Highlights from the American College of Cardiology Annual Scientific Sessions 2002: March 17 to 20, 2002☆☆☆

Session: Late-breaking clinical trials III 

Study: The Danish Multicenter Randomized Trial on Thrombolytic Therapy Versus Acute Coronary Angioplasty in Acute Myocardial Infarction (DANAMI-2) Trial

Presenter: Dr Henning Rud Andersen, Aarhus University Hospital, Denmark

Background: The primary goal of all reperfusion therapies is rapid and complete reopening of occluded coronary arteries, a concept confirmed by the results of GUSTO-I. Coronary patency, defined as the restoration of normal blood flow (grade 3 flow according to the Thrombolysis In Myocardial Infarction, TIMI classification) in the infarct-related vessel, has been shown to preserve myocardial tissue and improve survival. Patency rates achieved with primary angioplasty cannot currently be obtained with thrombolytic agents. Primary angioplasty, when performed by experienced clinicians, restores normal blood flow in 80% to 95% of patients compared with 50% to 70% achieved by thrombolysis. Successful primary angioplasty has also been proven to reduce the incidence of reocclusion of the infarct-related artery to <10% compared with reocclusion rates up to 30% after successful reperfusion by thrombolytic agents. These findings suggest that primary angioplasty is superior to thrombolytic therapy in achieving early and sustained infarct-related artery patency. Although it has been demonstrated that thrombolytic therapy reduces early mortality rate by 20% to 30%, little is known about the reduction in mortality rates when using primary percutaneous coronary intervention (PCI). Several meta-analytic studies that included >20,000 patients from randomized trials have demonstrated short- and long-term superiority of primary PCI over thrombolytic therapy in the rates of death, nonfatal reinfarction, and stroke after acute myocardial infarction. This observation was reported only in several small randomized and nonrandomized clinical trials that included between 200 and 1100 patients, the GUSTO-IIb angiographic study being the largest randomized trial.

The DANAMI-2 trial was a prospective study of 1575 patients with ST-elevation myocardial infarction (STEMI) randomly assigned to fibrinolysis (100 mg front-loaded tPA) versus primary PCI and stenting. The primary end points were death, reinfarction, or disabling stroke within 30 days. The inclusion criteria were ST elevation >4 mm, random assignment within 12 hours of symptoms, and transfer time <3 hours from the referral hospital to the catheterization laboratory. There was no upper age limit. Patients with contraindications to fibrinolysis, cardiogenic shock, or persistent life-threatening arrhythmias were excluded. The randomization took place in 2 kinds of medical centers: referral hospitals (<3 hours from a PCI facility) and invasive centers.

The median age of patients was 64 years (range 23-96); 52.4% of patients in the fibrinolysis arm and 53.2% of patients in the PCI arm had an anterior wall myocardial infarction. A clear benefit for PCI caused enrollment to be halted.

Results: There was a >40% relative risk reduction for the 30-day primary end point in patients treated with primary PCI (n = 790) compared with the thrombolytic-treated group (n = 782) (8.0% vs 13.7%, P = .0003). The benefit in the primary composite end point result was predominantly driven by a lower rate of recurrent infarction among patients treated with thrombolysis compared with primary PCI (1.6% for PCI vs 6.3% for tPA, P < .0001). Although not statistically significant, there was a trend toward decreased rates of death (6.6% vs 7.6%, P = .35) and disabling stroke (1.1% vs 2.0%, P = .15) for PCI versus tPA, respectively. These benefits were observed in all patients treated with primary PCI irrespective of the site of enrollment. No deaths occurred during transportation from the referral to invasive centers. The rate of arrhythmic complications during transportation was 6.4%. The mean time from the onset of symptoms to arrival at the hospital was 120 minutes for all the patients. The door-to-needle time in the thrombolytic group was <60 minutes (referral centers 54 minutes, invasive centers 45 minutes). The door-to-balloon time was <120 minutes (117 minutes for referral centers and 90 minutes for invasive centers). The mean time of transportation was 25 minutes. The mean time difference between door-to-needle and door-to-balloon for patients from referral centers was 56 minutes.

Interpretation: The results of the DANAMI-2 trial show primary PCI to be superior at 30 days to thrombolysis in patients with STEMI treated with reperfusion within 2 hours of the time of presentation. These findings are consistent with previous observations from several meta-analyses and many small studies. Many investigators agree that primary PCI should be the treatment of choice for acute STEMI because it is believed to achieve reperfusion more rapidly and more completely in a higher number of patients than thrombolysis. However, the relation between death and time to treatment reported in randomized trials has been largely debated. It has been thought that the reperfusion results achieved with PCI in randomized trials conducted in tertiary centers may not be reproducible in community hospitals where the door-to-balloon time is longer. DANAMI-2 is the first large, randomized trial to compare the transfer of patients with large, acute myocardial infarction for PCI with lytic therapy on-site using 100 mg front-loaded tPA. The time interval between arrival at the referral center and the first balloon inflation in the current study (117 minutes) was significantly longer than reported in previous randomized trials (including GUSTO-IIb, in which the median time was 76 minutes). This observation reflects the delay frequently encountered in current clinical practice in hospitals other than tertiary centers. DANAMI-2 demonstrated the superiority of PCI over thrombolysis in patients treated with PCI within 2 hours of presentation as well as the safety of transportation of patients with acute STEMI to undergo PCI. According to the results of DANAMI-2, patients presenting with acute STEMI in non-PCI centers achieve greater benefit when transported and treated with primary PCI compared with those receiving thrombolysis on-site despite 56 minutes of mean door-to-treatment time difference.

Should we then transport patients with acute STEMI for PCI instead of treating them rapidly with thrombolytics if the time difference between door-to-needle and door-to-balloon can be <1 hour? Although questions regarding differences in primary end points in patients who had longer versus shorter transfer times remain to be addressed, the DANAMI-2 results will have a worldwide impact on current clinical practice and will trigger the conduction of other large, randomized, international studies in patients with high-risk acute STEMI.

Numerous series have demonstrated an inverse relationship between death from cardiovascular causes and the number of cardiac procedures performed by individual practitioners or hospitals. Primary PCI for the treatment of acute STEMI is one of the complex technical procedures that requires experienced personnel and optimized hospital logistics. The question of whether primary angioplasty would still be superior to fibrinolytic therapy when performed in low-volume hospitals and by physicians with a low volume of procedures has been constantly debated. The information provided by the DANAMI-2 investigators about the limited expertise of the interventional cardiologists and the limited experience of participating PCI centers has made the success of PCI in this trial more striking. Finally, the study demonstrated an early (30 days) benefit of treatment with PCI over lytic therapy for acute STEMI mainly achieved through significant reduction of reinfarction at 30 days. It remains to be established in subsequent studies whether treatment modality will improve long-term clinical outcome by its superiority in reducing the incidence of reocclusion of the infarct-related artery over time.

Session: Late clinical breaking trial 

Study: INTERACT (INTegrilin and Enoxaparin Randomized assessment of Acute Coronary syndrome Treatment)

Presenter: Dr Shaun Goodman, Canadian Heart Research Centre, Toronto, Ontario, Canada

Background: Several studies have described the importance of anticoagulation in acute coronary syndrome (ACS). In this population, 2 randomized, clinical trials have shown enoxaparin, a low molecular-weight heparin, to be superior to conventional unfractionated heparin (UFH) in preventing the composite of death or cardiac ischemic events. Antithrombotic therapy in ACS is constantly evolving and now includes agents such as the glycoprotein (GP) IIb/IIIa receptor inhibitors. Studies have established the additive benefit of GP IIb/IIIa inhibitors to UFH in ACS. There remain questions regarding the safety and efficacy of the combination of low molecular-weight heparin and GP IIb/IIIa antagonists.

The primary goal of INTERACT was to evaluate the risk of noncoronary artery bypass surgery-related major bleeding at 96 hours with enoxaparin in conjunction with eptifibatide (a GP IIb/IIIa antagonist) compared with UFH with eptifibatide. Secondary end points included the incidence of ischemia at 48 and 96 hours as measured by continuous Holter electrocardiographic monitoring and the composite of death, myocardial infarction, or recurrent ischemia at 30 days.

To be eligible, patients were required to have ischemic chest pain at rest lasting ≥10 minutes within 24 hours of enrollment, associated with either electrocardiographic ST-segment changes (ST depression or transient ST elevation ≥0.1 mV in ≥2 contiguous leads) or cardiac biomarker positivity (troponin I or T ≥3× reference limit or creatine-kinase MB greater than normal). Patients were to take 160 mg aspirin followed by 80 to 325 mg daily with 180 μg/kg eptifibatide IV bolus followed by a 2 μg/kg per minute infusion for at least 48 hours. They were then randomly assigned to receive either 1 mg/kg enoxaparin subcutaneously every 12 hours or 70 U/kg UFH IV bolus (max 5000 U) followed by 15 μg/kg per hour (max 1000 μg/h), adjusted to maintain a targeted aPTT 1.5 to 2.0 times the normal (approximately 50 to 70 seconds) for a minimum of 48 hours. Decisions regarding cardiac catheterization and revascularization were left to the discretion of the investigators.

Results: Seven hundred forty-six patients from Canadian centers participated in the study. The baseline characteristics of the cohort were typical for ACS trials and similar between the 2 study groups. Notably, there were approximately 80% with myocardial infarction by cardiac enzyme definitions. The mean time from symptom onset to random assignment was approximately 6 hours and similar between the 2 groups. Angiography was performed in about two thirds of the patients (64.7% and 62.1% for the enoxaparin and UFH arms, respectively), and percutaneous coronary intervention was performed in approximately 30% and coronary artery bypass graft (CABG) in approximately 12% for both groups. The mean time to angiography was about 100 hours for both arms. Use of concurrent medications was similar among the study groups, with 95% use of aspirin, 15% use of clopidogrel, and 80% use of oral β-blockers.

The primary end point of non-CABG major bleeding by use of the TIMI scale at 96 hours was significantly lower for enoxaparin compared with UFH (1.8% vs 4.6%, P = .03). A difference was apparent at 48 hours (1.1% vs 3.8%, P = .014). There was an increase in minor bleeding with enoxaparin (32.5% vs 24.9%, P = .024), mostly related to ecchymoses at the local injection site and therefore not clinically significant. The incidence of ischemia detected by continuous electrocardiographic monitoring was lower at 48 hours with enoxaparin versus UFH (14% vs 25.1%, P = .0002). Similar ischemic protective effects were found between 48 and 96 hours (12.7% vs 25.9%, P < .0001). The composite of death or reinfarction at 30 days was reduced with enoxaparin compared with UFH (5% vs 9%, P = .031). When adding recurrent ischemia with electrocardiographic changes, the composite event rate continued to favor enoxaparin (8.4% vs 12.6%) but was not statistically significant (P = .064).

Interpretation: INTERACT demonstrated the safety of the up-front use of the combination of enoxaparin and eptifibatide in ACS. These results are similar to the Assessment of Cardioversion Using Transesophageal Echocardiography II (ACUTE-II) trial, which evaluated the safety of enoxaparin with tirofiban, another GP IIb/IIIa antagonist. In high-risk patients with non-ST-elevation myocardial infarction, these results further suggest an improved clinical efficacy with enoxaparin accomplished by the reduction of myocardial ischemia (death, reinfarction, ST changes by Holter electrocardiography monitoring). An analysis of the periprocedural (percutaneous coronary intervention, CABG) anticoagulation treatment of this cohort may provide further insights. Since the design and conduct of INTERACT, an early invasive strategy (cardiac catheterization) in high-risk patients with ACS has been advocated after the publication of Treat angina with Aggrastat and determine Cost of Therapy with an Invasive or Conservative Strategy-Thrombolysis In Myocardial Infarction (TACTICS-TIMI-18). The evaluation of the safety and efficacy of enoxaparin with this early invasive strategy is being addressed in the Superior Yield of the New Strategy of Enoxaparin Revascularization GlYcoprotein IIb/IIIa inhibitors (SYNERGY) trial, currently in progress.

Session: Late-breaking clinical trials I 

Study: Weekly Intervention with Zithromycin for Atherosclerosis and Is Related Disorders (WIZARD): Preliminary Results

Presenters: Dr Michael Dunne, Pfizer Central Research, Groton, Conn, and Dr Christopher M. O'Connor, Duke University Medical Center, Durham, NC

Background: Epidemiologic studies link antibody to Chlamydia pneumoniae with coronary artery disease. This organism has been previously described in atherosclerotic plaques. Animal models suggest initiation and progression of early plaque formation after infection. Further animal studies and pathology suggest that arterial infection with C pneumoniae generates an inflammatory condition consistent with atherogenesis. Azithromycin is a macrolide antibiotic, a dose of which (1.5 g given over the course of 3-5 days) is effective in treating C pneumoniae. Several small studies such as the ROXithromycin Ischemic Syndromes (ROXIS)1 and recently released CLarithromycin in Acute Coronary Syndrome Patients in Finland (CLARIFY)2 studies have shown a benefit of macrolide antibiotics in patients with unstable angina. The Azithromycin in Coronary Artery Disease Elimination of Myocardial Infection with Chlamydia (ACADEMIC) trial,3 however, did not show a reduction in cardiovascular events in a cohort of patients with coronary artery disease treated prospectively with azithromycin. Given these conflicting results, a large-scale outcome trial of treating chlamydia pneumoniae infection with antibiotics was undertaken.

The purpose of this trial was to see if recurrent coronary events could be prevented with antibiotic treatment against C pneumoniae. The trial was a multicenter, randomized, double-blind trial of azithromycin versus placebo. Initial azithromycin administration was a 600 mg daily dose for 3 days, followed by weekly administration for 11 weeks. Inclusion criteria stipulated adults at >6 weeks after myocardial infarction (diagnosed by electrocardiography or elevated creatine phosphokinase) with an IgG titer against C pneumoniae of >1:16. Women were required to be of nonchildbearing age or taking birth control. Exclusion criteria included any revascularization procedure in the prior 6 months or any chronic condition requiring antibiotics.

The primary composite end point included all-cause death, recurrent myocardial infarction, hospitalization for angina, and revascularization procedure. The secondary end points individually tracked time to cardiovascular death, recurrent myocardial infarction, hospitalization for unstable angina/ischemia, and revascularization. The sample size calculations assumed an event rate of 8% in the placebo arm and 90% power to detect a treatment difference of 25% (α = 0.05). The study would conclude with 520 end points. The study protocol was amended in June 2000, and the percentage of treatment difference was lowered to 18.5%. This changed the end point target to 1038 end points.

Results: The trial screened 11,451 patients and enrolled 7724 patients (3868 in the azithromycin arm and 3856 in the placebo arm); 82% of the patients were male, and the average age was 62 years; 21% of patients had diabetes, 55% had hypertension, and 57% had hypercholesterolemia. Family history of coronary artery disease was seen in 33% of the azithromycin group and 31% of the placebo group (P < .05); 23% had myocardial infarction in the last year, and 11% had angina at the time of enrollment; 67% had prior revascularization procedures. There was no baseline difference in the use of aspirin (87%), HMG CoA reductase inhibitors (67%), ACE inhibitors (42.5%), or β-blockers (60.5%). The mean follow-up was 2.1 years.

Primary end point analysis revealed a hazard ratio of 0.93 (0.83-1.05) in the azithromycin group compared with placebo (P = .23). In subgroup analysis, there was a trend toward a positive treatment effect in males, patients with diabetes, and smokers, but no subgroup had a benefit that was statistically significant (with all hazard ratios crossing 1.0). Subgroup analysis included baseline C pneumoniae titer.

Post hoc analyses suggested a possible early treatment benefit with fewer deaths and myocardial infarctions at 6 months. This was not sustained over the observation period of 4 years. The azithromycin group had statistically more gastrointestinal adverse events (diarrhea) and a lower incidence of respiratory infection.

Interpretation: The adjunctive treatment of patients after myocardial infarction with a 12-week course of azithromycin did not reduce major cardiovascular events over a 4-year follow-up. There was no statistically significant benefit of treatment in any subgroup. This trial neither supports nor contradicts the role of C pneumoniae in the pathogenesis of atherosclerosis. A 3-month course of azithromycin was safe and well tolerated. There are ongoing trials of patients with coronary artery disease using macrolide antibiotics for a longer duration of therapy.

Session: Late-breaking clinical trials 

Studies: (1) Atrial Fibrillation Follow-up Investigation of Rhythm Management (AFFIRM) Trial, (2) Rate Control versus Electrical Cardioversion for Persistent Atrial Fibrillation (RACE) Study

Presenters: (1) Dr D. George Wyse, University of Calgary, Calgary, Alberta, Canada, (2) Dr Harry J. Crijns, University Hospital Maastricht, Maastricht, The Netherlands

Background: Atrial fibrillation (AF) is the most common arrhythmia encountered in clinical practice, accounting for >30% of hospitalizations for cardiac rhythm disturbance. It is estimated that approximately 2.2 million Americans have paroxysmal or persistent AF. Cross-sectional studies have found the prevalence of AF increases with age, ranging from <1% among those aged <60 years to >6% among patients aged >80 years. AF is often associated with underlying structural heart disorders, but a substantial proportion of patients with AF have no detectable heart disease. Hemodynamic impairment and thromboembolic events related to AF result in significant morbidity, mortality, and cost.

Among patients with persistent AF, there are 2 fundamental treatment strategies: to restore and maintain sinus rhythm (“rhythm control”), or to allow AF to continue with control of the ventricular rate (“rate control”). Although this decision must be faced often by clinicians because AF is so common, there have been no previous prospective, controlled trials comparing these 2 treatment strategies. The AFFIRM and RACE studies were designed to help clarify the treatment of this common clinical problem.

Results: The prospective, randomized, multicenter AFFIRM study (sponsored by the NHLBI) was conducted at 213 centers in the United States and Canada. Four thousand sixty patients with AF (present for ≥6 hours within the last 6 months, with qualifying episode within 12 weeks of enrollment) were randomly assigned to a rate control or rhythm control strategy. All patients had at least one risk factor for stroke, were eligible for either treatment strategy, and received warfarin anticoagulation (to maintain INR [international normalized ratio] 2.0 to 3.0) at the time of enrollment. Cardioversion was permitted before random assignment; patients were excluded if it was unsuccessful. The trial was conducted on an intention-to-treat basis, and the primary end point was all-cause death.

Among patients randomly assigned to the rhythm control strategy, the choice of antiarrhythmic drug was left to the discretion of the primary treating physician. Changes in doses and/or classes of agents were permitted, and repeat cardioversion(s), focal AF ablations, and pacemakers were allowed. The most common antiarrhythmic drugs used were amiodarone (39%), sotalol (33%), and propafenone (10%). There were frequent changes in the antiarrhythmic drug regimen, and amiodarone was ultimately administered to 60% of study patients. Anticoagulation could be discontinued if sinus rhythm was maintained for 1 month.

Among patients randomly assigned to the rhythm control strategy, therapy consisted primarily of digoxin (51%), β-blockers (49%), and calcium-channel antagonists (41%).

Median duration of follow-up was 3.5 years. At the end of follow-up, 60% of patients in the rhythm control arm and 35% of patients in the rate control arm were in sinus rhythm. Successful rate control was achieved in 80% of patients in the rate control arm. The frequency of warfarin use was high in both arms: >80% and >65% in the rate and rhythm control arms, respectively.

All-cause death was not significantly different between the 2 treatment strategies; however, there were fewer deaths in the group randomly assigned to the rate control strategy (306 in rate control arm vs 356 in rhythm control arm, P = .058). There were no significant differences between groups in several secondary end points, including functional status and quality of life. Patients in the rhythm control arm had higher rates of stroke (7.3% vs 5.7%), hospitalization, and ventricular arrhythmias compared with patients in the rate control arm.

RACE, a prospective, randomized-controlled trial, conducted at 35 clinical centers in The Netherlands, randomly assigned 522 patients with persistent AF to a strategy of electrical cardioversion for rhythm control or pharmacologic rate control. To qualify for enrollment, AF had to be present for ≥24 hours within 12 months of random assignment. The primary end point of the study was a composite of several components: cardiovascular death, hospitalization for heart failure, thromboembolic complications, severe bleeding, pacemaker implantation, and severe drug side effects. All patients received anticoagulation during the study period. The primary hypothesis of the study was that rate control was not inferior to rhythm control in the treatment of patients with AF.

Among patients randomly assigned to the electrical cardioversion and rhythm control strategy, the choice of maintenance antiarrhythmic medication was dictated by the study protocol. Initial treatment used sotalol, with the stepwise substitution of flecanide, propafenone, and amiodarone if necessary. Patients could be “rechallenged” with a previously administered antiarrhythmic drug if a late recurrence of AF occurred after cardioversion.

The median age of patients in the study was 68 years, and >60% were male. The frequency of hypertension was 43% and 55% in the rate and rhythm control groups, respectively.

After a mean of 3 years' follow-up, the frequency of sinus rhythm was 10% in the rate control arm and 40% in the rhythm control arm. The rate of the primary end point was 17.2% in the rate control arm and 22.6% in the rhythm control arm, respectively (absolute difference of 5.4%). Cardiovascular mortality rates were 7.0% and 6.7% for the 2 groups, respectively. Patients in the rhythm control arm had higher frequencies of thromboembolism, heart failure, and adverse drug reactions compared with those in the rate control arm. Among patients with hypertension at the time of enrollment, there was a particularly high rate of the composite primary end point among those randomly assigned to the electrical cardioversion and rhythm control strategy (31% vs 19% in the rate control arm).

The study investigators concluded that a rate control strategy for patients with AF is not inferior to a strategy of electrical cardioversion and antiarrhythmic medications.

Interpretation: These 2 prospective, randomized studies provide important information to guide the clinician in treating patients with AF. Both studies suggest that a strategy of rate control with anticoagulation is acceptable and safe and is at least equivalent to the strategy of cardioversion and maintenance of sinus rhythm with antiarrhythmic medications. The latter strategy may be associated with a higher rate of stroke, which is probably related to the discontinuation of therapeutic anticoagulation therapy after evidence of sinus rhythm.

In the past, reasons for restoration and maintenance of sinus rhythm among patients with AF included the relief of symptoms, prevention of embolism, improvement in quality of life, and prevention of left ventricular dysfunction related to the tachycardia. The results of these trials suggest no advantage (and possible harm) with this strategy, with a similar quality of life and lower rate of thromboembolism among patients receiving rate control and anticoagulation.

These results may be less applicable to younger patients and those with intolerable symptoms related to their AF. Refractory symptoms and the prospect of very long-term anticoagulation may make the strategy of restoration of sinus rhythm more attractive in these patient groups. However, both studies suggest that long-term anticoagulation is necessary to prevent thromboembolic events among patients in whom sinus rhythm is restored. The optimal duration of this anticoagulation is unknown.

For patients with AF, the strategy of rate control and anticoagulation is safe and effective and does not appear to be associated with a poor quality of life and an excessive rate of thromboembolism. For patients who are eligible for both rate control and rhythm control strategies, the use of rate control and anticoagulation is supported by the findings of these trials.

Session: Modeling outcomes and cost 

Study: Does the Evidence Support the Verdict of High Profile Clinical Trials? A Conventional Versus Bayesian Perspective

Presenter: Dr Sanjay Kaul, Cedars-Sinai Medical Center, Los Angeles, Calif

Background: To the baseball savvy, the distinction between Bayes theorem and the more conventional (frequentist) statistical analysis can be explained in the context of forecasting the Cincinnati Reds' Ken Griffey Jr's batting average for the upcoming season. A Bayesian statistician might combine his beliefs (based on, for instance, Griffey's batting average in previous seasons, injury status, age, and so forth) with current data (eg, his spring training average) to arrive at an estimate. On the other hand, a conventional statistician is likely to base his estimate on the current spring training average alone. Although this is a simplistic description of the 2 methodologies, the ongoing debate about their relative merits and limitations in analyzing clinical trials data grows increasingly complex.

To compare the results of conventional statistical analyses with those of Bayesian analyses, Dr Sanjay Kaul and colleagues examined data from 7 high-profile clinical trials in the last 5 years.

Results: The Table I summarizes the results based on the 2 methods. The results were concordant in only 2 of 7 trials, with the conventional and Bayesian analyses converging at a value of P ≤ .001.

Table I. Results of clinical trials using conventional versus Bayesian analyses

	Trial	No.	End point	OR (95% CI)	P (χ2)	P (b)
	CAPRIE	19413	Death/MI/Stroke 1-3y	0.91 (0.83-1.00)	.047	0.07
	EPISTENT	1603	Death 1y	0.42 (0.18-0.97)	.037	0.12
	ESSENCE + TIMI-11B	7081	Death/MI 14 d	0.79 (0.60-0.90)	.016	0.20
	MIRACL	3085	Death/MI/Ischemia 4m	0.83 (0.68-1.00)	.048	0.18
	CAPRICORN	1959	Death at 1.3 y	0.74 (0.57-0.97)	.025	0.31
	HOPE	9297	Death/MI/Stroke 5 y	0.75 (0.68-0.84)	<.001	>0.99
	Lyon diet Heart Study	423	Death/MI 46 m	0.25 (0.13-0.47)	<.001	>0.99

P (χ2), Conventional 2-sided P value based on the χ2 statistic; P (b), posterior probability that the differences are truly different based on a point estimate prior probability of 0.5; MI, myocardial infarction; CAPRIE, Clopidogrel versus Aspirin in Patients at Risk of Ischemic Events; EPISTENT, Evaluation of IIb/IIIa Platelet Inhibitor for STENTing; ESSENCE + TIMI-11B, Efficacy Safety Subcutaneous Enoxaparin in Non-Q-wave Coronary Events + Thrombolysis In Myocardial Infarction 11B; MIRACL, Myocardial Ischemia Reduction with Aggressive Cholesterol Lowering; CAPRICORN, CArvedilol Post infaRct survival COntRol evaluatioN; HOPE, Heart Outcomes Prevention Evaluation.

Interpretation: There is growing consensus among mainstream researchers that conventional P values need to be interpreted with caution. Although Bayesian analyses have the inherent limitation that prior probabilities must be supplied and estimates might not be available in observed data, using them as an adjunct to standard statistical methods may strengthen interpretations of the results.

Session: Late-breaking clinical trials II 

Study: A Prospective, Blinded Trial of B-Type Natriuretic Peptide as a Diagnosis Test for the Emergency Diagnosis of Heart Failure: The Breathing Not Properly (BNP) Multinational Study

Presenter: Dr Peter A. McCullough, Truman Medical Center, Kansas City, Mo, and Dr Alan S. Maisel, San Diego VA Healthcare, La Jolla, Calif

Background: B-type natriuretic peptide is a normally occurring hormone that is secreted by the cardiac atria and ventricles. In pathologic conditions, production of B-type natriuretic peptide rises strongly. In particular, B-type natriuretic peptide elevations have been noted in conditions of increased cardiac wall stress such as congestive heart failure (CHF).1 Previous work has demonstrated the utility of a B-type natriuretic peptide assay for the diagnosis of CHF in the primary care setting.2 B-Type natriuretic peptide elevations are not specific to CHF and have also been reported in acute coronary syndromes, pulmonary embolism, and chronic renal failure.1 To assess the diagnostic accuracy of B-type natriuretic peptide in the diagnosis of CHF, the Breathing Not Properly (BNP) Multinational Study enrolled adult patients with acute dyspnea to the emergency departments of 7 medical centers. Patients were excluded from the study if they were dialysis dependent, presenting with acute myocardial infarction, or presenting with an overt cause of dyspnea such as chest wall trauma. All patients had B-type natriuretic peptide measured with a bedside assay at the time of enrollment and underwent an emergency department physician's estimation for the probability of CHF as the cause of dyspnea. Two independent, blinded cardiologists adjudicated all presentations for the gold standard of CHF. Patients were thus categorized as having dyspnea caused by CHF, dyspnea not caused by CHF despite a history of heart failure, and dyspnea without any evidence of heart failure.

Results: One thousand five hundred eighty-six patients were enrolled from April 1999 through December 2000. The mean age was 64 years; 56% of patients were male and 49% were white. The blinded cardiologists classified 744 patients as having dyspnea caused by CHF, 72 patients with dyspnea not caused by CHF despite depressed left ventricular function, and 770 patients with dyspnea not caused by CHF. Approximately one third of the patients with CHF had preserved systolic function. B-type natriuretic peptide was able to identify patients with dyspnea caused by CHF with a high degree of accuracy. By use of a cutoff of 100 pg/mL, this assay demonstrated 90% sensitivity and 76% specificity. For the overall receiver operator characteristics curve, the area under the curve was 0.91. B-Type natriuretic peptide was more accurate (84%) than the Framingham heart failure criteria (73%). In the multivariable analysis, a B-type natriuretic peptide level of ≥100 pg/mL had an odds ratio of 42 for dyspnea caused by CHF.

Interpretation: This study demonstrates that a bedside assay for B-type natriuretic peptide is an accurate test for identifying dyspnea caused by CHF in adult patients presenting to the emergency department. B-Type natriuretic peptide added significantly to other variables in a multivariable analysis for identifying CHF. This study adds to the growing literature on B-type natriuretic peptide for the diagnosis of CHF and suggests that this assay may have widespread clinical applicability for this purpose.

Session: Current dilemmas in valve surgery 

Study: Should CABG Patients with Mild to Moderate Aortic Stenosis Undergo Valve Replacement? A Markov Decision Analysis

Presenter: Dr William T. Smith IV, Duke University Medical Center, Durham, NC

Background: More than 350,000 patients undergo coronary artery bypass grafting (CABG) each year in the United States. A significant proportion of these patients have asymptomatic mild to moderate valvular aortic stenosis (AS). The decision to perform concomitant aortic valve replacement (AVR) at the time of CABG is often a difficult one to make, and little information is available to guide the practicing clinician faced with this common scenario. Previously reported case series include small numbers of patients and have reached conflicting conclusions regarding the need for AVR at the time of CABG. This quandary is reflected in the most recent American College of Cardiology/American Heart Association guidelines for the care of patients with valvular heart disease, in which the lack of evidence to guide decision-making in this setting is acknowledged.

The investigators performed a Markov model decision analysis to compare the long-term outcomes of patients with mild to moderate AS undergoing CABG or CABG/AVR. Age-specific procedural morbidity and mortality risks with CABG, CABG/AVR, and AVR after CABG were based on data from the Society of Thoracic Surgeons database containing clinical information on over 1.3 million patients who have undergone thoracic surgery. Data from 1995 to 2000 were included in this analysis. Input variables for the model describing the rate of symptom development, the probability of valve-related morbidity, and the rate of progression of AS were obtained from the current literature. Several assumptions of the model included the absence of any other life-threatening comorbid events, a constant rate of progression of AS (7 mm Hg/year), AVR with mechanical valves only, and a trivial rate of valve failure.

Results: CABG mortality rates increased from 1.3% for patients <55 years of age to 5.7% for those >75 years of age. CABG/AVR mortality rates ranged from 3.9% to 8.7% over a similar age span. The corresponding range of mortality rates for AVR after previous CABG was 4.2% to 11.3%. Incorporating the aforementioned assumptions, the optimal model decision was most strongly affected by baseline patient age and peak aortic valvular gradient by echocardiography. For example, among patients aged <70 years with a peak gradient of >25 mm Hg, average life expectancy was longer with combined CABG/AVR than with CABG alone. As age increased beyond 70 years, CABG/AVR was the preferred strategy only if peak gradient was increased (to approximately 50 mm Hg by age 85 years). Subsequent sensitivity analysis demonstrated that the rate of AS progression influenced outcomes, with CABG favored for most patients with slow (2 mm Hg/year) progression of AS and CABG/AVR favored for most with more rapid rates of stenosis progression (14 mm Hg/year).

Interpretation: This study is an important tool to help clinicians make decisions regarding the need for aortic valve replacement at the time of CABG for patients with mild/moderate valvular aortic stenosis. The most important variables to consider are the patient's age and baseline gradient by echocardiography. The rate of stenosis progression is an important component of the model, although this may not be known in all patients.

Clinicians should keep in mind several limitations of this decision model. Patients with significant comorbidities were excluded; thus, the application of this model to clinical practice may be limited. Additionally, only the use of mechanical valves was considered, and many elderly patients with the combination of CAD and mild/moderate calcific AS may be excellent candidates for bioprosthetic valves. Nevertheless, this is an important contribution to the literature and can help guide decision-making until more prospective data are available.

Session: Late-breaking clinical trials III 

Study: Lescol Intervention Prevention Study (LIPS)

Presenter: Dr Patrick Serruys, Thoraxcenter, Rotterdam, The Netherlands

Background: HMG CoA (3-hydroxy-3methylglutaryl coenzyme A) reductase inhibitors have revolutionized cardiology, with several trials showing an impressive benefit on morbidity and mortality rates. These trials have randomly assigned individuals in both primary and secondary prevention settings, based on baseline cholesterol. Advances in percutaneous intervention have occurred in parallel to those in the treatment of lipidemia. Professor Serruys et al1 have published favorable long-term (10-year) outcome data from patients undergoing attempted angioplasty at the Thoraxcenter (Rotterdam, The Netherlands). From these data, they generated a postulate that their favorable outcomes could be further improved with adjunctive treatment with an HMG CoA reductase inhibitor after intervention. The Lescol Intervention Prevention Study (LIPS) was then designed to evaluate the benefit of treatment with an HMG CoA reductase inhibitor after a first percutaneous intervention (PCI).

This study was a randomized, double-blind trial of 80 mg of fluvastatin versus placebo in patients undergoing their first PCI. The trial began in 1996 and enrolled through 1998. The sample size was calculated on the basis of an event rate of 25% in the placebo group and a target relative risk reduction of 25%. The α was 0.05 and the power 90%. The data were analyzed on an intention-to-treat basis, and the average length of follow-up was 3 to 4 years. Inclusion criteria were patients between the ages of 18 and 80 years undergoing their first successful PCI. Baseline cholesterol values were 135 to 270 mg/dL and triglycerides <400 mg/dL; patients could not have had a prior exposure to HMG CoA reductase inhibitors.

The primary end point was time to major adverse cardiac event (MACE), defined as death, myocardial infarction (MI), or revascularization (repeat PCI or CABG). Prespecified secondary end points included cardiac death, noncardiac death, death/MI, and noncardiac death/MI. MI was defined as the presence of new Q waves or creatine phosphokinase >2 times the upper limit of normal.

Results: One thousand six hundred seventy-seven patients were randomly assigned to fluvastatin versus placebo. Baseline characteristics were equal. The average age was 60 years, and 84% of the patients were male; 50% of patients had unstable angina, and 10% had evidence of silent ischemia; 89% had angina, and only 44% had evidence of a prior MI; 14.2% of the placebo arm had a history of diabetes compared with 9.8% of the fluvastatin arm (a difference that was not statistically significant). The average follow-up was 3.9 years. Average time-to-treatment after PCI was 2.7 days; 55% to 56% of patients received stents and 35% to 38% had multivessel disease. The average baseline low-density lipoprotein (LDL) in the trial was 132 mg/dL. The average treated LDL fell to <100 mg/dL, an effect that was consistent through the entire length of follow-up.

Analysis of the primary end point revealed a 22% reduction in MACE in the treated group compared with the placebo group (P = .0127). In an analysis of secondary end points, there was a 26.7% event rate in the placebo group compared with a 21.4% rate in the treated group, a relative risk reduction of 20% (P = .006). Prespecified subgroup analyses addressed the primary end point in subgroups of diabetics and in those with multivessel disease. In diabetic patients, the MACE rate was 37.8% in the placebo group and 21.7% in the treated group, a relative risk reduction of 43% (P = .022). In those with multivessel disease, there was an event rate of 33.9% in the placebo group compared with 23.0% in the treated group, a relative risk reduction of 32% (P = .008). Of interest is that when repeat PCI for restenosis was excluded from the MACE primary end point analysis, the risk reduction improved to 34% (P = .002).

Fluvastatin had an excellent safety profile. There were no liver enzyme level elevations in either group. Creatine kinase elevations (defined as levels 10 times normal) were observed in 0.4% of the placebo group and 0% of the fluvastatin-treated group; 33% of those in the placebo group crossed over to treatment with an HMG CoA reductase inhibitor, and 70% of those in the treated group continued to receive fluvastatin throughout the trial follow-up.

Interpretation: This is the first randomized, double-blind, placebo-controlled trial of outcomes in a population of patients undergoing their first PCI. When compared with prior lipid trials, this population had the lowest baseline LDL level (average LDL = 132). In relation to other secondary prevention trials (such as CARE [Cholesterol And Recurrent Events], LIPID [Long-term Intervention with Pravastatin in Ischemic Disease], and 4S [Scandinavian Simvastatin Survival Study]), this population had more unstable angina and less history of myocardial infarction. The calculated number needed to treat is 19 to prevent one primary end point. Subgroup analysis stratifying according to baseline LDL level is forthcoming. When subgroup analyses are available, we hope this trial will confirm the results of the Health Prevention Study that revealed benefit regardless of baseline LDL level.