The level of ambient platelet activation prior to undergoing percutaneous coronary intervention (PCI) has been shown to predict subsequent adverse clinical outcomes. An analysis of hemostatic predictors of stent thrombosis demonstrated that patients in the highest quartile of IIb/IIIa receptor expression have the greatest risk for stent thrombosis. In fact, GP IIb/IIIa receptor expression was a more powerful predictor of adverse outcomes than was prothrombin fragment F-1.2, which is a marker for thrombin production.
Another study demonstrated that patients with activated platelets as reflected by CD62 and CD63 expression had worse outcomes following angioplasty. In addition, the angioplasty procedure further activates platelets secondary to arterial injury and exposes a thrombogenic milieu. For example, following angioplasty, there is an increase in both GP IIb/IIIa as well as CD11b/18 receptor expression. The CD11b/18 (MAC 1) receptor modulates white cell adhesion, white cell platelet interactions and the inflammatory response to both stent-vessel injury and distal embolization.
When one traumatizes the artery and leaves a stainless steel prosthesis in place, the stent itself further activates the coagulation system. For years, the platelet was inadequately treated prior to and during PCI by aspirin and heparin alone. We now know that pre-procedural Plavix improves outcomes following PCI. This observation suggests that by treating the platelet more aggressively, we can reduce periprocedural ischemic events. When one gives a GP IIb/IIIa receptor blocking agent in optimal concentrations, more potent inhibition of platelet aggregation is observed and is associated with yet better clinical outcomes, specifically a reduction in the incidence of periprocedural myocardial infarction and abrupt coronary closure.
The Inflammatory Response to Vessel Injury
The inflammatory response to stenting is two-fold, involving both a local response at the site of arterial injury as well as a subsequent response in the distal myocardial bed downstream due to athero-embolization and microvascular obstruction. Stents elicit a localized inflammatory response, the magnitude of which is dependent on the degree of focal media injury. This response is largely mediated by macrophage infiltration, which drives the process of late in-stent restenosis. In fact, if one experimentally depletes macrophages with lyposomal chlodronate in the animal model (rabbit) of carotid balloon angioplasty, the macrophage-depleted animals have little or no scar tissue formation and the control animals have an abundance. The pivotal role of the macrophage in the process of neointimal proliferation in response to balloon injury is evident.
Time Course of the Inflammatory Response
The time course of the inflammatory response to stenting based on post-mortem histology is shown in Figure 1. Lymphocyte-macrophage (mononuclear cell) infiltration is prevalent both early and late through the sequence of thrombus formation, inflammation and neointimal proliferation. The time course of the inflammatory response to stenting as reflected by C-reactive protein (CRP) levels is shown in Figure 2. These levels appear to be maximal at 48-72 hrs post stenting. If the CRP exceeds 0.5 mg/L at 72 hours after stenting, an increase in adverse outcomes, including death, myocardial infarction (MI), and recurrent angina can be expected. Exciting new data suggest that patients who manifest an exaggerated inflammatory response to stenting (CRP > 0.5 mg/d at 72 hrs post-stent) may benefit from anti-inflammatory treatment. Patients defined by having a CRP > 0.5 mg/L were randomly assigned to either prednisone (45 days) or placebo. At 6 months follow-up, patients who received prednisone had enhanced event-free survival and a reduction in angiographic late loss and restenosis (Figure 3).
Drug-Eluting Stents and the Inflammatory Response
Although drug-eluting stents may suppress local plaque inflammation and the inflammatory response to stent-mediated injury, they do not affect the generalized inflammatory arteritis that characterizes acute coronary syndromes. Furthermore, they do little for inflammation in response to distal athero-embolization, which occurs at least as often with drug-eluting stents. In the SIRIUS trial, periprocedural CK-MB elevation exceeding 3X upper limit of normal was similar in incidence between Cypher (Cordis) and bare metal stents. Concern exists that drug-eluting stents may be associated with platelet activation and aggregation or with white cell-platelet aggregate formation. The coated stents may not be inert.
A novel and informative observation on the mechanism of stent restenosis was made by Dirk H. Walter (Munich, Germany). Patients who were on statin medications before having coronary stent deployment had lower rates of subsequent stent restenosis. Although this experience is observational and nonrandomized, the results could not be explained by either total cholesterol or change in cholesterol levels. How, then, do statins affect restenosis if not explained by cholesterol?
Dan Simon and his group at the Brigham and Women’s Hospital have provided an intriguing answer. They found that simvastatin reduces neointimal thickening in LDL- deficient mice following balloon angioplasty. Simvastatin reduced the white cell inflammatory response to balloon injury at one week with a subsequent reduction in neointimal proliferation by 28 days and no change in plasma lipids. Thus, simvastatin appears to directly alter the white cell inflammatory response by a mechanism distinct from LDL-cholesterol.
Drug-eluting stents provide targeted site-specific drug delivery, and for the time duration that a specific drug is present, will likely suppress the localized inflammatory response to stent injury. The inflammatory response to stenting (especially neointimal macrophage content) drives late neointimal proliferation. However, when the drug is completely eluted (no longer present) and the polymer delivery system remains, the polymer may exert the opposite effect. For example, in the QuaDS-QP2 system device, comprised of mechanical drug reservoirs containing the drug Taxol ® (paclitaxel), restenosis was reduced at 6 months post-deployment. However, when the paclitaxel was no longer present, the reservoir device actually induced an aggressive proliferative response. At one year following deployment, an increased incidence of diffuse restenosis was observed. Another such example was the Actinomycin D Eluting Stent Improves Outcomes by Reducing Neointimal Hyperplasia (ACTION) trial with actinomycin-D. Although 30-day data in animals looked quite promising, this polymer-based drug delivery system (TETRA stent) was associated with increased late in-stent restenosis compared with the bare metal TETRA stent alone (no drug or polymer). As demonstrated by the adverse late outcomes in these scenarios, either the drug itself, the polymer or the delivery platform may have adverse late effects that counter the presence of an early-antiproliferative action.
Acute Coronary Syndromes: A Generalized Disease Process
Patients with ACS who present with an elevated CRP and who then undergo angioplasty have an increased risk of death or MI at 6 months compared with patients who have a normal CRP. In fact, CRP levels prior to early revascularization for non-ST elevation ACS correlate with late (5-year) mortality.
We now realize that elevated systemic CRP correlates with white cell (neutrophil) activation. This reflects a generalized inflammatory arteritis. For example, patients with unstable angina who have a right coronary target stenosis, as well as those with a left coronary target stenosis, will demonstrate a myeloperoxidase gradient which signifies neutrophil activation between the aorta and great cardiac vein (Figure 4). The great cardiac vein drains the left anterior descending coronary distribution. Why would there be white cell activation in the left anterior descending coronary distribution if the patient has a right coronary target lesion? The obvious answer is that these patients have a multi-lesion, multi-vessel, multi-centric inflammatory process. As recently demonstrated by a multi-vessel intravascular ultrasound (IVUS) study of patients presenting with acute coronary syndromes, 80% of patients had at least one extra plaque rupture distinct and separate from the culprit plaque rupture which prompted their hospital admission. Indeed, the ultimate culprit in the pathogenesis of this process is macrophage infiltration of plaque. The degree of macrophage infiltration of plaque (inflammatory infiltrate) correlates with and parallels the level of coronary disease activity (Figure 5). If one deploys a drug-eluting stent at the site of a culprit plaque, how will it passivate the other non-culprit plaques? Remember, as demonstrated by IVUS, 25% of ACS patients have at least two other non-culprit ruptured plaques present. Acute coronary syndromes represent a generalized inflammatory disease process. Elevated levels of CRP observed in these syndromes reflect the degree of underlying vascular inflammation. Targeted drug delivery (drug-eluting stents) will not effectively treat a systemic, diffuse, inflammatory process.
Reducing Restenosis vs. Treating A Diffuse Disease Process
At this time, drug-eluting stents have been demonstrated only as reducing restenosis. GP IIb/IIIa inhibiting agents have not been conclusively shown to reduce restenosis. Although data in diabetic patients from the EPISTENT trial suggest that abciximab administered at the time of stent deployment is associated with a reduction in late target vessel revascularization, as a class the GP IIb/IIIa blocking agents have not been shown to reduce restenosis.
So what benefits do GP IIb/IIIa blocking agents provide?
1. First, they reduce periprocedural myocardial infarction. Remember that in the SIRIUS trial, the Cypher stent did not reduce the occurrence of periprocedural myocardial infarction. Indeed, if drug-eluting stent technology drives an increase in complex, multi-vessel PCI, one might predict that the incidence of periprocedural myocardial infarction will rise in the absence of adequate periprocedural platelet GP IIb/IIIa blockade.
2. Secondly, periprocedural subacute stent thrombosis is increased with long or multiple stents, stent overlap and suboptimal stent deployment. Abciximab has been demonstrated to reduce abrupt coronary closure and subacute stent thrombosis. Thus, the potential utility of periprocedural platelet GP IIb/IIIa blockade to reduce subacute stent thrombosis following stenting in more complex patient and vessel subsets is clear.
Abciximab Anti-Inflammatory Therapy
Abciximab has a potent, direct and sustained anti-inflammatory effect. In a retrospective analysis of the EPIC trial, patients randomly allocated to receive abciximab had a highly significant suppression in periprocedural (24-48 hrs) levels of interleukin 6 (IL 6) and CRP (Figure 6). Furthermore, there was no evidence for re-elevation or rebound in levels of these cytokines at 4 weeks. In part, the durability of abciximab’s anti-inflammatory effect may be explained by redistribution pharmacokinetics. A single bolus of abciximab has been demonstrated to exert a 2-week treatment phase. At 15 days following abciximab bolus, there remains greater than 10% platelet GP IIb/IIIa receptor occupancy by abciximab (Figure 7). Abciximab has been demonstrated to block IL 6 production by monocytes from patients who have unstable angina.
The double bolus plus infusion dose regimen of eptifibatide provides potent antiplatelet therapy that may have secondary anti-inflammatory effects. The CRP profile of PCI during treatment with eptifibatide (Figure 8) is shown. During bolus and infusion, CRP is suppressed from baseline values. However, following eptifibatide discontinuation, the CRP level is actually higher (48 hours) than baseline. Thus, eptifibatide appears to suppress inflammation only during the infusion. To date, no durable or sustained suppression of inflammation following eptifibatide has been demonstrated. Furthermore, we know that inflammation is not only at the stent site but also in the distal myocardial bed secondary to microvascular atheroembolism. In addition, the inflammatory response to stenting is maximum at 72 hours post-procedure.
Following discontinuation of eptifibatide, what anti-inflammatory effect will be present at 48-72 hours post procedure? This time course is critical in the inflammatory response to stenting, as reflected by CRP measurement. An antiplatelet therapy may have secondary anti-inflammatory effects. For example, at very high levels of GP IIb/IIIa receptor antagonism (> 70-80%), the release of CD40L from the alpha granule of the platelet can be suppressed. CD40L is a pro-thrombotic, pro-inflammatory mediator which is stored in the platelet alpha granule and released with other alpha-granule contents (PDGF, PF4, BTGF, TSP) during platelet aggregation. If platelet aggregation is blocked by GP IIb/IIIa receptor inhibition, so is the release of alpha granule contents. Suppressed release of soluble CD40L thus reduces inflammation and thrombosis. If platelet inhibition is the primary mechanism by which inflammation is suppressed, when drug (short-acting small molecule) infusion is discontinued, both antiplatelet and anti-inflammatory effects should resolve quickly.
In fact, data that longer is better have already been presented by the Enhanced Suppression of the Platelet IIb/IIIa Receptor with Integrilin Therapy (ESPRIT) study group (Figure 9). Patients who received a less than 16-hour infusion of eptifibatide had significantly worse clinical outcomes than those patients who received more than a 22-hour infusion used as a reference. With a hazard ratio exceeding 3.0 for Diabetes: An Inflammatory State
Diabetes is an inflammatory state, marked by increased macrophage infiltration of plaque (Figure 10). William O’Neill and the group at Beaumont Hospital believe that the degree of inflammation, more than diabetes per se, determines clinical outcomes. They analyzed patients with ACS for cardiac death at one year by diabetic status and terile of CRP (Figure 11). It was the degree of vascular inflammation, as reflected by CRP, that determined mortality more than diabetes status per se. The common denominator between these inflammatory states (ACS, diabetes, vascular response to injury) is the presence of increased macrophage infiltration in plaque. Since platelet inhibition may suppress inflammation, what effect will GP IIb/IIIa inhibitor agents have on diabetics with ACS? GP IIb/IIIa inhibitors reduce mortality in diabetic patients with ACS (Figure 12).
However, an interesting observation can be made from these data. The GUSTO IV trial was not the largest trial included in the analysis shown and was a negative trial (did not achieve primary endpoint) for the trial as a whole. Yet it is the ONLY trial which achieves a statistically significant reduction in mortality for diabetic patients and where the confidence intervals for abciximab effect do not cross the midline (Figure 12). This observation is consistent with the effect of abciximab on late mortality in the Evaluation of IIb/IIIa Platelet Inhibitor for Stenting (EPISTENT) trial (Figure 13). In both trials, abciximab therapy significantly reduced mortality in diabetic patients. Why should abciximab be so effective in diabetic patients? The answer may have little to do with platelets. Abciximab also demonstrates affinity for the CDIIb/18 (MAC1) receptor. This receptor is upregulated on neutrophils and monocytes following PCI and is inhibited by abciximab. CDIIb/18 modulates white cell adhesion, white-cell platelet interactions and the inflammatory response to stent-vessel injury.
Furthermore, we have demonstrated that this receptor is integral in mediating monocyte-induced smooth muscle death. For example, we know that patients with ACS have increased macrophage infiltration of plaque (Figure 5). Macrophages produce monocyte colony stimulating factor (MCSF). We have demonstrated that the combination of macrophage, binding to vascular smooth muscle cells in the presence of MCSF, results in smooth muscle death (Figure 14). The marked degree of vascular smooth muscle killing associated with this scenario is inhibited by abciximab in physiologic concentrations. The addition of physiologic concentrations of platelet GP IIb/IIIa inhibitor agents (eptifibatide or tirofiban) provided no cytoprotective effect or reduction in smooth muscle cell death (Figure 15). This observation is very consistent with what we know about plaque stability or instability from post-mortem histologic studies. It is well established that plaques prone to rupture which cause sudden cardiac death or myocardial infarction are rich in monocyte/ macrophage content and have a reduction in vascular smooth muscle cells in the cap of the plaque (Figure 16). Thus, the macrophage, in combination with MCSF, mediates vascular smooth muscle cell death through mechanisms which involve the CD 11b/18 (MAC 1) receptor. This process erodes the structural integrity of the plaque and promotes plaque instability and rupture.
Abciximab (but not eptifibatide or tirofiban) exerts a cytoprotective, plaque-stabilizing effect by inhibiting the CD 11b/18 receptor. Clinical evidence in support of this mechanism may be gleaned from the 7-year follow-up of the Evaluation of c7E3 for the Prevention of Ischemic Complications (EPIC) trial (Figure 17). At 7 years follow-up, patients who had received bolus-only abciximab had a similar reduction in mortality when compared with patients who were administered a bolus plus 12 hour infusion of abciximab (when compared to placebo). Both abciximab treatment regimens were associated with long-term survival benefit, although only the bolus plus 12 hour infusion significantly suppressed early platelet-mediated ischemic events (death, myocardial infarction or urgent revascularization to 30 days). The long-term survival benefit of bolus-only abciximab suggests a plaque-stabilizing effect with this drug when administered systematically and is not likely to be achieved by targeted (localized) drug administration via a drug-eluting stent.
Indeed, there are no data as yet to support the concept that drug-eluting stents provide a long-term survival advantage. However, from the recent pooled analysis of the EPIC, EPILOG, and EPISTENT trials, abciximab administration (vs. placebo) was associated with a significant (p=0.030) 22% relative reduction in mortality at 3 years follow-up (Figure 18). These observations with abciximab suggest that it may influence the natural history of the atherosclerotic disease process by suppressing multicentric vascular inflammation. Such systemic treatment of a generalized disease process will not accompany targeted drug delivery via a drug-eluting stent.
Treating the Underlying Disease Process
IIb/IIIa agents, particularly abciximab, treat the disease process. These agents treat the coronary syndrome, rather than just the target lesion. In general, all IIb/IIIa inhibitors, including tirofiban and epifibatide, reduce periprocedural myocardial infarction. To date, periprocedural MI has not been reduced by drug-eluting stents. Therefore, GP IIb/IIIa anti-platelet agents, specifically abciximab, will benefit patients who get drug-stents:
1. By reducing periprocedural myocardial infarction;
2. By reducing sidebranch occlusion (also linked to periprocedural MI);
3. By providing a durable anti-inflammatory effect (abciximab) which may help all plaques and favorably influence the nature history of the disease. For example, 80% of patients with ACS have at least one additional ruptured plaque distinct and separate from the culprit plaque rupture which prompted their admission.
Drug-eluting stents reduce restenosis, but certainly do not eliminate it for many populations, including diabetics and patients with small vessels.
The Cost of Drug-Eluting Stents
Interventional cardiologists have always had a fascination with stents. With drug-eluting stents, will the relative cost increment be validated by the subsequent reduction in restenosis? One point of note is that restenosis usually doesn’t kill the patient, so it may be difficult to impossible to demonstrate that DES reduce mortality. DES will likely reduce morbidity related to restenosis. In analyzing incentives of our health care system, it should be noted that restenosis is reimbursable and profitable to most hospitals. Hospital systems will lose a large portion of this downstream revenue as a result of drug-eluting stents.
Recently, the Centers for Medicare and Medicaid Services (CMS) granted unprecedented approval for procedural reimbursement (DRG) prior to FDA approval of the investigational device. CMS created DRG 526 for DES with acute myocardial infarction, and DRG 527 for DES without myocardial infarction. The geographically and institutionally weighted incremental reimbursement (above that provided for bare metal stent by DRG 517) for DRG 527 will be about $1800.
Yet, what happens if the drug-eluting stent is priced at $3000? The hospital pays full fare and CMS pays a remarkable discount while also enjoying the downstream cost offsets secondary to a reduction in coronary bypass surgery and repeat PCI due to restenosis. The hospital loses on the index procedure as well as on the downstream restenosis and surgical business. This is clearly a no-win proposition for the U.S. hospital system under the currently proposed reimbursement scheme.
Private insurance payers present yet another issue. Assumptions that private payers will provide the CMS DRG level of reimbursement may not be correct. Many payers currently reimburse less than CMS and may take up to 6-12 months following FDA approval to initiate payments. Private payers may empirically and arbitrarily refuse to pay for new technology which they refer to as experimental. For example, one year after vascular brachytherapy was FDA-approved, certain private payers still rejected legitimate claims. We have done a cost analysis for the first year following the introduction of DES. One major question is whether DES will be selectively or indiscriminately deployed. How will these decisions be made? Many doctors plan to deploy them only in high-risk patients. How is high risk defined? Does this refer to diabetics, smokers, and patients with small vessels and longer lesions? Which randomized trial DES-enrolled diabetic smokers with long lesions in small vessels? In other words, this is not an evidence-based decision. In fact, restenosis rates are higher with the Cypher stent in small vessels and diabetics, especially those that are insulin-treated. Restenosis will clearly not be eliminated. Furthermore, what are we going to tell the patient who says, I want one of those drug stents? Do we answer, No, you’re not high-risk? It creates a ludicrous situation.
In our calculations for the cost impact of DES, we assumed no incremental stent use and reimbursement at the CMS DRG level. We calculated the impact of both a 50 and 100% conversion to DES. We estimate that our cardiovascular system will lose between 5.2 (100% conversion) and 3.0 (50% conversion) million dollars in the first year following DES release if a CMS DRG level payment exists for each patient. Obviously, losses will be much greater if a portion of the patient population has no reimbursement.
At the Crossroads With a Better Mousetrap
Drug-eluting stents are a better mousetrap. What, if any, are our alternatives? The new thin strut cobalt chromium stents, including the Multi-Link Vision (Guidant), the (formerly S8) Driver (Medtronic) and the Steeplechase (Cordis) may also represent an advance in technology. For the Vision stent, late in-stent restenosis was 15.7% and in-segment restenosis was only 16.7%. If the restenosis burden is reduced by these stents, those patients who develop restenosis can receive effective treatment with beta radiation. This strategy will be both clinically effective and cost effective. In most American hospitals, cardiovascular business supports many other hospital programs. Hospitals may have to discontinue obstetrics/gynecology, psychiatry or other services if the cardiovascular unit becomes non-profitable. Indeed, programs and even hospitals may close. Drug-eluting stents come at a time of escalating costs for technology, most notably implantable defibrillators and bi-ventricular pacemakers. How will hospitals provide these technologies in the absence of adequate reimbursement?
Our system is primed for a paradigm shift in its reimbursement mechanism. In France or Germany, the government pays a certain stipend amount for every citizen’s care. While they don’t have 40 or 50 million citizens who are uninsured (as in the U.S.), the standard allotment may provide for a bare metal (non-DES) stent. If the patient wants a more advanced technology (DES), the patient is responsible for paying for it. The basic tenet that high technology health care is a right, not a privilege, is challenged by the U.S. healthcare system if adequate reimbursement is not provided. Our current system also obligates care of uninsured patients. Clearly, other sources of funds will be required to provide cutting-edge expensive technology to the mass of American people, without bankrupting the American hospital system.
Recommended ReadingBuffon A, Biasucci LM, Liuzzo G, D'Onofrio G, Crea F, Maseri A. Widespread coronary inflammation in unstable angina. N Engl J Med 2002; 347: 5-12.
Chen Z, Fukutomi T, Zago AC, Ehlers R, Detmers PA, Wright SD et al. Simvastatin reduces neointimal thickening in low-density lipoprotein receptor-deficient mice after experimental angioplasty without changing plasma lipids. Circulation 2002; 106: 20-3.
Cusack MR, Marber MS, Lambiase PD, Bucknall CA, Redwood SR. Systemic inflammation in unstable angina is the result of myocardial necrosis. J Am Coll Cardiol 2002; 39: 1917-23.
Davies MJ, Richardson PD, Woolf N, Katz DR, Mann J. Risk of thrombosis in human atherosclerotic plaques: Role of extracellular lipid, macrophage, and smooth muscle cell content. Br Heart J 1993; 69:377-381.
Farb A, Sangorgi G, Carter AJ, et al. Pathology of acute and chronic coronary stenting in humans. Circulation 1999;99:44-52.
Lincoff AM, Kereiakes DJ, Mascelli MA, Deckelbaum LI. Abciximab suppresses the rise in levels of circulatory inflammatory markers after percutaneous coronary revascularization. Circulation 2001: 104: 163-167.
Moreno PR, Falk E, et al. Macrophage infiltration in acute coronary syndromes: Implications for plaque rupture. Circulation 1994; 90:775-778.
Moreno PR, Bernardi VH, LÃ pez-Cuellar J, Palacios IF, Gold HK, Fuster V, Fallon JT: Macrophage infiltration predicts restenosis after coronary intervention in patients with unstable angina. Circulation 1996; 94: 3098-3102.
Moreno et al. Circulation 2001:104;II-56. American Heart Association Scientific Sessions 2001, Anaheim, CA, U.S.A., November 11-14, 2001. Circulation 2001;104(17)
(Supplement II):II-56, Abstract.Rioufol G, Finet G, Ginon I, Andre-Fouet X, Rossi R, Vialle E et al. Multiple atherosclerotic plaque rupture in acute coronary syndrome: a three-vessel intravascular ultrasound study. Circulation 2002; 106: 804-8.
Seshiah PN, Kereiakes DJ, Vasudevan SS, Lopes N, Su BY, Flavahan NA, Goldschmidt-Clermont PJ. Activated monocytes induce smooth muscle cell death: role of macrophage colony-stimulating factor and cell contact. Circulation 2002 Jan 15;105(2):174-80.
Topol EJ, Lincoff AM, Kereiakes DJ, Kleiman NS, Cohen EA, Ferguson JJ et al. Multi-year follow-up of abciximab therapy in three randomized, placebo-controlled trials of percutaneous coronary revascularization. Am J Med 2002; 113: 1-6.
Versaci F, Gaspardone A, Tomai F, et al. Immunosuppressive therapy for the prevention of restenosis after coronary artery stent implantation (IMPRESS study) J Am Coll Cardiol 2002;40:1935-42.
Walter DH, Schachinger V, Elsner M, Mach, S, Auch-Schwelk W, Zeiher A M. Effect of statin therapy on restenosis after coronary stent implantation. Am J Cardiol 2000; 85: 962-968.