Association between BMI trajectories and metabolic syndrome in postmenopausal women: a cross-sectional study from NHANES 1999–2018 Authors: Yingying Liu, Zhixiang Liang, Xingying Li, Juan Xiong Published: 2025 ABSTRACT BackgroundObesity is an established risk factor for metabolic syndrome (MetS), with both conditions being particularly prevalent among postmenopausal women. However, research examining the relationship between body mass index (BMI) changes across the lifespan and MetS in this specific population remains limited. This study aims to explore the association between BMI trajectories and MetS risk in postmenopausal women.MethodsThis cross-sectional study included postmenopausal women aged ≥ 60 years from the National Health and Nutrition Examination Survey 1999–2018. MetS was defined by 2009 Joint Interim Statement criteria. BMI trajectories over time were identified using latent class growth mixture modeling. Multivariable logistic regression analyses were performed to evaluate the association between distinct BMI trajectories and MetS risk. Subgroup analyses based on reproductive characteristics were conducted to investigate potential effect modification in the relationship between BMI trajectories and MetS risk.ResultsAmong 4061 eligible participants, the overall prevalence of MetS was 71.57% (n = 3070). Three distinct BMI trajectories were identified: stable trajectory (n = 3047, 76.34%), increasing trajectory (n = 540, 12.30%), and increasing–decreasing trajectory (n = 474, 11.36%). Compared with the stable trajectory group, participants in the increasing trajectory group had a higher risk of MetS (OR = 5.09, 95% CI: 3.19–8.13), while those in the increasing–decreasing trajectory group had the highest risk of MetS (OR = 5.89, 95% CI: 3.73–9.31). Subgroup analyses revealed significant effect modification of the association between BMI trajectories and MetS risk by early menopause (P = 0.059), early age at first live birth (P < 0.05), and high parity (P < 0.05).ConclusionsThe increasing and increasing–decreasing BMI trajectory groups are associated with a higher risk of MetS in postmenopausal women. This association is significantly modified by reproductive characteristics. Maintaining a long-term stable BMI may benefit metabolic health.Supplementary InformationThe online version contains supplementary material available at 10.1186/s12889-025-25480-4. Background Metabolic Syndrome (MetS) is characterized by a cluster of metabolic disorders, including abdominal obesity, elevated triglycerides, reduced high-density lipoprotein cholesterol, elevated blood pressure, and elevated fasting blood glucose [1]. MetS significantly increases the risk of cardiovascular disease (CVD) and type 2 diabetes mellitus, posing a major global health burden [1]. Epidemiological studies reveal a sustained increase in the global prevalence of MetS among adults, with age-standardized prevalence varying significantly by definition, region, and demographic factors. Higher rates are often observed in urban populations and among women [2]. Menopause, typically occurring between the ages of 45 and 55, marks the cessation of ovarian function and a decline in estrogen levels [3]. This estrogen deficiency affects lipid metabolism, exacerbates insulin resistance, and promotes visceral adiposity [4]. Consequently, postmenopausal women face a 3- to 5-fold higher risk of MetS compared with premenopausal women [5]. This risk increases further with age, reaching up to 6-fold higher in women aged ≥ 60 compared with those aged 20–39 [6]. Weight gain and obesity during the menopausal transition are significant contributors to the rising prevalence of MetS in postmenopausal women [7]. Approximately two-thirds of women of perimenopausal age are overweight or obese [8], with excess body fat leading to profound metabolic consequences. While the risk of MetS among women is more than 5-fold higher for those who are overweight compared with normal-weight peers and dramatically elevated for those with obesity [6], this threat is exacerbated during the menopausal transition driven by hormonal shifts that favor visceral fat accumulation. Current research paradigms fail to adequately capture the dynamic nature of weight change during menopause. Most studies rely on single-point BMI measurements, which fail to account for the cumulative effect of weight change over time [9–11]. Even studies that use BMI at two time points inadequately model weight change, assuming a linear progression and overlooking the complex, non-linear patterns that may differentially impact MetS risk [12]. A life-course perspective, using BMI trajectories derived from multiple time points, provides a more nuanced understanding of how long-term weight change patterns contribute to disease risk [13–17]. This helps distinguish subgroups that share a similar current BMI but differ in cumulative and temporal profiles that may map onto different MetS components. However, this advanced approach has predominantly been applied to child, adolescent, and general adult populations. Notably, the relationship between distinct BMI trajectories spanning the menopausal transition and subsequent MetS risk in postmenopausal women remains unexplored. To fill this gap, this study aims to clarify the complex association between long-term BMI patterns and MetS risk in postmenopausal women by employing latent class growth mixture modeling (LCGMM) to identify distinct BMI trajectories across the lifespan of US postmenopausal women aged ≥ 60 years and exploring the association between these trajectories and the prevalence of MetS. Subgroup analyses will also be used to evaluate potential effect modification by reproductive factors on this association. The findings from this study will enhance our understanding of MetS pathogenesis in postmenopausal women and inform the development of more targeted, life-course-based prevention strategies. Methods Study population This study utilized data from 10 consecutive cycles of the National Health and Nutrition Examination Survey (NHANES) 1999–2018, with a total original enrollment of 101,316 participants. NHANES is a series of national cross-sectional survey conducted by the National Center for Health Statistics (NCHS) to assess the health and nutritional status of the US population, which used a stratified, clustered, multi-stage, and multi-year sampling method to ensure representativeness. NHANES obtained approval from the NCHS Ethics Review Board, and all participants provided written informed consent. A total of 51,423 women were initially identified. The following exclusion criteria were applied: (1) age < 60 years at assessment; (2) having at least one menstrual period within the past 12 months or absence of menstrual period due to pregnancy or breastfeeding; (3) history of simple premenopausal hysterectomy; (4) age of menopause < 30 or > 60 years; (5) insufficient data to determine MetS or non-MetS; and (6) fewer than three available BMI records. Ultimately, 4061 postmenopausal women were included in the final analysis. The participant selection process is detailed in Fig. 1.Fig. 1Flow chart of the study population Flow chart of the study population Definition of menopausal status Menopause is defined as the permanent cessation of menstruation resulting from ovarian failure, which leads to estrogen deficiency. Menopause is primarily classified into two types: natural menopause and surgical menopause. Natural menopause is diagnosed after 12 consecutive months of amenorrhea without other physiological or pathological causes, such as pregnancy, breastfeeding or hormonal contraceptives [18]. Surgical menopause occurs due to bilateral oophorectomy (with or without hysterectomy) or medically induced ovarian failure, such as chemotherapy or radiotherapy [18]. Notably, in cases of simple premenopausal hysterectomy, where menstruation ceases but ovarian function persists, such women are not classified as menopausal [19]. Postmenopausal women were identified using the NHANES reproductive health questionnaire through a sequential process. First, women were asked whether they had at least one menstrual period in the past 12 months. Those who responded “yes” were excluded. Those who responded “no” were asked the reason. Responses indicating pregnancy or breastfeeding were excluded, while responses such as “periods usually irregular”, “menopause”, “medical conditions/treatment”, “hysterectomy”, “change of life”, or “other” were retained. Subsequently, women were asked whether they underwent a hysterectomy or bilateral oophorectomy and the age at the time of the procedure. Women who underwent a hysterectomy before menopause, with both ovaries preserved or removed after menopause, were excluded from the postmenopausal classification; those who underwent a bilateral oophorectomy at or before menopause, with or without hysterectomy, were classified as having surgical menopause; and those who underwent these procedures after menopause, or with no history of hysterectomy or oophorectomy, were classified as having natural menopause [20, 21]. Assessment of BMI BMI at the time of assessment was derived from measured weight and height obtained from the Body Measures file within the Examination Data. Historical weight data, including weight 1 year ago, weight 10 years ago, weight and height at age 25, maximum weight, and age at maximum weight, were self-reported and obtained from the Weight History questionnaire. Heights 1 year ago and 10 years ago were assumed equal to height at assessment [22]. Height at the time of maximum weight was estimated using either the height at assessment or the height at age 25, depending on temporal proximity to the reported age at maximum weight. BMI for all time points was calculated as weight (kg) divided by height squared (m2). Our BMI assessment was consistent with previous NHANES studies [17, 22, 23]. Furthermore, the reliability of self-reported weight is supported by a validation study, which shows moderate to high correlations between recalled and measured weight values [24]. “Trajectory” was defined as a latent BMI pattern constructed from recalled and measured time points. Definition of MetS MetS was defined according to the harmonized criteria established in the 2009 Joint Interim Statement (JIS) [1]. MetS required meeting ≥ 3 of the following 5 criteria in women: (1) elevated waist circumference (WC): ≥ 88 cm; (2) elevated triglycerides (TG): ≥ 150 mg/dL or current use of lipid-lowering medications; (3) reduced high-density lipoprotein cholesterol (HDL-C): < 50 mg/dL or current use of HDL-C-raising medications; (4) elevated blood pressure (BP): systolic ≥ 130 mmHg and/or diastolic ≥ 85 mmHg, or current use of antihypertensive medications in individuals with a history of hypertension; and (5) elevated fasting blood glucose (FBG): ≥ 100 mg/dL or current use of antidiabetic medications, with FBG values harmonized across survey cycles using regression equations provided by the NCHS. Covariates The demographic, socioeconomic, and lifestyle factors that could influence the outcomes were selected based on previous literature [6, 25–31], including age, ethnicity, education, poverty-income ratio (PIR; ≤ 1, > 1 to < 4, and ≥ 4), smoking status (current, former, and never), alcohol consumption (current, former, and never), dietary inflammation index (DII; Q1, Q2, Q3, and Q4), physical activity, and sleep status. Additionally, reproductive factors were considered [32–35], including age at menarche, age at menopause, reproductive lifespan, age at first live birth, number of live births, menopause type (natural or surgical), and hormone therapy (yes or no). For model adjustment and subgroup analysis, these variables were further categorized as follows: early menarche (< 12 years); early menopause (< 45 years); short reproductive lifespan (< 30 years); early age at first live birth (< 20 years); and high parity (≥ 4 live births) [36]. Statistical analysis LCGMM was used to identify distinct trajectories of BMI over time among postmenopausal women [37]. LCGMM calculates the posterior probability of each individual belonging to each trajectory class and assigns individuals to the class with the highest posterior probability. To determine the optimal shape and number of trajectories, we fitted models with varying trajectory shapes and class numbers. The optimal model was selected based on the following criteria: (1) a lower Bayesian Information Criterion (BIC); (2) average posterior probabilities of assignment (AvePP) ≥ 0.7; (3) odds of correct classification (OCC) > 5; and (4) class sizes of the population ≥ 5%. In addition to statistical fit, the clinical interpretability of the trajectories was also considered in the model selection process. Continuous variables were presented as means with standard errors (SEs), and categorical variables as counts with percentages. Group comparisons were conducted using analysis of variance (ANOVA) or t tests for continuous variables and chi-square tests for categorical variables. Logistic regression models were employed to evaluate the association between BMI trajectory groups and MetS risk by odds ratios (ORs) and 95% confidence intervals (CIs). Model 1 was adjusted for age; Model 2 was additionally adjusted for ethnicity, education, PIR, smoking status, alcohol consumption, DII, physical activity, and sleep status; and Model 3 was further adjusted for early menarche, early menopause, early age at first live birth, high parity, menopause type, and hormone therapy. Subgroup analyses were performed to assess the effect consistency across all reproductive covariates. Sensitivity analyses were conducted to assess the robustness of the primary results by: (1) excluding participants with any missing data on MetS components; and (2) using the International Diabetes Federation (IDF) 2005 criteria for MetS. Missing values were imputed using the random forest method. Both descriptive and inferential analyses incorporated NHANES sampling weights and clustering effects using the R “survey” package (version 4.4.2), with the exception of counts for categorical variables, which were reported as unweighted values to provide direct transparency on the actual sample size. All statistical analyses were performed using R software (version 4.4.3), and a two-sided P-value < 0.05 was considered statistically significant. Results Characteristics of participants This study included 4061 postmenopausal women. The mean age at assessment was 69.95 years, and the overall prevalence of MetS was 71.57% (n = 3070). The characteristics of the participants according to the MetS status are summarized in Table S1. Significant differences were observed between MetS groups for all variables except smoking status, hormone therapy, age at menopause, and reproductive lifespan. Figure 2 illustrates the three distinct BMI trajectories identified using LCGMM, labeled as “stable” (n = 3047, 76.34%), “increasing” (n = 540, 12.30%), and “increasing–decreasing” (n = 474, 11.36%). The detailed statistical indices for all evaluated models, including the BIC, AvePP, OCC, and class sizes of the population are presented in Table S2, with the quadratic 3-class model identified as providing the best fit. Comparisons of characteristics according to the three BMI trajectory groups are described in Table 1. Participants in the increasing or increasing–decreasing trajectory groups were characterized by younger age at assessment, lower economic status, a more pro-inflammatory diet, less hormone therapy use, a higher proportion of surgical menopause, earlier age at menarche and first live birth, and a greater number of live births.Fig. 2BMI trajectories and 95% CIs in postmenopausal women aged ≥ 60 from NHANES (1999-2018)Table 1Characteristics of the study participants by BMI trajectoriesVariablesOverall(N = 4061)Stable(N = 3047)Increasing(N = 540)Increasing–decreasing(N = 474)P-valueMetS status (%)< 0.001 Non-MetS991 (28.43)906 (34.33)40 (9.93)45 (8.80) MetS3070 (71.57)2141 (65.67)500 (90.07)429 (91.20)Ethnicity (%)< 0.001 Mexican American612 (3.81)431 (3.37)110 (5.75)71 (4.65) Non-Hispanic White2128 (80.21)1685 (81.27)235 (76.85)208 (76.68) Non-Hispanic Black659 (7.41)397 (5.84)121 (10.91)141 (14.16) Other ethnicity662 (8.57)534 (9.51)74 (6.49)54 (4.50)Education (%)0.198 Less than high school1260 (19.39)924 (19.48)193 (21.70)143 (16.29) High school1075 (29.31)814 (28.94)122 (26.72)139 (34.57) Some college or above1726 (51.30)1309 (51.58)225 (51.58)192 (49.14)PIR (%)0.025 ≤ 1717 (10.74)491 (9.59)111 (12.90)115 (16.16) > 1 to < 42466 (59.12)1868 (59.48)323 (58.65)275 (57.22) ≥ 4878 (30.14)688 (30.93)106 (28.45)84 (26.62)Smoke (%)0.300 Never2513 (58.34)1914 (58.90)313 (53.84)286 (59.43) Former1137 (31.49)816 (30.59)177 (36.00)144 (32.70) Current411 (10.17)317 (10.51)50 (10.16)44 (7.87)Alcohol (%)0.161 Never1762 (39.14)1354 (40.59)214 (33.36)194 (35.69) Former1180 (27.26)864 (26.66)172 (29.76)144 (28.64) Current1119 (33.59)829 (32.76)154 (36.87)136 (35.68)DII (%)0.001 Q1986 (28.47)788 (29.89)99 (21.16)99 (26.83) Q21017 (26.38)790 (27.24)119 (23.47)108 (23.75) Q3981 (23.64)727 (22.97)150 (29.02)104 (22.28) Q41077 (21.52)742 (19.90)172 (26.36)163 (27.14)Physical activity (%)0.092 Inactive2522 (58.08)1857 (56.53)352 (66.14)313 (59.81) Insufficient502 (12.60)390 (12.84)63 (11.98)49 (11.68) Moderate546 (15.20)414 (15.40)69 (12.26)63 (17.06) High491 (14.11)386 (15.23)56 (9.63)49 (11.45)Sleep status (%)0.167 Insufficient1203 (24.77)862 (23.83)181 (28.07)160 (27.46) Sufficient2098 (55.09)1620 (56.11)262 (54.49)216 (48.92) Excessive760 (20.14)565 (20.06)97 (17.44)98 (23.62)Hormone therapy (%)< 0.001 No2484 (53.82)1807 (51.86)335 (53.82)342 (67.00) Yes1577 (46.18)1240 (48.14)205 (46.18)132 (33.00)Menopause type (%)0.054 Natural3184 (76.67)2429 (77.69)413 (75.67)342 (70.92) Surgical877 (23.33)618 (22.31)127 (24.33)132 (29.08)Age, years69.95 (6.98)70.63 (7.13)67.33 (5.46)68.25 (6.49)< 0.001Age at menarche, years12.84 (1.68)12.94 (1.66)12.57 (1.67)12.40 (1.65)< 0.001Age at menopause, years48.10 (6.44)48.11 (6.34)48.01 (7.17)48.12 (6.29)0.981Reproductive lifespan, years35.26 (6.60)35.16 (6.53)35.45 (7.24)35.72 (6.37)0.344Age at first live birth, years22.70 (4.46)22.91 (4.45)22.20 (4.54)21.79 (4.24)< 0.001Number of live births, times3.67 (2.32)3.57 (2.05)3.93 (3.01)4.06 (3.03)0.001Continuous variables were presented as means with SEs and categorical variables as counts (%). Group comparisons were performed using ANOVA for continuous variables and chi-square tests for categorical variables. Both descriptive and inferential analyses were weighted; categorical counts were unweighted BMI trajectories and 95% CIs in postmenopausal women aged ≥ 60 from NHANES (1999-2018) Table 1Characteristics of the study participants by BMI trajectoriesVariablesOverall(N = 4061)Stable(N = 3047)Increasing(N = 540)Increasing–decreasing(N = 474)P-valueMetS status (%)< 0.001 Non-MetS991 (28.43)906 (34.33)40 (9.93)45 (8.80) MetS3070 (71.57)2141 (65.67)500 (90.07)429 (91.20)Ethnicity (%)< 0.001 Mexican American612 (3.81)431 (3.37)110 (5.75)71 (4.65) Non-Hispanic White2128 (80.21)1685 (81.27)235 (76.85)208 (76.68) Non-Hispanic Black659 (7.41)397 (5.84)121 (10.91)141 (14.16) Other ethnicity662 (8.57)534 (9.51)74 (6.49)54 (4.50)Education (%)0.198 Less than high school1260 (19.39)924 (19.48)193 (21.70)143 (16.29) High school1075 (29.31)814 (28.94)122 (26.72)139 (34.57) Some college or above1726 (51.30)1309 (51.58)225 (51.58)192 (49.14)PIR (%)0.025 ≤ 1717 (10.74)491 (9.59)111 (12.90)115 (16.16) > 1 to < 42466 (59.12)1868 (59.48)323 (58.65)275 (57.22) ≥ 4878 (30.14)688 (30.93)106 (28.45)84 (26.62)Smoke (%)0.300 Never2513 (58.34)1914 (58.90)313 (53.84)286 (59.43) Former1137 (31.49)816 (30.59)177 (36.00)144 (32.70) Current411 (10.17)317 (10.51)50 (10.16)44 (7.87)Alcohol (%)0.161 Never1762 (39.14)1354 (40.59)214 (33.36)194 (35.69) Former1180 (27.26)864 (26.66)172 (29.76)144 (28.64) Current1119 (33.59)829 (32.76)154 (36.87)136 (35.68)DII (%)0.001 Q1986 (28.47)788 (29.89)99 (21.16)99 (26.83) Q21017 (26.38)790 (27.24)119 (23.47)108 (23.75) Q3981 (23.64)727 (22.97)150 (29.02)104 (22.28) Q41077 (21.52)742 (19.90)172 (26.36)163 (27.14)Physical activity (%)0.092 Inactive2522 (58.08)1857 (56.53)352 (66.14)313 (59.81) Insufficient502 (12.60)390 (12.84)63 (11.98)49 (11.68) Moderate546 (15.20)414 (15.40)69 (12.26)63 (17.06) High491 (14.11)386 (15.23)56 (9.63)49 (11.45)Sleep status (%)0.167 Insufficient1203 (24.77)862 (23.83)181 (28.07)160 (27.46) Sufficient2098 (55.09)1620 (56.11)262 (54.49)216 (48.92) Excessive760 (20.14)565 (20.06)97 (17.44)98 (23.62)Hormone therapy (%)< 0.001 No2484 (53.82)1807 (51.86)335 (53.82)342 (67.00) Yes1577 (46.18)1240 (48.14)205 (46.18)132 (33.00)Menopause type (%)0.054 Natural3184 (76.67)2429 (77.69)413 (75.67)342 (70.92) Surgical877 (23.33)618 (22.31)127 (24.33)132 (29.08)Age, years69.95 (6.98)70.63 (7.13)67.33 (5.46)68.25 (6.49)< 0.001Age at menarche, years12.84 (1.68)12.94 (1.66)12.57 (1.67)12.40 (1.65)< 0.001Age at menopause, years48.10 (6.44)48.11 (6.34)48.01 (7.17)48.12 (6.29)0.981Reproductive lifespan, years35.26 (6.60)35.16 (6.53)35.45 (7.24)35.72 (6.37)0.344Age at first live birth, years22.70 (4.46)22.91 (4.45)22.20 (4.54)21.79 (4.24)< 0.001Number of live births, times3.67 (2.32)3.57 (2.05)3.93 (3.01)4.06 (3.03)0.001Continuous variables were presented as means with SEs and categorical variables as counts (%). Group comparisons were performed using ANOVA for continuous variables and chi-square tests for categorical variables. Both descriptive and inferential analyses were weighted; categorical counts were unweighted Characteristics of the study participants by BMI trajectories Continuous variables were presented as means with SEs and categorical variables as counts (%). Group comparisons were performed using ANOVA for continuous variables and chi-square tests for categorical variables. Both descriptive and inferential analyses were weighted; categorical counts were unweighted Association between BMI trajectories and MetS The prevalence of MetS varied significantly across trajectory groups (P < 0.001): 65.67% in the stable trajectory group, 90.07% in the increasing trajectory group, and 91.20% in the increasing–decreasing trajectory group. As shown in Table 2, BMI trajectories were strongly associated with MetS risk. In the age-adjusted model (Model 1), compared with the stable trajectory group, the ORs for MetS were 5.69 (95% CI: 3.52–9.21, P < 0.001) for the increasing trajectory group and 6.29 (95% CI: 4.17–9.47, P < 0.001) for the increasing–decreasing trajectory group. This association persisted after sequential adjustments. In the fully adjusted model (Model 3), the ORs for MetS remained significantly elevated at 5.09 (95% CI: 3.19–8.13, P < 0.001) for the increasing trajectory group and 5.89 (95% CI: 3.73–9.31, P < 0.001) for the increasing–decreasing trajectory group.Table 2Logistic regression for the association between BMI trajectories and MetS riskStableModel 1OR (95% CI)P-valueModel 2OR (95% CI)P-valueModel 3OR (95% CI)P-valueIncreasing5.69 (3.52–9.21)< 0.0015.27 (3.30–8.41)< 0.0015.09 (3.19–8.13)< 0.001Increasing–decreasing6.29 (4.17–9.47)< 0.0016.00 (3.88–9.29)< 0.0015.89 (3.73–9.31)< 0.001Model 1: adjusted for age;Model 2: adjusted for variables in model 1 plus ethnicity, education, PIR, smoking status, alcohol consumption, DII, physical activity, and sleep status;Model 3: adjusted for variables in model 2 plus early menarche, early menopause, early age at first live birth, high parity, menopause type, and hormone therapy Table 2Logistic regression for the association between BMI trajectories and MetS riskStableModel 1OR (95% CI)P-valueModel 2OR (95% CI)P-valueModel 3OR (95% CI)P-valueIncreasing5.69 (3.52–9.21)< 0.0015.27 (3.30–8.41)< 0.0015.09 (3.19–8.13)< 0.001Increasing–decreasing6.29 (4.17–9.47)< 0.0016.00 (3.88–9.29)< 0.0015.89 (3.73–9.31)< 0.001Model 1: adjusted for age;Model 2: adjusted for variables in model 1 plus ethnicity, education, PIR, smoking status, alcohol consumption, DII, physical activity, and sleep status;Model 3: adjusted for variables in model 2 plus early menarche, early menopause, early age at first live birth, high parity, menopause type, and hormone therapy Logistic regression for the association between BMI trajectories and MetS risk Model 1: adjusted for age; Model 2: adjusted for variables in model 1 plus ethnicity, education, PIR, smoking status, alcohol consumption, DII, physical activity, and sleep status; Model 3: adjusted for variables in model 2 plus early menarche, early menopause, early age at first live birth, high parity, menopause type, and hormone therapy The individual components of MetS were compared across BMI trajectory groups. As illustrated in Fig. 3, compared with the stable trajectory group, both the increasing and increasing–decreasing trajectory groups showed elevated levels of WC and FBG, and reduced HDL-C levels, while only the increasing trajectory group showed significantly elevated TG levels. Notably, WC was higher in the increasing trajectory group than in the increasing–decreasing trajectory group. For blood pressure measurements, no significant differences were observed among any of the groups.Fig. 3MetS components across BMI trajectory groups A waist circumference (WC); B triglycerides (TG); C high-density lipoprotein cholesterol (HDL-C); D fasting blood glucose (FBG); E systolic blood pressure (SBP); F diastolic blood pressure (DBP) MetS components across BMI trajectory groups A waist circumference (WC); B triglycerides (TG); C high-density lipoprotein cholesterol (HDL-C); D fasting blood glucose (FBG); E systolic blood pressure (SBP); F diastolic blood pressure (DBP) Subgroup analysis by reproductive characteristics Table 3 presents subgroup analyses assessing effect modification by reproductive characteristics. For the increasing trajectory group, the association with MetS was potentiated in women with early menopause (OR = 11.39, 95% CI: 4.29–30.19) compared to those without early menopause (OR = 4.56, 95% CI: 2.71–7.65). In contrast, for the increasing–decreasing trajectory group, this association was attenuated in women with early menopause (OR = 2.47, 95% CI: 1.08–5.67) compared to those without early menopause (OR = 8.20, 95% CI: 4.73–14.22). This interaction showed borderline significance (P for interaction = 0.059). Similarly, early age at first live birth and high parity significantly modified the association between BMI trajectories and MetS risk (both P for interaction < 0.05). However, no significant effect modification was observed for the other reproductive factors (all P for interaction > 0.05).Table 3Subgroup analysis of BMI trajectories for MetS riskVariablesN (%)LevelsOR (95% CI)P-valueP for interactionEarly menarche0.975 No3322 (81.64)StableIncreasing8.84 (5.05–15.48)< 0.001Increasing–decreasing5.90 (3.52–9.87)< 0.001 Yes739 (18.36)StableIncreasing2.01 (0.92–4.40)0.083Increasing–decreasing5.88 (1.98–17.46)0.002Early menopause0.059 No3046 (75.38)StableIncreasing4.56 (2.71–7.65)< 0.001Increasing–decreasing8.20 (4.73–14.22)< 0.001 Yes1015 (24.62)StableIncreasing11.39 (4.29–30.19)< 0.001Increasing–decreasing2.47 (1.08–5.67)0.035Short reproductive lifespan0.130 No3246 (80.67)StableIncreasing4.82 (2.89–8.05)< 0.001Increasing–decreasing7.27 (4.39–12.02)< 0.001 Yes815 (19.33)StableIncreasing7.54 (2.87–19.79)< 0.001Increasing–decreasing2.63 (1.11–6.23)0.031Early age at first live birth0.014 No2874 (74.48)StableIncreasing4.53 (2.60–7.89)< 0.001Increasing–decreasing8.60 (5.24–14.11)< 0.001 Yes1187 (25.52)StableIncreasing7.91 (3.60–17.40)< 0.001Increasing–decreasing3.27 (1.47–7.26)0.004High parity0.012 No2009 (56.33)StableIncreasing4.58 (2.48–8.44)< 0.001Increasing–decreasing11.68 (6.37–21.41)< 0.001 Yes2052 (43.67)StableIncreasing6.22 (3.34–11.59)< 0.001Increasing–decreasing3.27 (1.73–6.16)< 0.001Menopause type0.841 Natural3184 (76.67)StableIncreasing4.91 (2.92–8.23)< 0.001Increasing–decreasing5.89 (3.58–9.71)< 0.001 Surgical877 (23.33)StableIncreasing5.83 (2.36–14.41)< 0.001Increasing–decreasing5.89 (2.23–15.61)0.001Hormone therapy0.967 No2484 (53.82)StableIncreasing3.53 (2.03–6.14)< 0.001Increasing–decreasing6.23 (3.65–10.62)< 0.001 Yes1577 (46.18)StableIncreasing9.65 (4.90–19.02)< 0.001Increasing–decreasing6.22 (2.76–14.01)< 0.001All models were adjusted for variables in Model 3 from Table 2, except for the stratification variable itself Table 3Subgroup analysis of BMI trajectories for MetS riskVariablesN (%)LevelsOR (95% CI)P-valueP for interactionEarly menarche0.975 No3322 (81.64)StableIncreasing8.84 (5.05–15.48)< 0.001Increasing–decreasing5.90 (3.52–9.87)< 0.001 Yes739 (18.36)StableIncreasing2.01 (0.92–4.40)0.083Increasing–decreasing5.88 (1.98–17.46)0.002Early menopause0.059 No3046 (75.38)StableIncreasing4.56 (2.71–7.65)< 0.001Increasing–decreasing8.20 (4.73–14.22)< 0.001 Yes1015 (24.62)StableIncreasing11.39 (4.29–30.19)< 0.001Increasing–decreasing2.47 (1.08–5.67)0.035Short reproductive lifespan0.130 No3246 (80.67)StableIncreasing4.82 (2.89–8.05)< 0.001Increasing–decreasing7.27 (4.39–12.02)< 0.001 Yes815 (19.33)StableIncreasing7.54 (2.87–19.79)< 0.001Increasing–decreasing2.63 (1.11–6.23)0.031Early age at first live birth0.014 No2874 (74.48)StableIncreasing4.53 (2.60–7.89)< 0.001Increasing–decreasing8.60 (5.24–14.11)< 0.001 Yes1187 (25.52)StableIncreasing7.91 (3.60–17.40)< 0.001Increasing–decreasing3.27 (1.47–7.26)0.004High parity0.012 No2009 (56.33)StableIncreasing4.58 (2.48–8.44)< 0.001Increasing–decreasing11.68 (6.37–21.41)< 0.001 Yes2052 (43.67)StableIncreasing6.22 (3.34–11.59)< 0.001Increasing–decreasing3.27 (1.73–6.16)< 0.001Menopause type0.841 Natural3184 (76.67)StableIncreasing4.91 (2.92–8.23)< 0.001Increasing–decreasing5.89 (3.58–9.71)< 0.001 Surgical877 (23.33)StableIncreasing5.83 (2.36–14.41)< 0.001Increasing–decreasing5.89 (2.23–15.61)0.001Hormone therapy0.967 No2484 (53.82)StableIncreasing3.53 (2.03–6.14)< 0.001Increasing–decreasing6.23 (3.65–10.62)< 0.001 Yes1577 (46.18)StableIncreasing9.65 (4.90–19.02)< 0.001Increasing–decreasing6.22 (2.76–14.01)< 0.001All models were adjusted for variables in Model 3 from Table 2, except for the stratification variable itself Subgroup analysis of BMI trajectories for MetS risk All models were adjusted for variables in Model 3 from Table 2, except for the stratification variable itself Sensitivity analysis The positive association between non-stable BMI trajectories and MetS remained statistically significant when the analysis was restricted to participants with complete data on all five MetS components, which was consistent with the primary analysis (Table S3). Consistent results were also observed when using the IDF 2005 criteria to define MetS (Table S4). These findings support the robustness of our primary results to variations in both data completeness and the definition of MetS. Discussion The relationship between BMI trajectories and MetS is an emerging field, with growing evidence linking long-term weight change patterns to metabolic health [14–16]. In particular, longitudinal studies across diverse populations consistently demonstrate this association. For instance, a 16-year study of Chinese adolescents identified three distinct BMI trajectories, finding that both low-fast and high-fast BMI trajectories were associated with a higher incidence of MetS in early adulthood compared with the low-slow BMI trajectory [14]. Similarly, a 30-year prospective cohort study confirmed that BMI trajectories from childhood to adolescence predicted MetS development in middle age, highlighting the importance of longitudinal weight change patterns over single-point measurements [15]. However, a 20-year study of American women demonstrated that the persistence of a high BMI from adolescence into adulthood is the true underlying risk factor for MetS in adulthood [16]. In our study of postmenopausal women aged ≥ 60 years, LCGMM identified three distinct BMI trajectories from adulthood through menopause to older age: stable, increasing, and increasing–decreasing. Both non-stable trajectory groups exhibited a significantly higher prevalence of MetS than the stable trajectory group. Notably, the strongest association with MetS risk was observed in the increasing–decreasing trajectory group, a pattern characterized by an earlier and steeper BMI rise followed by late-life decline that mirrors the established physiological changes during aging. Consistent with a global macro-epidemiological pattern, which shows a peak in obesity prevalence among women aged 50–65 [8], our findings add to this observation by revealing that this peak arises from the convergence of two distinct subgroups: the “increasing–decreasing” individuals, who were already obese, and the “increasing” individuals, who newly transitioned into obesity, thereby driving the overall prevalence to its highest level. Menopause plays a significant role in reshaping body composition across all BMI trajectory groups. The decline in estrogen triggers visceral abdominal fat accumulation and lean mass reduction. This fat-lean mass counterbalance was most evident in the stable trajectory group, where BMI remained relatively constant despite metabolic changes [38–40]. However, without lifestyle modifications during this period, women are prone to weight and visceral fat gain [41], potentially placing their BMI on an increasing trajectory toward obesity. These physiological changes highlight menopause as a pivotal period for abdominal adiposity development, a primary driver of MetS pathogenesis. Indeed, our study confirms that participants with MetS generally exhibited an enlarged WC, second in prevalence only to hypertension among MetS components. This was particularly evident among women in the increasing trajectory group, whose BMI began to rise during menopause, showing significantly higher WC than those in the increasing–decreasing trajectory group. The obesity threshold (BMI ≥ 30 kg/m2) [42] effectively stratified MetS risk. Participants in the non-stable trajectory groups progressively exceeded this threshold, reinforcing the positive correlation between obesity duration and MetS risk. Although statistically non-significant, the trend toward a higher OR in the increasing–decreasing trajectory group compared with the increasing trajectory group may reflect greater cumulative exposure to obesity, as the former group experienced an earlier and steeper BMI increase. This finding aligns with existing evidence indicating that the rate of BMI change may serve as a more sensitive predictor of cardiometabolic outcomes than static BMI levels [43]. Our study highlights the significant association between obesity and MetS at the population level. This finding is consistent with established pathophysiological mechanisms, among which chronic low-grade inflammation serves as a critical bridge between obesity and metabolic diseases [44]. Obesity fundamentally reprograms the immune microenvironment of adipose tissue: whereas lean adipose tissue maintains insulin sensitivity via anti-inflammatory M2 macrophages, obesity promotes a shift toward pro-inflammatory M1 macrophages. This transition generates mediators such as TNF-α that sustain inflammation. Mechanistically, this shift is driven by at least three key pathways: first, increased intestinal permeability allows gut-derived lipopolysaccharides to activate TLR4 receptors; second, elevated free fatty acids engage TLR4/TLR2 via fetuin A, activating NF-κB and JNK1 pathways; and third, tissue expansion creates hypoxia, which induces HIF-1α-mediated inflammatory signaling. These mechanisms converge to impair insulin signaling and establish a self-perpetuating inflammatory network across multiple metabolic organs that contributes to insulin resistance in the liver, disrupts insulin secretion in pancreatic islets, and alters energy homeostasis in the hypothalamus [44]. This framework of multi-organ inflammation provides a compelling pathophysiological explanation for the robust obesity-MetS link we observed in our study. We also observed that reproductive factors, particularly early menopause, early age at first live birth, and high parity, significantly modified the relationship between BMI trajectories and MetS risk. In subgroup analyses, women with early menopause, early age at first live birth, or high parity exhibited the strongest MetS risk when their BMI trajectory followed a persistently increasing pattern, transitioning into obesity during menopause. Notably, this contrasted with the overall analysis, where the increasing–decreasing trajectory group conferred the highest MetS risk. This shift in the risk pattern suggests that these reproductive factors may amplify metabolic vulnerability during hormonal transitions, potentially by exacerbating body composition changes. In particular, central obesity during menopause is a well-known driver of MetS, underscoring the importance of weight management in this period. These findings highlight that both the timing and the frequency of reproductive events critically shape long-term MetS risk. Epidemiological studies confirm that early menopause, early age at first live birth, or high parity independently increases susceptibility to weight gain and MetS [45–50]. Mechanistically, early menopause extends exposure to estrogen deficiency, which promotes visceral adiposity, insulin resistance, and inflammation. Similarly, early age at first live birth induces metabolic stress at a physiologically immature stage, potentially leading to persistent adverse metabolic effects. Likewise, high parity imposes a cumulative gestational load, characterized by repeated episodes of pregnancy-induced insulin resistance and dyslipidemia, which can lead to lasting adverse cardiometabolic remodeling and postpartum weight retention. Consequently, women with early or frequent reproductive milestones warrant tailored metabolic monitoring to mitigate these risks. To the best of our knowledge, this is the first study to evaluate the association between BMI trajectories and MetS risk in postmenopausal women. Furthermore, our subgroup analyses stratified by reproductive characteristics provide novel insights into how BMI dynamics influence metabolic health in this population. The key strengths of our study include the use of a nationally representative sample, accurate classification of menopausal status, comprehensive adjustment for metabolic and reproductive confounders, and longitudinal BMI assessment from adulthood to postmenopause. Several limitations should be noted. First, the significant time span between BMI assessments posed challenges to precise trajectory tracking, highlighting the need for more frequent BMI assessments in future studies. Second, the use of self-reported historical weight and the assumption of constant height to calculate past BMI may introduce recall bias and systematic bias, underscoring the importance of objectively measured longitudinal data. Third, the cross-sectional NHANES design further limits causal inference despite longitudinal trajectory modeling. Fourth, our findings from a US population may lack generalizability to other populations, and external validation is strongly recommended to confirm our results. Finally, the current study may not have fully explored the biological mechanisms underlying BMI fluctuations, and future research could integrate genomics, metabolomics and behavioral data to provide a more comprehensive understanding of the biological basis of BMI changes. Conclusions In conclusion, our study demonstrated that long-term BMI trajectories from adulthood to postmenopause exhibit distinct patterns that were significantly associated with MetS risk in postmenopausal women. Specifically, both the increasing and increasing–decreasing trajectory groups were linked to substantially higher risk of MetS compared with the stable trajectory group. Reproductive factors were found to modulate the relationship between BMI trajectories and MetS risk. These findings enhance our understanding of the long-term metabolic impact of BMI patterns and underscore the importance of incorporating long-term BMI monitoring alongside reproductive history assessment for early MetS risk stratification. Tailored interventions for high-risk subgroups are essential to preventing the development of MetS. Supplementary Information Supplementary Material 1: Table S1. Characteristics of the study participants by MetS status. Table S2. Model fit statistics for competing latent class growth mixture models. Table S3. Logistic regression for the association between BMI trajectories and MetS risk (Complete MetS data, JIS 2009 criteria). Table S4. Logistic regression for the association between BMI trajectories and MetS risk (Complete MetS data, IDF 2005 criteria). Supplementary Material 1: Table S1. Characteristics of the study participants by MetS status. Table S2. Model fit statistics for competing latent class growth mixture models. Table S3. Logistic regression for the association between BMI trajectories and MetS risk (Complete MetS data, JIS 2009 criteria). Table S4. Logistic regression for the association between BMI trajectories and MetS risk (Complete MetS data, IDF 2005 criteria).