Predictors of survival and functional outcomes in natalizumab-associated progressive multifocal leukoencephalopathy

Abstract

Natalizumab, a highly effective therapy for relapsing-remitting multiple sclerosis, is associated with a risk of progressive multifocal leukoencephalopathy (PML). The objective of this analysis was to examine factors predicting survival in a large natalizumab-associated PML global population. Patients with natalizumab-associated PML identified through postmarketing surveillance were followed up for up to 24 months using a structured questionnaire completed by treating physicians. Demographic and clinical characteristics, JC viral load, magnetic resonance imaging (MRI) results, and Expanded Disability Status Scale (EDSS) and Karnofsky Performance Scale (KPS) scores were compared in survivors and nonsurvivors. Kaplan-Meier analysis was used to model survival function. Among the 336 patients included in this analysis, 76 % survived, with mean follow-up time from PML diagnosis of 16.1 months for survivors; mean time from diagnosis to death was 4.7 months for nonsurvivors. Survivors were significantly younger at diagnosis, had significantly lower EDSS scores and higher KPS scores prior to PML diagnosis, and had significantly lower cerebrospinal fluid JC viral load at the time of diagnosis. Patients with less extensive disease on MRI at diagnosis had a higher survival rate than those with widespread disease. Survivors generally had less functional disability pre-PML, at PML diagnosis, and in subsequent months. In survivors, functional disability appeared to stabilize approximately 6 months post-PML diagnosis. In this analysis, younger age at diagnosis, less functional disability prior to PML diagnosis, lower JC viral load at diagnosis, and more localized brain involvement by MRI at the time of diagnosis appeared to predict improved survival in natalizumab-associated PML.

Keywords: Expanded Disability Status Scale; Karnofsky Performance Scale; Natalizumab; Progressive multifocal leukoencephalopathy; Survival.

Fig. 1

Log JC viral load (copies/mL) for PML patients according to survival status (p < 0.0001). p value adjusted for age. White horizontal line, median; horizontal bars, range

Fig. 2

Functional outcomes as measured by a EDSS and b KPS scores for PML patients according to survival status. Each symbol represents a single patient measurement at a single time point. The light gray lines represent polynomial regression trend lines (LOWESS curves) for survivors; the dark gray lines represent polynomial regression trend lines (LOWESS curves) for nonsurvivors. EDSS and KPS scores were not available for all patients at all time points

Fig. 3

Kaplan-Meier estimate of overall survival after PML diagnosis