Ten RCTs were included in the review (n=891 patients). Jadad scores ranged from 1 to 3. Three trials scored 3 points, four trials scored 2 points, and three trials scored 1 point. Follow-up ranged from six to 12 months for induction treatments, 39 to 72 months for maintenance treatments, and 12 to 60 months for low-dose versus high-dose cyclophosphamide treatments.
Induction treatments (six RCTs, 662 patients)
There was no statistically significant difference between mycophenolate mofetil and cyclophosphamide induction regimes for complete remission (five RCTs), partial remission (four RCTs), overall response rate (six RCTs), end-stage renal disease (two RCTs), death (three RCTs), herpes infection (four RCTs), and any other infection (four RCTs). Mycophenolate mofetil was associated with a non-significant trend towards a lower risk of amenorrhoea (RR 0.166, 95% CI 0.020 to 1.343; two RCTs) and leukopenia (RR 0.412, 95% CI 0.149 to 1.135; four RCTs) than cyclophosphamide. Heterogeneity was found for analyses of complete remission (I2=54%), response rate (I2=49%) and other infections (I2=79%). The authors stated in the text that heterogeneity was found for the analysis of amenorrhoea, but the table reported an I2 value of 0%. Egger’s test showed evidence of publication bias for response to induction therapy.
Maintenance treatment (two RCTs, 91 patients)
There was no statistically significant difference between mycophenolate mofetil and azathioprine maintenance regimes in response (two RCTs) or end-stage renal disease (two RCTs). Mycophenolate mofetil and azathioprine regimens were associated with significantly higher six-year event-free survival rates for death (p=0.05) and renal failure (p=0.009) than cyclophosphamide regimes. The mycophenolate mofetil group had a higher relapse-free survival than the cyclophosphamide group (p=0.002) in one trial (32 patients).
Low-dose versus high-dose cyclophosphamide (two RCTs, 138 patients)
The cumulative doses of cyclophosphamide differed between the two included trials. Low-dose cyclophosphamide regimes were associated with a significantly reduced relapse rate (RR 0.465, 95% CI 0.261 to 0.830), a significantly reduced infection rate (RR 0.688, 95% CI 0.523 to 0.905) and a reduced treatment failure rate of marginally significance (RR 0.451, 95% CI 0.202 to 1.009). Heterogeneity was found for the analysis of infection rate (I2=54%).