Coronary obstruction (CO) due to TAVR is a rare (0.7%) but frequently catastrophic complication (mortality 40%). Predicting CO is critical, but predictions based on coronary height and/or sinus of Valsalva width are imprecise.    

Two large global registries (22 centers, 13 countries) of TAVR patients suffering from CO (95 patients) were analyzed, including 60 patients with CCTA analysis. Comparisons were made to 1381 patients who underwent TAVR without CO. 78% of patients had left CO, 17% had right CO, and 5% had both. In hospital mortality was 27%.  

Predictors of CO included smaller coronary heights and sinus of Valsalva widths, with optimal cutoffs for the LCA of 11 mm and 31 mm, respectively, and for the RCA of 15 mm and 29 mm, respectively, with AUCs ranging from 0.71 to 0.85. Additionally, a smaller residual sinus diameter (sinus of Valsalva minus valve size) predicted CO, with an optimal cutoff of 5 mm for both RCA and LCA (AUC 0.76 and 0.78, respectively).

Advanced CT analysis revealed cusp height > coronary height to be very sensitive for coronary obstruction (97%). In addition, valve to coronary distance (VTC), measured using a “virtual valve” of < 4 mm or leaflet calcium volume of > 600 mm3, was 96% sensitive. However, specificity could not be determined due to the lack of control group measurements.  

In conclusion, these registries demonstrated CO has a high mortality of 27%, albeit lower than previously reported. Optimal cutoffs of coronary height (11 mm for LCA and 15 mm for RCA) and sinus of Valsalva width (31 mm for LCA and 29 mm for RCA), and the more valve-specific residual sinus diameter (5 mm for both) for predicting CO were identified. The more advanced measurements of cusp height < coronary height, as well as VTC distance > 4 plus leaflet calcium volume < 600 mm3, likely have powerful negative predictive value for CO.