Effectiveness of rituximab treatment in primary Sjögren's syndrome: A randomized, double-blind, placebo-controlled trial


  • ClinicalTrials.gov identifier: NCT00363350.



To study the efficacy and safety of B cell depletion with rituximab, a chimeric murine/human anti-CD20 monoclonal antibody, in patients with primary Sjögren's syndrome (SS) in a double-blind, randomized, placebo-controlled trial.


Patients with active primary SS, as determined by the revised American–European Consensus Group criteria, and a rate of stimulated whole saliva secretion of ≥0.15 ml/minute were treated with either rituximab (1,000 mg) or placebo infusions on days 1 and 15. Patients were assigned randomly to a treatment group in a ratio of 2:1 (rituximab:placebo). Followup was conducted at 5, 12, 24, 36, and 48 weeks. The primary end point was the stimulated whole saliva flow rate, while secondary end points included functional, laboratory, and subjective variables.


Thirty patients with primary SS (29 female) were randomly allocated to a treatment group. The mean ± SD age of the patients receiving rituximab was 43 ± 11 years and the disease duration was 63 ± 50 months, while patients in the placebo group were age 43 ± 17 years and had a disease duration of 67 ± 63 months. In the rituximab group, significant improvements, in terms of the mean change from baseline compared with that in the placebo group, were found for the primary end point of the stimulated whole saliva flow rate (P = 0.038 versus placebo) and also for various laboratory parameters (B cell and rheumatoid factor [RF] levels), subjective parameters (Multidimensional Fatigue Inventory [MFI] scores and visual analog scale [VAS] scores for sicca symptoms), and extraglandular manifestations. Moreover, in comparison with baseline values, rituximab treatment significantly improved the stimulated whole saliva flow rate (P = 0.004) and several other variables (e.g., B cell and RF levels, unstimulated whole saliva flow rate, lacrimal gland function on the lissamine green test, MFI scores, Short Form 36 health survey scores, and VAS scores for sicca symptoms). One patient in the rituximab group developed mild serum sickness–like disease.


These results indicate that rituximab is an effective and safe treatment strategy for patients with primary SS.

Sjögren's syndrome (SS) is a systemic autoimmune disease that is characterized by chronic inflammation of the salivary and lacrimal glands, resulting in xerostomia and keratoconjunctivitis sicca in ∼95% of patients (1). These symptoms are frequently accompanied by extraglandular manifestations such as Raynaud's phenomenon, arthritis, arthralgia, and myalgia, and 85% of patients experience severe fatigue. Moreover, B cell hyperactivity, reflected by increased serum levels of IgG and IgM rheumatoid factor (RF) and the presence of anti-SSA and anti-SSB autoantibodies, is a common finding in SS. Furthermore, SS has a large impact on health-related quality of life, employment, and disability, as reflected by lower Short Form 36 (SF-36) health survey scores, reduced employment rates, and higher rates of disability in patients with SS compared with the general population (1).

To date, no targeted systemic treatment has been available for SS. In pilot trials, however, it has been shown that rituximab, a chimeric murine/human anti-CD20 monoclonal antibody that binds to the B cell surface antigen CD20, might improve subjective and objective symptoms related to primary SS for at least 6–9 months (2, 3). On the basis of these promising results, a randomized, double-blind, placebo-controlled trial was performed to investigate the efficacy and safety of rituximab in the treatment of patients with primary SS.


Study design.

This was a prospective, single-center, randomized, double-blind, placebo-controlled study. The study protocol was approved by the institutional review board of the University Medical Center Groningen. All patients provided their written informed consent.


All patients were age ≥18 years and fulfilled the American–European Consensus Group criteria for primary SS (4). Eligibility criteria were a rate of secretion of stimulated whole saliva of ≥0.15 ml/minute and positivity for autoantibodies (IgM-RF ≥10 IU/ml and anti-SSA and/or anti-SSB autoantibodies). In addition, results from a salivary gland biopsy performed within 12 months before inclusion and showing the characteristic features of SS had to be available (5). During the study, patients were asked to use reliable methods of contraception. Patients with either primary or secondary SS who had been treated previously with other monoclonal antibodies were excluded. Treatment with prednisone and hydroxychloroquine had to be discontinued at least 1 month before baseline, and treatment with methotrexate, cyclophosphamide, cyclosporine, azathioprine, and other disease-modifying antirheumatic drugs had to be discontinued at least 6 months before baseline. Patients were allowed to use artificial tears and artificial saliva, but the regimen had to remain identical during followup. The use of these substitutes had to be stopped 1 day prior to each assessment.

All patients underwent electrocardiography and chest radiography at baseline. Patients with a history of any malignancy or with underlying cardiac, pulmonary, metabolic, renal, or gastrointestinal conditions, chronic or latent infectious diseases, or immune deficiency were excluded.

Drug administration.

Twenty patients were treated with intravenous (IV) infusions of 1,000 mg rituximab (Roche, Woerden, The Netherlands) and 10 patients were treated with IV infusions of placebo on days 1 and 15. To minimize side effects (infusion reactions, serum sickness), all patients were pretreated with methylprednisolone (100 mg IV), acetaminophen (1,000 mg orally), and clemastine (2 mg IV), and received 60 mg oral prednisone on days 1 and 2, 30 mg on days 3 and 4, and 15 mg on day 5 after each infusion.

Outcome parameters.

Definition of end points.

The primary end point was defined as a significant improvement in the secretion of stimulated whole saliva (flow rate, in ml/minute) in the rituximab group compared with the placebo group. Secondary end points were measurements of salivary/lacrimal function and immunologic and subjective variables. All variables were assessed at baseline (within 4 weeks before treatment) and at 5, 12, 24, and 48 weeks after treatment.

Determination of salivary gland function

Whole, parotid, and submandibular/sublingual saliva samples were collected in a standardized manner and at a fixed time of the day (in this study, between 1:00 and 4:00 PM), in order to minimize fluctuations related to a circadian rhythm of salivary secretion (6, 7) and composition. Glandular saliva was collected from both individual parotid glands by use of Lashley cups, and submandibular/sublingual saliva was collected simultaneously by syringe aspiration from the area with the orofices of the submandibular excretory ducts. Unstimulated saliva was collected in the first 5 minutes, followed by collection of stimulated saliva after the salivary glands had been stimulated for 10 minutes. The salivary glands were stimulated with citric acid solution (2%), applied with a cotton swab to the lateral borders of the tongue every 30 seconds. Flow rates were calculated and the composition of the saliva was analyzed according to the methods described by Burlage et al and Kalk et al (8–10).

Determination of lacrimal gland function

Lacrimal gland function was evaluated by performing the Schirmer's test, the lissamine green test, and breakup time (BUT) (11). Schirmer's test I (without anesthesia) was carried out by placing a filter strip in the lower fornix of the conjunctiva of the eye. The amount of wetting was measured after 5 minutes. The lissamine green test was performed by instillation of 1% lissamine green in both eyes. After 1 or 2 full blinks, the intensity of staining of both medial and lateral bulbar conjunctiva and the cornea was scored, with a maximum score of 9 points (up to 3 points for each section [1 = sparsely scattered, 2 = densely scattered, 3 = confluent]). The BUT is the interval between a complete blink and the appearance of the first randomly distributed dry spots and is assessed by instilling a 1% fluorescein solution in the fornix of both eyes. The patient was asked to blink a few times, after which the intervalin seconds between the last blink and the first break in the tear film was measured.

Laboratory assessments

Laboratory assessments included serum biochemical analyses and determination of the complete blood cell count. Levels of immunoglobulins (IgG, IgA, and IgM) and IgM-RF were measured by nephelometry. Numbers of circulating CD19+, CD4+, and CD8+ T cells were quantified with the use of a FACSCalibur flow cytometer in TruCOUNT tubes (Becton Dickinson, Mountain View, CA). The absolute T cell number was determined by comparing the number of cellular events with that of bead events, analyzed using CellQuest software (Becton Dickinson).

Subjective assessments

Patients completed the Multidimensional Fatigue Inventory (MFI) (12) and the SF-36 health survey (13). In addition, a 100-mm visual analog scale (VAS) was used for rating oral and ocular sicca symptoms.

Extraglandular manifestations

Arthralgia, arthritis, renal involvement, esophageal involvement (confirmed by esophageal scintigraphy), polyneuropathy, Raynaud's phenomenon, tendomyalgia, and vasculitis were defined as extraglandular manifestations. At each visit, extraglandular manifestations were scored as present or not present, according to protocol.

Definition of serum sickness

Serum sickness was defined as the development of fever, lymph node swelling, purpura, myalgia, arthralgia, thrombocytopenia, and proteinuria, as well as a decrease in complement levels. Serum sickness–like disease was defined as the development of some of these symptoms of serum sickness.

Sample size.

Based on a formal sample size calculation, 30 patients were included, of whom 20 were assigned to receive rituximab and 10 to receive placebo. The patients were randomly assigned by staff in the pharmacy department to 1 of the 2 treatment arms in a 2:1 ratio (rituximab:placebo) in blocks of 3, using a random-number generator on a computer. The study investigators (who also provided care and assessed the outcome variables) and patients were blinded to the assigned study medication. The code was revealed to the investigators after followup of all patients was completed. Because of the double-blind design, we assumed a 5% rate of false-positive findings among the patients in the placebo group who displayed clinical signs of serum sickness. This resulted in an obligation to terminate the trial if 2 patients developed clinical signs of serum sickness after the first or second infusion within the first 9 patients, or if 3 patients developed clinical signs of serum sickness after the first or second infusion within the first 29 patients. If, for any reason, the protocol was terminated, patients were not replaced.

Statistical analysis.

All data analyses were carried out according to a preestablished plan. To compare treatment effects over time between the 2 treatment groups, repeated-measures analysis of variance was performed. To determine whether an improvement had occurred over time relative to baseline, repeated-measures analysis of covariance was performed to evaluate changes from baseline. Statistical analyses performed on secondary end points were considered to be explorative in nature, and therefore no corrections were made for multiple comparisons. The assumptions on data homogeneity were met. If data were not normally distributed, a log-transformation was performed on the data prior to statistical analysis, or a distribution-free alternative was used.


Patient distribution.

Between August 2006 and September 2007, 30 patients were randomly assigned to a treatment group (Figure 1). The baseline characteristics of the patients are summarized in Table 1. Six patients were taking medication that had to be discontinued before study inclusion, in accordance with the inclusion criteria.

Figure 1.

Randomization of patients with primary Sjögren's syndrome to 1 of the 2 treatment groups in the randomized, double-blind, placebo-controlled trial of rituximab. Of a cohort of 300 patients, a preselection of 61 patients was made, based on last available sialometry, IgG, anti-SSA positivity, anti-SSB positivity, and rheumatoid factor data.

Table 1. Baseline characteristics of the patients in the rituximab and placebo treatment groups*
VariablePlacebo (n = 10)Rituximab (n = 20)
  • *

    Except where indicated otherwise, values are the number (%) of patients. RF = rheumatoid factor.

  • P < 0.05 versus placebo.

Age, mean ± SD years43 ± 1743 ± 11
No. female/no. male1019/1
Disease duration, mean ± SD months67 ± 6363 ± 50
IgG, mean ± SD gm/liter21 ± 723 ± 8
IgM-RF, mean ± SD IU/ml221 ± 245102 ± 79
Anti-Ro/SSA positive10 (100)20 (100)
Anti-La/SSB positive8 (80)14 (70)
Parotid gland swelling10 (100)17 (85)
Whole saliva flow, ml/minute  
 Unstimulated0.06 ± 0.090.17 ± 0.19
 Stimulated0.42 ± 0.260.70 ± 0.57
Extraglandular manifestation  
 Arthralgia5 (50)15 (75)
 Arthritis0 (0)6 (30)
 Renal involvement0 (0)2 (10)
 Esophageal involvement1 (10)0 (0)
 Peripheral polyneuropathy0 (0)1 (5)
 Raynaud's phenomenon6 (60)11 (55)
 Tendomyalgia8 (80)17 (85)
 Vasculitis3 (30)6 (30)
 Thyroid dysfunction0 (0)1 (5)
Use of artificial tears8 (80)14 (70)
Use of artificial saliva2 (20)2 (10)

Efficacy (Table 2).

Salivary gland function.

The stimulated whole saliva flow rate (the primary end point) (Figure 2A) significantly improved in the rituximab group (P = 0.018 at week 5 and P = 0.004 at week 12, versus baseline), while in the placebo group these values significantly decreased from baseline, which is consistent with the natural progression of the disease. A significant difference in the mean change from baseline to week 12 in the stimulated whole saliva flow rate was found between the groups (P = 0.038). The unstimulated whole saliva flow rate (Figure 2B) and the submandibular/sublingual flow rate (results not shown) also significantly increased from baseline in the rituximab group.

Table 2. Results of laboratory, functional, and subjective assessments in the rituximab and placebo treatment groups*
VariableBaselineWeek 5Week 12Week 24Week 36Week 48
  • *

    Values are the mean ± SD (median). Due to missing data, the differences between mean values in this table differ slightly from the means of differences shown in Figure 2. RF = rheumatoid factor; MFI = Multidimensional Fatigue Inventory; SF-36 = Short Form 36; VAS = visual analog scale.

  • †, ‡

    Data are not normally distributed.

  • P < 0.05 versus baseline in the same treatment group, by analysis of covariance.

  • §

    P < 0.05 versus change from baseline in the placebo group, by analysis of variance.

Whole saliva flow, ml/minute            
 Unstimulated0.06 ± 0.09 (0.03)0.17 ± 0.19 (0.08)0.09 ± 0.07 (0.08)0.24 ± 0.22 (0.20)0.05 ± 0.05 (0.04)0.23 ± 0.22 (0.19)0.08 ± 0.08 (0.09)0.22 ± 0.25 (0.14)0.07 ± 0.09 (0.02)0.16 ± 0.15 (0.11)0.05 ± 0.04 (0.04)0.18 ± 0.18 (0.13)
 Stimulated0.42 ± 0.26 (0.36)0.70 ± 0.57 (0.47)0.41 ± 0.24 (0.37)0.84 ± 0.71 (0.48)0.28 ± 0.17 (0.25)0.87 ± 0.87 (0.56)§0.36 ± 0.28 (0.24)0.74 ± 0.60 (0.52)0.29 ± 0.18 (0.26)0.64 ± 0.58 (0.44)0.28 ± 0.21 (0.22)0.66 ± 0.71 (0.42)
Lacrimal gland function            
 Schirmer's test, mm/5 minutes7 ± 9 (3)11 ± 11 (7)7 ± 11 (4)10 ± 9 (10)6 ± 5 (5)11 ± 10 (11)8 ± 8 (6)12 ± 12 (5)7 ± 7 (5)11 ± 10 (7)5 ± 5 (6)10 ± 11 (7)
 Lissamine green test4 ± 1 (4)3 ± 2 (4)5 ± 1 (5)3 ± 2 (3)4 ± 2 (4)3 ± 2 (3)4 ± 2 (4)2 ± 2 (2)4 ± 2 (4)2 ± 2 (2)4 ± 2 (4)2 ± 3 (1)
 Tear breakup time, seconds3 ± 2 (3)6 ± 2 (6)3 ± 1 (3)6 ± 3 (6)3 ± 2 (3)5 ± 3 (5)5 ± 2 (6)6 ± 3 (7)5 ± 3 (5)7 ± 3 (8)4 ± 3 (4)6 ± 3 (8)
B cells, 109/liter0.27 ± 0.12 (0.28)0.21 ± 0.17 (0.18)0.20 ± 0.09 (0.17)0.00 ± 0.00 (0.00)§0.25 ± 0.10 (0.27)0.01 ± 0.03 (0.00)§0.28 ± 0.11 (0.26)0.05 ± 0.08 (0.03)§0.28 ± 0.12 (0.31)0.10 ± 0.08 (0.08)§0.33 ± 0.15 (0.37)0.17 ± 0.10 (0.15)§
IgM-RF, IU/ml221 ± 245 (108)102 ± 79 (83)162 ± 175 (96)55 ± 36 (53)156 ± 138 (102)44 ± 30 (36)§258 ± 260 (113)45 ± 34 (32)§253 ± 256 (119)71 ± 68 (54)§225 ± 199 (126)103 ± 103 (72)
MFI, general fatigue14 ± 5 (17)16 ± 4 (18)11 ± 5 (12)15 ± 4 (16)13 ± 5 (14)13 ± 4 (13)12 ± 5 (12)13 ± 4 (12)14 ± 4 (14)14 ± 4 (14)14 ± 6 (17)15 ± 4 (16)
SF-36 total score64 ± 17 (65)52 ± 20 (53)70 ± 17 (70)56 ± 18 (52)67 ± 15 (71)63 ± 15 (65)72 ± 16 (82)67 ± 16 (70)63 ± 16 (65)60 ± 17 (64)62 ± 17 (62)55 ± 18 (55)
VAS score, oral dryness59 ± 28 (62)55 ± 28 (61)50 ± 28 (53)47 ± 27 (53)53 ± 30 (60)40 ± 27 (40)64 ± 27 (74)34 ± 27 (46)68 ± 26 (79)51 ± 28 (61)69 ± 25 (76)50 ± 28 (53)
VAS score, ocular dryness65 ± 27 (63)59 ± 29 (68)55 ± 28 (52)49 ± 28 (51)61 ± 25 (54)48 ± 29 (47)68 ± 24 (74)41 ± 28 (43)70 ± 27 (72)46 ± 27 (52)§76 ± 19 (80)46 ± 28 (55)§
Figure 2.

Mean values of absolute change from baseline (A, B, andD–I) and mean absolute number of B cells (C) in the rituximab group compared with the placebo group. The primary end point was A, the rate of secretion of stimulated whole saliva, while the secondary end points were B, the rate of secretion of unstimulated whole saliva, C, absolute number of B cells, D, rheumatoid factor levels, E, Multidimensional Fatigue Inventory (MFI) score for general fatigue, F, Short Form 36 (SF-36) total score, G, visual analog scale (VAS) score for oral dryness, H, VAS score for ocular dryness, and I, mean number of extraglandular manifestations (EGM) per patient. ∗ = P < 0.05 versus baseline.

Lacrimal gland function.

The LG test showed significant improvement in lacrimal gland function in the rituximab group from baseline to weeks 5–48. However, the Schirmer's test and BUT test revealed no significant changes in lacrimal gland function in either group.

Changes in laboratory variables.

B cells were completely depleted after the first infusion in patients treated with rituximab (Figure 2C). In contrast, no significant changes in the mean absolute number of B cells were found in the placebo group. In the patient who developed serum sickness (see data on safety assessments below), who received only 1 infusion of rituximab, B cells reappeared within 12 weeks after treatment. In the other 19 rituximab-treated patients, B cells returned within 24–48 weeks after treatment, although B cell levels still had not returned to baseline values by week 48. Significant differences in the mean change in absolute B cell count from baseline to weeks 5, 12, 24, 36, and 48 were found between the groups (each P < 0.05). No significant changes were found in the levels of CD4+ and CD8+ T cells in either the rituximab group or placebo group.

Levels of RF (Figure 2D) decreased significantly in the rituximab group over week 5 to week 36, whereas in the placebo group, the RF levels decreased significantly only at week 5. Significant differences in the mean change in RF levels from baseline between the groups were found at weeks 12, 24, and 36 (each P < 0.05). The same patterns of change were found for the levels of IgG, IgM, and IgA in each group (results not shown).

Changes in subjective measurements.

The MFI and SF-36 scores showed the strongest improvements in the rituximab group (Figures 2E and F). Compared with that in the placebo group, patients receiving rituximab showed a significant change in the MFI score, showing decreased scores for reduced activity from baseline to week 36 (P = 0.023) and for reduced motivation from baseline to week 12 (P = 0.039). In addition, in patients receiving rituximab, there was significant improvement in the SF-36 score for vitality from baseline to week 36 (P = 0.013). Moreover, all VAS scores for oral and ocular sicca symptoms improved in the rituximab group (Table 2 and Figures 2G and H), whereas VAS scores in the placebo group only showed a significant improvement at week 5. Significant differences in the mean change in VAS scores from baseline were observed between the groups, in that patients receiving rituximab reported improvement in the ratings for dry mouth during the night at weeks 24, 36, and 48 and in the ratings for dry eyes at weeks 36 and 48 in the rituximab group (each P < 0.05).

Extraglandular manifestations.

At baseline, there were no differences in the number of extraglandular manifestations between the rituximab group and placebo group (Figure 2I). The number of reported extraglandular manifestations (number reported as present) significantly decreased in the rituximab group compared with the placebo group for tendomyalgia at weeks 12 and 36 (P = 0.029) and for vasculitis at week 24 (P = 0.030). In addition, there was a strong tendency toward a significant decrease in the number of reported symptoms of Raynaud's phenomenon (P = 0.057), tendomyalgia (P = 0.074), and arthralgia (P = 0.058) from baseline to week 24 in patients receiving rituximab. Six patients in the rituximab group had symptoms of arthritis at baseline; this resolved in 4 patients during followup. In the placebo group, no patients had symptoms of arthritis at baseline; however, 3 patients developed symptoms during followup. One patient with decreased thyroid function before rituximab treatment showed a normalization of thyroid function without additional thyrostatic supplementation. Renal function remained stable during followup (2 patients had renal tubular acidosis, and both were treated with rituximab). Clinical symptoms of polyneuropathy (in 1 patient in the rituximab group) improved after 12 weeks of followup.

Safety (Table 3).

Serum sickness.

One female patient with diabetes developed a mild serum sickness–like disease, which was identified 14 days after the first infusion of rituximab. The patient developed fever, purpura on both legs, and arthralgia, and she was admitted to the hospital in order to control her serum glucose levels during IV administration of corticosteroids and nonsteroidal antiinflammatory drugs. She recovered completely within a few days, without developing human antichimeric antibodies. The second infusion of rituximab was not administered. This patient had not been treated with any immunosuppressive drug previously. None of the 6 patients who had discontinued immunosuppressive drugs 1–6 months prior to rituximab treatment developed serum sickness–like disease.

Table 3. Adverse events observed in patients following treatment with rituximab as compared with placebo*
EventPlacebo (n = 10)Rituximab (n = 20)
  • *

    Values are the number (%) of patients.

Early infusion reaction02 (10)
Late infusion reaction02 (10)
Serum sickness01 (5) 
Infections within 2 weeks after infusion  
 Upper airway infection01 (5) 
 Parvovirus01 (5) 
Infections during 48 weeks of followup  
 Otitis media02 (10)
 Upper airway infection4 (40)4 (20)
 Recurrence of ocular toxoplasmosis01 (5) 
 Parotid gland infection03 (15)
 Recurrence of herpes zoster1 (10)0
 Epstein-Barr virus1 (10)0
 Rubella1 (10)0


A total of 12 infections were reported by 11 patients in the rituximab group, while 4 patients in the placebo group reported a total of 7 infections. The rates of infection were 76 and 65 events per 100 patient-years for the placebo and rituximab groups, respectively. None of the infections required hospitalization. No opportunistic infections were observed.


This study showed that rituximab-induced B cell depletion can be considered an effective and safe treatment strategy for patients with primary SS. B cell depletion resulted in improvement of objective and subjective parameters of disease activity in patients with primary SS for at least 6–9 months. Among the end points, salivary gland function improved, fatigue diminished, and the number of extraglandular manifestations was reduced.

Rituximab has already been shown to be a safe and effective treatment for rheumatoid arthritis (RA), as shown by a decrease in disease activity, diminished radiologic progression of the disease, and improved quality of life in patients with RA (14–16). Previously, the utility of rituximab for the treatment of SS had only been investigated in a few open-label, phase II studies and in 1 randomized, double-blind, placebo-controlled study. Results from open-label studies, in terms of objective and subjective variables, were promising (2, 3), as was the improvement of systemic features (17). Although the duration of the treatment effect differed between the trials, in all trials a significant effect occurred 12–24 weeks after treatment. In a previous randomized, double-blind, placebo-controlled study of rituximab treatment of SS, a significant improvement in fatigue (the primary end point) was noted as compared with the values at baseline in the rituximab group, but there were no significant changes in the secondary end points assessing glandular manifestations (unstimulated salivary flow rate and Schirmer's test results) (18). Moreover, the study by Dass et al (18) used an objective eye test for lacrimal gland function that was less accurate (the Schirmer's test); the rose bengal score and LG test are considered to be more accurate (11). This fact, together with the small number of patients included in that trial (8 receiving rituximab, 9 receiving placebo), might explain the lack of significant improvement in glandular manifestations following rituximab treatment.

In our trial, most significant improvements in the end points associated with rituximab treatment were observed between 12 weeks and 36 weeks following treatment. In contrast, improvement of most of the variables observed in patients in the placebo group occurred 5 weeks after the first infusion. We hypothesize that the improvements observed after placebo treatment were related to the treatment with prednisolone, which had been administered before and during the days after the infusions, although data are inconclusive regarding the effect of prednisolone on SS symptoms. Although results of a previous study indicated a significant increase in whole saliva flow during the use of low-dose prednisolone (19), other studies noted no significant improvement in glandular function (20, 21).

The flow rate of stimulated whole saliva provides a general indication of overall salivary glandular function, which is an important outcome in a disease that specifically affects the salivary glands. Pijpe et al (3) reported the occurrence of a significant increase in the stimulated whole saliva flow rate in rituximab-treated patients with primary SS whose stimulated salivary flow rate was >0.10 ml/minute at baseline. These patients also showed significant improvement in such subjective parameters as mouth dryness, arthralgia, physical functioning, vitality, and most domains of the MFI. In other words, patients with some residual secretory potential may benefit the most from rituximab treatment. The secretory potential at baseline might even be used to identify those patients who would be considered to be a good responder to rituximab treatment. Therefore, the stimulated whole saliva flow rate was chosen as the primary end point of our study. As a cutoff value, a stimulated whole saliva flow rate of ≥0.15 ml/minute was chosen, since this is a flow rate that discriminates patients showing increasing disease activity (e.g., progressive loss of secretory function) and patients with end-stage primary SS (21). In our study, we observed an increase in salivary flow in the rituximab group that exceeded the intrapatient variability observed for repeated collections of saliva (8). This increase was also reflected in the improvements in subjective scores for dry mouth, which indicates that these changes were clinically meaningful in the patients. The nonsignificant baseline difference between the groups for the stimulated whole saliva flow rate was caused by high salivary flow rates in a few patients before inclusion. All patients in the study were required to have a stimulated whole saliva flow of ≥0.15 ml/minute. This meant that all patients had a clinically relevant functional secretory salivary gland capacity. Our pilot study revealed that no relevant improvement in salivary gland function can be expected in patients with little or no secretory potential at baseline.

In clinical trials of rituximab in patients with RA, the number of reported (serious) infections and infusion reactions is within the range expected for patients with RA treated with biologic agents. Therefore, the risk:benefit ratio is considered to be good regarding rituximab treatment of RA (22). In clinical trials of rituximab treatment for other autoimmune diseases (including SS), the reported numbers of infusion reactions and infections have varied widely; this is possibly due to variability in how these adverse events are defined or to the small numbers of patients. The incidence of infusion reactions and infections reported for the rituximab group in this trial was largely comparable with that in the placebo group, and was lower or within the same range as that reported previously (23). Moreover, the rate of infections per 100 patient-years was lower compared with the previously reported rate in RA patients treated with rituximab. This might be explained by the fact that our patients did not have any other immunosuppressive therapy (24).

When compared with patients with lymphoma, patients with RA, and patients with systemic lupus erythematosus (SLE) treated with rituximab, patients with primary SS treated with rituximab develop serum sickness–like disease more frequently (6–27%) (25). A therapy-related explanation for this phenomenon might be that patients with RA and those with SLE usually receive or have received higher doses of steroids and/or other immunosuppressive drugs, in addition to rituximab, whereas our patients with primary SS received no other medication, except a 5-day period of steroids after IV administration of rituximab. Another therapy-related explanation is that patients with RA and those with SLE often have been exposed to intensive immunosuppressive regimens, including treatment with biologic agents, before they undergo treatment with rituximab, whereas our patients with primary SS were far more likely to have never taken such medications at the time of rituximab treatment. The higher susceptibility for serum sickness could also be inherent to the disease itself. The patients with primary SS in this trial, as well as in our pilot trial (3), who developed serum sickness were more likely to have an active, early, and progressive form of SS. It is possible that such patients with primary SS are more prone to develop serum sickness. Furthermore, hypergammaglobulinemia is common in primary SS, which could make these patients prone to the development and deposition of immune complexes and, thus, to serum sickness–like disease (18).

Because of the higher risk of developing serum sickness–like disease in patients with SS, we decided to increase the steroid dose. Only 1 patient in the current study developed serum sickness–like disease (5%), which is considerably lower than the incidence reported in our open-label study (27%) (3). Based on these findings, we would recommend administering 100 mg methylprednisolone immediately prior to each infusion of rituximab. The oral regimen of prednisolone in the days following each infusion is a point of interest and should be explored in future trials. The administration of higher doses of prednisolone in the days following infusion, such as is performed during lymphoma treatment, should also be considered.

This study indicates that rituximab treatment could be effective for patients who have active primary SS and remaining salivary gland secretory potential, as well as for primary SS patients with extraglandular manifestations. Future trials of rituximab treatment for patients with primary SS are warranted, in which larger groups of patients should be included and less-strict inclusion criteria (e.g., no restriction to those with salivary gland function ≥0.15 ml/minute and autoantibody positivity) should be used, in order to be able to extrapolate the results to a larger group of patients with primary SS. In addition to the defined inclusion criteria, attention should be given to the criteria used for response to treatment. Activity scores for primary SS have now been developed and need validation. These scores should be included in the response criteria to be used in future trials.

Based on the promising results of this study and our prior study on retreatment with rituximab, which resulted in a beneficial effect comparable with that of the first treatment with this biologic agent (26), a maintenance therapy with rituximab infusions every 6–9 months may be a reasonable approach. Advantages of maintenance therapy might be a reduction or even arrest of disease progression and improvement of quality of life for a long period. This improvement will be a great achievement in patients with SS, since SS has a large impact on health-related quality of life, employment, and disability (1). A threat might be the long-term side effects (thus-far unknown) of repeated B cell depletion. The timing of retreatment could be based on return of symptoms; however, retreatment just before return of symptoms would even be better.

In conclusion, the results of this study indicate that rituximab could be an effective and safe treatment strategy for patients with primary SS. B cell depletion resulted in improvement of the primary end point, the rate of stimulated whole saliva secretion. Explorative analyses also showed improvements, of at least 6–9 months' duration, in the objective and subjective secondary end points of disease activity. Since primary SS has a great impact on health-related quality of life, employment, and disability (1), it is worthwhile to further explore the role of rituximab in a large-size, randomized, controlled trial.


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. Vissink 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. Meijer, Vissink, Spijkervet, Brouwer, Kallenberg, Bootsma.

Acquisition of data. Meijer, Meiners, Spijkervet, Abdulahad, Kamminga, Brouwer, Bootsma.

Analysis and interpretation of data. Meijer, Meiners, Vissink, Abdulahad, Kallenberg, Bootsma.


We are grateful to Janita Kuiper, Philip M. Kluin, Jaqueline E. van der Wal, Khaled Mansour, Gustaaf W. van Imhoff, and Justin Pijpe for their support and meaningful discussions.