To do or not to do? plasma exchange and timing of steroid administration in progressive multifocal leukoencephalopathy

Objective To retrospectively analyze the effect of plasma exchange (PLEX; yes = PLEX + , no = PLEX – ) and steroids administration timing (prophylactically [proST] or therapeutically [therST]) on the longitudinal clinical course of patients with natalizumab‐related progressive multifocal leukoencephalopathy (PML) and full‐blown immune reconstitution inflammatory syndrome (PML‐IRIS). Methods Clinical and radiological data of 42 Italian patients with PML were analyzed. Patient's data are available until 12 months after PML diagnosis. PLEX and steroids treatment as time‐dependent covariates were entered in: (1) a Cox model to investigate their impact on full‐blown PML‐IRIS latency; (2) an analysis of variance ANOVA to investigate their impact on IRIS duration; and (3) a linear mixed model to assess their impact on the longitudinal clinical course (measured by means of Expanded Disability Status Scale [EDSS]). Results Treatment with PLEX was not associated to PML‐IRIS latency (hazard ratio [HR] = 1.05; p = 0.92), but once IRIS emerged, its duration was significantly longer in patients who underwent PLEX (101 vs 54 days in PLEX + and PLEX – patients; p = 0.028). Receiving proST versus therST was not associated to IRIS latency (HR = 0.67; p = 0.39) or duration ( p = 0.95). Patients who underwent proST had a significantly higher EDSS increase during PML (0.09 EDSS points per month; p = 0.04) as compared to those who had therST. Interpretation This study highlights that: (1) caution on the use of PLEX should be considered as the current data do not support a beneficial effect of PLEX and (2) caution on the early use of steroids is suggested because their prophylactic use to prevent full‐blown PML‐IRIS seems to negatively impact on the longitudinal disability course. Ann Neurol 2017;82:697–705