Correlates and outcomes of late and very late drug-eluting stent thrombosis: results from DESERT (International Drug-Eluting Stent Event Registry of Thrombosis).

Objectives: The aim of this study was to identify clinical, procedural, and angiographic correlates of late/very late drug-eluting stent (DES) thrombosis as well as to determine the clinical outcomes of these events.

Background: Late/very late DES thromboses are a poorly studied phenomenon, partly due to the relative infrequency of these events, even in large cohort studies.

Methods: In the DESERT (International Drug-Eluting Stent Event Registry of Thrombosis), a retrospective, case-control registry, 492 cases of late/very late definite DES thrombosis from 21 international sites were matched in a 1:1 fashion with controls without stent thrombosis (ST). Controls were matched according to 2 criteria: same enrolling institution and date of initial DES implantation. Baseline and procedural variables were collected, and clinical follow-up was obtained for patients with ST as long as 1 year after the event. Offline quantitative coronary angiography was performed for a subset of 378 case-control pairs.

Results: The majority of ST events occurred after 1 year (75%) and continued to occur for as long as 7.3 years. The clinical presentation of late/very late ST events was mainly myocardial infarction (66.7% ST-segment elevation myocardial infarction and 22.0% non-ST-segment elevation myocardial infarction); in-hospital mortality was 3.8%. A minority of patients (30%) with ST were receiving dual-antiplatelet therapy at the time of the event. Independent clinical correlates of late/very late ST were younger age, African-American race, current smoking, multivessel disease, longer stented length, overlapping stents, and percutaneous coronary intervention of vein graft lesions. Independent angiographic correlates for late/very late ST were lesions within the left anterior descending artery or a bypass graft, thrombus, and a larger residual diameter stenosis after the initial DES implantation. Despite the large sample of ST cases, all identified correlates of late/very late ST had weak associations with subsequent ST (all odds ratios <2.5).

Conclusions: Despite a large sample of ST cases and use of limited matching to maximize the identification of predictive factors associated with late/very late ST, the variables associated with the development of late/very late ST were only weakly predictive of subsequent events. Additionally, a relatively low observed mortality rate of ST in this series may reflect a different pathophysiology of these late/very late events compared with acute/subacute ST. (Drug Eluting Stent Registry of Thrombosis [DESERT]; NCT00812552).