Author: Ruhan Wei, Ph.D., Director of Duke Health Clinical Laboratories Clinical Chemistry and Duke Central Automated Laboratory, Assistant Professor of Duke University School of Medicine.
Accurate laboratory test results are prominent for high-quality patient care. Although laboratory-related errors could occur at any phase of laboratory testing (preanalytical, analytical, and postanalytical), most errors occur in the preanalytical phase. One common preanalytical error is sample turbidity, not due to lipemia. Poor preanalytical sample handling practices, such as rough sample transportation and poor centrifugation, could produce suspended cells or debris in a specimen. The turbidity of the specimen could cause erroneous results or unnecessary sample canceling. Typically, this can be resolved by clarifying the specimen with standard centrifugation, whereas truly lipemic specimens require the centrifugation at higher speeds and for a longer time. Another common preanalytical error is the delayed separation of the serum/plasma from the cellular material after centrifugation. Pseudohyperkalemia or pseudohypoglycemia could occur due to cellular potassium leakage and in vitro glycolysis.
Laboratories could utilize autoverification to detect preanalytical errors. Autoverificaoitn is a process where software-based algorithms perform actions on laboratory results without laboratory staff intervention. Results are automatically released based on the laboratory-defined acceptance parameters if they fall within the predetermined criteria. If results fail to meet the criteria, they can be reviewed by the laboratory staff before reporting.
In this study, two autoverification rules are used to assess the aforementioned preanalytical errors. In the first example, automated lipemia/turbidity indices (LI) are utilized to distinguish the lipemia or turbidity of the specimen. If the index exceeds the designed threshold(false lipemia), the specimen will be aliquoted and centrifuged to remove turbidity. If the LI is not lowered below the threshold after standard centrifugation(true lipemia), the aliquot will be spun at a higher speed for a longer centrifugation time to remove lipemia. In the second example, the preanalytical error is determined by monitoring the specimentype, collection and analysis times, and potassium and glucose concentrations. If the sample is received as whole blood and the time is greater than 4 hrs from the collection with critically high potassium (>6 mmol/L) or low glucose concentrations (<40 mg/L; <2.2 mmol/L), the order will be canceled with an explanatory comment.
This study analyzed the frequency of true and false lipemia in 3 common analytes with relatively low LI thresholds (direct bilirubin, aspartate transaminase, and alanine transaminase) and determined the frequency of pseudohyperkalemia/pseudohypoglycemia at a large academic hospital. The data showed 96% of direct bilirubin, 95% of aspartate transaminase, and 98% of alanine transaminase turbidity/lipemia alarms were due to sample turbidity, rather than lipemia. Of the total potassium results >6.0 mmol/L or glucose results <40 mg/dL (2.2 mmol/L), 30% and 50%, respectively, were noted to have delayed sample separation.
In conclusion, the study showed that a turbidity/lipemia algorithm noted the overwhelming majority of LI alarms were from sample turbidity unrelated to lipemia. Reduction in preanalytical sample turbidity would be of benefit to the laboratory. Furthermore, a pseudohyperkalemia/pseudohypoglycemia rule prevented the release of results with prolonged contact with cellular material. Implementing quality-based autoverification rules may minimize potential errors and improve the value of laboratory results.
利用检验结果自动验证的功能来监测分析前错误
为了提供高质量的临床服务，准确的实验室检测结果非常重要。尽管实验室相关错误可能发生在实验室检测的任何阶段（分析前、分析中和分析后），但大多数错误发生在分析前阶段。其中非脂血症类的样品浑浊是常见的分析前错误。分析前样品的不适当处理，例如颠簸的样品运输和离心分离不当，可能会在样品中产生悬浮细胞或碎片， 导致样品浑浊。样品的浊度可能导致错误的检查结果或不必要的样品取消。通常情况下，标准化离心可以解决这一问题，而真正的脂血标本需要更高速度更长时间的离心。另一个常见的分析前错误是离心后血清/血浆与细胞的分离延迟。这种情况下，由于细胞钾渗漏和体外糖酵解，很可能会导致假性高钾血症或假性低血糖症。
实验室可以利用检验结果自动验证的功能来发现分析前错误。 检验结果自动验证的功能是基于软件的算法，可以在没有实验室工作人员干预的情况下对实验室结果执行操作。如果结果符合预定标准，则根据已经设定的参数自动汇报检测结果。如果结果不符合要求，检测结果可以由实验室工作人员人工审查再汇报。
这项研究里有两个自动验证功能用来评估上述分析前错误。在第一个示例中，自动脂血/浊度指数 (LI) 用于区分样本的脂血或浊度。如果指标超过预定的阈值，样品将被分装并用标准离心以去除混浊（假性脂血）。如果在标准离心后 LI 没有降低到阈值以下，高速旋转以及增加离心时间来去除血脂（真性脂血）。在第二个示例中，我们通过监测标本类型、采集和分析时间以及血钾和葡萄糖浓度来确认分析前误差。如果样本收到的时候是全血，并且从样品采集到收到时间超过 4 小时，且血钾浓度极高（>6 mmol/L）或血糖浓度极低（<40 mg/L；<2.2 mmol/L），样品会被取消并附上解释。
本研究分析了 LI 阈值相对较低的 3 种常见检查项目（直接胆红素、谷草转氨酶和谷丙转氨酶）中真假性脂血症的概率。我们同时也对一家大型学术医院的假性高钾血症/假性低血糖症发生概率进行了分析。我们的数据显示 96% 的直接胆红素、95% 的谷草转氨酶和 98% 的谷丙转氨酶浊度/脂血症警报是由于样本浊度而并非真正的脂血症。对血钾 >6.0 mmol/L 或血糖 <40 mg/dL (2.2 mmol/L) 的样品分析，我们发现有 30% 的高血钾和 50% 的低血糖样品是由于分离延迟造成的。
总之，该研究表明，绝大多数 LI 警报是由于与脂血症无关的样品浊度引起的。因此降低样品浑浊度有利于提高实验室检查准确度。此外，由于细胞分离延迟导致的假性高钾血症/假性低血糖症，预设的该功能可以防止错误检查报告。检验结果自动验证功能可以最大限度地减少潜在错误的发生，并提高实验室结果的质量。
Reference:
Wei R, Légaré W, and McShane AJ. Autoverification-based algorithms to detect preanalytical errors: Two examples. Clinical Biochemistry (2022). https://doi.org/10.1016/j.clinbiochem.2022.06.010