RHEUMATOLOGY

Rheumatology 2011;50:14241430 doi:10.1093/rheumatology/ker101 Advance Access publication 16 March 2011

Original article
Assessment of a lupus nephritis cohort over a 30-year period
Sara C. Croca1, Teresa Rodrigues2 and David A. Isenberg3
Abstract
Objective. LN is a common and potentially severe complication of SLE. Our aim was to characterize a LN cohort followed at a single centre for 30 years and assess its evolution over time. Methods. Between 1975 and 2005, 156 LN patients were followed up at University College Hospital London. We divided them into three groups depending on the date of recognition of renal involvement (197585, 198695 and 19962005) and compared them in terms of clinical, demographic and serological characteristics and disease outcome. Results. Clinical characteristics were comparable between groups; however, the proportion of Afrocaribbean (AC) patients and the prevalence of anti-ENA antibodies rose significantly over time. The 5-year mortality rate decreased by 60% between the first and second decades, remaining stable over the third decade. The 5-year end-stage renal disease (ESRD) rate remained constant. An increasing number of renal transplants (RTx) was performed with encouraging results. AC origin was associated with a poorer overall prognosis. Conclusion. LN, a common complication of SLE, is associated with increased mortality and morbidity, particularly among AC patients. Despite encouraging results for RTx, once ESRD is established the prognosis is relatively poor and no improvement in preventing its development was achieved. Moreover, although a significant decrease in mortality was observed, it has remained stable for 10 years. These results suggest that we have maximized the benefits of conventional therapies and if further improvement is to be achieved, novel drug regimens must be developed.

Downloaded from http://rheumatology.oxfordjournals.org/ by guest on March 1, 2013

CLINICAL SCIENCE

Key words: Lupus nephritis, 30-year assessment, Mortality, End-stage renal disease.

Introduction
SLE is associated with a high risk of developing renal disease (RD), with proteinuria and hypertension as its most prominent features. Clinically evident RD occurs in up to roughly half of the patients [1]. Clinical findings alone can, however, underestimate the true incidence of renal involvement as some patients have significant pathological abnormalities without relevant clinical signs of renal involvement (silent LN) [2].

Internal Medicine Department I, Hospital de Santa Maria, Biomathematics Department, Faculdade de Medicina de Lisboa, Lisbon, Portugal and 3Department of Medicine, Centre for Rheumatology, University College London, London, UK.
2

Submitted 12 August 2010; revised version accepted 31 January 2011. Correspondence to: Sara C. Croca, Servico de Medicina 1, Hospital de Santa Maria, Avenida Egas Moniz, 1649-035 Lisbon, Portugal. E-mail: saracrocahsm@gmail.com

There are various types of SLE-associated renal involvement although IC-mediated glomerular diseases are the most common [3]. Most renal abnormalities emerge within 35 years after SLE diagnosis [4]. Although eventually up to one-third of SLE patients will develop elevated serum creatinine, evidence of significantly impaired renal function is an uncommon early manifestation. There are many variations in the prevalence and disease course of SLE-associated RD and many factors, both clinical and demographic, have been shown to influence the outcome [5, 6]. During the past 30 years, an improved understanding of the disease mechanisms and optimizing various treatment approaches have shown that the development of renal failure can be reduced to <5% of lupus patients overall [5]. However, notably related to ethnicity and or socioeconomic disadvantage, 2030% of patients with renal lupus may develop renal failure [59]. Overall, the prognosis for LN patients is improving [10, 11].

! The Author 2011. Published by Oxford University Press on behalf of the British Society for Rheumatology. All rights reserved. For Permissions, please email: journals.permissions@oup.com

Assessment of LN cohort

This study is an assessment of the outcome of RD in a cohort of lupus patients over a 30-year period. We have reviewed our LN cohort followed, in detail, at a single UK centre between 1975 and 2005 (n = 156). The cohort was divided into those diagnosed with LN during each of three decades and the groups compared with respect to demographic, clinical, serological and histological features as well as to outcome focusing on the evolution to end-stage RD (ESRD) and mortality rate.

to contingency tables larger than k 2 when one variable was in an ordinal scale. Where appropriate the MannWhitney U test or/and median test was used to compare data. Odds ratio homogeneity was analysed with extended MantelHaenszel chi-square test for linear trend.

Results
General characterization
Patients presenting during the three decades were compared in terms of sex, ethnicity, age at SLE and RD diagnosis, mean time from SLE to RD and WHO class (Table 1). There were no significant differences between the three decades with respect to sex distribution, age at SLE diagnosis, age at RD, WHO class distribution and renal function values (i.e. serum creatinine, serum urea and riadata not shown). proteinu Looking at the time between the SLE diagnosis and RD, there is a decrease in the number of patients diagnosed with RD during the first year when comparing the first decade with the subsequent. However, the vast majority of patients develop RD within 5 years of SLE diagnosis. When assessing the ethnic composition of the population over time, some patterns emerge (Fig. 1). Throughout the decades, there was a predominance of Caucasians (C). However, from the second decade onwards, there is a trend towards an increase in the proportion of AfroCaribbean (AC) patients and a fall in the percentage of C. In contrast, the proportion of Asian (A) patients (including those from India, Pakistan and Bangladesh) changed very little. When comparing the ratio AC : nonAC, there is a non-significant trend towards an increase in AC patients over the three decades (P = 0.087) that becomes more apparent upon dichotomization of the time period considered (before and after 1990) (P = 0.03) (Fig. 1). Apart from a clear association between AC patients and WHO Class V (P = 0.03), no further statistically significant associations were found.

Patients and methods

Between January 1975 and December 2005, 156 patients were diagnosed with LN at the Middlesex Hospital/ University College Hospital London, as part of the cohort of 400 patients diagnosed with SLE during this period. For this study, only those patients with a minimum 5-year follow-up were selected (n = 154). The patients were considered to have LN if they had a biopsy consistent with the World Health Organization (WHO) classification criteria (n = 133) or, when a biopsy had not been performed (n = 21), if there was unequivocal clinical, serological and urinary protein evidence of renal involvement (for further details on inclusion criteria, see [5]). Of this cohort, 30 patients developed ESRD, defined as requiring dialysis and/or renal transplant (RTx). All the patients were characterized in terms of sex, ethnicity, age at the diagnosis of SLE and when RD became evident, WHO class, serum C3 and autoantibody status, notably anti-dsDNA, anti-ENA (ENAab) and aPL. The C3 levels were measured by laser nephelometry; anti-dsDNA antibodies detected by Crithidia and ELISA (Sheila Diagnostics, Dundee, UK) and the anti-ENA antibodies (anti-ENAabs) by ELISA (as described elsewhere [12]) in all the patients including those diagnosed earlier in the cohort (when necessary, stored frozen sera were retested). The ESRD patients were further analysed with regard to the time from RD to ESRD; the type and number of years in dialysis and, in some cases, the type and outcome of RTx. We noted the time of development of ESRD and mortality rate at 5 years after RD. To assess the evolution of these variants over time, we divided the population into three groups corresponding to those diagnosed with LN during three decades: 197585, 198695 and 19962005.

Downloaded from http://rheumatology.oxfordjournals.org/ by guest on March 1, 2013

Serological differences
Virtually all the patients had a positive ANA (n = 154) and raised anti-dsDNA antibody levels (n = 152); defined as Crithidia positive or ELISA assay levels more than twice the upper limit of normal on at least two occasions, with no statically significant differences throughout. There was no significant difference between the three decades with respect to the presence/absence of APS or aPL. In contrast, there was a definite trend towards an increasing prevalence of anti-ENAabs overall over time (P < 0.001) as well as of the individual ENAab subtypes with the exception of anti-La antibodies (Table 2). A strong association between ENAab prevalence and ethnicity was identified, with patients of AC origin being the most likely to be ENAab positive (P < 0.001). A nonsignificant slightly lower prevalence of ENAab among A patients was noted (Table 3). Based on analysis of blood samples taken at the time of diagnosis, antibodies to Ro,

Statistical analysis
All calculations were performed using statistical analysis supported by SPSS 18.0 (SPSS Inc., Chicago, IL, USA). When assessing the statistical significance of the different associations made, Fishers exact test or FreemanHalton extension to Fishers exact test was used. On occasion, adjusted residual analysis completed the study when it was considered convenient. Actuarial (life table) analysis was used to compare the three decades in terms of the 5-year mortality and ESRD rates. The significance threshold for P-value was established at 5%. However, due to the relatively small size of the population, when the result was considered for clinical significance, borderline values were taken into consideration (5% < Pvalue < 10%). Chi-square test for linear trend was applied

www.rheumatology.oxfordjournals.org

1425

Sara C. Croca et al.

TABLE 1 Cohort characterization divided by decades of RD diagnosis

197585 (n = 29) Sex Male Female Ethnicity A AC C Other ethnicity Age at SLE diagnosis, mean (S.D.), (min; max) Age at RD diagnosis, mean (S.D.), (min; max) Time from SLE to RD diagnosis, years 41 1 to 45 5 to 410 >10 WHO class I II III IV V VI Mortality No. of deaths 5 years after RD diagnosis, % Time from RD to death, mean (S.D.) (min; max) Age at death, mean (S.D.), (min; max) ESRD Patients in ESRD 5 years after RD, n (%) Age at ESRD, mean (S.D.), (min; max) Time from RD diagnosis to ESRD, years 41 1 to 45 5 to 410 >10 No. of years in dialysis 41 1 to 45 5 to 410 >10 Type of dialysis CAPD HD CAPD and HD No. of RTx Related donor Cadaver donor NA Post-transplant immunosuppressant regimen CSA Tacrolimus NA: not available. 198695 (n = 41) 19962005 (n = 86)

1 29 5 4 18 2 28.3 (10.1) (11; 46) 30.7 (10.1) (11; 66) 18 8 2 1 1 1 3 11 1 0 5 (17.2) 9.9 (8.2) (<1; 15) 44.0 (15.6) (24; 81) 2 (6.9) 44.0 (15.6) (24; 81) 1 1 2 6 3 2 3 2 0 4 2 0

5 36 4 9 20 8 24.6 (7.9) (10; 45) 27.5 (8.1) (14; 45) 19 8 12 2 1 2 5 18 9 1 3 (7.7) 6.7 (3.93) (<1; 11) 35.5 (11.5) (19; 49) 3 (7.7) 30.5 (10.8) (17; 47) 1 1 2 6 0 1 2 1 4 1 3 4 0 3 1 3 1

8 78 14 27 35 10 26.5 (13.6) (11; 75) 30.3 (13.3) (11; 75) 44 26 9 7

Downloaded from http://rheumatology.oxfordjournals.org/ by guest on March 1, 2013

0 3 15 51 16 1 4 (4.7) 5.8 (2.4) (<1; 9) 40.0 (17.5) (19; 68) 7 (8.1) 37.2 (14.3) (17; 68) 4 3 9 0 4 10 2 0 7 7 2 12 6 4 2 3 9

La, Sm and RNP were all more common in the AC patients (only with anti-Sm were the differences not statistically significant) (Table 3). However, when dividing the patients into two categories, AC and non-AC, anti-Sm antibodies were significantly associated with AC patients (P = 0.02).

Outcome analysis
Mortality There was no significant difference between the decades with respect to the mean (S.D) age at death, which was 40.3 (15.7) years (min 19, max 81), and the mean (S.D.) time

1426

www.rheumatology.oxfordjournals.org

Assessment of LN cohort

FIG. 1 Ethnic distribution over time: overall evolution (A) and comparison between AC and non-AC patients (B). Others: other ethnicity.

A 40
35 30 25 20 15 10 5 0 A 197585 198695 19962005 AC C Others n

B 100
Percentage of within group 90 80 70 60 50 40 30 20 10 0 1990 AC Non-AC Downloaded from http://rheumatology.oxfordjournals.org/ by guest on March 1, 2013 >1990 P = 0.03

TABLE 2 ENAab and ENAab subtypes distribution during time

Decade of LN diagnosis 197585 (% within decade) 198695 (% within decade) 19962005 (% within decade) P-value (chi-square test) ENAab positive 17.2 48.8 69.0 <0.001

Ro 17.2 39.2 50.2 0.002

La 3.4 17.1 11.9

RNP 17.2 34.1 41.7

Sm 17.2 9.8 25.0

0.409 <0.001 <0.001

P-values refer to comparison between the first and third decades.

from RD to death, which was 7.8 (6.0) years (min <1, max 24). Comparing the three decades, 5 years after RD the mortality rates were 17.2, 7.7 and 4.7%, respectively. Reviewing the mortality rate over time and considering the first decade as the baseline, there is an obvious reduction in the mortality rate (P = 0.0274) (Table 4). A very striking finding is that the mortality rate for the AC patients in our cohort was increased compared with the A and C patients (P = 0.029) (Table 5). When we dichotomized the sample in AC vs non-AC patients, this association was even stronger (P = 0.0013). There was no statistically significant association between the overall ENAab status or the different ENAab subgroups and mortality. No association was found between the different WHO classes and mortality. ESRD Out of 154 patients, 30 (19.5%) went into ESRD. Comparing the decades, no significant differences were found. Almost all were women (n = 29) and no significant

differences in the ethnic distribution were found (A: 23.3%, AC: 36.7%, C: 26.7%, others: 13.3%). The mean (S.D) age at ESRD was 37.7 (13.86) years (range 1761 years). WHO Class IV was the commonest (72.0%). On average, the time from RD to ESRD was 7.5 years (S.D. 7.0, years range 124 years), 43.3% during the first 5 years of RD, 36.7% between 5 and 10 years and 20.0% after 10 years. The rate of evolution to ESRD 5 years after RD was 6.9% during the first decade, 7.7% during the second and 8.1% during the third. These differences are not statistically significant, with no difference in the risk of developing ESRD over time (P = 0.885) (Table 4). Patients of AC origin have a higher ESRD rate within 5 years of RD (P = 0.001) (Table 5). Reviewing the type of renal replacement therapy, most patients had either continuous ambulatory peritoneal dialysis (CAPD, n = 11) or haemodialysis (HD, n = 12), with a few having had both, usually due to CAPD failure (n = 7). RTx was performed in 16 patients, mostly after 2000 (n = 11), using living related donors (n = 6) or cadaver non-related (n = 7) generally within 5 years of ESRD diagnosis (n = 14). Only four failed RTx, three of which were cadaver grafts. One patient had recurrent LN without hindering the graft function. Overall, the mean duration of a functioning graft is 5.06 years (range 018 years). Before 2000, the commonest immunosuppressive drug was CSA (4/5), with tacrolimus subsequently used more frequently (7/11). Of the 30 patients that developed ESRD, 14 died. There was a significant association between AC origin and higher mortality (P = 0.027). No association between ENA status and type of renal replacement therapy was found. On average, death occurred 4.3 years after ESRD diagnosis (S.D. 3.7 years, range <113 years), with a mean (S.D.) age at death of 41.6 (12.5) years (range 1961 years). The commonest direct causes of death were infection (n = 5), renal (n = 3) and cardiovascular (n = 2) events.

www.rheumatology.oxfordjournals.org

1427

Sara C. Croca et al.

TABLE 3 Association between ENAab status and ethnicity

ENAab profile A AC C Other P (chi-square test)

Overall ENAab status, n (%) within ethnicity Positive 6 (26.1) Negative 17 (73.9) ENAab subtype, n (%) within ethnicity Ro Positive 4 (17.4) Negative 9 (82.6) La Positive 0 Negative 23 (100.0) RNP Positive 5 (21.7) Negative 18 (78.3) Sm Positive 4 (17.4) Negative 9 (82.6)

30 (81.1) 7 (18.9)

32 (44.4) 40 (55.6)

15 (75.0) 5 (25.0)

<0.001

22 (57.9) 16 (42.1) 8 (21.1) 30 (78.9) 19 (50.0) 19 (50.0) 11 (28.9) 27 (71.1)

23 (31.5) 50 (68.5) 6 (8.2) 67 (91.8) 17 (23.3) 56 (76.6) 8 (11.0) 65 (89.0)

14 (70.0) 6 (30.0) 4 (20.0) 16 (80.0) 8 (40.0) 12 (60.0) 2 (10.0) 18 (90.0)

<0.001

0.028

0.021

<0.104

Downloaded from http://rheumatology.oxfordjournals.org/ by guest on March 1, 2013

P-values refer to comparison between ethnicity and ENAab status overall and subtypes.

TABLE 4 A 30-year evolution of mortality and ESRD development rates within 5 years of RD diagnosis
Mortality and ESRD Mortality analysis Dead Alive Odds ratio (chi-square test) Cause of death 197585 (n = 29) 198695 (n = 41) 19962005 (n = 86)

Mean age at death, years Time from RD to death, years ESRD analysis ESRD Non-ESRD Odds ratio Time from RD to ESRD, years Time from ESRD to death, years CV: cardiovascular.

5 24 1 One CV, two renal failures, one cancer, one infection 33.8 2.2 2 27 1 1.9 7

3 37 0.38 Two CV, one suicide

34.0 3.3 4 36 1.46 2.4 7.8

4 82 0.34 (P = 0.027) Two renal failures, one infection, one cancer 33.6 6.25 7 79 1.24 (P = 0.885) 2.3 8.25

Discussion
During the past 30 years, much has changed in the way LN patients are managed. To our knowledge, very few LN cohorts have been assessed in detail over a 30-year period [1318] and even fewer have correlated varying ethnicity and serology with outcome. Indeed, a notable feature of the population that we have studied is the major change in its ethnic distribution, with a definite trend towards an increased prevalence of patients of AC origin that has not previously been reported. These patients are liable to more aggressive lupus-associated RD with a known poorer response to CYC than other ethnic groups [19, 20]. Our data concur with these findings, showing a clear increase in mortality among the

AC patients (P = 0.029), as well as ESRD development (P = 0.001). In our cohort, we found a strong association between AC patients and increased prevalence of Class V type nephritis (P = 0.01). Type V nephritis in its pure form occurs in 1020% of patients with LN [21]. These patients usually present with heavy proteinuria and hypoalbuminaemia and the majority develop hypertension, which is, in some cases, difficult to control [22]. Typically, the proliferative types of nephritis (Classes III and IV) are associated with a poorer prognosis. The fact that our AC patients had a worse outcome despite having a slightly lower incidence of proliferative nephritis suggests that there are other factors besides WHO class that put them at greater risk. The presence of nephritic-range proteinuria

1428

www.rheumatology.oxfordjournals.org

Assessment of LN cohort

TABLE 5 Association between ethnicity, mortality and ESRD development rate within 5 years of RD diagnosis
Mortality and ESRD P (chi-square test)

AC

Mortality analysis (within decade), n (%) Dead 0 9 (31.0) 2 (6.2) Alive 6 (100.0) 20 (69.0) 30 (93.8) ESRD analysis (within decade), n (%) ESRD 1 (16.7) 9 (30.0) 0 Non-ESRD 5 (83.3) 21 (70.0) 32 (100.0)

0.029

0.001

For statistical analysis reasons related to the size of the sample, only three ethnic groups were considered, with no effect on results.

is a potent cardiovascular risk factor that could, at least to some extent, account for the higher mortality in this ethnic group. Other explanations have been proposed to explain the less favourable outcome of AC patients, including genetic polymorphisms [21], socioeconomic factors and issues related to adherence to treatment [5]. Regarding the serological characterization of the population, the only significant difference was related to ENA prevalence, which increased with time (P = 0.01). This result is likely to be associated with the increasing number of AC patients, as reported in previous studies [23, 24]. The role of ENAab in the pathogenesis of LN is still not clear, with conflicting evidence arising periodically [2528]. Anti-Ro, anti-La and anti-Sm antibodies have been detected in renal biopsy specimens obtained from patients with lupus at autopsy [29]. This finding suggests that these antibodies may be involved in renal pathology though the case for anti-dsDNA/anti-nucleosome is much stronger [30]. We have said relatively little about the drug treatment of our LN patients partly because this topic has been reported elsewhere [31, 32] and partly because there are gaps in the records of some of our patients. In broad terms, our philosophy until 2003 was to use steroids and AZA in the first instance, switching to the i.v. CYC regimen advocated by the National Institutes of Health [33]. From 2003 onwards, we have preferred to use MMF to CYC and when this failed, rituximab. A more detailed analysis of these different approaches is pending. The mean age at death was 40.3 years with an average of 7.5 years between RD and death, showing the marked impact on the life expectancy of these patients. AC patients, in particular, have a greater mortality rate (P = 0.029). However, it is encouraging to see a clear reduction in the mortality rate over time (P = 0.027). Despite the overall reduction in mortality, barely any improvement was seen in the past 10 years with virtually no difference between the second and third decades (Table 5). The mortality reduction between the first and second decades is probably due to multiple issues, namely the widespread use of more aggressive immunossupressant regimens (CYC and AZA) and a better awareness of potential disease and treatment complications such as infection and

cardiovascular disease allowing for closer monitoring and earlier recognition. Additionally, recognition of the benefits of angiotensin-converting enzyme inhibitors in terms of renal protection [34] and their generalized use in the late 1990s as well as the use of lipid-lowering drugs, particularly statins, may also have contributed to the improvement in mortality rate. The evolution to ESRD in our group was 19.5%. Most patients developed ESRD within 10 years of RD (80%). Disappointingly, there was no reduction in the rate of evolution to ESRD during these 30 years (P = 0.885), which could be accounted for, at least partially, by the increase in prevalence of AC patients. The type of renal replacement therapy did not influence the outcome in these patients. The use of RTx increased, particularly after 2000, with encouraging results. Only four grafts failed, most of which were cadaver non-related donors. The mortality rate in the ESRD group was 46.7%, with death usually occurring within 5 years of ESRD diagnosis. This result emphasizes the notion that ESRD must be avoided, or at least postponed, to minimize the impact upon survival. In summary, LN is a relatively common manifestation of SLE associated with significant mortality and morbidity and reduction in life expectancy of >20 years. AC patients in particular have a poorer prognosis. There is encouraging evidence that during the past 30 years, the outcome for LN patients has improved; however, there is definitely room for improvement. The absence of further improvement in both mortality and ESRD rates suggests that we have maximized the benefits of traditional drug regimens. Novel biological therapies may be the next step in achieving rapid control of inflammation, thus preventing irreversible damage. Rheumatology key messages LN-associated mortality and morbidity have not been significantly reduced over a 30-year period. . Improving prognosis for LN patients will require a change in the current therapeutic approach.
.

Downloaded from http://rheumatology.oxfordjournals.org/ by guest on March 1, 2013

Disclosure statement: The authors have declared no conflicts of interest.

References
1 Cervera R, Khamashta MA, Font J et al. Systemic lupus erythematosus: clinical and immunological patterns of disease expression in a cohort of 1,000 patients. Medicine 1993;72:113. 2 Zabaleta-Lanz M, Vargas-Arenas RE, Tapanes F, Daboin I, Atahualpa Pinto J, Bianco NE. Silent nephritis in systemic lupus erythematosus. Lupus 2003;12:2630. 3 Baranowska-Daca E, Choi YJ, Barrios R, Nassar G, Suki WN, Truong LD. Nonlupus nephritides in patients with systemic lupus erythematosus: a comprehensive clinicopathological study and review of the literature. Hum Pathol 2001;32:1125.

www.rheumatology.oxfordjournals.org

1429

Sara C. Croca et al.

4 Seligman VA, Lum RF, Olson JL, Li H, Criswell LA. Demographic differences in the development of lupus nephritis: a retrospective analysis. Am J Med 2002;112: 726. 5 Adler M, Chambers S, Edwards C, Neild G, Isenberg D. An assessment of renal failure in an SLE cohort with special reference to ethnicity over a 25-year period. Rheumatology 2006;548. 6 Mok CC. Prognostic factors in lupus nephritis. Lupus 2005;14:3944. 7 Patel M, Clarke AM, Bruce IN, Symmons DP. The prevalence and incidence of biopsy-proven lupus nephritis in the UK: evidence of an ethnic gradient. Arthritis Rheum 2006;54:29639. 8 Rivera TL, Belmont HM, Malani S et al. Current therapies for lupus nephritis in an ethnically heterogenous cohort. J Rheumatology 2009;36:298305. 9 Cooper GS, Treadwell EL, St Clair EW, Gilkeson GS, Dooley MA. Sociodemographic associations with early disease damage in patients with systemic lupus erythematosus. Arthritis Rheum 2007;57:9936. 10 Lionaki S, Kapitsinou PP, Iniotaki A, Kostakis A, Moutsopoulos HM, Boletis JN. Kidney transplantation in lupus patients: a case-control study from a single centre. Lupus 2008;17:6705. 11 Ward MM. Changes in the incidence of end-stage renal disease due to lupus nephritis in the United States, 1996-2004. J Rheumatol 2009;36:63. 12 Isenberg DA, Garton M, Reichlin MW, Reichlin M. Long term follow up of autoantibody profiles in black female lupus patients and clinical comparison with Caucasian and Asian patients. Brit J Rheum 1997;36:22933. . The very long-term 13 Bono L, Cameron JS, Hicks JA prognosis and complications of lupus nephritis and its treatment. Q J Med 1999;92:2118. 14 Donadio JV, Hart GM, Bergstralh EJ, Holley KE. Prognostic determinants in lupus nephritis: a long term clinicopathologic study. Lupus 1995;4:10915. 15 Moroni G, Quaglini S, Gallelli B, Banfi G, Messa P, Ponticelli P. The long-term outcome of 93 patients with proliferative lupus nephritis. Nephrol Dial Transplant 2007; 22:25319. 16 Mok CC, Wong RW, Lau CS. Lupus nephritis in Southern Chinese patients: clinicopathologic findings and long term outcome. Am J Kidney Dis 1999;34:31523. 17 Esdaile JM, Abrahamowicz M, MacKenzie T, Hayslett JP, Kashgarian M. The time-dependence of long-term prediction in lupus nephritis. Arthritis Rheum 1994;37:35968. 18 Mok CC, Ying KY, Yim CW, Ng WN, Wong WS. Very long term outcome of pure lupus membranous nephropathy treated with glucocorticoids and azathioprine. Lupus 2009;18:10915. 19 Dooley MA, Hogan S, Jennette C, Falk R. Cyclophosphamide therapy for lupus nephritis patients: poor renal survival in black Americans. Kidney Int 1997;11: 8895. 20 Ward MM, Pyun E, Studenski S. Long-term survival in systemic lupus erythematosus. Patient characteristics

associated with poorer outcomes. Arthritis Rheum 1995; 38:27483. 21 Salmon JE, Millard S, Schachter LA et al. Fc gamma alleles are heritable risk factors for lupus in African Americans. J Clin Invest 1996;97:134854. 22 Austin HH, Illei GC, Braun MJ, Balow JE. Randomized controlled trial of prednisone, cyclophosphamide and cyclosporine in lupus membranous nephropathy. J Am Soc Nephrol 2009;20:901. 23 Sloan RP, Schwartz MM, Korbet SM, Borok RZ. Long-term outcome in systemic lupus erythematosus membranous glomerulonephritis. J Am Soc Nephrol 1996; 7:299. 24 Tikly M, Gould T, Wadee AA, Van der Westhuizen E, Mokgethwa BBN. Autoantibodies in black South Africans with systemic lupus erythematosus: spectrum and clinical association. Clin Rheumatol 1996;15:2615. rrez-Uren a S et al. a CO, Molina JF, Gutie 25 Garc Autoantibody profile in African-American patients with lupus nephritis. Lupus 1996;5:6025. panes FJ, Va squez M, Ram rez R, Matheus C, 26 Ta guez MA, Bianco N. Cluster analysis of anti-nuclear Rodr autoantibodies in the prognosis of SLE nephropathy: are anti-extractable nuclear antibodies protective? Lupus 2000;9:43744. 27 Koren E, Koscec M, Wolfson-Reichlin M et al. Murine and human antibodies to native DNA that cross-react with the A and D SnRNP polypeptides cause direct injury of cultured kidney cells. J Immunol 1995;154: 485764. 28 Hitchon CA, Peschken CA. Sm antibodies increase risk of death in systemic lupus erythematosus. Lupus 2007;16: 18694. 29 Mannik M, Merrill CE, Stamps LD, Wener MH. Multiple autoantibodies form the glomerular immune deposits in patients with systemic lupus erythematosus. J Rheumatol 2003;30:1495504. 30 Manson J, Isenberg DA. The origin and pathogenic consequences of anti-dsDNA antibodies in systemic lupus erythematosus. Expert Rev Clin Immunol 2006;2: 37785. 31 MacGowan JR, Ellis S, Griffiths M, Isenberg DA. Retrospective analysis of outcome in a cohort of patients with lupus nephritis treated between 1977 and 1999. Rheumatology 2002;41:9817. 32 Lu TY, Ng KP, Cambridge G et al. A retrospective seven year analysis of the use of B-cell depletion therapy in systemic lupus erythematosus at University College Hospital. The first 50 patients. Arthritis Care Res 2009;61: 4827. 33 Boumpas DT, Austin HA, Balow JE et al. Controlled clinical trial of methylprednisolone versus two regimens of pulse cyclophosphamide in severe lupus nephritis. Lancet 1992; 340:74150. 34 Chiurchiu C, Remuzzi G, Ruggenenti P. Angiotensin-converting enzyme inhibitoin and renal protection in non-diabetic patients: the data of the meta-analysis. J Am Soc Nephrol 2005;16:S5863.

Downloaded from http://rheumatology.oxfordjournals.org/ by guest on March 1, 2013

1430

www.rheumatology.oxfordjournals.org