Abstract: Objective To analyze the characteristics and regularity of occurrence of drug-induced kidney injury, and to explore the risk signals of kidney injury caused by anti-infective drugs so as to provide reference for safe clinical use.Methods Valid reports of kidney injury collected between 2016 and 2018 in the ADR database of the PLA were retrospectively analyzed, and risk signals mining was carried out using four data mining methods: the reporting odds ratio, proportional reporting ratio, combination x² test-PRR measure used by the Medicines and Healthcare Products Regulatory Agency, and Bayesian Confidence Propagation Neural Network.Results Among the 968 cases of drug-induced kidney injury, 633 cases (65.39%) involved males and 335 cases (34.61%) females, with an average age of 58.60±19.55 years. The most-frequently reported drugs were anti-infective drugs in 269 cases (27.79%), and the top three drugs were vancomycin in 43 cases (4.44%), mannitol in 30 cases (3.10%), and iodixanol in 29 cases (3.00%). After data mining, eleven types of anti-infective drugs, including vancomycin, generated positive risk signals of kidney injury.Conclusion There are many kinds of drugs involved in drug-induced kidney injury, among which the frequency of anti-infective drugs is the highest. In clinical application, the awareness of relevant drug vigilance should be enhanced and the use of drugs with high risk of nephrotoxicity standardized.

Key words: drug-induced kidney injury, adverse drug reaction, pharmacovigilance, data mining, anti-infective drugs