Background/Purpose:

Anti-TNF drugs are an effective treatment option for rheumatoid
arthritis (RA) patients (pts) but have been associated with an increased
incidence of serious infectious events (SIEs). High disease activity has been suggested
to be a predictor of SIEs in RA. Here we examine whether improvement in disease
activity over time is associated with a reduced risk of SIEs in certolizumab pegol
(CZP)-treated RA patients.

Methods:

CZP safety data from bio-naïve
patients were pooled from the RAPID1 and RAPID2 randomized clinical trials
(RCTs; NCT00152386/NCT00160602) and their open-label extensions (OLEs; NCT00175877/NCT00160641).
Post-hoc analyses included all SIEs with onset after the first CZP dose and up
to 84 days after the last CZP dose or withdrawal. A multivariate Cox
proportional hazards model (selection criteria: entry p≤0.2, stay
p≤0.25) was used to estimate the relative risk (hazard ratio [HR]) of baseline
(BL) and time-dependent pt covariates to SIEs. BL covariates selected by the
final Cox model were age (<65; ≥65 yrs) and medically-treated
comorbidity (diabetes, COPD, cardiac disorder hypertension,
hyperlipidemia, transient ischemic attack or cerebrovascular disorder). Time-dependent
covariates were measured at each pt visit (≤12 week intervals) and included
DAS28(CRP), HAQ-DI and flare (increase >1.2 in DAS28[CRP] between visits). Pts
were categorized by decrease in DAS28(CRP) between specific time points and the
assessment performed before the first CZP dose (absolute decrease: <0.6;
0.6–1.2; 1.2–2.6; ≥2.6); pts with >1 DAS28(CRP) assessment in a time
interval were categorized using the largest decrease in that interval. Incidence
rates (IRs) of SIEs were calculated per 100 patient-years (PY), with 95%
confidence intervals (CIs).

Results:

1506 bio-naïve CZP pts
were included, with a total exposure of 5778.6 PY (max pt exposure: 6.40 PY;
median exposure per pt: 4.79 PY). In total, 201 pts reported ≥1 SIE (IR
3.66/100 PY [3.17–4.21]) over ~6 yrs of CZP treatment. According to the Cox
model, pts experiencing a flare had more than double the risk of SIEs (HR 2.21
[1.41–3.47]); also, an increment of 0.125 in HAQ-DI was associated with a 27%
increase in the risk of SIEs (HR 1.27 [1.03–1.56]). By contrast, a one unit
decrease in DAS28(CRP) was associated with a 17% reduction in the risk of SIEs (HR
0.83 [0.75–0.92]). Furthermore, the observed incidence of SIEs over time was generally
lower among pts with an absolute decrease ≥2.6 in DAS28(CRP) (Table).

Conclusion:

Over time, lower disease
activity was associated with a reduced risk of SIEs in CZP-treated RA pts and,
conversely, flare was associated with increased risk. This study suggests that,
in line with current treat-to-target guidelines, reducing disease activity may help
to decrease the long-term risk of SIEs associated with anti-TNF drugs.