Author + information
- Received July 29, 2002
- Revision received March 28, 2003
- Accepted April 4, 2003
- Published online April 21, 2004.
- Jeffrey W Moses, MD, FACC,
- Roxana Mehran, MD, FACC*,
- George D Dangas, MD, PhD, FACC*,* (, )
- Yoshio Kobayashi, MD, FACC*,
- Alexandra J Lansky, MD, FACC*,
- Gary S Mintz, MD, FACC*,
- Eve D Aymong, MD, MSc*,
- Martin Fahy, MSc*,
- Gregg W Stone, MD, FACC* and
- Martin B Leon, MD, FACC*
- ↵*Reprint requests and correspondence:
Dr. George D. Dangas, Cardiovascular Research Foundation, Lenox Hill Heart and Vascular Institute, 55 East 59th Street, 6th Floor, New York, New York 10022, USA.
Objectives The present study evaluated clinical outcomes in diabetic patients after multivessel stenting.
Background Multivessel angioplasty studies have reported decreased survival in diabetic patients undergoing conventional balloon angioplasty compared with coronary artery bypass graft surgery (CABG). However, several studies have demonstrated excellent procedural success and acceptable clinical outcomes after multivessel stenting.
Methods Multivessel stenting was performed in 689 patients with 1,639 native coronary lesions. Patients were classified into three groups according to diabetes mellitus (DM) status: 1) no DM (501 patients/1,200 lesions); 2) DM treated with oral agents (102 patients/235 lesions); and 3) DM treated with insulin (86 patients/204 lesions).
Results Procedural success was high overall. In-hospital CABG was higher in diabetics treated with insulin compared with the other two groups (3.5% vs. 0.4% vs. 1.0%, p = 0.02). There were no significant differences in the incidence of in-hospital cardiac death and myocardial infarction. Diabetic patients treated with oral agents or insulin had higher one-year target lesion revascularization rates than non-diabetic patients (25% vs. 35% vs. 16%, p < 0.001). Lower one-year survival was observed in diabetic patients treated with either oral agents or insulin, compared with non-diabetic patients (85% vs. 86% vs. 95%, p < 0.001). On multivariable analysis, DM was an independent predictor of one-year mortality, myocardial infarction, and target lesion revascularization after multivessel stenting.
Conclusions Despite a high technical success rate of multivessel stenting, diabetic patients, especially those treated with insulin, have higher in-hospital CABG, higher subsequent revascularization rates, and lower one-year survival than non-diabetic patients.
The Bypass Angioplasty Revascularization Investigation (BARI) showed that there was no significant difference in the overall five-year survival rate between percutaneous transluminal coronary angioplasty (PTCA) and coronary artery bypass graft surgery (CABG) groups with multivessel disease (1). However, in the subgroup of BARI patients with medication-treated diabetes mellitus (DM), there was a lower survival rate in the PTCA group (2). This finding was not replicated in the BARI registry (3). More importantly, coronary stents were not used in the BARI study.
Previous randomized trials (4,5)have demonstrated lower restenosis rates with coronary stents compared with conventional balloon angioplasty. This has resulted in an exponential increase in stent use (6). Recently, several studies (7–10)have shown high procedural success and an acceptable long-term outcome after multivessel stenting. However, there is little information on short- and long-term results of multivessel stenting in diabetic patients, which is the group addressed in BARI. The present study evaluated in-hospital and one-year outcomes after multivessel stenting in diabetic patients.
The prospective database of the Cardiovascular Research Foundation was reviewed to identify consecutive patients who underwent multivessel stenting in native coronary lesions (1993 to 1999). Multivessel stenting was defined as stenting in two or more major native coronary artery territories. Lesions treated with coil stents were excluded because higher event rates after stenting with coil stents have been demonstrated (11). Patients with an acute myocardial infarction (MI) within the previous 48 h and lesions in arterial or vein grafts were also excluded. There were 689 consecutive patients who underwent multivessel stenting in 1,639 native coronaries. Patients were classified into three groups according to the status of DM. Patients were considered to have DM if they received insulin or oral hypoglycemic drugs at the time of study entry. The 25 patients with a history of diabetes, but who were not taking oral hypoglycemic agents or insulin, were included in the patient group without DM, similar to the BARI study (2). There were 501 patients (1,200 lesions) with no DM, 102 diabetic patients (235 lesions) treated with oral agents, and 86 diabetic patients (204 lesions) treated with insulin.
All procedures were performed after written, informed consent was obtained. All patients underwent stenting according to current guidelines (12). Several slotted-tube, corrugated, and mesh stents were used, depending on lesion length, location, proximal tortuosity, calcification, operator preferences, and/or stent availability. After stent implantation, angiographic optimization was performed using high-pressure balloon dilation to achieve an acceptable angiographic result with <20% residual stenosis by a visual estimate.
All patients received aspirin 325 mg/day at least 24 h before the procedure and continued indefinitely. Patients were treated concomitantly with an additional antiplatelet agent: either ticlopidine, 250 mg twice daily, or clopidogrel, 75 mg/day for 4 weeks.
All cineangiograms were analyzed using a computer-assisted, automated edge-detection algorithm (ARTREK, Quantitative Cardiac Systems, Ann Arbor, Michigan) by a core laboratory that was blinded to the clinical outcomes. Standard qualitative and quantitative definitions and measurements were used (13). The outer diameter of the contrast-filled catheter was used as the calibration, and the minimum lumen diameter was obtained from the single view that revealed the worst vessel stenosis.
Clinical definitions and follow-up
A dedicated Data Coordinating Center performed all data management and analyses. Prespecified clinical and laboratory demographic data were obtained from hospital charts that were reviewed by independent research personnel who were unaware of the objectives of the study; accumulated data were then entered prospectively in the data base. Q-wave MI (QMI) was defined by the presence of new pathologic Q waves on the electrocardiogram. Non-QMI after stenting was defined as creatine kinase-MB enzyme elevation at least three times the upper normal value without new Q waves. Post-procedural renal failure was an increase of ≥25% over baseline serum creatinine levels, with or without dialysis. Long-term (one-year) clinical follow-up was performed by either telephone contact or an office visit. The occurrence of major late clinical events was recorded, including death (all-cause), MI, and target lesion revascularization (TLR); these were adjudicated by accompanying source documentation.
Statistical analysis was performed using Statistical Analysis Systems (SAS Institute Inc., Cary, North Carolina) by a dedicated data analysis center. Data were expressed as proportions or mean values ± SD. Continuous variables were compared using analysis of variance. Categorical variables were compared with chi-square statistics. A p value <0.05 was considered significant.
Kaplan-Meier estimates were used to report cumulative rates of survival, MI, TLR, and event-free survival. Kaplan-Meier curves were compared using the log-rank test. Stepwise logistic regression with entry and stay criteria of 0.10 was used to find independent predictors of one-year mortality, MI, and TLR. Diabetic status was forced into all models; candidate variables for inclusion in the models were female gender, age in years, current smoking, history of cerebrovascular event, preexisting chronic renal insufficiency, post-procedural renal failure, previous PCI, previous MI, hypertension, peripheral vascular disease, and left anterior descending coronary artery and ostial lesion location. Data on reference vessel diameter was incomplete in 30% of patients; therefore, it was not included in the models.
Baseline and procedural characteristics
Patient demographics are presented in Table 1. There were more women in the diabetic group than in the non-diabetic group. A higher proportion of diabetic patients had a history of hypertension, chronic renal failure (baseline serum creatinine value ≥2.0 mg/dl or receiving dialysis), peripheral vascular disease, and cerebrovascular events. A higher incidence of congestive heart failure and a lower left ventricular ejection fraction were observed in diabetic patients treated with oral agents, as compared with the other two groups.
Table 2shows lesion and procedural characteristics. There were more left main and ostial lesions in diabetic patients treated with oral agents. Glycoprotein IIb/IIIa inhibitors were used in <10% of all groups. More than half of non-diabetic patients and diabetic patients treated with oral agents and three-fourths of insulin-treated diabetics underwent stenting of three or more lesions.
Quantitative angiographic results are presented in Table 3. There was a significant difference in reference vessel diameter (smaller in diabetic patients treated with insulin). Before and after the procedure, the minimum lumen diameter, diameter stenosis, and lesion length were not significantly different among the three groups.
Procedural and in-hospital outcomes
Overall procedural success was high in all three groups (Table 4). There were no significant differences in the incidence of QMI and non-QMI. A higher in-hospital CABG rate was observed in diabetic patients treated with insulin, as compared with the other two groups. There was no significant difference in the incidence of cardiac death among the three groups.
Lower one-year survival was observed in diabetic patients treated with either oral agents or insulin, as compared with non-diabetic patients (85% vs. 86% vs. 95%, p < 0.001) (Fig. 1). There was a trend toward more MIs during follow-up in diabetic patients treated with either oral agents or insulin (9.2% vs. 8.1% vs. 3.9%, p = 0.057). Diabetic patients treated with oral agents and insulin had higher one-year TLR rates, as compared with non-diabetic patients (25% vs. 35% vs. 16%, p < 0.001). Lower event-free survival, defined as freedom from death, QMI, and TLR, was observed in diabetic patients treated with either oral agents or insulin (63% vs. 60% vs. 79%, p < 0.001) (Fig. 2).
Multivariate analysis demonstrated that DM (p < 0.001), post-procedural renal failure (p < 0.01), old age (p < 0.01), left anterior descending coronary artery lesions (p = 0.04), and peripheral vascular disease (p = 0.04) were independent predictors of one-year mortality after multivessel stenting. Diabetes mellitus (p = 0.02) and previous CABG (p = 0.02) were independent predictors of MI during follow-up. Multivariate predictors of TLR were DM (p < 0.001), hypertension (p = 0.02), and a previous percutaneous coronary intervention (p = 0.03); baseline renal failure and post-intervention renal failure were not significant.
The present study demonstrated high technical success rates for multivessel stenting in diabetic patients treated with either oral agents or insulin. However, a higher incidence of in-hospital CABG, a higher one-year revascularization rate, and a lower one-year survival rate were observed in diabetic patients, especially those treated with insulin, compared with non-diabetic patients.
Percutaneous coronary intervention in diabetic patients with multivessel disease
Previous randomized trials (1,14–16)demonstrated that there was no significant difference in overall survival rates between the PTCA and CABG groups of patients with multivessel disease. In the BARI study, 1) the in-hospital event mortality rates for PTCA and CABG were 1.1% and 1.3%, respectively (p = NS); 2) the respective in-hospital QMI rates were 2.1% and 4.6% (p < 0.01); and 3) the overall five-year survival rates were 89% in patients assigned to CABG and 86% in those treated with PTCA (p = 0.19) (1). However, in those studies (1,14–16), the TLR rate was higher in the PTCA group (range 37% to 55%) than in the CABG group (range 6% to 13%). In addition, the need for in-hospital CABG in the PTCA group was relatively high (range 5% to 10%).
Diabetes mellitus has been shown to have a negative impact on mortality and morbidity after catheter-based interventions, as well as CABG, although diabetic patients constitute up to 26% of patients presenting for such procedures (2,17,18). Previous studies (18)demonstrated that balloon angioplasty to treat coronary artery disease in diabetic patients has been associated with an increased rate of acute complications and a higher restenosis rate, compared with non-diabetic patients. In a subgroup analysis of the BARI trial in diabetic patients treated with insulin or oral hypoglycemic agents (2), the respective in-hospital event rates for PTCA and CABG were 0.6% and 1.2% for mortality (p = NS) and 1.8% and 5.8% for QMI (p = NS). However, those patients had a significantly lower five-year survival rate with multivessel angioplasty compared with CABG (66% vs. 81%, p = 0.003), and the corresponding cardiac mortality rates were 20.6% and 5.8%. This led to the notion that coronary angioplasty may not be a suitable therapy for multivessel disease in patients with DM (19). However, those randomized trials were conducted before the stent era.
Stenting in diabetic patients with multivessel disease
Since randomized studies (4,5)have demonstrated higher early success and lower restenosis rates in selected lesions, stents are used in up to 80% of interventional procedures worldwide (6). Recently, several studies (7–10)have shown high early success and acceptable long-term results with multivessel stenting. Moussa et al. (7)demonstrated a procedural success rate of 97% in 100 patients who underwent multivessel stenting. The in-hospital mortality rate was 1%, the in-hospital CABG rate was 2%, and the QMI rate was 2%. During follow-up, mortality and CABG rates were 4% and 2%, respectively, and angiographic restenosis was observed in 22% of lesions and 37% of patients. In 398 patients with multivessel stenting, Kornowski et al. (8)demonstrated a procedural success rate of 96% and a hospital mortality rate of 0.5%. In-hospital CABG was performed in 2% of patients, and QMI was observed in 0.9%. During follow-up, the mortality and CABG rates were 0.7% and 7.9%, respectively. Target vessel revascularization was necessary in 15% of lesions.
There is little information on short- and long-term outcomes after multivessel stenting in diabetic patients. The present study demonstrated that, despite a high angiographic success rate, multivessel stenting in diabetic patients, especially those treated with insulin, was associated with higher in-hospital CABG, long-term mortality, and TLR rates, as compared with non-diabetic patients. The rate of peri-procedural non-QMI was 28% in the entire population and did not differ by diabetic status. This likely reflects the fact that this was a high-risk population with a large number of lesions stented. In addition, the low use of glycoprotein IIb/IIIa agents may have contributed to the large number of peri-procedural non-QMIs. Despite this, in-hospital QMI rates were acceptable.
It has long been appreciated that DM is a major risk factor for the development of atherosclerosis, resulting in a wide variety of cardiovascular dysfunction and complications (20). Based on high angiographic success and low acute closure rates, the higher in-hospital CABG rate may be associated with incomplete revascularization due to diffusely diseased coronary arteries. The high long-term adverse cardiac event rate may reflect the development of new lesions, progression of untreated less severe lesions, and incomplete revascularization due to diffusely diseased coronary arteries (21). It may also be associated with rates of co-morbidities, including chronic renal failure, which is associated with a high event rate after percutaneous coronary intervention (22).
Coronary arteries of treated diabetic patients tend to be smaller, probably due to more diffuse disease (18,23). This may partly explain a high restenosis rate in diabetic patients (17,23). Furthermore, high restenosis rates in diabetic patients probably result from complex hormonal and biochemical alterations associated with DM (24). These may result in accelerated intimal hyperplasia after stenting (25).
Recently, randomized studies have demonstrated long-term outcomes in patients with multivessel disease treated with multivessel stenting compared with CABG (9,10). The Arterial Revascularization Therapies Study (ARTS) showed no difference in mortality rate at one-year follow-up between the two groups (9). On the other hand, the Estudio Randomizado Argentino de angioplastia vs CIrugia (ERACI II) demonstrated a lower one-year mortality rate in the multivessel stenting group compared with the CABG group (10). In both studies, repeat revascularization procedures were performed more frequently in the multivessel stenting group.
The baseline characteristics in the present study were worse than those in the multivessel angioplasty or stenting arm of BARI and ARTS (Table 5). Because this study was a real-world clinical study, not a randomized study, the physician, patient, and family might choose multivessel stenting for patients with more complex lesions than the patients included in randomized studies. Thus, it may not be justified to compare the results after multivessel stenting in diabetic patients in the current report with those after CABG in randomized studies. Furthermore, in the BARI registry, there were no significant differences in the long-term mortality of diabetic patients between the PTCA and CABG groups (3). Finally, drug-eluting stents may alter this balance in favor of percutaneous revascularization (26).
First, this is a retrospective analysis rather than a prospective, randomized clinical trial designed to assess the efficacy of stenting in patients with multivessel disease. Therefore, baseline characteristics were worse in patients with DM, especially those treated with insulin. Offsetting this limitation, the data were collected prospectively by independent monitors and entered into a dedicated database, and an independent core laboratory interpreted all angiograms. Furthermore, this represents conditions of real-world experience, rather than the conditions typically included in randomized trials. Second, detailed information on diabetic status, such as fasting glucose values or levels of glycosylated hemoglobin, was not collected. Previous studies (27,28)demonstrated the importance of glycemic control in reducing cardiovascular mortality and morbidity in diabetic patients. In addition, detailed information on lipid status was not available, although the incidence of hypercholesterolemia was not different among the three groups. Third, glycoprotein IIb/IIIa inhibitors were used at a rate of <10% in all groups. Recently, lower long-term mortality and TLR rates have been shown in diabetic patients who received abciximab during percutaneous coronary intervention (29)or stenting (30). This may partially explain the high event rates during follow-up in this study. Fourth, the current analysis does not include patients undergoing multivessel stenting in saphenous vein grafts. Because this patient population may differ in their baseline demographics and clinical outcomes, we have elected to address the issue of multivessel stenting in vein grafts in a separate study. Finally, the reference vessel diameter was not assessed in 30% of patients. Diabetics often have significantly smaller vessels, and this may have been an important confounder in the higher rate of TLR reported.
Despite a high technical success rate of multivessel stenting, diabetic patients, especially those treated with insulin, have higher in-hospital CABG, lower survival, and higher subsequent revascularization procedure rates at one-year clinical follow-up, compared with non-diabetic patients. If multivessel stenting is performed in treated patients with DM, meticulous post-procedural scrutiny and medical management are warranted.
- Bypass Angioplasty Revascularization Investigation
- coronary artery bypass graft surgery
- diabetes mellitus
- myocardial infarction
- percutaneous transluminal coronary angioplasty
- Q-wave myocardial infarction
- target lesion revascularization
- Received July 29, 2002.
- Revision received March 28, 2003.
- Accepted April 4, 2003.
- American College of Cardiology Foundation
- The BARI Investigators
- Detre K.M,
- Guo P,
- Holubkov R,
- et al.
- Rodriguez A,
- Ambrose J
- Moussa I,
- Reimers B,
- Moses J,
- et al.
- Kornowski R,
- Mehran R,
- Satler L.F,
- et al.
- Rodriguez A,
- Bernardi V,
- Navia J,
- et al.,
- the ERACI II Investigators
- Lansky A.J,
- Roubin G.S,
- O'Shaughnessy C.D,
- et al.
- Popma J.J,
- Bashore T.D
- Abizaid A,
- Kornowski R,
- Mintz G.S,
- et al.
- Kip K.E,
- Faxson D.P,
- Detre K.M,
- et al.
- O'Neill W.W
- Rosenman Y,
- Sapoznikov D,
- Mosseri M,
- et al.
- Schofer J,
- Schluter M,
- Rau T,
- et al.
- Aronson D,
- Bloomgarden Z,
- Rayfield E.J
- Kornowski R,
- Mintz G.S,
- Kent K.M,
- et al.
- Laakso M. Glycemic control and the risk for coronary heart disease in patients with non-insulin-dependent diabetes mellitus. The Finnish Studies. Ann Intern Med 1996;124:127–30.
- Bhatt D.L,
- Marso S.P,
- Lincoff A.M,
- et al.