Association of Spousal Diabetes Status and Ideal Cardiovascular Health Metrics With Risk of Incident Diabetes Among Chinese Adults

Key Points Question Is there an association between spousal diabetes status, ideal cardiovascular health metrics, and incident diabetes? Findings In this cohort study with 34 821 Chinese adults, spousal diabetes status with uncontrolled glycated hemoglobin level was independently associated with increased risk of incident diabetes. Achievement of ideal cardiovascular health metrics attenuated this risk. Meaning These findings suggest the potential benefit of couple-based lifestyle or pharmaceutical interventions for diabetes.


Introduction
The diabetes epidemic represents an escalating challenge worldwide. 1China also has witnessed a substantial increasing trend of diabetes over the past decades. 2The prevalence of diabetes in China has exceeded 11% since 2010, representing more than 100 million adults with diabetes, which indicated the urgent need for diabetes prevention and control not only globally, but especially in China. 3family history of diabetes is a critical contributor to diabetes risk and is frequently involved in diabetes risk scores. 4,5Previous epidemiologic studies have observed evident aggregation of patients with diabetes within households. 6,7This phenomenon could be attributable to heritable factors, but much of it remains yet to be explained, which may be captured by studies on spousal concordance.Spouses are generally genetically unrelated, yet they may share common nonheritable factors, such as living environments, social habits, and health behaviors. 8For this reason, cohabiting couples are often at risk of the same diseases, including cardiovascular disease, metabolic syndrome and diabetes. 9evious studies suggested that individuals with a spouse with diabetes were at an increased risk of developing type 2 diabetes, 10,11 whereas some studies did not find a significant spousal correlation of diabetes prevalence. 12The heterogeneity of research conclusions may be due to the difference in study design, participants, follow-up periods, and methods for diabetes diagnosis.Most of those studies were cross-sectional, 12 had a small sample size, 11 or indirectly ascertained diabetes (based on participants' self-reports or electronic medical records rather than systematic glucose testing). 13Thus, evidence from large-scale cohort studies using standard criteria for diabetes diagnosis is still required.Moreover, it remains unclear how to avoid or reduce the association of spousal diabetes status with individuals' risk of diabetes.In 2010, the American Heart Association (AHA) proposed 7 ideal cardiovascular health metrics (ICVHMs) to guide healthy behaviors and ideal metabolic parameters to lower cardiovascular disease (CVD) risk. 14Recently, the AHA updated the definition of ICVHMs to Life's Essential 8, which included diet, physical activity, nicotine exposure, sleep health, body mass index (BMI [calculated as weight in kilograms divided by height in meters squared]), blood pressure, blood glucose levels, and blood lipid levels. 15Multiple previous studies indicated that the components of ICVHMs were closely associated with a reduction of diabetes incidence, 16,17 but the association of comprehensive ICVHM status with the interaction between spouses' metabolic status requires further investigation.
In the cross-sectional analysis we launched in 2016, spouses of individuals with diabetes were found to have a higher prevalence of diabetes, obesity, metabolic syndrome, and CVD. 18We aim to further investigate the diabetes incidence of individuals whose spouses were diagnosed with diabetes and the association of the comprehensive ICVHM profile and spousal diabetes status with diabetes risk to examine the potential value of couple-based diabetes interventions.

Study Design and Population
The China Cardiometabolic Disease and Cancer Cohort (4C) Study is a multicenter, nationwide, prospective cohort study consisting of community-dwelling adults aged 40 years or older.The study protocol and informed consent were approved by the Committee on Human Research at Ruijin Hospital, affiliated with Shanghai Jiao Tong University School of Medicine, Shanghai, China.All participants provided written informed consent.This study followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline.
The details of the 4C Study design were described previously, 19,20 and more details of study visits and statistical methods are included in the eAppendix in Supplement 1.Among 193 846 participants examined at baseline in 2011 to 2012, 32 196 couples (64 392 individuals) participated, and 58 254 individuals (90.5%) attended the in-person follow-up visit in 2014 to 2016, with a mean (SD) follow-up time of 3.6 (0.9) years.The median (IQR) follow-up time was 3.2 (2.9-4.5)years.
Overall, 9.5% of the baseline participants (6138) were lost to follow-up.Of the remaining 58 254 individuals, 40 010 individuals without diabetes were selected by excluding 13 166 participants with diagnosed or screen-detected diabetes at baseline and 5078 participants with indeterminate diabetes status due to missing baseline information on plasma glucose measurement of their own or their spouse's.After excluding 5189 participants who did not complete fasting blood glucose or oral glucose tolerance test (OGTT) measurements at follow-up, 34 821 individuals entered the current analysis (eFigure 1 in Supplement 1).The distribution of study participants is shown in eFigure 2 in Supplement 1.

Data Collection
Data collection was conducted in local community clinics by trained staff according to standard protocols at baseline and follow-up visits.A questionnaire collecting information about demographic characteristics, lifestyle patterns (including smoking, drinking habits, physical activity, dietary habits, and sleep patterns), chronic diseases history, and medication usage was administered by trained personnel.Skilled nurses measured height, weight, and blood pressure of participants according to standard protocols.All participants underwent an OGTT after an overnight fast for at least 10 hours, and blood samples were collected at 0 and 2 hours during the test.Plasma fasting glucose (FPG) and 2-hour postload glucose (2h-PG) concentrations were evaluated at local hospitals using the glucose oxidase or hexokinase method.The levels of glycated hemoglobin (HbA 1c ) and lipids were examined at the central laboratory.HbA 1c was tested using finger capillary whole blood by high-performance liquid chromatography (VARIANT II Systems [Bio-Rad]).Serum total cholesterol, low-density lipoprotein cholesterol (LDL-C), high-density lipoprotein cholesterol (HDL-C).and triglycerides were tested using an autoanalyzer (Abbott Laboratories) at the central laboratory.Data types and definitions of covariates involved in the study are listed in eTable 1 in Supplement 1.

Definition of ICVHMs
The updated Life's Essential 8 proposed by the AHA included a point scoring system for each metric (ranging from 0-100 points). 15Accordingly, 7 metrics reaching 100 points were defined as ICVHMs: ideal smoking status, physical activity at goal, healthy sleep habits, ideal BMI (<23), untreated blood pressure of less than 120 mm Hg systolic blood pressure and less than 80 mm Hg diastolic blood pressure, untreated non-HDL-C of less than 130 mg/dL (to convert to millimoles per liter, multiply by 0.0259), and untreated fasting plasma glucose of less than 100 mg/dL (to convert to millimoles per liter, multiply by 0.0555) or HbA 1c of less than 5.7% (to convert to proportion of total hemoglobin, multiply by 0.01).The food frequency questionnaire of the previous 12 months was used to characterize dietary metrics and defined healthy dietary habits according to the modified components of the Mediterranean Eating Pattern for Americans (MEPA) score, which were more consistent with the dietary habits of the Chinese population (eTable 2 in Supplement 1).

JAMA Network Open | Diabetes and Endocrinology
Spousal Diabetes Status, Ideal Cardiovascular Health Metrics, and Risk of Diabetes

Diabetes Ascertainment
According to the 2010 American Diabetes Association criteria, 21 incident diabetes was diagnosed if (1) fasting plasma glucose was 126.1 mg/dL or greater, and/or (2) 2h-PG was 200.0 mg/dL or greater, and/or (3) HbA 1c level was 6.5% (Ն48 mmol/mol) or greater.Self-reported diagnosis by clinicians of diabetes were also included.

Statistical Analysis
The baseline characteristics of individuals according to their spousal diabetes status were presented as proportions in categorical variables, means and SDs in normally distributed variables, and medians and IQRs in nonnormally distributed variables.One-way analysis of variance was used to compare continuous variables, and χ 2 tests were used to compare categorical variables.
The cumulative incidence of diabetes was calculated for a mean follow-up of 3.6 years.The association of incident diabetes with spouse's diabetes status was examined using Cox proportional hazard analysis.Participants who did not have spouses with diabetes at baseline were the reference group.To investigate whether spousal HbA 1c level affected the association, the analysis was further performed in subgroups categorized by HbA 1c level of less than 7.0% and 7.0% or greater.Model 1 was adjusted for age, sex, high school education or greater, family history of diabetes, local personal income, and urban or not.Model 2 was further adjusted for obesity, hypertension, dyslipidemia, and prediabetes status based on model 1.Model 3 was further adjusted for diet score, physical activity, sleep, and smoking status based on model 2. Model 4 was further adjusted for 7 spousal ICVHM factors (obesity [BMI Ն23] hypertension, dyslipidemia, diet, physical activity, sleep and smoking status, excluding glucose status) based on model 3.
To comprehensively analyze the association of CVH status or ICVHM profile and spousal diabetes status with incident diabetes, the study population was stratified in 3 ways, according to the participant's ICVHM components and number, the spouse's ICVHM number, and a comparison of the numbers of ICVHMs of a participant and the spouse.The association between spousal diabetes status and incident diabetes was estimated in these strata.To compare the CVH status of the couples, 3 groups were created: (1) participant has more ICVHMs than spouse, (2) participant and spouse have same number of ICVHMs, and (3) participant has fewer ICVHMs than spouse.These 3 levels reflected the CVH status in a participant who had better, equal, or worse cardiovascular health markers than their spouse, respectively.To demonstrate possible interactions of ICVHM and spousal diabetes status in the development of diabetes, interaction terms were created using the cross products of spousal diabetes status with components or number of ICVHMs.The interaction was tested using the likelihood ratio test by comparing the full model including the interaction term with the reduced model excluding the interaction term.Six sensitivity analyses were performed.To test sex disparities, the main analysis was performed separately in women and men (eTables 3 and 4 and eFigures 3 and 4 in Supplement 1).To examine the robustness of the results, the main analysis was performed in a model further adjusting for alcohol consumption (eTable 5 in Supplement 1) or replacing prediabetes status with HbA 1c level (eTable 6 in Supplement 1).The validity of the main results was also tested in multiple imputation data sets imputed for missing baseline information (eTables 7-10 in Supplement 1) or outcome (eTable 11 in Supplement 1).To address the center effect, the major analysis was performed by random-effects models (eTable 12 in Supplement 1).SAS software version 9.4 (SAS Institute) was used for all statistical analyses from July 2022 to November 2022.All reported P values are nominal.Statistical significance was a 2-tailed P < .05.

Baseline Characteristics of the Study Population
Overall, 34 821 individuals were included, with a mean (SD) age of 56.4 (8.3) years and 16 699 (48.0%) male participants.As shown in the total cholesterol and LDL-C levels were higher in participants with spouses diagnosed with diabetes than in those whose spouses were not diagnosed with diabetes.There was no significant difference in FPG levels among them, possibly because elevated 2h-PG levels might develop earlier than elevated FPG levels at the early stage of metabolic disease in the Chinese population. 19Among participants with spouses who had uncontrolled blood glucose, we observed higher systolic blood pressure and lower HDL-C levels.The characteristics of individuals who developed and did not develop diabetes were also significantly different (eTable 13 in Supplement 1).Spouses had high concordance among most components of ICVHMs (eTable 14 in Supplement 1).

Association Between Spousal Diabetes Status and Incident Diabetes
During the follow-up of 3.6 years, 2896 individuals experienced incident diabetes (2564 [88.5%] had spouses without diabetes and 332 [11.5%] had spouses with diabetes).We analyzed the association between spousal diabetes status and incident diabetes (Figure 1).After multivariate without spouses diagnosed with diabetes, while the risk of incident diabetes in participants whose spouses had an HbA 1c level of less than 7.0% was not increased significantly (HR, 1.10; 95% CI, 0.92-1.30).The associations remained similar in the sensitivity analyses (eTables 5-12 in Supplement 1).
The analysis was performed separately in women and men.When husbands had diabetes with high HbA 1c levels, wives tended to have a higher risk of incident diabetes (HR, 1.25; 95% CI, 1.03-1.52),but when wives had diabetes, husbands tended not to have a higher risk of diabetes (HR, 1.14; 95% CI, 0.90-1.44)(eTables 3 and 4 in Supplement 1).

Associations of Individual ICVHMs and Spousal Diabetes Status With Incident Diabetes
As shown in Figure 2, there was no significant interaction between individual ICVHM components and spousal diabetes status for incident diabetes, with all P values for interaction greater than .05.

Associations of Spousal ICVHMs and Spousal Diabetes Status With Incident Diabetes
As Figure 3B shows, the risk of incident diabetes attributable to spousal diabetes diagnosis varied with spousal CVH status.If the spouses with diabetes had intermediate to ideal CVH status (Ն4 ICVHMs), the association between spousal diabetes diagnosis and incident diabetes was attenuated (4 ICVHMs: HR, 1.23; 95% CI, 0.79-1.90),but the association persisted if spouses had poor CVH status (<4 ICVHMs) (3 ICVHMs: HR, 1.39; 95% CI, 1.03-1.87).The results were similar in spouses with different composite CVH scores (eFigure 6 in Supplement 1).

Associations of Comparison of Numbers of ICVHMs Between Couples and Spousal Diabetes Status With Incident Diabetes
As Figure 4 shows, if a participant had better CVH status or a higher number of ICVHMs than their spouse, spousal diagnosis of diabetes was not significantly associated with risk of diabetes (HR, 1.17; 95% CI, 0.94-1.45);however, if a participant had an equal number of or fewer ICVHMs than their spouse, spousal diagnosis of diabetes was associated with incident diabetes (equal ICVHMs: HR, 1.58;   95% CI, 1.17-2.14;fewer ICVHMs: HR, 1.71; 95% CI, 1.31-2.24),the interaction of which was also significant (P = .04).The results were similar with results using categories of composite CVH scores (eFigure 7 in Supplement 1).The variation of association was more evident in wives than in husbands, with a P for interaction of .02 in wives compared with .24 in husbands (eFigures 3 and 4 in Supplement 1).

Discussion
In this nationwide prospective cohort study in China, we found that incident diabetes was associated with spousal diabetes status, and the risk was higher if spouses had HbA 1c levels of 7.0% or greater.
However, this association varied with the composite CVH status of husbands and wives, assessed by Life's Essential 8.The association between spousal diabetes diagnosis and incident diabetes was attenuated if either the participant or spouse had intermediate to ideal CVH status, or at least 4 ICVHMs.Better CVH status in participants than in their spouses also attenuated the association.To our knowledge, our study is one of the large-scale population-based cohort studies in the East Asian population to reveal the association of spousal diabetes status and composite cardiometabolic profiles with incident diabetes, diagnosed by comprehensive measurements of FPG, 2h-PG, and HbA 1c levels.
The association of spousal diabetes status with incident diabetes has been studied extensively.
The Atherosclerosis Risk in Communities (ARIC) study found a positive association between spousal diabetes status and the development of diabetes (HR, 1.20; 95% CI, 1.02-1.41). 10The authors also performed a meta-analysis of 17 studies, and the association remained significant.In a cross-sectional analysis we conducted in 2016, we also observed that having a spouse diagnosed with diabetes was associated with a 33% higher risk of diabetes in men and 35% in women. 18The current study confirmed in a large-scale Chinese cohort that spousal diabetes was associated with a 15% higher risk of incident diabetes after adjustment of important covariates, which is in line with previous findings.
There are several possible mechanisms for the association. 10,22One is assortative mating.Another refers to spouses converging in their lifestyles over marriage due to engaging in similar activities and habits.Further research on underlying mechanisms is needed.
Moreover, our study found that the risk of incident diabetes increased 20% when spouses had uncontrolled HbA 1c levels but did not increase if spouses had adequate HbA 1c control.The study suggested that good control of blood glucose could reduce rates of diabetes-related comorbidity and protect spouses of patients with diabetes from increased risk of diabetes.Strict control of glucose levels in patients with diabetes appears to be essential to benefit both the patients and their spouses.
The study also found that the diabetes risk attributable to spousal diabetes diagnosis was associated with healthy behaviors and metabolic profiles in couples.To reduce CVD morbidity and mortality, the AHA defined 7 components of ideal cardiovascular health in 2010 and updated the ICVHM to 8 components recently, named Life's Essential 8.The ARIC study defined participants with 0 to 2, 3 to 4, and 5 or more ICVHMs as having poor, intermediate, and ideal cardiovascular health, respectively. 23It was suggested in multiple studies that these metrics were associated with lower diabetes incidence.

Figure 3 .B
Figure 3. Association of Spousal Diabetes Status and Number of Individual Ideal Cardiovascular Health Metrics (ICVHMs) With Incident Diabetes

Figure 4 .
Figure 4. Association of Comparison of Numbers of Individual Ideal Cardiovascular Health Metrics (ICVHMs) Between Participants and Spouses and Spousal Diabetes Status With Incident Diabetes

JAMA Network Open | Diabetes and Endocrinology Spousal
Table, the group of participants whose spouses had been Diabetes Status, Ideal Cardiovascular Health Metrics, and Risk of Diabetes were older, had more education, and tended to live in urban areas than individuals whose spouses had not been diagnosed with diabetes.The lifestyle patterns of participants with spouses diagnosed with diabetes were generally less healthy than those whose spouses were not diagnosed with diabetes.BMI, systolic blood pressure, 2h-PG, HbA 1c level, and JAMA Network Open.2023;6(6):e2319038.doi:10.1001/jamanetworkopen.2023.19038(Reprinted) June 23, 2023 4/13 Downloaded From: https://jamanetwork.com/ on 09/29/2023 diagnosed with diabetes

Table .
Baseline Characteristics of Participants Without Diabetes by Spouse's Diabetes Status Diabetes was defined by FPG of 126.1 mg/dL or greater, and/or 2h-PG of 200.0 mg/dL or greater, and/or HbA 1c of 6.5% or greater, and/or self-report.Diabetes was defined by fasting plasma glucose level of 126.1 mg/dL or greater and/or a 2-hour plasma glucose level of 200.0 mg/dL or greater (to convert to millimoles per liter, multiply by 0.0555) and/or a hemoglobin A 1c level of 6.5% (to convert to proportion of total hemoglobin, multiply by 0.01), and/or self-report.Model 1 was adjusted for age, sex, high school education or greater, family history of diabetes, local personal income, and urban residence.Model 2 was further adjusted for obesity, hypertension, dyslipidemia and prediabetes status based on model 1.Model 3 was further adjusted for diet score, physical activity, sleep, and smoking status based on model 2. Model 4 was further adjusted for 7 spousal ICVHM factors (obesity, hypertension, dyslipidemia, diet, physical activity, sleep, and smoking status, excluding glucose status) based on model 3. Reference group was participants with spouses without previously diagnosed diabetes at baseline.HR indicates hazard ratio.diabetes diagnosis was associated with a 15% higher risk of incident diabetes (hazard ratio [HR], 1.15; 95% CI, 1.03-1.30).Notably, individuals whose spouses had an HbA 1c level of 7.0% or greater had a 20% higher risk of incident diabetes (HR, 1.20; 95% CI, 1.04-1.39)than people a d Obesity was defined as a BMI of 23 or greater.e Dyslipidemia was defined as a total cholesterol level of 240 mg/dL or greater and/or an LDL-C level of 160 mg/dL or greater and/or HDL-C level of 40 mg/dL or greater and/or triglyceride level of 200 mg/dL or greater and/or use of lipid-lowering medications.a P < .001.b P < .05. c P < .01.JAMA Network Open | Diabetes and Endocrinology Spousal Diabetes Status, Ideal Cardiovascular Health Metrics, and Risk of Diabetes JAMA Network Open.2023;6(6):e2319038.doi:10.1001/jamanetworkopen.2023.19038(Reprinted) June 23, 2023 6/13 Downloaded From: https://jamanetwork.com/ on 09/29/2023 adjustment, spousal However, variation was observed in the association with different ICVHM numbers or CVH status.As shown in Figure3A, if participants had intermediate to ideal CVH status (Ն4 ICVHMs), their diabetes risk associated with spousal diabetes status would be attenuated (4 ICVHMs: HR, 1.01; 95% CI, Model 1 was unadjusted; model 2 was adjusted for high school education or above, family history of diabetes, local personal income, and urban residence or not.HR indicates hazard ratio.

25 JAMA Network Open | Diabetes and Endocrinology 24.
16,17However, few longitudinal studies of couples have investigated whether the composite CVH profile was associated with spousal similarities in diabetes risk.In our study, the association between spousal diabetes diagnosis and incident diabetes varied with ICVHMs in couples.Our results support the potential benefits of couple-based lifestyle or pharmacologic interventions to promote metabolic health.It is advised that both individuals and their spouses with diabetes adopt a healthier lifestyle to reduce the diabetes risk in individuals.Recently, a randomized Jeemon P, Harikrishnan S, Ganapathi S, et al.Efficacy of a family-based cardiovascular risk reduction intervention in individuals with a family history of premature coronary heart disease in India (PROLIFIC): an openlabel, single-centre, cluster randomised controlled trial.Lancet Glob Health.2021;9(10):e1442-e1450. doi:10.1016/S2214-109X(21)00319-3 25. Liu Y, Xiao X, Peng C, et al.Development and implementation of couple-based collaborative management model of type 2 diabetes mellitus for community-dwelling Chinese older adults: a pilot randomized trial.Front Public Health.2021;9:686282.doi:10.3389/fpubh.2021.686282Hagedoorn M, Sanderman R, Ranchor AV, Brilman EI, Kempen GI, Ormel J. Chronic disease in elderly couples: are women more responsive to their spouses' health condition than men?J Psychosom Res.2001;51(5): 693-696.doi:10.1016/S0022-3999(01)00279-328.Pinquart M, Sörensen S. Gender differences in caregiver stressors, social resources, and health: an updated meta-analysis.J Gerontol B Psychol Sci Soc Sci.2006;61(1):33-45.doi:10.1093/geronb/61.1.P33 29.Stimpson JP, Peek MK.Concordance of chronic conditions in older Mexican American couples.Prev Chronic Dis.2005;2(3):A07.Data Types and Definitions of Covariates Involved in the Study eTable 2. Definitions of ICVHM eTable 3. Association Between Husband's Diabetes Diagnosis and Incident Diabetes in Women eTable 4. Association Between Wife's Diabetes Diagnosis and Incident Diabetes in Men eTable 5. Association Between Spousal Diabetes Diagnosis and Incident Diabetes in Models Further Adjusted for Alcohol Consumption eTable 6. Association Between Spousal Diabetes Diagnosis and Incident Diabetes in Models Replacing Prediabetes Status with Baseline HbA1c Level eTable 7. Association Between Spousal Diabetes Diagnosis and Incident Diabetes in Multiple Imputation Data Set Imputed for Missing Baseline Information eTable 8. Association of Individual ICVHMs and Spousal Diabetes Diagnosis With Incident Diabetes in Multiple Imputation Data Set Imputed for Missing Baseline Information eTable 9. Association of Spousal ICVHMs and Spousal Diabetes Diagnosis With Incident Diabetes in Multiple Imputation Data Set Imputed for Missing Baseline Information eTable 10.Association of Comparison of Numbers of ICVHMs Between Couples and Spousal Diabetes Diagnosis With Incident Diabetes in Multiple Imputation Data Set Imputed for Missing Baseline Information eTable 11.Association Between Spousal Diabetes Diagnosis With Incident Diabetes in Multiple Imputation Data Set Imputed for Outcome eTable 12. Association Between Spousal Diabetes Diagnosis With Incident Diabetes Estimated by Random-Effects Models eTable 13.Baseline Characteristics of Individuals Categorized by Diabetes Diagnosis at Follow-up and Spousal Diabetes Status at Baseline eTable 14.Percentage of Spousal Concordance by the ICVHMs eFigure 1. Participant Flow Diagram eFigure 2. Study Sites and Participant Distribution eFigure 3. Association of Comparison of Numbers of ICVHMs Between Couples and Spousal Diabetes Diagnosis With Incident Diabetes in Men eFigure 4. Association of Comparison of Numbers of ICVHMs Between Couples and Spousal Diabetes Diagnosis With Incident Diabetes in Women eFigure 5. Association of Individual CVH Score Categories and Spousal Diabetes Diagnosis With Incident Diabetes eFigure 6. Association of Spousal CVH Score Categories and Spousal Diabetes Diagnosis With Incident Diabetes eFigure 7. Association of Comparison of CVH Scores Between Couples and Spousal Diabetes Diagnosis With Incident Diabetes clinical trial (PROLIFIC) in the families of individuals with premature coronary heart disease found that a family-based CVD risk reduction intervention could effectively reduce total CVD risk in those patients.24Clinicaltrials are required to examine the efficacy of couple-targeted collaborative lifestyle and metabolic modification programs for type 2 diabetes.