Author + information
- Received February 19, 1998
- Revision received May 13, 1998
- Accepted May 20, 1998
- Published online September 1, 1998.
- Alexandre Abizaid, MDa,
- Ran Kornowski, MDa,
- Gary S Mintz, MD, FACCa,
- Mun K Hong, MD, FACCa,
- Andrea S Abizaid, MDa,
- Roxana Mehran, MDa,
- Augusto D Pichard, MD, FACCa,
- Kenneth M Kent, MD, PhD, FACCa,
- Lowell F Satler, MD, FACCa,
- Hongsheng Wu, PhDa,
- Jeffrey J Popma, MD, FACCa and
- Martin B Leon, MD, FACCa,* ()
- ↵*Address for correspondence: Martin B. Leon, MD, Director, Cardiovascular Research, Washington Cardiology Center, 110 Irving Street NW, Suite 4B1, Washington, DC 20010
Objectives. We compared the clinical outcomes following coronary stent implantation in insulin-treated diabetes mellitus (IDDM), non-IDDM patients, and nondiabetic patients.
Background. Diabetic patients have increased restenosis and late morbidity following balloon angioplasty. The impact of diabetes mellitus (DM), especially IDDM, on in-stent restenosis is not known.
Methods. We studied 954 consecutive patients with native coronary artery lesions treated with elective Palmaz-Schatz stents implantation using conventional coronary angiographic and intravascular ultrasound methodology. Procedural success, major in-hospital complications, and 1-year clinical outcome were compared according to the diabetic status.
Results. In-hospital mortality was 2% in IDDM, significantly higher (p <0.02) compared with non-IDDM (0%) and nondiabetics (0.3%). Stent thrombosis did not differ among groups (0.9% in IDDM vs. 0% in non-IDDM and 0% in nondiabetics, p >0.1). During follow-up, target lesion revascularization (TLR) was 28% in IDDM, significantly higher (p <0.05) compared with non-IDDM (17.6%) and nondiabetics (16.3%). Late cardiac event–free survival (including death, myocardial infarction [MI], and any coronary revascularization procedure) was significantly lower (p = 0.0004) in IDDM (60%) compared with non-IDDM (70%) and nondiabetic patients (76%). By multivariate analysis, IDDM was an independent predictor for any late cardiac event (OR = 2.05, p = 0.0002) in general and TLR (odds ratio = 2.51, p = 0.0001) in particular.
Conclusions. In a large consecutive series of patients treated by elective stent implantation, IDDM patients were at higher risk for in-hospital mortality and subsequent TLR and, as a result, had a significantly lower cardiac event–free survival rate. On the other hand, acute and long-term procedural outcome was found to be similar for non-IDDM compared with nondiabetic patients.
Diabetic patients have an increased risk of restenosis after balloon angioplasty (PTCA) (1–10), resulting in increased late morbidity and mortality compared with nondiabetic patients. Recently, intracoronary stents have been shown to improve procedural outcome and decrease restenosis in focal, de novo native coronary lesions compared with PTCA (11,12). However, even with stents, diabetic patients seem to have increased late loss index and restenosis compared with nondiabetic patients (13,14), although stents may still improve results compared with PTCA in diabetic patients (15,16). Furthermore, in the stent era few studies have examined the acute procedural and long-term outcome of insulin-dependent diabetes mellitus (IDDM) and non–insulin-dependent diabetes mellitus (non-IDDM) versus nondiabetic patients following coronary interventions.
The purpose of the current study was to determine the influence of IDDM and non-IDDM on in-hospital and long-term outcomes following stent implantation in native coronary arteries. We compared in-hospital results and 1-year clinical outcome in diabetics compared with a concurrent series of nondiabetic patients.
The patient cohort includes a consecutive series of 954 patients (1,304 lesions) in the Cardiology Research Foundation Angioplasty Database, treated with Palmaz-Schatz stents (Cordis) in native coronary arteries between January 1994 and January 1996. Patients were divided into three groups according to their diabetic status. There were 706 (74%) patients without diabetes, 151 (15.8%) diabetic patients treated with oral hypoglycemic drugs (non-IDDM), and 97 (10.2%) diabetic patients treated with insulin (IDDM). Indications included planned elective stent use and planned provisional stent implantation use following a suboptimal result with a nonstent balloon or other primary device. Patients were excluded from the current analysis if they were treated by Gianturco-Rubin stents, investigational (non–FDA-approved) stents, if the stents were implanted for bail-out indications or to treat saphenous vein graft lesions, or if stent implantation was preceded by atheroablative device use.
Baseline clinical demographics and in-hospital events were confirmed by independent hospital chart review. Long-term outcome data were obtained by serial telephone interviews conducted by research nurses. Late clinical events (death, nonfatal myocardial infarction, and any revascularization) and target lesion revascularization (TLR) were adjudicated and corroborated by primary source documentation. All patients were treated and studied after giving informed consent. These protocols have the ongoing approval of the Washington Hospital Center Institutional Review Board.
Stent implantation techniques
Details of the stent implantation procedure have been previously described (11,12). Following the initial balloon angioplasty, coronary or “biliary” Palmaz-Schatz stents were implanted over a 0.014-in. guidewire. Coronary stents were used whenever possible; the larger “biliary” version was reserved for vessels ≥4 mm in diameter. Adjunct high-pressure PTCA was performed after initial stent deployment (routinely to ≥16 atmospheres) in all cases. Angiographic and intravascular ultrasound (IVUS) assessment was obtained pretreatment, post–stent implantation, and post–adjunct PTCA. IVUS was performed at all time points in 94% of cases. Optimal stent implantation was carefully monitored using an interactive technique with prespecified IVUS endpoints and additional high-pressure inflations as needed. IVUS was used to optimize stent apposition, expansion (defined as minimal stent area 80% of the average of the proximal and distal references lumen area), and lesion coverage. In addition, IVUS enabled the detection of inflow-outflow obstruction and residual dissection at both stent margins. The poststent anticoagulation regimens included aspirin (325 mg/d) and ticlopidine (250 mg BID for 1 month) for all patients. Low molecular weight heparin (for 2 weeks) was administered in particularly high-risk patients (usually patients treated with ≥3 stents).
Cineangiograms were reviewed by our Angiographic Core Laboratory at the Washington Hospital Center by an observer unaware of the clinical outcome. Standard morphologic criteria were used for the identification of lesion location, length (“shoulder-to-shoulder”), eccentricity, irregularity, fluoroscopic calcification, and ulceration.
Quantitative angiographic analysis was performed using selected end-diastolic cineframes demonstrating the stenosis in its most severe and nonforeshortened projection. Cineframes were optically magnified (2.4:1) and digitized using a cine-video converter. Using the contrast-filled guiding catheter as the calibration standard, reference and minimal lumen diameters (MLD) were determined before and after stent implantation, and post–final adjunct PTCA using a validated, automated edge-detection algorithm (Cardiovascular Measurement Systems; Medis). Based upon these measurements, percent diameter stenoses were determined.
IVUS studies were performed using a commercially available system (Boston Scientific Corporation/Cardio Vascular Imaging System, Inc.). This system incorporated a single-element 30-MHz transducer mounted on the tip of a flexible drive shaft within a 3.2-F short monorail polyethylene imaging sheath. The transducer was rotated 1800 rpm to form planar images in real time and withdrawn mechanically at 0.5 mm/s with a motorized transducer pullback device to perform the imaging sequence. IVUS studies were recorded on 0.5-in. high-resolution s-VHS tapes for off-line analysis. To perform the imaging sequence, the imaging catheter was advanced approximately 10 mm distal to the lesion, the motorized transducer pullback device was activated, and a complete uninterrupted imaging run was performed back to the aorto-ostial junction.
Cross-sectional measurements of external elastic membrane (EEM), stent, lumen, and plaque + media (P + M) cross-sectional areas (CSA) by IVUS have been validated (17,18). The term EEM is short-hand for the media-adventitia border, which is a reproducible measure of the total arterial CSA. Because media thickness cannot be measured accurately, P + M CSA was used as a measure of the amount of atherosclerotic plaque. The term cross-sectional narrowing (CSN) has also been called the plaque burden or percent plaque area by other investigators. When the tissue encompassed the catheter, the lumen was assumed to be the physical (not acoustical) size of the imaging catheter. Therefore, 1.0 mm was the smallest MLD and 0.8 mm2was the smallest lumen CSA that could be measured preintervention.
Using computerized planimetry (TapeMeasure; Indec Systems, Inc.), the target lesion was assessed preintervention by measuring: (1) EEM CSA (mm2); (2) lumen CSA (mm2); (3) MLD (mm); (4) P + M CSA (mm2, equals EEM CSA minus lumen CSA); and (5) CSN (%, equals P + M CSA divided by EEM CSA). The target lesion was also assessed postintervention. Final IVUS measurements included: (1) stent CSA (mm2), (2) lumen CSA (mm2), and (3) MLD (mm).
The target lesion was normalized for the reference segment. Reference segment dimensions were calculated as the mean of the proximal and distal reference. Each reference segment was selected as the most normal-looking cross section within 10 mm proximal or distal to the target lesion, but before a major side branch. If either the proximal or distal reference segment could not be analyzed (e.g., ostial lesion location, diffuse disease, or major side branch close to the lesion), then only one reference segment measurement was used.
Statistical analysis was performed using SAS software (19). Continuous variables are presented as mean ± 1 standard deviation. Categorical data are presented as percent frequency and compared between groups using chi-square statistics. Multivariate analysis was performed by SAS (SAS Institute) Logistic Regression Statistics. Survival curves were calculated and displayed using the SAS LIFETEST procedure. The Wilcoxon test was used for survival comparison between groups (IDDM vs. non-IDDM vs. nondiabetics). The means of nominal values were compared using analysis of variances (ANOVA). A p value <0.05 was considered as statistically significant.
lists the baseline characteristics of all treated patients, grouped according to the diabetic status. Overall, demographics in patients with non-IDDM and nondiabetics were similar. Patients with IDDM were more often female (50.5%), hypertensive (73.3%), and had a previous CABG surgery (32.6%) (p <0.05 compared with the other two groups). Left ventricular ejection fraction was similar for all three groups. The average number of stents per procedure was not different among the three groups (1.9 ± 1.0, 2.2 ± 1.5, and 1.8 ± 1.0 in nondiabetic, non-IDDM, and IDDM patients, respectively, p >0.1).
Qualitative and quantitative angiographic results (Tables 2 and 3)
⇑⇓Lesion location was similar among the groups. Likewise, the qualitative lesion characteristics were also similar among the groups. Quantitative measurements preprocedure showed that the reference vessel size was smaller in the IDDM group compared with nondiabetics; however, postprocedure measurements were similar in all three groups.
IVUS analysis (Tables 4 and 5)
Postintervention, the instent lumen area and minimum lumen diameter were significantly smaller in the IDDM group compared with the other two groups.
In-hospital events (Table 6)
Overall procedural success (<50% final diameter stenosis and TIMI flow >3 without major in-hospital complications) was uniformly high: 97.2% for IDDM, 96.8% for non-IDDM, and 98.3% for nondiabetic patients (p >0.1). Similarly, combined major in-hospital complications (death, Q-wave myocardial infarction, and emergent coronary artery bypass surgery) did not differ significantly among the groups. However, in-hospital mortality was 2.0% in IDDM, significantly higher compared with 0% in non-IDDM and 0.3% in nondiabetics (p <0.05 for each comparison). Subacute stent thrombosis was less than 1% and did not differ significantly among the three groups.
Long-term outcomes (Table 7)
One-year clinical follow-up was available in 950 of 954 (99.6%) patients. The overall coronary event rate (death, nonfatal myocardial infarction, TLR, and any revascularization) during follow-up was higher in IDDM (39.8%) compared with non-IDDM and nondiabetic patients (p <0.05). There was no statistically significant difference in either death or nonfatal myocardial infarction among the groups. TLR was 28% in IDDM, significantly higher compared with 17.6% in non-IDDM and 16.3% in nondiabetics (p <0.05 for each comparison). Moreover, revascularization procedures (TLR and non–TLR-related procedure) were more common in IDDM (37.6%), compared with non-IDDM (28.2%) and nondiabetics (22.7%) (p <0.05 for each comparison). Actuarial event-free survival curves for any coronary event are shown in Figure 1. Coronary event–free survival at 1 year was 60% in IDDM, significantly lower compared with 70% in non-IDDM and 76% in nondiabetics (p = 0.0004). Similarly, TLR-free survival (Fig. 2)was significantly lower in IDDM, compared with the other two groups (p = 0.0001).
Logistic regression analysis was used to identify the independent predictors of any adverse cardiac event and TLR at 1 year of follow-up. IDDM (OR = 2.05, p = 0.0002), QCA final procedural MLD (OR = 0.44, p = 0.01), and IVUS reference site plaque + media CSA (OR = 1.24, p = 0.01) were found to predict any adverse event at 1 year. IDDM (OR = 2.51, p = 0.0001), final procedural MLD by QCA (OR = 0.47, p = 0.023), IVUS final minimum lumen CSA (OR = 0.72, p = 0.013), IVUS preprocedural reference lumen CSA (OR = 0.57, p = 0.011), and IVUS preprocedure lesion site plaque + media (OR = 2.04, p = 0.003) were found to predict TLR at 1 year.
In this large consecutive series of patients treated with IVUS-guided elective stent implantation in native coronary arteries, IDDM patients had worse procedural and long-term clinical outcomes compared with non-IDDM or nondiabetic patients. In this study, patients with IDDM (compared with non-IDDM and nondiabetic patients) manifested: (1) higher in-hospital mortality; (2) higher cardiac event rate during 1-year follow-up, which is primarily due to more coronary revascularization procedures (both TLR and non-TLR); and (3) significantly lower event-free survival. Importantly, IDDM was found to be an independent predictor for coronary events and TLR during follow-up. Conversely, after coronary stenting, non-IDDM patients had a similar in-hospital mortality and late TLR compared with nondiabetic patients. Nevertheless, overall event-free survival for non-IDDM patients was lower than nondiabetic patients, primarily due to more non-TLR revascularization procedures (10.6%) from progression of disease in untreated sites. By multivariate analysis, we identified independent clinical, angiographic, and IVUS predictors of future clinical events. These predictors were IDDM, final MLD by QCA, and IVUS plaque + media at the reference sites. The independent predictors of TLR were final MLD by QCA, final IVUS lumen CSA, and preprocedure plaque + media at lesion site.
PTCA in diabetic patients
The role of DM as a clinical predictor for an adverse procedural outcome and higher restenosis rate after PTCA has been well documented (1–10). Prior to the stent era, these studies reported restenosis rates of 47–71% after PTCA among diabetic patients, compared with 30–40% in nondiabetics. The majority of these studies did not distinguish between IDDM patients versus non-IDDM. Stein et al. (10)were the only investigators who distinguished between IDDM and non-IDDM. They reported that IDDM patients had similar procedural results, but worse long-term outcome compared with non-IDDM. These data strongly indicate that restenosis and progression of coronary disease are more frequent among diabetic patients. More recently, the Emory experience has found that at 5 and 10 years, IDDM patients had lower survival rates after balloon PTCA compared with to the entire diabetic population (72% and 31% vs. 78% and 45%, respectively) (20). A recent subanalysis of The Bypass Angioplasty Revascularization Investigation (BARI) Trial revealed a worse 5-year survival rate in diabetic patients with multivessel coronary disease treated by balloon angioplasty compared with coronary bypass surgery (21). These findings highlighted again the problem of managing DM patients by the transcatheter interventional approach.
Stent treatment in diabetic patients
Carrozza et al. (13)found increased late loss and a greater incidence of restenosis among diabetics after coronary stenting compared with nondiabetics (55% vs. 20%, p = 0.001). This study had a relatively small number of diabetic patients (37 patients), the majority of whom (70%) were non-IDDM. Furthermore, 54% of the lesions were in saphenous vein grafts. Thus, although this study indicated that diabetics had a higher restenosis rate, it could not distinguish between IDDM and non-IDDM, and suggested that intimal hyperplasia is more abundant in DM within the stented site. A recent subanalysis of the STRESS trial revealed a significant decrease in restenosis among DM patients treated by stents implantation compared with balloon angioplasty (24% vs. 60%, p <0.001), suggesting a major benefit for stents among DM patients (15). Likewise, Van Belle et al. found a lower angiographic restenosis rate in diabetic patients who were treated with stents compared with those treated with balloon angioplasty (25% vs. 63%) (16). In the same study, the restenosis rates were similar in diabetics and nondiabetics treated with stents (25% vs. 27%, respectively). In none of these studies was there a distinction between IDDM and non-IDDM patients.
Mechanisms for restenosis in DM
Recent serial (postintervention and follow-up) IVUS observations from our laboratory have demonstrated the cause for increased restenosis in DM (22). It was found that diabetic patients manifested an increased proliferative response after transcatheter coronary interventions in both stented and nonstented lesions. Because of small patient numbers, no distinction could be made between IDDM and non-IDDM patients.
The distinction between IDDM and non-IDDM patients became apparent from recent pre-intervention IVUS observations: there were major differences in lesion and reference vessel adaptive remodeling characteristics in IDDM patients versus non-IDDM and nondiabetic patients (23). According to these IVUS findings, coronary atherosclerosis in patients with IDDM manifested as smaller vessels; there was less atherosclerotic plaque burden and less positive remodeling for a given degree of plaque accumulation. These lesion site characteristics might attenuate the benefit of coronary stenting in IDDM patients. Because of the smaller EEM dimensions, contemporary stenting (maximizing balloon to artery ratio and high pressure balloon expansion) might exaggerate vascular injury and hence evoke a more aggressive proliferative response within coronary stents among IDDM patients. Moreover, final stent CSA was smaller in IDDM patients compared with their counterparts. This variable was also found to be an independent predictor for TLR in the multivariate model.
Excessive neointimal reaction in diabetes probably resulted from complex hormonal and biochemical alterations associated with DM in general and with insulin treatment in particular (24). These might result in accelerated smooth muscle cell proliferation after coronary stenting (25–27). Experimental studies have found that vascular smooth muscle cells are sensitive to the growth stimulatory action of insulin and insulin-like growth factors (25–27). Insulin therapy may elevate the vascular sympathetic tone (28). In addition, the proatherogenic effect of insulin therapy might be responsible for accelerated atherosclerosis in untreated less severe lesions, resulting in more non-TLR interventions in IDDM patients (29,30).
This was a retrospective analysis of the clinical, angiographic, and IVUS data derived from a large group of consecutively studied patients with careful and nearly complete follow-up data. The relatively small group of IDDM patients analyzed may have resulted in Type II statistical errors when comparing the clinical outcome to those in non-IDDM and non-DM groups.
Differences in baseline demographics and lesion characteristics could have produced the current findings. For example, lesions in smaller vessels containing unfavorable plaque morphology might result in unfavorable stent treatment results. However, angiographic morphology and lesion location in the IDDM patients were similar to those in the other two groups.
Although, in our study, insulin treatment was found to predict unfavorable clinical outcomes after stenting, insulin use may be a surrogate rather than a cause of such adverse outcomes. Insulin-treated patients might be at more advanced stage of coronary atherosclerosis and therefore more prone to cardiovascular complications.
In a large consecutive series of patients treated with stent implantation, IDDM patients had higher in-hospital mortality and morbidity and increased 1-year coronary event rates following the procedure. Consequently, a significantly lower event-free survival was found in IDDM compared with non-IDDM and non-DM patients. By contrast, non-IDDM patients have a similar procedural success and long-term TLR compared with nondiabetic patients. Stents appear to equalize procedural results and TLR events for non-IDDM patients (compared with nondiabetics), but IDDM patients are still at increased risk for acute and long-term adverse outcomes.
☆ This study was supported in part by the Cardiology Research Foundation, Washington, DC.
- cross-sectional area
- cross-sectional narrowing
- external elastic membrane
- insulin-dependent diabetes mellitus
- intravascular ultrasound
- minimum lumen diameter
- P + M
- plaque + media
- balloon angioplasty
- target lesion revascularization
- Received February 19, 1998.
- Revision received May 13, 1998.
- Accepted May 20, 1998.
- American College of Cardiology
- Vandormeal M.G,
- Deligonul U,
- Kern M.J,
- et al.
- Hollman J,
- Badhwar K,
- Beck G.J,
- Granco I,
- Simpfendorfer C
- Deligonul U,
- Vandormeal M,
- Kern M.J,
- Galan K.L
- Lambert M,
- Bonan R,
- Cote G,
- et al.
- Weintraub E.S,
- Kosinski A.S,
- Brown C.L,
- King S.B
- Kip K.E,
- Faxon D.P,
- Detre K.M,
- Yeh W,
- Kelsey S.F,
- Curier J.W
- Stein B,
- Weintraub W.S,
- Gebhart S.S.P,
- et al.
- Elezi S,
- Schuhlen H,
- Wehinger A,
- et al.
- Savage M.P,
- Fischman D.L,
- Slota P,
- et al.
- Van Belle E,
- Bauters C,
- Hubert E,
- et al.
- Tobis J.M,
- Mallery J,
- Mahon D,
- et al.
- Potkin B.N,
- Bartorelli A.L,
- Gessert J.M,
- et al.
- ↵Ray AA. SAS User’s Guide. Cary, NC: SAS Institute, 1986.
- Weintraub W.S,
- Stein B,
- Kosinski A,
- et al.
- Kornowski R,
- Mintz G.S,
- Kent K.M,
- et al.
- Aronson D,
- Bloomgarden Z,
- Rayfield E.J
- Stout R.W,
- Bierman E.L,
- Ross R
- Bornfeldt K.E,
- Raines E.W,
- Nakano T,
- Graves T.N,
- Krebs E.G,
- Ross R
- Baron A.D
- Sobel B.E