Background/Purpose:

There are limited observational data to guide physician decision-making when choosing to begin a patient on a new biologic treatment. Understanding real-world trends in persistency on therapy and factors related to persistency would aid in this process. We sought to identify predictors of persistence on adalimumab (time to discontinuation and change in treatment) in a large US registry of RA patients.

Methods:

The CORRONA registry is a large, multicenter, longitudinal database of RA patients enrolled from > 90 academic and private practices across the USA. RA patients enrolled in CORRONA between March 2002 and September 2011 who initiated adalimumab and had at least one follow up visit post-initiation were included in this analysis. Discontinuation of adalimumab and treatment change were modeled separately. Discontinuation was defined as stoppage of adalimumab for any duration, while treatment change included discontinuation of adalimumab, change in adalimumab dose/frequency, or adding /discontinuing a DMARD. Kaplan-Meier curves were used to estimate persistence quartiles. Associations between patient characteristics and persistence were estimated using univariate and multivariate proportional hazards (PH) regression models.

Results:

1639 patients initiated adalimumab and had at least one follow-up visit, 1003/1639 patients discontinued adalimumab, with a Kaplan-Meier estimate of median days to discontinuation: 632 (95% CI: 564.5 – 717). A multivariable model showed risk of discontinuation was associated with prior use of another biologic, number of swollen joints, duration of RA and prednisone use at initiation (Table 1). 1358/1639 patients experienced some form of treatment change. Median days to treatment change: 276 (95% CI: 258 – 298). Likelihood of treatment change increased with prior use of biologics, number of swollen joints and prednisone use at initiation; it decreased with concomitant use of MTX at initiation (Table 1). Hazard Ratios represent the baseline risk association.

Conclusion:

In the CORRONA registry, persistence on adalimumab was inversely correlated with factors suggestive of high disease severity. Patients using concurrent MTX with adalimumab at initiation were less likely to experience a change in their treatment compared to non-MTX users.

Table 1.

SJC, swollen joint count

*time varying association. HR estimates initial risk association; association decreases over time

+time varying association. HR estimates initial risk association; association increases over time