Association between childhood weight status and timing of puberty: a systematic review and meta-analysis Authors: Ting Zhou, Feng Xiong Published: 2026 ABSTRACT AbstractBackgroundChildhood overweight and obesity have increased worldwide alongside a trend toward earlier pubertal onset, particularly in girls. Excess adiposity may affect pubertal timing through metabolic and endocrine pathways, but observational findings remain heterogeneous. This study systematically evaluated childhood weight status and pubertal timing, mapped the evidence, quantified early-puberty risk associated with overweight/obesity, and explored dose-response relationships and sex differences.MethodsWe systematically searched PubMed/MEDLINE, Embase, Web of Science, China National Knowledge Infrastructure (CNKI), and Wanfang databases from inception to January 2025. Observational studies using childhood weight/body mass index (BMI) as exposure variables and pubertal development indicators as outcomes were included. Two researchers independently screened literature, extracted data, and assessed study quality. The Newcastle-Ottawa Scale was used to evaluate risk of bias. Effect sizes were pooled using the DerSimonian-Laird random-effects model, with heterogeneity assessed via the I2 statistic. Subgroup analyses were conducted by weight category and outcome type. An evidence map was constructed to describe the distribution and direction of evidence across different exposure and outcome types.ResultsForty-one studies were included in the systematic review, involving a total sample size of 416,789 participants. Studies were published between 2000 and 2025 and originated from 18 countries/regions. The evidence map indicated that quantifiable evidence primarily focused on BMI/weight classification and weight trajectory in relation to outcomes such as “onset/precipitation of puberty (including precocious puberty risk)” and “menarche/spermatogenesis onset time”. Studies examining outcomes such as pubertal progression/duration and bone age were relatively scarce. From studies with consistent exposure grouping and extractable odds ratio (OR)/hazard ratio (HR)/relative risk (RR) and 95% confidence interval (CI), 24 effect size data points (from 3 studies) were extracted for meta-analysis. The random-effects model revealed that compared with normal-weight children, overweight/obese children had a significantly increased risk of early puberty onset (pooled OR/HR =1.56, 95% CI: 1.44–1.69, P<0.001; I2=96.0%). Subgroup analysis demonstrated a clear dose-response relationship: Overweight (OR/HR =1.46, 95% CI: 1.31–1.63), obesity (OR/HR =1.55, 95% CI: 1.36–1.77), and severe obesity (OR/HR =1.69, 95% CI: 1.37–2.07). Analysis by outcome type revealed the greatest impact on menarche timing (OR/HR =2.24, 95% CI: 1.82–2.75), followed by Tanner stage (OR/HR =2.06), breast development (OR/HR =1.47), pubic hair development (OR/HR =1.47), and testicular development (OR/HR =1.22).ConclusionsChildhood overweight/obesity is significantly associated with increased risk of early puberty onset, exhibiting a dose-response relationship. Female pubertal markers, particularly menarche, exhibited greater sensitivity to weight status. These findings underscore the importance of childhood weight management in promoting normal pubertal development. Future studies should adopt standardized definitions of exposure and outcomes to enhance the comparability of evidence.Keywords: Obesity, overweight, pubertal development, menarche, Tanner stage Highlight box Key findings • Childhood overweight and obesity are linked to earlier puberty, especially in girls, with a clear dose-response relationship. • Effect sizes differ by pubertal marker, and studies show high heterogeneity due to body mass index measures, outcome definitions, and confounder adjustment. What is known and what is new? • Global childhood obesity and earlier puberty have risen in parallel, supported by biological mechanisms involving leptin and estrogen, with observational studies confirming positive associations. • This study provides the first comprehensive meta-analysis quantifying effects across weight categories and pubertal markers, includes an evidence map identifying research gaps, and systematically explores heterogeneity to clarify the true effect size. What is the implication, and what should change now? • Early puberty raises long-term health risks, so preventing childhood obesity and monitoring pubertal development can help delay puberty and reduce related harms. • Public health efforts should enhance early obesity prevention, future research needs standardized prospective design with more males, and pediatric guidelines should include puberty timing as a routine indicator. Introduction Background Over the past four decades, the prevalence of overweight and obesity among children has risen sharply worldwide. According to the World Health Organization, approximately 390 million children and adolescents aged 5–19 years were overweight or obese in 2022, accounting for over 18% of this age group—a marked increase from the 4% recorded in 1975 (1,2). In China, the combined prevalence of overweight and obesity among school-aged children surged from 1.1% in 1985 to approximately 24% in 2019, making it a pressing public health concern (3). In parallel, multiple epidemiological studies have documented a secular trend toward earlier pubertal onset, most notably a decline in the age at menarche among girls (4,5). The temporal co-occurrence of these two trends has spurred interest in their potential causal relationship. Rationale and knowledge gap Puberty is a critical period of growth and development, whose initiation and tempo are governed by a complex interplay of genetic, nutritional, endocrine, and environmental factors (6). Altered pubertal timing may carry substantial health consequences. Early puberty has been linked to adult-onset obesity, type 2 diabetes mellitus, cardiovascular disease, and metabolic syndrome (7,8), as well as to an elevated risk of hormone-sensitive cancers, including breast and endometrial cancer (9,10). In addition, early-maturing adolescents are at increased risk for depression, anxiety, behavioral difficulties, and lower academic attainment (11,12). Understanding the modifiable determinants of pubertal timing is therefore of considerable public health importance. Adipose tissue functions not only as an energy reservoir but also as an active endocrine organ that secretes adipokines and sex steroids involved in pubertal regulation (13). Leptin, produced by adipocytes, is regarded as a key permissive signal for pubertal initiation, acting on hypothalamic gonadotropin-releasing hormone (GnRH) neurons to modulate reproductive axis activation (14,15). Furthermore, aromatase within adipose tissue converts androgens to estrogens; elevated peripheral estrogen levels in obese children may accelerate breast development and advance menarche in girls (16). Obesity-associated insulin resistance and hyperinsulinemia may further influence pubertal timing by reducing sex hormone-binding globulin (SHBG) concentrations and increasing bioavailable sex steroids (17). A growing number of observational studies have examined the relationship between childhood weight status and pubertal development, yet their findings remain heterogeneous. This variability likely reflects differences in study design, exposure and outcome definitions, population characteristics, and the extent of confounder adjustment (18). Some studies relied on a single BMI measurement at one time point, while others employed trajectory modeling to capture longitudinal weight changes (19,20). Although several reviews and meta-analyses have recently addressed obesity and pubertal timing, most were restricted to boys (21-23) or were narrative in nature (24,25), and none combined an evidence-mapping framework with a quantitative synthesis that simultaneously spans multiple weight categories (overweight, obesity, and severe obesity) and the full spectrum of male and female pubertal markers. To date, a comprehensive synthesis and quantitative evaluation of this evidence base is lacking. Objective Accordingly, the present study aimed to: (I) identify and critically appraise all observational studies on the association between childhood weight status and pubertal development; (II) construct an evidence map depicting the distribution and direction of findings across different exposure and outcome types; (III) quantitatively estimate the effect of overweight/obesity on early puberty risk through meta-analysis; (IV) examine dose-response relationships by weight category (overweight, obese, severely obese); (V) compare effect sizes across pubertal markers (menarche, breast development, pubic hair development, testicular enlargement, Tanner stage); and (VI) explore sources of heterogeneity and assess the robustness of findings through sensitivity analyses. We present this article in accordance with the PRISMA reporting checklist (available at https://tp.amegroups.com/article/view/10.21037/tp-2026-0362/rc) (26). Methods The protocol was prospectively registered in PROSPERO (Registration number: CRD420261366929). Inclusion and exclusion criteria Eligibility criteria were defined using the PECO framework. Population: (I) general pediatric or adolescent populations; (II) age range spanning childhood (at the time of weight/exposure measurement) to adolescence (at outcome assessment); (III) populations with conditions known to affect growth and development (e.g., congenital adrenal hyperplasia, precocious puberty syndromes, Turner syndrome, Prader-Willi syndrome, and other endocrine or genetic disorders) were excluded. Exposure: (I) weight or BMI as the primary exposure variable; (II) weight/BMI categorized (e.g., normal, overweight, obese) or analyzed as a continuous variable; (III) single-time-point measurement or multi-time-point trajectory analysis; (IV) objectively measured (stadiometer and calibrated scale) or self-reported; (V) overweight/obesity defined according to internationally recognized standards [World Health Organization (WHO), International Obesity Task Force (IOTF), Centers for Disease Control and Prevention (CDC)] or local population-specific reference values. Comparison: normal-weight children served as the reference group. Outcome: (I) pubertal development indicators, including but not limited to age at menarche, Tanner staging (breast, pubic hair, genital development), testicular volume, bone age, and age at peak height velocity (APHV); (II) outcomes ascertained via clinical examination, self-report questionnaire, or biomarkers; (III) outcomes defined with sufficient clarity to permit extraction of quantitative effect sizes. Study design: (I) prospective or retrospective cohort studies; (II) birth cohort studies; (III) cross-sectional studies (included in the systematic review only, not the meta-analysis); (IV) case-control studies; (V) no language restriction. Exclusion criteria: (I) studies limited to adult populations; (II) case reports, narrative reviews, commentaries, or conference abstracts; (III) studies lacking full text or providing insufficient data for effect-size calculation; (IV) duplicate publications (the report with the largest sample or longest follow-up was retained); (V) control-group data from intervention trials (pharmacological or surgical weight-loss studies). Literature search strategy The following electronic databases were searched from inception through January 15, 2025: PubMed/MEDLINE, Embase, Web of Science Core Collection, CNKI, and Wanfang Database. No restriction on publication date was imposed; searching from database inception was intended to capture the full body of evidence and to reflect the long-term secular trends in both childhood adiposity and pubertal timing that motivate this review, in keeping with PRISMA guidance for comprehensive evidence synthesis. Studies published before 2016 were therefore retained, and the robustness of the pooled estimate was further examined through the pre-specified sensitivity and subgroup analyses described below. The English-language search strategy combined Medical Subject Headings (MeSH) with free-text terms. Taking PubMed as an example: ((“Body Mass Index”(MeSH) OR “Overweight”(MeSH) OR “Obesity”(MeSH) OR “Pediatric Obesity”(MeSH) OR BMI(tiab) OR “body mass index”(tiab) OR overweight(tiab) OR obesity(tiab) OR obese(tiab) OR “weight status”(tiab) OR “weight trajectory”(tiab) OR “BMI trajectory”(tiab) OR adiposity(tiab)) AND (“Puberty”(MeSH) OR “Sexual Maturation”(MeSH) OR “Menarche”(MeSH) OR puberty(tiab) OR pubertal(tiab) OR menarche(tiab) OR “Tanner stage”(tiab) OR “breast development”(tiab) OR “testicular volume”(tiab) OR “sexual maturation”(tiab) OR “pubertal timing”(tiab) OR “age at puberty”(tiab)) AND (“Child”(MeSH) OR “Adolescent”(MeSH) OR child*(tiab) OR adolescen*(tiab) OR pediatric(tiab) OR paediatric(tiab))). Equivalent search terms were used for the Chinese-language database searches. Supplementary search strategies included: (I) manual screening of reference lists from included studies; (II) backward and forward citation tracking of relevant systematic reviews and meta-analyses; (III) citation tracing via Google Scholar; and (IV) correspondence with domain experts to identify unpublished studies or grey literature. Study selection and data extraction Study selection proceeded in two stages. In the first stage, two investigators independently screened titles and abstracts to identify potentially eligible records. In the second stage, full texts of candidate studies were retrieved and assessed against the eligibility criteria. Disagreements were resolved by consensus, with a third investigator adjudicating unresolved cases. Reference management and deduplication were performed using EndNote X9. Data were extracted onto a standardized form encompassing the following items: (I) study characteristics (first author, publication year, country/region, journal, study design, data source/cohort name, study period, follow-up duration); (II) population characteristics (total sample size, analytic sample size, sex distribution, baseline age, age at outcome assessment, eligibility criteria); (III) exposure variables (weight/BMI measurement method, measurement timing and frequency, overweight/obesity definition criteria, trajectory modeling approach where applicable, category definitions and group sizes); (IV) outcome variables (type of pubertal indicator, measurement method, threshold for precocious puberty); (V) effect estimates [type—odds ratio (OR), hazard ratio (HR), relative risk (RR), mean difference (MD), β—point estimate and 95% confidence interval (CI), P value, confounders adjusted for, statistical model]; and (VI) quality-related information (loss to follow-up rate, management of selection bias, adequacy of confounder adjustment). Quality assessment Methodological quality of included studies was evaluated using the Newcastle-Ottawa Scale (NOS) for observational studies (27). The NOS appraises three domains: selection of study groups (four items, maximum four stars), comparability of groups (one item, maximum two stars), and ascertainment of outcomes (three items, maximum three stars), yielding a total score of zero to nine stars. Studies scoring ≥7 were classified as high quality, 4–6 as moderate quality, and <4 as low quality. Two investigators independently assessed each study. For the selection domain, the representativeness of the exposed cohort, selection of the comparison group, ascertainment of exposure, and demonstration that the outcome of interest was not present at baseline were evaluated. For comparability, studies received one star for controlling age and sex, and an additional star for adjusting for at least one further confounder (e.g., socioeconomic status, ethnicity, or birth weight). For the outcome domain, method of outcome assessment (clinical examination vs. self-report), length and adequacy of follow-up, and completeness of follow-up were appraised. Discrepancies in scoring were resolved through discussion or, when necessary, by consultation with a third reviewer. Quality scores were used descriptively and in sensitivity analyses; they did not serve as a basis for excluding studies from the review. Statistical analysis Effect-size extraction and conversion From each eligible study, effect estimates and their 95% CIs were extracted. For dichotomous outcomes (e.g., precocious vs. non-precocious puberty), OR, HR, or RR were extracted; for continuous outcomes (e.g., age at menarche), MD or standardized mean difference (SMD) were recorded. The present meta-analysis focused on pooling ratio-type measures (OR, HR, RR). Eligible effect estimates required exposure groups categorized as “overweight vs. normal”, “obese vs. normal”, or “severely obese vs. normal”, with complete 95% CI reporting. Effect measures were log-transformed prior to pooling: ln(OR) or ln(HR). Standard errors were derived from 95% CIs: SE = [ln(CI upper limit) – ln(CI lower limit)]/(2×1.96). Meta-analytic approach Given anticipated clinical and methodological heterogeneity, the DerSimonian-Laird random-effects model was used to pool effect sizes (28). This approach incorporates between-study variance (τ2) into the weighting: Wi =1/(SEi2 + τ2). The pooled effect estimate and its 95% CI were computed as a weighted average across studies. Heterogeneity assessment Between-study heterogeneity was evaluated using Cochran’s Q test and the I2 statistic. A Q-test P<0.10 was taken to indicate significant heterogeneity. The I2 statistic was interpreted as follows: <25%, low; 25–75%, moderate; >75%, high (29). The between-study variance τ2 was also reported. Subgroup analyses Pre-specified subgroup analyses were performed to explore potential sources of heterogeneity: (I) by weight category (overweight, obese, or severely obese vs. normal weight); (II) by outcome type (menarche, breast development, pubic hair development, testicular enlargement, Tanner stage); (III) by sex (girls, boys, mixed); (IV) by study design (prospective, retrospective, or birth cohort); and (V) by geographic region (North America, Europe, Asia, other). Sensitivity analysis To evaluate the robustness of the pooled estimates, the following sensitivity analyses were conducted: (I) leave-one-out analysis, sequentially removing each study and recalculating the pooled effect; (II) restriction to high-quality studies (NOS ≥7); (III) restriction to prospective designs; and (IV) restriction to studies adjusting for three or more confounders. Publication bias When ten or more studies contribute to a pooled estimate, publication bias is conventionally assessed by visual inspection of funnel plots and Egger’s regression test (30). Because only three studies provided effect estimates eligible for the present meta-analysis, formal publication-bias assessment was not performed. Results Literature screening process Under the search timeframe from 2000 to present, the initial search yielded 33,152 articles (PubMed: 7,509; Embase: 15,424; Web of Science: 8,639; CNKI: 773; Wanfang: 807). After deduplication via EndNote, 13,500 records proceeded to title and abstract screening. Of these, 13,278 records were excluded because they did not meet the inclusion criteria. Full-text assessments were conducted on 222 potentially relevant reports, of which 181 were excluded. Ultimately, 41 studies were included in the systematic review, and 24 effect-size estimates from 3 studies were included in the quantitative meta-analysis. Primary reasons for exclusion included: lack of relevant outcome data (n=84), inadequate exposure reporting (n=39), insufficient data for effect size calculation (n=41), overlapping populations (n=4), specialized disease populations (n=12), and unavailability of full text (n=1). Here, “overlapping populations” denotes reports that analyzed the same underlying cohort or data source as another, more complete study that was already included—for example, earlier, smaller, or partial analyses of the same birth-cohort or multi-cohort consortium data. To avoid double-counting participants, only the report with the largest analytic sample or longest follow-up from each source population was retained, in accordance with our duplicate-publication rule (Section Inclusion and exclusion criteria). Ultimately, 41 studies were included in the systematic review (qualitative analysis). A total of 279 effect size records were extracted from these studies, including 114 with OR/HR/RR data accompanied by complete 95% CI. After screening for exposure groups categorized as “overweight vs. normal weight”, “obese vs. normal weight”, or “severely obese vs. normal weight”, 24 effect size data points (from 3 studies) were finally included in the meta-analysis (quantitative synthesis). The literature screening process is detailed in the PRISMA flow diagram (Figure 1). PRISMA flow diagram of the study selection process. A total of 33,152 records were identified from seven electronic databases (PubMed, Embase, Web of Science, CNKI, and Wanfang). After removal of duplicates, approximately 13,500 records were screened by title and abstract; 222 full-text articles were assessed for eligibility. Forty-one studies met the inclusion criteria for the systematic review. Twenty-four effect-size estimates from three studies met the criteria for quantitative synthesis (meta-analysis), requiring consistent exposure grouping (overweight, obese, or severely obese vs. normal weight) and extractable ratio-type effect measures (OR/HR/RR) with 95% confidence intervals. Supplementary searches, including manual screening of reference lists, backward and forward citation tracking, Google Scholar citation tracing, and correspondence with domain experts, were conducted but yielded no additional eligible studies (n=0). CNKI, China National Knowledge Infrastructure; HR, hazard ratio; NA, not applicable; OR, odds ratio; RR, relative risk. Basic characteristics of included studies The 41 studies included in the systematic review were published between 2000 and 2025 and originated from 10 countries/regions: China (14 studies, including 1 from Taiwan), the United States (9 studies), Sweden (5 studies), Australia (4 studies), Denmark (3 studies), Spain (2 studies), Finland (2 studies), Romania (1 study), and Zimbabwe/Malawi (1 study).Nine of the 41 studies (22.0%) were conducted in the United States. Because the single largest cohort (Aghaee et al., n=133,208) was US-based, participants from studies classified as US-based in Table S1 accounted for approximately 55.7% of the summed sample among studies with a reported numerical sample size. Study designs included: prospective cohort/follow-up studies [9], longitudinal cohort/follow-up studies [8], retrospective cohort studies [5], birth cohort studies/analyses [9], pooled or multi-cohort studies [3], secondary analyses [3], randomized/intervention studies [3], and studies with the design not reported [1]. Among the 40 studies with a numerical sample size reported in Table S1, the summed sample size was 296,075 individuals, with individual study sample sizes ranging from 71 to 133,208 participants (median 1,940; interquartile range, 524–7,495). Study populations included female-specific, male-specific, and mixed-sex populations; however, a reliable study-level classification into all-female. Regarding exposure variable definitions: included single-time-point BMI or weight-status measures, continuous anthropometric indices, growth-related measures, and longitudinal or trajectory-based indicators, but it did not provide a study-level classification that allowed the numbers of single-time-point BMI and multi-time-point trajectory studies to be reliably recalculated. Outcome variables comprised: menarche-related outcomes (8 studies), Tanner-stage outcomes (6 studies), breast development (6 studies), pubic hair development (6 studies), testicular development/volume (5 studies), bone age (1 study), and peak height velocity–related outcomes, including PHV, APHV, or YAPHV measures (7 studies). Some studies reported multiple outcome measures. Detailed characteristics are presented in Table S1. Evidence map (exposure type × outcome type) To systematically present the evidence distribution for the “weight/body fat exposure—pubertal development outcome” focus in the research protocol, we categorized all effect sizes extracted from the included studies, which are summarized in Table S1. Evidence maps were constructed based on exposure type (BMI/weight classification, trajectory/growth pattern, continuous BMI/body fat indicators, other body fat indicator classifications) and outcome type (risk of early puberty/precocious puberty, pubertal onset indicators, menarche/ and age at first ejaculation, growth spurt/APHV, pubertal progression/duration, bone age, and other outcomes) (Table 2). Among studies reporting calculable effect sizes with 95% CIs, 15 studies provided quantitative evidence for at least one outcome category. In Table 2, “Early” indicates that higher weight/body fat is associated with earlier pubertal development or increased risk of precocious puberty; ‘Late’ indicates later development or reduced risk; “No/Mixed” indicates no statistically significant association or inconsistent results within the same study. Table 2Distribution of studies and direction of evidence by exposure type × outcome typeExposure typeRisk of early/precocious pubertyPubertal onset indicatorsMenarche/Spermarche timingGrowth spurt/APHVPubertal progression/durationBone age and other outcomesBMI/weight categories1 (earlier 0, later 0, null/mixed 1)4 (earlier 3, later 0, null/mixed 1)2 (earlier 2, later 0, null/mixed 0)1 (earlier 0, later 0, null/mixed 1)00Trajectories/growth patterns2 (earlier 2, later 0, null/mixed 0)3 (earlier 2, later 1, null/mixed 0)1 (earlier 1, later 0, null/mixed 0)01 (earlier 1, later 0, null/mixed 0)0Continuous BMI/adiposity indicators2 (earlier 1, later 0, null/mixed 1)2 (earlier 2, later 0, null/mixed 0)1 (earlier 1, later 0, null/mixed 0)000Other categorized adiposity indicators1 (earlier 1, later 0, null/mixed 0)00001 (earlier 0, later 1, null/mixed 0)Each cell shows the number of studies that reported quantifiable effect estimates (with 95% confidence intervals) for the corresponding “exposure type × outcome type” combination. Values in parentheses indicate the direction of evidence: earlier = earlier timing/increased risk; later = later timing/decreased risk; null/mixed = non-significant or inconsistent results. APHV, age at peak height velocity; BMI, body mass index. Distribution of studies and direction of evidence by exposure type × outcome type Each cell shows the number of studies that reported quantifiable effect estimates (with 95% confidence intervals) for the corresponding “exposure type × outcome type” combination. Values in parentheses indicate the direction of evidence: earlier = earlier timing/increased risk; later = later timing/decreased risk; null/mixed = non-significant or inconsistent results. APHV, age at peak height velocity; BMI, body mass index. Study quality assessment Based on the Newcastle-Ottawa Scale, the overall quality of included studies ranged from moderate to high. Twenty-eight studies (68.3%) were high quality (NOS ≥7 points), twelve (29.3%) were moderate quality (4–6 points), and one (2.4%) was low quality (<4 points). Regarding selection bias, most studies (35, 85.4%) had well-represented exposure cohorts, and 39 studies (95.1%) used objective methods (height and weight measurements) to determine exposure. Regarding comparability, 33 studies (80.5%) adjusted for age and sex, while 24 studies (58.5%) additionally adjusted for important confounders such as socioeconomic status, ethnicity, or birth weight. Regarding outcome assessment, 28 studies (68.3%) used clinical evaluations or medical records to determine outcomes, and 31 studies (75.6%) demonstrated good follow-up completeness (loss to follow-up <20%). Key quality concerns included: inadequate adjustment for potential confounders in some studies (particularly maternal age at menarche and birth weight), reliance on self-reported methods for assessing pubertal development outcomes, and high rates of loss to follow-up. Detailed quality assessment results for the studies are presented in Table S1 (31-50). The corresponding results for the remaining included studies are also presented in Table S1 (51-71). Meta-analysis results Overall effect The characteristics and effect estimates of the three studies contributing to the quantitative synthesis are summarized in Table 1. Pooling 24 effect-size estimates from these three studies, a random-effects meta-analysis revealed that compared with normal-weight children, overweight/obese children had a significantly increased risk of early puberty (pooled OR/HR =1.56, 95% CI: 1.44–1.69, P<0.001). High statistical heterogeneity existed among studies (I2 =96.0%, Q=580.11, df =23, P<0.001; τ2=0.043). The forest plot is shown in Figure 2. Table 1Characteristics of included studies and effect sizesAuthorYearCountryDesignSampleOutcome (study)Exposure (MA)Outcome (MA)SexTypeEffect95% CIP valueAghaee S, et al. (31)2022United States (Northern California)Prospective longitudinal cohort study133,208Testicular enlargementSevere obesity vs. normalTesticular enlargementMixed/unknownHR1.231.15–1.320.98 (earlier by ~3 months)Testicular enlargementObesity vs. normalTesticular enlargementMixed/unknownHR1.221.17–1.270.98Testicular enlargementOverweight vs. normalTesticular enlargementMixed/unknownHR1.151.11–1.190.98Testicular enlargementUnderweight vs. normalTesticular enlargementMixed/unknownHR0.830.77–0.881.02 (later)Pubic hair developmentSevere obesity vs. normalPubic hair developmentMixed/unknownHR1.441.34–1.550.96Pubic hair developmentObesity vs. normalPubic hair developmentMixed/unknownHR1.341.29–1.40.97Pubic hair developmentOverweight vs. normalPubic hair developmentMixed/unknownHR1.241.19–1.280.98Pubic hair developmentUnderweight vs. normalPubic hair developmentMixed/unknownHR0.850.79–0.91.02Breast developmentSevere obesity vs. normalBreast developmentFemaleHR1.631.5–1.770.94Breast developmentObesity vs. normalBreast developmentFemaleHR1.481.41–1.540.96Breast developmentOverweight vs. normalBreast developmentFemaleHR1.351.3–1.390.97Breast developmentUnderweight vs. normalBreast developmentFemaleHR0.790.74–0.841.03Pubic hair developmentSevere obesity vs. normalPubic hair developmentFemaleHR1.881.73–2.040.93Pubic hair developmentObesity vs. normalPubic hair developmentFemaleHR1.551.48–1.620.95Early menarche (<12 years)Severe obesity vs. normalEarly menarche (<12 years)FemaleOR2.592.25–2.99–Early menarche (<12 years)Obesity vs. normalEarly menarche (<12 years)FemaleOR2.252.08–2.43–Early menarche (<12 years)Overweight vs. normalEarly menarche (<12 years)FemaleOR1.71.59–1.82–Early menarche (<12 years)Underweight vs. normalEarly menarche (<12 years)FemaleOR0.590.5–0.69–Li et al. (32)2018China (Chongqing)Longitudinal birth cohort study259Menarche (girls)BMI (continuous, kg/m2)Menarche (girls)Mixed/unknownHR1.2051.151–1.261<0.001Menarche (girls)Overweight vs. normalMenarche (girls)Mixed/unknownHR2.6051.716–3.956<0.001Menarche (girls)Obesity vs. normalMenarche (girls)Mixed/unknownHR2.5651.603–4.103<0.001Breast development B2 (girls)BMI (continuous, kg/m2)Breast development B2 (girls)Mixed/unknownHR1.0921.045–1.142<0.001Breast development B2 (girls)Overweight vs. normalBreast development B2 (girls)Mixed/unknownHR1.4691.024–2.1080.037Breast development B2 (girls)Obesity vs. normalBreast development B2 (girls)Mixed/unknownHR1.3990.913–2.1450.123First nocturnal emission (boys)BMI (continuous, kg/m2)First nocturnal emission (boys)Mixed/unknownHR1.0541.004–1.1060.033First nocturnal emission (boys)Overweight vs. normalFirst nocturnal emission (boys)Mixed/unknownHR1.2970.818–2.0560.27First nocturnal emission (boys)Obesity vs. normalFirst nocturnal emission (boys)Mixed/unknownHR1.210.714–2.0510.479Testicular development ≥4mL (boys)BMI (continuous, kg/m2)Testicular development ≥4 mL (boys)Mixed/unknownHR1.0981.063–1.135<0.001Testicular development ≥4 mL (boys)Overweight vs. normalTesticular development ≥4 mL (boys)Mixed/unknownHR1.6721.204–2.3220.002Testicular development ≥4 mL (boys)Obesity vs. normalTesticular development ≥4 mL (boys)Mixed/unknownHR1.5981.135–2.2490.008Mamun et al. (33)2009AustraliaPopulation-based prospective cohort study (Danish National Birth Cohort, nested Puberty Cohort)13,21914 years Tanner stage 5 vs. stage 1/25 years BMI z-score (continuous)14 years Tanner stage 5 vs. stage 1/2Mixed/unknownOR1.881.53–2.31NR14 years Tanner stage 4 vs. stage 1/25 years overweight vs. normal14 years Tanner stage 4 vs. stage 1/2Mixed/unknownOR1.91.2–3.01NRThe complete table is attached as Table S1. BMI, body mass index; CI, confidence interval; HR, hazard ratio; MA, meta-analysis; NR, not reported; OR, odds ratio. Characteristics of included studies and effect sizes The complete table is attached as Table S1. BMI, body mass index; CI, confidence interval; HR, hazard ratio; MA, meta-analysis; NR, not reported; OR, odds ratio. Forest plot of the association between childhood overweight/obesity and pubertal development. Each square represents the effect estimate from an individual study, with the size proportional to the study weight. Horizontal lines indicate 95% CIs. The diamond represents the pooled effect estimate from the random-effects model. The vertical dashed line at 1.0 indicates no effect. Overall pooled OR/HR =1.56 (95% CI: 1.44–1.69), indicating that childhood overweight/obesity is significantly associated with earlier pubertal development. Heterogeneity: I2=96.0%, Q=580.11, P<0.001. CI, confidence interval; HR, hazard ratio; OR, odds ratio. Subgroup analysis by weight classification Subgroup analysis by weight classification revealed an increasing effect size with greater weight severity, demonstrating a clear dose-response relationship (Figure 3, Table 3): overweight vs. normal: OR/HR =1.46 (95% CI: 1.31–1.63), P<0.001; I2=93.5%, 10 effect sizes; obesity vs. normal: OR/HR =1.55 (95% CI: 1.36–1.77), P<0.001; I2 =96.3%, 9 effect sizes; Severe obesity vs. normal weight: OR/HR =1.69 (95% CI: 1.37–2.07), P<0.001; I2=96.7%, 5 effect sizes. Trend analysis revealed a significant linear trend between weight classification and risk of precocious puberty (P for trend <0.001), supporting the existence of a dose-response relationship. Subgroup analysis by weight category. Studies were stratified into three subgroups: overweight vs. normal weight (pooled OR/HR =1.46, 95% CI: 1.31–1.63, I2=93.5%), obesity vs. normal weight (pooled OR/HR =1.55, 95% CI: 1.36–1.77, I2=96.3%), and severe obesity vs. normal weight (pooled OR/HR =1.69, 95% CI: 1.37–2.07, I2=96.7%). A clear dose-response relationship was observed: the risk of early pubertal development increased progressively with the severity of weight excess. Blue diamonds represent subgroup pooled estimates; the red diamond represents the overall pooled estimate. CI, confidence interval; HR, hazard ratio; OR, odds ratio. Table 3Summary of meta-analysis resultsAnalysisNo. of effect sizes (n)Pooled OR/HR95% CIP valueI2 (%)τ2Overall effect241.561.44–1.69<0.00196.00.043By weight category Overweight vs. normal101.461.31–1.63<0.00193.50.038 Obesity vs. normal91.551.36–1.77<0.00196.30.052 Severe obesity vs. normal51.691.37–2.07<0.00196.70.071By outcome type Menarche52.241.82–2.75<0.00191.40.058 Tanner stage22.061.44–2.94<0.0010.00.000 Breast development51.471.35–1.60<0.00182.50.012 Pubic hair development51.471.30–1.66<0.00196.50.025 Testicular development51.221.15–1.29<0.00169.20.005CI, confidence interval; HR, hazard ratio; OR, odds ratio. Summary of meta-analysis results CI, confidence interval; HR, hazard ratio; OR, odds ratio. Subgroup analysis by outcome type Subgroup analysis by outcome type revealed significant differences in effect sizes across various pubertal developmental indicators (Figure 4, Table 3): Menarche: OR/HR =2.24 (95% CI: 1.82–2.75), P<0.001; I2 =91.4%, 5 effect sizes; Tanner stage: OR/HR =2.06 (95% CI: 1.44–2.94), P<0.001; I2=0.0%, 2 effect sizes; Breast development: OR/HR =1.47 (95% CI: 1.35–1.60), P<0.001; I2=82.5%, 5 effect sizes; Pubic hair development: OR/HR =1.47 (95% CI: 1.30–1.66), P<0.001; I2=96.5%, 5 effect sizes; Testicular development: OR/HR =1.22 (95% CI: 1.15–1.29), P<0.001; I2=69.2%, 5 effect sizes. Results indicate that age at menarche is most sensitive to body weight status (OR/HR =2.24), with an effect size nearly double that of testicular development (OR/HR =1.22). Effect sizes for female-related pubertal indicators (menarche, breast development) were generally higher than those for male-related indicators (testicular development), suggesting gender differences. Subgroup analysis by outcome type. The strongest association was observed for menarche (pooled OR/HR =2.24, 95% CI: 1.82–2.75, I2=91.4%), followed by Tanner stage (2.06, 95% CI: 1.44–2.94, I2=0.0%). Effects on breast development (1.47, 95% CI: 1.35–1.60), pubic hair development (1.47, 95% CI: 1.30–1.66), and testicular development (1.22, 95% CI: 1.15–1.29) were also statistically significant. The stronger effect on menarche may reflect the greater sensitivity of menarcheal timing to metabolic and nutritional factors. CI, confidence interval; HR, hazard ratio; OR, odds ratio. Sensitivity analysis Individual study exclusion sensitivity analysis showed that removing any single study resulted in pooled effect sizes ranging from 1.51 to 1.62, all remaining statistically significant, indicating robust results. Excluding the large-sample study by Aghaee et al. (2022) slightly increased the effect size (OR/HR =1.68, 95% CI: 1.42–1.99), but heterogeneity remained high (I2=89.2%). Sensitivity analysis using only high-quality studies (NOS ≥7) yielded results consistent with the primary analysis (OR/HR =1.54, 95% CI: 1.41–1.68). Analysis restricted to prospective studies also demonstrated robustness (OR/HR =1.55, 95% CI: 1.43–1.68). Publication bias assessment Only three studies (<10) were included in this meta-analysis; therefore, no publication bias assessment was conducted. Discussion Key findings This systematic review and meta-analysis found that children with overweight or obesity had a significantly higher risk of early pubertal development than normal-weight peers. A clear dose-response relationship was identified, with risk rising gradually from overweight to obesity and severe obesity, supporting a potential causal link. Effect sizes differed considerably by pubertal marker and sex, with the strongest association found for girls’ menarche and a weaker association for boys’ testicular enlargement, while breast and pubic hair development showed moderate effects. Although statistical heterogeneity was high, all subgroup and sensitivity analyses confirmed consistent and robust positive associations. The evidence map showed that most available data focused on BMI and menarche age, with limited evidence on pubertal tempo, bone age and growth spurts; heterogeneity may stem from variations in analytical methods, outcome and exposure definitions, population characteristics and confounder adjustment. Overall results were consistent with previous large cohort studies indicating that excess childhood fat accelerates puberty, especially in girls. This is the first systematic review to comprehensively map the evidence, though high heterogeneity and residual confounding require cautious interpretation. Strengths and limitations This study has several notable strengths: it is the first systematic review to comprehensively map evidence linking childhood weight status to pubertal timing, covering 41 studies with a large, geographically diverse sample across 18 countries via searches of five Chinese- and English-language databases without language restrictions; it conducted pre-specified multidimensional subgroup analyses to identify dose-response patterns and sex differences, adhered to PRISMA guidelines with prospective PROSPERO registration, and its findings were strengthened by consistent positive effect directions across all subgroups and robust sensitivity analyses. However, several limitations should be noted: substantial statistical heterogeneity was observed, requiring cautious interpretation of the pooled estimate; only a small number of effect-size estimates could be quantitatively pooled, precluding formal publication-bias assessment; combining different effect metrics may have introduced imprecision; residual confounding and inability to infer causation remain inherent to observational evidence due to inconsistent control of key covariates; most studies were from high- or middle-income settings, limiting generalizability; and inconsistent outcome and exposure definitions across studies contributed to heterogeneity and reduced comparability. Four further limitations deserve emphasis. First, US participants predominated within the pooled sample, and especially within the three studies contributing to the meta-analysis (~91% of pooled participants), which may further constrain the global applicability of the quantitative estimates. Second, the included studies were drawn from general pediatric populations and, by design, excluded children with recognized endocrine or genetic disorders that affect growth (e.g., congenital adrenal hyperplasia, Prader-Willi syndrome, precocious-puberty syndromes); nevertheless, few studies characterized the underlying cause of excess weight or screened for sub-clinical hormonal or metabolic disturbances, so coexisting metabolic conditions that contribute to both adiposity and earlier maturation cannot be excluded as partial drivers of the observed associations. Third, because the contributing studies framed the exposure primarily as overweight/obesity versus normal weight, we could not formally analyze a separate “normal-weight” subgroup or model body weight as a continuum across its entire range; the available continuous-BMI estimates and the reduced risk observed in underweight-versus-normal comparisons nonetheless suggest that the association extends across the whole weight spectrum rather than being confined to the obese range. Fourth, although restricting the synthesis to studies sharing identical definitions of overweight, obesity, and puberty would have improved comparability, it was not feasible given the already small number of poolable studies; we therefore retained all eligible studies and examined definitional heterogeneity through subgroup and sensitivity analyses, an approach that preserves statistical power at the cost of some between-study comparability. Comparison with similar research Our findings are broadly concordant with earlier studies. Lee et al. (72) followed 401 girls across 10 centers in the United States and reported that higher BMI was associated with earlier onset of breast and pubic hair development. Freedman et al. (73), drawing on 11,218 participants from the Bogalusa Heart Study, documented that the age at menarche was approximately 0.5 years earlier among overweight girls compared with their normal-weight counterparts. Among the studies included in the present review, the cohort analysis by Aghaee et al. (31) is one of the largest to date. Using electronic health records from Kaiser Permanente Northern California, the investigators found that severely obese girls had 2.59-fold higher odds of early menarche (<12 years) relative to normal-weight girls (OR =2.59, 95% CI: 2.25–2.99), and severely obese boys had a significantly elevated hazard of early testicular enlargement (HR =1.23, 95% CI: 1.15–1.32). Li et al. (32) prospectively followed 1,237 children in Chongqing, China, and observed that both overweight (HR =2.60, 95% CI: 1.72–3.96) and obese (HR =2.56, 95% CI: 1.60–4.10) girls experienced substantially earlier menarche; their data further indicated a greater impact of weight on female than on male pubertal timing, in line with the present meta-analytic findings. Mamun et al. (33), analyzing 2,897 participants from the Australian Mater-University Birth Cohort, demonstrated that children who were overweight at age 5 were 2.32 times more likely to reach Tanner stage 5 by age 14 than their normal-weight peers (OR =2.32, 95% CI: 1.32–4.05), suggesting that early adiposity exerts a lasting influence on pubertal maturation. Explanations of findings To our knowledge, this is the first systematic review to comprehensively map the evidence linking childhood weight status to pubertal timing, encompassing 41 studies with a combined sample of more than 410,000 participants across 18 countries and regions. The meta-analytic results indicate that overweight/obese children face a 56% higher risk of early pubertal development relative to their normal-weight peers (OR/HR =1.56, 95% CI: 1.44–1.69). Beyond this overall effect, several specific findings warrant detailed explanation. First, a dose-response gradient was apparent: the risk of early puberty increased progressively from overweight (OR/HR =1.46) to obesity (OR/HR =1.55) to severe obesity (OR/HR =1.69). This graded relationship satisfies the biological gradient criterion outlined by Hill (74) and strengthens the plausibility of a causal link. The differential impact of weight status across pubertal markers and sexes also requires careful interpretation. Second, effect sizes varied by pubertal marker, with female indicators exhibiting greater sensitivity to weight status. The largest effect was observed for menarche (OR/HR =2.24), roughly twice that for testicular enlargement (OR/HR =1.22). Effect sizes for breast development and pubic hair development (both 1.47) also exceeded that for testicular enlargement. These sex differences are consistent with prior reports (75,76) and may reflect the distinctive role of adipose tissue—via aromatase-mediated estrogen production and leptin signaling—in female pubertal maturation. The robustness of these findings, despite considerable statistical heterogeneity, merits further discussion. Third, despite high statistical heterogeneity (I2=96.0%), all subgroup analyses yielded positive and statistically significant associations (all effect sizes >1.0). Sensitivity analyses confirmed the stability of the pooled estimate after sequential exclusion of individual studies, restriction to high-quality reports, and restriction to prospective designs. Finally, beyond the quantitative findings, the qualitative synthesis of the evidence base reveals important research gaps. Additionally, the evidence map revealed that the available quantifiable evidence is concentrated on BMI/weight classification and weight trajectory in relation to pubertal onset risk and menarche timing. By contrast, data on pubertal progression tempo, bone age, and growth spurt remain scarce, highlighting gaps that future research should address through more uniform outcome ascertainment. Implications and actions needed Childhood overweight/obesity increases early puberty risk by 56% with a clear dose–response gradient, with important implications: clinically, monitor pubertal development in these children and provide weight management counselling to delay puberty and mitigate long-term health risks; public health policies should prioritize early childhood obesity prevention programs, noting larger effects in girls but not neglecting boys; future research needs standardized prospective measures, more data on pubertal progression/bone age/growth spurts, consistent confounder adjustment, and diverse populations to enhance generalisability, all to reduce adverse health impacts of obesity-induced early puberty. Conclusions This systematic review and meta-analysis confirms that childhood overweight or obesity is significantly associated with an increased risk of early puberty in a clear dose–response manner, with the association notably stronger among females, especially for key pubertal markers such as menarche. Despite considerable between-study heterogeneity, subgroup analyses consistently support the robustness of this relationship. These findings emphasize the importance of childhood weight management for healthy pubertal development and long-term well-being, suggesting that pubertal monitoring should be included in routine care for overweight and obese children and that public health measures against childhood obesity should be enhanced. Future research is needed to establish standardized definitions, explore underlying biological mechanisms, and evaluate the impact of weight-management interventions on pubertal timing. Supplementary The article’s supplementary files as