IMPORTANCE Lupus nephritis (LN) is typically treated with intravenous cyclophosphamide (IVCY), which is associated with serious adverse effects. Tacrolimus may be an alternative for initial treatment of LN; however, large-scale, randomized clinical studies of tacrolimus are lacking. OBJECTIVE To assess efficacy and safety of tacrolimus vs IVCY as an initial therapy for LN in China. DESIGN, SETTING, AND PARTICIPANTS This randomized (1:1), open-label, parallel-controlled, phase 3, noninferiority clinical trial recruited patients aged 18 to 60 years with systemic lupus erythematosus and LN class III, IV, V, III+V, or IV+V primarily from outpatient settings at 35 centers in China. Inclusion criteria included body mass index of 18.5 or greater to less than 27, 24-hour urine protein of 1.5 g or greater, and serum creatinine of less than 260 mu mol/L. Of 505 patients screened, 191failed screening (163 ineligible, 25 withdrawn consent, and 3 other reasons). Overall, 314 were randomized. The first patient was enrolled March 10, 2015, and the study finished September 13, 2018. The follow-up period was 24 weeks. Data were analyzed from December 2019 to March 2020. INTERVENTIONS Oral tacrolimus (target trough level, 4-10 ng/mL) or IVCY for 24 weeks plus prednisone. MAIN OUTCOMES AND MEASURES Complete or partial response rate at week 24 (prespecified). RESULTS A total of 314 patients were randomized (158 [50.3%] to tacrolimus and 156 [49.7%] to IVCY). Overall, 299 patients (95.2%) were treated (tacrolimus group, 157 [52.5%]; IVCY group, 142 [47.5%]). Baseline demographic and clinical characteristics were generally similar between groups (mean (SD] age, 34.2 [9.5] years; 262 [87.6%) female). Tacrolimus was found to be noninferior to IVCY for LN response at week 24. There was a complete or partial response rate of 83.0% (117 of 141 patients) in the tacrolimus group and 75.0% (93 of 124 patients) in the IVCY group (difference, 7.1%; 2-sided 95% Cl, -2.7% to 16.9%; lower limit of 95% CI greater than -15%). At week 24, least-square mean change in Systemic Lupus Erythematosus Disease Activity Index score was -8.6 with tacrolimus and -6.4 with IVCY (difference, -2.2; 95% CI, -3.1to -1.3). Changes in other immune parameters and kidney function were generally similar between groups. Serious treatment-emergent adverse events (TEAEs) were reported in 29 patients in the tacrolimus group (18.5%) and 35 patients in the IVCY group (24.6%). Most common serious study drug-related TEAEs were infections (14 [8.9%) and 23 [16.2%), respectively). Seven patients in each group withdrew due to AEs. CONCLUSIONS AND RELEVANCE In this study, oral tacrolimus appeared noninferior to IVCY for initial therapy of active LN, with a more favorable safety profile than IVCY. Tacrolimus may be an alternative to IVCY as initial therapy for LN.