High Plasma Heparin Cofactor II Activity Is Associated With Reduced Incidence of In-Stent Restenosis After Percutaneous Coronary Intervention

Thrombin plays a critical role in the development of atherosclerosis and restenosis after percutaneous coronary intervention (PCI) via activation of thrombin receptors (PAR-1 and -4) on endothelial cells, smooth muscle cells, macrophages, and fibroblasts.^1,2 Heparin cofactor II (HCII) and antithrombin (AT) serve as physiological inhibitors of thrombin actions, and inactivation of thrombin by these 2 serine protease inhibitors is accelerated >1000-fold by binding to glycosaminoglycans.³ Among glycosaminoglycans, dermatan sulfate consists of ≈9% of all glycosaminoglycans in healthy human coronary arteries, and a significant increase in its distribution was detected in advanced atherosclerosis.⁴ HCII binds to dermatan sulfate with high affinity and efficiently inactivates thrombin, whereas AT cannot bind to dermatan sulfate. These findings are consistent with the assumption that HCII exerts its antithrombin action by binding to dermatan sulfate at injured arterial walls, where large amounts of thrombin are generated on the surface of smooth muscle cells, macrophages, and fibroblasts.³

Several families of congenital heterozygous HCII deficiency have been reported. Clinical characteristics of these family members were indicative of a mild increase in the risk of venous thrombosis.^5–7 Recently, we found a patient with congenital heterozygous deficiency of HCII caused by a missense mutation of the HCII gene showing multiple and severe atherosclerotic lesions, including coronary artery stenosis, stenosis of the internal carotid artery, renal artery stenosis, and abdominal aortic aneurysm.⁸ Matsuo et al⁹ also reported a patient with congenital heterozygous HCII deficiency manifesting severe coronary artery disease with recurrent angina pectoris after repeated percutaneous transluminal coronary angioplasty. Both of these patients were >60 years old and had multiple risk factors for atherosclerosis. By analyzing clinical features of patients with congenital HCII deficiency reported thus far, we found that most of probands <40 years old suffered from deep vein thrombosis, whereas older probands >40 years old tended to manifest atherosclerotic disorders, including angina pectoris and cerebral infarction.¹⁰ From these observations, we hypothesized that patients with HCII deficiency have decreased capability of thrombin inactivation, which leads to enhanced actions of thrombin in smooth muscle cells, macrophages, and fibroblasts at injured arterial walls. Thus, if a patient with HCII deficiency has risk factors for atherosclerosis, the insufficient antithrombin actions of HCII at injured arterial walls accelerate the development and aggravation of atherosclerotic lesions.

Stent implantation has become a major therapeutic procedure for percutaneous myocardial revascularization. Despite excellent immediate outcomes, 7% to 37% of patients with coronary stent implantation are reported to develop angiographic restenoses and require additional treatments.¹¹ Recent histopathologic studies after coronary stenting have shown that in-stent restenosis (ISR) results from smooth muscle cell proliferation and extracellular matrix formation.^12,13 Therefore, it is plausible to assume that higher plasma HCII activity represents higher inhibitory activity of HCII against thrombin at injured arterial walls, which leads to a reduced incidence of ISR after coronary stenting. Thus, the present study was undertaken to clarify whether high plasma HCII activity can reduce ISR after stent implantation.

Methods

Subjects

A consecutive series of patients who had had symptomatic coronary arterial disorders successfully treated with stenting and had undergone follow-up angiography in the period between April 1998 and November 2002 at Health Insurance Naruto Hospital were enrolled in this study. Exclusion criteria included target lesions with chronic total obstruction, coronary aneurysm, saphenous vein graft, and inability to perform follow-up angiography. All patients were treated with stenting using NIR stents (Boston Scientific Corp). The study followed the institutional guidelines of the University of Tokushima and Health Insurance Naruto Hospital, and informed consent was obtained from all patients according to the Declaration of Helsinki.

Intervention Procedure

All patients received bolus heparin injections (10 000 U) before procedures, followed by continuous infusion of heparin supplements (500 U/h) for ≥24 hours. Patients were given a daily dose of aspirin (81 or 100 mg) before the procedure and an additional daily dose of ticlopidine (200 mg) for 1 month after the procedure. Intervention procedures were performed by means of the femoral approach, with arterial introducers of size 6F to 8F. Stent placement was performed according to standard methods. Success of the procedures was defined as a reduction in stenosis to <50% by quantitative coronary angiography, with a TIMI grade 3 flow without dissection.

Quantitative Coronary Angiography

After isosorbide dinitrate administration, matched views were selected from angiograms recorded before (baseline), immediately after (post), and at 6 months (follow-up) after intervention procedures. Films at the procedure and after 6 months were read in a blinded fashion using the Cardiovascular Measurement System (CCIP-310/W, Cathex Co) at the Tokushima University Hospital. The 6-month angiographies were analyzed by side-by-side projection of treatment catheters to ensure accurate identification of stents. Target sites were defined as segments of stent injury necessary to treat lesions. Reference diameter, minimal lumen diameter (MLD), and percent diameter stenosis (%DS) of a target site were determined. Acute lumen gain was defined as the difference in MLD between measurement at the end of intervention and before balloon dilatation. Late lumen loss was calculated as the difference in MLD between measurements immediately after procedures and at 6-month follow-up. Late loss index was expressed by dividing late lumen loss by acute lumen gain. Angiographic restenosis was defined by %DS≥50%.

End Points

The primary end point was the stenosis rate defined as %DS≥50% measured by quantitative coronary angiography at 6-month follow-up. Secondary end points included major adverse clinical events, such as all-cause deaths, myocardial infarction, stroke, and target vessel revascularization (PCI or CABG) during the 6-month follow-up period. Myocardial infarction was defined by the presence of new Q waves or the elevation of the MB isoenzyme of creatine kinase level to twice the upper limit of normal values.

Measurements of Plasma HCII and AT Activities

Blood was drawn from all patients at 6-month follow-up, collected into a tube containing 1/10 volume of 3.8% sodium citrate, and centrifuged at 2000g for 20 minutes. Plasma was stored at −80°C until use. Plasma HCII activity was measured on the basis of AT activity in the presence of dermatan sulfate with a Stachrom HCII kit (Diagnostica Stago). The intra-assay and interassay coefficients of variation of this kit were 3.9% and 4.3%, respectively. Plasma AT activity was measured by a chromogenic method based on AT activity in the presence of heparan sulfate using a Testzyme ATIII-2 kit (Kabi Diagnostica AB).

Statistical Analyses

Statistical analyses were performed with the StatView statistical package (StatView 5.0, SAS Institute Inc). Comparisons of continuous variables between groups were performed by repeated-measures 1-way ANOVA or Fisher’s protected least significant difference test. Comparisons of noncontinuous variables between groups were performed with the Kruskal-Wallis test. Subgroup comparisons of categorical variables were assessed by the χ² test. Multivariate analysis (logistic regression) was performed to determine parameters that predict angiographic restenosis (%DS≥50%).

Results

Study Patients

Because stent thrombosis occurred in 1 patient (1 coronary lesion) and this patient was treated with additional stent implantation, this case was excluded from the study. Therefore, a total of 166 treated lesions in 134 consecutive patients were analyzed in this study. During the 6 months, there was no clinical event such as myocardial infarction, stroke, or death.

Plasma HCII activities in 134 patients measured in this study ranged from 36.9% to 212.3% (97.7±24.8%, mean±SD). We tentatively classified these patients as follows, because plasma HCII activity >20% and <20% (approximately ±1 SD) of the mean value is ≈110% and 80%, respectively: Patients with HCII activity >110% were classified as the high-HCII group, those with <80% as the low-HCII group, and those from 80% to 110% as the normal-HCII group. By this classification, the high-, normal-, and low-HCII groups consisted of 36 patients (45 lesions), 66 patients (81 lesions), and 32 patients (40 lesions), respectively. Baseline clinical characteristics of the 3 groups were well matched with regard to various parameters other than gender difference (Table 1).

Procedural Comparison

As shown in Table 2, quantitative coronary angiography data obtained at baseline and immediately after procedure were not significantly different among the 3 groups. However, %DS at follow-up, which indicates the degree of angiographic restenosis, was significantly lower (P=0.046) in the high-HCII group (18.7±24.3%) than in the normal-HCII (30.3±24.8%) or low-HCII (29.0±29.0%) group. In addition, the ISR rate in the high-HCII group was only 6.7% (3 of 45), whereas the values in the normal-HCII and low-HCII groups were 18.5% (15 of 81) and 30.0% (12 of 40), respectively (Figure 1). Thus, the difference of ISR rate between the high-HCII and low-HCII groups was statistically significant (P=0.0039), and that between the high-HCII and normal-HCII groups was nonsignificant (P=0.068).

• Download figure
• Download PowerPoint

Multivariate Analysis

As shown in Figure 2, multivariate analysis of risk factors for ISR demonstrated that plasma HCII activity is an independent factor in reducing the incidence of angiographic restenosis (odds ratio, 0.953; 95% CI, 0.911 to 0.998; P=0.042). This value, 0.953, represents a value of a 1% increase of plasma HCII activity; thus, it is notable that the odds ratio of HCII as a predictor for ISR is calculated to be 0.786 (95% CI, 0.627 to 0.990) and 0.618 (95% CI, 0.394 to 0.980) when plasma HCII activity increases by 5% and 10%, respectively. Diabetes mellitus was significantly associated with the progression of ISR (odds ratio, 5.972; 95% CI, 1.074 to 33.221; P=0.041). No association was observed in gender, age, hypercholesterolemia, hypertension, smoking, or plasma level of AT activity.

• Download figure
• Download PowerPoint

Discussion

Among risk factors identified for ISR, female gender, MLD immediately after PCI, lesion length, and presence of diabetes mellitus are well-known predictors for ISR.^14–16 In addition, high plasma levels of homocysteine, lipoprotein (a), insulin, insulin-like growth factor I, and soluble CD40 ligand are also reported to be associated with ISR.^11,17–19 All the above predictive factors for ISR are prorestenotic when their plasma levels are elevated. In the present study, prospective follow-up of patients successfully treated with stent implantation demonstrated that plasma HCII activity >110% is associated with a reduced risk for ISR at 6 months after PCI. These results suggest that HCII may be a hitherto unrecognized negative predictor for ISR.

Multivariate analysis also showed that plasma AT activity has no relationship with ISR, further supporting the notion that the locations of action of these 2 thrombin inhibitors in vascular tissues are different.²⁰ In fact, AT/heparin complex can inactivate soluble thrombin but not surface-bound thrombin, whereas HCII/dermatan sulfate complex can inhibit both forms of thrombin.³ After vascular injury, expression of tissue factor on vascular smooth muscle cells, macrophages, and fibroblasts in both media and intima is upregulated, leading to overproduction of surface-bound thrombin.^1,21–23 In addition, vascular smooth muscle cells and fibroblasts also secrete dermatan sulfate. Because binding of HCII to dermatan sulfate greatly enhances its antithrombin action, the presence of these types of cells but not endothelial cells can activate HCII and inhibit thrombin action in arterial walls. When plasma HCII level is elevated, more HCII is expected to bind to dermatan sulfate in arterial walls. The formation of HCII/dermatan sulfate complex leads to an efficient inactivation of surface-bound thrombin, which is expected to reduce ISR by inhibiting proliferation of vascular smooth muscle cells and fibroblasts.

In our preliminary studies, we analyzed difference in restenosis rate between the high-, normal-, and low-HCII groups of patients after balloon angioplasty, which resulted in negative findings, with no significant influence of plasma HCII activity on coronary restenosis (data not shown). The pathogenesis of restenosis after balloon angioplasty is reported to be a combination of negative vascular remodeling and intimal hyperplasia. In contrast, a histopathologic study of coronary arteries from patients undergoing coronary stenting has shown that ISR is caused primarily by intimal hyperplasia composed of proliferated smooth muscle cells and increased extracellular matrix without negative vascular remodeling.^12,13 In addition, ISR tissue is composed of larger areas of smooth muscle cells and proliferative cells with smaller areas of macrophages and collagen deposition compared with restenotic tissues after balloon angioplasty.²⁴ Thus, it is conceivable that the reason why we could not detect a significant influence of plasma HCII activity on coronary restenosis after balloon angioplasty is a difference in the pathogenesis of restenosis after the 2 interventional procedures.

In an injury model of rabbit carotid artery, Buchanan and Brister²⁵ demonstrated that acute infusion of dermatan sulfate prevented neointimal hyperplasia by 45% and 47% after the first and second injuries, respectively. Although dermatan sulfate infusion also decreased vessel wall occlusion by 47% and 33% after the first and second injuries, respectively, heparin treatment showed no effectiveness on the prevention of neointimal hyperplasia or vessel wall occlusion.²⁵ Moreover, direct administration of HCII in a rat model prevented photochemically induced acute thrombus formation, and intimatan (dermatan 4,6-di-O-sulfate) administration in a canine model also significantly suppressed arterial thrombotic occlusion induced by electrolytic injury, whereas low-molecular-weight heparin had no effect.^26,27 In homozygous HCII-deficient mice, the time of formation of occlusive thrombus in the carotid artery after photochemically induced endothelial injury was shorter than that of wild-type mice.²⁰ Taken together, these results are consistent with the notion that thrombin activates and enhances proliferation of vascular smooth muscle cells, leading to neointimal hyperplasia, and that surface-bound thrombin can be inactivated by HCII/dermatan sulfate complex but not AT/heparin complex.

The possibility remains that the relation between high HCII activity and low incidence of ISR might be an epiphenomenon. To date, elevation of plasma HCII levels has been documented only in such physiological and pathological conditions as pregnancy, acute inflammation, nephrotic syndrome, and hemorrhagic stroke or in women taking oral contraceptives.¹⁰ No patients with these conditions were included in this study; however, unknown factors involved in the reduction of ISR that concomitantly elevate plasma HCII activity may be present. If this is the case, finding factors that upregulate plasma HCII activity may lead to the identification of novel therapeutic targets for reducing ISR.

Several drugs, including probucol, cilostazol, troglitazone, valsartan, and pemirolast, have been reported to be effective in preventing ISR in randomized clinical trials.^28–32 In addition, recent studies with drug-eluting stents have shown excellent outcomes in reducing the restenosis rate of de novo lesions to <5%.³³ In particular, clinical trials with sirolimus (rapamycin)-eluting stents showed long-term inhibition of neointimal hyperplasia for up to 2 years.^34,35 In a rabbit model, everolimus, an orally active sirolimus, is also effective in inhibiting in-stent neointimal proliferation.¹² However, use of pharmacological agents is frequently associated with adverse events and drug resistance. Because HCII acts as a physiological endogenous inhibitor of in-stent neointimal proliferation and restenosis, the present results warrant further studies to examine the effect of HCII-eluting or -coated stents in inhibiting ISR. Animal studies to clarify these issues are under way.

Footnotes