Effect of rosiglitazone on restenosis after coronary stenting in patients with type 2 diabetes☆
Article Outline
Abstract
Background
Thiazolidinediones have been shown to have an antiproliferative vascular effect in experimental models. We sought to study the effect of rosiglitazone on in-stent restenosis in patients with established type 2 diabetes.
Methods
Patients with treated type 2 diabetes (mean duration 5.5 ± 7.5 years) referred for coronary stenting were randomized in a double-blind fashion to receive oral rosiglitazone or placebo for 6 months. Quantitative coronary angiography and intravascular ultrasound data were obtained at baseline and follow-up. Plasma plasminogen activator inhibitor-1 levels were prospectively measured.
Results
Sixteen patients were enrolled. There were no significant differences in follow-up in-stent luminal diameter stenosis measured by quantitative coronary angiography or in-stent luminal area stenosis and neointimal volume index obtained by intravascular ultrasound, nor were there any differences in plasma plasminogen activator inhibitor-1 levels after long-term use despite improvement in diabetes control and insulin sensitivity.
Conclusions
Rosiglitazone, given at the time of stent implantation in treated diabetics, did not reduce in-stent restenosis in this small series. The vascular biological effects of this agent await further clarification in humans and evaluation in larger clinical trials.
About 17 million Americans are affected by diabetes mellitus.1 Of these, 90% have type 2 diabetes, a disorder associated with accelerated atherosclerosis and a 4- to 6-fold increase in cardiovascular morbidity and mortality. Furthermore, compared with nondiabetics, the long-term benefit from revascularization in diabetics is attenuated.2, 3
Diabetes has been repeatedly shown to be a strong clinical predictor of restenosis after percutaneous coronary intervention.4, 5, 6, 7, 8, 9 Even with the widespread use of coronary stenting, diabetes remains the single most significant clinical risk factor for in-stent restenosis and repeat revascularization.10, 11, 12 The insulin resistance syndrome and its related pathophysiologic alterations appear to be important modulators of the biological events that promote abnormal vessel response to injury.13, 14, 15, 16, 17, 18, 19, 20 The thiazolidinediones (TZDs) substantially improve target organ sensitivity to insulin. In addition, these drugs appear to have significant direct vascular effects. As potent agonists of the nuclear receptor peroxisome proliferator activated receptor-γ (PPAR-γ), they have been shown in animal and in vitro studies to inhibit growth factor–stimulated vascular smooth muscle cell proliferation and migration.21 In addition, troglitazone may reduce the abnormally elevated plasminogen activator inhibitor-1 (PAI-1) levels in the plasma and vascular tissue of type 2 diabetics.22 Increased PAI-1 expression has been linked to accelerated atherosclerosis23 as well as restenosis after balloon vascular injury.24 A few small studies on human diabetic subjects have demonstrated a favorable effect of troglitazone on the reduction of carotid artery intima-medial thickness25 and the amount of neointimal hyperplasia after coronary stent implantation.26 Such data in the case of rosiglitazone are unavailable. In this study we sought to examine the effect of rosiglitazone, one of the TZDs currently in clinical use, on in-stent restenosis and PAI-1 plasma levels in type-2 diabetic patients.
Methods
Study patients
Written informed consent was obtained from all study participants and the study was approved by the Institutional Review Board at the University of Texas Medical Branch at Galveston. Inclusion criteria mandated an established diagnosis of type 2 diabetes according to the revised American Diabetes Association criteria27 in patients undergoing coronary angiography. Exclusion criteria were patients already on a TZD, abnormal baseline liver function studies or chronic liver disease, ejection fraction<30% or heart failure, serum creatinine >2.5mg/dL, life expectancy of <12 months, ostial or bifurcation lesions, total occlusions, or lesions with reference vessel diameter <2.5 cm. Although metformin is also an “insulin sensitizer,” it acts by a different mechanism than TZDs and does not have known actions on the vascular smooth muscle cell.28, 29 As such, patients on this agent were included in the study.
After stent implantation, patients were randomly assigned, in a double-blinded fashion, to receive long-term oral rosiglitazone or placebo. Treatment was started within 6 hours of stenting. Patients were followed at 1-, 3-, and 6-month intervals. At 6 to 9 months from enrollment, patients underwent follow-up coronary angiography and intravascular ultrasound (IVUS). Blood samples were obtained at baseline and at 3 and 6 months after randomization.
Stent implantation
The choice of stent was left to the discretion of the operator. All patients received a loading dose of clopidogrel (300 mg) after stent implantation and 75 mg daily for 4 weeks thereafter. All patients were placed on aspirin indefinitely.
Quantitative angiographic analysis
Quantitative coronary angiography (QCA) was performed in a blinded manner for baseline and follow-up angiograms using Camtronics QCA System (Hartland, Wisc). A contrast-filled catheter was used for calibration. Minimal lumen diameter, reference vessel diameter, and percent diameter stenosis were measured. Measurements were obtained from multiple projections, and the results of the “worst” views were recorded. Angiographic restenosis was defined as >50% stenosis by QCA at follow-up study.
Intravascular ultrasound imaging
The IVUS imaging was performed after the index interventional procedure (following the final balloon inflation) and at follow-up. After administration of 200 μg of intracoronary nitroglycerin, the 20 MHz IVUS catheter (Boston Scientific, Maple Grove, Minn) was advanced in the coronary artery distal to the target lesion. Continuous imaging was performed as the IVUS catheter was automatically withdrawn at 0.5 mm/sec using a motorized pullback device. The images were recorded on a VHS videotape for off-line analysis. IVUS measurements were performed by an independent core laboratory at Stanford University.
Quantitative ivus measurement
Quantitative IVUS measurements were performed at a core laboratory blinded to treatment assignment. Computer-assisted planimetry was utilized. Standard IVUS measurements of diameters, areas, and volumes were obtained using previously validated methods.30, 31, 32 When IVUS was not performed at baseline the stent and lumen areas were considered equal (neointimal area imputed to be zero).
Use of rosiglitazone
Patients were started on the study drug (rosiglitazone or placebo) within 6 hours of stent implantation. A starting oral dose of 4 mg daily was used. The dose was increased to 8 mg daily at the 1-month visit and was continued for the entire period of treatment (6 months). Compliance to the treatment was verified at each visit and any adverse effects were noted. Patients' self-monitoring of blood glucose was encouraged. Occasionally, adjustment of other hypoglycemic medications was needed at follow-up visits as necessary to optimize glycemic control.
Biochemical markers
Blood specimens for measurement of plasma plasminogen activator inhibitor PAI-1, fasting insulin, and hemoglobin A1c (HgbA1c) were drawn at baseline and at 3- and 6-months. Care was taken to draw the fasting specimens at similar times in the mornings to avoid the effect of diurnal variations in PAI-1 levels. HgbA1c and insulin were measured using standard clinical laboratory methods. Blood for PAI-1 level was collected in citrated tubes and centrifuged immediately. Plasma was stored at −80°C for future analysis. The specimens were then thawed and PAI-1 levels were determined using standard ELISA assay (Biopool Inc, Ventura, Calif). Additionally, liver enzymes were assayed at baseline, 1-month, 3-month, and 6-month time points for safety monitoring.
Statistical analysis
The important continuous variables generated in each of the 2 groups are expressed as mean ± SD. Primary end points were IVUS-measured maximal in-stent area stenosis and volume stenosis index at follow-up. Secondary outcomes included QCA-determined percent diameter stenosis at follow-up, change in PAI-1 plasma level from baseline, and major adverse cardiac events (death, myocardial infarction, or target vessel revascularization). A 2-sample t test was used to assess the extent to which the changes in the treatment group differ from those in the control group. A P value of < 0.05 was considered significant. All analyses are by intention-to-treat method.
Data from a recent study examining a similar patient population33 showed that control subjects had a mean in-stent maximal area stenosis of 49% ± 15% compared to subjects treated with troglitazone (mean 28% ± 12%). Our initial goal was to enroll 12 patients in each group based on the power analysis of the preliminary results of that study, which showed that a sample size of 12 patients in each group would have 80% power to detect a difference in the means of an in-stent maximal area stenosis of 21%, assuming that the common SD is 17 using a 2-group t test with a .05 2-sided significance level.
Results
Patient population
A total of 16 type II diabetic patients (10 females and 6 males) were enrolled between June 2000 and November 2001. A total of 17 lesions were treated, with 18 stents used (10 Guidant Multilink, Santa Clara, Calif, 3 Medtronic S670, Minneapolis, Minn, and 5 Boston Scientific NIR). Compliance to study drug was monitored during follow up visits at 1, 3, and 6 months. Follow-up coronary angiography was obtained in 11 patients. The study was terminated short of target recruitment due to slow enrollment. Selected demographic and clinical characteristics are shown in Table I. There were no statistically significant baseline differences between the groups.
Table I. Baseline patient and treated lesion characteristics
| Rosiglitazone (n = 8) | Placebo (n = 8) | |
|---|---|---|
| Mean age (range) y | 53.5 (40–64) | 57.3 (38–79) |
| Sex M/F | 1/7 | 5/3 |
| Mean duration of diabetes (y) | 7.75 | 3.25 |
| Hypertension | 5 | 6 |
| Smoking | 6 | 2 |
| Insulin dependence | 3 | 2 |
| Clinical presentation | ||
| Angina | 7 | 4 |
| Myocardial infarction | 1 | 4 |
| LAD lesions | 5 | 1 |
| RCA lesions | 2 | 7 |
| LCx lesions | 1 | 1 |
| Mean stent length (range) mm | 17.6 (12–28) | 19.3 (12–38) |
| Mean stent diameter (range) mm | 3.3 (2.5–4.0) | 3.1 (2.5–3.5) |
Drug safety
The study drug was well tolerated and no adverse reactions were reported during follow-up. There were no increases in liver enzymes detected in any patient at 1, 3, and 6 months.
Coronary interventions
Fourteen lesions were pre-dilated and 3 were directly stented. Procedural success was obtained in all lesions. Vessels treated and stent sizes are shown in Table I.
Quantitative coronary angiography
QCA was performed at baseline in all patients and at follow-up for 11 patients. The other 5 patients, despite providing informed consent for follow-up angiography, were unwilling to be restudied. Results of QCA are shown in Table II. At baseline, there were no significant differences in treated lesion lengths, diameter stenoses, or reference vessel diameter between the 2 groups. The immediate results were similar. At follow-up, the 2 groups had similar minimal lumen diameter (1.4 ± 0.96 mm vs 1.5 ± 1.18 mm, P = .82) and percent diameter stenosis (57.7 ± 28.3 mm vs 55.4 ± 26.1, P = .89).
Table II. Results of baseline and follow-up QCA analysis
| Rosiglitazone | Placebo | |
|---|---|---|
| Baseline | ||
| RVD (mm) | 3.15 ± 0.8 | 3.19 ± 0.4 |
| MLD (mm) | 0.50 ± 0.3 | 0.65 ± 0.3 |
| DS (%) | 83.6 ± 8.2 | 80.2 ± 10.2 |
| Poststenting MLD | 3.3 ± 0.8 | 3.1 ± 0.5 |
| Follow-up | ||
| RVD (mm) | 3.05 ± 0.8 | 3.18 ± 0.4 |
| MLD (mm) | 1.5 ± 1.1 | 1.4 ± 0.9 |
| DS (%) | 55.4 ± 26.1 | 57.7 ± 28.3 |
IVUS measurements
Of the 11 patients who underwent follow-up angiography IVUS data were obtained from 9 patients; it was not performed on the other 2 patients because of inability to cross the lesions with IVUS catheter due to severe restenosis. At baseline, postintervention IVUS measurements showed no significant differences in average stent area (SA) and lumen area (LA). At follow-up, the 2 groups had similar average SA, LA, neo-intimal area (NIA), and SA stenosis. Furthermore, volumetric analysis showed no significant differences in baseline plaque volume, follow-up neo-intimal volume (NIV), and follow-up volume stenosis index (Table III).
Table III. Results of area and volumetric analyses by IVUS
| Rosiglitazone | Placebo | P | |
|---|---|---|---|
| Postintervention average SA mm2 | 10.68 ± 4.0 | 7.8 ± 2.6 | .26 |
| Postintervention average LA mm2 | 10.68 ± 4.0 | 7.8 ± 2.6 | .26 |
| FU average SA (mm2) | 10.61 ± 3.9 | 7.7 ± 2.4 | .27 |
| FU average LA (mm2) | 6.8 ± 5.5 | 4.3 ± 2.1 | .46 |
| FU average NIA (mm2) | 3.1 ± 2.2 | 3.6 ± 3.7 | .86 |
| FU average SA stenosis (%) | 37 ± 30 | 42 ± 28 | .29 |
| FU minimal LA (mm2) | 5.2 ± 4.7 | 2.6 ± 1.9 | .26 |
| FU maximal SA stenosis (%) | 53 ± 27 | 64 ± 24 | .75 |
| Baseline plaque volume (mm3) | 243 ± 122 | 342 ± 233 | .68 |
| FU stent NIV (mm3) | 48 ± 27 | 128 ± 110 | .48 |
| FU volume stenosis index (%) | 24 ± 14 | 41 ± 28 | .28 |
Biochemical markers
The results of HgbA1c, fasting insulin levels, and PAI-1 plasma levels are shown in Table IV.
Table IV. Effect of rosiglitazone on biochemical parameters
| Rosiglitazone | Placebo | |||
|---|---|---|---|---|
| Baseline | FU | Baseline | FU | |
| Hgb A1c (%) | 10.3 ± 3.2 | 7.4 ± 2.1 | 8.7 ± 1.9 | 8.2 ± 1.3 |
| Insulin (μU/mL) | 15.7 ± 8.1 | 11.5 ± 7.8 | 16.3 ± 7.2 | 19.8 ± 12.5 |
| PAI-1 (ng/mL) | 62.3 ± 23.9 | 55.5 ± 23.5 | 80.9 ± 23.6 | 60.9 ± 24.5 |
| HgbA1c % change | −24.1 ± 17.2* | −1.1 ± 13.3 | ||
| Insulin % change | −30.1 ± 45.3† | 36.6 ± 56.5 | ||
| PAI-1 % change | −7.23 ± 52.1 | −17.26 ± 45.5 | ||
∗ P = .025 versus placebo. |
† P = .04 versus placebo. |
Compared with placebo group, patients randomized to rosiglitazone had a statistically significant percentage reduction of their HgbA1c levels at follow-up relative to baseline (24.1 ± 17.2% vs 1.1 ± 13.3%, P = .025). This was accompanied by a 30.1% ± 45.3% reduction in fasting plasma insulin levels in the active drug group (vs an increase of 37% in the placebo group, P = .04). These findings are consistent with the insulin-sensitizing effect of rosiglitazone. On the other hand, there were no statistically significant differences in plasma PAI-1 levels between baseline and follow-up time points in either group nor in the percentage change in PAI-1 levels between the 2 groups (P = 0.7).
Major adverse cardiac events
Over the 6-month follow-up period there were 6 major adverse cardiac events. Target vessel revascularization was needed in 5 patients (3 placebo, 2 rosiglitazone). One patient in the rosiglitazone group died suddenly 3 days after enrollment. There was no difference in major cardiac events between the groups.
Discussion
This randomized, placebo-controlled study showed that rosiglitazone improved diabetes control, as reflected in lower long-term HgbA1C levels and decreased fasting insulin levels compared with placebo treatment. However, at this therapeutic dosage, oral rosiglitazone started at the time of coronary stent implantation in patients with type 2 diabetes did not appear to reduce the burden of in-stent restenosis. These data are in contrast with the study by Takagi et al, who showed that administration of troglitazone to diabetic patients reduces neointimal proliferation after coronary stenting.26 There are several potential explanations for these conflicting findings. First, the sample size of our study is small. However, the use of sensitive IVUS measures of restenosis in our study, including volumetric analysis, failed to show significant differences or noteworthy trends between the groups. Using volumetric IVUS end points, relatively small sample sizes have been shown to sufficiently demonstrate effectiveness of strategies to reduce in-stent restenosis.30 Second, the population studied by Takagi et al represented carefully selected patients with relatively early diabetes as opposed to our “real world” population with significantly more advanced disease. It is possible that patients with more advanced stages of diabetic atherogenesis have a higher burden of mitogenic stimuli and a vascular milieu that would promote an aggressive restenotic response.34 Third, patients in the troglitazone study were pretreated with study drug at least 2 days prior to stenting. In contrast, study drug was initiated after stent implantation in our study. This may have allowed the critical early proliferative response to mechanical injury to proceed unchecked. Fourth, the favorable antirestenotic result seen with troglitazone may not be a class effect. Finally, the 2 groups in this study were not well matched in all potential confounders of restenosis. Longer duration of diabetes, more left anterior descending lesions, and more smoking in the rosiglitazone group might have obscured the true effect of the drug.
Elevated PAI-1 plasma level and activity have been well described in various insulin resistance syndromes.35, 36, 37 This has been at least partly attributed to induction of PAI-1 expression by insulin and its precursors.23 Troglitazone has been shown to reduce PAI-1 plasma and tissue levels.22 An unexpected finding in our study was that despite lowering of plasma insulin levels by rosiglitazone, PAI-1 levels were not significantly altered. It is possible that this study was not adequately powered to demonstrate that result. However, it could also be postulated that the PAI-1 lowering action of troglitazone is principally mediated by mechanisms other than insulin reduction38, 39 and, therefore, rosiglitazone may not possess that effect.
The expected biochemical effects of rosiglitazone on insulin resistance were readily shown in our study patients randomized to the drug. To the extent that the effectiveness of the drug as an antidiabetic was demonstrated, the observed lack of benefit in reducing in-stent neointimal proliferation and PAI-1 levels appears to be real.
In conclusion, in a small population with treated type 2 diabetes, systemic rosiglitazone therapy initiated after coronary stent implantation does not appear to significantly reduce in-stent restenosis or PAI-1 plasma level. The exact effects of the thiazolidinedione group of drugs, especially the currently available clinically, on human vascular tissue and the overall processes of atherogenesis and thrombogenesis await further careful evaluation in in vitro as well as large clinical studies.
References
- . Implications of population growth on prevalence of diabetes (a look at the future). Diabetes Care. 1992;S1:6–9
- . Comparison of CABG with PTCA in patients with multivessel disease. N Engl J Med. 1996;33:217–225
- . Influence of diabetes mellitus on early and late outcome after PTCA. Circulation. 1995;91:979–989
- . Can restenosis after PTCA be predicted from clinical variables?. J Am Coll Cardiol. 1993;21:6–14
- . Multi-lesion coronary angioplasty (clinical and angiographic outcome). J Am Coll Cardiol. 1987;10:246–252
- . Repeat PTCA and predictors of recurrent restenosis. Am J Cardiol. 1989;63:409–413
- . Multiple coronary angioplasty (a model to discriminate systemic and procedural factors related to restenosis). J Am Coll Cardiol. 1988;12:310–314
- . Recurrence of restenosis after PTCA. Heart Lung. 1986;15:585–587
- . Luminal narrowing after PTCA. Circulation. 1993;88:975–985
- . Clinical and angiographic outcome after directional coronary atherectomy. Am J Cardiol. 1993;72:55–64E
- . Restenosis after arterial injury caused by coronary stenting in patients with diabetes mellitus. Ann Intern Med. 1993;118:344–349
- . Relation of restenosis after excimer laser angioplasty to fasting insulin levels. Am J Cardiol. 1994;73:323–327
- . Large platelets circulate in an activated state in diabetes mellitus. Semin Thromb Hemost. 1991;17:433–439
- . Increased PAI-1 levels (a possible link between insulin resistance and atherosclerosis). Diabetologia. 1991;34:462–475
- . Factors responsible for impaired fibrinolysis in obese subjects and NIDDM patients. Diabetes. 1994;43:104–109
- . Diabetes mellitus induces accelerated growth of aortic smooth muscle cells (association with overexpression of PDGF β-receptors). Eur J Clin Invest. 1993;23:84–90
- . Enhanced arterial intimal thickening after balloon catheter injury in diabetic animals accompanied by PDGF β-receptor overexpression of aortic media. Eur J Clin Invest. 1994;24:377–381
- . Insulin, insulin-like growth factor-I, and PDGF interact additively in the induction of proto-oncogene c-myc and cellular proliferation in culture bovine aortic smooth muscle cells. Mol Endocrinol. 1989;3:1183–1190
- . Expression of TGF-β is elevated in human and experimental diabetic nephropathy. Proc Natl Acad Sci USA. 1993;90:1814–1818
- . Glucose and glucosamine regulate growth factor gene expression in vascular smooth muscle cells. Proc Natl Acad Sci USA. 1992;89:8150–8154
- . An antidiabetic thiazolidinedione is a high affinity ligand for PPAR-γ. J Biochem. 1995;270:12953–12956
- . Elucidation of mechanisms responsible for normalization by troglitazone of plasma activity of PAI-1. J Am Coll Cardiol. 1999;3(Suppl A):301A
- . Augmentation of the synthesis of PAI-1 by precursors of insulin (a potential risk for vascular disease). Circulation. 1994;89:321–330
- . Increased intramural expression of PAI-1 after balloon injury (a potential progenitor of restenosis). J Am Coll Cardiol. 1994;24:1742–1748
- . Potent inhibitory effect of troglitazone on carotid arterial wall thickness in type 2 diabetes. J Clin Endocrinol Metab. 1998;83:1818–1820
- . Troglitazone reduces intimal hyperplasia after coronary stent implantation in patients with non-insulin dependent diabetes. J Am Coll Cardiol. 2000;36:1529–1535
- . Report of the expert committees on the diagnosis and classification of diabetes mellitus. Diabetes Care. 1997;20:1183–1197
- Differential inhibitory effects of antidiabetic drugs on arterial smooth muscle cell proliferation. Am J Hypertens. 1996;9:188–192
- . Cardiovascular risk continuum (implications of insulin resistance and diabetes). [review] Am J Med. 1998;105:4–14S
- . Validation of the in vivo intravascular ultrasound measurement of in-stent neointimal hyperplasia volumes. J Am Coll Cardiol. 1998;32:794–799
- . Small stent size and intimal hyperplasia contribute to restenosis (a volumetric intravascular ultrasound analysis). J Am Coll Cardiol. 1995;26:720–724
- . Patterns and mechanisms of in-stent restenosis (a serial intravascular ultrasound study). Circulation. 1996;94:1247–1254
- . Troglitazone reduces intimal hyperplasia after coronary stent implantation in patients with type 2 diabetes mellitus (a serial ultrasound study). J Am Coll Cardiol. 1999;3(Suppl A):100A
- . Potential mechanisms promoting restenosis in diabetic patients. J Am Coll Cardiol. 1996;27:528–535
- . Factors responsible for impaired fibrinolysis in obese subjects and NIDDM patients. Diabetes. 1994;43:104–109
- . Effect of diabetes on the coagulation and fibrinolytic systems and its implications for atherogenesis. Coron Artery Dis. 1992;3:26–32
- . Tissue type plasminogen activator antigen and plasminogen activator inhibitor in diabetes mellitus. Arteriosclerosis. 1988;8:68–72
- . Differential regulation by troglitazone of plasminogen activator inhibitor type 1 in human hepatic and vascular cells. J Clin Endocrinol Metab. 2000;85:1563–1568
- . Troglitazone reduces plasminogen activator inhibitor-1 expression and secretion in cultured human adipocytes. Diabetologia. 2000;43:377–383
☆ Studies were conducted in part on the General Clinical Research Center (GCRC) at the University of Texas Medical Branch at Galveston, funded by grant M01 RR 00073 from the National Center for Research Resources, NIH, USPHS.
PII: S0002-8703(03)00868-8
doi:10.1016/j.ahj.2003.12.006
© 2004 Elsevier Inc. All rights reserved.
