Author + information
- Published online April 3, 2017.
- Sorin J. Brener, MD∗ ()
- Department of Medicine, Division of Cardiology, New York Presbyterian Brooklyn Methodist Hospital, Brooklyn, New York
- ↵∗Address for correspondence:
Dr. Sorin J. Brener, New York Presbyterian Brooklyn Methodist Hospital, Department of Medicine, Division of Cardiology, 506 6th Street, Brooklyn, New York 11215.
Diabetes mellitus (DM) type 2 is an increasingly prevalent metabolic disease with vast implications for cardiovascular (CV) health in the developed and developing worlds. Patients with DM are 2- to 4-fold more likely than patients without diabetes to develop coronary artery disease and to die from it (1,2). Besides increasing risk of death from CV diseases, DM also increases risk of cancer-related deaths, which highlights its general contribution to morbidity and mortality (3). The nearly 2-decades-old observation that having DM is prognostically equivalent to having already had a myocardial infarction (4) remains strongly entrenched (maybe less justifiably so now) in guidelines for prevention of coronary artery disease (5).
In this issue of JACC: Cardiovascular Interventions, Faggioni et al. (6) very elegantly address a different aspect of the potential negative impact of DM. Using the well-designed and analyzed PARIS (Patterns of Non-Adherence to Antiplatelet Regimens in Stented Patients) registry (7), they investigated whether discontinuation of dual-antiplatelet therapy (DAPT) has differential effects in patients with and without DM (i.e., whether patients with diabetes interrupting DAPT after percutaneous coronary intervention [PCI] with drug-eluting stents are more likely to experience adverse events compared with patients without diabetes stopping DAPT). Among 4,207 patients treated predominantly (>80%) with the less thrombogenic second-generation drug-eluting stents, 34% had DM, a rate consistent more with U.S. than European registries. As expected, patients with DM had a worse combination of baseline characteristics, comorbidities, and prior CV manifestations than patients without diabetes, except for less frequent presentation with acute coronary syndromes and current smoking. Because of their comorbidities, they were more likely to be already treated with DAPT (almost exclusively clopidogrel-based regimens) on presentation.
Because PARIS dealt primarily with frequency and modes of DAPT discontinuation, it is striking to note how infrequent DAPT discontinuation was in the first year after PCI: only ∼20% in those with and without DM. Between 1 and 2 years after PCI, patients with DM were more likely to continue DAPT than patients without diabetes, and the main difference in rates of cessation resulted from physician-directed discontinuation, rather than from temporary interruption for procedures or discontinuation because of bleeding or noncompliance. By a factor of 2 to 3, physician-directed discontinuation was more common than interruption because of procedures or disruption because of bleeding or noncompliance, regardless of diabetic status.
Not surprising again, patients with DM had much higher rates of death and myocardial infarction than those without DM, both at 1 year (when nearly all patients were on DAPT) and at 2 years (when more patients with DM were on DAPT). This effect was independent of DM treatment (insulin or oral medication; Online Figure 2 in Faggioni et al. ).
The investigators also found a highly significant excess in major adverse cardiac events (composite of cardiac death, myocardial infarction, definite or probable stent thrombosis, and clinically indicated target vessel revascularization) in patients with diabetes compared with patients without diabetes. I would caution against using such an endpoint in the context of this manuscript because it can confound the pattern of DAPT use (revascularization may lead to restarting previously stopped DAPT). The other composite endpoint used, major adverse cardiac events-2, which does not include revascularization, confirms these observations.
Thus, the consistent and important detrimental impact of DM is confirmed once again in this contemporary dataset, in which patients received overall excellent medical care. Major bleeding was similar, and rare, in patients with and without diabetes. It did not seem that treatment of diabetes (oral medications or insulin) affected the patterns of DAPT cessation or the bleeding rates. Importantly, DAPT cessation, disruption in particular, was associated with a heightened risk of ischemic events, but this excess risk was not modulated by the presence or absence of diabetes.
Although the investigators were very careful to analyze the impact of the different modes of DAPT discontinuation on clinical outcomes, I was particularly impressed by the following less discussed observations:
1. Even though it is unfortunate that discontinuation of DAPT caused by bleeding and noncompliance were lumped in 1 category, it is quite clear that this type of DAPT cessation, dubbed disruption, was not predominantly caused by major bleeding. Regardless of definition of major bleeding, it occurred at most in 2% to 3% at 1 year and in 3% to 4% at 2 years, whereas the respective rates of disruption were ∼8% at 1 year and 13% at 2 years, without major difference between the 2 groups on the basis of diabetic status. Whether the predominant factor in disruption was minor/nuisance bleeding or noncompliance is not disclosed in the paper.
2. By 1 year, while ∼80% of both groups are on DAPT, cardiac death (2.6% vs. 0.7%; p < 0.001) and myocardial infarction (3.2% vs. 1.2%; p < 0.001) are already 2 to 3 times higher in patients with than without diabetes. These differences escalate further at 2 years, when roughly 50% of both groups are still on DAPT. This pattern of use essentially refutes the putative protective effect of DAPT in patients with diabetes (predicated on their heightened risk for thrombotic events because of hyperactive platelets), at least at this interval from PCI, and negates the possibility to comment on shorter durations of DAPT than 1 year. Whether the differences in clinical outcomes at 1 year between patients with and without diabetes reflect imbalances in baseline characteristics or an effect of DM not mediated by platelets, and thus unaffected by antiplatelet therapy, remains uncertain. I favor the latter, although it is also conceivable that without DAPT (or maybe because DAPT did not include the more effective novel antiplatelet agents) patients with diabetes would have fared even worse. It is important to understand that the hazard ratio provided by the authors (Faggioni et al. , Figure 3) represents the adjusted excess risk of DAPT cessation in patients with and without diabetes separately, compared with uninterrupted DAPT.
Thus, the negative impact of DM is yet again confirmed; so, where is the silver lining? One can find solace in the lack of excess bleeding in patients with diabetes after PCI and in the lack of interaction between DM and risk of events after DAPT cessation. The number of events is quite small, however, and the confidence intervals for the hazard ratio provided by the investigators are quite wide, suggesting that these observations may not be very robust.
Editorialists are supposed to provide critique, and not just praise and interpretation. This analysis is making this task difficult because of its quality. Yet, as the investigators recognize, the duration of cessation is not entered in the statistical adjustment model, nor do we know how many patients restarted DAPT after physician-directed discontinuation or disruption. At least intuitively, longer periods off DAPT may be more dangerous than shorter ones, particularly if occurring in a vulnerable period after PCI. We do not know how many patients had multiple types of cessation patterns and in which combinations, but Figure 2 in Faggioni et al. (6) suggests that this happened in ∼7% in each group. Maybe the most glaring deficiency is the lack of analysis of adjusted risk of any DAPT cessation versus uninterrupted use, rather than breaking it down by modes of cessation. This would have provided both a more definitive answer to whether continuous DAPT is beneficial in general, and reveal a potential interaction with diabetic status (absent in the analyses by mode of cessation).
Altogether, the findings in this subanalysis from PARIS should bolster efforts to prevent development of DM in the first place. Once it occurred, DAPT is clearly not the answer to the quest to prevent its CV consequences, silver lining or not.
↵∗ Editorials published in JACC: Cardiovascular Interventions reflect the views of the authors and do not necessarily represent the views of JACC: Cardiovascular Interventions or the American College of Cardiology.
Dr. Brener has reported that he has no relationships relevant to the contents of this paper to disclose.
- 2017 American College of Cardiology Foundation
- Preis S.R.,
- Pencina M.J.,
- Hwang S.J.,
- et al.
- Preis S.R.,
- Hwang S.J.,
- Coady S.,
- et al.
- Goff D.C. Jr..,
- Lloyd-Jones D.M.,
- Bennett G.,
- et al.
- Faggioni M.,
- Baber U.,
- Sartori S.,
- et al.