Skip to main content
NCBI home page
UKPMC Funders Author Manuscripts logoLink to UKPMC Funders Author Manuscripts
. Author manuscript; available in PMC: 2025 Aug 15.
Published in final edited form as: Eur J Prev Cardiol. 2025 Aug 5;32(10):857–866. doi: 10.1093/eurjpc/zwaf051

Reassessing the association between age at menarche and cardiovascular disease: Observational and Mendelian Randomisation analyses

Lena Tschiderer 1,*, Sanne AE Peters 2,3,4, Yvonne T van der Schouw 2, Stephen Burgess 5,6,7, Janneke Luijken 2, Cheyenne Bijmolt 2, Houda Soliman 2, Adam Butterworth 5,6,8,9,10, Angela M Wood 11,12,13,8,10, Tammy YN Tong 14, Christina C Dahm 15, Lisa Seekircher 1, Anne Tjønneland 16,17, Lene Mellemkjær 16, Matthias B Schulze 18,19, Giovanna Masala 20, Sabina Sieri 21, Salvatore Panico 22, Carlotta Sacerdote 23, Jolanda MA Boer 24, WM Monique Verschuren 24,2, Carlota Castro-Espin 25,26, Dafina Petrova 27,28,29, Sandra M Colorado-Yohar 30,29,31, Conchi Moreno-Iribas 32,29,33, Elisabete Weiderpass 34, Alicia K Heath 35, Ioanna Tzoulaki 35,36, Peter Willeit 1,5, N Charlotte Onland-Moret 2
PMCID: PMC7618008  EMSID: EMS207595  PMID: 40099743

Abstract

Aims

Observational studies have shown a U-shaped association between age at menarche (AAM) and cardiovascular disease (CVD). We assessed non-linearity of the observational association and the potential causal relationship between AAM and CVD using data from EPIC-CVD and the UK Biobank.

Methods

We included women without pre-existing myocardial infarction (MI) or stroke at baseline. We estimated hazard ratios for incident MI, ischaemic and haemorrhagic stroke later in life using Cox regression in observational analyses and conducted non-linear Mendelian Randomisation (MR) based on fractional polynomials and linear MR based on inverse-variance weighted regression.

Results

We analysed 283,210 women with a median AAM of 13 (IQR 12-14) years in both EPIC-CVD and the UK Biobank, of which 8,468 experienced MI, 5,501 an ischaemic and 1,887 a haemorrhagic stroke. The association between AAM and MI and ischaemic stroke was U-shaped with higher risks in women aged ≤12 and ≥16 compared to those aged 13 years at menarche. Our MR analyses found no evidence for non-linearity between genetically proxied AAM and any CVD endpoint later in life, but each year higher genetically proxied AAM was related to a lower risk MI (hazard ratio 0.92 [95% CI 0.86-0.99]), but not to ischaemic and haemorrhagic stroke.

Conclusion

This study supported non-linear observational associations between AAM and myocardial infarction and ischaemic stroke. MR analyses suggested a causal relationship between higher AAM and risk of MI without an indication for non-linearity. There was no support for a potential causal link with ischaemic and haemorrhagic stroke.

Keywords: age at menarche, cardiovascular risk, individual participant data meta-analysis, Mendelian Randomisation, non-linearity

Introduction

Average age at menarche (AAM) varies from country to country, ranging from 12 to 16 years,1 and a secular decline in age at pubertal onset has been shown worldwide.2 Factors that have been related to the onset of menarche include several pre-pubertal anthropometry- and adiposity-related measures such as body fat and body size.3,4 In addition, a younger AAM has been linked to several health-related traits including lower lung function, higher risk for psychosocial and mental health problems5 and increased risk for breast and endometrial cancer.6

Several observational studies have reported an association between AAM and cardiovascular disease (CVD) such as coronary heart disease, stroke, and peripheral artery disease.4 In a large-scale individual participant data meta-analysis, both early and late menarche were associated with higher cardiovascular risk suggesting a non-linear U-shaped association.7 Moreover, traditional cardiovascular risk factors such as blood pressure and body mass index (BMI) have been suggested to act as potential mediators on the association between AAM and CVD.8,9

Multiple Mendelian Randomisation (MR) studies have investigated whether the relationship between AAM and cardiovascular risk factors and CVD events is potentially causal. Younger genetically proxied AAM has been associated with higher fasting blood glucose levels,10 higher adult BMI,11 and a higher risk of ischaemic heart disease,12 coronary artery disease, and heart failure.9 Despite the non-linear association seen in observational studies, previous MR studies have assumed a log-linear relationship between genetically proxied AAM and cardiovascular risk. However, it is crucial to know the specific shape of the causal relationship between AAM and CVD and whether both early and late menarche cause CVD later in life. This would enhance our understanding of CVD in women and provide deeper insights into the underlying mechanisms driving the progression of the disease.

We examined whether the shape of association between AAM and CVD risk later in life is non-linear in an observational and a MR analysis setting using data from the European Prospective Investigation into Cancer and Nutrition – Cardiovascular Disease (EPIC-CVD) study and the UK Biobank.

Methods

Results are presented according to the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE)13 and the STROBE-MR Statement.14 The STROBE and STROBE-MR checklists are provided in Table S1 and Table S2.

Study design and data sources

We included data from the EPIC-CVD study and the UK Biobank. Details of these studies have been published previously.15–17 The EPIC study recruited more than 500,000 individuals from 23 centres across Europe between 1992 and 2000.16 EPIC-CVD is a case-cohort study nested in the EPIC study comprising individuals from a random sub-cohort and all other EPIC participants with incident coronary heart disease and stroke. The UK Biobank is a prospective study conducted in the general population of the UK and included >500,000 participants aged between 40 and 69 years who were recruited between 2006 and 2010.15 Women without a history of myocardial infarction (MI) or stroke at baseline were eligible to be included in the present study.

For the MR analysis, we used genetic variants associated with the exposure, i.e., AAM, from a meta-analysis of genome-wide association studies (GWASs).18 To obtain genetic associations with the outcome, we used imputed genetic data that passed quality control from EPIC-CVD and the UK Biobank. In EPIC-CVD, genotyping was performed using the Human Core Exome array, Omni Exome Express array, and Illumina 660 Quad array and genotype imputation was based on the Haplotype Reference Consortium.19 In the UK Biobank, participants were genotyped with the Affymetrix UK BiLEVE Axiom array and the Affymetrix UKB Axiom Array;15,20 genotype imputation was performed using the Haplotype Reference Consortium as well as the UK10K haplotype reference panel.21

The GWAS on AAM identified 389 single nucleotide polymorphisms (SNPs).18 Of these, we excluded 61 SNPs because no data on these SNPs or any proxy variant were available in EPIC-CVD and nine SNPs because they were palindromic and had a minor allele frequency >0.45 (Figure S1). We harmonised beta coefficients of the SNPs to represent the same effect and non-effect alleles. For EPIC-CVD, we used proxy variants for 21 SNPs that were in high linkage disequilibrium (R2>0.8) (Table S3).

AAM was defined as self-reported age at first menstrual periods in both EPIC-CVD and the UK Biobank. In both studies, cardiovascular outcomes were defined using ICD codes. The cardiovascular outcomes studied in this analysis comprised MI (I21-I25 in EPIC-CVD; I21-I23, I24.1, and I25.2 in the UK Biobank), ischaemic stroke (I63-I64), and haemorrhagic stroke (I60-I61). In EPIC-CVD centres, events were ascertained and validated using different methods such as follow-up questionnaires and linkage with morbidity, hospital, and death registries. In the UK Biobank, we used algorithmically defined outcomes (field IDs 42000, 42008, 42010, and 42012) that occurred until September 30, 2021 and were based on hospital admission and death register data. Descriptions of how additional variables were assessed are provided in the Supplementary Methods.

Statistical analyses

We applied two-sided statistical tests and deemed P-values <0.05 as statistically significant. All analyses were conducted using R 4.0.5 (The R Foundation, Vienna, Austria).

Descriptive data

We summarised categorical variables as numbers (percentages) and continuous variables as medians (interquartile ranges [IQRs]). For EPIC-CVD, we provide summarised descriptives only in women within the sub-cohort due to the case-cohort design of the study, as those best reflect baseline characteristics of the entire EPIC cohort.

Observational analysis

In our observational analysis, we multiply imputed missing values in the exposure variable and in the covariates based on chained equations (14 datasets, 30 iterations) using the R-package mice v3.14.0.22 Details about the imputation process are provided in the Supplementary Methods. We categorised AAM into <12, 12, 13, 14, 15, and ≥16 years. For all analyses, we used women aged 13 years at menarche as the reference group as this was the median AAM in both the EPIC-CVD study and the UK Biobank. We report hazard ratios in each category using quasi variances, which allows head-to-head comparisons of the individual categories of AAM.23 We modelled the association between AAM and risk of MI and stroke subtypes using Cox regression analysis. To take the case-cohort design into account, we implemented Prentice-weighted Cox regression for EPIC-CVD.24 Follow-up time was defined as time to first MI, ischaemic stroke, or haemorrhagic stroke, death, or loss to follow-up, whichever occurred first. Consequently, if a person experienced more than one CVD event, e.g. both a MI and an ischaemic stroke, only the first event would contribute to any of the Cox regression analyses. We obtained hazard ratios in each study separately and combined them using multivariate random-effects meta-analysis.25 Moreover, we adjusted for age, education (high, medium versus low), smoking status (current, ex versus never), and BMI (kg/m2). For EPIC-CVD, we additionally stratified by country.

In a sensitivity analysis, we restricted our primary observational analyses to women who were included in the MR study to enhance comparability.

MR study

In the MR study, we included all women with (1) genetic data on the SNPs used in our instrumental variable that passed quality control and (2) data on AAM. We calculated F-statistics and R2 for each study by regressing the genetic instruments on AAM using women from the random sub-cohort for EPIC-CVD.

To study the distribution of genetically proxied AAM across several participant characteristics, we calculated a genetic risk score (GRS) based on the GWAS summary-level data on genetic associations with AAM.18 This GRS contained all variants included in the instrumental variable used in the MR analysis weighted by the corresponding effect size obtained from the GWAS on AAM.18 We then compared thirds of the GRS across the variables age, hypertension, diabetes mellitus, BMI, height, smoking status, education, total cholesterol, C-reactive protein, use of oral contraceptive pills, menopausal status, and age at menopause. For EPIC-CVD, we included sub-cohort participants only to enhance comparability.

We applied non-linear MR analysis based on fractional polynomials26 using the R-package SUMnlmr27 based on the doubly-ranked stratification method as suggested by Tian et al., which ranks individuals into pre-strata based on their level of the GRS, before stratifying individuals by ranking them within each pre-stratum according to their phenotypic AAM.28 The number of pre-strata depends on the sample size of the study and the number of strata selected (sample size = number of strata × number of pre-strata).28 For the present analysis, we selected ten strata. In each of the strata, we conducted Cox regression analysis in the UK Biobank and Prentice-weighted Cox regression analysis in EPIC-CVD24 to obtain genetic associations with MI, and ischaemic and haemorrhagic stroke based on the GRS. We used age as the underlying time scale. Moreover, in each of the strata, we performed linear regression analysis regressing AAM on the GRS to obtain genetic associations with AAM. In EPIC-CVD, we adjusted for age, genotype array, and the first 10 genetic principal components and stratified by country. In the UK Biobank, we adjusted for age, genotype array, and the first 16 genetic principal components. Study-specific estimates in each stratum were combined using fixed-effect meta-analysis. We used fixed-effect meta-analysis because we assumed that the studies would estimate a common true effect size. Fractional polynomials were fitted using the frac_poly_summ_mr function from the R-package SUMnlmr.27 Non-linearity was assessed by a fractional polynomial non-linearity test comparing a non-linear fractional polynomial model with a linear model.26,27 In order to study selection bias potentially introduced by stratification, we assessed the relationship of the strata on AAM with several other traits as suggested previously.29

For situations where no evidence of non-linearity was found, we also report results of linear MR (using the R-package MendelianRandomization v0.6.030) performing inverse-variance weighted regression as main analysis and simple and weighted median regression, MR-Egger, and MR-PRESSO31 as sensitivity analyses. To obtain genetic associations with MI, ischaemic and haemorrhagic stroke, we used the same methods as described above, i.e., we performed Prentice-weighted Cox regression in EPIC-CVD24 and Cox regression analysis in the UK Biobank using age as the underlying time scale adjusting (and stratifying) for the same set of variables and combining study-specific hazard ratios using fixed-effect meta-analysis.

Results

Participants

The study flow diagram is shown in Figure 1 and numbers of participants excluded are described in detail in the Supplementary methods. 283,210 women (15,927 from EPIC-CVD [9,516 from the EPIC-CVD sub-cohort], 267,283 from the UK Biobank) were included in the observational analysis and 262,308 women (12,329 from EPIC-CVD [7,370 from the EPIC-CVD sub-cohort], 249,979 from the UK Biobank) were included in the MR analysis.

Figure 1. Flow diagram.

Figure 1

Open in a new tab

Abbreviations: CHD, coronary heart disease; EPIC-CVD, European Prospective Investigation into Cancer and Nutrition – Cardiovascular Disease; MI, myocardial infarction; QC, quality control.

Descriptive data

An overview of the participant characteristics is provided in Table 1. Median baseline age was 52.1 years (IQR 45.5-59.0) and 57.0 years (IQR 50.0-63.0) in the EPIC-CVD sub-cohort and the UK Biobank, respectively. The median AAM was 13.0 years (IQR 12.0-14.0) in both the EPIC-CVD sub-cohort and in the UK Biobank.

Table 1. Baseline characteristics.

EPIC-CVDa (n=9,516) UK Biobank (n=267,283)
Characteristic nb Median [IQR], n (%) nb Median [IQR], n (%)
Age, years 9,516 52.1 [45.5-59.0] 267,282 57.0 [50.0-63.0]
Age at menarche, years 9,081 13.0 [12.0-14.0] 258,630 13.0 [12.0-14.0]
Hypertension 9,445 3,239 (34.3%) 266,172 116,493 (43.8%)
Diabetes mellitus 8,779 227 (2.6%) 266,121 9,609 (3.6%)
Body mass index, kg/m2 9,447 25.1 [22.7-28.4] 265,907 26.1 [23.4-29.7]
Height, cm 9,481 161.2 [156.8-166.0] 266,189 162.0 [158.0-167.0]
Smoking status 9,439 265,833
     Never 5,267 (55.8%) 159,230 (59.9%)
     Ex 1,990 (21.1%) 83,107 (31.3%)
     Current 2,182 (23.1%) 23,496 (8.8%)
Education 9,349 262,098
     Low 4,135 (44.2%) 43,977 (16.8%)
     Medium 1,353 (14.5%) 66,861 (25.5%)
     High 3,861 (41.3%) 151,260 (57.7%)
Total cholesterol, mmol/L 9,138 5.8 [5.1-6.6] 249,012 5.8 [5.1-6.6]
C-reactive protein, mg/L 9,134 1.1 [0.5-2.5] 248,590 1.4 [0.6-3.0]
Ever use of OCP 9,154 5,067 (55.4%) 265,913 215,982 (81.2%)
Postmenopausal 9,516 5,292 (55.6%) 267,282 194,224 (72.7%)
Age at menopause, years 4,754 49.0 [45.0-52.0] 182,351 50.0 [45.0-52.0]
Open in a new tab

aIncluding data from the EPIC-CVD sub-cohort, bNumber of non-missing values. Abbreviations: EPIC-CVD, European Prospective Investigation into Cancer and Nutrition – Cardiovascular Disease; IQR, interquartile range; OCP, oral contraceptive pill.

Median time to event or end of follow-up was 12.9 years (IQR 10.8-14.3) in the EPIC-CVD sub-cohort and 12.6 years (IQR 11.8-13.3) in the UK Biobank. 8,468 women experienced a MI (3,736 in EPIC-CVD [of whom 184 in the EPIC-CVD sub-cohort], 4,732 in the UK Biobank), 5,501 women experienced an ischaemic stroke (2,542 in EPIC-CVD [of whom 137 in the EPIC-CVD sub-cohort], 2,959 in the UK Biobank), and 1,887 women experienced haemorrhagic stroke (725 in EPIC-CVD [of whom 44 in the EPIC-CVD sub-cohort], 1,162 in the UK Biobank).

Observational analysis

Results of our observational analysis are provided in Figure 2 and Table S4. The associations between AAM and risks of MI and ischaemic stroke were U-shaped, with higher risks for women aged ≤12 years or ≥16 years at menarche as compared to those aged 13 years. We found no significant association between AAM and risk of haemorrhagic stroke. As depicted in Table S5 results remained largely robust when restricting the observational analysis to women who were included in the MR analysis.

Figure 2. Observational results for the associations between age at menarche and risks of myocardial infarction and ischaemic and haemorrhagic stroke.

Figure 2

Open in a new tab

The models were adjusted for age, education (high, medium versus low), smoking status (current, ex versus never), and body mass index (kg/m2). For EPIC-CVD, models were stratified by country. Results are presented based on quasi variances using 13 years at menarche as reference category.

MR analysis

SNP-specific genetic associations with cardiovascular outcomes are provided in Table S3. The R2 of our genetic instrument was 10.8% and 5.6% in EPIC-CVD and the UK Biobank, respectively. The F-statistic was 2.67 in EPIC-CVD and 46.00 in the UK Biobank. As depicted in Table S6, the participant characteristics were largely similar across thirds of the GRS.

Figure 3 shows the results of the non-linear MR analysis. We found no evidence for statistically significant deviation from linearity in the associations between genetically proxied AAM and MI, ischaemic and haemorrhagic stroke (all P-values for non-linearity >0.05). When analysing the relationship of genetically proxied AAM and other traits across strata of AAM, we found no specific patterns of association across the traits (Figure S2). In addition, there was no clear evidence for a relationship with other traits, except for BMI, for which we found inverse associations across all strata of AAM, and for height, for which we found a positive association across all strata.

Figure 3. Mendelian Randomisation analysis for the relationships between age at menarche and risks of myocardial infarction and ischaemic and haemorrhagic stroke based on fractional polynomials.

Figure 3

Open in a new tab

The P-value for non-linearity is based on a fractional polynomial non-linearity test comparing a non-linear fractional polynomial model with a linear model.26,27 For EPIC-CVD, models were adjusted for age, genotyping array, and the first 10 genetic principal components, and stratified by country. For the UK Biobank, models were adjusted for age, genotyping array, and the first 16 genetic principal components.

When conducting linear MR analysis using inverse-variance weighted regression (see Figure S3, and Figure S4 and Figure S5 for study-specific results), a 1-year higher genetically proxied AAM was related to a lower risk of MI with a hazard ratio of 0.92 (95% CI 0.86-0.99; P=0.036). No significant associations were found between genetically proxied AAM and risks of ischaemic and haemorrhagic stroke, with hazard ratios of 1.00 (0.91-1.09; P=0.929) and 1.02 (0.89-1.17; P=0.805), respectively. Findings were consistent when applying simple and weighted median regression although they did not reach statistical significance for MI and MR-Egger suggested no evidence for directional pleiotropy. MR-PRESSO detected only one significant outlier when studying the relationship between genetically proxied AAM and MI in EPIC-CVD. However, when excluding this SNP from the analysis, the result was almost identical. MR-PRESSO detected no statistically significant outliers for any other analysis.

Discussion

In this study, we found U-shaped associations between AAM and risks of MI and ischaemic stroke later in life based on observational data, but no significant association with haemorrhagic stroke. This non-linear shape of association was not confirmed by MR analysis. However, we found a statistically significant relation per year younger genetically proxied AAM with higher risk of MI but not with ischaemic or haemorrhagic stroke.

Comparison to findings from the literature

A previous large-scale individual-participant data meta-analysis including data on more than 300,000 participants from twelve studies (including data from the UK Biobank) reported similar results to our observational analysis.7 In that meta-analysis, the association between AAM and risk of CVD was U-shaped. Compared to women aged 13 at menarche, those aged ≤10, 11, and 16 years at menarche had a statistically significantly higher risk for CVD with hazard ratios of 1.16 (95% CI 1.04-1.29), 1.15 (1.07-1.22), and 1.15 (1.06-1.24), respectively.7 Shapes of associations were also U-shaped for coronary heart disease and stroke.7

To the best of our knowledge this is the first MR study that investigated the specific shape of association between genetically proxied AAM and risk of CVD later in life as previously conducted MR studies assumed a linear association. In the present analysis, we found no evidence for deviation from a linear relationship between genetically proxied AAM and risk of any CVD endpoint. Consequently, the linearity assumption of previous MR studies seems to be valid. These studies also reported similar findings to those herein. A MR analysis based on summary-level data reported significant associations between genetic liability to older AAM and lower risk of coronary artery disease with an odds ratio of 0.91 (95% CI 0.88-0.94) per year older genetically proxied AAM, but no statistically significant relationship with the risks of ischaemic stroke and total stroke.9 Findings were similar in a sensitivity analysis restricted to women from the UK Biobank for genetic association with outcomes, with an odds ratio for coronary artery disease of 0.91 (0.85-0.96) per year older genetically proxied AAM.9 Another summary-level MR analysis reported older genetically proxied AAM to be related to a lower risk of ischaemic heart disease with an odds ratio of 0.80 (0.72-0.88).12 In an additional linear MR analysis, we also corroborate previous findings by reporting a hazard ratio for MI of 0.92 (0.86-0.99) per year older genetically proxied AAM and no statistically significant relationships with ischaemic stroke. We did not find significant relationships with haemorrhagic stroke. Contrarily, another summary-level MR study reported a potential causal relationship between older genetically proxied AAM and lower risk of intracerebral haemorrhage.32

Later menarche and risk of MI

Our observational analysis revealed a U-shaped association between AAM and risk of MI. The corresponding MR analysis suggested no non-linearity but a significant relationship with higher risk of MI per year younger genetically proxied AAM. This raises the question of why women with later AAM are at higher risk for MI in observational studies and whether observational findings are affected by confounding. Several pre-pubertal adiposity measures have been suggested to be causally related to later AAM, including lower BMI, lower total body fat, and lower waist-height ratio.3 A GWAS on childhood body fatness also reported a genetic correlation between AAM and childhood body fatness.33 In addition, smaller early life body size (self-reported variable, asking whether individuals considered themselves as being thinner, plumper, or average at age ten) and lower childhood BMI have been causally related to a lower risk of MI and coronary artery disease, respectively.34,35 Potential additional confounding factors in the observational analyses are childhood undernutrition and anorexia nervosa, which have been related to both late menarche36,37 as well as to cardiovascular complications.38,39 A systematic review on non-genetic determinants of AAM additionally highlighted potential implications of psychological factors, such as adverse childhood experiences and stressful family situations, and environmental factors on the timing of menarche.40 Although there are several theories on causes of later menarche, the specific mechanisms leading to higher observational risk of MI in women with late menarche still have to be elucidated. Future studies that investigate shared risk factors for both later menarche and risk of MI are needed to better understand the potential driving factors for the non-linear observational association between age at menarche and risk of MI.

Implications

According to our study, women with earlier menarche are at higher risk for MI and ischaemic stroke and earlier menarche appears to be causally related to MI. Several factors have been proposed to lie on the causal pathway between AAM and MI. Adult BMI has been suggested as a main driver of higher risk of coronary artery disease in women with lower genetically proxied AAM.9,41 In addition, type 2 diabetes mellitus (to which BMI is also causally linked42) has been proposed as a mediator on the effect of genetically proxied AAM on risk of coronary artery disease.9 Another MR mediation analysis suggested systolic blood pressure as a main mediator of the effect of genetically proxied AAM on coronary artery disease with a proportion mediated of 29%.43 In addition, a recent MR analysis found that 37.5% of the relationship between AAM and MI was mediated by genetically proxied current smoking behaviour, and levels of glycated haemoglobin, systolic blood pressure, and triglycerides.8 These results suggest that early menarche may cause cardiometabolic alterations on various levels, ultimately potentially leading to MI. However, the majority of these alterations may be traced back to obesity-related factors as that the relationship between AAM and hypertension44,45 and diabetes mellitus46 is also (partly) mediated by BMI. The reason for the causal association being specific to MI – but not ischaemic stroke – needs to be clarified. Prior MR studies have suggested a minor role of adiposity in the development of stroke.47 As BMI may be one of the main drivers of the association between genetically proxied AAM and risk of MI, this could be a potential reason for the apparent lack of causal relationship with stroke risk. However, it is clear that cardiometabolic risk profiles of women with early menarche require specific attention. Raising awareness of elevated cardiovascular risk and implementing strategies to reduce the risk to develop clinically manifest disease, such as MI, is, therefore, important to provide the opportunity for a long-term healthy life in women.

Strengths and limitations

Our study has several strengths. It comprises meta-analyses of both observational and MR analyses. In addition, it is the first MR analysis on AAM considering a non-linear shape of association. Moreover, our genetic instrument is based on findings from a large-scale GWAS that meta-analysed data from multiple studies.18 Furthermore, we used individual-level data from EPIC-CVD and the UK Biobank, which allowed us to restrict the data to women when obtaining genetic associations with cardiovascular outcomes rather than using sex-combined effect estimates. In addition, there have been criticisms of non-linear MR that can introduce bias.29 Therefore, we applied the doubly-ranked method, which has been shown to produce more robust results.28,29 Moreover, we have conducted a sensitivity analysis investigating the association between genetically proxied AAM and various traits across strata, which revealed no specific shape of association. Our study also has limitations. Both our exposure and outcome data were based on women who are mainly of European ancestry limiting the generalisability of our findings to women of other ancestries. Consequently, additional studies are needed in order to study the specific shape of association between AAM and CVD risk in populations of non-European ancestry. Previous observational studies indicated that associations between AAM and CVD risk in populations of Asian ancestry could be different from populations of European ancestry. For instance, a large-scale study conducted in Korea including >1 million postmenopausal women suggested a linear association between AAM and risk of MI and later menarche was linked to a higher risk of MI.48 Similarly, in an analysis of >1 million pre-menopausal Korean women, later menarche was related to a higher risk of CVD.49 Moreover, a systematic review highlighted that previous studies in individuals of Asian ancestry reported conflicting results.4 In addition, age at pubertal onset has been shown to decline secularly.2 Therefore, we cannot draw any conclusions about women born in decades other than those included in EPIC-CVD and the UK Biobank. Moreover, AAM was self-reported and may consequently be prone to recall bias, especially, given that AAM is an early life exposure. However, an analysis within the Newton Girls Study demonstrated a significantly high correlation between original and recalled AAM after 33 years.50 Furthermore, data on pre-pubertal cardiovascular risk factors were not available. Consequently, we could not account for these factors in our observational analyses. Another limitation is that the GWAS we used for genetic associations with AAM included data from EPIC and the UK Biobank preventing us from performing two-sample MR analysis with non-overlapping samples. In addition, as EPIC-CVD is a case cohort study on coronary heart disease and stroke events by design, it was not possible to analyse additional cardiovascular endpoints such as, for instance, heart failure and peripheral artery disease. Further studies are needed in order to investigate the shape of association between AAM and risk of additional cardiovascular outcomes. Moreover, although we aimed to harmonise endpoint definitions, the definition for MI differed slightly between the studies. When conducting MR analysis, there are three core assumptions on the genetic instrument: (1) it is associated with AAM, (2) it is not associated with AAM via confounding pathways, and (3) it can influence cardiovascular risk only via AAM. We checked whether our genetic instrument is likely to fulfil these assumptions. To fulfil the first assumption and select SNPs for our instrumental variable that are related to AAM, we used data from a large-scale GWAS on AAM.18 In addition, we assessed the association of our instrumental variables with phenotypic AAM. While in the UK Biobank the F-statistic was relatively high with 46.00, it was only 2.67 in EPIC-CVD due to the lower sample size. This could have resulted in weak instrument bias in EPIC-CVD. As the samples of the genetic associations with the exposure and outcome were overlapping, weak instrument bias could bias the results towards the observational association. However, R2 was 5.6% in the UK Biobank and was even higher with 10.8% in EPIC-CVD. The second assumption is usually very likely to hold as genetic variants are determined at conception. This assumption could be violated by population stratification. However, our MR analysis was mainly based on individuals of European ancestry and we additional adjusted our analyses for genetic principal components. It can usually not be excluded that the third assumption is violated in MR analysis. Nevertheless, we have conducted a range of sensitivity analysis in order to investigate the violation of this assumption. We studied the distribution of participant characteristics across the GRS for AAM and found that participant characteristics were distributed homogeneously across thirds of the GRS for AAM, which indicates that our instrumental variable is not related to these potential confounding factors. Moreover, MR-Egger identified no direct pleiotropy (all P-values of intercepts >0.05) and no significant outliers were detected by MR-PRESSO, suggesting that our findings are unlikely to be driven by pleiotropic effects.

Conclusions

In this large European study, we found U-shaped observational associations between AAM and risks of MI and ischaemic stroke later in life. MR analysis suggested no non-linear relationship between genetically proxied AAM and any cardiovascular outcome later in life, yet each year of higher genetically proxied AAM was related to a lower risk of MI.

Supplementary Material

Supplementary Material

Lay summary.

This study reassessed the relationship between age at menarche and risk of cardiovascular disease based on observational analyses and Mendelian Randomisation studies.

  • In observational analyses, the association between age at menarche and myocardial infarction and ischaemic stroke was U-shaped, which was not confirmed by Mendelian Randomisation studies.

  • This study contributes valuable insights into the shape of association between age at menarche and risk of cardiovascular disease, which ultimately leads to an improved understanding of the development of cardiovascular disease in women.

Acknowledgements

This research has been conducted using the UK Biobank Resource (Application Number 29916 and 101819). We are extremely grateful to all participants in UK Biobank. We thank all EPIC participants and staff for their contribution to the study, the laboratory teams at the Medical Research Council Epidemiology Unit for sample management and Cambridge Genomic Services for genotyping, Sarah Spackman for data management, and the team at the EPIC-CVD Coordinating Centre for study coordination and administration. Where authors are identified as personnel of the International Agency for Research on Cancer / World Health Organization, the authors alone are responsible for the views expressed in this article and they do not necessarily represent the decisions, policy or views of the International Agency for Research on Cancer / World Health Organization.

Funding

This research was funded by the Austrian Science Fund (FWF) [T 1253]. For the purpose of open access, the author has applied a CC BY public copyright licence to any Author Accepted Manuscript version arising from this submission. SAEP is supported by a VIDI Fellowship from the Dutch Organisation for Health Research and Development (ZonMw) (09150172010050). SB is supported by the Wellcome Trust (225790/Z/22/Z) and the United Kingdom Research and Innovation Medical Research Council (MC_UU_00002/7, MC_UU_00040/01). AMW is supported by the BHF Data Science Centre (HDRUK2023.0239) and as an NIHR Research Professor (NIHR303137). This work was supported by core funding from the: British Heart Foundation (RG/18/13/33946), NIHR Cambridge Biomedical Research Centre (BRC-1215-20014; NIHR203312) [*], Cambridge BHF Centre of Research Excellence (RE/18/1/34212), BHF Chair Award (CH/12/2/29428) and by Health Data Research UK, which is funded by the UK Medical Research Council, Engineering and Physical Sciences Research Council, Economic and Social Research Council, Department of Health and Social Care (England), Chief Scientist Office of the Scottish Government Health and Social Care Directorates, Health and Social Care Research and Development Division (Welsh Government), Public Health Agency (Northern Ireland), British Heart Foundation and Wellcome. The EPIC-CVD coordinating centre was supported by core funding from the European Commission Framework Programme 7 (HEALTH-F2-2012-279233), European Research Council (268834), Novartis, UK Medical Research Council (G0800270; MR/L003120/1), British Heart Foundation (SP/09/002; RG/13/13/30194; RG/18/13/33946) and NIHR* Cambridge Biomedical Research Centre (NIHR203312). The establishment of the study sub-cohort was supported by the EU Sixth Framework Programme (FP6) (grant LSHM_CT_2006_037197 to the InterAct project) and the Medical Research Council Epidemiology Unit (grant MC_UU_00006/1). The coordination of EPIC is financially supported by International Agency for Research on Cancer (IARC) and also by the Department of Epidemiology and Biostatistics, School of Public Health, Imperial College London which has additional infrastructure support provided by the NIHR Imperial Biomedical Research Centre (BRC). The national cohorts in EPIC are supported by: Danish Cancer Society (Denmark); Ligue Contre le Cancer, Institut Gustave Roussy, Mutuelle Générale de l’Education Nationale, Institut National de la Santé et de la Recherche Médicale (INSERM) (France); German Cancer Aid, German Cancer Research Center (DKFZ), German Institute of Human Nutrition PotsdamRehbruecke (DIfE), Federal Ministry of Education and Research (BMBF) (Germany); Associazione Italiana per la Ricerca sul Cancro-AIRC-Italy, Compagnia di SanPaolo and National Research Council (Italy); Dutch Ministry of Public Health, Welfare and Sports (VWS), LK Research Funds, Dutch Prevention Funds, Dutch ZON (Zorg Onderzoek Nederland), World Cancer Research Fund (WCRF); Health Research Fund (FIS) - Instituto de Salud Carlos III (ISCIII), Regional Governments of Andalucía, Asturias, Basque Country, Murcia and Navarra, and the Catalan Institute of Oncology - ICO (Spain); Swedish Cancer Society, Swedish Research Council and County Councils of Skåne and Västerbotten (Sweden); Cancer Research UK (C864/A14136 to EPIC-Norfolk; C8221/A29017 to EPIC-Oxford), Medical Research Council (MR/N003284/1, MC-UU_12015/1 and MC_UU_00006/1 to EPIC-Norfolk; MR/M012190/1 to EPIC-Oxford). (United Kingdom). *The views expressed are those of the authors and not necessarily those of the NIHR or the Department of Health and Social Care.

Footnotes

Ethics approval

This work uses data from EPIC-CVD and the UK Biobank. The EPIC-CVD study was approved by the local ethics committees of the participating centres and the Institutional Review Board of the International Agency for Research on Cancer (IARC, Lyon). All EPIC-CVD participants gave written informed consent. The UK Biobank study was approved by the North West Multi-Centre Research Ethics Committee and all participants provided written informed consent.

Competing interests

The authors declare that they have no competing interests.

Authors’ contributions

LT, SAEP, YTvdS, and NCO-M contributed to the conception and design of the work, the statistical analysis, and the drafting of the manuscript. All authors contributed to the interpretation of the data, revised the manuscript critically for important intellectual content, and approved the final manuscript.

Data availability

The data underlying this article can be requested via the UK Biobank website (https://www.ukbiobank.ac.uk/enable-your-research/apply-for-access).

Literature Cited

  • 1.Thomas F, Renaud F, Benefice E, de Meeüs T, Guegan JF. International variability of ages at menarche and menopause: patterns and main determinants. Hum Biol. 2001;73:271–290. doi: 10.1353/hub.2001.0029. [DOI] [PubMed] [Google Scholar]
  • 2.Eckert-Lind C, Busch AS, Petersen JH, Biro FM, Butler G, Bräuner EV, Juul A. Worldwide Secular Trends in Age at Pubertal Onset Assessed by Breast Development Among Girls: A Systematic Review and Meta-analysis. JAMA Pediatr. 2020;174:e195881. doi: 10.1001/jamapediatrics.2019.5881. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3.Fan H-Y, Huang Y-T, Hsieh R-H, Chao JC-J, Tung Y-C, Lee YL, Chen Y-C. Birthweight, time-varying adiposity growth and early menarche in girls: A Mendelian randomisation and mediation analysis. Obes Res Clin Pract. 2018;12:445–451. doi: 10.1016/j.orcp.2018.07.008. [DOI] [PubMed] [Google Scholar]
  • 4.Luijken J, van der Schouw YT, Mensink D, Onland-Moret NC. Association between age at menarche and cardiovascular disease: A systematic review on risk and potential mechanisms. Maturitas. 2017;104:96–116. doi: 10.1016/j.maturitas.2017.07.009. [DOI] [PubMed] [Google Scholar]
  • 5.Magnus MC, Guyatt AL, Lawn RB, Wyss AB, Trajanoska K, Küpers LK, Rivadeneira F, Tobin MD, London SJ, Lawlor DA, Millard LAC, et al. Identifying potential causal effects of age at menarche: a Mendelian randomization phenome-wide association study. BMC Med. 2020;18:71. doi: 10.1186/s12916-020-01515-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Lv Y, Xia X, Lei L, Xiang W, Wu X, Xie S, Li J. Health outcomes of age at menarche in European women: a two-sample Mendelian randomization study. Postgrad Med J. 2023 doi: 10.1093/postmj/qgad023. [DOI] [PubMed] [Google Scholar]
  • 7.Mishra SR, Chung H-F, Waller M, Dobson AJ, Greenwood DC, Cade JE, Giles GG, Bruinsma F, Simonsen MK, Hardy R, Kuh D, et al. Association Between Reproductive Life Span and Incident Nonfatal Cardiovascular Disease: A Pooled Analysis of Individual Patient Data From 12 Studies. JAMA Cardiol. 2020;5:1410–1418. doi: 10.1001/jamacardio.2020.4105. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Zheng J, Chen K, Huang T, Shao C, Li P, Wang J, Wang W, Zhang K, Meng X, Gao J, Wang X, et al. Genetically Determined Lifestyle and Cardiometabolic Risk Factors Mediate the Association of Genetically Predicted Age at Menarche With Genetic Predisposition to Myocardial Infarction: A Two-Step, Two-Sample Mendelian Randomization Study. Front Cardiovasc Med. 2022;9:821068. doi: 10.3389/fcvm.2022.821068. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.Ardissino M, Slob EAW, Carter P, Rogne T, Girling J, Burgess S, Ng FS. Sex-Specific Reproductive Factors Augment Cardiovascular Disease Risk in Women: A Mendelian Randomization Study. J Am Heart Assoc. 2023;12:e027933. doi: 10.1161/JAHA.122.027933. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Au Yeung SL, Jiang C, Cheng KK, Xu L, Zhang W, Lam TH, Leung GM, Schooling CM. Age at menarche and cardiovascular risk factors using Mendelian randomization in the Guangzhou Biobank Cohort Study. Prev Med. 2017;101:142–148. doi: 10.1016/j.ypmed.2017.06.006. [DOI] [PubMed] [Google Scholar]
  • 11.Gill D, Brewer CF, Del Greco MF, Sivakumaran P, Bowden J, Sheehan NA, Minelli C. Age at menarche and adult body mass index: a Mendelian randomization study. Int J Obes (Lond) 2018;42:1574–1581. doi: 10.1038/s41366-018-0048-7. [DOI] [PubMed] [Google Scholar]
  • 12.Chen J, Chen H, Zhu Q, Liu Q, Zhou Y, Li L, Wang Y. Age at menarche and ischemic heart disease: An update mendelian randomization study. Front Genet. 2022;13:942861. doi: 10.3389/fgene.2022.942861. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.von Elm E, Altman DG, Egger M, Pocock SJ, Gøtzsche PC, Vandenbroucke JP. The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: guidelines for reporting observational studies. Epidemiology. 2007;18:800–804. doi: 10.1136/bmj.39335.541782.AD. [DOI] [PubMed] [Google Scholar]
  • 14.Skrivankova VW, Richmond RC, Woolf BAR, Yarmolinsky J, Davies NM, Swanson SA, VanderWeele TJ, Higgins JPT, Timpson NJ, Dimou N, Langenberg C, et al. Strengthening the Reporting of Observational Studies in Epidemiology Using Mendelian Randomization: The STROBE-MR Statement. JAMA. 2021;326:1614–1621. doi: 10.1001/jama.2021.18236. [DOI] [PubMed] [Google Scholar]
  • 15.Sudlow C, Gallacher J, Allen N, Beral V, Burton P, Danesh J, Downey P, Elliott P, Green J, Landray M, Liu B, et al. UK biobank: an open access resource for identifying the causes of a wide range of complex diseases of middle and old age. PLoS Med. 2015;12:e1001779. doi: 10.1371/journal.pmed.1001779. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Riboli E, Hunt KJ, Slimani N, Ferrari P, Norat T, Fahey M, Charrondière UR, Hémon B, Casagrande C, Vignat J, Overvad K, et al. European Prospective Investigation into Cancer and Nutrition (EPIC): study populations and data collection. Public Health Nutr. 2002;5:1113–1124. doi: 10.1079/PHN2002394. [DOI] [PubMed] [Google Scholar]
  • 17.Fortuin-de Smidt MC, Sewe MO, Lassale C, Weiderpass E, Andersson J, Huerta JM, Ekelund U, Aleksandrova K, Tong TY, Dahm CC, Tjønneland A, et al. Physical activity attenuates but does not eliminate coronary heart disease risk amongst adults with risk factors: EPIC-CVD case-cohort study. Eur J Prev Cardiol. 2022;29:1618–1629. doi: 10.1093/eurjpc/zwac055. [DOI] [PubMed] [Google Scholar]
  • 18.Day FR, Thompson DJ, Helgason H, Chasman DI, Finucane H, Sulem P, Ruth KS, Whalen S, Sarkar AK, Albrecht E, Altmaier E, et al. Genomic analyses identify hundreds of variants associated with age at menarche and support a role for puberty timing in cancer risk. Nat Genet. 2017;49:834–841. doi: 10.1038/ng.3841. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Zheng J-S, Luan J, Sofianopoulou E, Imamura F, Stewart ID, Day FR, Pietzner M, Wheeler E, Lotta LA, Gundersen TE, Amiano P, et al. Plasma Vitamin C and Type 2 Diabetes: Genome-Wide Association Study and Mendelian Randomization Analysis in European Populations. Diabetes Care. 2021;44:98–106. doi: 10.2337/dc20-1328. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Collins R. What makes UK Biobank special? Lancet. 2012;379:1173–1174. doi: 10.1016/S0140-6736(12)60404-8. [DOI] [PubMed] [Google Scholar]
  • 21.Bycroft C, Freeman C, Petkova D, Band G, Elliott LT, Sharp K, Motyer A, Vukcevic D, Delaneau O, O’Connell J, Cortes A, et al. Genome-wide genetic data on ∼500,000 UK Biobank participants. 2017 [Google Scholar]
  • 22.van Buuren S, Groothuis-Oudshoorn K. mice : Multivariate Imputation by Chained Equations in R. J Stat Soft. 2011;45 [Google Scholar]
  • 23.Firth D. Quasi-variances. Biometrika. 2004;91:65–80. [Google Scholar]
  • 24.Onland-Moret NC, van der A DL, van der Schouw YT, Buschers W, Elias SG, van Gils CH, Koerselman J, Roest M, Grobbee DE, Peeters PHM. Analysis of case-cohort data: a comparison of different methods. J Clin Epidemiol. 2007;60:350–355. doi: 10.1016/j.jclinepi.2006.06.022. [DOI] [PubMed] [Google Scholar]
  • 25.White IR. Multivariate Random-effects Meta-analysis. The Stata Journal. 2009;9:40–56. [Google Scholar]
  • 26.Staley JR, Burgess S. Semiparametric methods for estimation of a nonlinear exposure-outcome relationship using instrumental variables with application to Mendelian randomization. Genet Epidemiol. 2017;41:341–352. doi: 10.1002/gepi.22041. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Mason AM, Burgess S. Software Application Profile: SUMnlmr, an R package that facilitates flexible and reproducible non-linear Mendelian randomization analyses. Int J Epidemiol. 2022;51:2014–2019. [Google Scholar]
  • 28.Tian H, Mason AM, Liu C, Burgess S. Relaxing parametric assumptions for non-linear Mendelian randomization using a doubly-ranked stratification method. PLoS Genet. 2023;19:e1010823. doi: 10.1371/journal.pgen.1010823. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Hamilton FW, Hughes DA, Spiller W, Tilling K, Smith GD. Non-linear mendelian randomization: evaluation of biases using negative controls with a focus on BMI and Vitamin D. medRxiv. 2023 doi: 10.1007/s10654-024-01113-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.Yavorska OO, Burgess S. MendelianRandomization: an R package for performing Mendelian randomization analyses using summarized data. Int J Epidemiol. 2017;46:1734–1739. doi: 10.1093/ije/dyx034. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31.Verbanck M, Chen C-Y, Neale B, Do R. Detection of widespread horizontal pleiotropy in causal relationships inferred from Mendelian randomization between complex traits and diseases. Nat Genet. 2018;50:693–698. doi: 10.1038/s41588-018-0099-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.Zou X, Wang L, Wang S, Le Zhang. Mendelian randomization study and meta-analysis exploring the causality of age at menarche and the risk of intracerebral hemorrhage and ischemic stroke. CNS Neurosci Ther. 2023 doi: 10.1111/cns.14245. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33.Warner ET, Jiang L, Adjei DN, Turman C, Gordon W, Wang L, Tamimi R, Kraft P, Lindström S. A Genome-Wide Association Study of Childhood Body Fatness. Obesity (Silver Spring) 2021;29:446–453. doi: 10.1002/oby.23070. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34.Richardson TG, Sanderson E, Elsworth B, Tilling K, Davey Smith G. Use of genetic variation to separate the effects of early and later life adiposity on disease risk: mendelian randomisation study. BMJ. 2020;369:m1203. doi: 10.1136/bmj.m1203. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35.Dong S-S, Zhang K, Guo Y, Ding J-M, Rong Y, Feng J-C, Yao S, Hao R-H, Jiang F, Chen J-B, Wu H, et al. Phenome-wide investigation of the causal associations between childhood BMI and adult trait outcomes: a two-sample Mendelian randomization study. Genome Med. 2021;13:48. doi: 10.1186/s13073-021-00865-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36.Rozé C, Doyen C, Le Heuzey M-F, Armoogum P, Mouren M-C, Léger J. Predictors of late menarche and adult height in children with anorexia nervosa. Clin Endocrinol (Oxf) 2007;67:462–467. doi: 10.1111/j.1365-2265.2007.02912.x. [DOI] [PubMed] [Google Scholar]
  • 37.Campisi SC, Humayun KN, Rizvi A, Lou W, Bhutta ZA. Later puberty onset among chronically undernourished adolescents living in a Karachi slum, Pakistan. Acta Paediatr. 2020;109:1019–1025. doi: 10.1111/apa.15053. [DOI] [PubMed] [Google Scholar]
  • 38.Sachs KV, Harnke B, Mehler PS, Krantz MJ. Cardiovascular complications of anorexia nervosa: A systematic review. Int J Eat Disord. 2016;49:238–248. doi: 10.1002/eat.22481. [DOI] [PubMed] [Google Scholar]
  • 39.Du R, Zheng R, Xu Y, Zhu Y, Yu X, Li M, Tang X, Hu R, Su Q, Wang T, Zhao Z, Xu M, et al. Early-Life Famine Exposure and Risk of Cardiovascular Diseases in Later Life: Findings From the REACTION Study. J Am Heart Assoc. 2020;9:e014175. doi: 10.1161/JAHA.119.014175. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 40.Yermachenko A, Dvornyk V. Nongenetic determinants of age at menarche: a systematic review. Biomed Res Int. 2014;2014:371583. doi: 10.1155/2014/371583. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 41.Cao M, Cui B. Negative Effects of Age at Menarche on Risk of Cardiometabolic Diseases in Adulthood: A Mendelian Randomization Study. J Clin Endocrinol Metab. 2020;105 doi: 10.1210/clinem/dgz071. [DOI] [PubMed] [Google Scholar]
  • 42.Larsson SC, Burgess S. Causal role of high body mass index in multiple chronic diseases: a systematic review and meta-analysis of Mendelian randomization studies. BMC Med. 2021;19:320. doi: 10.1186/s12916-021-02188-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 43.Fan H-Y, Huang Y-T, Chen Y-Y, Hsu JB, Li H-Y, Su T-C, Lin H-J, Chien K-L, Chen Y-C. Systolic blood pressure as the mediator of the effect of early menarche on the risk of coronary artery disease: A Mendelian randomization study. Front Cardiovasc Med. 2022;9:1023355. doi: 10.3389/fcvm.2022.1023355. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 44.Werneck AO, Oyeyemi AL, Cyrino ES, Ronque ERV, Szwarcwald CL, Coelho-E-Silva MJ, Silva DR. Association between age at menarche and blood pressure in adulthood: is obesity an important mediator? Hypertens Res. 2018;41:856–864. doi: 10.1038/s41440-018-0079-4. [DOI] [PubMed] [Google Scholar]
  • 45.Zhang L, Li Y, Zhou W, Wang C, Dong X, Mao Z, Huo W, Tian Z, Fan M, Yang X, Li L. Mediation effect of BMI on the relationship between age at menarche and hypertension: The Henan Rural Cohort Study. J Hum Hypertens. 2020;34:448–456. doi: 10.1038/s41371-019-0247-2. [DOI] [PubMed] [Google Scholar]
  • 46.Zhang L, Li Y, Wang C, Mao Z, Zhou W, Tian Z, Dong X, Zhang H, Yang X, Fan M, Li L. Early menarche is associated with an increased risk of type 2 diabetes in rural Chinese women and is partially mediated by BMI: the Henan Rural Cohort Study. Menopause. 2019;26:1265–1271. doi: 10.1097/GME.0000000000001385. [DOI] [PubMed] [Google Scholar]
  • 47.Kim MS, Kim WJ, Khera AV, Kim JY, Yon DK, Lee SW, Shin JI, Won H-H. Association between adiposity and cardiovascular outcomes: an umbrella review and meta-analysis of observational and Mendelian randomization studies. Eur Heart J. 2021;42:3388–3403. doi: 10.1093/eurheartj/ehab454. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 48.Jeong S-M, Yoo JE, Jeon KH, Han K, Lee H, Lee D-Y, Shin DW. Associations of reproductive factors with incidence of myocardial infarction and ischemic stroke in postmenopausal women: a cohort study. BMC Med. 2023;21:64. doi: 10.1186/s12916-023-02757-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 49.Jeong S-M, Jeon KH, Jung W, Yoo JE, Yoo J, Han K, Kim JY, Lee D-Y, Lee Y-B, Shin DW. Association of reproductive factors with cardiovascular disease risk in pre-menopausal women: nationwide population-based cohort study. Eur J Prev Cardiol. 2023;30:264–273. doi: 10.1093/eurjpc/zwac265. [DOI] [PubMed] [Google Scholar]
  • 50.Must A, Phillips SM, Naumova EN, Blum M, Harris S, Dawson-Hughes B, Rand WM. Recall of early menstrual history and menarcheal body size: after 30 years, how well do women remember? Am J Epidemiol. 2002;155:672–679. doi: 10.1093/aje/155.7.672. [DOI] [PubMed] [Google Scholar]
  • 51.IPAQ Research Committee. Guidelines for data processing analysis of the International Physical Activity Questionnaire (IPAQ) - Short and long forms. 2005. https://sites.google.com/view/ipaq/score .
  • 52.Peters T, Brage S, Westgate K, Franks PW, Gradmark A, Tormo Diaz MJ, Huerta JM, Bendinelli B, Vigl M, Boeing H, Wendel-Vos W, et al. Validity of a short questionnaire to assess physical activity in 10 European countries. Eur J Epidemiol. 2012;27:15–25. doi: 10.1007/s10654-011-9625-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 53.Burgess S, White IR, Resche-Rigon M, Wood AM. Combining multiple imputation and meta-analysis with individual participant data. Stat Med. 2013;32:4499–4514. doi: 10.1002/sim.5844. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary Material
EMS207595-supplement-Supplementary_Material.docx (1MB, docx)

Data Availability Statement

The data underlying this article can be requested via the UK Biobank website (https://www.ukbiobank.ac.uk/enable-your-research/apply-for-access).

RESOURCES