Enhancing Comparative Effectiveness Research By Combining Observational and Randomized Trial Data to Personalize the Choice Between Methotrexate and Triple Therapy for Methotrexate-Naïve Patients with Early Rheumatoid Arthritis

Glen S. Hazlewood^1,2, Cheryl Barnabe³, Gilles Boire⁴, Carol Hitchon⁵, Edward C. Keystone⁶, Boulos Haraoui⁷, J Carter Thorne⁸, Diane Tin⁹, Janet E. Pope¹⁰, Daming Lin¹¹, VP Bykerk¹² and CATCH investigators, ¹Institute of Health, Policy, Management and Evaluation, University of Toronto, Toronto, ON, Canada, ²Medicine, University of Calgary, Calgary, AB, Canada, ³Cumming School of Medicine, University of Calgary, Calgary, AB, Canada, ⁴Department of Medicine/Division of Rheumatology, Université de Sherbrooke, Sherbrooke, QC, Canada, ⁵University of Manitoba, Winnipeg, MB, Canada, ⁶Mount Sinai Hospital, University of Toronto, Toronto, ON, Canada, ⁷Institut de Rhumatologie, Montreal, QC, Canada, ⁸Southlake Regional Health Centre, Newmarket, ON, Canada, ⁹The Arthritis Program, Southlake Regional Health Centre, Newmarket, ON, Canada, ¹⁰University of Western Ontario, London, ON, Canada, ¹¹Rheumatology, Mount Sinai Hospital, University of Toronto, Toronto, ON, Canada, ¹²Hospital for Special Surgery, Weill Cornell Medical College, New York, NY

Meeting: ACR/ARHP Annual Meeting 2015

Date of first publication: September 29, 2015

Keywords: Comparative effectiveness and harms, Disease-modifying antirheumatic drugs, methotrexate (MTX) and rheumatoid arthritis (RA)

Session Information

Date: Sunday, November 8, 2015

Title: Rheumatoid Arthritis - Clinical Aspects I - Treatment Advances and Strategies

Session Type: ACR Concurrent Abstract Session

Session Time: 4:30PM-6:00PM

Background/Purpose: Randomized controlled trials (RCTs) are considered the gold standard for comparing efficacy of treatments, but the results may be less generalizable to clinical practice than observational studies. We have recently completed a Cochrane network meta-analysis that included all available RCT evidence and found a statistically significant increase in the odds of ACR50 response for triple therapy relative to methotrexate (MTX) alone in MTX-naïve patients. The objective of this study was to convert these relative treatment effects (odds ratios) into the expected probability of response for patients with different baseline characteristics, using data from an observational cohort. The overall aim was to enhance the knowledge required to achieve informed decision-making in practice.

Methods: We estimated the probability of an ACR50 response at 6 months for patients with ERA (2010 ACR criteria, symptoms ≤1-year), initiating methotrexate therapy (oral or subcutaneous) for patients in the multicenter, prospective Canadian Early Arthritis Cohort (CATCH). A multivariable logistic regression model with a candidate list of variables feasible to collect in clinic visits was developed to determine patient characteristics associated with ACR50 response. The final model was used to estimate the probability of ACR50 response with MTX monotherapy for patients with different baseline characteristics. We then determined the probability of ACR50 response for triple therapy, using the odds ratio derived from a completed Cochrane network meta-analysis of 150 trials.

Results: The population for the multivariable model included 666 patients with ERA: mean age 53 yrs, 74% female, mean DAS28 5.5, mean HAQ 1.1, swollen joint count 9.4. The mean starting dose of methotrexate was 19 mg/week. In the multivariable model, younger age, lower HAQ-DI and higher swollen joint counts were each independently associated with a higher odds of ACR50 response. The expected probability of an ACR50 response with MTX alone versus triple therapy for different patient characteristics is shown in Table 1. As the baseline chance of an ACR50 response increased, the absolute difference between MTX and triple therapy also increased.

Conclusion: By combining a predictive model from a large, prospective cohort with treatment effects from a network meta-analysis of all available RCTs, probabilities of treatment response with MTX and triple therapy can be estimated and may help inform evidence-based treatment choices for individual patients.

Table 1: Estimated probability of ACR50 response for methotrexate and triple therapy based on variable baseline characteristics

Change in baseline variables	Expected probability of ACR50 response		Absolute difference (95% CI)
Change in baseline variables	Methotrexate	Triple therapy (95% CI)	Absolute difference (95% CI)
Age increased (HAQ=1; SJC=10)
Age 20	32%	53% (37 to 70)	21% (5 to 38)
Age 40	25%	43% (28 to 61)	18% (4 to 36)
Age 60	18%	33% (21 to 51)	16% (3 to 33)
Age 80	13%	26% (15 to 42)	13% (2 to 29)
SJC increased (Age=55, HAQ=1)
SJC 5	14%	26% (16 to 43)	13% (2 to 29)
SJC 10	20%	36% (22 to 54)	16% (3 to 34)
SJC 15	27%	46% (31 to 64)	19% (4 to 37)
SJC 20	37%	57% (41 to 74)	20% (4 to 37)
HAQ increased (Age=55, SJC=10)
HAQ 0.5	24%	42% (28 to 61)	18% (4 to 36)
HAQ 1.0	20%	36% (23 to 54)	16%(3 to 34)
HAQ 1.5	16%	30% (18 to 47)	14% (3 to 31)
HAQ 2.0	12%	24% (14 to 40)	12% (2 to 28)

Disclosure: G. S. Hazlewood, None; C. Barnabe, Roche, Amgen, Abbott, 5; G. Boire, None; C. Hitchon, None; E. C. Keystone, Janssen Inc., 2,Abbott/AbbVie, 5,Amgen, 2,Bristol-Myers Squibb, 5,Janssen Inc., 5,Hoffmann-La Roche, Inc., 5,Janssen Inc., 2,Janssen Inc., 5,Merck Pharmaceuticals, 5,Merck Pharmaceuticals, 5,Pfizer Pharmaceuticals, 5,Pfizer Pharmaceuticals, 5; B. Haraoui, AbbVie, Amgen, BMS, Celgene, Janssen, Pfizer, Roche, UCB, 2,AbbVie, Amgen, BMS, Celgene, Janssen, Pfizer, Roche, UCB, 9,AbbVie, Amgen, BMS, Celgene, Janssen, Pfizer, Roche, UCB, 8; J. C. Thorne, Abbvie, 5,Amgen, 5,AstraZeneca, 5,Bristol-Myers Squibb, 5,Celgene, 5,Centocor, Inc., 5,Genzyme Corporation, 5,Hospira, 5,Pfizer Inc, 5,Roche Pharmaceuticals, 5,Genzyme Corporation, 8,Medac Pharma Inc, 8,Antares, 8,Abbvie, 2,Amgen, 2,AstraZeneca, 2,Celgene, 2,Eli Lilly and Company, 2,Novartis Pharmaceutical Corporation, 2,Pfizer Inc, 2; D. Tin, None; J. E. Pope, Abbott/AbbVie, 2,Abbott/AbbVie, 5,Amgen, 2,Amgen, 5,Actelion Pharmaceuticals Ltd., 2,Actelion Pharmaceuticals Ltd., 5,Bristol-Myers Squibb, 2,Bristol-Myers Squibb, 5,Hoffmann-La Roche, Inc., 2,Hoffmann-La Roche, Inc., 5,Hospira, 5,Janssen Inc., 2,Novartis Pharmaceutical Corporation, 5,Pfizer Pharmaceuticals, 2,Pfizer Pharmaceuticals, 5,UCB, 2,UCB, 5; D. Lin, None; V. Bykerk, Amgen, Pfizer, Janssen, BMS, UCB, Roche, Medexus, 2,Amgen, Pfizer, Janssen, BMS, UCB, Roche, Medexus, Antares, Regeneron, 5,Biogen Idec, 3,Biogen Idec, 1.

To cite this abstract in AMA style:

Hazlewood GS, Barnabe C, Boire G, Hitchon C, Keystone EC, Haraoui B, Thorne JC, Tin D, Pope JE, Lin D, Bykerk V. Enhancing Comparative Effectiveness Research By Combining Observational and Randomized Trial Data to Personalize the Choice Between Methotrexate and Triple Therapy for Methotrexate-Naïve Patients with Early Rheumatoid Arthritis [abstract]. Arthritis Rheumatol. 2015; 67 (suppl 10). https://acrabstracts.org/abstract/enhancing-comparative-effectiveness-research-by-combining-observational-and-randomized-trial-data-to-personalize-the-choice-between-methotrexate-and-triple-therapy-for-methotrexate-naive-patients-with/. Accessed .

« Back to ACR/ARHP Annual Meeting 2015

ACR Meeting Abstracts - https://acrabstracts.org/abstract/enhancing-comparative-effectiveness-research-by-combining-observational-and-randomized-trial-data-to-personalize-the-choice-between-methotrexate-and-triple-therapy-for-methotrexate-naive-patients-with/