Objective: To explore the optimal dose of L-arginine to prevent acute high-dose (HD)-PDD nephrotoxicity. Methods: 128 cases using PDD with the dosage of 100 mg/m 2 within two days (D1, 2) in combination with L-arginine were randomly divided into 3 groups of A, B and C. The dosages of Larginine in the 3 groups were 5 g/(m 2 ·d), 10 g/(m 2 ·d) and 15 g/(m 2 ·d), respectively. Each patient received 2 cycles chemotherapy to form self control: 1 cycle combined with L-arginine, while 1 cycle chemotherapy alone. β2-MG in urine, BUN, Cr and uric acid in blood were detected just 24 h before and after using PDD. The changes of each index in the three groups were observed in the presence or absence, and the therapeutic effects were compared among the three groups. Results: There was no significant difference in BUN, Cr and uric acid in blood before and after chemotherapy in the presence or absence, showing these indexes could not be used as markers of early acute nephrotoxicity. Urine β2-MG values in the presence and absence were 0.9120±0.6618 vs 1.5167±0.7908 (P <0.05), 0.5404±0.5810 vs 1.4616±0.8120 (P <0.01), 0.4998±0.6210 vs 1.5210±0.7710 (P <0.01) in the groups A, B and C respectively. The excellent effective rate and total effective rate in groups A, B and C were 40.9% and 59.1%, 68.2% and 90.9%, and 77.5% and 97.5%, respectively. There was significant difference in the excellent effective rate and total effective rate between groups A and B, but not between groups B and C. Conclusion: The optimal dose of Larginine to prevent acute HD-PDD nephrotoxicity is 10 g/(m 2 ·d). Increased dosage can't improve the effect accordingly.