Study population
KNHANES is a nationally representative cross-sectional study conducted by the Korea Disease Control Agency (KDCA) and the database has been annually collected from the non-institutionalized men and women aged more than1 year since 1998. Details of the KNHANES have been previously described and published elsewhere [24]. Briefly, a complex, multistage probability sampling design was used in the KNHANES to include a representative sample of the Korean population. The participants in this study were never-smoking adult (≥ 19 years) cancer survivors enrolled from the sixth and seventh KNHANES (2013–2018) with history of any type of cancer based on the self-reported questionnaire on cigarette smoking status and cancer diagnosis records, respectively. Of the 1,161 never-smoking adult cancer survivors, we excluded 331 participants with missing information on SHS exposure, cardiometabolic health or covariates to investigate the association of SHS exposure with cardiometabolic health in never-smoking adult cancer survivors. The final study population included 830 never-smoking adult cancer survivors (Fig. 2). The Institutional Review Board (IRB) of KCDA, which is in accordance with the guidelines of the Declaration of Helsinki approved the protocols of the research and data release for the KNHANES (2013-12EXP-03-5C). Prior to participating in the KNHANES, all subjects provided informed consent. The KCDA grants researchers access to the KNHANES database for research purpose upon approval (www.kdca.go.kr).
Secondhand smoke (SHS) exposure
Information on SHS for never-smoking adult cancer survivors was derived from the following self-report questionnaires in the KNHANES: “During the past week, were you ever exposed to SHS at work?”, “During the past week, were you ever exposed to SHS at home?”, and “During the past week, were you ever exposed to SHS at public places (e.g. public institution buildings, schools, libraries, public transportation, concert halls, tourist accommodations, restaurants, and etc.) except for the designated smoking areas?” Based on the responses to the self-report questionnaires, SHS exposure was defined as any SHS exposure at work, home or in public place. Assessment of SHS status from the self-report questionnaires in the KNHANES has been described in previous studies [25, 26].
Components of cardiometabolic health
Based on the blood pressure and anthropometric measurements, blood samples (after at least 8 h of fasting) and self-report measures on medication in never-smoking adult cancer survivors, cardiometabolic health were categorized as the following: hypertension (2017 American Heart Association, AHA/American College of Cardiology, ACC criteria), general and abdominal obesity (Korean Society for the Study of Obesity criteria modified from the World Health Organization definition for the Asian population), hyperlipidemia, hypertriglyceridemia, reduced high density lipoprotein cholesterol (HDL-C), and impaired fasting glucose (American Diabetes Association, ADA/World Health Organization, WHO criteria) as suggested in previous studies [27,28,29,30]. Hypertension (2017 AHA/ACC) was defined as systolic blood pressure (SBP) ≥ 130 mmHg and diastolic blood pressure (DBP) ≥ 80 mmHg or taking antihypertensive drugs. General and abdominal obesity was defined as body mass index (BMI) ≥ 25.0 kg/m2 and waist circumference (WC) ≥ 90 cm for men and WC ≥ 85 cm for women, respectively. Hyperlipidemia, hypertriglyceridemia, and reduced HDL-C were defined as total cholesterol ≥ 240 mg/dL or taking cholesterol lowering drugs, triglyceride ≥ 200 mg/dL or taking cholesterol lowering drugs, and HDL-C ≤ 40 mg/dL for men and ≤ 50 mg/dL for women, respectively. Impaired fasting glucose was defined as fasting serum glucose of 100–125 mg/dL.
Other variables
Self-report questionnaires and 24-h dietary recall in the KNHANES were used to obtain sociodemographic factors (age, sex, education level, household income, insurance type, marital status, residential area, occupation type), health behavior (regular aerobic exercise, muscle strengthening exercise, and alcohol consumption), dietary intake (total energy intake), cancer sites, and family history of cardiovascular disease (coronary heart disease and total stroke). Education level was defined as the highest education completed and was categorized as elementary school, middle school, high school, and ≥ college/university. Household income was calculated by dividing the sum of the income contributed from the total household members by the square root of the number of household members and categorized into quartiles. Insurance type was categorized as self-employed insured, employee insured, and medical aid according to the nationwide health insurance system (i.e. National Health Insurance Service) implemented in the Republic of Korea. Marital status was categorized as married and single (no history of marriage), widowed, divorced, or separated. Residential area was categorized into urban/metropolitan and rural based on the administrative units of the participants according to a previous study. Occupation type was categorized as manager, professional, manual labor, or unemployed based on the classification of job roles from a previous study. Physical activity was classified as regular aerobic exercise (engaging in at least 1.25 to 2.5 h of moderate to vigorous aerobic exercise per week) and muscle strengthening exercise (categorized as none, 1–2 times, 3–4 times, and ≥ 5 times per week) in accordance to the Physical Activity Guidelines for Americans, 2nd edition [31]. Alcohol consumption was categorized into none (non-drinkers) and habitual alcohol consumption (consuming at least one standard drink, which is equal to 10.0 g of alcohol). Cancer sites were classified as gastric cancer, colorectal cancer, breast cancer, cervical cancer, thyroid cancer, and other types of cancer (i.e. other than the above-mentioned cancer types) according to the self-report questionnaire [32]. Family history of cardiovascular disease was defined as family history of coronary heart disease or total stroke.
Statistical analysis
Survey regression analysis and Rao-Scott F-adjusted chi-square tests were used to compare the characteristics (sociodemographic factors, health behavior, dietary intake, cancer sites, and family history of cardiovascular disease as continuous and categorical variables) for never-smoking adult cancer survivors according to SHS exposure status. These characteristics of the participants were calculated using mean (standard error) for continuous variables and number (weighted percentage) for categorical variables using sampling weights in the KNHANES. For each dichotomous outcome of hypertension (2017 AHA/ACC), general obesity, abdominal obesity, hyperlipidemia, hypertriglyceridemia, reduced HDL-C, and impaired fasting glucose (ADA/WHO), we used multivariable logistic regression adjusted for sociodemographic factors (age, sex, education level, household income, marital status, residential area, and occupation type) for Model 1 to estimate the odds ratio (OR) and 95% confidence intervals (95% CI) for association of SHS exposure with cardiometabolic health. In addition, Model 2 was constructed from adding health behaviors (aerobic exercise, muscle strengthening exercise, alcohol consumption, and total energy intake) into the variables included in Model 1 and Model 3 was additionally adjusted for family history of cardiovascular disease including variables from the Model 2. All of the regression models were accounted for sampling weights in the KNHANES for the participants. In addition, we carried out the F-adjusted mean residual goodness-of-fit test and tested multicollinearity between the independent variables for each model (Supplemental Table 1).
For subgroup analyses, we stratified the participants into age (< 65 years and ≥ 65 years), sex (male and female), education level (university/college and < high school), household income (upper half and lower half), and occupation status (manager, professional, manual labor, and unemployed) and conducted multivariable logistic regression to investigate the association of SHS exposure and components of cardiometabolic health in each subgroup included in Model 3 using the sampling weights. In addition, we performed log likelihood test for interaction effects for variables used in the subgroup analyses. Data collection and statistical analyses were performed with SAS software version 9.4 (SAS Institute., NC, USA). P-values were two-sided and p-values less than 0.05 was considered statistically significant.