Background: Progressive multifocal leukoencephalopathy (PML) was reported to have developed in three patients treated with natalizumab. We conducted an evaluation to determine whether PML had developed in any other treated patients. Methods: We invited patients who had participated in clinical trials in which they received recent or long-term treatment with natalizumab for multiple sclerosis, Crohn's disease, or rheumatoid arthritis to participate. The clinical history, physical examination, brain magnetic resonance imaging (MRI), and testing of cerebrospinal fluid for JC virus DNA were used by an expert panel to evaluate patients for PML. We estimated the risk of PML in patients who completed at least a clinical examination for PML or had an MRI. Results: Of 3417 patients who had recently received natalizumab while participating in clinical trials, 3116 (91 percent) who were exposed to a mean of 17.9 monthly doses underwent evaluation for PML. Of these, 44 patients were referred to the expert panel because of clinical findings of possible PML, abnormalities on MRI, or a high plasma viral load of JC virus. No patient had detectable JC virus DNA in the cerebrospinal fluid. PML was ruled out in 43 of the 44 patients, but it could not be ruled out in one patient who had multiple sclerosis and progression of neurologic disease because data on cerebrospinal fluid testing and follow-up MRI were not available. Only the three previously reported cases of PML were confirmed (1.0 per 1000 treated patients; 95 percent confidence interval, 0.2 to 2.8 per 1000). Conclusions: A detailed review of possible cases of PML in patients exposed to natalizumab found no new cases and suggested a risk of PML of roughly 1 in 1000 patients treated with natalizumab for a mean of 17.9 months. The risk associated with longer treatment is not known.