SAT0047 Predicting maintenance of response based on disease characteristics and early clinical response in rheumatoid arthritis patients upon withdrawal of adalimumab. (15th June 2017)

• Record Type: Journal Article
• Authors: Smolen, JS; Emery, P; Zhang, H; Wang, X; Suboticki, J; Sainsbury, I; Kavanaugh, A
• Is Part Of: Annals of the rheumatic diseases. Volume 76(2017)Supplement 2
• Page Start: 785
• Page End: 785
• Publication Date: 2017-06-15
• Subjects: Rheumatism -- Periodicals; 616.723005
• DOI: 10.1136/annrheumdis-2017-eular.2503
• Languages: English
• ISSNs: 0003-4967

Abstract

Abstract : Background: Some patients (pts) with rheumatoid arthritis (RA) achieve low disease activity (LDA) after treatment with adalimumab (ADA) plus methotrexate (MTX) and can maintain LDA after ADA withdrawal 1 . However, others experience a flare in disease activity. The factors associated with loss or maintenance of response are not understood. Objectives: To identify pt disease characteristics and early clinical responses, which predict maintenance of LDA upon ADA withdrawal in individual RA pts. Methods: Data from the OPTIMA trial were used in this post hoc analysis. In period 1 (P1), pts were treated for 26 weeks (wks) with ADA+MTX or placebo (PBO) +MTX. Pts on ADA+MTX who achieved DAS28-CRP <3.2 at wks 22 and 26 (responders) were randomized to ADA withdrawal, or ADA+MTX continuation up to Wk 78. Responders to PBO+MTX in P1 continued on PBO+MTX up to Wk 78 (PBO continuation). Data from the ADA withdrawal arm were used to predict LDA at Wk 78 by DAS28-CRP (≤3.2) or SDAI (≤11). Potential factors including baseline (BL) disease characteristics and Wk 26 responses, including DAS28-CRP, SDAI, ACR score components, modified total sharp score (mTSS) and joint space narrowing score (JSN), were screened by the LASSO method 2, which performs variable selection by penalizing unduly complicated models, with/without incorporating the speed of DAS28-CRP or SDAI response as an individual predictor. Logistic regression on the LASSO-selected factors yielded coefficients used to derive individual scoring equations and prediction scores for Wk 78 outcomes (fig footnote). Prediction score cutoffs were established by the regression tree method 3 . The results were validated in data from the PBO continuation arm. Results: For the prediction of DAS28-CRP LDA at Wk 78, BL physician global assessment (PhGA) and health-assessment questionnaire-disability index (HAQ-DI), and Wk 26 DAS28-CRP, HAQ-DI, JSN and CRP were selected by LASSO, and used to calculate the prediction score. Including speed of response did not affect the predictors chosen. Out of 9 pts predicted not to have DAS28-CRP LDA at Wk 78, 0 had LDA (NPV=100%) (fig 1 ). Out of 66 pts predicted to have DAS28-CRP LDA at Wk 78, 63 predictions were correct (PPV=96.5%). Results were comparable for most cutoff categories in the validation arm (PPV=82%); however, no pts were predicted to have a non-response at Wk 78. For the prediction of SDAI LDA at Wk 78, the NPV was 86% (1/7 predictions incorrect); PPV was 95% (39/41 predictions correct); in the validation arm, the PPV was 82%. Conclusions: DAS28-CRP LDA at 78 wks could be individually predicted for up to 63% pts in OPTIMA who withdrew ADA/continued PBO+MTX with 96.5% accuracy, based on demographics and clinical outcomes at 6 months. This instrument could help identify pts who may be able to maintain LDA upon ADA withdrawal. References: Smolen et al, 2013. Lancet;383. Tibshirani, R, 1996. J Royal and Stat Society; 58. Breiman L., et al, 1984. Classification and Regression Trees. Wadsworth. Acknowledgements: AbbVie: study sponsor, contributed to design, data collection, analysis, interpretation; and writing, reviewing, approval of final version. Medical writing: Naina Barretto of AbbVie. Disclosure of Interest: J. Smolen Grant/research support from: AbbVie, Consultant for: AbbVie, P. Emery Grant/research support from: from Pfizer, MSD, AbbVie Inc., BMS, UCB, Roche, Novartis, Samsung, Sandoz and Lilly., Consultant for: from Pfizer, MSD, AbbVie Inc., BMS, UCB, Roche, Novartis, Samsung, Sandoz and Lilly., H. Zhang Employee of: AbbVie, X. Wang Employee of: AbbVie, J. Suboticki Employee of: AbbVie, I. Sainsbury Employee of: AbbVie, A. Kavanaugh Grant/research support from: AbbVie Inc., Amgen, Astra-Zeneca, BMS, Celgene, Centocor-Janssen, Pfizer, Roche, and UCB., Consultant for: AbbVie Inc., Amgen, Astra-Zeneca, BMS, Celgene, Centocor-Janssen, Pfizer, Roche, and UCB.