Can 11-OH-THC Glucuronide in Serum or Urine Prove Recent Cannabis Impairment? Evidence From a Synthesized Reference Standard

Answer: No. Detecting a glucuronidated metabolite of 11-hydroxy-THC can help characterize cannabinoid metabolism and support analytical identification, but it does not, by itself, establish when cannabis was used or whether a particular person was impaired. In a 2020 original forensic analytical study, researchers chemically synthesized a reference standard for the alcoholic glucuronide of 11-hydroxy-Δ9-tetrahydrocannabinol (11-OH-THC). They matched that standard with an in-vitro liver-fraction product and with material in authentic cannabis-user specimens. The alcoholic conjugate was identifiable in eight of ten serum specimens. Only two of ten urine specimens showed the main hydroxylated-THC glucuronide signals under the study’s different urine procedure. The authors also described a probable phenolic glucuronide, but did not establish its identity using a synthesized phenolic reference standard.

Those findings support a specific chemical conclusion: more than one attachment site for glucuronic acid needs consideration when interpreting 11-OH-THC-related peaks. They do not create a validated serum-to-urine conversion, a cannabis-use clock, or an individualized driving-impairment test. The primary evidence is Hassenberg and colleagues’ original article, “Investigation of phase II metabolism of 11-hydroxy-Δ9-tetrahydrocannabinol and metabolite verification by chemical synthesis of 11-hydroxy-Δ9-tetrahydrocannabinol-glucuronide”.

What forensic question did the experiment actually answer?

The central question was structural identity, not intoxication. The metabolite 11-OH-THC has two chemically different locations where a glucuronic-acid group might attach: an alcohol-bearing site at carbon 11 and a phenolic hydroxyl site on the cannabinoid ring system. Products formed at different sites can have the same elemental composition and nominal or exact mass. A mass-spectrometric signal of the expected mass therefore cannot, without more evidence, tell an analyst which chemical structure produced it.

The investigators examined products generated by a pooled human-liver S9 fraction exposed to Δ9-THC, compared the signals with authentic urine and serum from cannabis users, and prepared a chemically defined alcoholic 11-OH-Δ9-THC glucuronide for reference. Each element answered a different part of the identification problem. The liver-fraction experiment indicated what metabolites could be generated in vitro. The authentic specimens addressed whether corresponding signals appeared in collected biological material. Chemical synthesis supplied a structural comparator for one particular signal. None of these components was a controlled study of crash risk, reaction time, or the duration of intoxication.

Three evidence streams—not 54 cannabis-dosed participants

The original experiment used a pooled human liver S9 fraction from 34 donors. These were contributors to a commercial enzyme preparation, not 34 people who were administered cannabis under observation. The research also examined ten urine specimens and ten serum specimens from cannabis users whose material had been submitted to a forensic toxicology laboratory. The paper does not establish that the twenty specimens represent twenty distinct people, paired collections, or a randomly sampled population. Counts of chromatographic peaks and repeated chemical measurements are not additional human subjects.

The third stream was bench chemistry: the team synthesized and characterized a specific 11-OH-THC glucuronide. Its identity was assessed using chemical characterization and its chromatographic and fragmentation behavior compared with the biological signals. That synthetic material is a measurement resource. It is not itself a biological exposure experiment, and a high synthesis yield is not a test of detection accuracy in a population of drivers.

Why alcoholic and phenolic glucuronides matter

Phase I oxidation of Δ9-THC produces 11-OH-THC and ultimately additional oxidized metabolites, including THC-COOH. Phase II glucuronidation adds a polar conjugate that can alter transport, excretion, extraction, chromatographic behavior, and which forms are detected by a particular analytical method. In this article, alcoholic refers to glucuronidation of the carbon-11 alcohol; phenolic refers to attachment at a phenolic hydroxyl group. It does not refer to beverage alcohol, ethanol exposure, or a breath-alcohol measurement.

That chemical distinction can matter whenever a laboratory targets “11-OH-THC glucuronide” as though it were a single unambiguous molecule. A method that does not separate or structurally confirm the isomers may produce a report with narrower interpretive support than the label suggests. The original authors could directly match the alcoholic form to their synthesized reference. Their assignments of other peaks as phenolic forms relied principally on chromatographic and tandem-mass-spectral evidence and remained provisional pending corresponding standards.

What the human-liver S9 assay showed

The investigators incubated Δ9-THC with pooled human liver S9 fraction in the presence of cofactors under controlled laboratory conditions, then analyzed the products with high-performance liquid chromatography coupled to high-resolution mass spectrometry (HPLC–HRMS), using heated electrospray ionization in both positive and negative modes. The incubation was a laboratory reaction, not evidence of a particular individual’s metabolic rate. The procedure included a positive glucuronidation control using paracetamol and a negative control that did not yield the corresponding products.

For hydroxylated-THC glucuronide signals, positive-mode extracted-ion chromatograms showed five peaks with the expected composition, at approximately 2.5, 2.9, 3.4, 3.6, and 5.0 minutes. The chemically synthesized alcoholic standard matched the approximately 3.4-minute peak. Other peaks were labeled G1 through G4. Their related mass spectra supported the existence of additional products, but the study did not prove that each unknown peak was a distinct, fully assigned human 11-OH-THC glucuronide isomer. Hydroxylation at other positions and instrumental selectivity must remain in the analytical interpretation.

Evidence table 1: Five in-vitro glucuronide-region peaks

The observed times and assignments above reproduce the original experiment’s scientific distinctions. They are not a universal reference table for different instruments, columns, gradient conditions, or urine and serum preparation methods. The expected positive-ion precursor was reported near m/z 507.2589; negative-mode examination used a related deprotonated precursor near m/z 505.2438. The five peaks should not be converted into a claim of five individually quantified human metabolites.

How the synthesized standard strengthened identification

To obtain an authenticated alcoholic conjugate, the authors developed chemical-synthesis steps using a Δ8-THC-related route before applying the strategy to 11-OH-Δ9-THC. They protected the phenolic position, formed the glucuronide at the remaining alcohol, and then removed protecting groups. The final deprotection was reported at 95% isolated yield for that step. This is a chemical synthesis yield, not 95% analytical sensitivity, recovery from clinical specimens, or population-level diagnostic accuracy.

The authenticated product aligned with a product of the S9 assay in retention time and tandem-mass-spectrum behavior. The paper reported a weighted spectral dot-product score of approximately 0.918 for the relevant paired spectra. That number describes spectral resemblance using the authors’ comparison method. It cannot legitimately be recast as a 91.8% probability that a driver was impaired, a 91.8% true-positive rate for cannabis use, or a forensic method’s total accuracy. Direct reference-standard matching and repeated fragment evidence are valuable, but those claims must remain about chemical identity.

What the ten urine specimens actually established

Researchers used a relatively simple dilution-and-injection preparation for urine: 50 microliters of urine were combined with ammonium-acetate/formic-acid buffer and acidified acetonitrile, then centrifuged before a small aliquot was injected for full-scan analysis. All ten selected urine specimens were reported positive for THC-COOH. Under this specific procedure, only two of the ten demonstrated the hydroxylated-THC glucuronide signals the investigators were studying.

Those positive urines showed the two dominant signals rather than the full five-peak pattern visible in the in-vitro S9 experiment. The alcoholic product matched the synthesized standard. The other prominent signal, G2, was consistent with a phenolic form but lacked a matching phenolic standard. The study found no evidence of a double-glucuronidated OH-THC product under its tested conditions. “Not observed” is narrower than proving that no such molecular species could ever exist or be detected with a different method.

One of the two positive urine specimens came from a daily medicinal-cannabis user. The other originated from a person undergoing evaluation of fitness to drive. Those descriptions are specimen context, not a controlled exposure-to-impairment experiment. No direct behavioral driving scores, individualized use-to-collection intervals, or validated urine-glucuronide concentrations were supplied that would permit a case-specific impairment or timing algorithm.

What the ten serum specimens showed—and what they did not

The serum workflow was deliberately different. For serum, the laboratory extracted a 500-microliter portion by solid-phase extraction and used a more sensitive parallel-reaction-monitoring approach rather than the simpler full-scan urine procedure. In eight of ten authentic serum specimens, the team identified the alcoholic 11-OH-THC glucuronide against the newly synthesized standard. Table 2 of the original article reported its individual serum concentrations between 0.56 and 5.5 ng/mL among detected specimens. Two were not detected.

The original serum table also reported parent Δ9-THC values of 5.2 to 21 ng/mL, unconjugated 11-OH-THC from 2.2 to 7.1 ng/mL, and THC-COOH from 29 to 266 ng/mL across these ten selected specimens. These are measured serum quantities, not legal whole-blood concentrations. The article gave a lower limit of detection of 0.05 ng/mL for the alcoholic 11-OH-THC glucuronide and 0.10 ng/mL for THC-glucuronide under the reported serum analysis. A detection limit is not necessarily a fully validated quantification limit.

There is an important contrasting null result: THC-glucuronide itself was not detected in any of the ten serum specimens, even though parent THC and several other cannabinoids were measured. This does not mean that these people had not consumed cannabis, nor does it establish that THC-glucuronide must always be absent from all serum. The observations are bounded by the ten selected specimens, the target, the method, and its analytical limits.

Evidence table 2: Authentic specimen findings, by physical matrix

The two denominators describe specimen sets, not independent random clinical cohorts. The authors did not perform a controlled head-to-head sensitivity trial using paired serum and urine under harmonized extraction and acquisition conditions. Accordingly, comparing 8/10 serum with 2/10 urine as though it proves one specimen type is four times “better” would be methodologically unsound.

Does a larger G2 peak prove that the phenolic conjugate predominates?

No quantitative predominance can be established from peak height alone here. The authors noted that in positive authentic specimens the presumed phenolic G2 signal could be larger than the alcohol-conjugate signal. They also observed different ionization efficiency between the two candidates, including a reversal in relative prominence when changing electrospray polarity. The study did not possess a matched authentic phenolic standard for a response-factor calibration or definitive structural confirmation.

In one illustrated serum, the alcoholic peak was almost as high as the G2 peak; in another it was about half as high. Those are qualitative chromatographic comparisons. Converting either into a percentage of total glucuronidation, a molar ratio in a person’s circulation, or a population metabolic preference would require standards and validation beyond this study. Even the exact identity of G2 remained an evidence-supported hypothesis rather than the same level of confirmation achieved for the alcoholic standard.

Why a related Δ8 signal did not explain G3

The authors examined whether G3, which eluted close to the chemically confirmed alcoholic Δ9 conjugate, might merely represent the corresponding Δ8 double-bond isomer. They had synthesized a Δ8 reference for comparison. Under their conditions the reference Δ8 glucuronide appeared near 3.5 minutes, while G3 appeared around 3.6 minutes; the compounds also differed in ammonium-adduct behavior. The evidence led the authors away from assigning G3 to that particular Δ8 conjugate.

This was a focused chromatographic selectivity experiment, not a universal declaration that every closely eluting THC-related peak can be resolved in every case. Closely spaced peaks require instrument-specific chromatographic separation, retention-time controls, characteristic ion ratios or spectra, matrix assessment, and appropriate authentic standards. The paper’s chemical insight makes identification more precise while also revealing why labels based only on nominal mass may be premature.

Serum, urine, whole blood, oral fluid and breath are not interchangeable

A major evidentiary risk is transferring a conclusion between biological compartments. Serum is the liquid obtained after blood has clotted and cellular material has been removed. The paper’s serum concentrations cannot be treated as if they were validated equivalents of cannabinoids in directly collected forensic whole blood. Urine contains excreted drug products and is affected by collection interval, dilution, renal handling, and the test’s analytical preparation. The study did not assay oral fluid, hair, vitreous humor, or exhaled breath for this newly synthesized reference target.

For context on the consequences of specimen selection, see the separately published biological-matrices evidence guide and the specimen-collection guide. Those pages provide interpretive frameworks, not a substitute for the study’s actual source data. The source’s two serum-versus-urine observations also came from different preparation and mass-spectrometry acquisition modes, which further prevents unvalidated cross-matrix comparison.

What a forensic analyst should check before making a claim

A laboratory result labeled 11-OH-THC glucuronide should prompt questions about its actual chemical target. Was a specific alcoholic or phenolic regioisomer intended? Was the reference standard a true matched structure? Were chromatography and ion fragmentation sufficient to separate possible isomers and matrix interference? Was the signal measured directly as a conjugate, or inferred following a hydrolysis step? What were the identification criteria, detection limit, quantification range, recovery, matrix effects, and quality-control outcomes for that exact matrix?

Those questions are distinct from whether the person had ever used cannabis, recently used it, or was impaired at a material time. The original study’s strong point was structural corroboration for one conjugate. It was not a validation of a forensic decision threshold. The broader distinction between specific identification and case interpretation is explained in the already published confirmatory drug-testing guide.

In litigation, a defensible description would identify the specimen as serum or urine, name the specific detected compound or qualified putative isomer, identify the analytical method and limits, and separate direct measurement from any inference concerning exposure, timing, effect, and causation. An evidence review should request the original instrument data and validation information rather than assume that a research chromatographic signal supports a new legal threshold.

Statistical and causal limitations of the original study

The ten-urine and ten-serum sets were selected analytical specimens, not a prospectively timed cannabis-dosing cohort. The paper did not report validated person-level sensitivity and specificity for the glucuronide to diagnose recent use, did not randomize participants to dose levels, and did not measure real or simulated driving performance tied to their specimens. It does not provide a regression equation relating 11-OH-THC glucuronide concentration to elapsed time since smoking, ingestion, functional impairment, tolerance, or the cause of an accident.

There was also no experimental head-to-head comparison establishing a causal explanation for the different urine and serum detection counts. Extraction methods, ionization polarity and acquisition modes differ, and biological excretion patterns may contribute. The authors’ suggestion that OH-THC conjugates in urine appear at higher metabolite concentrations was a limited interpretation of their small observed set, not a validated cutoff or universal pharmacokinetic rule.

Finally, a reference standard for the phenolic conjugate was absent. Structural confirmation of the alcoholic isomer represents progress in analytical specificity, but does not automatically establish the identities or relative quantities of all other peaks, or make the method applicable to other specimens without independent validation. A negative finding is bounded by the tested procedure, and an apparent positive must be interpreted with the appropriate standard and matrix controls.

Seven focused forensic FAQs

1. Does detection of alcoholic 11-OH-THC glucuronide show that the person was impaired?

No. The paper chemically verified a metabolite signal in selected clinical/forensic specimens. It did not establish an individual blood-concentration-to-performance relationship, assess driving, or create an impairment classifier. Metabolite identification and functional impairment are separate propositions.

2. Is 11-OH-THC glucuronide the same as the psychoactive 11-OH-THC metabolite?

No. The glucuronide is a conjugated chemical derivative of 11-OH-THC. The parent metabolite and its glucuronide were measured or identified as separate analytical targets. The study did not directly quantify psychoactive effects attributable to the conjugate, so one cannot import effects of unconjugated 11-OH-THC into an interpretation of the glucuronide signal.

3. Why were eight serum specimens positive but only two urine specimens showed the conjugate signals?

The original used different physical matrices and different preparation/acquisition methods: serum solid-phase extraction with targeted parallel reaction monitoring, versus diluted urine with full-scan analysis. The counts describe these selected samples, not equal-sensitivity head-to-head testing. The study does not validate serum as universally four times more sensitive.

4. Can a negative urine OH-THC-glucuronide result rule out cannabis use?

No. All ten tested urine samples were THC-COOH positive, although the glucuronide-region signals were detected in only two. A negative finding for one metabolite under one analytical method is not a blanket declaration that other cannabis markers or exposure are absent.

5. Did the researchers conclusively identify the phenolic 11-OH-THC glucuronide?

They proposed a phenolic identity, especially for G2, based on chromatography and fragmentation patterns. They did not possess a synthesized matching phenolic reference standard for the same direct confirmation they achieved for the alcoholic form. Therefore, the relative identification certainty differs.

6. Can the 0.918 mass-spectrum match be quoted as 91.8% clinical accuracy?

No. The value is a weighted spectral-similarity statistic for comparing the synthesized standard and an experimental product. It is not a sensitivity, specificity, false-positive probability, likelihood of recent use, or measure of impairment.

7. Can the reported serum ng/mL values be converted into legal whole-blood THC concentrations or time since use?

Not from this experiment. It assayed selected sera for several distinct parent and conjugated cannabinoids. It did not supply a validated matrix-conversion factor, serial timed dosing data, or an individual impairment-time model. Any case-specific extrapolation would require independent scientifically supported data.

Bottom line: better chemical identification, not a cannabis impairment clock

The 2020 study supplied a chemically synthesized authentic reference standard for alcoholic 11-OH-Δ9-THC glucuronide and demonstrated its presence in the studied human serum and urine materials. Five related in-vitro peaks indicated that mass alone cannot resolve all structural possibilities. A proposed phenolic product remained incompletely confirmed, double glucuronidation was not observed, and selected serum and urine detections differed under distinct methods.

The useful forensic takeaway is a disciplined one: report what was directly detected, in the actual physical specimen, using the validated analytical procedure and the correct degree of structural confidence. None of the measured peaks or the newly synthesized standard establishes a person’s legal whole-blood THC level, time of cannabis exposure, subjective experience, driving impairment, or a specific event’s cause. Primary source: Hassenberg C, Clausen F, Hoffmann G, Studer A, Schürenkamp J. International Journal of Legal Medicine (2020), DOI 10.1007/s00414-020-02387-w.