In this longitudinal study of middle-aged Korean women, we recruited participants between 2014 and 2018 from the Kangbuk Samsung Health Study, a cohort study of Korean men and women who underwent annual or biannual comprehensive health examinations at Kangbuk Samsung Hospital Total Healthcare Center clinics in Seoul and Suwon, South Korea. The eligibility criteria for enrollment in this longitudinal study included: (1) age 42–52 years; (2) no history of hysterectomy, oophorectomy, or hormone replacement therapy; (3) at least one menstrual period in the three months prior to the health screening check-up and no amenorrhea lasting for ≥ 60 days; and 4) no history of a chronic disease that may affect menstrual cycles (malignancy, renal failure, and hypo- or hyperthyroidism). Among the 5,230 women initially enrolled, 283 women in the early or late menopausal transition or postmenopausal stages were excluded because we only included women in the premenopausal stage in this longitudinal study. We also excluded participants who withdrew (n = 194), had no information on VMSs or lipid profiles (n = 59), used lipid-lowering drugs (n = 76), and had serum TG levels > 400 mg/dL (n = 3). In order to investigate longitudinal associations, women who had VMSs at baseline (n = 1,029), did not receive follow-up examinations (n = 1,043), and had missing information on VMSs (n = 3) during the follow-up period were excluded (Fig. 1). Therefore, 2,540 participants were ultimately included in this study.
This study was approved by the Institutional Review Board of Kangbuk Samsung Hospital (IRB No. KBSMC 2022–01-014). All methods in this cohort study were performed in accordance with relevant guidelines and regulations. All study participants provided written informed consent.
Information regarding demographic characteristics, health-related behaviors, medication usage, and reproductive factors was obtained using a standardized, structured, and self-administered questionnaire. Regarding health-related lifestyle factors, smoking status was categorized as never smokers and ever-smokers. Because the proportion of former/current smokers was very low in women, we did not differentiate former from current smokers. Never smokers were defined as women who had smoked less than 100 cigarettes in their lifetime and ever-smokers were defined as women who had smoked equal to or more than 100 cigarettes during their lifetime6. Levels of alcohol intake were categorized as < 10 g/day, 10–40 g/day, and > 40 g/day. Physical activity was assessed using the validated Korean version of the International Physical Activity Questionnaire short form and was divided into three groups: inactive, minimally active, and health-enhancing physical activity27,28. Education status was dichotomized as less than university graduates and equal to or greater than university graduates.
Reproductive factors included age at menarche, parity, and menopausal status. Parity, defined as the number of pregnancies, including live births and stillbirths, was divided into four groups: no child, one child, two children, and three or more children. Menopausal stages were classified based on the criteria of the Stages of Reproductive Aging Workshop + 10 as (1) premenopause; (2) early menopausal transition, defined as a persistent difference of ≥ 7 days in the length of consecutive cycles; (3) late menopausal transition, defined as amenorrhea of ≥ 60 days; and (4) postmenopause, defined as amenorrhea of ≥ 12 months29,30.
Participants wore a lightweight hospital gown and no shoes, and their height, weight, and body composition were measured by trained experts. Body mass index (BMI) was calculated as weight in kilograms divided by height in meters squared (kg/m2). Hypertension was defined as a systolic blood pressure (SBP) ≥ 130 mmHg, diastolic blood pressure (DBP) ≥ 80 mmHg31, or ongoing use of blood pressure-lowering drugs.
Blood samples were collected from the antecubital vein after at least 10 h of fasting. Fasting blood measurement included high-sensitivity C-reactive protein, glucose, hemoglobin A1c and insulin. Diabetes mellitus was defined as fasting glucose levels ≥ 126 mg/dL, glycated hemoglobin ≥ 6.5% (48 mmol/mol), or current use of insulin or glucose-lowering medication. Insulin resistance was evaluated by the homeostatic model assessment of insulin resistance (HOMA-IR) according to the following formula: HOMA-IR = fasting serum insulin (µIU/mL) × fasting serum glucose (mg/dL)/405 32. Serum total cholesterol and TG concentrations were determined using an enzymatic colorimetric assay. HDL and LDL cholesterol levels were measured directly using a homogenous enzymatic colorimetric assay on a cobas 8000 c702 (Roche Diagnostics, Tokyo, Japan). The classification of lipoprotein lipid levels was based on the National Lipid Association Recommendations for Patient-Centered Management of Dyslipidemia33. We categorized total cholesterol levels into three groups: < 200 mg/dL, 200–239 mg/dL, and ≥ 240 mg/dL. Serum LDL cholesterol levels were divided into three groups: < 100 mg/dL, 100–129 mg/dL, and ≥ 130 mg/dL. HDL cholesterol concentrations were categorized as groups of low (< 50 mg/dL), borderline (50–59 mg/dL), and high (≥ 60 mg/dL); non-HDL-cholesterols were divided into three groups: normal (< 130 mg/dL), borderline (130–159 mg/dL), and high (≥ 160 mg/dL). As there were only 6% of women with serum triglyceride n with serum triglyg/dL 0 oups: ndL), and high ((130 0 odL)mg/dL) n < 100 mg/dL, 100–149 mg/dL, and ≥ 150 mg/dL.
The VMSs included hot flashes and night sweats. To determine the presence and degree of VMSs, the validated Korean version of the Menopause-Specific Quality of Life (MENQOL) questionnaire was administered at baseline and at each follow-up visit34,35. Study participants indicated whether they had experienced VMSs during the past month and described how bothersome the symptoms were on a seven-point Likert scale; from “not bothered at all” (0) to “extremely bothersome” (6)34,36. For the statistical analysis, the raw scores of VMSs intensity were recoded to an eight-point grading system including zero: the answer “No” was rescored as zero and “Yes, but not bothered at all” was converted to one. The increasing degree of VMSs severity ranging from 1 to 6 was rescored from 2 to 7. If the participant responded “No” to hot flashes or night sweats, we considered the participant as not having VMSs. Women who answered “Yes” and experienced hot flashes or night sweats were considered as having VMSs. We also considered groups who had ≥ 3 recoded points as having moderate/severe VMSs, and women with 1 or 2 points as having mild VMSs. Early-onset VMSs was defined as the occurrence of VMSs before menopause.
Descriptive statistics were used to summarize the baseline characteristics of the study participants by early-onset VMSs. The differences in demographic and clinical characteristics at baseline between women with and without early-onset VMSs were analyzed using the t-test for continuous variables and Chi-squared test for categorical variables. The primary outcome was incident early-onset of VMSs. Based on self-reported surveys, we received participants’ information on VMSs and menopausal status at each visit. At baseline and during follow-up, VMSs, menopausal stages, and the last two menstruation dates were assessed based on the self-administered, structured questionnaire including the MENQOL. Each participant was followed from the time of the baseline visit to the time of first report of VMSs occurrence, the time of menopause, or the last time the questionnaire survey was completed, whichever came first. Each participant was followed from the time of the baseline visit to the time of VMSs occurrence, the time of menopause, or the last time the questionnaire survey was completed, whichever came first. If a woman was in the postmenopausal stage at the time of the first VMS report, follow-up ended at the preceding visit where MENQOL assessment was completed prior to menopause. Cox proportional hazards regression model was used to estimate hazard ratios (HRs) with 95% confidence intervals (CIs) for incident early-onset VMSs according to lipid profiles. For sensitivity analysis, we repeated the analyses using moderate/severe VMSs as an endpoint. Potential confounders included age, BMI, BP, diabetes, educational status, parity, physical activity, smoking status, and alcohol intake based on previous findings (also see Supplementary Figure 2)37,38,39,40,41. Confounding variables were chosen for inclusion in the multivariable models if they met the following criteria: (1) were associated with the outcome (incident early-onset VMS) and (2) were associated with the exposure (lipid profiles), but 3) were not intermediate variables in the causal pathway between the exposure (lipid profiles) and the outcome (incident VMS). For a linear trend test, the median value of each category was included as a continuous variable in the model. All statistical analyses were conducted using Stata version 17.0 (Stata Corp LP; College Station, TX, USA). Statistical significance was defined as a two-sided P-value of less than 0.05.