Impact of Dialysis on the Prognosis of Patients Undergoing Transcatheter Aortic Valve Implantation.


Department of Cardiology, Medical Clinic IV, Municipal Hospital Karlsruhe, Academic Teaching Hospital of the University of Freiburg, Germany. Electronic address: [Email]


End-stage renal disease (ESRD) affects approximately 2% to 4% of patients with severe aortic stenosis. It is because these patients have been excluded from clinical trials, the impact of transcatheter aortic valve implantation (TAVI) in this patient group has not been thoroughly investigated. Between April 2008 and March 2015, 2,000 patients (dialysis group, n = 56 [2.8%]) were consecutively enrolled when diagnosed with severe aortic stenosis and eligible to undergo TAVI. Procedural and longer-term outcomes were analyzed and adjusted for differences in baseline characteristics. Patients on dialysis had a higher periprocedural mortality (10.7% vs 1.7%; adjusted odds ratio [adjOR] 5.65, 95% confidence interval [CI] 1.91 to 16.67; p = 0.002) and a lower Valve Academic Research Consortium (VARC)-II (VARC) defined device success (adjOR 0.34, 95% CI 0.15 to 0.79; p = 0.012). At 30 days, there was an increased rate of all-cause mortality (21.4 vs 4.8%; adjOR 4.90, 95% CI 1.96 to 12.26; p = 0.001), cardiovascular (adjOR 3.67, 95% CI 1.43 to 9.41; p = 0.007) and noncardiovascular mortality (adjOR 6.28, 95% CI 1.36 to 9.41; p = 0.019), myocardial infarction (adjOR 9.39, 95% CI 1.84 to 48.03; p = 0.007), bleeding (adjOR 2.48, 95% CI 1.06 to 5.83; p = 0.036) as well as the VARC-II defined early safety combined end point (adjOR 2.97, 95% CI 1.28 to 6.90; p = 0.012) associated with dialysis. Dialysis was associated with poor survival at one (57.1% vs 84.2%) and 3 years (26.8% vs 66.9%) with or without the consideration of the first 72 hours (p <0.001; adjusted p <0.001). Although, in the multivariable regression analysis, reduced ejection fraction, peripheral arterial disease, pulmonary hypertension (PH), frailty and dialysis were associated with 1-year mortality, only PH (>60 mm Hg) remained significant in an analysis restricted to the dialysis patients (adjusted hazard ratio 2.68; 95% CI 1.18 to 5.88; p = 0.018). PH had a sensitivity of 45.8%, a specificity of 81.3%, and a positive predictive value of 64.7%. In conclusion, dialysis is an independent predictor of mortality in patients who underwent TAVI. Long-term mortality in dialysis patients appears to be largely determined by the kidney disease and/or dialysis itself whereas VARC-II defined complications are largely unaffected. An increased short-term mortality still calls for (pre-) procedural optimization.