Recent clinical trials have provided evidence for the efficacy of low-dose intravenous cyclophosphamide and mycophenolate mofetil as induction treatment for patients with proliferative lupus nephritis in comparative trials with standard-dose intravenous cyclophosphamide. Trials of maintenance treatments have had more variable results, but suggest that the efficacy of mycophenolate mofetil may be similar to that of quarterly standard-dose intravenous cyclophosphamide and somewhat more efficacious than azathioprine. Differential responses to mycophenolate mofetil based on ethnicity suggest that it may be more effective in black and Hispanic patients. Rituximab was not efficacious as an adjunct to induction treatment with mycophenolate mofetil.
Low-dose, short-course intravenous cyclophosphamide and mycophenolate mofetil have efficacy similar to that of standard high-dose intravenous cyclophosphamide for induction treatment of proliferative lupus nephritis, but these findings may depend on the ethnicity of the patients studied.
Mycophenolate mofetil is likely more effective than azathioprine as maintenance treatment of proliferative lupus nephritis.
Alternatives to high-dose intravenous cyclophosphamide have a lower risk of ovarian failure, but risks of other toxicities are not different.
Lack of recent improvement in long-term renal outcomes suggests that new treatment approaches are needed.
Lupus nephritis (LN) is an autoimmune-mediated glomerulonephritis and tubulointerstitial disease that is among the most common and serious manifestations of systemic lupus erythematosus (SLE). LN develops in 30% to 60% of patients with SLE, most often occurring early in the course of SLE. Common accompaniments include anemia, hypertension, hypocomplementemia, and autoantibodies to double-stranded DNA and the extractable nuclear antigens ribonucleoprotein and Sm. Rarely, LN may be the presenting manifestation of SLE in an otherwise asymptomatic patient. Persons of black African or Hispanic ancestry are 2 to 3 times more likely to develop LN, and tend to have more severe disease, than white counterparts. Although improvements in treatment have transformed severe LN from a near universally fatal condition to a manageable chronic disease, patients with LN are at more than twice the risk of death than patients without LN, and those with chronic kidney disease have more than 3 times the risk of death. Both the high prevalence of LN and the intensity of clinical care required by patients with LN make it the most costly manifestation of SLE.
Course of Lupus Nephritis
LN is largely an asymptomatic condition, and requires active screening to be detected. Clinical and laboratory features reflect the type of renal histologic lesions present in a given patient. Proteinuria, microscopic hematuria, urinary casts, and hypertension (occasionally severe) are common in patients with proliferative LN, and renal insufficiency may develop. Nephrotic syndrome with peripheral edema, serous effusions, wasting, and hypercoagulability can occur in patients with membranous or proliferative LN. LN can present with acute renal failure, or can be rapidly progressive as a result of severe glomerular inflammation, cellular crescents, and fibrinoid necrosis, but these presentations are rare. In most patients LN is a subacute or chronic condition, with treatment-induced remissions and spontaneous relapses being fairly common.
Prognosis of LN is closely related to the renal histologic class, and the degree of active inflammation and chronic damage incurred. Patients with mesangial LN have only mild laboratory abnormalities and very low risk of chronic kidney disease. Patients with membranous LN often have morbidity related to nephrosis but more than 90% achieve remission, and, excluding transitions to a proliferative subtype, fewer than 10% of patients progress to end-stage renal disease after 15 years. In patients with proliferative LN the prognosis is poorer in those with diffuse than with focal glomerular involvement, and in those with crescents and extensive chronic damage. End-stage renal disease develops in 25% to 40% of patients with proliferative LN after 15 years, predominantly among those with diffuse involvement. Advanced sclerosing glomerulonephritis represents an end-stage inactive lesion with more than 90% of glomeruli showing global sclerosis, with a high risk of end-stage renal disease.
Additional poor prognostic factors include delayed initiation of immunosuppressive treatment, incomplete response to induction therapy, occurrence of nephritic relapses, and poor control of hypertension.
Goals of Treatment
The goals of treatment of LN are to prevent end-stage renal disease and to decrease the risk of chronic kidney disease and its atherosclerotic and metabolic consequences. Treatment can also facilitate control of blood pressure and help avoid the vascular complications of hypertension. In addition, treatment of nephrotic syndrome can lessen the morbidity associated with fluid overload, hypoalbuminemia, and the risk of thrombosis.
It is important to recognize that, apart from proteinuria, these goals are not typically measured as outcomes in controlled clinical trials in LN. Chronic kidney disease and end-stage renal disease are often late manifestations, and impractical to measure in short-term clinical trials. Even 5 years of follow-up may not be sufficient to see differences in risks of these outcomes. The large functional reserve of kidneys makes measures of renal function insensitive to even moderate degrees of dysfunction. Doubling of the serum creatinine level has questionable validity for predicting long-term renal outcomes in patients with good renal function at baseline. Recent clinical trials have used improvement or normalization of proteinuria, improvement in urine sediment, and stabilization or improvement in serum creatinine levels as end points. Although the rationale that reductions in signs of active inflammation should translate into lower risks of kidney damage and dysfunction in the future is sound, the definitions of response in clinical trials have not been validated as good surrogate markers of late renal outcomes.
Traditional Treatment Approach
A shift from using glucocorticoids alone to using cytotoxic drugs together with glucocorticoids to treat serious LN occurred in the 1970s and 1980s with evidence from clinical trials of improved renal outcomes from combination therapy. A central role for cyclophosphamide (CYC) was solidified based on trials conducted at the National Institutes of Health (NIH), which showed numerically better outcomes after 10 years among patients with active (mostly proliferative) LN treated with intravenous CYC in comparisons with those treated with azathioprine or prednisone alone. Trials showing that prolonged treatment with CYC resulted in fewer nephritic flares and lower risks of end-stage renal disease ushered in the notion of induction and maintenance phases of treatment of LN, and established the “NIH regimen” of monthly intravenous CYC for 6 months followed by quarterly infusions for 24 additional months as acceptable for the treatment of severe LN. However, treatment-related complications including infection, ovarian failure, hemorrhagic cystitis, bladder cancer, and treatment-related mortality emerged as concerns.
The need for treatments with less toxicity and equal or better efficacy has motivated studies of new medications, and new methods of administering CYC. This article reviews recent clinical trials of new medications or treatment strategies for LN. The review is limited to medications that are currently available in the clinic because these can be used by clinicians today, and is further limited to published full-length studies of adults with LN.
Trials of alternative dosing of cyclophosphamide
Euro-Lupus Nephritis Trial
This randomized controlled trial compared a short-course regimen of low-dose intravenous CYC with a longer regimen of higher-dose intravenous CYC as induction treatment of active proliferative LN. The experimental group was given 500 mg of intravenous CYC every 2 weeks for 6 doses, whereas the control group was treated with 0.5 g/m 2 body surface area of intravenous CYC monthly for 6 doses, followed by 2 additional doses at 9 months and 12 months. In the control group the dose of CYC was sequentially increased, based on nadir peripheral blood leukocyte counts, to a maximum of 1500 mg. Both groups received pulse methylprednisolone at the start of treatment, followed by tapering doses of prednisone, and received azathioprine 2 mg/kg daily for maintenance treatment at the end of the CYC course.
The primary end point was treatment failure, defined as either persistent renal insufficiency or nephrotic syndrome, glucocorticoid-resistant flare, or doubling of the serum creatinine level. End points were assessed in a time-to-event analysis with a median follow-up of 41 months. The occurrence of the primary end point did not differ between treatment groups ( Table 1 ). The probability of renal remission (71% vs 54%) and renal flare (27% vs 29%) over time also did not differ between the low-dose and high-dose groups. Two patients in the low-dose group died and 1 progressed to end-stage renal disease while 2 patients in the high-dose group progressed to end-stage renal disease. Severe infections were more common in the high-dose group while the frequencies of nonserious infections, leukopenia, and ovarian failure were similar in the 2 treatment groups.
|Study, Ref. Year||Design||Intervention||N||Primary End Point a||Outcomes (%)|
|Euro-Lupus, 2002||Superiority||CYC 0.5 g every 2 wk × 6||CYC 0.5 g/m 2 monthly × 6, then quarterly × 2||44||46||Treatment failure||16||20|
|Petri et al, 2010||Superiority||CYC 50 mg/kg daily × 4||CYC 0.75 g/m 2 monthly × 6, then quarterly for 2 y||10||12||Complete response||20||64|
|Sabry et al, 2009||Superiority||CYC 0.5 g every 2 wk × 6||CYC 0.5 g/m 2 monthly × 6, then quarterly × 2||20||26||Treatment failure||5||3.4|
|Mitwalli et al, 2011||Superiority||CYC 5 mg/kg monthly × 6, then every 2 mo × 18||CYC 10 mg/kg monthly × 6, then every 2 mo × 6||44||73||Patient and renal survival||100 at 6 mo, no results for long term||100 at 6 mo, no results for long term|
The investigators concluded that the low-dose CYC regimen was comparable in efficacy with the higher-dose regimen in patients with proliferative LN, and that a high-dose regimen was not required as treatment for all patients with proliferative LN. However, the investigators noted that although most patients had diffuse proliferative LN, the majority did not have severe disease, with 22% having an elevated serum creatinine and 28% with nephrotic syndrome at study entry. The trial included few ethnic minorities. The results may not be generalizable to patients with more severe LN. Although the low-dose regimen was anticipated to have less toxicity, and was associated with fewer serious infections, other adverse events were similar between the treatment groups.
Several points are worth noting. Although presented as a superiority trial, an a priori power calculation was not apparent, and consequently the interpretation of statistical differences is difficult. A larger sample would be needed to convincingly demonstrate noninferiority between the treatments. The primary end point consisted of an absence of worsening for most patients, rather than demonstration of improvement. Therefore the possibility of persistent LN activity among “responders” also raises questions about how the results should be interpreted. Renal outcomes in both groups were excellent at 10 years, suggesting that residual activity did not affect long-term outcomes, although most patients still required immunosuppressive treatment. The high-dose CYC regimen was shorter than the typical NIH regimen, which somewhat complicates comparisons among studies. However, the results of this trial suggest that the intensity of immunosuppression can be tailored to the severity of disease among patients with proliferative LN.
This randomized trial compared the efficacy of a myeloablative regimen of intravenous CYC with that of monthly intravenous CYC followed by quarterly infusions for 3 years, following the NIH regimen. The rationale for testing myeloablative treatment was to attempt to eliminate autoreactive lymphocytes and reset the immune system. Patients with treatment-refractory SLE were eligible, and although the study was not limited to patients with LN, data on those with LN were reported separately. Complete response required a normal serum creatinine, normal creatinine clearance, normal urinary sediment, and proteinuria less than 500 mg/d. Complete responses were rare at 6 months, but were more common among those treated with the NIH regimen than with the myeloablative regimen at 30 months (see Table 1 ).
Sabry and colleagues replicated the Euro-Lupus Nephritis Trial in a 12-month randomized trial in 46 patients with proliferative LN. No differences in outcomes were observed, but the power of the study was limited. Mitwalli and colleagues found no difference in patient and renal survival between patients randomized to short-term high-dose intravenous CYC and a longer course of lower-dose CYC.
Trials of azathioprine, cyclosporine, or tacrolimus as alternatives to cyclophosphamide as induction treatment
Dutch Working Party Trial
This randomized trial compared treatment with azathioprine, 2 mg/kg daily along with 3 courses of pulse methylprednisolone, with the NIH regimen of intravenous CYC in patients with proliferative LN. At entry, 56% of patients had renal impairment and 53% had nephrotic-range proteinuria. Over a median follow-up of 5.7 years, 4 times as many patients in the azathioprine group as in the CYC group met the primary end point of doubling of serum creatinine ( Table 2 ). Although the likelihood of complete remission was similar in the 2 treatment groups over the first 2 years of the study, relapses were higher in the azathioprine group (7.1 per 100 patient-years vs 1.1 per 100 patient-years). Infections, particularly herpes zoster, were more common in the azathioprine group. On extended follow-up to a median of 9.6 years, the azathioprine group had numerically higher frequency of doubling of serum creatinine (16% vs 8%), mortality (16% vs 10%), and renal relapses (38% vs 10%). Although the reports emphasized the lack of significant differences in end points other than relapses, the power of the study to detect differences in other outcomes was limited, and the weight of evidence suggests greater efficacy for intravenous CYC than for azathioprine.
|Study, Ref. Year||Design||Intervention||N||Primary End Point a||Outcomes (%)|
|Dutch Working Party, 2006||Superiority||Azathioprine 2 mg/kg daily and pulse methylprednisolone × 9||CYC 0.75 g/m 2 monthly × 6, then quarterly for 21 mo||37||50||Doubling of serum creatinine||16||4|
|Chen et al, 2011||Noninferiority||Tacrolimus 0.05 mg/kg daily, titrated to trough levels||CYC 0.75 g/m 2 monthly × 6||42||39||Complete remission||52||38|
|Li et al, 2012||Superiority||Tacrolimus 0.08–0.1 mg/kg daily, titrated to trough levels||CYC 0.5–0.75 g/m 2 monthly × 6||20||20||Complete or partial remission||75||60|
|Cyclofa-Lune, 2010||Superiority||4–5 mg/kg daily × 9 mo||CYC 10 mg/kg × 8 over 9 mo||19||21||Complete remission||26||24|
Trials of Calcineurin Inhibitors
Chen and colleagues compared outcomes at 6 months between patients with active LN (89% proliferative; 11% membranous) treated with either tacrolimus (0.05 mg/kg daily, titrated to serum trough levels of 5–10 ng/mL) or intravenous CYC by the NIH regimen in a randomized noninferiority trial (see Table 2 ). At entry, 10% of patients had elevated serum creatinine and 43% had nephrotic-range proteinuria. Complete remission at 6 months was seen in 52% of tacrolimus-treated patients and 38% of CYC-treated patients, with most failures attributable to persistent proteinuria. The study enrolled only one-half of the projected sample size needed to adequately test noninferiority. In a smaller trial of similar design, Li and colleagues reported complete or partial remission in 75% of tacrolimus-treated patients and 60% of CYC-treated patients at 6 months. Responses after 9 months of treatment were also similar in a small trial that compared cyclosporine with intravenous CYC. Notable in these studies is the rapid improvement in proteinuria and generally good responses regardless of treatment arm, but their small sizes limit the strength of any comparisons.
Trials of mycophenolate mofetil as induction treatment
Seven trials have compared mycophenolate mofetil (MMF) with CYC as induction treatment for LN ( Table 3 ). Five trials included fewer than 50 patients, and 4 of these 5 trials reported similar treatment effects at 6 months. Bao and colleagues studied combined treated with MMF and tacrolimus versus intravenous CYC in patients with LN of mixed class V and IV, and found that combined treatment led to more complete remissions. The 2 largest studies are reviewed in more detail.