SPECIAL ARTICLE www.jasn.org Pathologic Classification of Diabetic Nephropathy Thijs W. Cohen Tervaert,* Antien L. Mooyaart,* Kerstin Amann,† Arthur H. Cohen,‡ ʈ H. Terence Cook,§ Cinthia B. Drachenberg, Franco Ferrario,¶ Agnes B. Fogo,** Mark Haas,‡ Emile de Heer,* Kensuke Joh,†† Laure H. Noe¨l,‡‡ Jai Radhakrishnan,§§ ʈʈ Surya V. Seshan, Ingeborg M. Bajema,* and Jan A. Bruijn,* on behalf of the Renal Pathology Society *Department of Pathology, Leiden University Medical Center, Leiden, The Netherlands; †Department of Pathology, University of Erlangen-Nuernberg, Erlangen, Germany; ‡Department of Pathology, Cedars-Sinai Medical Center, Los ʈ Angeles, California; §Department of Histopathology, Hammersmith Hospital, London, United Kingdom; Department of Pathology, University of Maryland, Baltimore, Maryland; ¶Renal Immunopathology Center, San Carlo Borromeo Hospital, Milan, Italy; **Department of Pathology, Vanderbilt University Medical Center, Nashville, Tennessee; ††Division of Pathology, Sendai-Shaho Hospital, Sendai City, Japan; ‡‡Department of Pathology, Hoˆpital Necker, Universite´ Rene´ ʈʈ Descartes, Paris, France; §§Department of Medicine, Columbia University, New York, New York; and Department of Pathology and Laboratory Medicine, Weill Cornell Medical College, New York, New York ABSTRACT Although pathologic classifications exist for several renal diseases, including IgA and improve clinical management and nephropathy, focal segmental glomerulosclerosis, and lupus nephritis, a uniform efficiency.5 classification for diabetic nephropathy is lacking. Our aim, commissioned by the In 1959, Gellman et al.6 first reported an Research Committee of the Renal Pathology Society, was to develop a consensus overview and clinical correlation of find- classification combining type1 and type 2 diabetic nephropathies. Such a classifi- ings on renal biopsies from patients with cation should discriminate lesions by various degrees of severity that would be DN. Before their study, the renal pathology easy to use internationally in clinical practice. We divide diabetic nephropathy into in patients with diabetes mellitus was only four hierarchical glomerular lesions with a separate evaluation for degrees of described at autopsy. Gellman proposed interstitial and vascular involvement. Biopsies diagnosed as diabetic nephropathy an elaborate systematic evaluation exam- are classified as follows: Class I, glomerular basement membrane thickening: ining glomeruli, tubules, arterioles, and the isolated glomerular basement membrane thickening and only mild, nonspecific interstitium that was unsuitable for practi- changes by light microscopy that do not meet the criteria of classes II through IV. cal use. More recently, attempts were made Class II, mesangial expansion, mild (IIa) or severe (IIb): glomeruli classified as mild to categorize patterns seen in DN after type or severe mesangial expansion but without nodular sclerosis (Kimmelstiel–Wilson 2 diabetes.7–10 Gambara et al.7 and Fioretto lesions) or global glomerulosclerosis in more than 50% of glomeruli. Class III, et al.8 made basic distinctions between typ- nodular sclerosis (Kimmelstiel–Wilson lesions): at least one glomerulus with nodular ical and atypical DN as well as other glo- increase in mesangial matrix (Kimmelstiel–Wilson) without changes described in merular diseases superimposed on DN.7,8 class IV. Class IV, advanced diabetic glomerulosclerosis: more than 50% global Although such schemes are useful for re- glomerulosclerosis with other clinical or pathologic evidence that sclerosis is at- search biopsies, they also are not practical tributable to diabetic nephropathy. A good interobserver reproducibility for the for clinical use. four classes of DN was shown (intraclass correlation coefficient ϭ 0.84) in a test of this classification. Published online ahead of print. Publication date J Am Soc Nephrol 21: 556–563, 2010. doi: 10.1681/ASN.2010010010 available at www.jasn.org. T.W.C.T. and A.L.M. contributed equally to this work. Diabetic nephropathy (DN) is a major IgA nephropathy,4 yet there is no uni- Correspondence: Dr. Antien L. Mooyaart, Depart- cause of ESRD, and the incidence of dia- form classification for DN. Classification ment of Pathology, Building 1, L1-Q, Leiden Univer- 1 sity Medical Center, PO Box 9600, 2300 RC Leiden, betes mellitus is rising rapidly. Patho- schemes lead to better communication The Netherlands. Phone: 0031715266574; Fax: logic classifications exist for several kid- between renal pathologists and clini- 0031715266952; E-mail: [email protected] 2 ney diseases such as lupus nephritis, cians, provide logistical structure for Copyright ᮊ 2010 by the American Society of focal segmental glomerulosclerosis,3 and prognostic and interventional studies, Nephrology 556 ISSN : 1046-6673/2104-556 J Am Soc Nephrol 21: 556–563, 2010 www.jasn.org SPECIAL ARTICLE We decided to classify DN due to type 1 Table 1. Glomerular classification of DN and type 2 diabetes together because there Class Description Inclusion Criteria is substantial overlap with respect to histo- I Mild or nonspecific LM changes and Biopsy does not meet any of the criteria 11–13 logic lesions and renal complications. EM-proven GBM thickening mentioned below for class II, III, or IV Our aim was to develop a uniform classifi- GBM Ͼ 395 nm in female and Ͼ430 nm cation system containing specific catego- in male individuals 9 years of age and ries that discriminate lesions with various oldera prognostic severities that would be easy to IIa Mild mesangial expansion Biopsy does not meet criteria for class use. This proposal was launched by the Re- III or IV Ͼ search Committee of the Renal Pathology Mild mesangial expansion in 25% of Society in 2006 in San Diego and further the observed mesangium discussed in Leiden in September 2008. IIb Severe mesangial expansion Biopsy does not meet criteria for class III or IV Presented here is a consensus classification Severe mesangial expansion in Ͼ25% of of DN developed by a group of interna- the observed mesangium tional experts. III Nodular sclerosis (Kimmelstiel– Biopsy does not meet criteria for class Wilson lesion) IV At least one convincing Kimmelstiel– CLASSIFICATION OF DN Wilson lesion IV Advanced diabetic Global glomerular sclerosis in Ͼ50% of It is essential to evaluate renal tissue us- glomerulosclerosis glomeruli ing appropriate standards for renal bi- Lesions from classes I through III opsy. These include hematoxylin and eo- LM, light microscopy. aOn the basis of direct measurement of GBM width by EM, these individual cutoff levels may be sin, periodic acid–Schiff (PAS), Masson considered indicative when other GBM measurements are used. trichrome, and periodic acid methena- mine silver stains for light microscopy. Biopsies should contain at least 10 glo- the absence of mesangial expansion, GBM and epithelial foot process efface- meruli,14 excluding incomplete glomer- nodular increases in mesangial matrix ment by EM have no influence on the uli along the biopsy edge. Immunofluo- (Kimmelstiel–Wilson lesions), and classification. Class I incorporates cases rescence requires the use of antibodies global glomerulosclerosis of more than that have been called “normal or near against IgA, IgG, IgM, C3, C1q, and 50% of glomeruli] the biopsy is as- normal DN” by Fioretto et al.,8 but in our kappa and lambda light chains to rule signed to class I (Table 1 and Figure 1), system, a certain degree of chronic and out other renal diseases. Electron mi- in which by direct measurements with other reactive changes (e.g., changes of croscopy (EM) must be performed; spe- EM the glomerular basement mem- arterionephrosclerosis, ischemic type cific guidelines are discussed below. All brane (GBM) on average is thicker changes, or interstitial fibrosis) are ac- of these methods are necessary for an ac- than 430 nm in males 9 years and older cepted as part of this category. Diagnos- curate diagnosis of DN. DN should never and thicker than 395 nm in females. ing DN in cases without characteristic be diagnosed without supportive clinical These cutoff levels are based on a devi- light microscopic glomerular lesions information, and a patient should carry a ation from normal GBM thickness plus may be difficult, especially when a clinical diagnosis of diabetes mellitus to 2 standard deviations as recently deter- thicker GBM is also seen with aging or apply the classification. Furthermore, mined by Haas.16 For children younger hypertension. The presence of arteriolar virtually any glomerular disease can ac- than 9 years old, we refer to Table 1 in hyalinosis may be helpful in these cases, company DN, postinfectious GN9,15 and the paper by Haas.16 Upper limits for although it is not a prerequisite. membranous glomerulopathy being the normal GBM thickness vary with the GBM thickening is a characteristic most common11; thus, any coexisting methods used to measure GBM width. early change in type 118–20 and type 2 disorders should also be described. Four For instance, using the orthogonal in- DN13 and increases with duration of dis- classes of glomerular lesions in DN are tercept method, upper limits of normal ease.21 GBM thickening is a consequence presented in Table 1. GBM thickness are 520 nm for adult of extracellular matrix accumulation, men and 471 nm for women.17 In indi- with increased deposition of normal ex- Classes of Glomerular Lesions vidual laboratories, the upper limits for tracellular matrix components such as Class I: Glomerular Basement Membrane normal GBM thickness