Journal of the American College of Cardiology
Effects of coronary artery bypass grafting using internal mammary arteries for diabetic patients
Author + information
- Takashi Hirotani, MDa,*,
- Tadashi Kameda, MDa,
- Takayuki Kumamoto, MDa,
- Shogo Shirota, MDa and
- Mototugu Yamano, MDa
- ↵*Reprint requests and correspondence: Dr. Takashi Hirotani, Department of Cardiovascular Surgery, Tokyo Saiseikai Central Hospital, 1-7-14 Mita, Minato-ku, Tokyo 108-0073, Japan
Abstract
OBJECTIVES
In our institute, internal mammary arteries (IMAs) have been preferred for coronary artery bypass grafting (CABG) in diabetic patients. The purpose of this study was to evaluate the influence of diabetes and IMA grafting on survival after CABG.
BACKGROUND
The influence of diabetes on the results of CABG is not well documented, and there is controversy about whether the use of IMAs conveys greater survival benefits to diabetic patients.
METHODS
A total of 420 consecutive patients who underwent CABG from April 1990 to July 1998 were reviewed; 211 of these patients had diabetes mellitus at the time of surgery. Internal mammary artery grafts have been used with increasing frequency, and bilateral IMAs have been used when possible since 1993. Internal mammary artery grafts were used in 164 nondiabetic patients (78%) and in 155 diabetic patients (73%). Seventy-eight nondiabetic patients and 74 diabetic patients received bilateral IMA grafts.
RESULTS
The postoperative mortality was 2.4% in the nondiabetic and 2.8% in the diabetic group. With regard to postoperative complications, diabetic patients had a significantly higher rate of chest wound infection (p < 0.05), irrespective of whether IMAs were used or not. The use of bilateral IMAs did not increase the risk of chest wound infection in nondiabetic or diabetic patients. Overall survival curve, cardiac death–free curve and cardiac event–free curve were not affected adversely by diabetes, and in diabetic patients, CABG with saphenous veins alone conveyed significantly (p < 0.01) less long-term benefit than did CABG with at least one IMA graft.
CONCLUSIONS
It was suggested that IMA grafts should be preferred in diabetic patients.
The influence of diabetes on outcome after coronary artery bypass grafting (CABG) is not well defined. The results of previous studies are conflicting, reporting either no effect (1,2)or adverse (3,4)effects of diabetes on the early and long-term results after CABG. Recently, the importance of internal mammary artery (IMA) use for CABG in diabetic patients has been reported (5). The IMA has been established as the preferred conduit for CABG. However, IMA use in diabetic patients may increase the risk of sternal infections. In our institute, IMAs have been preferred for CABG in diabetic patients because the patients usually have diffuse coronary artery disease involving distal vessels (6), and saphenous vein grafts to such vessels may not remain patent. Furthermore, we have used bilateral IMAs whenever possible since 1993, although many authors reported that the use of bilateral IMAs should be avoided in diabetic patients because of prevalence of chest wound infection (7). This report is a retrospective comparative study of our recent CABG surgery for patients with, or without, diabetes mellitus. Particular attention was given to operative morbidity when IMAs were used.
Methods
Between April 1990 and July 1998, 420 consecutive patients who underwent CABG had their charts reviewed. Patients with valvular heart disease requiring valve repair or replacement, postinfarction ventricular septal perforation or congenital heart disease were excluded. At the time of surgery, 211 patients (50%) had diabetes mellitus that had been diagnosed by the endoclinologists at our Tokyo Saiseikai diabetic center. Patients were defined as diabetic based upon the standard published criteria, having fasting serum glucose equal to or greater than 140 mg/dl. Of these patients, 96 (45%) were being treated with insulin and 58 (27%) with oral hypoglycemic medications.
Indications for CABG were based on standard clinical and angiographic criteria. Grafting was attempted on all vessels 1.5 mm or greater in diameter with 75% or greater obstruction. Coronary endarterectomy was generally avoided. All operations were undertaken through median sternotomy. Cardiopulmonary bypass was conducted under moderate hypothermia with antegrade crystalloid cardioplegia. Both proximal and distal anastomoses were performed during a single period of aortic occlusion. Internal mammary artery grafts have been used with increasing frequency, and bilateral IMAs have been used when possible since 1993. Right IMAs were used for the revascularization of left anterior descending arteries anterior to the heart, left circumflex arteries through the transverse sinus and right coronary arteries as pedicled or free grafts. Diabetes was not a reason for excluding the use of IMAs.
With regard to operative complications, myocardial infarctions were determined by persistent creatine phosphokinase, MB fraction enzyme elevations, new Q waves, or ST elevation on electrocardiograms. Arrhythmias were defined as complications if they were considered life-threatening or required medication. Wounds were defined as being infected if purulent material was discharged from the wound, with or without a positive culture. Minor chest wound infections included those infections limited to the subcutaneous tissues. Major chest wound infections included all cases where tissues were infected down to the sternal wire or beyond, and which required reexploration and refixation of the sternum. Leg wound infections were not stratified based on depth of infectious involvement. Strokes were diagnosed from physical findings and documented using brain computerized tomography. Respiratory failure was recorded for all patients who required mechanical ventilatory support for more than 48 h.
Follow-up data were obtained from each patient’s hospital record. The data were completely reviewed in 93% of the patients.
Limitations
This study is a nonrandomized retrospective study. Fewer patients with saphenous vein grafts alone underwent CABG during the earlier period than patients with IMA grafts. Furthermore, bilateral IMAs were used for CABG since 1993.
Data analysis
The differences between the nondiabetic and diabetic groups in preoperative characteristics were tested for statistical significance by tand chi-square tests as appropriate. Survival curves were estimated using the Kaplan–Meier method. The relationships between discrete risk factors to survival were investigated with log-rank tests.
Results
The differences between the nondiabetic and diabetic patient groups in age, gender, anatomic extent of coronary lesions, left ventricular ejection fraction, preoperative use of intra-aortic balloon pumping, frequency of urgent operations, history of old myocardial infarction, history of the previous CABG and presence of cerebrovascular complications are shown in Table 1. The preoperative baseline characteristics were similar in the nondiabetic and diabetic patient groups, and there was no statistically significant difference.
Preoperative Patient Characteristicslegend
Arterial grafts were used in 323 patients (77%). A single IMA was used in 86 nondiabetic and 81 diabetic patients. Bilateral IMAs were used in 78 nondiabetic and 74 diabetic patients. Other than IMAs, right gastroepiploic arteries were used as arterial conduits in 21 nondiabetic and 15 diabetic patients. The differences in the conduits used for coronary revascularization between the nondiabetic and diabetic patient groups are shown in Table 2. The mean number of distal anastomoses was 3.0 (range 1 to 7) in the nondiabetic and 3.2 (range 1 to 6) in the diabetic patient group. The mean aortic cross-clamp time was 110 (range 27 to 226) min in the nondiabetic and 116 (range 24 to 220) min in the diabetic patient group. These operative variables were remarkably similar.
Conduits Used for Coronary Revascularizationlegend
There were five operative deaths in the nondiabetic and six in the diabetic patients. The postoperative mortality was 2.4% in the nondiabetic and 2.8% in the diabetic patients, respectively, and the difference in mortality was not significant. The causes of operative death are shown in Table 3.
Causes of Operative Death
The operative complications and their rates are shown in Table 4for nondiabetic and diabetic patients. Diabetic patients had a significantly higher rate of chest wound infection (p < 0.05). It was then analyzed whether there were significant associations between the occurrence of chest wound infection and conduits used (Table 5). The incidence of chest wound infection was not different irrespective of whether IMAs were used or not, and the use of bilateral IMAs did not increase the risk of chest wound infection in nondiabetic or diabetic patients.
Operative Morbiditylegend
Conduits Used for Coronary Revascularization and Occurrence of Chest Wound Infectionslegend
The angiographic investigations conducted at about 3 weeks after surgery in 395 patients showed that the graft patency rate was 93.9% in nondiabetic and 96.3% in diabetic patients, respectively. The patency rate of IMA grafts was 98.7% in nondiabetic patients and 96.6% in diabetic patients. There was no difference in the graft patency rate between nondiabetic and diabetic patients. In the nondiabetic group, the IMAs had a higher patency rate than the saphenous veins (p < 0.05).
Overall survival curves and cardiac death–free curves for the nondiabetic and diabetic patients were compared for the entire population (Fig. 1). When cardiac death, acute myocardial infarction, coronary intervention and reoperation for coronary ischemia were included as cardiac events, cardiac event–free curves were also compared between these groups (Fig. 1). In each comparison, no adverse effects of diabetes were apparent. With regard to conduits used for coronary revascularization, the overall survival curves, cardiac death–free curves and cardiac event–free curves demonstrated that IMA grafts conferred no long-term benefit to nondiabetic patients (Fig. 2). However, diabetic patients who underwent CABG with at least one IMA graft received more benefit than those who had only saphenous veins. This difference was statistically significant (p < 0.01) (Fig. 3). The benefit of the bilateral use of IMAs, however, was not apparent in nondiabetic or diabetic patients.
Overall survival curves, cardiac death–free curves and cardiac event–free curves were compared by using the Kaplan–Meier method between nondiabetic and diabetic patients. DM = diabetes mellitus.
Overall survival curves, cardiac death–free curves and cardiac event–free curves in nondiabetic patients were compared according to conduits used for coronary revascularization. Internal mammary artery (IMA) grafts conferred no long-term benefit to nondiabetic patients.
Overall survival curves, cardiac death–free curves and cardiac event–free curves in diabetic patients were compared according to conduits used for coronary revascularization. In each of these curves, it was demonstrated that diabetic patients who underwent coronary artery bypass grafting with at least one internal mammary artery (IMA) graft received significantly more benefit than those with saphenous veins alone (p < 0.01).
Discussion
Operative mortality
It remains unclear whether or not diabetes mellitus is associated with increased operative mortality in patients undergoing CABG. Morris et al. analyzed a large series of diabetic patients undergoing CABG and demonstrated that diabetes was a significant, independent risk factor and that the use of IMA grafts conferred a significant survival benefit (3). In the analysis by Cosgrove et al. (7)of 8,000 patients undergoing CABG surgery, diabetes did not emerge as a predictor of operative mortality. Other studies have shown that diabetes does not increase the operative mortality, although the long-term mortality appears to be increased (8,9). In our study, there was no significant difference in operative mortality between nondiabetic and diabetic patients, and no adverse effects of diabetes were apparent from the overall survival curve, the cardiac death–free curve or the cardiac event–free curve.
Effects of using IMAs
In our institute, diabetes has not been a reason to refuse using IMAs for coronary revascularization. Bilateral IMAs have been used whenever possible, even in diabetic patients, since 1993. At least one IMA graft was used in 73% of diabetic patients, and 48% of them received bilateral IMAs grafting. The rates were similar in nondiabetic patients. In the recent Bypass Angioplasty Revascularization Investigation study, it was demonstrated that in diabetic patients, the relation of the presence of an IMA graft to cardiac mortality was particularly striking and that the survival benefit of CABG was limited to the use of IMA grafts (5). Diffuse coronary artery disease involving distal vessels is usually more common in diabetic patients, although it was not quantitated in our study. These angiographic characteristics of coronary arteries in diabetic patients may affect the relative benefit attributable to IMA grafting. The present study demonstrated the long-term benefit of IMA use in diabetic patients (Fig. 2 and 3). Therefore not only the early results but also the long-term results of our CABG surgery in diabetic patients were comparable with those in nondiabetic patients. This outcome may be attributed to our aggressive use of IMA grafts. Morris et al. described the importance of IMA use in diabetic patients, but they used IMA grafts in 67% of their diabetic patients. In our study, 73% of diabetic patients underwent CABG using at least one IMA, and about half of them received bilateral IMAs grafting. The rate of IMA use in our study was considered to be high, and the frequency of bilateral IMAs use was considered to be exceedingly high among the recent reports.
Operative morbidity
Many authors have reported increased morbidity in diabetic patients (10,11). Among the postoperative complications, chest wound infections are of great concern to surgeons, especially when IMA grafts are used, because dissection of the IMA would devascularize the sternum (12). In our study, the rate of chest wound infection was 5.7% in nondiabetic and 10.0% in diabetic patients. The rate of minor chest wound infection was 8.1% in diabetic patients and was significantly higher than 3.8% in nondiabetic patients, although the rate of major chest wound infection that required reexploration and refixation of the sternum was not significantly different in the two groups. Higher infection rates in diabetic patients after CABG have been documented previously. Fietsam et al. reported that the rate of wound infection was 7.5% in diabetic versus 0.89% in nondiabetic subjects (10). Farrington et al. found that 19% of diabetic patients had major chest wound infection, compared with 2% of nondiabetic patients (11). Furthermore, the prevalence of chest wound infection in patients receiving bilateral IMAs grafting has been of great concern to surgeons because devascularization of the sternum would be more severe than in patients receiving single IMA grafting. In our study, the rate of chest wound infection was not significantly different for single IMA, bilateral IMAs or saphenous veins alone, as shown in Table 5. These results were similar to those reported by Galbut et al. (13). The similar frequency of chest wound infection among patients who received no, single or bilateral IMAs in the present series may reflect our effective hemostasis and drainage around the area from which the IMAs were harvested. We suggest that the avoidance of hyperglycemia during the postoperative period and close observation of and infection control for the chest wound can protect against major sternal infections in diabetic patients.
Conclusions
Although the bilateral use of IMAs was no more beneficial than the single use of IMAs in the present study, the follow-up for a longer period may demonstrate better outcomes from the bilateral use of IMAs than from single use in diabetic patients.
It was obvious that the rate of chest wound infection was higher in diabetic patients, but the rate did not increase by using IMA grafts. We suggest that IMA grafts should be used in diabetic patients because of their excellent ability to remain patent for a long time.
Footnotes
☆ This study was not supported by any grants.
- Abbreviations
- CABG
- coronary artery bypass grafting
- IMA
- internal mammary artery
- Received November 2, 1998.
- Revision received March 10, 1999.
- Accepted April 21, 1999.
- American College of Cardiology
