How Long Does Oral CBD Persist in Calf Plasma—and Can Seven Calves Establish a Cattle Withdrawal Time?

Answer: In a 2022 primary pharmacokinetic study of seven young male Holstein calves given a single oral dose of cannabidiol (CBD) oil at 5 mg/kg, the geometric mean peak concentration of CBD measured in separated blood plasma was 0.05 µg/mL (50 ng/mL). The geometric mean time to that peak was 7.5 hours, and the estimated terminal elimination half-life was 23.02 hours. The investigators sampled plasma through 48 hours, but their estimated total exposure required substantial extrapolation beyond the last observation. These findings demonstrate measurable and variable systemic exposure after oral CBD oil; they do not establish a food-animal tissue withdrawal interval, milk residue safety, a regulatory clearance threshold, or a prediction for an individual animal.

The primary experimental source is Meyer K, Hayman K, Baumgartner J, and Gorden PJ, “Plasma Pharmacokinetics of Cannabidiol Following Oral Administration of Cannabidiol Oil to Dairy Calves,” Frontiers in Veterinary Science 9:789495 (2022), doi:10.3389/fvets.2022.789495. This article examines what was directly measured, what was mathematically estimated, and what would still need to be tested before a food-safety or case-specific conclusion could follow.

What did the seven-calf study actually establish?

All seven enrolled calves received the same nominal weight-based oral CBD dose and completed the study without an observed adverse drug reaction or behavioral change reported by the authors. Serial plasma samples revealed that CBD reached systemic circulation after oral administration. Both the measured peaks and the derived pharmacokinetic parameters varied appreciably across calves. The study is descriptive: it did not randomize animals to a placebo group, compare multiple doses, or test changes in productivity, pain, neurological performance, or disease outcomes.

The immediate forensic lesson is narrower than a general claim that CBD remains in cattle for a fixed number of hours. A measured plasma concentration at a specified interval is one type of evidence. A statistical summary of a concentration-time curve is another. Neither automatically provides the concentration in liver, kidney, muscle, fat, or milk, nor does it specify when an animal or product becomes acceptable for food use.

Which animals were studied, and why does age matter?

The researchers studied seven male Holstein calves, averaging 19.43 days of age (standard deviation 5.29 days) and 48.61 kg body weight (standard deviation 4.72 kg). These were pre-ruminating animals, not adult lactating dairy cows. They were housed at a university dairy facility, fed milk replacer twice daily, and provided water and calf starter. They were monitored for demeanor, hydration, and appetite.

Seven animals means seven biological observations, regardless of how many tubes, time points, injections into the analytical instrument, or fitted parameters were collected. The same calves contributed repeated specimens over two days; those serial measurements are not independent calves. A small uniform male cohort may be useful for exploring concentration-time behavior, but it cannot represent the full distribution among adult cows, replacement heifers, different breeds, diseased cattle, or lactating animals.

The authors explicitly noted that forestomach development could change oral absorption. A calf that relies primarily on milk does not necessarily process a lipophilic oil formulation like a mature ruminant. That biological distinction prevents direct extrapolation of this pilot result to a herd-wide withdrawal policy.

What was administered: CBD oil, not THC, hemp feed, or milk?

Each calf received 5 mg of CBD per kilogram of body weight as an oil formulation labeled 25 mg/mL. The dose was delivered orally using a syringe while the calf suckled, followed by a brief period of continued suckling intended to reduce loss from the mouth. At the cohort’s mean weight, 5 mg/kg corresponds arithmetically to about 243 mg of nominal CBD and roughly 9.7 mL of 25 mg/mL oil; those are illustrative calculations from group-average weight, not individual recorded administered amounts.

Oral dosing is not inhalation, injection, a cannabinoid-contaminated feed trial, or topical exposure. The investigation did not administer a THC-rich cannabis product. Its pharmacokinetic analyte of interest was cannabidiol itself, rather than psychoactive delta-9-THC, 11-hydroxy-THC, THC-COOH, or a CBD metabolite. An exposure reported in mg/kg cannot be substituted for a measured plasma concentration in ng/mL: mass given and mass detected in a particular biological compartment answer different questions.

Which biological matrix was measured?

Investigators collected blood through jugular venous catheters into sodium-heparin tubes. The tubes were cooled and transported to the laboratory, where they were centrifuged. The separated plasma was harvested and frozen before analysis. The reported results are plasma CBD concentrations. They are not direct measurements in unseparated venous whole blood, serum, urine, oral fluid, exhaled breath, vitreous humor, milk, or edible tissues.

This distinction matters because a reported plasma concentration cannot be numerically converted to whole blood without appropriate data on the relevant matrix distribution. Nor may a plasma curve be transformed into a tissue depletion curve by assuming that every tissue declines in parallel. CBD is lipophilic, and the actual distribution to specific food-producing tissues was not quantified in these seven calves.

For a general account of why matrix identity controls toxicological interpretation, see the site’s independently published guide to biological matrices in forensic toxicology.

How was CBD identified and quantified in plasma?

The method used liquid-liquid extraction into methyl tert-butyl ether followed by ultra-high-pressure liquid chromatography coupled to high-resolution Orbitrap mass spectrometry. A deuterated CBD internal standard helped account for laboratory processing and analytical variability. The detector operated in positive electrospray mode with precursor and characteristic fragment ions. These are specific laboratory measurements of the target compound in processed bovine plasma, rather than an immunoassay-positive screening category.

The original calibration included fortified blank plasma from 0.20 to 2,000 ng/mL, with multiple quality-control levels. The printed method reported a limit of detection of 0.2 ng/mL and a limit of quantification of 1.0 ng/mL. A detector’s ability to register evidence of a compound at the detection limit is not the same as validated numeric quantification below the quantification limit. Reported calibration correlation exceeding 0.999 likewise speaks to the fitted analytical response over the stated range, not to clinical specificity for intoxication or the accuracy of an individual withdrawal-date calculation.

Values below the quantification limit were handled differently in the noncompartmental calculation depending on their position relative to the observed peak: the analysts assigned zero before the peak and half the quantification limit after the peak. This modeling convention affects calculated exposure estimates and should be disclosed when a result is used to support further inferences. For additional method context, consult the site’s published page on confirmatory drug testing.

What did the serial collection schedule cover?

The investigators obtained a predose baseline and sampled after administration at approximately 0.25, 0.5, 1, 2, 4, 6, 8, 12, 18, 24, 30, 36, 42, and 48 hours. This design captures early absorption as well as the decline after the observed peak, but it does not directly observe every animal to complete disappearance from plasma. The study’s last planned sample was at 48 hours.

The source’s Figure 1 plots the geometric mean plasma CBD concentration across these time points with confidence intervals on a logarithmic concentration axis. The mean profile rises after dosing and is still visible in the plotted range at the end of observation. Importantly, the line is a summary of different calves, not an individually validated dosing clock. An animal whose peak occurred at two hours had a different trajectory from one peaking at eighteen hours.

What were the original measured and derived results?

Table 1 of the original paper presents each calf’s pharmacokinetic estimates and the geometric mean. The evidence table below reorganizes the reported statistics by what each number means. Concentration peaks and their time points were read from the measured plasma curves; exposure areas, terminal slopes, half-life, and apparent disposition terms required mathematical analysis.

Unit check: 0.05 µg/mL equals 50 ng/mL by unit conversion; it is not a separate measurement. All entries derive from the original authors’ numerical Table 1, not a pooled analysis of independent trials or a human blood threshold.

Why does a 23-hour half-life not mean elimination in 23 hours?

The terminal half-life is a fitted characteristic of the estimated late log-linear decline. In a simple setting, one half-life corresponds to halving an applicable concentration, not eliminating the compound completely. Even that general rule applies to the fitted terminal phase under its assumptions, not necessarily to the changing absorption and distribution occurring earlier. It is incorrect to equate an animal’s time-to-peak with its elimination half-life.

The reported 23.02 hours is a geometric mean across seven fitted animal curves, with individual estimates ranging from 13.83 to 47.14 hours. The range is wide relative to the mean. Two calves could have different peak timing and estimated terminal slopes despite receiving the same weight-adjusted nominal dose. Sampling ended at 48 hours, so a half-life estimated partly near the boundary of observation is not a certified time when plasma becomes analytically negative.

It would also be inaccurate to double this average to produce a universal 46-hour waiting period, or multiply it by a fixed number of half-lives and announce a food-animal withdrawal rule. Such a rule would require product-specific target tissues, analytical residue limits, an appropriate depletion model, adequate sampling after the terminal phase, representative animal populations, and regulatory review.

How much exposure was extrapolated rather than measured?

For the seven calves, the original Table 1 reported that an estimated geometric mean 22.44% of AUC0–∞ arose beyond the last sampled plasma point. The individual extrapolated fractions ranged from 10.62% to 48.64%. This is not a minor bookkeeping issue. Almost half the total predicted exposure for one calf depended on extension of the fitted tail beyond the direct sampling window, rather than on later chemical measurements.

AUC to the last measurement was 0.70 h·µg/mL as a geometric mean, whereas AUC extrapolated to infinity was 0.95 h·µg/mL. Although those values describe different constructed exposure summaries, the gap reinforces that the sampling protocol did not directly encompass all estimated exposure. The investigators recognized the high extrapolated fraction as a weakness and called for longer sampling in future research.

When a veterinary or forensic report discusses an estimated half-life or projected terminal concentration, the fraction of the model unsupported by direct later samples should be stated explicitly. The site also maintains a live discussion of pharmacokinetic models and interpretive limits.

How variable were the calves?

The seven individual maximum plasma CBD concentrations were approximately 0.09, 0.09, 0.01, 0.08, 0.03, 0.04, and 0.04 µg/mL. One calf therefore had a reported maximum about one-ninth of the upper observed values after the same nominal weight-based dose. Individual Tmax values were 6, 18, 8, 2, 12, 8, and 8 hours. These data are more informative than a single central estimate because they show the extent to which a one-animal prediction could depart from the geometric mean.

Calf 6523 had the lowest peak and AUC, and particularly high estimated apparent clearance and volume of distribution. The authors discussed several potential reasons, including low oral bioavailability. Because the calf was observed receiving the dose, they did not consider an incomplete administration likely, but the data cannot prove one mechanism. The study did not include intravenous CBD to estimate absolute systemic bioavailability, and a single unusual animal cannot establish a breed or disease phenotype.

Statistically, these are seven observed profiles. There were no independent repeated randomized study arms to determine whether the variation is attributable to product formulation, calf maturation, feed timing, individual metabolism, sampling variation, or analytical uncertainty. Causal explanations remain hypotheses.

Are apparent clearance and distribution volume direct physiological measurements?

No. The original pharmacokinetic software returned an apparent oral clearance CL/F of 5.29 L/kg/h as the geometric mean and an apparent terminal distribution volume Vz/F of 175.86 L/kg. The notation F represents the unknown fraction reaching systemic circulation after oral administration. When oral bioavailability is not separately measured, CL/F is not equal to intrinsic hepatic clearance, and Vz/F is not an anatomical volume within the calf.

Some fitted apparent values look numerically large because they incorporate uncertain F. The extremely high value in the low-exposure calf does not imply the animal literally contained hundreds of liters of circulating blood per kilogram. A practitioner should preserve the “apparent” qualifier and report the original route of administration. These modeled terms should not be inserted into a human whole-blood back-extrapolation formula or treated as direct tissue concentrations.

What direct results did the researchers obtain about adverse effects?

The authors reported that the seven enrolled calves completed the trial without observed adverse drug reactions or behavioral changes. Daily monitoring included demeanor, hydration, and appetite. This limited observation is useful, but it is not equivalent to a blinded safety comparison against placebo, continuous instrumented behavioral testing, or a robust assessment of rare effects. Seven animals and one dose cannot exclude infrequent adverse reactions or long-term developmental consequences.

There were no formal between-treatment statistical comparisons, because the study had one treatment and no concurrent untreated arm. Consequently, the source did not establish a statistically nonsignificant clinical efficacy result, an absence of sedation by validated testing, or a null milk residue result. Descriptive absence of an observed event must not be rewritten as a controlled test with a p-value. Likewise, no reported toxicity at this dose is not a generally validated safe exposure limit for food animals.

What was not measured—and why does that change the forensic conclusion?

The investigation is properly classified as a single-dose, descriptive calf plasma CBD pharmacokinetic study. It did not measure cannabinoid residues in meat, liver, kidney, fat, or milk, or determine elimination from a tissue when that tissue is the compliance matrix. The animals were male calves, not milk-producing cows. Neither collected calf plasma nor model-derived AUC establishes the distribution of CBD into milk from adult lactating cattle.

It also did not quantify parent delta-9-THC in legal venous whole blood, establish THC intoxication, observe a driving task, diagnose any disease, or examine a particular litigation event. CBD and THC must remain chemically distinct. The absence of a study on THC in cattle is not evidence that CBD behaves like THC, nor that CBD can be converted directly into THC concentrations using a fixed ratio.

Can the reported plasma half-life be used to set a milk or meat withdrawal interval?

No. The authors explicitly identified future tissue-depletion research as necessary. A withdrawal interval concerns the time after administration when residues in the relevant edible product meet an applicable safety and regulatory standard. It is not simply the elapsed time from the last dose to a selected mean plasma concentration or to an arbitrary number of terminal half-lives. Tissue-specific distribution, metabolism, local regulations, analytical procedures, and variation in adult food-producing animals matter.

The source’s cohort was excluded from the food chain, and the researchers were studying pharmacokinetics rather than authorizing production use. Their findings should therefore be used to design future experiments, not to justify releasing an animal to slaughter or milk into commerce. Any current legal decision requires an independently checked jurisdiction-specific regulatory position, approved-product status, and competent veterinary and food-safety guidance; this article does not substitute for them.

Can plasma CBD in a calf be equated with a human THC blood result?

No. A veterinarian’s bovine plasma CBD test and a forensic laboratory’s human whole-blood THC result differ in species, compound, sample matrix, route, dosing context, and interpretive purpose. CBD is not THC. Centrifuged plasma is not unseparated venous whole blood. An oral single-dose pharmacokinetic curve in pre-ruminating calves is not an acute human impairment study. Even a confirmed concentration cannot by itself prove an animal’s behavior or an individual’s criminal act.

When comparing published cannabinoid studies, evaluate the original method and independently sampled biological units before placing numerical concentrations side by side. The proper comparison might concern analytical sensitivity, sampling windows, or biological variability, but it should never erase the difference between measured analytes or invent a species-to-species conversion factor.

How should an expert describe the findings in a scientific or legal report?

A defensible description identifies the exact primary article, seven young male Holstein calves, the single oral 5 mg/kg CBD oil dose, serial 48-hour venous collection followed by plasma separation, mass-spectrometric parent-CBD analysis, and the original geometric mean and range of Cmax, Tmax, half-life, and exposure extrapolation. It should then state what that design cannot address: adult cattle, milk and food tissues, repeat dosing, statutory residue limits, and case-specific causation.

A poorly supported summary would claim the authors “proved cattle are drug-free after 23 hours,” “established a 48-hour milk withholding period,” or “determined impairment from a cannabinoid blood level.” Each invents an endpoint absent from the original Methods and Results. A properly bounded summary instead treats the study as hypothesis-generating pharmacokinetic evidence that could inform the duration of sampling and design of future tissue-depletion work.

Frequently asked questions about calf CBD plasma testing

1. Did all seven calves receive 5 mg/kg CBD?

Yes. Each calf received the study’s nominal weight-adjusted single oral 5 mg/kg dose of a 25 mg/mL CBD oil formulation. That is not a quantified THC dose and not an experiment in hemp-fed cattle.

2. Was the 50 ng/mL result measured in whole blood?

No. The geometric mean Cmax was 0.05 µg/mL, equivalent to 50 ng/mL, in separated bovine plasma. The original did not report the same figure as unseparated whole-blood THC or CBD.

3. Does a 23.02-hour half-life mean that CBD was gone at 23 hours?

No. The figure is the geometric mean fitted terminal half-life, not a disappearance time. Individual half-life estimates ranged from about 14 to 47 hours, and the last sample was obtained 48 hours after dosing.

4. Why could plasma CBD still be measurable at the final sampling time?

The original concentration-time plot showed a slow late decline, and the model estimated meaningful exposure beyond the final sample. Its AUC extrapolation exceeded 20% as a geometric mean, showing why longer studies are needed to characterize terminal persistence.

5. Did researchers test CBD residues in cow’s milk or edible tissue?

No. The animals were male calves, and the quantified cannabinoid matrix was plasma. Milk, muscle, liver, kidney and fat depletion were not directly analyzed, so this experiment does not establish a withdrawal interval for any of those products.

6. Were serious adverse drug effects or a lack of efficacy statistically established?

The investigators reported no observed adverse drug reaction or behavioral change among seven calves. The study did not have a placebo group, formal comparative efficacy endpoint, or a powered rare-event safety design. Absence of a reported event is not a nonsignificant randomized treatment comparison.

7. Can this animal study predict whether a person with a THC blood result was impaired?

No. It studied oral CBD and directly assayed calf plasma CBD, not human THC whole blood, real-world driving, or validated impairment. No blood-to-behavior or calf-to-human forensic threshold was tested.

Bottom line: a useful preliminary plasma study, not a residue-clearance rule

The strongest evidence is specific: seven pre-ruminating bull calves given one 5 mg/kg oral CBD oil dose yielded a geometric mean plasma CBD peak of 50 ng/mL at a geometric mean 7.5 hours and a fitted geometric mean terminal half-life of 23.02 hours. Animal-to-animal variation was substantial, and substantial late exposure remained model-extrapolated after a 48-hour sampling window. These findings are valuable for planning better pharmacokinetic sampling, but they do not determine milk, meat or tissue withdrawal intervals, adult cattle dosing recommendations, individual intoxication, or human forensic impairment.

Primary source: Meyer K, Hayman K, Baumgartner J, Gorden PJ. Plasma Pharmacokinetics of Cannabidiol Following Oral Administration of Cannabidiol Oil to Dairy Calves. Front Vet Sci. 2022;9:789495. https://doi.org/10.3389/fvets.2022.789495.