文献速递质量管理与危急值临床生化

基于患者的指数加权移动平均质控模型在三种抗精神病药物及其代谢物LC-MS/MS检测中的开发与验证

Development and validation of patient-based exponentially weighted moving average quality control models for three antipsychotic drugs and their metabolites by LC-MS/MS.
PubMed 文献 · Clin Chim Acta · 2026-07-02

本研究开发并验证了基于指数加权移动平均(EWMA)的基于患者实时质控(PBRTQC)模型,用于阿立哌唑、氯氮平、喹硫平及其主要代谢物的LC-MS/MS治疗药物监测。模型采用1.5 SD截断和λ=0.03,在15%分析偏倚下多数分析物平均检测前患者数(ANPed)小于50,30%偏倚下均小于20,可实时补充传统室内质控。

要点

  • EWMA-PBRTQC模型可实时监测抗精神病药物LC-MS/MS检测的分析偏移。
  • 多数分析物最优模型为1.5 SD截断结合λ=0.03。
  • 15%偏倚下喹硫平ANPed<20,30%偏倚下所有分析物ANPed<20。

核心结论

  • 成功建立了针对抗精神病药物监测的EWMA-PBRTQC模型,可实时补充传统IQC并早期检测分析偏移。

研究设计与样本

  • 按时间顺序将浓度数据分为训练集和验证集,通过截断和缩尾优化截断限,评估不同权重因子λ的多个EWMA模型,基于假阳性报警率(FAR)和平均检测前患者数(ANPed)选择模型,并用模拟恒定和比例误差验证。

主要发现

  • 多数分析物最优模型采用1.5 SD截断和λ=0.03。
  • 15%分析偏倚下,ANPed在喹硫平<20,去烷基喹硫平和阿立哌唑<40,脱氢阿立哌唑<50,氯氮平/去甲氯氮平<30。
  • 30%偏倚下所有分析物ANPed均<20。

关键数值与推荐

  • 最优模型参数:截断1.5 SD,λ=0.03(多数分析物)。
  • 15%分析偏倚下ANPed:喹硫平<20,去烷基喹硫平和阿立哌唑<40,脱氢阿立哌唑<50,氯氮平/去甲氯氮平<30。
  • 30%分析偏倚下所有分析物ANPed<20。

临床怎么用

  • 可将EWMA-PBRTQC模型作为传统IQC的实时补充,用于抗精神病药物LC-MS/MS检测的早期分析偏移检测。

以上解读依据论文摘要(全文未开放获取或许可不允许翻译转载)。

原文

BACKGROUND: Therapeutic drug monitoring of antipsychotics is essential for personalized dosing. Traditional internal quality control (IQC) lacks real-time capability, potentially delaying error detection. Patient-Based Real-Time Quality Control (PBRTQC) can overcome this by continuously analyzing patient data. However, its application to antipsychotic drug assays remains underexplored. OBJECTIVE: To develop and validate novel PBRTQC models based on the exponentially weighted moving average (EWMA) for the real-time monitoring of three first-line antipsychotics (aripiprazole, clozapine, quetiapine) and their major metabolites-addressing a critical gap in laboratory quality assurance for psychiatric pharmacotherapy. METHODS: Chronologically ordered concentration data were split into training and validation sets. Truncation limits were optimized via trimming and winsorization. Multiple EWMA models with different weighting factors (λ) were evaluated. Model selection was based on the false positive alarm rate (FAR) and the average number of patients before error detection (ANPed), validated using simulated constant and proportional errors. RESULTS: Optimal models for most analytes employed trimming at 1.5 standard deviation (SD) with λ = 0.03. The introduced PBRTQC system demonstrated high sensitivity: for a 15% analytical bias, the ANPed was <20 for quetiapine, <40 for desalkylquetiapine and aripiprazole, <50 for dehydroaripiprazole, and < 30 for clozapine/norclozapine. All ANPeds fell below 20 at a 30% bias, confirming effective error detection. CONCLUSION: EWMA-based PBRTQC models specifically tailored for antipsychotic drug monitoring were successfully established. The models provide a practical, real-time complement to traditional IQC, enabling the early detection of analytical shifts and enhancing the reliability of laboratory data for clinical decision-making in psychiatry.

来源:Clin Chim Acta,查看原文。版权归原作者所有。

引用

  1. Guo Q, Pu Y, He J, et al. Development and validation of patient-based exponentially weighted moving average quality control models for three antipsychotic drugs and their metabolites by LC-MS/MS. Clin Chim Acta. 2026. PMID: 42392485. doi:10.1016/j.cca.2026.121216

供执业医师工作参考,不替代临床判断;数值与分级以指南原文为准。