Thirteen autopsy cases of patients with clinical criteria of the Guillain-Barré syndrome were investigated for pathological changes and cellular composition of inflammatory infiltrates in the CNS and PNS. The survival times from the onset of neurological symptoms until death ranged from 1 day to 12 months. In the CNS, degeneration of spinal posterior tracts was seen in three cases. Mononuclear infiltrates consisted of evenly proportioned lymphocytes and macrophages in cases with survival of 1 and 2 days, whereas macrophages predominated in cases with survival of 5 days and longer. Infiltrates presented as nodular clusters around blood vessels and neurons, or were scattered diffusely. They were found within the spinal cord in eight out of 13 cases, within the medulla oblongata in eight out of 12 cases, within the pons in five out of nine cases, and in one out of four midbrains. Activation of microglia, either focal or diffuse, was found in various degrees in 11 out of 13 cases, involving the spinal cord (six out of 13 cases), the medulla oblongata (10 out of 12 cases), the pons (five out of nine cases) or as subependymal rims along the walls of the ventricular system and the central canal of the spinal cord (seven out of 13 cases). In the PNS, myelin loss (12 out of 13 cases), axonal degeneration (six out of 13 cases) and mononuclear cell infiltrates (13 out of 13 cases) were seen in segmental and cranial nerves, spinal ganglia and spinal roots in varying distribution and severity. Mononuclear cell infiltrates were composed of macrophages and T lymphocytes, with even distribution in cases with short survival (1 and 2 days), and predominance of macrophages in cases with protracted clinical course. T lymphocytes were equally composed of OPD4+ and CD8+ cells without obvious differences between cases of short and long duration, or between PNS and CNS infiltrates in 11 out of 12 cases, whilst two cases had a dominant OPD4+ subset. We conclude that CNS pathology is frequent in patients with Guillain-Barré syndrome. It involves axons with secondary myelin impairment, microglial activation and inflammatory infiltration. In this series, primary demyelination is not found in the CNS. Changes such as degeneration of spinal posterior tracts are secondary to pathology in the PNS. Inflammatory cell reactions in the CNS are similar to those in the PNS and to CNS pathology in experimental allergic neuritis. This inflammation might reflect CNS immune activation in the absence of the relevant antigen, in addition to cellular reactions accompanying secondary CNS changes. The presence of distinct pathology in the CNS is in contrast with other recent studies on the pathology of Guillain-Barré syndrome which, unlike this study, may have been influenced by recently introduced treatments.