Introduction

Systemic lupus erythematosus (SLE) starts with renal symptoms in about 30–50% of patients. After 10 years, over 70% of them have renal complications [1]. Such renal disease continues to cause major morbidity and some mortality for around 30–40% of patients with SLE [2]. The diffuse proliferative glomerulonephritis (class IV of the World Health Organization, WHO, classification) is the most frequent and severe form of renal involvement in SLE. The prognosis of this form of lupus nephritis (LN) was ominous until the early 1960s with a reported 17% 5-year survival in the era 1953–1969. This percentage, however, has considerably improved over the years, reaching 82% in the early 1990s, and continues to improve [3].

Several factors have contributed to this improvement in the prognosis: earlier diagnosis of SLE, detection of milder forms of SLE nephritis, and more appropriate prevention and management of extrarenal complications. There are few doubts, however, that the tremendous improvement in the prognosis of LN is mainly due to the refinement of the specific treatment [4].

Proliferative LN is characteristically a protracted disease with a waxing and waning course. The long-term follow-up of affected patients has demonstrated that focal and diffuse proliferative forms of glomerulonephritis (WHO class III and IV, respectively) progress to chronic renal failure (CRF) in the absence of appropriate immunosuppressive treatment [5]. Immunosuppressive regimens of glucocorticoids combined with cytotoxic drugs, particularly cyclophosphamide (CYC), were reported to be effective for the treatment of severe proliferative LN [6].

Open-label and controlled clinical trials have shown that CYC is superior to corticosteroids in preserving renal function; only a minority of patients treated with this drug reach end-stage renal disease (ESRD) [7]. Nevertheless, immunosuppressive therapy can be associated with adverse effects that are not evident during short-term clinical trials [8]. Such complications must be weighed against morbidity from uncontrolled disease and the need for extended therapy.

Several investigators have, however, raised some concerns about the indiscriminate use of the so-called “National Institute of Health (NIH) regimen” to treat all LN patients [9]. First, the results of the NIH studies, as well as that of a recent meta-analysis of all randomized trials in LN [10], failed to demonstrate that an extended course of IV CYC was superior in terms of renal outcome and survival to other regimens of oral or IV cytotoxic drug(s). Second, high-dose IV CYC treatment is highly toxic; up to 25% of patients developed herpes zoster infection, up to 26% experienced other serious infection, and up to 52% of women were at risk of developing ovarian failure [11]. Thirdly, clinically milder cases of biopsy proven proliferative nephritis—for which lesser aggressive treatment might be justified—are now frequently diagnosed due to prompt assessment of early renal involvement. On the other hand, long-term therapy with CYC enhanced renal survival in patients with proliferative forms of LN [12].

Further controlled trials have shown superiority of monthly IV pulses of CYC over monthly IV pulses of methylprednisone (MP), and a more protective effect of the long-term administration of IV pulses of CYC over a short-term treatment [13]. A controlled study showed that the combined administration of pulses of CYC and MP was more effective than individual pulses of either MP or CYC. With the combined regimen, however, side-effects such as osteonecrosis, infections and ovarian failure were also more frequent [14].

Data on mycophenolate mofetil (MMF) are limited to small case series. It may be an alternative to CYC as initial therapy, particularly among patients who refuse or cannot tolerate CYC. Several prospective trials have addressed this issue, with MMF being, at least, not inferior to CYC [15]. The findings from these studies suggested that a MMF regimen provides similar or perhaps greater efficacy and fewer adverse effects than a CYC regimen. However, these benefits may not be universally accepted.

Although CYC remains the ‘gold standard’ treatment for severe organ threatening SLE, including LN, its use is still controversial and many issues remain debated, such as the timing and length of treatment, the route of administration and the ideal dosage. Consequently, the objective of this study is to evaluate the initial efficacy, toxicity and short-term outcome of intermittent LD-CYC versus HD-CYC therapy in patients with severe LN.

Patients and methods

Study design and population

Patients meeting the American Rheumatism Association [16] diagnostic criteria for systemic lupus erythematosus and who had undergone a kidney biopsy were screened at Mansoura Urology and Nephrology Center, Mansoura University, Egypt. Those who were 18 years of age or older and had received a histological diagnosis of proliferative LN (WHO class IV) and a 24-h proteinuria of >500 mg were included. They were enrolled in this randomized, prospective study between August 2005 and September 2006. Patients who had taken CYC or AZA during the previous year or had taken >15 mg/day of prednisolone (or the equivalent) during the previous month were excluded (except for a course of glucocorticoids for a maximum of 10 days before referral). Other exclusion criteria were: renal thrombotic microangiopathy, pre-existing CRF, pregnancy, previous malignancy, diabetes mellitus, previously documented severe toxicity to immunosuppressive drugs, and anticipated poor compliance with the protocol. To avoid selection bias, all nephritis patients who met the inclusion criteria were randomized into the trial, except for a few patients who declined to participate.

The final number of patients who fulfilled the inclusion criteria and had not been dropped from the study was 46. The study was approved by our local Medical Research and Ethics Committee and a written informed consent was obtained from every patient.

Treatment

We recruited, from the Outpatient Clinic Department, every patient who was 18 years of age or older and who had received a histological diagnosis of proliferative LN and 24-h proteinuria of >]500 mg to enrolled to the study. Randomization was done manually without the use of a computer algorithm to allocate every other patient to either group which was then assigned to receive one of two regimens:

  1. (1)

    High-dose group (HD); 26 patients received eight IV CYC pulses within a year (six monthly pulses followed by two quarterly pulses). The initial CYC dose was 0.5 g/1.73 m2 body surface area intravenously. Subsequent doses were increased by 250 mg according to the white blood cell count nadir measured on day 14, with a final increment to reach a maximum dose of 1 g/1.73 m2 of body surface area.

  2. (2)

    Low-dose group (LD); the remaining 20 patients received six fortnightly IV CYC pulses at a fixed dose of 500 mg.

Of note, the six patients with the severest form of LN who were allocated to the group of HD-CYC regimen.

In both treatment arms, both prednisone and AZA were given as previously recommended for class IV LN (1). All patients initially received oral prednisone (0.5 mg/kg of body weight daily) for 4 weeks. The prednisone dose was then given every other day after being tapered by 5 mg each week to the minimal dose required to control extrarenal manifestations of SLE or to 0.25 mg/kg every other day, whichever was greater. AZA (2 mg/kg/day) was started 2 weeks after the last CYC injection and continued to the end of the study. For patients who developed AZA-related toxicity, the dosage was reduced to 1 mg/kg/day and those who did not tolerate such lowered AZA dosage (none) were planned to drop from the trial.

Cyclophosphamide administration

IV CYC was given in 100–500 ml saline over 1 h accompanied by adequate oral hydration to promote frequent voiding of dilute urine in the first 24 h after drug administration. Further, to reduce bladder toxicity, 100 mg of Mesna was initially given IV in 0.9% sodium chloride at the start of the CYC infusion followed by 200 mg orally at 4 and 8 h.

Discontinuation

The criteria for treatment discontinuation included any of the following: leucopenia (white cell count <2,000 per mm3), thrombocytopenia (platelet count <50,000 per mm3), a hemoglobin concentration of less than 8 g/dl, an increase in the serum creatinine concentration (by more than 0.6 mg/dl, regardless the serum anti-double-stranded DNA antibody (anti-ds DNA) titer) or doubling of the serum creatinine concentration , fall in C3 concentration after 2 weeks of treatment, failure of serum anti-ds DNA antibody to fall or C3 concentration to rise after 4 weeks or life-threatening complications such as organic brain syndrome or serious infection. Of note, all the study’s subjects reached the end of this study and none was withdrawn from it or lost from evidence.

Follow-up

The duration of the study was ∼1 year. Patients were evaluated weekly for 4 weeks, then every other week for 8 weeks and monthly thereafter. At each follow-up visit, we evaluated them for clinical manifestations of LN and for any adverse effects of therapy. Blood pressure was measured, urinalysis was performed, renal and liver functions were evaluated, and serum anti-ds DNA antibodies as well as serum C3 were measured. Urinary protein excretion was measured in 24 h collections obtained before initiation and at the end of the study.

Anti-hypertensives

Hypertension was treated with calcium-channel blocker, and beta-blocker was added if necessary. The target systolic blood pressure was less than 130 mmHg and the target diastolic pressure less than 80 mmHg for those with proteinuria of 1 g/day or less [17 ] and to a level as low as 125/75 for patients with proteinuria in excess of 1 g/day [18]. Angiotensin converting enzyme inhibitors and angiotensin II–receptor antagonists were not used because of their possible effects on urinary protein excretion and on renal function.

End-points

The primary end point of the study was treatment failure. It was defined as urinary protein excretion that remained at or above 3 g per 24 h, and/or doubling of serum creatinine or severe flare that was resistant to increased glucocorticoid dose.

The secondary end points of the study were renal relapse, as defined by a doubling of the urinary protein excretion or by an increase in the serum creatinine level by 50% or more for more than 1 month.

Criteria for remission

Complete clinical remission was arbitrarily defined as the complete absence of clinical and laboratory evidence of disease activity for at least 3 months. Partial remission was defined as clear evidence of lowered disease activity with at least 50% improvement in laboratory parameters for at least 3 months. Patients who did not meet these criteria were considered as having treatment failure.

Assessment of renal biopsy samples

Kidney biopsy specimens were assessed by light and immunofluorescence studies. Treatment with CYC and consequently inclusion in the study was decided according to the biopsy classification for those with WHO class IV LN [16]. The pathologic changes were seen in more than 50% of the glomeruli present in the examined renal biopsy samples. It was mainly in mesangial and subendothelial distribution and was frequently associated with crescent formation and some necrosis. We detected immunoglobumin G (IgG), IgA, and IgM as well as complement three in the pathologic deposits. Briefly, the activity index (maximum score 24) represents the sum of the scores of glomerular hyperactive and active lesions. Hyperactive lesions (hematoxylin bodies, necrosis, circumferential crescents) and active lesions (endocapillary proliferation, wire-loops, hyalin thrombi and acute tubulointerstitial lesions) were scored on a scale of 0–3, where 0 = absent, 1 = mild, 2 = moderate, and 3 = severe. The chronicity index (maximum score 12) was derived by summing the glomerular obsolescence score, fibrous crescents and tubulointerstitial fibrosis. Glomerular obsolescence was scored on a scale of 0–3, where 0 = absent, 1 = 1–29% of the glomeruli, 2 = 30–59% of the glomeruli, and 3 = 60% of the glomeruli. The extent of vascular sclerosis and tubulointerstitial fibrosis was scored on a scale of 0–3, where 0 = absent, 1 = small, 2 = large, and 3 = diffuse.

Statistics

Variables are given as mean ± standard deviation (SD) unless otherwise stated. To avoid bias from excluding cases with missing values in our analysis, we imputed them as the average of values occurring immediately before and after the missing ones. These values were 2.9 g/dl for serum albumin and 47 for anti-ds DNA. Chi-squared test was used to compare the prevalence of non-parametric variables while differences between variables were analyzed by paired Student’s t-test, multivariable analysis or Mann-Whitney test as deemed appropriate. The relation between measurements was assessed by Spearman’s correlation for binary variables. Logistic regression, by general linear model, is applied to detect potential risk factors for poorer response to either regimen of CYC treatment on renal outcome such serum creatinine deterioration and nephritis remission. A P of <0.05 was considered statistically significant. All analysis was performed using the Statistical Package for Social Science (SPSS, Chicago) version 10.0 for windows.

Results

Baseline data

The majority of the patients in both groups were females. They constituted 90% in the LD-CYC group and 84.6% of the patients in the HD-CYC group. Both groups were also comparable regarding their age.

Lupus nephritis in HD-CYC group of patients was of longer, yet non-significant, disease duration (65 ± 38.9 vs 40.9 ± 62.7 months, P = NS) and of higher chronicity index (3.6 ± 1.7 vs 2.0 ± 1.1, P = 0.012) compared to the LD-CYC group, with non-significant difference in the biopsy activity index (10.1 ± 4.2 vs 9 ± 5.4, P = NS). Further, the degree of proteinuria was more severe in the HD-CYC patients (5 ± 3.2 vs 3.09 ± 1.2 g/day, P = 0.023) with lower serum albumin levels (1.8 ± 1.1 vs 2.3 ± 0.5 mg/dl, P = 0.05) compared to the LD-CYC group (Table 1).

Table 1 Patients’ characteristics in low- and high-dose groups
Full size table

Markers of SLE activity such as anti-double stranded DNA antibodies (60.1 ± 22 vs 65.9 ± 43 U/l), complement 3 (C3) (0.35 ± 0.27 vs 0.41 ± 0.32 g/l), hemoglobin level (9.7 ± 2.2 vs 10.2 ± 2.4 g/dl), and white blood cell count (WBCs 9.1 ± 5.3 vs 5.2 ± 0.1×103/mm3) were comparable (P = NS) in both groups of CYC regimens. The only marker that was of significant difference was the erythrocyte sedimentation rate (ESR). It was significantly higher in HD-CYC compared to LD-CYC group of patients (96 ± 56 vs 39.7 ± 19 mm, P = 0.033).

Of note, using multivariate analysis for variables such as patient’s age, disease duration, serum creatinine levels and anti-ds DNA antibody showed a statistical level that persisted at the non-significance side of difference.

By the end of the study

There was no loss of either patients or renal survival by the end of the study in both groups. The patients were all alive and none experienced end-stage renal disease. Gastrointestinal side effects which were severe enough to need extra medication have not developed in any patient. New onset hematuria was observed in only one (1/26) patient of the HD-CYC group of patients. None of our patient complained of menstrual abnormalities. Proteinuria was reduced in LD-CYC patients to a significantly lower level than its base-line one (3.09 ± 1.2 vs 2.1 ± 1.6 g/day, P = 0.035) with consequently a significant rise in the serum albumin level (2.3 ± 0.5 vs 2.8 ± 0.6 g/dl, P = 0.006). In the HD-CYC, although the reduction in the severity of proteinuria was not significant (5 ± 3.2 vs 2.9 ± 1.5, P = NS) yet the resulting elevation of serum albumin was significant (1.8 ± 1.1 vs 2.9 ± 0.7, P = 0.0001). Overall, the changes in serum creatinine level were insignificant in both groups; the LD-CYC (1.3 ± 0.7 vs 1.3 ± 0.9, P = NS) as well as the HD-CYC (1.5 ± 0.7 vs 1.3 ± 0.8, P = NS).

Of the observed SLE activity markers that included anti-ds DNA, ESR, hemoglobin and WBCs, only the anti-ds DNA was the one that was significantly reduced in both LD-CYC (60.1 ± 22 vs 46.8 ± 19 U/l, P = 0.026) and HD-CYC patients (65.9 ± 43 vs 30.8 ± 21 U/l, P = 0.015).

For patients in the LD-CYC regimen, the prevalence of serum creatinine improvement, proteinuria reduction, and disease activity betterment was significant (Chi-squared P = 0.025, 0.002, and 0.025, respectively). However, in the HD-CYC group of therapy participants, the only marker of decreased disease activity was implied from significant prevalence of anti-ds DNA level reduction (Chi-squared P = 0.009). Using Mann-Whitney, the renal outcome in the form of reduction of proteinuria severity was better in the LD-CYC compared to HD-CYC regimen (P = 0.043).

Correlates of low-dose CYC

The LD-CYC regimen correlated significantly with improved renal outcome as inferred from reduced severity of proteinuria (r = 0.301, P = 0.042) as well as with betterment of SLE activity as implied from the SLEDAI score (r = 0.512, P = 0.0001). Unfortunately, the percentage of complications was higher in the LD-CYC (35%) compared to the HD-CYC regimen (15.4%), although the difference was statistically not significant. The most clinically significant complications took the form of chest infection and deep vein thrombosis in the LD-CYC group while in the HD-CYC group it was infection only.

Renal relapse was not reported in the LD-CYC group of patients, while in the HD-CYC group it was noticed in 3 of its 26 (11.5%) patients (nephritic in type). Further, treatment failure was seen in 1/20 patients (5%) of the LD-CYC group and in 1/26 (3.8%) of the HD-CYC group and was in the form of doubling of proteinuria. Importantly, with Mann-Whitney testing there was no statistical difference between the two groups regarding either parameters of renal outcome.

Predictors of unfavorable response

As seen in Table 2, multivariate regression analysis in a model that included as dependent variables female gender, regimen of CYC therapy, SLEDAI score and anti-ds DNA titer, showed that the monthly regimens of CYC as well as the high anti-ds DNA titer were unfavorable for improvement of serum creatinine in patients with LN.

Table 2 Predictors of poor renal outcome response to CYC therapy
Full size table

Discussion

In this study, we have used CYC as a remission-inducing drug for severe LN, as data from the NIH had shown that patients with LN who were treated with glucocorticoids and a prolonged course of cytotoxic agents have better long-term renal function than those treated with glucocorticoids alone [19]. Also, intermittent pulse CYC is widely used in renal [20] and major extrarenal complications of SLE [21], and its use in high doses without stem cell rescue led to long-lasting remission in a substantial number of patients, despite a history of resistance to multiple immunosuppressive regimens in the past [22]. The CYC regimens we applied here were the monthly HD-CYC and the fortnightly LD-CYC as previously described in the Euro-Lupus Nephritis Trial [23].

In our study, no difference was observed in both renal and patients' survival. None of our patients died and end-stage renal disease (ESRD) was not noted in any of them, while in the Euro-Lupus Nephritis Trial [23] two deaths were reported in the low-dose group. A difference in follow-up duration between the two studies may explain this inconsistency. Also, our results disagree with that of Mok and associates [24] who reported ESRD in 6.5% of their patients.

Biochemical markers of LN such as proteinuria and serum creatinine were both comparable in both regimens of therapy. However, proteinuria was remarkably reduced in LD-CYC regimen and not in the HD-CYC one. A higher level of proteinuria at study commencement may explain this difference in treatment response between the two regimens. Our results are in concordance with those reported earlier by the Euro-Lupus Nephritis Trial (ELNT) [23], and support the notion that Cyclophosphamide plus steroids reduced the risk of doubling of serum creatinine [25].

Our results are in contrast with those reported by Williams et al. [24] who concluded that intravenous pulse cyclophosphamide for SLE and diffuse proliferative glomerulonephritis is an ineffective form of treatment. Differences in genetic factors between the Egyptian patients included in our study and the Jamaican patients who have African descent in that study [26] might be the explanation. It was previously hypothesized that SLE has an important genetic background [27], and it can be assumed that difference in the genetics of different populations may be responsible for the difference in clinical presentation and response to different therapeutic modalities.

The overall disease activity, as inferred from the SLEDAI score, was improved in the whole study cohort, although this was more substantial with LD-CYC therapy than with HD-CYC. Cyclophosphamide is capable of decreasing disease severity in SLE patients with severe form of LN because of its efficacy as noticed by Petri [28]. Longer disease duration at our study commencement might be the reason for such difference in response between the two regimens of CYC therapy.

We have not noted renal relapse in the LD-CYC group of patients, although it was observed in 11.5% of patients given the HD-CYC. Cyclophosphamide was substantially superior to azathioprine with regard renal relapses, as was previously reported in the Dutch LN study which enrolled 87 proliferative LN patients over a 2-year follow-up period [29].

The higher relapse rate in the HD-CYC regimen patients could be explained by the higher activity index in their renal biopsy as suggested by Mok et al. [30].

Although Sidiropoulos et al. [31] concluded that renal flares are an important feature of the natural history of LN and that a substantial proportion of SLE patients (27–66%) will flare despite the use of immunosuppressive therapy, we have not noticed a significant incidence of treatment failure in this study. Similarly, a favorable short-term result was observed in children following intravenous pulse CYC for severe LN [32].

In LN, the long-term renal outcome represents the additive effect of multiple confounding factors, including the efficacy of immunosuppressive therapy, pre-existing renal reserve and scarring, the number, severity and treatment of subsequent relapses of nephritis, and the adequacy of blood-pressure control. Importantly, no remarkable difference in terms of treatment failure was seen with both regimens of CYC therapy, and this was the case with the ELNET [23]. It was at much lower percentages in the study under discussion.

Predictors of less favorable response to CYC therapy in our study were: the HD regimen as well as the anti-ds DNA titer at the start of treatment. Mok et al. [30] suggested that increasing serum creatinine level, higher histologic activity scores on renal biopsy, and a lower dose of CYC are the predictors of treatment failure.

Unfortunately, the percentage of complications was higher in the LD-CYC (35%) compared to the HD-CYC regimen (15.4%), although the difference was statistically not significant. The most clinically significant complications took the form of chest infection and deep vein thrombosis. Patients with SLE may have intrinsically increased risk for infection that is augmented by immunosuppressive therapies, particularly CYC [33].

Our findings are in agreement with the study of Calguneri et al. [34] who observed no serious infection, but not in agreement with what has been reported in the ELNT [23], since in this study episodes of severe infection were more common among the high dose group, although not statistically significant. This could be explained by the difference in follow-up period, since in ELNT it was extended for a median period of 73 months which is much longer than in our study. We have not experienced serious infection in any of our patients, as this risk is enhanced by CYC-induced reductions in the total WBC count <3,000 cells/μl and by sequential IV and oral CYC therapy as previously described [35], which was not the case in our study.

Furthermore, long term follow up is essential to observe differences that can develop much later [36] such as premature ovarian failure, because the cumulative dose rather than the route of CYC administration determines the ovarian toxicity [37]. This was reported to be <50% in women below 30 years and 60% between 30 and 40 years as reported by the prospective randomized study on protection against gonadal toxicity (PREGO-Study) [38].

The many advantages of the low dose regimen should be emphasized. A 500-mg pulse of CYC can be administered, with excellent immediate tolerance, on an outpatient basis as a 30-min infusion, without the need for either IV anti-emetics or forced hydration. Also, the costs of therapy are less as well as the cumulative dose-dependent.

Our study, however, has some limitations, including the small sample size which reduces the generalizability of our findings, and our short-term observations that cannot detect long-term side effects, such as premature ovarian failure, which is of concern since most of our patients are females in their child-bearing period.

Conclusion

In summary, a remission-inducing regimen of low-dose IV CYC given in an outpatient clinic practice achieves clinical results comparable to those obtained with a high-dose regimen in Egyptian patients with proliferative LN without serious complications.