Division of Biomedical Sciences (Genetics), Faculty of Medicine, Memorial University of Newfoundland, St John's, NL, Canada.
Discipline of Medicine, Faculty of Medicine, Memorial University of Newfoundland, St John's, NL, Canada.
Rheumatology (Oxford). 2021 Jun 18;60(6):2735-2744. doi: 10.1093/rheumatology/keaa693.
To identify endotypes of osteoarthritis (OA) by a metabolomics analysis.
Study participants included hip/knee OA patients and controls. Fasting plasma samples were metabolomically profiled. Common factor analysis and K-means clustering were applied to the metabolomics data to identify the endotypes of OA patients. Logistic regression was utilized to identify the most significant metabolites contributing to the endotypes. Clinical and epidemiological factors were examined in relation to the identified OA endotypes.
Six hundred and fifteen primary OA patients and 237 controls were included. Among the 186 metabolites measured, 162 passed the quality control analysis. The 615 OA patients were classified in three clusters (A, 66; B, 200; and C, 349). Patients in cluster A had a significantly higher concentration of butyrylcarnitine (C4) than other clusters and controls (all P < 0.0002). Elevated C4 is thought to be related to muscle weakness and wasting. Patients in cluster B had a significantly lower arginine concentration than other clusters and controls (all P < 7.98 × 10-11). Cluster C patients had a significantly lower concentration of lysophosphatidylcholine (with palmitic acid), which is a pro-inflammatory bioactive compound, than other clusters and controls (P < 3.79 × 10-6). Further, cluster A had a higher BMI and prevalence of diabetes than other clusters (all P ≤ 0.0009), and also a higher prevalence of coronary heart disease than cluster C (P = 0.04). Cluster B had a higher prevalence of coronary heart disease than cluster C (P = 0.003) whereas cluster C had a higher prevalence of osteoporosis (P = 0.009).
Our data suggest three possible clinically actionable endotypes in primary OA: muscle weakness, arginine deficit and low inflammatory OA.
通过代谢组学分析确定骨关节炎（OA）的表型。
研究参与者包括髋/膝关节 OA 患者和对照者。对空腹血浆样本进行代谢组学分析。采用常见因子分析和 K 均值聚类对代谢组学数据进行分析，以确定 OA 患者的表型。利用逻辑回归确定导致表型的最重要代谢物。检查与确定的 OA 表型相关的临床和流行病学因素。
纳入 615 例原发性 OA 患者和 237 例对照者。在测量的 186 种代谢物中，162 种通过质量控制分析。615 例 OA 患者分为 3 个聚类（A，66；B，200；和 C，349）。聚类 A 的患者丁酰肉碱（C4）浓度显著高于其他聚类和对照组（均 P<0.0002）。C4 升高被认为与肌肉无力和消耗有关。聚类 B 的患者精氨酸浓度显著低于其他聚类和对照组（均 P<7.98×10-11）。聚类 C 的患者溶血磷脂酰胆碱（含棕榈酸）浓度显著低于其他聚类和对照组（均 P<3.79×10-6）。此外，聚类 A 的 BMI 和糖尿病患病率高于其他聚类（均 P≤0.0009），冠心病患病率也高于聚类 C（P=0.04）。聚类 B 的冠心病患病率高于聚类 C（P=0.003），而聚类 C 的骨质疏松症患病率高于聚类 C（P=0.009）。
我们的数据表明原发性 OA 可能存在三种有临床意义的表型：肌肉无力、精氨酸缺乏和低炎症性 OA。
