Of 40 patients with chronic inflammatory demyelinating polyradiculoneuropathy (CIDP), 28 completed a controlled three-month trial of prednisone. Prednisone was shown to cause a small but significant improvement over no treatment in scored neurological disability, some measures of computer-assisted sensory detection threshold, graded muscle strength, and some attributes of nerve conduction. No subset of patients was more likely than another to be responsive to prednisone; those with a progressive course were as likely to be responsive as recurrent cases. This finding provides further justification for classifying progressive with recurrent cases as CIDP and demonstrates that prednisone treatment should not be withheld from patients with progressive disease.