Interactions between ALDH2 rs671 polymorphism and lifestyle behaviors on coronary artery disease risk in a Chinese Han population with dyslipidemia: A guide to targeted heart health management

We conducted a nested case-control study in a cohort of dyslipidemia subjects who participated in routine health examination at local community health centers during April and July 2013 in Yinzhou District of Ningbo, Zhejiang Province, China. Based on the electronic health records data set from Yinzhou Centers for Disease Control, we originally included 1359 unrelated participants, aged 40 to 80, who were diagnosed with dyslipidemia, and had not taken antihypertensive or cholesterol-lowering medications. Here, dyslipidemia was diagnosed in light of the Chinese guidelines on prevention and treatment of dyslipidemia 2007 version, according to which individuals with total cholesterol (TC) greater than 5.18 mmol/L, and/or triglyceride (TG) greater than 1.70 mmol/L, and/or high-density lipoprotein cholesterol (HDL-C) less than 1.04 mmol/L, and/or low-density lipoprotein cholesterol (LDL-C) greater than 3.37 mmol/L were defined as dyslipidemia.

Informed consent was obtained from all individual participants included in the study, and the study protocols were approved by the Medical Ethical Committee of the Affiliated Hospital of Hangzhou Normal University. After a 3-year follow-up period, in August 2016, 161 newly diagnosed CAD cases, and 495 control cases, matched for age and gender, were recruited from the cohort of dyslipidemia population. The CAD was confirmed according to the diagnostic criteria for coronary atherosclerotic heart disease of China, which is approved by the National Health and Family Planning Commission of the People’s Republic of China [34]. This study was limited to Han population in eastern China.

When the cohort of dyslipidemia population was confirmed in 2013, the demographic characteristics, lifestyle behaviors, and medical history information was obtained using a self-made questionnaire by a face-to-face interview. The dietary intake preference was assessed by a semi-quantitative food frequency questionnaire.

In this survey, the current smoking was defined as smoking ≥ 1 cigarette per day for ≥ 1 year; the current alcohol intake was defined as drinking ≥ 500 g of beer, 150 g of wine, or 50 g of distilled spirit per day for ≥ 1 year; and the physical activity intensity was classified into four sorts, which were sedentary, light, moderate, and heavy, respectively. The sedentary-intensity physical activity was defined as activities of office staff who was doing almost no daily exercise. The light-intensity physical activity was defined as activities of waiters, salesmen, and teachers, which demanded the least amount of effort. The moderate-intensity physical activity was defined as activities which required prolonged period of walking, pushing or pulling objects. The heavy-intensity physical activity was defined as activities which involved tough effort and frequent extensive body movement, such as activities for non-mechanized farming, construction workers, dockworkers, and dancers. The frequency of fried food and dessert intake was classified into two sorts, which were never and regular intake, respectively. The frequency for the never intake was defined as ≤ once a week, and for the regular intake was defined as ≥ 1 to 4 times per week.

The anthropometric data, including waist, body mass index (BMI), systolic blood pressure (SBP), diastolic blood pressure (DBP), and TC, TG, HDL-C, LDL-C levels was assessed in the physical examination by professional staff in accordance with a standard protocol.

5 ml peripheral blood samples were collected from the participants in both cases and controls in 2013. Genomic DNA was extracted from the peripheral blood samples using TIANamp Blood DNA Kits (Tiangen Biotech, Beijing, China) following the manufacturer’s protocol. ALDH2 rs671 polymorphism genotyping was performed using polymerase chain reaction-ligase detection reaction (PCR-LDR) method (Generay Biotech Company, Shanghai, China). Primer sequences and reaction conditions used were described previously [17]. For quality control, we randomly chose 10% of the samples to re-genotype, and the concordance was 100%.

The analysis of demographic characteristics, as well as the interaction between ALDH2 rs671 polymorphism and lifestyle factors on the risk of CAD, was conducted in SPSS 24.0 software (SPSS Inc., Chicago, IL, USA). Specifically, the differences in variables between CAD cases and control subjects were evaluated by using Pearson chi-squared test for categorical variables expressed as frequencies (percentages, %), Student’s t test for normally distributed continuous variables expressed as mean (standard deviation, SD), and Mann-Whitney U test for non-normally distributed continuous variables expressed as median (interquartile range, IQR). The interaction between ALDH2 rs671 polymorphism and lifestyle factors on CAD was evaluated by using unconditional logistic regression under a dominant model after adjustment for potential confounders [35]. In the logistic regression analysis, the dummy variables which represented the subgroups of the ALDH2 rs671 polymorphism and lifestyle factors were created. In addition, in PLINK 1.9 software incorporated into gPLINK (version 2.050), the association of ALDH2 rs671 polymorphism with the risk of CAD was evaluated by using logistic regression after adjusting for potential confounders in three genetic models, which are additive, dominant, and recessive models respectively [36]. Finally, tests of interaction between ALDH2 rs671 polymorphism and lifestyle factors on an additive scale were conducted by using relative excess risk due to interaction (RERI), as described by Knol et al. [29, 37]. The statistical significance level was set at P value< 0.05 (two-sided), and the Benjamini–Hochberg procedure with a false discovery rate at 0.12 was used to adjust for multiple comparisons concerning all the 46 tests presented in this study [38].

Table 1 shows that CAD cases and controls did not differ with respect to age. In addition, the two study groups were similar with respect to waist, body mass index (BMI), diastolic blood pressure (DBP), total cholesterol (TC) level, high-density lipoprotein cholesterol (HDL-C) level, low-density lipoprotein cholesterol (LDL-C) level, current smoking status, current alcohol intake status, as well as the history of diabetes mellitus. Whereas, the systolic blood pressure (SBP) and total cholesterol (TC) level were higher in the controls compared with the CAD cases (145 mmHg versus 142 mmHg, P value = 0.029; 1.58 mmol/L versus 1.43 mmol/L, P value = 0.012). Besides, there were significant differences in physical activity intensity, fried food intake, dessert intake, and distribution of rs671 genotypes. These results were still significant using the Benjamini–Hochberg procedure with the false discovery rate at 0.12. Other lifestyle behavior characteristics of the participants are presented in the Additional file 1: Table S1.

The distribution of the ALDH2 rs671 genotypes in controls was in Hardy–Weinberg equilibrium. Table 2 demonstrates the association of ALDH2 rs671 polymorphism with the risk of CAD in three genetic models. In additive model, carriers with an increasing number of A alleles appeared to have an elevated risk for CAD compared with individuals with G alleles (adjusted OR = 1.483, 95% CI = 1.114–1.974, P value = 0.007). Therefore, the A allele was regarded as the risk allele for CAD, and the G allele was considered as the non-risk allele in our study. In the dominant model, carriers of the rs671 A/G and A/A genotypes tended to have a higher risk for CAD than carriers of G/G genotype (adjusted OR = 1.492, 95% CI = 1.036–2.148, P value = 0.032). Similarly, in the recessive model, individuals of the rs671 A/A genotype tended to have a higher risk for CAD than individuals of A/G and G/G genotypes (adjusted OR = 2.161, 95% CI = 1.139–4.101, P value = 0.018). All the ORs were adjusted for potential confounders including age, gender, BMI, TG level, TC level, HDL-C level, SBP, current smoking status, current alcohol intake status, and the history of diabetes mellitus. These associations maintained significance when using the Benjamini–Hochberg procedure to adjust for multiple comparisons.

Table 3 indicates the joint analysis on the interaction between ALDH2 rs671 polymorphism and dessert intake on the CAD risk. After adjusting for potential confounders, compared with individuals with G/G genotype and no dessert intake preference, individuals with G/G genotype and regular dessert intake preference, as well as individuals with A/A and A/G genotypes and regular dessert intake preference had an elevated CAD risk (adjusted OR = 2.095, 95% CI = 1.113–3.944, P value = 0.022; adjusted OR = 2.539, 95% CI = 1.335–4.829, P value = 0.004, respectively). Within the A/A and A/G genotypes stratum, regular dessert intake did not significantly increased the risk for CAD, compared with no dessert intake. No significantly positive interaction between the A/A and A/G genotypes and the regular dessert intake was found on the additive scale.

Table 4 presents the interaction between ALDH2 rs671 polymorphism and fried food intake on the risk for CAD. After adjustment for potential confounders, compared with individuals with G/G genotype and no fried food intake preference, there was an increased CAD risk among individuals with both A/A and A/G genotypes and regular fried food intake preference (adjusted OR = 1.859, 95% CI = 1.074–3.217, P value = 0.027). Within the A/A and A/G genotypes stratum, regular fried food intake was associated with an elevated risk for CAD, compared with no fried food intake (adjusted OR = 1.949, 95% CI = 1.117–3.400, P value = 0.019). In addition, within the regular fried food intake stratum, carriers with A/A and A/G genotypes had an increased risk for CAD, compared with carriers with G/G genotype (adjusted OR = 1.715, 95% CI = 1.057–2.780, P value = 0.029). These associations remained significant by the Benjamini–Hochberg procedure. Borderline interaction between the A/A and A/G genotypes and regular fried food intake was found (RERI = 0.821, 95% CI = − 0.007 to 1.649, P value = 0.052).

Table 5 shows the interaction between ALDH2 rs671 polymorphism and physical activity intensity on the CAD risk. After adjustment for potential confounders, compared with individuals with G/G genotype and moderate/heavy physical activity, there was an elevated risk for CAD among individuals with both A/A and A/G genotypes and sedentary/light physical activity (adjusted OR = 2.071, 95% CI = 1.013–4.234, P value = 0.046). Within the A/A and A/G genotypes stratum, sedentary/light physical activity conferred an increased risk of CAD, compared with the moderate/heavy physical activity (adjusted OR = 2.612, 95% CI = 1.202–5.674, P value = 0.015). Besides, within the sedentary/light physical activity stratum, individuals with the A/A and A/G genotypes had a higher CAD risk than those with the G/G genotype (adjusted OR = 1.540, 95% CI = 1.012–2.344, P value = 0.044). Significantly positive interaction between the A/A and A/G genotypes and the sedentary/light physical activity was found (RERI = 0.933, 95% CI = 0.021 to 1.845, P value = 0.045). These associations maintained significance using the Benjamini–Hochberg correction for multiple comparisons.