Can Exercise or Fasting Make Residual Cannabis Look Like New Use? A Six-Person Serum and Urine Study

Answer: In a small monitored study of six daily cannabis users who had stopped using cannabis, 45 minutes of moderate treadmill exercise and 24 hours without food did not produce a consistent surge in creatinine-normalized urinary THC-COOH. Urine metabolite-to-creatinine ratios fell in five of six participants around exercise and in all six around fasting. Some measured serum cannabinoid values rose transiently, including a substantial change in one participant, but these findings were not uniform and did not establish renewed cannabis ingestion. The experiment cannot rule out every circumstance in which stored cannabinoids redistribute; it does challenge the simplistic assertion that ordinary exercise or skipping food reliably explains an upward urine drug-test pattern.

The underlying original human experiment is Westin, Mjønes, Burchardt, Fuskevåg and Slørdal, “Can Physical Exercise or Food Deprivation Cause Release of Fat-Stored Cannabinoids?” (2014). Its most useful forensic contribution is the side-by-side measurement of serum THC, serum THC-COOH and urinary THC-COOH normalized to creatinine over a seven-day monitored interval. Those measurements answer related but different questions; none directly measures intoxication or observed driving behavior.

What question does this experiment actually answer?

After chronic cannabis use, cannabinoid-related compounds may be detectable after the last reported dose. A contested urine finding might therefore prompt two very different explanations: continued elimination of a prior exposure or a new episode of use. A third claim sometimes arises when a person reports exercise, fasting, dieting or sweating between collections: perhaps activity mobilized drug stored in body fat and manufactured a new apparent positive. The Norwegian investigators addressed only part of that proposition. They studied whether two defined activities coincided with increases in laboratory-measured cannabinoid concentrations in monitored people who had recently stopped chronic use.

The important distinction is between biochemical plausibility and observed magnitude. THC is lipophilic, but the fact that a substance partitions into tissue does not establish that a treadmill workout produces a large, predictable rise in a subsequently collected urine specimen. The experiment tested this narrower prediction directly. It did not identify the anatomical source of circulating molecules with a tracer study, quantify fat release, calculate a legal probability of new use or observe a workplace testing program’s actual adjudications.

Who participated, and what were the exposure conditions?

Seven consecutively admitted patients were invited; six completed the protocol. Participants were five men and one woman, aged 25–35 years, with a reported mean age of approximately 30. They had smoked cannabis daily until admission and described about 5–30 grams of hashish or marijuana-equivalent material per week during their prior use. These self-reported grams describe plant or resin material, not a verified amount of absorbed Δ9-THC. Four of six also acknowledged some other illicit-drug use in the preceding month.

The study occurred in a residential detoxification ward. Belongings were searched at entry, rooms were separate, and staff surveillance was intended to support abstinence. These precautions improve the interpretability of declining or fluctuating residual excretion but do not constitute continuous chemical confirmation that every exposure possibility was eliminated. The researchers also acknowledge that participants themselves collected urine. There was no random population sampling, no obesity-enriched subgroup and no placebo or non-exercising concurrent control cohort.

How were the two interventions conducted?

Researchers collected urine during seven days after admission. The exercise condition required 45 minutes of treadmill running or jogging at an intended moderate intensity, 60–75% of an age-estimated maximum pulse. The food-deprivation condition lasted 24 hours, during which water was permitted but food was withheld. Conditions were conducted around days 3 and 5 in randomized order within the same six people. That within-person structure avoids treating six exercisers and six fasters as 12 independent volunteers.

Blood was drawn immediately before and after the interventions and on study days 1 and 6. Urine was collected at each natural void, rather than at exactly identical clock times for every person. Because the intervention order and elimination history varied, a before/after contrast also sits on top of several days of abstinence-associated change. The intervention was not a controlled new cannabis dosing experiment, and neither condition established how much THC remained in individual adipose stores.

What did the investigators measure in serum rather than whole blood?

Collected blood was centrifuged and the serum fraction was stored frozen at approximately −20°C. Serum was analyzed using liquid chromatography–tandem mass spectrometry (LC–MS/MS) for parent Δ9-THC and 11-nor-9-carboxy-THC, abbreviated THC-COOH or THCCOOH. The reported units were ng/mL of serum. Parent THC and THC-COOH are distinct analytes, and a serum concentration is not a direct headspace-GC or LC–MS/MS measurement of venous whole-blood THC.

The assay’s reported detection limits were 0.2 ng/mL for THC and 0.17 ng/mL for THC-COOH; the article identifies 0.6 ng/mL as the THC quantification threshold for the displayed exercise and fasting data. Values marked below detection or below quantification should not be silently replaced with zero or added as fully quantified observations. At study entry, parent THC was detectable in five of six participants, while serum THC-COOH and urinary THC-COOH were present in every participant. At admission, the investigators reported a parent-THC serum range from below detection to 3.6 ng/mL; this admission range should not be mistaken for the maximum of every subsequent measurement.

For the foundational reason that results must remain tied to the tested matrix, see the site’s published biological matrices guide. It is especially important here because an expert’s statement about “blood THC after exercise” can conceal that this experiment separated and assayed serum instead of legal-case whole blood.

How was the urinary metabolite result standardized?

Every collected urine specimen was tested for THC-COOH, not directly for the participant’s current psychoactive THC concentration. Before LC–MS/MS analysis, the team used β-glucuronidase treatment to address the conjugated THC-COOH fraction. Creatinine was determined separately by a photometric laboratory method. For each positive urine specimen, the THC-COOH concentration in ng/mL was divided by creatinine in mg/mL, producing the THC-COOH/creatinine ratio in ng/mg.

That normalization matters because an unadjusted urine concentration can increase or decrease simply when urine becomes more or less dilute. The ratio helps account for concentration differences between voids, but it is not a magical clock: urinary creatinine itself varies, a ratio may fluctuate during abstinence, and no single universal ratio proves exactly when cannabis was used. The investigators collected 188 urine specimens and 34 serum specimens overall. Those are specimen counts; they are not 222 independent human subjects or 222 independent tests of the intervention effect.

What happened around the 45-minute exercise intervention?

Table 2 of the original article reported the displayed mean serum THC concentration rising from 1.6 to 2.0 ng/mL, a difference of 0.4 ng/mL or 25%. Mean serum THC-COOH increased from 39.4 to 42.3 ng/mL, a difference of 2.9 ng/mL or about 7%. These descriptive averages must be interpreted against missing/censored data: one participant had no post-exercise blood draw because of difficult venous access, so the rows do not represent six complete paired measurements for every serum analyte.

The corresponding displayed average urinary THC-COOH/creatinine ratio decreased from 80.9 to 62.7 ng/mg, a reported change of −18.2 ng/mg (−23%). Five participants’ urine ratios fell; participant A’s ratio increased from 157.2 to 190.1 ng/mg (+21%). One participant’s “before exercise” urine came from the preceding day, not immediately before the treadmill session. Participant B’s serum THC-COOH increased from 23.6 to 44.0 ng/mL, an 86% increase; it would be misleading to conceal that outlying observation beneath the group mean.

The proper summary is therefore not “exercise had absolutely zero biochemical effect.” Rather, this particular moderate workout produced small average serum changes, some heterogeneous individual shifts, and no uniform exercise-triggered rise in creatinine-normalized urine metabolite results. No actual driving, intoxication or occupational error measure was collected alongside the laboratory specimens.

What happened during the 24-hour food-deprivation period?

For fasting, Table 3 reported the mean quantified serum THC values at 2.6 ng/mL before and 1.7 ng/mL after, but several individuals’ values were below detection or quantification. One participant’s displayed 0.6 ng/mL result was collected on the following day because an immediately post-fasting result was below quantification. Thus the displayed row is not a clean synchronized six-person THC mean at two exact time points.

By comparison, reported mean serum THC-COOH increased from 27.5 to 33.9 ng/mL, a change of +6.4 ng/mL or +23%. The urine ratio moved in the opposite direction: its displayed mean fell from 89.5 to 37.4 ng/mg, a decline of 52.1 ng/mg or 58%. All six participants’ creatinine-normalized urinary THC-COOH ratios decreased between the source’s stated before-and-after fasting observations. Participant F showed an 85% rise in serum THC-COOH but a 78% decline in the corresponding urinary ratio. That contrast exemplifies why serum and urine should not be described as if they measure the same biological compartment.

These data apply to one water-allowed 24-hour fast in a small group with relatively low body mass indices. They do not supply experimental observations for prolonged fasting, dehydration, rapid weight loss, obesity, endurance competitions, high-heat exercise or uncontrolled cannabis exposure.

What are the exact outcomes—and what was not statistically established?

The authors concluded there were no major concentration changes that would generally frustrate drug-testing interpretation. The source’s principal intervention tables display individual measurements and descriptive changes; they do not supply a conventional inferential p-value or confidence interval supporting a universal “no effect” finding for each analyte. Accordingly, it would be incorrect to advertise this study as proving equivalence, to invent a significance test or to use a nominal 23% or 58% decrease as a universal personal elimination constant.

Why can serum THC rise while the urine ratio falls?

The two assays capture different kinetic processes and time windows. Serum Δ9-THC reflects parent analyte present in the serum fraction at the draw. Serum THC-COOH reflects a carboxy metabolite at that draw. Urine THC-COOH/creatinine reflects metabolite output into the bladder between voids, downstream excretion and urine concentration, not an instantaneous serum equilibrium. Changes in distribution, metabolism, renal handling, collection interval and ongoing abstinence can alter these trajectories independently.

The study did not trace molecules leaving individual fat deposits and did not administer isotopically labeled THC to identify the origin of a transient serum rise. A serum increase is therefore not direct experimental proof that adipose release caused that change, and it is certainly not proof of a second inhalation or edible exposure. Conversely, a falling urinary ratio does not establish that serum THC must be falling at every second. The temporal behavior of a biological marker cannot be reduced to a single cross-matrix conversion factor.

The published pharmacokinetics guide provides a useful framework for discussing distribution and elimination without pretending that a population tendency is an individualized drug-use timeline.

Can a positive urine test be attributed to a workout alone?

Not from this experiment. All six monitored participants already had urinary THC-COOH at admission, and all stayed positive across the observation period. The study did not demonstrate a transition from truly metabolite-negative to a new metabolite-positive state caused by exercise. It also did not administer a confirmed new cannabis dose after abstinence and calculate sensitivity or specificity for detecting that dose. Calling this a forensic “false-positive test” experiment would misidentify its research design.

A urine sample can be analytically positive because a metabolite is present while the underlying exposure occurred earlier. That is not necessarily an analytical false positive. Whether a testing policy characterizes an increase between specimens as possible new use requires the laboratory’s actual quantitative method, concentration thresholds, sampling interval, creatinine values and decision procedure. The Westin data challenge the assumption that moderate exercise or one day of fasting reliably causes an important rise in the normalized urine metabolite; they cannot decide an unrelated person’s disputed pair of tests without the original case measurements.

How should an expert examine an alleged new-use pattern?

A case review should first secure the original laboratory reports rather than start with a theory of metabolic redistribution. Relevant records include specimen collection and voiding times, urine creatinine for each specimen, confirmatory assay identification and quantification limits, measurement uncertainty, internal controls, specimen integrity, storage history, documented cannabinoid exposures and the applicable rule for serial-test interpretation. A ratio should be recomputed from recorded numbers, not reconstructed from a screenshot of a positive/negative result.

One should also distinguish a metabolite screen from an appropriately validated confirmatory measurement. The site’s published confirmatory drug testing guide describes why a definitive identification, a cutoff-based screen and a clinical or legal conclusion are separate steps.

What limits the strength of this original experiment?

Six people represent a narrow, low-powered sample. Recruitment ended early because of staffing changes, and the participants had generally low BMI values, limiting inference to people with greater adiposity. There was no direct measure of adipose THC depletion, no independent non-intervention control arm, no controlled cannabinoid re-dose, no random cross-sectional population sample and no blinded determination of legal “new use.” An observational series over abstinence can characterize actual trajectories without proving the mechanism behind each fluctuation.

Specimen-completeness limits matter: participant C had difficult venous access and only three serum draws; participant E initially provided only one urine sample daily, and the urine specimen preceding exercise was from the prior day. One fasting THC result was obtained the next day after an immediately post-intervention value was below quantification. Treating these entries as a fully synchronized experiment would create precision that the original publication does not contain.

Finally, the tested workout was moderate, not exhaustive, and the fast lasted 24 hours, not weeks. The authors explicitly recommended further investigation in people with higher fat mass, other exercise modalities and longer deprivation. Those boundaries should remain visible even when the study’s general conclusion is useful in contesting a claim that sweating or skipping meals necessarily explains a quantitative urine THC-COOH change.

What is the bottom-line interpretation for a forensic dispute?

Westin and colleagues provide limited but direct experimental evidence that ordinary moderate activity and a single-day fast did not generate a predictable creatinine-corrected urinary THC-COOH surge in six recently abstinent daily cannabis users. They also document why small serum changes and transient individual spikes must be reported rather than erased. The laboratory record is nuanced: a positive urine metabolite finding may reflect prior exposure, but exercise is not thereby established as the cause of an observed increase, much less proof that a new use event did not occur.

The defensible endpoint is narrower than either advocacy extreme. This primary study cannot establish a universal “exercise never changes THC” rule, and it cannot justify an automatic “exercise explains every positive” defense. It offers measurements, laboratory methods, sample limitations and a reason to evaluate serial creatinine-normalized results instead of conflating serum parent THC with urinary metabolite excretion. Individual impairment or case-event causation must be supported by separate case-specific evidence, if it can be supported at all.

Frequently asked questions

1. Can exercise release THC stored in fat?

A redistribution mechanism is biologically plausible, and Westin et al. measured small transient serum changes in some participants. The experiment did not trace fat-derived molecules directly or show that exercise consistently generated a major urine THC-COOH rise. It cannot settle extreme exercise or every body composition.

2. Did any participant’s cannabinoid level rise after exercise?

Yes. The reported serum THC group mean rose from 1.6 to 2.0 ng/mL among available quantified observations, and one participant’s serum THC-COOH increased from 23.6 to 44.0 ng/mL. Those findings should be disclosed alongside the five-of-six decline in the urine metabolite/creatinine ratio.

3. Did fasting produce a positive urine test in someone previously negative?

No such conversion was demonstrated. Every study participant already had detectable urinary THC-COOH and remained positive during observation. The before/after fasting comparison showed a lower normalized urine ratio in each of the six participants, not a new negative-to-positive transition.

4. Is urine THC-COOH the same as blood THC?

No. Urinary THC-COOH is a metabolite measured after renal excretion and, in this study, corrected for creatinine concentration. Serum Δ9-THC is measured parent compound in a different specimen matrix. Neither result is interchangeable with a separately collected and analyzed venous whole-blood THC result.

5. How many people and specimens were analyzed?

Six adult participants completed both interventions. The study reported 188 urine specimens and 34 serum specimens, some with missing, delayed, below-detection or below-quantification entries. Repeated specimens from one person do not create extra independent participants.

6. Does a decreasing urine ratio prove no new cannabis use?

No. A declining ratio can be consistent with residual elimination, but new-use assessment depends on actual exposure information, timed paired measurements, laboratory methods and a defensible interpretation protocol. This study did not validate an individualized new-use classification algorithm or error rate.

7. Does the study establish whether a driver was impaired?

No. The six participants were not tested with an on-road driving task, standardized psychomotor impairment battery or collision-causation protocol. Laboratory analyte concentrations and ratios alone cannot establish what a particular driver could safely do at a specified event time.

Original primary research: Westin AA, Mjønes G, Burchardt O, Fuskevåg OM, Slørdal L. Can Physical Exercise or Food Deprivation Cause Release of Fat-Stored Cannabinoids? Basic & Clinical Pharmacology & Toxicology. 2014;115(5):467–471. doi:10.1111/bcpt.12235. This is a research-evidence explanation, not a finding about any identified litigant or an adoption of the conclusions as a personally authored expert opinion.