| |

Received: 10 October 2018    Revised: 8 January 2019    Accepted: 22 January 2019

DOI: 10.1111/1756-185X.13511

ORIGINAL ARTICLE

Real‐world effectiveness and safety of rituximab in the

treatment of rheumatoid arthritis: A single‐center experience in
Taiwan

Kai‐Chun Wang1,2  | Hsien‐Tzung Liao1,2,3  | Wei‐Sheng Chen1,2 | Chien‐Chih Lai1,2,4  |

1,2 1,2 1,2
Chung‐Tei Chou  | Ming‐Han Chen  | Chang‐Youh Tsai

1
Division of Allergy, Immunology
and Rheumatology, Department of Abstract
Medicine, Taipei Veterans General Hospital, Aim: To assess the real‐world effectiveness and safety of rituximab (RTX) at
Taipei, Taiwan
2 24 months in patients with established rheumatoid arthritis (RA) and to identify pre‐
Faculty of Medicine, National Yang‐Ming
University School of Medicine, Taipei, dictors of low disease activity/remission and a good European League Against
Taiwan
Rheumatism (EULAR) response.
3
Division of Allergy, Immunology and
Rheumatology, Department of Internal
Methods: Seventy RTX‐treated RA patients were enrolled. Predictors for low disease
Medicine, School of Medicine, College of activity/remission and a good EULAR response at 24 months were identified by mul‐
Medicine, Taipei Medical University, Taipei,
Taiwan
tivariate analyses.
4
Institute of Clinical Medicine, National Results: At 24 months, the mean Disease Activity Score of 28 joints—erythrocyte
Yang‐Ming University School of Medicine, sedimentation rate (DAS28‐ESR) decreased from 6.88 
± 0.85 at baseline to
Taipei, Taiwan
3.47 ± 0.85. Twenty‐nine patients (41.4%) reached low disease activity/remission,
Correspondence while all patients had a moderate/good EULAR response. After adjustment by multi‐
Ming‐Han Chen and Chang‐Youh Tsai,
Division of Allergy, Immunology, and variate analyses, we found concomitant methotrexate at a dosage >10 mg/week
Rheumatology, Department of Medicine, (odds ratio [OR] 5.17; 95% CI 1.34‐19.93; P = 0.017) predicted low disease activity/
Taipei Veterans General Hospital, Taipei
City, Taiwan. remission, and baseline DAS28 ≤6.5 (OR 4.97; 95% CI 1.22‐20.30; P = 0.026) pre‐
Emails: mhchen6@vghtpe.gov.tw; cytsai@ dicted good EULAR response at 24 months. The most common adverse events were
vghtpe.gov.tw
infusion‐related (5.7%), and there was no incidence of malignancy or mortality during
the treatment.
Conclusions: RTX was effective and safe in real‐life management of RA patients with
high disease activity. Patients taking concomitant methotrexate and with lower base‐
line DAS28‐ESR were more likely to benefit from RTX.

KEYWORDS
CD20, DAS‐28, rheumatoid arthritis, rituximab

1 |  I NTRO D U C TI O N strategy for RA, among which methotrexate (MTX) is the cornerstone
unless contraindicated or not tolerated. Short‐term glucocorticoids
Rheumatoid arthritis (RA), a systemic autoimmune disease targeting would be bridging therapy when switching csDMARDs. Biologic
synovial tissues, causes longstanding inflammation and catastrophic DMARDs (bDMARDs) and targeted synthetic DMARDs (tsDMARDs)
sequelae in multiple joints. The resultant progressive deformity and can be added to csDMARDs when remission cannot be achieved.1
physical disability impair quality of life. Conventional synthetic disease‐ A presence of rheumatoid factor (RF) and/or anti‐citrullinated
modifying antirheumatic drugs (csDMARDs) are the primary treatment protein antibody (ACPA) could be detected years before onset of

Int J Rheum Dis. 2019;1–9. wileyonlinelibrary.com/journal/apl   © 2019 Asia Pacific League of Associations for |  1
Rheumatology and John Wiley & Sons Australia, Ltd
 |
2       WANG et al.

clinical symptoms of RA, 2 and their titers have been shown to pre‐ Rheumatology Division in Taipei Veterans General Hospital, a ter‐
dict functional and radiological outcomes. In addition to producing tiary referral center in Taiwan, from October 2008 to January 2016.
these antibodies, B cells are also highly efficient in antigen presen‐ All patients fulfilled the ACR 1987 revised criteria or ACR/EULAR
tation and can activate auto‐reactive T cells. Rituximab (RTX), a bD‐ 2010 criteria for RA classification,11,12 and had previously failed
MARD of chimeric murine/human monoclonal antibody targeting standard csDMARDs. There were no disease activity or comorbid‐
CD20‐positive B cells, has thus been proven effective in randomized ity exclusion criteria. All patients received RTX as accredited by this
clinical trials (RCTs) in patients with active established RA resistant country's health administration in a 6‐month fixed repetitive proto‐
to csDMARDs and anti‐tumor necrosis factor α (anti‐TNFα) bD‐ cols (two 1000 mg intravenous doses, 14 days apart) as the 2nd line
MARDs.3-5 A single course of RTX with two 1000 mg infusions given bDMARD with the reasons for switching to RTX being recorded. This
2 weeks apart significantly ameliorates disease burden. As shown study was approved by the institutional ethics committee of Taipei
previously, at week 24, 51% of patients had an American College Veterans General Hospital (IRB‐TPEVGH No. 2017‐05‐002AC).
of Rheumatology 20% improvement (ACR20) response, 50% of pa‐
tients had a European League Against Rheumatism (EULAR) mod‐
2.2 | Clinical evaluations and laboratory tests
erate response and 15% of patients had a EULAR good response.5
However, the populations in RCTs may not accurately reflect the Clinical parameters including 28 tender joint count (TJC), 28 swol‐
characteristics of the real‐world patients as strict inclusion and ex‐ len joint count (SJC), and DAS28 using erythrocyte sedimentation
clusion criteria were applied to increase the internal validity but not rate (DAS28‐ESR) were assessed at baseline and at 6, 12, 18 and
external generalizability. A systemic review has demonstrated that 24 months after initiating RTX. The numbers and types of the previ‐
characteristics of patients from real‐world observational studies are ous anti‐TNFα agents, and numbers and dosages of the concomitant
substantially different from those enrolled in RCTs, leading to an DMARDs and glucocorticoids were also documented. Laboratory
efficacy‐effectiveness gap.6 As a result, the real‐world evidence is tests including ESR, RF and ACPA were periodically assessed. RF
vital because it is not only generated from investigations of hetero‐ was examined using nephelometry (Toshiba TBA™‐c16000 device,
geneous populations with diverse behaviors in routine practices, but Toshiba Medical Systems Corporation, Tochigi‐ken, Japan; positive if
also can validate the results in RCTs. Nevertheless, most of the RCTs >30 IU/mL) and ACPA was measured using enzyme‐linked immuno‐
and observational studies showing outcomes of RTX treatments sorbent assay (Phadia EliA™ kit, Phadia AB, Uppsala, Sweden; posi‐
have been conducted in Western countries and very few studies tive if >10 U/mL).
have examined the effectiveness and safety of RTX for RA in ethnic
Asians.7,8
2.3 | Definition of clinical response and remission
Several authors have also attempted to identify specific pre‐
dictive factors for RTX treatment response in RA patients to help The clinical remission status was defined by DAS28‐ESR ≤2.6, low
clinical rheumatologists take a more personalized and precise thera‐ disease activity by DAS28‐ESR >2.6 but ≤3.2, moderate disease
peutic approach. The presence of RF and/or ACPA, number of previ‐ activity by DAS28‐ESR >3.2 but ≤5.1 and high disease activity by
ous bDMARDs, baseline Disease Activity Score of 28 joints (DAS28), DAS28‐ESR >5.1. Responses to RTX therapy were assessed using
serum immunoglobulin level and biomarkers including type I in‐ the EULAR criteria.13 Patients were classified as responders (good
terferon and myeloid‐related protein (MRP)8/14 levels, have been or moderate) or non‐responders according to DAS28‐ESR levels and
shown to affect treatment responses in previous studies.9 On the their improvement from the baseline. The clinical remission status
other hand, a low body mass index has not yet been proven to be a and responses were evaluated at 6, 12, 18 and 24 months after the
reliable index to predict RTX effectiveness although it has been used 1st dose of RTX.
so in anti‐TNFα therapy.10
Due to lack of studies addressing the real‐life outcomes of RA
2.4 | Safety issues
patients undergoing RTX therapy in Asia, we carried out the pres‐
ent investigation to characterize the real‐world use, effectiveness, Adverse events including infusion reactions, infections, serious in‐
and safety of RTX in patients with RA. Furthermore, we identified fections necessitating admission, malignancies, and mortalities were
possible predictors for achieving low disease activity/remission recorded.
or a EULAR good response in these RTX‐treated RA patients at
24 months of treatment.
2.5 | Statistical analysis
The baseline characteristics were compared with the Chi‐squared
2 | PATI E NT S A N D M E TH O DS
test or Fisher's exact test for categorical variables and compared
with Mann‐Whitney U tests for the quantitative variables. DAS28‐
2.1 | Patients
ESR, ESR, TJC and SJC were compared with corresponding values
In the present retrospective study, we enrolled 70 adult patients with at baseline using the paired sample t test. Univariate logistic regres‐
RA who underwent RTX treatment at the Allergy, Immunology and sion was performed to identify predictive factors for the low disease
WANG et al.       3 |
TA B L E 1   Demography and characteristics of RA patients (CI). A P value <0.05 was considered significant. Data were analyzed
undergoing rituximab therapy with SPSS software, version 22 (IBM, Armonk, NY, USA).

RA patients
Characteristics (n = 70)
3 | R E S U LT S
Female gender, n (%) 67 (95.7)
Age at diagnosis, mean (SD), y 54.1 (10.6) 3.1 | Demographic and clinical characteristics
Age at RTX initiation, mean (SD), y 58.9 (10.6)
A total of 70 patients, including 67 women (95.7%), were en‐
Disease duration, mean (SD), y 4.9 (1.8)
rolled in this study. The mean age at RTX initiation was 58.9 years.
Baseline TJC, mean (SD), n 16.5 (4.8)
Demographic characteristics of patients are shown in Table 1. At
Baseline SJC, mean (SD), n 9.7 (4.8) baseline, patients had a mean disease duration of 4.9 years and a sig‐
Baseline ESR, mean (SD), mm/h 59.1 (33.6) nificant disease burden with a mean DAS28‐ESR high at 6.88 ± 0.85,
Baseline DAS28‐ESR, mean (SD) 6.88 (0.85) mean SJC of 9.7 ± 4.8, and mean TJC of 16.5 ± 4.8. Four patients
RF positive, n (%) 63 (90.0) (5.7%) overlapped with systemic lupus erythematosus, 20 patients
ACPA positive, n/total (%) 38/69 (55.1) (28.6%) had Sjögren's syndrome, and 7 patients (10%) had interstitial
Comorbidity, n (%) lung disease at baseline.
Systemic lupus erythematosus 4 (5.7) Sixty‐six patients (94.3%) had previously used at least 1 anti‐

Sjögren's syndrome 20 (28.6)

TNFα agent. The remaining 4 patients took RTX as the first‐line bD‐
MARD due to history of malignancy (3 had breast cancer and 1 had
Interstitial lung disease 7 (10.0)
schwannoma). During the 24 months of RTX treatment, 69 patients
Number(s) of previous anti‐TNFα, n (%)
(98.6%) received at least 1 concomitant csDMARDs including MTX,
None 4 (5.7)
leflunomide, sulfasalazine, hydroxychloroquine, and cyclosporine;
One 59 (84.3)
in average, 1.9 types of DMARDs were used. Forty‐nine patients
≥2 7 (10.0)
(70.0%) received MTX during the 24 months at a mean dosage of
Item(s) of previous anti‐TNFα (specified), n (%) 10.71 ± 3.85 mg/wk. There were also 26 patients (37.1%) taking
Etanercept 49 (74.2) prednisolone along with RTX at a mean dosage of 7.69 ± 2.60 mg/d.
Adalimumab 10 (15.2)
Etanercept and adalimumab 7 (10.6)
3.2 | Improvement of laboratory and clinical
Concomitant medication in 24 mo
parameters by RTX
csDMARDs, mean (SD), n 1.9 (0.6)
csDMARDs other than MTX, mean (SD), n 1.2 (0.6) After initiation of RTX, we observed a significant improvement of mean
DAS28‐ESR at 6, 12, 18, 24 months as compared with baseline levels
MTX usage, n (%) 49 (70.0)
(P < 0.001 for each comparison), and the mean decrement of DAS28‐
MTX dosage, mean (SD), mg/wk 10.71 (3.85)
ESR at 24 months was 3.41 ± 0.93 (Figure 1A). The mean ESR level at
Mean ≤7.5 mg/wk, n/total (%) 25/49 (51.0)
24 months significantly decreased when compared to that at baseline
Mean ≤10 mg/wk, n/total (%) 28/49 (57.1)
(19.7 ± 17.3 vs 59.1 ± 33.6; P < 0.001; Figure 1B). Similarly, the mean
Prednisolone usage, n (%) 26 (37.1)
SJC and TJC were also significantly lower at 24 months, compared to
Mean prednisolone dosage, mean (SD), mg/d 7.69 (2.6) those at baseline (P < 0.001, P < 0.001, respectively; Figure 1C,D).
ACPA, anti‐citrullinated protein antibody; CRP, C‐reactive protein; csD‐ All patients were in high disease activity at baseline. After
MARDs, conventional synthetic disease‐modifying anti‐rheumatic drugs; 24 months, none were in high disease activity: 12 patients (17.1%)
DAS28‐ESR, Disease Activity Score of 28 joints—erythrocyte sedimen‐
achieved remission, 17 patients (24.3%) reached low disease activity,
tation rate; ESR, erythrocyte sedimentation rate; MTX, methotrexate;
RF, rheumatoid factor; RTX, rituximab; SD, standard deviation; SJC, and 41 patients (58.5%) had moderate disease activity (Figure 2A).
swollen joint count; TJC, tender joint count; TNF, tumor necrosis factor.

3.3 | Predictors for clinical remission at 24 months

activity/remission and good EULAR responders at 24 months.
Variables with a P value ≤0.15 in the univariate analysis were taken Comparing the 41 patients (58.5%) with moderate disease ac‐
into multivariate logistic regression models using backward selec‐ tivity to the 29 patients (41.4%) with low disease activity/re‐
tions to estimate odds ratios (ORs) for clinical remission status and mission at 24 months, those who were in low disease activity/
EULAR responses. The multivariate models were also adjusted for remission at 24 months had a longer duration of disease (5.3 ± 2.1 vs
age, gender, and known predictive factors for RTX response accord‐ 4.6 ± 1.4 years; P = 0.041), and a significantly greater proportion of
ing to a literature review, which included the presence of RF and/or them took concomitant MTX at a mean dosage >10 mg/wk (59.1% vs
ACPA, baseline DAS28‐ESR levels, and numbers of previous biolog‐ 29.6%; P = 0.038; Table S1). There were no significant differences in
ics.9 Results are expressed as ORs with a 95% confidence interval age at diagnosis, age at RTX initiation, gender, TJC, SJC, DAS28‐ESR,
 |
4       WANG et al.

F I G U R E 1   Changes in clinical and laboratory parameters in 70 patients with rheumatoid arthritis treated with rituximab, at baseline, 6,
12, 18, and 24 mo. (A) The Disease Activity Score of 28 joints‐ erythrocyte sedimentation rate (DAS28‐ESR) levels; (B) ESR; (C) swollen joint
count (SJC); (D) tender joint count (TJC). All parameters were significantly lower than the corresponding baseline levels (P < 0.001)

presence of RF, presence of ACPA, autoimmune comorbidities, pro‐ (17.1%) and 58 moderate responders (82.9%).13 The good respond‐
portion of no more than 1 previous anti‐TNFα, refractoriness to anti‐ ers had a longer disease duration (5.5 ± 1.2 vs 4.8 ± 1.8 years;
TNFα, and concomitant uses of csDMARDs and prednisolone. P = 0.043), lower baseline DAS28‐ESR levels (41.58 ± 33.30 vs
We then used univariate logistic regression to find predictors for 62.72 ± 32.75; P = 0.043), and a baseline DAS28‐ESR ≤6.5 (66.7%
low disease activity/remission at 24 months of RTX therapy. It was vs 25.9% of respective cohort; P = 0.014; Table S2). No significant
identified that a concomitant MTX use (OR 1.17; 95% CI 1.002‐1.37; differences were observed between good and moderate respond‐
P = 0.047), especially at a level of more than 10 mg/week (OR 3.43; ers regarding the age at diagnosis, age at RTX initiation, gender,
95% CI 1.05‐11.22; P = 0.041) was a predictor for low disease activity/ TJC, SJC, RF/ACPA seropositivity, autoimmune comorbidities,
remission (Table 2). The multivariate logistic regression model, which concomitant use and adjustment of csDMARDs and predniso‐
was adjusted for gender, age, disease duration, DAS28‐ESR level, pres‐ lone, previous anti‐TNFα of no more than 1 and refractoriness to
ence of RF/ACPA, no more than 1 previous anti‐TNFα, and concomitant anti‐TNFα.
prednisolone usage revealed the correlation between concomitant We analyzed global DAS28‐ESR as well as categorized DAS28‐
mean MTX dosage of more than 10 mg/wk and low disease activity/re‐ ESR at 6, 6.5 and 7 using univariate logistic regression to identify
mission at 24 months (OR 5.17; 95% CI 1.34‐19.93; P = 0.017; Table 2). predictors of a good EULAR response and found that global DAS28‐
ESR (OR 0.28; 95% CI 0.11‐0.69; P = 0.005) and categorization of
DAS28‐ESR ≤6.5 (OR 5.73; 95% CI 1.51‐21.82; P = 0.001) were good
3.4 | Predictors for EULAR response at 24 months
predictors. We then analyzed these data further with a multivari‐
As shown in Figure 2B, all patients were EULAR responders at ate logistic regression model, which was adjusted by gender, age,
24 months after RTX treatment, including 12 good responders RF/ACPA seropositivity, and number of previous anti‐TNFα ≤1, and
WANG et al. |
      5

F I G U R E 2   (A) Percentages of clinical remission status according to Disease Activity Score of 28 joints‐ erythrocyte sedimentation rate
(DAS28‐ESR) levels and (B) percentages of responses according to European League Against Rheumatism criteria at months 6, 12, 18 and 24
in 70 patients with rheumatoid arthritis receiving rituximab

found that baseline DAS28‐ESR ≤6.5 remained a statistically signifi‐ IGRA received prophylaxis with isoniazid prior to the administration
cant predictor (OR 4.97; 95% CI 1.22‐20.30; P = 0.026; Table 3). of RTX. No reactivation of tuberculosis occurred. At baseline, nine
patients (12.9%) had chronic hepatitis B as defined by the presence
of hepatitis B surface antigen (HBsAg); no reactivation of hepatitis
3.5 | Safety outcomes
B was noted within 24 months. In addition, no patient in our cohort
None of the enrolled patients discontinued RTX therapy during the developed malignancy or died in the 24‐month follow‐up.
follow‐up. There were 4 events of infusion‐related reactions. One
serious infection event occurred, which was cellulitis complicated
with osteomyelitis caused by methicillin‐sensitive Staphylococcus 4 | D I S CU S S I O N
aureus and coagulase‐negative Staphylococcus. Thirteen patients
underwent interferon‐γ release assay (IGRA) at baseline; among In the present observational investigation from a tertiary refer‐
them, six patients had a positive result and two patients had an ral center in Taiwan, RTX was effective in the treatment of RA pa‐
indeterminate finding. Three out of the six patients with positive tients with high clinical disease activities regardless of response to
 |
6       WANG et al.

TA B L E 2   Predictors of low disease

Univariate OR Multivariate ORb
activity/remissiona at 24 mo
Characteristics (95% CI) P value (95% CI) P value

Female gender 1.44 (0.12‐16.62) 0.772 0.97 (0.04‐26.85) 0.973

Age at RTX initiation, y 0.98 (0.94‐1.03) 0.407 0.99 (0.91‐1.07) 0.752
Disease duration, y 1.25 (0.93‐1.68) 0.145 1.40 (0.97‐2.02) 0.076
DAS28‐ESR 0.58 (0.32‐1.06) 0.076 0.54 (0.19‐1.59) 0.265
RF positive 0.49 (0.10‐2.40) 0.381 0.12 (0.01‐1.28) 0.078
ACPA positive 1.47 (0.56‐3.90) 0.437 1.09 (0.26‐4.57) 0.904
Previous anti‐TNFα ≤1 0.94 (0.19‐4.54) 0.936 2.90 (0.20‐41.48) 0.432
Refractory to anti‐TNF‐α 0.57 (0.13‐2.51) 0.455
Concomitant medication in 24 mo
csDMARDs, n 1.25 (0.58‐2.70) 0.566
csDMARDs excluding 0.96 (0.43‐2.18) 0.929
MTX, n
MTX usage 1.63 (0.56‐4.74) 0.37
Mean dosage of MTX, mg/wk 1.17 (1.002‐1.37) 0.047*
>10 mg/wk, n/total 3.43 (1.05‐11.22) 0.041* 5.17 (1.34‐19.93) 0.017*
Prednisolone usage 0.39 (0.14‐1.05) 0.063 0.59 (0.13‐2.61) 0.485
Mean dosage of predniso‐ 0.84 (0.62‐1.15) 0.286
lone, mg/d

ACPA, anti‐citrullinated protein antibody; CRP, C‐reactive protein; csDMARDs, conventional syn‐
thetic disease‐modifying anti‐rheumatic drugs; DAS28‐ESR, Disease Activity Score of 28 joints—
erythrocyte sedimentation rate; ESR, erythrocyte sedimentation rate; MTX, methotrexate; RF,
rheumatoid factor; RTX, rituximab; TNF, tumor necrosis factor.
a
There are no patients with high disease activities at 24 mo.
b
Factors including age, gender, RF, ACPA, previous number of anti‐TNFα ≤1 and variables with
P < 0.15 in univariate logistic regression were analyzed by multivariate logistic regression with back‐
ward selection. ESR was excluded as it is a component of DAS28‐ESR.
*
P < 0.05

anti‐TNFα. Patients concomitantly taking MTX at a mean dosage of a good EULAR response, in contrast to the 41.1% who reached low
more than 10 mg/wk were more likely to be in low disease activity or disease activity/remission, which may also cause this inconsistency.
remission at 24 months. For EULAR response at 24 months, patients Our mean dosage of MTX has been lower than those prescribed in
with a baseline DAS28‐ESR ≤6.5 were more likely to be good respond‐ Western countries, partly in that Orientals may have a poorer tol‐
ers than to be moderate responders. These outcomes were robust erance to MTX than Caucasians. As higher dosages are commonly
even after adjustment in the multivariate logistic regression models. not tolerated due to adverse events, the maximum dosage of MTX is
The present investigation demonstrated that MTX of more than 16 mg/wk in Japan and 20 mg/wk in China.16-18 Thus, these findings,
10 mg/wk predicted lower disease activity at 24 months. These suggesting maintaining MTX dosage above 10 mg/wk in combina‐
results are important as no clinical trials have demonstrated thera‐ tion with RTX can decrease disease activity more effectively, are
peutic superiority of RTX monotherapy to combination of RTX with important for Orientals receiving RTX therapy.
csDMARDs up to now, and various clinical guidelines have recog‐ We have also shown that patients with lower baseline DAS28‐
nized MTX as a preferred csDMARD, either in itself or in combina‐ ESR were more likely to be good responders at 24 months. These re‐
tion with other csDMARDs, auxiliary to bDMARDs or tsDMARDs.1,14 sults were compatible with those reported previously which showed
The chimerically designed RTX has been reported to induce anti‐ 2019 RA patients from the CERERRA initiative, exhibited a lower
drug antibodies in 3%‐4% of RA patients,4 which can adversely af‐ baseline DAS28‐ESR level as a predictor of a good EULAR response
fect its efficacy, pharmacokinetics and safety profiles. Combination at 6 months.19 We further stratified our patients with different base‐
with MTX may potentially decrease this immunogenicity in a pattern line DAS28‐ESR levels (at 6, 6.5, and 7) and confirmed with multivar‐
similar to those seen in studies with anti‐TNFα.15 Although this com‐ iate analysis that those with baseline DAS28‐ESR ≤6.5 were more
bination only predicts low disease activity/remission but not a good likely to be good responders.
EULAR response in our study, we believe this divergence reflects The presence of RF and ACPA in serum has been reported to pre‐
the inherent differences of these endpoints, which are composed dict better response to RTX in previous studies. 20-22 However, the
of different factors. In addition, only 17.1% of our patients reached present study failed to show such correlation. The reason for this is
WANG et al. |
      7

TA B L E 3   Predictors of good EULAR

Univariate OR Multivariate ORa
response at 24 mo
Characteristics (95% CI) P value (95% CI) P value

Female gender 0.39 (0.03‐4.72) 0.461 0.25 (0.02‐4.11) 0.333

Age at RTX initiation, y 1.05 (0.98‐1.12) 0.176 1.04 (0.96‐1.12) 0.329
Disease duration, y 1.24 (0.90‐1.72) 0.192
Baseline DAS28‐ESR 0.28 (0.11‐0.69) 0.005*
DAS28‐ESR ≤6.5 5.73 0.010* 4.97 (1.22‐20.30) 0.026*
(1.51‐21.82)
RF positive 0.22 (0.04‐1.16) 0.075 0.20 (0.03‐1.21) 0.080
ACPA positive 1.52 (0.40‐5.78) 0.535 1.72 (0.36‐8.17) 0.494
Previous n of anti‐TNFα ≤1 0.47 (0.08‐2.78) 0.406 0.91 (0.07‐12.41) 0.941
Refractory to anti‐TNFα 0.48 (0.08‐2.81) 0.415
Concomitant medication in 24 mo
csDMARDs, n 1.01 (0.37‐2.72) 0.988
csDMARDs excluding 0.44 (0.13‐1.42) 0.168
MTX, n
MTX usage 5.79 0.104 3.72 (0.40‐34.14) 0.246
(0.70‐48.11)
MTX mean dosage, mg/wk 1.08 (0.91‐1.29) 0.390
Prednisolone usage 0.36 (0.09‐1.46) 0.151
Prednisolone mean 0.84 (0.53‐1.35) 0.475
dosage, mg/d

ACPA, anti‐citrullinated protein antibodies; csDMARDs, conventional synthetic disease‐modifying

anti‐rheumatic drugs; DAS28‐ESR, Disease Activity Score of 28 joints—erythrocyte sedimentation
rate; ESR, erythrocyte sedimentation rate; EULAR, European League Against Rheumatism; MTX,
methotrexate; RF, rheumatoid factor; RTX, rituximab; TNF, tumor necrosis factor.
a
Factors including age, gender, RF, ACPA, Number of previous anti‐TNFα ≤1 and variables with
P < 0.15 in univariate logistic regression are analyzed by multivariate logistic regression with back‐
ward selection. Swollen joint count (SJC) and tender joint count (TJC) were excluded as they are
components of DAS28‐ESR.
*
P < 0.05.

unclear but might be related to a relatively small sample size, a differ‐ immunity resulting from long‐term administration of RTX is un‐
ent ethnicity and the baseline comorbidities. Also, the average age of known and needs further studies. In the present investigation, we
our patients (58.9 ± 10.6 years) was higher than those enrolled in the have not found reactivation of latent tuberculosis infection (LTBI)
clinical trials. 20 Detection of RF increases as patients become older although some of the patients did have positive IGRA and under‐
and are exposed to accumulating kinds of microbes or environmental went prophylactic therapy with isoniazid. Although RA patients
factors, decreasing its accountability for RA. Further studies with have a 2.7‐fold increase in the risk of developing TB, 24 our findings
longer observation times, larger patient numbers, as well as prospec‐ are in agreement with those reported previously in that there were
tive data collection in a real‐world situation would be necessary to no increased risks of TB infection after RTX therapy. 25 Since RTX
demonstrate their predictive usefulness. Similarly, several studies has no significant effect on release of interferon‐γ in RA patients
have suggested that patients previously treated with no more than 1 with LTBI, it is conceivable that RTX exerts it effect through a signal
anti‐TNFα have a better response and are more likely to achieve low transduction pathway other than Th1 inhibition or innate immune
disease activity or remission.19,23 In contrast, the present study did suppression.8
not show similar results. Nevertheless, we adjusted this factor with On the other hand, although our previous study has shown the
a presence of RF and ACPA in the multivariate logistic regression as risk of reactivation of hepatitis B is significantly increased among
they were recognized predictors. RTX‐treated RA patients (adjusted hazards ratio, 16.51; 95% CI
RTX treatment appeared to be generally safe in our study. The 1.82‐149.67; P = 0.01)26 no reactivation of hepatitis B was docu‐
most frequent adverse events were infusion‐related, which were mented in the present investigation. This was largely because most
relieved by lowering infusion rates and the administration of glu‐ of our patients (43, 61.4%) had anti‐hepatitis B surface antibod‐
cocorticoids and antihistamines. Only one serious infection event ies (anti‐HBs) and a negligible HBV viral load at baseline and pos‐
happened by the end of the observation and there was no fatality. sibly also because the follow‐up period was not sufficiently long
Whether the infection was predisposed by interference of acquired (24 months only).
 |
8       WANG et al.

Although our present results are encouraging there are still 3. Edwards JC, Szczepański L, Szechiński J, et al. Efficacy of B‐cell‐tar‐
limitations in the investigation. These include its retrospective and geted therapy with rituximab in patients with rheumatoid arthritis.
N Engl J Med. 2004;350:2572‐2581.
observational design, a relatively small sample size, and a possible
4. Emery P, Fleischmann R, Filipowicz‐Sosnowska A, et al. The effi‐
selection bias with a strong female preponderance in the RA cohort cacy and safety of rituximab in patients with active rheumatoid
from only a single referral center. Nevertheless, to our knowledge, arthritis despite methotrexate treatment: results of a phase IIB ran‐
the present investigation is the 1st study in ethnic Asians, trying to domized, double‐blind, placebo‐controlled, dosage‐ranging trial.
Arthritis Rheum. 2006;54:1390‐1400.
explore the real‐world data regarding the effectiveness and safety
5. Cohen SB, Emery P, Greenwald MW, et al. Rituximab for rheumatoid
of RTX in active RA patients who failed anti‐TNFα treatment. In ad‐ arthritis refractory to anti‐tumor necrosis factor therapy: results
dition, it also provided detailed clinical information in a substantially of a multicenter, randomized, double‐blind, placebo‐controlled,
long time of observation, as well as a clear and segregated method‐ phase III trial evaluating primary efficacy and safety at twenty‐four
weeks. Arthritis Rheum. 2006;54:2793‐2806.
ology of evaluation on the effect of RTX in a complicated scenario of
6. Kilcher G, Hummel N, Didden EM, et al. GetReal Work Package 4.
tedious and variable combinations of csDMARDs. Rheumatoid arthritis patients treated in trial and real world set‐
tings: comparison of randomized trials with registries. Rheumatology
(Oxford). 2018;57:354‐369.
5 |  CO N C LU S I O N 7. Tan BE, Lim AL, Kan SL, et al. Real‐world clinical experience of bio‐
logical disease modifying anti‐rheumatic drugs in Malaysia rheuma‐
toid arthritis patients. Rheumatol Int. 2017;37:1719‐1725.
This real‐world study confirms the effectiveness and safety of RTX 8. Chen YM, Chen HH, Lai KL, et al. The effects of rituximab therapy
in well‐documented active RA patients regardless of their prior anti‐ on released interferon‐γ levels in the QuantiFERON assay among
TNFα refractoriness. Our results also suggest patients may benefit RA patients with different status of Mycobacterium tuberculosis in‐
fection. Rheumatology (Oxford). 2013;52:697‐704.
more from RTX if maintaining the concomitant MTX at a dosage of
9. Wijbrandts CA, Tak PP. Prediction of response to targeted treat‐
more than 10 mg/wk. ment in rheumatoid arthritis. Mayo Clin Proc. 2017;92:1129‐1143.
10. Ottaviani S, Gardette A, Roy C, et al. Body mass index and re‐
sponse to rituximab in rheumatoid arthritis. Joint Bone Spine.
AC K N OW L E D G E M E N T S 2015;82:432‐436.
11. Arnett FC, Edworthy SM, Bloch DA, et al. The American Rheumatism
The authors would like to thank Dr. Yi‐Syuan Sun for his valuable Association 1987 revised criteria for the classification of rheuma‐
advice and assistance with the statistical analysis of this study. This toid arthritis. Arthritis Rheum. 1988;31:315‐324.
study was supported by the grants from the Ministry of Science & 12. Aletaha D, Neogi T, Silman AJ, et al. 2010 Rheumatoid arthritis clas‐
sification criteria: an American College of Rheumatology/European
Technology (MOST‐107‐2314‐B‐075 ‐051 ‐MY3) and Taipei Veterans
League Against Rheumatism collaborative initiative. Arthritis
General Hospital (V107C‐084), Taiwan. Rheum. 2010;62:2569‐2581.
13. van Gestel AM, Prevoo ML, van 't Hof MA, , et al. Development
and validation of the European League Against Rheumatism re‐
C O N FL I C T S O F I N T E R E S T sponse criteria for rheumatoid arthritis. Comparison with the pre‐
liminary American College of Rheumatology and the World Health
The authors declare no conflicts of interest. Organization/International League Against Rheumatism Criteria.
Arthritis Rheum. 1996;39:34‐40.
14. Singh JA, Saag KG, Bridges SL Jr, et al. 2015 American College of
ORCID Rheumatology Guideline for the treatment of rheumatoid arthritis.
Arthritis Rheumatol. 2016;68:1‐26.
Kai‐Chun Wang  https://orcid.org/0000-0003-2037-6681
15. Krieckaert CL, Nurmohamed MT, Wolbink GJ. Methotrexate re‐
Hsien‐Tzung Liao  https://orcid.org/0000-0001-5930-4840 duces immunogenicity in adalimumab treated rheumatoid ar‐
thritis patients in a dose dependent manner. Ann Rheum Dis.
Chien‐Chih Lai  https://orcid.org/0000-0002-0546-8695 2012;71:1914‐1915.
Ming‐Han Chen  https://orcid.org/0000-0001-7822-2613 16. Atsumi T, Yamamoto K, Takeuchi T, et al. The first double‐blind,
randomised, parallel‐group certolizumab pegol study in methotrex‐
Chang‐Youh Tsai  https://orcid.org/0000-0002-4154-4018 ate‐naïve early rheumatoid arthritis patients with poor prognostic
factors, C‐OPERA, shows inhibition of radiographic progression.
Ann Rheum Dis. 2016;75:75‐83.
REFERENCES 17. Li R, Zhao JX, Su Y, et al. High remission and low relapse with pro‐
longed intensive DMARD therapy in rheumatoid arthritis (PRINT):
1. Smolen JS, Landewé R, Bijlsma J, et al. EULAR recommendations a multicenter randomized clinical trial. Medicine (Baltimore).
for the management of rheumatoid arthritis with synthetic and bi‐ 2016;95:e3968.
ological disease‐modifying antirheumatic drugs: 2016 update. Ann 18. Kameda H, Fujii T, Nakajima A, et al. Japan College of Rheumatology
Rheum Dis. 2017;76:960‐977. guideline for the use of methotrexate in patients with rheumatoid
2. Nielen MM, van Schaardenburg D, Reesink HW, et al. Specifc arthritis. Mod Rheumatol. 2019;29:31‐40.
autoantibodies precede the symptoms of rheumatoid arthritis: 19. Chatzidionysiou K, Lie E, Nasonov E, et al. Highest clinical effec‐
a study of serial measurements in blood donors. Arthritis Rheum. tiveness of rituximab in autoantibody‐positive patients with rheu‐
2004;50:380‐386. matoid arthritis and in those for whom no more than one previous
WANG et al. |
      9

TNF antagonist has failed: pooled data from 10 European registries. 26. Chen MH, Chen MH, Liu CY, et al. Hepatitis B virus reactivation
Ann Rheum Dis. 2011;70:1575‐1580. in rheumatoid arthritis patients undergoing biologics treatment. J
20. Sellam J, Hendel‐Chavez H, Rouanet S, et al. B cell activation Infect Dis. 2017;215:566‐573.
biomarkers as predictive factors for the response to rituximab in
rheumatoid arthritis: a six‐month, national, multicenter, open‐label
study. Arthritis Rheum. 2011;63:933‐938.
21. Tak PP, Rigby WF, Rubbert‐Roth A, et al. Inhibition of joint damage S U P P O R T I N G I N FO R M AT I O N
and improved clinical outcomes with rituximab plus methotrexate
in early active rheumatoid arthritis: the IMAGE trial. Ann Rheum Dis. Additional supporting information may be found online in the
2011;70:39‐46. Supporting Information section at the end of the article. 
22. Isaacs JD, Cohen SB, Emery P, et al. Effect of baseline rheumatoid
factor and anticitrullinated peptide antibody serotype on rituximab
clinical response: a meta‐analysis. Ann Rheum Dis. 2013;72:329‐336.
How to cite this article: Wang K‐C, Liao H‐T, Chen W‐S, et al.
23. Harrold LR, Reed GW, Shewade A, et al. Effectiveness of ritux‐
imab for the treatment of rheumatoid arthritis in patients with Real‐world effectiveness and safety of rituximab in the
prior exposure to anti‐TNF: results from the CORRONA registry. J treatment of rheumatoid arthritis: A single‐center experience
Rheumatol. 2015;42:1090‐1098. in Taiwan. Int J Rheum Dis. 2019;00:1–9. https://doi.
24. Liao TL, Lin CH, Chen YM, et al. Different risk of tuberculosis and
org/10.1111/1756-185X.13511
efficacy of isoniazid prophylaxis in rheumatoid arthritis patients
with biologic therapy: a nationwide retrospective cohort study in
Taiwan. PLoS ONE. 2016;11:e0153217.
25. Arkema EV, Jonsson J, Baecklund E, et al. Are patients with rheuma‐
toid arthritis still at an increased risk of tuberculosis and what is the
role of biological treatments? Ann Rheum Dis. 2015;74:1212‐1217.