Lower Mortality with Drug-Eluting Stents than Bare-Metal Stents
Abstract & Commentary
By Andrew J. Boyle, MBBS, PhD
Source: Mauri L, et al. Long-term clinical outcomes after drug-eluting and bare-metal stenting in Massachusetts. Circulation. 2008;118:1817-1827.
The dramatic reduction in in-stent restenosis (ISR) afforded by drug-eluting stents (DES) led to their widespread use throughout the world. Subsequently, more patients and lesions could be treated percutaneously with less thought about ISR. With the advent of a more widespread adoption of percutaneous coronary intervention, registry data began suggesting an increased risk of late and very late stent thrombosis with DES compared to bare-metal stents (BMS). Autopsy studies showed incomplete endothelialization of DES struts, giving rise to a plausible explanation for the suggested increase in stent thrombosis. However, it remained uncertain whether the increased rates of stent thrombosis were due to the stents themselves or to the more complex nature of the patients and lesions undergoing stenting with DES.
Clinical trials of DES vs BMS have not been powered to detect small differences in mortality. Meta-analyses of pooled data from the pivotal trials that led to approval of DES showed no difference in mortality or stent thrombosis in this relatively low-risk patient group. However, data are lacking on the efficacy and safety of DES compared to BMS in "real-world" settings, including both low- and high-risk patients and lesions. Accordingly, Mauri et al performed a retrospective analysis of patients undergoing coronary artery stenting in Massachusetts from April 2003 to September 2004. Drug-eluting stents were introduced to the market in 2003, and they demonstrated a rapid uptake of this new technology.
All patients undergoing PCI in Massachusetts at non-government hospitals were included. Using data compiled for the state by each center, describing patient and procedural details, they cross-linked with data from the Massachusetts Registry of Vital Statistics, the Social Security Death Index, and the hospital discharge data from the Massachusetts Division of Healthcare Finance and Policy. After excluding patients who were not residents in the state, those who were not linked to administrative files, and those receiving both DES and BMS, they had comprehensive data on 11,556 patients receiving DES and 6,237 receiving BMS. Significant differences existed between the groups at baseline, with patients receiving BMS being older and having co-morbidities more frequently, including cancer, gastrointestinal bleeding, heart failure, lung disease, renal insufficiency, cardiogenic shock, and acute myocardial infarction at presentation.
In the absence of randomization, Mauri et al used complicated statistical methods to overcome these baseline differences, minimize confounders, and make the most valid comparisons between patients receiving DES and those receiving BMS.
A propensity-score matched analysis was performed to compare groups receiving DES and BMS.
One-to-one matching was performed with a logistic regression model created from 63 variables to identify pairs of DES and BMS patients with similar baseline characteristics. In total, 5,549 patients receiving DES were compared with 5,549 patients receiving BMS. The validity of this approach was confirmed using several sensitivity analyses. The result was a large cohort of relatively well matched patients followed up for two years.
After propensity matching, DES use was associated with a reduction in mortality compared to BMS use (9.8% vs 12.0%; p = 0.0002). DES use also was associated with a reduction in myocardial infarction (8.3% vs 10.3%; p = 0.0005) and target-vessel revascularization (11.0% vs 16.8%; p < 0.0001). Furthermore, when patients presenting with myocardial infarction were excluded, DES use was still associated with reduction in total mortality and target vessel revascularization of approximately the same magnitude. Mauri et al conclude that DES use is associated with lower mortality, myocardial infarction, and target vessel revascularization, but that longer-term follow-up is needed to provide further reassurance of preserved safety of these devices.
Mauri et al have analyzed a large contemporary cohort of patients undergoing PCI with robust statistical techniques to minimize confounders. However, it is important to acknowledge that this is a retrospective study, rather than a prospective, randomized study. Despite the inherent limitations of such an analysis, the large numbers of patients enrolled and the techniques used to match the cohorts give us substantial confidence that DES use is not associated with increased events or mortality. In fact, these data are consistent with several other recent real-world registry studies demonstrating reduced mortality in patients receiving DES compared to BMS.
One possible explanation for the improved clinical outcomes in the DES group is the extended duration of dual anti-platelet therapy in this group. At the time these procedures were performed, dual anti-platelet therapy was recommended for one month for BMS and for 3-6 months in patients receiving DES. The cumulative incidence curves for all outcomes separated early, within the first 3-6 months, during the time of dual anti-platelet therapy, suggesting that this may have been a factor. If that is the case, we may see, for more recent recommendations for prolonged dual anti-platelet therapy for patients receiving DES and for those who present with acute coronary syndromes, a reduction in events in both BMS and DES, perhaps being even more in favor of patients receiving DES.
Mauri et al combined sirolimus-eluting stents and paclitaxel-eluting stents together, potentially missing an effect of one or the other device. In addition, there may have been confounders that they did not account for in their statistical analyses. Despite these limitations, Mauri et al present a strong argument for the safety and efficacy of DES use in varied patient groups.