The stroke rate after coronary artery bypass grafting (CABG) compared to percutaneous coronary intervention (PCI) is generally considered high because cardiopulmonary bypass and aortic manipulations are often associated with cerebrovascular complications. However, an increasing number of CABGs performed without cardiopulmonary bypass (OPCAB) may improve those outcomes. Of 6,323 patients with multivessel and/or left main coronary artery disease, 3,877 patients underwent PCI, 1,381 conventional on-pump CABG, and 1,065 OPCAB. Median follow-up was 3.4 years. Stroke types were classified as early (onset of stroke within 24 hours after revascularization), delayed (within 30 days), and late (after 30 days). Propensity score analysis showed that the incidences of early, delayed, and late stroke did not differ between PCI and OPCAB (0.65, 95% confidence interval 0.08 to 5.45, p = 1.00; 0.36, 0.10 to 1.29, p = 0.23; 0.81, 0.52 to 1.27, p = 0.72, respectively). In contrast, incidence of early stroke after on-pump CABG was higher than after OPCAB (7.22, 1.67 to 31.3, p = 0.01), but incidences of delayed and late stroke were not different (1.66, 0.70 to 3.91, p = 0.50; 1.18, 0.83 to 1.69, p = 0.73). In conclusion, occurrence of stroke was not found to differ in patients after PCI versus OPCAB regardless of onset of stroke. Occurrence of early stroke after OPCAB was lower than that after on-pump CABG, yet occurrences of delayed and late strokes were similar for the 3 revascularization strategies.
Off-pump coronary artery bypass grafting (CABG; OPCAB) was developed to decrease complications associated with cardiopulmonary bypass and aortic manipulations. Although the long-term survival benefit of OPCAB compared to conventional on-pump CABG has been a source of controversy, several studies have reported the potential of OPCAB in decreasing stroke-related mortality and morbidity compared to on-pump CABG. In addition, an increasing number of OPCAB procedures may improve the overall outcome of CABG. Therefore, an investigation of the impact of OPCAB for risk stratification and modification of strokes after coronary revascularization has important implications. The purpose of this study was to identify occurrences and preoperative risk factors of strokes after percutaneous coronary intervention (PCI), on-pump CABG, or OPCAB and to investigate the impact of OPCAB on stroke prevention.
Methods
The Coronary REvascularization Demonstrating Outcomes Study in Kyoto (CREDO-Kyoto) is a multicenter registry in Japan that enrolls consecutive patients undergoing first PCI or CABG and excludes those patients with acute myocardial infarction within 1 week before the index procedure. This study was approved by the institutional review boards or ethics committees of all participating institutions. Because the subjects were enrolled retrospectively, written informed consent was not obtained in accord with guidelines for epidemiologic studies issued by the Ministry of Health, Labor and Welfare of Japan. However, 73 patients were excluded because of their refusal to participate in the study when contacted for follow-up.
From January 2000 through December 2002, 9,877 patients were identified as having undergone PCI (6,878 patients) or CABG (2,999 patients) without previous coronary revascularization. Of these, patients with multivessel and/or left main coronary artery disease were included in the present study. Four hundred eighty-four patients undergoing concomitant valvular, left ventricular, or major vascular surgery were excluded from the present analysis. Patients with 1-vessel disease without left main coronary artery disease (3,001 patients with PCI and 65 patients with CABG) were also excluded. Therefore, the study group consisted of 6,323 patients undergoing first coronary revascularization (3,877 patients with PCI, 2,446 patients with CABG). Demographic, angiographic, and procedural data were collected from hospital charts or databases in each center by independent clinical research coordinators according to prespecified definitions. Baseline clinical characteristics such as myocardial infarction, heart failure, diabetes, hypertension, current smoker status, atrial fibrillation, chronic obstructive lung disease, and malignancy were regarded as present when these diagnoses were recorded in hospital charts. Left ventricular ejection fraction was measured by contrast left ventriculography or echocardiography. Chronic kidney disease was regarded as present when creatinine clearance estimated by the Cockcroft–Gould formula was <60 ml/min. Anemia was defined as a blood hemoglobin level <12 g/dl as previously described.
Stroke was defined as any new permanent global or focal neurologic deficit that could not be attributed to other neurologic or medical processes. In most patients, strokes were diagnosed by neurologists and confirmed by computed tomographic or magnetic resonance imaging head scans. Documentation of previous stroke required verification by each patient’s primary care physician, review of medical records, and review of results of computed tomography and magnetic resonance imaging if available. Follow-up data were obtained from hospital charts or by contacting patients or referring physicians and were closed at December 2006 with a survey for all patients. An independent clinical events committee adjudicated all events. If sufficient follow-up data were unavailable, investigators contacted patients by telephone or letter. If a patient was deceased at time of contact, the investigators tried, to the furthest extent possible, to obtain data from the family regarding the patient’s death including nonfatal events such as stroke that may have occurred before time of death.
The primary end point determined type of stroke. A stroke that occurred within 24 hours after coronary revascularization was regarded as an “early” stroke, that after 24 hours and within 30 days as a “delayed” stroke, and that after 30 days as a “late” stroke. An independent clinical events committee adjudicated events. Baseline characteristics of patients in the 3 groups are presented in Table 1 . The on-pump CABG and OPCAB groups generally included more high-risk patients such as those with a history of stroke, left ventricular dysfunction, heart failure, previous myocardial infarction, chronic kidney disease, and anemia. Patient with diabetes were more commonly found in the on-pump CABG and OPCAB groups. Regarding the complexity of coronary artery anatomy, the on-pump CABG and OPCAB groups included more complex patients such as those with 3-vessel disease, left main coronary artery disease, involvement of the proximal left anterior descending coronary artery, and total occlusion. All continuous variables were expressed as mean ± SD. Differences in baseline characteristics across the 3 groups were examined by analysis of variance of chi-square test. Logistic regression and Cox proportional hazard models were used to identify risk factors for early, delayed, and late stroke. Proportional hazard assumption was checked using a log-log plot. Odds or hazard ratios, 95% confidence intervals [CIs], and p values were reported. Propensity scores, which identified the probability that a patient would undergo PCI or on-pump CABG, were calculated for each patient. Propensity scores were estimated separately with a multivariable polytomous logistic regression model. Confounding factors in the logistic regression included age, gender, body mass index, emergency procedure, critical preoperative state (ventricular tachycardia/ventricular fibrillation or aborted sudden death, preoperative cardiac massage, preoperative ventilation before arrival in the anesthesia room, preoperative inotropes or intra-aortic balloon pumping, preoperative acute renal failure), previous myocardial infarction, congestive heart failure, stroke, peripheral arterial disease, carotid artery disease, atrial fibrillation, chronic obstructive pulmonary disease, malignancy, hypertension, diabetes, hemodialysis, chronic kidney disease, anemia, current smoker status, left ventricular ejection fraction, total occlusion, proximal left anterior descending coronary artery disease, 3-vessel disease, and left main coronary artery disease. The c-statistics of the propensity score for PCI was 0.941 (95% CI 0.936 to 0.946) and that for on-pump CABG was 0.896 (95% CI 0.888 to 0.903).
| Variables | PCI | On-Pump CABG | OPCAB | p Value ⁎ |
|---|---|---|---|---|
| (n = 3,877) | (n = 1,381) | (n = 1,065) | ||
| Age (years) | 68.3 ± 10.0 | 66.3 ± 9.3 | 68.6 ± 9.3 | <0.01 |
| Age >75 years | 1,081 (27.9%) | 260 (18.8%) | 306 (28.7%) | <0.01 |
| Male gender | 2,704 (69.7%) | 1,007 (72.9%) | 751 (70.5%) | 0.08 |
| Body mass index (kg/m 2 ) | 23.7 ± 3.3 | 23.5 ± 3.2 | 23.7 ± 3.2 | 0.03 |
| Number of narrowed coronary arteries | 2.35 ± 0.53 | 2.58 ± 0.73 | 2.55 ± 0.74 | <0.01 |
| 3 | 1,461 (37.7%) | 958 (69.4%) | 705 (66.2%) | <0.01 |
| Left main coronary artery | 165 (4.3%) | 410 (29.7%) | 330 (31.0%) | <0.01 |
| Chronic total occlusion | 1,301 (33.6%) | 672 (48.7%) | 455 (42.7%) | <0.01 |
| Proximal left anterior descending coronary artery disease | 2,831 (73.0%) | 1,179 (85.4%) | 909 (85.4%) | <0.01 |
| Old myocardial infarction | 1,006 (25.9%) | 489 (35.4%) | 338 (31.7%) | <0.01 |
| Heart failure | 569 (14.7%) | 316 (22.9%) | 302 (28.4%) | <0.01 |
| Ejection fraction (%) | 62.1 ± 13.6 | 58.6 ± 15.0 | 61.3 ± 13.7 | <0.01 |
| Ejection fraction <40% | 279 (7.2%) | 174 (12.6%) | 86 (8.1%) | <0.01 |
| Peripheral artery disease | 367 (9.5%) | 239 (17.3%) | 248 (23.3%) | <0.01 |
| Carotid artery disease | 105 (2.7%) | 114 (8.3%) | 158 (14.8%) | <0.01 |
| Stroke | 607 (15.7%) | 237 (17.2%) | 286 (26.9%) | <0.01 |
| Atrial fibrillation | 254 (6.6%) | 80 (5.8%) | 60 (5.6%) | 0.41 |
| Chronic pulmonary disease | 83 (2.1%) | 30 (2.2%) | 22 (2.1%) | 0.98 |
| Malignancy | 321 (8.3%) | 80 (5.8%) | 79 (7.4%) | 0.01 |
| Hypertension | 2,810 (72.5%) | 918 (66.5%) | 801 (75.2%) | <0.01 |
| Diabetes mellitus | 1,651 (42.6%) | 642 (46.5%) | 498 (46.8%) | <0.01 |
| Hyperlipidemia | 1,955 (50.4%) | 710 (51.4%) | 608 (57.1%) | <0.01 |
| Chronic kidney disease | 1,411 (36.4%) | 532 (38.5%) | 424 (39.8%) | 0.08 |
| Dialysis | 167 (4.3%) | 69 (5.0%) | 53 (5.0%) | 0.45 |
| Hemoglobin (mg/dl) | 13.1 ± 2.0 | 12.7 ± 2.0 | 12.6 ± 2.0 | <0.01 |
| Hemoglobin <12 mg/dl | 972 (25.1%) | 452 (32.7%) | 350 (32.9%) | <0.01 |
| Emergency | 191 (4.9%) | 77 (5.6%) | 75 (7.0%) | 0.03 |
| Critical preoperative state † | 40 (1.0%) | 25 (1.8%) | 24 (2.3%) | <0.01 |
⁎ Comparison among percutaneous coronary intervention, on-pump off-pump coronary artery bypass grafting, and off-pump coronary artery bypass grafting by analysis of variance and chi-square test.
† Ventricular tachycardia/ventricular fibrillation or aborted sudden death, preoperative cardiac massage, preoperative ventilation before anesthesia, preoperative inotropes or intra-aortic balloon pumping, preoperative acute renal failure.
Outcomes after PCI, on-pump CABG, or OPCAB were compared by logistic regression or Cox proportional hazard models stratified by quartiles of propensity scores. This analysis was performed according to the intent-to-treat principle. There were few patients with the opposite treatment in the first quartile of the propensity score for PCI and the fourth quartile of the propensity score for on-pump CABG, suggesting a systematic treatment selection; thus, we excluded these subsets. Propensity score–adjusted odds or hazard ratios, 95% CIs, and p values were reported. The p values for multiple comparisons, namely PCI versus OPCAB and on-pump CABG versus OPCAB, were adjusted by Bonferroni correction. Further, we conducted propensity score–adjusted logistic regression and Cox regression analyses with random effects for center-to-center differences as sensitivity analyses.
All reported p values were 2-sided. All analyses were conducted by a statistician using SAS 9.2 (SAS Institute, Cary, North Carolina) and S-Plus 7.0 (Insightful Corp., Seattle, Washington). The authors had full access to the data and take responsibility for their integrity. All authors have read and agreed to the report as written.
Results
Clinical follow-ups were completed for 98% of patients at 1 year and for 95% of patients at 2 years. Median follow-up period was 1,287 days. Thirty-day mortalities were 0.85%, 2.2%, and 0.83% in the PCI, on-pump CABG, and OPCAB groups and overall mortalities were 11.7%, 11.2%, and 11.7%, respectively.
Occurrences of stroke after each revascularization procedure for each period are presented in Table 2 . Kaplan–Meier curves for stroke within 30 days are shown in Figure 1 . Unadjusted occurrences of stroke varied among the 3 groups (p <0.01). Kaplan–Meier curves for overall stroke are shown in Figure 2 . Occurrences of stroke at 30 days, 1 year, and four years were 0.23%, 1.6%, and 5.4% for PCI, 2.5%, 4.2%, and 8.5% for on-pump CABG, and 0.94%, 2.5%, and 6.5% for OPCAB, respectively. Unadjusted occurrences of stroke varied among the 3 groups (p <0.01).
| Variables | Patients | Early (≤24 hours) | Delayed (>24 hours to ≤30 days) | Late (>30 days) | Total |
|---|---|---|---|---|---|
| Percutaneous coronary intervention | 3,877 | 3 (0.08%) | 6 (0.15%) | 182 (4.69%) | 191 (4.93%) |
| On-pump coronary artery bypass grafting | 1,381 | 18 (1.30%) | 17 (1.23%) | 72 (5.21%) | 107 (7.75%) |
| Off-pump coronary artery bypass grafting | 1,065 | 2 (0.19%) | 8 (0.75%) | 54 (5.07%) | 64 (6.01%) |
| Total | 6,323 | 23 (0.36%) | 31 (0.49%) | 308 (4.87%) | 362 (5.73%) |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
