Abstract
INTRODUCTION: Both mycophenolate mofetil (MMF) and intravenous cyclophosphamide (CYC) have been recommended in the induction therapy of lupus nephritis (LN) for years; nevertheless, their effectiveness and safety in a real-world setting are extremely lacking. Therefore, we decided to conduct this real-world study. METHODS: A total of 195 Chinese patients with LN who were initially treated with MMF (n = 98), or intravenous CYC (n = 97) as induction therapy were enrolled. All of the patients were followed up to 12 months. Complete renal remission (CRR) was defined as 24-h urinary protein (24 h-UTP) < 0.5 g, and partial renal remission (PRR) was defined as ≥ 50% reduction in 24 h-UTP to the subnephrotic level, however > 0.5 g, both with a change of serum creatinine (SCr) within 10% from baseline. The proportions of CRR, PRR, and total renal remission (TRR), as well as adverse events, were compared by Chi-square test and Kaplan-Meier analysis (log-rank test). Inverse probability of treatment weighting (IPTW) was used for propensity score matching and multivariable logistic regression analyses were employed. RESULTS: The cumulative proportion of TRR in 6 months (79.4 vs. 63.8%, p = 0.026) and CRR in 12 months (72.8 vs. 57.6%, p = 0.049) in MMF group were significantly higher than CYC group, and the above conclusions were further confirmed by IPTW. The proportions of PRR, CRR, and TRR at other time points were equivalent between two groups. Further subgroup analysis in 111 patients with biopsy-proven III-V LN also showed a significantly higher proportion of TRR at 6 months in the MMF group than in the CYC group (78.3 vs. 56.9%, p = 0.026). In the Kaplan-Meier analysis and after IPTW, the MMF group showed better TRR and CRR responses than CYC group in 12 months. Multivariable logistic regression analyses revealed that MMF use was the only predictor of CRR (HR 2.12, 95% CI 1.90-4.09, p = 0.026), while low complement level was also a predictor, albeit risk was reduced (HR 0.38, 95% CI 0.17-0.86, p = 0.019). Moreover, compared to the CYC group, MMF group patients were more likely to have significantly lower SCr (μmol/l) [72.5 (62.5, 86.5) vs. 79.0 (71.1, 97.5), p = 0.001] and daily dose of prednisone (mg/day) (15.7 ± 5.2 vs. 18.6 ± 11.3, p = 0.022) at 6 months; lower 24 h-UTP (g) [0.1 (0.1, 0.3) vs. 0.2 (0.1, 0.9), p = 0.005] and daily dose of prednisone (mg/day) (9.6 ± 3.3 vs. 11.2 ± 5.5, p = 0.023) at 12 months. Infection was the most common adverse event. Pneumonia and gastrointestinal discomfort were more frequently observed in the CYC group. CONCLUSIONS: Real-world data are a key component of the evidence supporting the effectiveness of drugs and are of interest to all stakeholders. Our comparative study demonstrated the effectiveness of MMF in LN induction therapy was at least equivalent to intravenous CYC, with superior tolerance.