Dr. Cacoub has received grant support from Schering-Plough, Roche, Servier, and Gilead, honoraria from Roche and Servier (more than $10,000 each) and from Bristol-Myers Squibb, Sanofi-Aventis, Gilead, and Schering-Plough (less than $10,000 each), and consulting or advisory fees from Roche and Servier (more than $10,000 each) and from Bristol-Myers Squibb, Sanofi-Aventis, Gilead, and Schering-Plough (less than $10,000 each).
Rituximab may form a complex with iGmκ mixed cryoglobulin and induce severe systemic reactions in patients with hepatitis C virus–induced vasculitis
Article first published online: 30 NOV 2009
Copyright © 2009 by the American College of Rheumatology
Arthritis & Rheumatism
Volume 60, Issue 12, pages 3848–3855, December 2009
How to Cite
Sène, D., Ghillani-Dalbin, P., Amoura, Z., Musset, L. and Cacoub, P. (2009), Rituximab may form a complex with iGmκ mixed cryoglobulin and induce severe systemic reactions in patients with hepatitis C virus–induced vasculitis. Arthritis & Rheumatism, 60: 3848–3855. doi: 10.1002/art.25000
- Issue published online: 30 NOV 2009
- Article first published online: 30 NOV 2009
- Manuscript Accepted: 24 AUG 2009
- Manuscript Received: 8 SEP 2008
To report on 6 cases of hepatitis C virus (HCV)–induced mixed cryoglobulinemia (MC) vasculitis in patients who developed severe systemic reactions after rituximab infusion, and to report the results of the in vitro analysis of the underlying immunologic mechanisms.
Twenty-two HCV-infected patients with MC vasculitis received rituximab infusions (a low-dose protocol cycle with 375 mg/m2/week for 4 consecutive weeks in 18 patients and a high-dose protocol cycle with 1,000 mg on days 1 and 15 in 4 patients). Systemic drug reactions following rituximab infusion were recorded and analyzed clinically and immunochemically.
Six of 22 patients (27.3%) experienced systemic drug reactions after rituximab infusion. Four patients developed a severe flare of MC vasculitis 1 or 2 days after rituximab infusion. Two patients developed serum sickness syndrome 7 and 9 days after the first 1,000 mg rituximab infusion. Compared with patients without drug reactions, those with drug reactions had higher mixed cryoglobulin levels (mean ± SD 1.4 ± 0.82 gm/liter versus 0.71 ± 0.77 gm/liter; P = 0.0475) and lower C4 levels (mean ± SD 0.02 ± 0.006 gm/liter versus 0.07 ± 0.07 gm/liter; P = 0.02), and more of them received 1,000 mg high-dose rituximab protocol (50% versus 6.25%; P = 0.046). In vitro immunochemical assays showed that rituximab formed a complex with the cryoprecipitating IgMκ that had rheumatoid factor (RF) activity. Moreover, the in vitro addition of rituximab to serum containing an RF-positive IgMκ type II mixed cryoglobulin was associated with visibly accelerated cryoprecipitation.
In HCV-associated MC vasculitis, rituximab may form a complex with RF-positive IgMκ, leading to accelerated cryoprecipitation and to severe systemic reactions. Rituximab should be administered with caution in MC vasculitis, with use of the 375 mg protocol and plasma exchanges prior to rituximab infusion in patients with high baseline levels of mixed cryoglobulin.
Rituximab is a chimeric IgG1κ immunoglobulin that targets CD20 molecules on B cells. Since 1997, its efficacy has been demonstrated in the treatment of B cell malignant proliferative disease (1). Rituximab is also used as a possible treatment of B cell–related autoimmune diseases, including autoimmune hemolytic anemia, autoimmune thrombocytopenia, rheumatoid arthritis (RA), Sjögren's syndrome (SS), and systemic lupus erythematosus (SLE) (2). Rituximab has recently demonstrated efficacy in mixed cryoglobulinemia (MC) vasculitis, which is mostly due to hepatitis C virus (HCV) infection (3, 4). The overall clinical and biologic tolerance is good. Reported systemic drug reactions include fewer than 20 cases of serum sickness syndrome (5–15). We describe herein 6 patients with HCV-associated MC vasculitis who presented with severe flares of vasculitis or serum sickness syndrome after rituximab infusion, and we report the results of our analysis of the underlying immunopathologic mechanisms.
PATIENTS AND METHODS
Twenty-two HCV-infected patients with biopsy-proven MC vasculitis received rituximab infusions (a low-dose protocol cycle with 375 mg/m2/week for 4 consecutive weeks [375 mg rituximab cycle] in 18 patients and a high-dose protocol cycle with 1,000 mg on days 1 and 15 [1,000 mg rituximab cycle] in 4 patients) and were followed up prospectively. Patients 5 and 6 had previously received a 375 mg rituximab cycle at 11 and 28 months, respectively, before the current 1,000 mg rituximab cycle. Premedication with intravenous methylprednisolone (40 mg), dexchlorpheniramine (8 mg), and acetamenophen (1 gm) was systematically given 30 minutes before each rituximab infusion. The main characteristics of the study population are shown in Table 1. Severe systemic drug reactions following rituximab infusions were recorded and analyzed. They included life-threatening flares of vasculitis and serum sickness syndrome. A life-threatening flare of vasculitis was defined as a flare of vasculitis symptoms that included internal organ involvement (cardiac, renal, digestive, or neurologic). Serum sickness syndrome was defined as the occurrence of a systemic reaction, including fever, arthralgia, rash, or lymphadenopathy, and raised levels of parameters of inflammation (erythrocyte sedimentation rate, C-reactive protein, fibrinogen) occurring 7–14 days after drug infusion (16).
|Age, years||56.7 ± 11.85|
|Women, no. (%)||14 (64)|
|HCV genotype, no. (%)|
|2 and 3||7 (31.8)|
|4 and 5||3 (13.6)|
|HCV viral load, log||5.75 ± 0.58|
|Cryoglobulin type, no. (%)|
|II IgMκ||18 (82)|
|Oligoclonal IgM||2 (9)|
|Cryoglobulin serum level, gm/liter||0.93 ± 0.84|
|C4 serum level, gm/liter||0.056 ± 0.065|
|RF positive, no. (%)||16/19 (84)|
|Organ involvement, no. (%)|
|Peripheral neuropathy||19 (86)|
|B cell NHL||8 (36)|
|Taking 1,000 mg rituximab, no. (%)†||4 (18)|
|Taking 375 mg rituximab, no. (%)‡||18 (81.8)|
Standard immunochemical tests.
Serum cryoglobulin detection and immunochemical typing were performed using a validated immunoblotting method (17). The cutoff for positivity was 0.05 gm/liter, as previously reported (18). Biologic parameters, including C3, C4, IgG, IgA, and IgM levels, were measured at 37°C using the nephelometric method (Dade-Behring, Marburg, Germany); CH50 activity was evaluated using a standardized hemolytic method (Diamedix, Miami, FL), and the rheumatoid factor (RF) titer was measured using an enzyme-linked immunoassay (FR Lisa; BMD, Marne la Vallée, France) with a cutoff for positivity of 20 IU/ml.
Sera stored at −30°C prior to rituximab treatment were used. Nitrocellulose strips were spotted with 1 μl of serial dilutions of rituximab (0.01–0.1 mg/ml). After blocking with skim milk, 1:100-diluted patient sera (i.e., from HCV-positive patients with an RF-negative mixed cryoglobulin, from an HCV-positive patient with no mixed cryoglobulin, from a patient with Waldenström's macroglobulinemia with an RF-negative IgMκ, and from a healthy [HCV-negative] patient) were incubated for 1 hour at 37°C. After washing, horseradish peroxidase–labeled anti-human IgM antibody (MP Biomedical, Solon, OH) or anti-human IgM F(ab′)2 fragment (Jackson ImmunoResearch, West Grove, PA) or anti-human κ light chain or anti-human λ light chain antibodies (whole molecule from Sigma, St. Louis, MO) were added to the nitrocellulose strips and incubated for 1 hour. After washing, 4-chloro-1-naphthol solution containing H2O2 was added, and the reaction was stopped with distilled water.
Western blot assays.
Sera stored at −30°C prior to rituximab treatment were used. Proteins were separated by electrophoresis at 37°C on thin-layer agarose gels (Paragon; Beckman, Gagny, France) and transferred onto nitrocellulose sheets (TransBlot Transfer medium; Bio-Rad, Hercules, CA) by pressure blotting (pressure of 1 kg and then 5 kg, each for 5 minutes). The blots were then oven-dried. After saturation with 50 mg/ml skim milk (Régilait, Lyon, France) in 0.15 moles/liter NaCl for 1 hour at 37°C, the blots were probed with polyclonal antibodies (rabbit antibodies, diluted 1:3,000) (Dako, Trappes, France) specific for light and heavy chains of human Ig. After four 5-minute washes in 0.15 moles/liter NaCl, the strips were incubated for 30 minutes with a 5,000-fold dilution of the appropriate detection antibody (goat anti-rabbit antibody labeled with alkaline phosphatase [EC 184.108.40.206]; Jackson ImmunoResearch). After further washes, enzyme activity was revealed using the appropriate substrate (BCIP/nitroblue tetrazolium; Sigma) prepared just before use. The reaction was stopped with distilled water.
In vitro cryoprecipitation tests with rituximab.
Sera stored at −30°C prior to rituximab treatment were used. We analyzed physical and immunochemical reactions following the addition of rituximab (at a final concentration of 0.25– 1 mg/ml) in tubes containing sera from the 6 patients who presented with a rituximab drug reaction and from 2 patients with RF-positive RA. For each patient, the serum tube in which rituximab was added (rituximab positive) was compared with the tube without rituximab (rituximab negative). In vitro cryoprecipitation was first examined within the first 2 hours following addition of rituximab and after a 7-day incubation at 4°C. IgM, C3, and C4 levels were measured in patient sera before the addition of rituximab and in supernatants of rituximab-negative and rituximab-positive tubes after the 7-day incubation. Cryoglobulin levels were also determined when a cryoprecipitate was evidenced in rituximab-negative and/or rituximab-positive tubes after the 7-day incubation, and results were compared with baseline data.
Categorical variables were compared using Fisher's exact test or chi-square test, and continuous variables were compared using the t-test or Mann-Whitney U test when appropriate. All tests were 2-tailed, and P values less than 0.05 were considered significant. All statistical analyses were performed using MedCalc version 220.127.116.11 (MedCalc, Mariakerke, Belgium).
Clinical and laboratory data.
Severe systemic rituximab-related drug reactions were experienced by 6 of the 22 patients (27.3%). There were 2 different clinical presentations: life-threatening flare of MC vasculitis (n = 4) and serum sickness syndrome (n = 2) (Table 2).
|Drug reaction||Patient||Age in years/sex||Baseline organ involvement||Cryoglobulin type||Cryoglobulin level, gm/liter||C4 level, gm/liter||RF, IU/ml||Rituximab infusion dose, mg||No. of rituximab infusions prior to drug reaction||Delay in drug reaction occurrence, day||Drug reaction organ involvement||Treatment|
|MC flare||1||69/F||P, K, NHL||II IgMκ||1.62||0.01||20||375||2||2||P, C, K, D, F, A||MP, PE, dialysis|
|MC flare||2||68/F||P, N||II IgMκ||0.78||0.02||67||375||2||1||N, A||MP, PE|
|MC flare||3||73/F||P, N, K, NHL||II IgMκ||1.98||0.02||130||1,000||1||1||P, C, K, D, F, A||MP|
|MC flare||4||58/M||P, N, K, D||II IgMκ||2.6||0.02||252||375||2||2||P, D, F, A||MP|
|SS||5†||47/M||N, C, NHL||II IgMκ||1.29||0.03||1,030||1,000||1||7||F, A, P||Spontaneous resolution|
|SS||6†||53/F||P, N||Oligo Mκ||0.33||0.02||20||1,000||1||9||F, A, P||Spontaneous resolution|
The 4 patients who developed a severe acute flare of cryoglobulinemic vasculitis presented with cutaneous (n = 3), cardiac (n = 2), renal (n = 2), digestive (n = 3), and neurologic (n = 1) involvement, 1 day (n = 2) and 2 days (n = 2) after the second 375 mg rituximab infusion (n = 3) or the first 1,000 mg rituximab infusion (n = 1). They were successfully treated with methylprednisolone pulses (n = 4) and plasma exchanges (n = 2) (Table 2).
Two patients (patients 5 and 6; 1 with type II IgMκ mixed cryoglobulin and 1 with oligo Mκ mixed cryoglobulin) developed a typical serum sickness syndrome (fever, arthralgia, purpura, increased serum parameters of inflammation) 7 and 9 days after the first 1,000 mg rituximab infusion, with a spontaneous recovery. Of note, these 2 patients had received 375 mg rituximab infusions at 11 and 28 months, respectively, before this cycle, and those had been well tolerated. Cryoglobulin and C4 serum levels were 1.29 gm/liter and 0.03 gm/liter, respectively, for patient 5 and 0.33 gm/liter and 0.02 gm/liter, respectively, for patient 6. The presence of human antichimeric antibodies was not evaluated.
Overall, the 6 patients who developed severe systemic rituximab-related drug reactions had received 1,000 mg rituximab infusions more often than the 16 patients who did not (3 of 6 [50%] versus 1 of 16 [6.25%]; P = 0.046) and had higher baseline levels of mixed cryoglobulin (mean ± SD 1.4 ± 0.82 gm/liter versus 0.71 ± 0.77 gm/liter; P = 0.0475) and lower baseline levels of C4 (mean ± SD 0.02 ± 0.006 gm/liter versus 0.07 ± 0.07 gm/liter; P = 0.02). Using receiver operating characteristic curve analysis, the best diagnostic performance of C4 serum levels for rituximab-related drug reaction was obtained with a baseline C4 level ≤0.03 gm/liter, leading to a sensitivity of 100%, a specificity of 69%, a positive predictive value (PPV) of 54.5%, and a negative predictive value (NPV) of 100%. The best diagnostic performance of mixed cryoglobulin serum levels was obtained for a baseline mixed cryoglobulin serum level >1.06 gm/liter, leading to a sensitivity of 67%, a specificity of 93%, a PPV of 80%, and an NPV of 87%. All 6 patients had either a mixed cryoglobulin serum level >1.06 gm/liter or a C4 serum level ≤0.03 gm/liter compared with 7 of 16 patients (43.75%) without a rituximab-related drug reaction (P = 0.046).
Underlying immune processes.
During their followup, patients with a rituximab-related drug reaction were assessed for mixed cryoglobulin serum levels before and 24 hours after rituximab infusions (data were not available for 1 patient). Surprisingly, a rapid and large decrease in serum levels of mixed cryoglobulin was evidenced after rituximab infusion (from 0.93 ± 0.7 gm/liter to 0.18 ± 0.14 gm/liter; P = 0.04). Because of the very short time interval, a direct effect of rituximab on mixed cryoglobulin production was thought to be improbable. These results suggested the formation of immune complexes between mixed cryoglobulin and rituximab, which may have rapidly decreased the measurable circulating level of mixed cryoglobulin. To confirm this hypothesis, immunochemical tests were performed to answer 2 questions. First, could rituximab be recognized by mixed cryoglobulin Ig? Second, did the RF activity, the mixed cryoglobulin serum levels, or the rituximab dose play a role?
To determine whether rituximab was recognized and fixed by the mixed cryoglobulin Ig, we used an immunodot method. Our results showed that only sera containing Ig with an RF activity, such as RF-positive IgMκ in type II mixed cryoglobulin and RA-associated RF-positive IgM, were able to recognize the rituximab IgG1κ. The immunodot staining intensity was positively correlated with the concentration of rituximab, suggesting that the recognition of rituximab by RF-positive Ig depends on the dose of rituximab (Figure 1). Sera without IgMκ type II mixed cryoglobulin (i.e., type III, oligoclonal, or type II IgGλ or type II IgMλ mixed cryoglobulin) from HCV-infected patients and from those with monoclonal IgMκ-associated Waldenström's macroglobulinemia were unable to recognize and bind rituximab in the absence of RF activity (Figure 1). These results suggested that the direct interaction between rituximab and mixed cryoglobulin was closely linked to the presence of RF activity. Anti-human IgM F(ab′)2 fragments were also used to demonstrate the presence of RF-positive IgM in the rituximab-positive spots with a significant immunochemical staining (results not shown).
To check whether the recognition of rituximab by RF-positive IgMκ was specific to rituximab, we tested human immunoglobulins and infliximab, which is another human chimeric IgG1κ Ig. Western blots showed that both human immunoglobulins (results not shown) and infliximab (Figure 2) were recognized by the RF-positive IgM from cryoglobulinemic sera or RA sera. This result suggests that the recognition of rituximab by RF-positive IgM is linked to the nonspecific activity of RF-positive IgM rather than to a rituximab-specific process.
Finally, we analyzed physical and immunochemical reactions following the addition of rituximab to tubes containing sera from the 6 patients who presented with a rituximab-related drug reaction and from 2 patients with RF-positive RA. The in vitro addition of rituximab to sera from RF-positive RA patients yielded unremarkable results, with no cryoprecipitation at the end of the 7-day incubation. The in vitro addition of rituximab to HCV-positive sera with an RF-positive IgMκ type II mixed cryoglobulin was associated with the occurrence of a visible accelerated cryoprecipitation (within less than 30 minutes for 2 patients) in rituximab-positive tubes compared with rituximab-negative tubes (Figure 3A). Second, at the end of the 7-day incubation at 4°C, a solid “pancake-like” deposit which hardly dissolved after rewarming at 37°C was evidenced at the bottom of the rituximab-positive tubes (Figure 3B). By Western blot assay, we showed that this “pancake-like” deposit was composed of the cryoprecipitating IgMκ and the rituximab IgG1κ (Figure 4). Third, after dissolution of the “pancake-like” deposit, a significant increase in cryoglobulin levels was evidenced in rituximab-positive tubes compared with rituximab-negative tubes (0.51 ± 0.17 gm/liter versus 0.29 ± 0.16 gm/liter; P = 0.01). Finally, in the supernatants of rituximab-positive tubes, a significant decrease in serum levels was evidenced for IgM (0.72 ± 0.01 gm/liter versus 1.07 ± 0.3 gm/liter; P = 0.03), C3 (0.74 ± 0.15 gm/liter versus 0.87 ± 0.18 gm/liter; P = 0.003), and C4 (0.018 ± 0.02 gm/liter versus 0.03 ± 0.02 gm/liter; P = 0.01).
Taking together the clinical data and the results of the immunodot and in vitro cryoprecipitation assays with rituximab, our results suggest that 1) the presence of an RF activity is the sine qua non condition for the recognition of rituximab by IgM, 2) the in vitro and in vivo formation of physical cryoprecipitating immune complex between RF-positive IgM and rituximab requires the presence of an RF-positive IgM type II mixed cryoglobulin, and 3) the intensity of the cryoprecipitating RF-positive IgM–rituximab immune complex and its harmful clinical effects are associated with the presence of high serum levels of cryoglobulin and with the use of high doses of rituximab.
We report here on the occurrence of severe systemic drug reactions in up to one-fourth of 22 HCV-infected patients with MC vasculitis treated with rituximab infusions. Three major factors were found to be associated with the occurrence of such rituximab-related drug reactions: high baseline levels of cryoglobulin, the 1,000 mg rituximab high-dose infusion, and a high level of complement activation. Immunochemical tests showed that the rituximab-associated drug reactions were associated with an “accelerated” formation of immune complexes between rituximab and RF-positive mixed cryoglobulin in a rituximab dose– and mixed cryoglobulin serum level–dependent manner.
So far, case reports of rituximab-associated systemic drug reactions mainly concerned serum sickness syndromes that typically occurred more than 5 days after rituximab infusion. These reactions mostly affected patients with autoimmune diseases, including SS (n = 2), SLE (n = 3, including 2 cases of autoimmune thrombocytopenia), primary autoimmune thrombocytopenia (n = 1), and autoimmune polyneuropathy (n = 1) (5, 7–10). All patients received the 375 mg rituximab protocol, and some of them were later retreated with rituximab infusion, with good tolerance. The presence of human antichimeric antibodies was evidenced in 1 patient who developed serum sickness syndrome during the second rituximab cycle (11).
In the 5 trials of rituximab in SS, which included 61 patients, 6 cases of serum sickness syndrome (9.8%) were reported (6, 11, 13, 19, 20). In the report by Seror et al (11), 1 patient was retreated with rituximab 11 months later with a good tolerance. Another case of serum sickness syndrome after retreatment with rituximab was reported by Meijer et al (21) in the extended followup of patients in their group's first study (13). In patients with RA, despite the wide use of rituximab in its high-dose protocol (1,000 mg on days 1 and 15), serum sickness syndrome has not been reported as a rituximab-induced adverse event, in contrast to acute reactions that occurred in up to 26% of patients after the first 1,000 mg rituximab infusion (22–24). In trials of rituximab use for pediatric autoimmune thrombocytopenia, the prevalence of serum sickness syndrome ranged from 6% to 13% (12, 14).
In cryoglobulinemic vasculitis, only 2 cases of rituximab-related serum sickness syndrome have been reported (9, 15). One case, previously reported by our group, occurred 2 days after the second 375 mg rituximab infusion and included severe heart and renal involvement that required intensive care procedures and dialysis (15). We now consider it to have been a rituximab-related flare of vasculitis, and we have reported it as such in the present article (patient 1). The life-threatening severity of the present cases, including 2 cases of heart involvement, 3 cases of digestive vasculitis, and 1 case of acute necrotizing mononeuropathy multiplex, has been scarcely reported in typical cryoglobulinemic vasculitis. The worsening of cryoglobulinemic vasculitis following rituximab administration may be misdiagnosed, since revealing symptoms are nonspecific and may be considered to represent the usual course of MC vasculitis.
Few cases of clinical worsening after rituximab infusion have been reported in IgM- and anti–myelin-associated glycoprotein–associated polyneuropathy (25–27). In Waldenström's macroglobulinemia, rituximab is known to induce a paradoxical increase in IgM serum levels in 30–70% of patients. The worsening occurs immediately after completing the rituximab course, and IgM levels return to baseline within 4 months (28). Noronha et al (29) described a patient with Waldenström's macroglobulinemia who experienced severe worsening of neuropathy the day after the first rituximab cycle. The proposed mechanisms of this “flare” phenomenon include B lymphocyte lysis, with resultant release of intracellular paraprotein, or CD20 signaling induced by rituximab. During treatment for non-Hodgkin's lymphoma, few cases of serum sickness syndrome have been reported (30, 31).
With regard to pathophysiologic mechanisms, our results showed that the presence of RF activity is necessary for the binding of rituximab by cryoprecipitating Ig. However, this condition is not sufficient for immune complex formation. Although patients with RF-positive RA are most likely receiving the high-dose 1,000 mg rituximab protocol, no cases of serum sickness syndrome have been reported thus far in RA patients. The in vitro addition of rituximab to sera from patients with RF-positive RA does not induce visible immune complex formation. Our results showed that the formation of immune complexes depends on the rituximab dose and the mixed cryoglobulin serum level.
The level of serum complement activation, appraised by measurement of C3, C4, and CH50 serum levels, may have played a role, since patients who presented with rituximab-related systemic drug reactions had very low serum C4 levels (≤0.03 gm/liter), and cryoglobulinemic vascular damage is mediated by the high affinity of RF-positive IgMκ mixed cryoglobulin for C1q and by the activation of the classical pathway of complement (32). Interestingly, rituximab, an IgG1 Ig, also presents a high affinity for C1q and mediates its therapeutic effects by activating the classical pathway of complement (33). This complement activation, initiated by C1q, leads to C3 and C4 cleavage and consumption and to the formation of the membrane attack complex, which may promote B cell depletion and vascular damage in cryoglobulinemic vasculitis. Hence, cryoglobulinemic vasculitis treated with rituximab may be the “perfect” situation in which all required conditions are present at the same time: high serum levels of RF-positive IgM mixed cryoglobulin, high level of complement activation, and IgG1 rituximab infusion.
Although rituximab infusion may be useful in the treatment of cryoglobulinemic vasculitis, from these clinical and laboratory data, we recommend careful use when mixed cryoglobulin serum levels are high (≥1.00 gm/liter) and C4 serum levels are low (≤0.03 gm/liter). The low-dose rituximab protocol (375 mg/m2 weekly for 4 consecutive weeks or even lower doses ) would be better to use for cryoglobulinemic vasculitis, and plasma exchanges should be performed to lower mixed cryoglobulin serum levels prior to rituximab administration in patients with high serum levels of mixed cryoglobulin.
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. Cacoub 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. Sène, Ghillani-Dalbin, Amoura, Musset, Cacoub.
Acquisition of data. Sène, Ghillani-Dalbin, Musset, Cacoub.
Analysis and interpretation of data. Sène, Ghillani-Dalbin, Musset, Cacoub.
- 23The efficacy and safety of rituximab in patients with active rheumatoid arthritis despite methotrexate treatment: results of a phase IIb randomized, double-blind, placebo-controlled, dose-ranging trial. Arthritis Rheum 2006; 54: 1390–400., , , , , , et al.
- 24Rituximab for rheumatoid arthritis refractory to anti-tumor necrosis factor therapy: results of a multicenter, randomized, double-blind, placebo-controlled, phase III trial evaluating primary efficacy and safety at twenty-four weeks. Arthritis Rheum 2006; 54: 2793–806., , , , , , et al.
- 27Clinical and immunological worsening in a patient affected with Waldenstrom macroglobulinemia and anti-mag neuropathy after treatment with rituximab. Haematologica 2006; 91 Suppl: ECR17., , , , , , et al.