Over a 10-year period, we followed up 60 patients (35 men and 25 women) with chronic inflammatory demyelinating polyradiculoneuropathy. Diagnosis was based on previously outlined criteria. Patients were treated in a uniform manner and the overwhelming majority, 56 (94.9%) of 59 treated patients, initially responded to immunosuppressive therapy. The time for initial improvement was 1.9 +/- 3.6 months while the time to reach a clinical plateau was 6.6 +/- 5.4 months. The course was monophasic in 32 patients (53.3%) and relapsing in 28 (46.6%). Despite the initial responsiveness, only 24 (40%) of 60 patients are in partial or complete remission, receiving no medication. Two patients died. We were unable to identify specific clinical or laboratory features at the time of diagnosis that predicted outcome. Our data analysis, along with previous reports, suggests that chronic inflammatory demyelinating polyradiculoneuropathy may be more heterogeneous than previously emphasized. In this light, we have proposed diagnostic criteria that allow for the heterogeneity but at the same time provide for a more consistent approach to better establish the natural history of this condition.