal onset; 2) hospital admissions for infections should be associated with infant or childhood growth, the potential mediators; 3) infant or childhood growth should be associated with age at pubertal onset; and 4) the association between hospital admissions and age at pubertal onset should be attenuated after adjustment for infant or childhood growth.

Confounders included birth weight, gestational age, birth order, breastfeeding, secondhand smoke exposure, maternal place of birth, maternal educational level, type of hospital at birth, income of household head, and Rutter score at 7 years of age, as categorized in Table 1. We used multiple imputation to predict missing confounders based on a flexible additive regression model with predictive mean matching (54), incorporating data on sex, birth weight, gestational age, birth order, method of delivery, breastfeeding, secondhand smoke exposure, parental ages, maternal place of birth, maternal educational level, maternal occupation, household income, type of hospital at birth, Rutter score at 7 years, maximum age at Tanner stage I, and minimum age at Tanner stage II (55). We summarized the results from 10 imputed data sets into single estimated time ratios with confidence intervals adjusted for missing data uncertainty (56). As a sensitivity analysis, we also performed a complete case analysis. Statistical analyses were performed using Stata, version 10 (Stata Corporation, College Station, Texas) and R, version 2.10.1 (R Development Core Team, Vienna, Austria).

Table 1.

Baseline Characteristics, by Number of Hospital Admissions, for Infections at 9 Days to 6 Months of Age in the Hong Kong Children of 1997 Birth Cohort, Hong Kong, China, 1997–2004

Characteristic Girls (n = 3,542)
 
Boys (n = 3,985)
 
No. 0 (n = 3,250)
 
1 (n = 239)
 
≥2 (n = 53)
 
No. 0 (n = 3,544)
 
1 (n = 336)
 
≥2 (n = 105)
 
Mean (SD) Mean (SD) Mean (SD) Mean (SD) Mean (SD) Mean (SD) 
Birth weight z score 3,542 −0.2 (1.0)  −0.3 (1.1)  −0.3 (1.1)  3,985 −0.3 (1.0)  −0.4 (1.1)  −0.3 (1.0)  
Birth weight, kg 3,542 3,147 (419)  3,103 (484)  3,102 (462)  3,985 3,237 (448)  3,202 (501)  3,221 (452)  
Mean gestational age, weeks 3,542 39.0 (1.6)  38.8 (1.9)  38.8 (1.9)  3,985 38.9 (1.6)  38.8 (2.0)  38.7 (1.7)  
Birth order               
    First 1,720  91.8  6.5  1.7 1,863  89.6  8.2  2.2 
    Second 1,436  91.5  7.4  1.1 1,695  88.8  8.6  2.6 
    Third or higher 386  92.5  5.4  2.1 428  86.9  8.7  4.4 
Breastfeeding               
    Never breastfed 2,011  91.1  7.3  1.5 2,279  87.6  9.4  3.1 
    Partially breastfed for any length of time or exclusively breastfed for <3 months 1,272  92.4  6.4  1.3 1,482  90.8  7.0  2.3 
    Exclusively breastfed for ≥3 months 258  93.7  3.9  2.3 224  90.6  8.8  0.6 
Exposure to secondhand smoke 
    None 993  93.9  5.0  1.0 1,072  90.1  7.2  2.7 
    Nonparental household smoking 1,346  91.6  7.2  1.3 1,496  90.4  7.7  1.9 
    Paternal smoking 1,017  91.1  7.2  1.8 1,233  86.4  10.2  3.5 
    Maternal smoking 187  85.1  10.6  4.3 185  87.6  10.0  2.4 
Maternal place of birth               
    Mainland China or elsewhere 1,305  90.6  7.5  2.0 1,484  85.3  10.6  4.0 
    Hong Kong 2,237  92.5  6.3  1.2 2,501  91.1  7.1  1.8 
Maternal educational level               
    Grade 9 or below 1,438  89.5  8.4  2.1 1,663  85.9  10.3  3.8 
    Grades 10–11 1,568  93.1  5.6  1.3 1,725  90.7  7.2  2.1 
    Grade 12 or above 536  93.8  5.6  0.6 597  92.1  6.9  1.0 
Type of hospital at birth               
    Private 2,565  90.2  7.8  2.0 2,858  86.4  10.2  3.5 
    Public 977  95.8  3.9  0.3 1,127  95.5  4.0  0.6 
Annual income of household head, dollarsa               
    Quintile 1: mean = 1,741 (SD, 408) 709  90.3  7.2  2.5 808  84.7  10.2  5.0 
    Quintile 2: mean = 2,846 (SD, 328) 745  87.4  10.1  2.5 833  86.5  10.2  3.4 
    Quintile 3: mean = 4,362 (SD, 556) 688  92.6  6.0  1.4 815  86.9  10.4  2.7 
    Quintile 4: mean = 6,805 (SD, 874) 697  94.2  5.0  0.7 760  92.6  6.3  1.1 
    Quintile 5: mean = 14,678 (SD, 15,497) 703  94.5  5.1  0.3 770  94.5  4.7  0.8 
Weight z score change at 0–3 months of age 3,542 0.3 (0.9)  0.3 (0.9)  0.0 (1.1)  3,985 0.3 (0.9)  0.3 (0.9)  0.3 (1.1)  
Weight z score change at 3–12 months of age 3,542 0.0 (0.6)  0.1 (0.7)  0.2 (0.8)  3,985 −0.1 (0.7)  0.0 (0.7)  0.0 (0.7)  
Body mass indexbz score at 7 years of age 3,542 0.0 (1.0)  0.1 (1.1)  −0.2 (1.0)  3,985 0.3 (1.3)  0.4 (1.4)  0.3 (1.3)  
Height z score at 7 years of age 3,542 −0.1 (0.9)  −0.1 (1.0)  −0.4 (0.8)  3,985 −0.1 (0.9)  −0.1 (1.0)  0.1 (0.9)  
Rutter score at 7 years of age               
    <13 points 2,793  92.0  6.6  1.4 2,739  89.1  8.2  2.7 
    >13 points 749  91.0  7.3  1.7 1,246  88.5  8.9  2.6 
Characteristic Girls (n = 3,542)
 
Boys (n = 3,985)
 
No. 0 (n = 3,250)
 
1 (n = 239)
 
≥2 (n = 53)
 
No. 0 (n = 3,544)
 
1 (n = 336)
 
≥2 (n = 105)
 
Mean (SD) Mean (SD) Mean (SD) Mean (SD) Mean (SD) Mean (SD) 
Birth weight z score 3,542 −0.2 (1.0)  −0.3 (1.1)  −0.3 (1.1)  3,985 −0.3 (1.0)  −0.4 (1.1)  −0.3 (1.0)  
Birth weight, kg 3,542 3,147 (419)  3,103 (484)  3,102 (462)  3,985 3,237 (448)  3,202 (501)  3,221 (452)  
Mean gestational age, weeks 3,542 39.0 (1.6)  38.8 (1.9)  38.8 (1.9)  3,985 38.9 (1.6)  38.8 (2.0)  38.7 (1.7)  
Birth order               
    First 1,720  91.8  6.5  1.7 1,863  89.6  8.2  2.2 
    Second 1,436  91.5  7.4  1.1 1,695  88.8  8.6  2.6 
    Third or higher 386  92.5  5.4  2.1 428  86.9  8.7  4.4 
Breastfeeding               
    Never breastfed 2,011  91.1  7.3  1.5 2,279  87.6  9.4  3.1 
    Partially breastfed for any length of time or exclusively breastfed for <3 months 1,272  92.4  6.4  1.3 1,482  90.8  7.0  2.3 
    Exclusively breastfed for ≥3 months 258  93.7  3.9  2.3 224  90.6  8.8  0.6 
Exposure to secondhand smoke 
    None 993  93.9  5.0  1.0 1,072  90.1  7.2  2.7 
    Nonparental household smoking 1,346  91.6  7.2  1.3 1,496  90.4  7.7  1.9 
    Paternal smoking 1,017  91.1  7.2  1.8 1,233  86.4  10.2  3.5 
    Maternal smoking 187  85.1  10.6  4.3 185  87.6  10.0  2.4 
Maternal place of birth               
    Mainland China or elsewhere 1,305  90.6  7.5  2.0 1,484  85.3  10.6  4.0 
    Hong Kong 2,237  92.5  6.3  1.2 2,501  91.1  7.1  1.8 
Maternal educational level               
    Grade 9 or below 1,438  89.5  8.4  2.1 1,663  85.9  10.3  3.8 
    Grades 10–11 1,568  93.1  5.6  1.3 1,725  90.7  7.2  2.1 
    Grade 12 or above 536  93.8  5.6  0.6 597  92.1  6.9  1.0 
Type of hospital at birth               
    Private 2,565  90.2  7.8  2.0 2,858  86.4  10.2  3.5 
    Public 977  95.8  3.9  0.3 1,127  95.5  4.0  0.6 
Annual income of household head, dollarsa               
    Quintile 1: mean = 1,741 (SD, 408) 709  90.3  7.2  2.5 808  84.7  10.2  5.0 
    Quintile 2: mean = 2,846 (SD, 328) 745  87.4  10.1  2.5 833  86.5  10.2  3.4 
    Quintile 3: mean = 4,362 (SD, 556) 688  92.6  6.0  1.4 815  86.9  10.4  2.7 
    Quintile 4: mean = 6,805 (SD, 874) 697  94.2  5.0  0.7 760  92.6  6.3  1.1 
    Quintile 5: mean = 14,678 (SD, 15,497) 703  94.5  5.1  0.3 770  94.5  4.7  0.8 
Weight z score change at 0–3 months of age 3,542 0.3 (0.9)  0.3 (0.9)  0.0 (1.1)  3,985 0.3 (0.9)  0.3 (0.9)  0.3 (1.1)  
Weight z score change at 3–12 months of age 3,542 0.0 (0.6)  0.1 (0.7)  0.2 (0.8)  3,985 −0.1 (0.7)  0.0 (0.7)  0.0 (0.7)  
Body mass indexbz score at 7 years of age 3,542 0.0 (1.0)  0.1 (1.1)  −0.2 (1.0)  3,985 0.3 (1.3)  0.4 (1.4)  0.3 (1.3)  
Height z score at 7 years of age 3,542 −0.1 (0.9)  −0.1 (1.0)  −0.4 (0.8)  3,985 −0.1 (0.9)  −0.1 (1.0)  0.1 (0.9)  
Rutter score at 7 years of age               
    <13 points 2,793  92.0  6.6  1.4 2,739  89.1  8.2  2.7 
    >13 points 749  91.0  7.3  1.7 1,246  88.5  8.9  2.6 

Abbreviation: SD, standard deviation.

a

US $1 = HK$7.80.

b

Weight (kg)/height (m)2.

Ethics approval

We obtained ethical approval from the University of Hong Kong-Hospital Authority Hong Kong West Cluster Joint Institutional Review Board and the Ethics Committee of the Department of Health, Government of the Hong Kong Special Administrative Region.

RESULTS

Of the 7,527 children (90% follow-up rate) for whom we had information about age at pubertal onset, 76% of the girls (2,695) and 42% of the boys (1,671) had experienced pubertal onset. Income was imputed for 10.3% of the children, maternal place of birth for 8.4%, secondhand smoke or breastfeeding for 2.4%, and maternal educational level or type of hospital at birth for 1.1%. The estimated mean age at pubertal onset was 11.7 years in boys (95% confidence interval: 11.6, 11.8) and 9.6 years in girls (95% confidence interval: 9.5, 9.6)—numbers that were very similar to what was previously observed in Hong Kong (9). In the first 6 months of life, 8% had 1 hospitalization for infections and 2% had ≥2; more boys (2.6%) than girls (1.5%) had at least 2 hospitalizations for infections. In the same period, 0.5% had hospitalizations for accidents.

Table 1 shows that children with ≥2 hospitalizations for infections in the first 6 months of life were more likely to never have been breastfed and to have been exposed to paternal or maternal smoking. They also had lower family socioeconomic positions and mothers from mainland China or elsewhere. Children hospitalized for infections in later infancy and childhood (6 months to 8 years of age) were also likely to be from families of lower socioeconomic position (Web Tables 1 and 2, available at http://aje.oxfordjournals.org/). Girls, but not boys, hospitalized for infections in early infancy tended to grow slowly at 0–3 months of age and show positive weight change at 3–12 months of age, but they were not thinner or shorter at 7 years of age. Higher BMI and height at 7 years of age were associated with younger age at pubertal onset in both sexes (Table 2). Weight gain at 0–3 months and at 3–12 months was associated with younger age at pubertal onset in boys but not in girls.

Table 2.

Unadjusted Associations Between Baseline Characteristics and Age at Pubertal Onset in the Hong Kong Children of 1997 Birth Cohort, Hong Kong, China, 1997–2004

Characteristic Girls (n = 3,542)
 
Boys (n = 3,985)
 
No. Time Ratio 95% CI No. Time Ratio 95% CI 
Birth weight z score 3,542 1.00 0.99, 1.00 3,985 1.00 0.99, 1.00 
Gestational age 3,542 0.99 0.99, 1.00 3,985 1.00 1.00, 1.00 
Birth order       
    First 1,720 Reference  1,863 Reference  
    Second 1,436 1.00 0.99, 1.01 1,695 1.00 0.99, 1.01 
    Third or higher 386 1.00 0.98, 1.02 428 1.00 0.98, 1.01 
Breastfeeding       
    Never breastfed 2,011 Reference  2,279 Reference  
    Partially breastfed for any length of time or exclusively breastfed for <3 months 1,272 1.01 1.00, 1.02 1,482 1.00 0.99, 1.01 
    Exclusively breastfed for ≥3 months 258 0.99 0.97, 1.01 224 1.00 0.98, 1.02 
Exposure to secondhand smoke       
    None 993 Reference  1,072 Reference  
    Nonparental household smoking 1,346 1.01 1.00, 1.02 1,496 1.00 0.99, 1.01 
    Paternal smoking 1,017 1.00 0.99, 1.02 1,233 1.00 0.99, 1.01 
    Maternal smoking 187 1.01 0.98, 1.03 185 1.02 1.00, 1.04 
Maternal place of birth       
    Mainland China or elsewhere 1,305 Reference  1,484 Reference  
    Hong Kong 2,237 1.02 1.01, 1.03 2,501 1.01 1.00, 1.02 
Maternal educational level       
    Grade 9 or below 1,438 Reference  1,663 Reference  
    Grades 10–11 1,568 1.02 1.00, 1.03 1,725 1.00 0.99, 1.01 
    Grade 12 or above 536 1.02 1.00, 1.03 597 0.99 0.98, 1.00 
Type of hospital at birth       
    Private 2,565 Reference  2,858 Reference  
    Public 977 1.01 1.00, 1.02 1,127 1.00 0.99, 1.01 
Annual income of household head, dollars       
    Quintile 1: mean = 1,741 (SD, 408) 709 Reference  808 Reference  
    Quintile 2: mean = 2,846 (SD, 328) 745 1.00 0.98, 1.01 833 0.99 0.98, 1.01 
    Quintile 3: mean = 4,362 (SD, 556) 688 1.00 0.99, 1.02 815 0.99 0.98, 1.00 
    Quintile 4: mean = 6,805 (SD, 874) 697 1.01 0.99, 1.03 760 1.00 0.98, 1.01 
    Quintile 5: mean = 14,678 (SD, 15,497) 703 1.03 1.01, 1.04 770 0.99 0.98, 1.01 
Weight z score change at 0–3 months 3,542 1.00 0.99, 1.00 3,985 1.00 0.99, 1.00 
Weight z score change at 3–12 months 3,542 0.99 0.98, 1.00 3,985 0.99 0.99, 1.00 
Body mass indexaz score at 7 years of age 3,542 0.97 0.97, 0.98 3,985 0.99 0.99, 1.00 
Height z score at 7 years of age 3,542 0.96 0.95, 0.96 3,985 0.98 0.97, 0.98 
Rutter score at 7 years of age       
    <13 points 2,793 Reference  2,739 Reference  
    ≥13 points 749 1.01 1.00, 1.02 1,246 1.00 0.99, 1.01 
Characteristic Girls (n = 3,542)
 
Boys (n = 3,985)
 
No. Time Ratio 95% CI No. Time Ratio 95% CI 
Birth weight z score 3,542 1.00 0.99, 1.00 3,985 1.00 0.99, 1.00 
Gestational age 3,542 0.99 0.99, 1.00 3,985 1.00 1.00, 1.00 
Birth order       
    First 1,720 Reference  1,863 Reference  
    Second 1,436 1.00 0.99, 1.01 1,695 1.00 0.99, 1.01 
    Third or higher 386 1.00 0.98, 1.02 428 1.00 0.98, 1.01 
Breastfeeding       
    Never breastfed 2,011 Reference  2,279 Reference  
    Partially breastfed for any length of time or exclusively breastfed for <3 months 1,272 1.01 1.00, 1.02 1,482 1.00 0.99, 1.01 
    Exclusively breastfed for ≥3 months 258 0.99 0.97, 1.01 224 1.00 0.98, 1.02 
Exposure to secondhand smoke       
    None 993 Reference  1,072 Reference  
    Nonparental household smoking 1,346 1.01 1.00, 1.02 1,496 1.00 0.99, 1.01 
    Paternal smoking 1,017 1.00 0.99, 1.02 1,233 1.00 0.99, 1.01 
    Maternal smoking 187 1.01 0.98, 1.03 185 1.02 1.00, 1.04 
Maternal place of birth       
    Mainland China or elsewhere 1,305 Reference  1,484 Reference  
    Hong Kong 2,237 1.02 1.01, 1.03 2,501 1.01 1.00, 1.02 
Maternal educational level       
    Grade 9 or below 1,438 Reference  1,663 Reference  
    Grades 10–11 1,568 1.02 1.00, 1.03 1,725 1.00 0.99, 1.01 
    Grade 12 or above 536 1.02 1.00, 1.03 597 0.99 0.98, 1.00 
Type of hospital at birth       
    Private 2,565 Reference  2,858 Reference  
    Public 977 1.01 1.00, 1.02 1,127 1.00 0.99, 1.01 
Annual income of household head, dollars       
    Quintile 1: mean = 1,741 (SD, 408) 709 Reference  808 Reference  
    Quintile 2: mean = 2,846 (SD, 328) 745 1.00 0.98, 1.01 833 0.99 0.98, 1.01 
    Quintile 3: mean = 4,362 (SD, 556) 688 1.00 0.99, 1.02 815 0.99 0.98, 1.00 
    Quintile 4: mean = 6,805 (SD, 874) 697 1.01 0.99, 1.03 760 1.00 0.98, 1.01 
    Quintile 5: mean = 14,678 (SD, 15,497) 703 1.03 1.01, 1.04 770 0.99 0.98, 1.01 
Weight z score change at 0–3 months 3,542 1.00 0.99, 1.00 3,985 1.00 0.99, 1.00 
Weight z score change at 3–12 months 3,542 0.99 0.98, 1.00 3,985 0.99 0.99, 1.00 
Body mass indexaz score at 7 years of age 3,542 0.97 0.97, 0.98 3,985 0.99 0.99, 1.00 
Height z score at 7 years of age 3,542 0.96 0.95, 0.96 3,985 0.98 0.97, 0.98 
Rutter score at 7 years of age       
    <13 points 2,793 Reference  2,739 Reference  
    ≥13 points 749 1.01 1.00, 1.02 1,246 1.00 0.99, 1.01 

Abbreviations: CI, confidence interval; SD, standard deviation.

a

Weight (kg)/height (m)2.

The associations between hospital admissions for infections and age at pubertal onset differed by sex in the first 6 months of life (P for interaction = 0.02) but not later (P value for interaction: at 6–<24 months of age P = 0.81, at 2–<5 years, P = 0.08, and at 5–8 years, P = 0.76). Table 3 shows that girls, but not boys, who had at least 2 hospitalizations for infections in the first 6 months of life had older age at pubertal onset than did those with no hospital admissions after adjustment for birth weight, gestational age, birth order, breastfeeding, secondhand smoke, maternal place of birth, maternal educational level, type of hospital at birth, household income per head, and Rutter score at 7 years. The number of hospital admissions for infections in later infancy or early childhood (6 months to 8 years of age) was not associated with age at pubertal onset in boys or girls. There was no association between number of hospital admissions for accidents and age at pubertal onset. The complete-case analysis produced similar results (Appendix Table 1).

Table 3.

Adjusteda Associations Between Number of Hospital Admissions for Infections and Accidents at Different Stages and Age at Pubertal Onset in the Hong Kong Children of 1997 Birth Cohort, Hong Kong, China, 1997–2004

No. of Hospital Admissions by Age Group and Cause Girls (n = 3,542)
 
Boys (n = 3,985)
 
No. Time Ratio 95% CI No. Time Ratio 95% CI 
9 days–<6 months       
    Infections       
        0 3,250 Reference  3,544 Reference  
        1 239 1.00 0.99, 1.02 336 1.01 0.99, 1.02 
        ≥2 53 1.08 1.04, 1.12 105 1.01 0.98, 1.03 
    Accidents       
        0 3,522 Reference  3,966 Reference  
        ≥1 20 0.97 0.91, 1.03 19 1.02 0.96, 1.08 
6–<24 months       
    Infections       
        0 2,959 Reference  3,096 Reference  
        1 440 1.01 0.99, 1.02 619 1.00 0.99, 1.01 
        ≥2 143 1.00 0.97, 1.02 270 1.00 0.98, 1.02 
    Accidents       
        0 3,468 Reference  3,893 Reference  
        ≥1 74 1.00 0.97, 1.03 92 0.97 0.94, 1.00 
2–<5 years       
    Infections       
        0 3,002 Reference  3,255 Reference  
        1 430 1.01 1.00, 1.03 526 0.99 0.98, 1.01 
        ≥2 110 0.99 0.96, 1.01 204 1.01 0.99, 1.03 
    Accidents       
        0 3,443 Reference  3,864 Reference  
        ≥1 99 1.03 1.00, 1.06 121 0.99 0.96, 1.01 
5–8 years       
    Infections       
        0 3,369 Reference  3,759 Reference  
        1 155 1.00 0.98, 1.03 195 0.99 0.97, 1.01 
        ≥2 18 1.01 0.95, 1.07 31 1.00 0.95, 1.06 
    Accidents       
        0 3,480 Reference  3,888 Reference  
        ≥1 62 1.01 0.97, 1.04 97 1.00 0.97, 1.03 
No. of Hospital Admissions by Age Group and Cause Girls (n = 3,542)
 
Boys (n = 3,985)
 
No. Time Ratio 95% CI No. Time Ratio 95% CI 
9 days–<6 months       
    Infections       
        0 3,250 Reference  3,544 Reference  
        1 239 1.00 0.99, 1.02 336 1.01 0.99, 1.02 
        ≥2 53 1.08 1.04, 1.12 105 1.01 0.98, 1.03 
    Accidents       
        0 3,522 Reference  3,966 Reference  
        ≥1 20 0.97 0.91, 1.03 19 1.02 0.96, 1.08 
6–<24 months       
    Infections       
        0 2,959 Reference  3,096 Reference  
        1 440 1.01 0.99, 1.02 619 1.00 0.99, 1.01 
        ≥2 143 1.00 0.97, 1.02 270 1.00 0.98, 1.02 
    Accidents       
        0 3,468 Reference  3,893 Reference  
        ≥1 74 1.00 0.97, 1.03 92 0.97 0.94, 1.00 
2–<5 years       
    Infections       
        0 3,002 Reference  3,255 Reference  
        1 430 1.01 1.00, 1.03 526 0.99 0.98, 1.01 
        ≥2 110 0.99 0.96, 1.01 204 1.01 0.99, 1.03 
    Accidents       
        0 3,443 Reference  3,864 Reference  
        ≥1 99 1.03 1.00, 1.06 121 0.99 0.96, 1.01 
5–8 years       
    Infections       
        0 3,369 Reference  3,759 Reference  
        1 155 1.00 0.98, 1.03 195 0.99 0.97, 1.01 
        ≥2 18 1.01 0.95, 1.07 31 1.00 0.95, 1.06 
    Accidents       
        0 3,480 Reference  3,888 Reference  
        ≥1 62 1.01 0.97, 1.04 97 1.00 0.97, 1.03 

Abbreviation: CI, confidence interval.

a

Adjusted for birth weight, gestational age, birth order, breastfeeding, secondhand smoke, maternal place of birth, maternal educational level, type of hospital at birth, income of household head, and Rutter score at 7 years of age.

Both infant and childhood growth only partially fulfilled the mediation criteria, because infant growth was not associated with age at pubertal onset in girls or with number of hospital admissions for infections in boys. Also, the number of hospital admissions for infections in the first 6 months of life was not associated with childhood growth in girls or boys. Moreover, the associations between hospital admissions for infections in the first 6 months of life and age at pubertal onset remained unchanged after additionally adjustment for infant and childhood growth separately or simultaneously (Table 4).

Table 4.

Associations Between Number of Hospital Admissions for Infections at 9 Days to 6 Months and Age at Pubertal Onset With Adjustment for Different Combinations of Growth Markersa, Hong Kong, China, 1997–2004

Model and No. of Hospital Admissions Girls (n = 3,542)
 
Boys (n = 3,985)
 
Time Ratio 95% CI Time Ratio 95% CI 
Model 1: unadjusted     
    0 Reference  Reference  
    1 1.00 0.98, 1.02 1.00 0.99, 1.02 
    ≥2 1.07 1.03, 1.12 1.00 0.98, 1.03 
Model 2: adjusted for weight z score change at 0–3 months of age     
    0 Reference  Reference  
    1 1.00 0.99, 1.02 1.01 0.99, 1.02 
    ≥2 1.07 1.03, 1.12 1.01 0.98, 1.03 
Model 3: adjusted for weight z score change at 3–12 months of age     
    0 Reference  Reference  
    1 1.01 0.99, 1.02 1.01 0.99, 1.02 
    ≥2 1.08 1.04, 1.13 1.01 0.98, 1.04 
Model 4: adjusted for weight z score change at 0–3 months and 3–12 months of age     
    0 Reference  Reference  
    1 1.00 0.99, 1.02 1.01 0.99, 1.02 
    ≥2 1.08 1.03, 1.12 1.01 0.98, 1.04 
Model 5: adjusted for BMIbz score at 7 years of age     
    0 Reference  Reference  
    1 1.01 0.99, 1.03 1.01 0.99, 1.02 
    ≥2 1.07 1.03, 1.11 1.01 0.98, 1.03 
Model 6: adjusted for height z score at 7 years of age     
    0 Reference  Reference  
    1 1.01 0.99, 1.02 1.01 0.99, 1.02 
    ≥2 1.06 1.02, 1.10 1.01 0.98, 1.04 
Model 7: adjusted for height and BMI z scores at 7 years of age     
    0 Reference  Reference  
    1 1.01 0.99, 1.02 1.01 0.99, 1.02 
    ≥2 1.06 1.02, 1.10 1.01 0.98, 1.04 
Model 8: adjusted for weight z score change at 0–3 months and 3–12 months of age and height and BMI z scores at 7 years of age     
    0 Reference  Reference  
    1 1.01 0.99, 1.03 1.01 0.99, 1.02 
    ≥2 1.06 1.03, 1.11 1.01 0.98, 1.04 
Model and No. of Hospital Admissions Girls (n = 3,542)
 
Boys (n = 3,985)
 
Time Ratio 95% CI Time Ratio 95% CI 
Model 1: unadjusted     
    0 Reference  Reference  
    1 1.00 0.98, 1.02 1.00 0.99, 1.02 
    ≥2 1.07 1.03, 1.12 1.00 0.98, 1.03 
Model 2: adjusted for weight z score change at 0–3 months of age     
    0 Reference  Reference  
    1 1.00 0.99, 1.02 1.01 0.99, 1.02 
    ≥2 1.07 1.03, 1.12 1.01 0.98, 1.03 
Model 3: adjusted for weight z score change at 3–12 months of age     
    0 Reference  Reference  
    1 1.01 0.99, 1.02 1.01 0.99, 1.02 
    ≥2 1.08 1.04, 1.13 1.01 0.98, 1.04 
Model 4: adjusted for weight z score change at 0–3 months and 3–12 months of age     
    0 Reference  Reference  
    1 1.00 0.99, 1.02 1.01 0.99, 1.02 
    ≥2 1.08 1.03, 1.12 1.01 0.98, 1.04 
Model 5: adjusted for BMIbz score at 7 years of age     
    0 Reference  Reference  
    1 1.01 0.99, 1.03 1.01 0.99, 1.02 
    ≥2 1.07 1.03, 1.11 1.01 0.98, 1.03 
Model 6: adjusted for height z score at 7 years of age     
    0 Reference  Reference  
    1 1.01 0.99, 1.02 1.01 0.99, 1.02 
    ≥2 1.06 1.02, 1.10 1.01 0.98, 1.04 
Model 7: adjusted for height and BMI z scores at 7 years of age     
    0 Reference  Reference  
    1 1.01 0.99, 1.02 1.01 0.99, 1.02 
    ≥2 1.06 1.02, 1.10 1.01 0.98, 1.04 
Model 8: adjusted for weight z score change at 0–3 months and 3–12 months of age and height and BMI z scores at 7 years of age     
    0 Reference  Reference  
    1 1.01 0.99, 1.03 1.01 0.99, 1.02 
    ≥2 1.06 1.03, 1.11 1.01 0.98, 1.04 

Abbreviations: BMI, body mass index; CI, confidence interval.

a

Models 2–8 were adjusted for birth weight, gestational age, birth order, breastfeeding, secondhand smoke, maternal place of birth, maternal educational level, type of hospital at birth, income of household head, and Rutter score at 7 years of age.

b

Weight (kg)/height (m)2.

DISCUSSION

In the present cohort, a contemporary, population-representative Hong Kong Chinese birth cohort with a high follow-up rate, girls hospitalized for infections in early infancy (<6 months of age) but not in later infancy or early childhood (6 months to 8 years of age) had older age at pubertal onset, consistent with a previously reported positive association of infections in infancy (3 months to 3 years of age) with late age of menarche (21) but not of childhood (birth to 8 years of age) infections (22). In boys, the number of hospital admissions for infections in the first 8 years of life was unrelated to age at pubertal onset. To our knowledge, this is the first study that clearly distinguishes the effect of infections in infancy from that of infections in childhood. In addition, our cohort was large enough to allow us to distinguish between the sexes. We found that the association was specific to girls, which is consistent with relatively later pubertal onset in girls than in boys with prenatally acquired human immunodeficiency virus infections (19). Additionally, we found no evidence of mediation by infant or childhood growth. There was also no association between hospital admissions for accidents and age at pubertal onset, which suggested minimal confounding by family socioeconomic position. Our study adds to the literature by showing that, consistent with our hypothesis, infections in early infancy but not later infancy or childhood were associated with later puberty in girls, independent of growth.

This large birth cohort, with its strength of case ascertainment based on public hospital discharge records, clinically assessed pubertal onset, and routinely measured anthropometric data, nevertheless has several limitations. First, we only assessed the effects of serious infectious morbidity that warranted hospitalizations, because these were more likely to have a long-term impact. Second, we cannot rule out the possibility of selection bias, given that undergoing check-ups provided by the Student Health Service is voluntary. However, there was a high follow-up rate (90% of the eligible cohort members), and children with follow-up were similar to those without it. Cohen effect sizes for sex (0.05), number of hospital admissions at different stages (<0.15), and family socioeconomic position (<0.21) were relatively small. Our results could have been biased if we had systematically excluded those children with serious infections in early infancy and pubertal onset at younger ages, though we could not see how that would be. Third, we acknowledge that collating a complete medical history for each child would provide a more comprehensive review of a child's history of infections. However, this is not possible in Hong Kong, where most primary care is privately provided. Although most hospital use is in the public sector, there is some limited private hospital use. However, parents who brought their infants to MCHCs and Student Health Service Centres were most likely to bring their children to public rather than private hospitals, as MCHCs and Student Health Service Centres were free and public hospitals were almost free. Moreover, we used accidents as a counterexample to give an indication of the level of residual confounding. Fourth, breast development misclassification by visual assessment is more likely in overweight girls (57). In our study, the prevalence of International Obesity Task Force-defined overweight (including obesity) at 7 years was 11.5% in girls; however, there was no association between childhood overweight and number of hospital admissions for infections (data not shown). Fifth, pubertal development was assessed by different physicians at 12 Student Health Service Centres. A standard guideline for staging and an orchiometer is available. Any random variation in pubertal assessment would reduce the precision of estimates. Finally, there may be differences in the use of hospitals by cause (e.g., accidents compared with infections), which would reduce the validity of using accidents as a control exposure.

There are several possible explanations for our observed association between specifically early life infections (in the first 6 months) and pubertal onset at an older age in girls. First, infant girls who were hospitalized before 6 months of age were smaller at birth, which could be associated with puberty at younger ages (58). However, we found an association between older age at pubertal onset adjusted for birth weight and gestational age. Second, infants hospitalized for infections might have congenital abnormalities that make them more vulnerable to infections. However, if congenital illness were the underlying cause, we would have expected the same association in boys. Third, several maternal characteristics potentially affect age of puberty, such as maternal age at menarche, prepregnancy BMI, and smoking during pregnancy (59). However, maternal adiposity and smoking, which tend to be positively associated with infant infections (60, 61) and negatively associated with the timing of puberty (59), are relatively uncommon in our population (62, 63) and thus may be less influential confounders. Further, although maternal age at menarche is a strong predictor of age of puberty (64), we have no reason to think that a mother's later menarche would increase specifically her infant daughter's susceptibility to infections. Moreover, sex-specific findings are unlikely to be due to residual or uncontrolled socioeconomic confounding. Finally, in this currently ongoing birth cohort, nearly half (42%) of the boys had pubertal onset compared with 76% of girls. We cannot rule out the possibility that the lack of association in boys could be because late-maturing boys are more susceptible to the effects of infections. However, 42% of boys had reached Tanner stage II.

Our findings may reflect the physiologic effects of infections during early infancy. One possible mechanism is growth faltering. Rapid infant weight gain is associated with earlier puberty (65). Infants with serious infectious morbidity may grow more slowly (34). However, we found that serious infections had little effect on infant growth in boys. Although girls with early infections grew slowly at 0–3 months and then faster at 3–12 months of age, we did not find any mediation by infant growth. Nonetheless, infectious morbidity may induce changes in body composition. Childhood overweight is associated with earlier puberty (66). Attaining a certain level of body weight or specifically body fat has been postulated as triggering pubertal onset (67, 68). Higher BMI at 7 years of age was associated with younger age at pubertal onset in our cohort, but there was no association between infections in early life and BMI at 7 years of age and also no mediation by childhood BMI. As such, our finding of an association between infections and puberty independent of infant or childhood growth suggests that the effect of infections is unlikely to operate via the somatotropic axis.

Alternatively, infections may suppress the gonadotropic axis. There are marked physiologic rises in levels of testosterone in boys and estrogen in girls together with marked genitalia development during the first 6 months of life (29, 30, 69), which indicates potential plasticity in early infancy. Our findings coincide with the physiologic pattern of gonadotropic axis activation; that is, associations with infections were only observed when the axis is active (“mini-puberty”), not when it is quiescent (“juvenile state”), suggesting infancy as a potentially critical period of exposure regulating pubertal onset. Experience of infections reduces sex steroid levels in animals (26) and humans (23–25), although the biologic mechanism by which the immune system in infancy permanently affects the gonadotropic axis in boys and girls remains to be elucidated. Women and men may have different life history strategies for maximizing their descendants. Women may gain more than men by playing a grandparent role (70, 71), which would make investing in the immune system at the expense of reproduction a better life history strategy for women than for men (72).

From a public health perspective, there are several implications of our findings. First, the reduced experience of and exposure to infections in early life that come with economic development may be additional factors that contribute to earlier pubertal onset in girls. Second, although secular trends of earlier puberty and taller height usually occur with improved living conditions (6, 73), they do not usually coincide (74, 75), and the biologic mechanisms that control pubertal onset may not be the same as those that control linear growth. This may explain why adopted girls have earlier puberty than do their counterparts in the country of origin but similar final heights (76, 77). Finally, the improved living conditions that accompany economic development, including more food and fewer infections, may produce higher pubertal sex steroid levels over generations (78), with corresponding implications for any related diseases, such as hormone-related cancers, suggesting that reduced exposure to infections may be one of the potential drivers of changes in patterns of adult disease with economic development.

Abbreviations

    Abbreviations
  • BMI

    body mass index

  • ICD-9-CM

    International Classification of Diseases, Ninth Revision, Clinical Modification

  • MCHCs

    Maternal and Child Health Centres

Author affiliations: Life Course and Lifestyle Epidemiology Group, School of Public Health, Li Ka Shing Faculty of Medicine, University of Hong Kong, Hong Kong Special Administrative Region, China (Man Ki Kwok, Gabriel M. Leung, Tai Hing Lam, C. Mary Schooling).

This work was a substudy of the Children of 1997 birth cohort, which was initially supported by the Health Care and Promotion Fund, Health and Welfare Bureau, Government of the Special Administrative Region, China (grant 216106). Since 2005, the birth cohort study has been funded by the Health and Health Services Research Fund (grant 03040771) and the Research Fund for the Control of Infectious Diseases in Hong Kong (grants 04050172 and 06060592), Government of the Hong Kong Special Administrative Region, and the University Research Committee Strategic Research Theme of Public Health, University of Hong Kong.

The authors thank colleagues at the Student Health Service and Family Health Service of the Department of Health for their assistance and collaboration. They also thank Dr. Connie Hui for her assistance with the record linkage and the late Dr. Connie O for coordinating the project and all of the fieldwork for the initial study in 1997–1998.

Conflict of interest: none declared.

APPENDIX

Appendix Table 1.

Complete-Case Analysis for Adjusteda Associations Between Number of Hospital Admissions for Infections and Accidents at Different Stages and Age at Pubertal Onset in the Hong Kong Children of 1997 Birth Cohort, Hong Kong, China, 1997–2004

No. of Hospital Admissions by Age Group and Cause Girls (n = 3,030)
 
Boys (n = 3,363)
 
No. Time Ratio 95% CI No. Time Ratio 95% CI 
9 days–<6 months       
    Infections       
        0 2,779 Reference  3,000 Reference  
        1 203 1.01 0.99, 1.03 277 1.01 0.99, 1.02 
        ≥2 48 1.07 1.03, 1.12 86 1.00 0.97, 1.04 
    Accidents       
        0 3,011 Reference  3,348 Reference  
        ≥1 19 0.98 0.92, 1.04 15 1.04 0.97, 1.12 
6–<24 months       
    Infections       
        0 2,528 Reference  2,608 Reference  
        1 375 1.00 0.99, 1.02 523 1.00 0.99, 1.01 
        ≥2 127 0.99 0.97, 1.02 232 1.00 0.98, 1.02 
    Accidents       
        0 2,973 Reference  3,290 Reference  
        ≥1 57 0.99 0.96, 1.03 73 0.97 0.94, 1.00 
2–<5 years       
    Infections       
        0 2,565 Reference  2,744 Reference  
        1 370 1.01 1.00, 1.03 452 1.00 0.98, 1.01 
        ≥2 95 0.99 0.96, 1.02 167 1.01 0.98, 1.03 
    Accidents       
        0 2,951 Reference  3,268 Reference  
        ≥1 79 1.03 1.00, 1.07 95 0.99 0.97, 1.02 
5–8 years       
    Infections       
        0 2,888 Reference  3,179 Reference  
        1 126 1.00 0.97, 1.03 161 0.99 0.97, 1.01 
        ≥2 16 1.00 0.94, 1.07 23 1.00 0.95, 1.06 
    Accidents       
        0 2,980 Reference  3,285 Reference  
        ≥1 50 1.01 0.97, 1.05 78 1.01 0.98, 1.04 
No. of Hospital Admissions by Age Group and Cause Girls (n = 3,030)
 
Boys (n = 3,363)
 
No. Time Ratio 95% CI No. Time Ratio 95% CI 
9 days–<6 months       
    Infections       
        0 2,779 Reference  3,000 Reference  
        1 203 1.01 0.99, 1.03 277 1.01 0.99, 1.02 
        ≥2 48 1.07 1.03, 1.12 86 1.00 0.97, 1.04 
    Accidents       
        0 3,011 Reference  3,348 Reference  
        ≥1 19 0.98 0.92, 1.04 15 1.04 0.97, 1.12 
6–<24 months       
    Infections       
        0 2,528 Reference  2,608 Reference  
        1 375 1.00 0.99, 1.02 523 1.00 0.99, 1.01 
        ≥2 127 0.99 0.97, 1.02 232 1.00 0.98, 1.02 
    Accidents       
        0 2,973 Reference  3,290 Reference  
        ≥1 57 0.99 0.96, 1.03 73 0.97 0.94, 1.00 
2–<5 years       
    Infections       
        0 2,565 Reference  2,744 Reference  
        1 370 1.01 1.00, 1.03 452 1.00 0.98, 1.01 
        ≥2 95 0.99 0.96, 1.02 167 1.01 0.98, 1.03 
    Accidents       
        0 2,951 Reference  3,268 Reference  
        ≥1 79 1.03 1.00, 1.07 95 0.99 0.97, 1.02 
5–8 years       
    Infections       
        0 2,888 Reference  3,179 Reference  
        1 126 1.00 0.97, 1.03 161 0.99 0.97, 1.01 
        ≥2 16 1.00 0.94, 1.07 23 1.00 0.95, 1.06 
    Accidents       
        0 2,980 Reference  3,285 Reference  
        ≥1 50 1.01 0.97, 1.05 78 1.01 0.98, 1.04 

Abbreviation: CI, confidence interval.

a

Adjusted for birth weight, gestational age, birth order, breastfeeding, secondhand smoke, maternal place of birth, maternal educational level, type of hospital at birth, income of household head, and Rutter score at 7 years of age.

References

1.
Golub
MS
Collman
GW
Foster
PM
, et al.  . 
Public health implications of altered puberty timing
Pediatrics
 , 
2008
, vol. 
121
 
suppl 3
(pg. 
S218
-
S230
)
2.
Lakshman
R
Forouhi
NG
Sharp
SJ
, et al.  . 
Early age at menarche associated with cardiovascular disease and mortality
J Clin Endocrinol Metab
 , 
2009
, vol. 
94
 
12
(pg. 
4953
-
4960
)
3.
Frontini
MG
Srinivasan
SR
Berenson
GS
Longitudinal changes in risk variables underlying metabolic syndrome X from childhood to young adulthood in female subjects with a history of early menarche: the Bogalusa Heart Study
Int J Obes Relat Metab Disord
 , 
2003
, vol. 
27
 
11
(pg. 
1398
-
1404
)
4.
Bellis
MA
Downing
J
Ashton
JR
Adults at 12? Trends in puberty and their public health consequences
J Epidemiol Community Health
 , 
2006
, vol. 
60
 
11
(pg. 
910
-
911
)
5.
Euling
SY
Herman-Giddens
ME
Lee
PA
, et al.  . 
Examination of US puberty-timing data from 1940 to 1994 for secular trends: panel findings
Pediatrics
 , 
2008
, vol. 
121
 
suppl 3
(pg. 
S172
-
S191
)
6.
Ong
KK
Ahmed
ML
Dunger
DB
Lessons from large population studies on timing and tempo of puberty (secular trends and relation to body size): the European trend
Mol Cell Endocrinol
 , 
2006
, vol. 
254-255
 (pg. 
8
-
12
)
7.
Cho
GJ
Park
HT
Shin
JH
, et al.  . 
Age at menarche in a Korean population: secular trends and influencing factors
Eur J Pediatr
 , 
2010
, vol. 
169
 
1
(pg. 
89
-
94
)
8.
Ma
HM
Du
ML
Luo
XP
, et al.  . 
Onset of breast and pubic hair development and menses in urban Chinese girls
Pediatrics
 , 
2009
, vol. 
124
 
2
(pg. 
e269
-
e277
)
9.
Huen
KF
Leung
SS
Lau
JT
, et al.  . 
Secular trend in the sexual maturation of southern Chinese girls
Acta Paediatr.
 , 
1997
, vol. 
86
 
10
(pg. 
1121
-
1124
)
10.
Günther
AL
Karaolis-Danckert
N
Kroke
A
, et al.  . 
Dietary protein intake throughout childhood is associated with the timing of puberty
J Nutr
 , 
2010
, vol. 
140
 
3
(pg. 
565
-
571
)
11.
Rogers
IS
Northstone
K
Dunger
DB
, et al.  . 
Diet throughout childhood and age at menarche in a contemporary cohort of British girls
Public Health Nutr
 , 
2010
, vol. 
13
 
12
(pg. 
2052
-
2063
)
12.
Buck Louis
GM
Gray
LE
Jr
Marcus
M
, et al.  . 
Environmental factors and puberty timing: expert panel research needs
Pediatrics
 , 
2008
, vol. 
121
 
suppl 3
(pg. 
S192
-
S207
)
13.
Schoeters
G
Den Hond
E
Dhooge
W
, et al.  . 
Endocrine disruptors and abnormalities of pubertal development
Basic Clin Pharmacol Toxicol
 , 
2008
, vol. 
102
 
2
(pg. 
168
-
175
)
14.
Schooling
CM
Hui
LL
Ho
LM
Lam
TH
Leung
GM
Cohort profile: “Children of 1997”: a Hong Kong Chinese birth cohort [published online ahead of print January 11, 2010]
Int J Epidemiol.
  
(doi: 10.1093/ije/dyq243)
15.
Crimmins
EM
Finch
CE
Infection, inflammation, height, and longevity
Proc Natl Acad Sci U S A
 , 
2006
, vol. 
103
 
2
(pg. 
498
-
503
)
16.
Yeung
C
Changing pattern of childhood diseases in Hong Kong: a rapidly developing community
Curr Paediatr.
 , 
1999
, vol. 
9
 
4
(pg. 
209
-
214
)
17.
Rajaratnam
JK
Marcus
JR
Flaxman
AD
, et al.  . 
Neonatal, postneonatal, childhood, and under-5 mortality for 187 countries, 1970–2010: a systematic analysis of progress towards Millennium Development Goal 4
Lancet
 , 
2010
, vol. 
375
 
9730
(pg. 
1988
-
2008
)
18.
Buchacz
K
Rogol
AD
Lindsey
JC
, et al.  . 
Delayed onset of pubertal development in children and adolescents with perinatally acquired HIV infection
J Acquir Immune Defic Syndr
 , 
2003
, vol. 
33
 
1
(pg. 
56
-
65
)
19.
De Martino
M
Tovo
PA
Galli
L
, et al.  . 
Puberty in perinatal HIV-1 infection: a multicentre longitudinal study of 212 children
AIDS
 , 
2001
, vol. 
15
 
12
(pg. 
1527
-
1534
)
20.
Ballinger
AB
Savage
MO
Sanderson
IR
Delayed puberty associated with inflammatory bowel disease
Pediatr Res.
 , 
2003
, vol. 
53
 
2
(pg. 
205
-
210
)
21.
Khan
AD
Schroeder
DG
, et al.  . 
Early childhood determinants of age at menarche in rural Guatemala
Am J Hum Biol.
 , 
1996
, vol. 
8
 
6
(pg. 
717
-
723
)
22.
Blell
M
Pollard
TM
Pearce
MS
Predictors of age at menarche in the Newcastle Thousand Families Study
J Biosoc Sci.
 , 
2008
, vol. 
40
 
4
(pg. 
563
-
575
)
23.
Ibrahim
II
Barakat
RM
Bassiouny
HK
, et al.  . 
Effect of urinary bilharzial infection on male pubertal development and endocrine functions
Arch Androl
 , 
1983
, vol. 
11
 
1
(pg. 
59
-
64
)
24.
Mavoungou
D
Lansoud-Soukate
J
Dupont
A
Steroid and gonadotropin hormone levels in young African women with filarial infection
J Steroid Biochem
 , 
1989
, vol. 
34
 
1–6
(pg. 
577
-
580
)
25.
Muehlenbein
MP
Alger
J
Cogswell
F
, et al.  . 
The reproductive endocrine response to Plasmodium vivax infection in Hondurans
Am J Trop Med Hyg
 , 
2005
, vol. 
73
 
1
(pg. 
178
-
187
)
26.
Larralde
C
Morales
J
Terrazas
I
, et al.  . 
Sex hormone changes induced by the parasite lead to feminization of the male host in murine Taenia crassiceps cysticercosis
J Steroid Biochem Mol Biol.
 , 
1995
, vol. 
52
 
6
(pg. 
575
-
580
)
27.
Boonekamp
JJ
Ros
AH
Verhulst
S
Immune activation suppresses plasma testosterone level: a meta-analysis
Biol Lett.
 , 
2008
, vol. 
4
 
6
(pg. 
741
-
744
)
28.
Nilsson
C
Jennische
E
Ho
HP
, et al.  . 
Postnatal endotoxin exposure results in increased insulin sensitivity and altered activity of neuroendocrine axes in adult female rats
Eur J Endocrinol
 , 
2002
, vol. 
146
 
2
(pg. 
251
-
260
)
29.
Ji
C
Huang
XW
Yang
RW
, et al.  . 
Gonadotropins and sex hormones in healthy Chinese infants
Indian Pediatr
 , 
2008
, vol. 
45
 
6
(pg. 
489
-
492
)
30.
Chemes
HE
Infancy is not a quiescent period of testicular development
Int J Androl
 , 
2001
, vol. 
24
 
1
(pg. 
2
-
7
)
31.
Beunen
GP
Rogol
AD
Malina
RM
Indicators of biological maturation and secular changes in biological maturation
Food Nutr Bull.
 , 
2006
, vol. 
27
 
suppl 4
(pg. 
S244
-
S256
)
32.
McDade
TW
Life history theory and the immune system: steps toward a human ecological immunology
Am J Phys Anthropol
 , 
2003
suppl 37
(pg. 
S100
-
S125
)
33.
McDade
TW
Life history, maintenance, and the early origins of immune function
Am J Hum Biol.
 , 
2005
, vol. 
17
 
1
(pg. 
81
-
94
)
34.
Panter-Brick
C
Lunn
PG
Langford
RM
, et al.  . 
Pathways leading to early growth faltering: an investigation into the importance of mucosal damage and immunostimulation in different socio-economic groups in Nepal
Br J Nutr
 , 
2009
, vol. 
101
 
4
(pg. 
558
-
567
)
35.
Leung
SS
Peng
CX
Xu
YY
, et al.  . 
Comparative study of growth of Chinese infants: Hong Kong versus Guangzhou
J Trop Pediatr
 , 
1994
, vol. 
40
 
3
(pg. 
166
-
171
)
36.
Family Health Service, Government of Hong Kong Special Administrative Region
Fees and Charges for Maternal & Child Health and Woman Health Services
 , 
2006
Hong Kong, China
Government of Hong Kong Special Administrative Region
 
37.
Lam
TH
Leung
GM
Ho
LM
The effects of environmental tobacco smoke on health services utilization in the first eighteen months of life
Pediatrics
 , 
2001
, vol. 
107
 
6
pg. 
E91
 
38.
Census and Statistics Department
Chapter 8: Health
 , 
2007
Hong Kong Special Administrative Region, China: Government Printer
 
39.
Census and Statistics Department
Chapter 8: Health
 , 
2005
Hong Kong Special Administrative Region, China: Government Printer
 
40.
Hui
LL
Schooling
CM
Wong
MY
, et al.  . 
Infant growth during the first year of life and subsequent hospitalization to 8 years of age
Epidemiology
 , 
2010
, vol. 
21
 
3
(pg. 
332
-
339
)
41.
Kwok
MK
Schooling
CM
Ho
LM
, et al.  . 
Early life second-hand smoke exposure and serious infectious morbidity during the first 8 years: evidence from Hong Kong's “Children of 1997” birth cohort
Tob Control
 , 
2008
, vol. 
17
 
4
(pg. 
263
-
270
)
42.
Tarrant
M
Kwok
MK
Lam
TH
, et al.  . 
Breast-feeding and childhood hospitalizations for infections
Epidemiology
 , 
2010
, vol. 
21
 
6
(pg. 
847
-
854
)
43.
Karlberg
JA
biologically-oriented mathematical model (ICP) for human growth
Acta Paediatr Scand Suppl
 , 
1989
, vol. 
350
 (pg. 
70
-
94
)
44.
Leung
GM
Ho
LM
Tin
KY
, et al.  . 
Health care consequences of cesarean birth during the first 18 months of life
Epidemiology
 , 
2007
, vol. 
18
 
4
(pg. 
479
-
484
)
45.
Diehr
P
Yanez
D
Ash
A
, et al.  . 
Methods for analyzing health care utilization and costs
Annu Rev Public Health
 , 
1999
, vol. 
20
 
1
(pg. 
125
-
144
)
46.
Hui
LL
Schooling
CM
Leung
SS
, et al.  . 
Birth weight, infant growth, and childhood body mass index: Hong Kong's Children of 1997 birth cohort
Arch Pediatr Adolesc Med
 , 
2008
, vol. 
162
 
3
(pg. 
212
-
218
)
47.
Kwok
MK
Schooling
CM
Lam
TH
, et al.  . 
Paternal smoking was associated with childhood overweight: evidence from the Hong Kong “Children of 1997”
Pediatrics
 , 
2010
, vol. 
126
 
1
(pg. 
e46
-
e56
)
48.
World Health Organization
Growth Reference 5–19 Years
 , 
2007
Geneva, Switzerland
World Health Organization
 
(http://www.who.int/growthref/en/). (Accessed April 10, 2010)
49.
Marshall
WA
Tanner
JM
Variations in the pattern of pubertal changes in boys
Arch Dis Child
 , 
1970
, vol. 
45
 
239
(pg. 
13
-
23
)
50.
Marshall
WA
Tanner
JM
Variations in pattern of pubertal changes in girls
Arch Dis Child
 , 
1969
, vol. 
44
 
235
(pg. 
291
-
303
)
51.
Peto
R
Experimental survival curves for interval-censored data
J R Stat Soc Ser C (Appl Stat)
 , 
1973
, vol. 
22
 
1
(pg. 
86
-
91
)
52.
Sun
X
Briel
M
Walter
SD
, et al.  . 
Is a subgroup effect believable? Updating criteria to evaluate the credibility of subgroup analyses
BMJ
 , 
2010
, vol. 
340
 pg. 
c117
 
53.
Baron
RM
Kenny
DA
The moderator-mediator variable distinction in social psychological research: conceptual, strategic, and statistical considerations
J Pers Soc Psychol
 , 
1986
, vol. 
51
 
6
(pg. 
1173
-
1182
)
54.
Harrell
FE
Regression Modeling Strategies: With Applications to Linear Models, Logistic Regression, and Survival Analysis
 , 
2001
New York, NY
Springer-Verlag
55.
Moons
KG
Donders
RA
Stijnen
T
, et al.  . 
Using the outcome for imputation of missing predictor values was preferred
J Clin Epidemiol
 , 
2006
, vol. 
59
 
10
(pg. 
1092
-
1101
)
56.
Schafer
JL
Multiple imputation: a primer
Stat Methods Med Res.
 , 
1999
, vol. 
8
 
1
(pg. 
3
-
15
)
57.
Wang
Y
Is obesity associated with early sexual maturation? A comparison of the association in American boys versus girls
Pediatrics
 , 
2002
, vol. 
110
 
5
(pg. 
903
-
910
)
58.
Persson
I
Ahlsson
F
Ewald
U
, et al.  . 
Influence of perinatal factors on the onset of puberty in boys and girls: implications for interpretation of link with risk of long term diseases
Am J Epidemiol
 , 
1999
, vol. 
150
 
7
(pg. 
747
-
755
)
59.
Maisonet
M
Christensen
KY
Rubin
C
, et al.  . 
Role of prenatal characteristics and early growth on pubertal attainment of British girls
Pediatrics
 , 
2010
, vol. 
126
 
3
(pg. 
e591
-
e600
)
60.
Håberg
SE
Stigum
H
London
SJ
, et al.  . 
Maternal obesity in pregnancy and respiratory health in early childhood
Paediatr Perinat Epidemiol
 , 
2009
, vol. 
23
 
4
(pg. 
352
-
362
)
61.
Håberg
SE
Stigum
H
Nystad
W
, et al.  . 
Effects of pre- and postnatal exposure to parental smoking on early childhood respiratory health
Am J Epidemiol
 , 
2007
, vol. 
166
 
6
(pg. 
679
-
686
)
62.
Janus ED
Table 6.3. Body mass index (BMI)
Hong Kong Cardiovascular Risk Factor Prevalence Study 1995–1996
 , 
1997
Hong Kong
China: Department of Clinical Biochemistry, Queen Mary Hospital of Hong Kong
63.
Tobacco Control Office, Government of Hong Kong Special Administrative Region
Statistics: Pattern of Smoking in Hong Kong
 , 
2008
Hong Kong, China
Government of Hong Kong Special Administrative Region
 
64.
Towne
B
Czerwinski
SA
Demerath
EW
, et al.  . 
Heritability of age at menarche in girls from the Fels Longitudinal Study
Am J Phys Anthropol
 , 
2005
, vol. 
128
 
1
(pg. 
210
-
219
)
65.
Dunger
DB
Ahmed
ML
Ong
KK
Early and late weight gain and the timing of puberty
Mol Cell Endocrinol
 , 
2006
, vol. 
254–255
 (pg. 
140
-
145
)
66.
Rosenfield
RL
Lipton
RB
Drum
ML
Thelarche, pubarche, and menarche attainment in children with normal and elevated body mass index
Pediatrics
 , 
2009
, vol. 
123
 
1
(pg. 
84
-
88
)
67.
Johnston
FE
Malina
RM
Galbraith
MA
, et al.  . 
Height, weight and age at menarche and the “critical weight” hypothesis
Science
 , 
1971
, vol. 
174
 
14
(pg. 
1148
-
1149
)
68.
Scott
EC
Johnston
FE
Critical fat, menarche, and the maintenance of menstrual cycles: a critical review
J Adolesc Health Care
 , 
1982
, vol. 
2
 
4
(pg. 
249
-
260
)
69.
Garagorri
JM
Rodríguez
G
Lario-Elboj
AJ
, et al.  . 
Reference levels for 17-hydroxyprogesterone, 11-desoxycortisol, cortisol, testosterone, dehydroepiandrosterone sulfate and androstenedione in infants from birth to six months of age
Eur J Pediatr
 , 
2008
, vol. 
167
 
6
(pg. 
647
-
653
)
70.
Lahdenperä
M
Lummaa
V
Helle
S
, et al.  . 
Fitness benefits of prolonged post-reproductive lifespan in women
Nature
 , 
2004
, vol. 
428
 
6979
(pg. 
178
-
181
)
71.
Jasienska
G
Reproduction and lifespan: Trade-offs, overall energy budgets, intergenerational costs, and costs neglected by research
Am J Hum Biol.
 , 
2009
, vol. 
21
 
4
(pg. 
524
-
532
)
72.
Rolff
J
Bateman's principle and immunity
Proc Biol Sci.
 , 
2002
, vol. 
269
 
1493
(pg. 
867
-
872
)
73.
Ji
CY
Chen
TJ
Secular changes in stature and body mass index for Chinese youth in sixteen major cities, 1950s–2005
Am J Hum Biol.
 , 
2008
, vol. 
20
 
5
(pg. 
530
-
537
)
74.
Karlberg
J
Secular trends in pubertal development
Horm Res.
 , 
2002
, vol. 
57
 
suppl 2
(pg. 
S19
-
S30
)
75.
Cole
TJ
Secular trends in growth
Proc Nutr Soc.
 , 
2000
, vol. 
59
 
2
(pg. 
317
-
324
)
76.
Proos
LA
Growth & development of Indian children adopted in Sweden
Indian J Med Res.
 , 
2009
, vol. 
130
 
5
(pg. 
646
-
650
)
77.
Tuvemo
T
Proos
LA
Girls adopted from developing countries: a group at risk of early pubertal development and short final height. Implications for health surveillance and treatment
Ann Med
 , 
1993
, vol. 
25
 
3
(pg. 
217
-
219
)
78.
Schooling
CM
Leung
GM
A socio-biological explanation for social disparities in non-communicable chronic diseases: the product of history?
J Epidemiol Community Health
 , 
2010
, vol. 
64
 
11
(pg. 
941
-
949
)