Elevated Levels of Homocysteine Predict Cardiovascular Death, Nonfatal Myocardial Infarction, and Symptomatic Bypass Graft Disease at 2-Year Follow-Up Following Coronary Artery Bypass Surgery


Nicolas W. Shammas, MD, MS, Genesis Heart Institute, Cardiovascular Medicine, PC, 1236 East Rusholme, Suite 300, Davenport, IA 52803
E-mail: shammas@mchsi.com


Severe graft disease occurs in patients at a rate of approximately 15% within the first year of coronary artery bypass grafting (CABG). In this study, the authors examined predictors of the combined end point of death, nonfatal myocardial infarction (MI), and bypass graft disease at 2-year follow-up after CABG. One hundred twenty-one consecutive patients were included in this study after informed consent was obtained. In univariate analysis, there was a significantly (P<.05) higher homocysteine level (11.0 ng/mol vs 9.7 ng/mol, P=.04) in patients who met the combined end point vs those who did not. There were no statistically significant differences in the following: low-density lipoprotein cholesterol, high-sensitivity C-reactive protein, and lipoprotein(a) values; age; body mass index; smoking and diabetes status; statin or aspirin use; creatinine level; hematologic markers; left ventricular ejection fraction; number of bypass grafts; and distribution of coronary artery disease. Logistic regression analysis modeling for low-density lipoprotein cholesterol, lipoprotein(a), fibrinogen, and homocysteine showed that homocysteine value (P=.016) was an independent predictor of the primary combined end point.

Earlier studies have indicated that 10% to 15% and 16% to 31% of all bypass grafts become severely diseased at 6-month and 1-year follow-up, respectively, following coronary artery bypass grafting (CABG).1–3 It is recognized that saphenous vein graft (SVG) closure within 1 month of surgery is due to acute thrombosis, whereas SVG disease up to 1 year is due to progressive intimal hyperplasia with a trend toward thrombotic events. Following the first year of surgery, atherosclerosis becomes the main mechanism for developing occlusive disease in SVG.4,5

Several studies indicate that graft handling, surgical technique, hematologic abnormalities, the use of antiplatelet drugs, and poor bypass targets can influence SVG patency.1,3,6–8 Recently, prebypass use of 3-hydroxy-3-methylglutaryl coenzyme A (HMGCoA) reductase inhibitors (ie, statins) has been shown to reduce cardiovascular events after bypass surgery.9 Studies also suggest that elevated levels of lipoprotein(a) (Lp(a))10–13 and homocysteine14–17 are risk factors for the development of atherosclerosis and/or vascular thrombosis. Lp(a), however, is not a consistent predictor of increased cardiovascular events18 or bypass graft disease.19 Homocysteinemia has been linked to an increase in cardiovascular events14–17 and SVG disease after CABG,20–22 but the association between homocysteine and SVG disease also has not been consistent in all studies.18 In this study, we prospectively evaluated clinical parameters and biomarkers as predictors of the combined end point of death, nonfatal myocardial infarction (MI), and angiographically documented severe symptomatic SVG disease at 2-year follow-up after CABG.


One hundred twenty-one patients undergoing elective bypass surgery for coronary artery disease were included in this study after informed consent was obtained. Patients were excluded if they had recent MI, established malignancy, emergent cardiac surgery, or cardiac intervention within 72 hours of bypass surgery.

The primary end point of the study was the predictor of occurrence of the combined end point of cardiovascular death, nonfatal MI, or angiographically documented significant vein graft disease (>50%) at 2-year follow-up (performed for symptom recurrence or abnormal stress test). Patients underwent a clinical follow-up at 1 month, 6 months, 1 year, and 2 years. Follow-up evaluations were conducted by routine office visits or by phone contact.

All patients had clinical, angiographic, and laboratory variables (obtained within 1 week before surgery). The laboratory variables included the following biomarkers: high-sensitivity C-reactive protein (hs-CRP), low-density lipoprotein cholesterol (LDL-C), Lp(a), total cholesterol, fibrinogen, and homocysteine. Clinical variables included age, sex, left ventricular ejection fraction, anginal class, history of smoking, diabetes mellitus, hypertension, hypercholesterolemia (treated or not treated with statins), previous MI, weight, height, body mass index, history of cerebrovascular accident or transient ischemic attack, congestive heart failure, and history of peripheral vascular disease. Angiographic variables included the number of significantly diseased coronary vessels, number of bypass grafts placed, number of bypass grafts with >50% obstructive disease, number of vessels bypassed, use of the left internal mammary artery, and on-pump vs off-pump surgery. Patients were divided into 2 groups: (1) asymptomatic after 2 years of CABG, requiring no further angiographic evaluation or no diseased bypass grafts on angiographic follow-up (group A) and (2) patients with cardiovascular death and nonfatal MI and those undergoing angiographic follow-up for symptoms or abnormal stress testing who had documented graft disease >50% in severity (group B).

Statistical Analysis

Bivariate and multivariate analyses of the data were performed. Baseline characteristics were analyzed with Fisher's exact test for dichotomous variables and t test for continuous variables. A logistic regression model was performed for predictors of the primary combined end point.


Baseline clinical characteristics are displayed in Table I. Univariate comparisons between groups A and B are shown in Table II. At 2 years, follow-up data were obtained on 90 patients (75%). Of the remaining 31 patients, 4 died of unknown causes, 6 died of noncardiac causes, 19 were lost to followup, and 2 refused to participate in the follow-up questionnaire. Twenty-nine patients (32%) met the primary composite end point (Table III). In univariate analysis, there was a significantly (P<.05) higher serum homocysteine level (11.0 ng/mol vs 9.7 ng/mol; P=.04) in patients who met the combined end point vs those who did not, respectively. The Figure also demonstrates that a higher percentage of patients who did not meet the primary combined end point had low and intermediate homocysteine levels compared with those who did. There were no statistically significant differences in LDL-C, hs-CRP, Lp(a), age, body mass index, smoking, diabetes mellitus, statin or aspirin usage, serum creatinine value, hematologic markers, left ventricular ejection fraction, number of bypass grafts, or distribution of coronary artery disease. In the entire cohort, patients receiving statins (n=74) had a mean LDL-C level of 86.7 mg/dL vs 108.5 mg/dL in those who were not receiving statins (n=47). Logistic regression analysis modeling for LDL-C, Lp(a), fibrinogen, and homocysteine showed that homocysteine (P=.016) was an independent predictor of the primary combined end point.

Table I.  Demographics and Clinical Characteristics of Cohort Enrolled
Age, y65.9±9.6
Body mass index29.6±5.6
Male, %78.5
Smoking status, %
 Current smoker15
Diabetes, %
Angina, %77.7
Hypertension, %81
Hyperlipidemia, %72.7
Statin use, %61.2
 >3-Mo presurgery, %86.5
Aspirin, %88.4
ACE inhibitor, %47.9
Previous MI, %42.1
Stroke/TIA, %5.8
PAD, %28.3
EF, %51.3±11.4
On-pump, %100
Abbreviations: ACE, angiotensin-converting enzyme; AODM, adult-onset diabetes mellitus; EF, ejection fraction; MI, myocardial infarction; PAD, peripheral arterial disease; TIA, transient ischemic attack.
Table II.  Univariate Analysis Between Patients Who Met the Combined Primary End Point vs Those Who Did Not
 No (n=61)Yes (n=29)P Value
Age, y64.8±10.266.7±8.5NS
Body mass index30.6±629.4±5.2NS
Male, %78.779.3NS
Smoking status, %NS
 Current smoker11.713.8 
Diabetes, %37.734.5NS
Angina, %
Hypertension, %8279.3NS
Hyperlipidemia, %75.469NS
Statin use, %5969NS
 >3-mo presurgery, %86.190NS
Aspirin, %9086.2NS
ACE inhibitor, %4155.2NS
Previous MI, %47.531NS
Stroke/TIA, %3.310.3NS
Ejection fraction, %49.5±10.353.9±12.1NS
No. of grafts3.8±1.44.0±1.1NS
Homocysteine, ng/mol9.7±2.511.0±2.9.04
Low-density lipoprotein cholesterol, mg/dL93.2±32.695.5±30.7NS
High-density lipoprotein cholesterol, mg/dL42.5±10.941.3±9.6NS
Total cholesterol, mg/dL171.9±36172±40.4NS
Lipoprotein(a), mg/dL46±53.732.8±35.3NS
Triglycerides, mg/dL181.9±99.1176.7±106NS
While blood cells7.7±2.78.4±2.7NS
Creatinine, mg/dL1.0±0.31.0±0.3NS
Postoperative creatinine kinase, U/L983.2±1267.3777.2±530.9NS
Creatinine kinase-MB fraction, ng/mL37.8±32.436.9±29.1NS
Abbreviations: ACE, angiotensin-converting enzyme; hs-CRP, high-sensitivity C-reactive protein; MI, myocardial infarction; NS, not significant; TIA, transient ischemic attack.
Table III.  Events Composing Primary Outcome
Cardiac death, %2
Nonfatal MI, %1.1
Bypass graft diseased, %25
Combined primary end point, %26.7
Abbreviation: MI, myocardial infarction.
Figure Figure.

The percentage of patients with low (3.5–7.9 ng/mol), intermediate (8.0–11.8 ng/mol), and high (>12 ng/mol) homocysteine levels in patients who did and did not meet the primary end point.


In this study, we have shown that baseline serum homocysteine level is an independent predictor of the combined end point of cardiovascular death, nonfatal MI, and vein graft disease 2 years after bypass surgery. The mechanism by which homocysteine promotes SVG disease is unknown but might be the result of inflammation and/or thrombosis or possibly accelerated atherogenesis in SVG. Homocysteine is known to reduce the availability of nitric oxide, induce endothelial cell dysfunction, and compromise antioxidative defenses within arterial walls.23 Our data are consistent with published studies indicating that homocysteine can accelerate bypass graft disease.13,20–22 In addition, recent studies have shown a significant correlation between homocysteine14–17 and the development of cardiovascular events.

Elevated LDL-C has been shown to correlate with increased rates of cardiovascular events and with SVG disease. Recent data suggest that patients undergoing bypass surgery and treated with statins have significantly fewer cardiovascular events than those who are untreated. An anti-inflammatory role of statins has been suggested as a possible mechanism for vascular protection independent of LDL-C lowering.9 In our study, the mean LDL-C level was low in the entire cohort studied and particularly in those treated with statins. This likely blunted the effect of LDL-C as a predictor of the primary end point of this study. In addition, hs-CRP was not a predictor, which was possibly related to the smaller sample size or the fact that these patients have already been aggressively treated with statins, aspirin, and lifestyle modification. Data suggest that statins blunt the impact of hs-CRP elevation on cardiovascular events.24

We conclude that in patients with well-controlled LDL-C, serum homocysteine levels are an independent predictor of the combined end point of cardiovascular death, nonfatal MI, and progression of graft disease at 2 years postsurgery. Recent studies have not been able to demonstrate reductions in risk for cardiovascular events when patients with hyperhomocysteinemia are treated with folate with or without supplementation with vitamins B6 and B12.25–27 Other trials evaluating the impact of serum homocysteine level reduction on risk for acute cardiovascular events are in progress. Rather than providing vitamin supplements to these patients, it is advisable to ensure that they are meeting national guideline-defined targets levels for lipids, blood pressure, and glycemic indices and encouraged to engage in cigarette smoking cessation, weight loss, and dietary modification to optimize risk reduction.

Study Limitations

The study was limited by a relatively small number of patients and the number of patients that lost to follow-up. Asymptomatic patients did not undergo routine angiography at 2 years and, therefore, it is unclear how asymptomatic graft disease would have affected our findings. However, the results of this study are compelling and warrant further investigation by a larger, prospective multicenter trial. Finally, this study does not address how treatment of high homocysteine levels affects patient outcomes.