Q J Med 2003; 96: 411-420
© 2003 Association of Physicians
Powerful morphometric indicator of prognosis in lupus nephritis
From the 1Department of Pathology, University of Birmingham, and 2Department of Nephrology, University Hospital NHS Trust, Birmingham, UK
Received 31 July 2002 and in revised form 19 March 2003
 |
Summary |
|---|
 Top Summary IntroductionMethodsResultsDiscussionReferences |
|---|
Background: Changes in renal biopsies in lupus nephritis have been analysed in many ways, but few have had prognostic value.
Aim: To see whether a morphometric measure of chronic renal damage that was a prognostic indicator in other conditions had similar value in lupus nephritis.
Design: Retrospective analysis of biopsies and study of outcome.
Methods: On sections of 260 biopsies from 182 consecutive patients with systemic lupus erythematosus, an image analysis system measured chronic damage as a proportion of cortical area, to give the index of chronic damage. This was related to survival (until death or onset of dialysis). Patients were followed for up to 20 years.
Results: The index of chronic damage ranged from 0 to 93%. Twenty-three patients (13%) died before dialysis, many from infection or myocardial infarction, and 40 (22%) went onto permanent dialysis. There were strong correlations between the index and time until death or dialysis (log rank test: 
2 = 51.08, three degrees of freedom [df], p < 0.001) and time to dialysis (log rank test: 2 = 72.88, 3df, p < 0.001), but there was no correlation with time until death before dialysis (log rank test: 2 = 0.36, 3df, p > 0.9). WHO class of nephritis had no major relation to outcome after the index was taken into account and after appropriate treatment of the different classes.
Discussion: The index was a strong indicator of risk of progression to renal failure in lupus nephritis, but not of risk of death before dialysis. This will be useful in clinical management and treatment trials.
|
Introduction |
|---|
|
TopSummaryIntroductionMethodsResultsDiscussionReferences |
|---|
Many methods have been described for assessing changes in renal biopsy specimens in lupus nephritis. Probably the most widely used is the World Health Organization (WHO) scheme for the classification of lupus glomerulonephritis.1 Grading of acute and chronic changes to give activity and chronicity indices was developed at the National Institutes of Health (NIH)2 and incorporated in the WHO scheme.1 The NIH/WHO activity and chronicity indices were found by some workers to predict progression to end stage renal failure,2–4 but others found they were of little value in prognosis,5–7 and their reproducibility was poor.8,9 Dissatisfaction with the indices has led to development of other methods of assessment, some of which are complex or require non-routine staining methods.10,11
A few clinical features have been shown to correlate with prognosis in lupus nephritis. These include serum creatinine concentration at the time of initial investigation of renal disease, development and persistence of the nephrotic syndrome, and racial background of the patient.5,12–14 Few if any pathological changes have been shown to have an equivalent prognostic value to these clinical features.
A simple morphometric technique was devised that allowed measurement of the amount of chronic damage in renal biopsy specimens stained routinely by the periodic acid–methenamine silver (PA-silver) method.15 The amount of damage in a specimen was expressed as a percentage of the cross-sectional area of cortex and called the index of chronic damage. This gave a strong indication of likely survival of renal function after biopsy in various conditions. Lupus was not examined in the previous study,15 but appeared worth investigation since the index of chronic damage was based on similar features to the NIH/WHO chronicity index.1,2 The hypothesis of the current study was that the index of chronic damage would be a prognostic indicator in a series of biopsies in lupus nephritis.
|
Methods |
|---|
|
TopSummaryIntroductionMethodsResultsDiscussionReferences |
|---|
Selection of patients and renal biopsy specimens
Consecutive patients were identified who had systemic lupus erythematosus by American Rheumatological Association criteria,16 had at least one renal biopsy in the 20-year period between April 1981 and March 2001, and had follow-up after the biopsy. One hundred and eighty–two patients fulfilled these criteria: 162 female and 20 male. Age at the time of the first or only biopsy ranged from 13 years to 73 years, mean 34.0 years, median 31 years. There were 260 biopsy specimens: one from each of 126 patients, two from each of 41 patients, three from each of nine patients, four from each of five patients and five from one patient. Specimens were reclassified using the latest WHO scheme,1 in which there are six classes that can be described briefly as: class 1, normal glomeruli; class 2, mesangiopathy; class 3, focal segmental glomerulonephritis; class 4, diffuse glomerulonephritis; class 5, membranous glomerulopathy; and class 6, advanced glomerulosclerosis. Treatment was not the same for each class, and in general was more aggressive for classes 3, 4 and 5. Serum creatinine concentration at biopsy was noted. Time was noted from biopsy to the first event of either death or onset of permanent dialysis, or from biopsy to the last time that the patient was known to be alive and not on dialysis. Cause of death was found if available. Some patients died or had a renal transplant after onset of permanent dialysis, but in these the only event timed for this study was the onset of dialysis.
Measurement of chronic damage
An interactive image analysis system (Aequitas, Dynamic Data Links) was used to measure the amount of chronic damage in biopsy specimens.15 A section routinely stained with PA-silver was examined on a microscope linked to a computer. On each image, cortex was outlined with the drawing facility, and the area measured. Then parts were outlined with chronic damage, defined as globally sclerosed glomeruli, atrophic tubules, interstitial fibrosis, occluded vessels and cysts (Figure 1). All the cortex was measured. For each specimen, the proportion of cortex with chronic damage was expressed as a percentage to the nearest integer. To test interobserver and intraobserver variation, 23 specimens were measured independently by two observers, one who had helped to develop the method (AJH) and one who would measure the series (NT), and 35 specimens were measured twice at intervals by one observer (NT). Agreement was assessed by the method of Bland and Altman after log transformation17 (necessary because the differences between measurements were proportional to the mean measurement). This method gave the bias, or mean difference between measurements, and limits of agreement, or 2 SDs either side of the mean, with 95%CIs for the bias and limits of agreement, all expressed as ratios when back-transformed. On each specimen from the patients with lupus, a slide was retrieved that had been routinely stained by PA-silver at the time of biopsy. These were examined without knowledge of the patient they were from, or the clinical course of the patient. The index of chronic damage was determined.
|
Statistical analysis
The index of chronic damage was initially divided into five groups for analysis: 0–9%, 10–19%, 20–39%, 40–59% and 60% and above, as in the previous paper,15 but preliminary study showed no differences between the groups of 0–9% and 10–19%, and these were combined. Survival to the end point, either death or dialysis, related to the index of chronic damage in the first or only renal biopsy, was examined by the Kaplan–Meier product limit method, and equality of survival between groups was tested by the log rank test to give the
2 statistic. Pairwise comparisons were not formally tested, but were assessed descriptively from Kaplan–Meier survival curves, to avoid problems of small sample sizes and risks of spuriously significant results from increased numbers of tests. Survival related to the index of chronic damage was also calculated by the life-table method. This gave slightly different values from the product limit calculations used in the Kaplan-Meier method. Survival was also examined by the Kaplan-Meier method related to other variables including the WHO class, sex, age divided into five groups, 13–19 years (n = 16), 20–29 years (n = 58), 30–39 years (n = 53), 40–49 years (n = 35), and 50–73 years (n = 20), and the serum creatinine concentration at biopsy, divided into groups of: <120 µmol/l (that is, within the normal range), 120–249 µmol/l, 250–499 µmol/l and 500 µmol/l and above. The Spearman rank correlation coefficient was used to study the correlation between the index of chronic damage in a biopsy specimen and the serum creatinine concentration at biopsy. The independent contribution of the index of chronic damage to the risk of death or dialysis after controlling for initial serum creatinine concentration in steps of 100 µmol/l was estimated by the Cox proportional hazards regression analysis model with censored data (Stata Statistical Software, Release 5.0). In this the index was categorized so that a change of one unit represented a change of 10% in the index. The Cox model was also used to investigate the independent effects of other variables on survival, namely WHO class, sex and age. The 2 test was used to study the distribution of groups of the index in the WHO classes. For all statistical tests, the conventional value of p < 0.05 was taken to be statistically significant. |
Results |
|---|
|
TopSummaryIntroductionMethodsResultsDiscussionReferences |
|---|
Agreement in measurement of the index of chronic damage
The mean agreement between two observers in measurement of the index of chronic damage, expressed as a ratio, was 1.03 (95%CI 0.99–1.07) and the limits of agreement were 0.88 (95%CI 0.82–0.93) and 1.22 (95%CI 1.15–1.30). The agreement between two measurements by the same observer was 1.01 (95%CI 0.97–1.06) and limits of agreement were 0.79 (95%CI 0.74–0.85) and 1.29 (95%CI 1.20–1.39). This meant that there was no significant bias in interobserver or intraobserver measurements and that virtually all measurements were well within the limits of one quarter below and above the initial measurement.
Correlations with survival: index of chronic damage
Twenty-three patients (13%) died before dialysis, 40 (22%) went onto permanent dialysis and 119 (65%) were alive without dialysis at last follow-up. There were autopsies on 11 of the 23 patients who died before dialysis. The cause of death was myocardial infarction in seven, infection in six, acute pancreatitis, hypertensive encephalopathy and carcinomatosis in one each, and unknown in seven. Death was at a median of 1 year 5 months after the first or only biopsy (range 3 days to 16 years). Permanent dialysis began at a median of 2 years 7 months after the first or only biopsy (range 2 days to 12 years 9 months). Follow-up on the 119 patients alive without dialysis was for a median period of 6 years 6 months (range 19 days to 20 years). Only five patients had <1 year of follow-up.
The index of chronic damage ranged from 0 to 93%. In the four groups of the index for the purposes of analysis, <20%, 20–39%, 40–59% and 
60%, the numbers of first or only biopsies were 122, 34, 16 and 10, respectively. There was a strong correlation between the index and survival until death or dialysis combined from the time of the first or only biopsy (log rank test: 2 = 51.08, three degrees of freedom [df], p < 0.001; Figure 2). There was a stronger correlation between the index and progression to dialysis alone when patients who died before dialysis were censored (log rank test: 2 = 72.88, 3df, p < 0.001; Figure 3). There was no correlation between the index and progression to death before dialysis when patients who went onto dialysis were censored (log rank test: 2 = 0.36, 3df, p > 0.9).
|
|
For the four groups of the index, respective cumulative survivals without death or dialysis combined, calculated by the life-table method, were: 81%, 63%, 53% and 20% at 5 years; 76%, 48%, 24% and 10% at 10 years; and 70%, 19%, 0 and 0 at 15 years. Respective cumulative survivals without dialysis alone, after patients who died were censored, were: 91%, 85%, 59% and 26% at 5 years; 89%, 64%, 27% and 13% at 10 years; and 85%, 25%, 0 and 0 at 15 years.
Correlations with survival: WHO class of nephritis
Survival was studied in relation to WHO class of nephritis in the first or only biopsy. Of 182 biopsies, 11 (6%) were WHO class 1, 28 (15%) class 2, 29 (16%) class 3, 73 (40%) class 4, 33 (18%) class 5 and eight (4%) class 6. There were significant relations between class and survival to death or dialysis combined (log rank test: 2 = 17.53, 5df, p < 0.01) and to dialysis alone (log rank test: 2 = 27.17, 5df, p < 0.001; Figure 4), but not between class and progression to death before dialysis (log rank test: 2 = 0.01, 5df, p > 0.9).
|
There was a significant relation between WHO class and the index of chronic damage (
2 = 66.96, 15df, p < 0.001; Table 1). Class 2 had the best prognosis (Figure 4) and also had no specimen with an index over 35%. Class 6 had the worst prognosis and had no specimen with an index under 30%. Class 2 had a significantly reduced risk of progression to dialysis alone and to dialysis or death combined, but not to death before dialysis, compared with all other classes combined (respectively, by log rank test: 2 = 7.02, 1df, p < 0.01; 2 = 3.90, 1df, p < 0.05; and 2 = 0.04, 1df, p > 0.8).
|
When only specimens with an index of chronic damage <20% were analysed (n = 122), there was no significant relation between class and survival to dialysis alone, death or dialysis combined, or death before dialysis (respectively, by log rank test:
2 = 8.55, 4df, 0.1 > p > 0.05; 2 = 7.06, 4df, p > 0.1; and 2 = 0.96, 4df, p > 0.5). This was true even when class 2 was compared with all other classes combined (respectively, by log rank test: 2 = 2.49, 1df, p > 0.1; 2 = 3.83, 1df, 0.1 > p > 0.05; and 2 = 1.49, 1df, p > 0.1). When only specimens with an index under 40% were studied (n = 156), there was no significant relation between class and survival to dialysis alone, death or dialysis combined, or death before dialysis when classes were kept separate (respectively, by log rank test: 2 = 4.73, 5df, p > 0.25; 2 = 2.31, 5df, p > 0.75; and 2 = 0, 5df, p > 0.9), but when class 2 was compared with all other classes combined, this class had a significantly reduced risk of progression to dialysis alone (log rank test: 2 = 4.11, 1df, 0.05 > p > 0.025) but not to death or dialysis combined (log rank test: 2 = 1.69, 1df, p > 0.1) nor to death before dialysis (log rank test: 2 = 0.03, 1df, p > 0.8).
Classes were combined to allow statistical analysis of specimens with an index of 40% and above (n = 26). There was no significant difference in survival between class 6 (n = 7), the class with the worst overall prognosis, and the other classes combined (n = 19), in survival to dialysis alone, death or dialysis combined, or death before dialysis (respectively, by log rank test: 2 = 0.2, 1df, p > 0.5; 2 = 0.04, 1df, p > 0.75; and 2 = 0.56, 1df, p > 0.25).
Only WHO classes 4 and 5, the two with the most specimens, had enough numbers in different groups of the index of chronic damage to allow analysis of the relation between the index and survival within a single class. In both classes there were significant correlations between the index and survival to death or dialysis combined (class 4, log rank test: 2 = 21.02, 3df, p < 0.001; class 5, log rank test: 2 = 13.78, 3df, p < 0.01) and to dialysis alone (class 4, log rank test: 2 = 27.82, 3df, p < 0.001, Figure 5; class 5, log rank test: 2 = 16.95, 3df, p < 0.001), but not to death before dialysis (class 4, log rank test: 2 = 0.13, 3df, p > 0.9; class 5, log rank test: 2 = 1.47, 3df, p > 0.5).
|
Correlations with survival: serum creatinine concentration and other factors
There were significant correlations between the index of chronic damage and the serum creatinine concentration both at the time of the first or only biopsy (n = 139, r = 0.63, p < 0.001) and at the time of all biopsies (n = 188, r = 0.64, p < 0.001; Figure 6). Serum creatinine concentration at the time of the first or only biopsy had correlations with survival until death or dialysis combined (log rank test:
2 = 66.83, 3df, p < 0.001; Figure 7), dialysis alone (log rank test: 2 = 58.68, 3df, p < 0.001) and death before dialysis (log rank test: 2 = 13.43, 3df, p < 0.01).
|
|
With the Cox proportional hazards model, after controlling for initial serum creatinine concentration, each 10% increase in the index of chronic damage in the first or only biopsy increased the risk of progression to dialysis alone with a hazard ratio of 1.31 (95%CI 1.10–1.56, p = 0.003) and to death or dialysis combined with a hazard ratio of 1.17 (95%CI 1.01–1.35, p = 0.04). There was no significantly increased risk of death before dialysis with each 10% increase in the index of chronic damage (hazard ratio 0.95, 95%CI 0.71–1.26, p > 0.7).
Inclusion of WHO class in the Cox model showed no significant contribution of class to the risk of progression to one end point or to either end point after controlling for initial serum creatinine concentration and index of chronic damage (for each class, p > 0.9 for dialysis alone, p > 0.5 for death alone and p > 0.15 for either death or dialysis).
There were no significant correlations between sex or age of the patient and survival until dialysis alone, death or dialysis combined or death before dialysis (for sex, respectively, by log rank test: 2 = 0.05, 1df, p > 0.75; 2 = 1.53, 1df, p > 0.1; and 2 = 3.02, 1df, 0.1 > p > 0.05; for age, respectively, by log rank test: 2 = 5.6, 4df, p > 0.1; 2 = 7.73, 4df, p > 0.1; and 2 = 2.19, 4df, p > 0.5). Similarly, sex and age showed no significant contribution in the Cox model to the risk of progression after controlling for initial serum creatinine concentration and index of chronic damage (for sex, p > 0.3, and for age, p > 0.1, for one or either end point).
Index of chronic damage in repeat biopsy specimens
There were 78 repeat specimens in 56 patients. When the index of chronic damage was considered in the groups used for survival analysis, only in seven biopsies in six patients did a later specimen show a reduction in the index from the first specimen. In 27 patients, the index stayed in the same group in repeat specimens and in 23 the index was in a higher group.
|
Discussion |
|---|
|
TopSummaryIntroductionMethodsResultsDiscussionReferences |
|---|
This study of 260 renal biopsies in 182 patients with systemic lupus erythematosus is one of the largest in lupus nephritis and has one of the longest lengths of follow-up (median of 6 years 6 months and up to 20 years after the first or only renal biopsy). This is comparable with the Mayo series (n = 439).18
The index of chronic damage was shown in the current study to be a powerful indicator of the outcome in patients with lupus nephritis, most of whom were young women. These had a high risk of development of renal failure. Of 182 patients, 40 (22%) went onto permanent dialysis, and the index correlated with risk of progression to end-stage renal failure. One implication of this study is that a major determinant of long-term renal function after biopsy in lupus is what has happened to the kidney before biopsy, and specifically how much chronic damage there has been. Patients with lupus nephritis also had a high risk of death within a few years of diagnosis, and 23 (13%) died before dialysis, at a median of 1 year 5 months after the first or only biopsy. In contrast to its predictive value for risk of end stage renal failure, the index of chronic damage was not correlated with risk of early death. This can be explained by the finding that death before dialysis was due to causes apparently unrelated to changes in the kidney, mainly infection and myocardial infarction.
The WHO class of lupus nephritis had a relation to outcome (Figure 4). This relation is not simple to analyse, mainly because treatment was not the same for each class and so the prognosis solely related to a particular class of lupus nephritis could not be separated from the prognosis related to treatment of that class. While this unknown effect of treatment should be kept in mind, there was a relation between class and the index of chronic damage (Table 1) and the relation of class to outcome appeared largely indirect after the index was taken into account. There are several pieces of evidence for this. Class had no independent influence on survival in the Cox proportional hazards model. Within WHO classes 4 and 5, there was a relation between the index and survival (Figure 5). Class 2 had the best prognosis but had relatively little chronic damage. This class had no significantly different outcome from other classes when specimens with an index under 20% were considered, although there was a significantly lower risk of progression to dialysis when specimens with an index under 40% were considered. This was the only statistically significant evidence that class may influence prognosis independently of the amount of chronic damage. Class 6 had the worst prognosis but had relatively severe chronic damage and had no significantly different outcome from other classes when specimens with an index of 40% and over were considered.
Several other studies also showed a weak or no effect of class.5,13,18 One problem is that the WHO class only reflects glomerular changes, although changes elsewhere in the kidney have more relation to renal function.19 Class is important as a guide to treatment, and in general classes 3, 4 and 5 are considered the most severe20 and were treated in a standard way during this study and more aggressively than classes 2 and 6. The findings in this paper cannot be taken to mean that without treatment there would be no difference in outcome between classes with a similar index of chronic damage. The surprising finding is that survival in class 1 was comparable to that in classes 3, 4 and 5, and this was because most patients with this class had the small-vessel vasculopathy of the type seen in systemic sclerosis, although glomeruli were normal.
The NIH/WHO activity and chronicity indices were not estimated in the present study because of evidence that they were not reproducible.8,9 Some have found the NIH/WHO activity index useful in prognosis.2–4,12 Others have not.5–7 Similarly, some have found the NIH/WHO chronicity index useful in prognosis2–4,21 but many have not.5–7,12,13,22 Modifications to the way that pathological changes in lupus nephritis are assessed subjectively are complex,11,23 and for this reason are unlikely to be used widely in determination of outcome.
Unlike many of the established methods of assessment of lupus nephritis, the index of chronic damage described by the current authors and used in the present study has prognostic value.15 This method is based on similar features to the NIH/WHO chronicity index, is simple and reproducible, and has the advantages of morphometric techniques over grading systems.24 It is done on routinely stained sections. The method should be applicable on most modern image analysis systems, found in many departments of pathology. Although an observer is required to interact with images, the process is fast: analysis of an image takes 2–3 min, and of most biopsies < 10 min. The index can range from 0 to 100%, which no automated method of measurement of chronic damage yet described can match.15
There was a correlation with serum creatinine concentration at the time of biopsy, but this was weak (Figure 6). Serum creatinine concentration is an indirect measure of renal function, and increases in both acute and chronic renal impairment. Acute impairment is associated with acute abnormalities in tubules rather than chronic changes, while the index is a measure only of chronic damage and indirectly of chronic renal impairment.15 As others had shown, serum creatinine concentration was an indicator of prognosis,12,13 although there was only a crude relation between serum creatinine concentration and outcome (Figure 7). Even when serum creatinine concentration at biopsy was taken into consideration, the index of chronic damage had a strong predictive value of risk of progression to end-stage renal failure in the Cox model.
Repeat biopsy specimens almost always showed either no change or a worsening of the amount of chronic damage. The few that showed an improvement may be explained by sampling of a process that is patchy in a kidney. Whether a renal biopsy specimen is representative of the whole kidney is always a problem, but since the sample is likely to be the only one available at any particular time and since the index of chronic damage is a statistically-proven indicator of prognosis in lupus (Figures 2, 3, 5) and other conditions,15 there is justification for the measurement.
The index of chronic damage is likely to be of use in individual patients in assessment of prognosis and decisions about clinical management. The index is also likely to be of use in clinical trials in lupus nephritis, allowing patients in different treatment groups to be matched by the extent of chronic renal damage.
|
Acknowledgments |
|---|
We are grateful to Ms L.J. Billingham from the Birmingham Cancer Research Campaign Trials Unit for statistical advice, to Mrs G.M. Richards for collection of data and to our colleagues in the Department of Nephrology: Dr J. Michael, Dr N.T. Richards, Professor C.O.S. Savage, Dr G.W. Lipkin and Dr P. Cockwell.
|
Footnotes |
|---|
Address correspondence to Dr A.J. Howie, Department of Pathology, The Medical School, Birmingham B15 2TT. e-mail: a.j.howie{at}bham.ac.uk
|
References |
|---|
|
TopSummaryIntroductionMethodsResultsDiscussionReferences |
|---|
1. Churg J, Bernstein J, Glassock RJ. Renal Disease: Classification and Atlas of Glomerular Disorders, 2nd edn. New York, Igaku-Shoin, 1995:152.
2. Austin HL, Muenz LR, Joyce KM, et al. Prognostic factors in lupus nephritis: contribution of renal histologic data. Am J Med 1983; 75:382–91.[CrossRef][ISI][Medline]
3. Austin HL, Muenz LR, Joyce KM, et al. Diffuse proliferative lupus nephritis: identification of specific pathologic features affecting renal outcome. Kidney Int 1984; 25:689–95.[ISI][Medline]
4. Austin HA, Boumpas DT, Vaughan EM, et al. Predicting renal outcomes in severe lupus nephritis: contributions of clinical and histologic data. Kidney Int 1994; 45:544–50.[ISI][Medline]
5. Appel GB, Cohen DJ, Pirani CL, et al. Long-term follow-up of patients with lupus nephritis: a study based on the classification of the World Health Organization. Am J Med 1987; 83:877–85.[CrossRef][ISI][Medline]
6. Schwartz MM, Bernstein J, Hill GS, et al. Predictive value of renal pathology in diffuse proliferative lupus glomerulonephritis. Kidney Int 1989; 36:891–6.[ISI][Medline]
7. Schwartz MM, Lan S, Bernstein J, et al. Role of pathology indices in the management of severe lupus glomerulonephritis. Kidney Int 1992; 42:743–8.[ISI][Medline]
8. Schwartz MM, Lan S, Bernstein J, et al. Irreproducibility of the activity and chronicity indices limits their utility in the management of lupus nephritis. Am J Kid Dis 1993; 21:374–7.[ISI][Medline]
9. Wernick RM, Smith DL, Houghton DC, et al. Reliability of histologic scoring for lupus nephritis: a community-based evaluation. Ann Intern Med 1993; 119:805–11.
10. Hill GS, Delahousse M, Nochy D, et al. A new morphologic index for the evaluation of renal biopsies in lupus nephritis. Kidney Int 2000; 58:1160–73.[CrossRef][ISI][Medline]
11. Daniel L, Sichez H, Giorgi R, et al. Tubular lesions and tubular cell adhesion molecules for the prognosis of lupus nephritis. Kidney Int 2001; 60:2215–21.[CrossRef][ISI][Medline]
12. Magil AB, Puterman ML, Ballon HS, et al. Prognostic factors in diffuse proliferative lupus glomerulonephritis. Kidney Int 1988; 34:511–17.[ISI][Medline]
13. Levey AS, Lan S, Corwin HL, et al. Progression and remission of renal disease in the lupus nephritis collaborative study: results of treatment with prednisone and short-term oral cyclophosphamide. Ann Int Med 1992; 116:114–23.[CrossRef][ISI][Medline]
14. Dooley MA, Hogan S, Jennette C, et al. Cyclophosphamide therapy for lupus nephritis: poor renal survival in black Americans. Kidney Int 1997; 51:1188–95.[ISI][Medline]
15. Howie AJ, Ferreira MAS, Adu D. Prognostic value of simple measurement of chronic damage in renal biopsy specimens. Nephrol Dial Transplant 2001; 16:1163–9.
16. Tan EM, Cohen AS, Fries JF, et al. The1982 revised criteria for the classification of systemic lupus erythematosus. Arthritis Rheum 1982; 25:1271–7.[ISI][Medline]
17. Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986; 1:307–10.[CrossRef][ISI][Medline]
18. Donadio JV, Hart GM, Bergstralh EJ, et al. Prognostic determinants in lupus nephritis: a long term clinicopathologic study. Lupus 1995; 4:109–15.
19. Risdon RA, Sloper JC, de Wardener HE. Relationship between renal function and histological changes found in renal biopsy specimens from patients with persistent glomerular nephritis. Lancet 1968; 2:363–6.[ISI][Medline]
20. Najafi CC, Korbet SM, Lewis EJ, et al. Significance of histologic patterns of glomerular injury upon long-term prognosis in severe lupus glomerulonephritis. Kidney Int 2001; 59:2156–63.[ISI][Medline]
21. Nossent HC, Henzen-Logmans SC, Vroom TM, et al. Contribution of renal biopsy data in predicting outcome in lupus nephritis: analysis of 116 patients. Arthritis Rheum 1990; 33:970–7.[ISI][Medline]
22. Valeri A, Radhakrishnan J, Estes D, et al. Intravenous pulse cyclophosphamide treatment of severe lupus nephritis: a prospective five-year study. Clin Nephrol 1994; 42:71–8.[ISI][Medline]
23. Hill GS, Delahousse M, Nochy 
, et al. Predictive power of the second renal biopsy in lupus nephritis: significance of macrophages. Kidney Int 2001; 59:304–16.[ISI][Medline]
24. Furness
. The use of digital images in pathology. J Pathol 1997; 183:253–63.[CrossRef][ISI][Medline]
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  Y Ioannou, S J Bowman, A Rahman, and I P Giles Review of presentations at the 6th European Lupus Meeting 3-5 March 2005 Lupus, June 1, 2005; 14(6): 467 - 478. [Abstract] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||