- Email: [email protected]
Tubular lesions determine prognosis of IgA nephropathy
ORIGINAL INVESTIGATIONS
Tubular Lesions Determine Prognosis of IgA Nephropathy Laurent Daniel, MD, Yannick Saingra, MD, Roch Giorgi, MD, Corinne Bouvier, MD, Jean-Franc¸ois Pellissier, MD, and Yvon Berland, MD ● To assess the prognostic value of histological classification for renal outcome, we did a multivariate analysis of 194 patients with immunoglobulin A (IgA) nephropathy between 1985 and 1995. We also evaluated semiquantitative scales of tubular lesions and vessel lesions. At the time of the biopsy, 65 patients (33.5%) had chronic renal failure. At the end of the follow-up, 33 patients (17%) required hemodialysis. The mean age of the patients was 37.8 ⴞ18.9 years with predominance of men (sex-ratio: 3.12). Patients were followed-up for a mean of 43.2 ⴞ 37.2 months. Univariate analysis showed that hypertension (P F 10ⴚ4), nephrotic syndrome (P ⴝ 0.01), and crescents (P ⴝ 0.02) were significant in predicting renal failure, unlike subendothelial topography of IgA deposits (P ⴝ 0.05) and proteinuria (P ⴝ 0.05). Hematuria was a protective factor (P ⴝ 0.03). Multivariate analysis showed that tubular grade 2 (relative risk [RR], 5.5) and tubular grade 3 (RR ⴝ 28.8) were the best factors to predict chronic renal failure. The histological classification of Haas was significant in the univariate analysis, but not in the multivariate analysis. Tubular grading predicted renal outcome better than did the other histological parameters. 娀 2000 by the National Kidney Foundation, Inc. INDEX WORDS: Glomerulonephritis; immunoglobulin A (IgA); prognosis; classification; tubular grading.
P
RIMARY immunoglobulin A (IgA) nephropathy is the most common glomerulopathy in Europe and France.1-3 Although initially reported as benign,1 the clinical course of IgA nephropathy is variable; it leads to chronic renal failure (CRF) or end-stage renal failure (ESRF) in one third of patients.4,5 Evidence has accumulated that more severe proteinuria and hypertension mean progression of the disease. Furthermore, at time of diagnosis, pathological features were predictors of renal outcome in IgA nephropathy.6 Histological grading systems have been used to evaluate the prognosis of IgA nephropathy.7-9 The classification of Haas seems to be more predictive than previous systems and is increasingly used.9 We evaluated the classification of Haas, tubular lesions, vascular lesions, IgA deposit topography, and crescents to determine which parameter best predicts renal outcome. We therefore retrospectively evaluated renal biopsy specimens in 194 patients with primary IgA nephropathy. PATIENTS AND METHODS
Patients We retrospectively retrieved all renal biopsies performed in Marseille, in the south of France, from January 1985 to December 1995. We selected 233 biopsies with the diagnosis of IgA nephropathy. A questionnaire was addressed to the correspondent nephrologists to know the clinical and biological parameters at the time of the biopsy and the outcome of patients. Biopsies were performed because of isolated pro-
teinuria, proteinuria with hematuria, or deterioration of the renal function without evident cause. Biopsies were not performed on patients with isolated hematuria. Proteinuria was considered significant if it was over 500 mg/d. The threshold for renal insufficiency in statistical analyses was 1.7 mg/dL. To be entered into the study, patients must have been free of steroidal treatment. Other treatments were only symptomatic.
Renal Biopsies We divided the biopsy specimen into two parts for direct immunofluorescence (IF) and light microscopy (LM). For LM, tissue was fixed 12 hours in 2.5% glutaraldehyde, and 2-µm-thick sections were stained with Gomori’s trichrome. Sections 0.2 µm thick were stained with Jones silver impregnation. For IF, tissue was fixed 24 hours in Michel liquid10,11 before being snap frozen in ⫺80°C isopentane, and then it was stored at ⫺80°C until use. Cryostat sections 5 µm thick were reacted with fluorosceinated rabbit antibodies specific for IgG, IgA, IgM, C1q, C3, fibrinogen, and immunoglobulin light chains. Antibodies were purchased from Binding System (Birmingham, UK). Sections were examined using a Reichert-Jung microscope equipped with epifluorescence
From the Service d’Anatomie Pathologique, C-H-U Timone, Marseille, France; the Service de Ne´phrologieDialyse, Hoˆpital de Sainte-Marguerite, Marseille, France; and the Service de l’Information Me´dicale, C-H-U Timone, Marseille, France. Received April 15, 1999; accepted in revised form August 3, 1999. Supported by institutional grants from AP-HM. Address reprint requests to Laurent Daniel, MD, Service d’Anatomie Pathologique (Pr Pellissier), C-H-U Timone, 264 rue Saint-Pierre 13385 Marseille cedex 05, France. E-mail: [email protected]
娀 2000 by the National Kidney Foundation, Inc. 0272-6386/00/3501-0003$3.00/0
American Journal of Kidney Diseases, Vol 35, No 1 (January), 2000: pp 13-20
13
14
illumination optics (Leica, Marseilles, France). The presence of subendothelial deposits was assessed by IF, and intensity was graded from 0 to 3⫹. Only intensities from 1⫹ to 3⫹ were considered significant for IgA and non-IgA deposits. Subendothelial deposits were also identified with semithin sections.
Histological Evaluation Classification. We graded renal biopsy specimens according to the classification of Haas.9 We considered only the first biopsy specimen and excluded renal allograft biopsy specimens. We required a minimum of three glomeruli for LM and for IF. Frozen specimen was then paraffinembedded for conventional examination to determine whether the specimen was homogenous. Such specimens were considered an adequate renal biopsy.12 The summarized classification is as follows: Subclass I: All glomeruli are normal or show a slight mesangial hypercellularity with or without detectable deposits. Subclass II: Some glomeruli show typical segmental sclerosis in a pattern resembling primary focal segmental glomerulosclerosis. A slight mesangial hypercellularity is accepted. Subclass III: Fifty percent or fewer of the glomeruli show mesangial or endocapillary substantial proliferation. Proliferation may be global or segmental. Crescents are present or not. Subclass IV: More than 50% or all the glomeruli show mesangial or endocapillary proliferation. Crescents are present or not. Subclass V: Fibrous lesions affect 40% of the glomeruli or 40% of the interstitium with tubular atrophy. Tubular grading. To assess tubular lesions, we examined serial sections of the cortical area except in the subcapsular area. The medulla was not examined. We considered as lesions atrophic tubules with wrinkled basal membrane, and dilated tubules. If tubules were not back-to-back, we established whether there was fibrosis causing this appearance. Atrophy and dilations were often associated in the same specimen. Hyaline or hematic casts were not considered lesions. The severity of lesions was graded by the grading of Katafuchi et al,13 modified as follows: Grade 0: Normal tubules Grade 1: Tubular lesions involving less than 25% of the cortical area Grade 2: Tubular lesions involving less than 50% but more than 25% of the cortical area Grade 3: Diffuse tubular lesions or involving more than 50% of the cortical area Vessel grading. We examined abnormalities of the arteriolar wall on semithin sections. The severity of lesions was graded by the grading of Katafuchi et al,13 modified as follows: Grade 0: Normal vessel with a ratio (R) of luminal diameter to outer diameter less than 0.5. Grade 1: Thin hyaline deposits or intimal sclerosis in the arteriolar wall with R higher than 0.25 and without
DANIEL ET AL
smooth muscle proliferation. Proliferation is more than two cells per cross-section. Grade 2: Thick hyaline deposits in the arteriolar wall with R less than 0.25 without smooth muscle proliferation. Lesions are not diffuse. Grade 3: Diffuse thick hyaline deposits in the arteriolar wall with R less than 0.25. Likewise, a biopsy specimen can show smooth muscle proliferation.
Statistical Analyses We established correlations between clinical data, biological data, and histological parameters by chi-squared test or Yates-modified chi-squared test (if theoretical counts were less than 5). Relative risk (RR) was calculated with 95% confidence interval (CI) for the patients with ESRF and those with CRF. Renal survival data were estimated by the Kaplan-Meier method and compared using the log-rank test. P values less than 0.05 were considered significant in univariate analysis. Significant variables in univariate analysis were used in multivariate analyses. Multivariate analyses were done by logistic regression and Cox proportional hazard regression with forward stepping. Thresholds for entering or removing variables were 0.10 and 0.15, respectively. Adjusted odds ratio (OR) was calculated with 95% CI. Data were analyzed by BMDP-system 1.1 software (Meridian Marketing Group, Sedona, AZ).
RESULTS
Clinical and Biological Parameters Between 1985 and 1995, we studied 4,568 biopsy specimens and found that 233 (5.1%) of them had IgA nephropathy. We excluded 39 patients because of coexisting disease such as Scho¨nlein-Henoch purpura, liver cirrhosis, and lupus erythematosus. Finally, we selected for analysis 194 patients with primary IgA nephropathy. Patients were 147 men and 47 women (sex ratio: 3.12). The mean age was 37.8 ⫾ 18.9 years (range, 4 to 81 years). Twenty-three cases occurred in childhood (range, 4 to 15 years). At the time of the biopsy, 117 patients (60.3%) evidenced proteinuria, which was at nephrotic range in 17 patients (8.8%). Hematuria was diagnosed in 145 patients (74.7%). Hypertension, defined as blood pressure exceeding 150/90 mm Hg, was observed in 63 patients (32.5%). At the time of the biopsy, 65 of the 194 patients (33.5%) had increased serum creatinine (⬎1.7 mg/dL). The mean follow-up was 43.2 ⫾ 37.2 months (range, 1 to 141 months). The appearance of CRF and its progression to ESRF varied; 61 of the 194 patients (31.4%) had an increased serum creatinine at the last consulta-
TUBULAR LESIONS OF IgA NEPHROPATHY
15
Fig 1. (A) Mesangial deposits (arrowheads) in a glomerulus with minimal changes (subclass I) (Jones, original magnification ⴛ1,000). (B) Glomerulus with focal and segmental lesion (arrow) of the floculus (subclass II) (Jones, original magnification ⴛ250). (C) Glomerulus with segmental obstruction of capillaries by proliferated endocapillary cells (arrows) (subclass III) (Jones, original magnification ⴛ400). (D) Glomerulus with segmental crescent (subclass IV) (Masson’s trichrome, original magnification ⴛ400).
16
DANIEL ET AL Table 1. Univariate Analysis: Risk Factors for CRF Variables
CRF⫹ (n; %)
CRF⫺ (n; %)
RR
CI95%
P
Proteinuria Nephrotic syndrome Hematuria Hypertension Crescents Subendothelial deposits Tubular index grade 1 Tubular index grade 2 Tubular index grade 3 Vessel index grade 1 Vessel index grade 2 Vessel index grade 3 Haas subclass II Haas subclass III Haas subclass IV Haas subclass V
43 (70.5) 10 (16.4) 39 (63.9) 42 (68.8) 18 (29.5) 15 (24.5) 7 (11.4) 25 (41) 19 (31.1) 9 (14.7) 21 (37.7) 14 (22.9) 1 (1.6) 7 (11.5) 10 (16.4) 27 (44.2)
74 (55.6) 7 (5.2) 106 (79.1) 21 (15.6) 16 (11.9) 18 (13.4) 26 (19.4) 16 (12) 4 (2.9) 16 (12) 16 (11.9) 4 (2.9) 7 (5.2) 16 (12) 2 (1.5) 13 (9.7)
1.6 2 0.6 4.6 2 1.6 3.1 8.9 12.1 2.9 4.5 6.2 1.1 2.8 7.6 6
1-2.5 1.3-3.2 0.4-0.9 2.9-7.2 1.3-3.1 1-2.5 1-8.5 4-20.1 5.5-26.8 1.4-6 2.5-8.2 3.6-11 0.2-7.8 1.2-6.4 4.9-14.1 3.4-11.2
0.05 0.01 0.02 ⬍10⫺4 0.002 0.05 0.04 ⬍10⫺4 ⬍10⫺4 0.015 ⬍10⫺4 ⬍10⫺4 NS 0.04 ⬍10⫺4 ⬍10⫺4
tion. Ten of 129 patients initially reported with normal renal function developed CRF; six of them required hemodialysis. In contrast, 14 of the 65 patients initially reported with an increased serum creatinine level had under 1.7 mg/dL at the last consultation. Last, 33 patients (17%) developed ESRF. No acute renal failure appeared during the study. Histological Features The five subclasses of the classification of Haas were distributed heterogeneously. The largest subclass was subclass I (101 patients; 55%), corresponding to minimal lesions (Fig 1A). Subclass II was observed in only eight patients (4.3%) (Fig 1B). Subclasses III (Fig 1C) and IV (Fig 1D), corresponding to the active proliferation, were found in 23 patients (12.5%) and 12 patients (6.5%), respectively. Subclass V was observed in 40 patients (21.7%). Thirty-four biopsy specimens (17.5%) had crescentic lesions. Only 13 biopsy specimens (6.7%) showed more than 20% of crescentic glomeruli. Subendothelial deposits were seen in 33 biopsy specimens (17%). One hundred forty biopsy specimens (72.1%) showed C3-associated deposits, whereas IgM, IgG, and C1q were seen in 35 (18%), 21 (10.8%), and 19 (9.8%) biopsy specimens, respectively. Clinicopathologic Correlations Correlations were calculated at the end of the follow-up period (Tables 1 and 2) and for the
group of patients requiring hemodialysis (Tables 3 and 4). At the end of the follow-up, the univariate analysis showed that most of the parameters, except hematuria (RR ⫽ 0.6, CI ⫽ 0.4 to 0.9) and Haas subclass II (RR ⫽ 1.1, CI ⫽ 0.2 to 7.8), were linked with CRF. Proteinuria (RR ⫽ 1.6, CI ⫽ 1 to 2.5; P ⫽ 0.05) and IgA deposit topography (RR ⫽ 1.6, CI ⫽ 1 to 2.5; P ⫽ 0.05) were nearly significant. Results were similar with proteinuria if patients with Haas subclass II were excluded (P ⫽ 0.047). Note that hypertension (OR ⫽ 3.6, CI ⫽ 1.2 to 10.7) was the only significant clinical risk factor in the multivariate analysis. Subclasses IV (P ⬍ 10⫺4) and V (P ⬍ 10⫺4) of the classification of Haas were correlated with CRF only in the univariate analysis. The multivariate analysis showed that tubular grade 2 (OR ⫽ 5.5, CI ⫽ 1.3 to 22.5) and tubular grade 3 (OR ⫽ 28.8, CI ⫽ 4.3 to 190.4) were significant risk factors. Crescents were correlated with CRF (OR ⫽ 11.2, CI ⫽ 2.8 to 59.6). Note also that at the time of the biopsy, crescents were correlated with neither nephrotic syndrome (P ⫽ 0.81) nor tubular grading (P ⫽ 0.9). High Table 2. Multivariate Analysis: Risk Factors for CRF Variables
OR Adjusted
CI95%
P
Tubular index grade 1 Tubular index grade 2 Tubular index grade 3 Crescents Hypertension
1.7 5.5 28.8 11.2 3.6
0.4-7.4 1.3-22.5 4.3-190.4 2.8-59.6 1.2-10.7
NS 0.01 0.001 0.01 0.01
TUBULAR LESIONS OF IgA NEPHROPATHY
17
Table 3. Univariate Analysis: Risk Factors for the Group of Patients With ESRF Variables
ESRF⫹ (n; %)
ESRF⫺ (n; %)
RR
CI95%
P
Proteinuria Nephrotic syndrome Hematuria Hypertension Crescents Subendothelial deposits Tubular index grade 1 Tubular index grade 2 Tubular index grade 3 Vessel index grade 1 Vessel index grade 2 Vessel index grade 3 Haas subclass II Haas subclass III Haas subclass IV Haas subclass V
20 (62.5) 9 (28.1) 18 (56.2) 24 (75) 10 (31.2) 10 (31.2) 2 (6.2) 14 (43.7) 12 (37.5) 5 (15.6) 12 (32.5) 8 (25) 1 (3.1) 4 (12.5) 5 (15.6) 15 (46.8)
97 (60.2) 8 (4.9) 127 (78.8) 39 (24.2) 24 (15.4) 23 (14.8) 31 (20) 27 (17.4) 11 (7.1) 20 (12.9) 25 (16.1) 10 (6.4) 7 (4.5) 19 (12.2) 7 (4.51) 25 (16.1)
1 3.9 0.4 5.5 2.2 2.1 2.6 15 22.9 4.2 6.8 9.3 3.16 4.39 10.52 9.47
0.5-1.9 2.2-7 0.2-0.7 2.7-11.2 1.2-4.3 1.1-4 0.4-18.2 3.6-63 5.5-95.4 1.3-13.4 2.5-18 3.44-25.3 0.4-25 1.2-16.2 3.26-33.92 3.35-26.80
0.95 0.004 0.003 ⬍10⫺4 0.02 0.02 0.3 ⬍10⫺4 ⬍10⫺4 0.02 ⬍10⫺4 ⬍10⫺4 0.3 0.04 0.0005 ⬍10⫺4
vessel grades (grades 2 and 3) (Fig 2A, B) were strongly correlated with hypertension at the time of the biopsy (P ⬍ 10⫺4). These results were similar for the 33 patients requiring hemodialysis (Tables 3 and 4). High tubular grades (grade 2, OR ⫽ 78, CI ⫽ 7.5 to 803.3; grade 3, OR ⫽ 105.3, CI ⫽ 10 to 1,106) were the most evident risk factors. No correlations were found between ESRF and crescents, or hypertension. Hematuria was a protective factor (RR ⫽ 0.4, CI ⫽ 0.1 to 0.9). Interestingly, vessel grading and high subclasses of the classification of Haas correlated with CRF only in the univariate analysis. In the multivariate analysis, nephrotic syndrome correlated with ESRF but not with CRF. Significant variables for renal survival with Cox’s model are reported in Table 5. The selected variables were still the tubular grading (Fig 3A), the nephrotic syndrome (Fig 3B), and hematuria (Fig 3C). Table 4. Multivariate Analysis: Risk Factors for the Group of Patients With ESRF Variables
OR Adjusted
CI95%
P
Nephrotic syndrome Hematuria Tubular index grade 1 Tubular index grade 2 Tubular index grade 3
10.2 0.3 12.7 78 105.3
3.5-29.5 0.1-0.9 0.8-190.2 7.5-803.8 10-1,106
0.06 0.15 NS 0.01 0.01
DISCUSSION
The results show that tubular grading is the highest predictive factor for CRF in IgA nephropathy and that the classification of Haas is not significant in the multivariate analysis. Our data are in agreement with others emphasizing the role of tubular lesions in the progression of IgA nephropathy.14-16 Tubular lesions were strongly correlated with CRF, as were some clinical parameters such as hypertension. In other glomerulopathies, it is often the intensity of accompanying tubular lesions that predicts decline in renal function.17 Tubular lesions reflect the ability of interstitial processes to damage normal glomeruli carrying kidney function. Some authors, nevertheless, excluded tubular lesions from their histological grading.18 The link between tubular lesions and glomerular lesions is probable because glomerular lesions induce tubular lesions through filtered toxic substances or hypertensive hemodynamic changes.18 Inflammatory mechanisms are involved in the pathogenesis of the tubular lesions through increased amounts of urinary cytokines or growth factors such as transforming growth factor , which was implicated in renal fibrosis.19 Our study confirms the male predominance in IgA nephropathy in France. That the mean age of our patients was higher than in other studies is probably due to the policy of our institution, where biopsies are not done for patients with
18
DANIEL ET AL
Fig 2. (A) Thick hyaline deposits (arrow) in an arteriolar wall (grade 3) (Jones, original magnification ⴛ400). (B) Smooth muscle proliferation (arrow) in an interlobular vessel wall (grade 3) (Jones, original magnification ⴛ400).
isolated hematuria. Only 17% of the patients developed ESRF, yet a third of them showed an increased serum creatinine level. Hypertension affected a third of the patients, and this study shows the link between the hypertension and vascular lesions. As previously described,20 our patients with hematuria tended to progress more slowly to CRF. Although this is difficult to explain, it might reflect that these glomeruli are better able Table 5. Multivariate Analysis of Renal Survival With Cox Regression Analysis Variables
RR
Tubular index grade 0 Tubular index grade 0 Tubular index grade 0 Tubular index grade 0 Nephrotic syndrome ⫺ Nephrotic syndrome ⫹ Hematuria ⫺ Hematuria ⫹
1 3.54 11.06 32.26 1 2.75 1 0.44
CI95%
P
0.0001 0.47-26.13 2.46-49.80 6.66-156.10 0.027 1.12-6.78 0.059 0.19-1.03
to cope with the hemodynamic consequences of the IgA deposits. Our data show that nephrotic syndrome was also a marker of progression, but it was observed only in the group of patients with ESRF. In the literature, the patients with focal and segmental sclerosing glomerulopathy-like IgA nephropathy usually have a nephrotic syndrome. The prognostic value of the focal and segmental hyalinosis (Haas subclass II) is debated.21,22 For Packham et al, these features were a marker of progression.22 Like Haas, we believe patients with such lesions have a better prognosis.21 The classification of Lee et al,8 modified by Haas,9 is proposed to predict the prognosis of IgA nephropathy.23 This classification is helpful for a global prognosis evaluation because of its reproducibility. In fact, only 4% of patients with Haas subclass I required hemodialysis, but 16.7% showed CRF at the last consultation. Proliferative lesions (subclasses III and IV) correlated strongly with CRF. The distinction between these two subclasses is justified because CRF at the last consultation was 43.5%
TUBULAR LESIONS OF IgA NEPHROPATHY
19
Fig 3. (A) Renal survival rate in patients with tubular lesions. The survival rate was significantly higher in patients with low-grade tubular lesions (P ⴝ 0.0001). (B) Renal survival rate in patients with or without nephrotic syndrome. Nephrotic syndrome was a significant RR for renal survival (P ⴝ 0.02). (C) Renal survival rate in patients with or without hematuria. Hematuria was a protective factor for renal survival (P ⴝ 0.05).
and 75% for subclass III and for subclass IV, respectively. Surprisingly, Haas subclass V was not significant in multivariate analysis, most likely because tubular grading masked the statistical power of this variable. Crescents were associated with poor prognosis in large series.24,25 This result was nevertheless not consistently found.17,26,27 In our series, 17.5% of the patients had crescentic glomerulonephritis. At the time of the biopsy, crescents were not usually accompanied by progressive scarring of the glomerular tuft. Crescents should account for interstitial lesions because within an adhesion the floculus merges with the interstitium and thus allows diffusion from capillaries toward the interstitium. D’Amico et al20 found that crescents were a negative prognostic factor by univariate analysis. We confirm this fact by a multivariate analysis. Finally, two limitations must be considered if one uses the classification
of Haas: tubulointerstitial analysis does not provide sufficient information, and the crescents are not precisely quantified in the high subclasses. The variability in the methods appreciating vascular lesions in IgA nephropathy raises doubt about the significance of such lesions.28,29 Our data emphasize the link between hypertension and vascular lesions in the progression to CRF. The glomerular sclerosis is nearly certainly enhanced by nonimmunologic factors such as intrarenal hemodynamic changes, but high vessel grades were not significant in our study. The prognostic value of subendothelial deposits in IgA nephropathy is still controversial. As in other studies, we found no prognosis value for this topography.24-26 Conversely, others reported that peripheral topography was a poor prognosis factor in univariate analysis.13,30 Subendothelial deposits indicate an extension of the immunologic injuries to the peripheral capillary walls,
20
DANIEL ET AL
and these deposits are responsible for lesions of the basement membrane, which are a marker of severity of most glomerulopathies.31 To conclude, we underline that tubular grading is essential to predict patient outcome in IgA nephropathy. We show that tubular grading is a stronger predictive factor than are glomerular lesions evaluated by the classification of Haas. We believe tubular grading should be applied in renal biopsy analyses of IgA nephropathy. ACKNOWLEDGMENT The authors thank Annie Fouquet for technical assistance and the physicians who provided follow-up data. We also thank Dr J. Gouvernet for statistical analyses and Prof Figarella-Branger for comments.
REFERENCES 1. Berger J, Hinglais N: Les de´poˆts intercapillaires d’IgAIgG. J Urol Nephrol 74:694-695, 1968 2. Simon P, Rame´e MP, Autuly V, Laruelle E, Charasse C, Cam G, Ang KS: Epidemiology of primary glomerular diseases in a French region: Variations according to period and age. Kidney Int 46:1192-1198, 1994 3. Alamartine E, Sabatier JC, Guerin C, Berliet JM, Berthoux F: Prognostic factors in mesangial IgA glomerulonephritis: An extensive study with univariate and multivariate analyses. Am J Kidney Dis 18:12-19, 1991 4. Ibels LS, Gyory AL: IgA nephropathy. Medicine 73:79102, 1994 5. Galla JH: IgA nephropathy. Kidney Int 47:377-387, 1995 6. Droz D, Kramar A, Nawar T, Noel LH: Primary IgA nephropathy: Prognostic factors. Contrib Nephrol 40:202207, 1984 7. Meadow SR, Glasgow EF, White RHR, Moncrieff MW, Cameron JS, Ogg CS: Scho¨nlein-Henoch nephritis. Q J Med 41:241-258, 1972 8. Lee SMK: Prognostic indicators of progressive renal disease in IgA nephropathy: Emergence of a new histologic grading system. Am J Kidney Dis 29:953-958, 1982 9. Haas M: Histologic subclassification of IgA nephropathy: A clinicopathologic study of 244 cases. Am J Kidney Dis 29:829-842, 1997 10. Michel B, Milner Y, David K: Preservation of tissuefixed immunoglobulins in skin biopsies of patients with lupus erythematosus and bullous diseases: Preliminary report. J Invest Dermatol 59:449-451, 1973 11. Croker BP, Dawson DV, Sanfilippo F: IgA nephropathy: Correlation of clinical and histologic features. Lab Invest 48:19-24, 1983 12. Madaio MP: Renal biopsy forum. Kidney Int 38:529543, 1990 13. Katafuchi R, Kiyoshi Y, Oh Y, Uesugi N, Ikeda K, Yanase T, Fujimi S: Glomerular score as a prognosticator in IgA nephropathy: Its usefulness and limitation. Clin Nephrol 49:1-8, 1998
14. Remuzzi G, Bertani T: Pathophysiology of progressive nephropathies. N Engl J Med 339:1448-1456, 1998 15. Bogenschutz O, Bohle O, Batz G, Whermann M, Pressler H, Kendziorra MV, Gartner HV: IgA nephritis: On the importance of morphological and clinical parameters in the long term prognosis of 239 patients. Am J Nephrol 10:137-147, 1990 16. Bennett WM, Walker RG, Kincaid-Smith P: Renal cortical interstitial volume in mesangial IgA nephropathy. Lab Invest 47:330-335, 1982 17. Abe S, Amagasaki Y, Iyori S, Konishi K, Kato E, Sakaguchi H: Significance of tubulointerstitial lesions in biopsy specimens of glomerulonephritic patients. Am J Nephrol 9:30-37, 1989 18. Koyama A, Igarashi M, Kobayashi M: Natural history and risk factors for immunoglobulin A nephropathy in Japan. Am J Kidney Dis 29:526-532, 1997 19. Border WA, Noble NA: Transforming growth factor- in tissue fibrosis. N Engl J Med 331:1286-1292, 1994 20. D’Amico G, Imbasciati E, Barbiano di Belgioioso G, Bertoli S, Fogazzi G, Ferrario F, Fellin G, Ragni A, Colasanti G, Minetti L, Ponticelli C: Idiopathic IgA mesangial nephropathy: Clinical and histological study of 374 patients. Medicine 64:49-60, 1985 21. Haas M: IgA nephropathy histologically resembling focal-segmental glomerulosclerosis: A clinicopathologic study of 18 cases. Am J Kidney Dis 28:365-371, 1996 22. Packham DK, Yan HD, Hewitson TD, Nicholls KM, Fairley KF, Kincaid-Smith P, Becker GJ: The significance of focal and segmental hyalinosis and sclerosis (FSHS) and nephrotic range proteinuria in IgA nephropathy. Clin Nephrol 46:225-229, 1996 23. Lee HS, Koh HI, Lee HB, Park HC: IgA nephropathy in Korea: A morphological and clinical study. Clin Nephrol 27:131-140, 1987 24. Kobayashi Y, Tateno S, Hiki Y, Shigematsu H: IgA nephropathy: Prognostic significance of proteinuria and histological alterations. Nephron 34:146-153, 1983 25. Radford MG, Donadio JV, Bergstralh EJ, Grande JP: Predicting renal outcome in IgA nephropathy. J Am Soc Nephrol 8:199-207, 1997 26. Kusumoto Y, Takebayashi S, Tagushi T, Harada T, Naito S: Long term prognosis and prognostic indices of IgA nephropathy in juvenile and in adult Japanese. Clin Nephrol 28:118-124, 1987 27. Magil AB, Ballon HS: IgA nephropathy: Evaluation of prognostic factors in patients with moderate disease. Nephron 47:246-252, 1987 28. Feiner HD, Cabili S, Baldwin DS, Schacht RG, Gallo GR: Intrarenal vascular sclerosis in IgA nephropathy. Clin Nephrol 18:183-192, 1982 29. Okada H, Suzuki H, Konishi K, Sakaguchi H, Saruta T: Histological alterations in renal specimens as indicators of prognosis of IgA nephropathy. Clin Nephrol 37:235-238, 1992 30. Mustonen J, Pasternack A, Helin H, Nikkila¨ M: Clinicopathologic correlations in a series of 143 patients with IgA glomerulonephritis. Am J Nephrol 5:150-157, 1985 31. Hricik DE, Chung-Park M, Sedor JR: Glomerulonephritis. N Engl J Med 339:888-899, 1998
Tubular Lesions Determine Prognosis of IgA Nephropathy Laurent Daniel, MD, Yannick Saingra, MD, Roch Giorgi, MD, Corinne Bouvier, MD, Jean-Franc¸ois Pellissier, MD, and Yvon Berland, MD ● To assess the prognostic value of histological classification for renal outcome, we did a multivariate analysis of 194 patients with immunoglobulin A (IgA) nephropathy between 1985 and 1995. We also evaluated semiquantitative scales of tubular lesions and vessel lesions. At the time of the biopsy, 65 patients (33.5%) had chronic renal failure. At the end of the follow-up, 33 patients (17%) required hemodialysis. The mean age of the patients was 37.8 ⴞ18.9 years with predominance of men (sex-ratio: 3.12). Patients were followed-up for a mean of 43.2 ⴞ 37.2 months. Univariate analysis showed that hypertension (P F 10ⴚ4), nephrotic syndrome (P ⴝ 0.01), and crescents (P ⴝ 0.02) were significant in predicting renal failure, unlike subendothelial topography of IgA deposits (P ⴝ 0.05) and proteinuria (P ⴝ 0.05). Hematuria was a protective factor (P ⴝ 0.03). Multivariate analysis showed that tubular grade 2 (relative risk [RR], 5.5) and tubular grade 3 (RR ⴝ 28.8) were the best factors to predict chronic renal failure. The histological classification of Haas was significant in the univariate analysis, but not in the multivariate analysis. Tubular grading predicted renal outcome better than did the other histological parameters. 娀 2000 by the National Kidney Foundation, Inc. INDEX WORDS: Glomerulonephritis; immunoglobulin A (IgA); prognosis; classification; tubular grading.
P
RIMARY immunoglobulin A (IgA) nephropathy is the most common glomerulopathy in Europe and France.1-3 Although initially reported as benign,1 the clinical course of IgA nephropathy is variable; it leads to chronic renal failure (CRF) or end-stage renal failure (ESRF) in one third of patients.4,5 Evidence has accumulated that more severe proteinuria and hypertension mean progression of the disease. Furthermore, at time of diagnosis, pathological features were predictors of renal outcome in IgA nephropathy.6 Histological grading systems have been used to evaluate the prognosis of IgA nephropathy.7-9 The classification of Haas seems to be more predictive than previous systems and is increasingly used.9 We evaluated the classification of Haas, tubular lesions, vascular lesions, IgA deposit topography, and crescents to determine which parameter best predicts renal outcome. We therefore retrospectively evaluated renal biopsy specimens in 194 patients with primary IgA nephropathy. PATIENTS AND METHODS
Patients We retrospectively retrieved all renal biopsies performed in Marseille, in the south of France, from January 1985 to December 1995. We selected 233 biopsies with the diagnosis of IgA nephropathy. A questionnaire was addressed to the correspondent nephrologists to know the clinical and biological parameters at the time of the biopsy and the outcome of patients. Biopsies were performed because of isolated pro-
teinuria, proteinuria with hematuria, or deterioration of the renal function without evident cause. Biopsies were not performed on patients with isolated hematuria. Proteinuria was considered significant if it was over 500 mg/d. The threshold for renal insufficiency in statistical analyses was 1.7 mg/dL. To be entered into the study, patients must have been free of steroidal treatment. Other treatments were only symptomatic.
Renal Biopsies We divided the biopsy specimen into two parts for direct immunofluorescence (IF) and light microscopy (LM). For LM, tissue was fixed 12 hours in 2.5% glutaraldehyde, and 2-µm-thick sections were stained with Gomori’s trichrome. Sections 0.2 µm thick were stained with Jones silver impregnation. For IF, tissue was fixed 24 hours in Michel liquid10,11 before being snap frozen in ⫺80°C isopentane, and then it was stored at ⫺80°C until use. Cryostat sections 5 µm thick were reacted with fluorosceinated rabbit antibodies specific for IgG, IgA, IgM, C1q, C3, fibrinogen, and immunoglobulin light chains. Antibodies were purchased from Binding System (Birmingham, UK). Sections were examined using a Reichert-Jung microscope equipped with epifluorescence
From the Service d’Anatomie Pathologique, C-H-U Timone, Marseille, France; the Service de Ne´phrologieDialyse, Hoˆpital de Sainte-Marguerite, Marseille, France; and the Service de l’Information Me´dicale, C-H-U Timone, Marseille, France. Received April 15, 1999; accepted in revised form August 3, 1999. Supported by institutional grants from AP-HM. Address reprint requests to Laurent Daniel, MD, Service d’Anatomie Pathologique (Pr Pellissier), C-H-U Timone, 264 rue Saint-Pierre 13385 Marseille cedex 05, France. E-mail: [email protected]
娀 2000 by the National Kidney Foundation, Inc. 0272-6386/00/3501-0003$3.00/0
American Journal of Kidney Diseases, Vol 35, No 1 (January), 2000: pp 13-20
13
14
illumination optics (Leica, Marseilles, France). The presence of subendothelial deposits was assessed by IF, and intensity was graded from 0 to 3⫹. Only intensities from 1⫹ to 3⫹ were considered significant for IgA and non-IgA deposits. Subendothelial deposits were also identified with semithin sections.
Histological Evaluation Classification. We graded renal biopsy specimens according to the classification of Haas.9 We considered only the first biopsy specimen and excluded renal allograft biopsy specimens. We required a minimum of three glomeruli for LM and for IF. Frozen specimen was then paraffinembedded for conventional examination to determine whether the specimen was homogenous. Such specimens were considered an adequate renal biopsy.12 The summarized classification is as follows: Subclass I: All glomeruli are normal or show a slight mesangial hypercellularity with or without detectable deposits. Subclass II: Some glomeruli show typical segmental sclerosis in a pattern resembling primary focal segmental glomerulosclerosis. A slight mesangial hypercellularity is accepted. Subclass III: Fifty percent or fewer of the glomeruli show mesangial or endocapillary substantial proliferation. Proliferation may be global or segmental. Crescents are present or not. Subclass IV: More than 50% or all the glomeruli show mesangial or endocapillary proliferation. Crescents are present or not. Subclass V: Fibrous lesions affect 40% of the glomeruli or 40% of the interstitium with tubular atrophy. Tubular grading. To assess tubular lesions, we examined serial sections of the cortical area except in the subcapsular area. The medulla was not examined. We considered as lesions atrophic tubules with wrinkled basal membrane, and dilated tubules. If tubules were not back-to-back, we established whether there was fibrosis causing this appearance. Atrophy and dilations were often associated in the same specimen. Hyaline or hematic casts were not considered lesions. The severity of lesions was graded by the grading of Katafuchi et al,13 modified as follows: Grade 0: Normal tubules Grade 1: Tubular lesions involving less than 25% of the cortical area Grade 2: Tubular lesions involving less than 50% but more than 25% of the cortical area Grade 3: Diffuse tubular lesions or involving more than 50% of the cortical area Vessel grading. We examined abnormalities of the arteriolar wall on semithin sections. The severity of lesions was graded by the grading of Katafuchi et al,13 modified as follows: Grade 0: Normal vessel with a ratio (R) of luminal diameter to outer diameter less than 0.5. Grade 1: Thin hyaline deposits or intimal sclerosis in the arteriolar wall with R higher than 0.25 and without
DANIEL ET AL
smooth muscle proliferation. Proliferation is more than two cells per cross-section. Grade 2: Thick hyaline deposits in the arteriolar wall with R less than 0.25 without smooth muscle proliferation. Lesions are not diffuse. Grade 3: Diffuse thick hyaline deposits in the arteriolar wall with R less than 0.25. Likewise, a biopsy specimen can show smooth muscle proliferation.
Statistical Analyses We established correlations between clinical data, biological data, and histological parameters by chi-squared test or Yates-modified chi-squared test (if theoretical counts were less than 5). Relative risk (RR) was calculated with 95% confidence interval (CI) for the patients with ESRF and those with CRF. Renal survival data were estimated by the Kaplan-Meier method and compared using the log-rank test. P values less than 0.05 were considered significant in univariate analysis. Significant variables in univariate analysis were used in multivariate analyses. Multivariate analyses were done by logistic regression and Cox proportional hazard regression with forward stepping. Thresholds for entering or removing variables were 0.10 and 0.15, respectively. Adjusted odds ratio (OR) was calculated with 95% CI. Data were analyzed by BMDP-system 1.1 software (Meridian Marketing Group, Sedona, AZ).
RESULTS
Clinical and Biological Parameters Between 1985 and 1995, we studied 4,568 biopsy specimens and found that 233 (5.1%) of them had IgA nephropathy. We excluded 39 patients because of coexisting disease such as Scho¨nlein-Henoch purpura, liver cirrhosis, and lupus erythematosus. Finally, we selected for analysis 194 patients with primary IgA nephropathy. Patients were 147 men and 47 women (sex ratio: 3.12). The mean age was 37.8 ⫾ 18.9 years (range, 4 to 81 years). Twenty-three cases occurred in childhood (range, 4 to 15 years). At the time of the biopsy, 117 patients (60.3%) evidenced proteinuria, which was at nephrotic range in 17 patients (8.8%). Hematuria was diagnosed in 145 patients (74.7%). Hypertension, defined as blood pressure exceeding 150/90 mm Hg, was observed in 63 patients (32.5%). At the time of the biopsy, 65 of the 194 patients (33.5%) had increased serum creatinine (⬎1.7 mg/dL). The mean follow-up was 43.2 ⫾ 37.2 months (range, 1 to 141 months). The appearance of CRF and its progression to ESRF varied; 61 of the 194 patients (31.4%) had an increased serum creatinine at the last consulta-
TUBULAR LESIONS OF IgA NEPHROPATHY
15
Fig 1. (A) Mesangial deposits (arrowheads) in a glomerulus with minimal changes (subclass I) (Jones, original magnification ⴛ1,000). (B) Glomerulus with focal and segmental lesion (arrow) of the floculus (subclass II) (Jones, original magnification ⴛ250). (C) Glomerulus with segmental obstruction of capillaries by proliferated endocapillary cells (arrows) (subclass III) (Jones, original magnification ⴛ400). (D) Glomerulus with segmental crescent (subclass IV) (Masson’s trichrome, original magnification ⴛ400).
16
DANIEL ET AL Table 1. Univariate Analysis: Risk Factors for CRF Variables
CRF⫹ (n; %)
CRF⫺ (n; %)
RR
CI95%
P
Proteinuria Nephrotic syndrome Hematuria Hypertension Crescents Subendothelial deposits Tubular index grade 1 Tubular index grade 2 Tubular index grade 3 Vessel index grade 1 Vessel index grade 2 Vessel index grade 3 Haas subclass II Haas subclass III Haas subclass IV Haas subclass V
43 (70.5) 10 (16.4) 39 (63.9) 42 (68.8) 18 (29.5) 15 (24.5) 7 (11.4) 25 (41) 19 (31.1) 9 (14.7) 21 (37.7) 14 (22.9) 1 (1.6) 7 (11.5) 10 (16.4) 27 (44.2)
74 (55.6) 7 (5.2) 106 (79.1) 21 (15.6) 16 (11.9) 18 (13.4) 26 (19.4) 16 (12) 4 (2.9) 16 (12) 16 (11.9) 4 (2.9) 7 (5.2) 16 (12) 2 (1.5) 13 (9.7)
1.6 2 0.6 4.6 2 1.6 3.1 8.9 12.1 2.9 4.5 6.2 1.1 2.8 7.6 6
1-2.5 1.3-3.2 0.4-0.9 2.9-7.2 1.3-3.1 1-2.5 1-8.5 4-20.1 5.5-26.8 1.4-6 2.5-8.2 3.6-11 0.2-7.8 1.2-6.4 4.9-14.1 3.4-11.2
0.05 0.01 0.02 ⬍10⫺4 0.002 0.05 0.04 ⬍10⫺4 ⬍10⫺4 0.015 ⬍10⫺4 ⬍10⫺4 NS 0.04 ⬍10⫺4 ⬍10⫺4
tion. Ten of 129 patients initially reported with normal renal function developed CRF; six of them required hemodialysis. In contrast, 14 of the 65 patients initially reported with an increased serum creatinine level had under 1.7 mg/dL at the last consultation. Last, 33 patients (17%) developed ESRF. No acute renal failure appeared during the study. Histological Features The five subclasses of the classification of Haas were distributed heterogeneously. The largest subclass was subclass I (101 patients; 55%), corresponding to minimal lesions (Fig 1A). Subclass II was observed in only eight patients (4.3%) (Fig 1B). Subclasses III (Fig 1C) and IV (Fig 1D), corresponding to the active proliferation, were found in 23 patients (12.5%) and 12 patients (6.5%), respectively. Subclass V was observed in 40 patients (21.7%). Thirty-four biopsy specimens (17.5%) had crescentic lesions. Only 13 biopsy specimens (6.7%) showed more than 20% of crescentic glomeruli. Subendothelial deposits were seen in 33 biopsy specimens (17%). One hundred forty biopsy specimens (72.1%) showed C3-associated deposits, whereas IgM, IgG, and C1q were seen in 35 (18%), 21 (10.8%), and 19 (9.8%) biopsy specimens, respectively. Clinicopathologic Correlations Correlations were calculated at the end of the follow-up period (Tables 1 and 2) and for the
group of patients requiring hemodialysis (Tables 3 and 4). At the end of the follow-up, the univariate analysis showed that most of the parameters, except hematuria (RR ⫽ 0.6, CI ⫽ 0.4 to 0.9) and Haas subclass II (RR ⫽ 1.1, CI ⫽ 0.2 to 7.8), were linked with CRF. Proteinuria (RR ⫽ 1.6, CI ⫽ 1 to 2.5; P ⫽ 0.05) and IgA deposit topography (RR ⫽ 1.6, CI ⫽ 1 to 2.5; P ⫽ 0.05) were nearly significant. Results were similar with proteinuria if patients with Haas subclass II were excluded (P ⫽ 0.047). Note that hypertension (OR ⫽ 3.6, CI ⫽ 1.2 to 10.7) was the only significant clinical risk factor in the multivariate analysis. Subclasses IV (P ⬍ 10⫺4) and V (P ⬍ 10⫺4) of the classification of Haas were correlated with CRF only in the univariate analysis. The multivariate analysis showed that tubular grade 2 (OR ⫽ 5.5, CI ⫽ 1.3 to 22.5) and tubular grade 3 (OR ⫽ 28.8, CI ⫽ 4.3 to 190.4) were significant risk factors. Crescents were correlated with CRF (OR ⫽ 11.2, CI ⫽ 2.8 to 59.6). Note also that at the time of the biopsy, crescents were correlated with neither nephrotic syndrome (P ⫽ 0.81) nor tubular grading (P ⫽ 0.9). High Table 2. Multivariate Analysis: Risk Factors for CRF Variables
OR Adjusted
CI95%
P
Tubular index grade 1 Tubular index grade 2 Tubular index grade 3 Crescents Hypertension
1.7 5.5 28.8 11.2 3.6
0.4-7.4 1.3-22.5 4.3-190.4 2.8-59.6 1.2-10.7
NS 0.01 0.001 0.01 0.01
TUBULAR LESIONS OF IgA NEPHROPATHY
17
Table 3. Univariate Analysis: Risk Factors for the Group of Patients With ESRF Variables
ESRF⫹ (n; %)
ESRF⫺ (n; %)
RR
CI95%
P
Proteinuria Nephrotic syndrome Hematuria Hypertension Crescents Subendothelial deposits Tubular index grade 1 Tubular index grade 2 Tubular index grade 3 Vessel index grade 1 Vessel index grade 2 Vessel index grade 3 Haas subclass II Haas subclass III Haas subclass IV Haas subclass V
20 (62.5) 9 (28.1) 18 (56.2) 24 (75) 10 (31.2) 10 (31.2) 2 (6.2) 14 (43.7) 12 (37.5) 5 (15.6) 12 (32.5) 8 (25) 1 (3.1) 4 (12.5) 5 (15.6) 15 (46.8)
97 (60.2) 8 (4.9) 127 (78.8) 39 (24.2) 24 (15.4) 23 (14.8) 31 (20) 27 (17.4) 11 (7.1) 20 (12.9) 25 (16.1) 10 (6.4) 7 (4.5) 19 (12.2) 7 (4.51) 25 (16.1)
1 3.9 0.4 5.5 2.2 2.1 2.6 15 22.9 4.2 6.8 9.3 3.16 4.39 10.52 9.47
0.5-1.9 2.2-7 0.2-0.7 2.7-11.2 1.2-4.3 1.1-4 0.4-18.2 3.6-63 5.5-95.4 1.3-13.4 2.5-18 3.44-25.3 0.4-25 1.2-16.2 3.26-33.92 3.35-26.80
0.95 0.004 0.003 ⬍10⫺4 0.02 0.02 0.3 ⬍10⫺4 ⬍10⫺4 0.02 ⬍10⫺4 ⬍10⫺4 0.3 0.04 0.0005 ⬍10⫺4
vessel grades (grades 2 and 3) (Fig 2A, B) were strongly correlated with hypertension at the time of the biopsy (P ⬍ 10⫺4). These results were similar for the 33 patients requiring hemodialysis (Tables 3 and 4). High tubular grades (grade 2, OR ⫽ 78, CI ⫽ 7.5 to 803.3; grade 3, OR ⫽ 105.3, CI ⫽ 10 to 1,106) were the most evident risk factors. No correlations were found between ESRF and crescents, or hypertension. Hematuria was a protective factor (RR ⫽ 0.4, CI ⫽ 0.1 to 0.9). Interestingly, vessel grading and high subclasses of the classification of Haas correlated with CRF only in the univariate analysis. In the multivariate analysis, nephrotic syndrome correlated with ESRF but not with CRF. Significant variables for renal survival with Cox’s model are reported in Table 5. The selected variables were still the tubular grading (Fig 3A), the nephrotic syndrome (Fig 3B), and hematuria (Fig 3C). Table 4. Multivariate Analysis: Risk Factors for the Group of Patients With ESRF Variables
OR Adjusted
CI95%
P
Nephrotic syndrome Hematuria Tubular index grade 1 Tubular index grade 2 Tubular index grade 3
10.2 0.3 12.7 78 105.3
3.5-29.5 0.1-0.9 0.8-190.2 7.5-803.8 10-1,106
0.06 0.15 NS 0.01 0.01
DISCUSSION
The results show that tubular grading is the highest predictive factor for CRF in IgA nephropathy and that the classification of Haas is not significant in the multivariate analysis. Our data are in agreement with others emphasizing the role of tubular lesions in the progression of IgA nephropathy.14-16 Tubular lesions were strongly correlated with CRF, as were some clinical parameters such as hypertension. In other glomerulopathies, it is often the intensity of accompanying tubular lesions that predicts decline in renal function.17 Tubular lesions reflect the ability of interstitial processes to damage normal glomeruli carrying kidney function. Some authors, nevertheless, excluded tubular lesions from their histological grading.18 The link between tubular lesions and glomerular lesions is probable because glomerular lesions induce tubular lesions through filtered toxic substances or hypertensive hemodynamic changes.18 Inflammatory mechanisms are involved in the pathogenesis of the tubular lesions through increased amounts of urinary cytokines or growth factors such as transforming growth factor , which was implicated in renal fibrosis.19 Our study confirms the male predominance in IgA nephropathy in France. That the mean age of our patients was higher than in other studies is probably due to the policy of our institution, where biopsies are not done for patients with
18
DANIEL ET AL
Fig 2. (A) Thick hyaline deposits (arrow) in an arteriolar wall (grade 3) (Jones, original magnification ⴛ400). (B) Smooth muscle proliferation (arrow) in an interlobular vessel wall (grade 3) (Jones, original magnification ⴛ400).
isolated hematuria. Only 17% of the patients developed ESRF, yet a third of them showed an increased serum creatinine level. Hypertension affected a third of the patients, and this study shows the link between the hypertension and vascular lesions. As previously described,20 our patients with hematuria tended to progress more slowly to CRF. Although this is difficult to explain, it might reflect that these glomeruli are better able Table 5. Multivariate Analysis of Renal Survival With Cox Regression Analysis Variables
RR
Tubular index grade 0 Tubular index grade 0 Tubular index grade 0 Tubular index grade 0 Nephrotic syndrome ⫺ Nephrotic syndrome ⫹ Hematuria ⫺ Hematuria ⫹
1 3.54 11.06 32.26 1 2.75 1 0.44
CI95%
P
0.0001 0.47-26.13 2.46-49.80 6.66-156.10 0.027 1.12-6.78 0.059 0.19-1.03
to cope with the hemodynamic consequences of the IgA deposits. Our data show that nephrotic syndrome was also a marker of progression, but it was observed only in the group of patients with ESRF. In the literature, the patients with focal and segmental sclerosing glomerulopathy-like IgA nephropathy usually have a nephrotic syndrome. The prognostic value of the focal and segmental hyalinosis (Haas subclass II) is debated.21,22 For Packham et al, these features were a marker of progression.22 Like Haas, we believe patients with such lesions have a better prognosis.21 The classification of Lee et al,8 modified by Haas,9 is proposed to predict the prognosis of IgA nephropathy.23 This classification is helpful for a global prognosis evaluation because of its reproducibility. In fact, only 4% of patients with Haas subclass I required hemodialysis, but 16.7% showed CRF at the last consultation. Proliferative lesions (subclasses III and IV) correlated strongly with CRF. The distinction between these two subclasses is justified because CRF at the last consultation was 43.5%
TUBULAR LESIONS OF IgA NEPHROPATHY
19
Fig 3. (A) Renal survival rate in patients with tubular lesions. The survival rate was significantly higher in patients with low-grade tubular lesions (P ⴝ 0.0001). (B) Renal survival rate in patients with or without nephrotic syndrome. Nephrotic syndrome was a significant RR for renal survival (P ⴝ 0.02). (C) Renal survival rate in patients with or without hematuria. Hematuria was a protective factor for renal survival (P ⴝ 0.05).
and 75% for subclass III and for subclass IV, respectively. Surprisingly, Haas subclass V was not significant in multivariate analysis, most likely because tubular grading masked the statistical power of this variable. Crescents were associated with poor prognosis in large series.24,25 This result was nevertheless not consistently found.17,26,27 In our series, 17.5% of the patients had crescentic glomerulonephritis. At the time of the biopsy, crescents were not usually accompanied by progressive scarring of the glomerular tuft. Crescents should account for interstitial lesions because within an adhesion the floculus merges with the interstitium and thus allows diffusion from capillaries toward the interstitium. D’Amico et al20 found that crescents were a negative prognostic factor by univariate analysis. We confirm this fact by a multivariate analysis. Finally, two limitations must be considered if one uses the classification
of Haas: tubulointerstitial analysis does not provide sufficient information, and the crescents are not precisely quantified in the high subclasses. The variability in the methods appreciating vascular lesions in IgA nephropathy raises doubt about the significance of such lesions.28,29 Our data emphasize the link between hypertension and vascular lesions in the progression to CRF. The glomerular sclerosis is nearly certainly enhanced by nonimmunologic factors such as intrarenal hemodynamic changes, but high vessel grades were not significant in our study. The prognostic value of subendothelial deposits in IgA nephropathy is still controversial. As in other studies, we found no prognosis value for this topography.24-26 Conversely, others reported that peripheral topography was a poor prognosis factor in univariate analysis.13,30 Subendothelial deposits indicate an extension of the immunologic injuries to the peripheral capillary walls,
20
DANIEL ET AL
and these deposits are responsible for lesions of the basement membrane, which are a marker of severity of most glomerulopathies.31 To conclude, we underline that tubular grading is essential to predict patient outcome in IgA nephropathy. We show that tubular grading is a stronger predictive factor than are glomerular lesions evaluated by the classification of Haas. We believe tubular grading should be applied in renal biopsy analyses of IgA nephropathy. ACKNOWLEDGMENT The authors thank Annie Fouquet for technical assistance and the physicians who provided follow-up data. We also thank Dr J. Gouvernet for statistical analyses and Prof Figarella-Branger for comments.
REFERENCES 1. Berger J, Hinglais N: Les de´poˆts intercapillaires d’IgAIgG. J Urol Nephrol 74:694-695, 1968 2. Simon P, Rame´e MP, Autuly V, Laruelle E, Charasse C, Cam G, Ang KS: Epidemiology of primary glomerular diseases in a French region: Variations according to period and age. Kidney Int 46:1192-1198, 1994 3. Alamartine E, Sabatier JC, Guerin C, Berliet JM, Berthoux F: Prognostic factors in mesangial IgA glomerulonephritis: An extensive study with univariate and multivariate analyses. Am J Kidney Dis 18:12-19, 1991 4. Ibels LS, Gyory AL: IgA nephropathy. Medicine 73:79102, 1994 5. Galla JH: IgA nephropathy. Kidney Int 47:377-387, 1995 6. Droz D, Kramar A, Nawar T, Noel LH: Primary IgA nephropathy: Prognostic factors. Contrib Nephrol 40:202207, 1984 7. Meadow SR, Glasgow EF, White RHR, Moncrieff MW, Cameron JS, Ogg CS: Scho¨nlein-Henoch nephritis. Q J Med 41:241-258, 1972 8. Lee SMK: Prognostic indicators of progressive renal disease in IgA nephropathy: Emergence of a new histologic grading system. Am J Kidney Dis 29:953-958, 1982 9. Haas M: Histologic subclassification of IgA nephropathy: A clinicopathologic study of 244 cases. Am J Kidney Dis 29:829-842, 1997 10. Michel B, Milner Y, David K: Preservation of tissuefixed immunoglobulins in skin biopsies of patients with lupus erythematosus and bullous diseases: Preliminary report. J Invest Dermatol 59:449-451, 1973 11. Croker BP, Dawson DV, Sanfilippo F: IgA nephropathy: Correlation of clinical and histologic features. Lab Invest 48:19-24, 1983 12. Madaio MP: Renal biopsy forum. Kidney Int 38:529543, 1990 13. Katafuchi R, Kiyoshi Y, Oh Y, Uesugi N, Ikeda K, Yanase T, Fujimi S: Glomerular score as a prognosticator in IgA nephropathy: Its usefulness and limitation. Clin Nephrol 49:1-8, 1998
14. Remuzzi G, Bertani T: Pathophysiology of progressive nephropathies. N Engl J Med 339:1448-1456, 1998 15. Bogenschutz O, Bohle O, Batz G, Whermann M, Pressler H, Kendziorra MV, Gartner HV: IgA nephritis: On the importance of morphological and clinical parameters in the long term prognosis of 239 patients. Am J Nephrol 10:137-147, 1990 16. Bennett WM, Walker RG, Kincaid-Smith P: Renal cortical interstitial volume in mesangial IgA nephropathy. Lab Invest 47:330-335, 1982 17. Abe S, Amagasaki Y, Iyori S, Konishi K, Kato E, Sakaguchi H: Significance of tubulointerstitial lesions in biopsy specimens of glomerulonephritic patients. Am J Nephrol 9:30-37, 1989 18. Koyama A, Igarashi M, Kobayashi M: Natural history and risk factors for immunoglobulin A nephropathy in Japan. Am J Kidney Dis 29:526-532, 1997 19. Border WA, Noble NA: Transforming growth factor- in tissue fibrosis. N Engl J Med 331:1286-1292, 1994 20. D’Amico G, Imbasciati E, Barbiano di Belgioioso G, Bertoli S, Fogazzi G, Ferrario F, Fellin G, Ragni A, Colasanti G, Minetti L, Ponticelli C: Idiopathic IgA mesangial nephropathy: Clinical and histological study of 374 patients. Medicine 64:49-60, 1985 21. Haas M: IgA nephropathy histologically resembling focal-segmental glomerulosclerosis: A clinicopathologic study of 18 cases. Am J Kidney Dis 28:365-371, 1996 22. Packham DK, Yan HD, Hewitson TD, Nicholls KM, Fairley KF, Kincaid-Smith P, Becker GJ: The significance of focal and segmental hyalinosis and sclerosis (FSHS) and nephrotic range proteinuria in IgA nephropathy. Clin Nephrol 46:225-229, 1996 23. Lee HS, Koh HI, Lee HB, Park HC: IgA nephropathy in Korea: A morphological and clinical study. Clin Nephrol 27:131-140, 1987 24. Kobayashi Y, Tateno S, Hiki Y, Shigematsu H: IgA nephropathy: Prognostic significance of proteinuria and histological alterations. Nephron 34:146-153, 1983 25. Radford MG, Donadio JV, Bergstralh EJ, Grande JP: Predicting renal outcome in IgA nephropathy. J Am Soc Nephrol 8:199-207, 1997 26. Kusumoto Y, Takebayashi S, Tagushi T, Harada T, Naito S: Long term prognosis and prognostic indices of IgA nephropathy in juvenile and in adult Japanese. Clin Nephrol 28:118-124, 1987 27. Magil AB, Ballon HS: IgA nephropathy: Evaluation of prognostic factors in patients with moderate disease. Nephron 47:246-252, 1987 28. Feiner HD, Cabili S, Baldwin DS, Schacht RG, Gallo GR: Intrarenal vascular sclerosis in IgA nephropathy. Clin Nephrol 18:183-192, 1982 29. Okada H, Suzuki H, Konishi K, Sakaguchi H, Saruta T: Histological alterations in renal specimens as indicators of prognosis of IgA nephropathy. Clin Nephrol 37:235-238, 1992 30. Mustonen J, Pasternack A, Helin H, Nikkila¨ M: Clinicopathologic correlations in a series of 143 patients with IgA glomerulonephritis. Am J Nephrol 5:150-157, 1985 31. Hricik DE, Chung-Park M, Sedor JR: Glomerulonephritis. N Engl J Med 339:888-899, 1998