Aims: This study aimed to investigate the relationship between physical activity (PA) and the risk of osteoarthritis (OA) and its subtypes.Methods: We included participants from the UK Biobank aged 37 to 73 years from February 2006 to June 2010. Baseline PA levels were categorized as high (≥ 3,000 metabolic equivalent of task (MET)-min/week), moderate (600 to < 3,000 MET-min/week), or low PA (< 600 MET-min/week) based on current public health guidelines. The associations between PA and OA and its main subtypes (hand, hip, and knee OA) were analyzed using Cox-proportional hazard models. The mediating role of BMI was tested under a causal counterfactual framework.Results: The median follow-up period was 12.50 years, with 25,036 incident total OA cases. Compared to low PA, we found no significant association between moderate PA and total OA risk. However, high PA significantly increased the risk of total OA by 19% (hazard ratio (HR) 1.19, 95% CI 1.15 to 1.23). This result is consistent with knee OA and hip OA, where HR was 1.25 (95% CI 1.19 to 1.31) for knee OA risk and 1.17 (95% CI 1.10 to 1.24) for hip OA risk, respectively. Moderate PA was associated with 14% (HR 0.86, 95% CI 0.76 to 0.97) reduction in the risk of hand OA. Moreover, we found that BMI mediated the association between PA and OA risk, with the mediating proportion for total OA, hand OA, knee OA, and hip OA, at 75.48%, 2.42%, 10.20%, and 51.39%, respectively.Conclusion: These findings suggest that high levels of PA increased the risk of total OA, as well as knee and hip OA, while moderate levels of PA significantly mitigate the risk of hand OA. BMI appears to mediate the relationship between PA and OA risk. Future research should elucidate the underlying mechanisms.Cite this article: Bone Joint Res 2025;14(7):656–665.

Osteoporosis (OP) has the characteristics of a systematically impaired bone mass, strength, and microstructure. Long non-coding RNAs (lncRNAs) are longer than 200 nt, and their functions in osteoporosis is yet not completely understood. We first harvested the bone marrow mesenchymal stem cells (BMSCs) from ovariectomy (OVX) and sham mice. Then, we systematically analyzed the differential expressions of lncRNAs and messenger RNAs (mRNAs) and constructed lncRNA–mRNA coexpression network in order to identify the function of lncRNA in osteoporosis. Totally, we screened 743 lncRNAs (461 upregulated lncRNAs and 282 downregulated lncRNAs) and 240 mRNAs (128 upregulated and 112 downregulated) with significantly differential expressions in OP compared to normal. We conducted Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) functional analyses to investigate the functions and pathways of the differential expression of messenger RNAs (mRNAs), a coexpressed network of lncRNA/mRNA. Quantitative PCR (qPCR) validated that the expressions of NONMMUT096150.1, NONMMUT083450.1, and NONMMUT029743.2 were all downregulated, whereas NONMMUT026970.2, NONMMUT051734.2, NONMMUT003617.2, and NONMMUT034049.2 were all upregulated in the OVX group. NONMMUT096150.1, as a key lncRNA in OP, was identified to modulate the adipogenesis of BMSCs. Further analysis suggested that NONMMUT096150.1 might modulate the adipogenesis of BMSCs via the peroxisome proliferator-activated receptor (PPAR) signaling pathway, AMPK signaling pathway, and the lipolysis regulation in adipocyte and adipocytokine signaling pathway. Our study expands the understanding of lncRNA in the pathogenesis of OP.