Is cisplatin a human carcinogen?

Abstract

Traditionally, cisplatin has not been regarded among chemotherapeutic drugs as a carcinogenic risk to humans because it is not a classical alkylating agent. A review of recently published experimental data indicates that cisplatin is mutagenic, clastogenic, capable of inducing cell transformation, able to act as an initiator in classical mouse skin initiation/promotion experiments, and carcinogenic in laboratory animals. Notably, it causes myeloid leukemia in BD IX rats. These observations demonstrate that cisplatin should be considered a potent carcinogen in experimental settings. A review of the literature identified 65 instances of subsequent cancer in patients receiving cisplatin-based chemotherapy for an initial malignancy. The majority of second cancers were acute nonlymphocytic leukemias or myelodysplasia. In only one instance was cisplatin the sole antineoplastic drug given to patients. The routine use of cisplatin in conjunction with other known or suspected human carcinogens makes it impossible to use these anecdotal reports as a basis for assessing cisplatin's carcinogenicity in humans. Two quantitative epidemiologic studies have addressed this question: One suggested that the combination of cisplatin and doxorubicin is leukemogenic in humans, while the other implicated etoposide rather than cisplatin as the leukemogen. Formal epidemiologic studies of appropriate cohorts of cisplatin-treated patients are needed to resolve the question of its carcinogenicity in humans.