Efficacy and Safety of Pyridoxal 5′-Phosphate (MC-1) in High-Risk Patients Undergoing Coronary Artery Bypass Graft Surgery: Title and subTitle BreakThe MEND-CABG II Randomized Clinical Trial FREE

Coronary artery bypass graft (CABG) surgery is one of the most important therapeutic options for relieving angina and improving survival and quality of life in patients with multivessel coronary artery disease.¹ It is the most commonly performed cardiac surgical procedure in the world, and in 2005, more than 250 000 CABG procedures were performed in the United States.² While operative mortality has declined over time, serious complications including myocardial infarction (MI), recurrent angina, ventricular failure, serious arrhythmias, renal insufficiency, stroke, and death continue to limit surgical revascularization, particularly in high-risk patients, who represent an increasing proportion of those referred for surgery.^{3 - 8} Many of these complications are at least partially attributable to the process of ischemia-reperfusion injury.⁹ If outcomes among CABG patients are to be improved, development of better strategies to reduce ischemia-reperfusion injury is imperative.

MC-1 (Medicure International Inc, Winnipeg, Manitoba, Canada) contains pyridoxal 5′-phosphate monohydrate, a naturally occurring metabolite of pyridoxine (vitamin B₆). MC-1, a purinergic (P2) receptor antagonist, prevents cellular calcium overload in preclinical and clinical studies of ischemia-reperfusion injury.^{10 - 13} In the MC-1 to Eliminate Necrosis and Damage in Coronary Artery Bypass Graft Surgery Trial (MEND-CABG), a phase 2, randomized, multicenter, placebo-controlled, double-blind trial in high-risk patients undergoing CABG surgery (n = 901), treatment with 250 mg/d of MC-1 resulted in a nonsignificant 14% reduction in the primary composite outcome of death, MI (creatine kinase [CK] MB fraction ≥50 ng/mL), or stroke (P = .31).¹⁴ In a post hoc analysis using more stringent definitions of MI (CK-MB ≥70 or ≥100 ng/mL), however, MC-1 was associated with larger, statistically significant reductions in the composite of death or MI.¹⁴

MEND-CABG II was undertaken to replicate the positive observations from MEND-CABG using the more stringent (CK-MB ≥100 ng/mL) definition of MI. The primary objective of MEND-CABG II was to assess the cardioprotective effect of MC-1, 250 mg/d, administered before and continued for 30 days after surgery, compared with placebo, on the incidence of 30-day cardiovascular death or nonfatal MI in high-risk patients undergoing CABG surgery. A secondary objective was to determine the safety of MC-1 administered in the peri-CABG setting.

The design of MEND-CABG II has been previously described in detail.¹⁵ Briefly, MEND-CABG II was a phase 3, multicenter, randomized, double-blind, placebo-controlled clinical trial to evaluate the effect of MC-1, 250 mg/d, given preoperatively and for 30 days postoperatively on cardiovascular death or nonfatal MI in high-risk patients undergoing CABG surgery. An academic steering committee provided scientific direction and oversight for the study. An independent data and safety monitoring board monitored the safety of the trial participants.

Patients were enrolled between October 2006 and September 2007 at 130 sites in Canada, the United States, and Germany. To be eligible, patients had to be at least 18 years of age, be scheduled to undergo CABG surgery with cardiopulmonary bypass, and have 2 or more features putting them at high risk of cardiovascular complications from surgery. As in MEND-CABG, high-risk features were defined as age 65 years or older; current or recent smoking; diabetes mellitus; left ventricular dysfunction (ejection fraction ≤45%) or congestive heart failure; history of stroke, transient ischemic attack, or carotid endarterectomy; requirement for urgent CABG surgery; history of recent (>48 hours but <6 weeks) MI; prior peripheral arterial revascularization; moderate (estimated creatinine clearance ≥30 mL/min but <60 mL/min) renal dysfunction; and asymptomatic stenosis of at least 50% in at least 1 carotid artery.¹⁴

Exclusion criteria included pregnancy; planned concomitant valve or other surgery; inability to have a screening visit at least 4 hours before surgery; Mini-Mental State Examination score of less than 24; cardiogenic shock; interventricular or papillary muscle rupture; uncontrolled diabetes mellitus (serum glucose ≥24 mmol/L or ≥432 mg/dL); acute (<48 hours) MI; estimated creatinine clearance of less than 30 mL/min or nephrotic syndrome; history of malignancy in the past 5 years; ongoing alcohol or drug abuse; any medical or psychiatric condition that, in the opinion of the investigator, made the patient an unsuitable candidate for the study; and participation in another investigational drug or device study within 30 days.

Institutional review board or ethics committee approval was obtained at all participating sites, and all participants gave written informed consent prior to enrollment.

Patients were screened prior to randomization and CABG surgery. Using an automated voice response system, participating patients were randomly assigned in a 1:1 ratio to receive either MC-1, 250 mg/d, or blinded matching placebo (Figure 1). Randomization was stratified by site. The first dose of oral study medication was administered 3 to 10 hours prior to CABG surgery. If surgery was delayed or rescheduled, a second preoperative dose was administered to ensure that all patients received study medication 3 to 10 hours before surgery. During surgery, the protocol required use of cardiopulmonary bypass, moderate systemic hypothermia (30°-35°C), and routine anticoagulation.

The first postoperative dose of study medication was administered 24 hours (±8 hours) following the preoperative dose. Subsequent doses of study medication were administered with food beginning 24 hours (±4 hours) after the first postoperative dose and were then continued once daily through postoperative day 30. Patients who were unable to swallow at the time of the first postoperative dose were administered an intravenous dose of MC-1 (5 mg) or blinded placebo by bolus injection 24 hours (±8 hours) following the preoperative dose. This intravenous dose is expected to achieve a plasma level similar to a 250-mg oral dose of MC-1. If the patient was then able to swallow, oral dosing commenced the following day. If the patient remained unable to swallow, intravenous dosing was continued once daily for a maximum of 4 days.

All other medical care was left to the discretion of the cardiac surgeon and other treating physicians. Use of guideline-based medical therapy and lifestyle modification was strongly recommended.

Patient information including demographics, medical history, electrocardiography and laboratory results, surgical procedural details, concomitant medications, clinical outcome events, and serious and nonserious adverse events was collected through 90 days using a standard paper case report form. Race data were collected to determine whether MC-1 had a differential effect in patients of different races. Race was self-reported by patients, recorded in the patient's medical record, and collected in the following categories: white, black/African American, North American Indian/native Alaskan/Inuit, Hispanic/Latino, Asian/Southeast Asian/native Hawaiian/other Pacific islander, or other.

Blood samples for CK-MB levels were systematically collected at baseline and at 8, 12, 16, 24, 36, 48, 72, and 96 hours and 30 and 90 days postoperatively and sent to a core laboratory (CIRION Central Laboratory, Laval, Quebec, Canada) for analysis. The Beckman Coulter CK-MB assay, with a normal range of 2.88 ng/mL for women and 4.94 ng/mL for men, was used. Routine electrocardiography was performed at baseline and postoperatively at 48 and 96 hours and at 30 and 90 days and sent to an electrocardiography core laboratory (Dynacare Kasper Medical Laboratories, Edmonton, Alberta, Canada).

Clinical follow-up for outcomes and adverse events was performed daily during the index hospitalization, at 96 hours, and at 30 (range, 30-40) and 90 (range, 76-104) days. A prespecified list of nonserious adverse events common to the postoperative CABG population, including incision pain, mild edema, anorexia or mild nausea/vomiting, constipation, mild insomnia or anxiety, mild hypokalemia or hypocalcemia, mild anemia, mild pleural effusion atelectasis, low-grade fever, hypotension during cardiopulmonary bypass, and mild hypertension, were not required to be reported.

The primary efficacy outcome was the incidence of cardiovascular death or nonfatal MI from randomization through postoperative day 30. Secondary efficacy outcomes included cardiovascular death or nonfatal MI through day 4; all-cause and cardiovascular death through days 4, 30, and 90; stroke through days 4, 30, and 90; new atrial fibrillation through days 4, 30, and 90; intensive care unit and index hospitalization length of stay; and CK-MB fraction area under the curve through 24 hours. The safety of MC-1 was assessed through the collection of reports of serious and nonserious adverse events.

An independent, blinded clinical events committee (Montreal Heart Institute, Montreal, Quebec, Canada) adjudicated all suspected MIs, strokes, and the cause of death for all deaths, and their assessment was used in the primary analysis. The diagnosis of MI was based on clinical information collected from the sites and CK-MB and electrocardiographic laboratory data from the core laboratories. Patients were diagnosed as having an MI if they had a peak CK-MB level of at least 100 ng/mL on any sample through postoperative day 4; a peak CK-MB level of at least 70 ng/mL on any sample through 96 hours with new 30-ms Q waves in 2 contiguous leads; a new peak CK-MB level of at least 25 ng/mL after 96 hours; or new 30-ms Q waves in 2 contiguous leads that were not present at 96 hours. Stroke was defined as a new, focal, nontraumatic, neurological deficit lasting at least 24 hours. All deaths without an identifiable noncardiovascular cause were considered cardiovascular.

The sample size of the study was based on a projected incidence of the primary outcome in the placebo group of between 14% and 17%, based on experience from MEND-CABG and other recent trials in similar populations.¹⁴ A sample size of 1500 patients per group had at least 80% power to detect a 25% relative reduction in the primary outcome with MC-1 compared with placebo based on a 2-sided χ² test at an α level of .05.

Demographic data, baseline characteristics, and surgical data are summarized by group using descriptive statistics. Categorical variables are expressed as percentages. Continuous variables are expressed as median (interquartile range [IQR]). All statistical analyses were performed using SAS software, version 8.2 (SAS Institute Inc, Cary, North Carolina).

The incidence of the primary efficacy outcome in the active treatment group was compared with the incidence in the placebo group with a χ² test in the intention-to-treat population. P<.05 was considered significant. Primary outcome events identified at the 30-day visit (day 30-40) with a reasonable possibility, as assessed by the clinical events committee, of having occurred prior to day 30 were included in the primary outcome analysis.

Prespecified ranked secondary outcomes included index hospital length of stay, intensive care unit length of stay, and cardiovascular death through day 90. A sequential closed testing procedure was used for these outcomes, with P<.05 considered significant. The overall study α level was maintained at .05. Hospital and intensive care unit length of stay were analyzed with time-to-event analyses. For these analyses, groups were compared using log-rank tests. Patients who died prior to hospital discharge were assigned a length of stay equal to 1 day longer than the maximum observed length of stay in the data.

All other analyses were performed based on the intention-to-treat population and were considered exploratory. Rates of categorical secondary outcomes were compared between the 2 groups using the χ² or Fisher mid P tests.¹⁶ The CK-MB area under the curve between 8 and 24 hours was calculated using the trapezoidal rule and compared using a t test after log transformation of the data. Estimated creatinine clearance was calculated using the formula described by Cockcroft and Gault¹⁷ and compared between groups using the Wilcoxon rank-sum test. Retrospectively, after observation of the primary results, the incidence of the primary outcome, cardiovascular death or nonfatal MI, was recalculated using alternate definitions of MI based solely on peak CK-MB values of at least 100 ng/mL, at least 70 ng/mL, and at least 50 ng/mL between 8 hours and 30 days.

The effect of MC-1 in prespecified subgroups based on age, diabetes mellitus, hypertension, renal dysfunction, region of enrollment, and use of intravenous study medication was also evaluated. Adverse events were codified by preferred term and by system organ class using MedDRA (Medical Dictionary for Regulatory Activities; http://www.meddramsso.com/MSSOWeb/index.htm), version 10.1. Certain adverse events were prespecified as being of particular interest, including infections and infestations, mediastinitis, pneumonia, sepsis, acute renal failure, and renal impairment. The percentage of patients with adverse events is reported by group.

This article reports the main results through postoperative day 30. As of March 7, 2008, complete data for the 90-day outcomes were not available. Patient flow through the study is shown in Figure 1. A total of 3023 patients were randomly assigned, 1519 to MC-1 and 1504 to placebo. Assessment of the primary outcome, cardiovascular death or nonfatal MI, was performed in 2996 (99.1%) of those randomized. Vital status through postoperative day 30 was known in all but 25 patients (0.8%).

The baseline characteristics of the study population are typical of intermediate- to high-risk patients undergoing CABG surgery and reflect the inclusion criteria of the trial (Table 1). The median patient age was 66 years, most patients were white, 46% had diabetes, 13% had renal dysfunction, 29% had a recent MI, 8% had a prior stroke, 24% had heart failure, and 36% had reduced left ventricular function.

Nearly all patients received study drug prior to surgery, with the first dose given a median of 5.1 hours before surgery (Table 2). Most patients received more than 4 doses of oral study drug while in the hospital following surgery. Intravenous study drug was given to 20% of patients, with most of these patients receiving just 1 intravenous dose. Adherence to study drug was good, with less than 10% of patients missing more than 10 postoperative doses.

Details of surgery are shown in Table 3. More than 98% of randomized patients underwent surgery with cardiopulmonary bypass support. The median duration of aortic cross-clamp was 1.0 hours. Most patients had an internal thoracic graft (90%) and vein grafts placed with 3 to 4 distal anastomoses. Other arterial grafts were used in 9% of patients.

Concomitant medical therapies during hospitalization and at discharge are shown in Table 4. Aprotinin and tranexamic acid were administered to 16% and 22% of patients, respectively. The use of evidence-based medications, including aspirin, β-blockers, angiotensin-converting enzyme inhibitors, and statins, was appropriately high given the risk of the population. Almost half of patients received some antiarrhythmic medication while in the hospital and more than 80% received insulin.

The primary efficacy outcome, cardiovascular death or nonfatal MI at 30 days, occurred in 140 of 1510 patients (9.3%) in the MC-1 group and 133 of 1486 patients (9.0%) in the placebo group (risk ratio, 1.04; 95% confidence interval, 0.83-1.30; P = .76). There was no difference in the primary outcome in any prespecified subgroup based on age, diabetes, hypertension, renal dysfunction, region of enrollment, or use of intravenous study medication (Figure 2).

Most patients had 9 (21%) or 10 (68%) blood samples collected at the 10 protocol-specified time points for CK-MB assays. The median 96-hour peak CK-MB level in the overall population was 22.8 (IQR, 14.7-39.3) ng/mL, with no difference between the treatment groups (median, 23.1 [IQR, 14.8–40.7] ng/mL for MC-1 vs 22.6 [IQR, 14.4-38.0] ng/mL for placebo). There was also no difference in the 8- to 24-hour CK-MB area under the curve between patients assigned to receive MC-1 and those assigned to placebo (median, 270 [IQR, 175-492] vs 268 [IQR, 170-456] hours × ng/mL; P = .11).

Other prespecified outcomes are shown in Table 5. Patients assigned to receive MC-1 had higher cardiovascular and all-cause mortality on day 4 than patients assigned to receive placebo (P = .03), but there was no difference in either cardiovascular or all-cause mortality at day 30. There was no difference in the incidence of nonfatal MI, postoperative stroke, atrial fibrillation, or renal function. Patients assigned to receive MC-1 and placebo had similar intensive care unit and hospital lengths of stay (Table 6). The duration of intubation and chest tube duration were also similar between groups. Rates of intra-aortic balloon pump support (2.4%), dialysis (1.0%), and cardioversion (3.2%) were low and similar between the groups. More than half of the patients assigned to either the MC-1 or placebo group received a blood transfusion following surgery.

Post hoc analyses based on recalculation of the primary outcome using alternate definitions of MI did not reveal any beneficial effect of MC-1 compared with placebo on cardiovascular death or nonfatal MI, using definitions of MI based on peak CK-MB levels of at least 100 ng/mL (8.1% vs 7.3%; P = .41), at least 70 ng/mL (12.7% vs 11.8%; P = .43), or at least 50 ng/mL (19.7% vs 17.3%; P = .09).

Selected prespecified adverse events are shown in Table 7. Overall rates of other serious and nonserious adverse events (including virtually all MedDRA system organ classes and preferred terms) were similar between patients assigned to receive MC-1 and placebo. Nausea and vomiting, the only known dose-related adverse effects of MC-1, were numerically more common among patients assigned to MC-1 than among those assigned to placebo.

In this population of intermediate- to high-risk patients undergoing CABG surgery, treatment with the purinergic receptor antagonist MC-1 had no effect on the composite of cardiovascular death or nonfatal MI. Although the placebo event rate (9.0%) was lower than expected, the results obtained exclude with 95% certainty a greater than 17% relative reduction in cardiovascular death or MI with MC-1 in this population. There was no beneficial effect of MC-1 seen in any prespecified subgroup or on other outcomes. The higher rates of cardiovascular and all-cause mortality seen with MC-1 on day 4 are concerning, but given the similar mortality rates observed at 30 days, most likely due to chance.

Although the placebo event rate was significantly higher in the first MEND-CABG trial, by design, the MEND-CABG and MEND-CABG II trials were similar. In light of these results, the positive post hoc findings from the first MEND-CABG trial are most likely due to chance.¹⁵ These findings illustrate the challenge and potential hazard of interpreting positive but non–statistically significant or post hoc findings from relatively small, phase 2 clinical trials. Although MC-1 was not effective at reducing myocardial injury in the broad population of intermediate- to high-risk patients undergoing CABG surgery, it is possible that this drug may be effective in selected populations at particularly high risk of ischemia-reperfusion injury, including patients with acute MI, shock, or prolonged valve surgery. In addition, other preliminary data suggest that MC-1, used in combination with the angiotensin-converting enzyme inhibitor lisinopril, may have beneficial effects on blood pressure and lipid metabolism in hypertensive patients with type 2 diabetes mellitus.^{18 - 19}

Despite progressively higher-risk patients being referred for CABG, the mortality associated with isolated CABG surgery has declined substantially over the last several decades.^{2 ,7 - 8} Nevertheless, myocardial injury remains a significant issue in patients undergoing CABG. The high rate of perioperative MI seen in MEND-CABG II is similar to that observed in other recent studies. In PREVENT-IV,⁴ GUARDIAN,²⁰ and PRIMO-CABG,²¹ the rates of perioperative MI were 9.7%, 9.1%, and 8.0%, respectively, with relatively similar definitions of MI. In MEND-CABG II, the definition of MI required a large CK-MB elevation to 20 or more times the core laboratory upper limit of normal.

These MI rates are higher than those reported clinically (2.4% in MEND-CABG II and 1.1% in the Society of Thoracic Surgeons National Cardiac Database).⁸ The higher rates of MI in clinical trials compared with registry data based on clinical practice are almost certainly the result of systematic surveillance and mandatory reporting. Better ascertainment of MI following CABG surgery with routine measurement of postoperative CK-MB or troponin in patients undergoing this procedure may be an important component of improving quality of care and outcomes for patients undergoing cardiac surgery.

Some degree of CK-MB elevation occurs in almost all patients undergoing CABG surgery, and the rate of MI depends on the threshold of CK-MB elevation used to define this event. In MEND-CABG II, more than 50% of patients had a peak CK-MB level of more than 5 times the upper limit of normal and more than 75% had a peak CK-MB level of more than 3 times the upper limit of normal. These data are consistent with those from the GUARDIAN trial,²⁰ in which the proportion of patients with postoperative CK-MB elevations of less than 5 times, 5 to less than 10 times, 10 to less than 20 times, and 20 or more times the upper limit of normal were 63.7%, 19.5%, 9.6%, and 7.0%, respectively.

Any increase in CK-MB after CABG surgery is suggestive of myocyte necrosis, and higher levels of CK-MB are likely to be associated with worse outcomes.^{22 - 23} A linear relation between postoperative CK-MB elevation and mortality has been previously reported, with postoperative peak CK-MB values of less than 5 times, 5 to less than 10 times, 10 to less than 20 times, and 20 or more times the upper limit of normal associated with 3.4%, 5.8%, 7.8%, and 20.2% 6-month mortality, respectively.²⁰ A recent consensus document recommended a definition of MI following CABG surgery based on a CK-MB elevation of at least 5 times the upper limit of normal during the first 72 hours following CABG surgery, associated with the appearance of new pathological Q waves or left bundle-branch block, angiographically documented new graft or native coronary artery occlusion, or imaging evidence of new loss of viable myocardium.²⁴

MC-1 is one of several agents that have been investigated for the prevention of the clinical complications associated with ischemia-reperfusion injury during cardiac surgery with unsuccessful results. Pexelizumab, a recombinant, single-chain anti-C5 monoclonal antibody, was studied in 2 multicenter, randomized clinical trials of patients undergoing CABG surgery with or without valve surgery on cardiopulmonary bypass.^{21 ,25} Patients were randomly assigned to receive placebo or pexelizumab bolus only or pexelizumab bolus plus infusion. There was no significant difference in the primary outcome of death, MI (CK-MB ≥60 ng/mL), or new central nervous system deficit between the pexelizumab and placebo treatment groups.²¹ Post hoc analysis suggested a reduction in the composite of death or large MI (CK-MB ≥100 ng/mL) in patients undergoing isolated CABG surgery in the bolus-plus-infusion group at 30 days (3% vs 9%; P = .004). A subsequent, larger trial in patients undergoing isolated CABG surgery, however, failed to confirm this finding (relative risk, 0.82; 95% confidence interval, 0.66-1.02).²⁵

Similarly disappointing results were observed with cariporide, a specific inhibitor of the sodium-hydrogen exchanger. In the large GUARDIAN trial,²⁶ 11 590 patients with unstable angina or non–ST-segment elevation MI undergoing percutaneous or surgical revascularization were randomly assigned to receive placebo or 1 of 3 doses of cariporide (20 mg, 80 mg, or 120 mg 3 times daily). No reduction in death or MI assessed after 36 days was observed with cariporide compared with placebo. However, the highest dose of cariporide was associated with a reduction in death or nonfatal MI in patients undergoing CABG surgery (relative risk, 0.75; 95% confidence interval, 0.59-0.97).²⁶ The subsequent large EXPEDITION trial²⁷ did confirm this effect of cariporide in reducing death or MI after CABG surgery; however, this benefit was offset by a significant increase in stroke.

A number of studies have investigated the role of adenosine in attenuating myocardial injury during CABG surgery.^{28 - 33} In a phase 2 randomized trial of adenosine added to cold-blood cardioplegia, higher doses of adenosine were associated with lower use of postoperative dopamine and better postoperative left ventricular function.³⁰ A larger, multicenter study supported the beneficial effects of adding adenosine to cardioplegia, demonstrating a reduction in postoperative vasopressor use and, although underpowered, a dose-dependent reduction in perioperative MI and other adverse outcomes.³¹ A meta-analysis of 5 trials including 4043 patients undergoing CABG surgery revealed a substantial beneficial effect of the purine nucleoside acadesine on the combined outcome of cardiovascular death, MI, or stroke (odds ratio, 0.73; 95% confidence interval, 0.57-0.93; P = .01).³² A subsequent randomized clinical trial in 2698 patients failed to demonstrate a reduction in MI with acadesine; however, among patients with perioperative MI, 2-year mortality was substantially lower in those receiving acadesine.³³ This is a promising area that deserves further study.

MEND-CABG II has several limitations. This trial enrolled intermediate- to high-risk patients undergoing isolated CABG surgery with cardiopulmonary bypass. Some patients at the highest risk of ischemia-reperfusion injury during CABG surgery, including those with acute MI, shock, or emergency surgery, those undergoing longer operations with concomitant valve surgery, and those with more advanced renal disease, were not included in this study. The effect of MC-1 in other populations, particularly those at higher risk of ischemia-reperfusion injury, might be different. In addition, MEND-CABG II investigated a specific dosing regimen of MC-1. Although the absorption and plasma half-life of MC-1 suggests that adequate levels were achieved, the intracellular kinetics of MC-1 are not well understood and it is possible that a different dose or different dosing strategy would produce different results. Although the follow-up in MEND-CABG II was more than 99% complete, a small number of patients (n = 27) withdrew consent or for other reasons were unable to complete 30-day follow-up.

MEND-CABG II demonstrates that among intermediate- to high-risk patients undergoing CABG surgery, MC-1, 250 mg/d, given immediately before and for 30 days following surgery did not reduce cardiovascular death or nonfatal MI. Myocardial injury remains a significant problem following CABG surgery. Effective therapies to reduce perioperative morbidity and mortality are needed but remain elusive.