Risk factors and impact of recurrent lupus nephritis in patients with systemic lupus erythematosus undergoing renal transplantation: Data from a single US institution




To determine the risk factors for recurrent lupus nephritis, allograft loss, and survival among patients with systemic lupus erythematosus (SLE) undergoing kidney transplantation.


The archival records of all kidney transplant recipients with a prior diagnosis of SLE (according to the American College of Rheumatology criteria) from June 1977 to June 2007 were reviewed. Patients who had died or lost the allograft within 90 days of engraftment were excluded. Time-to-event data were examined by univariable and multivariable Cox proportional hazards regression analyses.


Two hundred twenty of nearly 7,000 renal transplantations were performed in 202 SLE patients during the 30-year interval. Of the 177 patients who met the criteria for study entry, the majority were women (80%) and African American (65%), the mean age was 35.6 years, and the mean disease duration was 11.2 years. Recurrent lupus nephritis was noted in 20 patients (11%), allograft loss in 69 patients (39%), and death in 36 patients (20%). African American ethnicity was found to be associated with a shorter time-to-event for recurrent lupus nephritis (hazard ratio [HR] 4.63, 95% confidence interval [95% CI] 1.29–16.65) and death (HR 2.47, 95% CI 0.91–6.71), although, with the latter, the association was not statistically significant. Recurrent lupus nephritis and chronic rejection of the kidney transplant were found to be risk factors for allograft loss (HR 2.48, 95% CI 1.09–5.60 and HR 2.72, 95% CI 1.55–4.78, respectively). In patients with recurrent lupus nephritis, the lesion in the engrafted kidney was predominantly mesangial, compared with a predominance of proliferative or membranous lesions in the native kidneys.


African American ethnicity was independently associated with recurrent lupus nephritis. Allograft loss was associated with chronic transplant rejection and recurrence of lupus nephritis. Recurrent lupus nephritis is infrequent and relatively benign, without influence on a patient's survival.

Lupus nephritis continues to be a major cause of morbidity and mortality in patients with systemic lupus erythematosus (SLE). Renal involvement affects up to 60% of patients, depending on the ethnic composition and geographic origin of those studied (1–4). Despite advances in understanding the pathogenesis of SLE and use of aggressive treatments for lupus nephritis, the disease in 5–22% of these patients will progress to end-stage renal disease (ESRD) (5, 6).

Although successful transplantation with deceased-donor renal allografts has been performed since 1959, there has been a concern that lupus patients may have poor outcomes as compared with other patient populations. In 1975, however, investigators reviewed the American College of Surgeons/National Institutes of Health Transplant Registry (7) and observed that lupus patients undergoing renal transplantation had outcomes comparable with those in nonlupus patients, opening the doors to transplantation for patients with lupus (8–10). To date, there are relatively limited data about the factors that may account for renal and patient outcomes among lupus patients undergoing renal engraftment. Recognized risk factors for allograft loss in the lupus patient include thrombotic events (11, 12), donor type (8), and early rejection of the kidney transplant (8, 13–15). The exact role of recurrent disease in the engrafted kidney, however, remains unknown.

Recurrent lupus nephritis, a major concern in the early years after kidney transplantation, occurs at variable rates, ranging from 0 to 30% (16–21). Differences in patient characteristics and in the indications for renal biopsy may, to a certain extent, account for this broad range of incidence rates. The factors associated with recurrence have not been identified in the studies published to date. Furthermore, whether recurrent lupus nephritis in the allograft may influence either allograft loss or patient survival has not been determined.

Utilizing a renal-transplant registry from a single center in the US, we identified all SLE patients who had undergone renal transplantation to 1) assess the rate of recurrence of lupus nephritis and the factors predisposing to it, 2) characterize the features of recurrent lupus nephritis, and 3) ascertain the risk factors for allograft loss and patient survival, with special emphasis on the impact of recurrent lupus nephritis on these outcomes. The results of this study provide the information necessary to close the gaps in the existing literature regarding the safety of renal transplantation for lupus nephritis.



Between June 1977 and June 2007, a total of 6,922 kidney transplantations were performed at the University of Alabama at Birmingham (UAB), a large tertiary academic medical center serving Alabama, southern Tennessee, western Georgia, Mississippi, and the Florida panhandle. Two hundred twenty of the procedures (3.2%) were done in 202 SLE patients with ESRD secondary to lupus nephritis. All patients met the classification criteria for SLE as established over the years by the American College of Rheumatology (22–24). This study was conducted with the approval of the Institutional Review Board at UAB. About 5 years ago, a significant effort was placed in compiling clinical, immunologic, and therapeutic variables for all of these patients, as well as for their donors (25).

For the purpose of this study, this database was expanded by conducting a review of all available archival records, including both hospital and outpatient encounters (in nephrology, renal transplantation, and rheumatology). Patients who had lost the allograft or died within 90 days of engraftment were excluded from these analyses, because they would not have had time to develop recurrent disease. For recipients who had undergone renal transplantation on more than one occasion, the first procedure was considered the target transplantation, except in 1 patient for whom the second transplantation was analyzed because recurrent lupus nephritis occurred in this graft rather than in the first. At our institution, biopsies were not routinely performed in all transplant recipients and were, rather, prompted by the detection of unexplained hematuria, proteinuria, or decline in renal function. For the criterion of proteinuria, the cutoff value has not been uniform over the 30-year study duration, but general guidelines have been followed; during the 1970s and 1980s, biopsies were more likely to be performed in patients with proteinuria of ∼1 gm per day, but in the recent 2 decades, there has been a lower threshold value of ∼0.5 gm per day. Unfortunately, information about the precise level of proteinuria that prompted the biopsy in each patient was not available.

Outcome variables.

The outcome variables examined in this study were recurrent lupus nephritis, which was defined histopathologically according to the World Health Organization (WHO) classifications of lupus nephritis introduced in 1974 and revised in 1982 and 1995 (including findings on histopathologic analyses, immunofluorescence staining, and electron microscopy) (26), allograft loss, which was defined as a loss of renal function requiring renal replacement therapy, and patient survival.

Disease-specific variables.

Variables recorded from the socioeconomic and demographic domains included age, sex, ethnicity, weight, and health-related behaviors (alcohol intake and smoking) at the time of the engraftment. Where available, information was compiled on the clinical manifestations of SLE, types of autoantibodies present (antinuclear, anti-Sm, and anti–double-stranded DNA), complement levels, viral serologic test results, and immunosuppressant drugs taken before transplantation. The histopathologic diagnosis attributed to the native kidney (according to the WHO classification [26]) and the type of renal replacement therapy (hemodialysis or peritoneal dialysis) were documented. The posttransplantation histopathologic findings were classified as normal, recurrent lupus nephritis, acute or chronic rejection, thrombotic changes, acute tubular necrosis, calcineurin inhibitor toxicity, or other. For those patients with recurrent lupus nephritis, the therapies instituted were also specifically investigated. Allograft survival or allograft loss as well as death and cause of death were rigorously researched using all available records. The Social Security Administration and the National Death Index databases, however, were not accessed.

Interval variables.

To perform our analyses, the following intervals were computed: time-to-ESRD, defined from the date of the diagnosis of lupus to ESRD onset, dialysis duration, defined from the date of ESRD to the date of transplantation, duration of followup, defined from the date of transplantation to the date of death, date of allograft loss, or date of last followup visit, and duration of disease, defined from the date of the diagnosis of lupus to the date of transplantation. Time-to-event intervals for recurrent lupus nephritis, allograft loss, and death were computed from the date of transplantation.

Donor variables.

The following donor variables were examined: age, sex, ethnicity, donor type (living-related, living-unrelated, or deceased), and HLA-antigen match. Occurrence of delayed graft function (defined as the need for dialysis within the first week of engraftment) was also recorded.

Posttransplantation medications.

The immunosuppressive medications used and the standard induction therapy have changed during the 30-year study period. Azathioprine was introduced in the late 1960s, but the immunosuppressive protocol was modified in 1984, when cyclosporine, a calcineurin inhibitor, was added. Tacrolimus and mycophenolate mofetil were not introduced in the standard protocol until the mid-1990s. The use of antibodies (OKT3, anti–interleukin-2, and thymoglobulin), as components of the induction regimen, and the use of glucocorticoids were also recorded. Cumulative dosage and treatment duration for each of these compounds, however, could not be precisely defined from the available records.

Statistical analysis.

For the outcomes of interest (recurrent lupus nephritis, allograft loss, and survival), time-to-event analyses were performed. In a small proportion of patients (fewer than 10%), important dates were missing in a nonsystematic manner, which might have precluded the inclusion of these patients in the study. However, these missing dates were imputed on the basis of the available data on the dates of either diagnosis, ESRD onset, or transplantation obtained from all other patients; the proportions for these intervals were calculated and then used to impute the missing dates.

The variables of interest were then examined by univariable Cox proportional hazards regression analyses in relation to the 3 outcomes of interest. Variables with P values less than or equal to 0.10 in these univariable regressions were entered into multivariable Cox regressions. Recurrent lupus nephritis was entered into the regression models with allograft loss, whereas both recurrent lupus nephritis and allograft loss were included in the regression models with patient survival. Variables with P values less than or equal to 0.05 in these analyses were considered to be statistically significant. The data are presented as hazard ratios (HRs) with their 95% confidence intervals (95% CIs). An HR less than 1 indicates a longer time-to-event, and an HR greater than 1 indicates a shorter time-to-event. In addition, the subset of patients in whom recurrent lupus nephritis developed was examined using descriptive statistical analyses. Data were analyzed using SAS version 9.1 (SAS Institute, Cary, NC).


Characteristics of the total cohort.

One hundred seventy-seven of the 202 patients with SLE undergoing renal transplantation met the inclusion criteria for this study. The number of renal transplantations conducted at UAB has gradually increased over the 30-year study period (Figure 1A), whereas in the late 1990s, the number of transplantations in SLE patients reached a peak (3.73%) and has tended to decrease somewhat in the current decade (3.01%); in comparison, 1.12% were conducted at UAB in the late 1970s to early 1980s and 1.95% in the late 1980s and 1990s (P < 0.0001) (Figure 1A). Only 12% of the SLE patients who had received a kidney engraftment were cared for at UAB before they underwent transplantation. In the total cohort, the patients were predominantly women (80%) and 65% of the patients were African American. The mean ± SD age of the total cohort was 35.6 ± 10.0 years (range 13–82 years), and the mean ± SD disease duration at the time of the transplantation was 11.2 ± 8.7 years, whereas the mean ± SD interval from SLE diagnosis to onset of ESRD was 8.1 ± 7.5 years. The mean ± SD duration of dialysis was 3.1 ± 2.9 years, and the mean ± SD followup interval after transplantation was 7.1 ± 5.5 years (range 0.3–26.8 years).

Figure 1.

Summary of renal transplantations performed at the University of Alabama at Birmingham over the past 30 years. Except for the most recent time interval in 2007 (January to June), all other time periods are represented by 2-year intervals. A, Number of renal transplantations performed in all patients with systemic lupus erythematosus (SLE) and non-SLE transplant recipients during the intervals. B–D, Number of SLE patients undergoing transplantation in the 2-year intervals in whom recurrent lupus nephritis (RLN) did or did not develop at a subsequent time, among all SLE patients (B), African American (AA) SLE patients (C), and Caucasian (CC) SLE patients (D).

Renal replacement therapy before engraftment was required in nearly all patients (97%); most patients received hemodialysis (76% versus 24% receiving peritoneal dialysis). The induction therapy protocols have changed throughout the years, as have posttransplantation medication regimens. Patients were treated with azathioprine (n = 68 [38%]), cyclosporine (n = 135 [76%]), mycophenolate mofetil (n = 131 [74%]), or tacrolimus (n = 66 [37%]). In some cases, more than 1 medication was used. All patients received glucocorticoids.

Characteristics of the donors.

Sixty percent of the donors were male and 66% were African American. Among the donor kidneys obtained, 104 (59%) of the grafts were obtained from deceased donors, while 60 (34%) were from living-related donors and 12 (7%) were from living-unrelated donors. For 1 donor, this information was not available.

Characteristics of the subset of patients with recurrent lupus nephritis.

Rate and predictors of recurrence.

Of the 177 patients who met the inclusion criteria for this study, recurrent lupus nephritis developed in 20 (11.3%). The rate of recurrence has varied over the years, from a low of 5.6% for those undergoing transplantations in 2001–2002 to a high of 20.8% for those undergoing transplantations in 1997–1998 (Figure 1B). The overall incidence of recurrence was 0.027 per patient-year. The overall rates of recurrence in African Americans (13.9%) (Figure 1C) and in Caucasians (6.9%) (Figure 1D) did not differ significantly (P = 0.2674).

The univariable and multivariable analyses of the predictors of time-to-occurrence of recurrent lupus nephritis are shown in Table 1. Age and sex were not associated with either a shorter or a longer interval to recurrence of lupus nephritis. African American ethnicity, in contrast, was associated with a shorter interval to recurrent lupus nephritis, although a statistically significant association was only observed in the multivariable analysis (HR 4.63, 95% CI 1.29–16.65). The effect of chronic rejection of the renal transplant on the occurrence of recurrent lupus nephritis could not be computed because there were no patients who exhibited both; thus, chronic transplant rejection does not appear to increase the risk of developing recurrent lupus nephritis.

Table 1. Univariable and multivariable Cox proportional hazards regression analyses of the predictors of time-to-occurrence of recurrent lupus nephritis*
VariableUnivariable analysisMultivariable analysis
  • *

    Estimates for chronic transplant rejection could not be computed because none of the patients in whom recurrent lupus nephritis developed had experienced chronic transplant rejection. HR = hazard ratio; 95% CI = 95% confidence interval.

Male sex1.490.54–4.120.44251.360.43–4.340.6058
African American ethnicity2.780.93–8.340.06834.631.29–16.650.0190
Posttransplantation medication      
 Cyclosporin A0.710.23–2.160.5395
 Mycophenolate mofetil7.061.01–57.130.04873.210.35–29.190.3003

The roles of various posttransplantation medications are also shown in Table 1. For patients receiving either cyclosporine or mycophenolate mofetil, neither of these medications was associated with either a longer or a shorter interval to recurrence of lupus nephritis. Use of azathioprine was associated with a longer interval to recurrence of lupus nephritis in the univariable analysis (HR 0.17, 95% CI 0.05–0.54), but this association did not quite reach statistical significance in the multivariable analysis (HR 0.28, 95% CI 0.08–1.02). Use of tacrolimus, conversely, had an opposite effect by exhibiting a shorter interval to recurrence of lupus nephritis, in both the univariable and multivariable analysis.

Description of patients with recurrent lupus nephritis.

The overall features of the 20 patients in whom recurrent lupus nephritis developed are listed in Table 2. These patients were predominantly women (75%) and the mean ± SD age was 33.5 ± 9.6 years; these features were comparable with those in patients in whom recurrent disease did not develop. The proportion of African Americans among the group of patients with recurrent lupus nephritis (80%) did not statistically significantly differ from that among the group of patients who did not experience recurrent disease (64%). Histopathologic data were not available for the native kidneys of 3 of the 20 patients. Of note, as shown in Table 3, the predominant glomerular lesion in the native kidneys was either proliferative (n = 10) or membranous (n = 6), whereas in the majority of the allografts, it was a mesangial lesion (n = 12); this difference in pattern of lesions was statistically significant (Yates' χ2 = 5.02, P = 0.0251). The relatively small number of patients available for analysis precluded performing additional analyses to determine the factors associated with the development of a proliferative lesion in the engrafted kidneys.

Table 2. Selected sociodemographic and clinical features of the 20 patients with biopsy-proven recurrent lupus nephritis*
  • *

    SLE = systemic lupus erythematosus; ESRD = end-stage renal disease; RLN = recurrent lupus nephritis.

 Age, mean ± SD years33.5 ± 9.6
 Sex, no. (%) women15 (75)
 Ethnicity, no. (%) 
  African American16 (80)
  Caucasian4 (20)
 Smoker, no. (%)1 (5)
 Alcohol use, no. (%)1 (5)
 Weight, mean ± SD kg65.4 ± 7.6
 Deceased, no. (%)2 (10)
Interval, mean ± SD years 
 Diagnosis duration (SLE–ESRD)11.3 ± 10.9
 Dialysis duration (ESRD–transplantation)3.4 ± 3.7
 Time-to-recurrence (transplantation–RLN)4.5 ± 3.7
 Followup time (transplantation–last visit)6.2 ± 4.1
 Disease duration (SLE–last visit)14.1 ± 4.1
Posttransplantation serologic feature, no. positive/total no. tested (%) 
 Anti-DNA antibodies3/13 (23)
 Antinuclear antibodies12/12 (100)
 Low complement6/15 (40)
Table 3. Histopathologic findings in the native and transplanted kidneys in the 20 patients with biopsy-proven recurrent lupus nephritis*
Histopathologic featureNativeTransplanted
  • *

    Values are the number of kidneys. Three patients for whom this information was not available were excluded from this computation (2 patients in whom a biopsy had not been obtained and 1 for whom this information was missing); these 3 patients had mesangial glomerulonephritis in the transplanted kidneys. WHO = World Health Organization.

  • Yates' χ2 = 5.02, P = 0.0251 versus native kidneys.

Mesangial glomerulonephritis (WHO class II)112
Focal proliferative glomerulonephritis (WHO class III)22
Diffuse proliferative glomerulonephritis (WHO class IV)81
Membranous glomerulonephritis (WHO class V)65

Even though 9 (45%) of the 20 patients with recurrent lupus nephritis eventually lost the allograft, this proportion was comparable with that among patients in whom recurrent disease did not develop (38%) (P = 0.558). Most patients with recurrent lupus nephritis underwent increases in their doses of immunosuppressive medications. In only 2 cases, cyclophosphamide was added to the treatment regimen because of the concurrent development of other major lupus manifestations (results not shown). Data on pretransplantation serologic markers and clinical manifestations were unavailable for the majority of patients in the archival records at our institution. Thus, we could not include any of these pretransplantation variables as potential risk factors for the examined outcomes.

Allograft loss.

At the time of the analyses, 69 patients (39%) had lost the allograft. In 51 of these patients (74%), a kidney biopsy had been performed (mean interval from transplantation to biopsy 4.7 years); 9 of the 51 patients in whom biopsies had been performed exhibited changes of recurrent lupus nephritis, whereas in the remaining 42 patients, the histologic pattern was predominantly chronic transplant rejection. Of all the variables examined, only recurrent lupus nephritis and chronic transplant rejection were associated with a shorter interval to allograft loss in the multivariable analyses (HR 2.48, 95% CI 1.09–5.60 and HR 2.72, 95% CI 1.55–4.78, respectively), while African American ethnicity was significantly associated with a shorter interval to allograft loss in the univariable analysis only (HR 2.24, 95% CI 1.28–3.93). Age, sex, smoking, and donor characteristics were not associated with the risk of experiencing allograft loss. Among the posttransplantation medications taken, the use of mycophenolate mofetil was significantly associated with a decreased risk of allograft loss in the univariable analysis, but not in the multivariable analysis. These data are shown in Table 4.

Table 4. Univariable and multivariable Cox proportional hazards regression analyses of the predictors of time-to–allograft loss in patients with lupus nephritis who had undergone kidney transplantation*
VariableUnivariable analysisMultivariable analysis
  • *

    HR = hazard ratio; 95% CI = 95% confidence interval.

  • Deceased donor is the reference group.

Male sex0.890.46–1.690.71280.990.46–2.150.9786
African American ethnicity2.241.28–3.930.00501.710.89–3.290.1075
Recurrent lupus nephritis1.520.75–3.100.24182.481.09–5.600.0295
Living-related donor0.790.46–1.350.3804
Living-unrelated donor0.530.13–2.180.3785
Chronic transplant rejection2.831.75–4.50<0.00012.721.55–4.780.0005
Posttransplantation medications      
 Cyclosporin A1.640.77–3.490.1986
 Mycophenolate mofetil0.600.36–1.020.05520.630.36–1.100.1039


Among the 177 patients studied, 36 patients (20%) died during the 30-year followup. In all 36 of these patients, allograft loss preceded death. However, the impact of allograft loss on death could not be computed statistically, because there were no deaths among those patients who had not lost the allograft. Thus, allograft loss should be considered to be a strong predictor of decreased survival among SLE patients, even though the HR could not be computed.

Of the deceased SLE patients, biopsies had been performed in 23 patients (64%), with a mean interval from transplantation to biopsy of 4 years. Only 2 patients in this group experienced a recurrence of lupus nephritis. Despite extensive efforts, it was not possible to ascertain the cause of death in 31% of these patients, but the leading causes of death in the remaining 69% of patients were infection (n = 10 [40%]) and cardiovascular disease (n = 5 [20%]). Of note, there was no death attributable to SLE. Deceased patients showed distributions of age and sex that were comparable with those in the group of surviving patients, but African Americans were overrepresented among the deceased patients as compared with the surviving patients (81% versus 62%; P = 0.0483).

African American ethnicity and chronic transplant rejection negatively affected survival in the univariable analyses (HR 2.48, 95% CI 1.19–6.28 and HR 2.68, 95% CI 1.39–5.10, respectively), but neither variable was statistically significantly associated with survival in the multivariable analysis, although the association of African American ethnicity with survival approached significance (HR 2.47, 95% CI 0.91–6.71; P = 0.0771). Age, sex, smoking, and transplant donor type were not found to impact survival in our analyses. Of note, lupus nephritis in the engrafted kidney did not have a negative impact on survival. In terms of medication effects, we found an increased risk of diminished survival in association with the use of azathioprine (HR 2.41, 95% CI 1.10–5.27) and a decreased risk in association with the use of mycophenolate mofetil (HR 0.29, 95% CI 0.14–0.58) in the univariable analysis. These data are shown in Table 5.

Table 5. Univariable and multivariable Cox proportional hazards regression analyses of the predictors of survival in lupus nephritis patients who had undergone kidney transplantation*
VariableUnivariable analysisMultivariable analysis
  • *

    Estimates for allograft loss could not be obtained because none of the patients who retained the allograft had died by the time that the study was conducted. HR = hazard ratio; 95% CI = 95% confidence interval.

  • Deceased donor is the reference group.

Male sex1.290.59–2.850.51851.640.65–4.160.2979
African American ethnicity2.481.19–6.280.01722.470.91–6.710.0771
Recurrent lupus nephritis0.580.14–2.450.46100.750.17–3.400.7116
Living-related donor0.570.26–1.270.1706
Living-unrelated donor0.450.06–0.310.4304
Chronic transplant rejection2.681.39–5.100.00331.850.84–4.050.1249
Posttransplantation medications      
 Cyclosporin A1.940.67–5.620.2210
 Mycophenolate mofetil0.290.14–0.580.00050.500.21–1.180.1128


In this study, we examined the largest single-center population of its kind for the risk factors that could influence outcomes in patients with lupus nephritis who had undergone renal transplantation. Our results showed that African American ethnicity is independently associated with recurrent lupus nephritis and, possibly, with diminished survival. Chronic rejection of the kidney transplant increases the risk of allograft loss but has no impact on survival. This study is the first to show that recurrent lupus nephritis negatively impacts allograft survival but has no impact on patient survival. In addition, we observed that lupus nephritis in the transplanted kidney is relatively milder than the original disease, which may account for its lack of effect on patient survival in our study.

The overall rate of recurrent lupus nephritis in our registry (11%) falls within the rates reported in the published literature (9, 16, 18–21, 27, 28). Even within our own registry, however, significant variability was observed (∼6–21% over 30 years). This broad range could be due, in part, to the variable ethnic composition of the patients who underwent renal transplantation at our institution over the span of 30 years, and also may be attributable to the fact that the criteria used to perform biopsies of the engrafted kidneys have changed over the years. Admittedly, the true rate of recurrence can be ascertained only when all patients with lupus nephritis undergoing renal engraftment have a scheduled renal biopsy at specified intervals, which was not the case in our study. Furthermore, immunosuppressive therapy may have ameliorated the signs of active nephritis, as well as other, nonrenal lupus manifestations, leading to the diminished recognition of recurrent lupus nephritis and, possibly, the lack of identification of lupus itself as the cause of death. Nevertheless, given that our data were derived from a single institution in which 2 of the senior authors (BAJ and GSA) have been on the faculty for almost as long as the study period, it is possible that less variability in the indication for a biopsy would be expected in this study as compared with data generated from more than one center.

We also characterized the disease in the allograft as relatively benign (predominantly WHO class II lupus nephritis) as compared with that in the native kidney. With the exception of 2 patients, whose regimen included the addition of cyclophosphamide after the diagnosis of recurrent disease, only minor adjustments in the immunosuppressive regimen to treat this manifestation were made. The histopathologic findings in our patients were consistent with the data reported by Stone et al and Goral et al (10, 19), and probably account for the observation that recurrent lupus nephritis did not lead to death in our patients. The effect of our newer immunosuppressive regimens administered pre- and posttransplantation may explain the fact that proliferative lesions rarely developed in the transplanted kidneys or that recurrent lupus nephritis was less likely to occur.

African American ethnicity has been shown to be a risk factor for allograft loss in SLE (29). Although this association with ethnicity was significant in our univariable analysis, the significance was not retained in the multivariable analysis, probably because of the close relationship between chronic transplant rejection and allograft loss (results not shown). African American ethnicity, however, was a significant risk factor for recurrent lupus nephritis, and the association approached significance when this ethnicity was analyzed as a risk factor for survival. The findings in a study by Dooley et al (30) showed that, compared with Caucasians, lupus nephritis developed more frequently in African Americans (60% versus 25%). Moreover, African Americans displayed a less favorable response to treatment, and thus the disease progressed to ESRD at higher rates (30). Other studies have shown that African American patients with SLE also die at increased rates, although when adjustments for socioeconomic factors are carried out, this difference in the mortality rate is no longer evident (31–34). Our findings underscore the well-known fact that patients from this ethnic group tend to exhibit more severe initial disease manifestations and less favorable outcomes (31, 35–37).

Thus, the findings reported herein are not totally unexpected and were probably related to both genetic (38–40) and socioeconomic factors (e.g., poverty, years of education, and compliance), which, unfortunately, could not be examined because these data were not part of the registry. Although we and others have also shown that Hispanic patients with lupus, particularly those of Amerindian background, also experience severe disease and have less-than-favorable outcomes (1, 41–43), this registry included very few Hispanics because, until recently, this ethnic group has been relatively underrepresented in our referral area (44).

Not surprisingly, we found that chronic rejection of the transplant had a deleterious effect on allograft survival, whereas recurrent lupus nephritis was a risk factor for allograft loss but had no impact on patient survival. Our data thus suggest that SLE patients are good candidates for renal transplantation.

Although use of deceased-donor allografts has, for the most part, led to worse outcomes than has been observed with living-donor allografts (8, 29, 45, 46), we did not observe such a relationship. In addition, when smoking was analyzed as a risk factor, the relationship did not quite reach statistical significance as a factor for allograft loss, despite the known deleterious effect of this behavior in the kidney, particularly in lupus patients (47, 48). Therefore, smoking cessation should constitute part of the pretransplantation preparation.

The true relationship between use of immunosuppressive medications and the patients' outcomes in our study should be considered carefully, since changes in transplantation care that have taken place over the past 30 years may account for some of our observed trends. We did not adjust for other confounders, such as doses or treatment duration. In our study, we demonstrated a protective effect of azathioprine and, conversely, a deleterious effect of tacrolimus in relation to the development of recurrent lupus nephritis, although neither of these medications had an impact on allograft loss or patient survival. Moreover, use of mycophenolate mofetil increased the risk of recurrent lupus nephritis, whereas a protective effect of this medication was observed in association with allograft survival and patient survival. This effect of mycophenolate mofetil has been reported previously (8). It should be noted that the doses used traditionally in the transplantation regimen are lower than would be used for treatment of lupus nephritis in the native kidney. The effect of glucocorticoids could not be properly examined, because this medication had been included in all immunosuppressive regimens, although in a nonuniform manner.

Some limitations of our study are worth noting. Detailed information specific to SLE was not captured uniformly for all patients, and thus this information could not be included in our analyses. Missing information included the extent of nonrenal lupus manifestations, autoantibody profiles, and complement levels. Moreover, important data regarding histopathologic lesions in the native kidneys were inconsistently recorded, and details on the biopsy results obtained were lacking (e.g., findings of crescents, thrombotic microangiopathic changes, or karyorrhexis). Not all of the patients underwent routine renal biopsy surveillance, and the magnitude of proteinuria was not available for all patients. Inexplicably, 26% of the patients who had lost the allograft had not undergone a biopsy. Thus, as already noted, it is conceivable that the actual incidence of recurrent lupus nephritis may have been higher than the 11% reported. Any of the associations observed between the outcomes and medications would be difficult to interpret, since this study did not control for these variables and because the precise doses and durations of the different medications used for induction and maintenance were not available.

Furthermore, the precise cause of death could not be properly examined because we were unable to obtain this information in about one-third of the patients. Finally, there was no provision for the storage of biologic specimens from all registry patients, and therefore genetic studies aimed at identifying potential markers of renal disease progression could not be conducted (39, 40).

Despite these limitations, we have presented solid data describing the outcomes of patients with lupus nephritis who undergo kidney transplantation. First, African American ethnicity is a consistent marker of worse outcomes in these patients. Second, the occurrence of recurrent lupus nephritis does not portend a poor prognosis in terms of patient survival. Finally, azathioprine exerts a beneficial effect in these patients in terms of reducing the rate of recurrent lupus nephritis, and mycophenolate mofetil may potentially have a beneficial effect both on allograft survival and on patient survival. A structured protocol that includes a more complete ascertainment of clinical, immunologic, and socioeconomic/demographic parameters, coupled with establishment of a biologic repository, would enhance the findings presented in our study and those previously reported in the literature. A different question worth addressing is why only some patients with lupus nephritis experience disease progression to ESRD. In short, much work remains to be done to advance this field for the betterment of patients with SLE.


All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Dr. Alarcón had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study conception and design. Burgos, Perkins, Pons-Estel, Liu, W. T. Kendrick, Julian, Alarcón, Kew.

Acquisition of data. Burgos, Perkins, Pons-Estel, S. A. Kendrick, Liu, W. T. Kendrick, Cook, Julian, Alarcón, Kew.

Analysis and interpretation of data. Burgos, Perkins, Pons-Estel, S. A. Kendrick, Liu, Cook, Julian, Alarcón, Kew.


The authors thank Jan Scott and her staff for their assistance with all aspects of medical records review, and Maria A. Tyson for her expert assistance in the preparation of the manuscript. We are also indebted to the numerous faculty, fellows, and residents of the UAB Departments of Surgery and Medicine who have, over the years, collected data on these patients, to the nephrologists who referred their patients to UAB for renal transplantation, and to Dr. William J. Koopman for his valuable comments on a previous version of the manuscript.