Abstract & Commentary
Adoption of Internal Mammary Artery Bypass Grafts in the United States
By Michael H. Crawford, MD, Editor
Source: Hlatky MA, et al. Adoption and effectiveness of internal mammary artery grafting in coronary artery bypass surgery among medicare beneficiaries. J Am Coll Cardiol 2014;63:33-39.
In 1986, Floyd Loop reported that the use of the internal mammary artery (IMA) as a conduit in coronary artery bypass surgery (CABG) reduced 10-year mortality by 38%. Other observational trials confirmed these results, but a randomized trial has never been done. Hlatky and colleagues evaluated the adoption of IMA graft use in CABG in Medicare beneficiaries undergoing multivessel CABG between 1988 and 2008. They excluded patients who had concomitant valve operations, prior revascularization, or had end-stage renal disease. In those having CABG after 1992, they also evaluated clinical outcomes. The primary outcome endpoints were all-cause mortality, acute myocardial infarction (MI), and repeat CABG. The CABG subjects with IMA grafts were propensity matched by baseline clinical characteristics to a group not receiving IMA grafting. Among 374,918 patients undergoing CABG, 69% received IMA grafts, whose use increased steadily over time from 31% in 1988 to 91% in 2008. Use of IMA grafts varied from region to region in the United States, but improved over time. Also, IMA use varied by age, being used more in those < 70 years of age as compared to those > 86 years of age, but this difference also decreased over time. In addition, those with IMA use were more often male and white and less often diabetic. Over a median follow-up of almost 7 years, mortality was 46% in the IMA patients and 53% in the matched controls (hazard ratio [HR], 0.77; 95% confidence interval [CI], 0.75-0.79; P < 0.001). Death or MI was also less with IMA (HR, 0.78; 95% CI, 0.76-0.80) as was repeat revascularization (IMA 8% vs 9%, P < 0.001). Five-year survival was higher with IMA use (76% vs 70%, P < 0.001). The survival advantage of IMA was attenuated by diabetes, older patients, women, and those with peripheral artery disease (PAD). The authors concluded that IMA grafting was slowly adopted and there were regional variations in use. Also, IMA use is associated with lower rates of death, MI, or repeat CABG.
It is remarkable that it took 20 years for IMA graft use in CABG to exceed 90%. By contrast, as the authors point out, it took only 6 months for drug-eluting stent use to reach 50%. The wide regional variation in the early years suggests that it was driven by surgical practice style. Having lived through this transition, I can confirm this impression. Rarely is the IMA unsuitable or is it harmed during harvest. The observation that in diabetics its use impaired sternal wound healing may have been a factor, but the main reason was that it took more time and for most surgeons, time is money. In the 1990s, health maintenance organizations were decreasing their use of specialists, so services like CABG were often contracted out. I heard about a surgical group based at one hospital during that time that was contracted by another hospital to do their cardiac surgery. The surgeons used a high percentage of IMA grafts at their main hospital, but at the other hospital the patients mainly received saphenous vein grafts. Why, because they wanted to complete surgery quickly at the other hospital, which paid them a fixed amount per case. If economics is the driver, why has usage finally increased substantially all over the country? Several years ago, IMA use became a CMS quality measure and this has accelerated its use.
Other factors probably played a role as well. For example, there was never a randomized, controlled trial of IMA grafting. However, several smaller observational studies done earlier all showed the superiority of IMA grafts. IMA use did not have a commercial entity supporting it as did drug-eluting stents. Also, the extra time needed for IMA use was considered risky in unstable patients and those with significant lung disease.
This large study permitted subgroup evaluation, and only those aged > 86 years favored no IMA use. In all other subgroups it was superior, but less so in some groups such as diabetics, women, and those with PAD. The 5-year survival of patients with IMA graft use was 73%, which is lower than the survival observed in large trials of CABG vs coronary artery stenting of about 90%. However, this study is an unselected Medicare population with surgery performed at a large cross-section of hospitals. Also, these types of database studies have no clinical details, such as where the grafts were placed and what the left ventricular function was. So, comparing these results to other studies is difficult. This study confirms the superiority of IMA graft use and supports employing this use as a quality measure. Thus, a randomized trial is unlikely to ever be done.