Can Wooden-Tip Mass Spectrometry Detect THC at Oral-Fluid Cutoffs? A Primary Laboratory Study

Short answer: Not with the unmodified wooden-tip electrospray ionization mass-spectrometry method evaluated in this experiment. Investigators reported a 40,000 ng/mL limit of detection for delta-9-tetrahydrocannabinol (THC) in both experimentally spiked urine and oral fluid. The oral-fluid cutoff shown for THC in the study’s comparison guidelines was 2 ng/mL. The same apparatus measured several other drugs much more sensitively. This is an analyte-specific validation result, not a clinical finding that cannabis was absent, a measured rate of missed real-world users, or an impairment determination.

The underlying primary investigation is Ng, So, Hu, and Yao, Rapid detection and quantitation of drugs-of-abuse by wooden-tip electrospray ionization mass spectrometry, Journal of Food and Drug Analysis 27 (2019): 428–438, published online in 2018; DOI: 10.1016/j.jfda.2018.09.002. The study used prepared analytical materials, not a cohort of cannabis-impaired drivers.

What question did the experiment actually test?

The authors asked whether a low-preparation, direct-ionization mass-spectrometric technique could detect and quantify several drugs in urine and oral fluid. Their targeted compounds included methamphetamine, MDMA, cocaine, benzoylecgonine, heroin, 6-acetylmorphine, morphine, THC, and the THC metabolite THC-COOH. The laboratory compared analytical signal, calibration, accuracy, precision, limits of detection, and limits of quantitation. These are measurements of a method’s performance under prescribed analytical conditions. They are not measurements of driving performance or toxicity in an exposed individual.

This distinction matters because a single instrument can perform adequately for one drug while being unsuitable for another. An impressive methamphetamine limit does not establish a THC limit. The proper question is not whether “mass spectrometry” generally detects cannabinoids, but whether this ionization configuration, this analyte, this specimen matrix, and this decision threshold were validated together.

How did a wooden tip become the ion source?

A sharpened commercial wooden tip, about 1.5–1.7 cm long, was placed at a nano-electrospray mass-spectrometer source. Researchers applied approximately 2 µL of prepared sample and a 3.5-kV potential to the tip. The instrument was a Waters Quattro Ultima triple-quadrupole mass spectrometer. Detection used multiple-reaction monitoring (MRM) transitions, with positive-ion mode for most targeted compounds and negative-ion mode for THC-COOH. The analyte signals appeared quickly, and a sample could usually be screened within minutes.

This technique dispensed with a conventional chromatographic separation step. It should not be described as a validated gas-chromatography or liquid-chromatography confirmation procedure merely because a mass spectrometer was present. Direct introduction also raises potential matrix-related interference and signal-suppression questions; the published method’s operating conditions and validation results determine what it can establish.

Were the specimens from cannabis users or drivers?

The validation prepared spiked blank urine and blank oral fluid by adding known analytical standards. Standard mixtures were combined with internal-standard solutions and acidified methanol before application to the wooden tip. Each sample solution could be run repeatedly, and the accuracy/precision work used low, middle, and high spiked concentrations and multiple individual wooden tips. Neither the reported number of technical measurements nor the number of tips represents a number of independent intoxicated people.

The paper did not report a prospective set of authentic THC-positive drivers compared against a reference method to estimate real-world clinical sensitivity, specificity, predictive value, or false-negative frequency. Thus, its 40,000-ng/mL THC analytical detection limit is striking but is not itself a directly observed percentage of people whom this technique would classify incorrectly.

Which biological matrices were measured?

The two validated matrices in this paper were urine and oral fluid. Urine testing may target metabolite evidence of prior exposure; oral-fluid procedures often focus on parent THC, although matrix properties and collection methods vary. The paper attempted both parent THC and THC-COOH, but its reported sensitivity and quantitative performance for those targets were inadequate. It did not directly assay forensic venous whole blood, separated serum, or separated plasma for THC concentration in living subjects. No breath-alcohol measurement, vitreous specimen, hair specimen, or postmortem body tissue formed part of this validation.

A number labeled ng/mL must remain attached to the matrix, drug species, analytical method, and prepared-sample conditions that generated it. A urine THC-COOH screening cutoff cannot be recast as a venous whole-blood parent-THC concentration or as a time-of-use estimate. For an overview of matrix distinctions, see the published biological matrices guide.

What were the actual detection limits?

The original Table 3 reports the following experimental limits of detection. These are method-specific LOD values, not measured concentrations in case specimens, and not legally controlling cutoffs for every jurisdiction or date. The guidelines printed alongside them provide historical comparators in the source publication.

For this experimental setup, the reported oral-fluid THC LOD was 20,000 times the 2-ng/mL comparator appearing in the paper. That ratio describes the separation between two numerical thresholds, not a measured 20,000-fold error rate, a patient risk, or the performance of every modern THC mass-spectrometric assay. The method performed very differently for methamphetamine: its 12.5-ng/mL oral-fluid LOD fell below the source’s 15-ng/mL comparator. MDMA and cocaine did not satisfy those particular oral-fluid comparator values with the same analytical margin.

How are LOD and LOQ different?

The paper defined a limit of detection using a signal-to-noise criterion near 3 and a limit of quantitation using a criterion near 10, with blank measurements incorporated into that evaluation. Detection addresses whether an analytical response is distinguishable from background at a stipulated criterion; quantitation asks whether a concentration can be reported with sufficiently characterized measurement performance. Neither guarantees identification without appropriate selectivity checks.

For methamphetamine in urine, the reported LOD was 25 ng/mL and LOQ was 50 ng/mL. For THC, a 40,000-ng/mL LOD was reported in the two matrices, but a usable LOQ and satisfactory accuracy/precision dataset were not available. It would be scientifically incorrect to invent a THC LOQ by multiplying the detection limit or by borrowing the methamphetamine ratio. Nor should a merely detectable THC signal be called reliable THC quantitation at low oral-fluid concentrations.

Why did the same technique work better for other drugs?

The authors discussed the chemical behavior of each compound at the wooden surface and under electrospray conditions. Methamphetamine and some other amine-containing drugs produced stronger signals. THC, a phenolic compound, protonated less efficiently under the experimental conditions, while its carboxylated metabolite THC-COOH did not gain enough usable sensitivity in the evaluated negative-ion method. Matrix interactions, retention on wood, and ion suppression without chromatographic separation were plausible analytical influences discussed by the researchers, but the paper did not isolate and quantify a single definitive cause for every failed cannabinoid result.

The lesson is analyte-specific: using a mass-selective detector does not automatically rescue a weak ion source. Method-development changes, including surface treatments and potentially additional separation or extraction, could improve performance, but such modifications were not validated for THC in this specific reported procedure. Any newer method must be judged using its own full validation, not by transferring the numerical limits in this older pilot study.

What did the noncannabinoid quantitative validation show?

The investigators used at least five calibration concentrations, labeled internal standards where available, and repeated spiked levels to assess accuracy and relative standard deviation (RSD). Source Table 2 gave specific examples in urine and oral fluid. The entries below are actual experimental spiked levels—not concentrations collected from real drug users.

These five examples illustrate both success and qualified performance. The source’s own discussion identified morphine as a weaker quantitative target, and THC/THC-COOH quantitative accuracy and precision data were unavailable because signals were poor. It would be erroneous to assign the favorable methamphetamine or cocaine RSD to THC. Technical repeats also do not establish between-person variability, diagnostic classification, or routine field robustness under all environmental and specimen conditions.

Did the mass transitions alone prove identification?

Multiple-reaction monitoring detects selected precursor-to-product ion transitions. Deuterated internal standards helped manage differences in absolute signal between wooden tips. In a separate methamphetamine example, the study showed further tandem mass spectral fragment information supporting an analyte assignment. These features improve analytical specificity, but do not prove that every possible endogenous or exogenous interferent was excluded. In particular, no chromatographic retention time separated potential isobaric or structurally related species in the main wooden-tip process.

For a case report, the analyst should inspect what identification criteria were actually met in the case method: relevant transitions, product-ion ratios or spectra, calibrators, independent controls, blanks, carryover and interference experiments, acceptance thresholds, and any confirmatory test. Read more about the general distinction between screening and chemical identification in the published confirmatory testing guide.

Could a negative result with this assay rule out THC?

No. Where the reported THC detection threshold is very high relative to the relevant analytical decision point, the absence of a reportable signal cannot establish absence of THC below the method’s LOD. The source did not produce an authentic clinical series defining its real-world false-negative rate, so it is equally inappropriate to attach a fabricated percentage or predictive value to a negative wooden-tip result.

The proper interpretation would depend on the validated version used, the actual specimen, how it was collected and stored, its THC target and any metabolites, the reportable detection limit, QC status, and an appropriate complementary method. This study specifically cautions against treating a prototype’s sensitivity for stimulants as proof of adequate cannabinoid testing.

Could a positive signal establish use time or impairment?

No. A positive signal generated from a spiked urine or oral-fluid test mixture establishes at most a laboratory response under that preparation. Even a fully confirmed positive in an authentic specimen would require distinct evidence to assess identity, exposure, time of last use, route, dose, present drug effect, or impaired driving. The experiment did not measure THC concentrations in a driver’s venous whole blood, behavioral performance, psychomotor tasks, crash outcomes, or clinical impairment scores.

Urine and oral-fluid measurements are not automatically interchangeable with separated plasma or serum, and none can be silently converted to legal whole-blood THC without appropriate validated evidence. A laboratory result also cannot determine a person’s intent or credibility simply because a witness disputed consumption. Collection and handling fundamentals are outlined separately in the published specimen collection guide.

Why is this not a validation of a roadside THC device?

The instrument was a laboratory triple-quadrupole mass spectrometer driven by a wooden ionization tip, not a standalone roadside analyzer operated on motorists. Although the workflow was designed to be fast and required little sample preparation, speed is only one element of field fitness. Operational validation would also need defined collection devices and volumes, specimen stability, temperature and humidity testing, interoperator and between-laboratory reproducibility, interference and cross-reactivity assessment, ruggedness, maintenance, quality assurance, and verified performance around the intended detection threshold.

The study did not provide those full field-performance data, and it plainly did not support the required THC analytical sensitivity at the benchmark it displayed. Therefore it cannot be used to claim either that all rapid mass-spectrometric devices fail for THC or that this particular prototype is suitable for routine cannabinoid roadside enforcement.

What evidence would matter in a forensic laboratory dispute?

First, identify the precise test and whether a specimen was urine, oral fluid, venous whole blood, separated plasma or serum. Second, establish which chemical species was targeted: parent THC, THC-COOH, or something else. Third, obtain the applicable laboratory validation rather than a brochure or a research-paper abstract. The laboratory should be able to document calibrated range, LOD, LOQ, selectivity, precision, bias, carryover, matrix effect, extraction or direct introduction, preservation conditions, and decision thresholds for the actual analyte and matrix.

Fourth, distinguish an exploratory spiked-material experiment from blinded testing of authentic positives and negatives. Fifth, review the contemporaneous run records and chain-of-custody documentation rather than inferring that published performance was identical at a later date. Finally, separate whether the method produced an analytically valid identification from what any particular individual’s activity or driving behavior was. Those are different evidentiary questions.

What are the strongest limitations of this primary study?

It was a method-development experiment using spiked laboratory materials; the number of independent donor specimens and performance in an authentic field population were not established. It did not report a validated THC quantitative range, THC quantitative precision, or THC-COOH detection limit suitable for interpretation. Chemical extraction and matrix behavior were limited to the study’s mixture preparation and direct-ionization platform; chromatographic separation was absent. Historical guideline cutoff examples are not a substitute for jurisdiction-specific current rules or validated laboratory reporting criteria.

The authors did not use a behavioral endpoint, did not administer cannabis to participants, and did not test actual road users. Even strong performance for certain stimulant targets should not be generalized to other analytes. Conversely, the poor reported THC performance should not be generalized to every form of electrospray or LC–MS/MS testing. It describes a particular wooden-tip configuration under the reported conditions.

Frequently asked questions

Did the investigators determine how many cannabis users this assay missed?

No. Their cannabinoid analytical limits were obtained with laboratory-spiked urine and oral-fluid preparations. They did not enroll a known-use clinical or driver population and did not publish a true-positive/false-negative patient classification table for THC. An assay’s detection limit and its observed clinical false-negative rate are different statistics.

Was 40,000 ng/mL a person’s measured THC concentration?

No. It was the study’s THC limit of detection in each tested spiked matrix. It is not a case concentration, a typical concentration, a finding in whole blood, or evidence of impairment. The 2-ng/mL oral-fluid number beside it was a guideline cutoff in the authors’ comparative table, not the measured amount in any participant.

Did THC-COOH perform better than parent THC?

No usable quantitative advantage was established. The paper reported inadequate signal for both cannabinoid targets, including attempts at THC-COOH in negative-ion mode. Table 3 supplied no THC-COOH LOD or LOQ, and Table 2 provided no reliable THC or THC-COOH precision dataset. A missing result is not zero concentration.

Why could the method detect methamphetamine far more sensitively?

Methamphetamine generated a stronger ionization response in this setup. Its reported urine and oral-fluid LODs were 25 and 12.5 ng/mL. THC’s phenolic chemistry and the study’s ionization conditions produced much poorer analytical signals. That contrast is a method-and-analyte finding, not evidence that people produce different drug quantities.

Can an MRM transition replace a full confirmation?

Not automatically. MRM is chemically selective, and the authors also demonstrated tandem-mass-spectral fragments for methamphetamine. A real case still requires the validated identification and quality-control criteria of the method used, and sometimes an independent or chromatographically separated confirmation. No single generic instrument name can settle every interference question.

Does a high oral-fluid THC result establish a blood THC value?

No. Oral fluid and venous whole blood are distinct physical matrices, with different collection dynamics and concentration relationships. This experiment did not derive or validate a blood-to-oral-fluid conversion. It also did not determine a person’s last-use time, dose, tolerance, real-world driving performance or impairment.

What can this paper reliably contribute to a forensic review?

It provides directly observed evidence that a rapid prototype performed very differently across analytes: some stimulant limits were comparatively favorable, while its parent-THC LOD was far above the oral-fluid cutoff in its own benchmark table. That supports a demand for analyte- and matrix-specific method validation, not a universal conclusion about all cannabis tests or any one person’s impairment.

Bottom line

The method was rapid, but rapid and analytically adequate are not synonyms. In the actual primary experiment, the unmodified wooden-tip MRM mass-spectrometric assay gave a 40,000-ng/mL THC LOD in both urine and oral fluid, did not establish useful THC-COOH detection or cannabinoid quantitation, and performed better for selected other drugs. Its principal forensic lesson is to examine the precise substance, matrix, validated sensitivity, and question before interpreting a result. The complete original source is available under DOI 10.1016/j.jfda.2018.09.002.