Trends in type 2 diabetes, prediabetes, and hyperglycemic recognition among menopausal women: A 20-year analysis of sociodemographic disparities

Abstract

Background

Menopausal hormonal and metabolic changes may increase susceptibility to type 2 diabetes (T2D), yet long-term national trends and socio-demographic disparities in both T2D burden and hyperglycemia recognition among menopausal women remain insufficiently characterized.

Objectives

To examine 20-year trends in T2D, prediabetes, and recognition of prediabetes- or diabetes-range hyperglycemia among U.S. menopausal women and to assess sociodemographic disparities associated with abnormal glycemic status.

Design

Repeated cross-sectional analysis of continuous National Health and Nutrition Examination Survey (NHANES) cycles (2003-2023).

Methods

The analytic sample included 3,574 naturally menopausal women aged ≥45 years with fasting plasma glucose, HbA1c, and diabetes questionnaire data. T2D and prediabetes were defined using ADA-aligned biomarker thresholds. Hyperglycemia was operationalized specifically as biomarker-defined prediabetes- or diabetes-range hyperglycemia (fasting plasma glucose ≥100 mg/dL or HbA1c ≥5.7%). Recognition status was categorized as recognized hyperglycemia, unrecognized hyperglycemia, or normoglycemia based on glycemic biomarkers and self-reported prior clinician communication of diabetes or borderline diabetes. Survey-weighted prevalence estimates, trend models across cycles, and multinomial/survey logistic regression assessed associations with race/ethnicity, education, marital status, employment, and poverty-to-income ratio (PIR).

Results

Overall crude prevalence was 13.7% for diabetes and 34.8% for prediabetes. From 2003 to 2023, diabetes rose modestly from 11.3% to 11.9% (p<.001), while prediabetes increased substantially from 24.3% to 35.2% (p<.001). Unrecognized hyperglycemia increased from 23.9% to 32.9% (p<.001). Higher odds of diabetes were observed for non-Hispanic Black (OR=4.03) and Hispanic/Latina women (OR=2.81) versus non-Hispanic White women, for ≤high school education (OR=2.77) versus ≥4-year degree, for disability (OR=1.46) and retirement (OR=1.50) versus employment, and for PIR<1.3 (OR=1.54) versus PIR>3.5; marital status was not associated. Similar patterns were observed for recognized and unrecognized hyperglycemia.

Conclusion

Among menopausal women, prediabetes and unrecognized hyperglycemia increased over two decades, with pronounced disparities by race/ethnicity and socioeconomic position. Integrating routine glycemic screening and tailored, culturally responsive prevention efforts into menopausal care may reduce unrecognized abnormal glycemic status and related inequities.

Plain language summary

This study looked at changes in type 2 diabetes, prediabetes, and high blood sugar among menopausal women in the United States over 20 years, from 2003 to 2023. Using national health survey data from 3,574 women aged 45 years and older who had gone through natural menopause, the study found that prediabetes became much more common over time, while diabetes increased slightly. Many women had blood sugar levels in the prediabetes or diabetes range but did not report that a health professional had previously discussed diabetes or high blood sugar with them. The study also found that these problems were not shared equally across all groups. Black and Hispanic/Latina women, women with lower education levels, women with lower incomes, and women who were disabled or retired were more likely to have diabetes or high blood sugar that had not been recognized. These findings suggest that blood sugar screening should be a routine part of health care for menopausal women. Clear communication, culturally appropriate education, and earlier support may help reduce missed cases and improve diabetes prevention.

Introduction

Type 2 diabetes (T2D) is recognized globally as a significant public health concern, characterized by insulin resistance or impaired insulin secretion.^1,2 According to the International Diabetes Federation (IDF), approximately 537 million adults worldwide were living with diabetes in 2021, and this figure is projected to rise dramatically to 643 million by 2030 and 783 million by 2045. ³ In the United States (US), over 37 million individuals (approximately 11.3% of the population) were reported to have diabetes in 2019, with the large majority diagnosed with T2D. ⁴ Typical symptoms of T2D include increased thirst, frequent urination, fatigue, blurred vision, weight loss or gain, and delayed wound healing. ⁵ Long-term complications associated with uncontrolled diabetes include cardiovascular disease, peripheral neuropathy, diabetic nephropathy, diabetic retinopathy, diabetic foot ulcers, and increased mortality risk. ⁶ Proper management involving lifestyle interventions, regular monitoring, and pharmacological treatment is essential for minimizing complications, improving quality of life, and reducing the economic burden associated with diabetes care. ⁷

Globally, diabetes prevalence among women has reached critical levels, with approximately 10.2% of adult women affected in 2021, representing about 240 million women worldwide. ⁸ In the US, roughly 14% of adult women live with diabetes. ⁵ Effective T2D management is particularly critical among middle-aged and older women because aging and the menopausal transition coincide with changes in body composition, insulin sensitivity, glucose metabolism, and cardiovascular risk.^9–11 These overlapping biological and life-course factors may increase susceptibility to T2D and accelerate age-related health complications.^12,13 Prior research has also demonstrated that women in the menopausal transition and postmenopausal women may be especially vulnerable to T2D due to the decline in estrogen levels, accumulation of visceral adiposity, alterations in glucose metabolism, and age-related reductions in physical activity, ¹⁴ all of which contribute substantially to the increased diabetes risk in this population.

Menopause is an inevitable and profound phase of biological transformation that every woman experiences once in her lifetime. ¹⁵ Menopause, defined as the permanent cessation of menstruation resulting from ovarian follicle depletion, ¹⁶ typically occurs between 45 and 55 years of age and represents a critical biological transition in a woman’s life.^17–19 This universal transition marks the end of reproductive capacity and ushers in a period of extensive hormonal, physiological, and metabolic changes that have lasting impacts on women’s health.^20,21 Notably, estrogen deficiency during menopause plays a key role in modulating insulin sensitivity and glucose homeostasis by promoting visceral fat accumulation and impairing pancreatic beta-cell function, ²² both of which are central mechanisms underlying the increased risk of T2D. ²³ Despite menopause being recognized as a critical life event with clear biological implications for metabolic health, existing research remains inadequate in fully understanding the complex interactions between menopause and areas related to T2D symptom recognition, disease progression, and the development of effective self-management strategies within this population.

Furthermore, extensive research has consistently demonstrated sociodemographic disparities in T2D prevalence both globally and within the United States,^24,25 revealing significant correlations with sociodemographic indicators such as race/ethnicity,^26,27 household income level,^28,29 and educational attainment.^30,31 Specifically, one prior study has found that lower socioeconomic status is consistently linked to a higher prevalence of T2D, increased severity of the disease, and greater incidence of diabetes-related complications such as cardiovascular disease, nephropathy, and neuropathy. ³² Individuals with lower educational levels often have limited health literacy, reduced capacity for effective self-management, and lower adherence to prescribed diabetes treatment and preventive health behaviors, consequently exacerbating disease outcomes.^33,34 However, a significant and persistent gap exists in the scientific literature concerning how these socioeconomic disparities specifically affect menopausal women’s vulnerability to and management of T2D. Despite acknowledgment of the combined impact of menopause-related hormonal changes, aging, and diabetes risk, current studies have not provided comprehensive analyses that explicitly address the interplay between menopause, sociodemographic status, and diabetes prevalence. This critical gap highlights the urgent necessity for targeted research addressing sociodemographic disparities explicitly within menopausal women populations to inform and refine public health interventions effectively.

This study aims to address a critical research gap by systematically analyzing 20-year trends in T2D and prediabetes prevalence among naturally menopausal women in the US. In addition, this study examines recognition of biomarker-defined prediabetes- or diabetes-range hyperglycemia by differentiating between participants with abnormal glycemic biomarkers who reported prior clinician communication of diabetes or borderline diabetes and those with abnormal biomarkers who did not. This recognition-status approach adds information beyond diabetes and prediabetes prevalence alone because it distinguishes the biological burden of abnormal glycemia from whether that abnormal glycemic status had been previously communicated or recognized in clinical care. This distinction is important given recent NHANES-based evidence showing that prediabetes awareness remains low among US adults and that most adults with prediabetes are unaware of their condition. ³¹ Lastly, by conducting subgroup analyses based on race/ethnicity, educational attainment, marital status, employment status, and poverty-to-income (PIR) ratio, this study seeks to clarify how sociodemographic factors are associated with both glycemic burden and recognition gaps among menopausal women. This investigation has potential to inform more precise public health strategies to improve disease detection and management among this vulnerable population.

Methods

Study design

This study utilized the continuous National Health and Nutrition Examination Survey (NHANES), an ongoing cross-sectional survey employing a complex, multistage, stratified, clustered sampling design, which is representative of the civilian, non-institutionalized US population.^35–37 The NHANES protocol was approved by the Institutional Review Board of the Centers for Disease Control and Prevention (CDC), and all participants provided written informed consent. Comprehensive details about the survey methods and procedures are available on the NHANES website. ³⁸ Each NHANES cycle includes a national sample of approximately 10,000 individuals across all age groups and genders, with oversampling of individuals aged 60 and older, African Americans, and Hispanics to ensure reliable estimates. The continuous NHANES is structured into 2-year data cycles; due to the coronavirus disease 2019 (COVID-19) pandemic, data collected from the 2019 to March 2020 cycle were merged with data from the 2017-2018 NHANES cycle to create a nationally representative sample of NHANES 2017-March 2020 pre-pandemic data. ³⁷ This study was reported in accordance with the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guideline for cross-sectional studies. ³⁹

Data source

This study used publicly available, de-identified NHANES public-use data files obtained from the Centers for Disease Control and Prevention/National Center for Health Statistics (CDC/NCHS) website. NHANES public-use datasets, questionnaires, codebooks, and documentation are available by survey cycle and can be downloaded without a formal application or fee. No restricted-use NHANES data were used in the present analysis.

Study sample

This study used 20 years of data from NHANES, which was conducted from 2003 to 2023, by incorporating menopausal women based on common variables. According to the previously published studies, the median menopause age was reported to be 54 years in Europe, 51.4 years in North America, 48.6 years in Latin America, and 51.1 years in Asia. ⁴⁰ Since the distribution of menopause age was normally distributed from 40 to 54 years in previous community-based studies,^17–19 it was generally found to cluster between 45 and 55 years of age. Therefore, based on these statistics and previous findings, this study considered 45 years old and older as the general natural menopause period (i.e., menopausal due to surgery was not included in this study). Natural menopause was operationalized using the NHANES reproductive health questionnaire. Participants were classified as naturally menopausal if they reported no regular menstrual periods during the past 12 months and selected menopause as the reason for not having regular periods. Women who continued to report regular menstrual periods, women in the perimenopausal transition who did not meet this operational definition, and women whose absence of regular periods was attributable to non-menopausal reasons were excluded. Women with a history of hysterectomy or bilateral oophorectomy were also excluded to focus on natural menopause.

Inclusion and exclusion criteria

The inclusion criteria for this study were: 1) women over 45 years of age who reported “No” to the question “Had regular periods in the past 12 months?” and select “Menopause” as the reason for not having regular periods was classified as experiencing natural menopause; 2) women who provided a diagnosis of T2D through questionnaire; and 3) women who provided fasting plasma glucose concentration or HbA1c levels. However, participants were excluded from the study based on the following criteria: 1) women who have type 1 diabetes; 2) participants who responded “Yes” to the questions “Had a hysterectomy?” or “Had both ovaries removed?”; and 3) a diagnosis of diabetes at an age younger than 30 years.

Participants who reported diabetes diagnosis before age 30 years were excluded to reduce potential misclassification of type 1 diabetes as T2D. This age-based criterion was used as a conservative epidemiologic approach rather than as a definitive clinical diagnostic rule. The ADA Standards of Care indicate that type 1 and type 2 diabetes can occur across age groups and that diabetes subtype classification ideally incorporates multiple clinical features, including age at diagnosis, body mass index (BMI), ketoacidosis, autoantibody status, C-peptide, and insulin requirement. ⁴¹ Since subtype-confirming information was not consistently available across all NHANES cycles, this exclusion criterion was applied in line with prior NHANES-based studies that used diagnosis before age 30 years, alone or with insulin-use information, to reduce potential type 1 diabetes misclassification.^42–45 Although early-onset T2D is increasingly recognized, this conservative criterion was selected to improve specificity for T2D classification in the present secondary analysis. Also, women who are pregnant or missing data on key variables were excluded from the data. Across the 2003-2023 NHANES cycles, 98,551 participants were examined. Of these, 15,079 had complete data for the primary variables. This study sequentially excluded participants aged <45 years (n = 7,151), women who were not menopausal (n = 2,807), those without natural menopause (history of hysterectomy and/or bilateral oophorectomy; n = 1,204), and individuals who did not meet the diabetes biomarker criterion or had type 1 diabetes (n = 343). The analyzed final sample comprised 3,574 menopausal women. The participant selection process is detailed in Figure 1.

Figure 1.

Flow chart of sample selection in national health and nutrition examination survey 2003-2023.

Sample size considerations

Since this study was a secondary analysis of publicly available NHANES data, the analytic sample size was determined by the number of participants who met the predefined eligibility criteria and had complete data for the primary exposure, outcome, and covariate variables, rather than by a priori sample size calculation. After applying the inclusion and exclusion criteria, the final analytic sample included 3,574 naturally menopausal women. This sample included 645 participants with diabetes, 1,353 with prediabetes, and 1,576 with normoglycemia. For the classification of hyperglycemia recognition, 699 participants were classified as having recognized hyperglycemia, 1,299 as having unrecognized hyperglycemia, and 1,576 as having normoglycemia. The adequacy of the analytic sample for multivariable survey logistic regression was assessed using the events-per-variable framework. Prior simulation studies have suggested that logistic regression models generally require an adequate number of outcome events per model parameter to reduce bias and instability in regression estimates, with 10 events per variable commonly used as a practical benchmark, although this threshold is not an absolute rule.^46,47 In the present study, the smallest primary outcome category was diabetes (n = 645). Considering the number of covariate parameters included in the multivariable models, the available number of outcome events exceeded this commonly cited benchmark, supporting the stability of the model estimates.

To assess representativeness, characteristics of the included analytic sample (n = 3,574) were compared with the excluded population (N = 11,505). The analytic sample was older on average (mean 65.3 vs 46.8 years; p < .001) and included a higher proportion of non-Hispanic White participants (74.9% vs 42.4%), with lower proportions of non-Hispanic Black (8.6% vs 21.7%) and Hispanic or Latino participants (10.1% vs 25.0%) (all p < .001). Marital status and labor force status also differed: the analytic cohort included more married or partnered respondents (57.4% vs 48.7%; p < .001), fewer never married (15.6% vs 21.4%; p < .001), fewer employed (43.0% vs 48.9%; p < .001), and more retired (37.7% vs 18.1%; p < .001). Annual household income distribution favored higher socioeconomic status in the analytic sample, with a greater share above a poverty to income ratio of 3.5 (40.4% vs 25.9%; p < .001), whereas educational attainment did not differ materially between groups (p = 0.175) (Table 1).

Table 1.

Descriptive statistics for subjects included and excluded from analysis.

Variable	Excluded sample N =11,505		Included sample n = 3,574		p
	Mean or N	SD or weighted %	Mean or n	SD or weighted %
Age (year)	46.8	19.8	65.3	9.4	<.001
Race/Ethnicity	​				<.001
Non-Hispanic White	4,605	42.4%	1,784	74.9%	​
Non-Hispanic Black	2,681	21.7%	591	8.6%	​
Hispanic/Latino	2,902	25.0%	862	10.1%	​
Others	1,317	11.0%	337	6.4%	​
Education Level	​				.175
High School or Less	4,211	39.6%	1,759	42.1%	​
Some High School/2-Year Degree/No degree	3,184	27.6%	978	28.8%	​
4-Year Degree or More	2,478	22.0%	835	29.1%	​
Marital Status	​				<.001
Married/Partnered	5,562	48.7%	1,783	57.4%	​
Never Married	2,559	21.4%	669	15.6%	​
Divorced/Separated/Widowed	2,314	22.8%	1,120	27.0%	​
Employment	​				<.001
Employed	6,120	48.9%	1,256	43.0%	​
Disabled	867	8.4%	403	8.6%	​
Retired	1,307	18.1%	1,428	37.7%	​
Not Employed	3,197	24.4%	481	10.8%	​
Poverty-to-income ratio	​		​		<.001
< 1.3	4,745	40.0%	1,289	25.6%	​
1.3-3.5	3,901	34.1%	1,241	34.0%	​
>3.5	2,859	25.9%	1,044	40.4%	​

Note. SD: Standard Deviation.

Measures

Type 2 diabetes (T2D) and prediabetes status

Complete blood counts, including hemoglobin, were analyzed from whole blood specimens collected at the NHANES mobile examination centers, which serve as traveling clinics. ⁴⁸ All assays were conducted using a High-Performance Liquid Chromatography (HPLC) system.^49,50 The National Center for Health Statistics has confirmed that all NHANES laboratories met the criteria set by the National Glycohemoglobin Standardization Program (NGSP).^49,50 Detailed procedures for specimen collection and processing are outlined in the NHANES Laboratory Procedures Manual. ⁴⁸

In T2D status, participants were classified into three groups, diabetes, prediabetes, and normoglycemic, based on biomarker criteria. Biomarkers included fasting plasma glucose (FPG) and glycosylated hemoglobin (HbA1c) levels. Following the American Diabetes Association (ADA) (2014) guidelines, individuals were categorized as prediabetic if their FPG ranged between 100 to 125 mg/dL and diabetic if their FPG was ≥126 mg/dL. ⁵¹ Similarly, HbA1c thresholds identified prediabetes (HbA1c levels between 5.7% and 6.4%) ⁵² and diabetes (HbA1c ≥6.5%). ⁵³ Participants who did not meet these biomarker thresholds were classified as normoglycemic.

Recognition status of prediabetes- or diabetes-range hyperglycemia

For the secondary recognition-status analysis, hyperglycemia was operationalized specifically as biomarker-defined prediabetes- or diabetes-range. This definition included fasting plasma glucose ≥100 mg/dL or HbA1c ≥5.7%, corresponding to ADA-aligned thresholds for prediabetes- or diabetes-range glycemia. ⁵¹ The present analysis was limited to hyperglycemia and did not assess hypoglycemia or glycemic variability.

Recognition status was determined by combining biomarker-defined hyperglycemia with self-reported prior clinician communication of diabetes or borderline diabetes. Self-reported clinician communication was assessed using NHANES DIQ010, which asks participants, “Other than during pregnancy, have you ever been told by a doctor or health professional that you have diabetes or sugar diabetes?” Response options included “Yes,” “No,” “Borderline,” “Refused,” and “Don’t know.” Participants with biomarker-defined prediabetes- or diabetes-range hyperglycemia who answered “Yes” or “Borderline” were classified as having recognized hyperglycemia. The “Borderline” response was grouped with “Yes” because it may reflect the participant reported some prior diabetes-related or elevated-glycemic-risk communication for recognition. Participants with biomarker-defined hyperglycemia who answered “No” were classified as having unrecognized hyperglycemia. Participants who answered “No” and did not meet biomarker criteria for prediabetes- or diabetes-range hyperglycemia were classified as normoglycemic. Responses of “Refused” or “Don’t know” were coded as missing and excluded from this analysis.

These categories should not be interpreted as direct measures of subjective awareness, diabetes knowledge, risk perception, or health literacy because DIQ010 does not assess participants’ understanding of glycemic risk. Rather, the categories represent recognition status based on the presence or absence of self-reported prior clinician communication among participants with biomarker-defined prediabetes- or diabetes-range hyperglycemia. Table 2 presents the classification of recognition status for biomarker-defined prediabetes- or diabetes-range hyperglycemia.

Table 2.

Classification of recognition status for biomarker-defined prediabetes- or diabetes-range hyperglycemia.

Biomarker-defined prediabetes- or diabetes-range hyperglycemia	Self-reported prior clinician communication (DIQ010)	Classification	Interpretation
FPG ≥100 mg/dL or HbA1c ≥5.7%	Reported	Recognized hyperglycemia	Abnormal glycemic biomarkers with self-reported prior clinician communication
FPG ≥100 mg/dL or HbA1c ≥5.7%	Not reported	Unrecognized hyperglycemia	Abnormal glycemic biomarkers without self-reported prior clinician communication
FPG <100 mg/dL and HbA1c <5.7%	Not reported	Normoglycemia	No biomarker-defined hyperglycemia and no self-reported prior diabetes/borderline diabetes

Note. FPG = fasting plasma glucose; HbA1c = glycated hemoglobin. DIQ010 asks whether participants had ever been told by a doctor or health professional that they had diabetes or sugar diabetes, other than during pregnancy.

Sociodemographic information

Age (45 years and older), race/ethnicity (non-Hispanic White, non-Hispanic Black, Hispanic/Latina, and other), education level (high school or less, some high school/2-year degree/no degree, and 4-year degree or more), marital status (married/partnered, never married, and divorced/separated/widowed), employment (employed, disabled, retired, and not employed), and PIR (<1.3, 1.3-3.5, and >3.5) were included to examine sub-group analysis.

Statistical analysis

This study included a cohort of adults from nine NHANES 2-year survey cycles (2003-2023) consisting of 3,574 menopausal women with clinical, demographic, and laboratory data, which were divided into diabetes, prediabetes, and normoglycemic groups, and also were divided into recognized hyperglycemia, unrecognized hyperglycemia, and normoglycemic groups. The prevalence of T2D and recognition of hyperglycemia, classified by T2D status (diabetes, prediabetes, and normoglycemic) and recognition status of biomarker-defined prediabetes- or diabetes-range hyperglycemia (recognized hyperglycemia, unrecognized hyperglycemia, and normoglycemic), was estimated as percentages with corresponding 95% confidence intervals (CI) using SAS software version 9.4.0 (SAS Institute, Cary, NC, USA). All analyses accounted for the effects of complex sample design, including multi-stage sampling, stratification, clustering, and the application of sampling weights. The standard error was estimated using the Taylor series (linearization) method, which incorporates the complexities of the sampling design. To test for equality in the prevalence of categorical variables, the Rao-Scott χ²-test, a design-adjusted version of the Pearson χ²-test, was employed. Mean differences for continuous variables were assessed using PROC ANOVA. The PROC SURVEYREG procedure, as outlined in NHANES tutorials, was used to examine the prevalence of T2D from 2003 to 2023, with the 2000 Census as the reference population.

Linear trends in the prevalence of T2D, prediabetes, and recognized and unrecognized hyperglycemia were analyzed using regression models, treating the 2-year survey cycle as a continuous variable, while adjusting for potential confounding variables such as age, education level, marital status, and household income. Then, multinomial survey logistic regression (PROC SURVEYLOGISTIC) models were used to examine sociodemographic factors (e.g., race/ethnicity, education level, marital status, employment status, and household income) associated with both T2D status and recognition status of biomarker-defined hyperglycemia. Statistical significance was set at a p-value of < .05.

Results

The crude prevalence of diabetes and prediabetes were 13.7% and 34.8%, respectively. The overall crude prevalence for diabetes and prediabetes increased significantly from 11.3% and 24.3% in 2003 to 11.9% (p < .001) and 35.2% (p < .001) in 2023, respectively (Figure 2(a)). Additionally, the crude prevalence of recognized and unrecognized hyperglycemia increased significantly from 11.7% and 23.9% in 2003 to 14.3% and 32.9% in 2023, respectively (both p < .001; Figure 2(b)).

Figure 2.

The prevalence of diabetes status (a) and recognition status of biomarker-defined prediabetes- or diabetes-range hyperglycemia (b), NHANES 2003-2023.

The characteristics of participants were stratified by T2D status (diabetes, prediabetes, and normoglycemic) (Table 3). Menopausal women with diabetes were significantly younger (65.8 ± 9.0 years) compared to those with prediabetes (66.2 ± 9.3 years) and normal glucose levels (64.5 ± 9.5 years; p < .001). Significant racial and ethnic disparities were observed, with non-Hispanic Black menopausal women exhibiting the highest prevalence of diabetes (27.4%), followed by Hispanic/Latina (20.4%), others (19.8%), and non-Hispanic White (10.7%; p < .001).

Table 3.

Sample characteristics based on type 2 diabetes status.

Variable	T2D status (N = 3,574) Unweighted N (weighted %)			p
	Diabetes (N = 645)	Prediabetes (N = 1,353)	Normoglycemic (N = 1,576)
Age (year; Mean ± SD)	65.8 ± 9.0	66.2 ± 9.3	64.5 ± 9.5	<.001
Race/Ethnicity	​			<.001
Non-Hispanic White	213 (10.7)	624 (33.4)	947 (55.9)	​
Non-Hispanic Black	163 (27.4)	248 (40.0)	180 (33.6)	​
Hispanic/Latina	201 (20.4)	343 (40.5)	318 (39.2)	​
Others	68 (19.8)	138 (37.2)	131 (43.1)	​
Education Level	​			<.001
High School or Less	419 (19.4)	684 (38.5)	656 (42.1)	​
Some High School/2-Year Degree/No degree	150 (12.2)	384 (34.6)	444 (53.3)	​
4-Year Degree or More	76 (7.0)	284 (29.8)	475 (63.3)	​
Marital Status	​			.003
Married/Partnered	285 (11.7)	673 (33.8)	825 (54.5)	​
Never Married	213 (16.1)	454 (37.4)	453 (46.5)	​
Divorced/Separated/Widowed	147 (16.9)	226 (34.0)	296 (49.1)	​
Employment	​			<.001
Employed	186 (12.0)	433 (29.9)	637 (58.1)	​
Disabled	111 (21.4)	136 (30.9)	156 (47.7)	​
Retired	250 (13.7)	588 (41.1)	590 (45.2)	​
Not Employed	97 (14.3)	194 (35.4)	190 (50.3)	​
PIR	​			<.001
< 1.3	308 (20.6)	503 (37.8)	478 (41.6)	​
1.3-3.5	219 (14.2)	502 (40.2)	520 (45.6)	​
>3.5	118 (8.9)	348 (28.4)	578 (62.7)	​

Note. T2D: Type 2 Diabetes, SD: Standard Deviation, PIR: Poverty-to-Income Ratio.

Education level showed significant differences, where participants with a high school education or less reported the highest diabetes prevalence (19.4%), and participants with a 4-year degree or more had the lowest (7.0%; p < .001). Individuals who were divorced, separated, or widowed exhibited a higher diabetes prevalence (16.9%) compared to those who were married or partnered (11.7%) and those who had never married (16.1%) (p = .003). Employment status revealed disabled menopausal women had the highest prevalence of diabetes (21.4%), compared to employed (12.0%), retired (13.7%), and not employed individuals (14.3%; p < .001). Annual household income levels based on the PIR indicated the highest diabetes prevalence in women with under 1.3 (20.6%; p < .001). See Table 3 for a detailed breakdown of the sample characteristics based on T2D status.

Additionally, menopausal women with recognized hyperglycemia had a mean age of 66.1 ± 9.0 years, significantly older compared to the normoglycemic women (64.5 ± 9.5 years; p < .001). Racial and ethnic differences were significant, with Hispanic/Latina (22.3%) and non-Hispanic Black (26.3%) participants exhibiting higher proportions of recognized hyperglycemia compared to non-Hispanic White (13.3%; p < .001). Education level differences were also significant, showing higher rates of recognized hyperglycemia among menopausal women with high school education or less (20.8%) compared to those with a 4-year degree or more (9.2%; p < .001).

Marital status was significantly associated with hyperglycemic recognition, showing the highest prevalence among divorced, separated, or widowed women (20.5%; p = .001). Employment status revealed significant differences, with disabled participants exhibiting the highest rates of recognized hyperglycemia (24.2%; p < .001). Annual household income levels showed the highest proportion of recognized hyperglycemia among individuals with a PIR <1.3 (22.2%; p < .001). A detailed summary of sample characteristics according to hyperglycemic recognition status (i.e., recognized hyperglycemia, unrecognized hyperglycemia, normoglycemic) is provided in Table 4.

Table 4.

Sample characteristics based on hyperglycemic recognition status.

Variable	Hyperglycemic recognition status (N = 3,574) Unweighted N (weighted %)			p
	Recognized hyperglycemia (N = 699)	Unrecognized hyperglycemia (N = 1,299)	Normoglycemic (N = 1,576)
Age (year; Mean ± SD)	66.1 ± 9.0	66.0 ± 9.3	64.5 ± 9.5	<.001
Race/Ethnicity	​			<.001
Non-Hispanic White	257 (13.3)	580 (30.8)	947 (55.9)	​
Non-Hispanic Black	159 (26.3)	252 (40.1)	180 (33.6)	​
Hispanic/Latina	211 (22.3)	333 (38.6)	318 (39.2)	​
Others	72 (23.1)	134 (33.8)	131 (43.1)	​
Education Level	​			<.001
High School or Less	432 (20.8)	671 (37.1)	656 (42.1)	​
Some High School/2-Year Degree/No degree	177 (15.9)	357 (30.9)	444 (53.3)	​
4-Year Degree or More	90 (9.2)	270 (27.5)	475 (63.3)	​
Marital Status	​	​	​	.001
Married/Partnered	302 (13.6)	656 (32.0)	825 (54.5)	​
Never Married	243 (18.5)	424 (35.0)	453 (46.5)	​
Divorced/Separated/Widowed	154 (20.5)	219 (30.4)	296 49.1)	​
Employment	​			<.001
Employed	181 (13.6)	438 (28.3)	637 (58.1)	​
Disabled	130 (24.2)	117 (28.1)	156 (47.7)	​
Retired	280 (16.4)	558 (38.5)	590 (45.2)	​
Not Employed	106 (17.2)	185 (32.5)	190 (50.3)	​
PIR	​			<.001
< 1.3	326 (22.2)	485 (36.2)	478 (41.6)	​
1.3-3.5	249 (17.9)	472 (36.5)	520 (45.6)	​
>3.5	124 (10.4)	342 (26.9)	578 (62.7)	​

Note. SD: Standard Deviation, PIR: Poverty-to-Income Ratio.

According to the findings from the survey logistic regression analyses identifying factors associated with T2D status (Table 5), non-Hispanic Black menopausal women demonstrated significantly higher likelihood of diabetes (OR = 4.03, p < .001), as did Hispanic menopausal women (OR = 2.81, p < .001), and those from other racial/ethnic groups (i.e., non-Hispanic Asian, multi-racial individuals, and others) (OR = 2.30, p < .001) compared to non-Hispanic White counterparts. Menopausal women having a high school education or less showed the greatest odds (OR = 2.77, p < .001), followed by those with some high school education, a 2-year degree, or no degree (OR = 1.79, p < .001), compared to those holding a 4-year degree or more. However, marital status did not show an association with diabetes (p > .05). Employment status was significantly associated with diabetes risk, where disabled (OR = 1.46, p < .01) and retired menopausal women (OR = 1.50, p < .05) were more likely to have diabetes compared to employed individuals. Annual household income levels also indicated increased diabetes risk, particularly among those with a PIR below 1.3 (OR = 1.54, p < .01).

Table 5.

Survey logistic regression based on type 2 diabetes status.

Variable	T2D status
	Diabetes			Prediabetes
	β	S.E.	OR (95% CI)	β	S.E.	OR (95% CI)
Race/Ethnicity
Non-Hispanic Black	1.39	1.20	4.03*** (3.11-5.22)	0.74	0.11	2.09*** (1.68-2.60)
Hispanic/Latina	1.03	1.20	2.81*** (2.23-3.54)	0.49	0.11	1.64*** (1.36-2.00)
Others	0.84	1.21	2.30*** (1.66-3.21)	0.47	0.12	1.60*** (1.23-2.07)
Non-Hispanic White	Ref.			Ref.
Education Level
High School or Less	1.02	0.16	2.77*** (2.04-3.76)	0.32	0.11	1.37** (1.11-1.69)
Some High School/2-Year Degree/No degree	0.58	0.16	1.79*** (1.30-2.46)	0.26	0.11	1.29* (1.05-1.60)
4-Year Degree or More	Ref.			Ref.
Marital Status
Never Married	-0.05	0.12	0.95 (0.76-1.20)	0.03	0.09	1.03 (0.87-1.23)
Divorced/Separated/Widowed	0.17	0.13	1.19 (0.92-1.53)	-0.14	0.11	0.87 (0.70-1.07)
Married/Partnered	Ref.			Ref.
Employment Status
Not Employed	0.19	0.16	1.21 (0.88-1.65)	0.23	0.13	1.26 (0.99-1.61)
Disabled	0.38	0.12	1.46** (1.15-1.85)	0.40	0.09	1.49*** (1.25-1.77)
Retired	0.41	0.16	1.50* (1.09-2.07)	0.02	0.14	1.02 (0.77-1.35)
Employed	Ref.			Ref.
PIR
<1.3	0.43	0.15	1.54** (1.15-2.04)	0.28	0.11	1.32* (1.07-1.63)
1.3-3.5	0.28	0.14	1.32* (1.00-1.73)	0.28	0.10	1.33** (1.09-1.61)
>3.5	Ref.			Ref.

Note. Reference group: normoglycemic group; T2D: Type 2 Diabetes, S.E.: Standard Errors; OR: Odd Ratios; CI: Confidence Intervals, PIR: Poverty-to-Income Ratio; ***p<.001, **p<.01, *p<.05.

In prediabetes outcomes (Table 5), significant associations were observed among non-Hispanic Black (OR = 2.09, p < .001), Hispanic (OR = 1.64, p < .001), and other racial/ethnic groups (i.e., non-Hispanic Asian, multi-racial individuals, and others) (OR = 1.60, p < .001). Education levels also showed significant associations with prediabetes; menopausal women with high school education or less had elevated odds (OR = 1.37, p < .01) and with some high school education, a 2-year degree, or no degree had increased likelihood of prediabetes (OR = 1.29, p < .05). However, similar to diabetes outcomes, there was no significant difference between marital status and prediabetes (p > .05). Employment status was significantly related only among disabled women (OR = 1.49, p < .001) compared to employed menopausal women. Annual household income levels remained significant, especially among those with PIR below 1.3 (OR = 1.32, p < .05).

The survey logistic regression outcomes comparing recognized hyperglycemia and unrecognized hyperglycemia among menopausal women (Table 6) were associated with several sociodemographic factors. Compared with non-Hispanic White menopausal women, non-Hispanic Black women had higher odds of recognized hyperglycemia (OR = 2.84, p < .001), as did Hispanic women (OR = 1.95, p < .001), and those from other racial/ethnic groups (i.e., non-Hispanic Asian, multi-racial individuals, and others) (OR = 2.29, p < .001). Education level was a significant factor for recognized hyperglycemia, with those having high school education or less demonstrating higher odds (OR = 2.25, p < .001). While there was no association between marital status and recognized hyperglycemia (p > .05), in employment status, disabled (OR = 1.66, p < .001), retired (OR = 1.89, p < .001), and unemployed menopausal women (OR = 1.42, p < .05) were more likely to be in the recognized hyperglycemia group compared to employed individuals. Lower annual household income status, particularly PIR below 1.3, was associated with increased recognized hyperglycemia probability (OR = 1.60, p < .001).

Table 6.

Survey logistic regression based on hyperglycemic recognition status.

Variable	Hyperglycemic recognition status
	Recognized hyperglycemia			Unrecognized hyperglycemia
	β	S.E.	OR (95% CI)	β	S.E.	OR (95% CI)
Race/Ethnicity
Non-Hispanic Black	1.04	0.14	2.84*** (2.18-3.71)	0.83	0.12	2.30*** (1.84-2.88)
Hispanic/Latina	0.67	0.12	1.95*** (1.53-2.48)	0.46	0.10	1.58*** (1.29-1.93)
Others	0.83	0.17	2.29*** (1.65-3.19)	0.60	0.14	1.82*** (1.39-2.38)
Non-Hispanic White	Ref.			Ref.
Education Level
High School or Less	0.81	0.15	2.25*** (1.69-3.00)	0.38	0.11	1.46*** (1.18-1.80)
Some High School/2-Year Degree/No degree	0.53	0.15	1.69*** (1.25-2.28)	0.26	0.11	1.29* (1.04-1.60)
4-Year Degree or More	Ref.			Ref.
Marital Status
Never Married	0.05	0.11	1.05 (0.85-1.31)	-0.01	0.09	0.99 (0.83-1.18)
Divorced/Separated/Widowed	0.16	0.13	1.18 (0.92-1.52)	-0.15	0.11	0.86 (0.70-1.06)
Married/Partnered	Ref.			Ref.
Employment Status
Not Employed	0.35	0.16	1.42* (1.05-1.93)	0.16	0.13	1.17 (0.92-1.50)
Disabled	0.51	0.12	1.66*** (1.32-2.09)	0.34	0.09	1.41*** (1.18-1.68)
Retired	0.64	0.16	1.89*** (1.39-2.57)	-0.16	0.15	0.86 (0.64-1.14)
Employed	Ref.			Ref.
PIR
<1.3	0.47	0.14	1.60*** (1.21-2.11)	0.26	0.11	1.30* (1.05-1.61)
1.3-3.5	0.39	0.14	1.47** (1.13-1.92)	0.24	0.10	1.27* (1.04-1.55)
>3.5	Ref.			Ref.

Note. Reference group: normoglycemic group; S.E.: Standard Errors; OR: Odd Ratios; CI: Confidence Intervals, PIR: Poverty-to-Income Ratio; ***p<.001, **p<.01, *p<.05

Additionally, there was also a significant association between unrecognized hyperglycemia among menopausal women and sociodemographic factors (Table 6). In racial/ethnic groups, non-Hispanic Black (OR = 2.30, p < .001), Hispanic (OR = 1.58, p < .001), and other racial/ethnic groups (i.e., non-Hispanic Asian, multi-racial individuals, and others) (OR = 1.82, p < .001) are more likely to be in unrecognized hyperglycemia compared to non-Hispanic White women. Educational attainment is associated with unrecognized hyperglycemia risk, with women having a high school education or less showing a higher likelihood of unrecognized hyperglycemia (OR = 1.46, p < .001) rather than other education statuses. However, marital status did not show any association with unrecognized hyperglycemia (p > .05). In employment status, only disabled menopausal women were linked with unrecognized hyperglycemia (OR = 1.41, p < .001) compared to employed menopausal women. PIR levels were also relevant to unrecognized hyperglycemia, particularly for those with PIR below 1.3 (OR = 1.30, p < .05) and with PIR between 1.3 to 3.5 (OR = 1.27, p < .05) rather than those with PIR over 3.5.

Discussion

The primary aim of this study was to investigate trends in the prevalence of T2D status (i.e., diabetes and prediabetes) and hyperglycemic recognition (i.e., recognized and unrecognized hyperglycemia) among menopausal women over a 20-year period. Additionally, this study sought to examine sociodemographic disparities (race/ethnicity, education level, marital status, employment, and PIR) associated with T2D prevalence and hyperglycemic recognition in this population. Findings demonstrated that the prevalence of both diabetes and prediabetes increased during the study period from 2003 to 2023, with prediabetes showing a particularly marked rise. Similarly, prevalence rates for both recognized and unrecognized hyperglycemia increased over time, with unrecognized hyperglycemia exhibiting a notably substantial upward trend. These results highlight the importance of strengthening public health interventions and hyperglycemic recognition initiatives tailored specifically for menopausal women to address the growing prevalence of diabetes within this population.

The current study found that diabetes prevalence among menopausal women increased modestly from 11.3% to 11.9% over the 20-year period. This modest increase may reflect improvements in healthcare accessibility, enhanced diagnostic techniques, and greater public awareness regarding glycemia; however, it also may indicate the persistent and intensifying diabetes burden among menopausal women.^3,54 Notably, the substantial increase observed in prediabetes prevalence from 24.3% to 35.2% is indicative of a rapidly expanding at-risk population. This marked rise aligns with previous studies related to metabolic vulnerabilities characteristic of menopause transitions, particularly heightened insulin resistance, alterations in glucose metabolism, and increased visceral adiposity due to estrogen deficiency.^13,16 Physiological changes during menopause, including hormonal shifts and increased abdominal fat accumulation, significantly elevate the risk of developing impaired glucose tolerance and insulin resistance, ⁵⁵ subsequently leading to prediabetes and, if untreated, progressing to overt diabetes.^56,57 Therefore, these pathophysiological mechanisms highlight the critical need for timely detection and targeted lifestyle interventions among prediabetic menopausal women. Importantly, this trend should not be attributed to menopause-related hormonal changes alone. The increase may reflect a combination of aging, secular changes in adiposity, socioeconomic conditions, healthcare access, and menopause-related changes in body composition and insulin sensitivity. Recent evidence suggests that menopausal timing is associated with T2D risk, although these relationships are complex and may be partly mediated by BMI and abdominal adiposity. ⁵⁸ In addition, obesity remains highly prevalent among U.S. middle-aged and older adults, underscoring adiposity as an important contextual factor in interpreting trends in prediabetes and hyperglycemia. ⁵⁹

The recognition-status analysis in this study added information beyond the prevalence of diabetes and prediabetes alone. Whereas the diabetes and prediabetes analyses estimate the burden of abnormal glycemic status, the recognition-status analysis identifies the proportion of menopausal women with biomarker-defined prediabetes- or diabetes-range hyperglycemia who did not report prior clinician communication of diabetes or borderline diabetes. The significant increases in both recognized hyperglycemia (11.7% to 14.3%) and, more strikingly, unrecognized hyperglycemia (23.9% to 32.9%) cases highlight critical inadequacies in detection and recognition of abnormal glycemic status for this demographic group. Although the increase in recognized hyperglycemia may reflect prior screening, clinical contact, or communication of elevated glycemic risk, the disproportionate growth in unrecognized cases compared with recognized cases still points toward considerable deficits in symptom recognition and inadequate clinical screening among menopausal women, emphasizing gaps within current healthcare frameworks. Complicating this scenario, symptom overlaps between menopause (e.g., fatigue, weight gain, mood changes) and diabetes can often lead to misattribution,^60,61 causing significant delays of recognition about hyperglycemia and missed opportunities for early intervention. To effectively address this recognition challenge, healthcare provider training programs should emphasize distinguishing glycemic abnormalities and diabetes-specific clinical markers from general menopausal symptoms. Also, incorporating routine diabetes screening protocols into standard menopausal healthcare services, coupled with tailored health literacy interventions specifically designed for menopausal women, may enhance early detection and management of diabetes in this population.

This study also identified significant sociodemographic disparities in T2D prevalence among menopausal women. Specifically, non-Hispanic Black women were 4.03 times more likely to develop diabetes compared to non-Hispanic White women, followed closely by Hispanic women (2.81 times). Also, participants categorized as other race/ethnicity also had elevated odds of diabetes, prediabetes, recognized hyperglycemia, and unrecognized hyperglycemia compared with non-Hispanic White women. This finding should be interpreted cautiously because the other category was heterogeneous and included non-Hispanic Asian, multiracial, and other racial/ethnic groups. Nevertheless, recent NHANES evidence similarly suggests that adults categorized in other racial/ethnic minority groups may have higher diabetes prevalence than non-Hispanic White adults. ⁶² One possible explanation is that conventional BMI thresholds may not capture cardiometabolic risk equivalently across racial/ethnic groups; for example, Asian populations may experience comparable T2D risk at lower BMI levels than White populations. ⁶³ Future studies with larger samples should disaggregate these heterogeneous groups to better characterize subgroup-specific risks among menopausal women. Collectively, these racial and ethnic disparities may be attributable to multiple intersecting factors, including socioeconomic disadvantages,^32,64 structural barriers to healthcare access,^65,66 and culturally ⁶⁷ influenced health behaviors and beliefs that contribute to disease risk and progression. Previous studies have suggested that systemic inequalities, including experiences of discrimination and chronic stress among non-Hispanic Black and Hispanic populations, significantly impact metabolic health through stress-mediated inflammatory pathways,^68,69 consequently elevating diabetes risk. ⁷⁰ Moreover, in educational disparities, women who attained a high school education or less showed 2.77 times increase in the probability of developing diabetes. The results related to education disparities in this study are consistent with previous studies that lower educational attainment often correlates with limited health literacy, ⁷¹ which adversely affects an individual’s ability to comprehend diabetes-related health information,^72,73 engage in effective self-management practices, ⁷⁴ and utilize preventive healthcare services. ⁷⁵ In employment status, women who were unable to work due to disability or those who were retired had significantly elevated the probability of developing diabetes, 1.46 times and 1.50 times, respectively. Disability status is commonly associated with physical limitations and mobility restrictions, ⁷⁶ potentially leading to decreased participation in preventive health behaviors, diminished PA, and consequently increased insulin resistance. Similarly, retirement can trigger lifestyle changes characterized by reduced daily PA, increased sedentary behavior, and nutritional shifts, ⁷⁷ all of which are recognized contributors to diabetes development. Lastly, lower annual household income was significantly associated with increased diabetes risk, underscoring the critical role of socioeconomic determinants in diabetes development and progression. Economic disadvantages limit access to health-promoting resources, including quality healthcare services, nutritious food options, safe environments for PA, and essential preventive screenings. ⁷⁸ Additionally, financial strain can perpetuate chronic stress, ⁷⁹ potentially exacerbating metabolic dysregulation and insulin resistance, thereby increasing diabetes susceptibility.^80,81 Therefore, these findings highlight that initiatives aimed at improving health literacy, enhancing healthcare accessibility, providing socioeconomic support, and facilitating healthy lifestyle behaviors among vulnerable menopausal women are essential. Implementing comprehensive, multifaceted strategies could substantially reduce disparities in diabetes prevalence and improve health outcomes in these at-risk populations.

Similarly, the subgroup analysis of hyperglycemic recognition status revealed critical disparities among menopausal women. Non-Hispanic Black women had 2.84 times higher odds of recognized hyperglycemia than their non-Hispanic White counterparts, and Hispanic women had 1.95 times higher. These disparities are likely driven by systemic barriers like inequities prevalent within racial and ethnic minority communities. For instance, previous research emphasizes that Black and Hispanic populations experience significant structural and interpersonal discrimination within healthcare systems,^82,83 contributing to lower utilization of preventive services and delayed diagnosis of chronic conditions such as diabetes. More concerning, however, non-Hispanic Black women and Hispanic women exhibit 2.30 times and 1.58 times increased likelihood, respectively, of remaining recognition of abnormal glycemic status compared to their non-Hispanic White counterparts. Structural determinants of inequity in health system have been linked to worse diabetes processes and outcomes, including poorer glycemic control and standards of care, via pathways such as segregated neighborhoods, differential resource allocation, and policy environments that shape access and quality. ⁸⁴ These upstream conditions intersect with interpersonal inequity in clinical settings, which is associated with care delays; recent large-cohort data show that perceived inequity is tied to postponing care, partly mediated by poorer patient-clinician communication. ⁸⁵ These forces can produce lower screening uptake, later detection, and lower engagement with preventive services in Black and Hispanic populations, even when nominal access exists.

Likewise, women with lower educational attainment (high school education or less) demonstrated 2.25 times higher odds of recognized hyperglycemia and 1.46 times higher odds of unrecognized hyperglycemia compared to women with higher education levels. Although recent national surveillance did not detect significant overall differences in unrecognized hyperglycemia by education, ⁸⁶ menopausal women may represent a subgroup in which low educational attainment is more tightly coupled to missed detection due to reduced screening uptake, lower health literacy/numeracy, and symptom misattribution (e.g., fatigue, sleep disturbance) that delays testing. Low health literacy is consistently associated with poorer diabetes knowledge, self-care, and glycemic control,^87–89 and low numeracy impairs risk comprehension and adherence.^90,91 Therefore, these mechanisms plausibly explain both higher recognized burden and higher odds of unrecognized hyperglycemia in lower-education menopausal women.

Additionally, both unemployment and lower household income were independently associated with increased likelihoods of recognized and unrecognized hyperglycemia; moreover, unrecognized hyperglycemia was linked to unemployment attributable to disability. Socially, unemployment and low income amplify risk through chronic stress, food insecurity, constrained opportunities for PA, and reduced access to preventive care and screening; these factors also contribute to under-recognition and gaps in guideline-recommended care. Loss of employment commonly entails loss of employer-sponsored insurance; navigating transitions to Medicaid or Marketplace plans can be administratively complex, and continuation or replacement coverage may be abruptly unaffordable for low-income households. Consequently, unemployed adults are less likely to be insured than employed adults, leading to unmet healthcare needs; lack of recent care and absence of insurance are, in turn, associated with unrecognized hyperglycemia. ⁹²

In summary, these results highlight severe deficiencies in preventive screening, health education, and hyperglycemic recognition within marginalized and socioeconomically disadvantaged populations. The intersectionality of menopausal hormonal shifts, increased insulin resistance, and socioeconomic vulnerabilities amplifies diabetes risk and complicates symptom recognition,^93,94 and especially, menopausal women from minority or low socioeconomic backgrounds frequently can misattribute nonspecific symptoms of abnormal glycemic status or diabetes to menopausal transitions, further exacerbating delayed recognition or missed clinical communication. Therefore, the present study highlights that culturally sensitive diabetes education and community outreach programs should be provided effectively for improving early recognition and glycemic screening and improving disease management outcomes among socioeconomically disadvantaged and minority groups.

Strengths and limitations

A primary strength of this study lies in its extensive 20-year analysis utilizing a nationally representative dataset, enabling robust examination of T2D/prediabetes prevalence and hyperglycemic recognition trends among menopausal women. Additionally, by incorporating subgroup analyses, this study provided detailed insights into sociodemographic disparities, offering critical information to guide targeted public health interventions. However, this study still has several limitations. First, the analysis primarily focused on T2D prevalence and did not examine diabetes-related complications. Second, excluding participants with missing data increased the representation of older and non-Hispanic White individuals in the analytic sample, and the limited availability of HbA1c and FPG markedly reduced the classification sample, potentially introducing bias and exaggerating some findings. Nevertheless, missing data were evenly distributed across all study groups, minimizing the likelihood of systematic bias affecting estimates of diabetes prevalence and hyperglycemic recognition. Third, despite utilizing objective biomarkers, misclassification of glycemic or hyperglycemic recognition status remains possible. Reliance on a single laboratory measurement may overestimate diabetes prevalence due to transient fluctuations in blood glucose levels. To mitigate this limitation, we employed a combined assessment of HbA1c and FPG, recognizing that this approach reduces susceptibility to short-term glycemic variability and provides a more stable indicator of diabetes status than FPG alone. Fourth, the recognition-status outcome should be interpreted cautiously. This study did not directly measure subjective awareness of prediabetes, diabetes risk perception, or diabetes-related health literacy. Recognition status was inferred from self-reported prior clinician communication of diabetes or borderline diabetes combined with biomarker-defined prediabetes- or diabetes-range hyperglycemia. Therefore, unrecognized hyperglycemia may reflect lack of screening, lack of clinician communication, limited recall, or changes in glycemic status after prior clinical encounters. Fifth, the present study focused specifically on naturally menopausal women and did not compare findings with men of similar age or with premenopausal women. Therefore, the analysis cannot determine whether the observed temporal trends are specific to menopausal women or reflect broader age-related patterns in the U.S. population. Future studies should directly compare menopausal women with men and premenopausal women to better distinguish menopause-specific patterns from general aging-related trends. Additionally, a priori power analysis was not conducted because this study used a repeated cross-sectional secondary analysis of NHANES data, and the sample size was determined by the number of eligible participants with complete data after applying the predefined inclusion and exclusion criteria. Although the final analytic sample and number of outcome events were sufficient for the planned survey-weighted regression analyses based on events-per-variable considerations, the absence of a priori sample size calculation should be considered when interpreting the statistical power and precision of subgroup estimates. Lastly, BMI and other measures of adiposity were not incorporated into the primary trend models. Given the established role of adiposity in hyperglycemia ⁹⁵ and the high prevalence of obesity ⁹⁶ among U.S. middle-aged and older adults, secular changes in body weight may partly explain the observed increase in prediabetes. Future studies should evaluate whether BMI, waist circumference, or body composition mediates temporal trends in hyperglycemia among menopausal women.

Conclusion

The findings of this 20-year analysis underscore significant upward trends in T2D, prediabetes, and hyperglycemic recognition among menopausal women, emphasizing an escalating public health challenge within this population. The substantial rise in prediabetes prevalence indicates a growing population of naturally menopausal women at elevated glycemic risk, reinforcing the need for targeted screening and early lifestyle interventions. Because this repeated cross-sectional analysis cannot establish causal mechanisms, the observed increase should be interpreted as a population-level trend that may reflect multiple age-related, sociodemographic, behavioral, and metabolic factors. Moreover, marked increases in unrecognized hyperglycemia cases reveal significant gaps in clinical recognition and healthcare practices, likely exacerbated by symptom misattribution between menopausal and diabetic conditions. Crucially, pronounced sociodemographic disparities emerged, with non-Hispanic Black and Hispanic women, those with lower educational attainment, disabled, retired, and lower-income individuals facing markedly higher likelihood of diabetes and unrecognized hyperglycemia. Consequently, public health strategies must prioritize culturally sensitive, accessible glycemic screening and educational interventions tailored to menopausal women, particularly focusing on minority and socioeconomically disadvantaged communities, to effectively mitigate disparities and improve overall diabetes prevention and glycemic health outcomes.

ORCID iD

Jeong-Hui Park https://orcid.org/0000-0003-1323-0254

Ethical considerations

The National Health and Nutrition Examination Survey (NHANES) protocol was approved by the Institutional Review Board of the Centers for Disease Control and Prevention.

Data Availability Statement

The datasets used and/or analyzed during the current study are available from the corresponding author upon reasonable request.*