• High-Dose Statins in Acute Coronary Syndromes
  JAMA
  Editorial
  September 15, 2004
  Steven E. Nissen, MD

Select Your Interests

Early Intensive vs a Delayed Conservative Simvastatin Strategy in Patients With Acute Coronary Syndromes:Phase Z of the A to Z Trial

Long-term therapy with statin drugs has been shown to reduce the riskfor death, myocardial infarction (MI), and stroke among patients with establishedcoronary artery disease, even when low-density lipoprotein (LDL) cholesterollevels are not elevated.^1-4 Mostof the landmark clinical trials evaluating statins for secondary preventionenrolled patients who were stable for at least several months after an indexacute coronary syndrome (ACS) event.^1-3 Morerecently, 2 randomized controlled trials have evaluated earlier initiationof statin therapy following an ACS event and have noted a corresponding earlyreduction in major cardiovascular events.^5,6

Phase Z of the A to Z trial was designed to evaluate a strategy of earlyinitiation of intensive treatment with simvastatin in ACS patients comparedwith a delayed, less intensive strategy.

Details of the study design have been reported previously.⁷ TheA to Z trial is an international trial consisting of 2 overlapping phases.Phase A was an open-label noninferiority trial comparing enoxaparin with unfractionatedheparin in patients with non–ST-elevation ACS who were treated withtirofiban and aspirin. Patients were required to have chest pain at rest lasting10 minutes or longer within the previous 24 hours, which was associated witheither ST elevation or depression of 0.5 mm or higher, or with elevated levelsof creatine kinase–MB or troponin. Results of phase A have been reported.⁸

Phase Z is a double-blind trial comparing 2 statin regimens in patientswith ACS. Patients between the ages of 21 and 80 years with either non–ST-elevationACS or ST-elevation MI were eligible for enrollment if they had a total cholesterollevel of 250 mg/dL (6.48 mmol/L) or lower. Initially, patients were enteredinto phase Z only if they presented with non–ST-elevation ACS, werestabilized during phase A of the trial for at least 12 consecutive hours within5 days after symptom onset, and met at least 1 of the following high-riskcharacteristics: age older than 70 years; diabetes mellitus; prior historyof coronary artery disease, peripheral arterial disease, or stroke; elevationof serum creatine kinase–MB or troponin levels; recurrent angina withST-segment changes; electrocardiographic evidence of ischemia on a predischargestress test; or multivessel coronary artery disease determined by coronaryangiography. Patients enrolled in phase A who did not meet stability and high-riskcriteria were not eligible for continuation to phase Z. All patients providedwritten informed consent and the protocol was approved by the local institutionalreview board of each participating hospital.

The protocol was amended to allow patients with non–ST-elevationACS who were not enrolled in phase A and patients with ST-elevation MI toenter directly into phase Z. Patients in the latter category were requiredto receive fibrinolytic therapy or primary percutaneous coronary intervention(PCI) if they presented within 12 hours of symptom onset and no reperfusiontherapy if symptom onset was longer than 12 hours prior to presentation. Patientswere also required to meet criteria for stability and have at least 1 high-riskfeature in addition to cardiac biomarker elevation.

Patients were excluded from enrollment if they were receiving statintherapy at the time of randomization, if coronary artery bypass graft surgerywas planned, or if PCI was planned within the first 2 weeks after enrollment.Patients also were excluded for having an alanine aminotransferase (ALT) levelhigher than 20% above the upper limit of normal (ULN); for having an increasedrisk for myopathy due to renal impairment (serum creatinine level >2.0 mg/dL[176.8 µmol/L]) or concomitant therapy with agents known to enhancemyopathy risk, such as fibrates, cyclosporine, macrolide antibiotics, azoleantifungals, amiodarone, or verapamil; or for having a prior history of nonexercise-relatedelevations in creatine kinase level or nontraumatic rhabdomyolysis.

All patients were encouraged to adopt an American Heart AssociationStep I diet. They were randomized to either an early intensive statin treatmentstrategy (40 mg/d of simvastatin for 30 days and then 80 mg/d of simvastatinthereafter) or a less aggressive strategy (placebo for 4 months and then 20mg/d of simvastatin thereafter). Each center was assigned 1 or more blocksof 4 allocation numbers and blinded study supplies corresponding to theseallocation numbers. Treatments were assigned randomly to the allocation numbersusing a blocked randomization scheme. A patient was randomized by being assignedto the next available allocation number at that site.

Clinical and laboratory assessments (lipid levels, high-sensitivityC-reactive protein serum chemistries, liver function tests, creatine kinaselevel, and urine pregnancy tests) were performed prior to study drug initiationand at months 1, 4, and 8 and every 4 months thereafter until trial completion.Patients were followed up for at least 6 months and up to 24 months. Patientswho had LDL cholesterol levels that were higher than 130 mg/dL (3.37 mmol/L)at month 8 or any subsequent visit were provided additional dietary, lifestyle,and compliance counseling. If after 6 weeks the LDL cholesterol level remainedhigher than 130 mg/dL (3.37 mmol/L), the investigator could either add a bileacid sequestrant or discontinue the study drug and initiate open-label statintherapy. The study drug was discontinued if the LDL cholesterol level was40 mg/dL (1.04 mmol/L) or lower.

Patients with levels of ALT or aspartate aminotransferase (AST) thatwere higher than 3 times the ULN and creatine kinase levels higher than 5times the ULN were required to have a repeat measurement within 3 days. Protocol-mandatedwithdrawal from study treatment was required for patients with any consecutiveelevations in ALT or AST levels higher than 3 times the ULN, a single measurementof creatine kinase level higher than 10 times the ULN with muscle symptoms,or a consecutive measurement of creatine kinase level higher than 10 timesthe ULN without symptoms.

The primary efficacy end point of the trial was a composite of cardiovasculardeath, nonfatal MI, readmission for ACS (requiring new electrocardiographicchanges or cardiac marker elevation), and stroke. All primary end points wereadjudicated by an independent clinical end point committee blinded to treatmentassignment. Secondary end points included individual components of the primaryend point, revascularization due to documented ischemia, all-cause mortality,new-onset congestive heart failure (requiring admission or initiation of heartfailure medications), and cardiovascular rehospitalization.

Adverse events were recorded at each study visit, and safety data werereviewed by an independent data and safety monitoring board. The primary safetyoutcomes were elevations in levels of AST or ALT higher than 3 times the ULNor myopathy (defined as creatine kinase level >10 times the ULN and associatedwith muscle symptoms).⁹ Rhabdomyolysis wasdefined as a creatine kinase level higher than 10 000 units/L with orwithout muscle symptoms.¹⁰

Sample size calculations were based on the following assumptions: (1)a 1-year event rate of 20% in the placebo plus 20 mg/d of simvastatin groupbased on the observed event rate between 4 and 180 days in the tirofiban plusunfractionated heparin group in the Platelet Receptor Inhibition in IschemicSyndrome Management in Patients Limited by Unstable Angina Signs and Symptoms(PRISM PLUS) study¹¹; (2) a 15% discontinuationrate; and (3) a 20% reduction in the primary end point rate in the simvastatinonly group (40 mg/d and then 80 mg/d). Based on these assumptions, 970 patientsexperiencing a primary end point event would yield 90% power at the 2-sided5% significance level and 80% power to detect a 17.5% reduction in the hazardratio.

The sample size of 4500 patients (2250 per treatment group) was selectedto yield 970 events within 1 year after enrollment. The protocol permitteda sample size adjustment to achieve the projected number of 970 total primaryend point events, but the executive committee elected to halt enrollment afterthe planned 4500 patients were enrolled. During the course of the trial, dataemerged to support statin initiation early after ACS,⁵ andpractice patterns in participating hospitals changed considerably, so thatit became increasingly difficult to enroll patients into the placebo-controlledportion of the trial. Moreover, the investigators believed that lengtheningthe follow-up period beyond 2 years to accrue more events would fundamentallyalter the hypotheses being tested in the trial.

All efficacy and safety analyses were performed on an intent-to-treatbasis. The cumulative incidence of the primary end point was determined usingthe Kaplan-Meier product-limit method. Patients who did not achieve expectedfollow-up were censored at the time they withdrew consent or were lost tofollow-up. Statistical comparison between treatment groups was performed usinga Cox proportional hazards model that included covariates for treatment groupand age. Although a significant treatment × time interaction (P = .03) was observed, all end point analyses were performedas prespecified using the Cox proportional hazards model. The primary endpoint analysis required a 2-sided α level of .046 for significance toaccount for 2 interim data and safety monitoring board analyses. All otheranalyses required a significance level of .05. Comparisons of adverse safetyevents were performed using the Fisher exact test. Statistical analyses wereperformed using SAS software (version 8.0, SAS Institute Inc, Cary, NC).

Between December 29, 1999, and January 6, 2003, 4497 patients were enrolledat 322 centers in 41 countries (Figure 1).The mean time from symptom onset to randomization in phase Z was 3.7 days.Baseline characteristics were similar between the 2 treatment groups (Table 1). Treatment with guideline-recommendedtherapies, such as angiotensin-converting enzyme inhibitors, β-blockers,and aspirin, was high and similar between the 2 treatment groups (Table 1). There were 1958 (44%) patientswho underwent PCI to treat the index ACS event prior to enrollment. Treatmentwas discontinued prematurely in 711 patients (32%) in the placebo plus 20mg of simvastatin group and 765 (34%) in the simvastatin only group (40 mg/dand then 80 mg/d). The median follow-up period was 721 days and 22 patientsin each treatment group were lost to follow-up (Figure 1).

In the placebo plus simvastatin group, median LDL cholesterol levelsincreased by 11% during the 4-month placebo period from 111 mg/dL (2.87 mmol/L)to 124 mg/dL (3.21 mmol/L) and then decreased to 77 mg/dL (1.99 mmol/L) atmonth 8 after the initiation of 20 mg of simvastatin (31% change from baseline).In the simvastatin only group, the median LDL cholesterol levels decreasedby 39% to 68 mg/dL (1.76 mmol/L) over the first month during treatment with40 mg/d of simvastatin and then decreased an additional 6% to 62 mg/dL (1.61mmol/L) at month 4 following the increase to the 80 mg/d of simvastatin (Table 2). Changes in lipid and C-reactiveprotein values are shown inTable 2.The latter were equivalent in the 2 groups at baseline and at month 1, butbecame significantly lower in the simvastatin only group subsequently.

The primary end point of cardiovascular death, MI, readmission for ACS,and stroke occurred in 343 patients (16.7%) in the placebo plus simvastatingroup compared with 309 (14.4%) in the simvastatin only group (hazard ratio[HR], 0.89; 95% confidence interval [CI], 0.76-1.04;P =.14;Figure 2). Outcomes for selectedsecondary end points are shown inTable3. Cardiovascular death occurred in 109 patients (5.4%) in the placeboplus simvastatin group compared with 83 (4.1%) in the simvastatin only group(HR, 0.75; 95% CI, 0.57-1.00;P = .05; number neededto treat to prevent 1 cardiovascular death, 77). No significant differenceswere observed between treatment groups with regard to the secondary end pointsof MI, readmission for ACS, revascularization due to documented ischemia,or stroke. New-onset congestive heart failure was reduced from 5.0% in theplacebo plus simvastatin group to 3.7% in the simvastatin only group (HR,0.72; 95% CI, 0.53-0.98;P = .04; number needed totreat to prevent 1 episode of new-onset congestive heart failure, 77;Table 3).

In a post hoc analysis, no difference between treatment groups in theprimary end point was evident over the first 4 months following randomization,which corresponded to the placebo-controlled comparison period (HR, 1.01;95% CI, 0.83-1.25;P = .89). However, from 4 monthsthrough the end of the study, the primary end point was reduced from 9.3%in the placebo plus simvastatin group to 6.8% in the simvastatin only group(HR, 0.75; 95% CI, 0.60-0.95;P = .02;Figure 3).

No significant treatment interactions were detected in subgroups definedby demographic variables, index diagnosis, baseline lipid and C-reactive proteinlevels, or use of early PCI. There was no evidence of a greater relative treatmenteffect among patients with higher baseline levels of LDL cholesterol (Figure 4).

Rates of adverse events according to treatment assignment and time sincerandomization are summarized inTable 4. The proportion of patients with consecutive elevations in ASTor ALT levels of more than 3 times the ULN was 0.4% (8/2068) in the placeboplus simvastatin group and 0.9% (19/2132) in the simvastatin only group (P = .05).

Discontinuation of the study drug due to a muscle-related adverse eventoccurred in 1.5% (34/2230) of patients in the placebo plus simvastatin groupand 1.8% (41/2263) in the simvastatin only group (P =.49). A total of 10 patients developed myopathy (creatine kinase level >10times the ULN with associated muscle symptoms); 1 patient was in the placeboplus simvastatin group and 9 patients were in the simvastatin only group (whiletaking the 80 mg/d dose) (P = .02). Three of the9 patients with myopathy had creatine kinase levels higher than 10 000units/L and met the definition for rhabdomyolysis. Of these 3 patients, 1patient had contrast-induced renal failure and 1 patient was receiving concomitantverapamil, which is a known inhibitor of CYP3A4. In addition, 1 patient receiving80 mg of simvastatin had a creatine kinase level higher than 10 times theULN without muscle symptoms, which was associated with alcohol abuse. Therewere no cases of myopathy when patients were taking 20 mg of simvastatin or40 mg of simvastatin.

Among patients with ACS who were stabilized with guideline-based acutetherapies, the early initiation of an intensive simvastatin regimen resultedin a trend toward reduction in the rate of the primary composite end pointof death, MI, readmission for ACS, and stroke when compared with the delayedinitiation of a less intensive simvastatin regimen. However, the trial didnot achieve the prespecified end point and the 11% relative (2.3% absolute)reduction in the rate of the primary end point in the early intensive statingroup was not statistically significant. While no differences were observedbetween treatment groups with regard to the secondary end points of MI andreadmission for ACS, the early intensive statin regimen was associated witha reduction in cardiovascular mortality of 25% (absolute reduction, 1.3%;P = .05) and congestive heart failure of 28% (absolutereduction, 1.3%;P = .04). These findings are qualitativelyconsistent with results from the Pravastatin or Atorvastatin Evaluation andInfection Therapy: Thrombolysis in Myocardial Infarction 22 (PROVE IT-TIMI22)⁶ and the Reversal of Atherosclerosis withAggressive Lipid Lowering (REVERSAL)¹² trials,which demonstrated improved clinical outcomes and reduced progression of atherosclerosiswith a more intensive statin regimen (80 mg/d of atorvastatin) compared witha less intensive statin regimen (40 mg/d of pravastatin).

Several factors may explain the absence of a statistically significantbenefit with respect to the primary end point in the early intensive statingroup of the A to Z trial. First, the clinical benefit was delayed and ofa smaller magnitude than was anticipated. Second, the trial had less statisticalpower than was originally planned due to a lower than expected number of endpoints and a higher than expected rate of study drug discontinuation. Althoughthe study reached its projected enrollment number of 4500 participants, theprimary end point rate was considerably lower than projected, resulting inonly 652 primary end point events, which is short of the 970 events requiredto provide adequate power. The 33% rate of study drug discontinuation in thepresent study is higher than in previous secondary prevention trials^1-4 but similarto the rate observed in PROVE IT, a trial that also enrolled patients earlyafter ACS and followed up patients for up to 2 years.⁶ Thesetrials highlight challenges in maintaining long-term adherence to even well-toleratedtherapies when they are initiated during the acute phase of illness.

Realizing that there are limitations when making direct comparisonsbetween clinical trials, the magnitude of clinical benefit observed at theend of the 2-year follow-up period in the A to Z trial appeared to be lessthan was observed in PROVE IT-TIMI 22, a finding that may relate to the smallerbetween-group difference in LDL cholesterol level lowering in the A to Z trialcompared with PROVE IT. While the LDL cholesterol levels achieved in the aggressivestatin treatment groups were similar between the 2 trials, the median LDLcholesterol level in the standard care treatment group of the A to Z trialwas 77 mg/dL (1.99 mmol/L) at 8 months compared with 95 mg/dL (2.46 mmol/L)in the group receiving 40 mg/d of pravastatin in PROVE IT. Thus, from months4 through 24, the median difference in LDL cholesterol levels between treatmentgroups was approximately 14 mg/dL (0.36 mmol/L) in the A to Z trial (18% relativedifference) compared with 33 mg/dL (0.85 mmol/L) in PROVE IT (33% relativedifference). This difference may in part explain the smaller relative reductionin the primary end point rate of 11% in the A to Z trial compared with 16%in PROVE IT. These findings are consistent with analyses included in the recentupdate to the National Cholesterol Education Program Adult Treatment PanelIII Guidelines, which support a direct proportional relationship between themagnitude of LDL cholesterol level lowering and coronary heart disease riskreduction.¹³

Until recently, information regarding the timing of initiation of statinagents following ACS had been limited to observational studies and post hocanalyses of clinical trials performed for other purposes.^14-17 Includingthe A to Z trial, 3 trials have evaluated early initiation of statins afterACS. The Myocardial Ischemia Reduction with Aggressive Cholesterol Lowering(MIRACL) study reported a 16% lower rate of death and nonfatal major cardiacevents 4 months after ACS in patients receiving 80 mg/d of atorvastatin comparedwith placebo (P = .05).⁵ Noearly differences in mortality or MI were evident in MIRACL, but rehospitalizationfor recurrent ischemia was significantly reduced—a finding that is incontrast to the A to Z trial, in which no effect of aggressive statin therapywas observed for the end point of readmission for ACS. Because MIRACL excludedpatients with recent or planned revascularization, the use of glycoproteinIIb/IIIa inhibitors and clopidogrel was low. In contrast to the MIRACL trial,PROVE IT included an active comparison group (40 mg/d of pravastatin), patientswith both ST-elevation MI and non–ST-elevation ACS, and patients whowere treated with PCI for their index ACS event. A significant benefit withregard to the primary end point of death, MI, and total revascularizationfavoring high-dose atorvastatin was evident within the first 6 months of PROVEIT and a strong trend was observed at 30 days.

In the A to Z trial, no early divergence in event rates was noted betweentreatment groups despite marked differences in LDL cholesterol levels. However,the primary end point rate was significantly lower in the aggressive simvastatingroup between 4 and 24 months. A number of factors related to patient characteristicsand concomitant treatment strategies may have contributed to the delay inclinical benefit observed in the A to Z trial. In contrast with the MIRACLstudy, glycoprotein IIb/IIIa inhibitor use was mandated in phase A of theA to Z trial and more than 50% of the patients were treated with an earlyinvasive management strategy.⁸ These therapiesmay have competed with statin therapy to reduce early nonfatal recurrent ischemicevents.

Differences between the A to Z trial and PROVE IT with regard to thetiming of patient enrollment and practice patterns in the enrolling sitesalso may have contributed to differences in early outcomes between the 2 trials.In PROVE IT, patients were enrolled an average of 7 days after ACS (largelyfrom centers in the United States) with the consequence that 69% of patientshad undergone revascularization to treat the ACS event prior to randomization.The culprit lesion and acute thrombotic process may have been stabilized priorto enrollment in PROVE IT, which is a hypothesis supported by the extremelylow cardiovascular mortality rate (<1.5% at 2 years) observed in both treatmentgroups in this trial.⁶ In contrast, in theA to Z trial, patients were enrolled 3 to 4 days earlier, were selected tohave higher risk features, and were less likely to undergo PCI prior to enrollment.These features may have resulted in more patients entering phase Z of thetrial with an ongoing active thrombotic process that was relatively less responsiveto statin therapy.

In contrast with the PROVE IT study,⁶ theC-reactive protein concentrations in the A to Z trial did not differ betweenthe treatment groups at 30 days despite marked differences in LDL cholesterollevels. The lack of a concurrent anti-inflammatory effect may also have contributedto the delayed treatment effect that was observed. Patients in the aggressivesimvastatin arm of the A to Z trial were not titrated up to 80 mg/d of simvastatinuntil after the first month. It is possible that more intensive therapy isrequired immediately after the onset of ACS during the period of greatestclinical instability to achieve a more rapid clinical benefit.

No difference in treatment effect was evident across subgroups definedby LDL cholesterol levels. These findings are consistent with those from theHeart Protection Study, which found no interaction based on initial LDL cholesterollevels in the reduction of major vascular events with 40 mg/d of simvastatin.⁴

The incidence of consecutive elevations in AST or ALT levels higherthan 3 times the ULN was low in both treatment groups. Muscle-related adverseevents occurred infrequently, including those that led to study drug discontinuation.Myopathy (creatine kinase level >10 times the ULN with muscle symptoms) occurredin 9 patients (0.4%) receiving 80 mg/d of simvastatin, which included 3 patientswho developed rhabdomyolysis (creatine kinase level >10 000 units/L);these rates are consistent with the long-term safety of this dose in patientswith hypercholesterolemia.⁹ There were no casesof myopathy in patients receiving 20 mg/d or 40 mg/d doses of simvastatin,which is also consistent with the overall safety of these doses reported inother long-term outcome trials.^1,4 Cliniciansshould be aware that the 80-mg/d dose of simvastatin is associated with ahigher risk of myopathy than lower dosages of the drug and should educatepatients receiving the 80-mg/d dose of simvastatin to pay close attentionto muscle-related symptoms.

The traditional approach to lipid management following ACS has beento begin with dietary management and then to initiate a statin agent at alow dose and increase the dose stepwise to achieve target LDL cholesterollevels. The findings from the A to Z trial, as well as from MIRACL and PROVEIT, support a strategy of aggressive LDL cholesterol lowering following ACSto prevent death and major cardiovascular events. Statins should be initiatedearly after ACS, with consideration of dosages well above the typical startingdose, and they should be down-titrated or discontinued if important adverseeffects, such as myopathy or significant liver abnormalities, develop.