Mri Defined OA as Outcome Measure for an Osteoarthritis Prevention Study; Practical Implications and Feasibility
Year: 2022
Type: article
Abstract: As MRI is more sensitive to early changes than conventional radiography, structural knee OA changes on MRI could be considered as outcome measures for urgently needed preventive RCTs. However, the practical implications and feasibility of such measures in a preventive setting remain uncertain. To 1) evaluate the baseline screening rate of MRI defined OA, the 30 months incidence of MRI OA, and the long-term predictive ability of incident MRI OA on traditional OA outcomes among a potential target population for a preventive RCT; overweight and obese, middle-aged women, free of knee symptoms at baseline, and 2) provide sample size calculation using these outcomes. Data from the PROOF preventive RCT included women with a BMI ≥27 kg/m2, between 50 and 60 years, and free of knee symptoms (clinical ACR-criteria). At baseline and at 30-month follow-up, multi-sequential MRI (1.5T) of both knees were obtained and scored using MOAKS, to define MRI OA in the tibiofemoral (TF) and patellofemoral (PF) compartments using the consensus-based OA definition (Hunter et al. 2011). Knee symptoms and standardized semi-flexed knee radiographs were obtained at baseline, 30 months, and 80 months to assess the incidence of clinical (combined ACR-criteria) and radiographic (KL-grade ≥2) knee OA. In total, 438 knees from 230 women had complete data over the 80 months of follow-up. Mean age was 55.7 ± 3.2 years and mean BMI was 31.7 ± 3.6 kg/m2. After telephone screening to exclude the presence of knee symptoms, the prevalence of MRI OA in the TF and PF compartments was 10.0% and 9.6%, respectively. Excluding any baseline MRI OA (TF and/or PF) resulted in the exclusion of 69 knees (15.8%) from the trial. Adding radiographic screening (KL0 knees only, 53.4% at baseline) prior to MRI assessments would result in the exclusion of 6.0% of the knees due to prevalent MRI OA. Among knees eligible after telephone screening without MRI OA, the incidence of MRI OA in TF only, PF only, or TF and/or PF compartment(s) after 30 months was 6.6%, 3.7%, and 7.3%, respectively. Among baseline KL0 knees, MRI OA incidence in TF only, PF only, or TF and/or PF compartment(s) was 4.7%, 2.1%, and 5.6%, respectively. After 80 months and among the knees free of any MRI OA at baseline, the incidence of clinical and radiographic knee OA was 9.9% and 8.5%. Among knees with incident MRI OA in the TF and/or PF compartment after 30 months, the incidence of clinical and radiographic knee OA after 80 months was 21.9% and 28.1%. Anticipating 7.3% incidence of TF and/or PF MRI OA after 30 months in the control group and a 20%, 30%, or 40% reduction in the intervention group, 4267, 1883, or 1025 knees would be required to obtain 80% power and α of 95%, respectively, without accounting for any dropout. Using MRI OA as an outcome measure in a preventive RCT requires the absence of MRI OA at baseline. Screening for the absence of MRI OA at baseline would lead to the exclusion of 15.8% of eligible knees. Adding radiographic screening for KL0 would reduce the number of unnecessary MRIs, but comes at the expense of a larger screening sample due to the prevalence of KL0, and of lower incidence rates of MRI OA after 30 months (7.3% among all and 5.6% among KL0 knees). Incident MRI OA after 30 months leads to a two to three times higher incidence of established knee OA after 80 months. Even in this high-risk population of middle-aged women with overweight/obesity and when using a more sensitive 30-months outcome measure, the required sample size for a future RCT with incident MRI OA as primary outcome is well beyond that used in previous disease progression randomized clinical trials. The PROOF study is funded by The Netherlands Organisation for Health Research and Development and the FP7 EU Program. All authors declare no conflicts of interest. CORRESPONDENCE ADDRESS: [email protected]asmusmc.nl
Source: Osteoarthritis Imaging
Language: English
Subfield: Rheumatology
Field: Medicine
Domain: Health Sciences
Open Access status: hybrid
Estimated APC paid: $1,400
OpenAlex ID: W4285307231