THU0425 COST-EFFECTIVE OA TRIALS REQUIRE ENROLMENT OF KNEES WITH DEFINITE JOINT SPACE NARROWING (KELLGREN LAWRENCE 3); DATA FROM 6,939 KNEES FROM THE OSTEOARTHRITIS INITIATIVE

Abstract

Background: The design of clinical trials for osteoarthritis is challenging; structural changes in tissues are quantitatively small and proceed very slowly. No clear guidance exists on how to optimise recruitment. We have previously shown that the use of radiographic joint space width of 2 to 4.5 mm is of major importance for improving responsiveness in clinical trials using MRI bone and cartilage outcomes. However, it can be technically challenging to screen for joint space width this carefully, so we considered whether Kellgren-Lawrence (KL) grade could be used as an enrichment strategy. As no other commonly used covariates have been shown to reliably increase responsiveness, it would be useful to know the numbers needed using ONLY an expert-read KL grade structural inclusion criterion. Objectives: To determine responsiveness of change in femur bone shape and cartilage thickness using a large observational dataset and calculate likely trial cohort sizes per arm for each KL grade. Methods: We used all knees from the Osteoarthritis Initiative which had MR images at baseline, 1 and 2 years, and a baseline KL grade (centrally read and adjudicated by 2 experienced radiologists). Quantitative 3D femur bone shape, and cartilage thickness in the central medial femoral region were used as outcome measures. Responsiveness was assessed using standardised response means “SRM” (CIs were assessed using the bootstrap method of Efron) and derived the number of patients per arm in a putative trial to demonstrate 50% change, at 80% probability, α=0.05. Results: 6,945 knees (3,667 subjects, 2,085 female) were included (KL 0: 2,798; KL1: 1,338 knees; KL2: 1,879 knees; KL3: 924 knees). Table 1 provides summary results and Figure 1 shows SRM and putative trial cohort numbers by KL grade for bone shape and cartilage thickness. Femur bone shape had higher SRM values at all timepoints, typically twice that of SRM for cartilage thickness in all KL groups. Cohort size when using cartilage thickness increased significantly between KL2 and KL3. Conclusion: Expert-read KL3 inclusion, using 2 independent radiologists, with strict attention to standards provides increased responsiveness for common MRI OA outcomes. It is worth noting that 80% of KL3 knees have OARSI JSN grade 3, and 78% of KL3 knees have radiographic JSW of 2 to 4.5 mm, and similar enrichment can be expected using these alternatives. Few clinical studies can afford cohort sizes of greater than 200 knees, and for these studies, using a large proportion of KL3 knees is critical, especially if cartilage thickness is to be used as the outcome. Detailed breakdown of SRM values, cohort size, mean change, and SD of that change for Figure 1. Units are in vector units for bone shape, and mm for cartilage thickness. Disclosure of Interests: Michael Bowes Shareholder of: Stryker Corporation, Employee of: Stryker Corporation, Alan Brett Shareholder of: Stryker Corp, Employee of: Stryker Corp, Philip G Conaghan Consultant for: Flexion Therapeutics, AbbVie, Medivir, Merck Serono, Novartis, GlaxoSmithKline