Abstract
Significance Statement
Nephrosclerosis (glomerulosclerosis, interstitial fibrosis, and tubular atrophy) is the defining pathology of both kidney aging and CKD. Optimal thresholds for nephrosclerosis that identify persons with a progressive disease are unknown. This study determined a young-age threshold (18–29 years) and age-based 95th percentile thresholds for nephrosclerosis on the basis of morphometry of kidney biopsy sections from normotensive living kidney donors. These thresholds were 7.1-fold to 36-fold higher in older (70 years or older) versus younger (aged 18–29 years) normotensive donors. Age-based thresholds, but not young-age threshold, were prognostic for determining risk of progressive CKD among patients who underwent a radical nephrectomy or a for-cause native kidney biopsy, suggesting that age-based thresholds are more useful than a single young-age threshold for identifying CKD on biopsy.
Background
Nephrosclerosis, defined by globally sclerotic glomeruli (GSG) and interstitial fibrosis and tubular atrophy (IFTA), is a pathology of both kidney aging and CKD. A comparison of risk of progressive CKD using aged-based thresholds for nephrosclerosis versus a single young-adult threshold is needed.
Methods
We conducted morphometric analyses of kidney biopsy images for %GSG, %IFTA, and IFTA foci density among 3020 living kidney donors, 1363 patients with kidney tumor, and 314 patients with native kidney disease. Using normotensive donors, we defined young-age thresholds (18–29 years) and age-based (roughly by decade) 95th percentile thresholds. We compared age-adjusted risk of progressive CKD (kidney failure or 40% decline in eGFR) between nephrosclerosis that was “normal compared with young,” “normal for age but abnormal compared with young,” and “abnormal for age” in patients with tumor and patients with kidney disease.
Results
The 95th percentiles in the youngest group (18–29 years) to the oldest group (70 years or older) ranged from 1.7% to 16% for %GSG, 0.18% to 6.5% for %IFTA, and 8.2 to 59.3 per cm2 for IFTA foci density. Risk of progressive CKD did not differ between persons with nephrosclerosis “normal compared with young” versus “normal for age but abnormal compared with young.” Risk of progressive CKD was significantly higher with %GSG, %IFTA, or IFTA foci density that was abnormal versus normal for age in both cohorts.
Conclusions
Given that increased risk of progressive CKD occurs only when nephrosclerosis is abnormal for age, age-based thresholds for nephrosclerosis seem to be better than a single young-age threshold for identifying clinically relevant CKD.
Keywords: clinical epidemiology, epidemiology and outcomes, histopathology, kidney biopsy, renal pathology, CKD non-dialysis
Introduction
CKD has been defined using a single threshold of eGFR <60 ml/min per 1.73 m2, regardless of age.1 This has been a source of significant debate,2,3 and recent studies suggest that when chronicity is present (eGFR based on three or more months), the <60 ml/min per 1.73 m2 threshold to define CKD is too low in young adults and too high in elderly adults.4,5 Distinguishing CKD severity in the setting of daily variation in eGFR and AKI is a known challenge.6 Chronicity of kidney disease can most clearly be determined by the detection of globally sclerotic glomeruli (GSG) and interstitial fibrosis and tubular atrophy (IFTA) on biopsy. Notably, these chronic changes of nephrosclerosis occur both with CKD and with kidney aging in healthy adults.7–9 However, a single-threshold approach to defining CKD would imply increases in nephrosclerosis with age are a disease. A thorough evaluation of nephrosclerosis is needed that compares the risk of progressive CKD between age-based thresholds versus a single threshold derived from young adults.
Current scoring of %GSG and %IFTA on kidney biopsy uses a 10% threshold for distinguishing none/minimal from mild to severe10 because this is practical for quick visual scoring. However, morphometric approaches are needed to accurately quantify severity of GSG and IFTA on kidney biopsy. We have previously developed age-based thresholds for morphometric %GSG based on 1847 healthy adults,11 and we and others have showed the utility of these thresholds for progressive CKD in patients with nephrotic syndrome or IgA nephropathy.12,13 A larger sample size to define morphometric IFTA and GSG thresholds that are tested across a spectrum of different patient populations is needed. Thus, we performed a study with two primary objectives. The first objective was to define an upper reference limit (95th percentile) for three measures of nephrosclerosis: %GSG, %IFTA, and IFTA foci density in healthy adults (kidney donors) using a threshold based on young adults (assumes age-related nephrosclerosis is a disease) and, separately, using a threshold for each age group (assumes age-related nephrosclerosis is not a disease). The second objective was to determine the risk of progressive CKD with age-based versus young-adult thresholds for nephrosclerosis in two other cohorts: patients with kidney tumor undergoing radical nephrectomy and patients with native kidney disease undergoing a for-cause kidney biopsy.
Methods
Study Populations
Three different cohorts with a baseline kidney biopsy within the Aging Kidney Anatomy study were used for all analyses. The first cohort was living kidney donors aged 18–77 years at Mayo Clinic Rochester (Minnesota), Mayo Clinic Arizona, and Cleveland Clinic sites with donation between 2000 and 2019 as previously described.14,15 At the time of kidney transplantation, a needle core biopsy of the renal cortex was obtained. The second cohort included patients at the Mayo Clinic Rochester who underwent a radical nephrectomy for a renal tumor between 2000 and 2019, as previously described.16 A wedge biopsy of the nontumor parenchyma away from the tumor was obtained from the radical nephrectomy specimen. The third cohort included adult residents of Olmsted County, Minnesota, who presented to Mayo Clinic Rochester for a needle core biopsy of their native kidneys to determine the cause of their kidney disease between 1993 and 2015 as previously described.17 There were 158 donor biopsies excluded for <2 mm2 of the cortex area, <4 glomeruli, or tissue compression/artifacts. All cohorts were studied with a waiver of consent and Mayo Clinic Institutional Review Board approval.
Clinical Characteristics
Baseline clinical characteristics were obtained from medical records before the biopsy as previously described. This included age, sex, race, body mass index, presence of diabetes (patients with kidney tumor and kidney disease cohorts only), presence of hypertension, measured GFR (kidney donors only), creatinine-based eGFR using the race-free 2021 Chronic Kidney Disease Epidemiology Collaboration equation,18 and 24-hour urine protein. Hypertension was defined by office blood pressure >140/90 mm Hg or use of antihypertensive medications in donors and by clinical diagnoses of hypertension on review of medical records in patients with tumor at the time of nephrectomy and patients with native kidney disease at the time of biopsy. For consistency, the 24-hour urine protein was determined primarily by estimation from spot protein-to-osmolality ratios in all three cohorts.19 In living kidney donors, if protein-to-osmolality ratio was not available, we estimated 24-hour urine protein from protein-to-creatinine ratios20 or used measured 24-hour urine protein values as available.
For patients with kidney tumor and patients with kidney disease, baseline was defined as occurring at 4 months after surgery (patients with tumor) or 4 months after biopsy (patients with kidney disease) and baseline eGFR was determined using the last available serum creatinine before this 4-month time point. This was performed in patients with tumor to study the eGFR loss relative to the new baseline after nephrectomy rather than eGFR loss due to nephrectomy. This was performed in patients with kidney disease because AKI often prompts kidney biopsy and a time period is needed to establish a baseline eGFR after AKI recovers.17 We excluded 149 patients with tumor who had kidney failure, cancer recurrence, or death in the 4 months after nephrectomy. We also excluded 93 patients with tumor with missing serum creatinine data and 33 with specific kidney diseases on biopsy as previously described.16 We excluded 80 patients with native kidney disease for kidney failure, death, or no further follow-up in the 4 months after biopsy. We also excluded 15 patients with kidney disease with missing serum creatinine data, 38 for inadequate biopsies, two for concurrent lymphoproliferative disease, and one who underwent a radical nephrectomy.
Nephrosclerosis Measures
Paraffin-embedded sections of kidney biopsies were stained with periodic acid–Schiff, and for donors and patients with kidney disease, an adjacent section was also stained with Masson trichrome. These stained sections were then digitized into a high-resolution (×20 or ×40) image using Aperio whole slide scanners (Aperio AT2 system scanner, Leica Microsystems, Inc., Buffalo Grove, IL; https://www.leicabiosystems.com/digital-pathology/). Digital images were magnified on a large touch-screen tablet using Aperio Image Scope Software (version 12.4.3.7009) to manually trace the cortex and each non–globally sclerosed glomerular profile, GSG profile, and IFTA focus as shown in Figure 1. Multiple physician annotators have contributed to data collection with the Aging Kidney Anatomy study over the past 13 years. A. Denic performed quality control reviews to ensure accurate and consistent data between different annotators. Annotations are corrected if errors are identified. From these annotations, %IFTA was calculated from the area of all IFTA foci divided by the cortex area. Each IFTA focus required at least three atrophic tubules with contiguous interstitial fibrosis. The IFTA foci density was calculated from counts of all IFTA foci divided by cortex area (per cm2). If IFTA foci were located on the edge of tissue biopsy (i.e., cut by the needle), they were counted as half a focus when calculating IFTA foci density (Supplemental Figure 1). The number of glomeruli was the sum of all glomerular profiles (sclerosed or not sclerosed). Empty capsules were also included in the counts of glomeruli. Nonsclerosed glomeruli (NSGs) bisected by the needle were counted as 0.675 (mean ratio of bisected NSG areas to complete NSG areas). The %GSG was calculated by the number of GSG divided by the number of glomeruli. This estimation was averaged between the consecutive trichrome- and periodic acid–Schiff-stained sections (donors and native kidney disease cohorts only) for a more precise estimate of %GSG with the limited tissue of needle core biopsy sections.
Figure 1.
Representative biopsy images. Images show increasing severity of nephrosclerosis from (A) low-risk living kidney donors to (B) medium-risk patients with kidney tumor to (C) high-risk patients with native kidney disease. For each image, the cortex is traced in green; non–globally sclerosed glomeruli are traced in blue; globally sclerosed glomeruli are traced in red; and each distinct IFTA foci are traced in black. GSG, globally sclerotic glomeruli; IFTA, interstitial fibrosis and tubular atrophy.
Other Kidney Biopsy Measures
Additional biopsy measures were assessed for their association with nephrosclerosis as described in more detail in the Supplemental Methods. Briefly, these included glomerular volume and density calculated with the Weibel and Gomez stereological models from all NSGs.21 We further calculated cortex per nonsclerosed glomerulus from the reciprocal of glomerular density within the non-IFTA cortex. This is a more accurate measure of nephron size because IFTA contributes to the area of cortex.15 Arteriosclerosis was assessed by artery %luminal stenosis from intimal thickening as calculated from the area of intima divided by the areas of intima and lumen using the most orthogonal artery to the plane of the biopsy section in the kidney donor and kidney disease cohort and the mean of the three most orthogonal arteries to the plane of the biopsy in patients with tumor.22
Progressive CKD Outcome
The patients with tumor were followed annually with serum creatinine values to determine eGFR as part of a postnephrectomy clinical protocol as previously described.23 The patients with kidney disease had follow-up per usual clinical care at Mayo Clinic Rochester (the only nephrology provider in Olmsted County and surrounding counties) as previously reported.17 For the patients with tumor and patients with kidney disease, a progressive CKD outcome was defined by dialysis, kidney transplantation, eGFR <10 ml/min per 1.73 m2, or a 40% decline in eGFR from the 4-month baseline after nephrectomy or after biopsy. We lacked the data needed to assess a progressive CKD outcome in living kidney donors.
Statistical Analyses
Thresholds for %GSG, %IFTA, and IFTA foci density were defined using only normotensive kidney donors (no antihypertensive medications and blood pressure<140/90 mm Hg) because hypertension associates with nephrosclerosis independent of age.9,11 The 95th percentile for these nephrosclerosis measures were calculated for each age group: 18–29 years (young adults), 30–39, 40–49, 50–59, 60–69, and 70+ years. There were no normotensive kidney donors older than 77 years, so the 70+ years age group was calculated using donors aged 70–77 years but applied to anyone 70 years or older in subsequent analyses. Because %GSG increases when there are fewer total glomeruli on a kidney biopsy,24 there was not a single %GSG threshold for each age group. Instead, we used quantile regression to estimate the 95th percentile for number of GSG by age group and by the number of glomeruli groups (1–4, 4.1–8, 8.1–16, 17.1–32, 32.1–48, 48.1–91.4) as previously described.11 Because the patients with tumor have much larger numbers of glomeruli than 91.4 (because of wedge sections), we converted age thresholds for the 48.1–91.4 glomeruli group into percent thresholds for use in patients with tumor by dividing the 95th percentile for number of GSG in this 48.1–91.4 glomeruli group by the mean number of glomeruli in that glomeruli group. We assessed the validity of counting IFTA foci on needle core biopsies that were bisected by the needle as half a focus. First, we identified a healthy subset of patients with tumor aged 50–79 years by excluding those with diabetes, hypertension, 24-hour urine protein >500 mg/d, eGFR <60 ml/min per 1.73 m2 for ages 50–59 years, and eGFR <45 ml/min per 1.73 m2 for ages 60–79 years. Then, we assessed whether the 95th percentile for IFTA foci density differed between normotensive kidney donors and healthy-subset patients with tumor for the age groups 50–59, 60–69, and 70–79 years with a test of interaction.
Figure 2 presents a conceptual graph of how these 95th percentile thresholds were used to classify patients by severity of nephrosclerosis measures into three groups: “normal compared with young,” “normal for age but abnormal compared with young,” and “abnormal for age.” Patients in the three cohorts were classified as normal compared with young if <95th percentile for the 18–29 years group, normal for age but abnormal compared with young if ≥95th percentile for the 18–29 years group but less than the 95th percentile for their age group, and abnormal for age if ≥95th percentile for their age group. The association of each nephrosclerosis measure with baseline clinical characteristics and other biopsy findings was assessed in each cohort. Analyses compared abnormal for age versus normal for age. Similar analyses compared normal compared with young versus abnormal compared with young among persons who were normal for age.
Figure 2.
Conceptual model of young- and age-based thresholds for nephrosclerosis measures defined using normotensive kidney donors. The 95th percentile for 18–29 years defines the abnormal compared with young threshold. The 95th percentile for 70–77 years defines the abnormal for age threshold for all persons 70 years and older. Using these thresholds, patients can be grouped into normal compared with young, normal for age but abnormal compared with young, and abnormal for age.
For patients with tumor and patients with kidney disease, the risk of the progressive CKD outcome from the 4-month postnephrectomy or postbiopsy baseline was assessed with Cox proportional hazard models. The risk of progressive CKD outcomes was first compared between those with nephrosclerosis normal compared with young, normal for age but abnormal compared with young, versus abnormal for age. Then, the risk of progressive CKD outcomes was compared between those with nephrosclerosis normal for age versus abnormal for age. All association analyses were age-adjusted (continuous). All statistical analyses were performed using BlueSky Statistics software version 7.40 (BlueSky Statistics LLC, Chicago, IL) and R (RStudio) version 3.4.2. A P value of <0.05 was considered significant.
Results
There were 3020 kidney donors, 1363 patients with tumor, and 314 patients with native kidney disease studied; their characteristics are provided in Table 1. Severity of proteinuria and of chronic changes on biopsy (larger nephrons and more nephrosclerosis) increased progressively from kidney donors to patients with tumor and from patients with tumor to patients with kidney disease (Table 1 and Figure 1). The 4-month baseline eGFR was lower in patients with tumor than patients with kidney disease; nonetheless, the progressive CKD event occurred much more frequently in patients with kidney disease than patients with tumor. Supplemental Table 1 presents the 95th percentile estimates for number of GSG estimated for each age group and each number of glomeruli group on the basis of quantile regression in normotensive kidney donors. Using the mean number of glomeruli in each group, we also converted the GSG-count 95th percentile to %GSG 95th percentiles. The %GSG 95th percentile based on 48.1–91.4 glomeruli Supplemental Table 1 was used for patients with tumor with many glomeruli on wedge biopsies. For example, a 55-year-old patient with tumor with 200 glomeruli on their wedge section needs to have <8.4% or <17 GSG to be normal for age. This is based on normotensive kidney donors aged 50–59 years with 48.1–91.4 (mean 59.8) glomeruli, where the 95th percentile for %GSG is 8.4% (5.0 of 59.8 glomeruli). All thresholds in Supplemental Table 1 were used in kidney donor and kidney disease cohorts based on the number of glomeruli on needle core biopsies.
Table 1.
Demographics and clinical characteristics of each cohort
| Clinical Characteristic | Kidney Donors (n=3020) | Patients with Kidney Tumor (n=1363) | Patients with Kidney Disease (N=314) |
|---|---|---|---|
| Mean (SD), Median (IQR), or n (%) | |||
| Demographic | |||
| Age, yr | 44.1 (12.1) | 64.0 (11.9) | 53.4 (18.0) |
| Men, % | 1235 (40.9) | 873 (64.0) | 167 (53.2) |
| Race | |||
| White or unknown | 2772 (91.8) | 1306 (95.8) | 289 (92.1) |
| Black | 92 (3.0) | 22 (1.6) | 11 (3.5) |
| American Indian/Alaskan Native | 48 (1.6) | 15 (1.1) | 1 (0.3) |
| Asian | 43 (1.4) | 4 (0.3) | 13 (4.1) |
| Other | 65 (2.2) | 16 (1.2) | 0 (0.0) |
| CKD risk factors at baseline | |||
| BMI, kg/m2 | 27.5 (4.8) | 30.8 (6.7) | 29.2 (6.9) |
| Diabetes mellitus, % | 0 (0) | 185 (13.6) | 73 (23.2) |
| Hypertension, % | 437 (14.5) | 919 (67.4) | 206 (65.6) |
| Kidney function at baseline | |||
| Measured GFR, ml/min per 1.73 m2 | 104.1 (19.7) | — | — |
| eGFR, ml/min per 1.73 m2 | 100.0 (16.2) | 51.1 (13.8) | 62.7 (31.8) |
| 24-h urine protein, mga | 63 (52–94) | 154 (87–307) | 1913 (583–4372) |
| Kidney biopsy findings at baseline | |||
| Total no. of glomeruli on biopsy | 20.4 (11.1) | 337.3 (155.5) | 11.0 (8.9) |
| Nephron size | |||
| Glomerular tuft volume, mm3 | 0.0026 (0.0011) | 0.0028 (0.0010) | 0.0042 (0.0029) |
| Nonsclerosed cortex per glomerulus, mm3 | 0.057 (0.028) | 0.071 (0.028) | 0.110 (0.113) |
| Nephrosclerosis | |||
| Globally sclerosed glomeruli, % | 2.9 (5.3) | 8.6 (8.7) | 21.1 (22.9) |
| IFTA, % | 0.3 (1.2) | 3.6 (7.5) | 15.3 (17.7) |
| IFTA density, foci per cm2 cortex | 4.4 (11.8) | 25.1 (20.9) | 79.1 (77.2) |
| Luminal stenosis (arteriosclerosis), %b | 35.1 (17.6) | 54.5 (15.5) | 43.5 (18.3) |
| Progressive CKD outcome | |||
| Kidney failure or 40% decline in eGFR | — | 126 (9.2) | 171 (54.5) |
BMI, body mass index; IFTA, interstitial fibrosis and tubular atrophy.
Twenty-four–hour urine protein missing in 42 donors, 178 patients with tumor, and four native biopsy patients.
Missing in 80 donors because of lack of an artery.
Table 2 presents the 95th percentile thresholds for %GSG, %IFTA, and IFTA foci density for each age group based on 2583 normotensive kidney donors. Between the young (18–29 years) and oldest (70+ years) age groups, the 95th percentile threshold increased 9.5-fold for %GSG (≥1.7% to ≥16%), 36-fold for %IFTA (≥0.18% to ≥6.5%), and 7.1-fold for IFTA foci density (≥8.2 to ≥59.3 per cm2). We confirmed that in the 50–79 years age range, the 95th percentile for IFTA foci density in the “healthy” subset of patients with tumor was similar to that in normotensive kidney donors (where IFTA foci cut by the needle were counted as half) (Supplemental Table 2). Figure 3 graphically shows these %GSG, %IFTA, and IFTA foci density thresholds, dividing normotensive kidney donors into three groups: normal compared with young, normal for age but abnormal compared with young, or abnormal for age. We then applied these thresholds to the full living kidney donor cohort (normotensive and hypertensive), the kidney tumor cohort, and the native kidney disease cohort. The %GSG, %IFTA, and IFTA foci density were abnormal for age among 5.4%, 6.0%, and 6.3% of all kidney donors; 23%, 24%, and 14% of patients with tumor; and 62%, 69%, and 55% of patients with kidney disease, respectively.
Table 2.
Age-based thresholds for nephrosclerosis measures (≥95th percentile) in 2583 normotensive kidney donors
| Nephrosclerosis Measures | %GSG | %IFTA | IFTA Foci Density | ||||
|---|---|---|---|---|---|---|---|
| Age Groups (Mean) | N | 95th Percentile Threshold, %a | Abnormal, n (%) | 95th Percentile Threshold, % | Abnormal, n (%) | 95th Percentile Threshold, per cm2 | Abnormal, n (%) |
| 18–29 yr (25 yr) | 396 | ≥1.7 | 27 (6.8) | ≥0.18 | 20 (5.1) | ≥8.2 | 20 (5.1) |
| 30–39 yr (35 yr) | 678 | ≥4.1 | 31 (4.6) | ≥0.53 | 34 (5.0) | ≥14.6 | 34 (5.0) |
| 40–49 yr (45 yr) | 792 | ≥6.6 | 36 (4.5) | ≥1.0 | 40 (5.1) | ≥18.4 | 40 (5.1) |
| 50–59 yr (54 yr) | 519 | ≥9.1 | 19 (3.7) | ≥2.0 | 26 (5.0) | ≥33.6 | 26 (5.0) |
| 60–69 yr (64 yr) | 180 | ≥12 | 8 (4.4) | ≥3.3 | 9 (5.0) | ≥47.2 | 9 (5.0) |
| 70+ yr (72 yr) | 18 | ≥16 | 1 (5.6) | ≥6.5 | 1 (5.6) | ≥59.3 | 1 (5.6) |
GSG, globally sclerotic glomeruli; IFTA, interstitial fibrosis and tubular atrophy.
Threshold shown for normotensive kidney donors with 48.1–91.4 glomeruli on biopsy; the 95th percentile threshold for % globally sclerotic glomeruli increased when fewer glomeruli were present on biopsy (See Supplemental Table 1 for 95th percentile thresholds used when 48 or fewer glomeruli are present on biopsy).24
Figure 3.
Thresholds for nephrosclerosis in normotensive living kidney donors based on the 95th percentile for young (18–29 years) and the 95th percentile for each age group. For (A) %GSG, (B) %IFTA, and (C) IFTA foci density, the abnormal compared with young threshold were 1.7%, 0.18%, and 8.2 per cm2, whereas the abnormal for age threshold ranged from 1.7% to 16%, 0.18%–5.6%, and 8.2–59.3 per cm2, respectively. GSG, globally sclerotic glomeruli; IFTA, interstitial fibrosis and tubular atrophy.
Table 3 shows the association of clinical and biopsy characteristics with %GSG, %IFTA, and IFTA foci density between persons who were normal for age versus abnormal for age in each of the three cohorts. Men were more likely to have abnormal for age %IFTA in all three cohorts. Women donors and women patients with tumor and Black donors and Black patients with tumor were more likely to have abnormal for age %GSG. Higher body mass index associated with abnormal for age %GSG and %IFTA only in patients with tumor. Diabetes, hypertension, lower GFR, and higher proteinuria generally associated with abnormal for age measures of nephrosclerosis, although some of these associations were not evident in kidney donors. Larger nephron size by non-IFTA cortex per glomerulus associated with abnormal for age %GSG in all three cohorts. Arteriosclerosis associated with abnormal for age measures of nephrosclerosis in patients with tumor and patients with kidney disease, but not kidney donors. Supplemental Table 3 shows the age-adjusted association of these same characteristics with %GSG, %IFTA, and IFTA foci density among persons who were normal for age divided between those who were abnormal compared with young versus normal compared with young. Similar associations were evident in patients with tumor, but there were fewer associations with clinical and biopsy characteristics in kidney donors and in patients with kidney disease.
Table 3.
Associations with nephrosclerosis abnormal versus normal for age in donors, patients with tumor, and patients with kidney disease (age-adjusted)
| Nephrosclerosis Measure | %GSG | %IFTA | IFTA Foci Density | ||||||
|---|---|---|---|---|---|---|---|---|---|
| Cohort | Kidney Donors | Patients with Kidney Tumor | Patients with Kidney Disease | Kidney Donors | Patients with Kidney Tumor | Patients with Kidney Disease | Kidney Donors | Patients with Kidney Tumor | Patients with Kidney Disease |
| Clinical characteristics | |||||||||
| Men | 0.66 (0.47 to 0.93) | 0.69 (0.52 to 0.90) | 1.21 (0.77 to 1.91) | 1.79 (1.31 to 2.45) | 1.36 (1.04 to 1.78) | 1.83 (1.13 to 2.98) | 1.51 (1.11 to 2.06) | 1.00 (0.73 to 1.37) | 1.76 (1.10 to 2.82) |
| Black | 2.47 (1.25 to 4.89) | 2.53 (1.03 to 6.23) | 1.80 (0.54 to 6.04) | 1.72 (0.82 to 3.64) | 0.31 (0.07 to 1.33) | 1.18 (0.31 to 4.56) | 1.25 (0.54 to 2.92) | 0.94 (0.27 to 3.20) | 1.36 (0.35 to 5.22) |
| BMI, per SD | 0.99 (0.85 to 1.17) | 1.17 (1.02 to 1.34) | 0.99 (0.80 to 1.23) | 1.04 (0.89 to 1.21) | 1.29 (1.14 to 1.46) | 1.00 (0.80 to 1.25) | 0.92 (0.78 to 1.08) | 1.05 (0.91 to 1.22) | 1.13 (0.91 to 1.41) |
| Diabetes | — | 1.83 (1.29 to 2.61) | 2.25 (1.31 to 3.89) | — | 2.21 (1.59 to 3.08) | 4.32 (2.03 to 9.22) | — | 1.20 (0.78 to 1.83) | 3.46 (1.75 to 6.86) |
| Hypertension | 1.70 (1.14 to 2.55) | 2.47 (1.73 to 3.53) | 2.35 (1.37 to 4.02) | 1.78 (1.20 to 2.64) | 2.15 (1.58 to 2.92) | 2.18 (1.27 to 3.75) | 1.91 (1.30 to 2.80) | 2.04 (1.40 to 2.97) | 3.37 (1.95 to 5.81) |
| GFR, per SDa | 1.10 (0.93 to 1.30) | 0.50 (0.43 to 0.59) | 0.57 (0.43 to 0.76) | 0.96 (0.81 to 1.15) | 0.56 (0.48 to 0.65) | 0.62 (0.46 to 0.83) | 0.88 (0.74 to 1.05) | 0.74 (0.63 to 0.87) | 0.85 (0.64 to 1.12) |
| 24-h protein, per doubling | 1.09 (0.88 to 1.34) | 1.22 (1.12 to 1.33) | 1.16 (1.04 to 1.29) | 1.35 (1.10 to 1.65) | 1.26 (1.16 to 1.36) | 1.18 (1.06 to 1.32) | 1.24 (1.01 to 1.51) | 1.08 (0.98 to 1.19) | 1.18 (1.06 to 1.32) |
| Nephron size | |||||||||
| NSG volume, mm3 | 0.92 (0.78 to 1.09) | 1.38 (1.21 to 1.56) | 1.03 (0.82 to 1.29) | 1.13 (0.98 to 1.31) | 1.41 (1.25 to 1.59) | 1.11 (0.86 to 1.43) | 1.04 (0.90 to 1.21) | 0.96 (0.82 to 1.12) | 1.16 (0.90 to 1.50) |
| Non-IFTA cortex per glomerulus, mm3 | 1.15 (1.00 to 1.32) | 1.62 (1.41 to 1.85) | 1.45 (1.10 to 1.91) | 1.02 (0.88 to 1.19) | 1.31 (1.16 to 1.48) | 1.40 (0.98 to 2.10) | 0.87 (0.73 to 1.04) | 0.90 (0.76 to 1.06) | 1.51 (1.04 to 2.19) |
| Arteriosclerosis | |||||||||
| % luminal stenosis | 1.06 (0.90 to 1.24) | 1.74 (1.47 to 2.07) | 1.43 (1.11 to 1.84) | 1.07 (0.92 to 1.24) | 1.59 (1.37 to 1.86) | 1.35 (1.03 to 1.76) | 0.98 (0.84 to 1.14) | 1.32 (1.11 to 1.58) | 1.64 (1.25 to 2.16) |
Data shown as odds ratios (95% confidence interval), per SD for continuous measures. GSG, globally sclerotic glomeruli; BMI, body mass index; NSG, nonsclerosed glomeruli; IFTA, interstitial fibrosis and tubular atrophy.
Measured GFR for donors and eGFR for patients with tumor and patients with kidney disease.
The patients with kidney tumor and patients with native kidney disease were followed for a median of 5.5 and 7.2 years, respectively. Table 4 summarizes the age-adjusted risk of progressive CKD in both cohorts for each nephrosclerosis measure. We initially evaluated normal compared with young, normal for age but abnormal compared with young, and abnormal for age. However, there was no evidence that normal for age but abnormal compared with young had a higher risk of progressive CKD than normal compared with young for either cohort with any nephrosclerosis measure (Table 4). Thus, these two categories were combined into one category of normal for age for comparison with abnormal for age (Table 4). Abnormal for age %GSG, %IFTA, and IFTA foci density associated with an increased risk of progressive CKD in patients with tumor and patients with kidney disease. Figure 4 shows the higher risk of progressive CKD for abnormal for age compared with normal for age for %GSG, %IFTA, and IFTA foci density in patients with kidney tumor and patients with kidney disease (log-rank P value <0.001 for all). Supplemental Figure 2 presents four clinical vignettes from the native kidney disease cohort that demonstrate practical application of %GSG and %IFTA thresholds when the nephrosclerosis severity is abnormal for age for younger patients but normal for age in older patients.
Table 4.
Globally sclerotic glomeruli, %IFTA, IFTA foci density categories as predictors of progressive CKD in 1363 patients with tumor and 314 patients with kidney disease (age-adjusted).
|
Nephrosclerosis Measure |
Patients with Kidney Tumor (N=1363) | Patients with Kidney Disease (N=314) | ||
|---|---|---|---|---|
| HR (95% CI) | P Value | HR (95% CI) | P Value | |
| Three categories | ||||
| Globally sclerotic glomeruli | ||||
| Normal compared with young | Ref | Ref | Ref | Ref |
| Normal for age but abnormal compared with young | 0.74 (0.38 to 1.47) | 0.39 | 1.40 (0.89 to 2.20) | 0.15 |
| Abnormal for age | 1.74 (0.85 to 3.57) | 0.13 | 2.12 (1.49 to 3.02) | <0.001 |
| %IFTA | ||||
| Normal compared with young | Ref | Ref | Ref | Ref |
| Normal for age but abnormal compared with young | 0.69 (0.34 to 1.39) | 0.30 | 1.42 (0.63 to 3.21) | 0.39 |
| Abnormal for age | 1.75 (0.87 to 3.53) | 0.12 | 3.37 (1.68 to 6.78) | <0.001 |
| IFTA foci density | ||||
| Normal compared with young | Ref | Ref | Ref | Ref |
| Normal for age but abnormal compared with young | 1.26 (0.70 to 2.27) | 0.44 | 1.47 (0.66 to 3.28) | 0.35 |
| Abnormal for age | 2.46 (1.29 to 4.71) | 0.007 | 4.08 (2.02 to 8.25) | <0.001 |
| Two categories | ||||
| Global glomerulosclerosis | ||||
| Normal for age | Ref | Ref | Ref | Ref |
| Abnormal for age | 2.28 (1.58 to 3.30) | <0.0001 | 1.87 (1.38 to 2.53) | <0.001 |
| %IFTA | ||||
| Normal for age | Ref | Ref | Ref | Ref |
| Abnormal for age | 2.41 (1.68 to 3.47) | <0.0001 | 2.65 (1.80 to 3.89) | <0.001 |
| IFTA foci density | ||||
| Normal for age | Ref | Ref | Ref | Ref |
| Abnormal for age | 2.03 (1.34 to 3.08) | 0.0009a | 3.11 (2.13 to 4.54) | <0.001 |
CI, confidence interval; IFTA, interstitial fibrosis and tubular atrophy; HR, Hazard Ratio.
Figure 4.
Risk of progressive CKD. The risk of kidney failure or a 40% decline in eGFR in patients with kidney tumor (A–C) and patients with kidney disease (D–F) by nephrosclerosis measures (%GSG, %IFTA, and IFTA foci density) normal for age versus abnormal for age. GSG, globally sclerotic glomeruli; IFTA, interstitial fibrosis and tubular atrophy.
Discussion
This study used healthy adults (normotensive kidney donors) to define 95% thresholds for distinguishing abnormal versus normal levels of nephrosclerosis (%GSG, %IFTA, and IFTA foci density) on kidney biopsy. Thresholds were defined based on young adults alone and age-based. Even in healthy young adults, some minimal degree of nephrosclerosis was common, but the severity of nephrosclerosis substantially increased with age among healthy adults. However, this finding alone does not clarify whether or not this increase in nephrosclerosis with age is due to a subclinical progressive CKD.
To address this, the risk of progressive CKD was compared with age-based thresholds and young-adult thresholds in both medium-risk patients with kidney tumor and high-risk patients with kidney disease. Age-based thresholds for %GSG, %IFTA, and IFTA foci density consistently identified persons at risk of progressive CKD among the kidney tumor patient and kidney disease patient cohorts. However, there was not an increased risk of progressive CKD with abnormal compared with young versus normal compared with young for %GSG, %IFTA, or IFTA foci density that was normal for age. Thus, these data support an age-based approach to classifying CKD.
These findings are particularly relevant for %GSG and %IFTA where scoring by pathologists with the commonly used 10% threshold fails to distinguish whether the severity of nephrosclerosis is consistent with what is expected with aging or is concerning for chronic changes from disease. This 10% threshold is also too high for %IFTA at all ages and for %GSG at ages younger than 60 years compared with the morphometry derived 95th percentile thresholds. Histological subtypes of GSG (e.g., solidification) and IFTA (e.g., with inflammation) help distinguish GSG and IFTA because of disease rather than aging. If a capsule is present on biopsy, subcapsular GSG is also more likely to be due to aging.25 However, GSG and IFTA may not always be distinguishable between an aging or disease etiology based on appearance. Thus, age-based thresholds provide a reasonable approach to assessing whether the nephrosclerosis exceeds the level expected for age. This is particularly relevant for older patients with kidney disease who undergo a kidney biopsy; age-related nephrosclerosis may be misintepreted as chronic kidney disease. Separate assessment of %GSG, %IFTA, and IFTA foci density also allows for a more multidimensional assessment of nephrosclerosis to help overcome some of the imprecision limitations with needle core biopsies.
These data found several clinical and biopsy characteristics that associate with nephrosclerosis abnormal for age across low-risk (kidney donor), medium-risk (patients with tumor), and high-risk (kidney disease) cohorts. The %IFTA and IFTA foci density was more likely to be abnormal in men, whereas %GSG was more likely to be abnormal in women. The reasons for these sex differences are not fully clear. The %GSG was more likely to be abnormal in Black adults. Black adults have a higher risk of proteinuria, progressive GFR decline, and kidney failure than White adults do.26–30 A glomerulopathy associated with APOL1 alleles may explain the higher likelihood of %GSG in Black adults.31 Risk factors of CKD (diabetes and hypertension), low GFR, and proteinuria had fairly consistent associations with abnormal nephrosclerosis across the cohorts. Low measured GFR did not associate with abnormal nephrosclerosis in kidney donors, but this is likely because of exclusion for low GFR during the donor selection process. Measures of larger nephron size showed some association with abnormal nephrosclerosis consistent with a compensatory response to the loss of functioning nephrons below levels appropriate for age.22 Among patients with tumor and patients with kidney disease, arteriosclerosis may lead to more GSG and IFTA via ischemia.32
While pathologists typically count glomeruli and GSG to determine %GSG, undercounting of glomeruli often occurs.33 Pathologists typically only estimate %IFTA at <10%, 10%–25%, 26%–50%, and >50% thresholds on the basis of visual inspection alone.10 Morphometric assessment to obtain accurate measures of IFTA severity is time-consuming and impractical for clinical care. Deep learning–based approaches show promise toward automating morphometric assessments of nephrosclerosis.34–40 When automated morphometry tools eventually become available, pathology reports will have the potential to provide more precise estimates of nephrosclerosis measures along with age-specific reference thresholds. This will help physicians meaningfully interpret whether the severity of GSG and IFTA on biopsy reflects an increased risk of progressive CKD or is just age-related nephrosclerosis.
There were limitations to this study. There were only 18 normotensive kidney donors older than 70 years and none older than 77 years available to define age-based thresholds. Our proposed approach used thresholds defined in persons aged 70–77 years for all persons 70 years and older. It is inherently difficult to define reference thresholds for ages older than 77 years because living kidney donation is rare. Our cohorts included predominantly White patients. We could still detect a higher proportion of Black adults with GSG abnormal for age in two different cohorts. To simplify our study, we used age groups to define age-specific thresholds. More precise thresholds for a specific age can be estimated with interpolation using the mean values in the provided tables. We were unable to assess the risk of progressive CKD in living kidney donors because we lacked long-term repeated follow-up eGFR measures that were available in the other two cohorts.
In conclusion, prognostic interpretation of nephrosclerosis severity on kidney biopsy requires age-based thresholds to define abnormal. This was validated across three different cohorts that varied substantially in the proportion that had nephrosclerosis abnormal for age and in their risk of progressive CKD. These nephrosclerosis thresholds can be used for research when kidney biopsy data are available for morphometric analysis. In the clinical setting, digital pathology with deep learning models may eventually automate quantifying nephrosclerosis on kidney biopsy images, but age-based reference ranges will still be needed for proper interpretation.
Supplementary Material
Disclosures
L. Barisoni reports Consultancy: Protalix, Sangamo, and Vertex; Honoraria: Protalix, Sangamo, and Vertex; Advisory or Leadership Role: Glomerular Disease Journal, Nature Review Nephrology, and Nephcure Scientific Advisory Board; and Other Interests or Relationships: Nephcure. B.C. Leibovich reports Ownership Interest: Pathright Medical; and Advisory or Leadership Role: Kidney Cancer Association. A.D. Rule reports Patents or Royalties: UpToDate; and Advisory or Leadership Role: JASN—Associate Editor, Mayo Clinic Proceedings—Section Editor, and NIDDK—Urological Diseases of America Contract Management Board. M.D. Stegall reports Consultancy: Aiosyn, eGenesis, Hansa, and Novartis; Research Funding: Janssen, Talaris, and Veloxis; and Advisory or Leadership Role: Aisosyn and eGenesis. All remaining authors have nothing to disclose.
Funding
This study was supported with funding from the National Institutes of Health, National Institute of Diabetes and Digestive and Kidney Diseases (R01 DK090358).
Author Contributions
Conceptualization: Andrew D. Rule.
Data curation: Muhammad S. Asghar, Joshua Augustine, Aleksandar Denic, Bradley C. Leibovich, Amr Moustafa, Andrew D. Rule, Mark D. Stegall, R. Houston Thompson.
Formal analysis: Muhammad S. Asghar, Aleksandar Denic, Aidan F. Mullan.
Funding acquisition: Andrew D. Rule.
Methodology: Aidan F. Mullan, Andrew D. Rule.
Project administration: Andrew D. Rule.
Supervision: Andrew D. Rule.
Validation: Aleksandar Denic, Aidan F. Mullan.
Writing – original draft: Muhammad S. Asghar, Andrew D. Rule.
Writing – review & editing: Mariam P. Alexander, Joshua Augustine, Laura Barisoni, Aleksandar Denic, Bradley C. Leibovich, Amr Moustafa, Mark D. Stegall, R. Houston Thompson.
Data Sharing Statement
All data are from patient medical records. Data can only be shared after a data use agreement and institutional review board approval.
Supplemental Material
This article contains the following supplemental material online at http://links.lww.com/JSN/E445.
Supplemental Table 1. Upper reference limit (95th percentile) for number of GSG (and mean %GSG) per section, on the basis of age and number of glomeruli per section in 2583 normotensive kidney donors.
Supplemental Table 2. Age-based thresholds for IFTA density (95th percentile) using 2583 normotensive kidney donors and 330 patients with tumor who are “healthy.”
Supplemental Table 3. Associations with nephrosclerosis abnormal versus normal for young in donors, patients with tumor, and patients with kidney disease who have nephrosclerosis normal for age (age-adjusted).
Supplemental Figure 1. Several biopsy examples showing IFTA.
Supplemental Figure 2. Clinical vignettes were abstracted from the medical record for four of the patients with native kidney disease to illustrate how these age-based 95th percentile thresholds can be applied practically.
References
- 1.Improving Global Outcomes (KDIGO) CKD Work Group. KDIGO 2012 clinical practice guideline for the evaluation and management of chronic kidney disease. Kidney Int Suppl. 2013;3(1):1–150. doi: 10.1038/kisup.2012.73 [DOI] [Google Scholar]
- 2.Delanaye P Jager KJ Bokenkamp A, et al. CKD: a call for an age-adapted definition. J Am Soc Nephrol. 2019;30(10):1785–1805. doi: 10.1681/ASN.2019030238 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.O'Hare AM, Rodriguez RA, Rule AD. Rule AD: overdiagnosis of chronic kidney disease in older adults-an inconvenient truth. JAMA Intern Med. 2021;181(10):1366–1368. doi: 10.1001/jamainternmed.2021.4823 [DOI] [PubMed] [Google Scholar]
- 4.Jonsson AJ, Lund SH, Eriksen BO, Palsson R, Indridason OS. The prevalence of chronic kidney disease in Iceland according to KDIGO criteria and age-adapted estimated glomerular filtration rate thresholds. Kidney Int. 2020;98(5):1286–1295. doi: 10.1016/j.kint.2020.06.017 [DOI] [PubMed] [Google Scholar]
- 5.Liu P Quinn RR Lam NN, et al. Accounting for age in the definition of chronic kidney disease. JAMA Intern Med. 2021;181(10):1359–1366. doi: 10.1001/jamainternmed.2021.4813 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Nishimoto M Murashima M Kokubu M, et al. Kidney function at 3 months after acute kidney injury is an unreliable indicator of subsequent kidney dysfunction: the NARA-AKI Cohort Study. Nephrol Dial Transplant. 2022;38(3):664–670. doi: 10.1093/ndt/gfac172 [DOI] [PubMed] [Google Scholar]
- 7.Denic A, Glassock RJ, Rule AD. The kidney in normal aging: a comparison with chronic kidney disease. Clin J Am Soc Nephrol. 2022. 10;17(1):137–139. doi: 10.2215/CJN.10580821 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Denic A, Rule AD, Glassock RJ. Healthy and unhealthy aging on kidney structure and function: human studies. Curr Opin Nephrol Hypertens. 2022;31(3):228–234. doi: 10.1097/MNH.0000000000000780 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Rule AD Amer H Cornell LD, et al. The association between age and nephrosclerosis on renal biopsy among healthy adults. Ann Intern Med. 2010;152(9):561–567. doi: 10.7326/0003-4819-152-9-201005040-00006 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Sethi S D'Agati VD Nast CC, et al. A proposal for standardized grading of chronic changes in native kidney biopsy specimens. Kidney Int. 2017. 10;91(4):787–789. doi: 10.1016/j.kint.2017.01.002 [DOI] [PubMed] [Google Scholar]
- 11.Kremers WK Denic A Lieske JC, et al. Distinguishing age-related from disease-related glomerulosclerosis on kidney biopsy: the Aging Kidney Anatomy study. Nephrol Dial Transplant. 2015;30(12):2034–2039. doi: 10.1093/ndt/gfv072 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Hommos MS Zeng C Liu Z, et al. Global glomerulosclerosis with nephrotic syndrome; the clinical importance of age adjustment. Kidney Int. 2018. 10;93(5):1175–1182. doi: 10.1016/j.kint.2017.09.028 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Chung CS Lee JH Jang SH, et al. Age-adjusted global glomerulosclerosis predicts renal progression more accurately in patients with IgA nephropathy. Sci Rep. 2020;10(1):6270. doi: 10.1038/s41598-020-63366-0 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Merzkani MA Denic A Narasimhan R, et al. Kidney microstructural features at the time of donation predict long-term risk of chronic kidney disease in living kidney donors. Mayo Clinic Proc. 2021;96(1):40–51. doi: 10.1016/j.mayocp.2020.08.041 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Issa N Vaughan LE Denic A, et al. Larger nephron size, low nephron number, and nephrosclerosis on biopsy as predictors of kidney function after donating a kidney. Am J Transplant. 2019;19(7):1989–1998. doi: 10.1111/ajt.15259 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Ricaurte Archila L Denic A Mullan AF, et al. A higher foci density of interstitial fibrosis and tubular atrophy predicts progressive CKD after a radical nephrectomy for tumor. J Am Soc Nephrol. 2021;32(10):2623–2633. doi: 10.1681/ASN.2021020267 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Denic A Bogojevic M Mullan AF, et al. Prognostic implications of a morphometric evaluation for chronic changes on all diagnostic native kidney biopsies. J Am Soc Nephrol. 2022;33(10):1927–1941. doi: 10.1681/ASN.2022030234 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Inker LA Eneanya ND Coresh J, et al. New creatinine- and cystatin C-based equations to estimate GFR without race. N Engl J Med. 2021;385(19):1737–1749. doi: 10.1056/NEJMoa2102953 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Wilson DM, Anderson RL. Protein-osmolality ratio for the quantitative assessment of proteinuria from a random urinalysis sample. Am J Clin Pathol. 1993;100(4):419–424. doi: 10.1093/ajcp/100.4.419 [DOI] [PubMed] [Google Scholar]
- 20.Marco Mayayo MP, Martinez Alonso M, Valdivielso Revilla JM, Fernandez-Giraldez E. A new gender-specific formula to estimate 24-hour urine protein from protein to creatinine ratio. Nephron. 2016;133(4):232–238. doi: 10.1159/000447604 [DOI] [PubMed] [Google Scholar]
- 21.Weibel ER, Gomez DM. A principle for counting tissue structures on random sections. J Appl Physiol. 1962;17(2):343–348. doi: 10.1152/jappl.1962.17.2.343 [DOI] [PubMed] [Google Scholar]
- 22.Denic A Lieske JC Chakkera HA, et al. The substantial loss of nephrons in healthy human kidneys with aging. J Am Soc Nephrol. 2017;28(1):313–320. doi: 10.1681/ASN.2016020154 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Ricaurte Archila L Denic A Mullan AF, et al. A higher foci density of interstitial fibrosis and tubular atrophy predicts progressive CKD after a radical nephrectomy for tumor. J Am Soc Nephrol. 2021;32(10):2623–2633. doi: 10.1681/ASN.2021020267 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Niznik RS Lopez CL Kremers WK, et al. Global glomerulosclerosis in kidney biopsies with differing amounts of cortex: a clinical-pathologic correlation study. Kidney Med. 2019;1(4):153–161. doi: 10.1016/j.xkme.2019.05.004 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Denic A Elsherbiny H Mullan AF, et al. Larger nephron size and nephrosclerosis predict progressive CKD and mortality after radical nephrectomy for tumor and independent of kidney function. J Am Soc Nephrol. 2020;31(11):2642–2652. doi: 10.1681/ASN.2020040449 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Poggio ED, Rule AD. A critical evaluation of chronic kidney disease--should isolated reduced estimated glomerular filtration rate be considered a 'disease. Nephrol Dial Transplant. 2008;24(3):698–700. doi: 10.1093/ndt/gfn704 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Tomlinson LA, Clase CM. Sex and the incidence and prevalence of kidney disease. Clin J Am Soc Nephrol. 2019;14(11):1557–1559. doi: 10.2215/CJN.11030919 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Bock F Stewart TG Robinson-Cohen C, et al. Racial disparities in end-stage renal disease in a high-risk population: the Southern Community Cohort Study. BMC Nephrol. 2019;20(1):308. doi: 10.1186/s12882-019-1502-z [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Norton JM Moxey-Mims MM Eggers PW, et al. Social determinants of racial disparities in CKD. J Am Soc Nephrol. 2016;27(9):2576–2595. doi: 10.1681/ASN.2016010027 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Chu CD Powe NR McCulloch CE, et al. Trends in chronic kidney disease care in the US by race and ethnicity, 2012-2019. JAMA Netw Open. 2021;4(9):e2127014. doi: 10.1001/jamanetworkopen.2021.27014 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Robinson TW, Freedman BI. The impact of APOL1 on chronic kidney disease and hypertension. Adv Chronic Kidney Dis. 2019;26(2):131–136. doi: 10.1053/j.ackd.2019.01.003 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Kasiske BL. Relationship between vascular disease and age-associated changes in the human kidney. Kidney Int. 1987;31(5):1153–1159. doi: 10.1038/ki.1987.122 [DOI] [PubMed] [Google Scholar]
- 33.Rosenberg AZ Palmer M Merlino L, et al. The application of digital pathology to improve accuracy in glomerular enumeration in renal biopsies. PLoS One. 2016;11(6):e0156441. doi: 10.1371/journal.pone.0156441 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Ginley B Jen KY Han SS, et al. Automated computational detection of interstitial fibrosis, tubular atrophy, and glomerulosclerosis. J Am Soc Nephrol. 2021;32(4):837–850. doi: 10.1681/ASN.2020050652 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Yi Z Salem F Menon MC, et al. Deep learning identified pathological abnormalities predictive of graft loss in kidney transplant biopsies. Kidney Int. 2022;101(2):288–298. doi: 10.1016/j.kint.2021.09.028 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Salvi M Mogetta A Gambella A, et al. Automated assessment of glomerulosclerosis and tubular atrophy using deep learning. Comput Med Imaging Graph. 2021;90:101930. doi: 10.1016/j.compmedimag.2021.101930 [DOI] [PubMed] [Google Scholar]
- 37.Kammardi Shashiprakash A Lutnick B Ginley B, et al. A distributed system improves inter-observer and AI concordance in annotating interstitial fibrosis and tubular atrophy. Proc SPIE Int Soc Opt Eng. 2021;11603:116030V. doi: 10.1117/12.2581789 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Hermsen M de Bel T den Boer M, et al. Deep learning-based histopathologic assessment of kidney tissue. J Am Soc Nephrol. 2019;30(10):1968–1979. doi: 10.1681/ASN.2019020144 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Marsh JN, Liu TC, Wilson PC, Swamidass SJ, Gaut JP. Development and validation of a deep learning model to quantify glomerulosclerosis in kidney biopsy specimens. JAMA Netw Open. 2021;4(1):e2030939. doi: 10.1001/jamanetworkopen.2020.30939 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Pesce F Albanese F Mallardi D, et al. Identification of glomerulosclerosis using IBM Watson and shallow neural networks. J Nephrol. 2022;35(4):1235–1242. doi: 10.1007/s40620-021-01200-0 [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
All data are from patient medical records. Data can only be shared after a data use agreement and institutional review board approval.