A recently published review on impact of body fat and timing of puberty concluded earlier onset of puberty in girls with increased body mass index (BMI) [4]. However, a study on Indian Bengali girls showed no correlation between menarcheal age and BMI [5]. Thus, there is paucity of data on impact of increasing obesity on pubertal development and progression in Indian children.

We therefore conducted this study to assess age of menarche, and impact of obesity on pubertal development in apparently healthy Indian school children, aged 6-17 years.

This cross-sectional study was carried out in seven schools selected from seven out of 11 districts of Delhi representative of all geographical regions (North, South, East, West and Central region) after obtaining permission from school management. All these schools were private schools catering to middle or upper socio-economic status, and were selected as per convenience. All seven schools which were approached agreed to participate. They were given detailed written information sheet regarding the study, to be sent to parents of all eligible children. Interaction with parents was undertaken to clear their queries about the study. Written informed consent from parents / guardians and verbal assent was taken from the children who participated in the study. The study was approved by Institutional ethics committee, All India Institute of Medical Sciences (AIIMS), New Delhi.

Apparently healthy girls aged between 6 and 17 years were recruited from January 2013 to December 2013. Children with history of any systemic illness or taking any treatment for more than one month in last three months were excluded from the study. Girls underwent tailored history and clinical examination, including anthropometry, followed by blood sample collection by trained phlebotomists. All measurements were made with participants dressed in minimal light clothing without footwear. Height was measured with portable Holtain’s stadiometer (Holtain Inc., Crymych, Pembs. UK). Weight was measured with the digital weighing machine. BMI was calculated and status of obesity and overweight was determined as per International Obestity Task Force (IOTF) criteria [6]. Pubertal stage assessment was using Tanner’s method carried out by trained professional women [7]. Status quo method was used to determine the median age of menarche. In status quo method, data for menarcheal age is obtained by asking subject of her ‘current menstrual status’ whether she has achieved menarche or not, and her date of birth [8]

Assuming a standard deviation of age at onset for B2 and G2 as 2 years, with absolute error of margin ± 0.25 year (3 months), it was calculated that in a two-sided 95% confidence interval, 246 girls would be needed. For other stages from B3 - B5 and G3 - G5 similar number were adopted. So considering all genital stages, a total of 2000 was calculated to be sufficient.

Statistical Analysis: Median and 95 % CI for pubertal age was calculated by probit analysis using SPSS Software version 20. For correlation of hormone data with breast stages, one-way ANOVA was used. Bonferroni correction was applied to find out level of significance between various pubertal stages, and gonadotropins and sex steroids. Log transformation of data was done to make data normally distributed. Data were adjusted for age and BMI to look for their effect on the gonadotropins and sex steroids level with relation to pubertal status. ROC tables and graph were obtained to find out threshold value of the gonadotropins and sex steroids level in relation to onset of puberty i.e. progression from stage 1 to stage 2. All tests were done keeping a probability of type 1 error at 5%.

A total of 2015 girls were recruited for the study. Exclusion of five girls for various reasons (non-availability of date of birth, and age >18 years) resulted in a final tally of 2010 girls. Details of all subjects in different age groups along with their height, weight and BMI are shown in Table I. Almost one-fourth (23.8%) girls were either overweight or obese.

TABLE I  Age and Anthropometric Characteristics of the Study Participants (N=2010)

Distribution of girls with different stages of breast and pubic hair development according to their BMI is shown in Table II. Table III shows the median age of onset of various breast and pubic hair stages. The median age of onset of thelarche (B2) was 10.8 years (95% CI -10.7- 10.9 years), and for pubarche it was 11.0 years (95% CI 10.8-11.2 y).

TABLE II	Distribution of Breast and Pubic Hair Stages According to BMI of Girls

TABLE III	Estimated Median Age (y) of Onset of Breast and Pubic Hair Stages Irrespective of BMI Status

TABLE IV Age (y) of Development of Breast and Menarche in Relation to Body Mass Index Status

A significant shift towards early onset of puberty by six months (P<0.05) and early progression from stage 2 to 5 was observed in overweight/ obese girls (Table IV). Similarly, appearance as well as progression of pubic hair was approximately 6-9 months earlier in overweight/ obese girls (P<0.05).

Overweight/obese girls attained menarche approximately four months earlier than girls with normal BMI (P<0.05). The time duration of approximately one and half years was seen between onset of thelarche and menarche, and was not different in overweight/ obese and normal BMI subjects (Table IV).

Blood samples were collected from 335 subjects with mean (SD) age 11.7 (0.9) years. Of these, 64 (19.1%) had BMI in overweight/ obese range. The mean (SD) levels of serum LH, FSH and estradiol were 4.21 (6.75) IU/L, 5.15 (2.48) IU/L and 64.5 (67.6) pg/mL, respectively. The distribution of LH, FSH and estradiol in different stages of breast development is shown in Web-Table I. Though serum LH progressively increased from B1 to B5, significant difference was observed between B1 and B2 only. Similarly, significant increase in serum FSH was seen between stage 1 and 2 only but the peak serum FSH levels were seen in stage 3, followed by a small fall in the consequent stages. Serum estradiol progressively increased from stage 1 to stage 5 with significant difference being noted between each stage till stage 4. Serum LH level of 0.88 IU/L showed sensitivity and specificity of 68.9% and 69.6% respectively for prediction of thelarche (B2). Similarly, sensitivity and specificity values for E2 >34.5 pg/mL were 69.9% and 69%, respectively (Web Table II). No impact of overweight/ obese was observed on serum levels of LH, FSH and estradiol levels.

The present study provides the normative data regarding age of pubertal onset in healthy Indian schoolgirls. We documented an early onset of puberty in obese and overweight girls. We also showed that serum levels of LH, FSH and estradiol correlate well with stage of puberty.

The present study has several limitations. The main limitation is use of status quo method for assessment of age of menarche. Prospective method by follow-up every 3-6 months would be more accurate [9]. Other limitations are: (a) limited number of subjects in the age group of 6-8 and 15-17 years; (b) blood samples for hormonal analysis were not collected from all girls; (c) all children were considered healthy on the basis of history alone; and (d) BMI, which may not truly represent percentage body fat, was used for defining obesity.

The present study reports the median age of thelarche to be 10.8 years which is comparable to the earlier Indian report of 1992 [4]. Among published studies in last 10 years, pubertal onset occurred earliest among Chinese girls (9.2 years) [10], followed by American girls [11,12], and then the Danish girls [13], which highlights that racial differences affect pubertal timing. The median age of onset of pubarche in our study was not different from other races, except black American population, where it is documented to occur earlier [11,12,14].

The mean/median age of menarche in studies from different parts of the world, including India, ranged from 12-13.4 years [10-15], similar to our findings. Menarche is also reported to occur earlier in black population than whites [11,14]. The duration of progression from thelarche to menarche in our study was shorter in contrast to 2.4-3 years as observed in some other studies [10,13,14]. A secular trend of decline in average menarcheal age which was reported earlier [1] has gradually either slowed down or stopped in last few decades. A recent report has even showed a modest increase in the average age of menarche [16].

The early onset of puberty among obese girls as documented in our study, has been previously reported [6, 16-19]. In our cohort, the gonadotropins and estradiol progressively increased with pubertal stages with FSH peak at stage 3 of puberty as documented earlier [20]. The best sensitivity and specificity cutoffs of LH and E2 are still prone to misclassify true thelarche if interpreted in absence of clinical examination. It would therefore be appropriate to combine clinical examination along with LH levels for prediction of stage B2, especially in overweight/ obese girls where clinical examination alone at times, may mislead thelarche. Though, the clinical observation in the present study suggest early onset of thelarche and menarche in overweight/obese girls, serum gonadotropin levels were lower than girls with normal BMI, as reported earlier [21]. This paradoxical observation has been suggested to result from excess adiposity, which in the absence of substantial sex steroid excess, may subtly suppress hypothalamic-pituitary-gonadal function in early pubertal girls [21].

The present study provides normative data for age of onest of puberty for Indian girls, and documents an earlier onset of puberty in overweight/obese girls. These data should be validated by longitudinal studies in different settings within the country.

Contributors: RK, RKM: conceived and designed thestudy, and supervised manuscript writing; VS, AD: designed the research plan, analyzed the data, and drafted the manuscript; NM: data collection; VS: analyzed the data and performed the statistical analysis; MGA: manuscript preparation; NG: supervised laboratory work. All authors read and approved the final manuscript.

Funding: None; Competing interest: None stated.