BONE-MUSCLE-FAT INTERACTIONS AND PERFUSIVE EVIDENCE OF INFLAMMATION IN NON-OVERWEIGHT POSTMENOPAUSAL WOMEN WITH EARLY KNEE OA

Abstract

Postmenopausal women are affected by a decline of estrogen resulting in an established insult on bone (loss), muscle (loss), and fat (gain). These effects are likely not limited to central sites. We were interested in potential damage to the knee in this process and how it contributes to known phenotypes observed in osteoarthritis within this population of women. To glean evidence of inflammation, we investigated perfusion kinetics within periarticular bone, muscle, and fat tissues. To investigate how periarticular muscle and fat relate to subchondral bone properties; and to understand how perfusion may mediate these relationships. This was a cross-sectional study of 52 postmenopausal women recruited by convenience sampling at the University Health Network (Toronto). Women 60-85 years old with evidence of knee pain were included if their BMI upon screening was <=25 kg/m 2 and had no renal disease or contraindications to MRI or Gadolinium for Dynamic Contrast-Enhanced (DCE)-MRI. Time resolution on DCE-MRI scans was 7 ms and Toft's pharmacokinetic model was used to estimate perfusion rate, Kep, within each knee compartment and the infrapatellar fat pad (IPFP) using Tissue4D (Siemens) (Fig 1). Periarticular inter- and intramuscular fat (IMF) was segmented on transaxial proton-density weighted MRI using the iterative threshold-seeking algorithm (ITSA) in Python (Fig 2). Subchondral BMD, periarticular muscle mass, and IPFP content were quantified from compartment-specific pQCT scans at 200 µm in-plane pixel size on Stratec software (Fig 3). Where relevant, compartment-specific associations were obtained using general linear or logistic regression models, adjusting for age and BMI at minimum, with 95% confidence. A lower periarticular (calf side) muscle mass (b= -0.58(-1.02,-0.15)per 1mg) and density (b= -2.76(-5.35,-0.17)per 1mg/cc) was significantly associated with higher medial and lateral (b= -4.57(-7.21,-1.93)per 1mg/cc) subchondral BMD (mg/cc) of the tibia. While the directionality was similar for MR measures of IMF on the calf side, it was only marginally significant. In contrast, a higher IMF volume on the thigh side was associated with lower subchondral BMD of medial femoral (b= -2.39(-4.09,-0.69)) and tibial (b= -1.58(-3.05,-0.11)) compartments. Both higher periarticular IMF(MRI) and lower muscle mass(CT) were correlates of greater perfusion (Kep) of the lateral tibia (b= 1.69 to 2.88 per 1unit, p<0.02). Correspondingly, a higher perfusion of the lateral tibia(b=1.25(0.34,2.15)) was associated with a co-localized higher subchondral BMD. On the other hand, greater perfusion in the IPFP was related to a lower co-localized subchondral BMD (b= -1.70 to -2.22, p<0.01). Although no formal mediation analysis was performed in this smaller pre-pandemic sample, there is evidence to suggest that periarticular muscle and fat may influence subchondral bone by triggering nearby inflammation. The translated effects appear to differ - with correlates on the calf/tibia side resulting in higher BMD, and correlates on the thigh/femur side resulting in lower BMD. It is possible that the observed tibial BMD may reflect the beginning of a collapsing architecture. CIHR PJT156274. Arthritis Society Ken Smith Stars Career Development Award. None of the authors have disclosures. CORRESPONDENCE ADDRESS: andy.wong@uhnresearch.ca