Study Design and Participants
The Kailuan study is a prospective cohort study designed to identify the risk factors for common noncommunicable disease, especially CVD(14, 15). The study protocol and informed consent were approved by Ethics Committees of both the Kailuan General Hospital and Beijing Tiantan Hospital. All participants signed the written informed consent.
The details of the Kailuan study design have been described previously(16). At baseline, active and retired employees aged ≥ 18 years of the Kailuan Group, Tangshan, China, were invited to participate in this study. Generally, 101,510 participants (81,110 men and 20,400 women) with an age ranging between 18 years and 98 years, were enrolled and completed survey at baseline between June 2006 and October 2007. All participants underwent face-to-face questionnaire measurements, physical examinations, and laboratory assessments in the 11 local health care hospitals. We performed re-examinations biennially to the end of the follow-up on December 31,2017.
In the current study, we excluded 3,238 participants without data for any metabolic component at baseline, 3,358 participants with missing data on lifestyle risk factors, 83 participants with a history of myocardial infarction (MI) or stroke at baseline, finally, a total of 94,831 participants was selected for the current analysis (Online Fig. 1 in Supplemental material).
Metabolic Health Status
Metabolic health status at baseline was determined based on the physical examinations and laboratory assessment by trained nurses and physicians. Systolic blood pressure (SBP) and diastolic blood pressure (DBP) were measured 3 times with the participants in the seated position at least 5 minutes using a mercury sphygmomanometer, and the average of 3 readings was used for further analysis(17). Blood samples were collected after an overnight fast (8-h to 12-h) and measured the fasting plasma glucose (FPG), total cholesterol (TC), triglycerides (TG), high-density lipoprotein (HDL) cholesterol and low-density lipoprotein (LDL) cholesterol levels by an automatic analyzer (Hitachi 747; Hitachi, Tokyo, Japan) at local hospitals.
We used Adult Treatment Panel-III (ATP-III) criteria to define metabolic health status in the current study, which has been widely used to determine metabolic syndrome in adults worldwide. In the ATP-III criteria is based on the five CVD risk factors(18): 1) central obesity: waist circumference ≥ 90 cm in men and ≥ 80 cm in women; 2) elevated TG: TG level ≥ 1.69 mmol/L; 3) low HDL cholesterol: HDL cholesterol level < 1.03 mmol/L in men and < 1.29 mmol/L in women; 4) elevated BP: SBP/DBP ≥ 130/85 mmHg or taking antihypertensive drugs or self-reported history of hypertension; 5) elevated FPG: FPG level ≥ 5.6 mmol/L or taking hypoglycemic medications or self-reported history of diabetes. The metabolic health status ranged from 0 to 5, with lower scores indicating normal healthy metabolic, and were subsequently classified into three categories based on the distribution in this population: low risk (0–2 components), medium risk (3 components), and high risk (4–5 components) (Online Table 1 in Supplemental material).
Lifestyle Health Status
Lifestyle health status at baseline was collected by trained nurses and physicians using a standardized questionnaire interview. Current smoking was defined as smoking at least the previous year. Current alcohol consumption was defined as the average daily strong spirit (alcohol content > 50%) consumption of 100 ml or more than 100 ml for at least the previous year. Physical activity level was categorized as 1) ideally active: ≥80 minutes/week moderate and vigorous intensity; 2) moderately active: <80 minutes/week; 3) inactive: none. Sedentary behavior was classified into three categories: 1) < 4 hour/day; 2) 4–8 hour/day; 3) ≥ 8 hour/day. Considering salt intake plays an important role in the prevention of CVD in previous reports(19, 20), salt intake was used as a surrogate of health diet. The healthy diet was categorized as 1) ideal: < 6 g/day; 2) intermediate: 6–10 g/day; 3) poor: ≥10 g/day.
We estimated lifestyle health status in the current study according to five lifestyle risk factors: 1) current smoking; 2) current alcohol consumption; 3) physical inactivity: <80 minutes/week or none; 4) sedentary behavior: sedentary time ≥ 4 hour/day; 5) unhealth diet: salt intake ≥ 6 g/day. The lifestyle health status ranged from 0 to 5, with higher scores indicating unhealthy lifestyle, and were recorded as three categories: most healthy lifestyle (0–1 risk factor), moderately healthy lifestyle (2 risk factors), and least healthy lifestyle (3–5 risk factors) (Online Table 2 in Supplemental material).
Outcome Ascertainment
The present study participants were followed-up from the baseline examination at 2006 or 2007 up to December 31, 2017 as the end of the follow-up period, or to the date of a CVD event, or death, whichever came first. CVD events were defined as a composite of nonfatal MI and stroke during follow-up(21, 22). To retrieve potential CVD events, the subjects were linked to the Municipal Social Insurance and Hospital Discharge Register. All medical records including emergency department or hospitalized in local hospital were collected and adjudicated centrally. Stroke was defined according to the World Health Organization criteria on the basis of clinical symptoms, images obtained by computed tomography or magnetic resonance imaging, and other diagnostic reports(23). MI was defined based on cardiac enzymes levels, symptoms, electrocardiogram (ECG) signs and necropsy(24). Additionally, information on mortality was collected from vital statistics offices, with the death certificate reviewed by the study clinicians(21).
Statistical Analyses
The baseline characteristics were presented as mean ± standard deviation (SD) or median with inter-quartile range (IQR), or frequencies with percentages. Baseline characteristics across metabolic and lifestyle health status were compared using the ANOVA or Kruskal-Wallis tests for continuous variables and chi-square test for categorical variables.
The incidence rate of CVD, stroke, MI and all-cause mortality were reported as per 1,000 person-years (PY) with 95% confidence intervals (CIs). The Kaplan-Meier curves and the log-rank test was used to visual and test the significance of differences in the cumulative-incidence of clinical outcomes by metabolic and lifestyle health status. The multivariable adjusted hazard ratios (HRs) and 95% CIs for CVD, stroke, MI and all-cause mortality were calculated using Cox proportional hazards regression analysis after adjustments for covariates. These included age (continuous, years), sex (categorical, male or female), the family average monthly income (categorical, <¥800” or “≥ ¥800”), body mass index (BMI, calculated as continuous) and education (categorical, literacy/primary or middle school, high school or college/university). We first separately explored the association between lifestyle and metabolic risk and each clinical outcome. Moreover, an interaction between lifestyle and metabolic risk was tested by the likelihood-ratio test, and analyses were stratified by different metabolic risk category. Lastly, we assessed the joint association by creating a product-term between lifestyle and metabolic health status, with most healthy lifestyle and low metabolic risk group as reference.
All analyses were conducted using SAS version 9.4 (SAS Institute Inc., Cary, NC, USA). All reported P values were based on two-sided test of significance, and P < 0.05 was consider statistically significant in the current study.