This report provides a comprehensive global safety and clinical database analysis of confirmed PML cases occurring in patients exposed to rituximab for the autoimmune indications of RA and GPA/MPA. The observation period spans the first exposure to rituximab in clinical trials (from 2002 for RA and 2009 for GPA/MPA) until the data cutoff date of 17 November 2015. Our findings indicate: (1) the reported occurrence of PML is very rare in both populations (2.56 cases per 100,000 patients in RA and < 1 case per 10,000 in GPA/MPA); (2) in the RA population, the estimated reporting rate of PML appears to have generally decreased since 2010 and stabilized despite increasing rituximab exposure in patients with RA receiving multiple treatment courses over time; and (3) all confirmed PML cases were associated with other risk factors, independent of rituximab treatment. Widespread use of rituximab for RA only began following its regulatory approval in 2006. The reduction and stabilization of PML cases in patients with RA after 2010 may perhaps be explained by an increased use of rituximab as opposed to other immunosuppressive therapies that more widely employed prior to its approval for these indications which may have carried a higher risk of PML.
As reported previously (2007; Clifford et al. 2011; Fleischmann 2009; Molloy and Calabrese 2012; Hashi et al. 2008), cases of PML in patients with RA treated with rituximab are associated with confounding PML risk factors, including prior and concomitant therapies, a history of malignancy, prior or concomitant SLE, and other immune disorders (leukopenia, lymphopenia). Notably, in four of the nine cases reported here, the patient also had Sjögren syndrome. It is presently not possible to separate the contribution of these factors and any contribution of rituximab to the development of PML in these cases.
Immunosenescence, the gradual deterioration of the immune system brought on by aging, has also been suggested as a contributor to PML occurrence in patients treated with immunomodulatory therapies, with patients ≥ 44 years showing a higher prevalence of “early onset” natalizumab-associated PML (defined as PML developed prior to completing 24 natalizumab courses) than patients < 44 years old (Prosperini et al. 2016). Of 9 patients developing PML with fingolimod, another disease-modifying therapy for multiple sclerosis, all but one were older than 48 years (Berger 2017). The mean age of patients reported herein who developed PML was relatively high compared both with the mean observed in the RA and GPA/MPA global clinical trial programs and with that in the general RA and GPA/MPA populations. However, the overall numbers of confirmed PML cases are low by comparison, and therefore definitive conclusions on the effect of age as a possible risk factor cannot be drawn.
While natalizumab-associated PML usually occurs after a characteristic latency period of 18 to 24 months (Chahin and Berger 2015), the number of PML cases among rituximab-treated patients with RA or GPA/MPA remains too small to assess whether any such characteristic latency period exists with rituximab for these indications. A correlation between the number of courses of rituximab and PML occurrence might be expected if rituximab were a main driver of PML in these cases. However, the number of rituximab courses was highly variable among the nine patients with RA and PML documented here (between one and nine courses) and therefore does not provide any such evidence of causality. Similarly, there was no obvious pattern in PML latency in the cumulative confirmed PML cases where the information was reported. In addition to the lack of an apparent latent period from the time of first initiation of rituximab to the development of PML, the occurrence of PML in patients whose autoimmune underlying diseases already predispose them to PML and the exceptionally low numbers of observed cases in RA and GPA/MPA, despite widespread usage of the monoclonal antibody since regulatory authority approval, further distinguish reports of PML following rituximab use from that seen with natalizumab (Chahin and Berger 2015).
Identifying the risk of PML in any individual patient receiving a given immunosuppressive biologic treatment requires an improved understanding of the barriers to the development of PML. Although evidence from patients with HIV suggests that cell-mediated immunity has a central role in controlling JCV (Koralnik et al. 2001), the role of B cells is less clear: they may act as a potential viral reservoir or contribute to the immune response controlling JCV infection (Durali et al. 2015). Rituximab robustly achieves peripheral blood B cell depletion (Leandro et al. 2006) while the risk for PML remains low, suggesting a minimal involvement of peripheral blood B cells in JCV reactivation. In addition, rituximab administration to a patient with PML neither aggravated the disease process nor prevented clearance of JCV despite a significant B cell depletion lasting 15 months, indicating that B cells are not essential for JCV clearance and recovery (Asztely et al. 2015). Future research should seek to better elucidate the factors resulting in the development of PML and how these might be affected by specific immunosuppressive therapies—including developing an improved understanding of the localities of persistent JCV, how the neurovirulent form of JCV arises during PML pathogenesis, and how JCV is regulated in white blood cells (Wollebo et al. 2015).
PML should be ruled out in any patient displaying neurological deficits, e.g., patient 1, prior to rituximab initiation. Impaired cognitive functions in patients with RA or GPA/MPA who are receiving biologics may be more than a manifestation of aging. PML should be considered as a possible diagnosis when patients treated with rituximab for autoimmune conditions present with new onset neurological manifestations, and consultation with a neurologist is advised as clinically indicated. On suspicion of PML, it is advised that further dosing of immunosuppressive therapy, including rituximab, be suspended until PML has been excluded. Upon diagnosis of confirmed PML, rituximab must be permanently discontinued and the case should be reported and followed up per standard pharmacovigilance practices (2014, 2016). Routine JCV antibody testing for virus exposure is not approved, nor warranted, for patients with RA or GPA/MPA due to the rarity of PML in these populations (Borie and Kremer 2015). To date, JCV antibody testing for PML risk reduction has been validated only with natalizumab because the observed incidence of PML is sufficiently high enough to enable estimated risk stratification in patients with multiple sclerosis (Lee et al. 2013).
Only reported and confirmed cases of PML in approved autoimmune indications (RA and GPA/MPA) are included in the analysis and estimated reporting rates. The PML case reports were limited to available safety data as reported by clinicians and may not, in all cases, contain all relevant information concerning exposure to all potential confounders. For example, high-dose corticosteroids and other immunosuppressive agents are widely used in autoimmune disorders, but information on duration of concomitant therapies was generally lacking.
In addition, the data discussed herein are largely based on spontaneous reporting of PML cases over 11 years (for RA) and 5 years (for GPA/MPA), and the number of unreported cases during these time periods is unknown. Thus, this evaluation reveals the importance of physician vigilance, prompt reporting, and long-term/continued safety follow-up for an accurate prevalence estimation of very rare adverse events such as PML.