Acute kidney injury after transcatheter aortic valve implantation (TAVI) has been associated with adverse outcomes; however, data are limited on the subacute changes in renal function that occur after discharge and their impact on clinical outcomes. This study investigates the relation between subacute changes in kidney function at 30 days after TAVI and survival. Patients from 2 centers who underwent TAVI and survived beyond 30 days with baseline, in-hospital, and 30-day measures of renal function were retrospectively analyzed. Patients were stratified based on change in estimated glomerular filtration rate (eGFR) from baseline to 30 days as follows: improved (≥15% higher than baseline), worsened (≤15% lower), or unchanged (values in between). Univariable and multivariable models were constructed to identify predictors of subacute changes in renal function and of 2-year mortality. Of the 492 patients who met inclusion criteria, eGFR worsened in 102 (22%), improved in 110 (22%), and was unchanged in 280 (56%). AKI occurred in 90 patients (18%) and in only 27% of patients with worsened eGFR at 30 days. After statistical adjustment, worsened eGFR at 30 days (hazard ratio vs unchanged eGFR 2.09, 95% CI 1.37 to 3.19, p <0.001) was associated with worse survival, whereas improvement in renal function was not associated with survival (hazard ratio vs unchanged eGFR 1.30, 95% CI 0.79 to 2.11, p = 0.30). Worsened renal function at 30 days after TAVI is associated with increased mortality after TAVI. In conclusion, monitoring renal function after discharge may identify patients at high risk of adverse outcomes.
In symptomatic patients with severe aortic stenosis at high risk, transcatheter aortic valve implantation (TAVI) is the preferred method for intervention and is an alternative to surgical aortic valve replacement for those at intermediate and low risk. Although TAVI relieves symptoms and extends life in the majority of patients, acute kidney injury (AKI) occurs in 8% to 42% of cases and has been associated with markedly diminished survival. Conversely, recent data have demonstrated that improvements in renal function are associated with improved in-hospital outcomes, possibly due to the relief of cardiorenal syndrome caused by the hemodynamic burden of aortic stenosis. Although these observations suggest a strong link between changes in renal function and subsequent clinical outcomes, limited data exist on the frequency with which renal function worsens, improves, or remains unchanged after discharge, and the clinical relevance of such changes. We therefore sought to characterize subacute changes in kidney function after TAVI and to determine the impact of these changes on subsequent survival.
Methods
The study cohort included patients that underwent TAVI at Barnes Jewish Hospital and Massachusetts General Hospital (MGH) from 2008 to 2015. The written consent requirement was exempt from the MGH Institutional Review Board. Candidacy and appropriateness of TAVI was determined at both institutions by multidisciplinary teams consisting of noninvasive cardiologists, interventional cardiologists, and cardiac surgeons. Baseline clinical and laboratory data were collected prospectively. Serum blood urea nitrogen and creatinine were collected at baseline, daily until discharge, and at 30 days and 1 year after TAVI. Patients were included if they underwent TAVI at MGH or Barnes Jewish Hospital and had preprocedural creatinine recorded. Patients who died within 30 days, did not have 30-day creatinine recorded, or were on renal replacement therapy before TAVI were excluded from the analysis. Estimated glomerular filtration rate (eGFR) was calculated using the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) equation and percent-change of eGFR at 30 days was used to derive groups. Patients were categorized based on change in eGFR from baseline to 30-day follow-up as follows: improved (≥15% higher than baseline), worsened (≤15% lower), or unchanged (values in between). AKI was identified using the Valve Academic Research Consortium-2 criteria, defined as an increase in serum creatinine to 150% to 199% compared with baseline or an absolute increase of 0.3 mg/100 ml within 7 days.
Continuous variables were summarized as means with standard deviations and compared using one-way analysis of variance or Kruskal-Wallis testing. Categoric variables were expressed as percent per group and compared using the chi-square test or Fisher’s exact tests. Kaplan-Meier estimates were constructed on the basis of all available follow-up data, and survival was compared using the log-rank test. Cox proportional hazard regression analysis determined predictors of all-cause mortality at 24 months after TAVI. Linear regression was used to identify predictors of 30-day eGFR, controlling for baseline eGFR. Multivariable models were constructed using stepwise selection, with variables being selected if they were of clinical interest or if they satisfied the entry criterion of p ≤0.05 in univariable analysis. Variables included age, gender, baseline eGFR, coronary artery disease, peripheral arterial disease, diabetes mellitus, left ventricular ejection fraction (LVEF), aortic valve mean gradient, access site, baseline hemoglobin level, periprocedural blood transfusion, contrast dose, lung disease, New York Heart Association (NYHA) class, AKI, and Society of Thoracic Surgeons (STS) score. Statistical significance was defined by a 2-sided p <0.05, and statistical analyses were performed using SAS version 9.4 (SAS Institute Inc, Cary Ca C, North Carolina) and R version 3.4.3 (https://www.R-project.org/).
Results
The study cohort included 492 subjects with a mean age of 83 ± 19 years. Mean eGFR was 56 ± 19 ml/min/1.73 m 2 , and 289 patients (59%) had CKD at baseline with eGFR <60 ml/min/1.73 m 2 .
At 30 days after TAVI, subacute worsening of renal function occurred in 22%, improvement occurred in 22%, and no change occurred in 56% of patients. There were no significant differences in mean age, weight, median STS score, mean aortic gradient, aortic valve area, procedural iodinated contrast usage, or LVEF ( Table 1 ). Baseline eGFR was lower (p <0.001) and New York Heart Association (NYHA) class IV functional status more frequent ( p = 0.04) in patients with renal function improvement at 30 days than in patients with unchanged and worsened renal function.
| Subacute Renal Function | |||||
| Variables | Overall Cohort (n = 492) | Worsened (n = 108) | Unchanged (n = 277) | Improved (n = 107) | P value |
| Age (years) | 83±19 | 87±8 | 84±8 | 84±10 | 0.09 |
| Women | 232 (47%) | 47 (44%) | 133 (48%) | 52 (49%) | 0.74 |
| White | 486 (99%) | 107 (99%) | 275 (99%) | 104 (97%) | 0.32 |
| Creatinine (mg/dl) | 1.23±0.52 | 1.17±0.50 | 1.12±0.41 | 1.35±0.43 | <0.0001 |
| Estimated glomerular filtration rate (ml/min/1.73 m 2 ) | 56±19 | 56±20 | 59±20 | 47±16 | <0.0001 |
| Weight (kg) | 76±20 | 74±20 | 77±20 | 77±21 | 0.19 |
| Hypertension | 451 (92%) | 100 (93%) | 253 (91%) | 98 (92%) | 0.9 |
| Diabetes Mellitus | 167 (33%) | 30 (28%) | 94 (34%) | 43 (40%) | 0.21 |
| Lung disease | 68 (14%) | 14 (13%) | 39 (14%) | 15 (14%) | 0.78 |
| Society of Thoracic Surgeons Predicted Risk of Mortality score | 8 (5,11) | 8 (6,11) | 7 (5,10) | 8 (5,11) | 0.32 |
| New York Heart Association class | |||||
| II | 56 (11%) | 7 (7%) | 38 (14%) | 11 (10%) | 0.04 |
| III | 343 (70%) | 83 (77%) | 193 (70%) | 67 (62%) | |
| IV | 84 (17%) | 16 (15%) | 41 (15%) | 27 (25%) | |
| Peripheral vascular disease | 261 (53%) | 59 (55%) | 140 (50%) | 62 (58%) | 0.3 |
| Coronary artery disease | 358 (73%) | 78 (72%) | 199 (71%) | 81 (77%) | 0.69 |
| Coronary artery bypass graft | 166 (34%) | 33 (31%) | 94 (34%) | 39 (36%) | 0.71 |
| Percutaneous coronary intervention | 160 (33%) | 35 (32%) | 86 (31%) | 39 (37%) | 0.6 |
| Hemoglobin (g/dL) | 12±2 | 12±2 | 12±2 | 12±2 | 0.42 |
| Blood transfusion | 206 (42%) | 49 (45%) | 106 (38%) | 51 (48%) | 0.21 |
| Femoral Access | 238 (48%) | 57 (53%) | 148 (53%) | 49 (46%) | 0.44 |
| Mean aortic valve gradient (mmHg) | 46±15 | 47±15 | 45±15 | 46±16 | 0.71 |
| Left ventricular ejection fraction (%) | 58±14 | 60±14 | 58±14 | 56±14 | 0.08 |
| Volume of contrast used (ml) | 84±43 | 89±47 | 83±39 | 80±51 | 0.09 |
In patients whose eGFR ultimately worsened, those who experienced AKI (n = 30) underwent an acute drop in renal function which plateaued over time (baseline: 48 ± 19.62 ml/min/1.73 m 2 ; 48 hours: 34.65 ± 16.97 ml/min/1.73 m 2 ; 30 days: 32.12 ± 14.88 ml/min/1.73 m 2 ), whereas subjects who did not experience AKI (n = 72) had a progressive worsening of eGFR after TAVI (baseline: 61.33 ± 16.79 ml/min/1.73 m 2 ; 48 hours: 61.2 ± 20.09 ml/min/1.73 m 2 ; 30 day: 46.27 ± 13.94 ml/min/1.73 m 2 ) ( Figure 1 ).
Within the subgroup of patients with CKD (eGFR<60 ml/min/1.73 m 2 ) at baseline, improved, worsened, and unchanged renal function at 30 days was seen in 29%, 22%, and 49% of patients, respectively. In the worsened group, mean eGFR was 45 ± 11 ml/min/1.73 m 2 at baseline (mean creatinine 1.37 ± 0.43 mg/100 ml), 42 ± 19 ml/min/1.73 m 2 at 48 hours (mean creatinine 1.60 ± 0.65 mg/100 ml), and 32 ± 10 ml/min/1.73 m 2 at 30 days (mean creatinine 1.83 ± 0.61 mg/100 ml). In the improved group, mean eGFR was 40 ± 10 ml/min/1.73 m 2 at baseline (mean creatinine 1.47 ± 0.39 mg/100 ml), 46 ± 19 ml/min/1.73 m 2 at 48 hours (mean creatinine 1.37 ± 0.39 mg/100 ml), and 54 ± 14 ml/min/1.73 m 2 at 30-days (mean creatinine 1.17 ± 0.40 mg/100 ml) ( Figure 1 ).
AKI occurred in 90 patients (18%) within the entire cohort. Baseline eGFR was 57±19 ml/min/1.73 m 2 in patients who did not experience AKI and 47 ± 20 ml/min/1.73 m 2 in patients who experienced AKI (p <0.001). AKI occurred in 27%, 11%, and 17% of patients within the worsened, improved, and unchanged renal function groups, respectively (p <0.001) ( Figure 2 ). Proportions of patients experiencing AKI were similar when restricted only to patients with CKD (eGFR <60 ml/min/1.73 m 2 ) at baseline ( Figure 2 ). Notably, 73% of patients with subacute worsening of renal function did not experience AKI.
