Introduction Despite major advances in coronary stent technology, stent thrombosis (ST) remains a significant complication of percutaneous coronary intervention (PCI) with a high rate of MI and death. The underlying mechanisms are multi-factorial and poorly understood. Previous studies have involved small sample sizes, with incomplete patient characterisation, a lack of intracoronary imaging or platelet function data.

Methods As part of the European multi-centre PRESTIGE project, we established and coordinated a ST study, collecting multi-source data from patients presenting with definite ST between April 2012 and October 2014 at 12 UK cardiac centres. Demographic and clinical data were collected and all angiographic images reviewed. Intracoronary thrombus was examined histopathologically and intravascular ultrasound (IVUS) and/or optical coherence tomography (OCT) performed whenever possible. We undertook platelet function testing (PFT) acutely, at 24 h and at 30–60 days, and collected blood for thrombin generation assays and DNA analysis. All non-invasive studies including PFT were also undertaken in an average of 2.6 matched controls per ST case.

Results A total of 138 patients with definite ST were recruited in the UK, with 353 controls. Of the ST cases, 4.3% were acute (within 24 h), 18.8% sub-acute (24 h to 1 month), 8.7% late (1–12 months) and 68.1% very late (> 1 year) after the initial PCI. PFT was carried out in 74.6% and 86.4% of the ST patients and controls respectively, and 89.1% of the ST cases and 98.0% of the control patients had a blood sample stored for DNA analysis. Of the ST cases, 63.8% had thrombus retained, 38.4% had IVUS and 21.7% had OCT imaging.

The ST cases and control groups were well matched for gender, hypertension, heart failure, renal impairment and indication for PCI. The ST group were younger (mean age 59.2 ± 11.6 vs 63.2 ± 10.5 years; p < 0.01), had more smokers (39.1% vs 20.7%; P < 0.01), previous MI (32.6% vs 21.4%; p < 0.01), diabetes mellitus (22.5% vs 14.2%; p = 0.03) and active malignancy (3.6% vs 0.6%; p = 0.01). The angiographic features were well matched between the two groups, apart from less circumflex artery involvement in the ST group (13.0% vs 24.9%; p < 0.01). Factors that differed between the ST cases and controls were use of aspirin on presentation in the ST group (84.1% vs 92.1%; p = 0.01) non-cardiac surgery in the 90 days prior to randomisation (6.5% vs 2.8%; p = 0.04), number of stents inserted (2.2 vs 1.7; p < 0;01), a higher proportion of first generation drug eluting stents (31.9% vs 19.0%; p < 0.01) and of two-stent bifurcation procedures (6.5% vs 2.8%; p = 0.05).

Conclusions Using a multi-centre approach, we collected multi-source data from patients presenting with coronary ST, along with matched controls. PFT, thrombin generation and OCT data have been analysed to identify important risk factors for ST and these will be presented.