Alcohol or cannabis: which is more harmful? The numbers

Before comparing two substances, you have to decide what to compare. The most cited method is multi-criteria decision analysis, in which experts score each drug on a list of harms and then weight the criteria. In 2010, David Nutt's British committee scored 20 drugs on 16 criteria: nine concerning the user and seven concerning other people and society. Alcohol ranked first with an overall score of 72 out of 100, ahead of heroin (55) and crack cocaine (54). Its lead comes mostly from harm to others, where it scored 46 points for injury, crime, economic costs and family adversity.

In 2015, 40 experts repeated the exercise for the European Union and reached almost the same result: alcohol scored 72, heroin 55 and crack 50, and no other drug, cannabis included, went above 38.

These rankings are expert judgements, and that is their limit: harm to society also depends on how many people use a substance. Because alcohol is so widespread, its collective costs rise mechanically. Other measures are needed to know what a substance does to the person taking it.

Toxicology takes a more detached view with the margin of exposure: the dose that becomes dangerous in animals divided by the dose a person actually takes. A smaller number means ordinary use sits closer to the toxic threshold. In 2015, Dirk Lachenmeier and Jürgen Rehm calculated this margin for alcohol, tobacco, cannabis and the main illicit drugs. For an individual drinker, alcohol falls below 10, in the "high risk" category alongside nicotine, cocaine and heroin. At population level, alcohol is the only substance still in that category. Cannabis exceeds 10,000, well beyond every other substance studied.

The measure is limited to acute toxicity. Accidents, mental health problems and diseases that take years to develop are left out, and for alcohol those diseases count for a lot. The World Health Organization classifies alcohol as an established carcinogen and attributes 2.6 million deaths worldwide to it in 2019. In France, Santé publique France counts 41,000 deaths a year linked to alcohol, the third cause of preventable death after tobacco and air pollution.

Both products can lead to dependence, at different speeds. The reference study remains the 2011 paper by Catalina Lopez-Quintero and colleagues, based on NESARC, a large US general-population survey. The authors estimated how likely a person who has started using is to become dependent one day.

Two points emerge from these figures. A drinker is about 2.5 times more likely than a cannabis user to become dependent. When cannabis dependence does develop, it comes quickly: half of cases appear within 5 years, against 13 years for alcohol. The OFDT, the French observatory on drugs, notes that problematic cannabis use and dependence are well documented and that early use raises the risk.

Mental health is where cannabis carries its main specific risk. The EU-GEI study, published in 2019 in The Lancet Psychiatry, compared 901 people after a first psychotic episode with 1,237 controls at 11 sites in Europe and Brazil. Daily cannabis use multiplied the risk of a psychotic disorder by 3.2 compared with people who had never used it. For products estimated at 10% THC or more, the factor reached 4.8. In Paris, one of the three sites where such products were most common, it was 4. Our article on cannabis and mental health covers this research in detail.

According to the OFDT, use can also lower memory and learning capacity in the medium term. How cannabis is taken counts as well: smoked with tobacco, it brings the harms of combustion and nicotine on top of its own. Our article on tobacco and cannabis goes further into that comparison.

Beyond "which one", the practical question is often "both at once", and the road gives the clearest answer. The ActuSAM study, published in 2016 by the French national road safety observatory (ONISR), examined the fatal crashes of 2011 in France.

Testing positive for cannabis multiplies a driver's risk of being responsible for a fatal crash by 1.65. Drink-driving multiplies it by 17.8 on average, ranging from 6.4 between 0.5 and 0.8 g/l to 44.4 above 2 g/l. For a driver positive for both, the factor is 29. Among responsible drivers under the influence of cannabis, more than one in two had also been drinking.

French law applies the same reasoning. The official service-public.fr page states that driving after drug use is punishable by 3 years in prison and a €9,000 fine, whatever the quantity. If alcohol is also involved, the penalty rises to 5 years and €15,000. Our article on CBD and driving explains why a full-spectrum oil can be enough for a positive test.

CBD sold legally in France falls outside this comparison. It comes from hemp varieties with no more than 0.30% Δ⁹-THC and has none of the psychoactive effect of THC; our article on what CBD is sets out that framework. Every study cited above concerns THC-rich cannabis.

CBD still calls for some precautions. It interacts with some medicines, and full-spectrum products contain traces of THC. Trials on using CBD to drink less remain few and small; our overview of CBD and alcohol sums up what has been measured. CBD is not a treatment for dependence.