2018 Volume 82 Issue 3 Pages 708-714
Background: There is sparse research on whether if early menarche is related to left ventricular (LV) diastolic dysfunction. The present study examined this relationship in Korean women.
Methods and Results: In a cross-sectional study we analyzed the records of 18,910 Korean women (≥30 years) who underwent echocardiography as part of a comprehensive health examination. Age at menarche was assessed using standardized, self-administered questionnaires. Presence of LV diastolic dysfunction was determined from the echocardiographic findings. Of the 18,910 women, 3,449 had LV diastolic dysfunction. Age at menarche was inversely associated with prevalence of LV diastolic dysfunction. In a multivariable-adjusted model, odds ratios (95% confidence interval) for LV diastolic dysfunction comparing menarche age to menarche at 15–18 years were 1.77 (1.38–2.27) for <12 years, 1.31 (1.11–1.54) for 12 years, 1.26 (1.11–1.43) for 13 years, and 1.03 (0.91–1.15) for 14 years (P for trend <0.001). Adjusting for body mass index or percent fat mass partially reduced these associations.
Conclusions: This large study found an inverse relationship between menarche age and LV diastolic dysfunction. Future prospective studies are needed to investigate potential causal relationships.
Heart failure is a major public health problem with a prevalence of more than 23 million people worldwide. It is related to significant morbidity, mortality and healthcare expenditures.1–4 Previous studies have demonstrated that subclinical left ventricular (LV) diastolic dysfunction predicts future development of clinical heart failure.5–8 Furthermore, LV diastolic dysfunction is related to increased risk of sudden death.9
Menarche represents a significant event in the reproductive life of a woman and average age at menarche has declined in recent years.10,11 Early menarche has received attention because of its implications for women’s health.12 Epidemiological studies have reported relationships between early menarche and obesity, metabolic syndrome (MetS), insulin resistance (IR), diabetes, nonalcoholic fatty liver disease, gallstones, cardiovascular disease, stroke, some types of cancer, and mortality.12–24 Recent epidemiological studies suggest a close link between LV diastolic dysfunction and metabolic disorders such as MetS, IR, and obesity,25,26 all of which are likely to develop in women who experience early menarche. Therefore, we hypothesized that the metabolic milieu of women who undergo early menarche may trigger pathophysiological processes associated with LV diastolic dysfunction. However, little research has examined if early menarche is associated with LV diastolic dysfunction.
Therefore, the purpose of this study was to examine the relationship between age at menarche and LV diastolic dysfunction in a large sample of Korean women. We explored whether observed associations were mediated by adiposity or IR.
The Kangbuk Samsung Health Study was a cohort study of Korean men and women who underwent a comprehensive annual or biennial examination at the Kangbuk Samsung Hospital Total Healthcare Center in Seoul and Suwon, South Korea.19,24 The study population was the subset of female participants (age ≥30 years) who underwent an echocardiography as part of their comprehensive health examination between March 2011 and December 2014 (n=23,417). Analysis was performed in 2016.
We excluded 4,507 participants for history of a malignancy (n=1,155); history of cardiovascular disease (n=194) based on standardized, self-administered questionnaires: systolic heart failure (ejection fraction <50%), hypertrophic or dilated cardiomyopathy, ischemic cardiomyopathy, postoperation or valvular replacement status, mitral valve stenosis or regurgitation (moderate to severe), aortic valve stenosis or regurgitation (moderate to severe), atrial fibrillation and congenital heart disease based on echocardiographic findings (n=1,273); missing data for age at menarche (n=2,038); or reported age at menarche >18 years, which may reflect an underlying pathological issue (n=207). Because some exclusion criteria overlapped, 18,910 patients were eligible. This study was approved by the Institutional Review Board of Kangbuk Samsung Hospital. As we only used nonidentified retrospective data routinely collected during the health screening process, the requirement for informed consent was waived.
MeasurementsAll examinations were conducted at Kangbuk Samsung Hospital Total Healthcare Center clinics in Seoul and Suwon. Data on demographic characteristics, medical history, medication use, education level, smoking status, and alcohol intake were collected by standardized, self-administered questionnaires as previously described.19,24 Information regarding reproductive factors was collected using self-administered, standardized questionnaires that asked about menopausal status, frequency and regularity of menstrual periods, and use of hormone replacement therapy (HRT) and oral contraceptives. Age at menarche was assessed by answers to the question, “At what age did your menstrual periods begin?” as previously described.19,24 We categorized age at menarche as <12, 12, 13, 14 and 15–18 years. We defined early menarche as onset of menstruation occurring before age 12.27 Parity was assessed from the number of reported live and still births. Physical activity levels were measured using the validated Korean version of the International Physical Activity Questionnaire Short Form and classified into 3 categories: inactive, minimally active, and health-enhancing physically active.28,29
Trained nurses measured height, weight, and body composition while participants wore a lightweight hospital gown and no shoes. Body mass index (BMI) was calculated as weight (kg) divided by height (m) squared (kg/m2) and classified according to criteria proposed for Asian populations.30 Percent fat mass was estimated using a multifrequency bioimpedance analyzer with 8-point tactile electrodes (InBody 720, Biospace Co., Seoul, South Korea), validated for reproducibility and accuracy for body composition.31 Blood pressure (BP) was measured using an automated oscillometric device (Welch Allyn 5300P) while participants sat with the arm supported at the same height as the heart. Hypertension was defined as systolic BP ≥140 mmHg, diastolic BP ≥90 mmHg, or current use of antihypertensive medication. The Laboratory Medicine Department at Kangbuk Samsung Hospital in Seoul, South Korea, is accredited by the Korean Society of Laboratory Medicine and the Korean Association of Quality Assurance for Clinical Laboratories; the laboratory participates in College of American Pathologists Survey Proficiency Testing. Measurement of serum biochemical parameters, including total cholesterol, triglycerides, alanine aminotransferase (ALT), low-density lipoprotein-cholesterol (LDL-C), high-density lipoprotein-cholesterol (HDL-C), glucose, insulin and hemoglobin A1c is described in detail elsewhere.19,24 IR was assessed with the homeostatic model assessment-IR (HOMA-IR) equation: fasting blood insulin (uIU/mL)×fasting blood glucose (mmol/L)/22.5. Diabetes was defined as a fasting serum glucose ≥126 mg/dL, hemoglobin A1c ≥6.5%, or current use of insulin or antidiabetic medications.
EchocardiographyConventional echocardiography was performed with ultrasound scanners (in Vivid 7 and E9, General Electric, Milwaukee, WI, USA) by trained sonographers and measurements used a standardized guideline.32 Linear measurements of left posterior wall thickness (PWT), intraventricular septum thicknesses (IVST), and LV cavity diameters at the end of diastole and systole were obtained with M-mode in the parasternal long-axis view. LV mass (LVM) was calculated using the following equation with measurements obtained in M-mode: LVM=0.8(1.04[LVEDD+IVST+PWT]3−LVEDD3)+0.6 g.33,34 LVM index (LVMI) was calculated as LVM/height2.7 and LV hypertrophy was defined as LVMI ≥45 g/m2.7.32 The anteroposterior (AP) diameter of the left atrium (LA) was measured for all participants. LA volume was estimated by the biplane area-length method if linear measurement of LA diameter was >40 mm.
To assess diastolic function, we sampled pulse-wave Doppler transmitral LV inflow in an apical 4-chamber view. Early diastolic mitral inflow peak velocity (E), late diastolic peak velocity (A) during atrial contraction, and the deceleration time of E velocity were measured. Early (E’) and late (A’) the tissue velocities were measured from tissue Doppler imaging of the septal mitral annulus. LV diastolic dysfunction was primarily defined as decreased E’ (<8 cm/s). We further graded diastolic dysfunction using E/E’, which reflects LV filling pressure and was available for all subjects. The cutoff values of E’ and E/E’ were derived from recent American Society of Echocardiography guidelines.35,36 Grade 1 diastolic dysfunction was defined as decreased E’ (<8 cm/s) plus decreased E/E’ (<10). Grade 3 diastolic dysfunction was defined as decreased E’ (<8 cm/s) and E/E’ >15. Remaining cases were categorized as grade 2 diastolic dysfunction.
Statistical AnalysisCharacteristics of study participants were explored by age at menarche, divided into the following groups: <12 (early menarche), 12, 13, 14, and 15–18 years (reference category). Age-adjusted estimated mean values (95% confidence interval [CI]) of echocardiographic characteristics were explored by categorical age at menarche. Category numbers were used as continuous variables in regression models to test for linear trends.
To evaluate the relationship of LV diastolic dysfunction across categorical age at menarche, we used a logistic regression model to estimate odds ratios with 95% CIs for LV diastolic dysfunction comparing categories of age at menarche to the reference category. We used 3 models with progressive adjustment for confounding variables. We initially adjusted for age, and further adjusted for study center, year of screening exam, alcohol intake, smoking, physical activity level, and educational level. We next included the reproductive factors parity, menopausal status, use of current oral contraceptives, and use of HRT. To assess if the relationship between menarche age and prevalence of LV diastolic dysfunction was mediated by BMI, percent fat mass, BP, HOMA-IR or lipid profiles, we included these variables in multivariable models.
We conducted a series of sensitivity analyses to test the robustness of our primary results. First, we examined if menarche age was associated with LV diastolic dysfunction after further excluding participants with hypertension or diabetes. Second, we also analyzed the records of 11,819 Korean women, aged 30 years or older, who underwent brachial-ankle pulse-wave velocity as part of a comprehensive health examination from March 2011 to December 2014 and examined the association between age at menarche and brachial-ankle pulse-wave velocity. Third, we made age-matched comparisons after excluding postmenopausal women. Finally, in order to examine the relationship between age at menarche and LV diastolic dysfunction grade, we also fit multinomial logistic regression models as outcomes categorized as LV diastolic dysfunction grade 1 and grades 2–3.
We performed stratified analyses in prespecified subgroups. Interactions between subgroups were tested using likelihood ratio tests comparing models with and without multiplicative interaction terms. All P-values were 2-tailed, and values of P<0.05 were considered statistically significant. We used STATA version 14.2 (Stata Corp., College Station, TX, USA) for data analysis.
Baseline characteristics of participants by categorical age at menarche are shown in Table 1. For the 18,910 women, mean age was 41.0 years (SD, 8.1), age at menarche was 13.9 years (SD, 1.5) and BMI was 22.0 kg/m2 (SD, 3.1). Prevalences were 2.7% for diabetes, 7.0% for hypertension, 1.6% for current smoking and 14.8% for obesity. Clear dose-response relationships were observed between some of the variables and age at menarche. Specifically, women with a younger menarche age were more likely to be younger at baseline. Systolic and diastolic BPs, total cholesterol, triglycerides, LDL-C, glucose, ALT, physical activity and parity were positively related to age at menarche. BMI, percent fat mass, HDL-C, HOMA-IR and education level were negatively related to age at menarche. The proportion of women with diabetes or hypertension or who used HRT also correlated positively with age at menarche. In an age-adjusted model, E/A ratio, septal E’ and septal A were positively related to age at menarche. Heart rate, E/E’ ratio, LVM and LVMI were negatively related to age at menarche (Table 2).
| Characteristics | Overall | Age at menarche (years) | P for trend |
||||
|---|---|---|---|---|---|---|---|
| <12 | 12 | 13 | 14 | 15–18 | |||
| No. of participants | 18,910 | 867 | 2,529 | 4,223 | 4,586 | 6,705 | |
| Age (years) | 41.0 (8.1) | 36.5 (5.4) | 37.6 (5.8) | 39.2 (6.7) | 40.9 (7.3) | 44.0 (9.4) | <0.001 |
| BMI (kg/m2)d | 22.0 (3.1) | 22.4 (3.2) | 22.1 (3.2) | 22.0 (3.2) | 21.8 (3.1) | 22.0 (3.0) | <0.001 |
| Percent fat mass (%)d | 29.7 (5.9) | 30.5 (6.0) | 29.9 (5.9) | 29.7 (5.9) | 29.4 (5.9) | 29.6 (6.0) | <0.001 |
| Weight at age 20 (kg)c,d | 50.6 (5.4) | 51.8 (5.7) | 51.4 (5.6) | 50.7 (5.3) | 50.3 (5.2) | 50.1 (5.2) | <0.001 |
| Obesity (%) | 14.8 | 18.4 | 14.9 | 15.0 | 13.8 | 14.7 | 0.06 |
| Current smoker (%) | 1.6 | 1.8 | 1.3 | 1.4 | 1.6 | 1.8 | 0.16 |
| Alcohol intake (%)a | 11.8 | 11.5 | 9.7 | 11.2 | 11.8 | 13.0 | <0.001 |
| HEPA (%) | 14.6 | 11.3 | 13.0 | 13.3 | 15.1 | 16.2 | <0.001 |
| Education level (%)b | 74.7 | 92.2 | 89.2 | 82.7 | 75.4 | 61.3 | <0.001 |
| Parity ever (%) | 87.3 | 80.0 | 82.1 | 85.9 | 88.7 | 90.3 | <0.001 |
| Menopause (%) | 14.6 | 4.9 | 6.9 | 9.0 | 12.9 | 23.7 | <0.001 |
| Use of current HRT (%) | 5.0 | 4.6 | 3.5 | 4.6 | 5.2 | 5.7 | <0.001 |
| Use of current oral contraceptive (%) |
0.4 | 0.1 | 0.5 | 0.7 | 0.5 | 0.3 | 0.07 |
| Diabetes (%) | 2.7 | 0.9 | 2.1 | 2.0 | 2.5 | 3.8 | <0.001 |
| Hypertension (%) | 7.0 | 4.2 | 4.2 | 5.5 | 5.9 | 10.3 | <0.001 |
| Systolic BP (mmHg)d | 102.3 (12.0) | 99.9 (11.2) | 100.6 (11.2) | 101.3 (11.2) | 102.0 (11.8) | 103.7 (12.8) | <0.001 |
| Diastolic BP (mmHg)d | 65.5 (9.0) | 64.1 (8.8) | 64.5 (8.7) | 65.1 (8.7) | 65.5 (9.0) | 66.3 (9.2) | <0.001 |
| Glucose (mg/dL)d | 92.0 (12.7) | 89.7 (9.1) | 91.2 (13.5) | 91.5 (12.5) | 92.0 (12.3) | 92.9 (13.1) | <0.001 |
| Total cholesterol (mg/dL)d | 190.2 (33.8) | 189.0 (32.2) | 188.5 (32.1) | 188.6 (33.4) | 190.4 (33.5) | 191.8 (35.1) | <0.001 |
| LDL-C (mg/dL)d | 113.0 (30.4) | 111.3 (28.5) | 110.6 (28.4) | 111.4 (29.7) | 112.9 (30.1) | 115.1 (31.8) | <0.001 |
| HDL-C (mg/dL)d | 64.4 (15.1) | 65.4 (14.9) | 65.1 (15.1) | 64.8 (15.0) | 64.8 (15.1) | 63.5 (15.0) | <0.001 |
| Triglycerides (mg/dL)e | 75 (57–103) | 74 (55–99) | 73 (56–100) | 73 (57–100) | 75 (57–103) | 77 (57–107) | <0.001 |
| ALT (u/L)e | 13 (10–17) | 12.5 (10–17) | 12 (10–17) | 13 (10–17) | 13 (10–17) | 14 (11–18) | <0.001 |
| hsCRP (mg/L)e | 0.3 (0.2–0.7) | 0.4 (0.2–0.9) | 0.4 (0.2–0.8) | 0.3 (0.2–0.7) | 0.3 (0.2–0.7) | 0.3 (0.2–0.7) | 0.011 |
| HOMA-IRe | 1.09 (0.74–1.61) |
1.12 (0.75–1.65) |
1.11 (0.75–1.64) |
1.12 (0.75–1.63) |
1.09 (0.73–1.61) |
1.07 (0.73–1.58) |
0.001 |
a≥10 g of ethanol per day. b≥College graduate. cAmong 16,576 participants for whom information on body weight at 20 years of age was available. Data are dmean (standard deviation), emedian (interquartile range), or percentage. ALT, alanine aminotransferase; BMI, body mass index; BP, blood pressure; HDL-C, high-density lipoprotein-cholesterol; HEPA, health-enhancing physical activity; hsCRP, high-sensitivity C-reactive protein; HOMA-IR, homeostasis model assessment of insulin resistance; HRT, hormone replacement therapy; LDL-C, low-density lipoprotein-cholesterol.
| Characteristics | Age at menarche (years) | P for trend |
||||
|---|---|---|---|---|---|---|
| <12 | 12 | 13 | 14 | 15–18 | ||
| Heart rate (beats/min) | 65.0 (64.4–65.6) | 65.1 (64.7–65.4) | 65.1 (64.8–65.3) | 65.0 (654.7–65.2) | 64.6 (64.4–64.8) | 0.01 |
| Ejection fraction (%) | 67.2 (66.8–67.6) | 67.4 (67.2–67.6) | 67.3 (67.1–67.4) | 67.2 (67.1–67.4) | 67.2 (67.1–67.4) | 0.51 |
| E (cm/s) | 78.8 (77.8–79.7) | 78.2 (77.6–78.8) | 78.0 (77.6–78.5) | 78.0 (77.6–78.4) | 77.8 (77.4–78.1) | 0.17 |
| A (cm/s) | 56.7 (43.1–70.3) | 56.1 (48.2–64.1) | 61.2 (55.1–67.3) | 54.2 (48.4–60.1) | 57.7 (52.7–62.6) | 0.86 |
| E/E’ | 7.96 (7.86–8.07) | 7.82 (7.75–7.88) | 7.75 (7.71–7.80) | 7.67 (7.62–7.72) | 7.71 (7.67–7.75) | <0.001 |
| E/A ratio | 1.49 (1.46–1.51) | 1.47 (1.46–1.49) | 1.50 (1.49–1.51) | 1.52 (1.50–1.53) | 1.52 (1.51–1.53) | <0.001 |
| Septal E’ (cm/s) | 10.2 (8.1–12.4) | 10.8 (9.6–12.1) | 10.6 (9.7–11.6) | 11.7 (10.7–12.6) | 11.9 (11.1–12.7) | <0.001 |
| Septal A’ (cm/s) | 7.6 (6.4–8.9) | 7.9 (7.2–8.7) | 7.8 (7.2–8.3) | 8.0 (7.5–8.6) | 8.4 (7.9–8.9) | 0.02 |
| LVEDD (mm) | 46.5 (46.2–46.7) | 46.5 (46.3–46.6) | 46.4 (46.3–46.6) | 46.4 (46.3–46.5) | 46.4 (46.3–46.5) | 0.44 |
| LVM (g) | 107.4 (105.9–108.9) |
107.7 (106.8–108.6) |
107.3 (106.6–108.0) |
106.1 (105.4–106.7) |
106.2 (105.6–106.8) |
<0.001 |
| LVMI (g/ht2.7, g/m2.7) | 30.7 (0.3–31.1) | 30.5 (30.2–30.7) | 30.2 (30.0–30.4) | 29.7 (29.5–29.9) | 29.8 (29.6–30.0) | <0.001 |
| LVMI (g/BSA, g/m2) | 67.0 (66.2–67.9) | 67.5 (67.0–68.0) | 67.4 (67.1–67.8) | 67.1 (66.7–67.4) | 67.3 (67.0–67.6) | 0.80 |
| LA diameter (mm) | 20.1 (19.9–20.2) | 20.1 (20.0–20.2) | 20.0 (19.9–20.1) | 20.0 (20.0–20.1) | 20.0 (20.0–20.1) | 0.20 |
Adjusted for age. Early diastolic mitral inflow peak velocity (E), late diastolic peak velocity (A) during atrial contraction and deceleration time of the E velocity were measured. Early (E’) and late (A’) tissue velocities were measured from tissue Doppler imaging of septal mitral annuls. BSA, body surface area; LV, left ventricular; LVEDD, left ventricular end-diastolic diameter; LVMI, left ventricular mass index.
The association between age at menarche and LV diastolic dysfunction is presented in Table 3. Of the 18,910 women, 3,449 had LV diastolic dysfunction. In particular, 1,256 women (6.6%) had grade 2 diastolic dysfunction, and 77 women (0.4%) had grade 3 diastolic dysfunction. Age at menarche was inversely related to the prevalence of LV diastolic dysfunction. In a multivariable model adjusting for screening center, year of screening exam, smoking, physical activity, alcohol intake, and educational level, odds ratios (95% CIs) for LV diastolic dysfunction comparing menarche age to menarche at 15–18 years were 1.78 (1.38–2.81) for <12 years (early menarche), 1.31 (1.11–1.54) for 12 years, 1.26 (1.11–1.43) for 13 years, and 1.03 (0.91–1.15) for 14 years (P for trend <0.001). After further adjustment for reproductive factors, age at menarche was still inversely related to prevalence of LV diastolic dysfunction with corresponding odds ratios of 1.77 (1.38–2.27) for <12 years, 1.34 (1.11–1.54) for 12 years, 1.26 (1.11–1.43) for 13 years, and 1.03 (0.91–1.15) for 14 years (P for trend <0.001).
| Age at menarche (years) | P for trend |
|||||
|---|---|---|---|---|---|---|
| <12 | 12 | 13 | 14 | 15–18 | ||
| n | 867 | 2,529 | 4,223 | 4,586 | 6,705 | |
| Cases of LV diastolic dysfunction (%) |
95 (11.0) | 272 (10.8) | 604 (14.3) | 759 (16.6) | 1,719 (25.6) | |
| Grade 1 | 62 (7.2) | 179 (7.1) | 402 (9.5) | 482 (10.5) | 991 (14.8) | |
| Grade 2 | 32 (3.7) | 88 (3.5) | 195 (4.6) | 268 (5.8) | 673 (10.0) | |
| Grade 3 | 1 (0.1) | 5 (0.2) | 7 (0.2) | 9 (0.2) | 55 (0.8) | |
| Age-adjusted OR | 1.46 (1.14–1.87) | 1.11 (0.94–1.30) | 1.13 (1.00–1.28) | 0.96 (0.85–1.07) | 1.00 (Ref.) | 0.003 |
| Multivariate-adjusted ORa | ||||||
| Model 1 | 1.78 (1.38–2.81) | 1.31 (1.11–1.54) | 1.26 (1.11–1.43) | 1.03 (0.91–1.15) | 1.00 (Ref.) | <0.001 |
| Model 2 | 1.77 (1.38–2.27) | 1.31 (1.11–1.54) | 1.26 (1.11–1.43) | 1.03 (0.91–1.15) | 1.00 (Ref.) | <0.001 |
aEstimated from logistic regression. Multivariable model 1 was adjusted for age, center, year of screening examination, smoking status, alcohol intake, physical activity level and educational level. Model 2 was model 1 plus adjustment for parity, use of HRT, use of oral contraceptives, menopausal status. CI, confidence interval; OR, odds ratio. Other abbreviations as in Tables 1,2.
To explore if the association of age at menarche with LV diastolic dysfunction was mediated by adiposity, IR, or metabolic profile including glucose, LDL-C, HDL-C, triglycerides, and systolic BP, we performed additional analyses adjusting for adult BMI, percent fat mass, HOMA-IR, glucose, LDL-C, HDL-C, triglycerides, and systolic BP. Our results did not qualitatively change after further adjustment for HOMA-IR or metabolic profile. Adjusting for adult BMI and percent fat mass partially reduced the identified associations, which still remained significant (Table 4).
| Age at menarche (years) | P for trend |
|||||
|---|---|---|---|---|---|---|
| <12 | 12 | 13 | 14 | 15–18 | ||
| Model 1: ORa (95% CI) | 1.38 (1.06–1.79) | 1.11 (0.94–1.31) | 1.12 (0.99–1.28) | 0.98 (0.87–1.11) | 1.00 (Ref.) | 0.011 |
| Model 2: ORa (95% CI) | 1.46 (1.13–1.88) | 1.15 (0.97–1.36) | 1.14 (1.00–1.30) | 0.99 (0.88–1.11) | 1.00 (Ref.) | 0.002 |
| Model 3: ORa (95% CI) | 1.65 (1.28–2.13) | 1.18 (1.00–1.40) | 1.20 (1.05–1.36) | 0.99 (0.88–1.11) | 1.00 (Ref.) | <0.001 |
| Model 4: ORa (95% CI) | 1.64 (1.27–2.13) | 1.22 (1.03–1.45) | 1.23 (1.08–1.40) | 1.00 (0.89–1.13) | 1.00 (Ref.) | <0.001 |
aEstimated from logistic regression. Multivariable model 1 was adjusted for age, center, year of screening examination, smoking status, alcohol intake, physical activity level, educational level, parity, use of HRT, use of oral contraceptives, menopausal status and BMI at recruitment. Model 2 was model 1 plus adjustment for percent fat instead of BMI at recruitment. Model 3 was model 1 plus adjustment for HOMA-IR instead of BMI at recruitment. Model 4 was model 1 plus adjustment for glucose, LDL-C, HDL-C triglyceride and systolic BP instead of BMI at recruitment. Abbreviations as in Tables 1,3.
The primary findings of the association of age at menarche with LV diastolic dysfunction were further corroborated by the results from several sensitivity analyses: First, age at menarche was significantly related to LV diastolic dysfunction even after excluding participants with hypertension or diabetes (Table S1). Second, we observed an inverse association between age at menarche and brachial-ankle pulse-wave velocity (P for trend 0.009) (Table S2). Third, after excluding postmenopausal women, 1,519 women with LV diastolic dysfunction and 1,519 age-matched controls were reanalyzed with a conditional logistic regression model. Results from age-matched comparisons were nearly the same as the primary findings (Table S3). Finally, the results from the multinomial logistic regression models were consistent with the binomial logistic regression models. Age at menarche was inversely associated with the prevalence ratios for LV diastolic dysfunction grade 1 and grades 2–3 (Table S4).
The relationship between age at menarche and LV diastolic dysfunction was similar across participant subgroups with no significant interactions between menopause (no vs. yes), age (<50 vs. ≥50 years), parity (never vs. ever), smoking status (never smoker vs. ex- or current smoker), alcohol intake (<10 vs. ≥10 g of alcohol per day), or BMI (<25 vs. ≥25 kg/m2) (data not shown).
This study was performed on a large sample of middle-aged women in whom we observed an inverse, dose-response association between age at menarche and LV diastolic dysfunction. Associations persisted even after multivariable adjustment for potential confounders, including reproductive factors.
A growing number of epidemiological studies suggest a relationship between age at menarche and adverse outcomes, including MetS, IR, and risk factors for cardiovascular disease.12–23 Specifically, several cohort studies reported that age at menarche is inversely related to diabetes, cardiovascular disease events, and cardiovascular disease mortality, and these associations are mediated by increased adiposity.15,37–39 However, little research has examined if age at menarche is related to LV diastolic dysfunction. This study showed an inverse association of age at menarche with prevalence of LV diastolic dysfunction. The Multi-ethnic Study of Atherosclerosis showed that subclinical atherosclerosis was associated with regional myocardial dysfunction.40–42 A recent study also reported a strong relationship between global pulse-wave velocity and LV diastolic dysfunction and hypertrophy in postmenopausal women.43 Under this background, we examined the association between age at menarche and brachial-ankle pulse-wave velocity as a sensitivity analysis. Results were nearly the same as the association between age at menarche and LV diastolic dysfunction. The relationship between early menarche and LV diastolic dysfunction has implications for public health and clinical research because of evidence showing that average age at menarche has declined in recent years. This decline coincides with an increasing health burden of heart failure.10,11,44
The mechanism by which age at menarche is inversely related to the prevalence of LV diastolic dysfunction has yet to be elucidated. IR might be 1 possible mechanism linking early menarche to LV diastolic dysfunction, because IR and MetS are strongly related to early menarche.16 These conditions are associated with abnormal LV diastolic dysfunction and structure.26,45 In our study, however, the association between age at menarche and LV diastolic dysfunction did not change after adjustment for HOMA-IR. In addition, further adjustment by metabolic profiles including glucose, LDL-C, HDL-C, triglycerides, and systolic BP did not qualitatively change our results.
Early menarche is associated with adult obesity.46–49 Because obesity produces changes in cardiac structure and function, including LV diastolic dysfunction, adiposity may be a possible mechanism of the relationship between early menarche and LV diastolic dysfunction.25 In our study, associations between age at menarche and LV diastolic dysfunction were partially attenuated after adjustment for adult BMI or percent fat mass, indicating that adult obesity plays a role in mediation of LV diastolic dysfunction in relation to early menarche. Increased adiposity is associated with an increase in both preload and afterload through hyperdynamic circulation, chronic volume overload, and increased peripheral resistance,50,51 resulting in the development of diastolic dysfunction. Early menarche is also associated with childhood obesity, higher sex hormone levels and decreased sex hormone-binding globulin; thus early menarche may be an indicator of early life exposure or prolonged exposure to excessive adiposity and/or differential exposure to sex hormones.14,48 It has therefore been suggested that the chronically abnormal hormonal and metabolic milieu in women with early menarche may predispose them to LV diastolic dysfunction. In our study, however, data on childhood BMI and other measures of sex hormones were unavailable. Future studies that assess childhood BMI, biomarkers and sex hormones are required to establish whether earlier menarche is an independent risk factor for LV diastolic dysfunction.
Study LimitationsFirst, the cross-sectional design limited our ability to establish temporal relationships and infer causality. Second, age at menarche was self-reported, so measurement error was inevitable. However, age at menarche is reported to be usually well recalled into adulthood.52,53 Third, for determining LV diastolic dysfunction in this study, we used peak mitral inflow velocity (E) and septal mitral annulus velocity (e’) by tissue Doppler imaging only during early diastole among the measurements recommended by the American Society of Echocardiography to assess LV diastolic function.35 For mitral annular velocity, measurement was only of the septal area. Septal E’ is more useful than lateral E’, because septal E’ reflects LV longitudinal myocardial relaxation and diastolic function.54,55 Also, the LA AP dimension does not exactly reflect the actual LA diameter, but an epidemiological study showed a correlation between actual LA diameter and AP diameter measured by M-mode.56 Because we had no further information about LA volume or tricuspid regurgitation velocity, and LV diastolic dysfunction was defined as only decreased E’(<8 cm/s), misclassification in our study was inevitable. However, a series of sensitivity analyses consistently showed the same results as our primary results. We assumed that this misclassification was non-differential and did not affect the overall findings.
Finally, our study population consisted of approximately middle-aged women in Asia; thus, our results may not be generally applicable to other populations. Nevertheless, the major strengths of our study are the large sample size, comprehensive use of potential confounding variables and laboratory procedures with extensive quality control measures.
In conclusion, our study showed an inverse association between age at menarche and LV diastolic dysfunction in a large sample of middle-aged Korean women. These results highlight the importance of understanding and targeting the childhood antecedents of adult disease. Future prospective studies are needed to investigate potential causal relationships between menarche age and development of LV diastolic dysfunction.
We thank J. Ahn (Kangbuk Samsung Hospital, Seoul, Korea) for his technical support with the gathering data.
The authors have no conflicts of interest to disclose.
Y.C. and S.R. planned and designed the study, and directed its implementation, including quality assurance and control. They also analyzed the data and designed the study’s analytic strategy. H.S. helped supervise the field activities. J.G.K., J.S., H.-S.J., K.E.Y., J.C., C.-W.K., M.-J.K., K.-H.K., K.-C.S. and J.-Y.K. helped conduct the literature review and prepare the Methods and the Discussion sections of the text. S.R. drafted the manuscript. Y.C., J.G.K., J.S., H.-S.J., K.E.Y., J.C., C.-W.K., M.-J.K., K.-H.K., J.-Y.K., H.S., K.-C.S. and S.R. interpreted the results. All authors contributed to critical revision of the manuscript.
None to declare.
Supplementary File 1
Table S1. Association of age at menarche with LV diastolic dysfunction among 17,290 participants without hypertension or diabetes
Table S2. Relationship between brachial-ankle pulse-wave velocity and age at menarche
Table S3. Age-matched conditional logistic regression after excluding postmenopausal women
Table S4. Prevalence ratios (95% CI) of LV diastolic dysfunction by age at menarche
Please find supplementary file(s);
http://dx.doi.org/10.1253/circj.CJ-17-0231
