| 邵宇飞,张健,杨良栋,等.基于倾向性评分匹配法探讨人工膝关节置换术后假体周围感染的危险因素[J].安徽医药,2025,29(11):2277-2281. |
| 基于倾向性评分匹配法探讨人工膝关节置换术后假体周围感染的危险因素 |
| Risk factors for periprosthetic joint infection after total knee arthroplasty using propensity score matching |
| |
| DOI：10.3969/j.issn.1009-6469.2025.11.034 |
| 中文关键词: 关节成形术，置换，膝 手术后并发症 假体周围感染 倾向性评分匹配法 危险因素 |
| 英文关键词: Arthroplasty, replacement, knee Postoperative complications Periprosthetic joint infection Propensity score match-ing method Risk factors |
| 基金项目: |
|
| 摘要点击次数: 497 |
| 全文下载次数: 160 |
| 中文摘要: |
| 目的采用倾向性评分匹配法控制混杂因素干扰，探讨人工膝关节置换术后假体周围感染( PJI)的危险因素。方法回顾性分析 2017年 1月至 2021年 12月在鞍山市中心医院行人工膝关节置换术的 1 557例病人的临床资料，根据术后有无发生 PJI将其分为 PJI组( 43例)和非 PJI组(1 514例)。采用倾向性评分匹配法，按照 1∶2匹配 PJI组和非 PJI组性别、年龄、术前诊断类型、美国麻醉医师协会( ASA)分级、麻醉方式等协变量，并采用单因素和多因素 logistic回归分析法分析匹配后人工膝关节置换术后 PJI的危险因素。结果倾向性评分匹配前， PJI组和非 PJI组性别、年龄、术前诊断比较差异有统计学意义( P<0.05)， ASA分级、麻醉方式比较差异无统计学意义( P>0.05)；倾向性评分匹配后， PJI组得到 43例病人，非 PJI组得到 86例病人，两组各协变量比较差异无统计学意义( P>0.05)。匹配后单因素分析结果显示， PJI组糖尿病占比［ 34.88%(15/43)比 16.28%(14/ 86)］、术前长期应用激素占比［ 27.91%(12/43)比 9.30%(8/86)］、手术时间［( 119.35±15.26)min比( 110.06±12.31)min］、术后引流时间 ≥48 h占比均高于非 PJI组，术前、术后次日血红蛋白水平［( 102.34±12.75)g/L比( 117.16±14.22)g/L］均低于非 PJI组( P<0.05)；排除共线性变量后行多因素 logistic回归分析，结果显示糖尿病、术前长期应用激素、术后次日血红蛋白水平较低、手术时间较长均是人工膝关节置换术后 PJI的独立危险因素( OR=2.35、5.20、0.72、1.72，P<0.05)。结论经倾向性评分匹配法控制组间混杂因素后，糖尿病、术前长期应用激素、术后次日血红蛋白水平较低、手术时间较长是人工膝关节置换术后 PJI的独立危险因素。 |
| 英文摘要: |
| Objective To investigate the risk factors for periprosthetic joint infection (PJI) after total knee arthroplasty using propen-sity score matching to control for confounding factors.Methods The clinical data of 1 557 patients who underwent total knee arthro-plasty at Anshan Central Hospital from January 2017 to December 2021 were retrospectively analyzed. Based on the occurrence of post-operative PJI, patients were divided into a PJI group (43 cases) and a non-PJI group (1 514 cases). Propensity score matching was ap-plied at a ratio of 1:2 to match covariates including sex, age, preoperative diagnosis, American Society of Anesthesiologists (ASA) classi-fication, and anesthesia method between the two groups. Univariate and multivariate logistic regression analyses were used to identifyrisk factors for PJI after matching.Results Before matching, significant differences were observed in sex, age, and preoperative diagno-sis between the PJI group and non-PJI group (P<0.05), whereas ASA classification and anesthesia method did not differ significantly (P> 0.05). After matching, 43 patients remained in the PJI group and 86 were included in the non-PJI group, with no significant differences in covariates between the two groups (P>0.05). Univariate analysis after matching indicated that the PJI group had a higher proportion of patients with diabetes [34.88%(15/43) vs. 16.28%(14/86)], a higher proportion of long-term preoperative steroid use [27.91%(12/43) vs. 9.30%(8/86)], longer operation time [(119.35±15.26) min vs. (110.06±12.31) min], and a higher proportion with postoperative drain-age duration ≥48 hours, while preoperative and postoperative day 1 hemoglobin levels [(102.34±12.75) g/L vs. (117.16±14.22) g/L] were lower compared to the non-PJI group (P<0.05). Multivariate logistic regression analysis after excluding collinear variables identified di-abetes, long-term preoperative steroid use, lower hemoglobin level on postoperative day 1, and longer operation time were as indepen-dent risk factors for PJI after total knee arthroplasty (OR=2.35, 5.20, 0.72, 1.72, P<0.05).Conclusion After controlling for intergroup confounding factors using propensity, diabetes, long-term preoperative steroid use, lower hemoglobin level on postoperative day 1, andlonger operation time are the independent risk factors of PJI after total knee arthroplasty. |
|
查看全文
查看/发表评论 下载PDF阅读器 |
| 关闭 |
|
|
|
