Daniel M. Keller, PhD

November 14, 2012

SAN DIEGO, California — Researchers from the US Renal Data System (USRDS) have shown that Medicare patients on dialysis who need cardiac revascularization have better 1-year survival if they receive drug-eluting stents (DES), but their long-term survival was better if they received a coronary artery bypass graft (CABG).

Charles Herzog, MD, director of the Cardiovascular Special Studies Center of the USRDS and professor of medicine at the University of Minnesota, Minneapolis, reported these findings here at Kidney Week 2012.

The best long-term survival was associated with use of internal mammary artery grafts and with a multivessel CABG, he added.

Dr. Herzog has published on the topic of survival of dialysis patients after revascularization, citing a 56% 2-year survival after CABG and 48% survival after percutaneous coronary intervention (PCI) with bare metal stents (BMS). However, little information existed on survival in the era of DES.

So from Medicare claims in the USRDS database he and a coauthor identified dialysis patients (94% hemodialysis) undergoing CABG without valve surgery (n = 6178) or PCI with BMS (n = 5011) or DES (n = 11,844) from 2004 through 2009. (Medicare covers end-stage renal disease [ESRD] patients as well as the elderly.)

The number of revascularization procedures for dialysis patients peaked in 2004-2005 at about 4400 per year. The proportion of procedures using DES peaked at about 62% in the years 2005 and 2006, after which time revascularization procedures dropped to just under 3500/year, and the proportion of DES use declined to about 42% to 45% of procedures in 2007-2009. During this latter period, the use of BMS increased to about 30% of procedures.

Best Long-term Survival With CABG, but Risk of Near-term Events

For the 2004-2009 period, patients receiving DES had the best survival (71%) within the first year after revascularization vs 70% for CABG and 63% for BMS. But at about 18 months, the survival curves crossed, and survival was best with CABG from then out to 6 years. The survival curves after 2 years were essentially parallel. At 6 years, survival of patients who had received CABG was 22%, 18% for DES, and 15% for BMS.

Survival was best for CABG patients if they received an internal mammary artery graft (about 75% of CABG procedures) compared with a different replacement vessel. This effect was evident by 6 months after revascularization (79% vs 72% survival, respectively) and persisted out to 6 years, at which time internal mammary graft recipients' survival was 24% vs 18% for other CABG patients (log rank, P < .001).

Survival also correlated with the number of vessels bypassed, and better survival occurred as the number of vessels increased from 1 to 4 or more (log rank, P < .001).

Independent Predictors of Death

In a comorbidity-adjusted Cox model using CABG, white race, male sex, age younger than 65 years, less than 2 years on dialysis, hemodialysis, and absence of other diseases as the reference, several factors emerged as significant independent predictors of death.

Some of the predictors varied between the periods of 2004-2006 and 2007-2009. For example, in the earlier period, the hazard ratio (HR) for DES vs CABG was 1.08 but dropped to 1.00 in the latter period. However, the risk for BMS vs CABG was essentially unchanged (HR, 1.27 vs 1.25, respectively).

Other predictors were also unchanged across time. Risk consistently increased with age, was elevated with diabetes as the cause of ESRD, time on dialysis for 6-10 years, being on peritoneal dialysis, and having congestive heart failure, stroke, or peripheral artery disease.

Women were at about the same risk for death as men, but black race appeared to be protective, with HR of 0.88 in the earlier period and 0.84 in the later period.

Predictors of death specifically with CABG were advancing age, diabetes, longer time on dialysis, peritoneal dialysis, heart failure, stroke, and peripheral artery disease. Again, female sex had no effect, and blacks had the same degree of lower mortality risk (HR, 0.88).

Dr. Herzog concluded that the total number of coronary revascularization procedures for dialysis patients decreased from 2004 to 2009 and that a major reduction in DES use happened in 2007.

Patients receiving DES had better 1-year survival, but those getting CABG had better unadjusted long-term survival, he said. Patients with internal mammary artery grafts and multivessel CABG did best.

Limitations of the study are that it is a retrospective analysis of an administrative database and is not a randomized controlled trial. As such, it is susceptible to selection bias and other confounding factors.

Dr. Herzog emphasized that the revascularization strategy has to be individualized for each patient, and he views this sort of data as useful for that purpose. "I don't think that this type of data is suitable for making a true comparative effectiveness analysis. I think it's very suitable for talking to patients."

He said in his experience patients have not been as concerned about a possible future need for a repeat procedure as they have been about getting out of the hospital sooner and having functional recovery with PCI, and balancing those benefits against the risk for immediate adverse events such as stroke or dying with CABG.

"I think the surgical discussion should only occur in patients with multivessel disease or 1 of the vessels is the left anterior descending coronary artery or left main coronary artery, and there are some caveats for a mammary graft," Dr. Herzog advised.

Session moderator Julie Ingelfinger, MD, senior consultant in pediatric nephrology at Massachusetts General Hospital and professor of pediatrics at Harvard Medical School in Boston, Massachusetts, identified another important limitation of this retrospective study in a comment to Medscape Medical News.

"The type of stents are evolving, and so when you look at a database study using stents that are really not specified, as he said, we just know the era in which this was done, stopping in 2009 — when there are more mTOR inhibitor-eluting stents now," she noted.

And she was more circumspect than Dr. Herzog about using the sort of information from this study to advise patients. "You can use a retrospective database study to say, 'This is what we've got at this point in time....' I think what we need to do responsibly as clinicians is say [to patients], 'We don't have a randomized trial to really share with you about this, but we do have some registry data,'" Dr. Ingelfinger said.

After explaining to patients what registry data are, her approach would be to say, "'Given your situation, we know that mortality is greatest in the CABG group early on, and you need to make a choice. We think you're going to live 10 years more, so that over the long haul your chances would be better.' But I always say, and I think most of us now do to patients, [that] each person is an N of 1. So I would not use these data as hard data. I would say the field is changing, and the techniques are changing for drug-eluting stents."

There was no commercial funding for the study. Dr. Herzog is employed by the Chronic Disease Research Group, which has research contracts with Affymax and Johnson & Johnson and a pending contract with Zoll. He has consultancy agreements with Abbott Labs, Affymax, Amgen, Cormedix, Fribrogen, Fresenius, Medtronic, and Merck. He has ownership interests in Boston Scientific, Cambridge Heart, Johnson & Johnson, and Merck. He is a scientific advisor or has membership relationships with Abbott Labs, Amgen, Fresenius, Medgenics, and RoFAR and has received honoraria from UpToDate. Dr. Ingelfinger was not involved in the study and has disclosed no relevant financial relationships.

Kidney Week 2012: Abstract HI-OR05. Presented November 3, 2012.