Progressive multifocal leukoencephalopathy associated with immunosuppressive therapy in rheumatic diseases: Evolving role of biologic therapies


  • Eamonn S. Molloy,

    Corresponding author
    1. St. Vincent's University Hospital, Elm Park, Dublin, Ireland
    • Department of Rheumatology, St. Vincent's University Hospital, Elm Park, Dublin 4, Ireland
    Search for more papers by this author
    • Dr. Molloy has received speaking fees from Bristol-Myers Squibb, MSD, Abbott Laboratories, and Roche, and honoraria for Advisory Board service from GlaxoSmithKline and Roche (less than $10,000 each).

  • Leonard H. Calabrese

    1. Cleveland Clinic Foundation, Cleveland, Ohio
    Search for more papers by this author
    • Dr. Calabrese has received consulting fees and speaking fees from Roche, Pfizer, Amgen, Bristol-Myers Squibb, and Centocor (less than $10,000 each) and from Genentech (more than $10,000).



To evaluate the association of progressive multifocal leukoencephalopathy (PML) with immunosuppressive therapy for autoimmune rheumatic diseases (ARDs).


A Freedom of Information Act request was submitted for all cases of PML within the Food and Drug Administration Adverse Event Reporting System database. ARD cases were selected for further analysis.


A total of 34 confirmed cases of PML in the setting of ARDs were identified: 17 had systemic lupus erythematosus, 10 had rheumatoid arthritis, 4 had vasculitis, and 3 had dermatomyositis. Fifteen of these patients were treated with one or more biologic agents: 14 received rituximab (RTX), 6 received anti–tumor necrosis factor (anti-TNF) therapy (5 treated with anti-TNF agent prior to RTX). Four RTX-treated patients were not receiving additional immunosuppressive therapy at the time of PML onset, other than an antimalarial drug and/or low-dose glucocorticoids; all others who were receiving a biologic agent were also receiving one or more synthetic disease-modifying agents. All but 1 patient receiving a biologic agent had at least 1 potential confounding factor for the diagnosis of PML. The remaining 19 confirmed cases of PML among ARD patients were treated with synthetic disease-modifying antirheumatic drugs only, 14 of whom had received an alkylating agent.


PML has been reported in patients with ARD treated with various immunosuppressive agents. The limitations of this study preclude definitive attribution of causality. While the paucity of confirmed cases recently exposed to anti-TNF therapy suggests a causal relationship is unlikely, a specific signal is emerging with regard to rituximab and PML. Although this is a rare adverse event associated with RTX therapy, the devastating nature of PML mandates continued vigilance, particularly in patients with current or prior exposure to an alkylating agent.

Progressive multifocal leukoencephalopathy (PML) is a rare, typically fatal, opportunistic infection caused by the JC virus. Practitioners caring for patients with human immunodeficiency virus (HIV) infection, cancer, and organ transplants are familiar with this devastating illness. More recently, PML has affected the practice of other medical specialists, especially those who prescribe biologic agents, such as natalizumab for conditions such as multiple sclerosis and Crohn's disease, efalizumab for psoriasis, and rituximab for rheumatoid arthritis (RA) and other autoimmune disorders. PML is a complex, rare, and poorly understood disorder; thus, ascribing risk to individual therapies has been problematic.

We have previously reported the occurrence of PML among patients with autoimmune rheumatic diseases (ARDs), identified from a systematic review of the published literature (1) as well as from a national hospital discharge database (2). Systemic lupus erythematosus (SLE) appears to be associated with a particular susceptibility to PML that cannot be entirely explained by the intensity of immunosuppressive therapy. These studies, however, were unable to adequately define the contribution of biologic agents to the risk of PML in patients with autoimmune rheumatic diseases. The aim of the current study was to examine the aggregate experience of PML reported in association with ARDs in the Food and Drug Administration (FDA) Adverse Event Reporting System (AERS) database.


A Freedom of Information Act request was submitted in October 2010 to obtain MedWatch forms for all cases of PML and/or JC virus infection within the FDA AERS database from November 1, 1997 to March 31, 2010. Due to a 3–6 month lag in data update in the AERS database, only reports through March 31, 2010 were available at the time of the request. All MedWatch forms were reviewed; those with identified autoimmune rheumatic diseases were selected for further analysis.

Definition of PML.

PML was classified as confirmed or unconfirmed. Confirmatory evidence could be obtained from either the AERS data or a peer-reviewed journal article associated with that case report source. Confirmed PML was defined as characteristic changes in brain tissue, as detected by magnetic resonance imaging (MRI) or computerized tomography (CT) and identification of JC virus in the cerebrospinal fluid via polymerase chain reaction (PCR) analysis or in the brain tissue (either through brain biopsy or at autopsy) via in situ hybridization, immunohistochemistry, or PCR. Characteristic neuroimaging features of PML are symmetric or asymmetric multifocal areas of white matter demyelination that do not conform to cerebrovascular territories and exhibit neither mass effect nor contrast enhancement. These lesions are typically hypodense on CT; on MRI, they appear as areas of decreased signal on T1-weighted images and increased signal on T2-weighted images. Involvement of the deep gray structures, including the basal ganglia and thalamus, can be found in a minority of cases, but is always accompanied by white matter disease. In particular, neuroimaging reports were screened for descriptions of atypical features, such as exclusive spinal cord involvement or evidence of acute infarct/ischemia.

Patient selection.

Patients were included in this study only if they had a diagnosis of an ARD and confirmed PML. Of 657 reports of PML contained in the AERS database, the large majority, as anticipated, were associated with HIV, cancer, and organ transplantation. The presence of an autoimmune rheumatic disease was suspected in 63 patients (9.6%), and these reports were selected for further analysis. Of these 63 cases, 18 (28.6%) were deleted because there was inadequate information to document the diagnosis of PML or the available information was not consistent with the diagnosis of PML (unconfirmed PML). Of the remaining 45 cases, 11 duplicate reports were deleted (24.4%). Therefore, the final number of confirmed cases of PML in the setting of ARDs was 34 (5.2% of the total number of PML reports).

Data collection.

Specific data abstracted from MedWatch forms included demographic features, known disease-association cofactors of PML (HIV, transplant, lymphoproliferative disease), autoimmune disease, use of immunosuppressive drugs, and other possible cofactors (e.g., CD4 cell counts, etc). Other information extracted included outcome of PML, use of antiviral therapy, and features suggestive of inflammatory PML (contrast enhancement or edema on neuroimaging and/or inflammatory infiltrates on brain biopsy).

Cofactors were defined as follows. 1) HIV/acquired immunodeficiency syndrome (AIDS) positivity was defined as the presence of any of the following terms, phrases, or laboratory results: HIV, HIV infection, AIDS, highly active antiretroviral therapy (HAART), antiretroviral treatment/therapy, HIV viral load >0 copies/ml, HIV RNA >0 copies/ml. 2) Transplant recipients were patients who had received any of the following transplanted organs: liver/hepatic, kidney/renal, stem cells/bone marrow/blood, skin, lungs, intestine, heart/cardiac. 3) Lymphoproliferative disease was defined as a diagnosis of any of the following conditions: lymphoma, leukemia, lymphoproliferative malignancy. Also documented were lymphocytopenia of <1,000 cells/μl and CD4+ cell counts that were <200 cells/μl, based on recorded values as documented on the MedWatch forms.


A total of 34 confirmed cases of PML in the setting of ARDs were identified (Tables 1 and 2). Summary details of the primary ARD in each case are provided in Table 3. Eight patients also had a diagnosis of a second ARD (6 had Sjögren's syndrome, 1 had RA, and 1 had dermatomyositis). No patient had a documented diagnosis of HIV/AIDS or an organ/stem cell transplant. Three patients had received chemotherapy and/or radiotherapy for prior malignancy; all 3 were in the group treated with biologic agents. In 11 cases (6 in the biologic agents group and 5 in the synthetic only group), inflammatory changes were noted on MRI (contrast enhancement) and/or brain biopsy at some point during the course of the disease, consistent with inflammatory PML.

Table 1. Confirmed cases of PML in patients with ARDs treated with one or more biologic agents*
Patient/age/sexARD diagnosis, durationImmunosuppressive agents at PML onsetPrevious drugs, dosage, time from last doseInflam. PML identifiedOutcome (interval)Antiviral drugsOther PML risk factorsComment
Biologic agent (no. of courses; time from first/last dose)Other agents, dosage, duration
  • *

    PML = progressive multifocal leukoencephalopathy; ARDs = autoimmune rheumatic diseases; inflam. = inflammatory; SLE = systemic lupus erythematosus; RTX = rituximab; HCQ = hydroxychloroquine; CYC = cyclophosphamide; AZA = azathioprine; NR = not reported; GCs = glucocorticoids (prednisone equivalents); cid. = cidofovir; IVIG = intravenous immunoglobulin; RA = rheumatoid arthritis; anti-TNF = anti–tumor necrosis factor; mef. = mefloquine; SS = Sjögren's syndrome; MTX = methotrexate; inflix. = infliximab; q8w = every 8 weeks; RT = radiotherapy; DM = dermatomyositis; ALC = absolute lymphocyte count (× 106/liter); qod = every other day; MRI = magnetic resonance imaging; ITP = immune thrombocytopenic purpura; CV = cryoglobulinemic vasculitis; IV = intravenous (pulse dose); etan. = etanercept; ada. = adalimumab; mirt. = mirtazapine; MMF = mycophenolate mofetil; chemo. = chemotherapy; LEF = leflunomide; SSZ = sulfasalazine; ana. = anakinra; plex. = plasma exchange.

B1/45/FSLE, 23 yearsRTX (3; 43 mos./5 mos.)HCQ 300 mg/day, 4 yearsOral CYC NR, 14 years; AZA NR, NR; GCs NR, 1 yearNoDied (4 mos.)Cid.Leukopenia, CD4 cells <200 × 106/literHistory of recurrent zoster, also given IVIG
B2/55/FRA, NRRTX (3; NR/ 6 mos.)Anti-TNF NR, NRNRSurvived (NR)Mef.Also plasmapheresis
B3/51/FRA, SS, 14 yearsRTX (4; 57 mos./17 mos.)MTX 20 mg/week, 6 years; GCs? NRHCQ NR, NR; inflix. 5 mg/kg q8w, 5 yearsBiopsyDied (1 mo.)NRLymphopeniaOral cancer 9 mos. previously, given RT, carboplatin, cetuximab
B4/41/FDM, RA?, 30 mos.RTX (2; 16 mos./0.5 mos.)MTX 25 mg/week, 29 mos.; GCs 20 mg/day, 30 mos.Inflix. 400 mg q8w, 3 mos.NoNRNoLymphopenia, ALC 3005 mos. of inflix. between 2 courses of RTX
B5/69/FSLE, NRRTX (1; 66 mos.)GCs 10 mg qod, NRIVIG NR, NR; AZA NR, NR; HCQ NR, NRMRI, biopsyNRNoAMG 531 for ITP
B6/55/FCV, 7 yearsRTX (2; 9 mos./ 1 mo.)AZA NR, NR; GCs 20 mg/day, NRIV CYC 25 gm, 7 years; MTX NR, NRNoDied (NR)Cid.No chronic viral infection
B7/72/FRA, SS, 30 yearsRTX (4; 26 mos./7 mos.)MTX 15 mg/week, 10 yearsEtan. NR, 5 years; ada. NR, 3 yearsNoDied (11 mos.)Mef., mirt.Chronic leukopenia, ALC <4005th course of RTX after onset of neurologic symptoms
B8/44/FCV, SS, 10 yearsRTX (2; 41 mos./29 mos.)MMF NR, NR; GCs NR, NRIV CYC ? ×12, 38 mos.; MTX NR, NRBiopsyNRNRChemo., RT for parotid lymphoma >4 years priorRTX given for mononeuritis multiplex
B9/73/FRA, 3 yearsRTX (1; 4 mos.)LEF 20 mg/day, 31 mos.; HCQ 400 mg/day, 16 mos.; GCs 2.5 mg/day, 32 mos.MRI, biopsyDied (4 mos.)Mef.Lymphopenia, ALC 718–1,411No prior anti-TNF
B10/62/FRA, 20 yearsRTX (3; 17 mos./5 mos.)MTX NR, 9 yearsHCQ NR, 9 years; GCs NR, 1 year; etan. NR, 4 years; ada. NR, NR; SSZ NR, 8 years; LEF NR, ?5 years; ana. NR, 2 years; gold NR, 2 yearsMRISurvived, stable (6 mos.)Mef., mirt., plex.Leukopenia4th cycle of RTX after likely onset of PML
B11/47/FSLE, 4 yearsRTX (1; 7 mos.)MTX? NR, NR; GCs? NR, NRAZA NR, NR; SSZ NR, NRNoNRMirt., IVIGDates of other drugs not reported
B12/66/FRA, SS, NRRTX (1; 15 mos.)GCs? NR, NRMTX NR, NR; HCQ NR, NRMRISurvived (9 mos.)Mirt., mef.Lymphopenia, ALC 600–900Breast cancer 3 years previously, given RT
B13/51/FSLE, 9 mos.RTX (1; 1 mo.)IV CYC 11 gm, 9 mos.; AZA NR, ?1 mo.HCQ NR, 1 mo.NoNRNROral CYC started after onset of PML
B14/35/FSLE, SS, >7 yearsRTX (3; 12 mos./0 mos.)IV CYC NR, 1 year; GCs <10 mg/day, NR; HCQ NR, 3 mos.MTX <15 mg/week, 3 yearsNoSurvived (17 mos.)NoAZA added after onset of PML
B15/69/MRV, RA, NRInflix. (NR; NR)Oral CYC 150 mg/day, NR; HCQ 400 mg/day, NR; GCs 5 mg/day, NRNoDied (NR)NoRheumatoid vasculitis diagnosed at autopsy
Table 2. Cases of PML in patients with ARDs treated with only synthetic (nonbiologic) immunosuppressive agents*
Patient/age/sexARD diagnosis, durationTreatment at PML onset, dosage, durationPrevious drugs, dosage, time from last doseInflam. PML identifiedOutcome (interval)Antiviral drugsOther PML risk factorsComment
  • *

    PML = progressive multifocal leukoencephalopathy; ARD = autoimmune rheumatic diseases; inflam. = inflammatory; SLE = systemic lupus erythematosus; AZA = azathioprine; GCs = glucocorticoids (prednisone equivalents); IFNα = interferon-α; MMF = mycophenolate mofetil; IV = intravenous (pulse dose); CYC = cyclophosphamide; MRI = magnetic resonance imaging; cid. = cidofovir; cytar. = cytosine arabinoside; NR = not reported; CSA = cyclosporin A; HCQ = hydroxychloroquine; CHL = chlorambucil; GPA = granulomatosis with polyangiitis (Wegener's); pred. = prednisone; RA = rheumatoid arthritis; CQ = chloroquine; pen. = penicillamine; DM = dermatomyositis; IVIG = intravenous immunoglobulin; risp. = risperidone; pero. = perospirone; LEF = leflunomide; mirt. = mirtazapine; MTX = methotrexate; SS = Sjögren's syndrome; bid = twice a day.

S1/41/MSLE, 6 yearsAZA 100 mg/day, 5 years; GCs NR, NRNoDied (7 mos.)IFNαLymphopeniaNR
S2/48/FSLE, 5 yearsMMF 2.5 gm/day, ?9 mos.; GCs NR, NRIV CYC 3.6 gm, ?9 mos.MRISurvived (12 mos.)Cid., cytar.
S3/51/FSLE, 8 years?GCs NRIV CYC 9.7 gm, 7 years; CSA NR, 5 yearsNoDied (?4 mos.)Cid.LymphopeniaDialysis × 4 years; MMF started after PML onset
S4/40/FSLE, 20 yearsGCs low dose, >2 yearsIV CYC ×12, 5 years; MMF NR, 2 years; AZA NR, 1 yearNoSurvived (51 mos.)Cid.Died of renal failure
S5/53/FSLE, 18 yearsGCs 10 mg/day, ?; HCQ 400 mg/day, ?; MMF 2 gm/day, 18 mos.IV CYC × 1, 8 yearsMRI, biopsySurvived (31 mos.)NoLymphopeniaMMF stopped at PML diagnosis
S6/62/FSLE, 10 yearsCHL 1 mg/day, 2 years; GCs 5 mg/day, 2 yearsNoDied (?)Cid., IFNαLeukopenia
S7/60/MGPA, 14 mos.Oral CYC 150 mg/day, 14 mos.; GCs 7.5 mg/day; 14 mos.NoSurvived (24 mos.)NoGPA remission with pred. 5 mg/day
S8/27/MSLE, 7 yearsOral CYC NR, 2 years; GCs low dose, NRAZA NR, 2 yearsNoDied (3 mos.)NoSplenectomy 15 years previouslyDiagnosis at autopsy
S9/59/FRA, 42 yearsCHL NR, 1 year; GCs NR, NRAZA NR, NR; CQ NR, NR; gold NR, NR; pen. NR, NRNoSurvived (5 years)NoPrevious 4 years CHL exposure for amyloidosis; died of gastric bleeding
S10/51/MSLE, 11 yearsAZA 100 mg/day, 2 years; GCs 5 mg/day, NRIV CYC NR, NRBiopsySurvived (5 years)NoBiopsy 3 mos. after onset
S11/57/FSLE, 6 yearsIV CYC 4.5 gm, 9 mos.MTX ×1, 4 years; AZA NR, 1 year; MMF NR, NRNoDeath (7 mos.)NoTotal CYC exposure 9.1 gm
S12/67/MDM, 4 yearsCYC NR, 4 years; IVIG NR, 4 years; GCs NR, 4 yearsNoDied (?∼3 mos.)Cid.Initial treatment with high-dose GCs plus sirolimus
S13/27/FSLE, 7 yearsCSA 150 mg/day, 1 year; GCs NR, NRIV CYC NR, ?1 yearNoDied (?)Risp., pero.IV CYC × ?3 years total
S14/66/FRA, 1 year; DM, 1 mo.AZA 100 mg/day, <1 mo.; HCQ 400 mg/day, 9 mos.; GCs 10 mg/day, 12 mos.MTX 10 mg/week, 1 yearBiopsySurvived (9 mos.)Cid.Onset within 1 mo. of starting AZA for DM
S15/68/MRA, NRLEF NR, NRAZA NR, NRNoSurvived (12 mos.)Cytar.
S16/60/MGPA, 8 yearsAZA 150 mg/day, 4 mos.; GCs 8 gm/day, NRIV CYC ×17, 2 years; Oral CYC 150 mg/day, 4 mos.BiopsySurvived (NR)NoLymphopenia, CD4 cells 177 × 106/literMonthly IVIG
S17/70/MRA, ?3 yearsMTX 20 mg/week, 3 years; CQ 2 gm/week, 3 yearsNoDied (3 mos.)Cytar., mirt.Normal T cell subsetsNR
S18/59/FSLE, NRMTX NR, 6 years; GCs NR, 6 yearsNoDied (6 mos.)NoT cells 178 × 106/literMTX taken daily in error, ?duration
S19/62/FSLE, SS, 22 yearsMMF 1 gm bid, 4 mos.; GCs 12.5 mg/day, NRIV CYC NR, NR; HCQ NR, NRNoDied (<8 mos.)NoMTX >20 years previously
Table 3. Primary ARDs among 34 patients who developed PML during immunosuppressive therapy for ARD*
Primary ARDTotal no. of patientsNo. taking biologic agentsNo. taking synthetic agents only
  • *

    The primary autoimmune rheumatic disease (ARD) was the disease for which immunosuppressive agents were primarily prescribed. Eight patients had a secondary rheumatic disease: Sjögren's syndrome in 6, rheumatoid arthritis (RA) in 1, and dermatomyositis (DM) in 1. PML = progressive multifocal leukoencephalopathy; SLE = systemic lupus erythematosus; CV = cryoglobulinemic vasculitis; GPA = granulomatosis with polyangiitis (Wegener's).


Table 1 shows the features of 15 cases of PML occurring in patients who received one or more biologic agents for the treatment of an autoimmune rheumatic disease. Six of these cases have previously been in the public domain, either as the subject of communications from the FDA/Genentech (patients B1, B3, B7, and B9) and/or as subjects described in the published medical literature (patients B1 [3], B3 [4], and B7, B9, B10, and B12 [5]). Fourteen patients were treated with rituximab (RTX) and 6 with an anti–tumor necrosis factor (anti-TNF) agent.

RTX was the most recent biologic agent for the treatment of ARD in all 14 confirmed cases of PML associated with its use; 5 patients had been treated with an anti-TNF prior to RTX. The primary indications for RTX were RA in 6 patients (3 of whom had secondary Sjögren's syndrome), SLE in 5, vasculitis in 2, and dermatomyositis in 1.

PML developed after a median of 2 courses of RTX (range 1–4). The median interval between the first infusion of RTX and the development of PML was 12 months (range 1–57 months), and the median interval between the last infusion of RTX and the development of PML was 5 months (range 0–29 months).

Potential confounders for each rituximab-treated case are summarized in Table 4. Four RTX-treated patients were receiving no other immunosuppressive agent at the apparent time of onset of PML (patients B1, B2, B5, and B12); 1 patient was receiving only hydroxychloroquine (patient B1), and 2 patients were receiving only low-dose glucocorticoids (patients B5 and B12). All other patients treated with biologic agents were also receiving 1 or more synthetic disease-modifying agents. Three biologic agent–treated patients were concomitantly receiving cyclophosphamide (CYC) at the time of PML onset: 2 SLE patients (patients B13 and B14) concomitantly treated with intravenous pulse cyclophosphamide and rituximab, and 1 patient (patient B15) with rheumatoid vasculitis treated with daily oral cyclophosphamide and infliximab. Three additional RTX-treated patients had previously received CYC for the treatment of cryoglobulinemic vasculitis (n = 2) and SLE (n = 1).

Table 4. Summary of potential confounders in patients who developed PML following rituximab treatment for ARDs*
  • *

    Results are for patients B1–B14 in Table 1. Progressive multifocal leukoencephalopathy (PML) developed during immunosuppressive therapy for autoimmune rheumatic diseases (ARDs). No patient had a documented diagnosis of human immunodeficiency virus or organ transplantation. SLE = systemic lupus erythematosus.

Current medications              
 Mycophenolate mofetilX
 Radiation therapyXXX

Three RTX-treated patients had a history of malignancy. Patient B3 was diagnosed as having oropharyngeal carcinoma 9 months prior to the development of PML. She underwent radiotherapy and a chemotherapeutic regimen incorporating carboplatin and cetuximab. Patient B8 was treated with radiotherapy and chemotherapy for parotid mucosa–associated lymphoid tissue–type lymphoma diagnosed at least 4 years prior to the development of PML. No further information was provided with regard to the chemotherapy regimen used in this case. Patient B12 had been treated with surgery and radiotherapy (without chemotherapy) for breast cancer 3 years prior to the development of PML.

Of the 6 confirmed cases of PML in patients who had previously received an anti-TNF agent, only 1 was receiving ongoing anti-TNF therapy at the time of the development of PML (patient B15). This patient had received infliximab and CYC concurrently for the treatment of rheumatoid vasculitis. All of the other 5 patients received RTX after the discontinuation of anti-TNF therapy. Two had been treated with infliximab, 2 had been sequentially treated with etanercept and then adalimumab. In these 4 patients, anti-TNF therapy had been discontinued a median of 3 years prior to the development of PML (range 3 months to 5 years). Details of anti-TNF therapy were not provided in the fifth case (patient B2). One patient with PML was previously exposed to anakinra (patient B10). No confirmed cases of PML were reported in association with the use of other biologic agents, such as tocilizumab or abatacept.

The remaining 19 confirmed cases of PML among ARD patients were treated with synthetic DMARDs only (Table 2). Of these patients, 14 had received an alkylating agent (12 CYC and 2 chlorambucil); in 6 of these cases, this was ongoing at the time of development of PML. Of the 34 confirmed cases of PML, 14 were treated with azathioprine and 6 with mycophenolate mofetil (MMF), prior to the onset of PML.


PML has had a significant impact on the development of biologic and other immunosuppressive agents, as it has been the subject of FDA Alert/Dear Healthcare Provider letters with regard to 4 immunosuppressive agents in recent years (natalizumab, RTX, efalizumab, and MMF) and led to the withdrawal from the market of two biologic agents (natalizumab and efalizumab). Natalizumab subsequently returned to the market after institution of a stringent risk-management strategy. Consequently, it is imperative to clearly understand the risk of PML associated with these therapies in order to better inform risk/benefit decision-making. The epidemiology of PML in such settings is difficult to clearly define, however, because PML is rare and likely underdiagnosed and because of the uncertain nature of the risk associated with the underlying indication for immunosuppressive therapy.

We have previously cataloged the reported cases of PML among patients with rheumatic diseases (1) and generated an estimate of the relative frequency of PML associated with specific rheumatic conditions in a nationwide hospital discharge database (2). Because of the limitations of these previous studies, we sought to examine the epidemiology of PML in patients with rheumatic diseases in the FDA AERS database. While this is complementary to our previous efforts and is a necessary exercise, this database has significant limitations. These include the dependence on voluntary reporting, with likely underreporting and reporting bias, the variable quality of the data, and the inability to calculate a true numerator or denominator. Potential confounding factors for the development of PML are identifiable in many of these cases. It is also conceivable that undiagnosed HIV infection or malignancy could have contributed to the risk of PML in some of these cases. Nevertheless, a number of points can be gleaned from these results.

Of the 34 confirmed cases of PML among patients with autoimmune rheumatic diseases, 17 had a diagnosis of SLE. These results are consistent with our previous observations (1, 2) that patients with SLE appear to have a particular susceptibility to the development of PML, relative to patients with other systemic ARDs. No clear mechanism has yet been elucidated for this putative increased risk of PML among SLE patients. While lymphopenia is common among patients with SLE and is related to the disease itself and/or the immunosuppressive therapy, it is neither necessary nor sufficient for the development of PML. Given the near ubiquity of latent JC virus infection, additional, as-yet ill-defined host factors are presumed to play a role in the development of PML.

In this study, 6 reports described the occurrence of PML in patients with ARDs treated with anti-TNF therapies. All of these cases were confounded by treatment with other immunosuppressive agents. Five of these patients received RTX subsequent to the discontinuation of anti-TNF therapy, which was a median of 3 years prior to the development of PML. The sixth case occurred in an infliximab-treated patient and was confounded by concomitant CYC treatment for rheumatoid vasculitis. While an accurate figure for the number of patients who have been treated with anti-TNF agents worldwide is not available, in the exposed population globally, it is estimated to easily exceed two million. Based upon information provided by Centocor Inc., more than 1.1 million patients worldwide have been treated with infliximab alone through February 2009. Consequently, the rarity of the association between anti-TNF therapy and PML observed in the present study is evidence against a causal role of anti-TNF therapy in the development of PML. However, the recent report of an apparently nonconfounded case of PML in an RA patient treated with infliximab sounds a note of caution (6).

The occurrence of PML in patients treated with RTX has been the subject of FDA alerts (7) and several published reports (4, 8–12). Many cases are confounded by the nature of the underlying illness (non-Hodgkin's lymphoma, SLE) and other therapies (other chemotherapeutic and immunosuppressive agents and stem cell transplantation). In a recent retrospective chart review of RTX-treated lymphoma patients, Tuccori et al (13) found an incidence rate of 289/100,000 patient-years, which exceeds the rate observed in patients with HIV infection or B cell chronic lymphocytic leukemia, traditionally considered at highest risk of PML. Our examination of the FDA AERS database revealed 14 confirmed cases of PML in ARDs patients treated with RTX, 8 of whom had not previously been described in the literature. While potential confounders exist for all but 1 of these cases, this discordant signal suggests that RTX, through an as-yet-undefined mechanism, may facilitate reactivation of JC virus in the form of PML.

In this study, 6 PML patients had received RTX for the treatment of RA, 5 of whom had previously been described in the literature (5). One additional patient (patient B4) was described as having RA, but RTX (and prior infliximab) treatment had been initiated for severe dermatomyositis rather than RA. While we are unable to precisely estimate the risk of PML in RTX-treated RA patients, based on the estimated 129,000 RA patients treated with this agent through May 2010 (5), we can estimate a cumulative reporting rate of ∼5/100,000 exposed patients. These data suggest the apparent risk of PML in RTX-treated RA patients currently fits the World Health Organization definition of a “very rare” adverse event (i.e., <1/10,000) (14). This risk is an order of magnitude lower than the risk of PML associated with natalizumab, which is ∼1/1,000 exposed patients (15), and is lower than the risk associated with efalizumab therapy (1/400 patients treated for >3 years) (16). Interestingly, unlike natalizumab and efalizumab, no detectable risk effect for the development of PML has yet been observed from increasing the dosage or the duration of exposure to RTX, further emphasizing our lack of understanding of the potential pathogenic mechanisms involved. Although it is not possible to attribute causality with certainty at this point, each additional case provides further concern for risk, and ongoing vigilance for the development of PML among patients treated with RTX should be maintained.

We identified no cases of PML in patients treated with newer biologic agents, such as tocilizumab or abatacept. However, 2 cases of PML have been recently reported in association with belatacept (17), a second-generation, higher-avidity version of abatacept. Belatacept is a recombinant soluble fusion protein of the extracellular domain of human CTLA-4 with a fragment of a modified Fc domain of IgG1 that differs from abatacept by only 2 amino acids. Belatacept was recently approved by the FDA for the prophylaxis of renal transplant rejection. One case of PML occurred in a patient following renal transplantation (17) and the other in a patient following liver transplantation. Both patients had been treated with other standard immunosuppressive agents for prophylaxis of organ transplant rejection, including MMF, in addition to belatacept.

MMF has also been the subject of a recent FDA alert concerning the risk of PML (18, 19). Adjudication of the risk of PML associated with MMF therapy is also hampered by the therapeutic indication (organ transplantation, SLE) and concomitant and prior immunosuppressive agents received (20). In the current study, 6 MMF-treated patients developed confirmed PML. Five of these had SLE; 12 other SLE patients who had not been treated with MMF also developed PML. All patients had also been treated with other immunosuppressive agents, including CYC. However, in 3 of the cases, MMF therapy was ongoing at the time PML was diagnosed. In summary, there was no signal in excess of that seen with similar immunosuppressive agents such as azathioprine.

It is notable that in 11 of the 34 cases (6 receiving biologic agents and 5 receiving only synthetic agents), inflammatory changes were noted on MRI (contrast enhancement) and/or brain biopsy. These findings are not compatible with the classic descriptions of PML, but are increasingly being recognized as representing an “inflammatory PML” and are well-described in the setting of immune reconstitution inflammatory syndrome (IRIS) following initiation of highly active antiretroviral therapy in the setting of HIV infection. Inflammatory PML generally occurs in the setting of a less profound immunosuppressive state and, consequently, portends a less dismal prognosis than is the case with classic PML. In some cases, however, the degree of inflammation and edema may be of sufficient severity to warrant intravenous glucocorticoid therapy. These findings may hamper the differentiation of PML from the neurologic manifestations of ARDs (e.g., neuropsychiatric lupus, cerebral vasculitis), emphasizing the importance of considering PML in the differential diagnosis of patients with rheumatic disease that develop subacute neurologic deficits. In such cases, PCR for JC virus in the cerebrospinal fluid should routinely be performed, and brain biopsy considered if PCR is repeatedly negative, especially in the setting of unexplained progressive neurologic decline in patients receiving immunosuppressive therapy.

PML is a reported complication of a variety of ARDs and is associated with both synthetic and biologic immunosuppressive agents. PML has been reported with both TNF inhibitors and RTX in the setting of autoimmune diseases. Definitive attribution of causality is not possible, given the small numbers of cases, the potential for reporting bias, and the existence of confounders in many cases. However, the relative paucity of confirmed cases in patients recently treated with anti-TNF agents, despite their widespread use, suggests that a causal relationship is unlikely. In contrast, there is an increasing, specific signal emerging with regard to the association between RTX and the development of PML. Although this is a rare adverse event associated with RTX therapy, the devastating nature of PML mandates continued vigilance, particularly in patients with current or prior exposure to an alkylating agent.


All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Dr. Molloy had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study conception and design. Molloy, Calabrese.

Acquisition of data. Molloy, Calabrese.

Analysis and interpretation of data. Molloy, Calabrese.


The authors acknowledge the assistance of Genentech in obtaining access to the raw data required for this study. Genentech did not participate in analysis of the data, interpretation of the results, or preparation of the manuscript.