Background/Purpose: B cell depletion with rituximab
(RTX) is an established treatment for RA. It was first introduced at UCL in
1998 and at this centre patients are followed up in a
dedicated weekly clinic. Initial treatment protocols included combination with
cyclophosphamide, but since 2001 patients are treated with cycles of 2x1g RTX.
Initial retreatment strategies based on retreatment at flare evolved in the
last 6 years patients to retreatment to avoid flare, based on the duration of
response to initial cycles. We aim to characterize RTX drug survival in our
cohort and determine 5 years drug retention predictors.

Methods: We conducted a retrospective cohort study of
RA patients treated with RTX at UCL between 1998 and 2015. Medical records of 248
RA patients were reviewed; demographic and clinical data was collected. All
patients fulfilled 1987 ACR classification criteria. Drug survival data was analysed using Kaplan-Meier estimates. Predictors
for 5-years drug retention were determined by a multivariate logistic
regression model.

Results: Of the 248 patients included, 81% were
female, 80.9% Caucasians. Mean age (SD) was 60.75±15.4yrs and mean age at
diagnosis was 39.25±16.2yrs. RF and ACPA were positive in 89% and 84.6% of
patients, respectively. History of smoking was present in 59.43%.  The most frequent comorbidities were hypertension
(30.8%), osteoporosis (20.1%) and thyroid disease (15.43%). Mean follow
up duration after RTX initiation was 1737±1206 days (maximum 6055 days). At 2
and 5 years, 78% and 61% patients remained on rituximab (figure 1).  Baseline DAS28 was on average 5.93±1.25,
CRP 2.2±2.9mg/L. Majority of patients received concomitant therapy with other
DMARDs/steroids, MTX 46%, SSZ 21.4%, leflunomide 6.85%,
hydroxichloroquine 14.1% and low-dose prednisolone
28.6%. Main reasons for RTX discontinuation were: inefficacy 49 patients (44
primary failures, 6 secondary failures), adverse events 18 patients, pregnancy
1 patient and death 5 patients.

In univariate analysis,
current age, disease duration, RF seropositivity, CRP
at baseline, number of previous biologics and concomitant use of prednisolone
were associated with long-term RTX retention (>5 years). In multivariate
analysis, only CRP at baseline kept statistical significance (p=0.014).

Conclusion: Overall, high drug retention rates were
observed, independent of concomitant use of conventional DMARDs. In
multivariate analysis, higher baseline CRP was associated with long-term drug
survival. We found that secondary failures were rare (8% of drop outs) with the
current retreatment strategy based on the prevention of flare.