INTRODUCTION: Osteoarthritis (OA), characterized by articular cartilage degradation and bone sclerosis, is a prevalent joint disease. Due to limited understanding of OA pathogenesis and lack of an accurate early diagnostic technique, OA is not confirmed until its late stage, necessitating joint replacement surgeries. The origin of OA has been under debate over the past decades. Conventional analyses of articular cartilage and bone such as histology, volume fraction (BV/TV) and elastic modulus from finite element analysis have demonstrated that articular cartilage degradation occur concurrently with changes in the subchondral bone. These measures, however, only offer a gross analysis of the tissue and cannot elucidate how and where the structural changes are initiated. Furthermore, recent biologic studies suggest an important role of the underlying subchondral bone in OA pathogenesis. For instance, uncoupled resorption in the subchondral bone has been shown to induce cartilage degradation and OA in the knee, and blocking bone resorption attenuates cartilage damage1. The paradox of how abnormal bone resorption, a process that lowers bone mass, can eventually lead to bone sclerosis remains unexplained. In this study, we hypothesize that changes in the plate-and-rod microstructure of subchondral bone play a crucial role in OA initiation and progression. We examined changes in subchondral trabecular plate-and-rod microstructure and articular cartilage in human knees with advanced OA and in guinea pigs that develop spontaneous OA. METHODS: Tibial plateaus were collected from knee OA patients undergoing knee arthroplasty (n=102) and from cadaver donors without arthritis or other bone diseases (n=20). Knee joints of Dunkin Hartley guinea pigs, which develop OA spontaneously at 4 months of age, and controls that do not develop OA, were collected at 1, 2 and 3 months of age (n=8 for each group). These studies have been reviewed and approved by the Institutional Review Board or Institutional Committee on Animal Use and Care at the Hong Kong University. All the specimens were scanned using micro computed tomography. For human specimens, five subchondral bone subregions (Medial, Central, Lateral, Anterior and Posterior) were extracted from both lateral and medial condyles; for animal specimens, one subregion was extracted from the center of medial condyle. Cartilage integrity was examined by histology and OARSI scores, and the subregions were divided into damaged cartilage regions and intact cartilage regions. These areas were subjected to individual trabecula segmentation (ITS) for plate and rod based morphological parameters and micro finite element (FE) analyses for mechanical properties. The specimens were then processed for cartilage evaluation histologically. RESULTS: ITS analyses of bone under intact cartilage of the lateral side discovered significant reduction in rod bone volume fraction (rBV/TV) but no change in plate bone volume fraction (pBV/TV) (Fig. 1B). In these regions, significant decrease in rod trabecular number (rTb.N) and thickening of both plates and rods (pTb.th and rTb.th) were observed. In subchondral bone under severely damaged cartilage, decrease in rBV/TV was also discovered, while increase in pBV/TV was significant. Decrease in the number of rod trabeculae and thickening of both plates and rods were similar to subchondral bone under intact cartilage. In the guinea pig study, while BV/TV, histology, and OARSI scores in OA strain was no different from control at month 2 and 3, ITS analyses found similar decrease in rBV/TV and rTb.N and increase in pBV/TV. DISCUSSION: 伀甀爀 results suggest that in OA, microstructural changes in the subchondral bone precede significant cartilage degeneration, and that ultimate bone sclerosis can be attributed to trabecular plate thickening. The trabecular rod and plate changes lead to uneven distribution of trabecular bone density, probably increasing focal shear stresses and thus damages in cartilage.