Symptom Clusters Predict Risk of Metabolic-Syndrome and Diabetes in Midlife: The Study of Women’s Health Across the Nation

Abstract

Purpose.

Women’s cardio-metabolic risk increases in midlife. Previous work relating menopause symptoms to diabetes/Metabolic Syndrome (MetS) does not consider the adverse impact of multiple concurrent physical and psychological symptoms in midlife.

Methods.

Data are from 3097 participants in the Study of Women’s Health Across the Nation, a cohort of midlife women followed for over 20 years. Baseline latent symptom classes (LSC) were derived from 58 questions assessing physical, psychological and menopausal symptoms. Six identified LSCs ranged from 1 (most symptoms present at high severity) to 6 (few symptoms present at low severity). Accelerated failure time models estimated time to onset of type 2 diabetes and MetS for each LSC.

Results.

Women with multiple high/moderate severity symptoms had an earlier onset of diabetes (13.2%, 14.1% and 20.8% earlier onset in LSCs 1, 2, 3, respectively) and MetS (15.5%, 13.0% and 19.7% earlier onset in LSCs 1, 2, 3, respectively) than women with few/low severity symptoms.

Conclusions.

Having multiple concurrent moderate to high intensity physical and psychological symptoms in midlife are associated with early onset of diabetes and MetS. Monitoring and subsequent intervention on a broad range of symptoms in midlife may significantly mitigate cardio-metabolic risk during this critical life stage.

Introduction –

Metabolic syndrome (MetS) is a cluster of cardio-metabolic risk factors that is associated with increased risk of diabetes and cardiovascular disease – two of the major causes of decreased lifespan in the United States¹. Women’s risk of metabolic syndrome increases sharply in their forties with their highest risk for diabetes onset in midlife (45-60 years old)^2–7, the age span during which most women experience menopause. Thus midlife has been highlighted as an important life stage where women are at high risk for cardiovascular disease⁸.

The menopause marks a period of endocrinologic and physiological change⁹ which is associated with vasomotor, somatic, mood and urogynecological symptoms^10–17. Previous work has documented the frequency of individual menopausal symptoms, such as hot flashes, and also characterized how the range of women’s menopausal and aging-related symptom experience changes across the midlife^{10,11,13–15}. For example, a longitudinal study in Australia identified four major symptom categories present among midlife women: somatic (e.g. headaches/migraine, tiredness, pain etc.), urogynecological (e.g. urination, bowel or hemorrhoid issues), physical (e.g. allergies/hay fever, breathing difficulty, chest pain) and vasomotor (e.g. hot flashes, night sweats)¹¹. Biologic changes during the menopausal transition have system-wide physical and psychological effects^10,18–20. Previous studies have documented an array of individual symptoms^19–24 including headache, irritability and mood swings¹⁹, depression^21,22, perceived stress²³, sleep^18,25, and pain²⁴ that vary during the menopausal transition. Aging-related changes, such as onset of osteoarthritis with its associated pain, also contribute to symptom burden during the midlife²⁶. Given the variety of potential symptoms associated with the midlife and the menopausal transition it may be that the grouped occurrence of symptoms, in their varying forms and severities, may be a more accurate indicator of potential for system-wide dysregulation and cardio-metabolic risk than s individual symptoms alone. Thus in the Study of Women’s Health Across the Nation (SWAN) researchers, using a data-driven approach^27,28, identified six latent symptom classes (LSCs) that characterized the burden of menopausal, physical and psychological symptoms in midlife¹⁰. The study established that the six LSCs were stable throughout midlife with the more severe LSCs associated with poorer self-rated health¹⁰.

A higher symptom burden may indicate higher risk for poor cardio-metabolic health outcomes such as diabetes and metabolic syndrome, especially given that common symptoms have been individually associated with each of these conditions^{13–15,17,29–32}. Specifically urogyneoclogical and vasomotor symptoms have been associated cross-sectionally with metabolic syndrome³⁰, the Framingham risk score, and obesity³¹. Furthermore, urogynecologic symptoms have been found to be associated longitudinally with obesity and diabetes^14,29 Menopausal stage, along with sex hormone levels, have been associated with the metabolic syndrome^33–35. In cross sectional analyses, objective indices of sleep were associated with prevalent metabolic syndrome³⁶ among midlife women. Moreover, a history of major depression and the presence of depressive symptoms predicted the development of metabolic syndrome and diabetes respectively^37,38. A broad range of physical and psychological symptoms appear to increase risk for system-wide dysregulation such as with diabetes and MetS^8,9,39 Therefore, it may be that monitoring total symptom burden may provide a more accurate picture of an individual’s cardio-metabolic risk. This paper extends previous work in SWAN¹⁰, and examines the association of previously identified¹⁰ latent symptom classes with metabolic syndrome and diabetes.

Methods –

Data was from Study of Women’s Health Across the Nation (SWAN)⁴⁰ with 3302 women enrolled at seven sites in 1996/1997, with each site enrolling White women and women of another race/ethnicity including Black women at four sites and Chinese, Japanese, or Hispanic at one site each. Eligibility for the cohort study included: aged 42–52 years, intact uterus, at least one menstrual period and no use of reproductive hormones in the past 3 months, and not pregnant or lactating. Following the baseline visit, women participated in up to 15 follow-up visits across 20 years with approximately 75.0% of surviving participants providing follow-up data. The Institutional Review Boards at each study site approved the protocol and participants provided written, informed consent at each visit.

Each visit included interviewer-administered and self-administered questionnaires. Waist circumference was measured with a measuring tape placed horizontally at the level of the natural waist or the narrowest part of the torso. Abdominal obesity was defined as ≥80 cm for Japanese and Chinese women and >88 cm for all other race/ethnic groups. Blood pressure was calculated as the average of two seated measurements using a standard mercury sphygmomanometer. A fasting blood sample was also obtained except at the two non-clinic follow-up visits 11 and 14.

Race/ethnicity was self-defined. Information on highest level of education attained (high school graduate /GED or less than high school versus at least some college), economic strain (“how hard is it to pay for basics (very hard, somewhat hard, or not hard)?), physical activity (Baecke questionnaire, scores range 3-15⁴¹), and smoking status (current, past, never) were obtained at baseline.

Assays.

Serum glucose levels were measured using a hexokinase-coupled reaction (Boehringer Mannheim Diagnostics, Indianapolis, IN). Heparin-2M manganese chloride was used to isolate high density lipoprotein cholesterol (HDL)⁴². Enzymatic methods using the Hatachi 747 analyzer (Boehringer Mannheim Diagnostics, Indianapolis, Indiana) were used to determine fasting triglyceride level⁴³.

Latent Symptom Classes (LSC).

The latent class analysis has been described previously¹⁰. Briefly, we used 58 questions from the baseline questionnaire including items from the SF-36 domains of role-physical, bodily pain, role-emotional, vitality and social functioning⁴⁴, the Center for Epidemiologic Studies Depression Scale (CES-D)^45,46, the 4-item Cohen’s Perceived Stress Scale ⁴⁷ as well as a 14-item list of general symptoms assessed in SWAN and other studies of the menopausal transition including how often in the past two weeks (from not at all to daily) they experienced vasomotor symptoms, mood symptoms, somatic symptoms, fatigue and vaginal dryness,^48–50, self-reported sleep quality^51,52, involuntary urine loss⁵³ and sexual desire⁵⁴. Women were excluded from this analysis if information on more than 10 of the included symptoms were missing at baseline. These data were used to construct LSC’s based on the probability of indicating a given level of symptom⁵⁵. Using Bayesian Information Criterion (BIC), which avoids overfitting by penalizing models with large numbers of classes, as the model selection criterion a total of six LSC’s were identified (figure from original article):

• LSC 1 (Many High Intensity Symptoms) is highly symptomatic with high intensity ratings. Women endorse most measured symptoms, including physical, psychological and pain symptoms, poor sleep quality and fatigue.

• LSC 2 (Many Moderate Intensity Symptoms) is similar to LC1 in that they endorse most of the measured symptoms but with more moderate intensity ratings.

• LSC 3 (Moderate Physical Symptoms with Interference) endorsed a narrower range of predominantly physical symptoms at moderate intensity, including pain, fatigue, sleep and physical health symptoms sufficient to interfere in life activities. They endorsed few psychological symptoms and those they endorsed were at low intensity.

• LSC 4 (Mild Psychological Symptoms) also endorsed a narrower range of symptoms, predominantly those that were fatigue-related and psychological symptoms, and with milder intensity ratings.

• LSC 5 (Mild Physical Symptoms) is similar to LC3 but did not endorse the physical health interference items and has lower intensity ratings on the other physical symptoms.

• LSC 6 (Asymptomatic) is essentially asymptomatic reporting only a few, low intensity symptoms related mostly to fatigue.

Type 2 diabetes.

At each visit, a woman was classified as having type 2 diabetes if she met at least one of the following criteria: 1) a fasting serum glucose level ≥7 mmol/L (not available for clinic visits 2, 9, and 15 or the interim visits 11 and 14), 2) current use of insulin or oral hypoglycemic medication, or 3) self-report of physician diagnosis of type 2 diabetes.

Metabolic Syndrome (MetS).

At each visit, a woman was classified as having metabolic syndrome (MetS) if she met at least three of the following criteria⁵⁶: 1) obesity defined by waist circumference, 2) fasting glucose ≥100 mg/dL⁵⁷ or diagnosed diabetes, 3) hypertensive defined by a systolic blood pressure ≥130 mmHg or diastolic blood pressure ≥85 mmHg or taking any blood pressure medication, 4) fasting HDL level <50 mg/dL, and 5) fasting triglycerides ≥150 mg/dL. Fasting glucose level at ≥100 mg/dL was based on guidelines from the National Cholesterol Education Program^57,58 to additionally capture pre-diabetes. Data were available to classify MetS status from baseline through visit 15. If information on MetS status was missing for a given visit, status from the previous visit was used.

Statistical Analysis.

We were able to assign 3097 of the 3302 (93.7%) SWAN participants to a baseline LSC. Ten of these women had insufficient information to define MetS status.

Frequencies and percentages were calculated to describe the characteristics of the eligible study sample overall and by LSC. Bivariate analyses were performed to determine the association between LSC and each covariate with prevalent baseline and with incident diabetes and MetS. Interval-censored Kaplan Meier survival curves were estimated for time, measured as years of age, to the first diagnosis of diabetes and to MetS stratified by baseline LSC. To model age of diabetes onset and MetS onset, we used parametric accelerated failure time (AFT) models which include both prevalent cases (at baseline) along with incident cases of diabetes and MetS, We selected a lognormal distribution that yielded the best fit to the data based on log-likelihood measures, allowing regression parameter estimates to be interpreted, after exponentiating, as a percent change in mean time to diabetes onset or MetS onset. For each outcome, initial models estimated the effect of each LSC on survival time, and subsequent models were adjusted for a priori selected¹⁰ confounders including race/ethnicity, education, smoking, financial strain, and physical activity. Likelihood ratio tests confirmed inclusion of the confounders in the final models. Analyses were conducted using the LIFEREG procedure in SAS (version 9.4).

Results –

Table 1 provides information on the distribution of latent symptom classes (LSC’s), demographic and lifestyle characteristics at baseline. Most women are in LSC 2 (many moderate intensity), 4 (mild psychological symptoms) and 6 (asymptomatic) (20.0, 23.2 and 20.6 percent, respectively) followed by classes 3 (moderate physical with interference), 5 (mild physical symptoms) and 1 (many high intensity symptoms) (13.0, 12.6, and 10.6, respectively).

Table 1 -.

Baseline Characteristics of the Study Population, Study of Women’s Health Across the Nation (SWAN) (n=3097)

	Latent Symptom Class [% (n)]
	Total	1 (many high intensity)	2 (many moderate intensity)	3 (moderate physical with interference)	4 (mild psychological)	5 (mild physical)	6 (asymptomatic)
Characteristic		10.6 (328)	20.0 (618)	13.0 (404)	23.2 (717)	12.6 (391)	20.6 (639)
Race/Ethnicity
Black	28.7	35.1	27	31.4	24.5	24.8	32.6
White	47.9	45.7	50.5	48.3	43.5	61.1	43
Chinese	7.9	4.3	5.2	6.7	12	6.1	9.9
Hispanic	6.5	12.5	9.1	6.9	7.7	0.8	3
Japanese	8.9	2.4	8.3	6.7	12.3	7.2	11.6
Financial Strain1
Very Hard	9.1	27.7	11.3	5.9	9.1	3.3	3
Somewhat Hard	29.3	40.5	39.2	29.7	29.8	22	17.7
Not hard	61.3	31.4	49.2	64.4	60.7	74.4	79
Education
Some college	76.4	61.9	73.5	79.5	75	86.4	80.1
HS or less	23.6	38.1	26.5	20.5	25	13.6	19.9
Smoking2
Never smoker	56.9	44.2	54.4	58.9	61.6	54	61
Past smoker	25.2	23.8	24.3	24	23.2	34.5	23.9
Current smoker	17	31.1	20.1	16.1	14.5	10.5	14.2
Physical Activity3, mean(sd)	7.7	7	7.5	7.5	7.8	8	8
Waist Circumference4 in cm, mean(sd)	86.1	92.7	87.7	90.1	83.1	85.2	82.6

¹9 women have missing information

²29 women have missing information

³91 women have missing information

⁴31 women have missing information

Among all women, the prevalence of MetS at baseline was 24% and another 21% developed MetS during follow-up. The prevalence of diabetes was 5% at baseline and another 12% developed diabetes during follow-up period. Women belonging to LSC1, 2 and 3 (many high, many moderate, moderate physical with interference, respectively) had higher prevalence of both conditions at baseline and by the end of the study, than women belonging to LSC4, 5 and 6 (mild psychological, mild physical, asymptomatic, respectively), with the highest prevalence observed for women in LSC1 and 3 (Table 2).

Table 2 -.

Prevalence of Diabetes and Metabolic Syndrome (MS) at Baseline and in Follow Up by Latent Symptom Class and Race/Ethnicity

	Prevalence of Diabetes		Prevalence of MS
Characteristic	Baseline (n = 3097)	Overall1 (n = 3097)	Baseline (n = 29912)	Overall1 (n = 3087)
Latent Symptom Class at Baseline
1 (many high intensity)	7.3	22.3	32.3	57.2
2 (many moderate intensity)	6.3	18.6	28.1	50.6
3 (moderate physical with interference)	7.4	21.3	29.4	56.4
4 (mild psychological)	3.2	12.1	17.7	40.1
5 (mild physical)	2.3	12.8	17.7	46.5
6 (asymptomatic)	3.3	12.5	14.2	40.7
Race/Ethnicity
Black	8.8	24.4	30.3	58.9
White	3.2	12	19.5	43
Chinese	1.2	10.6	11.9	35.4
Hispanic	7.9	20.8	31.2	52
Japanese	0.4	10.1	12.1	38

¹Overall prevalence of diabetes/MetS from baseline (1996/1997) through follow-up visit 15.

²MetS status was not available for 96 women at baseline.

Figure 1 presents the Kaplan Meier survival curves for each condition by LSC, with the vertical dotted lines representing the youngest age of enrollment into the cohort, 42 years. For MetS, a risk segregation between LSCs 1-3 (many high, many moderate, moderate physical with interference, respectively) and LSC 4-6 (mild psychological, mild physical, asymptomatic, respectively) is clearly evident at the onset of the study and worsens over the course of follow-up. For diabetes, the risk segregation emerges in the forties and widens as women age.

Figure 1 -.

Interval Censored Kaplan-Meier Survival Curve for Diabetes and Metabolic Syndrome by Latent Symptom Class

Note: Latent symptom classes 1 – 6 are many high intensity, many moderate intensity, moderate physical with interference, mild psychological, mild physical and asymptomatic.

Table 3 presents the Accelerated Failure Time (AFT) models for MetS and diabetes. In unadjusted models, women in LSC1, 2 and 3 (many high, many moderate, moderate physical with interference, respectively) developed diabetes 29.0%, 20.9% and 26.3% earlier in life, respectively, than women in the asymptomatic LSC6, age at onset did not differ among women in LSC 4 and 5 (mild psychological and mild physical) compared to LSC 6. Difference in age at onset of diabetes was attenuated to 13.2%, 14.1% and 20.8% earlier after adjustment for confounders. This translates to highly symptomatic classes (LSC 1-3) estimated to develop diabetes approximately 5-8 years earlier than the asymptomatic class (LSC 6). Differences were diminished further and were not statistically significant when additionally adjusting for baseline waist circumference. For METs, women in LSC1, 2 and 3 (many high, many moderate, moderate physical with interference, respectively) developed MetS 29.4%, 19.9% and 24.9% earlier in life, respectively, than women in the asymptomatic LSC6. The difference in age at onset of MetS was attenuated to 15.5%, 13.0% and 19.7% after adjustment for confounders. This translates to highly symptomatic classes (LSC 1-3) estimated to developed MetS approximately 4-7 years earlier than the asymptomatic class (LSC 6). Figure 2 graphically illustrates the adjusted hazards of diabetes and MetS.

Table 3 -.

Percent Reduction in Time to Diabetes and Metabolic Syndrome (MetS) by Latent Symptom Class from Accelerated Failure Time Models (versus latent symptom class 6).

	Diabetes (n = 2944)			Metabolic Syndrome (n = 2961)
Predictor	Model 1	Model 2A	Model 2B	Model 1	Model 2
Latent Symptom Class
1 (many high intensity)	29.0 (17.8, 38.7)	13.2 (−0.5, 25.0)	0.4 (−13.8, 12.9)	29.4 (21.9, 36.2)	15.5 (6.5, 23.6)
2 (many moderate intensity)	20.9 (10.4, 30.2)	14.1 (3.1, 23.9)	7.5 (−3.1, 17.1)	19.9 (13.0, 26.3)	13.0 (5.7, 19.7)
3 (moderate physical with interference)	26.3 (15.5, 35.7)	20.8 (9.9, 30.5)	8.7 (−2.6, 18.7)	24.9 (17.5, 31.6)	19.7 (12.2, 26.5)
4 (mild psychological)	2.5 (−10.6, 14.1)	0.2 (−12.6, 11.6)	0.8 (−10.6, 11.0)	2.3 (−5.9, 10.0)	0.3 (−7.8, 7.8)
5 (mild physical)	1.7 (−13.9, 15.2)	4.0 (−10.5, 16.6)	-1.3 (−15.1, 10.8)	7.6 (−1.5, 15.9)	7.8 (−0.9, 15.7)
Race/Ethnicity
Black		24.3 (17.2, 30.8)	16.5 (9.6, 22.9)		15.3 (10.0, 20.3)
Chinese		−9.7 (−28.9, 6.7)	15.4 (2.1, 26.9)		−18.6 (−31.5, −6.9)
Hispanic		12.7 (−2.3, 25.5)	22.6 (10.7, 32.9)		5.2 (−6.3, 15.4)
Japanese		−10.5 (−28.4, 4.9)	20.8 (9.1, 31.0)		−11.6 (−22.5, −1.6)
Education
HS or less		13.0 (4.9, 20.4)	9.4 (1.9, 16.3)		9.5 (3.6, 15.0)
Smoking
Past smoker		5.9 (−3.0, 14.1)	−0.8 (−9.5, 7.1)		4.3 (−1.8, 10.0)
Current smoker		4.3 (−6.1, 13.6)	6.5 (−2.5, 14.7)		11.4 (4.9, 17.5)
Financial Strain
Very Hard		16.5 (4.4, 27.1)	5.0 (−7.4, 16.1)		10.4 (1.2, 18.7)
Somewhat Hard		6.3 (−2.1, 14.1)	−0.1 (−8.2, 7.4)		8.0 (2.5, 13.3)
Physical activity		−6.6 (−9.0, −4.1)	−1.4 (−3.5, 0.6)		−6.8 (−8.4, −5.2)
Waist Circumference			2.1 (1.8, 2.4)

Note: Significant effect estimates (α = 0.05) are bold

Figure 2 -.

Percent Reduction in Time to Diabetes and Metabolic Syndrome (MetS) by Latent Symptom Class from Selected Adjusted Accelerated Failure Time Models (versus latent symptom class 6).

Note: Latent symptom classes 1 – 6 are many high intensity, many moderate intensity, moderate physical with interference, mild psychological, mild physical and asymptomatic.

Discussion –

This study found that midlife women who experienced multiple high/moderate severity symptoms across a range of physical and mental health domains, not exclusive to menopausal symptoms, were at higher risk for earlier onset of MetS and diabetes than women with a lesser symptom burden. Specifically, women in LSC’s experiencing a range of moderate to high severity physical and psychological symptoms (LSC 1 and 2), as well as those experiencing moderate to high severity physical health symptoms and low intensity psychological symptoms (LSC 3) were at highest risk of earlier onset of MetS and diabetes. These results are consistent with previous SWAN findings that women in the most symptomatic classes (LSC 1-3) had poorer self-rated health than women in less symptomatic classes¹⁰. LSC 3 (moderate physical with interference), which comprised 13% of the sample, was at highest adjusted risk of both diabetes and MetS: This symptom class is characterized by fewer and lower intensity psychological symptoms but moderate intensity symptoms across multiple other domains including vasomotor, pain, fatigue, sleep and physical health symptoms that interfere with life activities¹⁰. These findings provide a more comprehensive assessment of the importance of symptom burden as a marker of cardio-metabolic risk in contrast to previous work which has focused on risk of a few, select menopausal symptoms rather than situating common menopausal symptoms (i.e. hot flashes, night sweats etc.) among other important physical and psychological health issues that may heighten with age and throughout the midlife transition^{11,14,15,17,29–32,59,60}.

Findings from this study extend previous research on cardio-metabolic risk associated with individual symptoms and suggest that in the midlife experiencing multiple physical and psychological symptoms concurrently may confer higher cardio-metabolic risk. Many longitudinal studies of midlife women have documented the system-wide ‘adverse changes in body fat distribution, lipids, lipoproteins, as well as structural and functional measures of vascular health’⁸ during midlife, increasing risk for conditions such as diabetes and MetS³⁹. Furthermore, decreased estrogen and progesterone production can stiffen blood vessels increasing risk of hypertension while also activating the renin-angiotensin aldosterone system increasing endothelial dysfunction which can increase risk for both diabetes and hypertension³⁹. Considerable work on individual menopausal symptoms, suggests that urogynecological and vasomotor symptoms are associated with MetS³⁰, Framingham risk score, diabetes and obesity^14,29,31. Previous work has shown that individual symptoms such as depression and poor sleep are associated with diabetes and MetS cross-sectionally³⁶ and longitudinally in midlife women^37,38. The variety of mid-life symptoms that have been found to have an association with diabetes and MetS, highlight the fact that the midlife is a critical life-stage during which a variety of symptoms can manifest. Results of this study further the hypothesis that the clustering of adverse physical and psychological symptoms in midlife may be a better, data driven indicator of chronic disease risk during this important life-stage¹⁰.

Abdominal obesity, in this study, was found to be a particularly important pre-disposing risk factor. It is a component of the MetS status classification and after adjusting for abdominal obesity, risk of diabetes did not differ across the symptom classes. This finding is consistent with prior findings that obesity was related to more severe urogenital symptoms, hot flushes, depression and muscle/joint pain^13,14,32. In the Australian Longitudinal Study on Women’s Health, obesity significantly attenuated the association between vasomotor symptoms and diabetes among women in the late vasomotor symptom class²⁹. Findings suggest that underlying obesity may increase risk for severe menopausal and other physical and psychological symptoms in early midlife and also increase risk for diabetes and MetS.

Results also highlight the importance of monitoring symptoms and cardio-metabolic risk early in midlife. As reported previously¹⁰, symptom burden appears to be established by early midlife and remains relatively stable over time in the absence of targeted interventions^61,62. Vasomotor symptoms commonly associated with menopause vary across the menopausal transition, in contrast to somatic, urogynecology and physical symptoms that tend to stay constant¹¹. The inclusion of prevalent cases in this analysis was crucial to correctly determining risk, as one-quarter of the cohort had MetS and one in 20 had diabetes at baseline. Research in younger cohorts is needed to establish temporality and to further disentangle the effect of symptoms occurring in midlife, including symptoms related to menopause and those related to aging, and the effect of persisting symptoms with onset in earlier life stages.

This study has some limitations. The SWAN cohort was left-truncated and likely excluded women in frail health or with numerous co-morbidities⁶³. We utilized accelerated failure time models, which include prevalent cases of diabetes and MetS at baseline along with incident cases during follow-up, in order to mitigate potential selection bias related to left-truncation induced from omitting prevalent cases at baseline. Additionally, some bias may exist due to right censoring but retention in the SWAN cohort is high. Only 4.5% of women were known to be deceased through the 15 ^th follow-up visit; those who died before observed onset of diabetes or MetS were treated as censored. The study has the additional strength of utilizing data from a community based, multiethnic cohort of over 3000 midlife women followed longitudinally.

In conclusion, this paper illustrates that midlife women who experience a substantial symptom burden are at high risk of earlier onset of MetS and diabetes. Nearly one-quarter of the midlife women in this study were highly or moderately symptomatic -- across a broad range of menopausal, physical and psychological symptoms -- as they entered midlife and started the menopausal transition. Monitoring total symptom burden across a broad range of symptoms may more fully capture potential cardio-metabolic risk. Devising interventions that address multiple symptoms simultaneously across the broad range of symptoms that manifest, early and through midlife may be critical to reducing cardio-metabolic risk during this pivotal life-stage.

Glossary

BMI

body mass index

CESD

Center for Epidemiologic Studies Depression Scale

Cl

confidence interval

Cm

centimeter

dL

deciliter

HDL

high density lipoprotein cholesterol

Kg

kilogram

LSC

latent symptom class

MetS

metabolic syndrome

Mg

milligrams

Sd

standard deviation

SWAN

Study of Women’s Health Across the Nation