Author + information
- Received July 9, 2002
- Revision received October 9, 2002
- Accepted October 31, 2002
- Published online February 19, 2003.
- ↵*Reprint requests and correspondence:
Dr. Steven E. Nissen, The Cleveland Clinic Foundation, F15, 9500 Euclid Avenue, Cleveland, Ohio, USA 44195.
Although an early invasive strategy (angiography and percutaneous coronary intervention) is the convention in acute coronary syndrome (ACS)/non–ST-segment elevation myocardial infarction (MI) in the U.S., a conservative pharmacologic approach is common in other countries. Trial evidence has demonstrated a modest benefit with an angiographically guided approach; but patients having negative troponin values or who were receiving aspirin showed little or no benefit, and those without ST-segment changes had slightly worse outcomes. Limitations of angiography are clinically important. Identification of hemodynamically significant stenoses may be confounded by coronary remodeling. Also, most plaques, particularly those responsible for acute events, are extraluminal. Assessment of the luminal diameter of a lesion, which requires comparison with a normal reference segment, may be impossible because of the diffuse nature of the disease. Percutaneous coronary intervention after plaque rupture may itself cause embolization and no-reflow phenomena, leading to severe complications. In addition, most ruptures may be clinically silent. Evidence of a systemic inflammatory component suggests that ACS patients are at risk for plaque rupture at multiple sites. The inability of angiography to depict the true extent of atherosclerosis is supported by necropsy and transplant donor studies. A metabolic approach to this systemic disease is the only strategy designed to influence the behavior of both the small number of angiographically visible lesions and the large number of occult plaques. Statins and other agents decrease the incidence of death and MI by stabilizing atherosclerotic plaques throughout the coronary bed, reducing inflammation, collagen degradation, tissue factor expression, and vasomotor tone.
Hospital admissions for acute coronary syndrome (ACS) or non–ST-segment elevation myocardial infarction (NSTEMI) occur in more than 1.4 million patients annually in the U.S. (1996 data) (1), but the optimal approach to its management remains controversial. In most American tertiary care centers, ACS patients are managed using an aggressive early invasive strategy, defined as angiography and percutaneous coronary intervention (PCI) within a few hours to a few days after presentation. However, most practice patterns and guidelines outside the U.S. promote an initial early conservative approach that uses anti-ischemic, anti-platelet, and anti-thrombotic drug therapy while they reserve urgent angiography and intervention for high-risk patients or those with refractory symptoms. With this approach, clinically stable patients undergo functional testing to stratify them for elective coronary angiography.
Advances in medical and interventional care have made comparison studies of conservative versus aggressive strategies difficult to interpret (2–8). Key advances in PCI include stenting with adjunctive use of glycoprotein (GP) IIb/IIIa inhibitors. Advances in medical therapy include GP IIb/IIIa inhibitors and/or clopidogrel, low-molecular-weight heparins, and aggressive use of lipid-lowering agents. Few trials have made comparisons between angiographically guided management and optimal medical therapy; however, a review of existing trials and an understanding of the pathophysiology of ACS provide a reasonable basis for making decisions. This review provides the rationale for approaching ACS as a systemic disease, rather than focusing on identification and interventional treatment of the focal stenosis.
Classical invasive versus conservative trials
The scientific case for an angiographically guided approach to ACS is not particularly well supported by traditional clinical trials, despite its popularity in the U.S. Before the development of stenting and GP IIb/IIIa inhibitors, the Thrombolysis In Myocardial Infarction (TIMI) IIIB study randomized 1,473 patients with unstable angina or NSTEMI and compared tissue plasminogen activator with placebo, and an early invasive with an early conservative strategy (9). Patients received what was then considered optimal medical therapy, including a beta-blocker, a short-acting calcium channel antagonist, nitrates, unfractionated heparin, and aspirin. In the tissue plasminogen activator versus placebo comparison, there were no significant differences in the six-week incidence of death, MI, or failure of initial therapy. Similarly, no differences were found between the invasive-versus-conservative arms in the six-week incidence of death, MI, or a poor performance on an exercise test (symptom-limited). However, patients treated with an aggressive strategy had significantly less recurrent ischemia requiring repeat hospitalization and needed fewer anti-anginal medications.
Conversely, the Veterans Affairs Non–Q-Wave Infarction Strategies in-Hospital (VANQWISH) trial suggested that an aggressive approach might actually worsen outcomes. This study randomized 920 NSTEMI patients to an early invasive or early conservative strategy (10). At hospital discharge, the invasive-strategy arm was associated with significantly higher rates of death (26% vs. 12%; p = 0.004) or non-fatal MI (15% vs. 4.9%; p = 0.007). These differences persisted to one month and one year.
The Fragmin and Fast Revascularization during InStability in Coronary artery disease (FRISC) II trial randomized 2,457 chest-pain patients with electrocardiographic changes or positive enzymes to the low-molecular-weight heparin, dalteparin, or placebo with either an invasive or conservative strategy (11). Stents were utilized in about two-thirds of PCI patients in both the invasive and conservative arms. At six months, there were no significant differences between dalteparin- and placebo-treated patients. However, death or MI was more common in the conservative than in the invasive group (12.1% vs. 9.4%; p = 0.03). Angina severity and need for re-admission were also less common in the invasive arm.
The Treat angina with Aggrastat and determine Cost of Therapy with an Invasive or Conservative Strategy (TACTICS) TIMI-18 study was organized to try to confirm whether an invasive approach was superior to a conservative strategy using contemporary pharmacotherapy and intervention techniques (12). All 2,220 NSTEMI/ACS patients received aspirin, heparin, and the GP IIb/IIIa inhibitor, tirofiban. In the invasive arm, angiography was performed 4 to 48 h after randomization. Revascularization was performed in 60% of patients in the invasive arm and 36% in the conservative arm, with at least one stent deployed in approximately 85% of PCIs. Medications included beta-blockers (82%), nitrates (94%), and lipid-lowering agents (52%).
The TACTICS TIMI-18 was the first contemporary trial to demonstrate a modest benefit from an angiographically-guided approach. The triple end point of death, non-fatal MI, or re-admission for ischemia at six months was significantly lower in the aggressive than in the conservative arm (15.9% vs. 19.4%; p = 0.025). The double end point of death or MI was also lower, but only marginally, in the invasive arm (7.3% vs. 9.5%; p = 0.05). Surprisingly, however, the two-thirds of patients taking prophylactic aspirin and the 59% of troponin T-negative patients (<0.1 ng/ml) showed little or no benefit from early invasive treatment. Furthermore, the 62% of subjects without ST-segment changes showed a slightly worse primary end point with early invasive strategy. Only the highest-risk ACS patients appeared to benefit, even with the use of GP IIb/IIIa inhibitor and stenting.
Limitations of interventional strategies
Stenosis evaluation using angiography
Although coronary angiography is the gold standard for defining coronary anatomy, radiographic imaging depicts only a rather poor representation of cross-sectional coronary anatomy from a simple planar silhouette or luminogram of the contrast-filled lumen. Angiography is confounded by observer variability, with differences in the estimation of stenosis severity approaching 50% (13,14). Also, functional testing often reveals discordance between the severity of angiographic lesions and physiologic effects (15,16). Necropsy studies and intravascular ultrasound (IVUS) demonstrate that coronary lesions, particularly after plaque rupture, are complex, with distorted luminal shapes (17)that are difficult to assess using a planar angiographic silhouette (Fig. 1).
Nevertheless, operators usually choose not to treat stenoses that do not appear hemodynamically significant, reserving intervention and PCI for visually severe lesions. However, intervention after plaque rupture and thrombosis is not without hazard. Embolization and no-reflow phenomena after successful angiographic intervention may lead to complications resulting in myocardial necrosis, unquestionably making some patients worse.
Trial data provide ample evidence that the limitations of angiography result in inappropriate treatment of lesions, resulting in adverse outcomes. Twice as many trial patients have undergone revascularization in early-invasive than in ischemia-guided arms. As mentioned above, TACTICS patients without ST-segment changes actually fared worse with an early invasive approach, despite receiving upstream GP IIb/IIIa inhibitors. Also, patients in the VANQWISH trial who were treated aggressively had a threefold increase in non-fatal MIs.
The diffuse nature of the disease
Angiography may also underestimate the extent of atherosclerosis (18,19). Both imaging and necropsy studies demonstrate that coronary artery disease (CAD) is rarely focal and typically involves a large portion of the arterial tree at the time of diagnosis (20,21). Indeed, with the appearance of the first luminal irregularity, at least 80% of the coronary tree is already atherosclerotic.
Angiographic assessment is confounded by a phenomenon known as coronary remodeling (22), which consists of outward displacement of the external elastic membrane (EEM) in atherosclerotic segments, opposing luminal encroachment and concealing the presence of disease (Fig. 2). Lesion assessment requires a comparison of the luminal diameter with an adjacent, normal reference segment. Because coronary atherosclerosis is usually diffuse, however, there may be no truly normal segment from which to determine diameter reduction. In the presence of diffuse disease, percent stenosis will always underestimate the severity of the disease. In extreme examples, the entire vessel may be symmetrically narrowed, appearing as a small artery with minimal irregularities (Fig. 3). Therefore, exclusive focus on a culprit stenosis ignores the diffuse nature of the atherosclerotic disease process.
Importance of the non-stenotic lesion
Abundant evidence demonstrates that high-grade stenoses (>70% obstruction) are rarely the source of ACS and that coronary occlusion and MI evolve most frequently from plaques that are only mildly to moderately obstructive on angiography (23). In clinical trials, about 20% of ACS patients undergoing early angiography exhibit no hemodynamically significant stenosis (24). Furthermore, because intra-lesional thrombus contributes to the severity of stenosis, delayed angiography often shows a less severe stenosis after an underlying thrombus has been resolved. In most clinical trials, a delay in angiography resulted in a reduction in the number of patients who underwent PCI.
The relationship between minimally obstructive stenoses and MI initially suggested that ACS evolves from the rupture of small, early atheromatous plaques. Recently, evidence from IVUS has offered a more complete explanation in stable versus unstable syndromes (25). We performed IVUS before coronary intervention in unstable (n = 85) and stable (n = 46) angina patients. A remodeling ratio (RR) was defined as the ratio of the EEM area at the lesion to that at the proximal reference site. Positive remodeling was defined as an RR >1.05 and negative remodeling as an RR <0.95. Surprisingly, plaque area (13.9 vs. 11.1 mm2; p = 0.005), EEM area (16.1 vs. 13.0 mm2; p = 0.004), and the RR (1.06 vs. 0.94 mm2; p = 0.008) were significantly greater at culprit lesions in patients with unstable syndromes. Positive remodeling was more frequent in unstable than in stable lesions (51.8% vs. 19.6%; p = 0.001) (Fig. 4).
Inflammation may be a common pathophysiologic link underlying the association between plaque vulnerability and remodeling. An IVUS and histologic study of the human femoral arteries showed an association between histologic markers associated with plaque inflammation and positive arterial remodeling (26). Comparative analysis of the pre-atherectomy ultrasound images and histology of the plaque tissue demonstrated an association between positive remodeling and intense stromolysin staining (an enzyme associated with degradation of the fibrous cap) (27), suggesting that degradation of the extracellular matrix plays a role in arterial remodeling.
Because bulky, non-stenotic lesions cause many, if not most, infarctions, the most severely stenotic lesion may not invariably be the culprit. Numerous ACS/NSTEMI patients have been identified in whom a comparatively minor stenotic lesion, distant from a ruptured plaque, was the apparent culprit lesion (Fig. 5). The morphologic characteristics of the vulnerable lesion are clearly more important in plaque vulnerability than percent stenosis. Plaque echolucency, which represents the lipid content of plaques, and a thin fibrous cap have been associated with ACS (28). Stable plaques characteristically have thick fibrous caps and small lipid cores. Fibrotic changes and vessel shrinkage, while causing more severe luminal stenosis, may render lesions more resistant to rupture.
Multiplicity of vulnerable plaques
Angioscopic and angiographic studies have identified lesions with characteristics associated with plaque vulnerability at sites other than the culprit lesion in ACS patients (29,30). Although rupture of a vulnerable atheroma is the major cause of ACS/NSTEMI, recent reports suggest that most ruptures are clinically silent and only occasionally cause acute symptoms (31). Presumably, after a rupture, the local balance between thrombosis and spontaneous thrombolysis prevents vessel occlusion. Recent evidence of a systemic inflammatory component also suggests that multiple sites are at risk for plaque rupture (32).
We hypothesized that ACS patients would exhibit more extensive plaque vulnerability, manifested by lesion ulceration, at sites other than the culprit lesion. We therefore examined the prevalence of ulcerated atherosclerotic lesions in vessel segments adjacent to culprit lesions in patients with stable versus those with unstable syndromes. Post-interventional IVUS examinations were performed in patients treated with emergent stenting for acute MI. Control patients were treated with elective stenting for either stable or unstable angina. Plaque ulceration distant from the infarct site was significantly more prevalent in the post-MI than the angina group (19% vs. 4%) (Schoenhagen P, et al., unpublished data).
These observations may explain the extraordinary difference in prognosis between patients with ACS/NSTEMI and chronic stable angina patients. Annual rates of death and MI average about 2% in chronic stable angina. The TACTICS trial, however, demonstrated a combined risk of death, non-fatal MI, and rehospitalization of approximately 16% at six months in ACS patients. Although the angiographically guided approach to ACS assumes that the focal lesion responsible for the unstable presentation can be readily identified and treated, this assumption is not invariably true.
Early onset of the systemic process
Although CAD patients typically become symptomatic after age 40, atherosclerotic changes in the vessel wall begin early in life (33–35). Gross inspection of coronaries of the American soldiers killed in the Korean and Vietnam wars demonstrated atherosclerosis in 77% and 45% of autopsies, respectively. In the Bogalusa Heart Study, the prevalence of CAD in 15-year-old children was 8%, reaching 69% in adults 26 to 40 years old.
Recently, we used IVUS in 262 heart transplant recipients to determine the presence, extent, and distribution of atherosclerosis in donor coronaries within weeks of transplantation (36). This method was designed to characterize the early atherosclerotic process, reflecting the fact that most donors are relatively young (Fig. 6). Donor atherosclerotic lesions were present in 136 patients, or 51.9% (Fig. 7). The prevalence of atherosclerosis varied from 17% in individuals <20 years old to 85% in subjects >50 years old. Because remodeling is evident in many early lesions, the lumen is usually preserved at first. Few lesions were detected by angiography, with stenoses identified in only 8% of patients. No abnormal angiograms were found in patients <30 years old, yet 28% of them had ultrasound evidence of atherosclerosis.
If young healthy transplant donors have extensive atherosclerosis, it is not difficult to imagine the plaque burden in patients with an ACS. Accordingly, it seems unrealistic to expect the placement of a stent in a culprit lesion to significantly alter a process that began decades earlier and involves most of the coronary arterial tree. A metabolic approach is the only strategy able to affect the behavior of both the small number of angiographically visible lesions and the large number of occult, extraluminal plaques that will ultimately result in future coronary events.
Plaque progression and vulnerability
The formation of atherosclerotic lesions is initiated by the recruitment of mononuclear leukocytes to the intima by adhesion molecules expressed on the surface of vascular endothelial cells (37). Chemoattractants facilitate entry into the intima, where macrophages accumulate lipid to become foam cells. A large, often eccentric, lipid pool forms, separated from the lumen by a fibrous cap (38). Positive remodeling, the outward expansion of the vessel at the lesion site, allows plaque accumulation without lumen obstruction (22). The accumulation of inflammatory cells and the changes in extracellular matrix macromolecules contribute to the changes in vascular morphology (39,40). These atheromata may remain silent for decades (41). Incompletely understood triggers may then initiate changes that lead to rupture and an ACS (42).
The border between the fibrous cap and normal vessel wall at the edge of a large lipid core (the so-called “shoulder area”) is particularly prone to rupture (43)(Fig. 8). The tensile strength of the fibrous cap depends on a balance between collagen biosynthesis by smooth muscle cells and collagen degradation by members of several proteinase families and their endogenous inhibitors (44,45). The secretion of these proteinases is induced by the accumulation and activation of macrophages and other inflammatory cells (46). Matrix-metalloproteinases (MMPs) play a central role in the degradation of the extracellular matrix of the fibrous cap (47).
When a plaque ruptures, platelets that accumulate on exposed plaque increase expression of the GP IIb/IIIa receptor complex, initiating further platelet aggregation and fibrin deposition and eventually forming a luminal thrombus (48). The eventual thrombus size is determined by the balance between the prothrombotic milieu and local, intrinsic thrombolytic activity. Some of these episodes of plaque rupture are clinically symptomatic, but most are silent. Further plaque development is characterized by fibrosis, leading to a thick fibrous cap and a small lipid core (45)(Fig. 9). Fibrotic changes and vessel shrinkage, while causing more severe luminal stenosis, render lesions more resistant to rupture and can be regarded as part of a healing process. The progression of lesions through repeated cycles of vulnerability and rupture, with subsequent lesion stabilization, is probably a common pattern.
If future events are to be avoided, atheroma formation, inflammation and rupture must be prevented. There is already evidence that lipid-lowering therapies can interrupt the atherosclerotic disease process, significantly influencing outcome in ACS patients.
Lipid-lowering and plaque progression
Percutaneous coronary intervention reduces the severity of angina by improving coronary blood flow in the target vessel. However, the effect of lipid-lowering agents on cardiac events is far greater than their effect on the severity of angiographic stenosis (49,50). Imaging studies have confirmed in vivo that statin therapy reduces plaque size and may change plaque composition without substantial changes in luminal area (20,51), suggesting that lipid-lowering therapy may regress or stabilize lipid-rich lesions. Unlike PCI, systemic therapies treat all of the potentially vulnerable plaques in the coronary arteries, not simply the focal lesion responsible for the ACS. Statins do not substantially change luminal area, but they decrease the incidence of death and MI by stabilizing atherosclerotic plaques throughout the coronary bed.
A small IVUS study examined the progression of atherosclerosis after three years of treatment with pravastatin or diet in mildly diseased coronaries (52). During follow-up, plaque area increased by 41% in the control group but decreased by 7% in the treatment group. In a more recent serial IVUS study, patients (n = 131) were randomized to treatment with high-dose atorvastatin or “usual care,” which could include statin therapy (51). After 12 months, mean low-density lipoprotein-cholesterol was reduced from 155 to 86 mg/dl in the atorvastatin group and from 166 to 140 mg/dl in the usual-care group. Mean absolute plaque volume showed a larger but insignificant increase in the usual-care group compared with the atorvastatin group (9.6 mm3vs. 1.2 ± 30.4 mm3). Plaque echogenicity increased to a larger extent in the atorvastatin group than in the usual-care group.
Larger, ongoing serial regression-progression trials include the Reversal of Atherosclerosis with Aggressive Lipid lowering (REVERSAL) trial, which is comparing two lipid-lowering regimens, and the Norvasc for Regression of Manifest Atherosclerotic Lesions (NORMALISE) trial, which is evaluating amlodipine versus enalapril, or placebo (50). The primary end point of these trials is plaque burden, not lumen size. Although the relationship between changes in plaque burden and clinical events remains to be established, the potential of anti-atherosclerotic therapies to limit plaque growth or induce regression is very promising.
Lipid lowering and plaque stability
The Myocardial Ischemia Reduction with Aggressive Cholesterol Lowering (MIRACL) trial randomized 3,086 patients with ACS or NSTEMI to atorvastatin (80 mg/day) or placebo initiated 24 to 96 h after hospital admission. In the atorvastatin group, mean low-density lipoprotein cholesterol declined from 124 mg/dl (3.2 mmol/l) to 72 mg/dl (1.9 mmol/l). The primary end point (16 weeks) was death, non-fatal acute MI, cardiac arrest with resuscitation, or recurrent symptomatic myocardial ischemia with objective evidence and requiring emergency rehospitalization. A primary end point event occurred in 228 patients (14.8%) in the atorvastatin group and 269 patients (17.4%) in the placebo group (relative risk, 0.84; 95% confidence interval, 0.70 to 1.00; p = 0.048). There were fewer strokes in the atorvastatin group than in the placebo group (12 vs. 24 events; p = 0.045).
How might anti-atherosclerotic therapy improve early outcomes in ACS? Benefits are probably not derived strictly through regression of fixed stenoses (49). Libby and colleagues has demonstrated that a low-fat diet or treatment with lipid-lowering agents can reduce macrophage numbers and MMPs and increase collagen contents, thereby limiting oxidized low-density lipoprotein in the vessel wall and limiting the transformation of plaques to an unstable morphology (53–55). In experimental animals, a low-cholesterol diet reduced both the amount and activity of tissue factor, the most thrombogenic component of plaques. Lowering lipids also reduces expression of noxious inflammatory stimuli (i.e., interferon, tumor necrosis factor) that have the potential to destabilize the atheroma and influence vasomotor tone through effects on nitric oxide synthase (55). In experimental animals, statins reduce macrophage accumulation, inflammation, and proteolytic activity (54). Similarly, a study of primates treated with statins found decreased coronary vasomotor tone and reduced numbers of intimal and medial macrophages (54).
The profound effects of anti-atherosclerotic therapies on plaque morphology and behavior render such treatments ideal for ACS patients. Reductions in atheroma inflammation, collagen degradation, tissue factor expression, and vasomotor tone have the potential to treat not only a single lesion, as in PCI, but also the entire systemic disease process. Entirely new classes of atherosclerosis-modulating therapies are under development, including agonists of the peroxisome proliferator-activated receptor (PPAR) system, including PPAR alpha agents that modulate lipid metabolism and PPAR gamma agonists that modulate both lipid and glucose metabolism. New anti-atherosclerotic therapies include inhibitors of acyl-CoA:cholesterol O-acyltransferase that prevent accumulation of cholesterol ester in the macrophage, and inhibitors of cholesterol ester transfer protein that can raise high-density lipoprotein cholesterol by 50% or more.
Ultimately, medical therapy and PCI should be considered as complementary rather than opposing strategies in the treatment and management of atherosclerosis and ACS. Patients are likely to benefit to the greatest extent from intervention when it is performed in combination with aggressive medical therapy.
☆ Please refer to the Trial Appendix at the back of this supplement for the complete list of Clinical Trials.
- acute coronary syndrome(s)
- coronary artery disease
- external elastic membrane
- intravascular ultrasound
- matrix metalloproteinase
- non–ST-segment elevation myocardial infarction
- percutaneous coronary intervention
- peroxisome proliferator-activated receptor
- remodeling ratio
- Received July 9, 2002.
- Revision received October 9, 2002.
- Accepted October 31, 2002.
- American College of Cardiology Foundation
- 1.↵National Center for Health Statistics. Detailed diagnoses and procedures: National Hospital Discharge Survey, 1996. Hyattsville, MD: National Center for Health Statistics; 1998:13; data from Vital and Health Statistics
- 4.Use of a monoclonal antibody directed against the platelet glycoprotein IIb/IIIa receptor in high-risk coronary angioplasty. The EPIC Investigation. N Engl J Med 1994;330:956–61
- 9.↵Effects of tissue plasminogen activator and a comparison of early invasive and conservative strategies in unstable angina and non–Q-wave myocardial infarction. Results of the TIMI IIIB trial. Thrombolysis In Myocardial Ischemia. Circulation 1994;89:1545–56
- Boden W.E.,
- O’Rourke R.A.,
- Crawford M.H.,
- et al.
- Zir L.M.,
- Miller S.W.,
- Dinsmore R.E.,
- Gilbert J.P.,
- Harthorne J.W.
- Galbraith J.E.,
- Murphy M.L.,
- Desoyza N.
- Vlaodaver Z.,
- French R.,
- van Tassel R.A.,
- Edwards J.E.
- Corti R.,
- Fayad Z.A.,
- Fuster V.,
- et al.
- Topol E.J.,
- Nissen S.E.
- Little W.C.,
- Constantinescu M.,
- Applegate R.J.,
- et al.
- Braunwald E.,
- Antman A.M.,
- Beasley J.W.,
- et al.
- Schoenhagen P.,
- Ziada K.M.,
- Kapadia S.R.,
- Crowe T.D.,
- Nissen S.E.,
- Tuzcu E.M.
- Pasterkamp G.,
- Schoneveld A.H.,
- van der Wal A.C.,
- et al.
- Schoenhagen P.,
- Vince D.G.,
- Ziada K.,
- et al.
- Yamagishi M.,
- Terashima M.,
- Awano K.,
- et al.
- Asakura M.,
- Ueda Y.,
- Yamaguchi O.,
- et al.
- Newby A.C.,
- Libby P.,
- van der Wal A.
- 32.↵Ridker PM. Role of Inflammatory biomarkers in prediction of coronary heart disease Lancet 2001;358:971–6
- Tuzcu E.M.,
- Hobbs R.E.,
- Rincon G.,
- et al.
- Falk E.
- Tuzcu E.M.,
- Kapadia S.R.,
- Tutar E.,
- et al.
- Falk E.,
- Shah P.K.,
- Fuster V.
- Loree H.M.,
- Kamm R.D.,
- Stringfellow R.G.,
- et al.
- Amento E.P.,
- Ehsani N.,
- Palmer H.,
- et al.
- Brown B.G.,
- Zhao X.Q.,
- Sacco D.E.,
- Albers J.J.
- Jukema J.W.,
- Bruschke A.V.,
- van Boven A.J.,
- et al.
- Schartl M.,
- Bocksch W.,
- Koschyk D.H.,
- et al.
- Aikawa M.,
- Rabkin E.,
- Okada Y.,
- et al.
- Williams J.K.,
- Sukhova G.K.,
- Herrington D.M.,
- Libby P.