Download this article's Factoid (PDF & PPT for Gold Subscribers)

In the very elderly, use of drug-eluting stents (DES) results in decreased target vessel revascularization (TVR) and similar mortality and bleeding compared to bare metal stents (BMS), according to a study published online November 19, 2013, ahead of print in Circulation: Cardiovascular Quality and Outcomes. The paper employed a sophisticated time-varying analysis to eliminate confounders.

Laura Mauri, MD, MSc, of Brigham and Women’s Hospital (Boston, MA), and colleagues analyzed data from 2,690  patients aged at least 85 years who underwent PCI with DES (n = 1,183) or BMS (n = 1,193) in nonfederal Massachusetts hospitals between April 1, 2003, and September 30, 2009.

Analysis Takes Into Account Changing DES Use Over Time

Quarterly DES usage changed dramatically over the study period, ranging from 15%, shortly after device approval to 88% in mid-2006.  Thereafter, DES use declined to less than 35% every quarter, reflecting safety concerns. 

Unadjusted 1-year mortality was markedly lower in the DES group than the BMS group (14.5% vs 23.0%; P < 0.0001), but the researchers described the difference as implausible based on prior randomized trial results. Unadjusted 1-year TVR was also lower in DES patients (4.3% vs. 9.3%; P < 0.0001), while bleeding rates were similar between DES-treated and BMS-treated patients (10.3% vs. 12.4%; P = 0.08).

In comparison, adjusted instrumental variable analysis found no difference in either mortality or bleeding, while the benefit of DES with regard to TVR persisted (table 1).

Table 1. Adjusted Instrumental Variable Outcomes at 1 Year^a

 ^a Using a 2-stage least squares approach adjusted for other patient covariables.

In contrast, traditional linear regression analysis and propensity-score matching showed DES to be associated with lower 1-year mortality (-5.8%; P = 0.001 and -4.0%; P = 0.02, respectively) compared to BMS, indicating that these methods are not effective at eliminating confounders, the investigators say. 

“Use rates of new coronary device technology, in this case of DES, served as a strong instrumental variable that allowed comparison of treatment effectiveness and safety within an unselected population of extremely elderly subjects,” the authors observe. “We found that DES were associated with similar mortality and a significant reduction in TVR compared with BMS in this population.”

Findings Important for an Aging Population

When DES were introduced in 2003, “we didn’t have significant evidence for their use in the extreme elderly population,” coauthor Robert W. Yeh, MD, MSc, of Massachusetts General Hospital (Boston, MA), told TCTMD in a telephone interview.  However, his paper suggests that DES effectively reduce TVR, are safe, and do not appear to increase the hazard for mortality. The findings are important because the elderly represent a growing population undergoing PCI, he said. 

The data indicate a positive outcome for sustained reduction in restenosis, Jeffrey W. Moses, MD, of Columbia University Medical Center/Weill Cornell Medical Center (New York, NY), said in a telephone interview with TCTMD.  “The take-away is that DES are safe and more effective than bare metal in this population.”

In a telephone interview with TCTMD, Ajay J. Kirtane, MD, also of Columbia University Medical Center, cited similar results from the XIMA trial, presented at the 2012 TCT scientific symposium, which found that DES drastically reduced repeat revascularization in the extreme elderly, although the devices missed the primary composite endpoint including mortality.

“But the very elderly die from things that have nothing to do with the stent procedure,” Dr. Kirtane noted. If clinicians are worried about restenosis and its impact on quality of life--and the patient’s life expectancy is good--then DES may be an option, he said, noting, however, that if the patient is not going to live much longer, this might be the threshold for avoiding the procedure.

A Better Methodology?

“One of the reasons we undertook an instrumental variable approach is that it’s less sensitive to confounding when trying to compare 2 treatments in a nonrandomized setting,” said Dr. Yeh. “If the reasons physicians choose treatment A over B are prognostically important, this can lead to confounding of traditional observational comparisons.”

Moreover, the very elderly represent  a group in which confounders could be the most significant, Dr. Yeh explained. Traditional regression or propensity score methods often find a significant mortality benefit with DES, but “we don’t believe this is the case.” Using an instrumental variable approach helps overcome confounders in a way that mimics randomization, he said, explaining that the rapid adoption of DES after 2003, and their subsequent decline after the stent thrombosis scare of 2006, effectively served as a  natural experiment.

The methodology is the most notable aspect of the paper, said Dr. Kirtane. “When you think about the people you treat in your practice with one stent versus another, the decision-making process is intangible and not caught in a database,” he said. For example, if a patient looks sick, that individual may receive one stent rather than another.

In addition, comparing patients receiving DES to those receiving BMS, which require different durations of dual antiplatelet therapy, is like “comparing apples to oranges with a conventional statistical approach,” Dr. Kirtane observed.  The time variable analysis tries to overcome conventional statistical limitations.

The study methodology “gives us a much more plausible outcome,” Dr. Moses agreed. “The marked variability in utilization [of stent types] without much change in the population helped them zero [in] on [the observed] outcomes.”

• Dr. Yeh reports no relevant conflicts of interest.
• Dr. Moses reports serving as a consultant for Abbott Vascular and Boston Scientific.
• Dr. Kirtane reports receiving institutional research funding from Acumed, Boston Scientific, Medtronic, and St. Jude Medical.

Related Stories:

• DES Show Better Outcomes vs. BMS in Elderly, Though Use Declining
• Improvements in PCI for Elderly Patients Yield Benefits
• No Excess Bleeding with EES vs. BMS in Octogenarians Despite Prolonged DAPT