Poster Presentations for Troubleshooting

Investigation of Noroxycodone Interference in LC-MS/MS Urine Drug Testing Jacob Rininger (Presenter) John Carroll University

Poster #1c View Map

This poster will be presented and discussed on Wednesday at 17:15 for 15 minutes in Montreal 3 (Track 2).

1. Problem
Routine LC-MS/MS urine drug testing identified multiple patient specimens as positive for noroxycodone with undetectable levels of the parent drug oxycodone and its secondary metabolite, oxymorphone. The mass spectrometry middleware program flagged the specimens for review based on a rule requiring at least two of the three compounds be detectable. Upon chart review, the positive noroxycodone results were inconsistent with the patients' controlled substance prescriptions and clinical histories. Further analysis by an external laboratory using an alternative analytical method yielded negative noroxycodone results for all specimens. This discrepancy prompted an investigation into a potential interference affecting noroxycodone quantification.

2. Method Information
• Urine specimens (100 µL) were hydrolyzed with 200 µL of Kura BGTurbo Enzyme solution (including the internal standard), followed by termination of the reaction with 300 µL of cold methanol. After centrifugation, 200 µL of the supernatant was combined with 300 µL of mobile phase, preparing the sample for injection.
• Waters Acquity UPLC-TQD Mass Spectrometer
• Mobile Phase A: 0.1% formic acid in water
• Mobile Phase B: 0.1% formic acid in acetonitrile (ACN)
• 4.5 minutes gradient LC program, with a flow rate of 0.6 mL/min
• LC gradient starts at 98:2 Mobile Phase A: Mobile Phase B
• Column: Waters Acquity UPLC BEH C18 1.7 µm, 2.1 x 100 mm
• Column oven temperature: 40°C
• Injection volume: 7.5 µL
• Quantitative MRM acquisition
• Noroxycodone transitions: Quantifier 302.1384 -> 186.9814; Qualifier 302.1384 -> 226.9733

3. Troubleshooting Steps
The investigation identified an interfering compound with the same ion mass and retention time as noroxycodone, which did not trigger ion ratio flags. To enhance specificity, a new qualifier transition (302.1384 → 284.1000) was evaluated and found to be exclusively detected in specimens truly positive for noroxycodone, effectively distinguishing authentic signals from interference.

4. Outcomes
Specimens confirmed to contain noroxycodone exhibited detectable peaks across all three monitored transitions, whereas those affected by interference lacked a peak for the third transition. Despite rigorous method validation, this rare interfering substance was only identified during clinical testing and was flagged by a rules-based algorithm in the middleware. These findings suggest that an unidentified compound contributed to a possible false-positive noroxycodone result being reported and emphasize the utility of enhanced data review using AI-based tools for improved performance of LC-MS/MS drug analysis.

Quantitative Sulfatide Measurement in Body Fluids: LC Optimization and Investigation of Non-Specific Adsorption Kaitlyn King (Presenter) The Children's Hospital of Philadelphia

Poster #1d View Map

This poster will be presented and discussed on Wednesday at 17:00 for 15 minutes in Montreal 3 (Track 2).

Background:
Metachromatic leukodystrophy (MLD) is an inherited lysosomal storage disorder caused by a deficiency in arylsulfatase A, leading to the accumulation of sulfatides in white matter. This accumulation results in progressive demyelination and neurodegeneration. Biochemically, MLD can be diagnosed based on the hyperexcretion of sulfatides in urine. Ex vivo gene therapy for MLD has been approved in US. Therefore, there is an urgent need to establish quantitative urinary sulfatide analysis for diagnosis and prognosis. Its measurement in CSF is also imperative for monitoring treatment efficacy.

Problem:
Previous studies have shown that the response of sulfatides is not consistent across multiple injections. Additionally, it has been reported that sulfatides are unstable after more than two freeze-thaw cycles. These findings have prompted us to investigate the underlying causes before proceeding with analytical and clinical validation.

Method Information:
We developed a UPLC-MS/MS assay to quantify 25 sulfatide species in urine and CSF.
• 20 uL urine or CSF extracted with 300 uL MeOH solution containing 3 isotope-labeled sulfatides as internal standards (IS)
• Waters ACQUITY UPLC coupled with TQ Absolute tandem mass spectrometer
• MPA: 50:50 Water:Acetonitrile, 0.1% formic acid
• MPB: 80:20 Acetonitrile:Isopropanol, 0.1% formic acid
• 7 min gradient with 0.7 mL/min flow rate
• Column: ACQUITY UPLC BEH C8 Column (1.7μm, 2.1mm x 50mm)
• Column temperature: 40 °C
• Injection volume: 10 uL
• Quantitative towards C16, C18, C24, C24:1 sulfatides. Semi-quantitative towards other
sulfatide species

Troubleshooting Steps:
We hypothesized that the unstable MS/MS response was due to the suboptimal LC condition. We conducted the LC-MS/MS analysis using two columns: the CORTECS C18 column and the ACQUITY UPLC BEH C8 column. Sulfatide samples in neat solution were repeatedly injected on both columns.
We hypothesized that the instability of sulfatide observed after multiple freeze-thaw cycles was primarily due to non-specific adsorption, rather than degradation of the sulfatide itself. This is
attributed to the high stability of sulfatide, particularly under neutral conditions, as well as the inactivity of the only sulfatide-degrading enzyme, ARSA, at neutral pH. Standard polypropylene and polystyrene containers used for the collection of urine and CSF create conditions for non- specific adsorption of analytes due to their electrostatic and hydrophobic properties. Furthermore, these matrices contain lower concentrations of proteins and lipids, which increases the possibility of non-specific adsorption during sample collection and processing. The number of freeze-thaw cycles the sample undergoes is another factor that could impact the degree of adsorption, which is an important consideration for sample storage conditions in a clinical setting.

We assessed potential non-specific adsorption in different matrices by performing sequential transfer and freeze-thaw experiments. Samples were tested in triplicate at three levels: baseline (no spike), low-level spike, and high-level spike. Standard mixes were spiked into pooled non- affected specimens, and each level was transferred across five separate primary sample collection containers for each matrix, followed by ten freeze-thaw cycles. Extractions were performed at transfers 1, 3, 5 and freeze-thaw cycles 1, 3, 5, and 10. Data were analyzed and compared across extractions for each matrix.

Results:
The sulfatide response, after normalization to the internal standard (IS), was found to be unstable when analyzed on the CORTECS C18 column. A 15–40% increase in response was observed after 50 injections. Surprisingly, the use of isotope-labeled IS did not mitigate this issue. On the contrary, the response was stable on the BEH C8 column across 200 injections.
We observed a significant decrease in sulfatide levels with each subsequent transfer in both urine and CSF collection containers. There was no evidence of analyte loss from multiple freeze-thaw cycles in either matrix.

Outcome:
Switching to the C8 column has stabilized the response, even for sulfatide species without an isotope-labeled internal standard. The non-specific adsorption observed in all tested collection containers for urine and CSF suggests that the addition of a blocking agent is critical for clinical diagnostics. Further studies are required to determine whether similar losses occur in plasma. If minimal loss is observed in plasma, bovine serum albumin (BSA) may be considered as a potential blocking agent due to its ability to compete for analyte binding. Alternatively, hexadecylpyridinium chloride monohydrate (HDP) should be tested as a blocking agent if plasma sulfatide loss is observed.

Start as You Mean to Go On; Controlling Batch Effects in Quantitative LC-SRM Measures Russell Grant (Presenter) Labcorp

Poster #2a View Map

This poster will be presented and discussed on Wednesday at 18:00 for 15 minutes in Montreal 3 (Track 2).

1. Problem
Conventionally, the start of a batch incorporates the fundamental parameters of assay determination, namely, calibrators for accuracy, IS response expectations and ion ratio response for outlier detection. This presentation will highlight issues and solutions we have implemented to overcome errors and drift when operating LC-MS/MS in clinical care.

2. Method Information
• Direct injection of diluted samples (TFC-LC) or SPE extracted samples (Plasma Catecholamines, PCats and Plasma Metanephrines, PMets)
• ARIA TX4 (4 streams).
• SCIEX API5000 and 7500
• MP-A: Load/Elute A: 0.1% Formic Acid
• Load B: 1:1 ACN:MeOH + 0.1% Formic acid (v/v), Elute B: 90:10 MeOH: H2O + 0.1% Formic acid (v/v)
• 3 min gradient LC program, 1-2 mL/min flow rate
• Column 1: TFC Cyclone P, 50 x 0.5mm, Column 2: Agilent XDB C18 50 x 2.1 mm, 5m 50z2mm Polymeric 0 x 3 mm, 2.6μm C18, with guard cartridge
• Injection volume 10-20 μL
• Quantitative SRM acquisition

3. Troubleshooting Steps
Initial runs included calibration curve re-injection at the end of a run, demonstrating unacceptable drift of back-calculated calibrator accuracy, IS response and/or Ion ratio. Repeat preparation and individual bracketing calibrators ruled out preparative errors leading to drift. Either a full plate of calibration curves or a single sample across a plate (prep/re-pool and re-plate) was used to study drift phenomenon as a function of source conditions (cold start) and/or multiplexing (switching between assays). Removal of the bypass valve (initiated pre and post analyte elution) failed to resolve changes in the ion source. Assay specific solutions were as follows.

<b>Clozapine/Norchlozapine</b>: While provisionally we accepted the 50% signal loss for running in parallel with identical chemistry for MPA/MPAG (different composition during elution) due to IS correcting drift and narrow analytical range. Implementation did not run in parallel due to IS recovery inaccuracies and included repeat priming injections (SST’s) prior to batch start (inefficient MS utilization). The assay was moved to another Mass spec to improve MS utilization and remove IS recovery bias.

<b>Fluconazole (antifungals)</b>: Repeat injection of calibration curves indicated 2 full curves required for system passivation to ensure calibrator accuracy and Ion ratio drift (used 3rd injection of curve for calibration) for launch of fluconazole assay. Eventually, commercial availability of a better IS (D4 – 3C13 change) ameliorated concentration and ion ratio drift.

<b>Chlorpromazine</b>: D3-Chlorpromazine response was stable but Chlorpromazine responses drifted after a cold start (quantitative error) even using the same neutral losses. Analysis of a single sample across a run and review of two analyte/IS transition pairs (mismatched neutral losses) for response drift correction of Chlorpromazine and D3 IS, indicating low energy CE transitions liable to batch effects that are not conserved between analyte and IS.

<b>Riboflavin</b>: Pre-screened Roboflavin and 13C4,15N2 IS transitions using 3 independent transitions indicated response drift for 1 of 3 transitions for the IS not observed by analyte to rule out batch effects by selecting transitions not liable to changes in response over the start of a run.

<b>Plasma Catecholamines</b>: Bypass in/out did not resolve disparate responses (>50% losses in parallel) for PCats assay, used PCats elution solvent on second pump added after PMets eluted on Pmets assay and introduced into the mass spectrometer to precondition source from PMets to PCats solvent chemistry prior Pcats sample analysis (response loss <10%).

4. Outcome
In each instance, drift in IS response (instrument sensitivity/outlier detection), ion ratio response (outlier detection) or analyte response drift that was not corrected by the IS (concentration error) was systematically resolved through multiple different approaches. This presentation will help the audience understand and resolve issues that we all have in assay design and operational utilization.

Distorted Oxymorphone Peak in Urine Drug Testing Panel Jaime Noguez (Presenter) University Hospitals Cleveland Medical Center/Case Western Reserve University

Poster #2b View Map

This poster will be presented and discussed on Wednesday at 17:30 for 15 minutes in Montreal 3 (Track 2).

1. Problem
Distorted peak shapes were observed in the oxymorphone quantification, qualification, and internal standard transition chromatograms on two LC-MS/MS systems, compromising the accuracy of quantitation.

2. Method Information
- A 100 µL urine sample was hydrolyzed with 200 µL of Kura BGTurbo Enzyme solution (including the internal standard), followed by termination of the reaction with 300 µL of cold methanol. After centrifugation, 200 µL of the supernatant was combined with 300 µL of mobile phase A, preparing the sample for injection.
- Waters Acquity UPLC-TQD Mass Spectrometer
- Mobile Phase A: 0.1% formic acid in water
- Mobile Phase B: 0.1% formic acid in acetonitrile (ACN)
- 4.5 minutes gradient LC program, with a flow rate of 0.6 mL/min
- LC gradient starts at 98:2 Mobile Phase A: Mobile Phase B
- Column: Waters Acquity UPLC BEH C18 1.7 µm, 2.1 x 100 mm
- Column oven temperature: 40°C
- Injection volume: 7.5 µL
- Quantitative MRM acquisition

3. Troubleshooting Steps
Oxymorphone was identified as an early-eluting compound within the panel, exhibiting consistent distortion on the left side of the peak. This observation suggested that the sample solvent might be the contributing factor.

Step 1: Further diluting the extracted sample (x2) with the mobile phase prior to injection resulted in improved peak shape.

Step 2: Reducing the injection volume also (3uL or 1uL) enhanced peak symmetry, whereas increasing the injection volume exacerbated peak distortion.

These steps were intended to minimize the concentration of organic solvent prior to injection or before the analytes reached the column, aligning conditions more closely with the initial LC gradient. The findings confirmed that the peak distortion was due to a high proportion of organic solvent in the extracted sample.

4. Outcome
Three approaches were identified to improve peak shape in this scenario:
- Replacing the tube connecting the autosampler and column with a longer, larger tube.
- Reducing the injection volume.
- Diluting the extracted sample with the mobile phase before injection or modifying the extraction procedure to lower the organic solvent content.

Contamination in 3MTO Internal Standard Peaks in a Quantitative LC-MS/MS Plasma Metanephrine Assay Sherrine Alleyne (Presenter) North West London Pathology

Poster #27a View Map

This poster will be presented and discussed on Wednesday at 17:45 for 15 minutes in Montreal 3 (Track 2).

1. Problem

A high level of background interference developed in the D4 3-methoxytyramine (3-MTO) Internal Standard (IS) transition (172.1>95.3) of our routine plasma metanephrines assay. This issue prevented the clinical reporting of 3-methoxytyramine.

2. Method Information
• 150 µl of patient plasma and 200 µl solution of D3 Metanephrine (MET), D3 Normetanephrine (NMET) and D4 3-MTO, IS solution mixed and extracted on a 96-well Oasis WCX µElution plate.
• Waters ACQUITY UPLC system
• Waters TQ-XS and TQ-S micro mass spectrometer
• Mobile Phase A: 100 mM ammonium formate, pH 3.0 with formic acid.
• Mobile Phase B : 100% acetonitrile (ACN)
• Waters Acquity UPLC BEH Amide 1.7um, with VanGuard Pre-Column
• UPLC gradient 2%A/98%B to 35%A/65%B over 4 minutes, then returns to the starting conditions , 2%A/98%B at 5.00 minutes at a flow rate of 0.20 ml/min & column temperature 35 oC.
• Injection volume: 15 µl on the TQ-XS and 5ul on the TQ-S micro.
• Quantitative SRM- MET – 180>148.3 NMET – 166.0>134.3 3MTO – 168.0>91.3

3. Troubleshooting Steps
Various troubleshooting steps were carried out to investigate the source of contamination.
Firstly, new internal standards made up and used in the assay, followed by a change of manufacturer for the ACN and the contamination was still present.

An Acid wash was carried out twice on the TQ-S micro, but each time the contamination was subsequently reintroduced into the system. After various experiments on the TQ-S micro it was established that the source of the contamination was mobile phase A. Therefore, new ammonium formate was ordered from a new supplier, but still the contamination remained. A new stock of the original supplier for the Ammonium formate was ordered and used in the mobile phase, but again still no change.

The in house LCMS grade H20 used for the mobile phase A was replaced with Honeywell bottled LCMS grade water, but there was still no change in the test assays.
Mobile phase A was made up with LCMS grade H20 and ammonium formate only which yielded a run without any contamination present. Thus the contamination was associated with the formic acid. Two other sources of formic acid were used to make up two separate mobile phase A, and surprisingly both runs showed contamination. When making up mobile phase A, a pH meter is used to titrate the pH of the mobile phase to 3, by slowly dripping the formic acid into the mobile phase using a plastic Pasteur pipette. Therefore, it was decided to not use a plastic Pasteur pipette and use a glass pipette to add the formic acid instead. When this was carried out and a test assay was run on both the TQ-XS and the TQ-S micro, no contamination was observed in the 3MTO IS peak. Although interference from the plastic Pasteur pipettes had not been observed previously, the laboratory had recently changed the supplier of these pipettes.

4. Outcome
Due to the plastic from the plastic Pasteur pipette leading to contamination of the formic acid, when carrying out the plasma metanephrines assay, plastic Pasteur pipettes are no longer used to avoid introducing the contamination to the assay and glass Pasteur pipettes are used instead.

Investigating Intermittent Retention Time Shifts Darcy Dochterman (Presenter) U.S. Department of Veterans Affairs

Poster #27b View Map

This poster will be presented and discussed on Wednesday at 16:45 for 15 minutes in Montreal 3 (Track 2).

1. PROBLEM
The isobaric cannabinoid metabolites delta-8-carboxy-tetrahydrocannabinol (THC) and delta-9-carboxy-THC have significant shifts to earlier retention times (RT) of 4.55 RT versus a normal RT of 4.77, and impaired chromatographic separation causing unreliable quantification of the assay target, delta-9-carboxy-THC. This is an intermittent problem affecting both patient and QC samples that was not observed during method development or validation.
2. METHOD INFORMATION
• 200ul urine treated with glucuronidase and extracted with solid phase extraction (SPE) using HLB uElution plate
• Waters Acquity LC with Column Manager
o Column heat: 45⁰C
o Injection volume: 10uL
o 2.1 x 50mm C18 1.7 um column with a 2.1 x 5mm 1.7 um C18 pre-column
o A 7.5 min method with a 4.0 minute isocratic flow of 0.3mL/min 29% aqueous 0.1% formic acid/0.5% ammonium acetate mobile phase A and 71% methanol mobile phase B
• Waters Xevo TQD
o MRM acquisition

3. TROUBLESHOOTING STEPS
The intermittent shift of retention time (RT) was seen across multiple sample preparations using multiple lots of reagents and the same method. Thus, a system problem was suspected. We eliminated the possibility of incorrectly prepared mobile phases by remaking them prior to sample injection. A new column and pre-column were installed. We noticed that when shutting off the binary solvent manager (BSM) and turning it back on, the RT did not shift from the normal 4.77 RT to 4.55 RT or earlier. Then upon investigation of the pressure trace files, a drop in psi was noted during the isocratic phase of the method. Constant 0.3mL/min flow of 29% mobile phase A and 71% mobile phase B2 should hold a steady psi. We then suspected the system leak was present, contaminating the mobile phases. The column manager contains two separate columns and separate mobile phases for each. The phenyl column uses mobile phase B1, acetonitrile/0.1% formic acid. The C18 column, what is used for the cannabinoid testing, uses mobile phase B2, methanol. Reversed-phase chromatography methods that use acetonitrile as the organic mobile phase have earlier RT since acetonitrile has higher elution strength. The accumulator and vent valve were checked for leaks, the mixer was changed and finally the solvent select valve was investigated. If the solvent select valve was leaking acetonitrile into the methanol mobile phase, the chromatographic peak of interest would shift earlier causing the decrease separation of delta-8-carboxy-THC and delta-9-carboxy-THC. When the peak tubing was placed together into mobile phase B2, primed, and a sample with known delta-8-carboxy-THC and delta-9-carboxy-THC standards was injected, the RT of the delta-9-carboxy-THC was 4.77 and separation between the isobaric compounds was optimal. Separation of the peak tubing into the respected mobile phase B1 and B2, priming, and injection of the known sample, shifted the RT earlier to 4.55. This proves that the acetonitrile B1 was leaking in the solvent select valve and contaminating the methanol mobile phase B2.
4. OUTCOME
The methanol mobile phase B2 was no longer contaminated by the acetonitrile mobile phase B1 once the select valve was replaced in the binary solvent manager of the LC/MS-MS. The RT was no longer shifted and the separation of chromatography between delta-8-carboxy-THC and delta-9-carboxy-THC was optimal for quantification of delta-9-carboxy-THC.