Can a Positive Urine THC Test After a Negative Mean New Cannabis Use? Evidence From 60 Monitored Users

Not necessarily. In a primary investigation of 60 cannabis users residing on a closely supervised research unit, urine cannabinoid tests sometimes became negative and then positive again without access to cannabis. The pattern was particularly persistent among participants whose first urine sample had a high creatinine-normalized concentration of 11-nor-9-carboxy-THC (THC-COOH). A urine result above a screening cutoff establishes an analytical finding in that specimen; it does not, without further evidence, establish a new cannabis episode, a precise ingestion time, an individual blood THC concentration, or impairment at an alleged event.

Goodwin and colleagues collected 6,158 individual urine specimens during continuously monitored abstinence for up to 30 days. Their original study, Urinary Elimination of 11-Nor-9-carboxy-Δ9-tetrahydrocannabinol in Cannabis Users During Continuously Monitored Abstinence, was published in the Journal of Analytical Toxicology in 2008. The primary publication (DOI: 10.1093/jat/32.8.562) supplies the numerical results discussed below. This is evidence about residual urinary metabolite excretion, not a controlled study of driving or acute intoxication.

The answer first: one negative urine screen is not a drug-use clock

The investigators organized participants by the concentration of THC-COOH divided by urinary creatinine in their first collected specimen. The three starting groups were 0–50, 51–150, and above 150 nanograms of THC-COOH per milligram of creatinine (ng/mg). The group’s starting concentration was strongly associated with how long positive screening results continued. In the highest group, the mean last-positive time in the observed records was approximately 15.4 days after admission, versus 4.6 days in the lowest group. Those are descriptive observations under this particular surveillance, not a validated formula to date a different person’s cannabis use.

The critical feature is intermittence. After a person’s first negative specimen, a later void could still exceed the 50 ng/mL cannabinoid immunoassay threshold. A reviewer should therefore not convert “negative on Tuesday, positive on Thursday” directly into “used cannabis on Wednesday.” That inference requires additional specimen history, confirmatory analysis, timing, dilution assessment, and an appropriately validated new-use interpretation procedure.

Who participated, and why does the monitored setting matter?

Sixty adults aged 18–45 who reported daily-to-weekly cannabis consumption participated. The group contained 46 men and 14 women; mean age was 26.9 years, with a reported standard deviation of 6.4. Participants reported averaging 9.4 joints per week and using cannabis on 10.6 of the previous 14 days, with substantial between-person variation. Self-report characterized earlier consumption; it was not a measured individual THC dose or a verified clock time of the final preadmission use.

Participants lived on a secure National Institute on Drug Abuse clinical unit under continuous medical supervision, with no access to unauthorized drugs. Researchers collected individual urine voids as they occurred, rather than taking only one scheduled specimen daily. All participants had screening evidence of prior cannabinoid exposure and normal eligibility kidney assessments. The design helps distinguish residual drug excretion from an opportunity for new use during observation.

But the study had no randomly assigned recent-cannabis dose, no unexposed control cohort designed to estimate specificity for new use, and different lengths of residence. An individual observed for four days cannot supply the same evidence about a 30-day excretion tail as someone who stayed for the full interval. A “last positive” is the last one observed during that person’s collection period, not necessarily the biologically final positive they could have produced.

What exactly was collected and measured?

The analytical specimen was urine. Each void was collected separately, promptly refrigerated, aliquoted, and frozen at −20°C on the collection day. After a single thaw for analysis, the screening result was classified as positive at 50 ng/mL or higher using cannabinoid immunoassay. The investigators separately described base hydrolysis followed by extraction and gas chromatography–mass spectrometry (GC–MS) quantification of urinary THC-COOH, with a stated quantification limit of 2.5 ng/mL. Creatinine was measured by a modified Jaffe procedure.

Those numbers have different functions. The 50 ng/mL value is an immunoassay screening classification threshold for the underlying serial positive/negative time course. The 2.5 ng/mL figure is an analytical quantification limit for the separate GC–MS procedure. The ng/mg values used to stratify participants are creatinine-normalized metabolite concentrations, not ng/mL and not a screening cutoff. The study’s group comparisons must not be rewritten as if all three measurements were the same test.

The study did not collect forensic venous whole blood to determine parent Δ9-THC at a driving event, nor did it determine an individual’s plasma or serum THC pharmacokinetic curve. A urinary metabolite persists after the acute parent-drug concentration and effects have changed. Neither a positive screening reaction nor a quantitative urine THC-COOH concentration is a direct chemical assay of contemporaneous brain effect.

The three baseline groups and the observed detection windows

The table below restates original Table 2 group summaries while preserving specimen, time, and concentration units. Each group comprises independent participants; the thousands of repeated urine voids do not become thousands of independent people. Values are reported means, not guarantees for any particular person.

There is a consequential limitation even within this table: a first-negative time cannot be observed in a participant who never produces a negative screen during residence, and a last-positive time is undefined for someone without a positive specimen. These group averages arise from the available defined measurements, with different observation periods and censored outcomes; they should not be used as a personalized time-to-clearance calculator. In the lower group some participants had no observed positive specimen after admission, while several participants in the upper groups had no observed negative specimen.

What did the statistically tested associations actually show?

Higher starting normalized THC-COOH group was statistically associated with the day of the first negative urine specimen (p = .0021), the day of the last positive specimen (p = .0003), and mean concentrations during the period between first negative and last positive (p = .0202). The baseline group was also associated with the last day showing at least 50% positive specimens (p = .0001) and the interval in days between first negative and last positive (p = .0002).

Average self-reported joints per day showed an association with a log-transformed maximum concentration in the intermittent interval (p = .0383). Body mass index showed an association with day of last observed positive urine specimen (p = .0489). Those are selected statistical associations from this sample, not demonstrations that BMI quantitatively measures stored THC or that a specific body’s adipose tissue was chemically sampled.

The researchers discussed lipophilic THC distribution and slow release as a plausible explanation for extended urinary metabolite elimination. But the study did not directly sample adipose tissue or demonstrate a person-specific depot-release rate, and a regression association with BMI is not a validated correction factor for an individual case.

A negative result could be followed by positive screens during verified abstinence

In the lowest normalized baseline group, the mean detection rate fell to zero by approximately day eight after the first negative test. The middle group declined more slowly and irregularly. The highest group had daily mean detection rates between roughly 60% and 100% at observed time points as late as day 28 following its first negative result. Figure 2 shows the fluctuating proportion of that day’s available urine voids exceeding the 50 ng/mL screen threshold, not the fraction of all original participants known to be positive on every subsequent day.

This distinction matters particularly when only a few participants remain under observation late in the study. The denominators for daily detection rates can change with discharge and available specimens. Such a graph cannot establish that all 20 people in the highest group remained positive for 28 days, much less that every frequent cannabis user will do so. It does show that recurrent positive results without observed renewed use were possible and that the initial normalized concentration helped distinguish shorter from longer tails.

Peak urinary metabolite need not occur in the first collected specimen

Sixty percent of participants had their maximum creatinine-adjusted THC-COOH concentration in the first sample taken during residence. The remaining 40% had peaks as late as 2.9 days after admission. That late peak arose within the monitored-abstinence setting; it cannot automatically be interpreted as new cannabis exposure.

Urine concentration is a combined product of prior exposure, changing excretion, elapsed collection interval, hydration, and the laboratory’s measurement procedure. Even creatinine adjustment reduces only part of the dilution problem. It does not transform a delayed peak into direct evidence that cannabis was consumed after admission.

Figure 3 depicts a relatively rapid reduction in mean normalized values over the first three days, followed by more gradual changes. Those trajectories are group averages. Individual results can oscillate around a decision threshold, producing a series of alternating positive and negative screening classifications despite an underlying overall decline.

What can be calculated, and what cannot be concluded?

The original experiment contained 6,158 separate urine collections from 60 people. A large specimen count yields fine-grained within-person histories, but not a large independent participant cohort. For a specific case, comparing a measured ng/mL screen with a later ng/mg creatinine-corrected GC–MS value without conversion would be a unit error. Likewise, applying a 50 ng/mL threshold to a curve expressed in ng/mg would change the question being asked.

A reliable case interpretation needs the raw chronological series, method identity and confirmed results, specimen creatinine, exact collection timestamps, and the stated program’s decision threshold. If a laboratory used a different immunoassay, cutoff, or confirmatory protocol, the study’s numerical screening detection windows should not simply be transferred. The study provides population evidence about residual elimination, not a single universal excretion half-life or an independent method for dating last use.

Two different statistical questions should not be confused

The next table separates what the investigators actually evaluated from what cannot be extracted from those tests. A significant association may be relevant to residual-excretion interpretation without validating a person’s new-use date. An unmeasured endpoint is not the same as a tested nonsignificant endpoint.

The work did not report standardized accuracy, sensitivity, specificity, receiver-operating-characteristic analysis, or cross-validated individual predictions of renewed use using these group means. Those missing validation measures should not be invented from the p values or from the large count of collected voids.

Screening immunoassay versus GC–MS confirmation

A cannabinoid urine immunoassay is a screening method that classifies a sample relative to an assay decision threshold. The chemical class recognized by the antibody and the specific calibrator matter. GC–MS confirmation after hydrolysis, by contrast, can target and quantify THC-COOH as an identified metabolite within the instrument method’s validated range. The study used these different techniques for complementary purposes; one should not describe the immunoassay’s serial positive/negative classifications as if every classification were an independently confirmed same-analyte quantitative result.

The confirmatory-drug-testing overview explains why specimen identity, chemical specificity, calibration and reporting purpose matter. Negative screening results are also threshold-dependent: a sample just below 50 ng/mL is not proof of the total absence of all cannabinoids or THC metabolites. An observed cutoff crossing may reflect an ordinary biological or analytical change rather than a discrete new exposure.

Why creatinine adjustment helps but does not solve exposure dating

Creatinine correction expresses THC-COOH mass relative to creatinine mass, commonly ng/mg. It can reduce some changes due to variable urine dilution, permitting more informative comparisons of serial specimens than raw ng/mL alone. It does not make excretion constant, and it should not be applied mechanically if renal function, specimen validity, or measurement quality differs materially from the controlled study.

Several features limit extrapolation: individual baseline concentration, chronic exposure history, available follow-up length, variation in void intervals, assay-specific measurement error, and uncertain exact last-use history before entering the unit. Although the residential design protected against new use during the stay, it did not experimentally randomize cannabis dose or independently validate a complete case-by-case prediction equation. The investigators cited earlier ratio methods developed elsewhere, but the current three-group study did not itself validate those prior numerical ratio cutoffs as universally correct.

What the results mean in probation, employment, and licensure matters

When a legal or administrative issue turns on alleged new cannabis exposure, an isolated urine cannabinoid positive after a previous negative is weak evidence of a precisely timed new use, particularly where a history of regular use and changing urinary concentration are plausible. A scientifically sound assessment identifies the actual assay and cutoff, whether the latter positive was confirmed, the relevant creatinine values, the time between voids, and whether there is a suitable validated serial-specimen interpretation procedure.

Conversely, the study does not prove that a later positive is always residual. New use is possible outside the monitored setting. The defensible conclusion is that the two explanations cannot necessarily be distinguished from a simple negative-then-positive sequence alone. Subject-specific facts and laboratory controls must determine which inferences remain supportable.

The specimen-handling and matrix distinction is explained in the verified specimen-collection guide and biological-matrices guide. Urine THC-COOH is not a substitute for measured parent THC in venous whole blood, and neither alone proves performance on a driving task, awareness, or causation of a collision.

Limits of the original evidence

All 60 participants had previous cannabis exposure, so this study did not estimate a general-population rate of cannabinoid-positive screens or compare users with a clean unexposed population. Their prior doses and exact last-use times were not independently administered or documented. The laboratory used a particular 50 ng/mL urine screening classification and a separately described GC–MS urine metabolite assay, not every modern screening platform or confirmatory testing scheme.

Follow-up lasted up to 30 days, but not every person completed 30 days. Some participants had no observed first-negative specimen; others had no observed later positive. At the 30-day boundary, continuing positivity in some participants is evidence of observation ending, not evidence that those participants necessarily became negative the next day. Group trend lines may include progressively fewer participants late in follow-up. There was no direct pharmacodynamic impairment experiment, no measured driving outcomes, and no individualized reliable timing model for exact last ingestion.

The central evidence remains valuable precisely within those boundaries: it demonstrates how residual urinary cannabinoid excretion can create fluctuating screen classifications despite controlled abstinence, and how baseline creatinine-adjusted concentration is associated with observed detection duration.

Frequently asked questions

Can a urine THC test turn positive again without someone smoking cannabis?

Yes. This study observed negative-then-positive urine cannabinoid screening patterns while participants resided under monitored abstinence. The result applies to urinary screening classification and residual metabolite elimination; it is not proof that every later positive in an unsupervised setting reflects old use.

How many people and specimens were actually studied?

There were 60 independent cannabis-using participants and 6,158 individual urine specimens. Repeated specimens are not 6,158 separate participants. The three baseline creatinine-normalized groups contained 19, 21, and 20 people.

Does a 50 ng/mL urine screen equal 50 ng/mg THC-COOH/creatinine?

No. Ng/mL is a concentration in urine used for the screening threshold. Ng/mg is the measured THC-COOH amount normalized to urinary creatinine. The thresholds, assay targets, denominators, and questions are different.

Does a confirmed urinary THC-COOH result establish current impairment?

No. THC-COOH is an inactive metabolite in urine. The research did not directly measure driving performance or contemporaneous parent THC concentration in forensic whole blood. A chemical exposure finding must not be equated with an individual’s impairment or a particular event’s cause.

Were all high-baseline participants positive for 28 days?

No. The reported 60%–100% late daily detection rates describe proportions of available urine specimens at observed follow-up times within the highest starting group. They do not mean all 20 participants produced positive specimens every day or that all were observed for an equal 28-day interval.

Does the first negative specimen identify the date of final THC elimination?

No. A specimen below a particular screen cutoff may be followed by another above it, depending on residual elimination, specimen dilution, and analytical variation. “First negative” is an operational event under a specified assay, not the moment all cannabinoids disappear from the body.

Can the study’s group averages prove when a specific person last used cannabis?

No. The study supplies population and repeated-specimen observations, not a validated individual date-of-use estimator. The full case history, confirmed assay results, creatinine, collection times, threshold and relevant uncertainty must be considered before drawing even a qualified conclusion about renewed exposure.

Primary source and practical conclusion

Goodwin RS, Darwin WD, Chiang CN, Shih M, Li SH, Huestis MA. Urinary Elimination of 11-Nor-9-carboxy-Δ9-tetrahydrocannabinol in Cannabis Users During Continuously Monitored Abstinence. Journal of Analytical Toxicology. 2008;32(8):562–569. Original primary study: DOI 10.1093/jat/32.8.562.

Bottom line: In monitored abstinence, residual urinary THC-COOH excretion generated both extended positive screens and negative-to-positive reversals. Baseline creatinine-adjusted THC-COOH was associated with detection persistence. The findings justify scrutiny of an alleged new-use inference based only on consecutive urine screening classifications; they do not establish that an individual used cannabis at a particular later time, had a particular whole-blood THC concentration, was impaired, or caused an event.