There have been surprisingly few controlled studies of the central nervous system actions of cytokines in humans, although many clinical trials have briefly reported side effects of these drugs. Clinical trials using cytokines have chiefly involved patients with cancer, often at an advanced stage, and prospective formal neurologic assessment has been infrequent. In patients with malignancy, a host of variables can confound neurologic evaluation, including nutritional status, metastatic deposits, metabolic derangements, effects of former treatments, and infection. The effects of the cytokines on the central nervous system are multiple and limit their clinical use. Careful prospective clinical study of these agents is essential and may yield insights into their functions in the CNS and increase their efficacy as clinical tools.