利用质谱检测人工关节假体周围感染。

Detecting Periprosthetic Joint Infection by Using Mass Spectrometry.

机构信息

Department of Orthopedics, The First Medical Center, Chinese PLA General Hospital, Beijing, People's Republic of China.

Department of Orthopedics, The Fourth Medical Center, Chinese PLA General Hospital, Beijing, People's Republic of China.

出版信息

J Bone Joint Surg Am. 2021 Oct 20;103(20):1917-1926. doi: 10.2106/JBJS.20.01944.

DOI:10.2106/JBJS.20.01944

PMID:34097653

Abstract

BACKGROUND

Novel methods for diagnosing periprosthetic joint infection (PJI) are currently being explored. Mass spectrometry (MS) is an approach that can detect whole-protein changes in synovial fluid and may represent a promising method.

METHODS

Between March 2017 and July 2018, we successively collected synovial fluid samples from patients who were undergoing diagnostic hip or knee aspiration because PJI was suspected. A PJI diagnosis was based on the modified Musculoskeletal Infection Society (MSIS) criteria. Cluster analysis and receiver operating characteristic (ROC) curves were used to evaluate the results, which were quantitatively confirmed with parallel reaction monitoring in another patient group who underwent aspiration between August 2018 and January 2019.

RESULTS

A total of 117 synovial samples, including 51 PJI and 66 non-PJI samples, were analyzed with liquid chromatography-tandem MS (LC-MS/MS). The cluster analysis sensitivity and specificity based on differentially expressed proteins were 0.961 (95% confidence interval [CI], 0.854 to 0.993) and 0.924 (95% CI, 0.825 to 0.972), respectively. Myeloid nuclear differentiation antigen (MNDA) and polymorphonuclear leukocyte serine protease 3 (PRTN3) were the 2 most important markers for detecting PJI. The areas under the curves (AUCs) of MNDA and PRTN3 were 0.969 (95% CI, 0.936 to 1.000) and 0.900 (95% CI, 0.844 to 0.956), respectively. When MNDA and PRTN3 were combined as variables of a predictive model to diagnose PJI, the AUC reached 0.975 (95% CI, 0.943 to 1.000). Our parallel reaction monitoring-based quantitative analysis of another 40 synovial samples confirmed this result.

CONCLUSIONS

MS could be a powerful tool for diagnosing PJI using proteome information or 2 specific markers, MNDA and PRTN3. The parallel reaction monitoring strategy simplified the PJI detection process and provided quantitative results with similar conclusions.

CLINICAL RELEVANCE

The clinical application of MS adds a new powerful tool for the diagnosis of PJI, and the parallel reaction monitoring strategy lays a foundation for the clinical application of MS.

LEVEL OF EVIDENCE

Diagnostic Level II. See Instructions for Authors for a complete description of levels of evidence.

摘要

背景

目前正在探索诊断人工关节假体周围感染（PJI）的新方法。质谱（MS）是一种可以检测滑液中全蛋白变化的方法，可能是一种很有前途的方法。

方法

我们在 2017 年 3 月至 2018 年 7 月间，连续收集因疑似 PJI 而行髋关节或膝关节抽吸术的患者的滑膜液样本。PJI 的诊断基于改良的肌肉骨骼感染学会（MSIS）标准。我们使用聚类分析和接收者操作特征（ROC）曲线来评估结果，并在 2018 年 8 月至 2019 年 1 月间进行抽吸术的另一组患者中使用平行反应监测进行定量确认。

结果

共分析了 117 份滑膜样本，包括 51 份 PJI 样本和 66 份非 PJI 样本，采用液相色谱-串联质谱（LC-MS/MS）。基于差异表达蛋白的聚类分析的敏感性和特异性分别为 0.961（95%置信区间 [CI]，0.854 至 0.993）和 0.924（95% CI，0.825 至 0.972）。髓样核分化抗原（MNDA）和多形核白细胞丝氨酸蛋白酶 3（PRTN3）是检测 PJI 的 2 个最重要的标志物。MNDA 和 PRTN3 的曲线下面积（AUC）分别为 0.969（95% CI，0.936 至 1.000）和 0.900（95% CI，0.844 至 0.956）。当将 MNDA 和 PRTN3 作为预测模型诊断 PJI 的变量结合使用时，AUC 达到 0.975（95% CI，0.943 至 1.000）。我们对另外 40 份滑膜样本进行的基于平行反应监测的定量分析证实了这一结果。