Predictive value of serum amyloid a levels for requirement of concomitant methotrexate in tocilizumab initiation: A post hoc analysis of the SURPRISE study

Abstract

Objectives: To identify predictive factors for remission by tocilizumab monotherapy in rheumatoid arthritis (RA) patients. Methods: This is a post hoc analysis of the SURPRISE study, a 2-year randomized, controlled study comparing the efficacy of tocilizumab with (ADD-ON) and without methotrexate (SWITCH). The primary endpoint was DAS28-ESR remission (<2.6) at week 24. The change in modified total Sharp score from baseline to week 52 (DmTSS/year) was also assessed as an endpoint. The effect of clinical parameters at baseline on remission was estimated by logistic regression analysis. Results: In SWITCH (n ¼ 96), CRP, SAA, RF, and DAS28 at baseline showed predictive value for DAS28 remission in unadjusted analysis. Adjusted analysis confirmed SAA and DAS28 as predictive factors, with SAA having the highest value (ROC-AUC ¼ 0.731). Furthermore, structural remission (DmTSS/year 0.5) rate was significantly higher in patients with SAA of < 50.0 lg/mL than other patients. In contrast, in ADD-ON (n ¼ 98), only DAS28 showed predictive value for DAS28 remission. In patients with SAA < 50.0 lg/mL, both DAS28 remission and structural remission rate were comparable between SWITCH and ADD-ON. Conclusion: RA patients with low SAA levels at baseline may benefit similarly from tocilizumab with and without methotrexate.

The growing role of precision medicine for the treatment of autoimmune diseases; results of a systematic review of literature and Experts’ Consensus

“…We identified 8 research articles that specifically evaluated if inflammatory markers could predict a response to tocilizumab (TCZ) [13,[15][16][17][18][19][20][21] and none for other IL-6 inhibitors. Five were observational studies [13,15,[19][20][21] and 3 post-hoc analysis of randomised controlled trials (RCTs) [16][17][18]. None of the post-hoc RCTs analyses found an association between baseline levels of inflammatory markers and response.…”

“…However, most of the studies had a small sample size and were of low quality. All of them except one [18] used univariate selection of covariates [13,15,[19][20][21] and/or stepwise selection [13,15,[19][20] for their regression models, which should be avoided to select covariates in prediction modelling [22][23][24]. Some did not present the number of outcome events [18,21] or did not control for other covariates [16][17].…”

“…All of them except one [18] used univariate selection of covariates [13,15,[19][20][21] and/or stepwise selection [13,15,[19][20] for their regression models, which should be avoided to select covariates in prediction modelling [22][23][24]. Some did not present the number of outcome events [18,21] or did not control for other covariates [16][17]. One study did not describe the baseline population [16] and most did not explain how they handled missing data [13,[15][16]21].…”

Title Current Status of the Search for Biomarkers for Optimal Therapeutic Drug Selection for Patients with Rheumatoid Arthritis

“…In contrast, there was no significant association between inflammatory markers at the baseline (CRP and ESR) and treatment response to IL-6 receptor inhibitors. In post hoc analyses of three RCTs [5][6][7] and a pooled analysis of five RCTs including 4186 RA patients, no association was found between levels of inflammatory markers before the administration of a therapeutic drug and response to IL-6 receptor inhibitors. One prospective, observational study focused on the association between levels of inflammatory markers and IL-6 receptor inhibitor retention, reporting that patients with higher levels of CRP at baseline were linked to lower discontinuation rates [8].…”