The current retrospective case–control study, comprising 200 unrelated Iranian patients with CAD, consisted of 124 males and 76 females (mean age of 62.21 ± 9.48). The subjects were selected consecutively from individuals who referred to the Rajaei Heart Hospital, Tehran, Iran between January 2018 until August 2019. As the control group, 100 healthy and unrelated Iranian individuals, consisting of 55 males and 45 females (mean age of 57.74 ± 10.5), who attended to Masoud Medical Laboratory, Tehran, Iran for a routine checkup, were included in the study. The healthy control subjects did not have autoimmune and inflammatory diseases, cancers, metabolite disorders, and immunodeficiencies, neither in themselves nor their immediate family members.

The diagnosis of CAD was accomplished in accordance with the visual evaluation of the coronary angiogram by cardiologist. A threshold of luminal narrowing greater than 50% in diameter seen in minimum one of the main coronary arteries or related main branches was regarded as validation for diagnosis of CAD. The clinical history of patients, such as risk factors related to the cardiovascular development, was obtained from CAD patients. The scores for blood pressure, glucose, and lipid profile were obtained from the medical records at the time of diagnosis, when the subjects were detected for CAD but were not under medication. Furthermore, there was a number of subjects that were newly diagnosed, which were also not under drug regimen for controlling blood pressure, glucose, and lipid profile. The patients were assessed for the presence of diabetes mellitus based on a raised fasting blood glucose (FBS) > 130 mg/dl. Individuals with a body mass index (BMI), which was determined through body weight divided by height squared (kg/m2), of higher than 30 was considered as obese patients. To determine the hypertension in the study participants, the seated blood pressure (BP) was measured (> 140/90 mmHg on two different occasions). Hypercholesterolemia was defined as a total cholesterol level above 200 mg/dl. Baseline features of the patients and healthy control individuals are described in detail in Table 1.

The local ethical committee of Iran University of Medical Sciences, Tehran, Iran approved the protocol of the study. Before sampling, all study participates signed the informed consent forms. Upon an overnight fasting, peripheral blood samples were obtained from all patients and control subjects in EDTA-treated vacuum tubes.

The DNA content was extracted from blood samples by the salting-out approach [14]. After dissolving the DNAs with RNase-free water, samples were stored at 4 °C for pending assays. In order to genotype eight HPA polymorphisms [including HPA-1 T196C (rs5918), HPA-2 C524T (rs6065), HPA-3 T2622G (rs5911), HPA-4 G526A (rs5917), HPA-5 G1648A (rs1062535), HPA-6 A1564G (rs13306487), HPA-9 A2603G (rs137852907), and HPA15 A2108G (rs10455097)], the polymerase chain reaction with sequence-specific primers (PCR-SSP) approach was employed, as already described [15]. The detection of two different alleles through PCR-SSP technique depends on the Taq polymerase inability to repair a single base mismatch located in the 3′-end of a primer. The amplification of the target region is happened if a complementary matching of the 3′-end of primer occurred to the sequence at the allelic polymorphism region. Otherwise, mismatching of the primer and sequence of the allelic variation leads to no amplification. Wild-type and variant HPA alleles for each locus were assigned as “a” and “b”, respectively. The thermocycling program of the PCR were: 1 cycle of 95 °C for 2 min, then followed by 10 cycles of 95 °C for 10 s and 65 °C for 1 min, then 20 cycles of 95 °C for 10 s, 61 °C for 50 s and 72 °C for 30 s, finally holding at 4 °C. The thermocycling conditions of the PCR were in accordance of the previous recommendation, with slight alterations on the annealing temperature of the primers (Table 2). In order to detect each allele of the HPA polymorphisms, two sets of primers, one for allele-specific and another for the common allele, were employed (Table 2) [16]. After performing electrophoresis of the amplified samples on the agarose gel (2% w/v) stained with DNA safe stain, discrimination of the alleles was conducted through direct observation of the products. Figure 1 illustrates the electrophoretic bands related to each allele of the determined HPA polymorphisms.

The gene-counting approach was used to determine the frequency of the alleles for each polymorphism of HPAs. The Fisher’s exact or Pearson’s chi-squared tests was implemented in order to determine the deviation of genotype frequencies in the control group from Hardy–Weinberg equilibrium (HWE). Univariate and then multivariate logistic linear regression analyses were employed to calculate the odds ratios (ORs) and corresponding 95% confidence intervals (CIs). The differences in the frequencies of the allele and genotype of HPA polymorphisms were determined using Fisher’s exact or Pearson’s chi-squared tests. The expectation maximization approach was carried out in order to estimate the HPA haplotypes using the online SHEsis software [17]. The correction of P values in more than 2 comparisons was conducted by the Bonferroni method. The relative risks were adjusted for the confounding variables, such as diabetes mellitus, hypertension, BMI, total cholesterol, and triglycerides. The difference between numerical variables was determined through Mann–Whitney U test to compare the means between CAD patients and healthy controls. A P value < 0.05 was regarded as the statistical level of significance. The data analysis was conducted using SPSS v.22 software. Presentation of data was done through mean ± standard deviation (SD) and percentage fraction of the total value for continuous and nominal variables, respectively.

Table 3 shows the frequency of the alleles and genotypes of each HPA polymorphism in both patient and control groups. No deviation from HWE was detected for the genotype distribution of the HPA polymorphisms in the control group.

The distribution of the HPA alleles and genotypes did not show significant differences between CAD patient and healthy control groups. It was observed that the HPA-15b allele frequency was the highest in patients and healthy control subjects. The homozygous HPA-15b/15b and HPA-3b/3b genotypes were detected in high prevalence in patient and healthy control groups. The frequencies of the alleles and genotypes of HPA-4b, HPA-6b, and HPA-9b polymorphisms were not seen to have any differences between patients and controls, and therefore were omitted from further analysis (Table 3).

The analysis of haplotype was based on three variants, including HPA1, HPA2, and HPA3. Among the eight identified HPA haplotypes, none of them indicated statistically significant differences between CAD patients and controls. In addition, upon performing the Bonferroni correction, it was observed that the differences were not significant for HPA haplotypes between patients and control subjects (Table 4).

The frequency of all potential haplotypes did not show statistically significant differences between CAD patients and healthy controls, according to the univariate regression analysis (Table 5). However, when the HPA-1a/2a/3a haplotype was considered as reference (OR = 1.00), there was a statistically significant association of HPA-1b/2a/3b (P = 0.015, OR = 2.792, 95% CI 1.45–8.59) with CAD risk in the multivariate regression analysis upon adjusting for the confounding factors (Table 5).