We are grateful to Yue-Hua Ruan and Ren-Qiang Yang for their interest in our Review (Predictors of stent thrombosis and their implications for clinical practice. Nat. Rev. Cardiol. 16, 243–256 (2019))1 and for their helpful comments (Relationship between stent fracture and thrombosis. Nat. Rev. Cardiol. https://doi.org/10.1038/s41569-019-0262-6 (2019))2. We absolutely agree that the relationship between stent strut fracture and thrombosis (as well as restenosis) requires further investigation.

No agreement exists on the true prevalence of strut rupture, which varies from <1% to 20% in the literature. This variability is influenced by factors such as time after implantation, stent type and implantation technique used, and vessel characteristics. Most importantly, the diagnosis of stent fracture is also influenced by the spatial resolution of the imaging method used. The resolution of either standard radiography or intravascular ultrasonography does not allow stent fractures to be systematically assessed. Optical coherence tomography also has limitations — for instance, in the setting of stent failure, when blood flushing is often incomplete, and late after stenting, when the visibility of the struts is reduced owing to the presence of neointima. Furthermore, invasive imaging methods provide little information when a stent fracture is not associated with dislodgement of the stent segments.

The discordance between reports also results from the lack of a standardized definition of stent fracture. Nakazawa and colleagues propose a classification for this entity on the basis of pathology findings, which unfortunately cannot be directly applied in routine clinical practice3. We agree with Ruan and Yang that such a classification (most likely on the basis of intracoronary imaging evidence) might help to define the strength and the mechanisms of the association between stent fracture and thrombosis (including whether only ‘major’ fractures lead to an increased risk of thrombotic or restenotic events).