nefropati diabetikum.pdf

Pathologic Classification of Diabetic Nephropathy

Thijs W. Cohen Tervaert,* Antien L. Mooyaart,* Kerstin Amann,† Arthur H. Cohen,‡

H. Terence Cook,§ Cinthia B. Drachenberg,� Franco Ferrario,¶ Agnes B. Fogo,**Mark Haas,‡ Emile de Heer,* Kensuke Joh,†† Laure H. Noel,‡‡ Jai Radhakrishnan,§§

Surya V. Seshan,� � Ingeborg M. Bajema,* and Jan A. Bruijn,* on behalf of the RenalPathology 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, LosAngeles, California; §Department of Histopathology, Hammersmith Hospital, London, United Kingdom; �Department ofPathology, University of Maryland, Baltimore, Maryland; ¶Renal Immunopathology Center, San Carlo Borromeo Hospital,Milan, Italy; **Department of Pathology, Vanderbilt University Medical Center, Nashville, Tennessee; ††Division ofPathology, Sendai-Shaho Hospital, Sendai City, Japan; ‡‡Department of Pathology, Hopital Necker, Universite ReneDescartes, Paris, France; §§Department of Medicine, Columbia University, New York, New York; and � �Department ofPathology and Laboratory Medicine, Weill Cornell Medical College, New York, New York

Diabetic nephropathy (DN) is a majorcause of ESRD, and the incidence of dia-betes mellitus is rising rapidly.1 Patho-logic classifications exist for several kid-ney diseases such as lupus nephritis,2

focal segmental glomerulosclerosis,3 and

IgA nephropathy,4 yet there is no uni-form classification for DN. Classificationschemes lead to better communicationbetween renal pathologists and clini-cians, provide logistical structure forprognostic and interventional studies,

and improve clinical management andefficiency.5

In 1959, Gellman et al.6 first reported anoverview and clinical correlation of find-ings on renal biopsies from patients withDN. Before their study, the renal pathologyin patients with diabetes mellitus was onlydescribed at autopsy. Gellman proposedan elaborate systematic evaluation exam-ining glomeruli, tubules, arterioles, and theinterstitium that was unsuitable for practi-cal use. More recently, attempts were madeto categorize patterns seen in DN after type2 diabetes.7–10 Gambara et al.7 and Fiorettoet al.8 made basic distinctions between typ-ical and atypical DN as well as other glo-merular diseases superimposed on DN.7,8

Although such schemes are useful for re-search biopsies, they also are not practicalfor clinical use.

Published online ahead of print. Publication dateavailable at www.jasn.org.

T.W.C.T. and A.L.M. contributed equally to thiswork.

Correspondence: Dr. Antien L. Mooyaart, Depart-ment of Pathology, Building 1, L1-Q, Leiden Univer-sity Medical Center, PO Box 9600, 2300 RC Leiden,The Netherlands. Phone: 0031715266574; Fax:0031715266952; E-mail: [email protected]

Copyright � 2010 by the American Society ofNephrology

ABSTRACTAlthough pathologic classifications exist for several renal diseases, including IgAnephropathy, focal segmental glomerulosclerosis, and lupus nephritis, a uniformclassification for diabetic nephropathy is lacking. Our aim, commissioned by theResearch Committee of the Renal Pathology Society, was to develop a consensusclassification combining type1 and type 2 diabetic nephropathies. Such a classifi-cation should discriminate lesions by various degrees of severity that would beeasy to use internationally in clinical practice. We divide diabetic nephropathy intofour hierarchical glomerular lesions with a separate evaluation for degrees ofinterstitial and vascular involvement. Biopsies diagnosed as diabetic nephropathyare classified as follows: Class I, glomerular basement membrane thickening:isolated glomerular basement membrane thickening and only mild, nonspecificchanges by light microscopy that do not meet the criteria of classes II through IV.Class II, mesangial expansion, mild (IIa) or severe (IIb): glomeruli classified as mildor severe mesangial expansion but without nodular sclerosis (Kimmelstiel–Wilsonlesions) or global glomerulosclerosis in more than 50% of glomeruli. Class III,nodular sclerosis (Kimmelstiel–Wilson lesions): at least one glomerulus with nodularincrease in mesangial matrix (Kimmelstiel–Wilson) without changes described inclass IV. Class IV, advanced diabetic glomerulosclerosis: more than 50% globalglomerulosclerosis with other clinical or pathologic evidence that sclerosis is at-tributable to diabetic nephropathy. A good interobserver reproducibility for thefour classes of DN was shown (intraclass correlation coefficient � 0.84) in a test ofthis classification.

J Am Soc Nephrol 21: 556–563, 2010. doi: 10.1681/ASN.2010010010

SPECIAL ARTICLE www.jasn.org

556 ISSN : 1046-6673/2104-556 J Am Soc Nephrol 21: 556–563, 2010

We decided to classify DN due to type 1and type 2 diabetes together because thereis substantial overlap with respect to histo-logic lesions and renal complications.11–13

Our aim was to develop a uniform classifi-cation system containing specific catego-ries that discriminate lesions with variousprognostic severities that would be easy touse. This proposal was launched by the Re-search Committee of the Renal PathologySociety in 2006 in San Diego and furtherdiscussed in Leiden in September 2008.Presented here is a consensus classificationof DN developed by a group of interna-tional experts.

CLASSIFICATION OF DN

It is essential to evaluate renal tissue us-ing appropriate standards for renal bi-opsy. These include hematoxylin and eo-sin, periodic acid–Schiff (PAS), Massontrichrome, and periodic acid methena-mine silver stains for light microscopy.Biopsies should contain at least 10 glo-meruli,14 excluding incomplete glomer-uli along the biopsy edge. Immunofluo-rescence requires the use of antibodiesagainst IgA, IgG, IgM, C3, C1q, andkappa and lambda light chains to ruleout other renal diseases. Electron mi-croscopy (EM) must be performed; spe-cific guidelines are discussed below. Allof these methods are necessary for an ac-curate diagnosis of DN. DN should neverbe diagnosed without supportive clinicalinformation, and a patient should carry aclinical diagnosis of diabetes mellitus toapply the classification. Furthermore,virtually any glomerular disease can ac-company DN, postinfectious GN9,15 andmembranous glomerulopathy being themost common11; thus, any coexistingdisorders should also be described. Fourclasses of glomerular lesions in DN arepresented in Table 1.

Classes of Glomerular LesionsClass I: Glomerular Basement MembraneThickening.If the biopsy specimen shows no oronly mild, nonspecific changes by lightmicroscopy that do not meet the crite-ria of classes II through IV [in effect, in

the absence of mesangial expansion,nodular increases in mesangial matrix(Kimmelstiel–Wilson lesions), andglobal glomerulosclerosis of more than50% of glomeruli] the biopsy is as-signed to class I (Table 1 and Figure 1),in which by direct measurements withEM the glomerular basement mem-brane (GBM) on average is thickerthan 430 nm in males 9 years and olderand thicker than 395 nm in females.These cutoff levels are based on a devi-ation from normal GBM thickness plus2 standard deviations as recently deter-mined by Haas.16 For children youngerthan 9 years old, we refer to Table 1 inthe paper by Haas.16 Upper limits fornormal GBM thickness vary with themethods used to measure GBM width.For instance, using the orthogonal in-tercept method, upper limits of normalGBM thickness are 520 nm for adultmen and 471 nm for women.17 In indi-vidual laboratories, the upper limits fornormal GBM thickness have usuallybeen established, and if other methodsthan direct GBM measurement areused, it is advised to use these locallyestablished cutoff points.

Light microscopic changes in the

GBM and epithelial foot process efface-ment by EM have no influence on theclassification. Class I incorporates casesthat have been called “normal or nearnormal DN” by Fioretto et al.,8 but in oursystem, a certain degree of chronic andother reactive changes (e.g., changes ofarterionephrosclerosis, ischemic typechanges, or interstitial fibrosis) are ac-cepted as part of this category. Diagnos-ing DN in cases without characteristiclight microscopic glomerular lesionsmay be difficult, especially when athicker GBM is also seen with aging orhypertension. The presence of arteriolarhyalinosis may be helpful in these cases,although it is not a prerequisite.

GBM thickening is a characteristicearly change in type 118 –20 and type 2DN13 and increases with duration of dis-ease.21 GBM thickening is a consequenceof extracellular matrix accumulation,with increased deposition of normal ex-tracellular matrix components such ascollagen types IV and VI, laminin, andfibronectin.22,23 Such accumulations re-sult from increased production of theseproteins, their decreased degradation, ora combination of the two. GBM thicken-ing may already be present in type 1 dia-

Table 1. Glomerular classification of DN

Class Description Inclusion Criteria

I Mild or nonspecific LM changes andEM-proven GBM thickening

Biopsy does not meet any of the criteriamentioned below for class II, III, or IV

GBM � 395 nm in female and �430 nmin male individuals 9 years of age andoldera

IIa Mild mesangial expansion Biopsy does not meet criteria for classIII or IV

Mild mesangial expansion in �25% ofthe observed mesangium

IIb Severe mesangial expansion Biopsy does not meet criteria for classIII or IV

Severe mesangial expansion in �25% ofthe observed mesangium

III Nodular sclerosis (Kimmelstiel–Wilson lesion)

Biopsy does not meet criteria for classIV

At least one convincing Kimmelstiel–Wilson lesion

IV Advanced diabeticglomerulosclerosis

Global glomerular sclerosis in �50% ofglomeruli

Lesions from classes I through IIILM, light microscopy.aOn the basis of direct measurement of GBM width by EM, these individual cutoff levels may beconsidered indicative when other GBM measurements are used.

SPECIAL ARTICLEwww.jasn.org

J Am Soc Nephrol 21: 556–563, 2010 Pathologic Classification of Diabetic Nephropathy 557

betes patients who are normoalbumin-uric.20,21 GBM thickening has even beendescribed as a “prediabetic” lesion: In pa-tients with proteinuria and isolated GBM

thickening but without overt diabetes,20% were positive on a blood test for di-abetes at the time of biopsy, whereas 44%were diagnosed with diabetes at 6

months, and 70% at 2 years after the bi-opsy was taken.24 Long-term glucosecontrol and urinary albumin excretion(UAE) correlate strongly with basementmembrane thickness.25

In 1979, Jensen et al.26 were amongthe first to measure GBM thickness usingthe orthogonal intercept method. Inbrief, a grid with eight evenly spaced in-tersecting lines (four horizontal and fourvertical) is placed over a photomicro-graph, and GBM measurements aremade at each point that a line on the gridintercepts an endothelial-GBM inter-face. Currently, some laboratories usecomputer-assisted measurements bywhich the mean width is calculated fromapproximately 50 measurements of theGBM at five different locations. TheGBM width is then compared with GBMwidth from normal subjects, as deter-mined previously by Steffes et al.27 andrecently updated by Haas.16 Ideally, glu-taraldehyde-fixed, plastic resin-embed-ded tissue should be used for EM, keep-ing in mind that other methods,particularly the reprocessing of paraffintissue for EM, may cause artifactualGBM thinning as recently reported byNasr et al.28 If computer-assisted mea-surements are not available, we recom-mend doing direct GBM measurementsas recently modified by Haas.16

Class II: Mesangial Expansion, Mild (IIa)or Severe (IIb).Class II encompasses those patients clas-sified with mild or severe mesangial ex-pansion but not meeting inclusion crite-ria for class III or IV (Table 1 and Figure1) and is analogous to the previouslyused term “diffuse diabetic glomerulo-sclerosis.” Mesangial expansion is de-fined as an increase in extracellular ma-terial in the mesangium such that thewidth of the interspace exceeds two mes-angial cell nuclei in at least two glomeru-lar lobules. The difference between mildand severe mesangial expansion is basedon whether the expanded mesangial areais smaller or larger than the mean area ofa capillary lumen. If severe mesangial ex-pansion is seen in more than 25% of thetotal mesangium observed throughoutthe biopsy, the biopsy is classified as IIb.

Figure 1. Representative examples of the morphologic lesions in DN. (A) Glomerulusshowing only mild ischemic changes, with splitting of Bowman’s capsule. No clearmesangial alteration. (B) EM of this glomerulus: the mean width of the GBM was 671 nm(mean taken over 55 random measurements). EM provides the evidence for classifyingthe biopsy with only mild light microscopic changes into class I. (C, D) Class II glomeruliwith mild and moderate mesangial expansion, respectively. In panel C, the mesangialexpansion does not exceed the mean area of a capillary lumen (IIa), whereas in panel Dit does (IIb). (E, F) In panel F is a class III Kimmelstiel–Wilson lesion. The lesion in panelE is not a convincing Kimmelstiel–Wilson lesion, therefore (on the basis of the findings inthis glomerulus) the finding is consistent with class IIb. For the purpose of the classifica-tion, at least one convincing Kimmelstiel–Wilson (as in panel F) needs to be present. Inpanel H, signs of class IV DN consist of hyalinosis of the glomerular vascular pole and aremnant of a Kimmelstiel–Wilson lesion on the opposite site of the pole. Panel G is anexample of glomerulosclerosis that does not reveal its cause (glomerulus from the samebiopsy as panel H). For the purpose of the classification, signs of DN should be his-topathologically or clinically present to classify a biopsy with global glomerulosclerosis in�50% of glomeruli as class IV.


558 Journal of the American Society of Nephrology J Am Soc Nephrol 21: 556–563, 2010

If this is not the case, but at least mildmesangial expansion is seen in morethan 25% of the total mesangium, the bi-opsy is classified as IIa.

Expansion of cellular and matrixcomponents in the mesangium is a hall-mark of type 1 and type 2 DN.13,18 It canbe detected in some patients within a fewyears after the onset of type 1 diabetes.20

When the mesangium expands, it re-stricts and distorts glomerular capillariesand diminishes the capillary filtrationsurface. In our classification, we do notdistinguish between mesangial hypercel-lularity, matrix expansion, or “mesan-giosclerosis”—any expansion of themesangium that conforms to our defini-tions given above and in Table 1 belongsto class II.

Various indices have been proposedto describe the amount of mesangial ex-pansion in DN. Mauer et al.18 definemesangial expansion by mesangial frac-tional volume or volume density (VV),defined as the fraction or percentage ofthe cross-sectional area of the glomerulartuft made up by mesangium, expressedin the formula: Vv(mes/glom).18 Usingthis formula, many correlations havebeen made between mesangial expansionand clinical parameters of DN, particu-larly showing highly inverse correla-tions exist between Vv(mes/glom) andGFR.18,29,30 There is also a relationshipbetween Vv(mes/glom) and UAE18,29

and blood pressure.31

Another index to express mesangialexpansion is the so-called “index of mes-angial expansion” (IME) for DN.18 TheIME is determined by a semiquantitativeestimate of the width of mesangial zonesin each glomerulus18: grade 0 is normal, 1is twice normal thickness, 2 is three timesnormal thickness, and so forth; halfgrades can also be assigned. The mean ofthe grades for each glomerulus for IMEcan thus be determined from a single bi-opsy. The IME closely correlates with theVv(mes/glom).18 Still, this is a ratherelaborate method.

In other classifications, mesangial ex-pansion is defined in more practicalways, such as in the new classification forIgA nephropathy in which it is defined asan increase in the extracellular material

in the mesangium such that the width ofthe interspace exceeds two mesangial cellnuclei in at least two glomerular lobules.4

Because interobserver agreement wastested in this study and found to be sat-isfactory using this definition, we de-cided to use the same definition for mes-angial expansion in our classificationfor DN.

Class III: Nodular Sclerosis (Kimmelstiel–Wilson lesions).If at least one convincing Kimmelstiel–Wilson lesion is found and the biopsyspecimen does not have more than 50%global glomerulosclerosis it is classifiedas class III (Table 1 and Figure 1). Kim-melstiel–Wilson lesions appear in type 1and type 2 diabetes as focal, lobular,round to oval mesangial lesions with anacellular, hyaline/matrix core, roundedperipherally by sparse, crescent-shapedmesangial nuclei.32

Paul Kimmelstiel and Clifford Wil-son, a German and an Englishman whomet at Harvard, first described nodularlesions in glomeruli from eight maturity-onset diabetes patients in 1936.33 Ac-cording to Cameron,34 they barely notedthe association with diabetes, and it wasArthur Allen who clarified the associa-tion in 105 patients with diabetes in1941.35 Nodular sclerotic lesions mayalso occur in the absence of DN that areclinically related to hypertension, smok-ing, hypercholesterolemia, and extrare-nal vascular disease.36

It is claimed that in the initial stage ofdeveloping nodular sclerotic lesions inDN, two important processes take place:lytic changes in the mesangial area calledmesangiolysis and detachment of endo-thelial cells from the GBM.37 Exactly howthese two processes relate remains uncer-tain. Paueksakon et al.38 detected frag-mented red blood cells in Kimmelstiel–Wilson lesions, which supports thetheory that microvascular injury con-tributes to the pathogenesis of these le-sions. Dissociation of endothelial cellsmay disrupt the connections between themesangial area and the GBM. This pro-cess precedes expansion of the Kimmel-stiel–Wilson lesion.37 These lesions con-sist of an accumulation of mesangial

matrix with collagen fibrils, small lipidparticles, and cellular debris.39 A com-pletely developed Kimmelstiel–Wilsonlesion destroys the normal structure ofglomerular tuft with a decrease in mes-angial cells, especially in the centralarea.37 In 1992, a graphic method of analy-sis of the position of Kimmelstiel–Wilsonlesions demonstrated the nodules were dis-tributed in a horseshoe-shaped area corre-sponding to the peripheral or intralobularmesangium,40 excluding the possibility ofhyperfiltration as being their main cause ofdevelopment.

The presence of at least one Kim-melstiel–Wilson lesion associates withlonger duration of diabetes and lessfavorable clinical parameters.10,41 In astudy of 36 patients with type 2 diabetes,patients with Kimmelstiel–Wilson le-sions had more severe overall retinopa-thy and higher serum creatinine concen-trations than those with mesangiallesions alone.10 In a study of 124 Chinesepatients with type 2 diabetes, patientswith at least one Kimmelstiel–Wilson le-sion had relatively long duration of dia-betes mellitus, a poor prognosis, and fre-quent evidence of diabetic retinopathy.41

Kimmelstiel–Wilson lesions are oftenfound in combination with mesangialexpansion. The occurrence of Kimmel-stiel–Wilson lesions is widely consideredtransitional from an early or moderatelyadvanced stage to a progressively moreadvanced stage of disease.41,42 Therefore,in our classification, the occurrence ofKimmelstiel–Wilson lesions implies aseparate class.

Class IV: Advanced Diabetic Glomerulo-sclerosis.Class IV implies advanced DN and des-ignates those biopsies with more than50% global glomerulosclerosis in whichthere is clinical or pathologic evidencethat the sclerosis is attributable to DN(Table 1 and Figure 1). Glomerulosclero-sis in DN is the end point of multifac-torial mechanisms that lead to excessiveaccumulation of extracellular matrixproteins such as collagen types I, III,and IV and fibronectin in the mesangialspace, which through stages of mesangialexpansion and development of Kimmel-



stiel–Wilson lesions finally result in glo-merulosclerosis.43 The clustering of scle-rotic lesions in columns perpendicular tothe kidney surface suggests that vascularfactors relating to the interlobular arter-ies also contribute.44

Designation of class IV lesions inour classification system is restricted tothose cases in which there is evidencefor DN. This evidence can come fromother lesions in the biopsy as describedfor classes I through III. The occur-rence of hyalinosis of the glomerularvascular pole or a capsular drop mayalso be taken as evidence for the pres-ence of DN. Alternatively, if DN is thelikely clinical diagnosis (e.g., by thepresence of retinopathy) a biopsy withextensive glomerulosclerosis can alsobe classified as class IV. Glomeruloscle-rosis without evidence of DN should bementioned as such in the conclusion ofthe pathology report but should not beassigned class IV.

Tubulointerstitial Lesions, VascularLesions, and NondiabeticGlomerular LesionsTubular Lesions.Concomitant tubular basement mem-brane thickening of nonatrophic tubules isapparent from the development of class IIglomerular diabetic lesions and becomesmore conspicuous in class III and IV,which is best seen in PAS or silver stains.

Interstitial fibrosis and tubular atro-phy (IFTA) follow glomerular changesin type 1 DN that ultimately lead toESRD.30 We score IFTA together as apercentage of the total involved area ofinterstitium and tubules (Table 2). Ascore of 0 is assigned when the biopsyspecimen shows no IFTA, a score of 1 isassigned when less than 25% IFTA ispresent, a score of 2 is assigned when atleast 25% but less than 50% of the biopsyhas IFTA, and finally, a score of 3 is as-signed when at least 50% IFTA is present,which is similar to the scoring in the re-cently published classification of IgA ne-phropathy.4

Presence of mononuclear cells in theinterstitium is a widely recognized find-ing in DN. Inflammatory interstitial in-filtrates comprise T lymphocytes and

macrophages.45 In Table 2, we score 0 ifinterstitial infiltrates are absent, 1 if theyonly occur around atrophic tubules, and2 if the inflammatory infiltrate is also inother areas than around atrophic tu-bules.

Vascular Lesions.According to Stout et al.,46 hyalinosis ofthe efferent arteriole is relatively specificfor DN, but hyalinosis of the afferentarteriole occurs in numerous other set-tings. Chronic cyclosporine nephropa-thy is a typical example in which arterio-lar hyalinosis occurs outside DN.47 Tracyet al. also report the presence of arteriolarhyalinosis in kidneys of young patientswith coronary heart disease.48 Efferentarteriolar hyalinosis is an important le-sion by which DN is distinguished fromhypertensive nephropathy.49 However,most studies relate arteriolar hyalinosisto clinical parameters, not distinguishingbetween efferent and afferent arterioles,showing clear correlations with UAE anddisease progression.50,51 In Table 2 wescore 0 if no arteriolar hyalinosis ispresent, 1 if one arteriole with hyalinosisis present, and 2 if more than one arte-riole is observed in the entire biopsy.

In addition to characteristic arterio-lar hyalinosis, relatively nonspecific ar-teriosclerosis may be present in the bi-opsy specimen. Bohle and colleagues45

found increases in vascular disease as-

sociate with more severe glomerulardisease. Osterby et al.52 use a so-called“matrix to media ratio” to investigatethe role of arteriosclerosis and find thisratio is increased in patients with mi-croalbuminuria, suggesting that arte-riolar matrix accumulation occursearly in the course of DN. In Table 2 wescore the most severely affected arteryin the biopsy and assign a score of 0 ifno intimal thickening is present, 1 ifintimal thickening is less than thethickness of the media, and 2 if intimalthickening is more than the thicknessof the media. Isolated or significantmedial thickness may be associatedwith concurrent hypertension.

Other Glomerular LesionsIn 1994, Stout et al.46 defined “insudativelesions” as consisting of intramural ac-cumulations of presumably imbibedplasma proteins and lipids within renalarterioles, glomerular capillaries, Bow-man’s capsule, or proximal convolutedtubules. Insudative lesions in Bowman’scapsule are called capsular drop lesions,and in afferent and efferent arteriolesthey are called hyalinized afferent and ef-ferent arterioles. In glomerular capillar-ies they are called fibrin cap lesions, al-though this term is considered obsoleteand moreover is a misnomer because thelesion does not contain fibrin; we preferthe term hyalinosis for these lesions.

Table 2. Interstitial and vascular lesions of DN

Lesion Criteria Score

Interstitial lesionsIFTA No IFTA 0

�25% 125% to 50% 2�50% 3

interstitialinflammation

Absent 0Infiltration only in relation to IFTA 1Infiltration in areas without IFTA 2

Vascular lesionsarteriolar hyalinosis Absent 0

At least one area of arteriolar hyalinosis 1More than one area of arteriolar hyalinosis 2

presence of largevessels

– Yes/no

arteriosclerosis (scoreworst artery)

No intimal thickening 0Intimal thickening less than thickness of media 1Intimal thickening greater than thickness of media 2



Capsular drops are mainly located be-tween the parietal epithelium and Bow-man’s capsule of the glomerulus.11 Cap-sular drops are prevalent in advancedDN53 and associate with disease progres-sion.54 The common belief, reviewed byAlsaad et al.,49 is that capsular drops arespecific but not entirely pathognomonicof DN. Stout et al.46 report a prevalenceof capsular drops in 5.3% of biopsieswithout diabetes. However, finding acapsular drop in a biopsy can help distin-guish DN from other causes of glomeru-losclerosis.

By light microscopy, glomerular hya-linosis describes the same staining char-acteristics as the capsular drop lesion butit occupies the capillary lumen instead ofbeing attached to Bowman’s capsule.This lesion is not a specific finding inDN, because similar lesions are recog-nized in focal glomerulosclerosis, arte-rionephrosclerosis, and lupus nephri-tis.55

Finally, there is increasing recogni-tion of abnormalities in the glomerulo-tubular junctions with focal adhesionscalled “tip lesions” and atrophic tubuleswith no observable glomerular opening(so-called “atubular glomeruli”). Theselesions are typically found in more ad-vanced stages of nephropathy associatedwith overt proteinuria.30

Interobserver ReproducibilityTo assess the reproducibility of our con-sensus classification, a pilot study wasperformed in which five pathologists in-dependently classified 25 renal biopsieswith DN using PAS stains only into classI, II, III, or IV. Two pathologists scoredall biopsies independently. Results of theraw data are given in Table 3. The repro-ducibility of the glomerular class scorewas evaluated using an intraclass corre-lation coefficient. Analyses were carriedout using SPSS software (version 16,SPSS, Inc., Chicago, IL). There was dis-agreement in seven cases: twice on a dif-ference between class I and II, twice on adifference between class II and III, andthree times on a difference between classIIa and IIb. Overall, the results seem sat-isfactory, resulting in an intraclass corre-lation coefficient of 0.84.

Flow Chart and Scoring FormA flow chart was devised to help distin-guish the four classes of DN (Figure 2).Supplemental Figure 1 shows the scoringform we recommend for classifying glo-meruli in DN and for scoring of extraglo-merular lesions or other features.

CONCLUSIONS

We developed a classification scheme forDN consisting of four progressive classessupported by international consensus. Theclassification is based on glomerular le-sions, with a separate evaluation for inter-stitial and vascular lesions. We chose a clas-sification scheme based on glomerularlesions because these are relatively easy torecognize with good interobserver agree-ment as shown by our pilot data and be-cause glomerular lesions best reflect thenatural course of progressive DN.49 Ofcourse, glomerular and interstitial lesionscontribute to the decline in renal function

in DN and may be independent factors inthe progression of DN13; however, manystudies also show that severity of chronicinterstitial and glomerular lesions closelyassociate.12,18,19,56

Although in some clinical practicesthere is a policy to only perform a renalbiopsy to exclude causes of renal diseasecharacterized by proteinuria other thanDN, there is increasing demand to classifythe severity of disease in those patients withpure DN. Our classification system for his-topathologic lesions in DN can be used forpatients with type 1 and type 2 diabetes,because it is now generally recognized thatsubstantial overlap exists between thesetwo types with respect to histologic lesionsand renal complications.12,13 Various stud-ies also report different proportions ofnondiabetic nephropathies in patientswith diabetes and proteinuria.7,57 The clas-sification system proposed here is only forDN, but it can also serve to classify DNwhen it is complicated by another super-imposed disease.

Table 3. Assessment of interobserver agreement for scoring of DN classes Ithrough IV

BiopsyNumber

Observer1

Observer2

Observer3

Observer4

Observer5

Agreement

1 IIA I N2 III III Y3 IIB IIA N4 IIB IIA N5 IIB IIA N6 III III Y7 IIB IIB Y8 IIB IIB Y9 IV IV Y

10 IIA IIA Y11 IVA IVA Y12 IIB IIB Y13 III III Y14 I I Y15 IV IV Y16 III III Y17 III III Y18 IIB I N19 IIB III N20 IIA IIA Y21 III III Y22 III IIA N23 III III Y24 III III Y25 IIA IIA Y

Y, yes; N, no; interclass coefficient � 0.84.



Little is known about the pathogene-sis of lesions developing in DN, but it issuggested that if we could unravel thevarious pathways ultimately leading toglomerulosclerosis in DN, this couldopen up new possibilities for interven-tion to prevent or forestall nephropa-thy.43 Most likely, intra- and extraglo-merular cells are involved in theprogressive accumulation of extracellu-lar matrix proteins in DN. We suggestthat progression evolves from GBMthickening to mesangial expansion,Kimmelstiel–Wilson lesions, and globalglomerulosclerosis, respectively, which isreflected in the four classes of our classi-fication system. Using our system toevaluate protocol biopsies of patientswith DN may further unravel the com-plex pathways of DN.

An important question for every his-tologic classification system is whetherit is predictive of clinical outcome. Asin other proposals for classifying renaldisease (e.g., for lupus nephritis2 andfocal segmental glomerulosclerosis3), wechose not to assess clinical outcome aspart of this proposal. We feel validationshould be done in separate prospectivestudies, preferably including protocol bi-opsies of patients with type 1 and type 2

diabetes and clearly defined clinical endpoints.

DISCLOSURESNone.

REFERENCES

1. The World Health Report 2006: Working To-gether with Health. Geneva, Switzerland,World Health Organization, 2006

2. Weening JJ, D’Agati VD, Schwartz MM, Se-shan SV, Alpers CE, Appel GB, Balow JE,Bruijn JA, Cook T, Ferrario F, Fogo AB, Gin-zler EM, Hebert L, Hill G, Hill P, Jennette JC,Kong NC, Lesavre P, Lockshin M, Looi LM,Makino H, Moura LA, Nagata M: The classi-fication of glomerulonephritis in systemic lu-pus erythematosus revisited. J Am SocNephrol 15: 241–250, 2004

3. D’Agati VD, Fogo AB, Bruijn JA, JennetteJC: Pathologic classification of focal seg-mental glomerulosclerosis: A working pro-posal. Am J Kidney Dis 43: 368–382, 2004

4. Cattran DC, Coppo R, Cook HT, Feehally J,Roberts IS, Troyanov S, Alpers CE, Amore A,Barratt J, Berthoux F, Bonsib S, Bruijn JA,D’Agati V, D’Amico G, Emancipator S,Emma F, Ferrario F, Fervenza FC, Florquin S,Fogo A, Geddes CC, Groene HJ, Haas M,Herzenberg AM, Hill PA, Hogg RJ, Hsu SI,Jennette JC, Joh K, Julian BA, Kawamura T,Lai FM, Leung CB, Li LS, Li PK, Liu ZH,

Mackinnon B, Mezzano S, Schena FP, To-mino Y, Walker PD, Wang H, Weening JJ,Yoshikawa N, Zhang H: The Oxford classifi-cation of IgA nephropathy: Rationale, clini-copathological correlations, and classifica-tion. Kidney Int 76: 534–545, 2009

5. Glassock RJ: Reclassification of lupus glo-merulonephritis: Back to the future. J AmSoc Nephrol 15: 501–503, 2004

6. Gellman DD, Pirani CL, Soothill JF, MuehrckeRC, Kark RM: Diabetic nephropathy: A clinicaland pathologic study based on renal biopsies.Medicine (Baltimore) 38: 321–367, 1959

7. Gambara V, Mecca G, Remuzzi G, Bertani T:Heterogeneous nature of renal lesions intype II diabetes. J Am Soc Nephrol 3: 1458–1466, 1993

8. Fioretto P, Mauer M, Brocco E, Velussi M,Frigato F, Muollo B, Sambataro M, Abater-usso C, Baggio B, Crepaldi G, Nosadini R:Patterns of renal injury in NIDDM patientswith microalbuminuria. Diabetologia 39:1569–1576, 1996

9. Mazzucco G, Bertani T, Fortunato M, Ber-nardi M, Leutner M, Boldorini R, Monga G:Different patterns of renal damage in type 2diabetes mellitus: A multicentric study on393 biopsies. Am J Kidney Dis 39: 713–720,2002

10. Schwartz MM, Lewis EJ, Leonard-Martin T,Lewis JB, Batlle D: Renal pathology patternsin type II diabetes mellitus: Relationship withretinopathy. The Collaborative Study Group.Nephrol Dial Transplant 13: 2547–2552,1998

11. D’Agati V, Jennette JC, Silva F: Diabeticnephropathy. In: Atlas of Nontumor Pathol-ogy: Non-Neoplastic Kidney Diseases, 4thEd., edited by King DW, Washington, DC,American Registry of Pathology, 2005, pp457–479

12. Osterby R, Gall MA, Schmitz A, Nielsen FS,Nyberg G, Parving HH: Glomerular structureand function in proteinuric type 2 (non-insu-lin-dependent) diabetic patients. Diabetolo-gia 36: 1064–1070, 1993

13. White KE, Bilous RW: Type 2 diabetic pa-tients with nephropathy show structural-functional relationships that are similar totype 1 disease. J Am Soc Nephrol 11: 1667–1673, 2000

14. Solez K, Colvin RB, Racusen LC, Haas M, SisB, Mengel M, Halloran PF, Baldwin W, BanfiG, Collins AB, Cosio F, David DS, Drachen-berg C, Einecke G, Fogo AB, Gibson IW,Glotz D, Iskandar SS, Kraus E, Lerut E, Man-non RB, Mihatsch M, Nankivell BJ, NickeleitV, Papadimitriou JC, Randhawa P, Regele H,Renaudin K, Roberts I, Seron D, Smith RN,Valente M: Banff 07 classification of renalallograft pathology: Updates and future di-rections. Am J Transplant 8: 753–760, 2008

15. Haas M: Postinfectious glomerulonephritiscomplicating diabetic nephropathy: A fre-quent association, but how clinically impor-tant? Hum Pathol 34: 1225–1227, 2003

NO

YES

YES

YES

NO

NO

NO

YES

YES

Global glomerulosclerosisin >50% of glomeruli with

lesions from class I-III?

Mesangium> capillary

lumen?

Advanced diabeticnephropathy

Nodular sclerosis

Severe mesangialexpansion

Mild mesangialexpansion

Mesangialexpansion

>25%?

Nodular sclerosisin glomeruli?

GBM thickeningGBM >395 nm in female and

>430 nm in male individuals of9 years and older at EM?

IV

III

II B

II A

I

Figure 2. Flow chart for classifying DN.



16. Haas M: Alport syndrome and thin glomer-ular basement membrane nephropathy: Apractical approach to diagnosis. Arch PatholLab Med 133: 224–232, 2009

17. Dische FE: Measurement of glomerularbasement membrane thickness and its ap-plication to the diagnosis of thin-membranenephropathy. Arch Pathol Lab Med 116: 43–49, 1992

18. Mauer SM, Steffes MW, Ellis EN, SutherlandDE, Brown DM, Goetz FC: Structural-func-tional relationships in diabetic nephropathy.J Clin Invest 74: 1143–1155, 1984

19. Osterby R, Gundersen HJ, Nyberg G, AurellM: Advanced diabetic glomerulopathy.Quantitative structural characterization ofnonoccluded glomeruli. Diabetes 36: 612–619, 1987

20. Drummond K, Mauer M: The early naturalhistory of nephropathy in type 1 diabetes: II.Early renal structural changes in type 1 dia-betes. Diabetes 51: 1580–1587, 2002

21. Perrin NE, Torbjornsdotter TB, Jaremko GA,Berg UB: The course of diabetic glomeru-lopathy in patients with type I diabetes: A6-year follow-up with serial biopsies. KidneyInt 69: 699–705, 2006

22. Falk RJ, Scheinman JI, Mauer SM, MichaelAF: Polyantigenic expansion of basementmembrane constituents in diabetic ne-phropathy. Diabetes 32[Suppl 2]: 34–39,1983

23. Kim Y, Kleppel MM, Butkowski R, Mauer SM,Wieslander J, Michael AF: Differential ex-pression of basement membrane collagenchains in diabetic nephropathy. Am J Pathol138: 413–420, 1991

24. Mac-Moune LF, Szeto CC, Choi PC, Ho KK,Tang NL, Chow KM, Li PK, To KF: Isolatediffuse thickening of glomerular capillarybasement membrane: A renal lesion in pre-diabetes? Mod Pathol 17: 1506–1512, 2004

25. Bangstad HJ, Osterby R, Dahl-Jorgensen K,Berg KJ, Hartmann A, Hanssen KF: Improve-ment of blood glucose control in IDDM pa-tients retards the progression of morpholog-ical changes in early diabetic nephropathy.Diabetologia 37: 483–490, 1994

26. Jensen EB, Gundersen HJ, Osterby R: De-termination of membrane thickness distribu-tion from orthogonal intercepts. J Microsc115: 19–33, 1979

27. Steffes MW, Barbosa J, Basgen JM, Suther-land DE, Najarian JS, Mauer SM: Quantita-tive glomerular morphology of the normalhuman kidney. Lab Invest 49: 82–86, 1983

28. Nasr SH, Markowitz GS, Valeri AM, Yu Z,Chen L, D’Agati VD: Thin basement mem-brane nephropathy cannot be diagnosed re-liably in deparaffinized, formalin-fixed tis-sue. Nephrol Dial Transplant 22: 1228–1232, 2007

29. Caramori ML, Kim Y, Huang C, Fish AJ, RichSS, Miller ME, Russell G, Mauer M: Cellularbasis of diabetic nephropathy: 1. Study de-

sign and renal structural-functional relation-ships in patients with long-standing type 1diabetes. Diabetes 51: 506–513, 2002

30. Najafian B, Kim Y, Crosson JT, Mauer M:Atubular glomeruli and glomerulotubularjunction abnormalities in diabetic ne-phropathy. J Am Soc Nephrol 14: 908 –917, 2003

31. Mauer SM, Sutherland DE, Steffes MW: Re-lationship of systemic blood pressure to ne-phropathology in insulin-dependent diabe-tes mellitus. Kidney Int 41: 736–740, 1992

32. Stout LC, Kumar S, Whorton EB: Focal me-sangiolysis and the pathogenesis of the Kim-melstiel–Wilson nodule. Hum Pathol 24: 77–89, 1993

33. Kimmelstiel P, Wilson C: Intercapillary le-sions of the glomeruli of the kidney. Am JPathol 12: 83–96, 1936

34. Cameron JS: The discovery of diabetic ne-phropathy: From small print to centre stage.J Nephrol 19[Suppl 10]: S75–S87, 2006

35. Allen A: So-called intercapillary glomerulo-sclerosis: A lesion associated with diabetesmellitus. Morphogenesis and significance.Arch Path 32: 33–51, 1941

36. Markowitz GS, Lin J, Valeri AM, Avila C, NasrSH, D’Agati VD: Idiopathic nodular glomer-ulosclerosis is a distinct clinicopathologicentity linked to hypertension and smoking.Hum Pathol 33: 826–835, 2002

37. Nishi S, Ueno M, Hisaki S, Iino N, Iguchi S,Oyama Y, Imai N, Arakawa M, Gejyo F: Ul-trastructural characteristics of diabetic ne-phropathy. Med Electron Microsc 33: 65–73,2000

38. Paueksakon P, Revelo MP, Ma LJ, Marcan-toni C, Fogo AB: Microangiopathic injuryand augmented PAI-1 in human diabetic ne-phropathy. Kidney Int 61: 2142–2148, 2002

39. Glick AD, Jacobson HR, Haralson MA: Mes-angial deposition of type I collagen in hu-man glomerulosclerosis. Hum Pathol 23:1373–1379, 1992

40. Sandison A, Newbold KM, Howie AJ: Evi-dence for unique distribution of Kimmel-stiel–Wilson nodules in glomeruli. Diabetes41: 952–955, 1992

41. Hong D, Zheng T, Jia-qing S, Jian W, Zhi-hong L, Lei-shi L: Nodular glomerular lesion:A later stage of diabetic nephropathy? Dia-betes Res Clin Pract 78: 189–195, 2007

42. Mason RM, Wahab NA: Extracellular matrixmetabolism in diabetic nephropathy. J AmSoc Nephrol 14: 1358–1373, 2003

43. Qian Y, Feldman E, Pennathur S, Kretzler M,Brosius FC III: From fibrosis to sclerosis:Mechanisms of glomerulosclerosis in dia-betic nephropathy. Diabetes 57: 1439–1445, 2008

44. Horlyck A, Gundersen HJ, Osterby R: Thecortical distribution pattern of diabetic glo-merulopathy. Diabetologia 29: 146–150,1986

45. Bohle A, Wehrmann M, Bogenschutz O,

Batz C, Muller CA, Muller GA: The patho-genesis of chronic renal failure in diabeticnephropathy. Investigation of 488 cases ofdiabetic glomerulosclerosis. Pathol ResPract 187: 251–259, 1991

46. Stout LC, Kumar S, Whorton EB: Insudativelesions—Their pathogenesis and associationwith glomerular obsolescence in diabetes: Adynamic hypothesis based on single viewsof advancing human diabetic nephropathy.Hum Pathol 25: 1213–1227, 1994

47. Bennett WM, DeMattos A, Meyer MM, An-doh T, Barry JM: Chronic cyclosporine ne-phropathy: the Achilles’ heel of immunosup-pressive therapy. Kidney Int 50: 1089–1100,1996

48. Tracy RE, Strong JP, Newman WP III, Mal-com GT, Oalmann MC, Guzman MA: Reno-vasculopathies of nephrosclerosis in relationto atherosclerosis at ages 25 to 54 years.Kidney Int 49: 564–570, 1996

49. Alsaad KO, Herzenberg AM: Distinguishingdiabetic nephropathy from other causes ofglomerulosclerosis: An update. J Clin Pathol60: 18–26, 2007

50. Fioretto P, Steffes MW, Sutherland DE,Mauer M: Sequential renal biopsies in insu-lin-dependent diabetic patients: Structuralfactors associated with clinical progression.Kidney Int 48: 1929–1935, 1995

51. Ruggenenti P, Gambara V, Perna A, BertaniT, Remuzzi G: The nephropathy of non-insu-lin-dependent diabetes: Predictors of out-come relative to diverse patterns of renalinjury. J Am Soc Nephrol 9: 2336–2343,1998

52. Osterby R, Asplund J, Bangstad HJ, NybergG, Rudberg S, Viberti GC, Walker JD: Neo-vascularization at the vascular pole region indiabetic glomerulopathy. Nephrol DialTransplant 14: 348–352, 1999

53. Bloodworth JMB: Re-evaluation of diabeticglomerulosclerosis 50 years after discoveryof insulin. Hum Pathol 9: 439–453, 1978

54. Takazakura E, Nakamoto Y, Hayakawa H,Kawai K, Muramoto S: Onset and progres-sion of diabetic glomerulosclerosis: A pro-spective study based on serial renal biop-sies. Diabetes 24: 1–9, 1975

55. Laufer A, Stein O: The exudative lesion indiabetic glomerulosclerosis. Am J ClinPathol 32: 56–61, 1959

56. Lane PH, Steffes MW, Fioretto P, Mauer SM:Renal interstitial expansion in insulin-depen-dent diabetes mellitus. Kidney Int 43: 661–667, 1993

57. Parving HH, Gall MA, Skott P, JorgensenHE, Lokkegaard H, Jorgensen F, Nielsen B,Larsen S: Prevalence and causes of albumin-uria in non-insulin-dependent diabetic pa-tients. Kidney Int 41: 758–762, 1992

Supplemental information for this article is availableonline at http://www.jasn.org/



nefropati diabetikum.pdf

Documents