Renal Function as Predictor of Mortality in Patients After Percutaneous Transcatheter Aortic Valve Implantation

Patients

Patients with severe, symptomatic aortic stenosis and no reasonable surgical option due to excessive perioperative risk underwent planned TAVI at our institution. Within this observational, prospective study, 77 consecutive patients were included into the study after written informed consent was obtained. Five patients receiving chronic dialysis (logistic European System for Cardiac Operative Risk Evaluation [EuroSCORE] 23.0 ± 12.4%, Society of Thoracic Surgeons [STS] mortality score 7.9 ± 3.0%, 30-day mortality: 0%, in-hospital mortality: 1 of 5, 1-year mortality: 1 of 5) were excluded from this analysis, because AKI would not be an issue in these patients. In preparation for TAVI, coronary anatomy and hemodynamic status were assessed by coronary angiography and left and right heart catheterization. Valvular anatomy and annulus dimension were evaluated with transthoracic and transesophageal echocardiography (including 3-dimensional reconstruction), contrast angiography of the aortic root, and multislice computer tomography of the thoracic aorta. The vascular access site was assessed by color-coded Doppler sonography and multislice computer tomography or contrast angiography of the abdominal aorta and the aortoiliofemoral system. Acceptance for TAVI required a consensus by a team of cardiac surgeons and cardiologists.

Median follow-up time was 283 days (interquartile range: 49 to 439 days). The clinical end point of the study was all-cause mortality 30 days and 1 year after TAVI. Information about the cause of death was obtained from the treating hospital or general practitioner charts. The study was approved by the local ethics committee of the University of Bonn.

Systemic inflammatory response syndrome (SIRS) was defined according to existing guidelines as fulfilling at least 2 of the following criteria (20): temperature ≤36.0°C or ≥38.0°C, heart rate ≥90 beats/min, respiratory rate ≥20/min or hyperventilation (PaCO₂ ≤33 mm Hg), leukocyte count ≥12 or ≤4 (10⁹/l).

Valve implantation

Procedures were performed with biplane fluoroscopy under local anesthesia in combination with a systemic sedative/analgesic treatment (fentanyl, midazolam). Patients were pre-medicated with 500-mg aspirin, 300 mg clopidogrel, and 2 g ceftriaxone. Patients with pre-procedural chronic kidney disease received intravenous hydration and N-acetylcysteine before and after TAVI and were prophylactically treated with intravenous bicarbonate to prevent contrast-induced nephropathy. Visipaque 320 (GE Healthcare, Munich, Germany) was used as contrast medium in patients with chronic kidney disease. Vascular access was obtained through a percutaneous 18-F sheath placed in the common femoral artery (1 patient received surgical preparation of the common femoral artery and consecutive surgical closure due to femoropopliteal bypass). After pre-dilation of the native aortic valve with a 22-mm or 25-mm NuCLEUS Percutaneous Transluminal Valvuloplasty catheter (NuMED, Inc., Baylis Medical, Montreal, Quebec, Canada) during right ventricular burst pacing, the self-expandable 18-F CoreValve prosthesis (CoreValve Revalving Technology, Medtronic, Minneapolis, Minnesota) was advanced retrogradely and deployed within the aortic annulus (26 mm or 29 mm prosthesis according to the recommendations of the manufacturer). The femoral puncture site was closed percutaneously with the use of the Prostar XL suture device (Abbott Vascular, Abbott Park, Illinois).

Laboratory methods

Serum creatinine levels were measured on the day before TAVI and on days 1, 2, 3, and 7 in all patients (ECREA Flex reagent cartridge, Dimension Vista, Siemens Healthcare Diagnostics GmbH, Munich, Germany). Estimated glomerular filtration rate (eGFR) was calculated by the simplified Modification of Diet in Renal Disease formula (21). Chronic renal failure was defined as an eGFR ≤60 ml/min; AKI as an absolute increase in serum creatinine of ≥0.3 mg/dl (≥26.4 μmol/l); percentage increase in serum creatinine of ≥50%; or reduction in urine output, defined as <0.5 ml/kg/h for more than 6 h, in <48 h after TAVI according to the Acute Kidney Injury Network classification (15). The N-terminal pro-hormone brain natriuretic peptide (NT-proBNP), C-reactive protein (CRP), troponin I, and lactate levels were routinely measured on the day before and the second day after TAVI (PBNP Flex reagent cartridge, CRP Flex reagent cartridge, CTNI Flex reagent cartridge, and LA Flex reagent cartridge; Dimension Vista, Siemens Healthcare Diagnostics GmbH, Munich, Germany).

Statistical analysis

Continuous variables are presented as mean ± SD if normally distributed and as median (interquartile range) if not normally distributed. Categorical variables are given as frequencies and percentages. Continuous variables were tested for differences with the Student t or Mann-Whitney U test, categorical variables with the chi-square test. Pearson's and Spearman's correlation coefficients were used to establish associations. The cumulative survival plot in relation to creatinine according to quartiles 1 to 4 and the incidence of AKI was estimated by the Kaplan-Meier method. Survival in groups was compared with the log-rank test. To identify predictors of death after TAVI, a Cox proportional hazard model was applied. Results are reported as adjusted hazard ratio (HR) with 95% confidence interval (CI). All analyses were conducted with SPSS Statistics version 17.0.0 (SPSS, Inc., Chicago, Illinois).

Baseline characteristics

The TAVI was performed in 77 patients with severe, symptomatic aortic stenosis unsuitable for open heart surgery due to multiple comorbidities. Baseline characteristics are given in Table 1. All patients were at high risk for conventional surgery as reflected by a mean STS score of 9.3 ± 6.1 and a mean logistic EuroSCORE of 31.2 ± 17.6%. Common comorbidities included coronary artery disease (65%), chronic renal failure (62%), peripheral artery disease (46%), pulmonary hypertension (42%), history of previous myocardial infarction (42%) or stroke (26%), and chronic obstructive pulmonary disease (COPD) (26%).

Baseline characteristics and outcome

Procedural success was achieved in 75 of 77 (97%) patients, with 1 periprocedural death (1%) due to cardiogenic shock after balloon valvuloplasty and 1 conversion to open heart surgery (1%) due to valve embolization into the ascending aorta without possibility to implant a second valve. A second valve was implanted in 3 patients (4%) due to valve malposition with severe peri-valvular prosthetic regurgitation. A total of 8 patients (10%) died within the first 30 days after TAVI. In-hospital mortality rate was 14% (11 of 77 patients). The 1-year mortality rate was 26% (20 of 77 patients). The causes of death for these patients were: pneumonia/septicemia (n = 7), cardiorenal syndrome with multi-organ failure (n = 7), congestive heart failure (n = 4), cardiogenic shock (n = 1), fatal myocardial infarction (n = 1), and intracerebral bleeding (n = 1).

Nonsurvivors had a higher logistic EuroSCORE (41.7 ± 16.5 vs. 27.6 ± 16.7; p = 0.002) and a higher STS mortality score (12.5 ± 6.6 vs. 8.2 ± 5.5; p = 0.006) and greater frequency of coronary artery disease (90% vs. 56%; p = 0.006), peripheral arterial disease (70% vs. 37%; p = 0.01), chronic renal failure (85% vs. 54%; p = 0.01), COPD (45% vs. 19%; p = 0.02), and pulmonary hypertension (60% vs. 35%; p = 0.05) (Table 1). On the second post-procedural day (48 h after TAVI), nonsurvivors suffered more frequently from periprocedural AKI (60% vs. 14%; p < 0.001), relevant peri-prosthetic regurgitation (30% vs. 11%; p = 0.03), and SIRS (50% vs. 25%; p = 0.03) (Table 2).

Renal function before and after TAVI

Serum creatinine levels and eGFR at baseline significantly differed between nonsurvivors and survivors (1.50 [1.22 to 1.91] mg/dl vs. 1.21 [1.02 to 1.44] mg/dl, p = 0.01, and 41.2 [31.6 to 55.6] ml/min vs. 53.3 [44.9 to 64.6] ml/min, p = 0.02, respectively). Furthermore, NT-proBNP levels (6,861 [3,790 to 10,814] pg/ml vs. 2,783 [952 to 9,513] pg/ml; p = 0.03) and CRP levels (20.9 [5.3 to 35.4] mg/l vs. 6.2 [2.9 to 19.0] mg/l; p = 0.01) were significantly higher in nonsurvivors (Table 3).

Forty-eight hours after TAVI, a significant decrease of serum creatinine and increase of eGFR in survivors compared with nonsurvivors (1.02 [0.86 to 1.26] mg/dl vs. 1.55 [1.25 to 2.58] mg/dl, p < 0.001; 66.8 [50.7 to 79.3] ml/min vs. 37.2 [25.6 to 59.2] ml/min, p < 0.001) was observed (Fig. 1, Table 4). Estimated GFR increased in 64% of all patients; 19%, 12%, and 4% of patients had eGFR decreases of <25%, 25% to 50%, and >50%, respectively (Table 2).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Renal Function Before and After Transcatheter Aortic Valve Implantation

Time course of serum creatinine (mg/dl) on days 0, 1, 2, 3, and 7 after transcatheter aortic valve implantation according to outcome. In the box-and-whisker plot, the central box represents the interquartile range (IQR); the middle line represents the median. Outside values (•) are smaller/larger than the lower/upper quartile ± 1.5× the IQR; far out values (*) are smaller/larger than the lower/upper quartile ± 3× the IQR.

Correlation of baseline characteristics with renal function

Pre-procedural serum creatinine level and eGFR were associated with the logistic EuroSCORE (r = 0.57, p < 0.001; r = −0.31, p = 0.009, respectively), the STS mortality score (r = 0.55, p < 0.001; r = −0.42, p < 0.001, respectively), and the pre-procedural NT-proBNP level (r = 0.48, p < 0.001; r = −0.34, p = 0.01, respectively). Furthermore, pulmonary hypertension (p = 0.03; p = 0.04) was associated with serum creatinine level and eGFR at baseline, respectively.

On the second post-procedural day, serum creatinine level and eGFR were associated with the leukocyte count (r = 0.26, p = 0.03; r = −0.37, p = 0.002, respectively). No correlation was observed between post-procedural serum creatinine and eGFR, and the amount of contrast medium (r = −0.09, p = 0.44; r = 0.06, p = 0.64, respectively).

Renal function at baseline and outcome

Pre-procedural serum creatinine levels were related to an increased 1-year mortality risk across quartiles: 10%, 16%, 26%, and 53%; p = 0.008 (Fig. 2). A pre-procedural serum creatinine level ≥1.58 mg/dl (quartile 4 vs. quartile 1 to 3) was associated with a 6-fold increased risk of 30-day mortality (HR: 5.9, 95% CI: 1.4 to 24.8; p = 0.02) and a 4-fold increased risk of 1-year mortality (HR: 3.9, 95% CI: 1.6 to 9.5; p = 0.002) (Table 5).

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

1-Year Outcome According to Baseline Serum Creatinine

Survival rate according to quartiles (Qs) of baseline serum creatinine (mg/dl). Q1: <1.08 mg/dl, Q2: 1.08 to 1.24 mg/dl, Q3: 1.25 to 1.57 mg/dl, Q4: ≥1.58 mg/dl.

AKI after TAVI

Acute kidney injury according to the Acute Kidney Injury Network classification (15) occurred in 20 patients (26%) after TAVI: 8 of 20 patients (40%) needed hemofiltration, and 12 of 20 patients (60%) with periprocedural AKI died during follow-up. The incidence of AKI was related to peripheral arterial disease (65% vs. 39%; p = 0.04); the occurrence of SIRS (60% vs. 21%, p = 0.002); and post-procedural, peri-prosthetic regurgitation ≥2+ (35% vs. 9%; p = 0.02) but independent of age, EuroSCORE, STS mortality score, left ventricular ejection fraction, prevalence of COPD, aortic valve area, hematocrit, platelet count, red blood cell transfusion, and the amount of contrast medium (Table 1). Acute kidney injury was not related to serum creatinine levels at baseline (p = 0.52).

Patients with AKI showed—besides significantly increased serum creatinine levels (1.70 [1.35 to 2.64] mg/dl vs. 1.00 [0.84 to 1.24] mg/dl; p < 0.001) and decreased eGFR (38.0 [23.8 to 48.2] ml/min vs. 66.8 [53.3 to 79.1] ml/min; p < 0.001)—higher NT-proBNP (3,959 [1,575 to 7,421] pg/ml vs. 1,128 [737 to 5,528] pg/ml; p = 0.03), lactate (1.1 [0.8 to 2.2] mmol/l vs. 0.8 [0.6 to 1.0] mmol/l; p < 0.001), and troponin I levels (1.20 [0.51 to 5.93] μg/l vs. 0.53 [0.25 to 1.76] μg/l; p = 0.05) on day 2 after TAVI (Table 4). In patients with AKI, CRP levels (90.5 [77.2 to 112.5] mg/l vs. 58.2 [20.4 to 93.0] mg/l; p = 0.005) and leukocyte count (13.8 ± 5.5 · 10⁹/l vs. 10.4 ± 3.8 · 10⁹/l; p = 0.004) were significantly higher 48 h after TAVI.

AKI and outcome

The occurrence of AKI after TAVI was strongly associated with 30-day (p = 0.02), 6-month (p < 0.001), and 1-year mortality (p < 0.001) (Fig. 3). The relative decrease of eGFR was associated with outcome after TAVI: after 1 year, the mortality rate in patients with an increase of eGFR was 12% versus 47%, 44%, and 67%, respectively, in patients with a decrease of eGFR of <25%, 25% to 50%, or >50%, respectively (p = 0.006) (Table 2).

• Download figure
• Open in new tab
• Download powerpoint

Figure 3

1-Year Outcome According to the Occurrence of Periprocedural Acute Kidney Injury

Survival rate according to the occurrence of periprocedural acute kidney injury (absolute increase in serum creatinine of ≥0.3 mg/dl [≥26.4 μmol/l]; percentage increase in serum creatinine of ≥50%; or reduction in urine output, defined as <0.5 ml/kg/h for more than 6 h, in <48 h after TAVI, according to the Acute Kidney Injury Network classification [15]).

Patients suffering from AKI had a 5-fold increased risk for 30-day mortality (HR: 4.9, 95% CI: 1.2 to 20.4, p = 0.03) and a 6-fold increased risk for 1-year mortality (HR: 5.9, 95% CI: 2.4 to 14.5, p < 0.001) (Table 5).

Renal function and mortality after TAVI

Renal function and the development of AKI are important factors for the outcome of patients after invasive procedures. Previously published data underline the importance of AKI after TAVI and describe an incidence of 12% to 28%, which is comparable to our study (17–19). Webb et al. (9) recently showed that chronic kidney disease is associated with a significant higher risk for cumulative late mortality. Bagur et al. (19) found that AKI after TAVI is associated with a more-than-4-fold increased post-procedural mortality risk, which is consistent with our data. Our study extends these findings, because not only does the mortality risk increase stepwise across quartiles of baseline serum creatinine but also the occurrence of AKI is related to an increased mortality risk in short- and mid-term follow-up—independent of whether renal function returns to baseline or not.

Mechanisms of AKI

An AKI after conventional cardiac surgery or percutaneous coronary intervention is related to prolonged hospital stay and mortality (12–15). An increase of serum creatinine is not simply a marker of illness severity but rather represents the onset of AKI, which acts as a causative factor for cardiovascular injury with concomitant activation of neurohormonal, immunological, and inflammatory pathways (24,25). Episodes of pre- and periprocedural hypotension (i.e., rapid burst pacing during balloon valvuloplasty, deployment of the valve prosthesis) might lead to sublethal endothelial injury, which could impair the production of vasodilatory substances such as endothelial nitric oxide and promote vasoconstriction resulting in tubular ischemia and injury (11,26).

Because the amount of contrast media was not associated with the occurrence of periprocedural AKI in our study, the nephrotoxic mechanisms of contrast-induced nephropathy should not be the major issue. Nonetheless, the amount of contrast media needed during the procedure is higher for the CoreValve prosthesis than for the Edwards-Sapien valve (17,19).

Because the incidence of AKI in our study was associated with the prevalence of peripheral arterial disease, advanced generalized atherosclerosis might play an additional role, as previously hypothesized by Aregger et al. (17). Bagur et al. (19) described red blood cell transfusion as an additional risk factor for AKI after TAVI. However, this was not confirmed in our study.

Another factor that might influence renal function after TAVI could be moderate-to-severe peri-prosthetic regurgitation with a prevalence of 7% to 19%, as published in recent data (6,7,22,27). A recent analysis of the German TAVI registry described an increased 30-day mortality risk in patients with moderate-to-severe peri-prosthetic regurgitation (22). In our study cohort, moderate peri-prosthetic regurgitation was found in 15% of the patients and was associated with the occurrence of AKI as well as increased mortality: in multivariate analysis, patients had a 5-fold-higher risk for 1-year mortality and/or AKI. Acute volume overload with increased left ventricular end-diastolic pressure due to relevant peri-prosthetic regurgitation might be responsible for hemodynamic changes with impairment of diastolic renal blood flow and, thus, deterioration of renal function with consecutive kidney injury—especially in patients with low left ventricular ejection fraction.

The development of SIRS might contribute to the occurrence of AKI. Aregger et al. (17) observed a pathological leukocyte count and fever without focus in patients who developed AKI after TAVI. In our study, 60% of the patients with AKI fulfilled the criteria of SIRS and showed significantly higher leukocyte counts and CRP levels 48 h after TAVI. Animal models of renal ischemia-reperfusion injury have demonstrated the pathologic role of interstitial inflammation and the elaboration of cytokines and reactive oxygen species in the production of acute tubular necrosis (28,29).

Furthermore, particulate emboli generated during valvuloplasty, catheter passage in the aorta, and deployment of the valve prosthesis not only might play a role for cerebral outcome after TAVI (30,31) but also might be in part responsible for the postoperative decrease in eGFR (32).

Thus, hemodynamic, inflammatory, atheroembolic, and nephrotoxic factors are involved and overlap each other in leading to an abrupt worsening of renal function, which might lead to the so-called cardiorenal syndrome with acute cardiac dysfunction (25,33,34). In our study, patients with AKI had higher NT-proBNP levels on day 2 compared with patients without AKI. The 30-day mortality was 44% (4 of 9) in patients with a combination of AKI and an NT-proBNP increase, compared with 6% (4 of 68) in all other TAVI patients (HR: 8.3, 95% CI: 2.1 to 33.6; p = 0.003). In these patients, the coincidence of acute renal and cardiac failure might have led to the development of a cardiorenal syndrome with very poor prognosis. Data from cardiac surgery studies point to a concerted injury with the involvement of multiple factors, which results in acute tubular necrosis (11,29). Excessive activation of the renin-angiotensin-aldosterone axis, renal dysfunction with sodium and water retention, and acute myocardial ischemia from an increase in myocardial oxygen demand related to sympathetic activation and peripheral vasoconstriction are the suggested pathways after cardiac surgery and presumably after TAVI, leading to a vicious circle with severe kidney damage followed by a systemic injury leading to the death of the patient (25).

Because the occurrence of periprocedural AKI is multifactorial and related to several (post-procedural) clinical characteristics (moderate-to-severe peri-prosthetic regurgitation, SIRS, and the like), every effort has to be made in TAVI patients to avoid or treat these complications. Especially in patients with multiple comorbidities and a high-frailty index (7,35), preprocedural renal function should be a major issue in the selection process for TAVI and could be a reason for exclusion, because it plays a pivotal role for outcome. The prevention and/or treatment of AKI should lead to a better long-term outcome after TAVI; but several issues, like the coincidence with relevant peri-prosthetic regurgitation and SIRS, are not yet fully understood. Whether a moderate-to-severe peri-prosthetic regurgitation predisposes for the development of AKI and acute cardiorenal syndrome or whether the development of SIRS contributes to adverse outcome of TAVI patients remains controversial and has to be elucidated in further studies about the pathophysiology of AKI after TAVI.

A larger controlled multicenter trial is needed to further validate our data, because it would have major implications on the selection process as well as preventive treatment of TAVI patients. For the time being, strategies to prevent AKI in TAVI patients remain an important challenge. Preventive strategies in TAVI patients at risk for AKI include pre- and periprocedural hydration, the potential use of N-acetylcysteine, and intravenous bicarbonate application. However, no specific pharmacological intervention has demonstrated conclusive efficacy in the prevention of AKI (11). To prevent nephrotoxic mechanisms, iso-osmolar contrast agents should be used, and their amount has to be kept low. After development of the valve prosthesis, additional balloon valvuloplasty or valve-in-valve implantation should be considered in patients with moderate-to-severe peri-prosthetic regurgitation to downgrade a relevant regurgitation and avoid negative hemodynamic effects on the patient, especially in case of markedly impaired left ventricular function. Data about prophylactic hemofiltration in patients with AKI after cardiac surgery are empirical (36), but hemofiltration might improve outcome of patients developing an acute cardiorenal syndrome after TAVI.

Study limitations

Sample size and the monocentric character are limitations of our prospective observational study. Furthermore, only univariate regression analysis has been presented, because the event count was not sufficient to support multivariate analysis with inclusion of all 10 mortality predictors in our study, and thus a multivariate model would have been overfitted. However, AKI was an independent predictor for 30-day and 1-year mortality when multivariate analysis was performed. With the increasing use of TAVI in high-risk patients, identification of clinical criteria is important to optimize patient selection and post-procedural management. Hypothesis-generating studies are of pivotal interest to design larger randomized studies. The type of AKI (ischemic, nephrotoxic, or atheroembolic) could not be determined in our study, because renal biopsies were not performed. Because of the nature of the patients and the performed procedures, AKI was assumed to be multifactorial. Thus, our data cannot establish whether the development of AKI is a marker of multisystem failure in critically ill patients or directly contributes to mortality or both.