Everolimus- and sirolimus-eluting stents in patients with and without ST-segment elevation myocardial infarction

The primary endpoint was a composite of cardiac death, MI and target vessel revascularisation (TVR). MI was defined as a rise of creatine kinase (CK) more than three times the upper limit of normal along with a rise in CK-MB with recurrent symptoms and/or new electrocardiographic changes. In acute coronary syndrome patients, re-infarction within 48 h after index procedure was defined as a re-elevation of CK of >1.5 times the previous value with elevation of CK-MB, along with recurrent symptoms and/or new electrocardiographic changes. MI around coronary artery bypass grafting required a CK rise of >5 times the upper limit of normal. TVR was defined as any repeat percutaneous or surgical intervention on any segment of the target vessel. Other secondary endpoints included the individual components of the composite endpoint, target lesion revascularisation (TLR) and definite or probable ST. TLR was defined as any repeat intervention or bypass grafting of the target lesion previously treated with stenting along with the 5 mm proximal or distal vessel. ST was defined in accordance with the Academic Research Consortium definitions [13].

Comparisons were made according to randomised treatment and presentation with or without STEMI. Continuous variables are presented as means with standard deviations or medians with interquartile range (IQR) and were compared using Student’s t-test. Categorical variables are expressed as counts and percentages and were compared by means of Pearson’s χ ² test. All statistical tests were 2-tailed and a p-value <0.05 was considered statistically significant. Time-to-event analyses were performed using Kaplan-Meier curves, which were compared using log-rank tests. To adjust for unbalanced baseline characteristics, Cox proportional hazards analyses were performed. The proportional hazards assumption was investigated visually. Adjusted effect sizes were calculated for primary and secondary endpoints with a p-value less than 0.10 as judged by log rank test. Adjustment was performed for characteristics significantly differing between groups (p < 0.05), which were incorporated in the multivariable models. Analyses were repeated with a variable stating the trial the patient originated from, to evaluate the influence of individual trials on the results.

To avoid dropping of events due to missing baseline information, multiple imputation was performed for the baseline variables that were included in the multivariable models: presence of heavy calcification was unknown in four patients (three EES and one SES) and total stent length was unknown in four patients (three EES and one SES). Reasons for missing data were unknown and assumed to be random. Total stent length was log transformed to meet the assumption for normal distribution. Missing data values were imputed for heavy calcification and total stent length using the following predictors: age, gender, cardiac risk factors, cardiac history, renal insufficiency, indication for PCI, target lesion, lesion type, number of vessel disease, heavy calcification, total stent length, max stent diameter, and number of stents per patient. Twenty imputed datasets were created and Cox proportional hazards analyses were performed on the pooled datasets [14]. Analyses were performed using IBM SPSS version 21.

STEMI patients were younger than patients without STEMI and more likely to smoke but had lower rates of comorbidity and other risk factors (Table 1). Coronary thrombus was more common in STEMI, but rates of heavy calcification, bifurcations and multivessel disease were lower. Stent length and number of stents used were also lower in STEMI, but stent diameter was slightly larger. Finally, glycoprotein IIb/IIIa inhibitor use was more common in STEMI while TIMI 3 flow after procedure was less often achieved. During 2-year follow-up, STEMI patients showed lower rates of the primary endpoint (7.2 % vs. 11.2 %, p = 0.007) and TLR (1.4 % vs. 4.1 %, p = 0.001) compared with patients without STEMI.

In the STEMI population, EES showed less calcified lesions and total stent length was shorter compared with SES (Table 2). During 2-year follow-up, randomisation to EES resulted in a similar primary endpoint rate (unadjusted HR 0.63, 95 % CI 0.36–1.09, p = 0.097, adjusted HR 0.66, 95 % CI 0.38–1.15, p = 0.141) compared with SES (Table 3, Fig. 1). A trend was observed for reduced early definite/probable ST in EES. However, long-term ST rates were low and similar (Fig. 2). At 1-year, aspirin (or coumadin) compliance was 94.8 % in EES versus 91.5 % in SES (p = 0.092). Thienopyridine compliance was 95.6 % in EES versus 91.8 % in SES (p = 0.040). Two patients were not on dual antiplatelet therapy at the time of ST: 1 EES and 1 SES patient, both suffering probable ST.

In the population without STEMI, EES patients showed lower rates of hypertension, prior PCI, bypass grafting and heavy calcification compared with SES patients (Tables 1 and 2). At 2 years, EES and SES showed similar rates of the primary endpoint (HR 1.10, 95 % CI 0.74–1.64, p = 0.637) (Table 3, Fig. 3). Other secondary endpoints were also balanced. Definite/probable ST rates were low and similar between the groups (Fig. 4). Aspirin compliance during 1 year was 97.6 % in EES and 99.3 % in SES (p = 0.047). Thienopyridine compliance was 96.8 % in EES and 97.5 % in SES (p = 0.518).

Our study is limited by its post hoc nature and therefore findings should be considered hypothesis generating. XAMI included only patients with acute MI, while APPENDIX-AMI included all-comer patients. Furthermore, the randomisation rate differed between XAMI (2:1) and APPENDIX-AMI (1:1) which created baseline misbalance between the groups, although multivariable corrections were performed to adjust for these differences. Finally, the analysis was underpowered to detect differences in the ST rates.