Published online 27 July 2007 | Nature | doi:10.1038/news070723-11


Medical opinion comes full circle on cannabis dangers

Frequent use more than doubles psychosis risk, says new large-scale analysis.

Cannabis may be responsible for 14% of UK psychotic illnesses.Cannabis may be responsible for 14% of UK psychotic illnesses.Getty

Frequent cannabis use more than doubles the risk of developing psychotic illnesses such as schizophrenia, according to the most rigorous analysis of the evidence to date.

The finding, which comes from a new study that combines results from 35 previous surveys, represents a significant U-turn from previous suggestions that cannabis is harmless to mental health. The analysis is published in medical journal The Lancet, which in 1995 began one of its issues with the sentence: "The smoking of cannabis, even long term, is not harmful to health."1

In fact, having used the drug even once increases your risk of developing psychotic problems by 41%, according to the new research2. This suggests that 14% of all psychotic illness in Britain is caused by cannabis use.

'Serious risks'

"The message that has to be made clear is there are potentially quite serious risks from using cannabis," says study author Stanley Zammit of Cardiff University, UK. "For psychotic outcomes there certainly is enough evidence to warn people of the risk."

“There is now sufficient evidence to warn young people that using cannabis could increase their risk of developing a psychotic illness later in life’”

Stanley Zammit, Cardiff University

Zammit adds that the new analysis is the "most thorough" to date. "This adds a certain robustness to the evidence," he says.

Previously there have been questions over whether cannabis actually causes psychotic illness, or whether people with psychiatric problems self-medicate with the drug. Although the new study cannot conclusively settle the argument, it offers the best evidence short of conducting a full randomized trial, in which participants are given either cannabis or a placebo.

Zammit and his colleagues combined data from 35 longitudinal trials, in which populations are observed over time. They found that, even after allowing for other factors, such as other substance use and intelligence, people who have taken cannabis are 41% more likely to develop schizophrenia or other psychotic problems than those who have never used it. Those who used cannabis most frequently were more than twice as likely to suffer problems.

It was less clear whether cannabis use was also linked to depression, suicidal thoughts or anxiety.

Drug law rethink

The increased risk of psychosis, particularly in long-term cannabis users, is worrying, despite the actual risk of developing schizophrenia only being 1% of the population overall3. "Those figures are frankly quite staggering," says Neil McKeganey, director of the University of Glasgow's Centre for Drug Misuse Research. "Because cannabis is our most widely used illegal drug, I think any increased risk we have to be greatly concerned with. One can't take any comfort in the fact they may be relatively small numbers."


The discovery will exacerbate the current controversy over Britain's cannabis laws. Several high-ranking politicians have recently admitted using the drug in the past. A review has also been announced of the 2004 change that saw cannabis downgraded from Class B to Class C, the least serious class of illegal drugs.

In an editorial4 accompanying the new study, Merete Nordentoft and Carsten Hjorthøj of Copenhagen University Hospital, Denmark, say: "In the public debate, cannabis has been considered a more or less harmless drug compared with alcohol, central stimulants [such as amphetamines], and opioids. However, the potential long-term hazardous effects of cannabis with regard to psychosis seem to have been overlooked."

Visit our opinioncomesfullcir.html">newsblog to read and post comments about this story.  

  • References

    1. Anon. The Lancet 346, 1241 (1995).
    2. Zammit, S. et al. The Lancet 370, 319-328 (2007).
    3. Schultz, S. et al. Am. Fam. Physician 76, 1821-1829 (2007).
    4. Nordentoft, M. et al. The Lancet 370, 293-294 (2007).