Progressive multifocal leukoencephalopathy in patients with follicular lymphoma treated with bendamustine plus rituximab followed by rituximab maintenance

Progressive multifocal leukoencephalopathy (PML) is a rare fatal disease caused by polyomavirus JC occurring in the context of a deep and prolonged immunosuppression. Rituximab is one of the monoclonal antibodies labelled with a black box warning for the risk of PML. The estimated incidence rate of PML ranges from 1·39 to 1·87 per 10·000 rituximab-exposed patients (Focosi et al., 2019). Although PML is classified as a ‘very rare’ complication of rituximab treatment, the combination with other drugs may significantly modify this risk. In the last decade, the combination of bendamustine and rituximab (BR) has become the preferred regimen chosen by many hematologists to treat patients with high tumour burden follicular lymphomas (FL). Similarly, maintenance therapy administering rituximab every two or three months for two years has become part of standard first line treatment of FL. Here we report our monocentric experience of three cases of PML occurring in patients with FL treated with BR followed by rituximab maintenance. This retrospective analysis included 47 patients with high tumour burden FL (22 males, 25 females; median age 59 years, range 28–82 years), diagnosed at our institution from 2014 to 2018, and treated with rituximab maintenance following six cycles of BR induction chemotherapy. All patients were HIV-negative. All patients treated with rituximab maintenance had at least a partial response to BR induction treatment (40 complete response and seven partial response). At last follow-up, 41 patients are alive and 6 patients had died. Among the causes of death, three patients died of PML, one of lymphoma, one of gastrointestinal bleeding and one of unspecified cause. The probability of OS at a median follow-up of 3 years was 91% with significantly lower probability for patients aged >70 years when compared to younger patients (78% vs. 95%, respectively; P = 0·001). Characteristics of the three patients developing PML are summarized in Table 1. The onset of PML symptoms occurred during rituximab maintenance in two patients and at the end of maintenance in a third patient. Time between onset of symptoms and definitive diagnosis of PML was approximately 1 month. Definitive diagnosis of PML was based on the demonstration of the JC virus using PCR in the CSF, combined with the typical clinical and imaging findings (Table 1 and Fig 1) according to consensus criteria of the American Academy of Neurology Neuroinfectious Disease Section (Berger et al., 2013). At the time of PML diagnosis, lymphocyte counts were lower than normal. Immunophenotyping of peripheral blood, available in two patients, showed a marked reduction of the CD4+ cell count <200/μl (Table 1). B cells detected by CD19 were almost absent. Our observation of three cases of PML in a cohort of 47 patients with FL treated with BR followed by R maintenance appears unusually high and, considering the limit of a small monocentric cohort, it does not allow to calculate actual incidence rates. A low and persistent CD4+ cell count in peripheral blood has been described as the major risk factor for PML in HIV-positive patients (Engsig et al., 2009). JC virus is ubiquitous in humans (50–70% of the general population has a positive serology) and remains in a latent state in several organs, including the brain. CD4+ cells appear to play a central role in the control of JC virus infection in the brain and drive CD8+ T-cell activation into JC virus-specific effector cells through a T Helper 1-type response (Pavlovic et al., 2018). Lymphocytopenia and CD4+ cell deficit are well-known sequelae of BR treatment (García Muñoz et al., 2014; Martínez-Calle et al., 2019). In a cohort of 295 patients with CLL treated with BR, median times to lymphocyte count recovery (≥1000/μl) and CD4+ recovery (≥200/μl) were 26 and 24 months, respectively (Martínez-Calle et al., 2019). In the GALLIUM study, bendamustine combined with