The Adverse Events and Coronary Artery Disease Progression (DECADE) cooperation is a pooled analysis of individual patient data from DES trials with 10-year follow-up.

We searched for randomized trials investigating 10-year clinical outcomes after DES implantation, across relevant electronic scientific databases. The final search was performed in October 2020. We identified 6 trials with full-length manuscripts.13–18 No further citations were identified after inspection of the reference lists from these eligible studies. The principal investigators of the studies were contacted to provide the individual data of participants assigned to DES implantation. One randomized trial was excluded, as the principal investigator did not agree to share patient-level data.14 Data from the remaining studies were transferred without patient identifiers and combined in a single pooled database. The included trials were ISAR TEST 4,19 ISAR TEST 5,20 SORT OUT III,21 SIRTAX22 and EXAMINATION.23 The final dataset was checked for completeness and consistency and compared with the results from prior publications. Principal investigators were directly contacted if there were inconsistencies with the original publications or requirements for additional data. Divergences were resolved by consensus. Data were analyzed according to the intention-to-treat principle.

The ISAR TEST 4 trial randomized 2603 patients to 3 DES treatment arms: new-generation, biodegradable-polymer sirolimus-eluting stents (SES) (n=1299), new-generation, permanent-polymer everolimus-eluting stents (EES) (n=652) and early-generation, permanent-polymer SES (n=652). The ISAR TEST 5 trial randomized 3002 patients to either new-generation, polymer-free sirolimus/probucol-eluting stents (n=2002) or new-generation, permanent-polymer zotarolimus-eluting stents (n=1000). The SIRTAX trial randomized 1012 patients to early-generation, permanent-polymer SES (n=503) or paclitaxel-eluting stents (n=509). The SORT OUT III trial randomized 2332 patients to new-generation, permanent-polymer zotarolimus-eluting stents (n=1162) or early-generation, permanent-polymer SES (n=1170). The EXAMINATION trial randomized 1498 patients to receive either a new-generation, permanent-polymer EES (n=751) or BMS (n=747). Ten-year results of the 5 individual trials included in the DECADE cooperation have already been published.13,15–18 The characteristics of the included trials are summarized in table 1. The key inclusion and exclusion criteria, primary endpoints and definitions of ST for the trials in the DECADE cooperation are summarized in table 1 of the supplementary data. Each study included in the present analysis was approved by the institutional review board or ethics committee at each participating center, and all patients signed informed, written consent before receiving the assigned treatment.

The primary endpoint of the current analysis was definite ST through to 10 years after PCI. Information regarding probable ST was not included in all trials and so we were unable to include this as an endpoint. For the analysis of this endpoint, patients were divided into 2 groups as per the generation of DES implanted, namely early DES and new DES. Stents included in the early DES group were permanent-polymer SES and paclitaxel-eluting stents. Stents included in the new DES group were permanent-polymer zotarolimus-eluting stents, biodegradable-polymer SES, permanent-polymer EES and polymer-free sirolimus/probucol-eluting stents. Data from the group of patients randomized to BMS in the EXAMINATION trial were excluded from this analysis, as they were not considered relevant to the study research question, which was focused on ST after DES implantation. In the SORT OUT III trial, ST data were available up to 5 years of follow-up. All endpoints were evaluated according to the definitions in the original trial protocols (table 1 of the supplementary data). ST was defined according to the Academic Research Consortium criteria in 4 of 5 trials included in the DECADE cooperation (ISAR TEST 4, ISAR TEST 5, SORT OUT III and EXAMINATION19–21,23). In the SIRTAX trial, definite ST was adjudicated in cases of acute coronary syndrome with angiographic documentation of either occlusion of the target lesion or thrombus within the previously stented segment.22 The definitions of ST used in the trials are summarized in table 1 of the supplementary data. We also provided data regarding mortality and myocardial infarction in patients treated with new- and early-generation DES, in order to provide more context to the definite ST results.

Individual participant data were analyzed using a 1-stage approach. Continuous data are presented as means±standard deviation or medians and interquartile ranges. Categorical data are presented as counts and proportions. Data distribution was tested for normality by using the Kolmogorov-Smirnov test for goodness-of-fit. Differences between groups were checked for significance using an analysis of variance test (ANOVA) for continuous data. Depending on the data distribution, the chi-squared test or Fisher exact test was used to check for differences between categorical variables. Here, a 2-tailed P value of <.05 was taken to confer statistical significance. Survival was analyzed by the Kaplan-Meier method to estimate the time to first event for the endpoint of interest and differences between the 2 groups were tested with the log-rank test. Hazard ratios (HR) and 95% confidence intervals (95%CI) were calculated using a Cox proportional hazards model after checking for fulfilment of the proportional hazards assumption as per the method of Grambsch and Therneau.24 The analysis of ST accounted for the competing risk of death using the cuminc function in the cmprsk package in R. Adjusted hazard ratios (HRadjusted) with pertinent 95%CI were reported. These were derived from a conventional multivariable analysis with adjustment for the following variables: age, sex, diabetes mellitus, hypertension, smoking, hypercholesterolemia, history of myocardial infarction, acute coronary syndrome, and vessel treated. We also performed a multivariate sensitivity analysis, which accounted for several angiographic and procedural variables in addition to patient characteristics with <5% missing values in the pooled dataset, including age, sex, diabetes, hypertension, smoking status, hypercholesterolaemia, previous myocardial infarction, acute coronary syndrome presentation, treated vessel, lesion complexity, balloon diameter and total stented length.

We performed a landmark analysis for ST for the following time periods: 0 to 30 days (accounting for acute and subacute ST), 30 days to 1 year (accounting for late ST), 1 to 5 years (accounting for VLST) and 5 to 10 years (accounting for VVLST). ST event rates were also calculated for these time periods, compared using the exact 2-sided Poisson test and expressed as a rate ratio (RR) and 95%CI. For ST from 0 to 30 days after PCI, rates of ST were expressed as the number of events per 1000 patient days of follow-up. From 30 days to 1 year, 1 to 5 years and 5 to 10 years after PCI, the ST rate was expressed as the number of events per 1000 patient years of follow-up. The statistical analysis was performed using the R 3.6.0 Statistical Package (The R Foundation for Statistical Computing, Austria).

Table 2 displays the baseline characteristics for the cohort as a whole and as per the stent generation implanted. Patients in the new DES group were older and were more frequently diabetic and hypertensive. Current smokers at the time of PCI were more frequent in the early DES group. In over half of patients in both the early DES and new DES groups, the indication for PCI was stable angina. Table 3 displays the baseline angiographic and procedural characteristics both for the cohort as whole and as per the stent generation implanted. There was a higher proportion of lesions with complex morphology and treated bifurcation lesions in the new DES group. In addition, the new DES group had a longer total stented length and a higher total number of implanted stents than the early DES group.

Definite ST occurred in 160 of 9700 patients (1.6%) through to 10-year follow-up. The incidence of definite ST through to 10 years as per each individual stent type is shown in table 2 of the supplementary data. In the new DES group, definite ST occurred in 69 of 6866 patients (1.0%) through to 10-year follow-up. In the early DES group, definite ST occurred in 91 of 2834 patients (3.5%) through to 10-year follow-up. The cumulative incidence of definite ST through to 10 years was lower in patients treated with new DES than in those treated with early DES in both the unadjusted (HR, 0.30; 95%CI, 0.22-0.41) and adjusted (HRadjusted, 0.32; 95%CI, 0.23-0.45) analyses. These data are presented in figure 1 and table 4.

Figure 1.

Ten-year cumulative incidence of definite stent thrombosis as per drug-eluting stent generation. The hazard ratio reported here was derived from a conventional multivariable analysis with adjustment for the following variables: age, sex, diabetes mellitus, hypertension, smoking, hypercholesterolemia, history of myocardial infarction, acute coronary syndrome, and vessel treated. 95%IC, 95% confidence interval; HR, hazard ratio.

(0.15MB).

Figure 2 and table 4 demonstrate the incidence of definite ST within 4 time-periods: 0 to 30 days, 30 days to 1 year, 1 to 5 years, and 5 to 10 years. Compared with the early DES group, the incidence of definite ST was lower in the new DES group from 1 to 5 years (0.3% vs 1.8%, HRadjusted, 0.16; 95%CI, 0.09-0.28) and from 5 to 10 years (0.2% vs 0.9%, HRadjusted, 0.25; 0.10-0.60) after PCI.

Figure 2.

Landmark analysis of definite stent thrombosis up to 10 years after percutaneous coronary intervention as per drug-eluting stent generation.

95%CI, confidence interval; HR, hazard ratio.

(0.22MB).

From 0 days to 30 days after PCI, the definite ST rate was 0.14 (95%CI, 0.10-0.20)/1000 patient days in the new DES group and 0.25 (95%CI, 0.15-0.38)/1000 patient days in the early DES group (RR, 0.57; 95%CI, 0.31-1.05). From 30 days to 1 year after PCI, the definite ST rate was 2.30 (95%CI, 1.26-3.85)/1000 patient years in the new DES group and 3.97 (95%CI, 1.90-7.29)/1000 patient years in the early DES group (RR, 0.58; 95%CI, 0.24-1.46). From 1 to 5 years, the definite ST rate was 0.69 (95%CI, 0.40-1.11)/1000 patient years in the new DES group and 4.80 (95%CI, 3.55-6.34)/1000 patient years in the early DES group (RR, 0.14; 95%CI, 0.08-0.26). From 5 to 10 years, the definite ST rate was 0.46 (95%CI, 0.21-0.87)/1000 patient years in the new DES group and 1.99 (95%CI, 0.99-3.57)/1000 patient years in the early DES group (RR, 0.23; 95%CI, 0.08-0.61). These data are presented in table 5.

In the multivariate sensitivity analysis (see Methods for details), new DES were also associated with a reduced risk of definite ST than early DES at 10 years (HR, 0.27; 95%CI, 0.19-0.39). The reduced risk of definite ST with new DES was confirmed in this model from 0 to 1 years (HR, 0.57; 95%CI, 0.34-0.94) and from 1 to 10 years (HR, 0.11; 95%CI, 0.06-0.21).

During follow-up, death occurred in 2004 of 6866 patients (30.4%) in the new DES group and in 765 of 2834 patients (28.3%) in the early DES group (HRadjusted, 0.97; 95%CI, 0.89-1.06). Myocardial infarction occurred in 461 of 6866 patients (6.9%) in the new DES group and in 291 of 2834 patients (10.7%) in the early DES group (HRadjusted, 0.66; 95%CI, 0.57-0.77). At 10 years, 31 of 74 patients who had experienced a definite ST event at 1 year died compared with 2738 of 9626 patients who did not experience definite ST at 1 year (41.9% vs 28.4%, P=.01).

This is a post hoc analysis of individual patient data from 5 RCTs. As such, it has the usual limitations associated with post hoc analyses and should be regarded as hypothesis generating. For the purposes of this analysis, stents implanted were categorized as early and new DES. However, there are differences between the stent platforms within these 2 categories and therefore this dichotomization could be considered overly reductionist or arbitrary. Of interest, the incidence of definite ST between the individual stent types grouped in the new and early DES cohorts appeared to support this dichotomization. Nevertheless, the present analysis is largely underpowered to assess differences in terms of ST within the new and early DES groups. In addition, patients were not randomized in all the included studies to treatment with either early or new DES and there were some differences in baseline and procedural features between the 2 groups. In light of these arguments, although risk estimates for the primary endpoint went in the same direction on both unadjusted and adjusted analyses, a potential bias due to residual confounding cannot be definitively ruled-out in this context. The risk of ST is determined not only by stent-related factors but also by other factors, including iteration of implantation techniques, patient selection and periprocedural and long-term antithrombotic regimens over time. All these factors may also have varied between studies and contributed to the observed differences.

Another potential limitation is that we did not have detailed information on long-term antiplatelet therapy or secondary prevention measures for patients enrolled in the original studies. It is also important to consider that the PCI procedures performed in the trials included in the DECADE cooperation may not reflect current practice. While event adjudication processes were similar between all trials included in the DECADE cooperation, events were not adjudicated centrally in this analysis and therefore we cannot exclude heterogeneity with regard to the reporting and adjudication of events between the individual trials.

Another important point is that the patients enrolled in the trials included in the DECADE cooperation represented a selected cohort and therefore may not be fully representative of patients encountered in clinical practice. Given that this analysis focused only on definite ST in a relatively selected cohort of PCI patients, it may have underestimated the true incidence of ST.