Changes in Glomerular Filtration Rate after Renal Revascularization Correlate with Microvascular Hemodynamics and Inflammation in Swine Renal Artery Stenosis

Abstract

Background

Selection of patients with renal artery stenosis (RAS) likely to improve glomerular filtration rate (GFR) after percutaneous transluminal renal angioplasty (PTRA) is difficult. We examined basal hemodynamic and inflammatory factors linked to improved stenotic kidney (STK) function after PTRA in swine RAS.

Methods and Results

Fifteen pigs after 6 weeks of hemodynamically-significant RAS were studied before and 4 weeks after technically-successful PTRA+stenting. STK and contralateral kidney (CLK) hemodynamics and function were evaluated by multi-detector computed-tomography before and after acetylcholine challenge. Single-kidney deoxyhemoglobin (R2*, reciprocal to blood relaxation) and energy-dependent tubular function were assessed using blood-oxygen-level-dependent magnetic resonance imaging (BOLD-MRI) before and after furosemide. Baseline renal vein (RV) and inferior vena cava (IVC) levels of inflammatory markers were measured, and their gradient and net release calculated. Baseline parameters were compared to normal (n=7) and sham RAS (n=7) pigs, and correlated with the change in STK-GFR after revascularization (ΔGFR). Four weeks after PTRA blood pressure was normalized in all animals, but STK-GFR improved in 10/15 (ΔGFR=+22.0±8.5mL/min). ΔGFR correlated inversely with basal STK-GFR, renal release of inflammatory markers, and medullary R2* response to furosemide, but directly with GFR response to acetylcholine. Basal CLK-GFR directly correlated with ΔGFR.

Conclusions

Low basal STK-GFR with preserved response to acetylcholine may predict benefit from revascularization in RAS, while renal inflammation and robust STK-R2* responses to furosemide (possibly reflecting avid tubular oxygen consumption) are associated with less favorable outcomes. These tools may be useful for identification of patients likely to improve renal function after revascularization.

Renal artery stenosis (RAS), an important cause of secondary hypertension in the elderly population¹, is frequently detected incidentally in patients with widespread peripheral artery disease or those undergoing coronary angiography^{2, 3}. RAS may ultimately lead to end-stage renal disease, and represents a sizeable fraction of new patients entering dialysis programs in the United-States⁴. Furthermore, patients with RAS have increased risk of death for cardiovascular events including myocardial infarction, stroke, and heart failure⁵.

Renal revascularization by percutaneous transluminal renal angioplasty (PTRA) and stenting has been recommended commonly for patients with RAS, but recovery of glomerular filtration rate (GFR) after revascularization is inconsistent. Recent prospective randomized controlled trials comparing PTRA combined with medical therapy to medical therapy alone identified no differences in blood pressure and renal function at follow-up between the two treatment groups^{6, 7}. This might reflect partly the difficulty in selecting patients whose renal function is likely to improve following revascularization. Considering the variable efficacy, cost, and potential complications of PTRA⁸, identification of good candidates for revascularization beforehand is critical.

Pre-treatment characteristics of the stenotic kidney (STK) likely determine its recovery potential better than the function of total renal mass, which the contralateral kidney (CLK) may mask. Renal flow and functional reserve (responses to vasoactive challenge) may provide indices of STK viability. The degree of histopathological STK tissue damage correlates with clinical outcomes in RAS patients⁹, and might be partially mediated by inflammation, which magnifies parenchymal damage in RAS¹⁰. Thus, assessment of post-stenotic kidney inflammation may constitute an important tool for evaluating RAS patients before vascular intervention.

The current study was therefore designed to test the hypothesis that STK inflammation and reduced functional reserve are associated with attenuated renal functional recovery after PTRA in a well-established swine model of RAS. For this purpose, we correlated basal hemodynamics, function, oxygenation, inflammation, and functional reserve in the STK and CLK with the change in STK-GFR after revascularization (ΔGFR).

Methods

Experimental design

Twenty-nine female domestic pigs (40–55 kg) were studied after approval of the institutional Animal Care and Use Committee. At baseline, we induced RAS in 22 animals, while 7 others (normal) underwent a sham procedure. Six weeks later, single-kidney function, oxygenation, and inflammatory marker release were measured, followed immediately by either a sham procedure (7 normal and 7 sham-RAS), or PTRA+stenting (15 RAS+PTRA). Single-kidney function was then measured again 4 weeks later.

Induction of RAS

Pigs were anesthetized with intramuscular telazol (5 mg/kg) and xylazine (2 mg/kg), intubated, and anesthesia maintained with intravenous ketamine (0.2 mg/kg/min) and xylazine (0.03 mg/kg/min). The left femoral artery was catheterized, followed by a heparin bolus (5000U). Under fluoroscopic guidance, a 7mm balloon catheter wrapped with an irritant copper-coil was placed into the proximal-middle right renal artery of RAS pigs and inflated to high pressure. The balloon was then deflated and removed, leaving the coil embedded in the vessel wall, as previously shown¹¹. Normal animals underwent a sham procedure. A telemetry transducer (Data Sciences International, Arden Hills, MN) was implanted in the femoral artery to continuously measure mean arterial pressure (MAP) until study completion. Animals were then allowed to recover.

Measurement of kidney oxygenation

Six weeks later, animals were similarly induced, and anesthesia maintained with inhaled 1–2% isoflurane. Blood-oxygen-level-dependent 3T magnetic resonance imaging (BOLD-MRI) scanning (Signa Echo Speed; GE Medical Systems, Milwaukee, WI) was performed in the normal, sham-RAS, and 9 of the 15 RAS+PTRA pigs, to assess renal oxygenation, as described^12–15. For data analysis (4–6 slices/kidney), regions of interest (ROI) were manually traced in the cortex and medulla, and their average MR signal computed. The BOLD signal (relaxivity index R2*) was then measured, which is directly proportional to the local concentration of deoxyhemoglobin¹². Following baseline acquisition, furosemide (0.05 ml/kg IV) was administered, flushed with saline, and BOLD measurements repeated 15 min later. The degree of change in R2* in response to furosemide was calculated. These pigs underwent sample collection, measurement of kidney function, and PTRA or sham 2–3 days later.

Sample collection

Prior to PTRA or sham, all pigs were again anesthetized, and under fluoroscopic guidance, catheters advanced into the STK and CLK renal vein (RV) to collect samples. Blood samples were also collected from the inferior vena cava (IVC) for measurement of plasma renin activity (PRA) and creatinine. Samples were centrifuged and plasma aliquots stored at −80°C until assay. IVC and RV levels of e-Selectin (Biotang, P4988), tumor necrosis-factor (TNF)-α (Invitorgen, Cat# KSC3011), interleukin (IL)-1β (R&D DY681), IL-10 (Invitrogen, Cat# KSC0101), IL-17 (Kingfisher Biotech, Cat# VS0260S-002), monocyte chemoattractant-protein (MCP-1), and interferon (IF)-γ (Kingfisher, VS0259S-002) were measured by ELISA. Then, given lack of arterio-venous gradient of inflammatory cytokines across the renal circulation^16,17(implying comparable arterial and IVC levels), we estimated cytokine gradient (RV-IVC) and net renal release (gradient × renal blood flow (RBF)) for each measured analyte, as recently shown¹⁷. In addition, urine samples were collected and micro-albumin concentration quantified by ELISA (Bethyl Laboratories, Texas).

Measurement of kidney function

Single-kidney volume, regional perfusion, RBF, and GFR were then evaluated using multi-detector computerized-tomography (MDCT, SOMOTOM Definition-64; Siemens, Forcheim, Germany), and again after a 10-min suprarenal arterial infusion of acetylcholine (Ach, 5 mg/kg/min). Ach induces endothelium-dependent microvascular vasodilation and diuresis, thereby increasing RBF and GFR¹⁸. MDCT images were analyzed with Analyze™ (Biomedical Imaging Resource, Mayo Clinic, MN). Tissue time-attenuation curves obtained in ROI selected from the aorta, renal cortex, and medulla were fitted by curve-fitting algorithms to obtain measures of renal function¹⁹. Cortical and medullary volumes were calculated by planimetry, and RBF as the sum of the products of cortical and medullary perfusions and volumes. GFR was calculated from the cortical proximal-tubular curve²⁰. The degree of stenosis was assessed as the decrease in renal arterial luminal diameter and area, at its most stenotic compared to a disease-free segment, in images acquired at 6mm slice thickness and 3mm overlap, reconstructed with BioF convolution kernel. The stenosis length was measured at high-magnification using a computer-caliper program.

Revascularization

Immediately after MDCT scanning, 15 RAS animals underwent PTRA+stenting using a balloon catheter mounted with a standard tantalum stent. The balloon was inflated to high pressure, and then deflated and removed, ^{11, 21, 22}. Renal artery patency was confirmed with repeat angiography. The other animals underwent a sham procedure with selective renal angiography.

Four weeks after revascularization, the pigs were similarly anesthetized, and MDCT studies repeated. We have previously demonstrated the reproducibility of single-kidney assessment of renal hemodynamics and function using MDCT over 1-month intervals¹⁸.

Statistical analysis

Statistical analysis was performed using JMP software package version 8.0 (SAS Institute, Cary, NC). We used the Shapiro-Wilk test to evaluate deviation from normality. Comparisons within groups were performed using paired Student’s t-test and among groups using ANOVA with Bonferroni correction and unpaired t-test. For data without a Gaussian distribution, comparisons within and among the groups used non-parametric tests (Wilcoxon and Kruskal Wallis, respectively). P-values <0.05 were considered significant.

Response to revascularization was evaluated by the change in MAP (ΔMAP) and STK-GFR (ΔGFR). Then, RAS+PTRA animals were classified as having improved (STK-GFR increased by ≥1ml/min or by>15% versus baseline), deteriorated (STK-GFR decreased by ≥1ml/min or by>15%), or remained stable (STK-GFR changed by <1ml/ml and <15%) GFR after revascularization, as described²³. A simple linear regression analysis was used to determine correlations between basal parameters and ΔGFR.

Results

Systemic baseline parameters and PTRA

Table 1 shows baseline parameters in normal, sham-RAS, and RAS+PTRA pigs with subsequent improved or deteriorated STK-GFR. Six weeks after induction, all RAS groups achieved similar hemodynamically significant decreases in the diameter and area of the stenotic renal artery (60–99%, p=0.94, ANOVA), and equivalent length suggested comparable stenosis morphology. Basal MAP and serum creatinine levels were similarly elevated in RAS pigs compared to normal, while urinary albumin and PRA were not different among the groups.

Table 1.

Baseline parameters (mean±SD) in normal, sham-treated pigs with unilateral renal artery stenosis (RAS), and RAS pigs that improved or deteriorated stenotic kidney glomerular filtration rate (STK-GFR) after revascularization.

	Normal	Sham RAS	Improved STK-GFR	Deteriorated STK-GFR
Body weight (Kg)	46.4±3.6	49.1±4.0	43.7±5.4	47.4±3.0
Decrease in lumen diameter (%)	0	69.7±14.7*	73.4±13.4*	68.1±5.0*
Area of stenosis (mm2)	0	89.8±10.2*	90.8±15.0*	89.9±11.1*
Length of stenosis (mm)	0	6.4 ±1.8*	6.6±2.0*	6.5±1.8*
Systolic blood pressure (mmHg)	113.0±8.2	144.6±24.2*	146.5±20.2*	156.2±4.5*
Diastolic blood pressure (mmHg)	85.3±6.9	116.0±24.1*	104.9±24.6*	118.4±20.8*
Mean Arterial Pressure (mmHg)	94.6±6.8	125.5±23.9*	118.7±22.4*	124.4±19.5*
Serum creatinine (mg/dl)	1.2±0.1	1.8±0.3*	1.8±0.2*	1.9±0.4*
Urinary albumin (µg/ml)	13.9±8.9	25.3±14.8	23.4±27.8	23.7±12.2
PRA IVC (ng/ml/hr)	0.26±0.09	0.22±0.21	0.33±0.49	0.25±0.20
PRA STK (ng/ml/hr)		0.27±0.10	0.24±0.11	0.22±0.17
PRA CLK (ng/ml/hr)		0.20±0.08	0.27±0.32	0.13±0.18

SD: standard deviation; PRA: plasma renin activity; IVC: inferior vena cava; STK: stenotic kidney; CLK: contralateral kidney.

^*p<0.05 vs. Normal,

^#p<0.05 vs. Sham RAS,

^†p<0.05 vs. Improved STK-GFR (none observed for # and † comparisons).

Subsequent PTRA was technically successful in all animals (0% stenosis 4 weeks later, Figure 1B). MAP remained elevated in sham-RAS (p<0.05 vs. normal), but decreased to baseline levels in PTRA-treated pigs (Figure 1C).

Figure 1.

3D computed-tomography image of the stenotic kidney (left) and renal angiography (right) showing peri-stenosis collateral formation (A, arrows), underscoring its chronic nature. Representative renal angiography from a pig with renal artery stenosis (RAS) showing successful restoration of renal artery patency after revascularization (B). Mean arterial pressure measured by telemetry similarly decreased in pigs with improved or deteriorated STK-GFR after percutaneous transluminal renal angioplasty (PTRA) (C). Single-kidney glomerular filtration rate (GFR) at baseline and 4 weeks after revascularization or sham (D).

Baseline STK-GFR was lower in sham-RAS (38.7±7.3ml/min) and RAS+PTRA (51.6±16.7ml/min) compared to normal. Ten of 15 pigs improved STK-GFR after PTRA (ΔGFR=+22.0±8.5ml/min, p<0.0001 vs. baseline), while in the other 5 GFR deteriorated (ΔGFR=−12.1±3.4ml/min, p=0.001 vs. baseline) (Figure 1D); none sustained stable STK-GFR. STK-GFR decreased in sham RAS pigs (ΔGFR=−12.4±4.3, p<0.05 vs. normal). No significant change in single-kidney GFR was observed in sham normal animals (ΔGFR=+2.6±1.3ml/min, p=0.36).

Renal function

Basal STK cortical volume, perfusion, and RBF were lower in RAS compared to normal, but not different among RAS groups (Table 2). They were also lower than their own CLK, except for cortical perfusion in pigs with deteriorated STK-GFR after PTRA. Notably, basal STK-GFR was lower in sham-RAS and pigs in which STK-GFR subsequently improved compared to those with subsequently deteriorated STK-GFR. Yet, those with subsequently improved STK-GFR showed more robust RBF and GFR responses to Ach compared to Sham-RAS and to those with subsequently deteriorated STK-GFR (Figure 2A–B). Contrarily, CLK-GFR was reduced in pigs with deteriorated STK-GFR compared to the other two RAS groups, but neither CLK-GFR nor RBF responses to Ach were different among the groups. Notably, Bonferroni–adjusted basal STK-RBF, STK-GFR, and their responses to Ach remained significantly different among these groups (all p<0.0001).

Table 2.

Single-kidney baseline renal functional characteristics and release of inflammatory mediators (mean±SD) in pigs with unilateral renal artery stenosis (RAS) before percutaneous transluminal renal angioplasty (PTRA) or sham.

	Normal	Sham RAS		Improved STK-GFR		Deteriorated STK-GFR
		STK	CLK	STK	CLK	STK	CLK
Renal function
Cortical Volume (cc)	91.4±8.6	48.4±13.8*	93.4±3.7‡	48.7±8.2*	91.0±12.3‡	50.5±12.4*	90.9±8.9‡
Medullary Volume (cc)	18.0±7.1	21.1±3.4	19.9±4.2	22.3±6.3	19.2±3.8	19.9±8.3	18.6±3.3
Cortical Perfusion (ml/min/cc tissue)	4.4±1.2	3.2±0.7*	4.4±0.8‡	3.4±0.4*	4.4±1.3‡	3.2±0.7*	4.1±1.7
Medullary Perfusion (ml/min/cc tissue)	3.5±1.4	2.4±0.5	2.7±0.5	2.4±1.5	3.6±1.6	3.3±1.3	4.8±2.5
§ RBF (ml/min)	595.9±125.3	370.8±34.7*	651.4±86.4‡	335.5±133.2*	673.9±56.8‡	357.3±73.6*	609.3±88.6‡
§ GFR (ml/min)	64.2±5.8	38.7±7.3*	63.8±2.4‡	37.5±9.7*	66.0±10.8‡	47.9±6.3*#†	53.2±3.8*#†‡
Renal oxygenation (BOLD-MRI)
Cortical R2* (1/s)	16.4±1.1	17.1±1.5	17.8±2.3	17.4±2.3	17.9±0.9	16.7±0.6	17.7±0.9
Medullary R2* (1/s)	20.1±1.7	27.2±2.4*	22.7±4.9‡	30.0±8.6*	23.0±2.8‡	26.7±7.3*	22.1±1.0‡
Inflammatory markers: net release
§ e-Selectin (µg/ml/min)	−7.5±4.4	6.7±3.5*	−3.9±1.0‡	2.5±3.9*#	−11.2±4.9#‡	9.0±2.1*†	−0.2±8.2†‡
§ TNF-α (µg/ml/min)	−4.6±2.0	15.5±5.0*	−9.3±9.2‡	8.3±13.4*#	−6.7±9.4‡	31.4±15.*†	−5.3±3.7‡
§ MCP-1 (µg/ml/min)	−12.5±9.4	34.5±11.0*	−3.3±3.4‡	8.3±6.9*#	−18.6±10.9#‡	37.9±25.3*†	−2.8±10.3†‡
IL-1β (µg/ml/min)	−1.7±0.8	2.2±1.0*	−4.6±2.4‡	3.4±2.8*	−1.8±7.2‡	4.6±2.2*	−3.8±5.7‡
IL-17 (µg/ml/min)	−0.8±0.9	0.7±1.8*	−0.5±0.8‡	0.4±0.4*	−0.2±0.1‡	0.5±0.4*	−0.3±0.1‡
IF-γ (pg/ml/min)	−5.1±7.5	3.8±7.1*	−3.9±3.5‡	1.3±1.6*	−3.0±1.7‡	3.1±4.4*	−3.8±2.3‡
IL-10 (µg/ml/min)	0.7±0.5	−0.9±0.8*	0.3±1.0‡	−0.3±0.4*#	0.4±0.8‡	−0.8±0.09*†	0.3±0.2‡

SD: standard deviation, STK: stenotic kidney, CLK: contralateral kidney, BOLD-MRI: blood-oxygen-level-dependent magnetic resonance imaging, TNF: tumor necrosis factor, MCP: monocyte-chemoattractant protein, IL: interleukin, IF: interferon, RBF: renal blood flow, GFR: glomerular filtration rate. Raw p-values:

^*p<0.05 vs. Normal,

^†p<0.05 vs. improved (between groups),

^#p<0.05 vs. Sham RAS,

^‡p<0.05 vs. STK (within groups).

^§p<0.003 (one-way ANOVA Bonferroni-adjusted).

Figure 2.

A: Renal blood flow (RBF, A) and glomerular filtration rate (GFR, B) responses to acetylcholine (Ach). Basal stenotic-kidney (STK)-GFR inversely correlated with ΔGFR (C), while basal contralateral-kidney (CLK)-GFR directly correlated with ΔGFR (D). RBF and GFR responses to Ach directly correlated with ΔGFR (E–F). The height of the bar represents the mean and the extension-line the standard deviation. Raw p-values: *p<0.05 vs. Normal, †p<0.05 vs. improved STK-GFR, #p<0.05 vs. Sham RAS, ‡p<0.05 vs. STK (within groups).

Correlations between baseline renal function and ΔGFR

ΔGFR correlated inversely with basal STK-GFR (Figure 2C), but directly with basal CLK-GFR (Figure 2D), STK-RBF and GFR responses to Ach (Figure 2E–F). No correlations were found between ΔGFR and either CLK-RBF or CLK-GFR responses to Ach (both, p>0.05).

Renal oxygenation and response to furosemide

STK and CLK cortical R2* values were similar in all groups. Medullary R2* values were higher in all STK compared to normal and CLK, suggesting decreased oxygenation, but did not differ among RAS groups (Table 2). Basal STK medullary R2* responses to furosemide were reduced in STK compared to normal and CLK, but were more attenuated in pigs with subsequent improved STK-GFR compared to sham RAS and to pigs with deteriorated STK-GFR after revascularization (p=0.03, Figure 3A–B). CLK medullary R2* values and their responses to furosemide were comparable between RAS groups. Furthermore, basal STK (but not CLK) medullary R2* response to furosemide inversely correlated with ΔGFR (Figure 3C).

Figure 3.

Representative BOLD-MRI images (reflecting the level of deoxyhemoglobin) from pigs with improved (left) or deteriorated (right) STK-GFR before (top) and after (bottom) intravenous furosemide. Hypoxic regions (red) decreased after furosemide in pigs with deteriorated, but remained unchanged in pigs that improved STK-GFR after revascularization. B: Medullary R2* response to furosemide. C: Basal stenotic kidney medullary R2* response to furosemide inversely correlated with ΔGFR (n=9). The height of the bar represents the mean and the extension-line the standard deviation. Raw p-values: *p<0.05 vs. Normal, †p<0.05 vs. improved STK-GFR, #p<0.05 vs. Sham RAS, ‡p<0.05 vs. STK (within groups).

Renal inflammation

STK net release of e-Selectin, TNF-α, MCP-1, IL-1 β, IL-17, and IF-γ were elevated, while release of the anti-inflammatory IL-10 was reduced in all RAS groups compared with normal and CLK (Table 2). Moreover, STK release of e-Selectin, TNF-α, and MCP-1 were lower and IL-10 higher in pigs with improved STK-GFR compared to sham-RAS and deteriorating RAS. CLK release of only e-Selectin and MCP-1 was higher in pigs with deteriorated STK-GFR after revascularization compared to those that improved. Bonfferoni-adjusted STK-net release of e-Selectin, TNF-α, and MCP-1 remained highly significant (all p<0.0001).

ΔGFR correlated inversely with basal STK release of e-Selectin, TNF-α, and MCP-1 (Figure 4A–C), directly with STK IL-10 release (Figure 4D), and inversely with CLK release of e-Selectin and MCP-1 (Figure 4E–F). Furthermore, STK-GFR responses to Ach inversely correlated with STK net release of TNF-α and MCP-1 (Figure 4G–H). However, no correlation was found between either STK or CLK release of IL-1β, IL-17 or IF-γ and ΔGFR.

Figure 4.

ΔGFR inversely correlated with stenotic kidney (STK) net release of the inflammatory biomarkers e-Selectin, tumor necrosis factor (TNF)-α, and monocyte-chemoattractant protein (MCP)-1 (A–C), but directly correlated with release of the anti-inflammatory IL-10 (D). Similarly, contralateral kidney (CLK) release of e-Selectin and MCP-1 inversely correlated with ΔGFR (E–F). STK release of TNF-α, and MCP-1inversely correlated with GFR responses to acetylcholine (Ach, G–H).

Finally, ΔMAP did not correlate with the stenosis characteristics, serum creatinine levels, or with STK or CLK basal hemodynamics, oxygenation, or inflammatory marker release.

Discussion

This study suggests that preserved endothelial function associated with lower basal STK-GFR may be linked to preserved recovery capacity of renal function after revascularization in RAS. Furthermore, lower release of inflammatory markers from both kidneys is associated with more favorable functional outcomes after PTRA. Contrarily, robust STK BOLD-MRI responses to furosemide (possibly reflecting ongoing tubular oxygen consumption) are associated with blunted improvement of renal function.

RAS remains an important cause of chronic renal disease for which therapeutic strategies remain controversial. In the current study, four weeks after revascularization blood pressure was restored to normal levels in all animals, whereas renal function was improved in only two thirds, underscoring the greater efficacy of PTRA to mitigate renovascular hypertension than STK dysfunction, as we previously showed^{11, 21}. This variability of renal response to PTRA in our otherwise homogeneous group of animals extends previous clinical observations that renal outcomes after revascularization differ between RAS patients with similar demographics and degree of stenosis^24–26. Similar heterogeneous adaptive reactions occur in pigs in development of collateral coronary circulation after experimental myocardial infarction²⁷.

Because revascularization often shows little additive benefit over medical therapy for improvement in renal function for entire groups of patients^{6, 7}, identification of basal parameters to predict which specific patients would respond to PTRA could be invaluable. Consequently, several clinical studies have sought predictive factors that could differentiate PTRA “responders” from “non-responders”. For instance, resistance index values <80 by Doppler ultrasound were associated with greater improvement in renal function after revascularization²⁶. In our study, morphological characteristics of the stenotic lesion (length, decrease in luminal diameter or area) did not distinguish pigs with different PTRA response, consistent with previous reports that parenchymal injury predicts PTRA success better than the degree of proximal narrowing²⁸. Indeed, selection of patients for revascularization based on angiographic parameters primarily does not guarantee improvement in renal function^{6, 7, 29–31}. A low³² or rapidly-declining GFR³³ before revascularization are associated with favorable renal outcomes in RAS patients. Alas, a recent meta-regression analysis of prospective studies could not identify consistent clinical predictors of favorable renal outcome after revascularization³⁴.

Because rapid deterioration in renal function³³ predicts favorable outcomes after PTRA, progressive parenchymal damage may reflect the hemodynamic significance of the stenosis. Our findings extend previous observations, and emphasize that while lower basal GFR is associated with a measurable improvement after revascularization, basal RBF and GFR reserve to Ach in these kidneys needs to be preserved. STK-RBF, STK-GFR, and their responses to Ach emerged as dominant hemodynamic determinants of ΔGFR. In agreement, attenuated endothelium-dependent vasodilation blunts improvement following percutaneous coronary interventions³⁵. In our study, medullary volume and perfusion did not correlate with renal response to PTRA, possibly due to compensatory mechanisms that ensure their preservation in chronic renovascular disease³⁶. Interestingly, STK-GFR, but not serum creatinine levels, negatively correlated with response to PTRA, implying greater sensitivity of selective STK assessment to ultimate recovery. This is likely secondary to the compensatory effect of the CLK, which was supported by a modest direct correlation between basal CLK-GFR and ΔGFR.

BOLD-MRI allows assessment of renal oxygen content in the cortex and medulla. Furthermore, their responses to furosemide offer the opportunity to assess the functional tubular mass, as blunted changes in medullary R2* imply reduced prevailing oxygen consumption related to tubular sodium transport. We observed that basal intra-renal oxygenation (R2*) did not predict response to PTRA. Similarly, basal R2* was similar in human RAS kidneys that improved after PTRA and those that did not, although their ratio with STK-GFR better-predicted renal functional outcome, possibly because STK-GFR was somewhat lower in subjects with subsequently improved STK-GFR²³. Indeed, the combination of R2* and STK-GFR might be a powerful index of renal recovery potential. Interestingly, we found that blunted response to furosemide was associated with better renal function after intervention, as were robust responses to Ach. In some respects, the ability to vasodilate in response to acetylcholine appears contradictory to the reduced tubular furosemide response. This ostensible discrepancy between the responses to vascular and tubular functional challenges perhaps attributable to the decreased basal STK-GFR in pigs that subsequently improved after PTRA, which in turn decreased the basal filtrate volume and response to furosemide. This post-stenotic “hibernating” kidney³⁷ decreases its workload and oxygen consumption³⁸, thus adequate recruitment of this protective mechanism may portend better outcomes. Therefore, while blunted responses to Ach (endothelial-dysfunction) reflect microvascular injury and are associated with little benefit from revascularization, decreased medullary R2* responses to furosemide might imply decreased tubular oxygen consumption that preserved tubular viability.

Recent data indicate that renal inflammation is associated with deleterious processes in the stenotic kidney, like fibrosis and microvascular damage^{17, 39–41}. Indeed, our study illustrates that higher basal renal release of inflammatory markers was associated with poorer PTRA outcomes in RAS. STK-net release of e-Selectin, TNF-α, and MCP-1 as appear to be key determinants of improvement in GFR after revascularization. These cytokines, possibly secreted by inflammatory cells, may also impair endothelial functional reserve in the STK and its response to revascularization. In line with this postulation, we have shown that endothelial function improved in renovascular hypertensive pigs treated with an MCP-1 inhibitor¹⁰. Furthermore, IL-6 and TNF-α levels in RAS patients are elevated shortly after PTRA, but decrease 1 month later⁴². Importantly, our study showed an inverse relationship between release of pro-inflammatory markers and renal functional recovery, as well as with basal GFR responses to Ach. Therefore, inflammation may impair response to revascularization by ameliorating STK endothelial function and contribute to systemic inflammation by releasing inflammatory markers.

Importantly, both basal levels and responses to Ach of STK-RBF and STK-GFR, as well as STK-net release of e-Selectin, TNF-α, and MCP-1, remained significant predictors of ΔGFR upon Bonferroni adjustment, underscoring their predominant role as chief determinants of response to revascularization in swine ARAS.

Interestingly, CLK release of the inflammatory markers e-Selectin and MCP-1 also inversely correlated with ΔGFR. These findings extend previous studies in experimental⁴³ and clinical⁴⁴ RAS showing CLK inflammation and fibrosis. Furthermore, unilateral CLK nephrectomy is associated with post-atrophic regeneration of the STK in experimental RAS⁴⁵, underscoring the contribution of the non-stenotic kidney to the pathophysiology of RAS.

Our study is limited by the short duration of the disease, relatively small group of young animals, and lack of comorbidities like atherosclerosis, essential hypertension, or diabetes that might modify parenchymal injury and revascularization outcomes. Hence, PTRA was more successful in restoring blood pressure control in our pigs than typically achieved in humans. Moreover, the majority of RAS patients are treated with statins and/or blockers of the renin-angiotensin-system, which may modulate both blood pressure and renal response to PTRA. The absence of correlation between urinary albumin and ΔGFR in our study may be related to the relatively early stage of the disease. Also, we cannot rule out the possibility that some aspects of the stenoses that were not fully captured by our measurements contributed to differences between the groups. In addition, BOLD-MRI data were available only in 9 of the revascularized RAS animals. The statistical relationship between basal STK-GFR and its change in response to PTRA (ΔGFR) should be interpreted with caution, because it might reflect mathematical artifacts⁴⁶. Although these limitations may affect the translational potential of our observations, our study represents an important first step in identification of parameters to predict renal response to PTRA in human RAS. The extent of renal tissue injury and functional decline in our model are similar to human kidneys^{47, 48}. Furthermore, our assessment of single-kidney hemodynamics and function is clinically applicable⁴⁷, and provides a unique opportunity to evaluate parameters that may help predict response to PTRA in RAS. Future studies need to confirm these results in humans.

In conclusion, lower basal STK-GFR, “tubular hibernation” (decreased STK tubular oxygen consumption) reflected by diminished response to furosemide, and preserved endothelial functional reserve to acetylcholine may predict recovery of GFR after revascularization in RAS. Conversely, elevated release of inflammatory markers from both kidneys is associated with attenuated renal functional recovery after PTRA. These tools may be potentially clinically applicable for identification of patients likely to improve renal function after revascularization.

Bulleted Points.

What is Known?

• Renal revascularization often shows little additive benefit over medical therapy for improvement in renal function in patients with renal artery stenosis

• Identification of basal parameters to predict which specific patients would respond to revascularization could be invaluable

What this Article Adds?

• This study shows that low basal stenotic-kidney glomerular filtration rate with preserved response to acetylcholine may predict benefit from revascularization

• We also found that renal inflammation and robust stenotic kidney-R2* responses to furosemide (possibly reflecting avid tubular oxygen consumption) are associated with less favorable outcomes

• These tools may be potentially clinically applicable for identification of patients likely to improve renal function after revascularization.