A Higher Foci Density of Interstitial Fibrosis and Tubular Atrophy Predicts Progressive CKD after a Radical Nephrectomy for Tumor

Abstract

Significance Statement 

A higher percentage of the kidney cortex area that exhibits interstitial fibrosis/tubular atrophy (IF/TA) is a feature of poor prognosis on biopsy. This study describes IF/TA density (the number of IF/TA foci per area of cortex) and shows it predicts progressive CKD, independent of both percentage IF/TA and clinical characteristics among patients who have undergone nephrectomy. Other patterns of IF/TA and inflammation were not predictive for progressive CKD and mortality beyond percentage IF/TA and clinical characteristics. At the same percentage IF/TA, a higher IF/TA density suggests more scattered and widespread damage to the nephrons, with a worse kidney prognosis. This information may be helpful for accurate interpretation of the severity of “chronic changes” observed on a kidney biopsy.

Background 

Chronic tubulointerstitial injury on kidney biopsy is usually quantified by the percentage of cortex with interstitial fibrosis/tubular atrophy (IF/TA). Whether other patterns of IF/TA or inflammation in the tubulointerstitium have prognostic importance beyond percentage IF/TA is unclear.

Methods 

We obtained, stained, and digitally scanned full cortical thickness wedge sections of renal parenchyma from patients who underwent a radical nephrectomy for a tumor over 2000–2015, and morphometrically analyzed the tubulointerstitium of the cortex for percentage IF/TA, IF/TA density (foci per mm² cortex), percentage subcapsular IF/TA, striped IF/TA, percentage inflammation (both within and outside IF/TA regions), and percentage subcapsular inflammation. Patients were followed with visits every 6–12 months. Progressive CKD was defined as dialysis, kidney transplantation, or 40% decline from the postnephrectomy eGFR. Cox models assessed the risk of CKD or noncancer mortality with morphometric measures of tubulointerstitial injury after adjustment for the percentage IF/TA and clinical characteristics.

Results 

Among 936 patients (mean age, 64 years; postnephrectomy baseline eGFR, 48 ml/min per 1.73m²), 117 progressive CKD events and 183 noncancer deaths occurred over a median 6.4 years. Higher IF/TA density predicted both progressive CKD and noncancer mortality after adjustment for percentage IF/TA and predicted progressive CKD after further adjustment for clinical characteristics. Independent of percentage IF/TA, age, and sex, higher IF/TA density correlated with lower eGFR, smaller nonsclerosed glomeruli, more global glomerulosclerosis, and smaller total cortical volume.

Conclusions 

Higher density of IF/TA foci (a more scattered pattern with more and smaller foci) predicts higher risk of progressive CKD after radical nephrectomy compared with the same percentage of IF/TA but with fewer and larger foci.

Interstitial inflammation in kidneys serves as a protective response to disease and promotes healing and repair. Still, ongoing inflammation, regardless of the etiology, promotes progressive interstitial fibrosis and tubular atrophy (IF/TA), which in turn represents a common final pathway leading to ESKD.¹ Although the percentage of cortex that is IF/TA can be visually scored by renal pathologists, morphometric methods are more accurate and precise in some settings.^2,3

Several studies have reported the association of IF/TA and inflammation with outcomes in different patient populations. We previously found that IF/TA predicted a measured GFR <60 ml/min per 1/73 m² among living kidney donors, shortly after kidney donation.⁴ Other studies have reported that total interstitial inflammation and inflammation within fibrotic areas predicted graft loss after kidney transplantation.^5678–9 Conversely, inflammation outside fibrotic areas predicted disease progression in patients with IgA nephropathy and lupus nephritis.^10,11 In a study of patients with a wide array of kidney pathologies, higher amounts of IF/TA and inflammation within regions of IF/TA predicted a subsequent loss of kidney function.¹² All of these studies have relied on needle core biopsies, which have limited precision for assessing tubulointerstitial pathology.

Patients that undergo a radical nephrectomy due to renal cancer represent a unique opportunity for assessing biopsy-outcome associations in a population not selected on conditions that warrant a kidney biopsy. Radical nephrectomy specimens also allow sampling full thickness sections of renal cortex away from the tumor for more precise and detailed morphologic characterization, including tubulointerstitial pathology.^13,14 Recently, we reported that both nephron hypertrophy and nephrosclerosis, including IF/TA, predicted progressive CKD, and that IF/TA alone predicted noncancer mortality in patients after radical nephrectomy.¹⁴ However, this study and other past studies in patients who underwent nephrectomy^{151617181920–21} have only characterized IF/TA as a percentage of cortical involvement (%IF/TA), without consideration of its pattern of distribution within the cortex.

Different patterns of IF/TA have been reported. A striped pattern of IF/TA has been observed with calcineurin inhibitor use in transplanted kidneys.²² It has been hypothesized that subcapsular fibrosis may occur from ischemia to the superficial cortex regions (more distal blood supply), particularly in older kidneys.²³ In support of this, age-related glomerulosclerosis occurs primarily in the superficial cortex.¹³ A higher number of discrete IF/TA foci in the cortex may reflect a more scattered and less focal disease process. More IF/TA foci has been found to associate with older age, hypertension, and a smaller total kidney cortex to medulla ratio among living kidney donors.²⁴ A distinction between patchy (high density) and regional (low density) IF/TA has also been described in the setting of renal allograft biopsies.²⁵

Thus, we performed a study to characterize different patterns of IF/TA and inflammation on kidney wedge sections, to determine their association with kidney function and CKD risk factors, and to determine if they prognostically predicted progressive CKD or noncancer mortality. We also specifically sought to determine whether patterns of IF/TA and inflammation were informative beyond the commonly used %IF/TA.

Methods

Study Population

The study included 936 patients from a previously described cohort within the Aging Kidney Anatomy study.¹⁴ These patients underwent a radical nephrectomy for a renal tumor between 2000 and 2015, and had no metastatic lesions or positive lymph nodes at the time of surgery. To assess for the long-term risks, instead of immediate surgical complications, we set the baseline to 4 months after nephrectomy, as previously detailed.¹⁴ Patients without any serum creatinine testing between nephrectomy and subsequent 4 months were excluded. Patients with cancer recurrence, kidney failure, or death, within 4 months after surgery, were excluded. Patients with a diffuse specific kidney disease on histology (other than mild-moderate diabetic nephropathy), severe and diffuse tubulointerstitial inflammation (involving most the cortex), severe ischemia with %IF/TA >90% of cortex, a large focal scar, or end-stage kidney histology (thinned and nearly completely scarred cortex) were also excluded.¹⁴ This study was approved by the Mayo Clinic Institutional Review Board.

Kidney Function and Risk Factors

Demographic and clinical data were obtained from electronic medical records. Height, body mass index (BMI), serum creatinine (corrected to standardized values if assayed before standardization), and 24 hour urine protein were recorded. Hypertension and diabetes were identified as diagnoses in the medical record. The eGFR was determined with the CKD Epidemiology Collaboration equation.²⁶ The 24 hour urine protein excretion was estimated from spot urine protein-to-osmolality ratios.²⁷ For patients who did not have postnephrectomy 24 hour urine protein, we used prenephrectomy 24 hour urine protein values (if available within 1 year before surgery).¹⁴

Kidney Microstructure

After nephrectomy, kidney specimens were formalin fixed. Full cortical thickness wedge samples were taken from the renal parenchyma distant to the tumor location. After paraffin embedding, 3 µm thick sections were cut and stained with periodic acid–Schiff (PAS). Slides were scanned into high-resolution (40×) digital images (Aperio XT system scanner, Leica Microsystems, Inc., Buffalo Grove, IL; http:/www.aperio.com) and were magnified with Image Scope software (version 12.2.2.5015 Aperio) on a large touch-screen tablet. Biopsy images were analyzed by investigators that were blinded to clinical characteristics. As previously described, the cortex was outlined, and at high magnification, every glomerular profile was traced to calculate the mean nonsclerotic glomerular tuft volume (mean glomerular volume) using the Weibel–Gomez stereological model.^24,28 The number of globally sclerotic glomeruli were divided by the total number of glomeruli to calculate the percentage of globally sclerotic glomeruli. The percentage of luminal stenosis (from intimal thickening) was determined from the area of intima divided by the area of intima and lumen.²⁹ The mean percentage of luminal stenosis in the three most orthogonal small to medium-sized arteries on the biopsy section was used for analyses. Cortical thickness (capsule to medulla) was averaged between both ends of the wedge section.

IF/TA Measurements

A renal pathologist (M.P.A.) reviewed each high-resolution digital image, visually estimated the percentage of total cortex area that showed IF/TA, and then scored it in the following categories: <10%, 10%–25%, 26%–50%, and >50% IF/TA. A second renal pathologist (M.L.S.) also scored 50 random biopsy images. In addition, on the same high-resolution digital images we manually traced all IF/TA foci (each discrete contiguous region of IF/TA) within the cortex. To calculate %IF/TA, the total area of all traced IF/TA foci was divided by the traced cortical area.¹³ The IF/TA density was defined by the number of distinct traced IF/TA foci divided by the traced cortical area. Figure 1 shows examples of low versus high IF/TA density at a similar %IF/TA. The percentage of subcapsular IF/TA was calculated by dividing the total length of all regions of capsule adjacent to IF/TA by the total length of capsule (Figure 2A). Striped IF/TA was defined as IF/TA foci that extend through at least 50% of the cortical thickness perpendicular to the capsule (Figure 2B). The number of striped IF/TA foci was counted and divided by the capsule length (to standardize for wedge sections of varying sizes).

Figure 1.: Example of wedge biopsy sections from two patients with similar %IF/TA, but different IF/TA density. The green line outlines the cortical area. Black lines outline foci of IF/TA. (A) This patient had %IF/TA of 5.3% and IF/TA density of 16 IF/TA foci per cm² cortex. (B) This patient had %IF/TA of 5.6% and IF/TA density of 100 IF/TA foci per cm² cortex.

Figure 2.: Schematic example of how %subcapsular IF/TA, striped IF/TA, and inflammation measures were calculated. (A) Percent subcapsular IF/TA was obtained by dividing the length of cortex affected by IF/TA (two segments labeled by arrows) by total subcapsular cortex length (dashed black line). (B) All IF/TA foci that extend at more than 50% of mean cortical depth were counted and indexed to the length of capsule. This example shows six distinct foci, but only two were counted as striped IF/TA foci. (C) The percent of total interstitial inflammation was calculated by dividing the sum of all areas of inflammation (dark gray foci) by cortex area. The area of inflammation within the fibrotic areas (three black foci within light gray foci) was divided by the total IF/TA area (all gray shaded areas) to calculate %inflammation–IF/TA. The area of inflammation in the non-fibrotic (normal) cortex (two black foci outside of IF/TA) was divided by nonfibrotic cortex area (total cortex minus area of all IF/TA foci) to calculate %inflammation-outside–IF/TA.

Inflammation Measurements

All foci with cortical inflammation were also manually traced (Figure 2C). The total percentage of cortical interstitial inflammation (%inflammation) was calculated by dividing the sum of all areas of inflammation by cortex area. Then, the area of inflammation within the fibrotic areas was divided by the total IF/TA area to calculate %inflammation–IF/TA, and the area of inflammation in the nonfibrotic (normal) cortex was divided by the non-IF/TA cortex area to calculate %inflammation-outside–IF/TA. The percentage of subcapsular inflammation (%subcapsular inflammation) was calculated by dividing the total length of all regions of capsule adjacent to inflammation by the total length of capsule.

Kidney Cortical Volume

Presurgery computerized tomography or magnetic resonance imaging images from the clinical evaluation were downloaded to a workstation for processing. The cortical volumes of the contralateral kidney (chosen because ipsilateral cortical volume was often distorted by a tumor) were segmented using a semiautomated image processing tool (ITK-SNAP software, version 2.2; University of Pennsylvania, Philadelphia, PA) as previously reported.²⁴ A past study in a kidney donor population found higher %IF/TA to correlate with decreased cortical volume to a similar extent in the ipsilateral compared with the contralateral kidney.²⁴ Among computerized tomography or magnetic resonance imaging scans with poor cortical-medullary differentiation, the total kidney parenchymal volume (cortex plus medulla) was segmented, and cortical volume was estimated from this parenchymal volume, as previously described.²⁴

Outcomes

After nephrectomy, patients returned for visits every 3–6 months during the first year and subsequently once or twice per year. Serum creatinine levels were measured during follow-up. The last follow-up visit for this study was April 6, 2019. Patients who were unable to return for a follow-up visit were contacted to determine vital status. If alive, patients were followed with surveys and external electronic medical records were reviewed to obtain information on dialysis, and kidney transplantation status and serum creatinine levels. Progressive CKD was defined as a ≥40% decline in eGFR, dialysis or kidney transplantation, or an eGFR <10 ml/min per 1.73m². Noncancer mortality was assessed through death certificates; if death certificates were not available, the cause of death was verified with the local physician or family.

Statistical Analysis

Inflammation and IF/TA measurements were compared with Spearman’s correlations. The prognostic risks for progressive CKD or noncancer mortality were assessed separately with Cox proportional hazards models. Cox models for progressive CKD censored at last eGFR, death, or cancer recurrence. Cox models for noncancer mortality censored at last follow-up or cancer death. The risk of each of these outcomes was assessed for each doubling of tubulointerstitial measurements (%IF/TA, IF/TA density, %subcapsular IF/TA, striped IF/TA foci per 1 cm cortex, %inflammation, %inflammation–IF/TA, %inflammation-outside–IF/TA, and %subcapsular inflammation). Models were unadjusted, adjusted for %IF/TA, and further adjusted for clinical characteristics (age, sex, BMI, hypertension, diabetes, smoking status, eGFR, and proteinuria). The C-statistics of predictive models for progressive CKD on the basis of pathology IF/TA score versus morphometric %IF/TA were compared by computing a chi-squared statistic on the basis of the difference in concordances (covariance estimated using a jackknife variance estimator). Agreement between the two renal pathologists’ IF/TA score was assessed with Cohen’s weighted κ. Finally, analysis assessed the correlation of IF/TA and inflammation measurements with concurrent clinical characteristics and other microstructural features after adjustment for %IF/TA (using partial correlations) or after adjustment for %IF/TA, age, and sex (using partial correlations).

Results

There were 1206 patients who underwent radical nephrectomy from 2000 to 2015. As previously reported, we excluded 135 patients due to missing serum creatinine within 4 months of surgery; 107 patients due to death, cancer recurrence, kidney transplantation or dialysis within 4 months of surgery, and 28 due to specific diffuse kidney pathology (diagnosed upon histologic evaluation).¹⁴ Thus, the final cohort included 936 patients. Clinical characteristics, tumor stage, kidney microstructural and tubulointerstitial features, and outcomes are summarized in Table 1. The median time to progressive CKD or last available eGFR was 5.3 years. After censoring for cancer recurrence, progressive CKD occurred in 91 patients during follow-up, and of these, 29 patients required dialysis or received a kidney transplant. The median time to death or last follow-up was 6.4 years. There were 299 deaths during follow-up, of which 183 were not attributed to cancer.

Table 1. - Baseline characteristics of 936 patients with renal tumors

Demographics
Age, yr, mean (SD)	63.7 (12.0)
Men, n (%)	593 (63.4)
White, n (%)	861 (92.0)
Body mass index, kg/m2, mean (SD)	30.7 (6.7)
Hypertension, n (%)	614 (65.6)
Diabetes mellitus, n (%)	119 (12.7)
Active smoker, n (%)	127 (13.6)
Tumor burden, n (%)
Tumor stages
Benign tumor or tumor could not be staged	69 (7.4)
1A	109 (11.6)
1B	220 (23.5)
2A	113 (12.1)
2B	64 (6.8)
3A	286 (30.6)
3B	55 (5.9)
3C	12 (1.3)
4	8 (0.9)
Tumor volume, cm3, median (25%–75%)	97 (40–226)
Kidney function
Pre-nephrectomy creatinine, mg/dl, mean (SD)	1.1 (0.3)
Pre-nephrectomy eGFR, ml/min/1.73 m2, mean (SD)	71.9 (18.3)
Post-nephrectomy baseline 24-h urine protein, mg, median (25%–75%) a	124 (72–272)
Post-nephrectomy baseline creatinine, mg/dl, mean (SD)	1.5 (0.4)
Post-nephrectomy baseline eGFR, ml/min/1.73 m2, mean (SD)	48.0 (13.2)
Kidney microstructural features, mean (SD)
Biopsy cortex area, mm2	141.2 (57.2)
Capsule length, mm	23.8 (19.5)
Glomerular volume, mm3	0.0028 (0.0011)
Cortex volume per glomerulus, mm3	0.076 (0.035)
Tubular cross-sectional area, µm2	5989 (1478)
Globally sclerotic glomeruli, %, mean (SD)	9.1 (9.3)
Artery luminal stenosis, %, mean (SD)	55.5 (15.1)
Tubulointerstitial features, mean (SD)
IF/TA, %	4.2 (8.2)
IF/TA density, foci per cm2 cortex	25 (20)
Number of striped IF/TA, per cm of capsule length	0.15 (0.56)
Subcapsular IF/TA, %	12.6 (21.6)
Inflammation, %	0.5 (1.9)
Inflammation-IF/TA, %	8.5 (11.6)
Inflammation-outside-IF/TA, %	0.2 (0.3)
Subcapsular inflammation, %	5.5 (9.6)
Contralateral kidney on CT or MRI
Cortex volume, cm3, mean (SD) b	128.7 (41.5)
Outcomes during follow-up, n (%)
Progressive CKD	91 (9.7)
Noncancer death	183 (19.6)
Cancer death	116 (12.4)
Cancer recurrence	236 (25.2)

CT, computed tomography; MRI, magnetic resonanceimaging.

^aProteinuria measures were available in 816 patients.

^bCortex volume was available in 764 patients (417 directly measured, 347 estimated from total parenchymal volume).

%IF/TA measured by morphometry modestly correlated with pathologist’s %IF/TA score (r_s=0.38, P<0.0001) (Figure 3A). The agreement of the pathologist’s %IF/TA score (0%–10%, 11%–25%, 26%–50%, >50%) with the same categories by morphometry was also modest (κ=0.40; 95% confidence interval [95% CI], 0.31 to 0.48, P<0.0001). Among 50 random biopsy images, there was better agreement between the two pathologists on the %IF/TA score (agreement on 34 out of 50, κ=0.59, P<0.0001). The morphometric %IF/TA was substantially more predictive of progressive CKD than the pathologist’s %IF/TA score (C-statistic, 0.70 versus 0.54, P<0.0001). Therefore, subsequent analyses were limited to morphometric %IF/TA.

Figure 3.: Correlation of Morphometry %IFTA, pathologist's %IFTA score, and IFTA foci density. (A) Morphometric measure of %IF/TA shows significant but modest correlation with pathologist’s visually estimated IF/TA score (r _s=0.38, P<0.0001). Gray shaded area and dotted lines represent the ranges for pathologic scores. (B) Morphometric measure of IF/TA density increase with morphometric %IF/TA until %IF/TA >20% (r _s=0.76, P<0.0001).

Morphometric %IF/TA correlated with all other morphometric IF/TA measurements (IF/TA density, r_s=0.76, P<0.0001; %subcapsular IF/TA, r_s=0.84, P<0.0001, and striped IF/TA, r_s=0.49, P≤0.0001). The %inflammation correlated with subcapsular inflammation (r_s=0.76, P<0.0001), and the two components of %inflammation (%inflammation–IF/TA and %inflammation-outside–IF/TA) were correlated with each other (r_s=0.57, P≤0.0001). Notably, at higher levels of %IF/TA (>20%), IFTA density decreased with further increases in %IF/TA (Figure 3B). %IF/TA and %inflammation correlated with each other (r_s=0.71, P≤0.0001). The %IF/TA also correlated with %inflammation–IF/TA (r_s=0.30, P≤0.0001), %inflammation-outside–IF/TA (r_s=0.46, P≤0.0001), and %subcapsular inflammation (r_s=0.58, P≤0.0001).

All IF/TA measurements predicted progressive CKD and noncancer mortality in an unadjusted analysis (Table 2). As previously reported,¹⁴ %IF/TA predicted progressive CKD and noncancer mortality after adjusting for clinical characteristics. The risk of progressive CKD also increased with higher IF/TA density (Figure 4). After adjusting for both %IF/TA and clinical characteristics, only higher IF/TA density associated with a higher risk of CKD progression. This fully adjusted finding did not vary in persons with ≥5% IF/TA versus <5% IF/TA (hazard ratio [HR] 1.27 versus 1.12, interaction P=0.57), or when %IF/TA × IF/TA density interaction was assessed as continuous (P=0.54). Higher IF/TA density was also associated with an increased risk of noncancer mortality after adjusting for %IF/TA, but not after further adjusting for clinical characteristics. The %IF/TA per doubling remained a predictor of CKD progression (HR, 1.26; 95% CI, 1.14 to 1.39) and noncancer mortality (HR, 1.24; 95% CI, 1.16 to 1.33) after adjusting for IF/TA density.

Table 2. - Different IF/TA measurements as predictors of CKD progression and noncancer mortality from 4 months postnephrectomy

IF/TA Measure Per Doubling	Unadjusted		Adjusted for %IF/TA		Further Adjusted for Clinical Covariates a
	HR	P value	HR	P value	HR	P value
	(95% CI)		(95% CI)		(95% CI)
Outcome: CKD progression
%IF/TA	1.32	<0.001	—	—	1.22	0.003
	(1.19 to 1.46)				(1.07 to 1.39)
IF/TA density	1.15	<0.001	1.19	0.003	1.19	0.006
	(1.08 to 1.23)		(1.06 to 1.34)		(1.05 to 1.36)
%subcapsular IF/TA	1.07	<0.001	0.98	0.53	1.00 (0.93 to 1.07)	0.94
	(1.03 to 1.12)		(0.93 to 1.04)
Striped IF/TA per 1 cm cortex length	1.26	<0.001	1.11	0.07	1.09	0.19
	(1.16 to 1.38)		(0.99 to 1.25)		(0.96 to 1.25)
Outcome: noncancer mortality
%IF/TA	1.28	<0.001	—	—	1.13	0.005
	(1.20 to 1.37)				(1.04 to 1.23)
IF/TA density	1.11	<0.001	1.13	0.005	1.03	0.54
	(1.07 to 1.17)		(1.04 to 1.23)		(0.93 to 1.15)
%subcapsular IF/TA	1.08	<0.001	1.00	0.85	0.98	0.31
	(1.05 to 1.11)		(0.97 to 1.04)		(0.94 to 1.02)
Striped IF/TA per 1 cm cortex	1.19	<0.001	1.01	0.75	1.07	0.17
	(1.11 to 1.27)		(0.93 to 1.11)		(0.97 to 1.17)

CKD progression defined as eGFR reduction by ≥40%, need for dialysis, or kidney transplantation. Sample size of 936 patients with 91 CKD progression events and 183 noncancer related deaths.

^aAge, sex, BMI, hypertension, diabetes, active smoker, eGFR, and proteinuria (n=816).

Figure 4.: The probability of developing progressive CKD (40% decline in eGFR from postnephrectomy baseline, dialysis, transplantation, or eGFR <10 ml/min per 1.73 m²) increased with a higher IF/TA density (foci per cm² cortex).

All inflammation measurements predicted progressive CKD and noncancer mortality in unadjusted analysis (Table 3). After adjusting for %IF/TA, %inflammation-outside–IF/TA predicted an increased risk of CKD progression, whereas %inflammation-outside–IF/TA and %subcapsular inflammation predicted an increased the risk of noncancer mortality. After further adjustment for clinical characteristics, none of the inflammation measures predicted progressive CKD or noncancer mortality.

Table 3. - Different inflammation measurements as predictors of CKD progression and noncancer mortality from 4 months postnephrectomy

Inflammation measure per doubling	Unadjusted		Adjusted for %IF/TA		Further Adjusted for Clinical Covariates a
	HR	P value	HR	P value	HR	P value
	(95% CI)		(95% CI)		(95% CI)
Outcome: CKD progression
%inflammation	1.22	<0.001	1.09	0.13	0.93	0.29
	(1.13 to 1.32)		(0.97 to 1.23)		(0.81, 1.06)
%inflammation within IF/TA	1.11	0.01	1.01	0.79	0.93	0.22
	(1.02 to 1.21)		(0.92 to 1.12)		(0.83 to 1.04)
%inflammation-outside–IF/TA	1.14	<0.001	1.07	0.03	1.00	0.97
	(1.07 to 1.23)		(1.00 to 1.15)		(0.94 to 1.07)
%subcapsular inflammation	1.17	<0.001	1.07	0.08	1.03	0.58
	(1.09 to 1.26)		(0.99 to 1.17)		(0.93 to 1.13)
Outcome: noncancer mortality
%inflammation	1.19	<0.001	1.06	0.19	0.99	0.90
	(1.13 to 1.26)		(0.97 to 1.15)		(0.91, 1.09)
%inflammation within IF/TA	1.11	<0.001	1.02	0.66	0.96	0.25
	(1.04 to 1.18)		(0.95 to 1.09)		(0.88 to 1.03)
%inflammation-outside–IF/TA	1.11	<0.001	1.05	0.049	1.02	0.48
	(1.06 to 1.16)		(1.00 to 1.09)		(0.97 to 1.06)
%subcapsular inflammation	1.16	<0.001	1.07	0.02	1.02	0.55
	(1.11 to 1.22)		(1.01 to 1.14)		(0.96 to 1.08)

CKD progression defined as eGFR reduction by ≥40%, need for dialysis, or kidney transplantation. Sample size of 936 patients with 91 CKD progression events and 183 noncancer related deaths.

^aAge, sex, BMI, hypertension, diabetes, active smoker, eGFR, and proteinuria (n=816).

We further evaluated how measures of IF/TA and inflammation associated with clinical characteristics independent of %IF/TA. After adjusting for %IF/TA, higher IF/TA density correlated with older age, female, lower BMI, hypertension, lower eGFR, lower proteinuria, smaller glomerular volume, more globally sclerotic glomeruli, more arteriosclerosis (artery luminal stenosis), and smaller kidney cortex thickness and volume. After further adjusting for age and sex, higher IF/TA density correlated with lower eGFR, smaller nonsclerosed glomeruli, more global glomerulosclerosis, and smaller kidney cortex volume (Table 4). Higher %IF/TA no longer correlated with older age after adjusting for IF/TA density (r_s=0.04, P=0.27). After adjusting for %IF/TA, both %inflammation-outside–IF/TA and %subcapsular inflammation correlated with older age, hypertension, lower eGFR, proteinuria, more globally sclerotic glomeruli, and more arteriosclerosis (artery luminal stenosis).

Table 4. - Spearman’s unadjusted and partial correlations of different IF/TA and inflammation measurements with clinical characteristics, kidney function, and other microstructural features

	Unadjusted				Adjusted for %IF/TA			Further Adjusted for Age and Sex
	%IF/TA	IF/TA density	%inflammation- outside–IF/TA	%subcapsular inflammation	IF/TA density	%inflammation- outside–IF/TA	%subcapsular inflammation	IF/TA density
	rs (P value)	rs (P value)	rs (P value)	rs (P value)	rs (P value)	rs (P value)	rs (P value)	rs (P value)
Clinical characteristics
Age	0.35	0.43	0.22	0.31	0.27	0.07	0.14	—
	(<0.0001)	(<0.0001)	(<0.0001)	(<0.0001)	(<0.0001)	(0.03)	(<0.0001)
Male sex	0.13	0.02	0.06	0.03	−0.13	0.00	−0.05	—
	(<0.0001)	(0.62)	(0.07)	(0.29)	(<0.0001)	(0.98)	(0.13)
BMI	0.08	0.01	0.00	0.03	−0.07	−0.04	−0.03	−0.02
	(0.01)	(0.68)	(0.99)	(0.44)	(0.02)	(0.20)	(0.43)	(0.65)
Hypertension	0.27	0.26	0.19	0.25	0.09	0.08	0.12	0.04
	(<0.0001)	(<0.0001)	(<0.0001)	(<0.0001)	(0.005)	(0.02)	(0.0002)	(0.23)
Diabetes	0.22	0.15	0.07	0.10	−0.02	−0.03	−0.03	−0.01
	(<0.0001)	(<0.0001)	(0.03)	(0.002)	(0.46)	(0.30)	(0.29)	(0.67)
Current smoker	−0.07	−0.09	−0.09	−0.06	−0.05	−0.07	−0.03	0.00
	(0.03)	(0.009)	(0.005)	(0.05)	(0.14)	(0.04)	(0.36)	(0.97)
Kidney function
Postnephrectomy eGFR	−0.35	−0.28	−0.32	−0.34	−0.20	−0.14	−0.13	−0.10
	(<0.0001)	(<0.0001)	(<0.0001)	(<0.0001)	(<0.0001)	(0.006)	(0.0001)	(0.003)
Estimated 24-hour urine protein a	0.29	0.17	0.23	0.26	−0.08	0.11	0.12	−0.07
	(<0.0001)	(<0.0001)	(<0.0001)	(<0.0001)	(0.03)	(0.001)	(0.0005)	(0.06)
Other microstructural features
Mean glomerular volume	0.14	0.02	0.11	0.11	−0.13	0.05	0.04	−0.08
	(<0.0001)	(0.62)	(0.001)	(0.0005)	(<0.0001)	(0.13)	(0.22)	(0.01)
% globally sclerotic glomeruli	0.60	0.59	0.44	0.55	0.25	0.22	0.31	0.15
	(<0.0001)	(<0.0001)	(<0.0001)	(<0.0001)	(<0.0001)	(<0.0001)	(<0.0001)	(<0.0001)
% artery luminal stenosis	0.30	0.27	0.23	0.27	0.07	0.11	0.12	−0.02
	(<0.0001)	(<0.0001)	(<0.0001)	(<0.0001)	(0.04)	(0.0004)	(0.0001)	(0.46)
Mean cortical thickness	0.00	−0.08	−0.08	−0.05	−0.13	−0.09	−0.07	−0.04
	(0.97)	(0.01)	(0.02)	(0.10)	(<0.0001)	(0.006)	(0.04)	(0.22)
Contralateral kidney on CT or MRI
Cortical volume b	0.02	−0.14	−0.02	−0.04	−0.23	−0.04	−0.06	−0.11
	(0.53)	(0.0001)	(0.55)	(0.27)	(<0.0001)	(0.32)	(0.07)	(0.002)

CT, computed tomography; MRI, magnetic resonance imaging.

^aProteinuria measures were available in 816 patients.

^bCortex volume was available in 764 patients.

Discussion

Patients who undergo a radical nephrectomy for a tumor typically lack an overt nephropathy that would warrant a kidney biopsy, but are enriched with comorbidities that may affect the kidney. Histologic assessment of their nontumor parenchyma is recommended by the International Collaboration on Cancer Reporting guidelines.³⁰ The %IF/TA is an important prognostic marker, especially when quantified by morphometry rather than visual inspection by a pathologist, as we confirmed. This study assessed three different IF/TA measures and four different inflammation measures beyond the commonly used %IF/TA. Although most of these tubulointerstitial patterns of IF/TA and inflammation did not predict noncancer mortality or progressive CKD independent of %IF/TA and clinical characteristics, a more scattered pattern of IF/TA (more IF/TA foci per area of cortex) was more predictive of progressive CKD than a pattern of fewer and larger foci of IF/TA (fewer IF/TA foci per area of cortex).

Modest correlation between visually estimated IF/TA with morphometrically traced IF/TA is consistent with similar analysis in needle core kidney biopsies.³¹ The wedge sections used in this study were about 20 times larger than usual needle core biopsy sections, allowing for a more precise estimate of %IF/TA by both visual inspection and morphometry. Although time consuming in patients with many IF/TA foci, manual morphometry is a more quantitative and precise method to calculate %IF/TA, rather than solely relying on subjective visual estimation, resulting in better reproducibility.^2,3,31 To make morphometric analysis of %IF/TA more feasible, there have been attempts to improve assessment of IF/TA on kidney biopsies, using special stains,^2,32 or computer-based color image analysis.³³ With the advancement of artificial intelligence methods (deep learning or convolutional neural networks), there are also recent efforts in developing tools to automate the detection of %IF/TA from biopsy images.^343536–37 Pathologists may benefit from formal calibration training of their %IF/TA scoring against manual morphometry as the reference method, but there are also limits to how precisely %IF/TA can be estimated by visual inspection.

As expected, there was some degree of positive correlation between the different IF/TA and inflammation measurements. Further, all measures predicted CKD progression and noncancer mortality in unadjusted analysis. Because of this, it is difficult to detect patterns of IF/TA and inflammation that have unique prognostic importance without accounting for the dominant effect of overall %IF/TA. However, with multivariable analysis, we could assess the prognostic significance of different IF/TA and inflammation measures independent of %IF/TA.

Striped IF/TA has been described in the context of calcineurin inhibitor nephrotoxicity,³⁸ and is associated with diabetes, hypertension, and arteriosclerosis.^{25,38394041–42} The proposed underlying mechanism is thickening and hyalinization of interlobular arteries, which causes accentuated ischemic injury in watershed regions surrounding the medullary rays. However, striped IF/TA may be of limited prognostic importance because it did not predict progressive CKD or noncancer mortality independent of overall %IF/TA.

Increased subcapsular IF/TA has been described in aging kidneys.²³ In this pattern, IF/TA is arranged in the outermost cortical region, under the kidney capsule, and is attributed to the marginal vascular supply of the superficial cortical areas.²³ In support of this, we have previously reported age-related global glomerulosclerosis and ischemic glomeruli to be increased in the superficial cortex of these patients who underwent radical nephrectomy.¹³ The subcapsular IF/TA pattern has also been described in hypertensive nephrosclerosis.^23,43 However, subcapsular IF/TA may be of limited prognostic importance because it did not predict progressive CKD or noncancer mortality independent of overall %IF/TA.

We assessed the scatter of IFTA (foci density), which showed substantial variation at the same level of %IFTA. Although %IF/TA and IFTA density were positively correlated, IFTA density predicted progressive CKD independent of %IF/TA, suggesting additive information on risk. Larger foci of IF/TA in fewer regions of the cortex may reflect a disease process that affects some but not all nephrons. Whereas more scattered small foci of IF/TA may reflect a disease process affecting all, or nearly all, nephrons. Higher IF/TA density at the same level of %IF/TA also associated with smaller nonsclerosed glomeruli with lower eGFR. This may suggest a pathophysiology that constrains, or at least does not stimulate, an increase in the size and filtration rate of remaining glomeruli, such as the aging process.^44,45

Higher IF/TA density at the same level %IF/TA further reflects a more longstanding process by its correlation with more globally sclerotic glomeruli, and less cortex by imaging or biopsy. Foci of IF/TA may gradually atrophy over time, contracting the cortex and increasing their density. Consistent with this hypothesis, higher IF/TA density correlated with older age independent of %IF/TA, but higher %IF/TA did not correlate with older age independent of IF/TA density. Progressive atrophy and even disappearance of IF/TA foci over time may also explain why the increase in %IF/TA with healthy aging is minimal despite the substantial loss of nephrons.⁴⁶ An important stereological principal to consider is that at the same two-dimensional density, a biopsy with smaller IF/TA foci has a higher three-dimensional density of IF/TA foci than a biopsy with larger IF/TA foci. Thus, stereological models to estimate the true three-dimensional density of IF/TA foci may be of value in future studies. Conversely, the %IF/TA in two dimensions reasonably approximates to %IF/TA in three dimensions.⁴⁷

Inflammation is primarily a protective mechanism against kidney injury, but unresolved inflammation will ultimately be detrimental. Infiltrating inflammatory cells interplay with intrinsic renal cells and promote the secretion of profibrotic cytokines and growth factors that eventually result in IF/TA.¹ Inflammation can be present within IF/TA foci or outside of IF/TA. In the transplant setting, inflammation outside fibrotic areas and inflammation within scarred areas and in the immediate subcapsular cortex has been associated with a decreased graft survival.^5678–9 Also, inflammation outside of IF/TA has been associated with disease progression in patients with IgA nephropathy and lupus nephritis.^10,11 Consistent with these studies, we found that %inflammation-outside–IF/TA still modestly predicted progressive CKD and predicted noncancer mortality after adjusting for %IF/TA. These findings indicate that %inflammation-outside–IF/TA could represent an ongoing active process of kidney injury that, if left untreated, can ultimately progress to CKD. We found in adjusted analysis, this risk with %inflammation-outside–IF/TA was already accounted for by clinical characteristics including kidney function assessments. Alternatively, a study of clinically indicated kidney biopsies for a wide array of kidney diseases have found that %inflammation-outside–IF/TA associated with a lower risk of CKD progression.¹² This discrepancy with our study may be related to the use of kidney biopsies by a nephrologist to guide treatment decisions, including use of immunosuppression in some patients with %inflammation-outside–IF/TA. However, the nephrectomy kidney sections in this study were generally not used to guide treatment decisions. Further, none of these prior studies adjusted for morphometric %IF/TA, kidney function (GFR and proteinuria), and common CKD risk factors to the extent done in this study.

Our study has several potential limitations. The cause of death relied on death certificates, which are not always accurate. However, it is reassuring that histologic findings predict noncancer mortality, but not cancer mortality, in this cohort.¹⁴ Proteinuria per 24 hours was estimated from urine protein to osmolality ratios; this still represents a more accurate assessment of proteinuria than the urine dipstick, which is affected by urine concentration. All wedge specimens were stained with PAS to quantitatively measure IF/TA and inflammation. Other stains such as Trichrome and Sirius red have been reported to be better for the evaluation of IF/TA.²⁵ Nonetheless, we would expect the differences with other stains to be modest and foci of IF/TA were readily identified on PAS staining. We also lacked specific immunohistochemical stains to further characterize the inflammatory infiltrate. Thus, these findings provided insights into general patterns of inflammation, but an assessment of the different types of inflammatory cells may identify high risk patterns this study failed to detect.

To the best of our knowledge, this is the first study to rigorously assess several different patterns of IF/TA and inflammation beyond %IF/TA for an association with progressive CKD and noncancer mortality in a population not selected for a clinically indicated kidney biopsy. Notably, among patients with the same degree of %IF/TA, those with smaller IF/TA foci more scattered in the renal parenchyma had an increased risk of progressive CKD not detected by baseline eGFR, proteinuria, or common CKD risk factors. A more quantitative assessment of IF/TA with automated artificial intelligence technologies may be required to make these measurements feasible for clinical care.

Disclosures

A.D. Rule reports being a scientific advisor or member as JASN Associate Editor, Mayo Clinic Proceedings Section Editor, and member of the National Institute of Diabetes and Digestive and Kidney Diseases CKD Biomarker Consortium External Expert Panel; reports other interests/relationships with UpToDate; and, because A.D. Rule is an editor of JASN, he was not involved in the peer review process for this manuscript and a guest editor oversaw the peer review and decision-making process. B. Leibovich reports being a scientific advisor or member of the Kidney Cancer Association. All remaining authors have nothing to disclose.

Funding

This study was supported by the National Institutes of Health, National Institute of Diabetes and Digestive and Kidney Diseases grant R01 DK090358 and the Intramural Mayo Clinic Robert W. Fulk Career Development Award Fund in Nephrology Research (to A. Denic).