Before 2004
Nobody wrote this reversal down for four thousand years.
Cannabis enters the medical record as the thing you take to stop being sick. There is no clear description of it doing the opposite until Adelaide, 2004.
The history
Three thousand years, and then this.
A plant used to stop vomiting since before written medicine started making people vomit uncontrollably in the last twenty years. This page runs that claim down properly: what the old record actually says, what a genuine primary source search found and did not find, and where that search is weaker than it sounds. The medical case for CHS itself lives on the evidence page. This page asks one narrower question.
The plant that stopped vomiting now causes it.
Cannabis is one of the oldest continuously used medicines on earth, and for most of that history people used it to make vomiting stop, not start. Ayurvedic medicine used bhang as a digestive aid roughly three thousand years ago. Physicians across the medieval Islamic world wrote it up explicitly as an antiemetic, a drug that settles the stomach. Dronabinol, a synthetic form of THC, is still an FDA approved prescription antiemetic for chemotherapy nausea today.[2]
The same compound, from the same plant family, now causes a distinct vomiting syndrome severe enough that people show up to the ER convinced they are dying, and find that the only thing that helps is standing under scalding water. That is the paradox. It is the reason this page exists, and it is true.
One warning belongs at the top, before the search even starts, because it is the honest way to hold the rest of this page. Saying nothing like this appears in the old record is an argument from silence, and a document that was never asked the question is weak evidence on its own. So this page does something most pages making this argument do not: it publishes the search itself, in full, rather than just the conclusion. If a claim here turns out to rest on nothing, that will be visible, not buried. The method behind how every source below was checked is written up separately on how this was made.
Antiemetic, for most of recorded history.
None of this is controversial background. Cannabis shows up in the written medical record early, and where it names a digestive effect at all, that effect runs the opposite direction from CHS.
- c. 1550 BCEAncient EgyptThe Ebers Papyrus lists cannabis as medicine, among the oldest written drug records that survive.[2]
- c. 1000 BCEAyurveda, IndiaBhang is used as a digestive aid, to settle the stomach and quicken digestion. A remedy for nausea, not a cause of it.[2]
- 8th to 18th centuryMedieval Islamic medicinePhysicians document cannabis in detail across several centuries and list it explicitly as an antiemetic.[2]
- 1839O'Shaughnessy brings it into Western medicineHis clinical description is later quoted directly by the Indian Hemp Drugs Commission, and it reports the opposite of a vomiting illness.
"The intoxication lasts about three hours, when sleep supervenes; no nausea or sickness of stomach succeeds, nor are the bowels at all affected next day."Quoted verbatim by the Indian Hemp Drugs Commission, 1894[1]
By the time cannabis reached its widest documented study anywhere, the 1893 to 1894 Indian Hemp Drugs Commission in British India, its reputation as a stomach settler was already centuries old. That is exactly why the Commission is the right place to look for the first sign of anything resembling CHS, and why the next chapter reads the actual text instead of a summary of it.
The largest study ever done, read directly this time.
Most places online, including earlier versions of this project, describe the Indian Hemp Drugs Commission the way people describe a book they have not opened: an appeal to its size and reputation rather than its content. The Report of the Indian Hemp Drugs Commission, 1893 to 94, Volume I, questioned roughly twelve hundred witnesses across British India and studied heavy daily users at a scale nothing since has repeated. For this page, the primary text was read directly rather than paraphrased, the whole of Volume I on archive.org, roughly forty eight thousand four hundred lines of period OCR.[1] Here is what that reading actually found.
What the Commission did record
Chapter XI, on physical effects, spends paragraphs 499 to 511 on the digestive consequences of habitual and excessive use at length: indigestion, lost appetite, wasting, constitutional decline. The Commission quotes its own medical witness, Dr Prain, directly.
"What does happen is that digestion is impaired, the daily consumption of such a dose of narcotic overburdens the alimentary and eliminating systems, the man's appetite goes, and the food he takes is not properly absorbed."Dr Prain, quoted in the Commission's Report, paragraph 511[1]
The Commission is not shy about naming harm where it found it, which is exactly why its silence on one specific thing carries weight. In paragraph 510 it rejects outright any causal link between cannabis and dysentery, and calls the wider evidence for physical harm from moderate use "in the highest degree defective," language sharper than most modern guideline authors would risk.[1] A witness from Hyderabad describes the same chronic pattern in plainer words.
"It impairs the constitution, injures digestion, and causes anaorexia. Smoking ganja has reduced many strong men to the lowest ebb."T. Streenivasa Rao, witness 9, Hyderabad[1]
The one mention of vomiting
One passage names vomiting at all, and it names it as an acute effect of a single dose, not a recurring illness.
"The man becomes peevish, stupefied, sees double, and occasionally it may cause vomiting."The Commission's Report, on acute intoxication[1]
That is a plausible, unremarkable description of an intoxicated stomach reacting to one dose. It is not a cyclic pattern, it is not tied to hot bathing, and it does not recur across the weeks or years the way CHS does. Reading it as an early sighting of CHS would be reaching well past what the sentence says.
What is not there
Search the full text for the words that would matter, and they are simply absent. Bath, bathing, hot water, cyclic, and periodic vomiting do not occur anywhere across all forty eight thousand four hundred lines of Volume I.[1]
Before you trust that silence, here is why it is weaker than it sounds.
An argument from silence is only as strong as the search behind it. Four honest limits belong right next to the claim above, not in a footnote a reader has to go find.
Four reasons the search has real limits
The Commission was never asked about episodic vomiting or bathing behavior in the first place. It was assembled to study intoxication, addiction, and the moral and social effects of hemp drugs, not to screen for a syndrome nobody had reason to look for yet.
Nineteenth century Indian bathing customs would have made hot bathing an unlikely thing to notice or record even if it happened. Daily bathing with heated water was not a universal or remarkable act the way a scalding shower is in a modern home, so a witness would have had little reason to mention it as a treatment.
The OCR is machine generated from a century old scan, and it contains visible errors on a straightforward reading. A garbled mention of bathing or cyclic symptoms buried in a misread word cannot be fully ruled out.
Volumes II through VII, the raw witness testimony behind Volume I's summary, were not sampled for this page. The text is available on archive.org. Reading it was simply not done here, and that is a limit of this project's search, not of the historical record itself.
The honest verdict: this is real negative evidence that nobody in the largest study of heavy chronic cannabis use ever conducted flagged a recognizable cyclic vomiting syndrome as a known effect of heavy use. It is not proof the syndrome did not exist.
Two studies would have caught this by the 1970s. Neither could be opened.
If CHS existed at 1970s potencies, two specific studies are exactly where it should have surfaced. Vera Rubin and Lambros Comitas followed chronic heavy daily cannabis users in Jamaica for years, with structured medical examinations, publishing the results as Ganja in Jamaica in 1975.[3] William Carter ran a similar long term study of heavy daily users in Cannabis in Costa Rica, published in 1980.[4] Both are the closest thing that exists to a real cohort study of the population CHS should hit hardest, run decades before anyone had a name for the syndrome or a reason to look for it.
What could not be done, said plainly
Neither monograph could be reached for this page. Both predate wide digitization, neither is available on archive.org in full text, and neither could be located through any access route tried. Their silence, if they are silent, is currently unverified. An unfinished search shown honestly is worth more than a finished one asserted, so this stays an open question rather than a claim.
The plant is not the plant either study measured.
If nothing like CHS shows up before 2004, something had to change. It did, and the honest version of that change is more interesting than a single climbing line, because part of it plateaued and part of it kept going.
The historical series
The longest running US measurement, drawn from DEA seized samples, tracked average THC in cannabis flower from about 4% in 1995 to about 12% in 2014.[5] Over the same span, CBD, the part of the plant that buffers THC's effects, was bred toward irrelevance.
1995
3.9%
average THC in flower
2014
12%
average THC in flower
1995
14 : 1
THC to CBD
2014
80 : 1
THC to CBD
The global rate of change
A 2021 global meta analysis pooling 12 studies from 1970 to 2017 found herbal cannabis THC rising at 0.29 percentage points per year worldwide, and cannabis resin rising faster still, at 0.57 points per year, while CBD held flat across the same stretch.[6] This page cites that rate of change and nothing more. The paper's absolute regional starting and ending percentages sit behind a publisher wall that five separate access attempts could not clear, so no specific "from this percent to that percent" figure is drawn from it here. When a source is behind a wall, the honest move is to cite exactly what was read, not to round up.
The modern measured numbers
Flower potency is not the whole story anymore. Concentrates, the dabs, waxes, and vape distillates that went mainstream in the 2010s, are measured lab data now, not an estimate off a factsheet. Ninety nine concentrate products tested across fifty two Colorado dispensaries averaged 70.7% THC.[7] A separate, independent Colorado sample found a median of 72.6%.[8] Washington's inhalable extracts averaged almost 85% in 2023, up from roughly 70% in 2017.[9]
Concentrates, measured in the lab
A potency with no natural analog anywhere in the plant's history.
The surprise: flower plateaued
This is the finding that cuts against the simple "potency just keeps climbing" story, and it deserves real weight rather than a passing mention. Washington's own market regulator measured flower holding roughly flat near 22% from 2017 straight through 2023.
Washington flower, average THC, 2017 to 2023
Roughly flat for six straight years, not still climbing.
Source: Washington Joint Legislative Audit and Review Committee, Cannabis Market Study Preliminary Report, May 2025, stating "no evidence that average THC concentration of flower has changed in that period."[9] Colorado's independent measurements land nearby: mean 20.8% in one 2025 sample, median 18.1% in a separate 2026 sample.[7][8]
Flower stopped climbing years ago. If flower alone drove CHS, the trend lines for the syndrome should have leveled off with it. They have not, which points toward concentrates and the form cannabis is consumed in as much as the raw number on a jar of flower.
The labels are wrong too
A site that shows the labels themselves are unreliable earns more trust than one that only repeats them, so here is what the same lab data found about accuracy.
2.2 to 4.25 pts
56.7%
of flower products fell within 15% of their printed label at all.[7]
30.3%
of flower products were over labeled, printing a higher THC number than the lab found inside.[7]
96%
of concentrate products, by contrast, were accurate within 15% of their label.[7]
If you have ever bought flower that felt weaker than the number on the jar, the lab data says you were probably right.
The product mix moved even where potency did not
Flower's share of the market fell even as its own potency held steady, and the shift toward concentrated forms is where the real change since the 1970s and the 1990s actually lives.
Washington cannabis sales by product type, 2023
Respondents' dollars, not people. A larger share moved toward concentrated forms.
Source: Washington Joint Legislative Audit and Review Committee, Cannabis Market Study Preliminary Report, May 2025.[9]
Total THC sold through flower still rose 70% from 2017 to 2023, simply because more flower moved. THC sold through extracts and edibles roughly tripled over the same period.[9]
2017 to 2023
70%
rise in total THC sold via flower
2017 to 2023
3x
roughly, rise in total THC sold via extracts and edibles
Neither cohort study in the last chapter ever saw a product like this. That is the plant that changed, and it changed in a form Rubin, Comitas, and Carter had no way to measure even if they had looked.
The rise
Then the emergency rooms started counting.
Ontario multiplied thirteen times. Massachusetts logged a fourteen-fold rise. The line climbed steeply, and in the most recent data it flattened.
No, it is not GMO. It is something breeders did.
The honest version
There is no genetically modified cannabis for sale anywhere. No transgenic strains, no CRISPR flower on a dispensary shelf. What actually happened is selective breeding, the same process that turned wild mustard into broccoli and cauliflower, layered with sinsemilla growing, feminized seed, controlled environment horticulture, and solvent extraction. Real gene editing cannabinoid science exists, but it lives in yeast vats and university labs, not in anything a person smokes.[10][11]
So the instinct behind "they must be genetically modifying it" is pointing at something true even though the word is wrong. The plant was pushed past anything three thousand years of cultivation ever produced: bred for maximum THC, stripped of CBD, extracted past 85%, and sold in forms no historical grower or 20th century researcher ever measured. Not genetically modified. Genuinely re-engineered.
Twenty years from first description to a diagnostic code.
- 2004The first descriptionAllen and colleagues in Adelaide describe 19 chronic heavy users with cyclical vomiting relieved by hot bathing. CHS enters the medical literature by name for the first time.[12]
- 2012The 98 patient seriesA Mayo Clinic case series sets out formal diagnostic criteria that are later folded into the Rome IV functional gastrointestinal disorder criteria.[13]
- 2024Guidelines arrive
- October 1, 2025It gets a codeICD-10-CM R11.16, "Cannabis hyperemesis syndrome," effective from the CDC's National Center for Health Statistics.[16]
Close on the code, because it matters more than a bureaucratic footnote. For twenty years, US administrative data had no way to count CHS directly. Every pre-2025 count in this project's own evidence page is built from a proxy, a broader vomiting code paired with a cannabis use disorder code, layered with inference. A diagnosis that could not be counted for two decades is the reason so much of the epidemiology is still argued over rather than settled, and it is the reason the next chapter has to be as careful as it is.
The potency story is a hypothesis with good circumstantial support, not a proof.
Potency rose. CHS appeared afterward. That timing is the best explanation available, and it is not proof. Researchers behind the Colorado post legalization data said outright that their own numbers cannot separate "greater use, improved reporting, or both."[17] No study has ever expressed CHS incidence over a denominator of heavy users tracked across time. No natural experiment comparing CHS rates across a clean THC potency policy boundary exists, despite real attempts to find one.
Two kinds of study would actually settle this. One measures incidence over a heavy user denominator across years, so a real rise can be told apart from more people using more heavily. The other compares CHS rates across a genuine potency policy boundary, one place capping THC hard, a comparable place not capping it at all, everything else held roughly constant.
The near ideal version of that second study already has a candidate: Quebec caps THC at 30% and bans most concentrates outright, while the rest of Canada does neither. That comparison cannot currently be built, because Quebec is entirely excluded from the national hospitalization database used for Canada's CHS trend data, "due to differences in collection and reporting systems," in the researchers' own stated words.[18] That is why the study you would actually want to run does not exist yet, and saying so plainly is more honest than leaving the gap unexplained.
One rough, secondary number is worth naming and immediately qualifying: an estimate of roughly 0.5 to 0.6% of frequent cannabis users aged 15 to 44 needing an ED visit for CHS in Alberta after legalization, reached only through a summary brief rather than the primary source directly.[19] Treat that as a rough signal about scale, not a settled per user rate.
What the search actually found.
Three thousand years of recorded cannabis use, and the largest chronic use study ever conducted, contain no clear description of a cyclic, hot bath relieved vomiting syndrome. Read directly rather than summarized, the record shows plenty of real harm from heavy use and one acute, single dose mention of vomiting, nothing more. That silence is real evidence. It is not proof, because the search that produced it has honest limits stated in full above, and two of the studies that should have settled the question could not be opened at all.
Set beside a plant whose flower held flat for six years while its concentrates kept climbing past anything measured before, the paradox in chapter one still stands. Something changed. The record says it was not the plant's age old use as medicine.
Where to go from here.
This page covers the search into the historical record and the plant's own change. The clinical case, the diagnostic code, the rising and plateauing case counts, and the honest counterarguments live on the evidence page.
Sources
Every number above is numbered below with a working link where one exists. How each source was graded, and the specific texts this project could not get past a paywall, a login wall, or simple non-digitization, are written up in full on how this was made.
- Report of the Indian Hemp Drugs Commission, 1893 to 94, Volume I. archive.org/details/in.ernet.dli.2015.102443. A genuine primary source, read directly in full OCR for this page, roughly forty eight thousand four hundred lines. The OCR is machine generated from a century old scan and can contain errors, and only Volume I of seven was searched. Volumes II through VII hold the raw witness testimony and were not sampled.
- "History of medical cannabis." en.wikipedia.org/wiki/History_of_medical_cannabis. A general encyclopedia entry used only for uncontested background on ancient and medieval antiemetic use and dronabinol's modern approval, not a primary source in its own right.
- Rubin V, Comitas L. Ganja in Jamaica: A Medical Anthropological Study of Chronic Marihuana Use. Mouton, 1975. Not digitized in full text. The primary monograph could not be located online for this project, so its content could not be checked directly.
- Carter WE, ed. Cannabis in Costa Rica: A Study of Chronic Marihuana Use. Institute for the Study of Human Issues, 1980. Not digitized in full text. The primary monograph could not be located online for this project, so its content could not be checked directly.
- ElSohly MA et al. "Changes in Cannabis Potency Over the Last 2 Decades." Biol Psychiatry 2016. doi 10.1016/j.biopsych.2016.01.004. The long running US Potency Monitoring Program on DEA seized samples. Seized product may not perfectly represent what regular users actually consume.
- Freeman TP et al. "Changes in delta-9-THC in cannabis over time: systematic review and meta-analysis." Addiction 2021;116(5):1000-1010. doi 10.1111/add.15253. A rigorous global meta analysis. This page cites only its rate of change figures; the paper's absolute regional percentages sit behind a paywall this project could not clear.
- Giordano G, Brook CP, Ortiz Torres M et al. Scientific Reports 2025;15:20822. doi 10.1038/s41598-025-03854-3. Directly measured lab data, 99 concentrate and flower products across 52 Colorado dispensaries. A single state, a moderate sample size.
- Limbacher SA et al. PLOS ONE 2026. doi 10.1371/journal.pone.0321832. An independent Colorado sample that largely confirms the Scientific Reports numbers. Small sample, n=10 for concentrates.
- Washington Joint Legislative Audit and Review Committee, Cannabis Market Study Preliminary Report, May 2025. State market regulator data, not an independent academic study, but the most complete product mix and potency tracking dataset available for any single legal market.
- "Sinsemilla." en.wikipedia.org/wiki/Sinsemilla. Background encyclopedia entry on selective breeding history, used to establish that potency gains came from breeding, not genetic modification.
- Luo X et al. "Complete biosynthesis of cannabinoids and their unnatural analogues in yeast." Nature 2019. doi 10.1038/s41586-019-0978-9. Real transgenic cannabinoid science. It has never touched a dispensary product and is cited only to show the actual gene editing work happens in a lab, not on a farm.
- Allen JH, de Moore GM, Heddle R, Twartz JC. "Cannabinoid hyperemesis: cyclical hyperemesis in association with chronic cannabis abuse." Gut 2004;53(11):1566-1570. pubmed.ncbi.nlm.nih.gov/15479672. The first description, 19 patients. Small and retrospective by nature of being first, but it has held up for two decades.
- Simonetto DA, Oxentenko AS, Herman ML, Szostek JH. "Cannabinoid Hyperemesis: A Case Series of 98 Patients." Mayo Clin Proc 2012;87(2):114-119. mayoclinicproceedings.org. The largest single center case series and the source of the formal criteria later folded into Rome IV. Retrospective, single institution.
- Rubio-Tapia A, McCallum R, Camilleri M. "AGA Clinical Practice Update on Diagnosis and Management of CHS." Gastroenterology 2024. doi 10.1053/j.gastro.2024.01.040. A specialty society's own clinical guidance. Guidance, not a primary trial.
- Borgundvaag B et al. "GRACE-4: cannabinoid hyperemesis syndrome management in the emergency department." saem.org/publications/grace/grace-4. A multi society emergency medicine guideline. Most of its own recommendations are graded low or very low certainty by the guideline authors themselves.
- ICD-10-CM R11.16, "Cannabis hyperemesis syndrome," effective October 1, 2025, CDC National Center for Health Statistics. icd10data.com. A regulatory fact, not a research finding. About as solid a citation as exists on this page.
- Kim HS et al. Acad Emerg Med 2015. doi 10.1111/acem.12655. One of the earliest post legalization datasets. The authors state outright their data cannot separate more use from more reporting, and this page quotes that limitation directly.
- Plebon-Huff S, Goodman S, Abramovici H. Frontiers in Public Health 2026. doi 10.3389/fpubh.2026.1740300. National Canadian hospitalization data. Quebec is excluded from the underlying database, a real, stated limitation the authors disclose themselves, which is why the ideal Quebec comparison in chapter nine cannot currently be built.
- Andrews et al 2022, reached only through a Canadian Centre on Substance Use and Addiction summary brief rather than the primary source directly. The Alberta per user estimate above is a secondary source read, not a primary one, and should be treated as a rough signal rather than a verified rate.