Pioglitazone Reduces Neointima Volume After Coronary Stent Implantation
A Randomized, Placebo-Controlled, Double-Blind Trial in Nondiabetic Patients
Background— Restenosis requiring reintervention limits the long-term success after coronary stent implantation. Thiazolidinediones, like pioglitazone or rosiglitazone, are oral antidiabetic drugs with additional antirestenotic properties. In a randomized, placebo-controlled, double-blind trial, we examined the effect of 6-month pioglitazone therapy on neointima volume after coronary stenting in nondiabetic coronary artery disease patients.
Methods and Results— Fifty nondiabetic patients after coronary stent implantation were randomly assigned to pioglitazone (30 mg daily; pio) or placebo (control) treatment in addition to standard therapy, and neointima volume was assessed by intravascular ultrasound at the 6-month follow-up. Both groups were comparable with regard to baseline characteristics, angiographic lesion morphology, target vessel, and length of the stented segment. In addition, there were no statistical differences in minimal lumen diameter before and after intervention, as well as reference diameter after stent implantation. In this study population of nondiabetic patients, pio treatment did not significantly change fasting blood glucose, fasting insulin, or glycosylated hemoglobin levels, as well as lipid parameters. In contrast, pio treatment significantly reduced neointima volume within the stented segment, with 2.3±1.1 mm3/mm in the pio group versus 3.1±1.6 mm3/mm in controls (P=0.04). Total plaque volume (adventitia-lumen area) was significantly lower at follow-up in the pio group (11.2±3.2 mm3/mm) compared with controls (13.2±4.2 mm3/mm; P=0.04). Moreover, the binary restenosis rate was 3.4% in the pio group versus 32.3% in controls (P<0.01).
Conclusions— Thus, 6-month treatment with pio significantly reduced neointima volume after coronary stent implantation in nondiabetic patients. These data bolster the hypothesis that antidiabetic thiazolidinediones, in addition to their metabolic effects, exhibit direct antirestenotic effects in the vasculature.
Received January 12, 2005; revision received June 6, 2005; accepted June 7, 2005.
Restenosis requiring reintervention is still a limitation of percutaneous coronary angioplasty. Despite the use of stents, the rate of restenosis remains 20% to 30% in all patients, making it a challenging problem in interventional cardiology.1,2 Migration and proliferation of medial vascular smooth muscle cells (VSMCs) is the predominant mechanism of neointima formation leading to in-stent restenosis.3 Various pharmacological strategies, such as the application of antiplatelet or antiinflammatory agents, have been shown to modulate these processes in vitro and were efficient in reducing restenosis in animal models. Still, in clinical trials, most of these attempts did not successfully limit neointima formation after coronary stenting.4–7
Editorial p 2759
Thiazolidinediones (TZDs), like pioglitazone (pio) or rosiglitazone, are a novel class of oral antidiabetic agents currently used to treat patients with type 2 diabetes mellitus. These agents increase insulin sensitivity and, as such, have favorable effects on blood glucose levels and the lipid profile in treated patients.8 Beyond their metabolic action, TZDs have been shown to exhibit antiinflammatory and antiatherogenic effects in vascular cells in vitro and to limit lesion development in various animal models of arteriosclerosis (reviewed in Marx et al9). Moreover, TZDs inhibit VSMC proliferation and migration, 2 critical processes in neointima formation after coronary stenting.10–12 Data from rodent models suggest that TZDs limit intimal proliferation after vascular injury,13 and in clinical studies with type 2 diabetic coronary artery disease (CAD) patients, TZDs have been shown to reduce neointima formation as well as restenosis after coronary stent implantation.14–17 Still, it remains unclear to what extend these effects depend on the metabolic action of these drugs and what might mainly be due to the improvement in glycemic control.
Therefore, we performed a randomized, placebo-controlled, double-blind trial in nondiabetic CAD patients and examined by intravascular ultrasound (IVUS) the effect of 6 months of pio treatment on neointima volume after coronary stent implantation.
Study Design and Patient Selection
Fifty nondiabetic patients with angina pectoris and/or exercise-induced ischemia in the presence of a significant stenosis in a native coronary artery suitable for stent implantation were included in this randomized, placebo-controlled, double-blind trial. The nondiabetic state was assessed by a negative history of diabetes mellitus, no treatment with antidiabetic drugs, and/or assessment of fasting blood glucose. All patients were recruited at the Department of Internal Medicine II, University of Ulm, Ulm, Germany. Exclusion criteria included the following: diabetes mellitus, acute ST-segment elevation myocardial infarction (MI), contraindications to treatment with platelet inhibitors or pio, impaired liver function (aspartate aminotransferase or alanine aminotransferase 2.5-fold above upper normal limits), renal insufficiency requiring hemodialysis, pregnancy, systemic inflammatory disease, and life expectancy <6 months. Patients scheduled for angioplasty were randomized to placebo or pio 30 mg daily after written, informed consent was obtained. The first dose was given before the procedure, and treatment was continued until the 6-month follow-up angiography. Study medication was given in addition to standard treatments, including aspirin 100 mg/d, β-blockers, angiotensin-converting enzyme inhibitors, and statins. In addition, patients received clopidogrel 75 mg/d for at least 4 weeks after coronary stenting. Patients were seen after 8 weeks for clinical follow-up and were scheduled for a repeated angiography, including IVUS, 6 months after the primary intervention. The study was conducted according to the principles of the Declaration of Helsinki and was approved by the local ethics committee.
The primary end point of this study was the extent of neointima volume after 6 months, as assessed by IVUS. A secondary end point was the mean diameter stenosis of the total segment after 6 months, as assessed by quantitative coronary angiography (QCA).
Measurement of Inflammatory Biomarkers
Tumor necrosis factor (TNF)-α and soluble CD40L (sCD40L) were determined by ELISA (R&D Systems) according to the manufacturer’s protocol. Fibrinogen and C-reactive protein (CRP) were measured as previously described.18
Angioplasty Procedure and IVUS
All patients were pretreated with at least 500 mg aspirin orally or intravenously. They received a preangioplasty heparin bolus that was adjusted according to the activated clotting time (>280 seconds). Lesions were treated by primary or direct stent implantation and IVUS guidance to ensure correct stent size and complete strut apposition. According to the study protocol, patients received only Express Stents (Boston Scientific Scimed, Inc), but the number of stents was not limited. Segments were examined by mechanical IVUS (UltraCross 2.9F, 30 MHz; Boston Scientific Scimed) with automated pullback at 0.5 mm/s. A coronary segment beginning 5 mm distal to and extending 5 mm proximal to the stented segment was also analyzed. A computer-based contour-detection program was used for automated 3-dimensional reconstruction of the segments (Medis Medical Imaging Systems BV).19 Total vessel volume, stent volume, and lumen volume were calculated. In the absence of neointimal formation, lumen volume was delineated by the boundaries of the stent struts. Total plaque volume, plaque volume behind the stent, and neointima formation were calculated as total vessel volume minus lumen volume, total vessel volume minus stent volume, and stent volume minus lumen volume, respectively. To account for differences in stented length, all IVUS parameters were calculated per millimeter of stent length for the stented segment and per millimeter of segment length for the proximal and distal adjacent segment.
Quantitative Coronary Angiography
QCA before and after stent implantation and at follow-up was performed in the same projections of the treated lesion after administration of intracoronary glycerol trinitrate. To assess the minimal luminal diameter, the most severe stenosis in 2 orthogonal views was measured. Angiographic measurements were done offline with Pie Medical software version 2.1 (Pie Medical Imaging) as previously described.20,21
Differences in metabolic parameters between groups and between treatment time points within a group were analyzed by the Mann-Whitney U test or Student t test, as appropriate. The primary end point was neointima volume in the stented segment as determined by IVUS at the 6-month follow-up. Samples size was calculated on the basis of results from a previous trial that had examined the effect of troglitazone on neointima formation in diabetic subjects, with neointima formation in the control group of 3.5±1.8 mm3/mm and in the troglitazone group of 2.0±0.9 mm3/mm, resulting in 24 lesions per group to achieve statistical significance (α=0.05, β=0.2, 2 tailed). To account for dropouts and an assumed 80% IVUS-follow-up, we planned to include 25 patients per group, with an assumed target-lesion rate of 1.4 per patient. The random allocation sequence was blocked for every 4 patients. The secondary end point was the mean diameter stenosis of the total segment after 6 months, as assessed by QCA. The total segment included edge effects 5 mm proximal and distal to the stented segment according to trials with brachytherapy.20 Furthermore, the occurrence of major adverse cardiac events, including death, MI, and need for reintervention (angioplasty or surgical revascularization), was analyzed. Group comparisons were done on a per-lesion basis. To account for repeated assessments within 1 patient, we performed a GEE linear regression (for continuous outcomes) or a GEE logistic regression (for binary outcomes) to estimate corrected probability values. Probability values of IVUS and QCA data were adjusted for established parameters that could influence follow-up results (reference diameter after procedure, minimal luminal diameter after procedure, length of stented segment, and the presence of acute coronary syndrome).22 GEE regression was performed with the SAS statistical software package (version 8.02 for Unix, SAS Institute Inc). Discrete variables were expressed as numbers and percentages and compared by the χ2 test. Summary values are expressed as mean±SD. Skewed data were reported as median (interquartile range); differences between means of continuous variables were analyzed by t test or the rank-sum test (Statistica version 6.0, StatSoft Inc). Statistical significance was assumed at the 5% α-error level (P<0.05).
Fifty nondiabetic patients with CAD requiring coronary intervention were enrolled in this study and randomized to receive either placebo (24 patients) or pio (26 patients) treatment in addition to standard therapy for CAD. One patient experienced acute cholecystitis on the day of intervention, and in 1 patient, stent implantation was not successful, leading to exclusion from the study (1 patient in each group). Both groups did not significantly differ with respect to baseline characteristics (Table 1), angiographic lesion morphology, and target vessel. There were no significant differences in procedural characteristics other than a significantly higher maximal inflation pressure in the placebo group (Table 2). Still, in both groups, high-pressure inflation was performed, and the groups did not significantly differ with respect to the balloon-to-artery ratio. There were no differences in minimal lumen diameter and reference diameter before and after stent implantation between groups. Stents were implanted in 34 stenoses in the placebo group and in 36 lesions in the pio group (1.5 lesions per patient). One patient in the placebo group (diagnosis of cancer during follow-up) and 4 patients in the pio group were lost to IVUS follow-up (1 patient owing to a diagnosis of cancer, 1 patient refusal of reangiography, and 2 patients withdrawn from the study because of noncompliance; Figure 1). No serious drug-related side effects were observed.
In this study population of nondiabetic patients, pio treatment did not significantly change fasting blood glucose (5.3 [4.9, 5.5] versus 5.4 [4.8, 6.7] mmol/L; P=0.52), fasting insulin (5.1 [3.9, 7.6] versus 7.9 [5.0, 12.1] mmol/L; P=0.08), or glycosylated hemoglobin (HbA1c; 5.6±0.3% versus 5.6±0.6%; P=0.88) levels compared with placebo at 6 months’ follow-up. In addition, pio did not significantly change total cholesterol, HDL cholesterol, or triglyceride levels (Table 3). Moreover, changes in the parameters from baseline to follow-up were not significantly different between the 2 groups (data not shown).
Because TZD treatment has been shown to modulate inflammatory biomarkers of arteriosclerosis, we also measured plasma levels of CRP, fibrinogen, TNF-α, and sCD40L. Only plasma levels of fibrinogen were significantly lowered by pio, but they were also decreased in the placebo group, leading to a nonsignificant difference between the groups at the 6-month follow-up. In addition, pio did not significantly affect plasma levels of CRP, TNF-α, or sCD40L.
IVUS and Angiographic Data
IVUS after intervention revealed complete stent apposition to the vessel wall in all lesions. One patient in the pio group had an asymptomatic stent thrombosis of the vessel with 2 stenoses and was not available for follow-up assessment of neointima volume by IVUS. This patient was also excluded from QCA follow-up.
With regard to the primary end point, pio significantly reduced neointima volume within the stented segment compared with placebo (2.3±1.1 mm3/mm in the pio group versus 3.1±1.6 mm3/mm in the placebo group; P=0.04). Moreover, after 6 months, total plaque volume in the stented area was significantly lower in patients treated with pio (11.2± 3.2 mm3/mm) compared with controls (13.2±4.2 mm3/mm; P=0.04; Figure 2 and Table 4). Similarly, in the adjacent segments proximal and distal to the stent (edge effects), total plaque volume was significantly lower in the pio compared with the placebo group (Table 4). The effect on neointima volume and total plaque volume remained statistically significant, with adjusted linear regression to account for repeated assessments within 1 patient (P values of 0.046 and 0.002, respectively). Neither inflation pressure nor insulin levels were correlated significantly with the primary end point (data not shown).
Results of QCA are detailed in Table 5. In the pio group, mean diameter stenosis in the target lesion and the total segment was significantly reduced compared with placebo (percentage of luminal diameter: pio, 22.1±12.7 versus placebo, 37.3±24.2; P=0.01), leading to a significant reduction in the angiographic restenosis rate by pio treatment (Figure 3). There was a nonsignificant reduction of late loss and late loss index in the pio group compared with placebo. With adjusted GEE linear regression analysis to account for repeated assessments within 1 patient, the effect of pio on mean diameter stenosis and binary restenosis remained significant (P=0.0004 and 0.002, respectively). Moreover, with this analysis, the higher minimal luminal diameter at follow-up (P=0.003 for target lesion and P=0.0004 for the total segment) in the pio group compared with placebo, as well as the effect on late loss (P=0.026 for target lesion and P=0.007 for the total segment) and late loss index (P=0.028 for target lesion and P=0.017 for the total segment), became statistically significant. The effect on restenosis rate was still preserved when analysis included the 1 patient with stent thrombosis and total occlusion of the vessel (target lesion restenosis, 9.7% in the pio group versus 32.3% in the placebo group; P=0.03; total segment restenosis, 9.7% in the pio group versus 38.7% in the placebo group; P=0.01). There were no deaths or MIs during follow-up. Target-vessel revascularization due to restenosis was performed in 9 of 31 (29.0%) lesions in the placebo group compared with 2 of 29 (12.9%) lesions in the pio-treated group (P=0.02).
The present randomized, placebo-controlled, double-blind trial demonstrates that 6-month treatment with pio reduced neointima volume after coronary stent implantation in nondiabetic CAD patients. Previous studies have shown that TZD treatment reduces restenosis and neointima formation after coronary stenting in patients with type 2 diabetes mellitus.14–17 Although some of the studies analyzed whether this effect was dependent on the glucose-lowering properties of these agents,17 it remained unclear to what extend the favorable results on restenosis were due to improvement in glycemic control or other metabolic parameters. The data presented here suggest that pio directly influences neointima formation independent of its metabolic action. First, our study was conducted in nondiabetic subjects, and we did not find any changes in blood glucose, insulin, or HbA1c levels after 6 months of pio treatment. This is consistent with previous findings that have shown that TZD treatment of nondiabetic subjects does not have an effect on glucose metabolism.23 In addition, we did not find changes in total cholesterol, HDL cholesterol, or triglyceride levels, making it unlikely that the effect on neointima formation resulted from changes in the lipid profile. Still, we did not perform oral glucose tolerance tests in our patients and cannot exclude the possibility that some of the patients might have had impaired glucose tolerance, which may have been influenced by pio treatment. However, the lack of an effect on blood glucose, insulin, or HbA1c levels argues against a causal role of major metabolic effects for the reduction in neointima formation.
The primary end point of this study was the assessment of neointima volume after coronary stenting by IVUS. Both this primary end point and total plaque volume in the stented area and the proximal and distal adjacent segments were significantly lower in the pio-treated patient group compared with patients receiving placebo. In the placebo group, a significantly higher inflation pressure was applied, but in both groups, high-pressure inflation (>14 atm) was performed, and the balloon-to-artery ratio, considered to be an index of coronary injury, was not statistically different in the 2 groups. In addition, in both groups, inflation pressure were not significantly correlated with the primary end point of neointima volume, making the difference in inflation pressure an unlikely explanation for the results observed on neointima volume. Furthermore, there was a nonsignificant trend to higher insulin levels in the placebo group compared with pio-treated patients, potentially reflecting a more insulin-resistant state in the placebo group with an increased risk for the development of restenosis. Because there was a trend to a lower body mass index, a very reliable marker of insulin resistance, in the placebo group, the increased insulin values were most likely due to chance. Moreover, baseline insulin levels were not correlated with neointima volume in both groups. In addition, insulin values fell more in the placebo group than in the pio group during the 6 months of treatment, making it unlikely that the difference at baseline accounted for the effect on neointima formation.
The secondary end point was the mean diameter stenosis of the total segment after 6 months, as assessed by QCA. Our study revealed a significant reduction in mean diameter stenosis of the total segment as well as a significant reduction of in-stent restenosis by pio treatment. This effect was paralleled by a nonsignificant trend to increased minimal luminal diameter, as well as decreased late loss and late loss index in the pio group compared with placebo. Still, when adjusted GEE linear regression analysis was performed to account for repeated assessments within 1 patient, the effect on these parameters was statistically significant.
The data obtained herein are consistent with experimental data on the effect of TZDs on processes involved in neointima formation after coronary stenting, like VSMC migration and proliferation. TZDs are activators of the nuclear transcription factor peroxisome proliferator–activated receptor (PPAR)-γ and, as such, are regulators of gene expression in various cell types. Several groups including our own have demonstrated that PPAR-γ is expressed in VSMCs in vivo and in vitro and that activation of this receptor by TZDs limits both VSMC migration and proliferation.10–12 In addition, very recent experimental data suggest that TZD treatment increases the number of endothelial progenitor cells, a mechanism considered important for endothelialization and reduction of restenosis after coronary stenting.24 Moreover, animal data have shown that TZD treatment reduces intimal hyperplasia after vascular injury.13 Our study extends the knowledge of TZDs’ effects on restenosis by showing that TZD treatment limits neointima volume independent of its metabolic effect in a nondiabetic patient population. These results are in line with previous reports showing that TZDs exhibit direct antiinflammatory and antiatherogenic properties in the vasculature. As such, TZDs reduce serum levels of inflammatory biomarkers, like CRP, fibrinogen, or soluble E-selectin, and modulate endothelial function independent of their metabolic action.25–27 Because inflammatory processes in the vessel wall may also contribute to a reduction in neointima formation, antiinflammatory TZD effects may explain the results observed here. Still, in our study, we did not find significant changes in CRP, fibrinogen, TNF-α, or sCD40L levels after pio treatment compared with placebo. This is most likely due to the small sample size as well as multiple confounding factors, such as hospitalization and the intervention itself, all known to modulate serum levels of these markers.28 However, the lack of a significant effect on inflammatory biomarkers does not exclude the possibility that the antiinflammatory action of pio contributed to the reduction in neointima formation.
Major limitations of the present study are the small sample size in both groups as well as the lack of mechanistic insight of pio’s effect on neointima formation. Therefore, larger studies are needed to further elucidate the antirestenotic effect of TZDs. In addition, such studies should include clinical end points like target-vessel revascularization, which was high (29%) in the placebo group in our study. If larger clinical trials can confirm the beneficial effects of TZD treatment on neointima volume, thus translating into a reduced need for target-vessel revascularization, this treatment with an orally taken drug may be a promising tool to modulate restenosis after stenting. This may also be important with respect to drug-eluting stents, for which the restenosis rate is significantly lower compared with bare metal stents. However, in more complex lesions, the binary restenosis rate is &15% and is as high as 31% in small vessels, despite the use of drug-eluting stents, as shown in the TAXUS-V study.29 In these subjects, systemic therapy with TZDs may be combined with drug-eluting stents to further improve clinical outcomes. In addition, TZD treatment may be used in patients with a history of gastrointestinal or intracranial bleeding who are otherwise unsuitable candidates for stenting with drug-eluting stents because of prolonged combined antiplatelet therapy.
Taken together, our study suggests a direct effect of TZD treatment on neointima volume after coronary stent implantation in nondiabetic CAD patients, promoting the concept that PPAR-γ–activating TZDs, independent of their metabolic action, may exhibit direct protective effects in the vasculature. Still, larger clinical trials should replicate these findings and determine whether the effects of TZD treatment on neointima formation also translate into clinical benefits, such as reduction of target-vessel revascularization.
This work was supported by grants from the Deutsche Forschungsgemeinschaft (SFB 451, project B9) as well as the Else-Kröner Fresenius-Stiftung to Dr Nikolaus Marx. The Department of Internal Medicine II, University of Ulm, has received an unrestricted grant from Takeda Pharma. We also thank Jens Baumert, MS, Institute of Epidemiology, GSF Research Institute, Neuherberg, Germany, for expert statistical advice, as well as Gerlinde Trischler, Miriam Grüb, Helga Bach, and Renate Durst for excellent technical assistance.
Dr Marx has received research grants from, served on the speakers’ bureaus of, and/or served as a consultant to Takeda Pharma and GlaxoSmithKline. Dr Höher has received a research grant from Takeda Pharma.
↵*The first 3 authors contributed equally to this study and should be considered first authors.
Betriu A, Masotti M, Serra A, Alonso J, Fernandez-Aviles F, Gimeno F, Colman T, Zueco J, Delcan JL, Garcia E, Calabuig J. Randomized comparison of coronary stent implantation and balloon angioplasty in the treatment of de novo coronary artery lesions (START): a four-year follow-up. J Am Coll Cardiol. 1999; 34: 1498–1506.
Mudra H, di Mario C, de Jaegere P, Figulla HR, Macaya C, Zahn R, Wennerblom B, Rutsch W, Voudris V, Regar E, Henneke KH, Schachinger V, Zeiher A. Randomized comparison of coronary stent implantation under ultrasound or angiographic guidance to reduce stent restenosis (OPTICUS Study). Circulation. 2001; 104: 1343–1349.
Leon MB, Baim DS, Popma JJ, Gordon PC, Cutlip DE, Ho KK, Giambartolomei A, Diver DJ, Lasorda DM, Williams DO, Pocock SJ, Kuntz RE. A clinical trial comparing three antithrombotic-drug regimens after coronary-artery stenting: Stent Anticoagulation Restenosis Study Investigators. N Engl J Med. 1998; 339: 1665–1671.
O’Neill WW, Serruys P, Knudtson M, van Es GA, Timmis GC, van der Zwaan C, Kleiman J, Gong J, Roecker EB, Dreiling R, Alexander J, Anders R. Long-term treatment with a platelet glycoprotein-receptor antagonist after percutaneous coronary revascularization: EXCITE Trial Investigators: Evaluation of Oral Xemilofiban in Controlling Thrombotic Events. N Engl J Med. 2000; 342: 1316–1324.
Marx N, Duez H, Fruchart JC, Staels B. Peroxisome proliferator-activated receptors and atherogenesis: regulators of gene expression in vascular cells. Circ Res. 2004; 94: 1168–1178.
Law RE, Goetze S, Xi XP, Jackson S, Kawano Y, Demer L, Fishbein MC, Meehan WP, Hsueh WA. Expression and function of PPAR-γ in rat and human vascular smooth muscle cells. Circulation. 2000; 101: 1311–1318.
Marx N, Schönbeck U, Lazar MA, Libby P, Plutzky J. Peroxisome proliferator activated receptor-γ activators inhibit gene expression and migration in human vascular smooth muscle cells. Circ Res. 1998; 83: 1097–1103.
Takagi T, Akasaka T, Yamamuro A, Honda Y, Hozumi T, Morioka S, Yoshida K. Troglitazone reduces neointimal tissue proliferation after coronary stent implantation in patients with non-insulin dependent diabetes mellitus: a serial intravascular ultrasound study. J Am Coll Cardiol. 2000; 36: 1529–1535.
Takagi T, Yamamuro A, Tamita K, Yamabe K, Katayama M, Mizoguchi S, Ibuki M, Tani T, Tanabe K, Nagai K, Shiratori K, Morioka S, Yoshikawa J. Pioglitazone reduces neointimal tissue proliferation after coronary stent implantation in patients with type 2 diabetes mellitus: an intravascular ultrasound scanning study. Am Heart J. 2003; 146: E5.
Choi D, Kim SK, Choi SH, Ko YG, Ahn CW, Jang Y, Lim SK, Lee HC, Cha BS. Preventative effects of rosiglitazone on restenosis after coronary stent implantation in patients with type 2 diabetes. Diabetes Care. 2004; 27: 2654–2660.
Marx N, Froehlich J, Siam L, Ittner J, Wierse G, Schmidt A, Scharnagl H, Hombach V, Koenig W. The antidiabetic PPAR-γ activator rosiglitazone reduces MMP-9 serum levels in type-2 diabetic patients with coronary artery disease. Arterioscler Thromb Vasc Biol. 2003; 23: 282–288.
Hoher M, Wohrle J, Wohlfrom M, Kamenz J, Nusser T, Grebe OC, Hanke H, Kochs M, Reske SN, Hombach V, Kotzerke J. Intracoronary β-irradiation with a rhenium-188-filled balloon catheter: a randomized trial in patients with de novo and restenotic lesions. Circulation. 2003; 107: 3022–3027.
Wohrle J, Grebe OC, Nusser T, Al-Khayer E, Schaible S, Kochs M, Hombach V, Hoher M. Reduction of major adverse cardiac events with intracoronary compared with intravenous bolus application of abciximab in patients with acute myocardial infarction or unstable angina undergoing coronary angioplasty. Circulation. 2003; 107: 1840–1843.
Hoffmann R, Mintz GS. Coronary in-stent restenosis: predictors, treatment and prevention. Eur Heart J. 2000; 21: 1739–1749.
Wang CH, Ciliberti N, Li SH, Szmitko PE, Weisel RD, Fedak PW, Al-Omran M, Cherng WJ, Li RK, Stanford WL, Verma S. Rosiglitazone facilitates angiogenic progenitor cell differentiation toward endothelial lineage: a new paradigm in glitazone pleiotropy. Circulation. 2004; 109: 1392–1400.
Sidhu JS, Cowan D, Kaski JC. Effects of rosiglitazone on endothelial function in men with coronary artery disease without diabetes mellitus. Am J Cardiol. 2004; 64: 151–156.
Haffner SM, Greenberg AS, Weston WM, Chen H, Williams K, Freed MI. Effect of rosiglitazone treatment on nontraditional markers of cardiovascular disease in patients with type 2 diabetes mellitus. Circulation. 2002; 106: 679–684.
Liuzzo G, Buffon A, Biasucci LM, Gallimore JR, Caligiuri G, Vitelli A, Altamura S, Ciliberto G, Rebuzzi AG, Crea F, Pepys MB, Maseri A. Enhanced inflammatory response to coronary angioplasty in patients with severe unstable angina. Circulation. 1998; 98: 2370–2376.
Stone GW, Ellis SG, O’Shaughnessy CD, Statler L, Martin S, McGarry T, Turco M, Kereiakes DJ, Jacobs WC, Russell ME. A prospective, multicenter randomized trial of the paclitaxel-eluting stent versus vascular brachytherapy in patients with coronary in-stent restenosis: the TAXUS-V ISR Trial. J Am Coll Cardiol. 2005; 45 (suppl A): 83A.