LC-MS/MS Quality Assurance in Production

Clinical laboratories excel at meeting quality goals through policies and procedures that make up a laboratory quality management system (QMS). Two critical components of any QMS are quality control (QC) and quality assurance (QA). The Clinical and Laboratory Standards Institute (CLSI) has developed several guidelines on establishing a versatile QMS for routine testing using Food and Drug Administration (FDA)-approved methods. However, until recently, labs had minimal guidance on post-implementation monitoring of liquid chromatography-tandem mass spectrometry (LC-MS/MS) for clinical diagnostics. While some general quality measures used for routine testing hold true for LC-MS/MS, labs must take additional measures to ensure that the time and effort they spend developing a reliable LC-MS/MS method is not wasted after implementation when they have difficulties maintaining quality.

A guiding principle of QA plans is that they be proactive—not reactive. For LC-MS/MS methods, labs must be vigilant and ensure that they catch potential problems. LC-MS/MS quality depends first on developing and verifying an accurate and robust method. For this, labs have several established guidelines, such as CLSI-C62A; FDA’s guidance, Bioanalytical Method Validation; the European Medicines Agency’s Guideline on Bioanalytical Method Validation; and the Scientific Working Group for Forensic Toxicology Standard Practices for Method Validation. While these offer ample guidance for development, it’s surprising that only CLSI-C62A provides best practices for QA and post-implementation monitoring. However, the QA guidelines typically advise that “an acceptable range/tolerance should be established during method development (by the laboratory),” suggesting that each method will call for its own quality measures.

Despite the lack of literature on what constitutes a QMS for LC-MS/MS testing, clinical laboratories often use several standard quality measures. As the English writer John Ruskin said, “Quality is never an accident. It is always the result of intelligent effort.” So how should labs intelligently design a QA plan for LC-MS/MS testing?

It is imperative that labs begin with plans that cover not only sample analysis performance but also instrument and assay performance over time. Often new issues arise with a method after a lab processes hundreds to thousands of samples and evaluates batch after batch. The key is catching new issues before they affect routine testing. Many labs use system suitability testing to monitor instrument performance. A system suitability test (SST) uses a reference solution to verify performance of the LC-MS/MS analytical system. This solution—including the analyte(s) and internal standard(s)—should be evaluated after instrument maintenance, a power outage, break in instrument vacuum, instrument tuning/calibration, and prior to sample analysis (Table 1).

Monitoring sample analysis performance—the largest component of an LC-MS/MS QA plan—can be subdivided into two categories: overall batch review parameters, and individual peak review parameters. The first focuses on calibration curve acceptance, internal standard (IS) recovery throughout the batch, and QC acceptance (Table 1). When generating the calibration curve, labs must use the same fit (linear or quadratic) and the same weighting factors (e.g. 1/x or 1/x2) as defined in their method validation. A change in the appropriateness of the fit should prompt the lab to investigate the root cause.

With the curve established, calibrator accuracy should be +15% for all points except the lowest calibrator (+20%). In addition, the curve slope should be r2 > 0.995. Changes in IS recovery within and between batches may indicate potential problems at multiple points in the analytical process, including instrument drift and charging, insufficient injection volume, poor sample preparative recovery, and most importantly, unacceptable ionization suppression or enhancement in an individual sample.

While labs should use standard clinical laboratory QC practice for LC-MS/MS methods, the number of QC samples analyzed during a batch should represent at least 5% of the total number of patient samples or at least six total (three concentrations analyzed in duplicate). Once the lab deems a batch acceptable, individual samples must meet pre-established ranges for retention time (Rt) and/or relative retention time (RRt) to the internal standard and ion ratio. Labs should calculate a mean Rt/RRt and ion ratio (qualifying ion peak area/quantifying ion peak area) for the calibrators in each batch, which should not vary significantly within or between runs. A failed ion ratio in a patient sample suggests a possible peak integration failure, interfering substance, and/or loss in assay sensitivity.

Finally, monitoring assay performance over time is essential. An LC-MS/MS QA plan should include procedures for daily, weekly, and periodic instrument maintenance. Special considerations to ensure optimal instrument performance for LC-MS/MS methods include monitoring the number of column injections and the vacuum and LC pressures. As with all clinical assays, proficiency testing and every 6 months verifying linearity, accuracy, and instrument correlations (if the laboratory is using more than one instrument for the same method) are not only required but key to providing quality results.

In addition, labs use consumable materials (sample preparation reagents, mobile phases, analytical column, calibration standards, QC material, IS) for LC-MS/MS methods, and often must manually prepare reagents. CLSI C62-A states that all new lots must be compared to current lots and run as unknowns to establish set-points. The same patient samples should be evaluated with both lots to ensure that the results are consistent.

Nothing is more frustrating than when a method is not performing to specifications and the lab can’t report results without extensive troubleshooting or instrument service. An intelligently designed QA plan cuts down significantly on instrument and method downtime. Each individual method in any given laboratory may not meet specific consensus QA guidelines; however, this does not negate the fact that these parameters are still essential to monitor. Future studies investigating the actual performance of clinical LC-MS/MS methods using QA metadata should provide laboratories with more guidance on defining a “high performing method” that we should all strive to achieve.

Kara L. Lynch, PhD, is assistant clinical professor and associate division chief of the clinical chemistry and toxicology laboratory at the University of California, San Francisco. +Email: [email protected]

CLN's Focus on Mass Spectrometry is supported by Waters Corporation.