|
Abstract INTRODUCTION:
Mass spectrometry has long since been the gold standard for urine drug confirmation. Its ability to measure large numbers of analytes with high sensitivity and specificity has been the mainstay of this method.
Five years ago, in our clinical laboratory at NLHS, we developed and validated a urine drug confirmatory screen using LC-MS/MS to measure 55 drugs and/or their metabolites. We used a simple dilute-and-shoot approach employing the enzymatic conversion of drug glucuronides using a native β-glucuronidase enzyme. This method was validated according to CLSI guidelines C-50, C-52, and C-62 and CSA guideline Z316.8-18.
Upon implementation of this method, limitations were identified in multiple phases of testing impacting our ability to provide a robust clinical service. These limitations, along with steps taken to overcome them, will be the subject of this presentation.
METHODS:
Throughout this process, we strove to keep as much of our analytical method the same, recognizing that more in-depth changes would need more extensive validations.
Post-implementation of our validated method we encountered issues when changing our lot number of the native β-glucuronidase enzyme, which necessitated the addition of a filter plate to filter particulate matter coming from the enzyme solution. This unfortunately caused the loss of an analyte from our method. Further to that, we experienced supply change issues forcing us to change the enzyme source itself (from native to recombinant), which subsequently eliminated the need to use the filter plate and allowed us to re-introduce the previously lost analyte.
During our change to a recombinant enzyme, we also halved the amount of urine in our sample preparation to reduce the amount of material entering the analytical system. We made changes to the MS method to minimize the acquisition windows and optimize dwell times of the 150+ transitions.
Our current final validated method combines 25 μL of urine with 85 μL of an enzyme MasterMix consisting of acetic acid, recombinant β-glucuronidase and an internal standard solution in methanol. This mixture is incubated at room temperature for 15 minutes, diluted with 10 % ACN in water, centrifuged and transferred to another 96 well plate for injection into the LC-MS/MS.
We evaluated all modifications to our analytical method on the same LC–MS/MS system we use for routine patient testing. We also examined retrospective data from our final software reports and our laboratory information system (LIS) to investigate how improvements to both our analytical and post-analytical processes affected turnaround times and reporting accuracy.
RESULTS:
Work is ongoing in the pre-analytical phase of specimen suitability testing to implement more relevant criteria for flagging adulterated samples. This includes increasing the urine pH value that would flag a sample as potentially adulterated and reducing the number of samples that would have been inaccurately reported as such.
In the analytical phase, we implemented (and subsequently removed) a filter plate to clean up our samples, validated a different source of β-glucuronidase, and removed an analyte from our assay and added it to its own assay. Implementation of a recombinant enzyme resulted in cleaner and more diluted samples which improved the time needed between maintenance cycles and resulted in more system uptime. It also shortened the sample preparation time, resulting in a simplified and more efficient preparation procedure.
We also strove to gain a better understanding of some sources of interference in our assay, including interferences from metabolites of drugs included in our assay and interferences from drugs not included in our assay. We also observed mass spectrometric interferences leading us to change some transitions, as well as cross talk from very highly concentrated samples. We developed flags for some interferences to avoid releasing false positive results. Understanding these interferences better also allowed us to avoid repeating samples on subsequent runs, therefore improving our turnaround time for these samples.
In the post-analytical phase, we implemented an interface whereby the assay results are transmitted directly from our instrument processing software file to our LIS, eliminating the need for manual result entry and double checking of thousands of results. This improved the overall turnaround time by decreasing the reporting time. We are also working on implementing more updated automatic interpretation comments to improve the accuracy of our reporting.
CONCLUSION:
Five years ago, in our clinical laboratory at NLHS, we developed and validated a urine drug confirmatory screen that measures 55 drugs and/or their metabolites that uses LC-MS/MS. Since then, we have made many updates to the assay that improved instrument uptime, decreased the patient sample turnaround time and increased accuracy of patient reporting. All of this demonstrates that work does not end when assay validation is over. Post-implementation assay improvement is an ongoing effort, and this work underscores the importance of a regulatory structure that allows for these changes.
|
= Discovery stage. (53.14%, 2025)
= Translation stage. (22.33%, 2025)
= Clinically available. (24.53%, 2025)
