Nephrology 17 (2012) 445–451
Original Article
Incidence, risk factors and prognosis of acute kidney injury after transcatheter aortic valve implantation nep_1593
1,3
WAI Y KONG,
2
445..451
1
GERALD YONG and ASHLEY IRISH
1 Department of Nephrology and Transplantation, Royal Perth Hospital, 2Interventional Cardiology & High Risk Aortic Stenosis State Service WA, Royal Perth Hospital, Perth, Western Australia, Australia; and 3Division of Nephrology, Department of Medicine, University Malaya Medical Centre, Kuala Lumpur, Malaysia
KEY WORDS: acute kidney injury, aortic stenosis, cardiorenal syndrome, health cost, transcatheter aortic valve implantation. Correspondence: Dr Wai Y Kong, Department of Nephrology and Transplantation, Level 6 South Block, Royal Perth Hospital, Perth, WA 6000, Australia. Email:
[email protected] Accepted for publication 27 February 2012. Accepted manuscript online 5 March 2012. doi:10.1111/j.1440-1797.2012.01593.x
SUMMARY AT A GLANCE The paper described the incidence and predictors of acute kidney injury after transcatheter aortic valve implantation. Acute kidney injury has previously been reported to be associated with mortality in several manuscripts.
ABSTRACT: Aim: Transcatheter aortic valve implantation (TAVI) poses a significant risk of acute kidney injury (AKI). Little is known of the impact of TAVI and AKI on long-term kidney function and health cost. We explored the predictive factors and prognostic implications of AKI following TAVI. Methods: Single-centre retrospective analysis of 52 elderly patients undergoing TAVI was conducted. The primary endpoint was renal outcome which included the incidence of AKI and 12-month renal function after TAVI. Secondary endpoints were mortality, the length of hospital stay (LOS) and cost. Results: AKI occurred in 15/52 (28.8%) patients (mean age 84 1 6) and three patients (6%) required dialysis. Patients with AKI (AKI+) had greater comorbidity (diabetes and cerebrovascular disease) and a trend towards reduced estimated glomerular filtration rate (eGFR) at baseline compared with those without AKI (56.6 vs AKI-: 65.7 mL/min per 1.73 m2, P = 0.07). Following TAVI, AKI- patients experienced an immediate improvement in eGFR, which remained significantly higher at all time points compared with AKI+ patients (70.4 vs 46.9 at 6 months and 73.7 vs 53.0 at 12 months, P < 0.001). Cumulative mortality for AKI+ versus AKI- group was 26.7% and 2.7% (P = 0.006). LOS doubled (P < 0.001) and average hospitalization cost per patient was 1.5 times higher in the AKI+ group (P < 0.001). Independent predictors of AKI were peri-procedural blood transfusion (OR: 2.4, 95% CI: 2.0–3.1), trans-apical approach (OR: 9.3, 95% CI: 4.3–23.7) and hypertension (OR: 6.4, 95% CI: 2.9–17.3). Conclusion: AKI developed in 28.8% of patients after TAVI and was associated with procedural technique and transfusion requirement, and an increased LOS and mortality. However, most patients achieved a significant and sustained improvement in eGFR.
Aortic stenosis (AS) is one of the commonest native cardiac valve disorders of the elderly, where surgical intervention is the most effective therapeutic option.1,2 Aortic valve replacement is not performed in two-thirds of patients with symptomatic disease because of an unacceptable high operative mortality due to multiple comorbidities. Conservatively managed symptomatic disease especially in elderly patients greater than 80 years old is associated with a poor prognosis (38% 1-year and 68% 5-year mortality).2 Transcatheter aortic valve implantation (TAVI), first reported in 2002, is a promising treatment alternative to surgery with © 2012 The Authors Nephrology © 2012 Asian Pacific Society of Nephrology
comparable 1-year patient survival to surgical valve replacement and is associated with more favourable outcome compared with medical treatment.3,4 The risk of acute kidney injury (AKI) after TAVI, however, is not well described and its effects upon short- and longterm renal function are uncertain. The occurrence of AKI not merely is an independent predictor of adverse outcome but also predisposes to the development of chronic kidney disease (CKD). AKI requiring dialysis has a poor prognosis (50% in-hospital mortality) and a significant proportion of patients will progress to end-stage kidney disease needing 445
WY Kong et al.
long-term renal replacement therapy.5–7 While improvement in renal function was observed within 1 week post-TAVI, whether this benefit is sustained remains unknown.8–11 The objectives of this study were to determine (i) the incidence, prognostic implications and predictors of AKI after TAVI, (ii) the impact of TAVI on mid- and long-term renal function stratified according to the occurrence of AKI and (iii) the impact of AKI on patient survival, health-care cost and resource usage.
MATERIALS AND METHODS Patient population Transcatheter aortic valve implantation was initiated for high-risk AS patients as a statewide service based in Royal Perth Hospital, Western Australia in 2009. Patients with symptomatic severe AS were offered TAVI if they were considered to have high operative risk (in general, defined as age ⱖ80 years old and/or logistic EuroSCORE ⱖ20% or the presence of other high-risk factors which are not included in the EuroSCORE).12 An interdisciplinary heart valve team with interventional cardiologists and cardiothoracic surgeons was involved in the selection of all patients. In addition to consenting for the therapeutic procedure, every patient provided informed consent for clinical and procedural data collection.
TAVI procedure and post-procedural monitoring Patients were admitted on the day before the procedure. All patients were given pre-procedural hydration with 0.5 L to 1 L of 0.9% normal saline (depending on cardiac status) and N-acetylcysteine 600 mg, two doses before the procedure and continued for another 2 days thereafter. TAVI was performed in the catheterization laboratory under fluoroscopy guidance with the use of non-ionic low-osmolar radio-iodine contrast media (iohexol) via trans-femoral (TF, n = 41) or trans-apical (TA, n = 11) approach depending on patients’ anatomical and vascular characteristics. Either a self-expandable (Medtronic CoreValve, Medtronic Inc., Minneapolis, MN, USA) or balloon-expandable (Edwards SAPIEN valve, Edwards Lifesciences, Irvine, CA, USA) valve prosthesis was used. Percutaneous vascular closure device (either Perclose ProGlide suture-mediated closure device, Abbott Laboratories, Abbott Park, IL, USA, n = 15 or Prostar XL percutaneous vascular surgical system Abbott Laboratories, n = 25) were used for all TF approach except one patient who needed vascular cut-down. There was no percutaneous closure device failure leading to bleeding complication. Daily full blood count and renal function were obtained till patient discharge or death. Thereafter, renal function was repeated at 1-month, 6-month and 12-month time points. 446
Data collection and study endpoints Medical records and databases were reviewed and the following information was gathered: baseline and periprocedural characteristics, comorbidities, laboratory parameters (full blood count, renal function and proteinuria) before and after TAVI. Estimated glomerular filtration rate (eGFR) was derived using simplified MDRD equation. The prevalence of proteinuria was determined with the use of urine dipstick (urine spot protein ⱖ30 mg/dL is equivalent to ⱖ500 mg of protein per day). The primary endpoint was renal outcome which included the incidence of AKI and 12-month renal function after TAVI. Secondary endpoints were the length of hospital stay (LOS), the incidence of combined life-threatening/major bleeding, health-care cost and resource usage, in-hospital and cumulative mortality. AKI was defined according to RIFLE classification based on changes in serum creatinine or eGFR.13 The time frame for the changes from baseline was set at 48 h to ensure the occurrence of AKI is related to the index procedure. LOS was the period between the day of the procedure and the day of discharge or in-hospital mortality. Overt bleeding either associated with a haemoglobin drop of ⱖ30 g/L or requiring red blood cell (RBC) transfusion of ⱖ2 units was considered as major bleeding.14
Statistical analysis Continuous data were expressed as mean 1 standard deviation (95% confidence interval, CI) or median with interquartile range (IQR). Categorical data were presented as the number of patients and percentage. Baseline, periprocedural and post-procedural characteristics and outcome measures between patients who experienced AKI (AKI+) and those without AKI (AKI-) were compared with the use of c2 test, Mann–Whitney test and t-test. The time course of renal function after TAVI between groups was analysed with longitudinal mixed-effects multivariate logistic regression model based on maximum likelihood estimation. The mortality analysis was performed using Kaplan–Meier method and the Log-rank test was used to assess the survival difference. Significance was evaluated using a two-sided P-value
