Nonneoplastic Renal Cortical Scarring at Tumor Nephrectomy Predicts Decline in Kidney Function Open Access

Hematoxylin and eosin (H&E)–stained slides from the nonneoplastic renal parenchyma of total and partial nephrectomies were retrospectively reviewed from the years 1998 to 2008. The main indication for surgery was the presence of a renal mass. All surgical procedures were performed and evaluated at our medical center. In total, 456 cases were reviewed: 118 partial and 338 total nephrectomies. Compressed parenchymal changes were frequently seen in the adjacent kidney tissue, up to 1 to 2 mm from the tumor, showing increased interstitial fibrosis and GS. Therefore, where possible, the representative nontumor sections were examined at a significant distance from the tumor (at minimum 1 cm away) from the cases of total nephrectomy. Partial nephrectomies, because of the limited amount of adjacent nonneoplastic renal tissue, were more challenging to evaluate. For these cases, it was critical to recognize tumor compression artifact and evaluate the section as far as possible beyond the tumor. Typically this measurement was around 0.5 cm, but it was variable from case to case depending on the size and location of the tumor. Seventy-five cases, 32 partial and 43 total nephrectomy specimens, were excluded from the evaluation because of lack of sufficient, viable nonneoplastic tissue for examination. The reason for exclusion was tumor compression artifact in 19 (6 total and 13 partial), absence of nonneoplastic tissue in 25 (6 total and 19 partial), embolization or infarction of the entire kidney in 22 (all total nephrectomies), and previous development of end stage renal disease in 9 (all total).

In 170 cases, additional diagnostic techniques were used in conjunction with routine H&E evaluation. Special stains were most commonly used: periodic acid–Schiff (PAS) in 155, Jones methenamine silver in 13, and trichrome in 15. Immunofluorescence microscopy was performed on 7 cases, 6 of which were brought to the attention of the renal pathologist in a fresh state at the time of the nephrectomy and 1 that was done retrospectively using standard immunofluorescence techniques on paraffin-embedded tissue. Electron microscopy was used in 11 cases: 7 had tissue placed in paraformaldehyde at the time of surgery and 4 were reprocessed from the paraffin block. With the exception of 32 cases that became clinically apparent and were referred for a formal diagnostic review by a renal pathologist, all the pathologic evaluation was blinded to the patients' clinical histories and laboratory data.

Pathologic parameters evaluated were percentage of GS, percentage of interstitial fibrosis/tubular atrophy (IF/TA) estimated to the nearest 5%, and vascular sclerosis of arteries and arterioles separately scored in a semiquantitative manner as none-to-mild, moderate, or severe. The arteriosclerosis was more commonly fibrointimal and scored based on the intimal and medial thickening leading to narrowing of the vascular lumen of up to 20 to 25 observed vessels: mild, less than 25% narrowing (Figure 1, A); moderate, 25% to 50% narrowing (Figure 1, B); and severe, more than 50% narrowing (Figure 1, C) (adapted from Gautam et al²  and Bijol et al⁴). Arteriolosclerosis, best scored on glomerular hilar arterioles, was typically medial hypertrophy with thickened layers of smooth muscle. Arteriolosclerosis was also scored based on degree of narrowing using the same parameters as arteries (Figure 1, D through F). Hyalinosis, where present, involved arteriolar intima and was separately described. Pathologic assessment of GS, IF/TA, and vascular sclerosis was compared in all cases where H&E and PAS methods were used.

Additional nonneoplastic diagnoses were established in 56 cases either at the time of surgery (22), subsequently because of worsening renal function (10), or at review for this study (24).

Clinical information was retrospectively obtained from the patients' records under an institutional review board protocol. The clinical parameters assessed were patient age, sex, comorbidities including hypertension and diabetes, preoperative serum creatinine level, presence of proteinuria or hematuria, blood pressure, and weight. Follow-up serum creatinine level was also evaluated for all patients with at least a 12-month follow-up period (156 cases). The decline in renal function was measured by change in creatinine level from preoperative to latest postoperative follow-up value.

After exclusion of 75 cases for compression, embolization, lack of adequate or preserved nonneoplastic tissue, or previous end-stage renal disease, 381 cases remained for pathologic evaluation: 86 partial and 295 total nephrectomies. The tumors were renal cell carcinoma, clear cell type (254), papillary type (62), or chromophobe type (29); oncocytoma (30); urothelial cell carcinoma (33); angiomyolipoma (14); and miscellaneous (27). In toto, 117 (30.7%) had minimal pathologic change as defined by less than 5% GS, less than 10% IF/TA, and mild or no significant vascular sclerosis (interlobular, small arteries, arterioles including hilar arterioles). The remainder of the cases (69.3%) displayed pathologic changes characterized by global GS (mean 11.5%, range, 0%–85%), IF/TA (mean 14%, range, 0%–100%), or vascular sclerosis (46.7% with at least moderate arteriosclerosis) with or without intimal hyalinosis.

Additional nonneoplastic diagnoses were rendered in 57 cases, 15.0% of all cases reviewed (Table 1). The additional diagnoses were most commonly diabetic nephropathy (28 cases) or hypertensive nephrosclerosis (11 cases). Diabetic nephropathy may have multiple histologic appearances, from isolated glomerular basement membrane thickening to mild or severe diffuse mesangial GS (class IIA and IIB, respectively) (Figure 2, A and B) to nodular mesangial GS (class III) (Figure 2, C).¹⁰  Exudative lesions (hyalinosis, fibrin cap, capsular drop) may also be seen (Figure 2, C). Focal segmental glomerulosclerosis was present in 7 cases (Figure 3, A), collapsing glomerulopathy in 2 (Figure 3, B), and granulomatous tubulointerstitial nephritis in 2 (Figure 3, C); and there was 1 case each of acute pyelonephritis, thrombotic microangiopathy, atheroembolic disease (Figure 3, D), reflux nephropathy, proliferative glomerulonephropathy, membranous glomerulonephropathy, and secondary amyloidosis (serum amyloid A type).

The 28 patients with diabetic nephropathy all had diabetes, and in limited data available on 19, proteinuria was present in 9 of 15 (range, 0–10.6 g/24 h) and creatinine level ranged from 0.7 to 4.5 mg/dL (mean 1.5 mg/dL; to convert mg/dL to μmol/L, multiply by 88.4).

Of the 7 patients with focal segmental glomerulosclerosis, we had additional data on 5. All 5 had proteinuria: 1 had proteinuria in the nephrotic range, 1 had proteinuria of 1.5 g/24 h; for the other 3, spot proteins were detected on urinalysis only. It is important to point out that the focal segmental glomerulosclerosis lesions were mainly perihilar variant because of presumed hyperfiltration injury in these patients. Additional clinical information on 1 of the patients with collapsing glomerulopathy revealed proteinuria of 1.3 g/24 h. For 2 patients with granulomatous tubulointerstitial nephritis, 1 had active ulcerative colitis and 1 had fungal forms identified. The patient with reflux nephropathy had documented reflux. Finally, the patient with membranous nephropathy had 9.4 g/24 h of proteinuria.

Kidneys with tumor as well as an additional nonneoplastic disease as listed above tended to have more severe vascular sclerosis (74% severe, 17% moderate, and 9% none-to-mild), higher global GS (mean 15%), and more IF/TA (mean 22%).

The degree of vascular sclerosis had a statistically significant, direct relationship with the percentage of global GS and percentage of tubular atrophy and interstitial fibrosis in the renal cortical tissue (Table 2). The cases with mild arteriolosclerosis had 5.56% ± 2.73% GS (range, 0%–20%) and 7% ± 6.6% IF/TA (0%–70%). Moderate arteriolosclerosis kidneys showed 9.2% ± 9.7% GS (0%–60%) with 15% ± 17.4% IF/TA (5%–100%). The cases with severe arteriolosclerosis had 23% ± 27.8% GS (5%–100%) and 29% ± 33.7% IF/TA (5%–100%).

Tumor size and location dictate eligibility for partial versus total nephrectomy, which has clinically been shown to impact prognosis on chronic kidney disease. Of the 456 consecutive cases in the study, 226 (49.5%) were American Joint Committee on Cancer stage T1a (≤4 cm). Of these, 60% underwent total nephrectomy and 40% partial nephrectomy. During the time period of the study, at our institution, partial nephrectomies for T1a tumors increased in frequency: 1998–2002, 24% partial; 2003–2005, 26% partial; and 2006–2008, 55% partial. Twenty-six of the nonneoplastic diagnoses listed above were from T1a tumors, 16 of which were removed by total nephrectomy.

Tumors of 4 to 7 cm (American Joint Committee on Cancer stage T1b) are also occasionally resectable via partial nephrectomy. An additional 117 T1b tumors were present in the study, 17% removed by partial nephrectomy. As for T1a, the T1b tumor partial nephrectomy frequency increased during the time period studies: 1998–2002, 9%; 2003–2005, 5%; and 2006–2008, 25%. An additional 16 cases of specific nonneoplastic disease were diagnosed in T1b tumor cases.

Follow-up of greater than 12 months (mean 49.2 ± 34.8 months) was available in 156 cases (Table 3). Overall, the mean change from preoperative to postoperative creatinine level was an increase of 0.4 mg/dL (range, −0.6 to 4.6 mg/dL). A statistically significant rise in creatinine level was observed for different degrees of arteriosclerosis and arteriolosclerosis from preoperative to postoperative period with a mean follow-up ranging from 45.6 to 52.8 months. The average change in creatinine level in cases with severe arteriosclerosis (0.9 mg/dL) or arteriolosclerosis (0.9 mg/dL) was significantly higher than in either mild (0.2–0.3 mg/dL) or moderate (0.3 mg/dL) vascular sclerosis, respectively. The type of surgery also correlated with statistically significant rise in the postoperative creatinine levels (0.5 mg/dL, 52.8 months mean follow-up) following total nephrectomy as compared with the partial nephrectomy cases (0.2 mg/dL, 39.6 months mean follow-up).

The percentages of both IF/TA and GS also correlated with long-term rise in creatinine level at cutoffs of more than 10% for IF/TA and more than 5% for GS (Table 3). When data were specifically separated into smaller groups with cutoffs of 10%, 25%, and 50%, there was no statistically significant difference in postoperative creatinine level change based on IF/TA.

Substantially higher changes in postoperative creatinine level were seen for total versus partial nephrectomies at all degrees of arteriolosclerosis: mild, total 0.3 mg/dL (31.2 months follow-up), versus partial, 0.1 mg/dL (32.4 months); moderate, total 0.3 mg/dL (36 months), versus partial, 0.1 mg/dL (40.8 months); severe, total 1.3 mg/dL (31.2 months), versus partial, 0.3 mg/dL (44.4 months). Several other clinical and pathologic variables were not associated with a measurable creatinine level change, including clinical diagnosis of diabetes, pathologic diagnosis of diabetic nephropathy, and older age (age groups younger than 50 years, 50–75 years, or older than 75 years).

Age was also compared in conjunction with the degree of arteriolosclerosis to determine the effect on the change in postoperative creatinine level at last follow-up (Table 4). The number of patients was small in each group, and the only statistically significant difference was that those patients between the ages of 50 and 75 with severe arteriolosclerosis had a higher mean rise in creatinine level compared with the same age group with either mild or moderate vascular disease. At the same degree of arteriolosclerosis, there was no difference in the change in creatinine level between the age categories.

Specific tumor types generally did not portend a statically significant worse postnephrectomy course, but oncocytomas (n = 11) did have a higher change in creatinine level, 0.7 mg/dL, compared with the other tumor types. The oncocytomas had a longer follow-up period, 70.8 months, compared with 32.4 to 52.8 months for other tumor types. They were more commonly found in older patients, mean 69.9 ± 8 years, compared with 61 ± 11 years (P = .01) for clear cell carcinomas and 57.7 ± 13.4 years (P = .004) for papillary renal cell carcinomas. The only statistically significant difference found in tumor types was between urothelial cell carcinoma (P = .02) and angiomyolipoma (P = .01), but not clear cell (P = .08), papillary (P = .41), chromophobe (P = .10), or the miscellaneous tumors (P = .75).

Three cases progressed to end-stage renal disease, 2 with hypertensive nephrosclerosis and 1 with diabetic nephropathy, from 1 to 4.5 years after surgery.

In the 155 cases where both H&E and PAS staining was used, there were only minimal differences in interpretation of GS, IF/TA, and vascular sclerosis based on the parameters described. The majority of the cases were identical for the different variables. In general, the assessment of global GS (4% difference) and vascular sclerosis was mildly higher by PAS staining, whereas IF/TA was scored mildly lower (5% difference). The difference in vascular sclerosis for arteries and arterioles evaluated by H&E and PAS stains ranged from 7% to 13% (155 cases). Based on PAS staining, in 4 cases arteries and in 9 cases arterioles were downstaged from severe to moderate and in 12 cases arteries and 13 cases arterioles were upstaged from moderate to severe. The changes in postoperative creatinine level were similar when patients were grouped by vascular sclerosis assessed by PAS compared with H&E methods, and the statistical significance was unchanged by the minimal PAS-assessed differences.

The goal of the examination of nonneoplastic disease in the setting of tumor nephrectomy is to provide prognostic information useful for predicting subsequent renal function decline, mainly the amount of global GS and interstitial fibrosis, the degree of vascular sclerosis, and any additional medical renal disease present. Knowledge of such information could then be used to tailor therapy in the postnephrectomy follow-up period. Herein, we report a 10-year retrospective review of medical renal lesions seen in the nonneoplastic portion of 456 consecutive tumor nephrectomy specimens with available follow-up data. Important prognostic factors for renal survival could be assessed in the majority of both partial and total nephrectomies. Exclusions were made in 27% of partial and 12.7% of total nephrectomies because of adjacent tissue compression, embolization/infarction, lack of nonneoplastic tissue, or previous development of end-stage renal disease. Of those evaluated, 69.3% of the cases showed more than minimal pathology involving 1 or more components, that is, glomerular sclerosis, interstitial fibrosis, or vascular sclerosis. In addition, 57 cases (15.0%) had definitive medical renal diagnoses, of which only 38.6% (22 of 57) were recognized and diagnosed at the time of initial resection. Follow-up of greater than 12 months in 156 cases (mean 48.8 months) demonstrated higher postoperative creatinine levels for those kidneys with pathologically detectable severe arteriosclerosis and arteriolosclerosis, more than 10% IF/TA, and more than 5% GS, as well as for those with non–nephron-sparing surgical techniques, that is, total or radical nephrectomies.

Various preexisting renal parenchymal diseases in the nonneoplastic portion of the kidney are not infrequently found in the nontumor portion of the kidney specimen and offer an opportunity to examine an adequate amount of tissue using appropriate means of testing such as special stains (PAS, Jones methenamine silver, trichrome), immunohistochemical stains, and/or electron microscopy for diagnosis.^3,4  A discussion of where to sample and appropriately evaluate the nonneoplastic kidney is difficult and may depend on the availability of evaluable tissue. Certainly, in total nephrectomies, the ideal section would be well away from the tumor in an entirely separate block. In partial nephrectomies, because of limited material present, sampling locations can be individually judged by determining how much tissue is available beyond the compression-related changes of interstitial fibrosis, global GS, and vascular sclerosis immediately adjacent to the tumor. Tumor size, location, and effects on adjacent vessels have a major impact in the extent of adjacent parenchymal compression. So, a specific value such as 0.5 cm may serve as a guideline, but may overestimate or underestimate the amount of relatively preserved tissue in some cases. Moreover, despite compression, certain lesions, when present, may be recognizable even in compressed parenchyma.

Our findings of 15.0% of cases with diagnostic nonneoplastic disease and 69.3% nonspecific, chronic glomerular, tubulointerstitial, and/or vascular changes are in keeping with previous reports that have demonstrated additional diagnoses in 9.8%³  and greater than 60% to up to 90% for nonspecific changes.^4,9 

Nonneoplastic disease was not recognized or mentioned in 61.4% (35 of 57) of cases at the time of reporting of the tumor diagnosis. Overlooking medical disease in the nonneoplastic tissue is not unusual, as emphasis for tumor nephrectomy specimens is placed on the diagnosis, classification, and staging of the tumors typically by a general surgical pathologist or genitourinary pathology specialist, who may not be familiar with medical renal pathologic changes.³  For this reason, following an initiative by the Renal Pathology Society in 2009, a proposal was made to the College of American Pathologists Cancer Committee to institute additional steps in the examination of tumor nephrectomies.

The College of American Pathologists checklist protocol has been subsequently modified to include the importance of recognizing medical renal disease in these specimens and its importance on subsequent renal function.¹¹  The recommendation is that an additional stain, PAS or Jones methenamine silver, be ordered up front for all cases on the nonneoplastic renal parenchymal section. The utility of either special stain is to enhance the review of the glomerular architecture as well as the renal parenchymal elements and best serve to address any abnormalities that may be identified. In addition, a systematic examination with semiquantitative assessment of the more-common pathologic changes such as global GS, arteriosclerosis, the presence and severity of diabetic nephropathy, and identification of additional glomerular, tubulointerstitial, or vascular lesions, in consultation with a nephropathologist if necessary, is recommended.

The importance of identifying and reporting the medical kidney disease, if present, in the tumor nephrectomy is to alert the physician of the potential for ongoing risk factors or causes of chronic renal insufficiency. Appropriate medical management and close clinical follow-up may then be initiated to minimize the risks and damage.

A patient with unilateral total or, to a lesser extent, partial nephrectomy sustains a loss of renal mass and nephron number that leads to functional and structural adaptations of the remaining kidney, including hyperfiltration, intraglomerular hypertension, glomerulopathy, proteinuria, and renal insufficiency.^12–14  Hyperfiltration effects are further complicated by preexisting vascular, diabetic, or hypertension-associated renal parenchymal changes, which independently contribute to progressive impairment of renal function.

Our group of patients with long-term follow-up (mean 48.8 months), 156 of greater than 12 months duration, is, to our knowledge, by far the largest group of consecutive pathologically assessed cases studied. The presence of severe arteriosclerosis and/or arteriolosclerosis, IF/TA (>10%), GS (>5%), and type of surgery performed (total or partial) significantly correlated with a greater increase in postoperative creatinine level. Additionally, a significant follow-up increase in creatinine level was found when the groups of partial versus total nephrectomy were compared separately for each degree of arteriolosclerosis: mild, moderate, or severe.

Review of the literature results in 2 types of studies on postnephrectomy renal function: large studies with only clinical variables and a few smaller studies with limited follow-up (Table 5). Clark et al⁷  studied the consequences of partial nephrectomy on renal function where chronic kidney disease stage 3 or greater developed postoperatively in nearly one-third of the patients with an estimated glomerular filtration rate more than 60 mL/min preoperatively, even with a relatively short follow-up, up to 18 months. They also found increased age, female sex, larger tumor size, and lower preoperative glomerular filtration rate to be independently associated with progression to chronic kidney disease stage III from chronic kidney disease stage I or II. Furthermore, Weight et al¹⁵  demonstrated in their large study of patients with renal masses who had undergone extirpative surgery, partial nephrectomy or radical nephrectomy, that partial nephrectomy offers cancer specific survival equivalent to that of radical nephrectomy and is technically feasible in at least 50% of patients with cT1b tumors. These authors¹⁵  also showed better preservation of renal function after partial nephrectomy compared with radical nephrectomy, and, perhaps most importantly, found an increased risk of cardiovascular death and overall death in their patients who progressed to chronic renal insufficiency following nephrectomy. This study, although it did not have an association with patient mortality, did confirm the previous findings of increased incidence of partial nephrectomy and better resulting outcomes compared with total nephrectomy for T1a and T1b tumors.

However, the larger studies with extended follow-up only assessed the predictability of preoperative clinical parameters in determining what variables affect postoperative renal function. The presence of conditions such as older age, hypertension, diabetes, or elevated preoperative creatinine level was used as surrogate for presumptive renal pathology. When pathologic parameters were assessed in studies with fewer cases and shorter follow-up, the postoperative renal function decline was shown to correlate with the amount of GS in 49 radical nephrectomies at a mean follow-up of 19.7 months (range, 3–60.8 months),²  a finding confirmed by this study. A second study showed worse renal follow-up with increased interstitial fibrosis, with more than 20% GS and diffuse diabetic glomerular disease in 39 patients followed for 6 months,⁴  also confirmed.

Our study is unique because it is, to our knowledge, the first to evaluate the pathologic parameters available at the time of nephrectomy with a large sample size, 156 cases with at least 12 months follow-up, as well as an extended follow-up period (mean, 48.4 months; range, 376 days to 11.7 years). In our cohort of both partial and radical nephrectomies, only the specific type of surgery and presence of severe vascular sclerosis, more than 5% GS, and more than 10% IF/TA irrespective of additional nonneoplastic disease were found to contribute toward long-term renal functional impairment. The presence of vascular hyalinosis, clinical diabetes, or diabetic nephropathy was not independently predictive of subsequent renal function decline. The number of patients with diabetic nephropathy and arteriolar hyalinosis was limited (11 and 19, respectively). More cases in the study might have made a difference in the statistical outcome for both variables. The percentage of GS and IF/TA was significantly higher in the kidneys with more advanced vascular disease, indicating an overlap in assessed variables.

Diabetes and hypertension, the 2 leading causes of end-stage renal disease in the United States, are more commonly seen in elderly patients who are also prone to developing renal malignancy. In fact, studies indicate an increase in renal cell carcinoma in patients with diabetes and hypertension as well as with other chronic kidney disease risk factors including smoking, obesity, and older age.^16–18  This suggests that there may be a noncoincidental, causative link and therefore an even higher likelihood of detecting medical kidney disease in patients with renal neoplasm. In addition, although rare, malignancy and specifically renal cell carcinoma have been associated with a number of paraneoplastic glomerular diseases including immunoglobulin A nephropathy, focal segmental glomerulosclerosis, membranous glomerulonephropathy, minimal change disease, secondary amyloidosis, and crescentic glomerulonephritis.^3,4,10,19  We have also identified kidney disease that is unrelated to the tumor and only incidentally present at the time of nephrectomy, such as atheroembolic disease and granulomatous tubulointerstitial nephritis. Often, it is the examination of the nephrectomy specimen that may alert the clinician to the existence of a second medical kidney disease along with the tumor.

The retrospective nature of this study is the principal limitation. Diagnostic kidney biopsies are routinely subjected to immunofluorescence and electron microscopy, which detect unsuspected immune complex deposits, glomerular basement membrane thickening or thinning, and other pathologic processes. Immunofluorescence was performed only in cases with a clinical history of medical renal disease at the time of nephrectomy. Electron microscopy can be performed on deparaffinized tissue, but was used only in cases with light microscopic abnormalities. For instance, early diabetic nephropathy (class I) consists of isolated glomerular basement membrane thickening that would not have been detected.¹⁰  Therefore, it is likely that more than the detected 15% of cases had definable pathology. Limitations of light microscopic examination, including thick sectioning and lack of routine special staining, were also encountered. The length of follow-up and treatment was also not possible to standardize. The increased change in creatinine level for the kidneys with oncocytoma likely reflects an extended follow-up period for those patients as well as a small sample size, with 2 of the 11 patients having diabetic nephropathy. Future prospective controlled studies with larger follow-up should alleviate these issues and yield additional information.

In conclusion, nephrectomized patients are at increased risk for postoperative decline in renal function potentially because of preexisting renal parenchymal changes, secondary to age-related changes, comorbid medical diseases, or unrelated conditions. Postoperative kidney impairment is associated with poor outcome including an increased risk of death, and is likely to rise as neoplasm detection, surgery, and adjunct treatment advances raise the survival of these patients. Therefore, a thorough preoperative clinical assessment and examination of the nonneoplastic portion of tumor nephrectomy specimens using the recommended special stain (PAS or Jones methenamine silver) or other studies (immunohistochemistry/electron microscopy) if necessary may provide valuable diagnostic and prognostic information for appropriate management.