Evaluation and management of aortic stenosis in chronic kidney disease

Abstract

The combination of aortic stenosis (AS) and chronic kidney disease (CKD) represents a formidable clinical challenge commonly encountered in practice. Specific guidelines addressing the management of AS in CKD/end-stage kidney disease (ESKD) patients are lacking. This review addresses the unique epidemiological features, caveats relevant to diagnostic modalities and specific management considerations pertinent for AS in patients with CKD/ESKD. Transthoracic echocardiography remains the imaging modality of choice for establishing an accurate diagnosis and assessing severity of AS in this population. However, the presence of anatomical and hemodynamic alterations along with co-prevalent valve disease often necessitates complementary imaging including dobutamine stress echocardiography, cardiac computed tomography and magnetic resonance imaging. Management includes blood pressure control, addressing CKD-related metabolic bone disease, and the selection of optimal dialysis modality. For aortic valve replacement (AVR), the choice between mechanical versus biological prosthesis, as well as between surgical versus transcatheter approaches requires careful consideration of patient-specific factors, including life expectancy, bleeding risk, risk of acute kidney injury; as well as prosthesis specific considerations including durability and need for long-term anticoagulation. Patients with CKD/ESKD encounter higher periprocedural risks as well as worse long-term outcomes; yet AVR is associated with improved survival compared to conservative management. Shared-decision making with the patient is essential regarding choice of prosthesis. A multidisciplinary heart-kidney team is recommended for optimizing perioperative planning, hemodynamic/volume status and preventing acute kidney injury to ensure a balanced and comprehensive patient-centered strategy for managing AS in the context of CKD/ESKD.

The combination of aortic stenosis (AS) and chronic kidney disease (CKD) represents a formidable combination encountered commonly in routine clinical practice. While there exist guidelines and recommendations for management of AS in the general population, several nuances and caveats are applicable in the evaluation and management of AS among individuals with CKD and end-stage kidney disease (ESKD). A recent scientific statement by the American Heart Association (AHA) has recommended a special emphasis on the “heart-kidney” team-based approach, adopting a shared-decision making approach with the patient, as well as the importance of considering the composite clinical outcomes of major adverse cardiac and renal endpoints (MACRE) that are meaningful for the patient. ¹ The current review will outline the unique epidemiological considerations of patients with AS and CKD, discuss the role of various non-invasive diagnostic modalities in establishing an accurate diagnosis of AS, as well as address specific management considerations in this population.

Epidemiology of AS in CKD/ESKD

While a higher prevalence of various echocardiographic abnormalities has been described in general among patients with chronic kidney disease (CKD), ² there are several noteworthy epidemiological considerations that are specific among patients with coprevalent AS and CKD. Firstly, AS is more prevalent among patients with CKD compared to the general population. Greater than mild AS has been described in nearly 10% of patients with CKD, which is almost 3-fold higher compared to patients without CKD. ² Secondly, the presence of CKD has been identified as an independent risk factor for progression in AS. In a retrospective study of patients with AS, there was an inverse correlation between kidney function and progression of AS. ³ Among patients with end-stage kidney disease (ESKD) on dialysis, AS has been shown to progress twice as rapidly as controls (decline in aortic valve area [AVA] −0.19 vs −0.07 cm² annually). ⁴ Thirdly, the presence of CKD is associated with higher all-cause and cardiovascular survival in AS after adjustment for other risk factors.^2,5 These distinctive epidemiological features are helpful for clinicians to not only recognize the high-risk nature of this population of co-prevalent CKD and AS, but also determine most optimal monitoring and management strategies.

Cardiovascular imaging considerations for AS in CKD

Role of echocardiography

Transthoracic echocardiogram (TTE) with Doppler imaging is the imaging modality of choice for the accurate diagnosis and assessment of the severity of aortic stenosis (AS). Other imaging modalities such as transesophageal echocardiogram (TEE), dobutamine stress echocardiography (DSE), aortic valve calcium scoring by computed tomography (CT) and invasive cardiac catheterization offer complementary imaging information and are typically recommended when reliable diagnostic data cannot be obtained using TTE or when there is a discrepancy between clinical presentation and echocardiographic findings. ⁶ During echocardiographic evaluation of AS, it is important to obtain a comprehensive evaluation of the aortic valve (AV) using multiple windows to capture accurate short-axis images for valve morphology and maximum Doppler signals to allow for accurate assessment of the AVA, gradients and velocities. The AVA is derived by using the continuity equation, based on the principle of conservation of mass. Pragmatically, the AVA is calculated using the formula: 0.785*LVOTd ² *LVOT VTI/AV VTI (LVOTd = left ventricular outflow tract diameter, VTI = velocity time integral). The LVOTd is measured in peak systole at the level of the aortic annulus and leaflet insertion point. Careful attention must be paid to the accuracy of this measurement since any error in its measurement can compound the impact on the AVA.^1,6 A zoomed view of the LVOT is recommended to faciliate accurate measurement. Based on the AVA, valve gradients and flow state, AS is categorized as typical high gradient AS (D1), low-flow, low-gradient (LFLG) AS with reduced ejection fraction (EF) (D2) or low-gradient AS with normal EF or paradoxical low-flow AS (D3). A threshold of stroke volume index (SVi), <35 mL/min/m² is typically the cutoff to diagnose a low-flow state. ⁶

Echocardiographic assessment of AS in CKD/ESKD can be challenging due to the anatomic, metabolic, and hemodynamic changes seen in this population.^1,3,5 For example, LVOTd measurement can be challenging due to poor acoustic windows, with a higher possibility of measurement errors due to left ventricular hypertrophy and higher burden of aortic annular and valve calcification owing to the unique metabolic milieu of secondary hyperparathyroidism. The presence of prominent basal septal hypertrophy negates an important assumption inherent to calculations in the continuity equation that the LVOT is circular. Additionally, septal hypertrophy could result in flow acceleration in the LVOT, thereby leading to measurement errors in the calculation of the LVOT VTI, directly impacting an accurate assessment of the AVA.^1,7,8 To address these concerns, experts recommend reviewing prior TTE images of the same patient for LVOTd measurement for consistency. Other solutions include relying on the dimensionless index (DI, <0.25 consistent with severe AS), calculating an estimated LVOTd ([5.7*BSA] + 12.1) or obtaining a TEE for more accurate assessment and planimetry.^1,8,9 Another challenge often encountered in the CKD/ESKD population is the impact of loading conditions and flow states on AS assessment. Frequently, uncontrolled hypertension, increased prevalence of low flow states, high flow states (especially in those with an arteriovenous fistula [AVF] for dialysis access) and prevalent anemia can pose challenges in the accurate assessment of flow dependent parameters of AS.^1,10 Systemic hypertension is a highly prevalent comorbidity in CKD/ESKD and can contribute to errors in accurate estimation of the severity of AS due to the resultant increase in aortic pressure causing aortic root expansion/ commissural separation, and increase in afterload resulting in increased systemic arterial resistance with resultant decrease in cardiac output (CO). ¹¹

To address this concern, experts recommend obtaining TTE for AS assessment when patients are at their hemodynamic baseline, eg, on a post-dialysis or interdialytic day to minimize the impact of hypertension and afterload. The valvuloarterial impedance (Zva) offers an instantaneous estimation of the composite afterload experienced by the LV. Zva is calculated using the formula: systolic blood pressure + transvalvular mean gradient divided by the Svi; Zva represents the entirety of the valvular and vascular afterload encountered by the LV. Elevated Zva (>3.5 mm Hg/mL/m²) has been associated with increased mortality in patients with AS.^12,13 However, specific data for prognostic value of Zva in CKD/ESKD are scant. ¹⁴ In addition to the issues discussed above, discordant grading secondary to the presence of low or high flow states is another commonly encountered challenge. It has been reported that LFLG severe AS is more commonly associated with CKD/ESKD; it is postulated that adverse ventricular remodeling, valvular regurgitation, and atrial fibrillation are responsible for low-flow states seen in CKD/ESKD population.. ¹⁵ Notably, patients with LFLG severe AS have worse prognosis compared to high gradient AS underscoring the importance of accurately confirming the severity of AS. ¹⁶ Low dose DSE is used to assess contractile reserve and distinguish severe and pseudo-severe AS. In cases where the results of the DSE are inconclusive, AV calcium score on CT can also be used to support the diagnosis of severe AS in LFLG patients.^1,6

Conversely, discordance may arise when transvalvular gradients are disproportionately higher compared to the AVA, this usually occurs in a high-flow state eg, in presence of a large AVF, anemia, pregnancy, thyrotoxicosis. High flow state in the presence of a large AVF could impact accurate assessment of AS in dialysis patients as well as post-transplant patients, caused by higher preload and decreased afterload. In an echocardiographic cohort of patients with coprevalent AS and ESKD, it was shown that TTE parameters of flow states and AS severity differ in those with AVF versus other dialysis accesses, and vary with progression in AS. ¹⁷ Although repeating the study when the high flow state/ underlying conditions have been corrected and when hemodynamics are more favorable can be attempted, this is often difficult to accomplish pragmatically.^1,18 Some authors have suggested routinely reporting the DI, SV, CO and noting the presence or absence, size, and position of an AVF in order to provide valuable context for the interpreting physician.^7,19 High output heart failure can occasionally occur as a consequence of the high flow state in the presence of a large AVF in CKD/ESKD or even post-renal transplant patients. ¹⁹ Data has shown that temporary occlusion of the AVF (Nicoladoni-Israel-Branham test) can normalize transvalvular flow and stroke volume; however, due to risk of fistula thrombosis, this manuever is not routinely recommended in clinical practice.^10,19,20

Presence of other valvular disease

Patients with CKD/ESKD have increased prevalence and rapid progression of valvular disease frequently affecting the aortic valve cusps and mitral annular calcification (MAC) leading to valve stenosis or regurgitation. The pathophysiology of accelerated calcification is thought to be related to the vascular effects of secondary hyperparathyroidism, hypocalcemia, increased shear stress, mineral bone disorders and chronic systemic inflammation prevalent in ESKD. ²¹ The presence of MAC with resultant stenosis can lead to a low flow state which in turn can impact the accurate assessment of AS. ²² In these patients, with mixed valve disease, it is often necessary to consider TEE to accurately evaluate associated valve disease to guide management decisions. ²³

Role of cardiac CT

There are specific instances where further evaluation with a cardiac CT provides incremental information to establish a diagnosis (eg, among patients with LFLG AS or paradoxical low-gradient AS). The extent of AV calcification and severity of AS are directly linked, since calcific deposits lead to valvular stiffness and narrowing. ⁷ Quantification of valve calcification performed with low radiation dose (<1 mSv) non-contrast CT with electrocardiographic gating and 3 mm slices using the modified Agatston method, has been shown to be a strong prognostic marker in predicting disease progression and survival. ²⁴ The 2021 European guidelines recommend the use of CT to determine the severity of AV calcification in patients with LFLG aortic stenosis with preserved ejection fraction. ²⁵ A score greater than or equal to 2000 Agatston units in men or 1200 Agatston units in woman is considered severe. Despite the fact that AV calcification is highly prevalent in CKD/ESKD patients there is no validated calcium scoring method or threshold for AS severity in CKD patients.^21,26 Extrapolating the thresholds from the general population to the CKD/ESKD population runs the risk of overestimating the hemodynamic severity of AS. In this population it is important that careful attention be paid to discriminating valvular and nonvalvular (LVOT, mitral annulus, aortic root, and coronary ostia) calcification to avoid overestimation of AS severity; AV calcium score should thus only include the regions of the valve cusps and aortic annulus. Contrast enhanced CT can further be utilized to assess AV morphology (especially discerning bicuspid from tricuspid valves), measure AVA by direct planimetry, and hybrid AVA (in which the LVOT area is measured by CT). ²⁷ Planimetry is often challenging in patients with CKD due to extensive cusp calcification and needs to be performed with careful attention. ⁷

Role of cardiac MRI

Cardiac MRI (CMR) is another noninvasive imaging modality that can provide complimentary information when evaluating a patient with AS. CMR provides a thorough functional evaluation and importantly, myocardial tissue characterization. ²⁸ CMR can assess LV mass and volume, as well as patterns of LV remodeling with greater precision than TTE. In the setting of AS, LV remodeling is a compensatory process that eventually progresses to a maladaptive response with myocyte hypertrophy, interstitial fibrosis, and apoptosis. ²⁹ Patients with CKD, even in the absence of AS, have evidence of LV hypertrophy due to systemic hypertension and maladaptive aldosterone regulation; the degree of LV remodeling and hypertrophy do not closely correlate with the severity of AV stenosis. ³⁰ In the setting of transcatheter aortic valve replacement (TAVR) assessment, reliable measurements can be obtained for the aortic annulus without the use of contrast, a point of interest in patients with relative contraindications to coronary CT angiography (CCTA). Cardiac MRI can be used to attain a ‘hybrid AVA’; however, it is not first line in the assessment of AS due to the time-consuming nature of the assessment and dependance of previously attained TTE parameters to accurately calibrate the cine velocity.^8,23 The versatility of CMR additionally allows for concomitant evaluation of dimensions of the entire aorta and direct quantitation of aortic insufficiency.

Myocardial fibrosis is a hallmark of severe AS and offers an important prognostic role. Focal LV fibrosis in a non-infarct pattern is independently associated with AS mortality. ³¹ Late gadolinium enhancement imaging (LGE) is the mainstay of CMRI assessment for myocardial fibrosis and infarct. ³² Until recently, CKD G4-G5D was a contraindication to gadolinium-based contrast media (GBCM) due to risk of nephrogenic systemic fibrosis (NSF) with group I agents. According to the most recent recommendations from the Canadian Association of Radiologists Group II and Group III GBCMs are associated with an exceedingly low (<0.07%) risk for nephrogenic systemic fibrosis.^32,33 A position statement by the American College of Radiology and the National Kidney Foundation advises that the risk of NSF is so low with group II GBCM, the potential benefits outweigh risks. ³⁴ Native T1 mapping is a CMRI technique that can detect diffuse (as opposed to focal) myocardial fibrosis without the use of gadolinium. Higher myocardial T1 values in patients with AS have been associated with an increased risk of all-cause death and heart failure hospitalizations. A recent systematic review and meta-analysis explored the utilization of CMR to detect fibrotic changes in patients with CKD/ESKD with preserved left ventricular ejection fraction and found that they have higher T1 values, indicating fibrosis. ³⁵ However, normal and pathologic T1 values in CKD/ESKD patients with AS have not been well established. Further studies are needed to determine normative values for T1 mapping for risk stratification in patients with AS and CKD, as well as evaluate the incremental utility of CMRI specifically in patients with advanced CKD.

Imaging prior to AVR

Obtaining a gated cardiac CT is standard practice for all patients anticipated to undergo TAVR for precise annular sizing to prevent immediate procedural and/or long-term complications (such as patient-prosthesis mismatch). This typically involves electrocardiographically synchronized (gated) evaluation of the aortic root, followed by non-gated acquisition of the aorto/ilio/femoral arterial tree. The severity of coronary artery disease (CAD) is also routinely evaluated prior to anticipated aortic valve replacement (AVR). Although typically accomplished with invasive angiography; assessment can also be performed with a high-quality CCTA. ³⁶ Mitigating the risk of periprocedural AKI is an important priority; it is important to be mindful to limit the use of contrast especially in patients with CKD, who have a higher risk of developing contrast induced nephropathy. In addition, it is important to ensure adequate hydration prior to dye exposure to minimize the risk of acute kidney injury (AKI). We await the results of the ENRICH trial that will compare the effectiveness of oral versus intravenous hydration to prevent AKI. ³⁷ Among patients with CKD, novel techniques such as low contrast volume protocol, transesophageal echocardiography for valve sizing and zero contrast or non-contrast TAVR have been described to reduce the risk of AKI.^38–40 It should be noted that the presence of a porcelain aorta is a significant risk factor for stroke in those undergoing surgical aortic valve replacement (SAVR); the diagnosis of a porcelain aorta can be established on a non-contrast CT.

Imaging following AVR

The typical recommendation is to obtain a baseline TTE immediately post-implantation following SAVR and TAVR to evaluate valve hemodynamics and ventricular function. ⁶ After bioprosthetic SAVR, TTE at 5-10 years is recommended, and yearly subsequently, to monitor for structural valvular deterioration (SVD). ⁶ High-risk populations such as those with CKD/ESRD may experience accelerated bioprosthetic valve degeneration occurring <5 years following implantation; hence a selective adoption of an earlier annual TTE screening may deserve consideration. ⁶ Earlier repeat TTE is further recommended when a change in clinical status occurs or if the patient has other risk factors associated with accelerated valve deterioration, including young age (<60 years) at implantation, smoking, diabetes mellitus, initial mean gradient ≥15 mm Hg, and certain valve types. In patients with TAVR, the recommendation is to obtain TTE immediately following implantation (before discharge or within 30 days) for establishing a baseline, at 1 year following implantation, and annually thereafter. ⁴¹ Particular attention should be focused on detection of SVD as well as other forms of bioprosthetic valve dysfunction (including non-SVD, thrombosis and endocarditis). Patients with younger age, lower EF, those not on dual antiplatelet therapy are at higher risk of SVD following TAVR. ⁴²

Management of aortic stenosis in CKD

Role of nephrology consultation

Interestingly, despite the consistent observation of more rapid progression of AS in patients with advancing CKD, the specific components that are responsible for this finding are not clear. Nevertheless, risk factors from the general population provide suggestions of areas of support that can be targeted by the nephrologist co-managing patients with AS. Nephrologists are centrally involved in many of the contributing components to the development or progression of AS. First and foremost is blood pressure control. Hypertension is a frequent finding in CKD and is an important contributor towards AS progression. A prospective analysis of patients with mild or moderate AS demonstrated that systolic hypertension was associated with a faster progression of AV calcification. ⁴³ This is particularly important for patients with CKD given the frequent occurrence of calcification. It is not clear whether more aggressive treatment of hypertension would retard progression of AS or improve outcomes in CKD. Agents preferentially used in hypertension control in CKD have variably been shown to be beneficial in AS too, including angiotensin converting enzyme inhibitors (ACEI) and angiotensin receptor blockers (ARB). The specific targets for hypertension management have not been identified in AS. ⁴⁴

As mentioned, calcification is a prime driver of AV deterioration, and this can be accelerated among patients with advanced CKD. ⁴⁵ The pathophysiology of calcification is complex and is in part related to dysregulation of mineral bone metabolism in CKD. This includes impaired phosphorus excretion, lower vitamin D production with resultant elevations in parathyroid hormone levels and demineralization of bone with subsequent liberalization of calcium. This pathophysiology is considered under the umbrella of CKD related metabolic bone disease (MBD). Early studies suggested that medications that potently suppress parathyroid hormone can lower AV calcification. ⁴⁶ More recently, chelation agents for removal of excess calcium deposits have demonstrated some early safety and efficacy data for reducing AV calcification; further definitive studies are anticipated. ⁴⁷

Preventing or slowing the decline in kidney function remains the primary goal of CKD management, and is a prime area of involvement of the nephrology collaborator in the heart-kidney team among patients with AS. The armementarium of agents that help preserve kidney function has grown over the last decade. Traditionally, renin angiotensin system inhibition (RASi) with ACEI/ARB formed the mainstay of therapy. ⁴⁸ In the contemporary era, the available agents for preventing CKD progression have expanded to include sodium glucose cotransporter 2 inhibitors (SGLT2i) and glucagon like peptide-1 receptor agonists (GLP-1RA), which have potent renal preservation characteristics.^49–52 ACEI/ARB are of particular interest as initial concern of potential detrimental hemodynamic effects in the presence of AS were raised. However, studies have demonstrated an overall beneficial effect of RASi on clinical outcomes in AS, independent of renal function. ⁵³ The influence of SGLT2i and GLP-1RA on AS outcomes in CKD has not been extensively evaluated in clinical studies. Preliminary studies suggest potential beneficial role of SGLT2i in reduction in AV calcification; ⁵⁴ further clinical studies are needed to validate these findings.

Finally, management of patients on dialysis with moderate to severe AS poses particular challenges. In addition to the medical management considerations mentioned, volume shifts have to be carefully considered. Peritoneal dialysis (PD) has qualities that may make it a more favorable kidney replacement modality in patients with hemodynamically significant AS for several reasons. First, PD results in the avoidance of AVF or arteriovenous grafts required for hemodialysis, and the associated alterations in CO/SV that ensue. Second, PD affords the ability to have lower fluid removal goals, thus decreasing the potential for hypotension induced exacerbation of systemic hypoperfusion in the setting of AS. ⁵⁵ These are, of course, challenging decisions to make if patient is already receiving hemodialysis but can be discussed with patient if symptoms worsen while on hemodialysis. Ultimately, shared decision making is required between cardiology, the patient/family/friends, and the nephrologist in deciding on the ideal dialysis modality while awaiting decisions on more definitive AS management.

Aortic valve replacement in CKD

Observations from a large cohort revealed that despite the presence of symptoms, AVR occurred less frequently in those with moderate-severe CKD. ⁵ It is unclear whether this observation reflects therapeutic nihilism (or “renalism”) on the part of clinicians ⁵⁶ or simply a reflection of how daunting this clinical combination of CKD and AS is. Yet, in this retrospective analysis, the performance of AVR was also associated with a significant reduction in all-cause and cardiovascular mortality as compared to conservative management only. ⁵ In fact, AVR was associated with better survival compared to conservative management at each stage of CKD.

Observational studies have demonstrated the higher risks associated with AVR among patients with advanced CKD and ESKD. In a study of the National Inpatient Sample of TAVRs performed 2016-2020, patients with advanced CKD (29.2% CKD, 3.7% ESKD) had higher in-hospital mortality, higher risk of cardiogenic shock, need for mechanical circulatory support, and vascular access complications. ⁵⁷ Patients with ESKD also had a higher risk of cardiac arrest and periprocedural myocardial infarction. Severe CKD has also been identified as an independent predictor of lack of functional improvement following TAVR. ⁵⁸ These observations outline the quagmire that clinicians face in the management of AS – although advanced CKD is associated with higher risks of mortality and higher procedural risks of AVR compared to non-CKD, AVR is associated with improved survival compared to medical management alone.

Choice of aortic valve prosthesis for AVR in CKD/ESKD

The landscape for AVR has evolved markedly after pivotal randomized clinical trials established the non-inferiority of TAVR compared to SAVR in several populations. With more widespread adoption of TAVR, isolated SAVR volumes decreased in the US, but interestingly, the observed mortality rates with SAVR also declined in the hospitals performing the highest number of TAVRs. ⁵⁹ Existing data on management of severe AS in advanced CKD/ESKD predominantly stem from retrospective studies comparing SAVR versus TAVR (Table 1). To date, no clinical trial has directly compared mechanical versus bioprosthetic valves or TAVR versus SAVR specifically in CKD/ESKD patients, though earlier pivotal trials have identified CKD/ESKD as a significant determinant of outcomes.

Table 1.

Summary table of key studies that compared transcatheter versus surgical aortic valve replacement in chronic kidney disease and end-stage kidney disease.

Author and year of publication	Subjects	Study design and sample size	Major outcomes
Nguyen TC et al, 60	Patients with GFR greater than 60 mL/min were compared to those with GFR 31-60 mL/min and GFR 30 mL/min or less.	Retrospective, observational (N = 1336)	Similar rates of in-hospital mortality rates for TAVR and SAVR (3.5% vs 4.1%, P = .6). Worsening preoperative renal failure associated with in-hospital mortality for SAVR but not TAVR.
D'Errigo P et al, 61	CKD stages 3b to 5	Retrospective (170 propensity matched pairs of CKD)	In-hospital mortality is similar between TAVR and SAVR (7.1% vs 2.9%, P = .09). At 2 years TAVR compared with SAVR: - All-cause mortality similar (31.2% vs 23.4%; P = .118) - Major cardiac and cerebrovascular events (37.2% vs 31%; P = .27) - Risk of dialysis lower (12.4% vs 21.2%; P = .052)
Doshi R et al, 62	Advanced kidney disease (CKD stages IV, V and ESKD)	Retrospective propensity matched (N = 2485 in each group)	SAVR versus TAVR had higher: - In-hospital mortality (12.9% vs 6.2%; P < .01) - Acute kidney injury (50.3% vs 33%; P < .01) - Dialysis requirements (26.8% vs 20.1%; P < .01)
Kumar N et al 63	Any diagnosis of CKD	Retrospective propensity-matched (N = 6874; 2820 TAVR, 4054 SAVR)	TAVR versus SAVR was associated with lower: - In-hospital mortality (OR 0.47, 95% CI 0.32-0.69, P < .001), - Rates of AKI (OR 0.18, 95% CI 0.14 to 0.22, P < .001) - Dialysis-requiring AKI (OR 0.30, 95% CI 0.20 to 0.44, P < .001) - Postoperative stroke (OR 0.27, 95% CI 0.13 to 0.53, P < .001).
Shavit L et al, 64	Any diagnosis of CKD	Prospective, observational. (N = 318; 119 TAVR, 199 SAVR)	SAVR versus TAVR had higher in-hospital mortality (OR 5.9; 95% CI 1.6-29.6, P = .02) and postoperative infection (OR 4.2, P = .005). Mortality at 1 and 2 years was similar in SAVR versus TAVR groups (P = .059).
Pineda AM et al, 65	Mild, moderate, and severe CKD except eGFR of ≤20 mL/min/1.73 m2	Posthoc analysis of CoreValve US randomized trial. (N = 797; 391 TAVR, 359 SAVR).	TAVR versus SAVR had a lower incidence of: - MACRE at 3 years (42.1% vs 51.0%, P = .04). - AKI (9.6% vs 18.2%, P = .001) - Life-threatening bleeding (20.7% vs 46.4%, P < .001).
Cheng X et al, 66	CKD	Metaanalysis of 10 observational studies	TAVR versus SAVR had lower risk of early all-cause mortality (6.1% vs 102%; OR 0.71, 95% CI 0.51-0.98) and stroke (1.1% vs 2.2%; OR 0.53, 95% CI 0.37-0.75). TAVR increased the risk of pacemaker implantation (OR: 2.06; 95% CI: 1.16-3.66), but reduced the risk of blood transfusion, AKI and AKI requiring dialysis.
Garcia S et al 67	Moderate to severe CKD (estimated GFR <60 mL/min/m2)	Posthoc analysis of the randomized PARTNER 2A trial and SAPIEN3 registry. (N = 1045 TAVR:512 SAPIEN XT, 533 SAPIEN 3 and 479 SAVR).	SAVR group had better composite MACRE outcome than SAPIEN XT TAVR group (52.8% vs 68.0%, HR 1.2; 95% CI: 1.01-1.44; P = .04). Composite outcomes for SAVR and SAPIEN 3 TAVR were similar (52.8% vs 58.7%, respectively; HR: 1.01; 95% CI: 0.84-1.22; P = .89).
Wei S et al, 68	Advanced CKD (stage 3-5)	Systematic review and metaanalysis; January 2000-October 2020	TAVR versus SAVR had lower risks of in-hospital mortality (OR 0.53; 95% CI: 0.36-0.78; P = .001], stroke (OR: 0.68; 95% CI: 0.47-0.96; P = .03), AKI (OR: 0.42; 95% CI: 0.34-0.52; P < .00001), AKI requiring dialysis (OR: 0.65; 95% CI: 0.58-0.73; P < .00001), bleeding (OR: 0.35; 95% CI: 0.31-0.38; P < .00001)
Mas-Peiro S et al, 69	Moderate to severe CKD (eGFR 15-60 mL/min/1.73 m2) eligible for either procedure	Observational registry data. N = 1078 (704 TAVR, 374 SAVR)	1 year survival was similar between TAVR and SAVR (CI 0.795-2.031, P = .316).
STUDIES EVALUATING TAVR VS. SAVR AMONG PATIENTS WITH ESKD
Condado JF et al, 70	ESKD on dialysis undergoing balloon aortic valvuloplasty (BAV), TAVR or SAVR	Retrospective observational (N = 85)	No significant difference in 30-day mortality between BAV, TAVR or SAVR (16.7% vs 10% vs 10%, P = .74). BAV treated patients had increased 1-year mortality compared to TAVR or SAVR (87.0% vs 32.0%, vs 36.7%, P = <.001).
Bhise, V et al, 71	ESKD on dialysis	Retrospective propensity matched (TAVR 119 vs SAVR 244)	No significant difference in in-hospital mortality between TAVR and SAVR groups post propensity match (9.2 vs 10.9, P = .784)
Alqahtani F et al, 72	ESKD on dialysis	Retrospective propensity matched (197 matched pairs)	Higher in-hospital mortality with SAVR versus TAVR (13.7% vs 6.1%, P = .021). Higher permanent pacemaker implantation with TAVR (13.2% vs 5.6%, P = .012) but lesser blood transfusion rates (43.7% vs 56.8%, P = .02).
Alkhalil A et al 73	ESKD on dialysis	Retrospective analysis with propensity matching (175 matched pairs).	In-hospital mortality similar between TAVR and SAVR (8% vs 10.3%, P = .58). TAVR associated with lower in-hospital complications (60.6% vs 76%, P = .003).
Khan et al, 74	ESKD on dialysis	Retrospective analysis with propensity matching. (N = 12 550; TAVR 5735, SAVR 6815).	In-hospital mortality higher with SAVR versus TAVR (13.9% vs 4.1%, P < .001) in unmatched cohorts. Except permanent pacemaker implantation, in-hospital complications higher with SAVR versus TAVR.
Färber G et al, 75	ESKD on dialysis	Retrospective analysis with propensity matching. (N = 1118; TAVR 661 or SAVR 457)	Mortality lower with TAVR versus SAVR at 30 days (8.6% vs 15.0%, P = .02. Mortality at 1 year was similar (TAVR: 33.4% vs SAVR 35%, P = .72).
Mentias et al, 76	ESKD on dialysis, Medicare beneficiaries	Retrospective propensity matched cohort. (N = 8107)	TAVR when compared with SAVR had a lower incidence of 30-day mortality (4.6% vs 12.8%, P < .01) and 30-day stroke (2.6% vs 3.7%, P = .02). At 1-year, there was no difference in adjusted mortality between the 2 groups (28.1% vs 31%, P = .1).
Ogami et al, 77	Dialysis dependent patients older than 70 years of age from the United States Renal Data System.	Retrospective registry data. (N = 4332; SAVR, 1280 TAVR).	In-hospital mortality lower for TAVR versus SAVR (OR 0.38, 95% CI 0.27-0.52; P < .001). TAVR had significantly higher risk of 3-year mortality versus SAVR (HR 1.24, 95% CI 1.01-1.39; P < .001).

Abbreviations: TAVR: transcatheter aortic valve replacement; SAVR: surgical aortic valve replacement; CKD: chronic kidney disease; ESKD: end-stage kidney disease; AKI: acute kidney injury; eGFR: estimated glomerular filtration rate; MACRE: major adverse cardiovascular and renal events rate; OR: odds ratio; CI: confidence intervals.

Choice of mechanical versus biological prosthesis

Notably, the choice between mechanical and bioprosthetic valves in CKD patients is a pivotal first step in assessment. In a large observational analysis of individuals undergoing AVR, receipt of a biological prosthesis was associated with a higher risk of mortality among those between 45-54 years of age, but not in those between 55-64 years of age. ⁷⁸ As a corollary, among younger CKD patients with acceptable surgical risk, life expectancy beyond 8-10 years and lower bleeding risk, mechanical SAVR may be preferable due to longevity of the valve which in turn reduces the downstream risk of reoperation. ¹ Yet, the placement of a mechanical valve associates with the attendant risks of warfarin use, including increased risk of bleeding.^79,80 Calcific uremic arteriopathy represents an uncommon but severe adverse occurrence in individuals with ESKD undergoing treatment with warfarin, markedly influencing both overall survival and quality of life.^81,82 Bioprosthetic valves are beneficial in regards to lower risk of bleeding related to warfarin, but have a higher risk of accelerated SVD resulting in the need for reoperation among younger patients with a longer life expectancy. Yet, bioprosthetic valves (especially those greater than 21 mm in size) could offer the future option of a valve-in-valve intervention in the contemporary era. ⁸³

Bioprosthetic valves have limited long term durability due to being prone to SVD as opposed to mechanical valves. SVD of a bioprosthetic valve is an acquired abnormality defined as the deterioration of the leaflets or supporting structures that result in tearing, calcification, thickening, or disruption of prosthetic valve materials resulting in hemodynamic dysfunction with stenosis or regurgitation. ⁸⁴ Among explanted transcatheter heart valves (THV), a time dependent degenerative process has been described on histological analysis, comprising early thrombus formation, endothelial hyperplasia, fibrosis, tissue remodeling, expression of matrix metalloproteinase and overt calcification. ⁸⁵ As a corollary, sevelamer hydrochloride, a phosphate binder used commonly to treat hyperphosphatemia in CKD, has been shown to attenuate bioprosthetic heart valve deterioration in animal studies. ⁸⁶ CKD, particularly ESKD, is a well-known risk factor for accelerated SVD. ⁸⁴ A number of potential contributors can be hypothesized for this observation. Valvular calcification due to altered calcium phosphate homeostasis and mineral metabolism are likely significant contributing factors. ⁸⁷ CKD patients have a high co-prevalence of arterial hypertension and persistent left ventricular hypertrophy, which associates with accelerated SVD. ⁸⁸ Leaflet thrombosis and hypoattenuated leaflet thickening (HALT) are more prevalent in TAVR valves, potentially contributing to reduced durability.

In a retrospective study of ESKD patients on dialysis from the US Renal Data System between 1978-1998, 2-year survival was comparable between mechanical and bioprosthetic valves (∼40%), balancing the risks of bleeding versus SVD respectively. ⁸⁹ It was sobering to note that the 5-year survival in this cohort of patients undergoing SAVR was only ∼14%–15%. ⁸⁹ In a metaanalysis of 24 studies involving 10 164 participants on dialysis, patients undergoing SAVR with a mechanical prosthesis were demonstrated to have better long-term survival, but also higher risk of bleeding complications. ⁹⁰ Conversely, another metaanalysis involving 16 studies with 8483 patients showed no differences in survival with a higher risk of bleeding and thromboembolism with mechanical valves. ⁹¹ Given the ambiguity in available observational data, ACC/AHA guidelines provide no specific recommendations about preferred valve type in ESKD patients. ⁶ Some authors have also suggested using alternative surgical approaches such as the Ross (pulmonary autograft in the aortic position) and Ozaki (aortic valve reconstruction using autologous pericardium) procedures in this high-risk population. ⁹²

Choice of TAVR versus SAVR in CKD/ESKD

Several observational studies have evaluated outcomes of AVR among CKD patients with common themes (Table 1). Firstly, importantly but not surprisingly, patients with CKD have higher risks of in-hospital, 30-day mortality with TAVR and SAVR compared to patients without CKD.^93,94 In retrospective subgroup analysis of randomized data as well as observational data, MACRE (in-hospital mortality, major bleeding, stroke, AKI) were lower with TAVR compared to SAVR, whereas rates of permanent pacemaker implantation were higher.^62,63,65 A comprehensive meta-analysis of previous retrospective observational studies comparing TAVR versus SAVR among CKD patients revealed notably higher rates of early mortality, stroke, and AKI in the SAVR group, whereas the TAVR cohort exhibited increased post-procedural pacemaker implantation rates. ⁶⁶ In regards to interpretation of these studies, two important factors need to be considered. Since these were retrospective analysis, the effects of unmeasured confounders and selection bias cannot be excluded; hence a randomized trial would be most appropriate for more conclusive guidance. Also, long-term data regarding durability of TAVR versus SAVR prosthesis are not available from these data.

Clinical evidence regarding incidence of SVD in THV among CKD patients is scant. In a pooled retrospective study of the PARTNER 2A trial (patients randomly assigned to SAVR vs SAPIEN XT TAVR) and the SAPIEN 3 registry, an evaluation for SVD was performed for patients with moderate to severe CKD. ⁶⁷ At a follow-up interval of 5 years, the composite primary endpoint of MACRE (death, stroke, rehospitalization and new hemodialysis) was significantly higher in SAPIEN XT TAVR versus SAVR (68% vs 52.8%, P = .04) but similar between SAVR and SAPIEN3 TAVR. Moreover, bioprosthetic valve failure and advanced SVD were similar for SAVR and SAPIEN 3 TAVR, but higher for SAPIEN XT TAVR. ⁶⁷ These findings demonstrate important differences in long-term durability based on generation of TAVR valves, and also emphasize that long-term follow-up data continue to be needed to guide clinicians in their choice of most optimal valve in patients with moderate to severe CKD.

Notably, the recent expansion of TAVR usage has also encompassed patients with ESKD undergoing hemodialysis. The management of AS among patients with ESKD on dialysis is particularly challenging, since outcomes following SAVR and TAVR are significantly worse in this population and there is a significant competing risk of all-cause mortality. Ultimately, in the absence of randomized controlled trials, clinicians have to rely on these observational data to make individualized decisions for this complex population. Several retrospective studies have evaluated outcomes of TAVR versus SAVR among dialysis patients (Table 1). In a retrospective analysis of large cohort of Medicare recipients with ESKD on hemodialysis with AS, 50% underwent TAVR, 31.6% SAVR and 17.4% were treated medically. Despite higher frailty in TAVR patients, 30-day mortality was lower with TAVR compared to SAVR (4.6% vs 12.8%). However, at 1 year, there was no difference in adjusted mortality between the 2 groups (28.1% vs 31.0%, P = .1). These data highlight not only the higher upfront risk of SAVR but also the exceedingly high 1-year mortality in this population. In an analysis of 72 631 patients from the STS/TVT registry treated with TAVR, 4.2% patients had ESKD; these were younger than non-dialysis patients but with higher STS predictive risk scores for mortality. Patients with ESKD had a higher risk of in-hospital (5.1% vs 3.4%) and 1 year mortality (36.8% vs 18.7%) as well as major bleeding (1.4% vs 1%). ⁹⁵ In view of the exceedingly high (∼37%) 1-year risk of mortality, the appropriate selection of dialysis patients that might improve after TAVR is critical. Experts recommend frank discussion with the patient regarding anticipated risks and potential alternatives, including the option of expectant or palliative management.. ⁹⁶ The American Heart Association (AHA) statement recommended consideration of palliative medicine consultation in patients with ESKD due to high risk of 1 year mortality. ¹

Another key consideration among patients with CKD undergoing AVR is the presence of concomitant mitral stenosis (MS). As outlined above, the presence of altered calcium, phosphorus and parathyroid metabolism can predispose to presence of significant coexisting mitral annular calcification also in addition to AS, resulting in functional mitral stenosis. In a study of patients with MS using data from the National Inpatient Sample (NIS), TAVR was associated with a lower risk of in-hospital mortality compared to SAVR. ⁹⁷ The need for concomitant mitral valve surgery with SAVR markedly increases the perioperative surgical risk.

AKI following AVR

The presence of AS in CKD has been described as a “progressive, intertwined death spiral of both the heart and the kidneys” since worsening CKD predisposes to worsening AS and worsening AS could contribute to increased venous congestion, reduced CO and renal perfusion, in turn worsening kidney function. ⁹⁸ Lower estimated glomerular filtration rate (eGFR) is associated with markers of high intracardiac filling pressures and pulmonary vascular resistance; thus cementing our understanding that high right atrial pressure/venous congestion or renal venous hypertension is a predictor and driver of renal dysfunction. ⁹⁹ Thus, cardiorenal syndrome has been implicated as a putative pathophysiological contributor towards the occurrence of AKI post TAVR.

In this context, it is posited that correction of AS could arrest the adverse hemodynamic consequences of the downward spiral of cardiorenal syndrome.^98,100 In a retrospective analysis of TAVR treated patients from the PARTNER trials (prevalent CKD stage >2 in 91%), CKD stage either improved or remained unchanged in the majority of patients, whereas progression to post-TAVR dialysis occurred in 2% of the population. ¹⁰¹ In a cohort study of 894 patients undergoing TAVR, AKI occurred in 11.1% of patients within 48 h and largely resolved by discharge. ¹⁰² Reassuringly, greater than 80% of patients in this cohort had stable or improved kidney function 1 month following TAVR, and those with improved kidney function at 1 month had 2-year mortality comparable to those with eGFR > 60 mL/min/1.73 m². ¹⁰² An analysis of the UK-TAVI registry demonstrated that the proportion of new patients needing dialysis (median follow-up of 625 days) fell from 6.1% in 2007-2008 to 2.3% in 2013-2014. ¹⁰³

By the same token, it has been clearly demonstrated that the development of AKI is certainly not benign. In a study of 503 patients undergoing AVR (72% SAVR, 28% TAVR), eGFR obtained prior to preprocedural cardiac catheterization was an independent predictor of mortality at a median follow-up duration of 1348 days post AVR. ⁹⁹ All-cause mortality nearly 3.7 years post AVR was greater than 2-fold higher in the first preprocedural eGFR quartile compared to other quartiles. This remarkable observation outlines the importance of eGFR as a long-term prognosticator post AVR. In another cohort, deterioration of eGFR by over 10% at 1 month following TAVR was associated with a 2-fold increased hazard of mortality at 2 years. ¹⁰² In a study of the NIS, the development of AKI and AKI needing dialysis associated with a several-fold higher risk of in-hospital mortality. ¹⁰⁴ Requirement of new dialysis post TAVR was associated with markedly higher hazards of in-hospital mortality (hazard ratio [HR] 6.66; CI 4.87-8.53) and at 4 years (HR 3.54, CI 2.99-4.19). ¹⁰³

In a metaanalysis including 19 954 patients, incidence of AKI was lower following TAVR versus SAVR (7.1% vs 12.1%) but the incidence of AKI needing dialysis was similar. ¹⁰⁵ When only low-intermediate risk patients were included, TAVR was associated with a lower risk of incident AKI as well as dialysis-requiring AKI. The pathogenesis of AKI post TAVR involves different risk factors, including demographic, intraoperative, and technique related factors including use of iodinated contrast (risk of contrast nephropathy), rapid pacing (risk of reduced CO reducing renal perfusion) and atheroembolism during valve deployment. ¹ Patients with higher CKD stages (ie with lower eGFR) are at higher risk of AKI post TAVR. Hypertension, high STS score, diabetes, and COPD has also been shown to increase the risk of AKI. ¹⁰⁶ Percutaneous coronary intervention prior to TAVR could potentially contribute to AKI. Similar to TAVR, pre-existing risk factors such as hypertension, diabetes, obesity, and CKD, are associated with AKI post SAVR. Intra-operative factors like renal blood flow perturbations during cardiopulmonary bypass, aortic cross-clamping, systemic inflammatory response post cardiac surgery, post-operative blood transfusions, and exogenous vasopressors contribute to AKI risk after SAVR. ¹⁰⁷

Heart-kidney team-based approach to management of AS in CKD/ESKD

Ultimately, the pragmatic choice of prosthesis among patients with severe AS in the context of CKD involves complex considerations with inherent tradeoffs. The AHA scientific statement recommends the use of a heart-kidney multidisciplinary team-based approach to selection of prosthetic valves in the management of these high-risk patients. ¹ The choice of valvular prosthesis is nuanced with multiple considerations that need to be balanced including age, risk of bleeding versus reoperation, life expectancy, need for dialysis and feasibility of valve in valve procedures, among others (Central Illustration). ¹ The heart-kidney team including the structural cardiologist, cardiovascular surgeon, nephrologist must carefully consider risks versus benefits factoring in several variables to assess perioperative risk including patient age, life expectancy, comorbidities, anticoagulation-related risks, and the presence of ESKD necessitating hemodialysis. Input from nephrology is particularly valuable in hemodynamic management, assessing the risk of bleeding as well as overall life expectancy and frailty, particularly among dialysis patients. Most importantly, given the complexities involved in the decision-making and lack of evidence from randomized data, it is critical to ensure we pursue a shared decision-making approach with the patient. Beyond the decision-making pertaining to the choice of mechanical versus bioprosthesis and TAVR versus SAVR, meticulous and immaculate periprocedural planning is crucial, since in-hospital and short-term mortality risk is significantly higher in this population. Planning pertaining to several periprocedural aspects is pertinent including hemodynamic and volume management, optimization of kidney function to prevent AKI and periprocedural dialysis needs and assessment of adequacy of dialysis access. ¹ The scientific community eagerly awaits the conduction of randomized rials to inform clinical decision-making in this high-risk population.