Methods

Study design and population

The study was a prospective multicentre registry of patients older than 18 years receiving OAT who underwent PCI-S between November 2003 and June 2006 from nine catheterisation laboratories of tertiary care teaching hospitals in Spain and one in the United Kingdom. Exclusion criteria were life expectancy <3 months and the impossibility of follow-up. The study was approved by the Vall d’Hebron University institutional review board as well as by each participant hospital institutional review board, and patients gave their informed consent.

At each hospital, the designated doctor recorded demographic and clinical data, the reason for OAT, risk of thromboembolic events, type of procedure performed, the therapeutic regimen prescribed after stent implantation as well as all adverse events that occurred during admission and at 6 months (bleeding events and cardiovascular events such as are described above). Since this was an observational study, the decision whether to implant a specific stent and the therapeutic regimens to be used at discharge were made at the discretion of the attending interventional cardiologist or clinician.

Definitions

Thromboembolic risk

Thromboembolic risk was stratified as low, moderate or high according to a previous risk stratification15 16: low risk: atrial fibrillation (AF) with a CHADS2 score <2,15 history of deep-vein thrombosis or dilated cardiomyopathy with ventricular dysfunction; moderate risk: history of pulmonary thromboembolism or left ventricular thrombus, AF with CHADS2 score 2–315 or mitral valve disease without AF; and high risk: AF with CHADS2 score >3,15 mechanical prosthesis or a history of stroke.

Adverse events

Bleeding events

Bleeding was considered major when one or both of the following occurred: (a) any bleeding that lowered Hb by ⩾4 g/dl and required transfusion of ⩾2 IU of blood or corrective surgery; (b) cerebral haemorrhage or documented retroperitoneal haemorrhage, or any combination of these events.17

Cardiovascular events

Cardiovascular events were defined as the presence of stent thrombosis, myocardial infarction, target vessel revascularisation, stroke/peripheral embolism or cardiovascular death. Stent thrombosis was defined according to the Academic Research Consortium statement.18

End points

The primary end points were defined as (a) safety: the absence of a bleeding event during follow-up; (b) efficacy: the absence of any cardiovascular event during follow-up.

Follow-up

Patients were followed up by phone call 1 and 6 months after discharge and asked specifically about medication adherence and the presence of complications: bleeding events, cardiovascular events or death. When the answer was affirmative, clinical records of patients with hospital readmissions were reviewed.

Statistical analysis

Descriptive analysis was made using mean (SD) and range for continuous variables. Absolute and relative frequencies of patients in each category were analysed for categorical variables. Comparison of continuous variables among the three treatment groups was made by analysis of variance (ANOVA) and χ² test for categorical variables. The sequential Bonferroni correction was used for multiple comparisons.

The probability of adverse events during follow-up for each therapeutic group was calculated by Kaplan–Meier analysis and compared by the Mantel–Cox log-rank test. A Cox proportional hazards model was performed to assess the association between the therapeutic regimen at discharge and cardiovascular events. This association was adjusted using the variables “thromboembolic risk” and “number of vessels” as confounders. A two-tailed value of p<0.05 was considered significant. Statistical analysis was performed using the statistical package SPSS 15.0.

Results

Four hundred and five patients (328 (81%) men and 77 (19%) women, mean (SD) age 71 (9) years) receiving OAT underwent PCI-S for a coronary event. Three treatment groups were identified according to therapeutic regimens received at discharge. Most patients received TT (n = 278, 68.6%). TT included patients receiving warfarin plus DAT (n = 248) and patients receiving low molecular weight heparin (LWMH) plus DAT (n = 30). Forty-six patients (11.4%) were receiving AC+AT at discharge. AC+AT included patients receiving warfarin plus clopidogrel (n = 37), patients receiving warfarin plus acetylsalicylic acid (ASA—aspirin; n = 6), patients receiving LMWH plus clopidogrel (n = 2) and patients receiving LMWH plus ASA (n = 1). Finally, 81 patients (20%) were discharged with DAT only.

Baseline and procedural characteristics

Table 1 shows the clinical characteristics, thromboembolic risk and reason for OAT in the entire cohort. One hundred and fifty-seven out of 405 patients (38.8%) were at low thromboembolic risk, 132 (32.6%) at moderate risk and 116 (28.6%) at high risk. In the group of patients receiving DAT, there were a higher proportion of patients at low risk for thromboembolic events than in the other two groups (33.8% receiving TT, 41.3% receiving AC+AT and 54.3% receiving DAT, p = 0.003). The most common reason for OAT was AF (table 1).

Table 2 presents the procedural characteristics, including number of stents, type of stent, access site, use of access-site devices and anticoagulation status. One or more drug-eluting stents (DES) were used in 187 patients (46.2%) in the overall population. One hundred and forty-seven of the 324 patients undergoing any anticoagulant therapy (45.3%) received DES.

Follow-up and adverse events

All patients completed follow-up at 6 months, except one patient receiving AC+AT who moved 2 months after stenting. Table 3 shows the overall adverse events in the whole cohort and in the different subgroups according to thromboembolic risk. The incidence of overall adverse events in the whole cohort was high (20.2%) with major bleeding in 4%, cardiovascular events in 7.4% and overall mortality in 6.2% at 6 months’ follow-up.

Safety

In the whole cohort of patients, bleeding events occurred in 52 cases (12.8%), 16 (4%) of which were major. Thirty-one patients (59.6%), presented with a bleeding event within the first month following the procedure (fig 1A).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Cumulative survival free from bleeding events. Stratified analysis was performed according to the three treatments used at discharge. (A) Overall cohort; (B) patients at low thromboembolic risk; (C) patients at moderate and high thromboembolic risk. AC+AT, any anticoagulant plus a single antiplatelet agent; DAT, double antiplatelet therapy—aspirin plus clopidogrel; TT, triple therapy—any anticoagulant plus aspirin plus clopidogrel.

Overall bleeding events were more common in patients receiving TT (15.5% receiving TT, 13% receiving AC+AT and 3.7% receiving DAT, p = 0.02; fig 1A; log-rank, p = 0.030), even though there were no differences in the incidence of major bleeding among the treatment groups.

In patients at low thromboembolic risk, bleeding events were more common in patients receiving TT (14.6% receiving TT, 11.8% receiving AC+AT and 0% receiving DAT, p = 0.033; table 3). Interestingly, no patient receiving DAT had a bleeding event (fig 1B), (log-rank, p = 0.05). However, in patients at moderate–high thromboembolic risk there were no differences in bleeding complications among the treatment groups (fig 1C) (log-rank, p = 0.419).

Efficacy

In the whole cohort of patients, cardiovascular events occurred in 30 patients (7.4%) and were more frequent in patients receiving AC+AT (7.9% receiving TT, 15.2% receiving AC+AT and 1.2% receiving DAT, p = 0.013; fig 2A; log-rank, p = 0.015), especially in those at moderate–high thromboembolic risk (fig 2C; log-rank, p = 0.006). Remarkably, in the overall cohort a higher incidence of stent thrombosis occurred in patients receiving AC+AT, especially in the moderate–high risk group (table 3).

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

Cumulative survival free from cardiovascular events. Stratified analysis was performed according to the three treatments used at discharge. (A) Overall cohort; (B) patients at low thromboembolic risk; (C) patients at moderate and high thromboembolic risk. AC+AT, any anticoagulant plus a single antiplatelet agent; DAT, double antiplatelet therapy—aspirin plus clopidogrel; TT, triple therapy—any anticoagulant plus aspirin plus clopidogrel.

In a multivariable Cox analysis, the therapeutic regimen (p = 0.013), thromboembolic risk (HR = 3.08; 95% CI 1.00 to –9.42, p = 0.049) and number of injured vessels (HR = 1.69; 95% CI 1.04 to 2.75, p = 0.032) were independent predictors of cardiovascular events (p = 0.002). Patients receiving TT (HR = 0.22; 95% CI 0.087 to 0.56; p = 0.002) or DAT (HR = 0.05; 95% CI 0.007 to 0.47; p = 0.019), had a lower risk for this outcome than patients receiving AC+AT.

Mortality

Twenty-five patients (6.2%) died during follow-up. Causes of death were: cardiovascular in 16 (4%), bleeding in four (1%) and others in five (1%). Mortality did not differ significantly among the treatment groups. No patient at low thromboembolic risk receiving DAT died. Cardiovascular death was significantly higher in patients at moderate–high thromboembolic risk receiving AC+AT (p = 0.046; table 3).

Discussion

This is the first prospective registry specifically designed to assess the safety and efficacy of therapeutic strategies used in daily clinical practice after coronary stent implantation in chronically anticoagulated patients. Triple therapy was the most common therapeutic regimen and used independently of the level of thromboembolic risk. Patients receiving TT showed a higher incidence of bleeding events (15.5%) in agreement with published retrospective studies which show an increased bleeding risk associated with this strategy.6 7 8 9 10

Our study illustrates that the incidence of bleeding events in patients at low thromboembolic risk receiving TT exceeded the incidence of cardiovascular events with DAT. Our most notable finding was that no patient receiving DAT at low risk presented with either bleeding or cardiovascular events. Furthermore, DAT led to similar 6-month efficacy outcomes as other therapeutic options in this subgroup of patients. In agreement with our data, the ACTIVE study19 supported the idea that withdrawal of OAT while patients are receiving DAT should be considered for a limited period after PCI-S in patients at low risk of stroke (CHADS2 <2 in AF (and extending this concept to a history of deep-vein thrombosis or dilated cardiomyopathy with ventricular dysfunction).

Our results are difficult to compare with those reported by Ruiz-Nodar et al describing a series of patients with AF who demonstrated improved survival and lower rate of adverse events in patients who were receiving OAT. In their study, patients were not stratified according to thromboembolic risk, although the authors describe their population as high risk. Additional important methodological differences included the retrospective nature of the study, the large number of patients without well-defined paroxysmal AF and the high proportion (12%) of patients lost to follow-up.20

The combination of an anticoagulant plus any single antiplatelet drug (AC+AT) was associated with the worst rates of efficacy (particularly in patients at moderate and high thromboembolic risk), even after adjustment for confounding variables using Cox multiple regression analysis. Although the number of patients receiving this therapeutic regimen, and the number of events observed were low, a higher incidence of stent thrombosis was noted in this therapeutic group. This observation agrees to some extent with that reported by Nguyen et al, who found a higher incidence of stroke in patients receiving OAT combined with a single antiplatelet agent than in those who received TT.21 Other previous studies have shown that DAT is the “gold standard” after PCI-S with greater safety and efficacy compared with aspirin alone or in combination with warfarin.1 2 3 4

In our series, patients receiving DAT also showed a low rate of bleeding and cardiovascular events in patients at moderate–high thromboembolic risk. However, as expected, few patients in these subgroups received DAT—only eight patients at high thromboembolic risk were allocated to DAT after PCI-S. In all of them, the reason for receiving OAT before PCI was a history of stroke without AF, and none of them had a cardiovascular event during follow-up. Only one patient at moderate risk receiving DAT had a cardiovascular event.

Overall mortality was high (6.2%), cardiovascular death being the leading cause (4%). Sudden death occurred in seven patients, five of whom were receiving TT. Bleeding events were the cause of death in 1%. Treatment discontinuation for different reasons (including bleeding events) may underlie thromboembolic events or death and this supports the importance of adequately tailoring the choice of therapeutic strategy to the individual patient, both with respect to the choice of stent and combination of anticoagulant and/or antiplatelet therapy.

In summary, our results suggest that the risk–benefit ratio favours a short-term dual antiplatelet regimen in patients at low thromboembolic risk. Triple therapy should be restricted to patients at moderate or high thromboembolic risk.

This study was a prospective observational study with limitations inherent in this type of design. The percentage of highest risk patients (those with mechanical valves) is relatively low. This could limit the extrapolation of our findings to this population and therefore constitutes a limitation of our study. Additionally, there is a possible other bias secondary to the presence of some cluster effect associated to centre. For example, the development of stent thrombosis may be influenced not only by antithrombotic management but also by operator technique and the choice of stent within the different participating centres.

Our conclusions should be considered as hypothesis generating rather than confirmatory and should thus be tested in further studies. However, while we await conclusive data the current findings may guide prescription decisions and the choice of stent in patients receiving long-term anticoagulant therapy undergoing PCI-S.