Association of Transcatheter Aortic Valve Replacement With 30-Day Renal Function and 1-Year Outcomes Among Patients Presenting With Compromised Baseline Renal Function

Key Points

Question

What is the effect of transcatheter aortic valve replacement on renal function and clinical outcomes among high-surgical risk or inoperable patients with severe aortic stenosis and baseline impaired renal function?

Findings

In this substudy of a clinical trial, baseline renal dysfunction was present in 72% of patients, 42% of whom experienced an improvement in renal function and 24% whose renal function worsened after transcatheter aortic valve replacement, and at 1 year, a worsening estimated glomerular filtration rate (eGFR) was associated with an increased mortality trend compared with those with an unchanged eGFR. Significant predictors of an improved eGFR were being female and nonsmoking; predictors of worsening eGFR were left ventricle mass, smoking, and age; and predictors of 1-year mortality were left ventricular ejection fraction, baseline eGFR, and worsening eGFR.

Meaning

Baseline renal dysfunction was frequent among patients who underwent a transcatheter aortic valve replacement and those with worsening renal function after the procedure trended toward increased mortality.

Abstract

Importance

The frequency of baseline renal impairment among high-risk and inoperable patients with severe aortic stenosis undergoing a transcatheter aortic valve replacement (TAVR) and the effect of TAVR on subsequent renal function are, to our knowledge, unknown.

Objective

To determine the effect of TAVR among patients with baseline renal impairment.

Design, Setting, and Participants

This substudy of patients with baseline renal impairment (estimated glomerular filtration rate [eGFR] ≤ 60 mL/min) and paired baseline and 30-day measures of renal function undergoing TAVR in the PARTNER 1 trial and continued access registries was conducted in 25 centers in the United States and Canada.

Main Outcomes and Measures

Patients were categorized with improved eGFR (30-day follow-up eGFR≥10% higher than baseline pre-TAVR), worsened eGFR (≥10% lower), or no change in renal function (neither). Baseline characteristics, 30-day to 1-year all-cause mortality, and repeat hospitalization were compared. Multivariable models were constructed to identify predictors of 1-year mortality and of improvement/worsening in eGFR.

Results

Of the 821 participants, 401 (48.8%) were women and the mean (SD) age for participants with improved, unchanged, or worsening eGFR was 84.90 (6.91) years, 84.37 (7.13) years, and 85.39 (6.40) years, respectively. The eGFR was 60 mL/min or lower among 821 patients (72%), of whom 345 (42%) improved, 196 (24%) worsened, and 280 (34%) had no change at 30 days. There were no differences in baseline age, body mass index, diabetes, chronic obstructive pulmonary disease, coronary artery disease, peripheral arterial disease, hypertension, pulmonary hypertension, renal or liver disease, New York Heart Association III/IV symptoms, transaortic gradient, left ventricular ejection fraction, or procedural characteristics. The group with improved eGFR had more women, nonsmokers, and a lower cardiac index. Those with worsening eGFR had a higher median Society of Thoracic Surgeons score and left ventricle mass. From 30 days to 1 year, those with improved eGFR had no difference in mortality or repeat hospitalization. Those with worsening eGFR had increased mortality (25.5% vs 19.1%, P = .07) but no significant increases in repeat hospitalization or dialysis. Predictors of improved eGFR were being female (odds ratio [OR], 1.38; 95% CI, 1.03-1.85; P = .03) and nonsmoking status (OR, 1.49; 95% CI, 1.11-1.01; P = .01); predictors of worsening eGFR were baseline left ventricle mass (OR, 1.00; 95% CI, 1.00-1.01; P = .01), smoking (OR, 1.51; 95% CI, 1.06-2.14; P = .02), and age (OR, 1.03; 95% CI, 1.00-1.05; P = .05); and predictors of 1-year mortality were baseline left ventricular ejection fraction (OR, 0.98; 95% CI, 0.97-0.99; P = .003), baseline eGFR (OR, 0.98; 95% CI, 0.96-0.99; P < .001), and worsening eGFR vs no change in eGFR (OR, 1.51; 95% CI, 1.02-2.24; P = .04).

Conclusions and Relevance

Baseline renal impairment was frequent among patients who underwent TAVR. While improved eGFR did not improve 1-year outcomes, worsening eGFR was associated with increased mortality.

Trial Registration

clinicaltrials.gov Identifier: NCT00530894

This cohort substudy of a clinical trial examines renal functions and outcomes of high-surgical-risk or inoperable patients with severe symptomatic aortic stenosis undergoing a transcatheter aortic valve replacement who had impaired baseline estimated glomerular filtration rates at 30 days.

Introduction

Transcatheter aortic valve replacement (TAVR) is an effective therapy for intermediate- to high-surgical risk and inoperable patients with severe symptomatic aortic stenosis. Patients undergoing TAVR often have an elevated burden of comorbidities, and 29% to 34% of those in large multicenter registries experience moderate to severe baseline renal dysfunction.

It has previously been shown that severe baseline renal dysfunction predicts increased 1-year rates of morbidity and mortality among patients undergoing TAVR. Furthermore, previous studies have reported acute kidney injury among 12% to 57% of patients undergoing TAVR (the risk being higher among those with impaired baseline renal function), and this is associated with a 2-to-6-fold increase in short- and long-term mortality rates. Therefore, concern has been raised about offering TAVR as a treatment option to these patients. However, these patients also have an increased risk of mortality and morbidity after surgical aortic valve replacement, and 1 in 3 patients with severe valvular heart disease are refused the opportunity to undergo surgery based on the severity of their baseline renal disease.

The causes of renal impairment among patients with severe symptomatic aortic stenosis (AS) are multifactorial and it is possible that among those with impaired baseline renal function, TAVR may improve the postoperative renal function by relieving AS and improving cardiac output, leading to improved renal perfusion, decreased renal venous pressure, and improvement in right ventricular dysfunction. However, the procedure itself may have an adverse effect on postprocedural renal function because of factors such as using iodinated contrast, hypotension during rapid pacing, athero-emboli, and procedural complications such as bleeding. To our knowledge, the effect of TAVR on subsequent renal function among patients presenting with reduced baseline renal function has not been rigorously evaluated and data are limited.

We hypothesized that among patients with severe symptomatic AS and reduced baseline renal function, TAVR may improve postprocedural renal function and affect all-cause mortality. This hypothesis was evaluated among patients undergoing TAVR in the Placement of Aortic Transcatheter Valve (PARTNER) I Trial and Continued Access (CA) Registry.

Methods

Patient Selection

The design of the PARTNER 1 trial, along with detailed selection criteria and operative methods, have been previously reported. Enrolled patients had severe, symptomatic trileaflet AS (aortic valve area of <0.8 cm² with either a mean aortic valve gradient of ≥40 mm Hg or a peak aortic jet velocity of ≥4.0 m/s), a New York Heart Association class II or greater heart failure symptoms, and at least a high surgical risk based on the Society for Thoracic Surgeons mortality risk score and other factors determined by the heart team. In PARTNER 1A, high-risk patients were randomized to surgical aortic valve replacement or TAVR. In PARTNER 1B, inoperable patients were randomized to TAVR or medical treatment with or without a balloon aortic valvuloplasty. All patients undergoing transfemoral TAVR received either a 23- or a 26-mm balloon-expandable Edwards SAPIEN transcatheter heart valve (Edwards Lifesciences). At the time of the PARTNER 1 trial and registry, a 29-mm valve was not available. Important exclusion criteria included bicuspid aortic valve disease, left ventricular ejection fraction less than 20%, severe renal insufficiency (renal insufficiency with serum creatinine >3.0 mg/dL and/or end-stage renal disease requiring chronic dialysis), severe mitral regurgitation, severe aortic regurgitation, recent gastrointestinal bleeding, or recent neurologic event.

All patients undergoing TAVR from the PARTNER 1A, PARTNER 1B (inoperable), and both the randomized and nonrandomized CA cohorts with baseline eGFR at 60 mL/min or fewer (using the modified diet in renal disease formula) were analyzed. (Figure). Of 2559 patients enrolled in the PARTNER 1 trial and CA registry, 1792 patients had a baseline eGFR of 60 mL/min or lower. Of these, 821 patients had paired creatinine measurements available at baseline and 30 days and composed the study population. Patients were categorized as having improved eGFR (30-day follow-up eGFR at least 10% higher than baseline preTAVR eGFR), worsened eGFR (30-day follow-up eGFR at least 10% lower than baseline preTAVR eGFR), or having experienced no change (not fitting either previous group). Since the Valve Academic Research Consortium criteria do not specify criteria for improvement in renal function after TAVR, we used a 10% change in eGFR as being clinically meaningful. An independent clinical events committee adjudicated all the clinical events. The institutional review board at each participating site approved the study, and all patients provided written informed consent.

Figure. Kaplan-Meier Estimates of 30-Day to 1-Year All-Cause Mortality Among Groups A, B, and C.

The overall log-rank P value for groups A (improvement in renal function), B (no change in renal function), and C (worsening of renal function) was .01. The log-rank P value for group A vs B was .22 (hazard ratio, 0.79; 95% CI, 0.54-1.16). The log-rank P value for group A vs C was .003 (hazard ratio, 0.55; 95% CI, 0.37-0.82). The log-rank P value for group B vs C was .07 (hazard ratio, 0.70; 95% CI, 0.47-1.03). Pairwise comparisons are also shown.

Study End Points

The frequency of worsening, no change, or improvement in eGFR at 30 days following the TAVR procedure was calculated. The frequency of 30-day to 1-year all-cause mortality and cardiovascular mortality, repeat hospitalization, any stroke, major stroke, death/major stroke, aortic valve reintervention, major bleeding, and renal failure requiring dialysis were reported according to a modified version of the Valve Academic Research Consortium-1. These end points were prespecified and adjudicated.

Statistical Analysis

Continuous variables were summarized as mean (SD) or as medians and quartiles and were compared using t tests or Mann-Whitney rank sum tests. Categorical variables were compared using χ² tests or Fisher exact tests. For patients who survived up to 30 days, landmark survival curves for time-to-event variables were constructed on the basis of all available follow-up data using Kaplan-Meier estimates, and comparisons were performed using the log-rank test. Univariate and multivariable analyses were performed to determine baseline clinical and echocardiographic characteristics that contributed to differences in all-cause mortality between the 3 groups. For the multivariable models predicting improvement or worsening in 30-day eGFR, we used the group who experienced no significant change in eGFR as the comparator population.

A Cox proportional hazard regression analysis was performed to determine predictors of 1-year all-cause mortality. Logistic regression models were used to determine the predictors of improvement in eGFR and of worsening in eGFR. The multivariable models were built by stepwise selection, with candidate variables being selected if they were of clinical interest or satisfied the entry criterion of P < .10 in the univariable analysis. Variables were entered with entry/stay criteria of 0.1/0.1 in a forward stepwise fashion.

A 2-sided α level of .05 was used for all statistical testing. All statistical analyses were performed using SAS, version 9.4 (SAS Institute Inc). An independent academic biostatistics group performed all data analyses.

Results

Of the 1142 patients who had paired baseline and 30-day follow-up serum creatinine levels available, 821 (72%) had impaired baseline renal function (eGFR ≤ 60 mL/min). Of these, 42% (n = 345) showed an improvement in renal function, 34% (n = 280) showed no change, and 24% (n = 196) showed worsening renal function following TAVR (eFigure in the Supplement).

Baseline Clinical, Echocardiographic, and Procedural Characteristics

Comparing baseline clinical characteristics between the 3 groups, there was no statistical difference in the mean age (mid 80s), body mass index (calculated as weight in kilograms divided by height in meters squared), or frequency of diabetes, chronic obstructive pulmonary disease, renal disease (baseline serum creatinine > 2 mg/dL), coronary artery disease, peripheral arterial disease, hypertension, pulmonary hypertension, New York Heart Association class 3/4, or liver disease. Women were more likely to have improved eGFR following TAVR (improved eGFR, 55.1%; no change in eGFR, 47.1%; worsening eGFR, 40.3%; P = .003). People who smoked were less likely to have improved eGFR following TAVR (improved eGFR, 39.4%; no change in eGFR, 48.0%; worsening eGFR, 54.4%; P = .003). The mean (SD) baseline eGFR was lowest among the people with improved GFR (39.7 [11.8] mL/min) compared with those who experienced no change in eGFR (41.6 [10.3]) and worsening GFR (41.9 [10.6]) (P = .03) (Table 1).

Table 1. Baseline and Procedural Characteristics.

Characteristic	eGFR			P Value
	Improved (n = 345)	No Change (n = 280)	Worsening (n = 196)
Baseline characteristics
Age, mean (SD), y	84.90 (6.91)	84.37 (7.13)	85.39 (6.40)	.42
Women, %	55.1	47.1	40.3	.003
BMI, mean (SD)	26.88 (6.64)	27.43 (6.26)	26.62 (5.06)	.26
Diabetes, %	38.0	35.1	36.4	.76
COPD, %	45.8	39.6	43.9	.3
Renal disease (sCr > 2 mg/dL), %	20.3	18.3	23.2	.43
CAD, %	79.4	78.5	81.5	.71
PAD, %	41.3	41.2	44.6	.71
Smoker, %	39.4	48.0	54.4	.003
HTN, %	92.5	92.8	94.9	.55
PHTN, %	42.5	37.5	38.9	.44
NYHA ¾, %	95.7	92.9	93.9	.31
Liver disease, %	2.6	2.9	0	.06
STS score, median (IQR)	10.90 (9.50-13.30)	10.60 (9.10-12.90)	11.10 (9.70-13.25)	.04
Baseline GFR, mean (SD)	39.7 (11.8)	41.6 (10.3)	41.9 (10.6)	.03
Baseline echocardiographic and hemodynamic characteristics, mean (SD)
Mean aortic valve gradient, mm Hg	43.75 (14.14)	43.77 (14.08)	42.31 (13.62)	.61
Cardiac index, L/min/m2	1.98 (0.60)	2.13 (0.65)	2.21 (1.67)	.02
Stroke volume, mL	63.2 (20.8)	65.1 (19.4)	69.3 (20.7)	.06
Stroke volume index	35.4 (10.3)	35.1 (9.4)	35.4 (10.5)	.90
LVEF, %	50.5 (13.7)	52.5 (12.4)	53.0 (12.4)	.18
LV mass, g	243.49 (68.85)	249.72 (84.02)	263.80 (72.02)	.01
Procedural characteristics
Successful valve implantation, %	98.3	97.5	97.4	.75
Hemodynamic support needed, %	5.5	3.6	8.2	.09
Conversion to open heart surgery, %	1.2	0.7	2.1	.42
Volume of contrast used, mean (SD), cm3	120.84 (173.89)	118.96 (98.32)	113.43 (71.45)	.79

Abbreviations: BMI, body mass index (calculated as weight in kilograms divided by height in meters squared); CAD, coronary artery disease; COPD, chronic obstructive pulmonary disease; eGFR, estimated glomerular filtration rate (determined by MDRD formula); GFR, glomerular filtration rate; HTN, hypertension; IQR, interquartile range; LV, left ventricular; LVEF, left ventricular ejection fraction; NYHA, New York Heart Association class; PAD, peripheral artery disease; PHTN, pulmonary hypertension; sCR, serum creatinine (to convert to micromoles per liter, multiply by 76.25); STS, Society of Thoracic Surgeons Predicted Risk of Mortality score.

Baseline echocardiographic parameters revealed no difference between the groups for the mean aortic valve gradient or baseline left ventricular ejection fraction. However, the mean (SD) cardiac index was significantly different between the groups, being lowest among the group with improved eGFR (1.98 [0.61] L/min/m²) and similar among the group who experienced no change in eGFR (2.13 [0.65]) and those with worsening eGFR (2.10 [0.66]) (P = .02 across groups). Left ventricle mass was highest among the group with worsening eGFR (263.80 [72.02] g; P = .01). Stroke volume tended to be lowest among the group with improved eGFR (63.2 [20.8] mL) and highest among the group with worsening eGFR (69.3 [20.7] mL) (P = .06 across groups) (Table 1).

Procedural characteristics, including successful valve implantation, conversion to open heart surgery or volume of contrast used, and the need for repeat aortic valve intervention, were similar between the 3 groups. There was a trend toward an increased need for hemodynamic support during the TAVR procedure among the group with worsening eGFR as compared with the other 2 groups (improved eGFR, 5.5%; no change in eGFR, 3.6%; worsening eGFR, 8.2%; P = .09) (Table 1).

Those who had more repeat hospitalizations or had acute kidney injury, as defined by needing renal dialysis between baseline and 30 days, were more commonly found among the group with worsening eGFR. The frequency of acute kidney injury by this definition was low among the groups with improved and unchanged eGFR (1.4% each) and higher (7.6%) among the group with worsening eGFR.

Data comparing patients with paired eGFR (those included in this analysis) and those without paired eGFR (those excluded from this analysis) are shown as supplemental data (eTables 1, 2, and 3 in the Supplement). Overall, the 2 groups were similar in terms of their baseline clinical characteristics except that there were fewer patients with baseline serum creatinine at greater than 2 mg/dL in the group without paired eGFR (with paired eGFR, 20.3%; without paired eGFR, 14.6%; P < .001). There were no differences between the groups in terms of echocardiographic or procedural characteristics.

30-Day to 1-Year Outcomes

From 30 days to 1-year follow-up, the rate of all-cause mortality correlated with the 30-day change in eGFR (improved eGFR, 15.4%; no change in eGFR, 19.1%; worsening eGFR, 25.5%; P = .01), as did the rate of cardiovascular mortality (improved eGFR, 4.7%; no change in eGFR, 6.8%; worsening eGFR, 11.1%; P = .02), the composite of death or major stroke (improved eGFR, 15.2%; no change in eGFR, 25.3%; P = .01), and the occurrence of renal failure requiring dialysis (improved eGFR, 1.4%; no change in eGFR, 3.5%; worsening eGFR, 5.9%; P = .01). There was no significant difference in the rate of repeat hospitalization between 30 days and 1 year (improved eGFR, 12.3%; no change in eGFR, 15.2%; worsening eGFR, 19.0%; P = .17).

These differences were primarily driven by differences between the groups with worsening and improved eGFR. Pairwise comparisons between the group with improved eGFR and those who had experienced no change in eGFR were not significant for all-cause mortality (15.4% vs 19.1%; P = .22), the need for repeat hospitalization (12.3% vs 15.2%; P = .35), or any other clinical end points. Compared with those who experienced no change in eGFR, the group with worsening eGFR trended toward increased all-cause mortality (25.5% vs 19.1%; P = .07) and renal failure requiring dialysis (5.9% vs 3.5%; P = .19), but there were no significant differences in other end points.

However, when compared with the group with improved eGFR, the group with worsening eGFR had significantly increased all-cause mortality rates (15.4% vs 25.5%; P = .003), cardiovascular mortality rates (P = .01), composite of death or major stroke (P = .003), and renal failure (P = .002). Comparative differences in clinical end points from 30 days to 1-year are shown in Table 2 and in the Figure.

Table 2. Landmark Analysis 30-Day to 1-Year Clinical Outcomes^a.

Outcome	eGFR			P Value	Improved vs No Change	Worsening vs No Change	Worsening vs Improved
	Improved (n = 345)	No Change (n = 280)	Worsening (n = 196)
Death
All-cause mortality	15.4	19.1	25.5	.01	0.22	0.07	0.002
Cardiovascular	4.7	6.8	11.1	.02	0.25	0.10	0.004
Repeat hospitalization	12.3	15.2	19.0	.17	0.35	0.31	0.06
Any stroke	0.9	2.0	3.3	.25	0.32	0.46	0.09
Major stroke	0.9	2.0	2.5	.44	0.32	0.74	0.20
Death or major stroke	17.7	19.0	25.3	.01	0.23	0.08	0.003
Aortic valve reintervention	0.03	1.8	1.1	.17	0.06	0.52	0.25
Major bleeding	6.2	5.9	8.2	.60	0.86	0.35	0.42
Renal failure (dialysis required)	1.4	3.5	5.9	.01	0.08	0.19	0.002

Abbreviation: eGFR, estimated glomerular filtration rate.

^aAll results presented are Kaplan-Meier estimate percentages.

Multivariable Predictors of Change in eGFR From Baseline to 30-Days PostTAVR

To determine correlates of improvement or worsening in 30-day post-TAVR eGFR, multivariable models were constructed including age, sex, smoking status, presence of diabetes, chronic obstructive pulmonary disease, hypertension, baseline eGFR, peripheral arterial disease, anemia, baseline left ventricular mass, baseline eGFR, any aortic valve reintervention, and any major bleeding event or bleeding requiring blood transfusion. In these models, female sex was associated with improvement in eGFR (odds ratio [OR], 1.38; 95% CI, 1.03-1.85; P = .03) as was being a nonsmoker (OR, 1.49; 95% CI, 1.11-2.04; P = .01). Factors associated with worsening eGFR at 30 days post-TAVR were baseline left ventricular mass (OR [per 10 g increment], 1.03; 95% CI, 1.01-1.06; P = .02), smoking (OR, 1.51; 95% CI, 1.06-2.14; P = .02), and age (OR, 1.03; 95% CI, 1.00-1.05; P = .05) (Table 3).

Table 3. Multivariable Predictors of Improved and Worsening eGFR After TAVR^a,b,c.

Multivariable Predictors	Hazard Ratio (95% CI)	P Value
Improved eGFR after TAVR
Female	1.38 (1.03-1.85)	.03
Nonsmoking	1.49 (1.11-2.04)	.01
Worsening eGFR after TAVR
Baseline LV mass	1.00 (1.00-1.01)	.01
Smoking	1.51 (1.06-2.14)	.02
Age	1.03 (1.00-1.05)	.05

Abbreviations: eGFR, estimated glomerular filtration rate; LV, left ventricular; TAVR, transcatheter aortic valve replacement.

^aModels were constructed using stepwise logistic regression with an entry/stay criteria of 0.1/0.1.

^bCandidate variables considered in both models were: age, sex, diabetes, smoking, hypertension, peripheral vascular disease, chronic obstructive pulmonary disease, anemia, baseline LV mass (per 10g), baseline eGFR, aortic valve reintervention, major bleeding, and transfusion.

^cFor the multivariable models predicting improvement or worsening in 30-day eGFR, we used the group with no significant change in eGFR as the comparator population.

Multivariable Predictors of All-Cause 1-Year Mortality

Adjustments were made for age, female sex, body mass index, Society of Thoracic Surgeons risk score, diabetes, baseline eGFR, cerebrovascular disease, chronic obstructive pulmonary disease, stroke or transient ischemic attack (last 6-12 months), baseline left ventricle ejection fraction, prior coronary artery bypass graft, changes in stroke volume index from baseline to 30 days, and changes in eGFR (in Model 1 as a continuous variable and in Model 2 as a categorical variable with no change as the reference category). In the model using change in eGFR as a continuous variable, the significant predictors of all-cause 1-year mortality were baseline left ventricle ejection fraction (hazard ratio [HR], 0.98; 95% CI, 0.97-0.99; P = .003), baseline eGFR (HR, 0.97; 95% CI, 0.96-0.99; P < .001), and delta eGFR (HR, 0.45; 95% CI, 0.25-0.81; P = .01). Examining changes in eGFR as a categorical variable revealed that worsening eGFR was significantly associated with an increased hazard of 1-year mortality (HR, 1.51; 95% CI, 1.02-2.24; P = .04), while increased eGFR showed no significant association (HR, 0.74; 95% CI, 0.50-1.09; P = .13) (Table 4).

Table 4. Multivariable Predictors of 30-Day to 1-Year All-Cause Mortality.

Variable	Hazard Ratio (95% CI)	P Value
Model 1: Delta eGFR (continuous)
Baseline ejection fraction	0.98 (0.97-0.99)	.003
Baseline eGFR	0.97 (0.96-0.99)	.001
Delta eGFR (30 d eGFR/Baseline eGFR)	0.45 (0.25-0.81)	.01
Model 2: eGFR Categories
Baseline ejection fraction	0.98 (0.97-0.99)	.004
Baseline eGFR	0.98 (0.96-0.99)	.003
eGFR
Improved vs no change	0.74 (0.50-1.09)	.13
Worsening vs no change	1.51 (1.02-2.24)	.04

Abbreviation: eGFR, estimated glomerular filtration rate.

Discussion

In a first such analysis, to our knowledge, of high-surgical risk or inoperable patients with severe symptomatic AS undergoing TAVR who had impaired baseline eGFR and paired eGFR measurements at 30 days, we found that (1) decreased baseline eGFR was frequent, being present among 72% of patients; (2) TAVR resulted in a 10% or greater improvement in eGFR among 42% of patients, an at least 10% worsening of eGFR among 24% of the patients, and no change in eGFR among 34% of patients; (3) independent baseline correlates of improvement in eGFR after TAVR were being female and nonsmoking status; (4) between 30 days and 1 year, changes in eGFR were associated with all-cause and cardiovascular mortality, the composite of death and major stroke, and the incidence of renal failure (need for dialysis); and (5) using the group who experienced no change in eGFR as a comparator arm, the group with improved eGFR showed no differences in all-cause mortality or the need for repeat hospitalization, but did trend toward a lower need for dialysis, while patients with worsening eGFR trended toward increased all-cause mortality.

The present analysis underscores the high prevalence of baseline renal insufficiency among high-risk and inoperable patients with severe symptomatic AS, the cause of which is likely to be multifactorial, including type 2 cardiorenal syndrome (decreased renal perfusion because of cardiac dysfunction) related to AS-induced reduced cardiac output, right ventricular dysfunction, and elevated renal venous pressure, all of which contribute to decreased renal perfusion. Additionally, these patients have frequent comorbidities such as hypertension, dyslipidemia, peripheral arterial disease, and older age, all of which contribute to renal dysfunction and are associated with elevated operative risk.

It is interesting that among this cohort of high-risk or inoperable patients with compromised baseline renal function undergoing TAVR, 76% of patients experienced either improvement (42%) or no change (34%) in their renal function after TAVR. It appears, therefore, that among these patients, the salutary hemodynamic effects of TAVR resulting in improved renal perfusion may have outweighed the adverse procedural features, such as using iodinated contrast, drops in blood pressure during rapid pacing at the time of valve deployment, and embolic debris. In the present analysis, those with improved eGFR after TAVR were more likely to have a lower baseline cardiac index and eGFR, suggesting that among these patients relief of AS with TAVR had a beneficial impact on type 2 cardiorenal syndrome. We found no differences between the groups in procedural characteristics such as the volume of contrast used, the rate of successful valve implantation, or the use of transfemoral or transapical access. There was a trend toward an increased need for hemodynamic support among the group with worsening eGFR compared with the group who experienced no change in eGFR, suggesting a more hemodynamically compromised group.

These results also suggest a more frequent rate of improvement in renal function (42% in the present study) than previously reported in a smaller single center study, which reported improvement in renal function among 15% of patients at 30 days (defined as an absolute decrease in serum creatinine ≥0.3 mg/dL) and included all patients irrespective of baseline renal function. In contrast, in the present study, we specifically studied patients who presented with compromised baseline renal function and found a substantial percentage with improved or stable postTAVR renal function. A selection bias may exist in the present analysis because we excluded patients without 30-day eGFR or those that had early mortality within 30 days. Because the Valve Academic Research Consortium definitions do not specify criteria to determine improvement in renal function after TAVR, we used a clinically meaningful change in eGFR of 10% rather than specifying a decline of serum creatinine of 0.3 mg/dL or more as in the other recent analysis.

Patients that had improvement in eGFR after TAVR showed no difference in all-cause mortality or the need for repeat hospitalization as compared with the group who experienced no change in eGFR. These results are surprising and may be because of the present study being underpowered to detect this difference. However, patients with worsening eGFR trended toward an increased 1-year all-cause mortality compared with those who experienced no change in eGFR. These results confirm the adverse effects of worsening renal function after TAVR.

We attempted to determine the predictors of improved eGFR and found that the predictive ability was limited. Our modeling did not reveal procedural characteristics such as repeat aortic valve interventions, major hemorrhage, or the need for a transfusion as being predictive beyond baseline clinical characteristics. Women were 38% more likely to have an improvement in post-TAVR eGFR. It has been previously shown in an analysis from the PARTNER trial that women undergoing TAVR in the PARTNER trial had a lower prevalence of comorbidities, including peripheral vascular disease, diabetes, and baseline renal dysfunction. Therefore, it is possible that the relief of type 2 cardiorenal syndrome is a dominant effect of TAVR among women, whereas among men this competes with baseline renal dysfunction driven by the more frequent occurrence of comorbidities. Smoking was predictive of worsening renal function (people who smoked were 51% more likely to worsen than people who did not). This was likely because of an increased incidence of peripheral arterial disease, endothelial dysfunction, baseline renal dysfunction, and possibly an increased risk of embolic debris during the procedure among people who smoked. While procedural characteristics are expected to affect post procedural eGFR, we found no significant difference between the groups in terms of procedural characteristics such as the volume of contrast used, procedural success, needing aortic valve reinterventions, major bleeding, vascular complications, or needing transfusions.

Careful preprocedural hydration, minimizing the duration of rapid pacing and associated hypotension at time of transcatheter valve deployment, bringing the blood pressure up if patient is hypotensive before placing the transcatheter valve across the aortic valve, and possibly using newer techniques such as the RenalGuard system (RenalGuard Solutions Inc) and renal embolic protection devices, among others, may all help to avoid post-TAVR worsening of renal function.

Of interest, independent predictors of worsening eGFR included history of smoking, baseline left ventricle mass and patient age. The association between smoking status and change in eGFR was commented on earlier. A higher baseline left ventricle mass was likely the result of longer standing AS and associated risk factors such as hypertension, while older patients were more likely to have a higher burden of comorbidities and renal dysfunction at baseline. The independent association of worsening eGFR at 30-day postprocedure with elevated 1-year mortality confirms results from previous studies.

Limitations

This study is a retrospective subanalysis of the PARTNER trial and is subject to the limitations inherently present in such analyses. A survival bias may exist in the present analysis because we excluded patients without 30-day eGFR or those that had early mortality within 30-day. Those with severe renal insufficiency (baseline serum creatinine >3 mg/dL) and those on dialysis were not included in the present analysis since they were excluded from the PARTNER trial and nonrandomized continued access registries. The impact of TAVR in those with baseline serum creatinine greater than 3.0 mg/dL is therefore unclear. The study is likely not adequately powered to detect differences in some clinical end points among the groups with improved or worsened eGFR compared with the group who experienced no change, despite event rates that follow a consistent pattern of being lowest among the group with improved eGFR, intermediate among the group who experienced no change, and highest among the group with worsened eGFR. We used a clinically meaningful definition of improvement or worsening of eGFR, but other definitions such an absolute increase or decrease of serum creatinine of 0.3 mg/dL could be used as an extrapolation from the Valve Academic Research Consortium definitions.

Conclusions

Decreased baseline eGFR was frequent among high-risk and inoperable patients with severe symptomatic AS. After undergoing TAVR, 76% of patients experienced either improved (42%) or unchanged (34%) renal function, diminishing the concern about offering this otherwise beneficial therapy to such patients. At 1 year, after adjusting for baseline risk factors, worsening eGFR was associated with an increased hazard of all-cause mortality.

Supplement.

eFigure. Study Flow.

eTable 1. Baseline Characteristics of Patients With vs. Without Paired eGFR Measurements at Baseline and 30 Days.

eTable 2. Baseline Echocardiographic and Hemodynamic Characteristics of Patients With vs. Without Paired eGFR.

eTable 3. Procedural Characteristics of Patients With vs. Without Paired eGFR.