Research Letters Multiple sclerosis is an inflammatory demyelinating disease of the CNS, often associated with relapses. Although current treatments are uniformly applied, clinical, radiological, and pathological features are heterogeneous. Previous studies have described four immunopathological patterns of demyelination in early multiple sclerosis lesions characterised by inter-individual heterogeneity, but intra-individual homogeneity. Pattern I is characterised by T-cell/ macrophage-associated demyelination; pattern II is characterised by antibody/complement-associated demyelination; pattern III is defined by a distal oligodendrogliopathy; and pattern IV is characterised by oligodendrocyte degeneration in periplaque white matter. 1 These observations could be clinically relevant since distinct patterns might need distinct treatments. Furthermore, since antibodies and complement contribute to multiple sclerosis pathology, 2 treatments to deplete them, such as therapeutic plasma exchange (TPE), could be effective. About 45% of patients with fulminant attacks of multiple sclerosis and other idiopathic inflammatory demyelinating diseases (IIDD) of the CNS that are unresponsive to corticosteroids, improve after TPE. 3 Factors variably associated with favourable response include male sex, retained reflexes, and early treatment. 4 The response to TPE seems to be all or none, which suggests that a single biological factor might explain much of the variation. We postulated that immunopathological heterogeneity could contribute to differences in TPE response. An inclusion criterion was TPE administered for a fulminant IIDD attack of the CNS resulting in major neurological deficit in one or more of the following: motor, cerebral, brainstem or cranial nerve, cerebellar, or sensory functions. Other criteria were sufficient clinical documentation to gauge response to therapy; tissue available from biopsy done before TPE to exclude alternative diagnoses, such as tumour, or from autopsy; pathology consistent with active inflammatory demyelination; and lesions immunopathologically classifiable into patterns I–IV, as previously described. 1 Exclusion criteria were: diagnoses of neuromyelitis optica, acute disseminated encephalomyelitis, or other non-multiple sclerosis disease on follow-up; and use of multiple concomitant immunosuppressive or immunomodulatory medication at time of TPE. Insufficient tissue was available for immuno- pathological classification for three patients. 23 individuals met inclusion criteria, of whom four were excluded (two acute disseminated encephalomyelitis, one lymphoma, one multiple concomitant medication). 19 were included (16 biopsies, two both biopsy and autopsy, and one autopsy). Long-term clinical follow-up was obtained for 16 patients via examination by a study investigator (n=14) or by contact with local physicians (n=2). TPE was undertaken with continuous-flow centrifugation in all patients. TPE response was assessed prospectively in four patients included in a prior double- blinded, sham-controlled trial 3 and retrospectively in 15 (11 via neurological examination done by a study investigator; four by chart review and physician contact). The response was assessed independently from, and blinded to, immunopathological classification for all patients. Patients were assessed at the Mayo Clinic (n=12), University of Vermont (n=1), or a European centre (n=6). Immunopattern classification was done with 100% consensus among three study investigators, as well as independently from, blinded to, and after clinical assessment and documentation of TPE response for all patients. The figure shows the characteristics of pattern II and III pathology. The primary outcome was TPE response according to an established grading scale 5 that defines neurological improvement as: none; mild (subjective or minimal, but no effect on function); moderate (gain in neurological status that affects function); or marked (important difference from baseline with major gain in function). Lancet 2005; 366: 579–82 See Comment page 526 Department of Neurology (M Keegan MD, B Weinshenker MD, M Rodriguez MD, C F Lucchinetti MD), Department of Health Sciences Research (R McClelland PhD), Department of Laboratory Medicine and Pathology (J Parisi MD), Department of Immunology (M Rodriguez), and Mayo Graduate School (Y Morales BA), Mayo Clinic College of Medicine, Rochester MN, USA; Institute of Neuropathology Göttingen, Germany (F König MD, W Brück MD); Ruppiner Kliniken GmbH, Neuruppin, Germany (A Bitsch MD); University of Vermont, Burlington VT, USA (H Panitch MD); and Brain Research Institute, University of Vienna, Vienna, Austria (H Lassmann MD) Correspondence to: Claudia F Lucchinetti, Department of Neurology, Mayo Clinic College of Medicine, Rochester, MN 55905, USA lucchinetti.claudia@mayo.edu www.thelancet.com Vol 366 August 13, 2005 579 Relation between humoral pathological changes in multiple sclerosis and response to therapeutic plasma exchange Mark Keegan, Fatima König, Robyn McClelland, Wolfgang Brück, Yazmín Morales, Andreas Bitsch, Hillel Panitch, Hans Lassmann, Brian Weinshenker, Moses Rodriguez, Joseph Parisi, Claudia F Lucchinetti Early, active multiple sclerosis lesions show four immunopathological patterns of demyelination. Although these patterns differ between patients, multiple active lesions from a given patient have an identical pattern, which suggests pathogenic heterogeneity. Therapeutic plasma exchange (TPE) has been successfully used to treat fulminant demyelinating attacks unresponsive to steroids. We postulated that patients with pattern II would be more likely to improve after TPE than those with other patterns since pattern II lesions are distinguished by prominent immunoglobulin deposition and complement activation. We retrospectively studied 19 patients treated with TPE for an attack of fulminant CNS inflammatory demyelinating disease. All patients with pattern II (n=10), but none with pattern I (n=3) or pattern III (n=6), achieved moderate to substantial functional neurological improvement after TPE (p0·0001). Patients with multiple sclerosis with pattern II pathology are more likely to respond favourably to TPE than are patients with patterns I or III.