Contemporary age at menarche among Thai girls: a nationwide cross-sectional study and its public health implications

Article information

Ann Pediatr Endocrinol Metab. 2026;31(5):292-300
Publication date (electronic) : 2026 October 31
doi : https://doi.org/10.6065/apem.2550462.231
1Faculty of Political Science, Chulalongkorn University, Bangkok, Thailand
2Department of Pediatrics, King Chulalongkorn Memorial Hospital, Bangkok, Thailand
3Division of Pediatric Endocrinology, Department of Pediatrics, Faculty of Medicine, Chulalongkorn University, Bangkok, Thailand
4Faculty of Sociology and Anthropology, Thammasat University, Bangkok, Thailand
5Valaya Alongkorn Rajabhat University Under the Royal Patronage, Pathumthani, Thailand
6Institute of Future Studies for Development, Bangkok, Thailand
Address for correspondence: Taninee Sahakitrungruang Division of Pediatric Endocrinology, Department of Pediatrics, Faculty of Medicine, Chulalongkorn University, Bangkok 10330, Thailand Email: taninee.p@chula.ac.th
*These authors contributed equally to this study as co-first authors.
Received 2025 November 17; Revised 2026 February 15; Accepted 2026 March 6.

Abstract

Purpose

Age at menarche has declined globally. However, contemporary population-level data from Southeast Asia remain limited. This study aimed to characterize menarcheal timing among Thai schoolgirls and identify associated factors using survival-based methods.

Methods

A nationwide cross-sectional study was conducted among 8,161 Thai schoolgirls in grades 3–9. Menarcheal status and age at menarche were assessed using questionnaires and verified by caregivers. Kaplan-Meier survival analysis was used to estimate the cumulative probability of menarche by age, accounting for right-censoring. Determinants of menarcheal timing were examined using Cox proportional hazards regression models.

Results

At assessment, 52.6% of participants had attained menarche. Kaplan-Meier analysis demonstrated that the cumulative probability of menarche rose from 5.56% at 10 years to 85.21% at 13 years, with a median age of 12.0 years (interquartile range, 11.2–12.7 years). Based on survival-derived percentiles, early menarche (approximately <5th percentile) was defined as occurring before 10 years of age. Significant regional differences in menarcheal timing were observed (log-rank test, P<0.001). In multivariable Cox regression analysis, overweight (hazard ratio [HR], 1.67), obesity (HR, 1.64), and high socioeconomic status (HR, 1.28) were independently associated with earlier menarche, whereas underweight status was associated with delayed menarche (HR, 0.45; all P<0.001). After adjustment, only the Southern region remained significantly associated with later menarche compared with Bangkok.

Conclusions

Nutritional status and socioeconomic factors are key factors associated with menarcheal timing in Thailand. Survival-based, percentile-defined thresholds provide a useful framework for identifying atypical pubertal timing and may inform population-level public health strategies.

Highlights

· This nationwide study of 8,161 Thai schoolgirls in grades 3–9 estimated a median age at menarche of 12.0 years using Kaplan-Meier survival analysis.

· Earlier menarche was independently associated with overweight/obesity and higher socioeconomic status, while menarche before age 10 years corresponded to approximately below the 5th percentile.

· These findings provide population-based reference data and support early puberty education and healthy weight promotion among school-aged girls.

Introduction

Puberty is a crucial developmental phase marked by significant physiological changes that prepare individuals for reproduction. Menarche, the onset of menstruation, signifies an advanced stage of female sexual development and is a key milestone of pubertal development. The secular trend of declining pubertal timing in girls is a well-documented global phenomenon. A comprehensive meta-analysis has shown that the age at pubertal onset, as marked by thelarche (breast development), has decreased by approximately 3 months per decade since the 1970s [1]. While this trend may be stabilizing in some high-income nations, it continues in many transitioning economies [2,3], underscoring the critical need for current, country-specific data from regions such as Southeast Asia.

This secular trend is particularly pronounced across Asia, as confirmed by a recent pooled analysis of over half a million women from 6 Asian countries [4]. Recent national data from South Korea, for example, report a decline of 0.71 years per decade, resulting in a mean menarcheal age of 12.45 years for recent birth cohorts [2], while the median age in urban China has fallen to 12.0 years [5]. In India, a similar decline has been confirmed, with a current mean age of 13.49 years [6]. However, comprehensive, nationally representative data from Southeast Asia, particularly the Indochinese Peninsula, remain limited. This study, with its large nationwide sample, aimed to address this gap.

The public health imperative to monitor pubertal timing is underscored by a growing body of evidence, including Mendelian randomization studies, suggesting that early menarche may causally increase the lifetime risk of numerous non-communicable diseases—such as type 2 diabetes, cardiovascular disease, and hormone-related cancers—as well as adverse mental health outcomes and accelerated biological aging [7-9].

The onset of puberty and menarche is influenced by a complex interplay of genetic and environmental factors, including lifestyle, socioeconomic status (SES), nutrition, and environmental exposures, all of which have changed over time. This study aimed to characterize current patterns of menarcheal timing among girls in Thailand.

Materials and methods

1. Participants

A nationally representative sample of 8,161 girls was enrolled in the study between December 2019 and March 2020. The inclusion criteria comprised schoolgirls in grades 3–9, attending either private or public schools across 5 regions of Thailand (Central, North, North-East, South, and Bangkok) from stratified samples by region, province, urban/rural, private/public school, and grade attended. Exclusion criteria included girls who declined participation due to discomfort with sensitive questions related to sexual behavior or hygiene. The questionnaire included specific questions regarding a history of precocious puberty and treatment with gonadotropin-releasing hormone (GnRH) analogs. However, other endocrine conditions or medical histories potentially affecting pubertal timing were not systematically assessed. Participants with a reported history of precocious puberty and GnRH analog treatment were retained in the analysis. Given their small proportion in the sample and the use of survival-based analytic methods, their inclusion was considered unlikely to materially influence population-level estimates.

2. Data collection and measurements

Data for this cross-sectional study were collected during the survey period from December 2019 to March 2020. Permission was obtained from school authorities. Chronological age and age at menarche were recorded in years and months, whereas weight (kg) and height (cm) were measured using standard equipment by the research team. Body mass index (BMI) was calculated using the formula BMI = weight (kg)/height (m)². The standard deviation scores (SDS) for weight, height, and BMI were calculated using World Health Organization growth curves. BMI SDS (z-scores) were classified into 4 categories: "underweight" (less than -2 standard deviation [SD]), "normal weight" (≥ -2 to ≤ +1 SD), "overweight" (> +1 to ≤ +2 SD), and "obese" (> +2 SD).

The primary outcome variable of this study was the age at menarche among schoolgirls in grades 3–9. Menarcheal status was assessed using a standardized questionnaire that asked whether the participant had attained menarche (yes/no). For those who responded "yes," age at menarche was recorded to the nearest month. Menarcheal age was verified through telephone interviews with mothers or primary caregivers. When discrepancies occurred, caregiver-confirmed age was used.

3. Sampling procedure

Sampling was stratified by school type and geographic area to account for potential regional differences in lifestyle. The sampling framework was as follows (Supplementary Table 1):

• Step 1: Stratified sampling by age range

The sample included students from grades 3–9, covering the period during which substantial pubertal changes occur.

• Step 2: Stratified sampling by region

The regions included Bangkok Metropolis and the North, Central, North-East, and Southern regions. Each region was then stratified by urban or rural area.

• Step 3: Stratified sampling by school type

The sample included girls from government and private schools.

• Step 4: Random sampling within selected schools

Random sampling was performed within selected schools according to the inclusion and exclusion criteria for the study.

4. Statistical analysis

Statistical analyses were performed using Stata 17 (StataCorp LLC, USA). Continuous variables are presented as means, and categorical variables as frequencies and percentages. In the Kaplan-Meier analysis, attainment of menarche was defined as the event, with age in months used as the time scale. For participants who had not yet attained menarche, observations were right-censored at their current age at the time of the survey (in months). The cumulative probability of menarche by age was estimated from the Kaplan-Meier nonparametric analysis. Differences in menarcheal timing across regions were evaluated using the log-rank test. Associations between age at menarche and potential determinants were examined using Cox proportional hazards regression models. Univariable analyses were first performed, followed by multivariable models adjusting for region, area of residence, nutritional status (BMI z-score category), and SES. Hazard ratios (HRs) with 95% confidence intervals (CIs) are reported. A 2-sided P-value<0.05 was considered statistically significant.

5. Ethical considerations

The study was approved by the institutional review board (IRB) of the Faculty of Medicine, Chulalongkorn University (IRB No. 579/62). Written informed consent was obtained from the parents or legal guardians of all participants prior to enrollment, in accordance with international guidelines for human research protection. These guidelines include the Declaration of Helsinki, the Belmont Report, the Council for International Organizations of Medical Sciences guidelines, and the International Conference on Harmonization Good Clinical Practice guideline.

Results

A total of 8,161 schoolgirls in grades 3–9 from 95 schools across 5 regions of Thailand were enrolled between December 2019 and March 2020. Of these, 4,293 girls (52.6%) had attained menarche at the time of assessment, whereas the remaining participants were premenarcheal and treated as right-censored observations. Fifty-six participants (0.69%) had a documented history of precocious puberty and were receiving or had previously received treatment with GnRH analogs; these participants were included in the analysis as part of the study sample. Demographic and clinical characteristics are summarized in Table 1. The overall mean BMI z-score of the cohort was 0.28, and 13.2% of participants were classified as obese.

Demographic data

1. Kaplan-Meier estimation of age at menarche

Kaplan-Meier survival analysis demonstrated a rapid increase in the cumulative probability of menarche between ages 10 and 13 years (Fig. 1A). The estimated cumulative probability of menarche was 5.56% at age 10 years, 55.83% at age 12 years, and 85.21% at age 13 years (Table 2). The cumulative probability of menarche exceeded 50% at 12 years, corresponding to a median age at menarche of 12.0 years (interquartile range [IQR], 11.2–12.7 years). By age 15 years, the estimated cumulative probability of menarche exceeded 99%, indicating that nearly all girls were estimated to have attained menarche by midadolescence.

Fig. 1.

Kaplan-Meier estimates of the cumulative probability of menarche by age. (A) Overall cumulative probability of menarche among Thai schoolgirls. (B) Cumulative probability of menarche stratified by geographic region (Bangkok Metropolis, Central, North, North-East, and South). Differences across regions were assessed using the log-rank test (P<0.001).

Probability of menarche at each age (in years and months) derived from Kaplan-Meier estimation

2. Regional differences in menarcheal timing

Kaplan-Meier curves stratified by geographic region revealed significant differences in menarcheal timing across regions (overall log-rank test, P<0.001). Girls residing in Bangkok had the earliest menarcheal timing, whereas those in the Southern region had the latest menarcheal timing (Fig. 1B; Supplementary Table 2).

3. Factors associated with menarcheal timing

1) Univariable Cox proportional hazards analysis

In univariable Cox regression analysis, rural residence was associated with a lower hazard of menarche compared with urban residence (HR, 0.93; 95% CI, 0.87–0.98). Compared with Bangkok, all other regions had significantly lower hazards of menarche. Nutritional status showed a strong association with menarcheal timing: overweight and obese girls had higher hazards of attaining menarche, whereas underweight girls had a lower hazard of menarche. Higher SES was also associated with earlier menarche (Table 3).

Univariable Cox proportional hazards regression for menarcheal timing

2) Multivariable Cox proportional hazards analysis

In the multivariable Cox proportional hazards model, which adjusted for all significant univariable variables, nutritional status and SES were significant independent predictors of menarcheal timing. Compared with normal-weight girls, overweight (HR, 1.67; 95% CI, 1.53–1.82) and obese girls (HR, 1.64; 95% CI, 1.48–1.82) had significantly higher hazards of menarche, while underweight girls experienced delayed menarche (HR, 0.45; 95% CI, 0.36–0.56). Girls from high-SES (HR, 1.28; 95% CI, 1.17–1.40) and moderate-SES (HR, 1.12; 95% CI, 1.03–1.21) households had significantly higher hazards of menarche than those from low-SES backgrounds.

After adjustment for nutritional status and socioeconomic factors, a regional difference persisted only for the Southern region, which showed a significantly lower hazard of menarche compared with Bangkok (HR, 0.74; 95% CI, 0.64–0.86). Area of residence (urban vs. rural) was no longer significantly associated with menarcheal timing after multivariable adjustment (Table 4).

Multivariable Cox proportional hazards regression for menarcheal timing

4. Kaplan-Meier-based definitions of early and late menarche

Based on the Kaplan-Meier cumulative probability estimates (Table 2; Fig. 1A), the cumulative probability of menarche increased from 3.20% at 9 years 11 months to 5.56% at 10 years. Accordingly, early menarche was defined as menarche occurring before 10 years of age, corresponding to approximately below the 5th percentile of the population distribution derived from survival analysis. This percentile-based definition appropriately accounts for right-censoring and characterizes the lower tail of the population distribution.

Late menarche was defined as menarche occurring at age 14 years or later, corresponding to approximately the 95th percentile of the Kaplan-Meier distribution. The median age at menarche (50th percentile) was estimated to be 12 years (IQR, 11.2–12.7 years), indicating that half of the study population had attained menarche by this age.

Discussion

The primary objective of this nationwide study was to characterize contemporary patterns of menarcheal timing among Thai adolescents using survival-based methods. By applying Kaplan-Meier analysis to account for substantial right-censoring, we demonstrated that the cumulative probability of menarche increased sharply between ages 10 and 13 years, with a median age at menarche of 12 years. The estimated cumulative probability of menarche exceeded 99% by age 15 years, indicating that menarche among Thai girls generally occurs within a relatively narrow window during early adolescence.

A major strength of this study is the use of survival analysis rather than mean-based estimates, which may be biased in populations with incomplete event ascertainment. Percentile-based definitions derived from Kaplan-Meier curves provide a more appropriate representation of population-level menarcheal timing. Using this approach, early menarche was defined as occurring before 10 years of age, corresponding to approximately below the 5th percentile of the distribution, while late menarche was defined at age 14 years or later (approximately above the 95th percentile). This framework avoids assumptions of normality and aligns with recent population-based studies in Asia that emphasize survival-based estimates over comparisons of mean age [10].

This nationally representative dataset is one of the largest and most recent assessments of menarcheal timing in Southeast Asia. Our findings are consistent with global evidence of a secular decline in pubertal timing reported across diverse regions, including high-and middle-income countries [1-4]. Within Asia, prior studies from Thailand, China, Korea, India, Malaysia, Indonesia, and Vietnam have documented substantial heterogeneity in menarcheal timing associated with socioeconomic and nutritional transitions [2,5,6,10-14]. Rather than ranking countries by mean age at menarche, our results highlight the importance of generating population-specific reference data using comparable methodologies.

Nutritional status was the factor most strongly associated with menarcheal timing in both univariable and multivariable Cox regression analyses. Overweight and obese girls experienced significantly earlier menarche, whereas underweight girls had delayed menarche, even after adjustment for region and SES. These findings are consistent with a large body of literature demonstrating associations of adiposity and energy availability with pubertal maturation [15-19]. The biological plausibility of this association has long been supported by the Frisch hypothesis regarding critical body weight and fat mass thresholds required for menarche, as well as by contemporary evidence implicating adipokines and peripheral estrogen production in the activation of the hypothalamic-pituitary-gonadal axis [18,19].

However, recent longitudinal studies suggest a bidirectional relationship between adiposity and pubertal timing, whereby earlier menarche may also contribute to higher BMI trajectories later in adolescence and adulthood [16]. While our cross-sectional design cannot establish causality or directionality, the strength and consistency of the association observed in this cohort underscore the importance of childhood nutritional status as a potentially modifiable factor associated with pubertal timing.

SES was independently associated with menarcheal timing, with girls from higher socioeconomic backgrounds experiencing earlier menarche. This finding likely reflects complex interactions between nutrition, lifestyle, healthcare access, and broader living conditions. Importantly, the relationship between SES and pubertal timing is known to vary by stage of economic development and by ethnicity. In high-income countries, earlier menarche is increasingly observed among lower socioeconomic groups, largely driven by higher childhood obesity prevalence [20-22]. Our findings suggest that Thailand may be in a transitional phase in which earlier menarche remains more common among higher socioeconomic groups, similar to patterns previously observed in China and other rapidly developing settings [5].

Regional differences in menarcheal timing were evident in Kaplan-Meier analyses; however, after multivariable adjustment, only the Southern region remained significantly associated with delayed menarche compared with Bangkok. This attenuation of regional effects after adjustment suggests that observed geographic variation is largely mediated by individual-level factors such as nutritional status and socioeconomic conditions rather than geographic region itself. Similar findings have been reported in studies from Korea and China, where urban-rural or regional gaps have narrowed or reversed over time [5,10].

Bangkok Metropolis exhibited the earliest menarcheal timing and the highest prevalence of obesity, although its prevalence of overweight was the lowest among the regions. After adjustment for nutritional status and socioeconomic factors, only the Southern region remained significantly associated with later menarche compared with Bangkok Metropolis. Other potential contributors to menarcheal timing include environmental exposures to endocrine-disrupting chemicals, dietary quality, and psychosocial stress. Exposure to chemicals such as bisphenol A, phthalates, and per- and polyfluoroalkyl substances has been associated with altered pubertal timing [23]. Proinflammatory dietary patterns [24], psychosocial stress, and early-life adversity [25,26] may also influence pubertal maturation. However, these factors were not directly assessed in our study, and their potential contributions remain hypothesis-generating.

Several limitations should be acknowledged. Age at menarche was self-reported and may be subject to recall bias despite caregiver verification. Potential age heaping at whole-year ages may also affect the precision of Kaplan-Meier estimates, particularly the median age of 12 years and the early-menarche cutoff of 10 years, which should be interpreted as approximate population-based estimates rather than precise clinical cutoffs. The cross-sectional design precludes causal inference, and detailed data on hormonal profiles, environmental exposures, and lifestyle factors were not available. Hormonal or endocrine disorders were not systematically assessed; although girls with treated precocious puberty were included, they represented a very small proportion of the cohort (0.69%), and their inclusion is unlikely to have materially influenced the population-level estimates. Nonetheless, residual confounding from unmeasured factors cannot be entirely excluded.

Despite these limitations, the large, nationally representative sample and the use of survival-based analytic methods enhance the robustness and generalizability of our findings.

Using Kaplan-Meier-based survival analysis, this study provides updated, population-level estimates of menarcheal timing among Thai girls and identifies nutritional status and socioeconomic factors as key factors associated with menarcheal timing. The percentile-based definitions of early (<10 years) and late (≥14 years) menarche offer a clinically meaningful framework for identifying atypical pubertal timing. Given the established links between early menarche and adverse metabolic, cardiovascular, and psychosocial outcomes, our findings underscore the need for integrated public health strategies that promote healthy weight trajectories, education about early puberty, clinical surveillance, and attention to social determinants of health among Thai girls.

Supplementary materials

Supplementary Tables 1–2 are available at https://doi.org/10.6065/apem.2550462.231.

Supplementary Table 1.

Distribution of the sample by grade, school type, and region

apem-2550462-231-Supplementary-Table.pdf
Supplementary Table 2.

Kaplan–Meier-estimated cumulative probability of menarche by age and geographic region (Fig. 1B)

apem-2550462-231-Supplementary-Table.pdf

Notes

Conflicts of interest

No potential conflict of interestrelevant to this article was reported.

Funding

This study was supported by the Thai Health Promotion Foundation (Grant No. 62-00-1061; Project No. 62-01620).

Data availability

The datasets generated and/or analyzed during the current study are not publicly available due to ethical restrictions and participant confidentiality requirements imposed by the institutional review board of the Faculty of Medicine, Chulalongkorn University. Deidentified data may be made available from the corresponding author upon reasonable request and with appropriate institutional approval.

Acknowledgments

Generative artificial intelligence (AI) technology was used to assist in refining the language, structure, and clarity of the manuscript, particularly in editing selected paragraphs in the Results and Discussion sections. The tool used was ChatGPT (GPT-4, model GPT-4o, developed by OpenAI), accessed via the ChatGPT web platform. This tool was employed under the authors' supervision and did not contribute to the generation or interpretation of original research data.

Author contribution

Conceptualization: PR, KS, AK, TS; Data curation: PR, KS, AK, WP, TS; Formal analysis: PR, KS, AK, PL, WP; Funding acquisition: PR, AK, TS; Methodology: PR, KS, AK, PL, WP, TS; Project administration: PR, AK, PL, TS; Visualization: PR, KS, AK, PL, WP; Writing - original draft: KS, TS; Writing - review & editing: PR, KS, AK, PL, WP, TS

References

1. Eckert-Lind C, Busch AS, Petersen JH, Biro FM, Butler G, Bräuner EV, et al. Worldwide secular trends in age at pubertal onset assessed by breast development among girls: a systematic review and meta-analysis. JAMA Pediatr 2020;174e195881.
2. Lee DH, Kim J, Kim HY. Temporal trend of age at menarche in Korean females born between 1927 and 2004: a population-based study. Front Endocrinol (Lausanne) 2024;15:1399984.
3. Leone T, Brown LJ. Timing and determinants of age at menarche in low-income and middle-income countries. BMJ Glob Health 2020;5e003689.
4. Abe SK, Nishio M, Huang HL, Leung CY, Islam MR, Rahman MS, et al. Age at menarche by birth cohort: a pooled analysis of half a million women in Asia. Public Health 2024;237:130–4.
5. Ma N, Shi D, Dang JJ, Zhong PL, Liu YF, Cai S, et al. Secular trends and urban-rural disparities in the median age at menarche among Chinese Han girls from 1985 to 2019. World J Pediatr 2023;19:1162–8.
6. Meher T, Sahoo H. Secular trend in age at menarche among Indian women. Sci Rep 2024;14:5398.
7. Sumi A, Iwase M, Nakamura U, Fujii H, Ohkuma T, Ide H, et al. Impact of age at menarche on obesity and glycemic control in Japanese patients with type 2 diabetes: Fukuoka Diabetes Registry. J Diabetes Investig 2018;9:1216–23.
8. Sun Y, Liu H, Mu C, Liu P, Hao C, Xin Y. Early puberty: a review on its role as a risk factor for metabolic and mental disorders. Front Pediatr 2024;12:1326864.
9. Collaborative Group on Hormonal Factors in Breast Cancer. Menarche, menopause, and breast cancer risk: individual participant meta-analysis, including 118 964 women with breast cancer from 117 epidemiological studies. Lancet Oncol 2012;13:1141–51.
10. Lee MH, Kim SH, Oh M, Lee KW, Park MJ. Age at menarche in Korean adolescents: trends and influencing factors. Reprod Health 2016;13:121.
11. Noipayak P, Rawdaree P, Supawattanabodee B, Manusirivitthaya S. Age at menarche and performance intelligence quotients of adolescents in Bangkok, Thailand: a cross-sectional study. BMC Pediatr 2016;16:87.
12. Kamarulzaman S, Mohamed PN, Ridzuan PM. Age at menarche and menstrual pattern among adolescences girls in Selangor. J Nat Ayurvedic Med 2019;3:000175.
13. Moelyo AG, Wulandari A, Imas O, Rahma UP, Hidayah N, Kesumaningtyas C, et al. Age at menarche and early menarche among healthy adolescents. Paediatr Indones 2019;59:33–7.
14. Nguyen TD, Nguyen TTH, Nguyen PV. A research study on sociodemographic determinants of different menarche in the female students in Southern Vietnam. Syst Rev Pharm 2020;11:333–6.
15. Joakimsen HL, Brendlien A, Furberg AS, Nielsen CS, Grimnes G, Evensen EK. Age at menarche and its association with preschool BMI among girls in Northern Norway. PLoS One 2025;20e0322986.
16. Gaml-Sørensen A, Brix N, Ernst A, Lunddorf LLH, Arah OA, Strandberg-Larsen K, et al. Pubertal timing and tempo and body mass index trajectories: investigating the confounding role of childhood body mass index. Am J Epidemiol 2025;194:2270–80.
17. Aghaee S, Deardorff J, Quesenberry CP, Greenspan LC, Kushi LH, Kubo A. Associations between childhood obesity and pubertal timing stratified by sex and race/ethnicity. Am J Epidemiol 2022;191:2026–36.
18. Frisch RE, Revelle R. Height and weight at menarche and a hypothesis of critical body weights and adolescent events. Science 1970;169:397–9.
19. Reinehr T, Roth CL. Is there a causal relationship between obesity and puberty? Lancet Child Adolesc Health 2019;3:44–54.
20. Krieger N, Kiang MV, Kosheleva A, Waterman PD, Chen JT, Beckfield J. Age at menarche: 50-year socioeconomic trends among US-born black and white women. Am J Public Health 2015;105:388–97.
21. James-Todd T, Tehranifar P, Rich-Edwards J, Titievsky L, Terry MB. The impact of socioeconomic status across early life on age at menarche among a racially diverse population of girls. Ann Epidemiol 2010;20:836–42.
22. Hiatt RA, Stewart SL, Deardorff J, Danial E, Abdiwahab E, Pinney SM, et al. Childhood socioeconomic status and menarche: a prospective study. J Adolesc Health 2021;69:33–40.
23. Demir A, Aydin A, Büyükgebiz A. Thematic review of endocrine disruptors and their role in shaping pubertal timing. Children (Basel) 2025;12:93.
24. Davis CP, Fest S, Cushing-Haugen K, Kensler TW, Chavarro JE, Harris HR. Dietary patterns and age at menarche in a prospective study of girls in the USA. Hum Reprod 2025;40:1087–93.
25. Liu S, Zhou Y, Wu D, Tian Y, Luo J, Zhang J, et al. Psychosocial stress trajectories are associated with puberty timing: findings from a 7-year cohort in Chongqing, China. J Clin Endocrinol Metab 2025;110:e945–52.
26. Gong C, Duan X, Su P, Wan Y, Xu Y, Tao F, et al. Heightened HPA-axis stress reactivity and accelerated pubertal progression predicts depressive symptoms over 4-year follow up. Psychoneuroendocrinology 2019;103:259–65.

Article information Continued

Fig. 1.

Kaplan-Meier estimates of the cumulative probability of menarche by age. (A) Overall cumulative probability of menarche among Thai schoolgirls. (B) Cumulative probability of menarche stratified by geographic region (Bangkok Metropolis, Central, North, North-East, and South). Differences across regions were assessed using the log-rank test (P<0.001).

Table 1.

Demographic data

Characteristic All (N=8,161) Bangkok (N=620) Central (N=2,203) North (N=1,380) North-East (N=2,600) South (N=1,358)
Age at survey (yr) 11.93 11.95 12.00 11.99 11.74 12.14
Attained menarche (%) 52.6 59.0 54.6 52.2 54.2 43.7
Age at menarche (yr)
 Mean* 11.57 11.48 11.54 11.60 11.63 11.54
 Median* 11.63 11.48 11.54 11.71 11.71 11.54
BMI z-score 0.28 0.45 0.31 0.26 0.25 0.23
Nutritional status
 Underweight (%) 4.5 3.1 5.5 3.8 4.0 5.1
 Normal weight (%) 64.6 64.0 61.8 67.5 66.0 63.9
 Overweight (%) 17.7 16.6 18.3 17.6 17.7 17.4
 Obese (%) 13.2 16.3 14.5 11.2 12.3 13.6
Area
 Urban (%) 63.5 100 60.2 49.4 56.9 79.2
 Rural (%) 36.5 0 39.8 50.6 43.1 20.8
Socioeconomic status†
 Low (%) 33.3 10.6 38.6 48.5 42.1 26.7
 Moderate (%) 33.3 30.4 31.6 30.0 36.9 37.6
 High (%) 33.3 59.0 29.8 21.5 21.0 35.6

BMI, body mass index.

*

Mean and median ages at menarche were calculated only among participants who had attained menarche at the time of the survey. These descriptive values differ from the Kaplan-Meier estimates, which account for right-censoring among participants who had not yet attained menarche.

†

Socioeconomic status (SES) was determined using a composite score derived from 3 parental characteristics: income, educational attainment, and occupation. Factor analysis was used to derive the composite SES score, which was scaled from 0.0 to 1.0. Participants were classified into 3 SES categories using fixed score intervals: low SES (0–0.3333), moderate SES (>0.3333–0.6667), and high SES (>0.6667–1.0000). These categories were defined using fixed cutoffs rather than sample-based tertiles.

Table 2.

Probability of menarche at each age (in years and months) derived from Kaplan-Meier estimation

Age Probability of menarche Age Probability of menarche Age Probability of menarche Age Probability of menarche
7 yr 7 mo 0.0001 10 yr 1 mo 0.0618 12 yr 7 mo 0.7480 15 yr 1 mo 0.9928
7 yr 8 mo 0.0001 10 yr 2 mo 0.0701 12 yr 8 mo 0.7628 15 yr 2 mo 0.9949
7 yr 9 mo 0.0001 10 yr 3 mo 0.0773 12 yr 9 mo 0.7801 15 yr 3 mo 0.9949
7 yr 10 mo 0.0002 10 yr 4 mo 0.0853 12 yr 10 mo 0.7929 15 yr 4 mo 0.9949
7 yr 11 mo 0.0002 10 yr 5 mo 0.0963 12 yr 11 mo 0.8099 15 yr 5 mo 0.9962
8 yr 0.0006 10 yr 6 mo 0.1100 13 yr 0.8521 15 yr 6 mo 0.9962
8 yr 1 mo 0.0006 10 yr 7 mo 0.1207 13 yr 1 mo 0.8663 15 yr 7 mo 0.9962
8 yr 2 mo 0.0006 10 yr 8 mo 0.1290 13 yr 2 mo 0.8782 15 yr 8 mo 0.9962
8 yr 3 mo 0.0007 10 yr 9 mo 0.1409 13 yr 3 mo 0.8897 15 yr 9 mo 0.9962
8 yr 4 mo 0.0011 10 yr 10 mo 0.1523 13 yr 4 mo 0.9001 15 yr 10 mo 0.9962
8 yr 5 mo 0.0015 10 yr 11 mo 0.1651 13 yr 5 mo 0.9113 15 yr 11 mo 0.9962
8 yr 6 mo 0.0018 11 yr 0.2332 13 yr 6 mo 0.9189 16 yr 0.9962
8 yr 7 mo 0.0020 11 yr 1 mo 0.2494 13 yr 7 mo 0.9266 16 yr 1 mo 0.9965
8 yr 8 mo 0.0025 11 yr 2 mo 0.2729 13 yr 8 mo 0.9299 16 yr 2 mo 0.9969
8 yr 9 mo 0.0032 11 yr 3 mo 0.2909 13 yr 9 mo 0.9372 16 yr 3 mo 0.9972
8 yr 10 mo 0.0036 11 yr 4 mo 0.3087 13 yr 10 mo 0.9423 16 yr 4 mo 0.9976
8 yr 11 mo 0.0042 11 yr 5 mo 0.3334 13 yr 11 mo 0.9475 16 yr 5 mo 0.9979
9 yr 0.0071 11 yr 6 mo 0.3567 14 yr 0.9610 16 yr 6 mo 0.9983
9 yr 1 mo 0.0083 11 yr 7 mo 0.3730 14 yr 1 mo 0.9664 16 yr 7 mo 0.9986
9 yr 2 mo 0.0088 11 yr 8 mo 0.3932 14 yr 2 mo 0.9706 16 yr 8 mo 0.9990
9 yr 3 mo 0.0106 11 yr 9 mo 0.4147 14 yr 3 mo 0.9760 16 yr 9 mo 0.9993
9 yr 4 mo 0.0114 11 yr 10 mo 0.4363 14 yr 4 mo 0.9781 16 yr 10 mo 0.9997
9 yr 5 mo 0.0137 11 yr 11 mo 0.4599 14 yr 5 mo 0.9810 16 yr 11 mo 1.0000
9 yr 6 mo 0.0156 12 yr 0.5583 14 yr 6 mo 0.9841 17 yr 1.0000
9 yr 7 mo 0.0174 12 yr 1 mo 0.5901 14 yr 7 mo 0.9854
9 yr 8 mo 0.0197 12 yr 2 mo 0.6206 14 yr 8 mo 0.9874
9 yr 9 mo 0.0226 12 yr 3 mo 0.6452 14 yr 9 mo 0.9887
9 yr 10 mo 0.0270 12 yr 4 mo 0.6697 14 yr 10 mo 0.9914
9 yr 11 mo 0.0320 12 yr 5 mo 0.7007 14 yr 11 mo 0.9919
10 yr 0.0556 12 yr 6 mo 0.7276 15 yr 0.9925

Table 3.

Univariable Cox proportional hazards regression for menarcheal timing

Variable HR (95% CI) P-value
Area
 Urban 1.000
 Rural 0.925 (0.869–0.984) 0.013
Region
 Bangkok 1.000
 Central 0.873 (0.777–0.982) 0.023
 North 0.799 (0.705–0.907) <0.001
 North-East 0.792 (0.706–0.889) <0.001
 South 0.660 (0.579–0.752) <0.001
Nutritional status
 Underweight 0.413 (0.339–0.503) <0.001
 Normal weight 1.000
 Overweight 1.689 (1.562–1.825) <0.001
 Obese 1.679 (1.530–1.842) <0.001
Socioeconomic status
 Low 1.000
 Moderate 1.145 (1.055–1.243) 0.001
 High 1.335 (1.231–1.449) <0.001

Event: attainment of menarche. Time scale: age (mo).

HR, hazard ratio; CI, confidence interval.

HR>1 indicates earlier menarche; HR<1 indicates later menarche.

Table 4.

Multivariable Cox proportional hazards regression for menarcheal timing

Variable HR (95% CI) P-value
Area
 Urban 1.000
 Rural 1.005 (0.933–1.083) 0.886
Region
 Bangkok 1.000
 Central 0.991 (0.871–1.128) 0.893
 North 0.899 (0.781–1.034) 0.136
 North-East 0.942 (0.826–1.074) 0.374
 South 0.739 (0.639–0.855) <0.001
Nutritional status
 Underweight 0.446 (0.359–0.555) <0.001
 Normal weight 1.000
 Overweight 1.668 (1.529–1.819) <0.001
 Obese 1.643 (1.484–1.818) <0.001
Socioeconomic status
 Low 1.000
 Moderate 1.116 (1.027–1.214) 0.010
 High 1.282 (1.173–1.400) <0.001

Event: attainment of menarche. Time scale: age (mo).

HR, hazard ratio; CI, confidence interval.

HR>1 indicates earlier menarche; HR<1 indicates later menarche.