American College of Rheumatology (ACR) response criteria is used to assess improvement in tender and swollen joint counts and in 3 of the 5 core measures (acute-phase reactant, physician global assessment, patient global assessment, pain, and physical function). From the clinical trial data on 5 approved biological products for the treatment of rheumatoid arthritis, population pharmacokinetic/pharmacodynamic models were developed to quantitatively describe the relationship between exposure and response rates of 3 individual components of ACR response criteria. The models were then used to simulate the clinical outcomes at various time points following different treatment regimens. The relative sensitivity of these criteria components was assessed using power analysis. As compared to the composite endpoints (ACR20/ACR50/ACR70), the individual ACR criteria components had adequate power and higher sensitivity in distinguishing treatment effects over placebo/methotrexate control. The 3 individual ACR criteria components appeared to have similar powers at different dose levels after long-term treatment. This research provides a unique approach to assess the relative sensitivity of the 3 binary components of ACR response criteria which would be useful to support future dose selection and trial design in the treatment of rheumatoid arthritis.
Keywords: American College of Rheumatology; binary outcome; exposure-response; modeling and simulation; power analysis; rheumatoid arthritis.
© 2017, The American College of Clinical Pharmacology.