Incidence, predictors, and outcomes associated with acute kidney injury in patients undergoing transcatheter aortic valve replacement: from the BRAVO-3 randomized trial.

Background Acute kidney injury (AKI) is not uncommon in patients undergoing transcatheter aortic valve replacement (TAVR). Objective We examined the incidence, predictors, and outcomes of AKI from the BRAVO 3 randomized trial. Methods The BRAVO-3 trial included 802 patients undergoing transfemoral TAVR randomized to bivalirudin vs. unfractionated heparin (UFH). The primary endpoint of the trial was Bleeding Academic Research Consortium (BARC) type ≥ 3b bleeding at 48 h. Total follow-up was to 30 days. AKI was adjudicated using the modified RIFLE (Valve Academic Research Consortium, VARC 1) criteria through 30-day follow-up, and in a sensitivity analysis AKI was assessed at 7 days (modified VARC-2 criteria). We examined the incidence, predictors, and 30-day outcomes associated with diagnosis of AKI. We also examined the effect of procedural anticoagulant (bivalirudin or unfractionated heparin, UFH) on AKI within 48 h after TAVR. Results The trial population had a mean age of 82.3 ± 6.5 years including 48.8% women with mean EuroScore I 17.05 ± 10.3%. AKI occurred in 17.0% during 30-day follow-up and was associated with greater adjusted risk of 30-day death (13.0% vs. 3.5%, OR 5.84, 95% CI 2.62-12.99) and a trend for more BARC ≥ 3b bleeding (15.1% vs. 8.6%, OR 1.80, 95% CI 0.99-3.25). Predictors of 30-day AKI were baseline hemoglobin, body weight, and pre-existing coronary disease. AKI occurred in 10.7% at 7 days and was associated with significantly greater risk of 30-day death (OR 6.99, 95% CI 2.85-17.15). Independent predictors of AKI within 7 days included pre-existing coronary or cerebrovascular disease, chronic kidney disease (CKD), and transfusion which increased risk, whereas post-dilation was protective. The incidence of 48-h AKI was higher with bivalirudin compared to UFH in the intention to treat cohort (10.9% vs. 6.5%, p = 0.03), but not in the per-protocol assessment (10.7% vs. 7.1%, p = 0.08). Conclusion In the BRAVO 3 trial, AKI occurred in 17% at 30 days and in 10.7% at 7 days. AKI was associated with a significantly greater adjusted risk for 30-day death. Multivariate predictors of AKI at 30 days included baseline hemoglobin, body weight, and prior coronary artery disease, and predictors at 7 days included pre-existing vascular disease, CKD, transfusion, and valve post-dilation. Bivalirudin was associated with greater AKI within 48 h in the intention to treat but not in the per-protocol analysis.