Effects of AGI-1067 and Probucol After Percutaneous Coronary Interventions
Background— AGI-1067, a metabolically stable modification of probucol, is an equipotent antioxidant to probucol but is pharmacologically distinct. In a multicenter trial, we studied whether AGI-1067 reduces restenosis assessed by intravascular ultrasound (IVUS) after percutaneous coronary intervention (PCI) compared with placebo and probucol used as a positive control.
Methods and Results— Two weeks before PCI, 305 patients were randomly assigned to 1 of 5 treatment groups: placebo, probucol 500 mg BID, or AGI-1067 70, 140, or 280 mg once daily. Patients were treated for 2 weeks before and 4 weeks after PCI. Baseline and 6-month follow-up IVUS were interpreted by a blinded core laboratory. Stents were used in 85% of patients. Luminal area at the PCI site at follow-up was 2.66±1.58 mm2 for placebo, 3.69±2.69 mm2 for probucol, 2.75±1.76 mm2 for AGI-1067 70 mg, 3.17±2.26 mm2 for AGI-1067 140 mg, and 3.36±2.12 mm2 for AGI-1067 280 mg (P=0.02 for the dose-response relationship; P≤0.05 for AGI-1067 280 mg and probucol versus placebo). There was a mean narrowing of 5.3 mm3 of reference segment lumen in the placebo group and an enlargement in the AGI-1067 140- and 280-mg groups at follow-up (P=0.05 for 140 mg). An increase in QTc interval >60 ms occurred in 4.8% of placebo patients, 17.4% of probucol patients, and 4.8%, 2.4%, and 2.5% of patients in the AGI-1067 groups (P=0.02).
Conclusions— AGI-1067 and probucol reduce restenosis after PCI. In contrast to probucol, AGI-1067 did not cause prolongation of the QTc interval and improved lumen dimensions of reference segments, suggestive of a direct effect on atherosclerosis.
Received August 13, 2002; revision received October 18, 2002; accepted October 20, 2002.
Systemic administration of a powerful antioxidant has the potential to prevent both restenosis and atherosclerosis progression.1 Countering oxidative stress may limit lipoprotein oxidation, endothelial dysfunction, and vascular inflammation.2 Antioxidant vitamins have failed in large randomized clinical trials, but there are several potential problems associated with their use, including their weak antioxidant properties and potential pro-oxidant effects.1,3 Probucol has been shown to reduce restenosis in several clinical trials,4–10 but its effect on HDL cholesterol and particularly the QT interval have limited its long-term use.11 In contrast, AGI-1067, the monosuccinic acid ester of probucol, is a metabolically stable modification of probucol that retains antioxidant properties equipotent to those of probucol (assessed with a leukomethylene blue assay)12 but did not cause HDL cholesterol decrease or QT-interval prolongation in animal studies. Furthermore, AGI-1067 has been effective at preventing atherosclerosis in all tested animal models. Accordingly, we conducted the Canadian Antioxidant Restenosis Trial (CART-1) to determine whether AGI-1067 reduces restenosis as assessed by intravascular ultrasound (IVUS) when administered for 2 weeks before and 4 weeks after percutaneous coronary intervention (PCI). Other study objectives included the assessment of AGI-1067 effects on the QT interval and on coronary segments not subjected to intervention.
Study Design and Population
CART-1 was a multicenter, double-blind, placebo-controlled, randomized trial with 5 groups. The protocol was approved by institutional review boards. Patients referred for elective PCI were evaluated ≥14 days before their scheduled procedures. Eligible patients were asked to provide written informed consent and underwent medical history, physical examination, electrocardiography, hematology, and clinical biochemistry. Patients were eligible if they were scheduled to undergo PCI with or without stenting on ≥1 native coronary artery and had ≥1 de novo target lesion with luminal narrowing ≥50%. We excluded subjects who had severe liver disease or serum creatinine ≥200 μmol/L, myocardial infarction ≤7 days ago, left main stenosis >50%, or ejection fraction <30%; had PCI for another lesion in the preceding 6 months; were being treated for a restenotic lesion; had scheduled atherectomy, brachytherapy, or PCI of bypass graft; required therapy with a class Ia/III antiarrhythmic agent; or had significant QTc lengthening (>0.470 seconds) before randomization.
Randomization and Drug Regimen
Patients were randomly assigned to receive AGI-1067 70, 140, or 280 mg once daily, probucol 500 mg twice daily, or placebo beginning 14 days before scheduled PCI. AGI-1067 or matched placebo was administered as 4 capsules given once daily and probucol or matching placebo as 2 tablets twice per day. Patients were thus treated with both capsules of AGI-1067 or matched AGI-1067 placebo once daily and tablets of probucol or matched probucol placebo twice daily. All patients also received an extra dose of AGI-1067 280 mg or probucol 1000 mg or matched placebos on the evening before PCI, according to random treatment assignment. After PCI, all patients were maintained on their assigned study regimen for 4 additional weeks.
PCI and Follow-Up Evaluation
PCI with or without stent placement and post-PCI management were performed according to current clinical practice. ECGs were obtained before PCI, immediately thereafter, and the morning after PCI. Creatine kinase, creatine kinase-MB fraction, and troponin I were measured on the evening after PCI and the following morning. Patients were discharged after PCI with 4 weeks of the study medication. Aspirin 325 mg/d was given from the time of recruitment and for the entire study duration. All patients treated with stents also received clopidogrel 75 mg/d for 30 days after PCI. Patients returned at 1 month for clinical evaluation and drug accountability and at 3 and 6 months. Patients were assessed for ischemic symptoms and adverse events, whether or not they were related to the study medication or PCI procedure. Blood chemistry values assessed at baseline were measured again at PCI discharge and at follow-up visits. Patients were readmitted for follow-up catheterization and IVUS 5 to 7 months after PCI. Those in whom catheterization was performed for clinical reasons before the fifth month returned for repeat IVUS examination at 5 to 7 months if no definite restenosis was present on ≥1 dilated site.
IVUS examinations were performed with 30-MHz, 3.2F ultrasound catheters (CVIS). IVUS studies were recorded before PCI whenever possible and were always performed after PCI (after final balloon inflation) and at follow-up (before any subsequent intervention). IVUS was always preceded by the intracoronary injection of nitroglycerin (0.3 mg). The IVUS catheter was advanced distal to the dilated site to a recognizable landmark that was noted and used for follow-up examination. Automated pullbacks were performed at a speed of 0.5 mm/s up to the guiding catheter, and the ultrasound images were recorded onto super VHS videotape for offline analysis, with a detailed audio commentary.
All IVUS images were interpreted at the Montreal Core Laboratory by experienced technicians supervised by a cardiologist, all of them blinded to treatment assignment. The preintervention, post-PCI, and follow-up studies were analyzed side by side. Care was taken to ensure that the identical anatomic slice was measured in serial IVUS studies. The use of a known pullback speed and reproducible landmarks allowed matching of the anatomic slice in serial studies, in combination with frame-by-frame review of the images. The cross section selected for serial 2D analysis was the one at the PCI site with the smallest lumen area (LA) at follow-up. The corresponding slice was then identified on the post-PCI and pre-PCI studies. Quantitative analysis was performed on digitized images with custom-developed software (INDEC). In stented lesions, analysis consisted of measurements of LA and stent area; neointimal area was defined as stent area minus LA. In nonstented lesions, LA and the area within the external elastic membrane (EEM) were measured, and plaque area was calculated as the difference between EEM and LA.
3D reconstructions of both the PCI site and reference segment were performed. For nonstented lesions, lumen and EEM borders at the PCI site were manually traced on every digitized cross section. For stented segments, lumen and stent borders were traced manually on slices at every 0.5 mm of axial length within the stent. Lumen, stent, and neointimal (or lumen/EEM/plaque for nonstented lesions) volumes were computed by summing areas of each cross section multiplied by the distance between neighboring slices. For nonstented lesions, a 15-mm segment centered on the target lesion was evaluated. To account for variable stent lengths, the volumetric results were divided by the length of the stent analyzed and multiplied by 15 to obtain volumes indexed for a 15-mm segment. 3D reconstruction of a 5-mm reference segment was also performed. This segment was centered on the most normal-looking cross section between 5 and 12 mm proximal to the PCI site but distal to any significant side branch. If the lesion was ostial in location, then a distal reference was used. For this reference segment, the lumen and EEM borders were traced on every digitized frame.
Quantitative Coronary Angiography
Control angiography after PCI and at follow-up was preceded by the intracoronary administration of nitroglycerin (0.3 mg). Quantitative analysis was performed to determine dichotomous restenosis rates.4 A PCI segment was defined as restenotic if diameter stenosis was ≥50% at follow-up, with an increase of ≥15% in the degree of stenosis compared with the post-PCI angiogram.
Analysis of ECGs
The ECGs were read in a core laboratory (EPIC) blinded to treatments. The QT interval was measured and corrected for heart rate (QTc) by dividing the measured QT interval by the square root of the R-R interval.13
We calculated that 290 patients were needed to be able to detect a 25% increase in the minimal LA on IVUS 6 months after PCI, with a power of 0.80 and a 2-tailed significance level of 0.05,6 which allowed a dropout rate of 15%. The follow-up minimal LA on IVUS, ie, the primary efficacy end point, was analyzed with a 2-way ANOVA (treatment, center). Interaction between treatment and center was tested and removed from the model if not significant at the 0.05 level. Uncorrected contrasts between groups followed where appropriate. Other continuous end points were analyzed similarly. Logistic regression with treatment and center as main effects was used for dichotomous end points. Measurements are reported as mean±SD. A probability value <0.05 was considered statistically significant.
We randomized 305 patients, who were distributed among the 5 groups as follows: 59, 64, and 61 patients received 70, 140, and 280 mg daily of AGI-1067; 60 received probucol; and 61 received placebo. There were no significant differences at baseline among groups (Table 1).
Stents were used in 85.1% of patients: 84.7% in the placebo group, 81.7% in the probucol group, and 87.9%, 86.6%, and 84.7% in the 3 AGI-1067 groups (P=NS). Overall, stents were inserted in 81.3% of PCI segments. IIb/IIIa receptor inhibitors were used in 75 patients (placebo 16, probucol 13; 16, 13, and 17 for the AGI-1067 groups). There were 13 PCI failures: 7 in the placebo group, 0 in the probucol group, and 2 patients per AGI-1067 group (P=0.03). Seventeen patients took <80% of dispensed capsules/tablets (2, 5, and 4 in the AGI-1067 groups, 2 with probucol, and 4 with placebo).
Analysis of PCI Segments
LA at the target lesion on IVUS was similar among study groups before PCI (P=0.79; Table 2). There was a significant AGI-1067 dose response for LA after PCI at the site of intervention (P=0.04). LA at follow-up was 2.66±1.58 mm2 for placebo, 3.69±2.69 mm2 for probucol, 2.75±1.76 mm2 for AGI-1067 70 mg, 3.17±2.26 mm2 for AGI-1067 140 mg, and 3.36±2.12 mm2 for AGI-1067 280 mg (P=0.046 for AGI-1067 280 mg versus placebo; P=0.01 for probucol versus placebo). There was a significant dose-response relationship among the AGI-1067 groups (P=0.02).
Volumetric assessment over the entire PCI segment supported the 2D results (Table 3; for lumen volume at follow-up: P=0.02 for AGI-1067 dose response, P=0.035 for AGI-1067 280 mg versus placebo, P=0.02 for probucol versus placebo). Among stented segments, lumen volume at follow-up was 64.0±25.2 mm3 for placebo, 80.0±43.1 mm3 for probucol, and 64.4±28.7, 72.9±33.6, and 77.4±33.5 mm3 for the AGI-1067 70-, 140-, and 280-mg groups, respectively (P=0.02 for AGI-1067 dose response; P=0.048 for AGI-1067 280 mg versus placebo; P=0.04 for probucol versus placebo). Restenosis rates were 37.5% in the placebo group, 25.5% in the probucol group, and 26% in the 3 AGI-1067 groups (P=0.85 for AGI-1067 versus probucol, P=0.09 for AGI-1067 versus placebo, P=0.24 for probucol versus placebo). When compliant patients were analyzed, restenosis rates were 37.7%, 25.9%, and 23.6% in the placebo, probucol, and AGI-1067 groups (P=0.03 for AGI-1067 versus placebo).
Analysis of Reference Segments
Lumen volumes of the reference segments were similar in all study arms at baseline at ≈110 mm3 (P=0.97; Table 4). The mean change in lumen volume in the reference segments from baseline to follow-up was −5.3 mm3 with placebo, −0.2 mm3 for probucol, and −2.4 mm3 with AGI-1067 70 mg but 3.5 mm3 with the 140-mg dose and 1.8 mm3 in the 280-mg group (P=0.05 for AGI-1067 140 mg versus placebo; P=0.077 for dose-response relationship).
Major Clinical End Points
There were no deaths and 11 myocardial infarctions (2 with placebo, 3 with probucol, and 1, 3, and 2 in the AGI-1067 groups) during the study. The rates of target lesion revascularization were 18.3% for placebo, 18.0% for probucol, and 16.9%, 14.1%, and 13.1% for the 3 AGI-1067 groups (P=NS). When only patients with stented lesions were considered, the rates were 17.4% for placebo, 15.6% for probucol, and 10.6%, 13.7%, and 6.2% for the AGI-1067 groups (P=0.09 for AGI-1067 280 mg versus placebo).
Safety and Tolerability
The QT intervals were similar among the treatment groups at baseline (Table 5). An increase in the QTc interval >60 ms was observed at least once during the study in 4.8% of placebo patients, 17.4% of patients treated with probucol, and 4.8%, 2.4%, and 2.5% of patients in the AGI-1067 groups (P=0.02). The corresponding increases in women were 14.3% for placebo, 40% for probucol, and 0% for AGI-1067 (P=0.12). Probucol reduced HDL cholesterol levels by 35.1% at the 1-month visit, whereas reductions of 0.2%, 4.4%, 9.0%, and 18.7% were observed in the placebo and the 3 AGI-1067 groups, respectively (P<0.01 versus placebo for probucol, AGI-1067 140 mg, and AGI-1067 280 mg). There were no significant changes of LDL levels in actively treated arms versus placebo during the study. Diarrhea was reported in 4.9% of placebo-treated patients, 15.0% of patients treated with probucol, and 8.5%, 7.8%, and 18.0% of patients in the AGI-1067 groups (P=0.11). Gastrointestinal disturbances occurred largely with therapy initiation, were generally self-limited, and resulted in 4 patients withdrawing from the study (1 with placebo, 1 in the AGI-1067 140-mg group, and 2 in the AGI-1067 280-mg group).
Previous trials have shown that probucol is effective in reducing restenosis after balloon angioplasty but could not address its value after stent deployment.4–10 Results from CART-1 demonstrate that both AGI-1067 and probucol improve lumen dimensions at the site of stent placement 6 months after the procedure. The benefit of AGI-1067 and probucol was present immediately after PCI. The cellular mechanisms for this finding have not yet been identified, but the larger lumen dimensions after PCI raise the possibility that countering oxidative stress may rapidly improve endothelial function and reduce vasomotor tone,14,15 thereby allowing for optimal response to balloon dilatation and stenting. Balloon injury indeed leads to the immediate release of reactive oxygen species,16,17 which may be inhibited by powerful antioxidants. Alternatively, changes in plaque content caused by these agents may have favored axial redistribution of plaque away from the dilated site toward the noninjured regions during PCI.18 The improvement in LA immediately after the procedure in the actively treated arms compared with placebo was more marked after stent deployment than in previous studies in which probucol was used before balloon angioplasty.4 This difference may be due to the higher inflation pressures used during stenting procedures. This early benefit correlated with reduced PCI failure rates, an effect that may become particularly relevant in patients with small coronary arteries or complex lesions. The early effects on LA persisted at follow-up, despite the fact that study medications had been stopped for 5 months.
Several observations support that this early benefit represents a true biological effect. First, the study groups were well balanced for baseline clinical characteristics. Second, IVUS showed that minimal LAs were distributed well among groups before PCI. Third, dimensions of reference segments at the time of PCI were also similar among study arms. The effect of 4 weeks of therapy after PCI on neointimal formation was not significant in stents. Because oxidative stress and inflammation may persist for >4 weeks after stenting, antioxidant therapy for the entire period of risk4,6,8 may lead to combined early and late benefits and better results.
AGI-1067 also had favorable effects on luminal dimensions of reference segments, which may represent the first clinical evidence of vascular protection for this new class of therapeutics that combine antioxidant properties with unique antiatherosclerotic activities (intracellular redox-signaling blockade and vascular cell adhesion molecule-1 inhibition).12 This effect was due to inhibition of negative remodeling and reduction of plaque burden, processes that, when unchecked, may contribute to progressive coronary artery narrowing. These clinical results are supported by demonstration of atherosclerosis prevention by AGI-1067 in apolipoprotein-E knockout and LDL receptor–deficient mice and the hyperlipidemic primate.12 This has potentially important implications, because PCI and local approaches to prevent restenosis, such as drug-coated stents, treat only short segments of the coronary circulation and are not expected to prevent atherosclerosis progression, myocardial infarction, and cardiovascular death. The effects of more prolonged treatment with AGI-1067 on neointimal formation and restenosis, as well as on the entire coronary tree and clinical outcomes, are being investigated in CART-2.
The rate and severity of QTc prolongation with probucol observed in CART-1 and other trials remain a long-term safety concern. In contrast, results with AGI-1067 suggest that a new class of vascular protectants may be efficacious while decreasing the risk of ventricular arrhythmias. Indeed, AGI-1067 was not different from placebo with respect to its effect on the QTc-interval in the present study. Probucol resulted in a mean lowering of HDL cholesterol levels of 35% in CART-1. In spite of significant HDL lowering in previous studies, prolonged therapy with probucol improved endothelial function and prevented atherosclerosis progression.14,19,20 AGI-1067 resulted in mean reductions in HDL cholesterol levels of 4.4% to 18.7%, which taken together with its positive effects on the reference segments reinforces the imperfect relationship between HDL changes and atherosclerosis.21 Furthermore, the other vascular protectant effects of AGI-1067 may outweigh a small reduction in HDL cholesterol, should this effect persist with long-term dosing. The relationship between lipid and atherosclerosis changes with AGI-1067 will be assessed in the ongoing CART-2 trial.
CART-1 was sponsored by AtheroGenics Inc.
Dr Glass holds AtheroGenics stock.
Tardif JC, Grégoire J, Bourassa MG. Antioxidants and restenosis: human studies.In: Tardif JC, Bourassa MG, eds. Antioxidants and Cardiovascular Disease. Boston, Mass: Kluwer Academic Publishers; 2000: 175–194.
Kunsch C, Medford RM. Oxidative stress as a regulator of gene expression in the vasculature. Circ Res. 1999; 85: 753–766.
Bowry VW, Ingold KU, Stocker R. Vitamin E: when and how this antioxidant becomes a pro-oxidant. Biochem J. 1992; 288: 341–344.
Rodés J, Tardif JC, Lespérance J, et al. Prevention of restenosis after angioplasty in small coronary arteries with probucol. Circulation. 1998; 97: 429–436.
Côté G, Tardif JC, Lespérance J, et al. Effects of probucol on vascular remodeling after coronary angioplasty. Circulation. 1999; 99: 30–35.
Kuzuya M, Naito M, Funaki C, et al. Probucol prevents oxidative injury to endothelial cells. J Lipid Res. 1991; 32: 197–204.
McLean LR, Thomas CE, Weintraub B, et al. Modulation of the physical state of cellular cholesteryl esters by probucol. J Biol Chem. 1992; 267: 12291–12298.
Kozarsky KF, Donahee MH, Glick JM, et al. Gene transfer and hepatic overexpression of the HDL receptor SR-B1 reduces atherosclerosis in the cholesterol-fed LDL receptor-deficient mouse. Arterioscler Thromb Vasc Biol. 2000; 20: 721–727.