Patients

Genetic analyses

Polymorphisms in genes coding for factors involved in lipid metabolism: apolipoprotein E (APOE) [4, 8], hepatic lipase (LIPC) [7], and cholesteryl ester transfer protein (CETP) [9], the NO-donor system: endothelial NO synthase (NOS3) [6], and the coagulation system: plasminogen activator inhibitor 1 (SERPINE-1) [10], coagulation factor V (F5) [11], and coagulation factor II (prothrombin (F2)) [12] previously described to play a role in the development of CAD were chosen. The genes investigated, their polymorphisms, function as well as the primer sequences are shown in Additional File 1.

Restriction Fragment Length Polymorphism (RFLP) Analysis

To investigate the selected seven polymorphisms, restriction fragment length polymorphism (RFLP) analysis was used. The patients genomic DNA was isolated from a 200-μl aliquot by means of the QIAamp blood mini kit (Qiagen, Hilden, Germany). The PCR reactions contained 5 μl PCR buffer (ABgene, Hamburg, Germany), 5 μl dNTPMix (Fermentas Life Sciences, St. Leon-Rot, Germany) as well as 1 μl forward and 1 μl backward primer (400 nmol; Thermo Electron Corporation, Ulm, Germany) in a volume of 50 μl containing about 150–200 ng of genomic DNA. Depending on the specific polymorphism, the PCR was comprised of an initial denaturation step (15 minutes at 95°C), 35 to 40 cycles of 95°C for 20 to 30 seconds, 58 to 72°C for 20 to 30 seconds, and of 72°C for 30 sec).

The restriction digests were performed in a 10-μl volume containing the restriction enzyme (Fermentas Life Sciences, St. Leon-Rot, and New England BioLabs, Frankfurt am Main, Germany), 1× restriction buffer, and the PCR product. The digests were incubated overnight at 37°C and analysed by electrophoresis in a 1,5–3% agarose gel which was subsequently stained with ethidium bromide. Details of the polymorphisms analyzed are summarized in Additional File 1.

Risk profile

Because SNPs had a relatively low prevalence in our limited patient cohort and CAD is a multifactorial disease, we constructed a risk profile. Carriers of the risk profile were defined as having one of the following polymorphisms: hetero- or homozygous for the variants of eNOS 4ab, the ApoE allele combination 2/4 or 4/4, homozygous expression of LIPC variant, homozygous expression of the CETP variant, and hetero- or homozygosity for the prothrombin G20210A variant. Patients had to be homozygous for PAI-1 5G insertion.

Statistics

For statistical data analysis Microsoft^® EXCEL 2002 and SPSS for Windows (Version 12.0, SPSS Inc., Chicago, IL, USA) were used. The primers for the PCR analyses were designed with the help of the computer program "PrimerExpress" (Applied Biosystems).

Values are expressed as mean and standard deviation of mean. The Student's t-Test was used to compare absolute quantitative values. Angina was censored at the day of follow-up, because a precise onset of angina could not be determined. Freedom from reoperation, reintervention, and the combined endpoints (reoperation, reintervention, or angina at follow-up) were calculated by the actuarial method and tested with the log-rank test. Every univariate parameter reaching or approaching significance (p < 0.2) was then tested in a Cox multivariate model using the conditional backward method. P < 0.05 was considered significant.

Results

Risk profile and occurrence of cardiac adverse events (Figure 1)

Patients with the risk profile had significantly more reoperations (0.012) and were more likely to have recurrent symptoms (0.0012). The incidence of percutaneous interventions were not different among groups (p = 0.38). However, multivariate Cox regression analysis revealed that only the risk profile had significant impact on the progression of CAD (p = 0.004).

ApoE Polymorphism

Several studies focused on the effects of ApoE after cardiopulmonary bypass and reported a higher incidence of inflammation [8] and nephropathy [13] in patients with the E4 allele; depending on the trial more or less neurological sequelae were seen with the E2 allele [14, 15]. In Kuukasjärvi et al. [16] investigated the ApoE E4 allele and found, in contrast to most of the studies, this polymorphism not to be a predictor for reoperation. Interestingly, in our study patients without the most frequent allele E3, thus only having the risk alleles E2 and E4, had a more aggressive type of CAD. In regard to the total population, there is no other gene product, which has such a big influence on the individual cholesterin level such as ApoE. Especially the ε4 allele is increasing the plasma level of LDL. In various investigations ApoE was associated with a higher incidence of atherosclerosis. Baroni et al conducted a study with 6 polymorphisms influencing the lipid system (ApoE, ApoAI, ApoCIII, ApoB, lipoprotein lipase LPL und LIPC [4]. One hundred and two patients with diagnosed CAD were enrolled and 104 healthy patients served as controls. Significant difference between CAD and healthy patients could be demonstrated for Apolipoprotein E, ApoB and HL. ApoE ε4 allele was associated with a manifest CAD. Results of our study were similar: patients with deletion of the most common allele ε3 that means they carry the rare combination of either ε2/4 or ε4/4 (isoform ε2/2 is extremely rare) had more frequently to undergo a reintervention. After incidence is so low (2–3% of general population), evaluation is limited. Therefore, a higher number of patients would be desirable.

eNOS 4ab

The investigated eNOS variant has been associated with a higher susceptibility to coronary lesions in smokers and NO metabolites were 20% decreased in patients with the 4aa variant [5]. For endothelial NOS at least 4 frequent polymorphisms (G894T, Glu298Asp, T786C, and the one used in this study) are well established and described to be risk factors for CAD. The argument of having no specific hypothesis to investigate just one and not all other polymorphisms is valid. Further more, some authors see differences in subgroups, i.e. smokers vs. non smokers and young adults. These oppositional results are frequently seen in genetic studies and may be due to the multi factorial origin of the disease. For instance, in japanese and caucasians similar patterns of eNOS alleles were observed [17], but in afroamericans the incidence is generally much lower. Especially environmental factors are difficult to record such as smoking, which is particularly important in eNOS. In our study, the a allele of eNOS 4ab polymorphism was associated with a significantly higher risk for reintervention and recurrent symptoms.

LIPC C202G

Enzyme activity of hepatic lipase plays a major role in regulating the lipid metabolism. In the above mentioned study from Baroni et al. in carriers of the g allele of C202G mutation, a decreased level of HDL cholesterol and increased level of triglycerides was observed [4]. Interestingly, the authors found another independent factor regarding clinical endpoints: carrier of homo- or heterozygous g allele had significantly more frequently a second adverse event. Also our results demonstrate a similar course: patients with homozygous expression of the g allele (n = 48) needed more frequently a reoperation. For this mutation, Murtomaki et al, demonstrated a binding imbalance towards additional LIPC polymorphisms such as L334F, T457T and C480T). Last is related to a low LIPC activity in CAD patients [18]. That means that C202G mutation is a simple marker for additional LIPC polymorphisms and their mutations.

Recently Taylor and associates [19] reported on the influence of lipoprotein lipase locus on the progression of atherosclerosis in coronary artery bypass grafts and identified the LPL-HINDIII 2/2 genotype as an independent risk factor.

Risk Profile

After single mutations had a relatively low prevalence in our limited patient cohort and CAD is a multifactorial disease we constructed a gene risk profile according to the definitions. It was a combination of the alleles with the highest incidence of either one of clinical endpoints: Included was hetero- or homozygous a-allele of eNOS 4ab polymorphism, because patients showed a significantly higher risk to get a rentervention or recurrent symptoms. Regarding the ApoE they had to carry ε4 allele, that means ε2/4 or ε4/4, as it is considered to be potentielly atherogenous, was we could also demonstrate. Homozygous carriers of hepatic lipase were prone to undergo a reoperation and recurrent symptoms. That's why part of the risk gene profile was homozygous expression of the G allele. Homozygous expression of CETP gene polymorphism was included. Additionally, hetero- or homozygous variants of G20210A mutation was chosen, because only 50 and 0% carriers of these variants were free from recurrent symptoms after 10 and 15 years, respectively. However, a simple gene risk profile constructed out of 7 randomly chosen polymorphisms was more predictive for the advancement of CAD than any cluster of classical risk factors.

Certainly medical therapy, particularly lowering of low-density lipoprotein cholesterol levels, has been proven to reduce the advancement of CAD after CABG [20]. Due to the design of our retrospective study, we were unable to prove the beneficial effects of medical therapy, because therapy after primary CABG was determined by the cardiologist and patients with recurrent symptoms received intensified medical therapy. In the investigated patient's cohort 87% received Aspirine, 70% were given beta blockers, 82% of patients received statins, 35% ACE inhibitors and 24% took calcium antagonists postoperatively. Likewise, the role of classical risk factors on the CAD progression is difficult to interpret. Risk factors at the time of primary CABG were medically treated as described above; patients stopped smoking and started training programs. Therefore, these classical risk factors lost partially their predictive value. Obviously, we investigated only long-term survivors of CABG surgery, who were willing to cooperate. This selection bias may also limit our results.

Knowledge of gene polymorphisms in the era of genomics and their influence on outcome in cardiac surgery is rapidly growing. However, most studies investigated the acute effects of polymorphisms outcome during the postoperative phase. Data on the progression of CAD after primary CABG is rare.

This study is only preliminary, because of its limitations in patient sample size as well as number and choice of investigated polymorphisms. More investigations are warranted and will most likely improve the predictive value of polymorphism tests. We proved the concept, that risk stratification by a simple gene test for the future advancement of CAD after primary CABG is possible. The concept is intriguing, because the detected gene variants give clues to the individual pathophysiology in every single patient in this multifactorial disease [18]. Therefore, this cheap diagnostic tool may hopefully lead to an individualized secondary prevention after primary CABG.