Background/Purpose

The PREDICT trial (NCT01255761) examined predictability of certolizumab pegol (CZP) treatment success at Week (Wk) 52 based on response at Wk12 assessed by RAPID3 and CDAI in RA patients (pts).¹ The objective of this post-hoc analysis was to evaluate whether a defined somatic comorbidity phenotype (SCP) influenced treatment response.

Methods

Pts were randomized to RAPID3 or CDAI for treatment response assessment and received CZP standard dosing regimen for 52 wks. Response at Wk12 was assessed (RAPID3 Responder, ≤6 or 20% improvement from Baseline [BL]; CDAI Responder, ≤10 or 20% improvement from BL), and pts with no improvement (<1 point CDAI improvement/no RAPID3 improvement) were to be withdrawn. The SCP, hypothesized to have suboptimal treatment response, was defined a priori as a diagnosis of depression, chronic pain, fibromyalgia or myalgias, and use of medications indicated for the treatment of depression, anxiety or neuropathic pain; insomnia diagnosis and narcotics were not included. The full analysis set (FAS; N=733), of which 313 pts had SCP at BL and ongoing during trial, is presented.

Results

At BL, 43% (313/733) of pts met the SCP classification: 23% due to medical diagnoses only (predominantly depression), 29% due to concomitant medications only (predominantly SSRIs, analgesics/antipyretics and other centrally acting agents), and the remaining 48% were due to both (predominantly depression and SSRI use). The proportion of pts with SCP was similar in the RAPID3 and CDAI arms; a similar proportion, with and without SCP, were withdrawn due to lack of efficacy by Wk12.

Among all pts randomized to CDAI (n=365), 79% without vs 73% with SCP were classified as Responders at Wk12 (Table); Wk12 Responders without SCP were approximately twice as likely to achieve LDA at Wk52 compared to those with the phenotype (41% vs 21%, respectively; Table). In addition, overall, pts without SCP were twice as likely to achieve LDA at Wk52 compared to those with the phenotype (32% vs 16%, respectively; Table).

In contrast, this phenotype was less differentiating among pts randomized to RAPID3 in the 3 outcomes examined (Table). Comparing CDAI to RAPID3, the likelihood of being classified as a Wk12 Responder was incrementally greater for those without (79% vs 63%) than those with the SCP (73% vs 68%) (Table).

Conclusion

In this large RA clinical trial population we have identified a potentially important phenotype that includes depression and chronic pain syndromes. Pts with this phenotype appear less likely to achieve LDA or be classified as a Responder when using CDAI as the outcome measure; however response rates were similar regardless of SCP when RAPID3 was used. Depending on the outcome measure used, enrolling large numbers of pts with this phenotype into an RA trial may affect the proportion able to achieve LDA/remission making it advisable to consider identifying such pts at screening.

References

1. Curtis J. Ann Rheum Dis 2014;73(S2):382