Introduction

Drug-eluting stents (DESs) have been a major advance in the treatment of patients with obstructive coronary artery disease, but their use has been associated with stent thrombosis (ST) and myocardial infarction (MI). Stent thrombosis after bare-metal stent (BMS) implantation typically occurs within 30 days, but it can occur later [1]. Stent thrombosis may occur years after implantation of drug-eluting stents (DESs), but the lifetime risk of ST for a first-generation DES does not seem to be higher than that for a BMS [2].

Dual antiplatelet therapy (DAPT), which is defined as the combination of aspirin and a platelet P2Y12 receptor antagonist, reduces ST after implantation of BMSs by 85 % (relative risk 0.15, 95 % confidence interval [CI] 0.05–0.43), as compared with aspirin alone [3]. Based on the findings in two randomized controlled trials (RCTs), namely PCI-CURE (a substudy of the Clopidogrel in Unstable angina to prevent Recurrent Events trial) [4] and CREDO (Clopidogrel for the Reduction of Events During Observation) study [5], and in observational studies reporting a persistent risk of ST risk beyond 6 months after stent implantation [6, 7], particularly in the context of DAPT cessation [8], the 2011 American College of Cardiology/American Heart Association guideline recommended a minimum duration of DAPT of at least 12 months after DES implantation [9]. In PCI-CURE, however, stenting comprised 80 % of the cases, and all stents were BMS. Similarly, in the CREDO trial, all implanted stents were BMS and only 63 % of patients assigned to clopidogrel completed 1 year of therapy [5].

After the publication of the 2011 guideline [9], 10 new RCTs have compared shorter courses (3–12 months) with longer courses (12–36 months) of DAPT in patients receiving newer generation DESs [10–19, 20•, 21•]. In order to identify the risks and benefits of prolonged DAPT in the 10 RCTs, the current analysis was performed to answer the question: As compared with 12–36 months of DAPT, did 3–12 months of DAPT after implantation of newer generation DESs result in differences in (1) mortality, (2) major hemorrhage, (3) MI, and (4) ST?

Methods

Aggregate data from 10 RCTs of patients undergoing newer generation DES implantation and randomized to either shorter or longer courses of DAPT [10–19, 20•, 21•] comprised the evidence base for the analysis (Table 1). Summary data from each trial were abstracted in triplicate and validated against published reports [21•, 22•]. Whenever possible, DES-enriched data sets [20•, 21•] were preferred over data sets from the same trials containing subjects treated with both BMSs and DESs [12, 23].

Table 1 Randomized controlled trials of dual antiplatelet therapy duration after implantation of drug-eluting stents
Full size table

The primary outcomes of the present analysis were all-cause mortality, major hemorrhage, MI, and definite or probable ST. Newer generation DESs were defined as everolimus-eluting stents (EESs), zotarolimus-eluting stents, and biolimus-eluting stents with biodegradeable polymers [24•].

Forest plots were created to convey the relative effectiveness of shorter versus longer durations of DAPT. A random-effects model was preferred to acknowledge the variations in study design, duration of treatment and differences in follow-up among the 10 RCTs. All analyses were intention-to-treat. Original meta-analyses were created with [R] 3.0.2 [25••] and library package meta 3.8–0 [26••]. Effect-size estimates and numbers needed to treat (NNTs) were calculated from the random-effects models using study weights, odds ratios (ORs), and 95 % CIs, using standard methods [27].

Results

The 10 RCTs [10–19, 20•] enrolled 31,666 patients, of whom 523 (1.7 %) died, 338 (1.1 %) had major hemorrhage, 597 (1.9 %) had an MI, and 177 (0.6 %) had definite or probable ST. The RCTs generally enrolled low-risk patients. Eight of 10 trials found lower than expected event rates (Table 1). The proportion of patients in each trial with prior MI ranged from 1.7 to 34.7 % (Table 1), and the overall pooled proportion of patients with prior MI was 20.0 %. However, the trials were judged to be of high quality with at least moderate freedom from bias and to have intermediate to high grades for relevance and fidelity [22•].

Relative Differences Between Shorter and Longer Durations of DAPT

A systematic analysis of the 10 RCTs suggested that shorter courses of DAPT of 3–6 months duration were associated with lower mortality (OR 0.83, 95 % CI 0.69–0.98) and major hemorrhage (OR 0.60, 95 % CI 0.48–0.75) but increased rates of MI (OR 1.34, 95 % CI 1.04–1.73) and ST (OR 1.75, 95 % CI 1.08–2.82) than were longer courses of 12–36 months duration (Fig. 1).

Fig. 1
figure 1

Forest plot of mortality rates (a), major hemorrhage (b), myocardial infarction (c), and definite or probable stent thrombosis (d) after shorter courses (3–12 months) or longer courses (12–36 months) of dual antiplatelet therapy for drug-eluting stent implantation. CI confidence interval, OR odds ratio

Full size image

Absolute Differences

A weighted analysis of the 10 RCTs suggested that shorter courses of DAPT of 3–6 months duration were associated with significantly lower annual rates of all-cause mortality (1.57 vs. 1.88 % [95 % CI 1.60–2.26 %]) and major hemorrhage (0.76 vs. 1.26 % [95 % CI 1.01–1.57 %]) but higher annual rates of MI (1.94 vs. 1.46 % [95 % CI 1.13–1.86 %]) and ST (0.59 vs. 0.34 % [95 % CI 0.21–0.55 %]) than were longer courses of DAPT of 12–36 months duration.

After shorter courses of DAPT, 16 patients per 1000 died annually compared with 19 (95 % CI 16–23) patients per 1000 who died in the same period of follow-up while being treated with longer courses of DAPT. After shorter courses of DAPT, 8 patients per 1000 had major hemorrhage annually compared with 13 (95 % CI 10–16) patients per 1000 who had hemorrhage annually during longer courses of DAPT. On the other hand, after shorter courses of DAPT, 19 patients per 1000 had MI annually compared with 15 (95 % CI 11–19) patients per 1000 who had MI annually with longer courses of DAPT, and 6 patients per 1000 had ST annually compared with 3 (95 % CI 2–5) patients per 1000 who had ST annually while under treatment with longer courses of DAPT during the same period of follow-up.

The NNT with shorter courses of DAPT to prevent 1 death was 325 (95 % CI 145–3237). The NNT with shorter courses of DAPT to prevent 1 major bleed was 199 (95 % CI 124–403). On the other hand, the NNT with longer courses of DAPT to prevent 1 MI was 210 (95 % CI 125–1314), and the NNT with prolonged courses of DAPT to prevent 1 ST was 396 (95 % CI 262–2435).

Discussion

This systematic review evaluated the outcomes in patients enrolled in 10 RCTs [10–19, 20•, 21•] after implantation of newer generation DESs and identified a tradeoff between shorter and longer courses of DAPT. Shorter courses (3–12 months) of DAPT resulted in lower mortality and major bleeding but caused more MIs and STs than did longer courses (12–36 months) of DAPT.

Mortality

The finding of decreased mortality with shorter courses of DAPT is consistent with results of other analyses [21•, 22•, 28•, 29•] but seems counterintuitive given that prolonged courses of DAPT reduced both MI and ST. The mortality finding may reflect the interplay of several factors. First, newer generation DESs have a lower risk of ST than first-generation stents, which may have diminished the ability of longer courses of DAPT to protect against late thrombotic events related to the stent itself. In support of this, Baber and colleagues [30] reported that the risk of ST after implantation of newer generation EESs was almost 50 % lower than that for first-generation DESs (OR 0.55 95 % CI 0.38–0.78), similar to the reduction in risk of ST for EESs compared with other stent types reported in the DAPT appendix [20•]. Second, the risk of a fatal outcome from ST declines over time. Whereas acute and subacute ST are associated with mortality rates approaching 50 %, late and very late ST are associated with mortality rates of about 10 % [31, 32]. Third, bleeding early after percutaneous coronary intervention (PCI) probably has a larger effect on late mortality than do ischemic complications [33, 34]. Emerging evidence suggests that giving transfusions after PCI strongly increases the risk of dying during follow-up [35], which may be mediated by transfusion-related lung injury, volume overload, systemic inflammation, platelet activation, microvascular plugging and reduced nitric oxide availability in the presence of aging erythrocytes. Thus, the overall survival advantage seen with shorter courses of DAPT may be explained by a declining risk of ST and MI associated with the use of newer generation DESs counterbalanced by a uniform risk of bleeding and its life-threatening consequences associated with longer courses of DAPT.

Clinical Implications

All antithrombotic regimens involve a “fragile balance between efficacy and adverse events” [36]. Many physicians will use longer courses of DAPT in patients who have not had a bleeding complication on DAPT and have an increased systemic risk of atherothrombotic events or increased risk of late ST. In selected patients with prior MI, longer courses of DAPT may be reasonable regardless of whether coronary stenting was used [37•]. In patients with high-risk angiographic findings such as multiple or overlapping stents, implantation of first-generation paclitaxel-eluting or sirolimus-eluting stents, prior stent thrombosis or impaired left ventricular function [38], longer courses of DAPT may be weighed as an option to prevent late ischemic events.

Conclusions

The findings in the current analysis may apply to populations similar to those enrolled in the RCTs [10–19, 20•, 21•] comparing shorter with longer durations of DAPT after implantation of predominantly newer generation DESs. In low-risk patients who have undergone newer generation DES implantation, a minimum DAPT duration of 3–12 months may be sufficient to prevent early and largely stent-related thrombotic events. Extension of DAPT beyond 12 months after implantation of newer generation DES entails a tradeoff. Patients at very low risk of bleeding or high risk of thrombotic events may derive a benefit from extension of DAPT beyond 12 months, but the inability to predict future bleeding limits the appeal of prolonged therapy.