Fatal Marijuana Overdose – Big Rehab’s Myth

“In man it is unlikely that Δ 9-THC plasma concentrations elevated enough to produce such acute toxic effects, could be reached after ingestion of cannabis preparations.” – Dr. Gabriel Nahas, 1972 report to UNODC.

Today I received an email from a reader looking for guidance on the topic of fatal marijuana overdose. Specifically, he needed to rebut someone in his circle who had stymied him with this little bit of pseudoscience quackery:

“Benno Hartung et al. (2014) report that the lethal half dose (LD50) for THC in humans is estimated to be around 30 mg/kg. This means that approximately 2 grams (.07 ounces) of pure THC has a 50% chance of killing a 150 pound man. This is the amount of THC found in .35 ounces of high grade weed with 20% THC content.”

From “Fatal Marijuana Overdose is NOT a Myth” by Kenneth Anderson, posted at

Well, you know I just had to go look this shit up. What follows is a look at just how shady Big Rehab can be in trying to scare up marijuana “addicts” to fill their coffers.

Benno Hartung’s fatal marijuana overdose data includes just two men’s deaths.

I looked up the study referred to by the author. It notes the case of two patients thought to have died from “Sudden unexpected death under acute influence of cannabis.”

In Case 1, “a 23-year-old male without known relevant illnesses suddenly
collapsed while using public transport” and “a small amount of marijuana was found in his pockets.” Subsequent autopsy “proved the acute influence of cannabis” because they found “THC 5.2 ng/ml” in his blood

Readers might know that the per se DUID law for cannabis in Washington State is the presence of 5 ng/ml of THC in the blood. Ever since passage of Washington’s I-502 initiative legalizing recreational marijuana, activists have pointed out the junk science behind that number. THC measurements that low hardly tell you anything about recent impairment, as a former marijuana critic showed when his THC test after abstinence and a full night’s sleep was over 15 ng/ml.

Furthermore, far higher concentrations of THC in the blood don’t even lead to much diminished driving capacity, as my favorite TV news driving test ever once showed.

Video “How high is too high? KIRO tests pot-smoking drivers to find out” from KIRO News, Seattle.

But we’re supposed to believe the 5.2 ng/ml of THC in this young man’s system was the cause of his fatal marijuana overdose?

In Case 2, “A 28-year-old male with a history of substance abuse (alcohol, amphetamines and cocaine until about 2 years before death; occasionally cannabis)… was found dead at home.” Some marijuana was found near him.

What, we’re going to ignore that this guy had abused booze, speed, and coke for at least five years, presuming, for no reason, that he started at age 21 and jump right to blaming the cannabis?

Yes, of course, because in his blood, they found 1.9 ng/ml of THC.

Yet that’s not the biggest leap of illogic in this quote, by far.

Benno Hartung was citing a 1972 study of rats

When I the followed the reference for Benno Hartung’s 2014 “report,” I discovered that it was citing “Toxicology and pharmacology of cannabis sativa with special reference to Δ 9-THC,” a report to the United Nations Office on Drugs and Crime from 1972.

Indeed, there is a reference to an established study of the 30 mg/kg lethal dose – not for humans, however, for mice. “The intravenous LD 50 in rats and mice is in the vicinity of 30 and 40 mg/kg respectively with death occurring in 15 minutes,” writes researcher Gabriel G. Nahas, MD, PhD.

And what of the potential harm to human beings, Dr. Nahas?

“In man it is unlikely that Δ 9-THC plasma concentrations elevated enough to produce such acute toxic effects, could be reached after ingestion of cannabis preparations.”

Dr. Gabriel G. Nahas, MD, PhD – “Toxicology and pharmacology of cannabis sativa with special reference to Δ 9-THC” 1972 report to UNODC

Where the author gets the 30 mg/kg lethal dose for humans comes from Dr. Nahas’ discussion of a 20-year-old French soldier who had gone into a four-day coma after trying to commit suicide.

“…the patient acknowledged having smoked nine to ten pipes of a mixture of tobacco and hashish with the intent of committing suicide, claiming that this method had been used by others. He admitted each pipe contained 15 to 20 gm of smoking mixture. Assuming (1) that the subject had smoked 180 gm; (2) 5% Δ 9-THC content for a potent hashish preparation; (3) a half-and-half mixture of tobacco and cannabis; (4) 50% absorption of the drug by the lungs, the lethal intravenous dose of Δ 9-THC in a 70 kg man would be of the order of 2,000 mg. or 30 mg/kg.

Dr. Gabriel G. Nahas, MD, PhD – “Toxicology and pharmacology of cannabis sativa with special reference to Δ 9-THC” 1972 report to UNODC

Just so you’re following this, a rehab website in 2019 is using a study from 2014 that cited the deaths of two young men from cannabis, one a long time hard drug user, neither of whom had substantial THC present in their bloodstream, and referred to a report from 1972 that proved how much intravenous THC would kill mice and cited French medical journal from 1971 that used a suicidal young soldiers’ self-report of how much hashish and tobacco he smoked, an assumption of its potency, and guess on its bioavailability to state definitively that 30 mg/kg of THC will kill half the humans who ingest it.

Sometimes, it’s what you don’t say…

The author also lets the reader do some of the heavy lifting in this scare. Note the three statements in his quote.

  1. The LD50 of THC is 30 mg/kg;
  2. Two grams THC would kill 150 lb (68 kg) man, half the time; and
  3. There’s 2 g THC in 0.35 oz (10 g) of 25% THC weed.

Oh my god, worries the reader, if I smoke ten grams of weed, there’s a 50/50 chance I’ll die!

But who said anything about smoking?

Remember, all this concern about the 30 mg/kg lethality is in regards to intravenous dosages of pure THC. Do you know anybody who’s tying off and injecting pure THC, like a stoner version of Trainspotting?

The author tries to up the stakes with the mention of “concentrated THC derivatives known variously as ‘wax,’ ‘dabs,’ or ‘butane hash oil’ consisting of up to 80% THC are now available on the streets; one tenth of an ounce of these would contain a lethal half dose of 2 grams of THC.”

Sure, and one-tenth of an ounce is over 2.8 grams. We’re talking about inhaling a nearly-three-gram dab. While I don’t recommend it, there are plenty of videos of young men at various cannabis expos dropping multi-gram dabs and surviving.

Because unlike intravenous ingestion, smoking or vaping marijuana is never going to deliver 100% of the THC to your lungs, and your lungs aren’t going to deliver all the THC to your bloodstream.

The rest of the article is worse

The 30 mg/kg lethality figure for THC is clearly based on some specious reasoning and leaps of logic. Reliable science tells us “the dose that kills animals range from 40 mg/kg to 130 mg/kg intravenously.” Could a large enough intravenous does of THC kill a human? Sure, but a large enough intravenous dose of just about anything can kill a human.

I went ahead and perused the rest of the article. In order to scare people into believing this fatal marijuana overdose is a real thing, the author consults with the CDC WONDER database. It consists of all the hospital-coded causes of death since 1999 in the United States. He then produces charts from the database to show how, yes, there have been cannabis poisoning deaths in America, and they’ve spiked since legalization.

“…according to this database there were 287 deaths due to cannabis or its derivatives in 2014. Numbers of deaths are calculated using UCD codes X40-44 (accidental drug poisoning), X60-64 (drug suicide), X85 (drug homicide), Y10-14 (drug poisoning of unknown intent) and MCD code T40.7 (Cannabis and its derivatives).”

From “Fatal Marijuana Overdose is NOT a Myth” by Kenneth Anderson, posted at

In other words, the author collected all the deaths from all drug poisonings, homicides, suicides, and unknown intent, and cross-referenced them with the code T40.7, indicating cannabis. He is fair about admitting that most of those deaths are from multiple drug causes, like mixing alcohol, opiates, and cannabis. But then he claims to have teased out the other drugs from the X&Y-coded deaths to leave only the T-coded marijuana deaths.

There’s a problem with that. Code T40.7 covers many reactions to cannabis. Specifically, according to ICD10Data, it covers, “Poisoning by, adverse effect of and underdosing of cannabis (derivatives).”

“Adverse effect of” means something causing distress, but not death. “Underdosing” is clearly not overdosing.

Fatal Marijuana Overdose chart from
Even if you believed this was true – and it isn’t – 300 deaths out of 30,000,000 marijuana smokers is a 0.001% death rate.

While I’ve been able to recreate the article’s first chart using the CDC WONDER database, I can’t get those marijuana-only deaths his second and third charts refer to. I’ll consult with some of my hospital coding and research science friends and update this piece with their take.

Bottom line… there is no realistic fatal marijuana overdose for humans

If this is the kind of science reporting Big Rehab thinks is going to scare more dollars into their coffers, I think the results of the last decade of opinion polls on legalization show you how well it’s working.

Unless you’re thinking of injecting grams of pure THC into your veins, I think you’re going to be fine. There’s no amount you can smoke, vape, eat, or dab that’s likely to be the cause of your death.