Drug-Induced Progressive Multifocal Leukoencephalopathy: A Comprehensive Analysis of the WHO Adverse Drug Reaction Database

Objective To identify safety signals concerning the association between the use of various drug classes and the onset of progressive multifocal leukoencephalopathy (PML). Methods All reports containing suspected or interacting PML-related or leukoencephalopathy-related drugs, held in the World Health Organization spontaneous individual case safety reports database as at 1 September 2014, were retrieved. We identified safety signals by analysing the drug-reaction pairs, using the reporting odds ratio as a measure of disproportionality. A safety signal was defined if a drug was reported more than twice in PML cases with a reporting odds ratio >2 and a lower 95 % confidence limit >1. Results We retrieved 2452 reports associated with PML (N = 1612), leukoencephalopathy (N = 835) or both (N = 5), corresponding to 343 different drugs. PML was reported similarly in male and female adults (18-64 years), and almost 30 % of the cases had a fatal outcome. The most frequent Anatomical Therapeutic Chemical (ATC) classification groups concerned antineoplastic agents (23.5 %), antivirals for systemic use (10.1 %) or immunostimulants (4.6 %). Significant disproportionality was found for 88 drugs in the overall analysis (of cases with 'progressive multifocal leukoencephalopathy' or 'leukoencephalopathy' as the Preferred Term), and a new safety signal was identified for 59 active substances (e.g. muromonab-CD3, basiliximab and antithymocyte Ig), as no information on a possible risk of PML was acknowledged in their Summary of Product Characteristics documents. Some safety signals were confirmed also after sensitivity analysis adjustment for several confounding factors (underlying diseases and considering only 'progressive multifocal leukoencephalopathy' as the Preferred Term). Conclusion We report a possible association between several drugs and PML that has not been previously described. In addition, we have confirmed previously reported signals in a number of drugs. We highlight the need for follow-up by regulatory agencies.