The clinical characteristics and biochemical parameters of the Slovenian subjects with T2DM are listed in Table 1. Cases (335 subjects with MI) had lower BMI, lesser waist circumference and better-controlled hypertension. Additionally, they had a higher total and LDL-cholesterol, triglycerides, and lower HDL-cholesterol. Moreover, cases had longer duration of T2DM. The two groups of subjects were well-matched with regard to age, gender, fasting glucose, HbA1c, hsCRP level and concomitant history of cerebrovascular insult (CVI) or transitory ischemic attack (TIA).
To address the issue of the oxidative stress status for CAD in subjects with T2DM without history of previous MI, the oxidative damage of DNA was evaluated by measuring the serum level of 8-OHdG. However, in the subgroup of 115 subjects with T2DM no significant difference in the median serum level was observed when 23 subjects with CAD (CAD +) and 92 subjects without CAD (CAD -) were compared [1.47 (range: 1.31 – 1.71) vs. 1.51 (1.3 – 1.73) ng/mL], respectively (p=0.80, Mann-Whitney U-test). In addition, among the 34 rs6060566 C allele carriers, median serum level was higher in 9 CAD + subjects when compared to 25 CAD - subjects [1.54 (range: 1.4 – 1.72) vs. 1.48 (1.33 – 1.69) ng/mL], respectively, but the difference was not statistically significant (p=0.48, Mann-Whitney U-test).
Furthermore, the ROC curve analysis was used to evaluate if the serum level of 8-OHdG could serve as a biomarker for CAD. In neither of the examined groups, 8-OHdG had the ability to correctly classify subjects with T2DM as having CAD or not. Specifically, in each of the two groups, 8-OHdG yielded AUC values of 0.483 (95% CI 0.35 – 0.61) and 0.655 (95% CI 0.37 – 0.79), respectively, with p-values far above the significance cutoff value (p=0.05, ROC curve analysis).
Moreover, we found that oxidative stress in 34 C allele carriers (CC + CT genotypes) was not significantly different from the 81 TT carriers [1.49 (range: 1.36 – 1.71) vs. 1.51 (1.26 – 1.71) ng/mL], respectively (p=0.70, Mann-Whitney U-test).
The genotype and allele frequencies of the ROMO1 rs6060566 polymorphism are shown in Table 2. Genotype distributions for both cases (subjects with MI) and controls (subjects without CAD) were in Hardy-Weinberg equilibrium (cases: p=0.70; controls: p=0.83, Pearson χ2 test; respectively). Moreover, in each of the studied groups, genotype (cases: p=0.40 and controls: p=0.21, Pearson χ2 test) and allele (cases: p=0.24 and controls: p=0.1, Pearson χ2 test) frequencies were not significantly different from those reported for the datasets in the 1000 Genomes Project Phase 3 European population.
Moreover, binary logistic regression analyses for different genetic models found no significant associations between different genotypes or alleles of the rs6060566 polymorphism and the risk of MI in Slovenian subjects with T2DM. Estimates of ORs were adjusted (AORs) (Table 3) for the variables (BMI, waist circumference, diastolic blood pressure, total cholesterol, HDL- and LDL-cholesterols, triglycerides, duration of DM in years) that were significant in the univariate analyses (Table 1).
Subjects who underwent coronary CTA did not show any significant differences in genotype and allele frequency distribution for the rs6060566 polymorphism (Table 4).
We performed multinomial logistic regression analyses to evaluate the association of the rs6060566 polymorphism with CAD in these subjects. Because of the low frequency of the minor C allele (Table 4) the analyses were performed assuming the dominant genetic model ([CC + CT] vs. TT). The final model is shown in Table 5. The dependent variables describing the severity of CAD were the number of diseased vessels and extent of stenosis (no diseased vessel and stenosis <50% were used as references, respectively). Independent variables included in the model were dominant genetic model (TT genotype was used as reference), age, gender, lipid parameters and duration of T2DM in years. We did not observe any interactions between the dominant genetic model and CAD without adjustment for the possible confounders (Table 5). Nevertheless, when well-known CAD risk factors (age, gender, lipid parameters and duration of T2DM in years) were fixed in the model, the association between carriers of the [CC + CT] genotypes and ≥50%≤75% cross-sectional area stenosis became statistically significant (p=0.025, multinomial logistic regression). The carriers of the C allele of the ROMO1 rs6060566 had a threefold increased likelihood of having coronary artery stenosis (AOR= 3.27, 95% CI 1.16 – 9.20, Table 5). However, in a full fitted model, only dominant genetic model with CC + CT genotypes (Wald=5.05, df=1, p=0.025, multinomial logistic regression) and triglycerides (Wald=3.77, df=1, p=0.05, multinomial logistic regression) were independently associated with CAD in subjects who were diagnosed with ≥50% - ≤75% stenosis. The model demonstrated the independent strength of the effect of the C allele; thus it appears that rs6060566 of the ROMO1 gene may contribute to CAD risk. Furthermore, the fit of the multivariate model was tested by ROC curve analysis. The predictive power of the final model, which included a dominant genetic model and additional risk factors – age, gender, lipid parameters and duration of T2DM – with the AUC of 0.729 (95% CI 0.61 – 0.85), was statistically significant (p<0.001, ROC curve analysis). In addition, likelihood ratio test as a measure of the statistical significance of all predictors in the model showed a statistically significant score with the χ2 value of 27.62 (df=16, p=0.035, likelihood ratio test).
A total of 128 subjects with T2DM underwent coronary CTA (Figure 1A to 1C). A single VD (1 VD) was observed in 22 (17%) subjects, two VD (2 VD) in 41 (32%) and three VD (3 VD) in 16 (13%) subjects. Moreover, in 49 (38%) subjects all major epicardial coronary arteries (LMCA, LAD, LCx and RCA) on CT angiograms (Figure 1A) were normal. Moreover, 97 (76%) subjects had non-obstructive CAD (cross-sectional area stenosis of <50%), in 28 (22%) subjects a cross-sectional area stenosis of ≥50% - ≤75% was detected while only 3 subjects (2%) had stenosis of >75% (Figure 1A). As shown in Figure 1A, subjects with 2 VD (14/41, 35%) and non-obstructive CAD (11/95, 11.3%) suffered from nonfatal MI more often than other subjects in both comparative groups (number of diseased vessels and percentage of cross-sectional area stenosis). Of note, there was a statistically significant difference (p=0.0096, Pearson χ2 test) in the frequency distribution between subgroups with and without MI with regard to the extent of the CAD (Figure 1A). In contrast, no difference (p=0.283; Fisher’s Exact test) was observed between subgroups with regard to the coronary cross-sectional area stenosis (Figure 1A). With regards to the number of the involved vessels, a significantly higher frequency (p=0.013; Pearson χ2 test) of MI was found in subjects with 2 VD. Interestingly, subjects with two affected epicardial coronary arteries showed a 3.72-fold risk for MI (OR= 3.72, 95% CI 1.27 – 10.84, Figure 1A).
Coronary CTA revealed that more than 50% of the 128 subjects had developed CAD in LAD (Figure 1B), while the remainder of the epicardial coronary arteries were spared of atherosclerotic disease more frequently. Atherosclerotic changes were noticed in LMCA in 39 subjects (30.6%), while a slightly higher percentage of atherosclerotic disease was seen in LCx and RCA (41%) (Figure 1B). Furthermore, the relationship between the presence or absence of CAD with regards to coronary arteries was statistically significant (p=0.055, Pearson χ2 test; Figure 1B).
As depicted in Figure 1C, subjects with CAD (CAD +) in LMCA or LAD had about 3.5-fold higher risk of experiencing MI (p=0.07 for LMCA and p=0.01 for LAD, Pearson χ2 test; respectively) compared with subjects without CAD (CAD -). However, in subjects with diseased LCx and RCA, MI occurred more frequently than in subjects with disease-free epicardial coronary arteries, although the difference was not statistically significant (p=1.0 for LCx and p=0.9 for RCA, Pearson χ2 test; respectively).
At the end of this study, the coronary artery segments, which were obtained by endarterectomy from subjects with advanced atherosclerosis, were examined with immunohistochemical staining. A statistically significantly higher numerical areal density of ROMO1 positive cells was found in 17 subjects with the C allele (Figure 2) in comparison with 23 subjects with ROMO1 TT genotype (wild type) (835 ± 215/mm2 versus 412 ± 153/mm2; p<0.001, Student̕ s t test).