NDT Advance Access originally published online on February 3, 2007
Nephrology Dialysis Transplantation 2007 22(4):1228-1232; doi:10.1093/ndt/gfl838
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Thin basement membrane nephropathy cannot be diagnosed reliably in deparaffinized, formalin-fixed tissue
1Department of Pathology and 2Department of Medicine, Division of Nephrology, Columbia University, College of Physicians and Surgeons, New York, NY, USA
Correspondence and offprint requests to: Samih H. Nasr, Department of Pathology, Columbia University, College of Physicians and Surgeons, 630 West 168th Street, VC14-224, New York, NY 10032. Email: sn386{at}columbia.edu
 |
Abstract |
|---|
 Top Abstract IntroductionMaterial and methodsResultsDiscussionReferences |
|---|
In diagnostic renal pathology, electron microscopy is ideally performed on glutaraldehyde-fixed, plastic resin-embedded tissue (EM-G). When no glomeruli are present in the portion of the biopsy fixed in glutaraldehyde, formalin-fixed, paraffin-embedded tissue can be reprocessed for electron microscopy (EM-F). The usefulness of this salvage technique for the diagnosis of thin basement membrane nephropathy (TBMN) has not been studied systematically. Here we compare the glomerular basement membrane (GBM) thickness by EM-G vs EM-F in 21 renal biopsies, including TBMN (eight patients), normals (two patients), minimal change disease (MCD) (six patients) and diabetic nephropathy (DN) (five patients). There was significant reduction of the GBM thickness by EM-F compared with EM-G across all diagnostic categories in all 21 cases. The mean percentage reduction in GBM thickness was 23% for the TBMN cases, 40% for the normal/MCD cases and 34% for the DN cases. Four patients with MCD had a mean GBM thickness by EM-F that fell below the defining threshold for diagnosis of TBMN. For the TBMN cases, the 99th percentile for GBM thickness by EM-F was 194 nm, suggesting that the diagnosis of TBMN by EM-F can be excluded with confidence if the GBM thickness is above 200 nm. No clear criteria could be established to diagnose TBMN by EM-F. Renal pathologists should be aware that reprocessing of paraffin tissue for EM causes artifactual GBM thinning that precludes accurate diagnosis of TBMN.
Keywords: formalin-fixed tissue; glomerular basement membrane thinning; thin basement membrane nephropathy
|
Introduction |
|---|
|
TopAbstractIntroductionMaterial and methodsResultsDiscussionReferences |
|---|
Glutaraldehyde is the ideal tissue fixative for ultrastructural examination of renal biopsies. When no glomeruli are present in the portion of a biopsy fixed in glutaraldehyde, electron microscopy (EM) can be performed on formalin-fixed paraffin-embedded tissue processed for light microscopy. Although this salvage technique often results in artifactual distortion of cellular and extracellular matrix architecture, it is usually adequate for detection of many renal ultrastructural alterations such as the presence and sites of electron-dense deposits and the degree of foot process effacement.
Thin basement membrane nephropathy (TBMN) is a glomerular disease that can only be diagnosed by ultrastructural analysis. We previously encountered a case of a teenage male who underwent nephrectomy for persistent gross haematuria due to vascular malformation of the kidney. EM was performed on glutaraldehyde-fixed tissue and showed a mean glomerular basement membrane (GBM) thickness of 280 nm (within normal range). One year prior, the patient had had a renal biopsy for haematuria in which EM was performed on deparaffinized, formalin-fixed tissue. By that technique the mean GBM thickness was 175 nm, suggesting a renal biopsy diagnosis of thin basement membrane nephropathy (TBMN). Our experience in this case prompted us to perform a systematic study of how EM on deparaffinized formalin-fixed tissue (EM-F) affects GBM thickness compared with EM on glutaraldehyde-fixed, Epoxy-embedded tissue (EM-G).
|
Material and methods |
|---|
|
TopAbstractIntroductionMaterial and methodsResultsDiscussionReferences |
|---|
Electron microscopy was performed retrospectively on archived formalin-fixed, paraffin-embedded tissue blocks from 21 renal biopsies (stored for up to 4 years). The processing protocol is detailed in Table 1. Ultrastructural evaluation was performed using a JEOL 100S electron microscope (JEOL, Peabody, MA, USA).
|
Electron microscopy on glutaraldehyde-fixed, plastic resin-embedded tissue (EM-G) had been performed previously for diagnostic purposes on all 21 biopsies. One- millimetre cubes of tissue were immediately fixed in 2.5% glutaraldehyde in 0.1 cacodylate buffer (pH 7.40) for 24 h. The tissue was then washed in cacodylate buffer, post-fixed in aqueous osmium tetroxide, dehydrated and processed as described for EM-F.
The 21 renal biopsies studied by both EM-F and EM-G consisted of eight biopsies of TBMN, two histologically normal (from patients without haematuria or diabetes), six minimal change disease (MCD) and five diabetic nephropathy (DN). A minimum of two glomeruli was studied ultrastructurally by EM-F and by EM-G. GBM measurements were performed using a magnification graticule (Graticules Limited, Tonbridge, Kent, England) placed over the electron micrographs. Glomeruli studied by EM-F frequently showed artifactual detachment of podocyte foot processes from the underlying GBM and/or poor preservation of the lamina rara interna and lamina rara externa of GBM. Therefore, only the central lamina densa of GBM, which was well preserved in most capillaries, was measured by EM-F and EM-G. For each biopsy, a total of 80 GBM measurements were made on EM-F and 80 measurements on EM-G. The measurements were performed at multiple representative points along randomly selected capillaries. The portions of GBM where it reflects over the mesangium, where it is cut tangentially, and where its detail is not well preserved (in the case of EM-F) were avoided. The GBM thickness in nanometres was then calculated according to the following equation:
|
|
Statistical analysis was performed using exact non-parametric methods. For pairwise comparisons, the Wilcoxon signed rank test was used. For comparison among multiple groups, the Kruskal–Wallis test was used. Statistical significance was assumed at P < 0.05.
|
Results |
|---|
|
TopAbstractIntroductionMaterial and methodsResultsDiscussionReferences |
|---|
Reprocessing and dehydration of formalin-fixed, paraffin-embedded tissue for EM caused significant artifactual GBM thinning in all 21 cases studied (Figure 1). The results of GBM morphometry are summarized in Table 2. When GBM thickness by EM-F was compared with EM-G for each biopsy, the mean percentage reduction in GBM thickness was 23% for the TBMN group (P = 0.004), 40% for the normal/MCD group (P = 0.004), and 34% for the DN group (P = 0.031).
|
|
For the TBMN cases, the 99th percentile for GBM thickness on EM-F was 194 nm. This means that for a GBM thickness above 194 nm by EM-F, there is <1% chance that the patient would have TBMN.
Importantly, out of the eight cases of normal/MCD studied, four (three males, one female) had a mean GBM thickness by EM-F that fell below the threshold previously established by our laboratory to define TBMN (250 nm in males, 220 nm in females) [1] (Table 2).
The percentage of GBM thickness reduction on EM-F compared with EM-G was statistically different between the normal/MCD group (40%) and the TBMN group (23%) (P < 0.001), but did not reach statistical difference between the DN group (34%) and the TBMN group (P = 0.093), and between the normal/MCD group and the DN group.
|
Discussion |
|---|
|
TopAbstractIntroductionMaterial and methodsResultsDiscussionReferences |
|---|
TBMN is a glomerular disease characterized clinically by isolated haematuria and frequently referred to as ‘benign familial haematuria’. Glomeruli appear unremarkable by light microscopy and no glomerular deposition of immunoglobulins or complement components is detected on immunofluorescence. The pathological diagnosis can only be established by electron microscopy. The defining ultrastructural feature is diffuse thinning of the GBM, in the absence of other significant glomerular alterations. The GBM attenuation primarily affects the lamina densa with relative preservation of the lamina rara interna and externa. The World Health Organization (WHO) uses a mean GBM thickness of <250 nm for adults and 180 nm for children aged 2–11 years to define TBMN [2].
GBM thickness varies widely in the general population and can be influenced by the method of measurement, fixation conditions and the embedding medium for electron microscopy [1]. Osawa et al. [3] found that tissue embedding in vestopal, a styrene-polyester-based resin of high viscosity results in thinner GBM compared with methacrylate and araldite resins. In animal studies, renal perfusion pressures also have been shown to influence GBM thickness [4]. In electron microscopic studies performed on rats subjected to different perfusion pressures, Yu et al. [4] found that the GBM became thinner when the perfusion pressures were increased from 100 cm H2O to 250 cm H2O.
In this report, we demonstrate that GBMs appear significantly thinner on EM-F compared with EM-G in all renal conditions studied, including TBMN, DN, MCD and normal controls. Thus, the differences in GBM thickness observed for EM-F and EM-G could not be attributed to intrinsic disease-related differences. In addition, these differences could not be explained by any specific technical modifications over the period of sample collection, because our laboratory's electron microscopy personnel and the reagents and protocols used for both routine electron microscopy and deparaffinization have been constant.
Our findings have major practical implications for the diagnosis of TBMN. Importantly, the mean GBM thickness on EM-F was far below 200 nm in four of the six cases of MCD studied. This measurement is below the threshold previously established in our laboratory to define thin GBM (250 nm in males and 220 nm in females) and below the level suggested by the WHO (250 nm in adults) [1,2]. If the tissue processing had not been taken into consideration, these four cases would have been incorrectly diagnosed as TBMN. Collar and Cattell [5] compared the GBM thickness on EM-G vs EM-F in a single patient with asymptomatic haematuria. They found a significant reduction in GBM thickness by EM-F compared with EM-G (230 nm vs 323 nm), in agreement with our findings, and concluded that EM-F is not suitable for diagnosis of TBMN. For the eight cases of TBMN studied herein, the 99th percentile for GBM thickness on EM-F was 194 nm. This indicates that for a GBM thickness above 194 nm by EM-F, there is <1% chance that the patient would have TBMN. Hence, based on our findings, the diagnosis of TBM by EM-F can be excluded with confidence if the GBM thickness is above 200 nm.
Sensitivity and specificity analyses were performed using the 80 individual measurements performed for each of the 16 cases in the categories of normal/MCD and TBMN. A cut-off of 165 nm in EM-F gave the best combination of sensitivity (74.8%) and specificity (74.1%) for a diagnosis of TBMN. All eight cases of TBMN had a mean GBM diameter below this value compared with only one of eight cases in the normal/MCD group. No cut-off gave both a sensitivity and specificity greater than 75% because of the large dispersion of the individual GBM measurements in any case. Furthermore, the positive predictive value of a GBM thickness <165 nm is only 74.3% while the negative predictive value of a GBM thickness >165 nm is 74.6%. Based on these results, we conclude that a cut-off of 165 nm in EM-F cannot be used as a diagnostically reliable basis for diagnosis of TBMN.
Why the GBM appears thinner on EM-F compared with EM-G is not clear. Formaldehyde is a less potent tissue fixative than glutaraldehyde because it has only one aldehyde group (–CHO) which binds to nitrogen atoms of amino acids, resulting in protein cross-linking. Therefore, some GBM components, particularly proteoglycans, could potentially be extracted and lost during the succeeding ethanol dehydration and xylene clearing. Wang and Minassian [6] compared the fine structures of tissue on EM-G vs EM-F in eight non-neoplastic and five neoplastic surgical specimens. They found that the least preserved structures on EM-F were the loose components of the cells such as lipid, glucose and ribosomes. They attributed these differences largely to fixation effects because they could find no alterations in fine structure resulting from the paraffin embedding itself. Another potential explanation for GBM thinning on EM-F is increased tissue dehydration, resulting in extraction of water molecules trapped within the GBM matrix. The procedure of EM-F includes an extra alcohol dehydration step before paraffin embedding.
In conclusion, we have demonstrated that the range of normal GBM thickness is lower in deparaffinized, formalin fixed tissue than glutaraldehyde fixed tissue. As a result, the usual criteria for diagnosis of TBMN cannot be applied to reprocessed tissue. Renal pathologists need to be aware of this potential diagnostic pitfall. EM-F can be used to exclude, but not to establish, a diagnosis of TBMN.
Conflict of interest statement. None declared.
|
References |
|---|
|
TopAbstractIntroductionMaterial and methodsResultsDiscussionReferences |
|---|
- Foster K, Markowitz GS, D’Agati VD. (2005) Pathology of thin basement membrane nephropathy. Semin Nephrol 25:149–158.[CrossRef][Web of Science][Medline]
- Churg J, Bernstein J, Glassock RJ. (1995) Renal Disease: Classification and Atlas of Glomerular Diseases 2nd edn, (Igaku-Shoin, New York, NY) pp. ; 411–412.
- Osawa G, Kimmelstiel P, Seiling V. (1966) Thickness of glomerular basement membranes. Am J Clin Pathol 45:7–20.[Web of Science][Medline]
- Yu Y, Leng CG, Kato Y, Ohno S. (1997) Ultrastructural study of glomerular capillary loops at different perfusion pressures as revealed by quick-freezing, freeze-substitution and conventional fixation methods. Nephron 76:452–459.[Web of Science][Medline]
- Collar J and Cattell V. (1995) Paraffin-processed material is unsuitable for diagnosis of thin-membrane disease. Nephron 69:187–188.[Web of Science][Medline]
- Wang NS and Minassian H. (1987) The formaldehyde-fixed and paraffin-embedded tissues for diagnostic transmission electron microscopy: a retrospective and prospective study. Hum Pathol 18:715–727.[CrossRef][Web of Science][Medline]
Accepted in revised form: 26.12.06
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||