Although the presence of uncovered struts may be associated with occurrence of stent thrombosis, the impact of uncovered struts detected routinely by optical coherence tomography (OCT) on subsequent long-term clinical outcomes remains unclear. The objective of this study was to determine the cut-off value of uncovered struts that predicted adverse clinical outcomes after drug eluting stent (DES) implantation. Major safety events (MSEs, a composite occurrence of cardiovascular death, myocardial infarction, and stent thrombosis) were evaluated in 489 DES-treated patients (535 lesions) during the median 851 days after follow-up OCT. MSEs occurred in six patients (four definite stent thrombosis and two sudden cardiac death). The best cut-off value of percentage of uncovered struts for predicting MSE was 5.9 % using the maximal χ² method: area under the receiver-operating characteristic curve = 0.779, 95 % confidence interval (CI) = 0.648–0.910, p = 0.019, a sensitivity of 83.3 % and a specificity of 70.3 %. Independent predictors for MSE were post-intervention minimal lumen diameter (odds ratio 0.019, 95 % CI = 0.001–0.513, p = 0.018) and percentage of uncovered struts ≥5.9 % (odds ratio 19.781, 95 % CI = 2.071–188.968, p = 0.010). A greater percentage of uncovered struts (the cut-off value of ≥5.9 % uncovered struts) might be significantly associated with occurrence of MSE after DES implantation.