Incidence and Predictors of Acute Kidney Injury Following Transcatheter Aortic Valve Replacement: Role of Changing Definitions of Renal Function and Injury

Abstract: Background. Acute kidney injury (AKI) following transcatheter aortic valve replacement (TAVR) is a known complication. The prospective validation of various AKI definitions and estimated baseline renal function equations in the context of TAVR remains an ongoing area of research. This study examined the Valve Academic Research Consortium (VARC) 1 and 2 criteria for AKI, and impact of three estimated glomerular filtration rate (eGFR) equations (CKD-EPI, MDRD, and Cockcroft-Gault) on AKI incidence in TAVR patients. Methods. Retrospective review of 120 consecutive TAVR procedures over a 4-year period was performed. AKI, including stage, was defined using the VARC 1 and VARC 2 criteria. Univariate and multivariate analyses were performed for association between AKI and known patient, hemodynamic, and procedural variables. Further logistic regression, stepwise logistic regression, and association plots were performed for the three different eGFR calculations. Results. AKI occurred in 22% of VARC 1 patients and 23% of VARC 2 patients. On multivariate analysis, baseline eGFR was predictive of stage 1 AKI by CKD-EPI classification (VARC 1: odds ratio [OR], 0.93; 95% confidence interval [CI], 0.88-0.99; P=.02; VARC 2: OR, 0.93; 95% CI, 0.87-0.99; P=.03) and MDRD (OR, 0.93; 95% CI, 0.88-0.99; P=.03). Non-transfemoral approach was predictive of stage 1 AKI by VARC 2 (OR, 33.33; 95% CI, 1.6-696.41; P=.02). Conclusions. The risk factor associations for AKI post TAVR vary by definitions used. Decreased GFR at baseline by both MDRD and CKD-EPI and non-transfemoral approach were associated with an increased risk of AKI post TAVR.

J INVASIVE CARDIOL 2020;32(4):138-141. Epub 2020 January 15.

Key words: acute kidney injury, TAVR

Acute kidney injury (AKI) is a common and morbid complication of transcatheter aortic valve replacement (TAVR). Specifically, AKI is associated with prolongation of hospital stays and increased mortality.^1-3 Challenges in understanding AKI incidence and risk are complicated by the variable definitions of kidney injury and heterogeneous methods of renal function reported in the literature. The Valve Academic Research Consortium (VARC) created a definition of AKI based on the modified Risk, Injury, Failure, Loss, End-Stage Renal Disease (RIFLE) criteria.⁴ This was later replaced by the current recommendations set by VARC 2, which adopted the newer Acute Kidney Injury Network (AKIN) criteria.⁵ The published literature describes a strong association between baseline renal function and the risk of AKI.^3,6,7 However, the impact of different equations used to determine estimated glomerular filtration rate (eGFR) on AKI incidence and risk association remains uncertain. In the current study, we examined the incidence of AKI and clinical risk factor associations using both the VARC 1 and 2 definitions, as well as the impact that various eGFR equations used to assess baseline renal function have on AKI incidence.

Methods

Patient selection and procedure. We performed a single-center, retrospective chart review of TAVR procedures performed at our institution (University Health System, San Antonio, Texas) over a 4-year period from May 2012 to December 2016. Prior to our data collection, we obtained Institutional Review Board approval from the University of Texas Health San Antonio. We excluded patients who were receiving renal replacement therapy. Patients were selected for TAVR by consensus from our valve heart team, which includes cardiothoracic surgeons and interventional cardiologists. Contrasted computed tomography (CT) was done for assessment of the aorta dimensions as well as iliac and femoral vessel anatomy (median 4 weeks prior to TAVR procedure). Based on vessel anatomy on CT, either the transfemoral, transapical, or transaortic approach were chosen for TAVR. Any coronary lesions thought to be significant on preprocedure left heart catheterization were evaluated and intervened on prior to TAVR (median 4 weeks prior to TAVR procedure). Sapien XT (Edwards Lifesciences) or CoreValve (Medtronic) devices were used for all procedures, as determined by operator preference.

Data. We recorded clinical data, including hydration volumes, procedural hemodynamics, and comorbidities, for analysis of associations with AKI incidence. We reviewed each patient’s serum creatinine at baseline obtained the morning of the procedure and followed it throughout the hospital stay. These creatinine values were the basis of the determination of AKI, including stage, using the definitions provided in the consensus documents for VARC 1^4,7 and VARC 2.⁵ The major differences between the two criteria are that VARC 2 prolongs the timing in which an AKI can be diagnosed post procedure from 3 days to 7 days, and also adds urine output criteria for diagnosis. We did not use urine output data from the medical record to diagnose AKI because of uncertainty regarding accuracy of these measurements in patients who were not catheterized for the entirety of their hospital stay. Clinical variables included total contrast and periprocedural hydration volumes, number of pacing runs, procedural changes in systolic blood pressures, right- and left-sided pressures in the heart, estimated blood loss, femoral versus non-femoral approach, loop diuretic or ACE inhibitor use, age, sex, coronary artery disease, and anemia. We assessed baseline renal function using two eGFR equations (the chronic kidney disease-epidemiology collaboration equation [CKD-EPI] and the modification of diet in renal disease study formula [MDRD]), and Cockcroft-Gault (CG) to estimate creatinine clearance (eCrCL).

Statistical analysis. Continuous variables are reported as mean ± standard deviation for normally distributed variables and as median (interquartile range) for non-normally distributed variables. Categorical variables are reported as percentages. Univariate analysis was used to calculate odds ratio (OR) for the clinical variables discussed above. Variables with P-values ≤.05 were then used for multivariate analysis to determine prognostic variables for AKI after TAVR. Results were reported as OR with a 95% confidence interval (95% CI). A P-value <.05 (2 tailed) was considered significant. The estimated Pearson correlation coefficients of CrCl, MDRD, and CKD-EPI were reported as an R-value. R statistical software (version 3.3.3; March 6, 2017) was used for this analysis.

Results

We evaluated 129 patients and included 120 after exclusion of those on dialysis. Mean age was 79.7 ± 8.9 years and 49% were women. Chronic kidney disease (CKD) was present in 42% of patients when defined using CKD-EPI. Sapien valves were used in 83% of cases. The most common approach used was transfemoral, followed by transaortic, then transapical (78%, 13%, and 8%, respectively).  There were no intraprocedural deaths, and 3 deaths occurred during the index hospitalization following the TAVR procedure. Rates of major bleeding complication (Bleeding Academic Research Consortium 3 or higher), major vascular complication, stroke, need for conversion to open surgery, and pacemaker requirement were 16.7%, 5.0%, 4.2%, 1.7%, and 11.7%, respectively. Further demographic information is shown in Table 1. AKI incidence was similar when using VARC 1 criteria vs VARC 2 criteria, with an overall incidence of 26 (22%) and 27 (23%), respectively (Figure 1).

The multivariate analysis demonstrated that baseline eGFR measured by either CKD-EPI or MDRD was predictive of AKI by both VARC 1 and VARC 2 definitions (Table 2). Alternatively, eCrCl was not predictive of AKI with either definition. We performed further correlation testing, which suggested that median values for renal function by CG, CKD-EPI, and MDRD were not significantly different (P=.61). However, the median eGFR values by CKD-EPI and MDRD calculations were highly correlated with each other, with an estimated Pearson correlation coefficient of r=0.92. Correlations between eCrCl by CG and MDRD and CKD-EPI were less robust (r=0.77 and r=0.79, respectively).

Use of transapical or transaortic approach (non-transfemoral) was predictive of AKI with VARC 2 but not VARC 1 criteria (Table 2). Contrast and periprocedural hydration volumes, number of pacing runs, valve type, medications, procedural changes in systolic blood pressure, and eGFR determined with CG were not associated with increased AKI rates using either criteria (P>.05).

Discussion

This retrospective study is the first to compare renal outcomes using VARC 1 and VARC 2 definitions of AKI. The key findings are more robust associations between clinical risk factors and AKI when using the VARC 2 criteria. In addition, estimation of renal function using the CG equation has a limited role in patients undergoing TAVR to predict subsequent development of AKI.

Many risk factors associated with AKIs in this population were evaluated; two were found to have statistically significant association, including CKD and procedural approach. Baseline renal function was found to be the strongest predictor of AKI. Further investigation of the different methods of measuring CKD suggested that both MDRD and CKD-EPI equations were similar and provided eGFRs that were clearly associated with AKI. Alternatively, in this study, the use of CG-calculated CrCl was of little value in prediction of AKI. This is consistent with the opinion of the National Kidney Foundation, which suggests that the CG formula no longer be used.  The CG formula is limited by the fact that it was validated using a homogenous group of elderly Caucasian men, which is not representative of other clinical populations.⁸ This formula has been shown to be additionally biased in estimating renal function in patients who are young, female, and those with increased body weight/body mass index.^9,10

The association between a non-transfemoral approach and increased risk of AKI is consistent with the literature.¹¹ Alternative access is generally performed if a patient has smaller femoral arteries or was found to have severe peripheral vascular disease (PVD). This association is thought to be due to more than just concomitant PVD and has other theorized causes. Patients with PVD are likely to have calcified aortas, which may lead to embolization of cholesterol crystals that are part of plaques that become dislodged during the procedure. The use of general anesthesia in non-transfemoral procedures could also lead to more hypotension procedurally.¹²

In the present analysis, evaluation of the hemodynamic variables obtained before and after TAVR failed to show any association with AKI using either definition. This was also the case for periprocedural intravenous hydration and pacing runs, which are variables that have not been well described in the literature in regard to AKI in the TAVR population. Anemia and blood loss have been found in other studies to be associated with AKI, ^13,14 but this was not the case in the present analysis. Interestingly, the amount of contrast dye used — although known to be deleterious to the kidneys in studies of coronary angiography¹⁵ — did not have an association with AKI in the present study. This is less surprising when taking into account that the body of literature has had difficulty in making the association between increasing contrast volumes and worsening kidney outcomes in the TAVR population. In fact, the studies have reached mixed conclusions regarding whether contrast volume plays a role in post-TAVR AKI.^16-20 This may be because the pathophysiology behind AKI in TAVR is different than in coronary angiography. Perhaps in TAVR, other factors such as procedural hypotension and embolization of cholesterol crystals are more important in the development of AKI. The TAVR population is also inherently different because they are all in various degrees of decompensated heart failure — and may, in fact, need diuresis and left ventricular unloading rather than volume expansion. Larger studies with registry data may be needed to finally be able to solidify contrast volume as a factor in post-TAVR AKI.

Study limitations. There are several limitations to the present study. The sample size is small and derived from a single-center, restrospective experience. The patient population largely received transfemoral procedures using the Sapien valve and were all done with general anesthesia. Extrapolation of these results to other TAVR centers may be less certain. We also lacked accurate urine output data — a variable that is difficult to obtain in many clinical settings.

Conclusion

Although limited by a smaller sample size and single-center design, the current study adds to the body of knowledge regarding AKI post TAVR. Specifically, we demonstrated that the VARC 2 definition may be the most associated with baseline renal function and implant approach. The broader clinical relevance of VARC 1 and VARC 2 criteria in the context of procedural variables and renal outcomes remains to be tested in a larger study.

From the ¹Division of Cardiology, Department of Medicine, University of Texas Health San Antonio, San Antonio, Texas; ²Department of Epidemiology and Biostatistics, UT Health San Antonio, San Antonio, Texas; ³Division of Nephrology, Department of Medicine, University of Texas Health San Antonio, San Antonio, Texas; and ⁴Henry Ford Health System, Division of Cardiology, Detroit, Michigan.

Funding: This work was funded by the Freeman Heart Association.

Disclosure: The authors have completed and returned the ICMJE Form for Disclosure of Potential Conflicts of Interest. The authors report no conflicts of interest regarding the content herein.

Manuscript submitted September 1, 2019, final version accepted September 9, 2019.

Address for correspondence: Anand Prasad, MD, FACC, FSCAI, RPVI, Interventional Cardiology and Vascular Medicine, Department of Medicine, Division of Cardiology, University of Texas Health Science Center at San Antonio, 7703 Floyd Curl Drive, MC 7872, San Antonio, TX 78229-3900. Email: anandprasadmd@gmail.com

1. Meneguz-Moreno RA, Ramos AI, Siqueira D, et al. Prognostic value of renal function in patients with aortic stenosis treated with transcatheter aortic valve replacement. Catheter Cardiovasc Interv. 2017;89:452-459.
2. Crowhurst JA, Savage M, Subban V, et al. Factors contributing to acute kidney injury and the impact on mortality in patients undergoing transcatheter aortic valve replacement. Heart Lung Circ. 2016;25:282-289.
3. Voigtlander L, Schewel J, Martin J, et al. Impact of kidney function on mortality after transcatheter valve implantation in patients with severe aortic valvular stenosis. Int J Cardiol. 2015;178:275-281.
4. Leon MB, Piazza N, Nikolsky E, et al. Standardized endpoint definitions for transcatheter aortic valve implantation clinical trials: a consensus report from the Valve Academic Research Consortium. J Am Coll Cardiol. 2011;57:253-269.
5. Kappetein AP, Head SJ, Genereux P, et al. Updated standardized endpoint definitions for transcatheter aortic valve implantation: the Valve Academic Research Consortium-2 consensus document. J Am Coll Cardiol. 2012;60:1438-1454.
6. Oguri A, Yamamoto M, Mouillet G, et al. Impact of chronic kidney disease on the outcomes of transcatheter aortic valve implantation: results from the FRANCE 2 registry. EuroIntervention. 2015;10:e1-e9.
7. Codner P, Levi A, Gargiulo G, et al. Impact of renal dysfunction on results of transcatheter aortic valve replacement outcomes in a large multicenter cohort. Am J Cardiol. 2016;118:1888-1896.
8. Cockcroft DW, Gault MH. Prediction of creatinine clearance from serum creatinine. Nephron. 1976;16:31-41.
9. Helou R. Should we continue to use the Cockcroft-Gault formula? Nephron Clinical Practice. 2010;116:c172-c185; discussion c186.
10. Michels WM, Grootendorst DC, Verduijn M, Elliott EG, Dekker FW, Krediet RT. Performance of the Cockcroft-Gault, MDRD, and new CKD-EPI formulas in relation to GFR, age, and body size. Clin J Am Soc Nephrol. 2010;5:1003-1009.
11. Thongprayoon C, Cheungpasitporn W, Srivali N, et al. Transapical versus transfemoral approach and risk of acute kidney injury following transcatheter aortic valve replacement: a propensity-adjusted analysis. Ren Fail. 2017;39:13-18.
12. Cheungpasitporn W, Thongprayoon C, Kashani K. Transcatheter aortic valve replacement: a kidney’s perspective. J Renal Inj Prev. 2016;5:1-7.
13. Arai T, Morice MC, O’Connor SA, et al. Impact of pre- and post-procedural anemia on the incidence of acute kidney injury and 1-year mortality in patients undergoing transcatheter aortic valve implantation (from the French Aortic National CoreValve and Edwards 2 [FRANCE 2] registry). Catheter Cardiovasc Interv. 2015;85:1231-1239.
14. Konigstein M, Ben-Assa E, Banai S, et al. Periprocedural bleeding, acute kidney injury, and long-term mortality after transcatheter aortic valve implantation. Can J Cardiol. 2015;31:56-62.
15. Tsai TT, Patel UD, Chang TI, et al. Contemporary incidence, predictors, and outcomes of acute kidney injury in patients undergoing percutaneous coronary interventions: insights from the NCDR Cath-PCI registry. JACC Cardiovasc Interv. 2014;7:1-9.
16. Sun Y, Liu X, He Y, et al. Meta-analysis of impact of diabetes mellitus on outcomes after transcatheter aortic valve implantation. Am J Cardiol. 2017;119:623-629.
17. Vontobel J, Possner M, Schutz P, et al. Early safety outcome following transcatheter aortic valve implantation: is the amount of contrast media used a matter of concern? Swiss Med Wkly. 2015;145:w14238.
18. Yamamoto M, Hayashida K, Mouillet G, et al. Renal function-based contrast dosing predicts acute kidney injury following transcatheter aortic valve implantation. JACC Cardiovasc Interv. 2013;6:479-486.
19. Giannini F, Latib A, Jabbour RJ, et al. The ratio of contrast volume to glomerular filtration rate predicts acute kidney injury and mortality after transcatheter aortic valve implantation. Cardiovasc Revasc Med. 2017;18:349-355.
20. Thongprayoon C, Cheungpasitporn W, Podboy AJ, Gillaspie EA, Greason KL, Kashani KB. The effects of contrast media volume on acute kidney injury after transcatheter aortic valve replacement: a systematic review and meta-analysis. J Evid Based Med. 2016;9:188-193.