Results of Percutaneous Transluminal Coronary Angioplasty in Unstable Angina and Non–Q-Wave Myocardial Infarction
Observations from the TIMI IIIB Trial
Background This report describes the results of percutaneous transluminal coronary angioplasty (PTCA) in the Thrombolysis in Myocardial Ischemia (TIMI) IIIB Investigation.
Methods and Results PTCA was performed before hospital discharge in 444 of 1473 patients with either unstable angina pectoris or non–Q-wave myocardial infarction (NQWMI) enrolled in TIMI IIIB. Angiographic success was observed in 96.1% of patients. For the entire cohort, the cumulative incidences of death and infarction at 1 year were 2.0% and 8.2%, respectively. The periprocedural incidence of myocardial infarction was 2.7%; emergency coronary bypass surgery, 1.4%; and death, 0.5%. By 1 year of follow-up, 122 patients (28.0%, Kaplan-Meier) had an additional revascularization procedure, 75 (61.5%) had PTCA only, 30 (24.6%) had coronary bypass surgery only, and 17 (13.9%) had both procedures. The results of PTCA were not improved by routine pretreatment with intravenous tissue plasminogen activator (TPA). Periprocedural myocardial infarction was more common among patients receiving TPA than placebo (odds ratio [OR], 2.19; P=.03) and among those with unstable angina than those with NQWMI (OR, 15.5; P=.007). No difference in outcome was observed when patients were analyzed according to age (OR, 1.06; P=.092) or sex (OR, 1.54; P=.51). Variables predictive of poor outcome were PTCA within the first 24 hours of enrollment, PTCA site being the left anterior descending coronary artery, and unsuccessful angiography.
Conclusions In TIMI IIIB, PTCA was performed for patients with unstable angina and NQWMI with a very high rate of angiographic success and a low incidence of complications. By 1 year, repeat revascularization was performed in 28.0% of patients. Routine pretreatment with thrombolysis did not enhance outcome.
Balloon angioplasty is recognized as a method of coronary revascularization for patients with acute myocardial ischemic syndromes. The feasibility of PTCA and its potential to normalize altered coronary circulatory dynamics and relieve ischemia has been well documented. Fewer data are available, however, that characterize PTCA for acute ischemia after recent advances in medical therapy and refinements in the technique of PTCA.
The TIMI IIIB Trial evaluated the value of intravenous thrombolytic therapy and a strategy of routine early cardiac catheterization with revascularization for patients presenting with unstable angina and NQWMI.1 No benefit was derived from thrombolysis. In comparison with a strategy of “watchful waiting,” the early use of an invasive strategy was associated with fewer days of hospitalization, with no influence on the incidence of death or recurrent MI.
PTCA was an integral component of both the invasive and conservative strategies for patients in TIMI IIIB. When technically suitable, the procedure was the preferred revascularization technique for patients assigned to the invasive strategy and in response to recurrent ischemia for those managed conservatively. The primary purpose of this report is to characterize the clinical and coronary angiographic characteristics of patients in TIMI IIIB having PTCA before hospital discharge and their in-hospital and 1-year clinical outcomes. In addition, the influence of selected clinical features, the extent of coronary disease, and pretreatment with thrombolytic therapy on clinical outcome are reported.
Details of TIMI IIIB have been published.1 In brief, between October 1989 and June 1992, the study enrolled 1473 patients who experienced acute chest discomfort at rest thought to be due to myocardial ischemia and lasting >5 minutes but <6 hours. This discomfort was associated with ECG changes compatible with acute myocardial ischemia but not transmural infarction and/or those that occurred in a patient with known coronary artery disease. Reasons for excluding patients included history of recent coronary angiography (30 days), MI (21 days), or PTCA (6 months); any CABG; or contraindications to thrombolytic therapy.
The study design for TIMI IIIB was a 2×2 factorial randomization to either thrombolysis or placebo and to either an invasive or a conservative management strategy. Patients assigned to thrombolysis received intravenous TPA (Activase, Genentech, Inc), 0.8 mg/kg (maximum, 80 mg) over a period of 90 minutes, with the initial one third of the total dose (not to exceed 20 mg) given as a bolus. Patients assigned to the invasive strategy routinely underwent diagnostic cardiac catheterization within 18 to 48 hours of enrollment. If coronary anatomy was suitable (see below), PTCA was performed at the time of catheterization. Revascularization by means of CABG was recommended for patients with significant coronary artery disease and anatomic features unsuitable for PTCA. Patients assigned to the conservative strategy did not undergo cardiac catheterization or revascularization unless they failed initial medical therapy.1
Cardiac Catheterization, PTCA, and CABG
Cardiac catheterization was performed according to a standardized protocol developed for the investigation. Once catheterization was completed, angiograms were reviewed immediately to determine the presence and extent of coronary artery disease. If significant (>60% diameter reduction by visual inspection) disease was present, an attempt was made to determine the culprit lesion(s), ie, the lesion(s) most likely responsible for the acute ischemic syndrome.2 This judgment was based on findings from the coronary angiogram, the left ventriculogram, the ECG, and any other potentially relevant clinical data.
The decision regarding the suitability of the patient for revascularization was then made. PTCA was considered the primary means of achieving revascularization when appropriate. In patients with single-vessel disease, PTCA was recommended if the diseased vessel was identified as the culprit vessel and exclusion criteria for PTCA were not met. If multivessel disease was present, PTCA was recommended if one or more lesions could be identified as culprit lesions and each was suitable for PTCA. Exclusion criteria for PTCA included (1) presence of a >50% left main stenosis, (2) the likely occurrence of hemodynamic collapse should the PTCA vessel become occluded, (3) presence of three-vessel disease in association with seriously impaired left ventricular systolic function (ejection fraction <40%), or (4) anatomic features of the PTCA vessel indicating the likelihood of an unsuccessful result or high-risk PTCA.
The objective of PTCA was to maximize the relief of ischemia while minimizing the chance of procedure-related complications. The precise manner by which PTCA was performed was left to operator discretion with commercially available balloon angioplasty equipment.
Definitions and End Points
Procedures were emergent if PTCA was performed as an emergency for an unstable patient who required immediate (within hours) revascularization. An urgent PTCA was one performed within 7 days for unstable patients or for stable patients having a complication within the previous 14 days considered to be at high risk for a future adverse event (eg, MI). The remaining PTCA procedures were classified as elective, that is, stable patients with no pressing clinical reason for PTCA to be performed immediately.
Coronary stenotic lesions were classified by the central angiographic core laboratory according to vessel location, severity (percent diameter reduction by electronic calipers), and TIMI flow grade.3 PTCA lesion outcome was classified as (1) complete success (a ≥20% improvement in luminal diameter narrowing and a residual stenosis <50% with no decline in TIMI flow grade), (2) partial success (either a ≥20% improvement in luminal diameter narrowing or a residual stenosis <50% with no decline in TIMI flow grade), or (3) failed (a <20% improvement in luminal narrowing and a residual stenosis ≥50% or a decline in TIMI flow grade).
Untoward events occurring during or after PTCA that were noted included death (from any cause), nonfatal MI, transient or sustained abrupt coronary artery closure, congestive heart failure, cardiogenic shock, cardiac tamponade, hemorrhage requiring transfusion, systemic arterial occlusion, hypotension requiring treatment, transient ischemic attack or stroke, hypersensitivity reaction, respiratory failure, renal failure requiring dialysis, and emergency CABG. Emergency CABG was any CABG performed within 24 hours of PTCA. MI associated with PTCA was defined as the occurrence of ischemic pain lasting >20 minutes, an elevation in CK to more than two times the upper limit of normal or in CK-MB above the upper limit of normal, or the development of new Q waves. Congestive heart failure was defined as the presence of dyspnea, pulmonary rales, and radiographic evidence of pulmonary vascular congestion.
Comparisons of distributions of baseline characteristics and clinical course variables were performed with χ2 tests for categorical data and t tests for continuous variables. Event rates were estimated by the Kaplan-Meier method,4 and comparisons were performed by use of the log-rank statistic.5 This statistic is sensitive to differences in the cumulative hazard rates in the groups being compared and has a tendency to weight “failure” at the end of the distribution more heavily than the front. The Gehan statistic was also used in these comparisons, but the difference in probability values between the two statistics was in the third decimal place and in no way changes the inferences derived from the log-rank test.
Multivariate analysis of categorical dependent variables was performed with a logistic regression model.6 A stepwise addition routine was used to determine the independent variables with unique associations with the dependent variables.
Multivariate comparisons of time-to-event variables were performed with Cox proportional-hazards models.7 Again, a stepwise addition routine was used to determine the independent variables with unique associations with the dependent variable.
Because of the large number of statistical comparisons performed in TIMI IIIB, values of P<.01 were considered to provide some evidence for an association and values of P<.001 to provide strong evidence of an association.
PTCA was attempted before hospital discharge in 444 of 1473 enrolled patients. Two hundred seventy-six of these procedures were performed on patients randomized to the invasive strategy, and 168 procedures were performed on patients randomized to the conservative strategy. The baseline clinical and angiographic features of these patients are displayed in Table 1⇓. For the most part, the patients were middle-aged males with a known history of coronary disease and prior angina. Abnormalities of the resting ECG were present in 93.7%. On average, left ventricular function was mildly depressed. The majority of patients had single-vessel coronary disease. A culprit artery was identifiable in 85.4% of patients and was most commonly the left anterior descending coronary artery. Total occlusion of the culprit artery (TIMI grade 0 or 1) was observed in 14.2% of patients.
PTCA Procedural Outcome
PTCA was attempted in 411 arteries (511 lesions) of the 444 patients undergoing the procedure (Table 2⇓). The highest percentage of lesions was in the left anterior descending artery, whereas the remainder were equally distributed between the circumflex and right coronary arteries. After PTCA, the mean value of arterial diameter narrowing in the treated lesion was reduced from 83.2±11.3% to 25.0±17.4%. Angiographic success (complete plus partial improvement) was achieved in 96.1% of the treated lesions. Complete success was achieved in 95.2% and partial success in 0.9%. PTCA was considered to be “failed” in 3.9%.
Within 24 hours of PTCA, 2 patients (0.5%) died. MI occurred and emergency CABG was performed in 2.7% and 1.4% of patients, respectively. No patient experienced a stroke within this interval. The rate of any serious complication (cumulative of death/MI/stroke) was 3.2%. By 1 year, the cumulative event rates were death, 2.0%; MI, 8.2%; and CABG, 10.8%. The combined rate of death/MI/stroke was 9.3%.
At 6 weeks and 1 year, 10% and 28% of patients, respectively, had an additional revascularization procedure. Repeat PTCA was performed in 21.2% by 1 year, and 10.8% of patients received CABG (17 patients underwent both PTCA and CABG).
There was no difference in the angiographic outcome of lesions treated with PTCA according to whether patients had routinely received intravenous TPA or placebo. At 42 days and at 1 year, however, the incidence of MI was greater in TPA patients (42 days: TPA, 8.4% and placebo, 3.5%; 1 year: TPA, 11.3% and placebo, 5.3%; P=.03). There were no differences in lesion or clinical outcome according to whether PTCA-treated patients were assigned to the conservative or invasive strategy.
Unstable Angina Versus NQWMI
Of the patients who underwent PTCA, 165 initially presented with NQWMI and 278 with unstable angina (Table 3⇓). Coronary anatomy was similar between the two groups, with two exceptions. First, the circumflex coronary artery was more commonly the site of the culprit lesion in NQWMI patients and was therefore more frequently the site of PTCA. In unstable angina patients, the most common site was the left anterior descending coronary artery. Second, total occlusion of the culprit lesion (TIMI flow grade 0 or 1) occurred more often in NQWMI patients.
Before the PTCA procedure, patients with NQWMI had culprit lesions that were more severe, with lower grades of antegrade flow, than those with unstable angina. After PTCA, however, the final flow grade and stenosis severity were similar. Regardless of the type of acute ischemic syndrome, PTCA angiographic success rates were similar and high, 94.8% in the NQWMI patients and 96.9% in the unstable angina patients.
By 24 hours after PTCA, MI was observed in 4.3% of unstable angina patients, compared with none of the NQWMI patients (P=.007). This difference persisted at 42 days (P=.018) but was less pronounced at 1 year (P=NS). There were no differences in the rates of death, stroke, or CABG between patients with NQWMI and unstable angina.
The majority (55.4%) of PTCA procedures were elective, 38.5% were urgent, and 6.1% were emergent. The clinical baseline characteristics were similar in these two groups. Total occlusion of the culprit artery, however, was more common among patients having emergent PTCA. Although rates of angiographic success were similar, significant differences were noted in clinical outcome. The incidences of death and MI were each higher in the emergent PTCA group than in the others at 42 days and 1 year. There was no additional death or infarction within the emergent PTCA group beyond 42 days. No differences in outcome were observed between patients having elective compared with urgent PTCA.
Neither the angiographic outcome of PTCA nor the frequency of adverse clinical events was influenced by age, sex, or the location of the culprit lesion. At 1 year, however, considerably more patients (16.6%) with left anterior descending culprit lesions compared with other sites required CABG (right coronary artery, 8.4%; left circumflex artery, 4.5%; P=.004). Also, the need to perform CABG was directly related to the extent of disease (P=.020). Revascularization by CABG was 8.9% for single-vessel disease, 12.0% for two-vessel disease, and 24.0% for three-vessel disease. Also, death within 42 days (3.8% versus 0.5%, P=.049) and at 1 year (7.7% versus 1.0%, P=.005) occurred more often in patients with left ventricular function ≤40% than in those >40%.
Fifty-three patients had total occlusion (TIMI flow grade 0 or 1) of their culprit artery before PTCA. These patients were more likely to have presented with NQWMI (60.4% versus 36.1%, P<.001) and demonstrated lower values of left ventricular ejection fraction (55.2% versus 59.8%, P=.017) than those with patent culprit arteries. PTCA was less often successful in patients with a totally occluded culprit artery than in those with a patent culprit artery (82.9% versus 97.9%, P<.001). Although no differences in early clinical outcome were present between these two groups, patients with closed culprit arteries were more than twice as likely to experience an MI by 1 year of follow-up.
Multivariate Predictors of Outcome
The Cox proportional-hazards model was used to detect those factors that were independently predictive of an untoward event after PTCA. Untoward events included death, MI, stroke, and CABG. Emergency PTCA was the most powerful predictor of experiencing an untoward event. Less influential but significant was PTCA being attempted on the left anterior descending coronary artery and a history of hypertension.
The variables listed in Tables 1 and 2⇑⇑ were included in a logistic regression procedure with a forward variable selection routine to determine factors associated with a successful angiographic outcome. TIMI flow grade before PTCA was the only independent predictor of this result.
The ability of PTCA to relieve coronary narrowing and myocardial ischemia in patients with unstable angina pectoris was first described in 1981.8 Since then, PTCA has become a well-established technique of coronary revascularization for this and other acute ischemic syndromes.9 10 11 12 13 14 15 TIMI IIIB provides a contemporary assessment of PTCA for patients with unstable angina and NQWMI. Furthermore, it evaluates the influence of routine pretreatment with thrombolysis and analyzes PTCA outcome for patients with NQWMI as well as those with unstable angina.
PTCA in TIMI IIIB was performed with a very high rate of angiographic success. Substantial relief of coronary arterial stenosis was achieved in 96.1% of patients. This rate of PTCA success exceeds that generally reported for large cohorts of patients with acute myocardial ischemia.9 10 11 12 13 14 15 Importantly, this high rate of angiographic success was accomplished with a low incidence of major cardiovascular complications. Reasons for these results may include recent refinements in procedural technology, standardized use of aspirin and heparin, and the selection of only PTCA operators who had considerable experience and records of excellence in performing PTCA.
Follow-up data through 1 year after study entry were also available. Major untoward cardiac events were uncommon. Cumulative 1-year mortality was only 2.0%. MI between 6 weeks and 1 year occurred in only 2.3% of patients who did not have a confirmed MI between enrollment and 6 weeks. Recurrent ischemia, on the other hand, was more frequent. Rehospitalization by 1 year was required for more than one third of patients. A repeat revascularization procedure was performed in 28% and was most often (21.2%) repeat PTCA. About 1 in 10 patients required CABG within 1 year. Coronary angiographic data were not available for the patients who required repeat revascularization. Given the chronology of their recurrent ischemia, however, recurrence of the angioplastied lesion was most likely responsible.16 Recently, intracoronary stents have been shown to reduce the incidence of recurrent arterial stenosis and ischemic events after PTCA.17 It is possible that newer advances such as this will impact the durability of PTCA and improve the late results observed in TIMI IIIB.
Acute thrombosis has been implicated as a component of coronary obstruction in patients with unstable angina and NQWMI.18 Furthermore, the presence of coronary thrombosis predisposes to suboptimal PTCA results.19 Accordingly, a reasonable expectation in TIMI IIIB was that patients assigned to receive TPA and heparin would demonstrate a better outcome after PTCA than those who received heparin alone. In this trial, however, no benefit in outcome was observed after TPA. Surprisingly, the prevalence of totally occluded culprit arteries and the degree of lesion severity before PTCA were similar in TPA and placebo groups. Also, patients pretreated with TPA were more likely to experience an MI after the procedure. This finding was especially true for patients presenting with unstable angina. The relationship between thrombolytics and periprocedural infarction is not unique to TPA; a similar association has now been reported with urokinase.20
A definitive explanation for the unfavorable influence of intravenous TPA is not apparent.1 20 Administration of thrombolytics with PTCA may increase the risk of plaque hemorrhage21 or stimulate new thrombus formation at the dilated site.22 These theories are consonant with the lower acute angiographic success rate observed in TIMI IIIB among TPA-treated patients compared with placebo-treated patients. Another possible explanation is that thrombi associated with unstable angina are composed mainly of platelets rather than fibrin.23 The favorable influence of 7E3, a monoclonal antibody directed against the platelet GP IIb/IIIa receptor, on PTCA outcome in unstable angina supports this concept.24 Such thrombi may not be responsive to thrombolytics. In fact, some have suggested that TPA can potentiate platelet aggregation.25 Finally, thrombolytics may predispose to embolization of thrombus originating from the culprit lesion and cause infarction by obstructing the distal coronary circulation.
It should be noted that the results of TIMI IIIB do not indicate that thrombolytics are valueless for patients with acute ischemia undergoing PTCA. Recent reports indicate that the effectiveness and safety of PTCA are enhanced in selected patients with visible coronary thrombus when PTCA is preceded by intracoronary thrombolysis.26 27
TIMI IIIB provided a unique opportunity to compare the results of PTCA for patients presenting with unstable angina with those with NQWMI. NQWMI patients more often had total occlusion of their culprit artery, and the culprit artery was more often the circumflex artery. This latter finding may not be surprising, because some NQWMI patients may have been patients with true posterior transmural MI, demonstrating anterior precordial lead ST-segment depression rather than elevation with no Q-wave formation on presentation.
Although NQWMI patients had more acute and severe coronary disease, the outcome of PTCA in NQWMI patients was similar to that of unstable angina patients but for one exception. A significant difference was noted in the rate of periprocedural infarction; patients with unstable angina had a 4.3% incidence, whereas no patient in the NQWMI group experienced a second infarction. The basis for this difference relates more to the infarction rate associated with NQWMI being low rather than the unstable angina rate being high. The rate of infarction among the unstable patients in TIMI IIIB is within the range of 2.5% to 10% previously reported for patients with unstable angina.9 12 13 14 15 One possible explanation for the low rate of infarction in the NQWMI patients may relate to the ability to cause and/or detect additional infarction once acute infarction has already occurred. Abrupt coronary closure or distal embolization arising from PTCA may not result in further infarction in a necrotic zone of myocardium.
In TIMI IIIB, the results of emergency PTCA were less favorable than PTCA in nonemergent circumstances. Although only a small fraction of PTCA procedures were emergent in TIMI IIIB, the rates of periprocedural death and infarction were considerably higher than in other TIMI IIIB patients. PTCA within 24 hours was an independent predictor of the occurrence of a subsequent cardiovascular event. This relationship does not appear to be attributable to technical failure of PTCA, because the angiographic success rate for emergent patients was 92.3%. A more likely explanation relates to the very high prevalence of total coronary occlusion present at the outset of the PTCA procedure in the emergency group. This finding suggests that many of these emergent patients had progressed to transmural infarction before PTCA and that PTCA was being performed as a “rescue” revascularization procedure. The high mortality rate after PTCA may thus relate to the acuteness and severity of the infarction rather than consequences of the PTCA procedure.
Sex has been identified as a factor capable of influencing the outcome of PTCA. The rates of abrupt closure, infarction, and death have been reported to be higher in women than men.28 29 30 Such an effect of sex on PTCA outcome was not observed in TIMI IIIB. Refinement in the techniques of PTCA and increasing familiarity with the use of PTCA in women may explain the difference between the results of PTCA in TIMI IIIB and those of prior reports. Also contrary to prior reporting, advanced age was not a risk factor for PTCA failure or complications in TIMI IIIB.28 31
Angiographic success and periprocedural complication rates were unaffected by the culprit coronary artery site and the extent of coronary disease. However, the need for subsequent CABG was significantly greater among patients having the culprit lesion located in the left anterior descending coronary artery and in those with multivessel disease than in others. These relationships may relate to the relatively high rate of restenosis associated with PTCA of the left anterior descending coronary artery.32 33 CABG during the year after PTCA was more common among patients with multivessel than those with single-vessel disease. This finding may relate to the inability of PTCA to achieve complete revascularization in patients with multivessel coronary disease34 or to the fact that multilesion PTCA presents a greater opportunity for restenosis to occur.
In TIMI IIIB, total occlusion of the culprit artery was the only angiographic feature associated with a reduced incidence of PTCA success. PTCA was successful in 97.9% of arteries not totally occluded compared with 82.9% of those that were totally occluded. This association between PTCA success and total coronary occlusion has been observed previously. In fact, reported success rates for totally occluded arteries typically range from 42% to 72%, lower than those observed in TIMI IIIB.35 Presumably, the higher rate of success for total occlusions in TIMI IIIB compared with those reported earlier may relate to a short duration of the total occlusion in patients with acute ischemia. Certainly, the PTCA success rate for totally occluded arteries in TIMI IIIB was not so low as to discourage the use of PTCA in this situation.
Impaired left ventricular function has been identified as an independent predictor of mortality in patients with coronary artery disease.36 This relationship was also observed in TIMI IIIB patients who underwent PTCA. Those with a left ventricular ejection fraction <40% had a 1-year mortality of 7.7% compared with those with higher values, in whom mortality was only 1.0%. Surgical revascularization has been demonstrated to reduce mortality rates in patients with advanced coronary artery disease and substantially impaired left ventricular function.37 Whether PTCA had a salutary or adverse effect on survival of patients with impaired left ventricular function cannot be determined from this investigation.
Before the results of PTCA in TIMI IIIB can be extrapolated to the noninvestigational clinical setting, certain issues need to be considered. First, PTCA in TIMI IIIB was performed by invasive cardiologists who were thoroughly experienced in the technique and who had met or exceeded the performance criteria stipulated by the study protocol. The trial did not address the issue of whether the results of TIMI IIIB could be achieved by physicians lacking these qualifications. Second, intravenous TPA was administered before PTCA in one half of patients. TPA administration was associated with an increased incidence of periprocedural infarction. It is thus possible that the rate of infarction observed in TIMI IIIB may be higher than in the absence of routine TPA administration.
The results of TIMI IIIB support the use of PTCA for patients with acute myocardial ischemia. The likelihood of the coronary stenosis responsible for the acute ischemic syndrome being relieved exceeds 90%, and the incidence of procedural cardiovascular complications is low. Furthermore, PTCA is applicable to a large proportion of patients hospitalized for this disorder. Physicians who are responsible for the longitudinal care of patients treated with PTCA should be alert to the possibility of recurrent ischemia and the need for further revascularization. Research directed toward enhancing the durability of the initial PTCA procedure should be encouraged.
Selected Abbreviations and Acronyms
|CABG||=||coronary artery bypass graft surgery|
|NQWMI||=||non–Q-wave myocardial infarction|
|PTCA||=||percutaneous transluminal coronary balloon angioplasty|
|TIMI||=||Thrombolysis in Myocardial Ischemia|
|TPA||=||tissue plasminogen activator|
The authors acknowledge the secretarial assistance of Arlene Grant in the preparation of this manuscript and of Nicholas Miele in assisting with the processing and interpretation of left ventricular and coronary cineangiograms.
*The TIMI IIIB Investigators are listed in Reference 1.
- Received April 14, 1996.
- Revision received September 3, 1996.
- Accepted September 9, 1996.
- Copyright © 1996 by American Heart Association
The TIMI IIIB Investigators. Effects of tissue plasminogen activator and a comparison of early and conservative strategies in unstable angina and non–Q-wave myocardial infarction: results of the TIMI IIIB Trial. Circulation. 1994;89:1545-1556.
Walker SH, Duncan DB. Estimation of the probability of an event as a function of several independent variables. Biometrika. 1967;54:167-179.
Cox DR. Regression models and life tables. J R Stat Soc B. 1972;34:187-200.
Bentivoglio LG, Detre KM, Yeh W, Williams DO, Kelsey SF, Faxon DP. Outcome of percutaneous transluminal coronary angioplasty in subjects of unstable angina pectoris: a report from the 1985-1986 National Heart, Lung and Blood Institute Percutaneous Transluminal Coronary Angioplasty Registry. J Am Coll Cardiol. 1994;24:1195-1206.
Fischman DL, Leon MB, Baim DS, Schatz RA, Savage MP, Penn I, Detre M, Veltri L, Ricci D, Nobuyosni M, Cleman M, Heuser R, Almond D, Teirstein PS, Fish RD, Colombo A, Brinker J, Mosses J, Shaknovich A, Hirshfeld J, Bailey S, Ellis S, Rake R, Goldberg. A randomized comparison of coronary-stent placement and balloon angioplasty in the treatment of coronary artery disease. N Engl J Med. 1994;331:496-501.
Sugrue DD, Holmes DR, Smith H-C, Reeder GS, Lane GE, Vlietstra RE, Bresnahan JF, Hammes LN, Piehler JM. Coronary artery thrombus as a risk factor for acute vessel occlusion during percutaneous transluminal coronary angioplasty: improving results. Br Heart J. 1986;56:62-67.
Ambrose JA, Almeida OD, Sharma SR, Torre SR, Marmur JD, Israel DH, Ratner DE, Weiss MB, Hjemdahl-Monsen CE, Myler RK, Moses J, Unterecker WJ, Grunwald AM, Garrell JS, Cowley MJ, Anwar A, Sobolski J. Adjunctive thrombolytic therapy during angioplasty for ischemic rest angina: results of the TAUSA Trial. Circulation. 1994;90:69-77.
Waller BF, Rothbaum DA, Pinkerton CA, Cowley MJ, Linnemeier JJ, Orr C, Ivons M, Helmuth RA, Wills ER, Aust C. Status of the myocardium and infarct related coronary artery in 19 necropsy patients with acute recanalization using pharmacologic (streptokinase r-tissue plasminogen activator) mechanical (percutaneous transluminal coronary angioplasty) or combined types of reperfusion therapy. J Am Coll Cardiol. 1987;9:785-801.
Aronson DL, Chang P, Kessler CM. Platelet-dependent thrombin generation after in vitro fibrinolytic treatment. Circulation. 1992;85:1706-1712.
Benedict CR, Refino CJ, Keyt BA, Pakala R, Paoni NF, Thomas GR, Bennett WF. New variant of human tissue plasminogen activator (TPA) with enhanced efficacy and lower incidence of bleeding compared with recombinant human TPA. Circulation. 1995;92:3032-3040.
Bredlau CE, Roubin GS, Leimgruber PP, Douglas JS, King SB, Gruentzig AR. In-hospital morbidity and mortality in patients undergoing elective coronary angioplasty. Circulation. 1985;72:1044-1052.
Kelsey SF, James M, Holubkov AL, Holubkov R, Cowley MJ, Detre KM. Results of percutaneous transluminal coronary angioplasty in women: 1985-1986 National Heart, Lung, and Blood Institute's Coronary Angioplasty Registry. Circulation. 1993;87:720-727.
Kelsey SF, Miller DP, Houlubkov R, Lu AS, Cowley MJ, Faxon DP, Detre KM. Results of percutaneous transluminal coronary angioplasty in patients greater than or equal to 65 years of age (from the 1985-1986 National Heart, Lung and Blood Institute's Coronary Angioplasty Registry). Am J Cardiol. 1990;66:1033-1038.
Leimgruber PP, Roubin GS, Hollman J, Cotsonis GA, Meier B, Douglas JS, King SB, Gruentzig AR. Restenosis after successful coronary angioplasty in patients with single-vessel disease. Circulation. 1986;73:710-717.
Williams DO, Baim DS, Bates E, Bonan R, Bost JE, Cowley M, Faxon DP, Feit F, Jones R, Kellet MA Jr, Kelsey SF, Sopko G, Stadius M, Topol EJ. Coronary anatomic and procedural characteristics of patients randomized to coronary angioplasty in the Bypass Angioplasty Revascularization Investigation (BARI). Am J Cardiol. 1995;75:27C-33C.
Meier B. Total coronary occlusion: a different animal? J Am Coll Cardiol. 1991;17(suppl B):50B-57B.
Mock MB, Ringquist I, Fisher D, Davis KB, Chaitman BR, Kouchoukos NJ, Kaiser GC, Alderman E, Ryan TJ, Russel RO, Mullin S, Fray D, Killip T. Survival of medically treated patients in the Coronary Artery Surgery Study (CASS) Registry. Circulation. 1982;66:562-568.