Nephrol Dial Transplant (1992) 7: 191-199 r 1992 European Dialysis and Transplant Association-European Renal Association
Nephrology Dialysis Transplantation
Original Article Glomerular basement membrane thinning in adults: Clinicopathological correlations of a new diagnostic approach A. L. C. McLay, R. Jackson, F. Meyboom and J. M. Boulton Jones University Department of Pathology and Renal Unit, Glasgow Royal Infirmary, Glasgow, UK
Abstract. We have studied glomerular basal laminar thickness in biopsy material, using a simple technique involving 16 selected measurements per case. Twentynine biopsied cases of adult glomerular haematuria were examined together with 'diseased' controls represented by a variety of glomerulopathies including minimal-change disease and IgA nephropathy. 'Normal' control populations were provided by 13 patients with acute-onset renal failure of non-glomerular origin and nine patients undergoing nephrectomy. Analysis of groups determined by the presence or absence of haematuria, the degree of proteinuria and presence or absence of a diagnostically characteristic immunofluorescence pattern showed that the nine patients with haematuria and proteinuria of less than 200 mg/24 h represented a distinct subpopulation with a mean membrane thickness of 225 nm compared to the control mean of 343 nm (P< 0.0001). All members of this subpopulation had mean values below an arbitrary cut-off value of 270 nm. Within other specific disease categories, sporadic cases had mean membrane thicknesses below this critical value, indicative of an overlap of pathologies. On shortterm follow-up there is no evidence that the 'pure' thin-membrane population are subject to any deterioration in rena! function. It is of further interest that eight of nine thin-membrane 'syndrome' cases were O Rh positive. This finding may provide a starting point for investigation of a specific genetic defect.
Correspondence and offprint ret/ut'.l. to: Dr A. L. C. McLay, University Department of Pathology, Glasgow Royal Infirmary, Cast!e Street, Glasgow G4 0SE, UK.
Key words: basement membrane; blood diagnosis; glomerulus; thinning
group;
Introduction Diffuse glomerular basal laminar thinning, whilst long recognized as being associated with benign haematuric states in children [1-4] has only recently been considered to be relevant to adult haematuric syndromes [5-11]. However, some of these authors [5,8,9] have challenged the benign nature of the process by identifying affected patients with deteriorating renal function. Where diagnosis is based on more than subjective assessment of ultrastructure it has been customary to employ numerous random measurements using the orthogonal intercept method [12] which may involve up to 250 measurements per case. However, in the course of a previous pilot study [13] we came to the conclusion that there were grounds for doubting whether average values based on multiple random measurements were necessarily representative of the intrinsic, undamaged laminar thickness for a given case. Moreover, as will be discussed later, we believed that in the particular case of the glomerular basal lamina, the orthogonal intercept method was demonstrably unnecessary' and might even fail to identify certain forms of clinically significant thinning. In any case of apparent glomerular haematuria the fundamental question of clinical relevance is whether there are sufficient grounds for presuming that haematuria can be attributed to an abnormality of the
192
mechanical properties of the capillary basal lamina. Any appropriate analytical approach should show clear differences between such cases and groups where the problem can reasonably be attributed to immunecomplex mediation. It would therefore be expected to identify significant glomerular basal laminar thinning in a group of patients with glomerular haematuria, normal renal function, no proteinuria, and negative immunofluorescence findings (conventional benign recurrent haematuria). Moreover, if basal laminar thinning invariably causes haematuria, patients with non-haematuric conditions such as minimal-change disease or tubulointerstitial disease would, in general, be expected to exhibit measurements that are not significantly different from normal values. Furthermore, patients with haematuria attributable to IgA nephropathy, for example, would also be expected to display essentially normal thickness values. We set out, therefore, using a small number of selected direct measurements, to ascertain whether anomalies of thickness thus determined would usefully correlate with other clinicopathological data and so meet the needs of a diagnostic service. This paper describes the simple technique employed and outlines the correlations obtained.
Subjects and methods Forty-five cases of glomerular disease were studied. These fell into three groups determined by clinical and pathological data. Proteinuria in excess of 200 mg/24 h was regarded as clinically significant.
Group I This comprised 15 haematuric cases with normal renal function biopsicd between January 1988 and March 1989 and in whom the immunofluorescence findings were negative (non-characteristic for any specific immune-complexmediated glomerulopathy). This group was subdivided. Group la comprised those cases (n = 9) without significant proteinuria, haematuria being the sole finding. Group lb comprised those with, in addition to haematuria, varying degrees of proteinuria (n = 6). Within this latter subgroup were four cases of minimal-change/IgM nephropathy and two cases whose only satisfactory designation may be 'idiopathic haematuria'.
Group 2 This comprised 14 haematuric cases with biopsy evidence of immune-complex-mediated disease (11 IgA nephropathy. 1 Henoch Schoenlein purpura. 2 systemic lupus erythematosus). All patients in this group also displa>ed some degree of proteinuria.
A. L. C. McLay eI al.
Group 3 In this group there were 16 non-haematuric patients biopsied on account of proteinuria and/or mild impairment of renal function. Subgroup 3a comprised eight cases displaying no evidence of immune-complex deposition (i.e. minimal-change nephropathy) whilst subgroup 3b (also eight cases) included a variety of immune-complex-mediated nephropathies (5 membranous, 2 IgA nephropathy. I Henoch-Schoenlein purpura).
Group 4 Control glomerular samples were derived from cases where no glomerular pathology was anticipated, and these were divided into three subgroups: Group 4a. This comprised five patients with acute-onset renal failure attributable to acute interstitial nephritis. Group 4b. In this group there were a further eight patients with acute-onset renal failure attributed to acute tubular necrosis. Group 4c. This comprised nine patients undergoing nephrectomy for tumour masses, presumptively normal glomeruli being selected from well-defined cortical tissue remote from the tumour mass.
Methods All material for ultrastructural study was fixed in 4% glutaraldehyde in phosphate buffer at pH 7.4, post-fixed in osmium tetroxide, embedded in EMIX resin (Emscope Ltd,) and stained with lead citrate and uranyl acetate. All specimens were examined in a Phillips 301G electronmicroscope at a nominal magnification of x 9800 The microscope magnification was calibrated using a crossed diffraction grating ruled at 2160 lines per millimetre. One glomerulus from each case was examined and electronmicrographs obtained from each of the four thinnest uniform peripheral loop elements. The basal lamina was measured at four representative points along a length of not less than 3.6 urn per loop at a final print magnification of x 27 000. All measurements were made from the overlying epithelial cell plasma membrane to the opposing endothelial cell plasma membrane. Furthermore, measurements were restricted to areas where both cell membranes were closely applied to the basal lamina, avoiding areas of focal attenuation and notching. Moreover, in cases of immunecomplex-mediated disease, measurements were restricted to peripheral loop elements which were not involved with electron-dense deposits. It was also considered important to select regions where endothelial fenestrations were clearly sectioned perpendicular to their diameter, thus further minimizing obliquity of cut. The arithmetical mean of 16 values obtained was then calculated. Measurements were undertaken 'blind' in so far as no reference was made to clinical data. However, it is clear that the micrographs themselves could provide the observer with specific diagnostic information (e.g. electron-dense deposits or foot process effacement). To test the variation in thickness measurement due to sampling bias, a second glomerulus was examined in three cases from group la and three from the control group 4c. Clinical correlation was based on detailed case sheet data analysis with particular reference to creatinine at presentation and to quantitation of haematuria and proteinuria. It was. moreover, considered that intrinsic red cell properties
Glomemlar basement membrane thinning
might influence haematuria. Blood group status was. therefore, also ascertained. Student's ' test was cmp!ovcd for evaluation of differences in laminar width between groups defined by clinical and immunoh'Stochcmical criteria.
Results The mean laminar thickness value for the whole control group, i.e. group 4 (n = 22) was 343 nm (SD 62 nm). The differences between normal laminae (Figure la) and markedly thinned laminae (Figure lb) are identifiable even on casual scrutiny, although subnormal values encompass a wide spectrum which extends from the thinnest to virtually normal. The measurements undertaken on a second glomemlus in three control cases (including one with a thin-membrane state) differed from the original mean only by + 7 to -lOnm. In the additional measurements carried out on cases from group la the deviations from the first glomerular mean covered a somewhat wider range, (-3 to -34 nm). All deviations here, however, only serve to lower the values of a group already distinguished by low values. The distribution of values for individual patients in each group is displayed in Figure 2. The range depicted refers to the highest and lowest average loop values. It is clear that the patients of group la form a distinct cohort in respect of laminar thickness. The mean value for this group is 225 nm (SD = 29 nm). The difference between this and the group 4 value is highly significant (P< 0.0001). No statistical difference is observable between groups lb, 2 or 3 and the control group, 4. It can also be seen that all the mean values for group la lie below an arbitrary value of 270 nm, intermediate between the 264 nm threshold of abnormality suggested by Tiebosch et al. [9] and that of 273 nm suggested by Saxena et al. [12]. These nine haematuric cases with thin laminae and no clinically significant proteinuria (group la) spanned a time period in which 17 cases of IgA nephropathy were diagnosed, suggesting an incidence of a similar order of magnitude. The clinical correlates and basal laminar measurements of this group are summarized in Table 1. All were normotensive and had unimpaired renal function. A clearly defined family history of haematuria was elicited from only one patient. Seven of the nine were female and two patients suffered episodes of macroscopic haematuria in addition to persistent microscopic haematuria. Light-microscopv showed glomeruli in this group to be normal or to be characterized by only mild focal segmental hyperce!!u!aritv. No significant tubulointerstitial changes were noted. Where immunoglobulin was identified on immunofluorescence examination only scanty focal mesangial deposits of
193
IgM were seen. Ultrastructura! examination often showed, in addition to laminar thinning, fairly prominent mesangial paramesangial wrinkling (Figure 3). No case, however, displayed the splitting lamellation characteristic of a proportion of cases of Alport's syndrome. Moreover, enquiry into family history failed to reveal any relatives with Alport's s>ndrome, and deafness was not clinically evident although formal audiometric studies were not undertaken. Blood group analysis of all cases revealed that in this thin-laminar population eight of nine cases were characterized by the O phenotype. This figure of 89% contrasts strongly with the local population norm of 55% and the figure of 62% for all other biopsy cases. This, therefore, raises the possibility of a significant association with blood group O. Group lb, characterized by proteinuria in addition to haematuria, fails, on the other hand, to display uniformly low values of laminar thickness. It is, moreover, associated with a wider spectrum of clinical patterns. The full data for this group are presented in Table 2. In two cases the thickness value lies below the arbitrary 270 nm datum. One such case (case 11) bore a diagnosis of IgM nephropathy whilst the other (case 10) is regarded as one of'idiopathic' haematuria but differs only marginally from group la cases in terms of proteinuria. The remainder of these cases in lb had normal laminar values despite suffering from haematuria which was not attributable to immunecomplex disease. Apart from cases 10 and 11 in group lb, individual cases in other groups, including controls, fall below the suggested critical value of 270 nm. The data for those associated with defined glomerulopathy are outlined in Table 3. These thus include cases of IgM nephropathy, FgA nephropathy, Henoch-Schoenlein purpura, minimal-change disease (Figure 4) and even membranous nephropathy (Figure 5). It is noteworthy that three such cases did not display haematuria despite the evidence of marked laminar thinning.
Discussion We have demonstrated here that patients with glomerular haematuria and no proteinuria are consistently characterized by subnormal laminar thickness values, as determined by a simplified measurement technique. Low values in individual cases are demonstrable within other clinicopathological groups, including controls, but, as populations, the mean values from these other groups fall within the normal range. Furthermore, whilst 13 of 29 haematuric cases displayed mean laminar thickness values below 270 nm, this was true for only three of 16 non-haematuric glomerulopathic cases and two of 22 controls. \*ore-
A. L. C McLay el al
194
31. M.
Jfl
Fig. 1. (a) N o r m a l lamina x 27 000 (b) M a r k e d l ) thinned lamina x 27 000.
Glomerular basement membrane thinning
195
° Femals • Ma!e 700 n
3a
3b
4b
4a
4c
Fig. 2. Laminar thickness values for individual patients. (Mean and interloop range shown.) la = Haematuria,no proteinuria—non-characteristic immunofluorescence; lb=Haematuria + proteinuria—non-characteristic lmmunofluorescence; 2 = Haematuria + proteinuria—characteristic immunofluorescence; 3a = Proteinuria only—negative immunofluorescence (minimal change); 3b = Proteinuria only—characteristic immunofluorescence; 4a = Controls—acute interstitial nephritis; 4b = Controls— acute tubular necrosis; 4c = Controls—nephrectomy. Table 1. Clinical correlates and glomerular basal laminar thicknesses (in nanometres) for group la Case
Sex
Age
Clinical details
Addis
QP
No.
1 2
F
3 4 5 6 7 8
F F F M F
33 58 54 52 23 36
9
M
32
F
F
41 56
Microhaematuria x 3 years Macro + Microhaematuria x 8 years MicTohaematuria x 1 'A years Microhaematuria x 1 year Microhaematuria Microhaematuria MicTohaematuria x 1 year Microhaematuria x 7 years Positive FHx Microhaematuria x 9/12
14 48 29 57 6
50 mg 0 0 0 0
Blood group
Basal lamina Mean
Range
Opos Apos
194 195
162-218 151-240
199 200 226 230 253 263
155-251 159-237 181-270 226-236 218-307 185-344 237-285
_ 200 200
50 mg 200 mg 50 mg
Opos Opos Opos Opos Opos Opos
18
50 mg
Opos
268
Blood group
Basal laminar dimensions
QP = Quantitative proteinuria; Addis = Number of red cells per high-power field. Table 2. Clinical correlates and glomerular basal laminar thicknesses for group lb. Case
Sex
Age
Clinical details
Addis
QP
No.
10 11
M
12 13 14
M F
M
39 55 21
15
M
62
F
21 37
MacTohaematuria x 4/12 Microhaematuria + proteinuria Nephrotic + haematuria Nephrotic + haematuria Microhaematuria x 1 year + macro + proteinuria Nephrotic + haematuria
over, within the limits of the small number of cases in which a second glomerulus was examined, there appears to be good evidence that the values obtained on one glomerulus are representative for any given
200 8
+ +
0.5 g 5.6 g 18g 13g
Mean
Range
Opos Opos
239 213
227-248 177-244
-
270 301 375
262-288 262 -329 297^133
401
314 527
+
0.7 g
Bpos Opos
+
7.87 g
Opos
case. We therefore believe that the technique employed is of practical value in confirming the presence of thin laminar states. Furthermore, when taken in conjunction with other clinical and patholo-
186
Fig. 3. Haematuric patient with thin laminae and non-charactenstic lmmunofluorescence There is conspicuous mcsangial/paramesangial wrinkling (Arrows), x 27 000. Table 3. Clinical correlates and GBM thicknesses for cases with 'significant' laminar thinning in groups 2, 3a, and 3b Case No
18 19 30 31 38
Group No.
2 2 3A 3A 3B
Sex
M F F F F
Age
32 36 64 52 18
Diagnosis
IgA nephropathy HSP Minimal change Minimal change Membranous
gical data, it may be used to identify those cases in which haematuria is most likely to be attributable to an intrinsic laminar abnormality. Issue may well be taken with the selective absolute measurement approach used in this study. However, we believe it is well justified on several counts. First of all, although the majority of previous studies of relevance have employed the orthogonal intercept method, careful scrutiny of the key reference [12] indicates that this latter method was devised for the general case of laminar measurement and then validated in the case of the glomerular basal lamina where absolute values could be determined with some accuracy since 'in this special case the membrane is delineated on both sides, allowing one to measure membrane width onl> at sites where the sectioning is within a known, narrow range of perpendicularity.'
Addis count
27 100 0 0 0
QP (g/day)
0.95 1 5 14.0 2.5
Blood group
B pos O pos OPos — A pos
Basal laminar dimensions Mean
Range
250 267 188 262 222
212-308 229 289 148 214 232 308 192-248
Areas of tangential section, and in pathological states of attenuation or thickening related to damage, are readily identifiable and easily excluded from morphometric assessment. Thus, direct measurement of architecturally normal laminae is eminently practical. Although the convention hitherto has been to consider only diffuse thinning as pathologically relevant, there are no a priori grounds for presuming this to be true. Indeed it seems particularly pertinent that in Alport's syndrome it has been observed [14,15] that irregular thickened laminar states evolved (presumably through recurrent damage) from laminae of subnormal thickness. Thick and thin elements may be observable within any given capillary loop so that average laminar thickness values may fall within or even exceed the normal range even although the intrinsic undamaged thickness is subnormal. In this
Glomerular basement membrane thinning
Fig. 4. Foot process effacement associated with thin laminar state in minimal-change nephropathy. x 27 000.
Fig. 5. Membranous nephropathy with thin laminae. Arrows identify subepithelial deposits, x 27 000.
197
198
context it is of considerable interest to note that in the study of Coleman et al. [8], one patient had 'an excess of thin GBM although the mean thickness was normal.' A case such as this serves to emphasize the fact that random averaged measurements may obscure the underlying abnormality. It seems likely that any thin-laminar condition could display similar features although severity and rate of progression would depend upon the specific molecular defect. We believe, therefore, that identification of a population with intrinsically thin membranes depends on the measurement, within any given case, of the thinnest elements which are not associated with identifiable architectural disturbance. Secondly, the control values in this study are close to those of recent studies using conventional epoxy resins [9,10] and generally comparable with the lower end of normality in the majority of studies. The important influence of the embedding medium on measured thickness is emphasized by Haynes [16]. Thirdly, as we have stated, the simple measurement technique employed here clearly identifies a specific haematuric population with no proteinuria and no evidence of immunologically mediated disease. This group meets the classical criteria for thin-basementmembrane 'syndrome' and the method would appear, therefore, to be entirely suitable for routine diagnostic practice. There would be no reason to believe, moreover, that the approach to measurement employed in this paper would necessarily identify populations coincident with those delineated by the harmonic mean of orthogonal intercepts. Thus, validation is provided by its clinicopathological correlations rather than by comparison with the orthogonal intercept method. Although it may be surprising to some, the finding of overlaps between thin-membrane states and other definable glomerulopathies confirms the impression gained in our pilot study [ 13] and is entirely consistent with the identified presence of thin membranes in the control population. This latter incidence corresponds with the recent observations of Dische et al. [17] in a transplant donor population. Moreover, more extensive attenuation of basal laminae in IgA nephropathy has been noted [18], although the incidence in that study seems unusually high whilst the magnification selected for measurement (x 2700) can be regarded as too low for the necessary degree of precision. We believe that where overlaps occur, phenomena other than laminar thinning are likely to have a greater influence on prognosis and it appears to us significant that many of the poorer prognosis cases in other studies had proteinuria in excess of 1 g/24 h and, or features suggestive of additional complicating factors such as malignant hypertension or deafness.
A. L C. McLay et al.
In view of the presence of haematuria in renal neoplasia we feel that no useful clinical correlations can be made with the thin-laminar states in Group 4c. However, it is of further interest to note that three of the seven cases showing both thin membranes (188, 222 and 262 nm) and other primary glomerular pathology did not suffer from haematuria despite membrane thicknesses as low as those in the haematuric group. We therefore share some of the concerns recently expressed [19] regarding arbitrary correlation between membrane thinning and haematuric glomerular disease. If one assumes that non-immunologically mediated haematuria is related to laminar fragility, then it would seem clear that thinness does not necessarily imply fragility. This suggests that laminar thinning may be a consequence of more than one molecular defect, a fact already implicit in the observation [20] that classical thin-membrane cases possess the Goodpasture antigen, unlike Alport cases, which may also display thinning [3,15]. It would therefore seem advisable to reserve the term 'thin-membrane disease (syndrome)' for haematuric cases with demonstrable laminar thinning but no evidence of immunologically mediated disease and with proteinuria of less than 200 mg/24 h. Within this group, in our series, no untoward trends in renal function have so far (time from biopsy 24-41 months) been observed. We believe that other cases should simply be regarded as exhibiting thin-laminar states. The demonstration within our thin-membrane disease group of such a high prevalence of blood group O is, in such a small sample, insufficient proof of a specific relationship, but seven of eight further cases identified since the completion of the formal study also belong to blood group O and the finding would appear therefore to merit further investigation. Whilst in the case of Alport's syndrome the laminar defect appears to relate to absence of Goodpasture epitopes, which are located in the non-collagenous globular domain of the carboxy terminal end of type IV collagen, [21], it may be that alterations of associated carbohydrate moieties might give rise to similar structural problems. Such alterations might be mediated by glycosyltransferases capable of influencing histoblood group antigen expression, and it is tempting therefore to speculate whether polymorphism of such a carbohydrate might correspond with known polymorphism in histo-blood group antigens [22]. Thus, if the blood group relationship suggested by the present study can be confirmed in a larger series, detailed investigation of phenomena such as secretor status and the expression of other blood group systems might provide a starting point for more precise evaluation of the nature of the presumed genetic defect involved. In conclusion, whatever the underlying defect, thin-
Glomerular basement membrane thinning
membrane states may be easily diagnosed in biopsy populations using this simple measurement technique. As others have indicated, thin-membrane disease is a common cause of persistent microscopic haematuria in non-uraemic adults. However, thin-laminar states overlap with other diagnostic categories and the latter may be of much more prognostic relevance in individual cases. Furthermore, a thin-laminar state does not necessarily result in glomerular haematuria. We believe, therefore, that thin-membrane disease with its implicit good prognosis should only be diagnosed in patients with haematuria, no proteinuria, and no evidence of a characteristic pattern of immunoglobulin deposition. Moreover, this phenomenon appears to have a strong but not invariable association with blood group O. Acknowledgements. The authors wish to thank Miss J. S. Munro for her skilled technical assistance throughout this study. They are also indebted to Mr C. Buck, Mr P. Paterson, and Mr R. Scott of the Urology Department, Glasgow Royal Infirmary for their assistance in obtaining fresh tissue from nephrectomies performed on patients under their care. Thanks are also due to Dr G. Murray of the University Department of Surgery, Glasgow Royal Infirmary, for his observations on certain statistical issues raised.
199
6. 7. 8.
9. 10. 11. 12. 13. 14. 15. 16. 17.
References 1. Rogers PW, Kurtzman NA, Bunn SM Jr, White MG. Familial benign essential hematuria. Arch Intern Med 1973; 131: 257-262 2. Yoshikawa N, White RHR, Cameron AH. Familial hematuria: clinico-pathological correlations. Clin Nephrol 1982; 17: 172-182 3. Yoshikawa N, Hashimoto H, Katayama Y, Yamada Y, Matsuo T. The thin glomerular basement membrane in children with haematuria. J Pathol 1984; 142: 253-257 4. Tina L, Jenis E, Josse P, Medani C, Papadopoulou Z, Cateagno P. The glomerular basement membrane in benign familial hematuria. Clin Nephrol 1982; 17: 1-4 5. Dische FE, Weston MJ, Parsons V. Abnormally thin glomerular basement membranes associated with hematuria, pro-
18. 19. 20.
21.
22.
teinuria or renal failure in adults. Am J Nephrol 1985; 5: 103-109 Abe S, Amagasaki Y, Iyori S et al. Thin basement membrane syndrome in adults. J Clin Pathol 1987; 40: 318-322 FujugaJd Y, Nagase M, Kobayashi S, Honda N", Muranaka Y. Alterations of glomerular basement membrane relevant to haematuria Virchows Arch [A] 1983; 413: 159-165 Coleman M, Haynes WDG, Dimopoulos P, Barratt LJ, Jarvis LR. Glomerular basement membrane abnormalities associated with apparently idiopathic hematuria. Hum Patliol 1986; 17: 1022-1030 Tiebosch ATMG, Frederik PM, van Breda Vriesman PJC el al. Thin basement-membrane nephropathy in adults with persistent hematuria. A' Engl J Med 1989; 329: 14-18 Aarons I, Smith PS, Davies RA, Woodroffe AJ, Clarkson AR. Thin membrane nephropathy: a clinico-pathological study. Clin Nephrol 1989; 32: 151-158 Saxena S, Davie DJ, Kirsner RLG. Thin basement membranes in minimally abnormal glomeruli. J Clin Pathol 1990; 43: 3 2 38 Jensen EB, Gundersen HJG, Osterby R. Determination of membrane thickness distribution from orthogonal intercepts. J Microsc 1979; 115: 19-33 McLay ALC, Meyboom F, Jackson R. Thin glomerular basal laminae and adult glomerular disease. J Pathol 1989; 157: 162A Antonovych TT, Deasy PF, Tina U, D'Albora JB, Hollerman CE, Calcagno PL. Hereditary nephritis: early clinical, functional and morphological studies. Pediatr Res 1969; 3: 545-556 Yum M, Bergstein JM. Basement membrane nephropathy. Hum Pathol 1983; 14: 996-1003 Haynes WDG. The normal human renal glomemlus. Virchows Arch B 1981; 35: 133-158 Dische FE, Anderson VER, Keane SJ, Taube D, Bewick M, Parsons V. The incidence of thin membrane nephropathy in kidney donors. Nephrol Dial Transplant 1990; 5: 306. (Abstract) Morita M, Sakaguchi H. A quantitative study of glomerular basement membrane changes in IgA nephropathy. J Pathol 1988; 154: 7-18 Editorial. Thin-membrane-nephropathy—how thin is thin? Lancet 1990; 336: 469-470 Dische FE, Brooke IP, Cashman SJ el at. Reactivity of monoclonal antibody PI with glomerular basement membrane in thin-membrane nephropathy. Nephrol Dial Transplant 1989; 4: 611-617 Butkowski RJ, Wieslander J, Wisdom BJ, Barr JF, Noelken ME, Hudson BG. Properties of the globular domain of type IV collagen and its relationship to the Goodpasture antigen. J Biol Chem 1985; 260: 3737-3739 Clausen H, Hakomori S. ABH and related histo-blood group antigens: Immunochemical differences in carrier isotypes and their distribution. Vox Sang 1989; 56: 1-20
Received for publication 15.1.91 Accepted in revised form 15.8.91
Nephrology Dialysis Transplantation
Original Article Glomerular basement membrane thinning in adults: Clinicopathological correlations of a new diagnostic approach A. L. C. McLay, R. Jackson, F. Meyboom and J. M. Boulton Jones University Department of Pathology and Renal Unit, Glasgow Royal Infirmary, Glasgow, UK
Abstract. We have studied glomerular basal laminar thickness in biopsy material, using a simple technique involving 16 selected measurements per case. Twentynine biopsied cases of adult glomerular haematuria were examined together with 'diseased' controls represented by a variety of glomerulopathies including minimal-change disease and IgA nephropathy. 'Normal' control populations were provided by 13 patients with acute-onset renal failure of non-glomerular origin and nine patients undergoing nephrectomy. Analysis of groups determined by the presence or absence of haematuria, the degree of proteinuria and presence or absence of a diagnostically characteristic immunofluorescence pattern showed that the nine patients with haematuria and proteinuria of less than 200 mg/24 h represented a distinct subpopulation with a mean membrane thickness of 225 nm compared to the control mean of 343 nm (P< 0.0001). All members of this subpopulation had mean values below an arbitrary cut-off value of 270 nm. Within other specific disease categories, sporadic cases had mean membrane thicknesses below this critical value, indicative of an overlap of pathologies. On shortterm follow-up there is no evidence that the 'pure' thin-membrane population are subject to any deterioration in rena! function. It is of further interest that eight of nine thin-membrane 'syndrome' cases were O Rh positive. This finding may provide a starting point for investigation of a specific genetic defect.
Correspondence and offprint ret/ut'.l. to: Dr A. L. C. McLay, University Department of Pathology, Glasgow Royal Infirmary, Cast!e Street, Glasgow G4 0SE, UK.
Key words: basement membrane; blood diagnosis; glomerulus; thinning
group;
Introduction Diffuse glomerular basal laminar thinning, whilst long recognized as being associated with benign haematuric states in children [1-4] has only recently been considered to be relevant to adult haematuric syndromes [5-11]. However, some of these authors [5,8,9] have challenged the benign nature of the process by identifying affected patients with deteriorating renal function. Where diagnosis is based on more than subjective assessment of ultrastructure it has been customary to employ numerous random measurements using the orthogonal intercept method [12] which may involve up to 250 measurements per case. However, in the course of a previous pilot study [13] we came to the conclusion that there were grounds for doubting whether average values based on multiple random measurements were necessarily representative of the intrinsic, undamaged laminar thickness for a given case. Moreover, as will be discussed later, we believed that in the particular case of the glomerular basal lamina, the orthogonal intercept method was demonstrably unnecessary' and might even fail to identify certain forms of clinically significant thinning. In any case of apparent glomerular haematuria the fundamental question of clinical relevance is whether there are sufficient grounds for presuming that haematuria can be attributed to an abnormality of the
192
mechanical properties of the capillary basal lamina. Any appropriate analytical approach should show clear differences between such cases and groups where the problem can reasonably be attributed to immunecomplex mediation. It would therefore be expected to identify significant glomerular basal laminar thinning in a group of patients with glomerular haematuria, normal renal function, no proteinuria, and negative immunofluorescence findings (conventional benign recurrent haematuria). Moreover, if basal laminar thinning invariably causes haematuria, patients with non-haematuric conditions such as minimal-change disease or tubulointerstitial disease would, in general, be expected to exhibit measurements that are not significantly different from normal values. Furthermore, patients with haematuria attributable to IgA nephropathy, for example, would also be expected to display essentially normal thickness values. We set out, therefore, using a small number of selected direct measurements, to ascertain whether anomalies of thickness thus determined would usefully correlate with other clinicopathological data and so meet the needs of a diagnostic service. This paper describes the simple technique employed and outlines the correlations obtained.
Subjects and methods Forty-five cases of glomerular disease were studied. These fell into three groups determined by clinical and pathological data. Proteinuria in excess of 200 mg/24 h was regarded as clinically significant.
Group I This comprised 15 haematuric cases with normal renal function biopsicd between January 1988 and March 1989 and in whom the immunofluorescence findings were negative (non-characteristic for any specific immune-complexmediated glomerulopathy). This group was subdivided. Group la comprised those cases (n = 9) without significant proteinuria, haematuria being the sole finding. Group lb comprised those with, in addition to haematuria, varying degrees of proteinuria (n = 6). Within this latter subgroup were four cases of minimal-change/IgM nephropathy and two cases whose only satisfactory designation may be 'idiopathic haematuria'.
Group 2 This comprised 14 haematuric cases with biopsy evidence of immune-complex-mediated disease (11 IgA nephropathy. 1 Henoch Schoenlein purpura. 2 systemic lupus erythematosus). All patients in this group also displa>ed some degree of proteinuria.
A. L. C. McLay eI al.
Group 3 In this group there were 16 non-haematuric patients biopsied on account of proteinuria and/or mild impairment of renal function. Subgroup 3a comprised eight cases displaying no evidence of immune-complex deposition (i.e. minimal-change nephropathy) whilst subgroup 3b (also eight cases) included a variety of immune-complex-mediated nephropathies (5 membranous, 2 IgA nephropathy. I Henoch-Schoenlein purpura).
Group 4 Control glomerular samples were derived from cases where no glomerular pathology was anticipated, and these were divided into three subgroups: Group 4a. This comprised five patients with acute-onset renal failure attributable to acute interstitial nephritis. Group 4b. In this group there were a further eight patients with acute-onset renal failure attributed to acute tubular necrosis. Group 4c. This comprised nine patients undergoing nephrectomy for tumour masses, presumptively normal glomeruli being selected from well-defined cortical tissue remote from the tumour mass.
Methods All material for ultrastructural study was fixed in 4% glutaraldehyde in phosphate buffer at pH 7.4, post-fixed in osmium tetroxide, embedded in EMIX resin (Emscope Ltd,) and stained with lead citrate and uranyl acetate. All specimens were examined in a Phillips 301G electronmicroscope at a nominal magnification of x 9800 The microscope magnification was calibrated using a crossed diffraction grating ruled at 2160 lines per millimetre. One glomerulus from each case was examined and electronmicrographs obtained from each of the four thinnest uniform peripheral loop elements. The basal lamina was measured at four representative points along a length of not less than 3.6 urn per loop at a final print magnification of x 27 000. All measurements were made from the overlying epithelial cell plasma membrane to the opposing endothelial cell plasma membrane. Furthermore, measurements were restricted to areas where both cell membranes were closely applied to the basal lamina, avoiding areas of focal attenuation and notching. Moreover, in cases of immunecomplex-mediated disease, measurements were restricted to peripheral loop elements which were not involved with electron-dense deposits. It was also considered important to select regions where endothelial fenestrations were clearly sectioned perpendicular to their diameter, thus further minimizing obliquity of cut. The arithmetical mean of 16 values obtained was then calculated. Measurements were undertaken 'blind' in so far as no reference was made to clinical data. However, it is clear that the micrographs themselves could provide the observer with specific diagnostic information (e.g. electron-dense deposits or foot process effacement). To test the variation in thickness measurement due to sampling bias, a second glomerulus was examined in three cases from group la and three from the control group 4c. Clinical correlation was based on detailed case sheet data analysis with particular reference to creatinine at presentation and to quantitation of haematuria and proteinuria. It was. moreover, considered that intrinsic red cell properties
Glomemlar basement membrane thinning
might influence haematuria. Blood group status was. therefore, also ascertained. Student's ' test was cmp!ovcd for evaluation of differences in laminar width between groups defined by clinical and immunoh'Stochcmical criteria.
Results The mean laminar thickness value for the whole control group, i.e. group 4 (n = 22) was 343 nm (SD 62 nm). The differences between normal laminae (Figure la) and markedly thinned laminae (Figure lb) are identifiable even on casual scrutiny, although subnormal values encompass a wide spectrum which extends from the thinnest to virtually normal. The measurements undertaken on a second glomemlus in three control cases (including one with a thin-membrane state) differed from the original mean only by + 7 to -lOnm. In the additional measurements carried out on cases from group la the deviations from the first glomerular mean covered a somewhat wider range, (-3 to -34 nm). All deviations here, however, only serve to lower the values of a group already distinguished by low values. The distribution of values for individual patients in each group is displayed in Figure 2. The range depicted refers to the highest and lowest average loop values. It is clear that the patients of group la form a distinct cohort in respect of laminar thickness. The mean value for this group is 225 nm (SD = 29 nm). The difference between this and the group 4 value is highly significant (P< 0.0001). No statistical difference is observable between groups lb, 2 or 3 and the control group, 4. It can also be seen that all the mean values for group la lie below an arbitrary value of 270 nm, intermediate between the 264 nm threshold of abnormality suggested by Tiebosch et al. [9] and that of 273 nm suggested by Saxena et al. [12]. These nine haematuric cases with thin laminae and no clinically significant proteinuria (group la) spanned a time period in which 17 cases of IgA nephropathy were diagnosed, suggesting an incidence of a similar order of magnitude. The clinical correlates and basal laminar measurements of this group are summarized in Table 1. All were normotensive and had unimpaired renal function. A clearly defined family history of haematuria was elicited from only one patient. Seven of the nine were female and two patients suffered episodes of macroscopic haematuria in addition to persistent microscopic haematuria. Light-microscopv showed glomeruli in this group to be normal or to be characterized by only mild focal segmental hyperce!!u!aritv. No significant tubulointerstitial changes were noted. Where immunoglobulin was identified on immunofluorescence examination only scanty focal mesangial deposits of
193
IgM were seen. Ultrastructura! examination often showed, in addition to laminar thinning, fairly prominent mesangial paramesangial wrinkling (Figure 3). No case, however, displayed the splitting lamellation characteristic of a proportion of cases of Alport's syndrome. Moreover, enquiry into family history failed to reveal any relatives with Alport's s>ndrome, and deafness was not clinically evident although formal audiometric studies were not undertaken. Blood group analysis of all cases revealed that in this thin-laminar population eight of nine cases were characterized by the O phenotype. This figure of 89% contrasts strongly with the local population norm of 55% and the figure of 62% for all other biopsy cases. This, therefore, raises the possibility of a significant association with blood group O. Group lb, characterized by proteinuria in addition to haematuria, fails, on the other hand, to display uniformly low values of laminar thickness. It is, moreover, associated with a wider spectrum of clinical patterns. The full data for this group are presented in Table 2. In two cases the thickness value lies below the arbitrary 270 nm datum. One such case (case 11) bore a diagnosis of IgM nephropathy whilst the other (case 10) is regarded as one of'idiopathic' haematuria but differs only marginally from group la cases in terms of proteinuria. The remainder of these cases in lb had normal laminar values despite suffering from haematuria which was not attributable to immunecomplex disease. Apart from cases 10 and 11 in group lb, individual cases in other groups, including controls, fall below the suggested critical value of 270 nm. The data for those associated with defined glomerulopathy are outlined in Table 3. These thus include cases of IgM nephropathy, FgA nephropathy, Henoch-Schoenlein purpura, minimal-change disease (Figure 4) and even membranous nephropathy (Figure 5). It is noteworthy that three such cases did not display haematuria despite the evidence of marked laminar thinning.
Discussion We have demonstrated here that patients with glomerular haematuria and no proteinuria are consistently characterized by subnormal laminar thickness values, as determined by a simplified measurement technique. Low values in individual cases are demonstrable within other clinicopathological groups, including controls, but, as populations, the mean values from these other groups fall within the normal range. Furthermore, whilst 13 of 29 haematuric cases displayed mean laminar thickness values below 270 nm, this was true for only three of 16 non-haematuric glomerulopathic cases and two of 22 controls. \*ore-
A. L. C McLay el al
194
31. M.
Jfl
Fig. 1. (a) N o r m a l lamina x 27 000 (b) M a r k e d l ) thinned lamina x 27 000.
Glomerular basement membrane thinning
195
° Femals • Ma!e 700 n
3a
3b
4b
4a
4c
Fig. 2. Laminar thickness values for individual patients. (Mean and interloop range shown.) la = Haematuria,no proteinuria—non-characteristic immunofluorescence; lb=Haematuria + proteinuria—non-characteristic lmmunofluorescence; 2 = Haematuria + proteinuria—characteristic immunofluorescence; 3a = Proteinuria only—negative immunofluorescence (minimal change); 3b = Proteinuria only—characteristic immunofluorescence; 4a = Controls—acute interstitial nephritis; 4b = Controls— acute tubular necrosis; 4c = Controls—nephrectomy. Table 1. Clinical correlates and glomerular basal laminar thicknesses (in nanometres) for group la Case
Sex
Age
Clinical details
Addis
QP
No.
1 2
F
3 4 5 6 7 8
F F F M F
33 58 54 52 23 36
9
M
32
F
F
41 56
Microhaematuria x 3 years Macro + Microhaematuria x 8 years MicTohaematuria x 1 'A years Microhaematuria x 1 year Microhaematuria Microhaematuria MicTohaematuria x 1 year Microhaematuria x 7 years Positive FHx Microhaematuria x 9/12
14 48 29 57 6
50 mg 0 0 0 0
Blood group
Basal lamina Mean
Range
Opos Apos
194 195
162-218 151-240
199 200 226 230 253 263
155-251 159-237 181-270 226-236 218-307 185-344 237-285
_ 200 200
50 mg 200 mg 50 mg
Opos Opos Opos Opos Opos Opos
18
50 mg
Opos
268
Blood group
Basal laminar dimensions
QP = Quantitative proteinuria; Addis = Number of red cells per high-power field. Table 2. Clinical correlates and glomerular basal laminar thicknesses for group lb. Case
Sex
Age
Clinical details
Addis
QP
No.
10 11
M
12 13 14
M F
M
39 55 21
15
M
62
F
21 37
MacTohaematuria x 4/12 Microhaematuria + proteinuria Nephrotic + haematuria Nephrotic + haematuria Microhaematuria x 1 year + macro + proteinuria Nephrotic + haematuria
over, within the limits of the small number of cases in which a second glomerulus was examined, there appears to be good evidence that the values obtained on one glomerulus are representative for any given
200 8
+ +
0.5 g 5.6 g 18g 13g
Mean
Range
Opos Opos
239 213
227-248 177-244
-
270 301 375
262-288 262 -329 297^133
401
314 527
+
0.7 g
Bpos Opos
+
7.87 g
Opos
case. We therefore believe that the technique employed is of practical value in confirming the presence of thin laminar states. Furthermore, when taken in conjunction with other clinical and patholo-
186
Fig. 3. Haematuric patient with thin laminae and non-charactenstic lmmunofluorescence There is conspicuous mcsangial/paramesangial wrinkling (Arrows), x 27 000. Table 3. Clinical correlates and GBM thicknesses for cases with 'significant' laminar thinning in groups 2, 3a, and 3b Case No
18 19 30 31 38
Group No.
2 2 3A 3A 3B
Sex
M F F F F
Age
32 36 64 52 18
Diagnosis
IgA nephropathy HSP Minimal change Minimal change Membranous
gical data, it may be used to identify those cases in which haematuria is most likely to be attributable to an intrinsic laminar abnormality. Issue may well be taken with the selective absolute measurement approach used in this study. However, we believe it is well justified on several counts. First of all, although the majority of previous studies of relevance have employed the orthogonal intercept method, careful scrutiny of the key reference [12] indicates that this latter method was devised for the general case of laminar measurement and then validated in the case of the glomerular basal lamina where absolute values could be determined with some accuracy since 'in this special case the membrane is delineated on both sides, allowing one to measure membrane width onl> at sites where the sectioning is within a known, narrow range of perpendicularity.'
Addis count
27 100 0 0 0
QP (g/day)
0.95 1 5 14.0 2.5
Blood group
B pos O pos OPos — A pos
Basal laminar dimensions Mean
Range
250 267 188 262 222
212-308 229 289 148 214 232 308 192-248
Areas of tangential section, and in pathological states of attenuation or thickening related to damage, are readily identifiable and easily excluded from morphometric assessment. Thus, direct measurement of architecturally normal laminae is eminently practical. Although the convention hitherto has been to consider only diffuse thinning as pathologically relevant, there are no a priori grounds for presuming this to be true. Indeed it seems particularly pertinent that in Alport's syndrome it has been observed [14,15] that irregular thickened laminar states evolved (presumably through recurrent damage) from laminae of subnormal thickness. Thick and thin elements may be observable within any given capillary loop so that average laminar thickness values may fall within or even exceed the normal range even although the intrinsic undamaged thickness is subnormal. In this
Glomerular basement membrane thinning
Fig. 4. Foot process effacement associated with thin laminar state in minimal-change nephropathy. x 27 000.
Fig. 5. Membranous nephropathy with thin laminae. Arrows identify subepithelial deposits, x 27 000.
197
198
context it is of considerable interest to note that in the study of Coleman et al. [8], one patient had 'an excess of thin GBM although the mean thickness was normal.' A case such as this serves to emphasize the fact that random averaged measurements may obscure the underlying abnormality. It seems likely that any thin-laminar condition could display similar features although severity and rate of progression would depend upon the specific molecular defect. We believe, therefore, that identification of a population with intrinsically thin membranes depends on the measurement, within any given case, of the thinnest elements which are not associated with identifiable architectural disturbance. Secondly, the control values in this study are close to those of recent studies using conventional epoxy resins [9,10] and generally comparable with the lower end of normality in the majority of studies. The important influence of the embedding medium on measured thickness is emphasized by Haynes [16]. Thirdly, as we have stated, the simple measurement technique employed here clearly identifies a specific haematuric population with no proteinuria and no evidence of immunologically mediated disease. This group meets the classical criteria for thin-basementmembrane 'syndrome' and the method would appear, therefore, to be entirely suitable for routine diagnostic practice. There would be no reason to believe, moreover, that the approach to measurement employed in this paper would necessarily identify populations coincident with those delineated by the harmonic mean of orthogonal intercepts. Thus, validation is provided by its clinicopathological correlations rather than by comparison with the orthogonal intercept method. Although it may be surprising to some, the finding of overlaps between thin-membrane states and other definable glomerulopathies confirms the impression gained in our pilot study [ 13] and is entirely consistent with the identified presence of thin membranes in the control population. This latter incidence corresponds with the recent observations of Dische et al. [17] in a transplant donor population. Moreover, more extensive attenuation of basal laminae in IgA nephropathy has been noted [18], although the incidence in that study seems unusually high whilst the magnification selected for measurement (x 2700) can be regarded as too low for the necessary degree of precision. We believe that where overlaps occur, phenomena other than laminar thinning are likely to have a greater influence on prognosis and it appears to us significant that many of the poorer prognosis cases in other studies had proteinuria in excess of 1 g/24 h and, or features suggestive of additional complicating factors such as malignant hypertension or deafness.
A. L C. McLay et al.
In view of the presence of haematuria in renal neoplasia we feel that no useful clinical correlations can be made with the thin-laminar states in Group 4c. However, it is of further interest to note that three of the seven cases showing both thin membranes (188, 222 and 262 nm) and other primary glomerular pathology did not suffer from haematuria despite membrane thicknesses as low as those in the haematuric group. We therefore share some of the concerns recently expressed [19] regarding arbitrary correlation between membrane thinning and haematuric glomerular disease. If one assumes that non-immunologically mediated haematuria is related to laminar fragility, then it would seem clear that thinness does not necessarily imply fragility. This suggests that laminar thinning may be a consequence of more than one molecular defect, a fact already implicit in the observation [20] that classical thin-membrane cases possess the Goodpasture antigen, unlike Alport cases, which may also display thinning [3,15]. It would therefore seem advisable to reserve the term 'thin-membrane disease (syndrome)' for haematuric cases with demonstrable laminar thinning but no evidence of immunologically mediated disease and with proteinuria of less than 200 mg/24 h. Within this group, in our series, no untoward trends in renal function have so far (time from biopsy 24-41 months) been observed. We believe that other cases should simply be regarded as exhibiting thin-laminar states. The demonstration within our thin-membrane disease group of such a high prevalence of blood group O is, in such a small sample, insufficient proof of a specific relationship, but seven of eight further cases identified since the completion of the formal study also belong to blood group O and the finding would appear therefore to merit further investigation. Whilst in the case of Alport's syndrome the laminar defect appears to relate to absence of Goodpasture epitopes, which are located in the non-collagenous globular domain of the carboxy terminal end of type IV collagen, [21], it may be that alterations of associated carbohydrate moieties might give rise to similar structural problems. Such alterations might be mediated by glycosyltransferases capable of influencing histoblood group antigen expression, and it is tempting therefore to speculate whether polymorphism of such a carbohydrate might correspond with known polymorphism in histo-blood group antigens [22]. Thus, if the blood group relationship suggested by the present study can be confirmed in a larger series, detailed investigation of phenomena such as secretor status and the expression of other blood group systems might provide a starting point for more precise evaluation of the nature of the presumed genetic defect involved. In conclusion, whatever the underlying defect, thin-
Glomerular basement membrane thinning
membrane states may be easily diagnosed in biopsy populations using this simple measurement technique. As others have indicated, thin-membrane disease is a common cause of persistent microscopic haematuria in non-uraemic adults. However, thin-laminar states overlap with other diagnostic categories and the latter may be of much more prognostic relevance in individual cases. Furthermore, a thin-laminar state does not necessarily result in glomerular haematuria. We believe, therefore, that thin-membrane disease with its implicit good prognosis should only be diagnosed in patients with haematuria, no proteinuria, and no evidence of a characteristic pattern of immunoglobulin deposition. Moreover, this phenomenon appears to have a strong but not invariable association with blood group O. Acknowledgements. The authors wish to thank Miss J. S. Munro for her skilled technical assistance throughout this study. They are also indebted to Mr C. Buck, Mr P. Paterson, and Mr R. Scott of the Urology Department, Glasgow Royal Infirmary for their assistance in obtaining fresh tissue from nephrectomies performed on patients under their care. Thanks are also due to Dr G. Murray of the University Department of Surgery, Glasgow Royal Infirmary, for his observations on certain statistical issues raised.
199
6. 7. 8.
9. 10. 11. 12. 13. 14. 15. 16. 17.
References 1. Rogers PW, Kurtzman NA, Bunn SM Jr, White MG. Familial benign essential hematuria. Arch Intern Med 1973; 131: 257-262 2. Yoshikawa N, White RHR, Cameron AH. Familial hematuria: clinico-pathological correlations. Clin Nephrol 1982; 17: 172-182 3. Yoshikawa N, Hashimoto H, Katayama Y, Yamada Y, Matsuo T. The thin glomerular basement membrane in children with haematuria. J Pathol 1984; 142: 253-257 4. Tina L, Jenis E, Josse P, Medani C, Papadopoulou Z, Cateagno P. The glomerular basement membrane in benign familial hematuria. Clin Nephrol 1982; 17: 1-4 5. Dische FE, Weston MJ, Parsons V. Abnormally thin glomerular basement membranes associated with hematuria, pro-
18. 19. 20.
21.
22.
teinuria or renal failure in adults. Am J Nephrol 1985; 5: 103-109 Abe S, Amagasaki Y, Iyori S et al. Thin basement membrane syndrome in adults. J Clin Pathol 1987; 40: 318-322 FujugaJd Y, Nagase M, Kobayashi S, Honda N", Muranaka Y. Alterations of glomerular basement membrane relevant to haematuria Virchows Arch [A] 1983; 413: 159-165 Coleman M, Haynes WDG, Dimopoulos P, Barratt LJ, Jarvis LR. Glomerular basement membrane abnormalities associated with apparently idiopathic hematuria. Hum Patliol 1986; 17: 1022-1030 Tiebosch ATMG, Frederik PM, van Breda Vriesman PJC el al. Thin basement-membrane nephropathy in adults with persistent hematuria. A' Engl J Med 1989; 329: 14-18 Aarons I, Smith PS, Davies RA, Woodroffe AJ, Clarkson AR. Thin membrane nephropathy: a clinico-pathological study. Clin Nephrol 1989; 32: 151-158 Saxena S, Davie DJ, Kirsner RLG. Thin basement membranes in minimally abnormal glomeruli. J Clin Pathol 1990; 43: 3 2 38 Jensen EB, Gundersen HJG, Osterby R. Determination of membrane thickness distribution from orthogonal intercepts. J Microsc 1979; 115: 19-33 McLay ALC, Meyboom F, Jackson R. Thin glomerular basal laminae and adult glomerular disease. J Pathol 1989; 157: 162A Antonovych TT, Deasy PF, Tina U, D'Albora JB, Hollerman CE, Calcagno PL. Hereditary nephritis: early clinical, functional and morphological studies. Pediatr Res 1969; 3: 545-556 Yum M, Bergstein JM. Basement membrane nephropathy. Hum Pathol 1983; 14: 996-1003 Haynes WDG. The normal human renal glomemlus. Virchows Arch B 1981; 35: 133-158 Dische FE, Anderson VER, Keane SJ, Taube D, Bewick M, Parsons V. The incidence of thin membrane nephropathy in kidney donors. Nephrol Dial Transplant 1990; 5: 306. (Abstract) Morita M, Sakaguchi H. A quantitative study of glomerular basement membrane changes in IgA nephropathy. J Pathol 1988; 154: 7-18 Editorial. Thin-membrane-nephropathy—how thin is thin? Lancet 1990; 336: 469-470 Dische FE, Brooke IP, Cashman SJ el at. Reactivity of monoclonal antibody PI with glomerular basement membrane in thin-membrane nephropathy. Nephrol Dial Transplant 1989; 4: 611-617 Butkowski RJ, Wieslander J, Wisdom BJ, Barr JF, Noelken ME, Hudson BG. Properties of the globular domain of type IV collagen and its relationship to the Goodpasture antigen. J Biol Chem 1985; 260: 3737-3739 Clausen H, Hakomori S. ABH and related histo-blood group antigens: Immunochemical differences in carrier isotypes and their distribution. Vox Sang 1989; 56: 1-20
Received for publication 15.1.91 Accepted in revised form 15.8.91
