Ideal cardiovascular health and mortality: pooled results of three prospective cohorts in Chinese adults : Chinese Medical Journal

Secondary Logo

Journal Logo

Original Article

Ideal cardiovascular health and mortality: pooled results of three prospective cohorts in Chinese adults

Zhang, Yanbo1; Yu, Canqing2; Chen, Shuohua3; Tu, Zhouzheng1; Zheng, Mengyi3; Lv, Jun2; Wang, Guodong3; Liu, Yan3; Yu, Jiaxin4; Guo, Yu5; Yang, Ling6; Chen, Yiping6; Guo, Kunquan7; Yang, Kun7; Yang, Handong7; Zhou, Yanfeng1; Jiang, Yiwen1; Zhang, Xiaomin1; He, Meian1; Liu, Gang1; Chen, Zhengming6; Wu, Tangchun1; Wu, Shouling3; Li, Liming2,5; Pan, An1

Author Information
Chinese Medical Journal ():10.1097/CM9.0000000000002379, January 20, 2023. | DOI: 10.1097/CM9.0000000000002379

Abstract

Introduction

In 2010, the American Heart Association (AHA) proposed the ideal cardiovascular health (ICH), also known as the Life's Simple 7, which is defined by the presence of both ideal health behaviors (including non-smoking, physical activity at recommended levels, body mass index [BMI] of <25.0 kg/m2, and healthy diets consistent with current guidelines/recommendations) and ideal medical factors (including untreated blood pressure, fasting blood glucose [FBG], and blood total cholesterol [TC] levels of <120/80 mmHg, 5.6 mmol/L, and 5.2 mmol/L, respectively).[1] Those behavioral and metabolic factors are major determinants of the global disease burden, including but not limited to cardiovascular disease (CVD).[2]

Increasing studies have investigated the association of ICH with mortality. Our recent meta-analysis of 16 studies found that the highest ICH score category was associated with 45% to 61% risk reductions in all-cause and CVD mortality.[3] However, most studies were conducted in high-income and western countries, and studies in middle- or low-income countries are lacking. This is important considering that socioeconomic factors could be upstream determinants of cardiovascular health, and implementation of health policies varied in different countries.[4] Besides, previous studies have reported the associations of ICH with risks of CVD subtypes, cancer, and respiratory diseases[3,5-7]; however, few studies have investigated associations of ICH with mortality from subtypes of CVD,[8,9] cancer,[7] and respiratory diseases, which are important to the tertiary prevention of major non-communicable diseases.

Moreover, the ICH score included both behavioral and medical factors, while their interaction and independent associations with mortality were seldom investigated, which could clarify the roles of behavioral and medical metrics in mortality and facilitate risk stratification in healthcare.[10] Thus, we investigated associations of ICH with all-cause and cause-specific mortality in three Chinese cohorts, and we investigated the interaction and independent associations of behavioral and medical scores with mortality. Given the different study designs, data collection tools, and population characteristics, all analyses were performed in each cohort first, and the results were then pooled by meta-analysis approach according to previous studies.[11,12]

Methods

Ethics approval

All participants provided their written informed consent. The study was approved by the Ethics Committees of the China National Center for Disease Control and Prevention (Approval No. 005/2004) and the Oxford University (Approval No. 025–04) for the China Kadoorie Biobank (CKB) study, the Dongfeng General Hospital and the Huazhong University of Science and Technology (Approval Nos. 2008–03 and 2012–10) for the Dongfeng-Tongji (DFTJ) cohort, and the Kailuan General Hospital for the Kailuan study (Approval No. 200608).

Study population

This study used three ongoing cohorts in China: the CKB study, DFTJ cohort, and Kailuan study. Briefly, the CKB study recruited 512,891 residents aged 30 to 79 years from ten regions including five rural counties (from Gansu, Henan, Hunan, Sichuan, and Zhejiang provinces) and five urban cities (from Guangxi Zhuang Autonomous Region, Hainan, Heilongjiang, Jiangsu, and Shandong provinces) between 2004 and 2008.[13] Approximately 5% of the participants from the CKB study were resurveyed in 2008 and 2013 to 2014, respectively.[14] The DFTJ cohort is a dynamic cohort consecutively recruiting retirees from the Dongfeng Motor Corporation, Shiyan, Hubei Province, China,[15] which was launched in 2008 and recruited 27,009 participants in 2008 to 2010. The first follow-up was conducted in 2013; meanwhile, 14,120 new retirees were additionally enrolled. The Kailuan study is a dynamic cohort in the Kailuan community in Tangshan, Hebei Province, China.[16] It was launched in 2006 and recruited 101,510 employees aged 18 to 98 years of the Kailuan Group in 2006 to 2007. Subsequent surveys were conducted every two years, and 25,337, 10,519, and 21,651 new participants were additionally recruited in 2008 to 2009, 2010 to 2011, and 2012 to 2013, respectively. Participants were invited to complete questionnaires and medical examinations, and blood was drawn at each survey. Each participant's baseline was set at the first survey when all ICH metrics and covariates were measured.

The DFTJ cohort and Kailuan study planned to measure all participants’ blood lipid levels; however, the CKB study only measured blood lipids from 37,287 participants. The current analysis was based on 237,433 participants (37,287 with measured blood lipid markers from the CKB study, 41,129 from the DFTJ cohort, and 159,017 from the Kailuan study). We excluded 17,972 participants (1993, 6147, and 9832, respectively, from the three cohorts) with missing data on ICH metrics and 6834 participants (116 from the DFTJ cohort and 6718 from the Kailuan study) with missing covariates including demographic characteristics and alcohol consumption. Considering that diagnoses and treatments of cancer and CVD could change patients’ ICH metrics and their future mortality risks, 14,463 patients with cancer or CVD (1628, 7886, and 4949, respectively, from the three cohorts) were excluded to mitigate reverse causation according to previous studies.[3] The final analysis included 198,164 participants (33,666, 26,980, and 137,518, respectively, from the three cohorts; Supplementary Figure 1, https://links.lww.com/CM9/B171).

Assessment of ICH

We created an ICH score according to the AHA's definition,[1] and some modifications were made to accommodate the methodology of three cohorts [Supplementary Table 1, https://links.lww.com/CM9/B171].

Current status of cigarette smoking and years since quitting smoking were self-reported. The AHA defined quitting smoking for ≥1 year as ideal level; however, a previous study found that only quitting smoking for ≥10 years was consistently associated with lower mortality risks from all causes, total and subtypes of CVD, respiratory disease, and smoking-caused or other cancers compared with current smoking.[17] Besides, two-thirds of individuals who tried to quit smoking relapsed to smoking in China,[18] and the relapse rate keeps increasing before 8 years of quitting smoking.[19] Accordingly, smoking was categorized into ideal, never smoking or quitting smoking for ≥10 years; intermediate, quitting smoking for <10 years; and poor, current smoking.

Weekly time spent on moderate and vigorous activities was obtained in the CKB study, while weekly time spent on walking, bicycling, dancing, Tai chi, gymnasium, balls, jogging, swimming, and climbing were obtained in the DFTJ cohort. Since most participants from the DFTJ cohort met the recommendation of weekly 150-min moderate-to-vigorous leisure-time physical activity, we divided participants into tertiles according to leisure-time physical activity (ideal, top; intermediate, medium; and poor, bottom) in the DFTJ cohort and CKB study to make the results comparable. In the Kailuan study, one question (i.e., “How often did you exercise in recent 5 years?”) with three options (ideal, >80 min/week; intermediate, 1–80 min/week; poor, none) was applied according to a previous study.[16]

Dietary habits were determined via simplified food frequency questionnaires (including rice, wheat, whole grain, red meat, poultry, seafood, egg, fresh vegetable, beans, pickles, fresh fruit, and dairy intakes) in the CKB study and DFTJ cohort. According to a previous study and current recommendations emphasizing increasing vegetable and fruit intakes while limiting red meat intakes,[20] a dietary metric composed of daily consuming vegetable, daily consuming fruit, and not daily consuming red meat was created: ideal, three components; intermediate, two components; poor, 0 to 1 component. According to a previous publication from the Kailuan study,[16] salt intakes obtained by “What's your daily salt intake? ideal, <6 g; intermediate, 6–12 g; poor, >12 g” were used as a surrogate for dietary quality in the Kailuan study since no other dietary information was available. A previous study from Kailuan study (2013–2014) found a strong association between salt intake and a healthy diet score consisting of vegetables, fruits, whole grains, fish, sweets, sugar-sweetened beverages, and sodium (the age- and gender-adjusted mean healthy diet score was 1.87, 0.98, and 0.70 among those with ideal, intermediate, and poor salt intakes, respectively).[21]

Standing height and body weight were measured in participants with light indoor clothing and without shoes. BMI was calculated as weight in kilograms divided by height in meters squared: ideal, 18.5 to 24.9 kg/m2; intermediate, 25.0 to 29.9 kg/m2; poor, <18.5 kg/m2 or ≥30.0 kg/m2.

Resting blood pressure was measured: ideal, systolic blood pressure (SBP)/diastolic blood pressure (DBP) <120/80 mmHg without self-reported treatment; intermediate, 120–139/80–89 mmHg, or <120/80 mmHg with treatment; poor, ≥140/90 mmHg.

FBG was measured using the Aeroset automatic analyzer (ABBOTT Laboratories) in the DFTJ cohort and Hitachi 747 automatic analyzer (Hitachi, Tokyo, Japan) in the Kailuan study. Random blood glucose was measured by rapid dipstick testing using the Johnson SureStep Plus in the CKB study, and FBG would be measured the next day when one's random blood glucose was 7.8 to 11.0 mmol/L. Blood glucose health was categorized into ideal, FBG <5.6 mmol/L or random blood glucose <7.8 mmol/L without self-reported treatment; intermediate, FBG 5.6 to 6.9 mmol/L or random blood glucose 7.8 to 11.0 mmol/L, or FBG <5.6 mmol/L or random blood glucose <7.8 mmol/L with treatment; poor, FBG ≥7.0 mmol/L or random blood glucose ≥11.1 mmol/L.

Blood lipid levels were measured using the AU680 Chemistry Analyzers (Beckman-Coulter, Brea, CA, USA) in the CKB study, ARCHITECTCi8200 automatic analyzer (ABBOTT Laboratories) in the DFTJ cohort, and Hitachi 747 automatic analyzer (Hitachi) in the Kailuan study. Since most studies found no associations between ideal TC and mortality,[10,22-27] we used low-density lipoprotein cholesterol (LDL-C), a better prognostic factor for incidence and prognosis of CVD,[28] to evaluate blood lipid: ideal, LDL-C <3.4 mmol/L without self-reported treatment; intermediate, 3.4 to 4.0 mmol/L, or <3.4 mmol/L with treatment; poor, ≥4.1 mmol/L.[28]

For each metric, 2, 1, or 0 points were assigned for ideal, intermediate, or poor level, and the score ranged between 0 and 14. Based on previous studies,[29,30] ICH score was equally divided into three groups according to the range, that is, scores 0–4, 5–9, and 10–14. However, there were too few individuals with scores 0–4 (1.1% in the CKB study, 1.1% in the DFTJ cohort, and 1.5% in Kailuan study), and participants were regrouped into inadequate (scores 0–6), average (scores 7–9), and optimum (scores 10–14) to increase statistical power, which was also consistent with a previous study.[29] We further used four behavioral metrics (i.e., cigarette smoking, physical activity, diet, and BMI) to construct an 8-point behavioral score and three medical metrics (i.e., blood pressure, blood glucose, and blood lipid levels) to construct a 6-point medical score. To ensure sufficient statistical power for interaction analyses, participants were divided into three groups with roughly similar sample sizes according to the behavioral score (i.e., 0–4, 5–6, and 7–8) and medical score (i.e., 0–3, 4, and 5–6), respectively.

Assessment of covariates

Age, gender, region (only in the CKB study), marital status, education level, income level (the CKB and Kailuan studies), occupation (the CKB study), and alcohol drinking were self-reported. Prevalent hypertension (defined as self-reported physician diagnosis or uses of antihypertensive medications, or SBP ≥140 mmHg or DBP ≥90 mmHg) and prevalent diabetes (defined as self-reported physician diagnosis or uses of anti-diabetic medications, or FBG ≥7.0 mmol/L or random blood glucose ≥11.1 mmol/L) were also considered for stratified analyses.

Assessment of outcomes

In the CKB study, vital status and causes of deaths were determined via official death certificates supplemented with medical records, residential records, active visiting local communities, hospital records, or verbal autopsies through December 31, 2018. The Dongfeng Motor Corporation's healthcare service system covering all retirees was used to determine the mortality status of participants through December 31, 2018. In the Kailuan study, deaths were ascertained through discharge lists from local hospitals, official death certificates, and active follow-ups, through December 31, 2019, while causes of deaths, which need extra time for adjudication by investigators especially during the coronavirus disease 2019 (COVID-19) pandemic, are currently updated to December 31, 2016; thus, the end of follow-up was set as December 31, 2019 for all-cause mortality and December 31, 2016 for cause-specific mortality in the Kailuan study. International Classification of Diseases Tenth Revision (ICD-10) codes were used to identify deaths from cancer (C00-C97), CVD (I00-I99), ischemic heart disease (IHD, I20-I25), stroke (I60, I61, I63, and I64), and respiratory disease (J00-J99).

Statistical analysis

Covariate distributions across baseline ICH scores were compared using the Kruskal–Wallis test for continuous variables and chi-squared test for categorical variables. We also compared covariate distributions between those included in and excluded from the analyses due to missing information in each cohort, and an absolute standardized difference of <0.1 indicated a negligible difference between two groups.[31]

We then used Cox regression models to investigate the associations of ICH score and its components with mortality, the independent associations of behavioral and medical scores with mortality, and the interaction between behavioral score and medical score. We also conducted several subgroup and sensitivity analyses. We controlled for age, gender, region (in the CKB study), marital status, education level, income level (the CKB and Kailuan studies), occupation (the CKB study), and alcohol drinking in all analyses, and these covariates were selected a priori according to previous studies.[3] All analyses were first performed within each cohort, and then we pooled hazard ratios (HRs) with confidence intervals (CIs) from three cohorts using random-effects models of meta-analysis according to previous studies[11,12] since the results from three cohorts were largely consistent. We chose random-effects models to allow heterogeneity in different cohorts, including differences in study design, participant characteristics, data collection, and variable definitions. Dose-response meta-analyses were conducted using the Greenland and Longnecker method to test both linear and nonlinear relations between the ICH score and mortality.[32]

Cox regression models were used to investigate the association of the ICH score with all-cause mortality, and competing risks were considered for cause-specific mortality. The ICH score was included in models as continuous and categorical variables, respectively. We also investigated the independent association between each ICH metric and mortality.

To illustrate independent associations of combined behavioral metrics and combined medical metrics, we mutually adjusted for the behavioral and medical scores in models. We further investigated associations between the behavioral score and mortality across participants with different medical scores, and associations between the medical score and mortality across participants with different behavioral scores.

Stratified analyses were conducted by age, gender, education level, alcohol consumption, and prevalent diabetes and hypertension. Meta-regressions were used to estimate P values for the difference between subgroups. Several sensitivity analyses were also conducted. First, deaths occurring within the first 3 years of follow-up were excluded to reduce reverse causation. Second, we applied the original AHA's definition of ICH metrics without modification (except for diet in three cohorts and physical activity in the Kailuan study; Supplementary Table 1, https://links.lww.com/CM9/B171). Third, we additionally excluded individuals with prevalent emphysema or chronic bronchitis at baseline given that respiratory disease could change patients’ smoking habits. However, the Kailuan study did not collect the information on respiratory disease history, so we excluded no participants from the Kailuan study in this sensitivity analysis. Fourth, we additionally adjusted for family history of diabetes, CVD, and cancer to account for the genetic background. Fifth, to eliminate the impact of non-responses, we performed the multiple imputation with five imputed datasets to impute missing covariates according to non-missing variables. Sixth, a 7-point ICH score was constructed by assigning one point to the ideal level while 0 points to the intermediate or poor level for each metric.

The partial population-attributable risk fraction was calculated to estimate the percentage of premature deaths that would be avoided if all people had been with optimum cardiovascular health (scores 10–14), assuming causal associations.[33] We used the pooled HRs estimated from three cohorts and population prevalence from the CKB study, considering that the CKB study recruited participants from ten regions of China but the other two cohorts each recruited participants from one city and with less representativeness. Analyses within cohorts were performed by SAS software version 9.4 (SAS Institute, Cary, NC, USA), and meta-analyses were conducted by STATA software version 14.0 (StataCorp, College Station, TX, USA). Two-sided P values <0.05 were considered statistically significant.

Results

Baseline characteristics

Table 1 shows baseline characteristics in three cohorts. In the Kailuan study, participants were younger, and females only contributed a small proportion of participants; accordingly, non-drinking and ideal behavioral metrics were less prevalent. However, the distributions of ICH score were similar across three cohorts [Supplementary Tables 2–4, https://links.lww.com/CM9/B171]. In total, there were 20,445 (10.3%), 93,826 (47.3%), and 83,893 (42.3%) participants with inadequate (scores 0–6), average (scores 7–9), and optimum (scores 10–14) cardiovascular health. Participants with higher ICH scores were more likely to be younger, female, with higher levels of education, and abstainers, and the ideal level of each ICH metric was more prevalent among those with optimum cardiovascular health. There were modest differences in the distribution of age, gender, education level, alcohol consumption, and ICH metrics between those included in and excluded from the analyses due to missing information (absolute standardized differences ranged between <0.01 and 0.24; Supplementary Table 5, https://links.lww.com/CM9/B171).

Table 1 - Baseline characteristics of study population in CKB study, DFTJ cohort and Kailuan study.
CKB study DFTJ cohort Kailuan study
Characteristics (N = 33,666) (N = 26,980) (N = 137,518)
Age (years) 57.8 ± 10.3 61.7 ± 8.1 48.9 ± 13.5
Female 18,872 (56.1) 15,065 (55.8) 26,530 (19.3)
Married 29,354 (87.2) 24,396 (90.4) 129,697 (94.3)
High school or higher 5743 (17.1) 11,032 (40.9) 35,248 (25.6)
Alcohol consumption
 Never drinking 27,149 (80.6) 19,212 (71.2) 81,892 (59.6)
 Former drinking 885 (2.6) 1200 (4.4) 3294 (2.4)
 Current drinking 5632 (16.7) 6568 (24.3) 52,332 (38.1)
Ideal cardiovascular health
 Cigarette smoking 22,824 (67.8) 20,700 (76.7) 84,837 (61.7)
 Physical activity 11,250 (33.4) 8715 (32.3) 19,704 (14.3)
 Diet 2304 (6.8) 13,049 (48.4) 14,891 (10.8)
 Body weight 20,235 (60.1) 15,810 (58.6) 70,659 (51.4)
 Blood pressure 6179 (18.4) 5518 (20.5) 28,680 (20.9)
 Blood glucose 29,418 (87.4) 12,746 (47.2) 96,501 (70.2)
 Blood lipid 32,156 (95.5) 18,449 (68.4) 120,948 (88.0)
Data are shown as mean ± standard deviation or n (%).
The ideal level of cigarette smoking was defined as never smokers or former smokers who quitted smoking for no less than 10 years. The ideal level of physical activity was defined as the highest tertile of exercise time in the CKB study and DFTJ cohort and >80 min/week in the Kailuan study. The ideal level of diet was defined as daily vegetable intake, daily fruit intake, and no daily red meat intake in the CKB study and DFTJ cohort and daily salt intakes of <6 g in the Kailuan study. The ideal level of body weight was defined as BMI of 18.5–24.9 kg/m2. The ideal level of blood pressure was defined as SBP of <120 mmHg and DBP of <80 mmHg without treatment. The ideal level of blood glucose was defined as FBG of <5.6 mmol/L or random blood glucose of <7.8 mmol/L without treatment. The ideal level of blood lipid was defined as LDL-C of <3.4 mmol/L without treatment.BMI: Body mass index; CKB: China Kadoorie Biobank; DBP: Diastolic blood pressure; DFTJ: Dongfeng-Tongji; FBG: Fasting blood glucose; LDL-C: Low-density lipoprotein cholesterol; SBP: Systolic blood pressure.

ICH metrics and mortality

During 1,968,854 person-years of follow-up, 20,176 deaths were recorded, and 8164 were from CVD (2565 IHD deaths and 4662 stroke deaths), 3289 were from cancer, and 1112 were from respiratory diseases. All ICH metrics were associated with all-cause mortality, except for LDL-C; nevertheless, the association of LDL-C with mortality was stronger than those of other lipid markers [Supplementary Table 6, https://links.lww.com/CM9/B171]. HRs (95% CIs) comparing participants with optimum vs. inadequate cardiovascular health were 0.52 (0.41–0.67) for all-cause mortality, 0.44 (0.37–0.53) for CVD mortality, and 0.54 (0.44–0.66) for respiratory mortality [Table 2]. Although a higher ICH score was associated with lower cancer mortality in the DFTJ cohort [Supplementary Table 7, https://links.lww.com/CM9/B171], the pooled HR across three cohorts was not statistically significant (HR 0.86; 95% CI 0.64–1.14). A higher ICH score was also associated with lower IHD and stroke mortality. Dose-response meta-analyses showed that the ICH score was nonlinearly associated with CVD and stroke mortality (χ2 ≥ 5.49, P for nonlinearity ≤0.019) while was linearly associated with all-cause, IHD, and respiratory mortality (χ2 ≤ 2.28, P for nonlinearity ≥0.130; Supplementary Figure 2, https://links.lww.com/CM9/B171). Given causal relations, 25.0% (17.3–32.4%), 31.9% (23.1–40.2%), and 24.8% (2.9–44.4%) of all-cause, CVD, and respiratory mortality would be avoided if all individuals followed optimum cardiovascular health (ICH scores 10–14).

Table 2 - Associations between ICH score and all-cause and cause-specific mortality: pooled results of three prospective cohorts in Chinese adults.
ICH score

Causes of mortality 0–6 points 7–9 points 10–14 points 1-point increase
All-cause mortality
 Person-years 201,739 934,890 832,225 1,968,854
 Number of events 3014 11,249 5913 20,176
 HR (95% CI) Reference 0.73 (0.60–0.88) 0.52 (0.41–0.67) 0.88 (0.84–0.92)
CVD mortality
 Person-years 141,745 655,686 589,738 1,387,169
 Number of events 1475 4674 2015 8164
 HR (95% CI) Reference 0.69 (0.58–0.82) 0.44 (0.37–0.53) 0.85 (0.83–0.87)
 IHD mortality
 Number of events 466 1422 677 2565
 HR (95% CI) Reference 0.67 (0.56–0.81) 0.44 (0.39–0.50) 0.84 (0.83–0.86)
 Stroke mortality
 Number of events 848 2750 1064 4662
 HR (95% CI) Reference 0.71 (0.64–0.79) 0.44 (0.40–0.49) 0.85 (0.83–0.86)
Respiratory disease mortality
 Number of events 195 612 305 1112
 HR (95% CI) Reference 0.70 (0.59–0.82) 0.54 (0.45–0.66) 0.89 (0.86–0.92)
Cancer mortality
 Number of events 412 1782 1095 3289
 HR (95% CI) Reference 0.96 (0.77–1.21) 0.86 (0.64–1.14) 0.96 (0.93–1.00)
Models were adjusted for age, gender, region (only adjusted in CKB study), marital status, education level, income level (adjusted in CKB and Kailuan studies), occupation (only adjusted in CKB study), and alcohol drinking. Results from three cohorts were pooled using random-effects models of meta-analyses. CI: Confidence interval; CKB: China Kadoorie Biobank; CVD: Cardiovascular disease; DFTJ: Dongfeng-Tongji; HR: Hazard ratio; ICH: Ideal cardiovascular health; IHD: Ischemic heart disease.

Stratified analyses showed consistent associations between ICH and mortality among subgroups [Supplementary Tables 8–11, https://links.lww.com/CM9/B171]. However, the HRs (95% CI) of CVD mortality comparing the optimum vs. inadequate cardiovascular health were 0.32 (0.28–0.36) among those aged <60 years, which was stronger than 0.53 (0.38–0.75) among those aged ≥60 years (t = 3.12, Pbetween-group = 0.036). Results remained comparable when excluding participants dying within the first three years of follow-ups, using the original AHA's definition of ICH score, excluding participants with prevalent emphysema or chronic bronchitis at baseline, adjusting for the family history of major noncommunicable diseases, using the multiple imputation to account for missing variables, or using the 7-point ICH score [Supplementary Table 12, https://links.lww.com/CM9/B171 and Supplementary Figure 3, https://links.lww.com/CM9/B171].

Interaction between behavioral and medical scores

Figure 1 shows that both higher behavioral and medical scores were independently associated with lower all-cause and CVD mortality, whereas only a higher behavioral score was associated with lower respiratory and cancer mortality. No significant interactions were found between the behavioral and medical scores (t < 1.99, Pbetween-group ≥ 0.086; Figures 2 and 3 and Supplementary Figure 4, https://links.lww.com/CM9/B171). Compared with those with behavioral scores of 0–4, participants with the behavioral scores of 7–8 were associated with lower all-cause mortality across all medical score groups, and the HRs (95% CIs) were 0.77 (0.64–0.93), 0.79 (0.62–1.01), and 0.73 (0.65–0.82) for those with the medical scores of 0–3, 4, and 5–6, respectively. Similar patterns were also observed in cause-specific mortality. Additionally, a higher medical score was associated with all-cause and CVD mortality across all behavioral score groups, for example, HRs (95% CIs) of all-cause mortality comparing participants with the medical scores of 5–6 vs. 0–3 were 0.62 (0.44–0.88), 0.59 (0.46–0.75), and 0.56 (0.38–0.83) for those with the behavioral scores of 0–4, 5–6, and 7–8, respectively. Results for IHD and stroke mortality were similar to those for CVD mortality.

F1
Figure 1:
Independent associations of behavioral score and medical score with all-cause (A), cardiovascular disease (B), respiratory disease (C), and cancer mortality (D): pooled results of three prospective cohorts in Chinese adults. The model adjusted for age, gender, region, marital status, education level, income level, occupation, and alcohol drinking, as well as behavioral score and medical score mutually. CI: Confidence interval; HR: Hazard ratio.
F2
Figure 2:
Associations of the behavioral score with all-cause (A), cardiovascular disease (B), respiratory disease (C), and cancer mortality (D) across participants with different medical scores: pooled results of three prospective cohorts in Chinese adults. P between-group was derived from the meta-regression, which indicated whether the difference in HR associated with a 1-point increase of the score across groups was statistically significant. CI: Confidence interval; HR: Hazard ratio.
F3
Figure 3:
Associations of the medical score with all-cause (A), cardiovascular disease (B), respiratory disease (C), and cancer mortality (D) across participants with different behavioral scores: pooled results of three prospective cohorts in Chinese adults. P between-group was derived from the meta-regression, which indicated whether the difference in HR associated with a 1-point increase of the score across groups was statistically significant. Among participants with behavioral score of 7-8 points, the HR of medical score of 5-6 points for respiratory disease mortality exceeded the upper range of the y-axis, and the HR (95% CI) is 1.56 (0.78-3.09). CI: Confidence interval; HR: Hazard ratio.

Discussion

In these three Chinese cohorts with about 2 million person-years of follow-up, optimum cardiovascular health was associated with lower all-cause, CVD, and respiratory mortality. Besides, better behavioral and medical health were independently associated with lower all-cause and CVD mortality; however, only better behavioral health was associated with lower cancer and respiratory mortality. No significant interactions between behavioral and medical health were observed.

The results were largely consistent with previous studies. A recent meta-analysis with >0.5 million participants reported that the highest ICH score category was associated with 45% and 61% lower all-cause and CVD mortality.[3] A previous publication from the Kailuan study also found that meeting more ICH metrics was associated with lower all-cause and CVD mortality; however, the study only included participants recruited in 2006 to 2007 and the follow-up duration was too short (median = 4.02 years).[27] Additionally, the associations of ICH with mortality from CVD subtypes or non-CVD diseases have not been extensively studied and the results were inconclusive. Two studies investigated the association between ICH and IHD mortality. The US National Health and Nutrition Examination Survey with 13,312 participants and 576 IHD deaths found the highest ICH score category was associated with 70% lower IHD mortality which was similar to our results,[9] while a Lithuania study with fewer IHD deaths (9209 participants with 342 IHD deaths) found no significant association.[8] For cancer mortality, the US National Health and Nutrition Examination Survey[24] and Cardiovascular Health Study[10] found individuals in the highest ICH score groups were associated with 43% to 58% lower cancer mortality; however, the Aerobics Center Longitudinal Study found no significant association.[22] Similarly, we found that optimum cardiovascular health was associated with a 36% lower risk of cancer mortality in the DFTJ cohort, whereas a null association was observed after pooling the results with the other two cohorts. When separating behavioral and medical metrics, we only found that a higher behavioral score was associated with lower cancer mortality. Moreover, a recent meta-analysis with five studies and >0.5 million participants showed that the highest ICH score groups were associated with a 29% lower cancer risk.[7] Future studies are still warranted to reach a reliable conclusion regarding the association between ICH and cancer mortality. Our study reported associations of ICH with stroke and respiratory mortality for the first time and found participants in the highest ICH score group were associated with 56% and 46% lower risks of stroke and respiratory mortality, respectively. Previous studies also found a higher ICH score was associated with lower risks of stroke[3] and respiratory diseases.[5,6] This is the largest study to investigate the associations of ICH with mortality from subtypes of CVD and non-CVD disease (about 0.2 million participants, 16 times larger than the previous largest study[24]), and the results are robust.

In our study, associations between ICH and mortality were largely consistent in multiple stratified analyses; however, the association between ICH and CVD mortality was stronger among those aged <60 years compared with those aged ≥60 years, which was consistent with previous studies.[3,9] Given the looser targets of blood pressure and blood glucose levels as well as obesity paradox among elders,[34-36] it is advisable to modify the definition of each ICH metric targeting elders.

Both our study and previous studies found that higher behavioral and medical scores were independently associated with lower all-cause and CVD mortality.[37,38] Besides, a previous longitudinal study of 3491 elders also found no interaction between the American Cancer Society behavioral score and AHA medical score on all-cause mortality[10]; however, the sample size was small. Other studies also found no interactions between behavioral and medical scores on incident CVD and diabetes.[39,40] Hence, all individuals should target healthy lifestyles and cardiometabolic conditions regardless of their medical and behavioral health.

With a sufficient sample size and long follow-up duration, we investigated the associations of ICH with all-cause and cause-specific mortality in three Chinese cohorts, as well as the interaction and independent associations of the behavioral and medical scores with mortality. However, several limitations should be acknowledged. First, the definitions of some ICH metrics were modified because of data availability and distribution of characteristics in our study population. Nevertheless, results remained consistent when we used the original AHA's definition. Besides, the definitions of ideal levels of physical activity and diet were different between DFTJ cohort/CKB study and Kailuan study due to data availability, and the results need cautious interpretation. Second, the study sample was not nationally representative, and thus the estimated population-attributable risk fraction should be interpreted cautiously. Third, although we adjusted for major demographic characteristics and alcohol consumption, residual confounding is still possible, and causal inference cannot be made due to the nature of the observational study. Fourth, the participants dying within a mean follow-up of 9.9 years might have serious comorbidities at baseline, which could influence both the ICH metrics and mortality. Although we excluded those with CVD or cancer at baseline and the results remained largely unchanged in the sensitivity analyses of excluding those dying within the first three years or excluding those with prevalent respiratory disease at baseline, there is still the possibility of residual confounding and reverse causation from other unmeasured disease histories. Fifth, we used single-measured ICH metrics and covariates, and demographic and lifestyle information was self-reported. Thus, results might be biased by measurement errors. Sixth, there were modest differences in demographic, behavioral, and medical factors between participants included in and excluded from the analyses due to missing information, and selection bias was possible; however, the results remained consistent when using the multiple imputation to impute missing information.

In conclusion, meeting more ICH metrics was associated with lower all-cause, CVD, and respiratory mortality in the Chinese. Healthier behaviors and medical conditions were independently associated with lower all-cause and CVD mortality, and no interaction was observed. Associations of ICH with mortality from non-CVD or subtypes of CVD and the interaction between behavioral and medical metrics need further validation in other populations.

Funding

The study was supported by grants from the National Nature Science Foundation of China (Nos. 81930124 and 82021005). The Dongfeng-Tongji cohort was supported by the National Key Research and Development Program of China (Nos. 2016YFC0900801 and 2017YFC0907504). The China Kadoorie Biobank study was funded by the National Natural Science Foundation of China (Nos. 82192901, 82192900, and 81390540), the National Key Research and Development Program of China (No. 2016YFC0900500), and the Chinese Ministry of Science and Technology (No. 2011BAI09B01).

Conflicts of interest

None.

References

1. Lloyd-Jones DM, Hong Y, Labarthe D, Mozaffarian D, Appel LJ, Van Horn L, et al. Defining and setting national goals for cardiovascular health promotion and disease reduction: the American Heart Association's strategic Impact Goal through 2020 and beyond. Circulation 2010;121:586–613. doi: 10.1161/circulationaha.109.192703.
2. GBD 2019 Risk Factors Collaborators. Global burden of 87 risk factors in 204 countries and territories, 1990-2019: a systematic analysis for the Global Burden of Disease Study 2019. Lancet 2020;396:1223–1249. doi: 10.1016/s0140-6736(20)30752-2.
3. Zhang YB, Pan XF, Chen J, Cao A, Xia L, Zhang Y, et al. Combined lifestyle factors, all-cause mortality and cardiovascular disease: a systematic review and meta-analysis of prospective cohort studies. J Epidemiol Community Health 2021;75:92–99. doi: 10.1136/jech-2020-214050.
4. World Health Organization. Noncommunicable diseases progress monitor, 2017. Geneva: World Health Organization, 2017. Available from: https://www.who.int/nmh/publications/ncd-progress-monitor-2017/en/. [Accessed on May 5, 2019].
5. Fan W, Lee H, Lee A, Kieu C, Wong ND. Association of lung function and chronic obstructive pulmonary disease with American Heart Association's Life's Simple 7 cardiovascular health metrics. Respir Med 2017;131:85–93. doi: 10.1016/j.rmed.2017.08.001.
6. Ogunmoroti O, Allen NB, Cushman M, Michos ED, Rundek T, Rana JS, et al. Association between Life's Simple 7 and noncardiovascular disease: the multi-ethnic study of atherosclerosis. J Am Heart Assoc 2016;5:e003954. doi: 10.1161/JAHA.116.003954.
7. Zhang YB, Pan XF, Chen J, Cao A, Zhang YG, Xia L, et al. Combined lifestyle factors, incident cancer, and cancer mortality: a systematic review and meta-analysis of prospective cohort studies. Br J Cancer 2020;122:1085–1093. doi: 10.1038/s41416-020-0741-x.
8. Tamosiunas A, Luksiene D, Baceviciene M, Bernotiene G, Radisauskas R, Malinauskiene V, et al. Health factors and risk of all-cause, cardiovascular, and coronary heart disease mortality: findings from the MONICA and HAPIEE studies in Lithuania. PLoS One 2014;9:e114283. doi: 10.1371/journal.pone.0114283.
9. Yang Q, Cogswell ME, Flanders WD, Hong Y, Zhang Z, Loustalot F, et al. Trends in cardiovascular health metrics and associations with all-cause and CVD mortality among US adults. JAMA 2012;307:1273–1283. doi: 10.1001/jama.2012.339.
10. Greenlee H, Strizich G, Lovasi GS, Kaplan RC, Biggs ML, Li CI, et al. Concordance with prevention guidelines and subsequent cancer, cardiovascular disease, and mortality: a longitudinal study of older adults. Am J Epidemiol 2017;186:1168–1179. doi: 10.1093/aje/kwx150.
11. Sotos-Prieto M, Bhupathiraju SN, Mattei J, Fung TT, Li Y, Pan A, et al. Association of changes in diet quality with total and cause-specific mortality. N Engl J Med 2017;377:143–153. doi: 10.1056/NEJMoa1613502.
12. Yuan J, He Q, Nguyen LH, Wong MCS, Huang J, Yu Y, et al. Regular use of proton pump inhibitors and risk of type 2 diabetes: results from three prospective cohort studies. Gut 2021;70:1070–1077. doi: 10.1136/gutjnl-2020-322557.
13. Chen Z, Lee L, Chen J, Collins R, Wu F, Guo Y, et al. Cohort profile: the Kadoorie Study of Chronic Disease in China (KSCDC). Int J Epidemiol 2005;34:1243–1249. doi: 10.1093/ije/dyi174.
14. Chen Z, Chen J, Collins R, Guo Y, Peto R, Wu F, et al. China Kadoorie Biobank of 0.5 million people: survey methods, baseline characteristics and long-term follow-up. Int J Epidemiol 2011;40:1652–1666. doi: 10.1093/ije/dyr120.
15. Wang F, Zhu J, Yao P, Li X, He M, Liu Y, et al. Cohort profile: the Dongfeng-Tongji cohort study of retired workers. Int J Epidemiol 2013;42:731–740. doi: 10.1093/ije/dys053.
16. Zhang Q, Zhou Y, Gao X, Wang C, Zhang S, Wang A, et al. Ideal cardiovascular health metrics and the risks of ischemic and intracerebral hemorrhagic stroke. Stroke 2013;44:2451–2456. doi: 10.1161/STROKEAHA.113.678839.
17. Kenfield SA, Stampfer MJ, Rosner BA, Colditz GA. Smoking and smoking cessation in relation to mortality in women. JAMA 2008;299:2037–2047. doi: 10.1001/jama.299.17.2037.
18. China CDC. Global Adult Tobacco Survey (GATS): China 2010 Country Report. 2019. Available from: https://extranet.who.int/ncdsmicrodata/index.php/catalog/252/related-materials. [Accessed on February 14, 2022].
19. Hughes JR, Peters EN, Naud S. Relapse to smoking after 1 year of abstinence: a meta-analysis. Addict Behav 2008;33:1516–1520. doi: 10.1016/j.addbeh.2008.05.012.
20. Lv J, Yu C, Guo Y, Bian Z, Yang L, Chen Y, et al. Adherence to healthy lifestyle and cardiovascular diseases in the Chinese population. J Am Coll Cardiol 2017;69:1116–1125. doi: 10.1016/j.jacc.2016.11.076.
21. Wu S, An S, Li W, Lichtenstein AH, Gao J, Kris-Etherton PM, et al. Association of trajectory of cardiovascular health score and incident cardiovascular disease. JAMA Netw Open 2019;2:e194758. doi: 10.1001/jamanetworkopen.2019.4758.
22. Artero EG, España-Romero V, Lee DC, Sui X, Church TS, Lavie CJ, et al. Ideal cardiovascular health and mortality: aerobics center longitudinal study. Mayo Clin Proc 2012;87:944–952. doi: 10.1016/j.mayocp.2012.07.015.
23. Ford ES, Greenlund KJ, Hong Y. Ideal cardiovascular health and mortality from all causes and diseases of the circulatory system among adults in the United States. Circulation 2012;125:987–995. doi: 10.1161/circulationaha.111.049122.
24. Han L, You D, Ma W, Astell-Burt T, Feng X, Duan S, et al. National trends in American Heart Association revised Life's Simple 7 metrics associated with risk of mortality among US adults. JAMA Netw Open 2019;2:e1913131. doi: 10.1001/jamanetworkopen.2019.13131.
25. Muntner P, Judd SE, Gao L, Gutiérrez OM, Rizk DV, McClellan W, et al. Cardiovascular risk factors in CKD associate with both ESRD and mortality. J Am Soc Nephrol 2013;24:1159–1165. doi: 10.1681/asn.2012070642.
26. Kim JY, Ko YJ, Rhee CW, Park BJ, Kim DH, Bae JM, et al. Cardiovascular health metrics and all-cause and cardiovascular disease mortality among middle-aged men in Korea: the Seoul male cohort study. J Prev Med Public Health 2013;46:319–328. doi: 10.3961/jpmph.2013.46.6.319.
27. Liu Y, Chi HJ, Cui LF, Yang XC, Wu YT, Huang Z, et al. The ideal cardiovascular health metrics associated inversely with mortality from all causes and from cardiovascular diseases among adults in a Northern Chinese industrial city. PLoS One 2014;9:e89161. doi: 10.1371/journal.pone.0089161.
28. Stone NJ, Robinson JG, Lichtenstein AH, Bairey Merz CN, Blum CB, Eckel RH, et al. 2013 ACC/AHA guideline on the treatment of blood cholesterol to reduce atherosclerotic cardiovascular risk in adults: a report of the American College of Cardiology/American Heart Association Task Force on practice guidelines. Circulation 2014;129 (25 Suppl 2):S1–S45. doi: 10.1161/01.cir.0000437738.63853.7a.
29. Mok Y, Sang Y, Ballew SH, Rebholz CM, Rosamond WD, Heiss G, et al. American Heart Association's Life's Simple 7 at middle age and prognosis after myocardial infarction in later life. J Am Heart Assoc 2018;7:e007658. doi: 10.1161/JAHA.117.007658.
30. Olson NC, Cushman M, Judd SE, McClure LA, Lakoski SG, Folsom AR, et al. American Heart Association's Life's Simple 7 and risk of venous thromboembolism: the reasons for geographic and racial differences in stroke (REGARDS) study. J Am Heart Assoc 2015;4:e001494. doi: 10.1161/JAHA.114.001494.
31. Austin PC. Using the standardized difference to compare the prevalence of a binary variable between two groups in observational research. Commun Stat Simul Comput 2009;38:1228–1234. doi: 10.1080/03610910902859574.
32. Greenland S, Longnecker MP. Methods for trend estimation from summarized dose-response data, with applications to meta-analysis. Am J Epidemiol 1992;135:1301–1309. doi: 10.1093/oxfordjournals.aje.a116237.
33. Spiegelman D, Hertzmark E, Wand HC. Point and interval estimates of partial population attributable risks in cohort studies: examples and software. Cancer Causes Control 2007;18:571–579. doi: 10.1007/s10552-006-0090-y.
34. American Diabetes Association. 12. Older adults: standards of medical care in diabetes-2020. Diabetes Care 2020;43 (Suppl 1):S152–S162. doi: 10.2337/dc20-S012.
35. Lavie CJ, De Schutter A, Milani RV. Healthy obese versus unhealthy lean: the obesity paradox. Nat Rev Endocrinol 2015;11:55–62. doi: 10.1038/nrendo.2014.165.
36. Whelton PK, Carey RM, Aronow WS, Casey DE, Collins KJ, Himmelfarb CD, et al. 2017 ACC/AHA/AAPA/ABC/ACPM/AGS/APhA/ASH/ASPC/NMA/PCNA guideline for the prevention, detection, evaluation, and management of high blood pressure in adults: a report of the American College of Cardiology/American Heart Association Task Force on clinical practice guidelines. Hypertension 2018;71:e13–e115. doi: 10.1161/HYP.0000000000000065.
37. Han C, Liu F, Yang X, Chen J, Li J, Cao J, et al. Ideal cardiovascular health and incidence of atherosclerotic cardiovascular disease among Chinese adults: the China-PAR project. Sci China Life Sci 2018;61:504–514. doi: 10.1007/s11427-018-9281-6.
38. Jin Y, Tanaka T, Bandinelli S, Ferrucci L, Talegawkar SA. Overall cardiovascular health is associated with all-cause and cardiovascular disease mortality among older community-dwelling men and women. J Aging Health 2017;29:437–453. doi: 10.1177/0898264316635590.
39. Dong C, Rundek T, Wright CB, Anwar Z, Elkind MS, Sacco RL. Ideal cardiovascular health predicts lower risks of myocardial infarction, stroke, and vascular death across whites, blacks, and hispanics: the northern Manhattan study. Circulation 2012;125:2975–2984. doi: 10.1161/circulationaha.111.081083.
40. Li M, Xu Y, Wan Q, Shen F, Xu M, Zhao Z, et al. Individual and combined associations of modifiable lifestyle and metabolic health status with new-onset diabetes and major cardiovascular events: the China Cardiometabolic Disease and Cancer Cohort (4C) study. Diabetes Care 2020;43:1929–1936. doi: 10.2337/dc20-0256.
Keywords:

China; Health behavior; Ideal cardiovascular health; Metabolic health; Mortality; Noncommunicable diseases; Body mass index; Smoking; Alcohol; Exercise

Supplemental Digital Content

Copyright © 2023 The Chinese Medical Association, produced by Wolters Kluwer, Inc. under the CC-BY-NC-ND license.