The most award winning
healthcare information source.
TRUSTED FOR FOUR DECADES.
Medical Therapy vs CABG in Patients with LV Dysfunction
Abstract & Commentary
By Andrew J. Boyle, MBBS, PhD, Assistant Professor of Medicine, Interventional Cardiology, University of California, San Francisco. Dr. Boyle reports no financial relationship relevant to this field of study.
Source: Velazquez EJ, et al. Coronary-artery bypass surgery in patients with left ventricular dysfunction. N Engl J Med 2011;364:1607-1616.
Patients with significant coronary artery disease (CAD) and left ventricular (LV) dysfunction are at high risk for death and hospitalization. Coronary-artery bypass graft (CABG) surgery has been the traditional treatment for those whose LV dysfunction was thought to be due to chronic ischemia or hibernating myocardium. However, CABG carries significant risk in patients with impaired LV function. Surgical and medical therapies have improved in recent years and the optimal strategy for treating these patients is not clear. To address this, the National Institutes of Health sponsored the surgical treatment for ischemic heart failure (STICH) trial, which enrolled patients with significant CAD and LV dysfunction, then randomized them to either CABG surgery + medical therapy or medical therapy alone. The overall study also included an arm comparing CABG alone vs CABG + surgical ventricular reconstruction, but this paper presents only the data from the medical therapy vs CABG randomized trial.
Inclusion criteria for the study were: age over 18 years, symptomatic heart failure class II-IV within 3 months of enrollment, LV ejection fraction ≤ 35% within 3 months of trial entry, and CAD suitable for revascularization. Exclusion criteria included: recent myocardial infarction or cardiogenic shock, valvular heart disease requiring repair or replacement, > 1 prior cardiac operation, refractory ventricular arrhythmias, any condition that significantly limited life expectancy or increased surgical risk, poor medical compliance, left main coronary stenosis ≥ 50%, and class III or IV angina. Medical therapy included angiotensin converting enzyme inhibitors (or angiotensin receptor blockers), beta-blockers, spironolactone, and antiplatelet therapy. Other medications, including statins and diuretics, were given for specific patient indications, and investigators were encouraged to avoid nonsteroidal anti-inflammatories and to manage concurrent conditions (hypertension, arrhythmias, angina, dyslipidemia) according to guidelines. Surgeons were guided to use at least one internal mammary artery if possible. The use of implantable defibrillators and biventricular pacemakers were at the treating physician's discretion.
The study randomized 1212 patients in 22 countries, including the United States, to medical therapy (n = 602) vs CABG + medical therapy (n = 610). There were no differences between groups in baseline demographics, LV function, heart failure class, angina class, or angiographic severity of lesions. The mean age was 60 years, 12% were female, 40% were diabetic, and 77% had prior MI. Initially, the investigators planned to enroll 2000 patients for 3 years, but due to slow enrollment they had to alter the number of patients. Instead they enrolled 1200 patients but followed them for longer (5 years) and determined this would give them similar statistical significance. The primary endpoint was all-cause mortality; secondary endpoints were cardiovascular mortality and the combined endpoint of death or hospitalization for cardiovascular causes.
Of those patients randomized to medical therapy, 17% crossed over to receive CABG; this was due to progressive symptoms or acute deterioration in symptoms in two-thirds and for patient or physician preference in one-third. Of those randomized to receive CABG, 9% crossed over and never received CABG; we are not told why. Adherence to medical therapy was high and similar in both groups. After a median follow-up of 56 months, the primary endpoint occurred in 41% of patients in the medical therapy group, and 36% in the CABG group (hazard ratio [HR] 0.86, P = 0.12), which failed to meet statistical significance. The secondary endpoints favored CABG, including lower rates of cardiovascular death (28% vs 33%, P = 0.05) and lower rates of death or hospitalization for cardiovascular cause (58% vs 68%, P < 0.001). Notably, the CABG group had a higher mortality rate in the first 30 days, which then reached equipoise with the medical therapy group around 2 years, and thereafter the CABG group tended to have lower mortality. The authors performed exploratory analyses of the per-treatment groups (those who actually received CABG, whether by randomization or crossover, vs those who never received CABG). This analysis showed CABG was associated with lower mortality (HR 0.70, P < 0.001). They also performed a per-protocol analysis that excluded the patients who crossed over. This also showed a lower mortality in the CABG group (HR 0.76, P < 0.01), but these analyses should be considered hypothesis-generating only. The authors conclude that there was no significant difference between medical therapy alone and medical therapy plus CABG with respect to the primary endpoint of all-cause mortality. Patients assigned to CABG had lower rates of death from cardiovascular causes and death from any cause or hospitalization from cardiovascular causes.
The authors are to be congratulated for this well-designed trial that addresses an important question. They screened their surgeons to find ones who had experience with this high-risk cohort and had achieved a surgical mortality < 5%. They achieved a very high adherence rate with medical therapy. Their results underscore the high mortality and rehospitalization rates in this patient group, despite optimal medical and/or surgical therapy. Several factors about the trial require clarification. The lack of statistical significance in the primary endpoint may be due to real differences in the treatments or to trial design factors. The intention-to-treat analysis is inherently biased against CABG. Firstly, because of the crossover rate, all patients in both groups received the same medical therapy, yet some analyzed in the medical therapy only group had actually received CABG. Secondly, because some patients were only followed up for 1 year, and some CABG patients waited up to 6 months from randomization to receive surgery, the early surgical mortality biases against CABG in those followed for shorter periods of time. Although the investigators did their best to account for this in the power calculation, an inherent bias remains.
Importantly, the slow recruitment necessitated changing the enrollment criteria during the study and this required changing the assumptions in the power calculations. The mortality curves clearly diverge and it is likely that with longer follow-up the CABG mortality would be statistically significantly lower. Thus, one must question whether the study was effectively underpowered for the event rates that were seen. Furthermore, the per-treatment and per-protocol analyses support CABG as the superior treatment; however, these must be considered exploratory only.
In addition, we are not given several important pieces of information, including the effects on LV function, the use of devices such as biventricular pacemakers or defibrillators in each group, whether or not there was inducible ischemia, and other clinical outcomes such as stroke, quality of life, and the cost-effectiveness of each approach. These will likely be the subjects of future manuscripts.
Should we abandon CABG based on the STICH trial results? Certainly not! When presented with an individual patient with CAD and LV dysfunction, medical therapy should be initiated and optimized. CABG remains the treatment of choice for patients with left main stenosis > 50% (who were excluded from this trial and thus medical therapy has not been tested in these patients). Other patients who have significant CAD should be referred for CABG if there is acute or chronic deterioration, as in this study. However, in the stable CAD patient with LV dysfunction, an initial strategy of medical therapy seems reasonable. More data are still needed.