Diabetes More Influential Than CAD Complexity in DES-Treated Patients

Patients with diabetes who receive DES are at higher risk for needing repeat treatment than their nondiabetic peers, according to new analysis of pooled data from earlier trials. The strength of that relationship does not appear vulnerable to whether patients had higher or lower SYNTAX scores, or whether they were implanted with newer-generation devices. 

“Together these findings indicate that diabetes mellitus per se, and not higher disease complexity among diabetic patients, is a driver of inferior efficacy and adverse PCI outcomes in the era of DES,” Stephan Windecker, MD, of Bern University Hospital (Bern, Switzerland), and colleagues write in a paper published online last week in Circulation: Cardiovascular Interventions. 

They looked at a patient-level pooled analysis of 6,081 patients from 4 all-comers trials—SIRTAX, RESOLUTE All Comers, BIOSCIENCE, and LEADERS—conducted in Europe between 2004 and 2013. There were no restrictions on number of treated lesions, treated vessels, lesion length, or number of stents implanted. Overall, 22% of patients had diabetes, and 75% of all DES used were new-generation.

Impact on Clinical Outcomes

Compared with their nondiabetic counterparts, diabetic patients had higher mean SYNTAX scores (13.9 vs 12.9; P < .001) and were more likely to be undergoing multivessel revascularization (25% vs 22%; P = .03).

At 2 years, the crude rate of MACE (primary endpoint; cardiac death, nonfatal MI, and clinically indicated TLR) was increased in diabetic patients, as were rates of TLR, TVR, and cardiac death. No difference by diabetes status was seen for MI.

After multivariable adjustment, the differences remained significant for MACE, TLR, and TVR but not for cardiac death. Again, MI risk was similar between groups. There also was no difference for Academic Research Consortium-defined definite stent thrombosis.

The findings held true in the subgroup of 4,554 patients who received newer-generation DES.

When stratified by SYNTAX score, rates of MACE were higher in diabetic vs nondiabetic patients, regardless of whether their scores were at or below the study population’s median score of 11 (11.8% vs 7.2%) or above it (16.7% vs 12.7%; P < .001 for both). Multivariable analyses also found no formal interaction between MACE risk and diabetes status by SYNTAX score, nor any interplay related to the outcomes of TLR, TVR, cardiac death, MI, or definite stent thrombosis.

Furthermore, in an exploratory analysis using a SYNTAX cutoff of 22, there was consistently no interaction between the score category and diabetes status for MACE, TLR, TVR, cardiac death, or MI.

Although multivessel revascularization during the index procedure was an independent predictor of 2-year MACE (P = .008), when the SYNTAX score was entered into the model, the significance disappeared (P = .15). Evaluation of SYNTAX score as a continuous variable to avoid potential confounding due to where the cutoffs were set also showed no interaction.

Reconsidering Disease Complexity

Windecker and colleagues point out that in the FREEDOM trial, SYNTAX score was not found to predict outcomes after PCI or CABG among diabetic patients. The SYNTAX trial, on the other hand, showed that adverse events increased incrementally across higher score tertiles, driven largely by more frequent need for revascularization. But diabetes was not independently associated with long-term mortality in the trial and subsequently was not included in the clinical variables of the SYNTAX score II.

The results of the new study also differ from another recent pooled analysis of 18 RCTs showing diabetic patients to have greater frequency of TLR and TVR within 1 year of DES PCI but only when treated for complex target lesions. The difference, they say, may be due to the fact that the earlier study did not account for disease complexity beyond the treated lesion, which is important in diabetic patients who typically harbor diffuse disease.

Diabetes “remains a major determinant of restenosis in the era of new-generation DES, even in patients with noncomplex anatomies,” Windecker et al note. They assert that their findings support the importance of preventing long-term PCI complications via meticulous attention to the acute PCI result and use of evidence-based adjunctive therapies in diabetic patients across the angiographic disease spectrum.

But Davide Capodanno, MD, PhD, of Ferrarotto Hospital (Catania, Italy), was skeptical of drawing too much from the study, saying he views the conclusion that diabetic status contributes to worse outcomes as a statistical artifact.

“It’s all about the variables you put in the multivariable model,” he wrote in an email. “Some variables may be predictors at univariate analysis (indeed, both the SYNTAX score and diabetes were univariate predictors of MACE and TLR in this study), but one may overshadow the other when they are both entered in the adjustment model.”

Capodanno also noted a number of limitations that affect interpretation of the study. In addition to different types of DES used, he observed that the mean Syntax score in this study was 13, while in the PCI arm of the SYNTAX trial it was much higher, at about 28. In addition, most of the patients in the pooled analysis would have been categorized in the lower tertiles of score if they had been enrolled in the SYNTAX trial. These factors, together with the disparate number of variables (9 in the pooled analysis, 16 in SYNTAX) accounted for in the respective multivariable models, “make the results hardly comparable with regards to independent predictors,” Capodanno concluded.

Disclosures:

Related Stories:

• Experts Debate Merits of Guideline-Directed Care for Patients With Multivessel Disease
• CAD Complexity Does Not Affect Dominance of Newer vs Older DES
• TUXEDO-India Affirms Superiority of EES in Patients With Diabetes