Sexual Minority Status Disparities in Life's Essential 8 and Life's Simple 7 Cardiovascular Health Scores: A French Nationwide Population‐Based Study

Abstract

Background

A higher burden of cardiovascular disease risk factors has been reported in sexual minority populations. Primordial prevention may therefore be a relevant preventative strategy. The study's objectives are to estimate the associations of Life's Essential 8 (LE8) and Life's Simple 7 (LS7) cardiovascular health scores with sexual minority status.

Methods and Results

The CONSTANCES is a nationwide French epidemiological cohort study that recruited randomly selected participants older than 18 years in 21 cities. Sexual minority status was based on self‐reported lifetime sexual behavior and categorized as lesbian, gay, bisexual, or heterosexual. The LE8 score includes nicotine exposure, diet, physical activity, body mass index, sleep health, blood glucose, blood pressure, and blood lipids. The previous LS7 score included 7 metrics without sleep health. The study included 169 434 cardiovascular disease–free adults (53.64% women; mean age, 45.99 years). Among 90 879 women, 555 were lesbian, 3149 were bisexual, and 84 363 were heterosexual. Among 78 555 men, 2421 were gay, 2748 were bisexual, and 70 994 were heterosexual. Overall, 2812 women and 2392 men declined to answer. In multivariable mixed effects linear regression models, lesbian (β=−0.95 [95% CI, −1.89 to −0.02]) and bisexual (β=−0.78 [95% CI, −1.18 to −0.38]) women had a lower LE8 cardiovascular health score compared with heterosexual women. Conversely, gay (β=2.72 [95% CI, 2.25–3.19]) and bisexual (β=0.83 [95% CI, 0.39–1.27]) men had a higher LE8 cardiovascular health score compared with heterosexual men. The findings were consistent, although of smaller magnitudes for the LS7 score.

Conclusions

Cardiovascular health disparities exist in sexual minority adults, particularly lesbian and bisexual women, who may represent a priority population for primordial cardiovascular disease prevention.

Nonstandard Abbreviations and Acronyms

Add Health

National Longitudinal Study of Adolescent to Adult Health

AUDIT

Alcohol Use Disorders Identification Test

CES‐D

Center for Epidemiological Studies Depression Scale

CNIL

Commission Nationale de l'Informatique et des Libertés

CNIS

Conseil National de l'Information Statistique

CNOM

Conseil National de l'Ordre des Médecins

CVH

cardiovascular health

GGMHS

Geneva Gay Men's Health Survey

IReSP

L'Institut pour la Recherche en Santé Publique

INSEE

Institut National de la Statistique et des Études Économiques

INSERM

L'Institut National de la Santé et de la Recherche Médicale

ISP

L'Institut de Santé Publique

LE8

Life's Essential 8

LS7

Life's Simple 7

SM

sexual minority

Clinical Perspective.

What Is New?

• The current nationwide study is the first to examine cardiovascular health disparities by sexual minority status.

• Lesbian and bisexual women had lower cardiovascular health as measured by Life's Essential 8 and Life's Simple 7 scores compared with heterosexual women.

• On average, gay and bisexual men had higher cardiovascular health scores compared with heterosexual men; however, rural‐residing sexual minority men had lower cardiovascular health compared with heterosexual men.

What Are the Clinical Implications?

• Lesbian and bisexual women represent a priority population for primordial cardiovascular disease prevention.

• Recognizing and overcoming barriers to health care access are essential to improve cardiovascular disease prevention and care provision in sexual minority individuals.

• Health care providers' cultural competency and awareness of cardiovascular disease risk factors in the sexual minority population can prompt more accurate physician‐patient risk communication.

The sexual minority (SM) population includes individuals who self‐identify as lesbian, gay, bisexual, or other nonheterosexual identities. Sexual orientation consists of several dimensions and is commonly assessed by self‐reported sexual identity, behavior, attraction, or a combination of these. ¹ , ² Data from the 2021 Gallup phone survey in a US population, which included a weighted sample of 1000 individuals, found that 1.5% of men identified as gay, 1.0% of women identified as lesbian, and 4.0% of all adults identified as bisexual. ³ In 2019, a nationally representative online survey in France including 994 lesbian, gay, and bisexual individuals of a sample of 12 137 individuals indicated that 3.2% identified as gay or lesbian, 0.9% as bisexual, 83.7% exclusively as heterosexual, and 12.2% as questioning or other categories. ⁴

Previous studies have reported lifestyle cardiovascular disease (CVD) risk factors disparities in SM individuals compared with heterosexual individuals. ⁵ , ⁶ , ⁷ , ⁸ Disparities in health factors, such as blood glucose, blood pressure (BP), and blood lipids, have been rarely investigated in SM populations. ⁵ , ⁹ , ¹⁰ , ¹¹ In addition to controlling existing CVD risk factors in SM individuals (ie, primary prevention), preventing their onset might be a relevant and complementary preventive strategy. This is the purpose of primordial prevention, which aims to prevent the onset of risk factors before they develop. ¹²

In 2010, the American Heart Association (AHA) devised the Life's Simple 7 (LS7) cardiovascular health (CVH) score, which assesses 7 modifiable and actionable CVH metrics: dietary intake, physical activity, nonsmoking, body mass index (BMI), fasting blood glucose, BP, and total blood cholesterol. ¹³ In 2022, the AHA released the Life's Essential 8 (LE8) score, an updated CVH score that integrates sleep health into the CVH construct. ¹⁴

While the distributions of LS7 and the more recent LE8 CVH scores have been reported for the general population, ¹⁵ , ¹⁶ to our knowledge, LE8 scores and particularly measured LE8 health factors have not been previously assessed in the SM population. Since a higher CVH score has been consistently associated with a lower CVD incidence in the general population, ¹⁵ such an evaluation is important to identify CVD prevention targets and vulnerable subgroups among SM individuals who could benefit from CVH score improvement. Therefore, the study's objectives are to examine the distribution of LE8 and LS7 CVH scores by SM status and to quantify the associations of SM status with LE8 and LS7 CVH scores while accounting for confounding factors.

METHODS

Data Access Policy

Access to the current study data was under the auspices of the competent legal and administrative regulators of the CONSTANCES study and is not publicly available. Research teams or public health organizations must submit an online request to the CONSTANCES study manager available via https://www.constances.fr/conduct‐project‐ongoing.php.

The CONSTANCES study is a nationwide population‐representative prospective cohort that examines multiple risk factors, including occupational and social determinants of chronic disease and aging outcomes. The cohort recruited from 2012 to 2020 individuals 18 years and older from the French national social security database. The sample was selected randomly and stratified by age, sex, socioeconomic status (SES), and region to represent the source population. ¹⁷ The participants attended 1 of 24 participating health examination centers in 21 cities distributed over the French European territory, where they underwent clinical examination and laboratory tests at inclusion in the cohort and then every 4 years. All participants gave their informed consent.

The current study conducted cross‐sectional analyses of baseline data. The CONSTANCES cohort was authorized by the National Data Protection Authority (CNIL [Commission Nationale de l'Informatique et des Libertés], n°910 486/n°1 881 675), and has received favorable opinions from the National Council for Statistical Information (CNIS [Conseil National de l'Information Statistique]), the National Medical Council (CNOM [Conseil National de l'Ordre des Médecins]), and the institutional review board of the French National Institute of Health and Medical Research (INSERM [L'Institut National de la Santé et de la Recherche Médicale]).

SM Status

Sex was self‐reported at inclusion as female or male. SM status was determined based on the participants' sex and their past lifetime sexual experiences(ie, behavior). The participants were asked the following questions: “How many partners have you had sex with during your life?” and “Were they: (1) men only?, (2) women only?, (3) men and women?, or (4) do not want to answer.” In the main analysis, SM status was categorized as heterosexual if the participants reported exclusive opposite‐gender partners, bisexual if they reported having both same and different‐gender partners, and gay/lesbian if they reported exclusively having same‐gender partners. Those who filled in that they did not wish to answer the lifetime sexual behavior questionnaire were considered in the main analyses as a separate category for each sex, labeled “declined to answer.” Participants with missing values for lifetime sexual experience were excluded.

CVH Score Definitions

The LE8 and LS7 scores were computed according to the AHA guidelines. ¹³ , ¹⁴ The LE8 score includes diet, physical activity, smoking, BMI, sleep health, BP, blood lipids, and blood glucose metrics. Table S1 lists each LE8 metric definition, cutoff, and modifications. The score for each metric ranged from 0 to 100. The LE8 was the average of all component scores, ranging from 0 to 100, and was categorized as low (0–49), moderate (50–79), or high (80–100). ¹⁴ For participants taking antihypertensive or antihyperlipidemic medications, 20 points were subtracted from their BP or blood lipids scores, respectively.

The LS7 score included the same metrics as the LE8 score except for the sleep health metric. The LS7 score (range 0–14) was calculated as the cumulative sum of 3 levels of each LS7 CVH metric score (0 is poor, 1 is intermediate, and 2 is ideal) and was categorized as low (0–8 points), moderate (9–11 points), or ideal (12–14 points) based on prior research. ¹⁸ The number of ideal LS7 CVH metrics was calculated as the sum of ideal metrics, ¹³ ranging from 0 to 7, and was categorized as low (0–2 ideal metrics), moderate (3–4 ideal metrics), or high (5–7 ideal metrics). ¹⁹ Participants treated to ideal goals of blood glycemia, BP, or blood lipids were downgraded to intermediate level for these metrics.

LE8 CVH Metrics Measurement

A food frequency questionnaire obtained weekly intakes of fruits/vegetables, soda and sweetened beverages, fish, whole grains, low‐fat dairy products, and red meat. ²⁰ Physical activity was self‐reported by a structured 6‐point scale questionnaire on off‐work physical activity from 0 (physically inactive/sedentary) to 6 (highly active), ²¹ and weekly duration of regular sports practicing (excluding trips, do‐it‐yourself projects, gardening, and housework) over the past 12 months. Current tobacco smoking status or period since quitting if former smoker were self‐reported at inclusion. BMI was calculated as weight in kilograms divided by height in meters squared (kg/m²). Weight and height were measured in health examination centers by a trained nurse. Weight was measured using a nonautomatic weighing instrument based on the International Organization of Legal Metrology guidelines (OIML R 76‐1, Edition 2006). Standing height was measured with a fixed stadiometer to the nearest 0.1 cm. Sleep health was computed from a self‐reported sleep duration questionnaire in hours per night.

BP was measured using an automated oscillometric sphygmomanometer (OMRON 705 CP‐II/OMRON 705IT) after a 5‐minute rest in a supine position. A measurement was made for each arm, and a third was made for the reference arm with the higher systolic and diastolic BP values. The average value was used for systolic or diastolic BP. Total and high‐density lipoprotein cholesterol, triglycerides, and fasting blood glucose were measured by standardized tests (COBAS Integra 400 Plus). Hemoglobin A_1c was self‐reported.

Covariates and Stratification Variables

The region of origin was assessed by asking, “Where are you from? (1) Metropolitan France, (2) Overseas France, (3) Europe, (4) North Africa, (5) Sub‐Saharan Africa, (6) Asia, (7) Other.”Educational attainment was coded as: (1) less than high school, (2) ≤2 years after high school, and (3) >2 years after high school. Socioprofessional status was coded as: (1) employee/manual worker, (2) self‐employed/managerial, and (3) other to indicate employment status. Cohabitation status was determined from responses to the question, “In the home where you live more often, do you live as a couple?” “yes/no.” Households with an annual poverty threshold of <60% of the national yearly median income at inclusion were categorized as economically disadvantaged based on National Institute of Statistics and Economic Studies (INSEE [Institut National de la Statistique et des Études Économiques]) recommendations. The residence type was defined as rural or urban using the 2012 census at the commune level. Family history of CVD was defined by reported known CVD in first‐degree family members (eg, parents and siblings). Menopausal status and parity in female participants were self‐reported at inclusion. Depressive symptoms were assessed using the 20‐item Center for Epidemiological Studies Depression Scale (CES‐D) score and defined as ≥20 for women and ≥16 for men. The Alcohol Use Disorders Identification Test (AUDIT) score assessed alcohol use modality with a cutoff of ≥8, indicating detrimental alcohol use. Sleep problems were defined by self‐reported limitation in performing routine activities, eg, at home or work, at least for the past 6 months compared with similar‐aged people due to sleep impairment.

Statistical Analysis

Main Analyses

Sex‐specific multivariable mixed‐effects linear regression models with random intercepts for inclusion period (years 2012–2020) and birth cohort (5‐year intervals) were used to separately estimate the associations between SM status and LE8 score (range, 0 to 100), LS7 score (range, 0–14), and the number of ideal LS7 metrics (range, 0–7), using heterosexual adults of the same sex as comparators (ie, reference group). In addition, multivariable mixed‐effects linear or logistic regression models with random intercepts for inclusion year and birth cohort were used to estimate the associations between SM status and either: (1) individual LE8 metrics score (range, 0–100) or (2) meeting the ideal criteria for individual LS7 metrics (ideal versus nonideal levels). Regression models were adjusted for age, educational attainment, employment status, cohabitation status, AUDIT score, CES‐D score, household poverty, sleep problems, residence type, and family history of CVD. The assumptions and validity of the models are detailed in Data S1. Missing CVH metrics (from 3745 [2.21%] for the BMI metric to 23 435 [29.87%] for the diet metric) and missing covariates (from 6 [0.004%] for residence type to 46 116 [27.22%] for employment status) were handled by adjusted multiple imputations by chained equations, using 50 imputation data sets. All tests were 2‐sided, with a CI level of 95% and an α threshold of 0.05. The analyses were conducted using Stata 17 (StataCorp LLC) and R version 4.1.2 (R Core Team [2021], R: A Language and Environment for Statistical Computing, R Foundation for Statistical Computing).

Sensitivity Analysis

A complete case analysis was conducted to address the impact of missing CVH metrics and covariates. The analyses were further stratified on whether individuals had a new partner in the past 12 months and whether they had a stable partner at inclusion to minimize classification bias of lifetime sexual behavior. The analyses were also stratified by rural/urban residence and pregnancy or postmenopausal status in women to evaluate their potential influence on the associations. Interactions were then evaluated by including the corresponding multiplicative interaction terms in the models. Furthermore, those who declined to answer the sexual life questionnaire were excluded from the analyses. Last, associations between SM status and sleep problems were analyzed to examine the links between SM status and suboptimal sleep.

RESULTS

Figure 1 shows the study sample flow chart. After excluding participants with prior CVD (n=13 716), those missing data for inscription date (n=6), SM status (n=21 076), and region of origin (n=825). The study included 169 434 participants recruited from 2012 to 2020 (53.6% women; mean age, 45.99 years). Among 90 879 women, 0.61% were lesbian, 3.47% were bisexual, 92.83% were heterosexual, and 3.09% declined to disclose information about their sexual behavior. Among 78 555 men, 3.08% were gay, 3.50% were bisexual, and 90.37% were heterosexual, while 3.05% declined to disclose information about their sexual behavior.

Figure 1. Study sample flow chart.

CVD indicates cardiovascular disease.

Table 1 reports the study sample characteristics by SM status. SM individuals were, on average, younger, held less self‐employed/managerial positions, had high educational attainment, were less likely to be cohabiting, were more frequently living in poverty and urban areas, and had more frequent detrimental alcohol use patterns compared with heterosexual individuals. SM individuals also more frequently reported past suicide attempts, depression symptoms, antidepressant prescriptions, and anxiety disorders compared with heterosexual individuals. However, the latter 3 health concerns were lower in lesbian women compared with heterosexual women. In both women and men, those who declined to provide their SM status were older, less frequently held self‐employment or managerial positions, less frequently attained high education levels, and more frequently lived in poor households compared with heterosexual individuals.

Table 1.

Study Sample Characteristics by SM Status

	Lesbian women n=555	Bisexual women n=3149	Heterosexual women n=84 363	Declined to answer n=2812	Gay men n=2421	Bisexual men n=2748	Heterosexual men n=70 994	Declined to answer n=2392
Mean age (SD), y	42.44 (13.37)	40.23 (12.31)	45.62 (13.13)	51.34 (13.06)	41.62 (11.64)	44.03 (12.33)	46.56 (13.04)	50.60 (13.16)
Employment status, %
Employee/worker	77.48	75.61	76.85	68.42	77.24	74.56	76.00	70.28
Self‐employed/managerial	17.30	18.07	18.95	26.03	17.27	19.80	20.24	23.79
Other	5.23	6.32	4.20	5.55	5.49	5.64	3.76	5.94
Educational attainment, %
Less than high school	13.69	10.70	18.55	35.53	11.19	5.14	25.83	40.84
≤2 y after high school	19.10	17.08	17.30	17.43	14.21	14.52	16.41	15.38
>2 y after high school	67.21	72.21	64.15	47.05	74.60	70.34	57.76	43.77
Cohabitant, %	65.05	56.84	72.25	66.86	52.21	54.91	78.79	68.31
Household poverty, %	6.85	8.45	4.50	8.50	5.12	5.93	3.78	7.48
Do not know/declined to answer, %	6.13	3.52	5.24	18.88	3.06	2.33	4.24	18.14
Residence type, %
Rural	21.08	16.54	21.59	18.24	8.72	12.66	21.19	17.47
Urban	78.92	83.46	78.41	81.76	91.28	87.34	78.81	82.53
Excessive alcohol use, %*	16.76	30.28	11.31	7.68	32.63	41.32	30.53	19.86
Depression symptoms, %†	14.95	23.12	16.97	20.55	24.82	26.06	15.46	20.44
Antidepressant prescription, %	16.76	25.31	18.41	20.27	16.44	17.39	9.49	11.54
Anxiety‐related conditions, %‡	3.24	6.54	3.40	4.59	4.54	5.20	2.39	3.22
Suicide attempts, %	3.60	6.00	2.22	3.02	3.43	3.78	1.01	1.84
Region, %
French European territory	94.23	89.74	90.20	83.53	88.52	89.16	89.46	82.32
Non‐French European territory	5.05	8.57	8.62	14.15	9.38	8.92	9.65	15.38
Other	0.72	1.56	1.12	1.74	2.07	1.82	0.82	1.38
Declined to answer	0.00	0.13	0.06	0.57	0.04	0.11	0.06	0.92

Values are expressed as arithmetic mean and SD for numerical variables and percentages for categorical variables. SM indicates sexual minority.

^*AUDIT (Alcohol Use Disorders Identification Test) score ≥8 denotes a detrimental alcohol use pattern.

^† Depressive symptoms were defined by an Epidemiologic Studies‐Depression Scale score of ≥20 for women and ≥16 for men.

^‡ Limitation in performing routine activities (eg, at home or work) due to anxiety or obsessive–compulsive disorders in the past 6 months.

The distribution of LE8 score levels by SM status is reported in Table 2 and Figure 2, while the distribution of LS7 score levels by SM status is reported in Table 3 and Figure 3, respectively. Irrespective of SM status, women had systematically higher LE8 and LS7 scores than men. In women, lesbian and bisexual women had greater percentages of high LE8 scores (80–100) compared with heterosexual women, whereas the distribution of LS7 score levels did not differ by SM status. In men, gay and bisexual men had higher LE8 and LS7 scores than heterosexual men. Of note, women and men who declined to answer the sexual life questionnaire systematically had the lowest percentages of high LE8 and LS7 score levels. The LE8 score distribution by birth cohort (5‐year intervals) is presented in Figure S1, which indicates improvement in mean LE8 scores across generations in SM groups in both sexes. In SM and heterosexual women, however, mean LE8 scores have plateaued since the birth cohort group in 1986 to 1991, while in SM and heterosexual men, such a trend is not observed. Consistent trends were observed with the LS7 score (Figure S2).

Table 2.

Distribution of LE8 Score Levels and Metrics by SM Status

	Lesbian women n=555	Bisexual women n=3149	Heterosexual women n=84 363	Declined to answer n=2812	Gay men n=2421	Bisexual men n=2748	Heterosexual men n=70 994	Declined to answer n=2392
LE8 score, median (IQR)	71.88	71.81	71.25	68.88	68.13	65.00	63.13	61.88
	(8.13)	(16.25)	(15.86)	(15.99)	(16.58)	(17.50)	(16.88)	(16.17)
LE8 score level (0–100), %
Low,0–49	6.31	6.48	5.79	8.64	9.13	13.65	14.50	16.56
Moderate,50–79	68.65	69.86	71.74	74.32	75.80	75.87	77.26	76.59
High,80–100	25.05	23.66	22.47	17.03	15.08	10.48	8.24	6.86
CVH metrics
Diet score(IQR)	43.37	47.36	47.20	47.89	41.83	41.17	38.07	39.80
	(28.03)	(28.70)	(26.44)	(25.16)	(28.09)	(27.19)	(26.23)	(24.65)
Physical activity score	63.92	65.84	65.97	66.51	63.17	64.11	65.19	62.50
	(25.65)	(24.55)	(24.72)	(25.37)	(24.73)	(24.64)	(25.73)	(26.88)
BMI score	85.96	88.30	86.80	83.46	87.86	86.48	82.76	80.40
	(21.74)	(20.33)	(20.61)	(22.39)	(18.88)	(18.18)	(18.97)	(20.68)
Nicotine exposure score	68.73	55.58	72.48	76.04	69.53	59.27	68.81	72.69
	(37.92)	(41.78)	(36.50)	(34.72)	(39.58)	(40.95)	(36.94)	(35.06)
Sleep health score	67.30	64.39	64.90	60.91	64.64	61.01	62.12	60.36
	(34.02)	(34.70)	(34.35)	(35.60)	(33.77)	(34.52)	(34.38)	(35.27)
Blood glucose score	90.79	91.43	89.28	85.39	86.41	84.38	81.03	75.85
	(22.75)	(22.76)	(24.85)	(27.42)	(23.18)	(25.42)	(27.93)	(30.62)
BP score	70.03	76.95	70.99	62.97	57.69	56.51	51.07	48.19
	(28.43)	(25.85)	(28.04)	(29.22)	(25.50)	(25.74)	(24.74)	(24.61)
Blood lipid score	68.49	71.29	63.59	57.51	64.93	60.22	55.18	53.75
	(30.89)	(30.23)	(30.86)	(30.88)	(30.49)	(31.87)	(30.21)	(30.13)

Life's Essential 8 (LE8) score was computed as the average of 8 modifiable metric scores, including a modified diet score based on the Dietary Approaches to Stop Hypertension, which includes vegetables, whole grains, low‐fat dairy products, soda and sweetened drinks, and red meat, physical activity, body mass index, nonsmoking, sleep health, blood glucose, blood pressure (BP), and cholesterol. ¹⁴ Percentages are reported for categorical variables and arithmetic means (SDs) for numerical variables, unless otherwise indicated. P values are for analysis of variance and χ² tests.

BMI indicates body mass index; CVH, cardiovascular health; IQR, interquartile range; and SM, sexual minority.

Figure 2. Distribution of Life's Essential 8 (LE8) score levels by sexual minority status.

LE8 cardiovascular health (CVH) score, ranging from 0 to 100, is categorized as low (0–49), moderate (50–79 ideal metrics), or high (80–100 ideal metrics). The overall LE8 score is the average of CVH metrics body mass index, nonsmoking, diet, physical activity, sleep health, blood glucose, blood pressure, and non–high‐density cholesterol. ¹⁴

Table 3.

Distribution of LS7 Number of Ideal Metrics, LS7 Score Levels, and LS7 Ideal Metrics by SM Status

	Lesbian women n=555	Bisexual women n=3149	Heterosexual women n=84 363	Declined to answer n=2812	Gay men n=2421	Bisexual men n=2748	Heterosexual men n=70 994	Declined to answer n=2392
No. of LS7 ideal CVH metrics (0–7), %
Low,0–2 ideal metrics	29.73	28.58	29.24	37.20	36.60	45.49	52.42	58.19
Moderate,3–4 ideal metrics	52.43	53.16	52.86	49.11	51.01	45.74	40.62	35.70
High,5–7 ideal metrics	17.84	18.26	17.90	13.69	12.39	8.77	6.96	6.10
LS7 score (0–14), median (IQR)
LS7 score	9 (3)	9 (3)	9 (3)	9 (3)	9 (3)	8 (3)	8 (3)	8 (3)
LS7 score level, %
Low,0–8 points	34.77	32.96	33.74	42.71	41.14	50.04	55.89	62.63
Moderate,9–11 points	53.87	54.65	53.95	48.29	49.69	43.96	39.47	33.32
High,12–14 points	11.35	12.38	12.31	9.00	9.17	6.00	4.64	4.06
LS7 CVH metrics, %
Ideal diet	17.84	16.01	18.47	21.94	16.94	17.36	14.54	13.71
Ideal physical activity	23.96	24.42	25.81	28.27	21.44	22.05	25.96	22.62
Ideal BMI	63.96	70.28	65.61	57.33	66.71	61.21	49.31	44.61
Ideal nonsmoking	44.86	30.17	50.11	56.22	49.69	33.19	40.91	45.86
Ideal blood glucose	85.59	87.84	84.54	76.60	71.95	69.25	64.26	55.77
Ideal BP	33.87	43.00	35.09	24.89	13.34	12.48	8.01	6.44
Ideal total cholesterol	53.33	55.76	44.16	32.68	54.61	48.14	40.02	37.50

Ideal cardiovascular health (CVH) metrics correspond to body mass index (BMI) <25 kg/m², never smoker or former smoker >12 months, having 3 diet items at an ideal level, including consumption of fruits/vegetables (twice or more per day), soda or sweetened beverages(fewer than once per week), and fish intake (2 to 3 or more times per week), an off‐work activity score ≥5 (range 0–6), untreated fasting blood glucose <100 mg/dL, untreated systolic blood pressure (BP) <120 mm Hg and diastolic BP <80 mm Hg, and total untreated blood cholesterol <200 mg/dL. P values were estimated from analysis of variance, Kruskal–Wallis, or χ² tests examining differences in CVH by sexual orientation. Heterosexual participants of the same sex were the reference group for all analyses. IQR indicates interquartile range; LS7, Life's Simple 7; and SM, sexual minority.

Figure 3. Distribution of Life's Simple 7 (LS7) cardiovascular health (CVH) score levels by sexual minority status.

The LS7 CVH score, ranging from 0 to 14, is categorized as low (0–8 points), moderate (9–11 points), or high (12–14 points).

Figure 4 presents the results of the associations between SM status and LE8 score. After adjustment for confounders, lesbian (β=−0.95 [95% CI, −1.89 to −0.02]) and bisexual (β=−0.78 [95% CI, −1.18 to −0.38]) women had significantly lower LE8 scores compared with heterosexual women. In contrast, gay (β=2.72 [95% CI, 2.25–3.19]) and bisexual (β=0.83 [95% CI, 0.39–1.27]) men had significantly higher LE8 scores compared with heterosexual men. Furthermore, the associations between LS7 score, the number of ideal LS7 metrics, and SM status were consistent with LE8 score findings (Figure 5). In both sexes, LS8 scores, LS7 scores, and the number of ideal LS7 metrics of SM participants who declined to provide their SM status did not differ from those of heterosexual men and women.

Figure 4. Association between sexual minority status and Life's Essential 8 (LE8) score.

Linear regression models with random intercepts for the inclusion period and birth cohort (5‐year intervals) were used to estimate the adjusted regression coefficients (β) for LE8 score (range 0–100). The models were sex‐specific and adjusted for chronological age, region of origin, educational attainment, employment status, cohabitation status, depressive symptomatology (Center for Epidemiological Studies Depression Scale [CES‐D] score), alcohol use pattern (Alcohol Use Disorders Identification Test [AUDIT] score), household poverty, residence type, and cardiovascular disease family history.

Figure 5. Association between sexual minority status and Life's Simple 7 (LS7) score.

Number of ideal cardiovascular health (CVH) metrics denotes the number of ideal LS7 metrics (0–7). Linear regression models with random intercepts for the inclusion period and birth cohort (5‐year intervals) were used to estimate the adjusted regression coefficients (β) for discrete outcomes: number of ideal LS7 CVH metrics (0–7 ideal metrics) and LS7 CVH score (0–14 points). The models were sex‐specific and adjusted for chronological age, region of origin, educational attainment, employment status, cohabitation status, depressive symptoms (Center for Epidemiological Studies Depression Scale [CES‐D] score), alcohol use pattern (Alcohol Use Disorders Identification Test [AUDIT] score), household poverty, sleep problems, residence type, and cardiovascular disease family history.

Table S2 reports the associations between SM status and individual LE8 metrics. Compared with heterosexual women, lesbian women had lower diet and BP scores, whereas bisexual women had higher diet score and higher nicotine exposure. Women who declined to answer had lower scores for physical activity, BMI, blood glucose, and BP but lower nicotine exposure. The sleep health score was not different in any group of SM women compared with heterosexual women. Compared with heterosexual men, SM individuals had higher scores for each LE8 metric, except for the lower physical activity metric in gay men and higher nicotine exposure in bisexual men, respectively. Gay men but not bisexual men had higher sleep health scores, while post hoc analyses suggest that bisexual men are more likely to report sleep problems (Table S3). Men who declined to answer had lower physical activity, BP, blood glucose scores, and nicotine exposure but higher diet and blood lipids scores.

Table S4 reports the results of the association between SM status and individual LS7 CVH metrics (ideal versus nonideal levels); the findings were compatible with those of LE8 metrics.

Sensitivity Analyses

The results of sensitivity analyses are reported in Tables S5 and S6. The complete case analyses, excluding missing CVH metrics and covariates, stratified analyses into having a new partner or not in the past 12 months or having a stable partner or not and were consistent with the main analyses in association and direction. Stratified analyses by residence type suggest that SM men had higher LE8 and LS7 scores in urban areas compared with heterosexual men. However, in rural areas, the opposite was found (P for interaction <0.05); no such effect modification was found in SM women. Stratified analyses by parity status showed that among women who had ever been pregnant, there were no longer any significant LE8 and LS7 score differences between lesbian and heterosexual women. In contrast, among those who had never been pregnant, the lower LE8 and LS7 scores in lesbian women compared with heterosexual women remained significant (P for interaction <0.05). No such interaction was detected in bisexual women. Menopausal status did not modify the association between SM status and CVH scores in lesbian or bisexual women.

DISCUSSION

This French nationwide population‐based study depicts CVH disparities evaluated by LE8 or LS7 scores according to SM status in women and men. Accordingly, multivariate analyses suggest that lesbian and bisexual women had lower LE8 CVH scores compared with heterosexual women, whereas gay and bisexual men had higher LE8 CVH scores compared with heterosexual men. Consistent findings were observed when examining LS7 score.

The current estimates of SM individuals—4.21% women and 6.79% men—are slightly higher compared with past population estimates. Of note, estimates vary slightly across countries depending on whether data are derived from epidemiological studies, surveys, or census data and whether gender identity is included. ²² , ²³ Still, it is likely that estimates of SM participants in the present study were underestimated, as 3.07% of the participants in the CONSTANCES study did not disclose their sexual behaviors.

The majority of prior studies have explored disparities in the distribution of single CVD risk factors among SM individuals, ⁵ , ⁹ whereas only 2 studies have assessed co‐occurring CVD risk factors. ²⁴ , ²⁵ , ²⁶ In the NHANES (National Health and Nutrition Examination Survey), SM women were at an increased CVD risk—defined as the difference between vascular age based on a Framingham risk score and the expected risk for the participants' age—compared with heterosexual women. ²⁴ However, no difference was detected in SM men compared with heterosexual men. ²⁵ In Add Health (National Longitudinal Study of Adolescent to Adult Health), among self‐identifying females, SM females had a higher CVD risk defined by a 30‐year Framingham‐based prediction score compared with heterosexual women, while no difference was detected in bisexual and homosexual women, or any category of SM men compared with heterosexual men. ²⁶

To the best of our knowledge, this is the first population‐based study to examine the distributions of LE8 or LS7 scores per se in the SM population. The findings that women had an overall higher CVH than men, irrespective of their SM status, are consistent with prior general population evidence. ²⁷ Furthermore, the lower CVH scores in SM women compared with their heterosexual counterparts in the current study are consistent with the increased CVD risk observed in SM women in the NHANES and Add Health studies. ²⁴ , ²⁶ Our sensitivity analyses suggest that pregnancy is an effect modifier of CVH in lesbian women. Increased contact with the medical community and risk factor control during pregnancy‐related counseling may explain why CVH scores did not differ between lesbian and heterosexual women who had been pregnant. Nonetheless, caution in interpretation is required owing to the relatively small subsample of lesbian women who reported being ever pregnant.

The current study found that SM men had higher CVH scores than their heterosexual counterparts, whereas the opposite was found in women. This may suggest that: (1) levels of exposure to stressors, such as discrimination, might differ between SM men and SM women compared with heterosexual individuals; and (2) differences in the available coping resources and resilience to buffer against minority stressors might exist in SM women and SM men. ²⁸ , ²⁹

It is noteworthy that gay or bisexual men had higher LE8 and LS7 scores than heterosexual men, despite their observed psychosocial and socioeconomic disadvantages. As the study was conducted in France, universal health care access to most residents may have buffered against the detrimental effects of low socioeconomic status on CVH access and CVD prevention in SM individuals.

The higher CVH score observed in SM men compared with heterosexual men may underly a compensatory mechanism that buffers against disparities or discrimination. However, this may not be enough; one possible consequence may be that the association between CVH score and CVD risk in SM men is different than what is classically observed in the general population(ie, higher CVH scores may not be related to lower CVD risk). Future prospective studies examining the association between CVH score and CVD risk in SM populations are thus needed and researchers should consider disparities and discrimination in their investigations.

The observed lower percentages of SM men compared with SM women who are rural‐residing may indicate a social selection process whereby discrimination and minority stressors are increased for SM men compared with SM women. In light of this hypothesis, it can be argued that rural‐residing SM men are: (1) driven out of rural areas due to increased societal pressure and stressors or (2) less empowered to disclose their sexual orientation or SM status in a health research context. However, there was no detectable difference in declining to answer about sexual life questionnaire between men or women in rural versus urban areas in the present study (data not shown), making the second hypothesis unlikely. Furthermore, sensitivity analysis results indicate that in rural areas, SM men have lower CVH compared with heterosexual men after adjusting for demographic and SES variables. Therefore, one may hypothesize that either SM men are subjected to more minority stressors in rural areas, and/or the statistical power to detect the same effect in women was insufficient(ie, due to smaller subsample size or decreased effect size in SM women).

The study's findings on behavioral CVH metrics are mostly concordant with existing evidence. For example, SM individuals are more likely to use tobacco products and smoke cigarettes. ⁵ , ⁹ , ³⁰ While lesbian and bisexual women have an increased risk for obesity (BMI ≥30), gay men are less likely to be obese (BMI ≥30) compared with their heterosexual counterparts. ⁵ , ⁷ , ⁸ Of note, gay and bisexual men are more likely to engage in disordered eating behaviors than their heterosexual peers, which may not be reflected by BMI or diet scores. ⁸ , ³¹ The mixed results for the diet component are consistent with the varied evidence on dietary intake in SM individuals. ⁶ Achieving a higher LE8 diet score or ideal LE7 diet metric was more likely in SM men, while it was significantly lower in lesbian women but higher in bisexual women for the LE8 diet score. In addition, a recent review points out that food insecurity, a known risk factor for an unhealthy diet, may influence CVD risk in SM individuals and thus may explain lower diet scores in lesbian women. ³² Consistent with evidence of decreased engagement of SM men in sports from a young age (12–22 years), ³³ SM men were less likely to achieve a high LE8 physical activity score or LS7 ideal physical activity metric compared with heterosexual participants. More research is needed to identify determinants of access to practicing sports and physical activity and their trajectories in SM individuals.

Previous literature on the links between sleep health and SM status is mixed, showing no difference or increased risk of impaired sleep health, either in quantity(ie, sleep duration), or quality(ie, initiating and sustaining sleep). ³⁴ The current study findings suggest that the LE8 sleep health score(ie, sleep quantity in hours), was only higher in gay men compared with heterosexual men, while no differences were detected between SM women and heterosexual women. However, in post hoc analyses, sleep problems(ie, daily life limitation due to impaired sleep), were more likely only in bisexual men, with no detectable difference in SM women (Table S3). Such findings may indicate that sleep quantity and quality may be varyingly associated with different SM groups. Future research is needed to investigate potential mechanisms or determinants and mediators of sleep health in SM individuals, such as social and financial capital, family rejection, and biased harassment. ³⁵

The current study reports the findings on health factors such as blood glycemia, BP, and blood lipids in SM individuals, which are scarce in previous literature. For example, the PATH (Population Assessment of Tobacco and Health) study in the United States found that SM men were more likely to report high BP and high cholesterol compared with heterosexual men. However, no differences were detected between SM and heterosexual women. ⁸ Moreover, analyses of data from GGMHS (Geneva Gay Men's Health Survey) suggest that gay men are more likely to have self‐reported high cholesterol. ¹⁰ In the current study, lesbian women had lower BP scores compared with their heterosexual counterparts, while SM men had higher blood glucose, BP, and blood lipids scores. Notably, most previous studies only included self‐reported health factors, whereas objective measures of health factors were used in the current study, potentially contributing to the discrepancies among the findings.

Furthermore, the findings that those who declined to answer the sexual life questionnaire in both sexes were older, less frequently attaining high education levels, and more frequently living in poverty may reflect that they are subjected to more societal stigmatization and are not empowered to be informed on links between SM status and CVH or CVD risk. Those factors may also partially contribute to the observation that SM individuals had lower BMI, blood glucose, and BP scores, especially in women.

Minority stress is the principal theory that explains health disparities among marginalized populations, including SM individuals. ³⁶ This model was recently extended to address CVH disparities in SM populations. ⁵ The minority stress model may unite the biological and psychosocial basis of CVH disparities in the SM population. Cumulative evidence indicates that the SM population experiences mental health inequalities, increased experiences of discrimination, and subjugation to bias‐motivated violence. ³⁶ , ³⁷ Those stressors may propagate unhealthy coping mechanisms, such as alcohol use, smoking, sedentary lifestyle, and unhealthy eating, which subsequently contribute to higher CVD risk and CVH disparities in the SM population. ³⁸ In addition, those stressors have been shown to have direct effects on chronic inflammation and heart rate variability, potentially explaining the previous findings on the increased vascular age in SM women. ²⁴ , ³⁹ , ⁴⁰

Limitations and Strengths

This study has several limitations. The cross‐sectional study design limits causal inferences. Residual confounding due to unmeasured variables cannot be ruled out. No data were available on gender identity at the time of analysis, preventing the investigation of CVH differences between transgender and cisgender individuals. The current study was conducted in a high‐income country with universal health care insurance; thus, generalizing the findings to other settings should be cautioned. Finally, only one dimension of SM status was investigated, and future studies investigating CVH should incorporate multidimensional data on sexual orientation, sex, gender identity, and gender expression. ⁴¹ The current study has several strengths, including a large representative sample, evaluation of LE8 and LS7 scores, permitting to evaluate the consistency of the findings, and analyzing CVH factors, which are rarely studied in SM populations. In addition, regression analyses accounted for potential temporal effects of age, inclusion period, and birth cohort.

Implications of the Study

The reported disparities in CVH among SM populations, especially in women, highlight the importance of primordial prevention in this population. SM individuals report more frequent negative health care experiences than their heterosexual comparators. ⁴² , ⁴³ Therefore, increasing health care providers' cultural competency and awareness of CVD risk factors in the SM population can achieve a 3‐fold goal of: (1) providing accurate CVD risk communication; (2) improving their health care experiences; and (3) promoting trust in health care providers. Moreover, primary care settings represent a contact point for promoting optimal BP, BMI, physical activity, and smoking cessation in SM individuals, particularly SM women. ⁴⁴ , ⁴⁵ However, prior work has shown that SM adults are less likely to access health care and more likely to delay health care than heterosexual adults. ⁴⁶ , ⁴⁷ , ⁴⁸ , ⁴⁹ Moreover, given that health care access depends on country‐specific health system organization, gatekeepers such as health care cost payment must be considered for planning effective CVH promotion and CVD risk prevention by taking into account the socioeconomic disadvantage observed in SM individuals across numerous studies. ⁴⁸ , ⁴⁹ More specifically, the benefits of interventions that promote healthy weight should be carefully weighed, given the gains in CVH, while considering potential harms such as eating disorders and body image disturbance, which disproportionately affect SM individuals. ³¹ , ⁵⁰ From a public health perspective, the current data suggest that addressing socioeconomic inequalities, employment conditions, and mental health issues may help advance primordial CVD prevention in SM individuals.

CONCLUSIONS

The current study demonstrates CVH disparities in SM adults in France. SM women were less likely to attain ideal CHV scores compared with their heterosexual counterparts, whereas the opposite was observed for SM men. Future longitudinal studies will need to examine the associations between CHV change and incident CHV outcomes in SM individuals.

Sources of Funding

The CONSTANCES cohort study was funded by the CNAM; the Ministry of Health; the Council of Île‐de‐France region; and the cohorts TGIR IReSP‐ISP INSERM (Ministère de la Santé et des Sports, Ministère délégué à la Recherche, INSERM, INCa (Institut National du Cancer), and CNSA (Caisse Nationale de Solidarité pour l'Autonomie). The CONSTANCES cohort study is an Infrastructure Nationale en Biologie et Santé and benefits from a grant from the French National Research Agency, Agence Nationale de la Recherche (ANR‐11‐INBS‐0002). O.D. is funded by a grant from the FRM (Fondation pour la Recherche Médicale). The funders had no role in the study conceptualization, data analyses, manuscript drafting, or decision to publish the study and publisher choice.

Disclosures

None.

Supporting information

Data S1

Tables S1–S6

Figures S1–S2