From January 2008 to January 2015, 819 consecutive patients with symptomatic severe aortic stenosis underwent TAVI and were included in a multicenter Brazilian registry.22 Twenty-two sites from different regions of Brazil participated in the study. For each patient, baseline characteristics, procedure details, and follow-up data were collected and recorded in a web-based case report form designed especially for this registry. To better understand the predictors and prognostic role of AKI following TAVI, we excluded 25 patients who died in the first 24hours after the procedure. Finally, 794 participants were included in this analysis. Informed written consent was obtained from the patients included prospectively. The study protocol was conducted in accordance with the Declaration of Helsinki and was approved by each institution's ethics committee.

The self-expandable CoreValve (Medtronic, Minneapolis, Minnesota, United States), the balloon-expandable SAPIEN XT (Edwards Lifesciences, Irvine, California, United States) or the Inovare (Braile Biomédica, São José do Rio Preto, São Paulo, Brazil) prosthesis were used. Transcatheter aortic valve implantation procedures were performed according to standard techniques. The type of anesthesia used and access site were left to the operator's discretion. Transfemoral vascular access was the first-choice approach. When transfemoral access was not feasible, transapical or transarterial approaches (transubclavian, direct transaortic, transapical or transcarotid) were used according to the preference of the Heart Team.

Patients were followed up and all adverse events were adjudicated by an independent committee, according to the VARC-2 definitions. The definition of AKI was based on the AKIN (Acute Kidney Injury Network) criteria as follows: stage 1 (increase in serum creatinine to 150%-199% compared with baseline or increase of 0.3mg/dL or urine output < 0.5mL/kg/h for > 6 but < 12h); stage 2 (increase in serum creatinine to 200%-299% compared with baseline or urine output < 0.5mL/kg/h for > 12 but < 24h); stage 3 (increase in serum creatinine to > 300% compared with baseline or serum creatinine of > 4.0mg/dL with an acute increase of at least 0.5mg/dL or urine output < 0.3mL/kg/h for > 24h or anuria for > 12h or renal replacement therapy). The baseline serum creatinine was measured at the hospital admission before the procedure and daily thereafter, up to 1 week or less if the patient was discharged before this period. In comparison with the original VARC definitions, the timing for the diagnosis of AKI was extended from 72hours to 7 days. Chronic kidney disease was defined as estimated glomerular filtration rate < 60mL/min. The definition of valve malpositioning included valve migration, valve embolization, and ectopic valve deployment.21 Device success was defined as the absence of procedural mortality and correct positioning of a single prosthetic heart valve into the proper anatomical location and intended performance of the prosthetic heart valve (mean aortic valve gradient < 20mmHg and no moderate or severe aortic valve regurgitation).21 The present study reports data on clinical outcomes at 30 days, 1 year, over the entire length of follow-up (up to 4 years), and starting at 1 year of follow-up (landmark analysis).

The baseline clinical and procedural characteristics of the study population are presented according to AKI status. Continuous variables are reported as mean ± standard deviation, or median and [interquartile range], and were compared using the Student t test or the Mann-Whitney test, as appropriate. Categorical variables are reported as frequencies and percentages, and were compared using the Pearson chi-square test.

Kaplan-Meier survival curves were created using different stages of AKI as a cutoff and the outcomes were compared using a log-rank test. Once AKI was associated with multiple factors including patient comorbidities, the procedure itself and its complications, a landmark survival analysis was used to explore the clinical impact of AKI on short- and long-term outcomes starting from 30 days and 1 year after the index procedure, respectively. Cox proportional hazards models were used to test the impact of AKI on all-cause death and cardiovascular mortality. Variables were selected if P < .2 in a bivariate analysis, and all variables with P < .05 remained in the model. The variables included in the models were age, sex, New York Heart Association functional class, chronic obstructive pulmonary disease, peripheral vascular disease, aortic balloon valvuloplasty, left ventricular ejection fraction, moderate or severe mitral regurgitation, pulmonary hypertension, access site (femoral vs nonfemoral), transoesophageal echocardiogram, predilatation, AKI, myocardial infarction, stroke, major or life-threatening bleeding, major vascular complication, and valve malpositioning.

A stepwise logistic regression analysis including all variables with P value < .2 in the univariate analysis was used to identify the predictors of AKI and 30-day all-cause and cardiovascular mortality. The variables tested in the models were age, chronic kidney disease, body mass index, diabetes mellitus, peripheral vascular disease, systemic hypertension, porcelain aorta, diuretics use, contrast media volume, creatinine clearance, valve malpositioning, major or life-threatening bleeding, and major vascular complications. All statistical tests were 2-sided, and the criterion for statistical significance was P < .05. All statistical analyses were performed using SPSS statistical software version 20.0.

The baseline characteristics of the study population are illustrated in Table 1. The overall mean age was 81.5 ± 7.3 years, 50.6% were women, 32.0% diabetics, the mean Society of Thoracic Surgeons score was 10.3 ± 7.9, and the mean EuroSCORE was 20.6 ± 14.7. The mean left ventricular ejection fraction was 58.6 ± 15% and the mean aortic gradient was 49 ± 15mmHg. The preprocedural estimated glomerular filtration rate was 48.4 ± 22.1mL/min and 77.1% of the patients had chronic kidney disease.

The incidence of AKI following TAVI was 18% (143/794). Of these, 11.1% (88) were classified as AKI stage 1, 2.4% (19) as stage 2, and 4.5% (36) as stage 3. Demographic characteristics were similar between patients with and without AKI (Table 1). The only exception was the Society of Thoracic Surgeons score, which was higher among patients with AKI (11.5 ± 8.3 vs 10 ± 7.7; P = .04). Length of hospital stay was significantly higher among patients with AKI (19.3 ± 29.1 days vs 11.2 ± 20.9 days; P < .001).

Transcatheter aortic valve implantation was performed via transfemoral access in most patients (93.6%), with CoreValve being the device most commonly used (72.9%). Most patients underwent general anesthesia. Procedural characteristics are shown in Table 2. We observed a higher prevalence of nontransfemoral access and predilatation among patients who developed AKI. The mean total contrast media used was 187 ± 110mL, with a nonsignificant trend toward less contrast use in the group who developed AKI (175 ± 73 vs 190 ± 116; P = .09). However, the amount of contrast media received was higher among patients with valve malpositioning (276 ± 219mL vs 183 ± 100mL; P = .028) in comparison with those without valve malpositioning.

Device success rate was lower and valve malpositioning was more frequently observed in patients who developed AKI following TAVI (Table 2).

On multivariate analysis, the independent predictors of AKI were age, diabetes mellitus, major or life-threatening bleeding, and valve malpositioning (Table 3), the latter being the strongest (odds ratio [OR], 4.90; 95% confidence interval [95%CI], 2.42-9.92; P < .001) predictor.

At 30 days, patients with AKI had higher rates of all-cause death (21% vs 2.9%; OR, 7.84; 95%CI, 3.70-16.58; P < .001] and cardiovascular mortality (18.2% vs 2.3%; OR, 8.55; 95%CI, 3.80-19.38; P < .001) than those without AKI. When an adjusted logistic regression model was used, AKI remained independently associated with 30-day all-cause mortality (OR, 2.71; 95%CI, 1.53-4.79; P = .001) and cardiovascular mortality (OR, 2.43; 95%CI, 1.32-4.46; P = .004).

Among patients with AKI, the 30-day all-cause death was 9.1% (OR, 3.33; 95%CI, 1.41-7.85) for those in stage 1, 36.8%(OR, 19.40; 95%CI, 6.87-54.78) in stage 2, and 41.7% (OR, 23.76; 95%CI, 10.63-53.12) in stage 3 (P < .001); 30-day cardiovascular mortality was 5.7% (OR, 2.55; 95%CI, 0.91-7.21), 36.8% (OR, 24.73; 95%CI, 8.54-71.64) and 38.9% (OR, 26.98; 95%CI, 11.61-62.71), in stages 1, 2 and 3, respectively (P < .001).

At 1-year of follow-up, we observed higher rates of all-cause (hazard ratio [HR], 4.17; 95%CI, 2.60-6.70; P < .001) and cardiovascular mortality (HR, 5.34; 95%CI, 2.98-9.57; P < .001) among patients with AKI compared with those without AKI (Figure 1).

Figure 1.

Kaplan-Meier curves for all-cause (A) and cardiovascular mortality (B) according to acute kidney injury status during the first year of follow-up and the unadjusted landmark analysis with follow-up starting at 30 days of all-cause mortality (A) and cardiovascular mortality (B). 95%CI, 95% confidence interval; HR, hazard ratio.

(0.45MB).

When only survivors at 30 days were considered in the landmark analysis, AKI remained associated with higher risk of all-cause (HR, 2.86; 95%CI, 1.55-5.28; P < .001) and cardiovascular mortality (HR, 4.04; 95%CI, 1.56-10.48; P < .001) (Figure 1).

The presence of AKI was associated with a 3-fold increase in the risk of all-cause mortality (HR, 3.06; 95%CI, 2.06-4.55; P < .001) and a 4-fold increase in the risk of cardiovascular mortality (HR, 4.24; 95%CI, 2.52-7.12; P < .001) at 4 years of follow-up (Figure 2). Importantly, we observed a stepwise increase in all-cause and cardiovascular mortality according to different stages of AKI over the entire length of follow-up (Figure 3). Acute kidney injury remained independently associated with all-cause (adjusted HR, 2.8; 95%CI, 2.0-3.9; P < .001) and cardiovascular mortality (adjusted HR, 2.9; 95%CI, 1.9-4.4; P < .001) mortality, after adjustment for possible confounding variables (Table 4).

Figure 2.

Kaplan-Meier curves for all-cause (A) and cardiovascular mortality (B) according to acute kidney injury status, over the entire length of follow-up and the unadjusted landmark analysis with follow-up starting at 1 year of all-cause mortality (A) and cardiovascular mortality (B). 95%CI, 95% confidence interval; HR, hazard ratio.

(0.37MB).

Figure 3.

Kaplan-Meier curves for all-cause mortality (A) and cardiovascular mortality (B) according to acute kidney injury network criteria (AKIN 1, AKIN 2 and AKIN 3) over the entire length of follow-up. AKIN, Acute Kidney Injury Network.

(0.27MB).

However, when considering only survivors at 12 months (428/794) in the landmark analysis, there was no difference in all-cause death (HR, 1.20; 95%CI, 0.60-2.39; P = .58) or cardiovascular mortality (HR, 1.35; 95%CI, 0.46-3.95; P = .4) among patients with and without AKI (Figure 2). Even after adjustment for important covariates, AKI was not associated with higher all-cause (adjusted HR, 1.2; 95%CI, 0.5-2.6; P = .71) and cardiovascular (adjusted HR, 0.7; 95%CI, 0.2-2.1; P = .57) mortality.

Our study has several limitations. First, data was self-reported by each center and, therefore, the occurrence of AKI might have been underreported, since on-site source document verification was randomly performed in only 20% of cases. Second, some patients may have been discharged from hospital earlier than 7 days; however, in our population, only 1% of the patients were discharged before this period. Third, although we appropriately adjusted for imbalance in baseline characteristics, residual confounding from unmeasured variables cannot be excluded and a cause and effect relationship between AKI and outcomes cannot be determined.