39

According to the National Safety Council, in 2021 in the US, 1159 people died of cannabis overdose, of which 1130 of the deaths were preventable:

enter image description here

Is this true?

I'm skeptical because the drug enforcement agency says on its Marijuana/Cannabis Factsheet "No deaths from overdose of marijuana have been reported".

DavePhD
  • 106,761
  • 24
  • 450
  • 479
  • 40
    I'm also very curious about the 29 unpreventable cannabis deaths due to overdose. – Ben Hocking Mar 08 '23 at 13:22
  • 2
    I did find this single example as to a THC-delivery device that can lead to overdose (cannabis contains THC), but it doesn't answer your question: https://www.newsweek.com/thc-overdose-death-marijuana-exposure-united-states-1442742 – Ben Hocking Mar 08 '23 at 13:25
  • 3
    Also, the CDC page on overdoses has a section on marijuana, and while it mentions negative effects, it doesn't actually mention overdoses, despite that being the primary subject matter of the page: https://www.cdc.gov/drugoverdose/deaths/other-drugs.html#marijuana – Ben Hocking Mar 08 '23 at 13:28
  • 2
    Additionally, while this is a Canadian legal source, it does point to the possibility of there being some slight legal distinction between cannabis and marijuana, which indirectly relates back to the Newsweek story: https://www.legalline.ca/legal-answers/what-is-the-difference-between-cannabis-and-marihuana/ – Ben Hocking Mar 08 '23 at 13:30
  • 12
    The chart says "select drugs". I wonder why they chose to select cannabis and not nicotine and alcohol. Alcohol Poisoning Deaths — CDC says that over 2000 people per year die of alcohol overdoses, and that's a direct cause, not simply contributing. – Ray Butterworth Mar 08 '23 at 19:24
  • 4
    @RayButterworth Contributing would clearly shoot that number sky high. DUI incidents are non-trivial, and there are a lot more other types of contributing deaths and injuries. – Nelson Mar 09 '23 at 02:15
  • 2
    Is cannabis listed also, even if the person dies of shotgun overdose? – Per Alexandersson Mar 09 '23 at 06:55
  • 1
    I LOLed. The Therapeutic index of morphine is 70. It's 15 for cocaine and ~10 for alcohol. As far as we know, it's ∞ for cannabis. – Eric Duminil Mar 10 '23 at 07:10
  • 2
    The lethal dose of THC is known - note, that is THC which is only a part of cannabis. Read https://en.wikipedia.org/wiki/Tetrahydrocannabinol; I don't think anybody knows how much actual cannabis you have to somehow ingest to get to a lethal dose, but I somehow doubt it is realistic to think that it can be achieved naturally. – j4nd3r53n Mar 10 '23 at 17:35

2 Answers2

81

That's not what this data says.

What's happening is that in cases of overdose deaths, both causative and contributing drugs get listed in this report. We have a hint in the description of the data, which states, "When comparing deaths by drug type, multiple drugs are often listed for one death." According to the data table, statistics are generated using ICD codes:

Deaths are classified using the International Classification of Diseases, 10th Revision. Preventable drug poisoning deaths are identified using underlying cause-of-death codes X40-X44, while all drug poisoning deaths are identified by codes X40-X44, X60-X64, X85, and Y10-Y14.

Among deaths with drug overdose as the underlying cause, the following multiple cause-of-death codes indicate the drug type(s) involved: Any opioid (T4NR-T40.4, T40.6), heroin (T40.1), methadone (T40.3), synthetic opioids other than methadone (including fentanyl, T40.4), natural and semisynthetic opioids (including oxycodone and hydrocodone, T40.2), benzodiazepines (T42.4), cocaine (T40.5), and cannabis (T40.7). Deaths involving more than one opioid category are counted in both categories.

Digging into the ICD codes a little deeper, however, we find this information, emphasis mine:

Codes for underlying cause of overdose death include the letter X or Y (for example, X40: accidental poisoning, X60: intentional self-harm).

Codes for opioid-related contributing cause of overdose death include a T (for example., T40.1: heroin, T40.4: other synthetic opioids).

https://mnprc.org/wp-content/uploads/2019/01/using-icd-10-codes-to-assess-opioid-related-overdose-deaths.pdf

So, cannabis is only uniquely identifiable in this dataset under code T40.7, which merely indicates a "contributing cause" but not an "underlying cause" of death. I can't find any ICD code that denotes cannabis as an underlying cause of death - one might infer from the lack of such a code that fatal cannabis overdose is exceedingly rare.

Overall, it seems this data indicates that there were 1,130 people that died from a drug overdose who had marijuana listed as a contributing - but not principally causative - factor in the death. The data does not indicate that there were 1,130 deaths whose underlying cause was cannabis overdose.

It is possible that there are deaths caused by cannabis overdose which appear in other categories, but from this data there is no way to ascribe a number specifically to cannabis. It's theoretically possible that over 1000 fatal cannabis overdoses appear in other categories, but it seems medically unlikely. While we can't conclude from this data that there aren't over 1000 fatal cannabis overdoses, we also can't conclude from this data that there's even 1 fatal overdose caused by cannabis.

Thomas Markov
  • 285
  • 1
  • 8
Nuclear Hoagie
  • 787
  • 6
  • 10
  • This says the T40 codes include "underdosing" https://www.icd10data.com/ICD10CM/Codes/S00-T88/T36-T50/T40-/T40.7 "Poisoning by, adverse effect of and underdosing of cannabis" If someone died from lack of cannabis, would that be included in the 1130? – DavePhD Mar 08 '23 at 17:16
  • 22
    @DavePhD That's not so ridiculous if you think about drugs which lead to strong physical dependence - during withdrawal, improper care could lead to death from complications of vomiting and diarrhoea, for instance. As far as I know, cannabis generally doesn't cause physical dependence of that sort, so "lack of cannabis" is unlikely to be even a "contributing cause". – IMSoP Mar 08 '23 at 17:29
  • 11
    And it also doesn't distinguish between cannabis ONLY and cannabis "found in system". If someone died of a multi-drug OD, cannabis will end up part of the "contributing" even though someone ODing and die from it alone is almost unheard of. – Nelson Mar 09 '23 at 02:18
  • 9
    Note that the non-existence of an ICD-Code meaning 'underlying cause of death is a cannabis overdose' implies that the WHO, who administers these codes currently believes this is not even potentially a valid diagnosis. The list of existing ICD-Codes gets updated every year and if the WHO believed there could be cases where such a code would be appropriate, then it would exist. – quarague Mar 09 '23 at 07:34
  • @IMSoP cannabis has a strong sedation effect that certainly can be habit forming. – Neil Meyer Mar 09 '23 at 16:15
  • 2
    @NeilMeyer I specifically said "physical dependence of that sort" - i.e. the sort of dependence that produces potentially dangerous withdrawal symptoms such as vomiting and diarrhoea, so I'm not sure how "a sedation effect that can be habit forming" is relevant. – IMSoP Mar 09 '23 at 16:37
  • 1
    @IMSoP "... death from complications of vomiting and diarrhoea ..." - Interestingly, there also exist more direct methods of death from drug withdrawal. For example, alcohol withdrawal can cause seizures, which can cause permanent brain damage and even death. – marcelm Mar 09 '23 at 16:55
  • @IMSoP as I recall cannabis does cause physical dependency. That dependency is roughly the level of Caffeine, ie so minor as to be mostly safe to ignore; but still technically a dependency. – dsollen Mar 09 '23 at 17:14
  • 1
    Its dependancy is on par with valium or xanax. Not safe to ignore at all. – Neil Meyer Mar 09 '23 at 17:22
  • 5
    @NeilMeyer that's utter nonsense. Not even the same planet as those two. Withdrawal from long term use of benzos requires significant medical intervention and careful tapering, possibly even the use of other medications. Meanwhile Snoop Dogg could quit tomorrow and while it would be unpleasant, there would be no other medical issues. Benzos are also vastly more addictive for most people, with physical dependency possible in literal weeks time – eps Mar 09 '23 at 19:32
  • 2
    @Nelson Right... and the problem there is cannabis is detectable months after use - whereas hard drugs tend to be undetectable after a few days sober. So if they smoked a joint 3 months ago and die from an OD of something else, cannabis is going to pop up in the screen. – Harper - Reinstate Monica Mar 10 '23 at 21:47
  • X42 includes cannabis in the category "narcotics and psychodysleptics [hallucinogens], not elsewhere classified" – Tim Sparkles Mar 13 '23 at 22:26
  • T40 says "Excl. intoxication meaning inebriation (F10-F19)" so it seems that the presumption is that cannabis can cause poisoning. OTOH it doesn't seem appropriate to diagnose a corpse with "Mental and behavioural disorders due to use of cannabinoids(F12)" It is unclear how a diagnosing doctor would determine "poisoning" given detection of cannabinoids in the system, nor whether these codes are applied in a standardized way given what we know about how long cannabis remains detectable after use. – Tim Sparkles Mar 13 '23 at 22:40
10

If you "Download Table" from https://injuryfacts.nsc.org/home-and-community/safety-topics/drugoverdoses/data-details/ there's a bit more detail on where these data came from:

Deaths are classified using the International Classification of Diseases, 10th Revision. Preventable drug poisoning deaths are identified using underlying cause-of-death codes X40-X44, while all drug poisoning deaths are identified by codes X40-X44, X60-X64, X85, and Y10-Y14.

X40-44 are all "accidental poisoning" causes, each number is a different category of drug. X60-64 are "intentional self-poisoning" equivalents. X85 is "assault by drug", Y10-Y14 are the "undetermined intent" equivalent of X40-44/X60-64. See https://mnprc.org/wp-content/uploads/2019/01/using-icd-10-codes-to-assess-opioid-related-overdose-deaths.pdf

None of these codes identify cannabis specifically. That is, while you could categorize a cannabis overdose under one of these codes, you would not be able to identify it as a cannabis overdose from the code. For example, X62 includes cannabis but also cocaine and heroin. The only way you'd specifically identify a cannabis overdose from these ICD codes is from the T40 range:

Among deaths with drug overdose as the underlying cause, the following multiple cause-of-death codes indicate the drug type(s) involved: Any opioid (T4NR-T40.4, T40.6), heroin (T40.1), methadone (T40.3), synthetic opioids other than methadone (including fentanyl, T40.4), natural and semisynthetic opioids (including oxycodone and hydrocodone, T40.2), benzodiazepines (T42.4), cocaine (T40.5), and cannabis (T40.7). Deaths involving more than one opioid category are counted in both categories.

It seems likely that any time T40.7 is present, some other code is also present that indicates the "actual" cause of death.

Bryan Krause
  • 2,129
  • 18
  • 19
  • Is the raw data available that would allow us to go from "likely" to "definitely" on your conclusion? – IMSoP Mar 08 '23 at 16:27
  • 1
    @IMSoP Possibly but I didn't see an explicit reference in the source to their source data. What is definitive is that it is not possible to ascribe a single cause using these codes. – Bryan Krause Mar 08 '23 at 16:53
  • 1
    I'm curious what subset of those numbers include (X62|X42|Y12)+T40.7... that might be as close as we can get to "cannabis may have been a significant factor in or potential underlying cause of death" – Doktor J Mar 08 '23 at 21:04
  • Possible Raw Data: https://wonder.cdc.gov/mcd.html – Dancrumb Mar 08 '23 at 21:58
  • 5
    @DoktorJ The T-codes are not mutually exclusive, So one would have to look for some X/Y code for death by drug overdose + T40.7 + no other T4xxx code to get cases where cannabis is the most significant factor. – quarague Mar 09 '23 at 07:40
  • X42 includes cannabis in the category "narcotics and psychodysleptics [hallucinogens], not elsewhere classified" – Tim Sparkles Mar 13 '23 at 22:26
  • @TimSparkles Includes but is not specific for, so there's no way to separate cannabis from, say, cocaine. – Bryan Krause Mar 13 '23 at 23:31
  • @BryanKrause My comment was addressed at "None of these codes mention cannabis specifically." These codes do not specifically mention cannabis specifically as the cause of death, but they do specifically mention it as a member of the category of drugs which was the cause of death. The remainder of that paragraph does I think sufficiently explain this, but the thesis is not quite correct in itself. – Tim Sparkles Mar 14 '23 at 22:06
  • 1
    @TimSparkles I changed "mentions" to "identifies". I provide an example in the next sentence of what is meant by this. None of these codes are for cannabis specifically, so you cannot use them to identify cannabis overdose. – Bryan Krause Mar 14 '23 at 22:08