Holochart

THC pharmacokinetics · eaten vs smoked and vaporised · sources read October 3, 2026

How long will these gummies last? ​

You feel them from about an hour in until hour four to six, depending on the dose. That is about as long as a comparable vaporised dose, shifted an hour later, with the impairment arriving later still. The THC itself outlasts the feeling by days: its half-life is about a day in studies that sample for three days and four days in studies that sample for two weeks, whichever way it was taken. A urine test can find one edible for two to four days.

1–6 h
when a 25 mg brownie was felt
Spindle 2021 · vaporised 20 mg: 0–5 h
6%
of an eaten dose reached the blood
Ohlsson 1980 · smoked: 18%
25–36 h
half-life, the same eaten or injected
Wall 1983 · 4 days in 2-week studies
2–4 days
urine test positive after one edible
Schlienz 2018 · 10 to 50 mg

The question has two answers because "last" means two things: how long the effect is felt, and how long THC stays in the body. The route changes the first a great deal and the second hardly at all. This page collects what controlled human studies report for each, for eaten THC against smoked and vaporised THC. Of the 35 sources, 14 were read as open full text; the older ones are behind paywalls and were read as abstracts, or quoted from open reviews that are named where they are used. The reference list marks which. Every chart is a Holochart example running live in your browser. None of this is medical or legal advice, and the studies used doses of 5 to 50 mg in small groups of healthy adults.

How long you feel it ​

The cleanest comparison is a 2021 crossover in which the same 20 infrequent users ate cannabis brownies and, on other days, inhaled vaporised cannabis, and were assessed every hour. A dot below is an assessment at which the measure differed significantly from baseline.

Loading example…

Brownies with 10 and 25 mg THC; vaporised cannabis with 5 and 20 mg. Nothing was assessed between hours 6 and 8. The 10 mg brownie and the 5 mg vaporised dose did not impair performance on any task. Source: Spindle 2021.

The windows are the same length and offset by an hour: the 25 mg brownie was felt from hour 1 to hour 6, the 20 mg vaporised dose from the first minutes to hour 5. The difference that matters is in the impairment rows. After vaporising, measured impairment was over within an hour or two on the performance tests, while people still felt the drug. After the brownie it did not begin until hour 2 and ran to hour 5. An edible is at its most impairing at about the time an inhaled dose would be wearing off.

Other studies from the same group agree on the shape. After brownies of 10 to 50 mg, effects were not apparent for 30 to 60 minutes, subjective effects peaked at 1.5 to 3 hours and cognitive impairment at 2 to 5 hours (Schlienz 2020); an earlier study describes effects lasting 6 to 8 hours (Vandrey 2017). After smoking or vaporising, effects peaked within half an hour (Spindle 2018). A meta-regression of 80 studies estimated that most driving-related skills recover within about 5 hours of inhaling 20 mg, and noted that recovery after oral doses may take longer (McCartney 2021).

Why eating is slower and weaker ​

Inhaled THC reaches the blood through the lungs within minutes. Eaten THC has to be absorbed from the gut and pass through the liver first. The blood shows it: inhaled doses peak in the first sample, eaten doses two to three hours later and far lower.

Loading example…

Whole blood, group means. After vaporising 20 mg, blood THC peaked at 37 ng/mL; after eating 25 mg, at 3 ng/mL. Sources: Spindle 2018, Spindle 2021, Zamarripa 2026.

The low peak is a matter of how much gets in. In the one study that gave the same people THC by vein, by cigarette and in a cookie, 18% of the smoked dose reached the circulation and 6% of the eaten dose. Heavy smokers got more from a cigarette than light smokers, which is usually put down to how they inhale.

Loading example…

Mean with the standard deviation printed in each abstract. Sources: Ohlsson 1980, Lindgren 1981, Ohlsson 1982, Wall 1983.

The dronabinol label explains the gap: 90 to 95% of an oral dose is absorbed from the gut, but only 10 to 20% survives the first pass through the liver. So milligrams are not comparable across routes, and the ratings bear that out. At every dose tested, people rated the effect of inhaled THC higher than the effect of the same amount eaten.

Loading example…

Mean peak rating on a 0 to 100 scale; each line is one study. Standard deviations are 20 to 37 points. Sources: Spindle 2018, Spindle 2021, Schlienz 2020, Zamarripa 2023, Zamarripa 2026.

The second drug ​

What the liver does to THC on first pass is turn much of it into 11-OH-THC, which is also psychoactive. After an intravenous dose there is one part of it to every ten or twenty of THC. After an oral dose there is about as much metabolite as THC (Wall 1983; the Marinol label says "approximately equal").

Loading example…

Circles are plasma, diamonds whole blood. The four Wall 1983 points are quoted from an open-access review (Poyatos 2020). Sources: Oh 2017, Ewell 2021, Zamarripa 2023, Zamarripa 2026.

This is the usual explanation for edibles feeling different and for blood THC understating their effect: after eating, about half of the active drug in circulation is not THC.

Gummies specifically ​

Most controlled work uses brownies or capsules. One study tested products off the shelf: five commercial 10 mg edibles, three of them gummies, in seven regular users who had fasted overnight.

Loading example…

Circles are gummies, diamonds other edibles. Plasma was sampled for 4 hours only. Source: Ewell 2021.

On an empty stomach, plasma THC peaked 36 to 62 minutes after the gummies, sooner than the two to three hours seen with brownies eaten after breakfast. The product sold as fast-acting was not the fastest, and the spread between people on the same product was larger than the gap between products. With seven participants, none of those differences should be read as established.

Food is the larger variable. In a study of 5 mg dronabinol capsules, a high-fat meal delayed the first appearance of THC in the blood from half an hour to two hours and the peak from 1.73 to 5.59 hours, and more than doubled total exposure, while leaving the peak height about the same.

Loading example…

Mean and standard deviation, plasma THC, 53 or 54 participants per arm. Source: Oh 2017.

A gummy after a large dinner is, on this evidence, a slower and longer dose than the same gummy before it. That is also the mechanism behind the familiar mistake of taking a second one because the first has not worked yet.

How long THC stays ​

Now the second meaning of "last". The terminal half-life is the time for the amount of THC in the body to halve once absorption and distribution are finished. It is a property of how the body clears THC, and it should not depend on how the THC got in. The studies agree: the one that measured it after both intravenous and oral doses found the same 25 to 36 hours for each.

What the half-life does depend on is how long the study kept looking.

Loading example…

Dots are reported means, bars reported ranges. Hover a row for the note on each value. The Hunt 1980 value is quoted from a review (Chiang 1987). Sources: Ewell 2021, Oh 2017, Syndros label, Lemberger 1971, Wall 1983, Johansson 1988, Johansson 1989, Johansson 1990, Huestis 1998, Gustafson 2004, Johansson 1989b.

Studies that sample for a few hours after an oral dose report half-lives of 2.5 to 10 hours. Studies that sample for three days report about a day. Studies that followed chronic users for two weeks, with labelled THC and a sensitive assay, report four days, and in two people followed for four weeks, ten and thirteen days.

These are not disagreements. THC is fat-soluble and leaves the blood quickly for fatty tissue, then returns from it slowly; it was still measurable in fat biopsies from heavy users four weeks after their last use (Johansson 1989c). Early on, the fall in blood THC is mostly distribution. The slow final phase only appears once the concentration is low, and a short study or an insensitive assay never sees it. The short oral half-lives at the top of the chart describe the first hours after a gummy, not how long THC stays.

The practical reading is that a single edible in an occasional user is mostly governed by the first day, and that with regular use the slow pool in fat fills up and the long half-life takes over.

In daily users ​

That long tail has been measured directly. Thirty men who smoked cannabis daily lived on a closed research unit for up to 33 days, with blood taken every day.

Loading example…

Redrawn from Table 2 of Bergamaschi 2013. Quantification limit 0.25 µg/L. Participants left at different times, so the percentage on later days is of fewer people.

On admission 90% had measurable THC. After a week 79.2% of the 24 still resident did, after two weeks 60% of 20, and on day 30 two of the five who remained. The concentrations were low, a median of 0.4 µg/L at two weeks, and only 5% were above 1 µg/L by day 12. The inactive metabolite THC-COOH was present in most participants throughout. None of this reflects intoxication; it is THC leaving storage.

Testing positive ​

Drug tests mostly look for THC-COOH in urine, or THC in oral fluid or blood, and each has its own window.

Loading example…

Hover a row for the cutoff used, which changes the answer. Sources: Vandrey 2017, Swortwood 2017, Huestis 1996, Schlienz 2018, Goodwin 2008, Bergamaschi 2013.

After a single brownie in people with no recent use, urine stayed positive at the 15 ng/mL confirmation cutoff for a mean of 45 hours after 10 mg, 52 hours after 25 mg and 93 hours after 50 mg, and two of the six who ate 50 mg were still positive on day 7. A single cigarette gives a similar window for a similar dose. Route matters much less here than dose and history: among 60 regular smokers who stopped, the last positive urine test at the common 50 ng/mL screening cutoff came after a mean of 5 to 15 days depending on how high they started, and for some at the end of the 30-day study.

Oral fluid tests mostly detect THC left in the mouth, not THC from the blood, and stayed positive for about a day in occasional users whether the dose was eaten or smoked.

How much people differ ​

Every number above is an average over a small group, and individuals vary more than the routes do. Smoking shows it most plainly, because the dose a person takes in depends on how they inhale.

Loading example…

Redrawn from the table printed with Figure 1 of Perez-Reyes 1990. The type of the printed plus-or-minus is not stated.

Within one group smoking the same 2.4% cigarette, peak plasma THC ran from 12 to 137 ng/mL. Edibles vary as much: the dronabinol label gives between-person variation of about 66% for peak concentration, and in the brownie studies the standard deviation of peak blood THC is about as large as the mean.

What this does not tell you ​

  • Higher doses. The controlled studies stop at 50 mg. Many retail products and most accidental overconsumption are above that.
  • Regular users. Most of the effect data comes from people who had not used cannabis for a month or more. In one study frequent users did not rate a 50 mg oral dose above placebo, while occasional users did (Newmeyer 2017). Regular use does not shorten how long THC is detectable; it lengthens it.
  • Newer formulations. No open controlled study of THC drinks or nano-emulsion edibles in humans was found. One abstract reports a microencapsulated edible reaching its peak about 30 minutes sooner than a standard one, with no difference in peak height (Conner 2026).
  • Whether you are fit to drive. Blood THC does not track impairment after eating: the peak is low and late, and half the active drug is a metabolite.

Methods and limits ​

Sources were gathered on October 3, 2026 from PubMed, PubMed Central, Europe PMC, journal sites where the article is free, DailyMed and US government archives of the NIDA research monographs. Studies were chosen for being controlled administrations in humans that report pharmacokinetic values or timed effects, favouring the most cited.

The main limit is access. None of the 1970 to 1992 disposition studies is open in full, so their values come from PubMed abstracts or from open reviews that tabulate them; the widely cited Huestis 2007 review could not be read at all and nothing is taken from it. Whole blood and plasma are different matrices (plasma THC is roughly twice whole-blood THC) and are not mixed within any chart except where marked. No values were read off figures, so there are no concentration-time curves here apart from the tabulated abstinence data. Details and the known disagreements between sources are in the source notes.

References ​

  1. Spindle 2021. Spindle TR, Martin EL, Grabenauer M, et al. J Psychopharmacol 2021;35(7):786-803. Crossover of cannabis brownies (10, 25 mg THC) and vaporised cannabis (5, 20 mg) in 20 infrequent users, assessed hourly for 8 hours. doi:10.1177/02698811211021583. PMID 34049452. Read as: full text.
  2. Spindle 2018. Spindle TR, Cone EJ, Schlienz NJ, et al. JAMA Netw Open 2018;1(7):e184841. Crossover of smoked and vaporised cannabis (10, 25 mg THC) in 17 infrequent users. doi:10.1001/jamanetworkopen.2018.4841. PMID 30646391. Read as: full text.
  3. Schlienz 2020. Schlienz NJ, Spindle TR, Cone EJ, et al. Drug Alcohol Depend 2020;211:107969. Crossover of cannabis brownies (10, 25, 50 mg THC) in 17 infrequent users. doi:10.1016/j.drugalcdep.2020.107969. PMID 32298998. Read as: full text.
  4. Zamarripa 2023. Zamarripa CA, Spindle TR, Surujunarain R, et al. JAMA Netw Open 2023;6(2):e2254752. Crossover of brownies with 20 mg THC, with and without 640 mg CBD, in 18 adults; plasma for 12 hours. doi:10.1001/jamanetworkopen.2022.54752. PMID 36780161. Read as: full text.
  5. Zamarripa 2026. Zamarripa CA, Lin S, Klausner M, et al. JAMA Netw Open 2026;9:e269842. Crossover of brownies (10, 25 mg THC) with and without alcohol in 25 adults. doi:10.1001/jamanetworkopen.2026.9842. PMID 42065887. Read as: full text.
  6. Ewell 2021. Ewell TR, Abbotts KSS, Williams NNB, et al. Pharmaceuticals (Basel) 2021;14(8):817. Crossover of five commercial 10 mg THC edibles in 7 regular users, fasted; plasma for 4 hours. doi:10.3390/ph14080817. PMID 34451914. Read as: full text.
  7. Oh 2017. Oh DA, Parikh N, Khurana V, Cognata Smith C, Vetticaden S. Clin Pharmacol 2017;9:9-17. Dronabinol 5 mg capsule fasted and after a high-fat meal, 54 healthy volunteers. doi:10.2147/CPAA.S119676. PMID 28138268. Read as: full text.
  8. Swortwood 2017. Swortwood MJ, Newmeyer MN, Andersson M, et al. Drug Test Anal 2017;9(6):905-915. Oral fluid after about 50.6 mg THC eaten, smoked and vaporised by the same 11 frequent and 9 occasional users. doi:10.1002/dta.2092. PMID 27647820. Read as: full text.
  9. Vandrey 2017. Vandrey R, Herrmann ES, Mitchell JM, et al. J Anal Toxicol 2017;41(2):83-99. Cannabis brownies (10, 25, 50 mg THC), 6 adults per dose; blood and oral fluid for 9 days. doi:10.1093/jat/bkx012. PMID 28158482. Read as: abstract.
  10. Schlienz 2018. Schlienz NJ, Cone EJ, Herrmann ES, et al. J Anal Toxicol 2018;42(4):232-247. Urine after the Vandrey 2017 brownies, collected for 9 days. doi:10.1093/jat/bkx102. PMID 29300962. Read as: full text.
  11. Newmeyer 2017. Newmeyer MN, Swortwood MJ, Abulseoud OA, Huestis MA. Drug Alcohol Depend 2017;175:67-76. Subjective and physiological effects of about 50 mg THC smoked, vaporised and eaten by frequent and occasional users. doi:10.1016/j.drugalcdep.2017.02.003. PMID 28407543. Read as: abstract.
  12. Conner 2026. Conner BT, Smith EE, DiCecco SM, et al. Cannabis Cannabinoid Res 2026;11(5):385-395. Fast-acting (microencapsulated) edible against a standard edible in 20 participants; two authors work for edible makers. doi:10.1177/25785125261441366. PMID 41958200. Read as: abstract.
  13. McCartney 2021. McCartney D, Arkell TR, Irwin C, McGregor IS. Neurosci Biobehav Rev 2021;126:175-193. Meta-regression of 80 studies of THC and driving-related cognitive skills. doi:10.1016/j.neubiorev.2021.01.003. PMID 33497784. Read as: abstract.
  14. Ohlsson 1980. Ohlsson A, Lindgren JE, Wahlen A, Agurell S, Hollister LE, Gillespie HK. Clin Pharmacol Ther 1980;28(3):409-16. Plasma THC after 5 mg intravenously, a 19 mg cigarette and a 20 mg cookie in 11 men. doi:10.1038/clpt.1980.181. PMID 6250760. Read as: abstract.
  15. Lindgren 1981. Lindgren JE, Ohlsson A, Agurell S, Hollister L, Gillespie H. Psychopharmacology (Berl) 1981;74(3):208-12. Intravenous and smoked THC in heavy and light users. doi:10.1007/BF00427095. PMID 6267648. Read as: abstract.
  16. Ohlsson 1982. Ohlsson A, Lindgren JE, Wahlen A, Agurell S, Hollister LE, Gillespie HK. Biomed Mass Spectrom 1982;9(1):6-10. Deuterium-labelled THC, intravenous and smoked, in 5 heavy and 4 light users; sampled 48 to 72 hours. doi:10.1002/bms.1200090103. PMID 6277407. Read as: abstract.
  17. Wall 1983. Wall ME, Sadler BM, Brine D, Taylor H, Perez-Reyes M. Clin Pharmacol Ther 1983;34(3):352-63. Intravenous and oral THC in 6 men and 6 women; excretion followed for 72 hours. doi:10.1038/clpt.1983.179. PMID 6309462. Read as: abstract.
  18. Lemberger 1971. Lemberger L, Tamarkin NR, Axelrod J, Kopin IJ. Science 1971;173(3991):72-4. Intravenous radiolabelled THC in chronic smokers and non-users. doi:10.1126/science.173.3991.72. PMID 5087483. Read as: abstract.
  19. Hunt 1980. Hunt CA, Jones RT. J Pharmacol Exp Ther 1980;215(1):35-44. Intravenous radiolabelled THC in 6 men before and after chronic oral dosing. PMID 6256518. Read as: secondary.
  20. Johansson 1988. Johansson E, Agurell S, Hollister LE, Halldin MM. J Pharm Pharmacol 1988;40(5):374-5. Smoked deuterium-labelled THC in chronic users; plasma for 13 days. doi:10.1111/j.2042-7158.1988.tb05272.x. PMID 2899638. Read as: abstract.
  21. Johansson 1989. Johansson E, Halldin MM, Agurell S, Hollister LE, Gillespie HK. Eur J Clin Pharmacol 1989;37(3):273-7. Smoked labelled THC in 8 heavy users; plasma for 10 to 15 days, 4 weeks in two. doi:10.1007/BF00679783. PMID 2558889. Read as: abstract.
  22. Johansson 1989b. Johansson E, Halldin MM. J Anal Toxicol 1989;13(4):218-23. Urinary THC-COOH in 13 heavy users over 4 weeks of abstinence. doi:10.1093/jat/13.4.218. PMID 2550702. Read as: abstract.
  23. Johansson 1990. Johansson E, Gillespie HK, Halldin MM. J Anal Toxicol 1990;14(3):176-80. Urinary metabolites after smoked and oral radiolabelled THC; 5-day collection. doi:10.1093/jat/14.3.176. PMID 2165199. Read as: abstract.
  24. Huestis 1996. Huestis MA, Mitchell JM, Cone EJ. J Anal Toxicol 1996;20(6):441-52. Urine after one 1.75% or 3.55% THC cigarette in 6 men. doi:10.1093/jat/20.6.441. PMID 8889681. Read as: abstract.
  25. Huestis 1998. Huestis MA, Cone EJ. Ther Drug Monit 1998;20(5):570-6. Urinary THC-COOH excretion half-life in the same 6 men. doi:10.1097/00007691-199810000-00021. PMID 9780137. Read as: abstract.
  26. Gustafson 2004. Gustafson RA, et al. J Anal Toxicol 2004;28(3):160-7. Urinary THC-COOH after oral THC three times a day for 5 days in 7 volunteers. doi:10.1093/jat/28.3.160. PMID 15107145. Read as: abstract.
  27. Smith-Kielland 1999. Smith-Kielland A, Skuterud B, Morland J. J Anal Toxicol 1999;23(5):323-32. Urinary cannabinoids in frequent and infrequent users after stopping. doi:10.1093/jat/23.5.323. PMID 10488918. Read as: abstract.
  28. Goodwin 2008. Goodwin RS, et al. J Anal Toxicol 2008;32(8):562-9. Urine from 60 regular smokers during up to 30 days of monitored abstinence. doi:10.1093/jat/32.8.562. PMID 19007504. Read as: full text.
  29. Bergamaschi 2013. Bergamaschi MM, et al. Clin Chem 2013;59(3):519-26. Daily whole blood from 30 male daily smokers during up to 33 days of monitored abstinence. doi:10.1373/clinchem.2012.195503. PMID 23449702. Read as: full text.
  30. Perez-Reyes 1990. Perez-Reyes M. NIDA Res Monogr 1990;99:42-62. Peak plasma THC after smoking cigarettes of nine potencies. PMID 2176276. Read as: full text.
  31. Johansson 1989c. Johansson E, et al. Biomed Chromatogr 1989;3(1):35-8. THC in fat biopsies from heavy users, before and 4 weeks after smoking. doi:10.1002/bmc.1130030109. PMID 2539872. Read as: abstract.
  32. Syndros label. Syndros (dronabinol) oral solution: US prescribing information, DailyMed set ID a7801c70-995d-46a2-91ee-141ef427c6b5, version of 2026-04-27. Single 4.2 mg oral dose, fasted. Read as: label.
  33. Marinol label. Marinol (dronabinol) capsules: US prescribing information, DailyMed set ID d0efeeec-640d-43c3-8f0a-d31324a11c68, version of 2023-02-15. Oral dronabinol capsules. Read as: label.
  34. Poyatos 2020. Poyatos L, et al. Medicina (Kaunas) 2020;56(6):309. Open-access review whose tables quote peak concentrations from Ohlsson 1980 and Wall 1983. doi:10.3390/medicina56060309. PMID 32585912. Read as: full text.
  35. Chiang 1987. Chiang CN, Rapaka RS. NIDA Res Monogr 1987;79:173-88. Review that quotes the half-life from Hunt and Jones 1980. PMID 2830536. Read as: full text.

Chart sources and data: examples/demos/thc-gummies.

Released under the MIT License.

Copyright © 2026 Holochart contributors