http://www.newscientist.com/channel/health/mg18524921.300I pulled the following story from newscientist.com only because I wasn't sure how long it would remain up.
"AT THE end of Jim van Os's street in the pleasant Dutch city of Maastricht there is a coffee shop. As with many such establishments in the Netherlands, "coffee shop" is something of a euphemism: most of its customers go there not to drink coffee but to buy and smoke dope. Van Os isn't too keen on the place. He doesn't like the shady characters it attracts. He doesn't like the fact that his children have to walk past it. And most of all he doesn't like that fact that the place breaks the law and sells marijuana to under-18s.
Van Os's fears are rooted in more than the usual parental angst. He is a psychiatrist at the University of Maastricht who investigates the effect of marijuana on people's brains - particularly adolescents' brains. And the findings of his research make him worry about the effects of all this dope smoking on the kids in his neighbourhood.
Over the past couple of years van Os and several others have been building the case that, for some teenagers, smoking cannabis leads to serious mental health problems in later life, including schizophrenia. Van Os claims that marijuana is responsible for up to 13 per cent of schizophrenia cases in the Netherlands. And with cannabis use among teenagers on the rise, the age at first use falling (see Graphic), and the strength of cannabis on the up, he says the figure can only increase.
It is a frightening conclusion, and one that is already starting to shape the debate over the legal status of cannabis. In the UK, van Os's findings have been seized upon by politicians, tabloid newspapers and mental-health lobby groups who want drug laws tightened up. But their case is far from made. Some researchers say there are potentially fatal flaws in the research, and that it would be a serious mistake to change the law based on an as-yet unproven theory. So who is right? The answer, it turns out, is more complex than it first appears.
Compared with substances like heroin and crack cocaine, cannabis is seen by many people as relatively harmless. Several European countries take a lenient view of its use, and in the UK marijuana was downgraded from a class B to a class C drug last year, meaning people caught with small quantities are not usually arrested.
But doctors have long known that taking a lot of cannabis over a short period can occasionally cause temporary symptoms of psychosis, one of the hallmarks of schizophrenia (see "Psychosis explained"). The question of whether the drug has long-term mental health consequences, however, is much more contentious.
The first hints of possible long-term harm came from Jamaica in the 1960s, where doctors noticed that Rastafarians who were heavy dope smokers seemed unusually prone to psychotic illnesses. The suspicion grew in the 1980s, when studies led by Robin Murray of the Institute of Psychiatry in London showed that patients with schizophrenia or some other kind of psychosis, were twice as likely as those without mental illness to be regular cannabis users.
But it wasn't until a major Swedish study was published in 1987 that a link between cannabis and long-term mental health problems was established. A team from the Karolinska Institute in Stockholm analysed the records of all the young men who had done national service in the Swedish army in 1969 and 1970 - 50,087 in total, representing about 97 per cent of the male population aged 18 to 20. The records included details of the men's experiences with drugs before being called up, including age at first use, what they had taken and how often.
The team then checked out each man's medical history up to the mid-1980s. They found that those who had smoked cannabis before being called up were six times as likely to end up in hospital with schizophrenia as non-users. This, the team concluded, provided clear evidence that smoking cannabis was a risk factor for psychosis.
Case closed? Not at all. The study certainly showed a correlation between cannabis use and psychosis, but, as many people pointed out, it by no means proved that the former caused the latter. For one thing, the analysis ignored several possible "confounding factors" - other differences between users and non-users of cannabis that could be the real explanation for the link with schizophrenia. For example, dope smokers could plausibly be more likely to take other drugs such as LSD, which is a hallucinogen. Perhaps it was these drugs that were actually causing schizophrenia. Another possibility was that there was some hidden factor - possibly genetic - that predisposed people both to smoking cannabis and developing psychosis.
More importantly, even if there were a causal relationship between cannabis use and psychosis, the researchers failed to establish in which direction it lay. Sure, it was possible that cannabis led to schizophrenia. But it was just as likely that schizophrenia was causing people to smoke cannabis. Perhaps they were "self-medicating" - taking the drug to alleviate their symptoms.
The self-medication hypothesis has proved a popular one, and was virtually orthodoxy in the 1980s. For one thing, patients with psychosis often report that cannabis makes them feel better. And recent research suggests that one of the chemical ingredients of cannabis, cannabidiol, exerts a powerful antipsychotic effect. Indeed the UK firm GW Pharmaceuticals, which is awaiting approval for an under-the-tongue cannabis spray for multiple sclerosis patients, plans to investigate cannabidiol as a treatment for schizophrenia.
But the idea that cannabis caused psychosis didn't go away. One prominent sceptic of the self-medication idea was Don Linszen, a psychiatrist at the University of Amsterdam, the Netherlands, who, as a mental-health practitioner in the cannabis capital of Europe, was in daily contact with dope-smoking schizophrenia patients. He noticed that the patients who used cannabis usually fared much worse than those who didn't. Studies by Murray and others confirmed Linszen's observations. If people with psychosis were self-medicating, it wasn't working very well.
“The more cannabis the young people smoked, and the earlier they smoked it, the worse their risk of psychosis”Meanwhile, several research groups were busy setting up long-term studies that followed large groups of young people over many years, keeping track of their drug use as teenagers and testing them for signs of psychosis later in life. They were similar in principle to the Swedish study, but were carefully designed to avoid its mistakes, particularly on the question of self-medication.
In the past couple of years results from a number of these studies have been reported, and together they make a decent case that the Swedish research got it right first time. "I don't think we can deny it any longer," says epidemiologist Mary Cannon of the Royal College of Surgeons in Ireland, based in Dublin. "Cannabis is part of the cause of schizophrenia."
Cannon helped carry out a study that followed 759 people born in Dunedin, New Zealand, in 1972 and 1973 (British Medical Journal, vol 325, p 1212). After carefully controlling for self-medication and other confounding factors, the researchers found that those who had smoked cannabis three times or more before the age of 15 were much more likely to suffer symptoms of schizophrenia by the time they were 26 - they had a 10 per cent chance compared with 3 per cent for the general population. The team concluded that there is a vulnerable minority of teenagers for whom cannabis is harmful. "We're not saying that cannabis is the major cause of schizophrenia," says Murray, who led the study. "But it's a risk factor."
Similar results have recently come in from a Greek study as well as a re-analysis of the original Swedish research. And late last year van Os and his team brought out further results. They followed a group of nearly 2500 14 to 24-year-olds living in and around Munich, Germany, over four years. After correcting for all the confounders they could think of, they found that, overall, smoking cannabis as an adolescent moderately raised the risk of developing signs of psychosis later on, from 16 per cent to 25 per cent. But when they focused on individuals who were known to be susceptible to psychosis - those who were showing signs of disturbed thought processes by age 11 - they found a much stronger link. Susceptible individuals who avoided cannabis had a 25 per cent chance of developing psychosis. Susceptible individuals who smoked it had a 50 per cent risk. And the more cannabis they smoked, and the earlier they smoked it, the worse the outcome.
According to van Os, Murray and others, there can no longer be any doubt that there is a small but significant minority of people who have a predisposition to psychosis and who would be well advised to steer clear of cannabis.
This message is already starting to filter out into society. In the Netherlands, the findings have fuelled a growing clamour for reform of the laws regulating drug use. In the UK, the mental-health charity Sane has called for the reclassification of cannabis to be reversed. And the British government recently acknowledged the link in its strongest terms yet, when it said in a press release that cannabis was an "important causal factor" in mental illness.
But for some researchers, such pronouncements are premature. "I'm not convinced," says Les Iversen, professor of pharmacology at the University of Oxford and a member of the UK Home Office's Advisory Council on the Misuse of Drugs. "I think the jury is still out on this one."
“Despite a steep rise in cannabis use among teenagers over the past 30 years, there has been no rise in schizophrenia”He points out that epidemiological studies are notoriously bad at proving cause and effect, in part because it is hard to identify all the confounding factors. Scientists are particularly wary of such research when the conclusions are based on small statistical differences - as in this case. In the New Zealand study, the number of people who had smoked dope on three occasions by the age of 15 was just 29, and only three went on to develop psychosis. "I can't help thinking that the conclusion is rather thin," says Iversen. "It makes you wonder. If they found another confounder, where would that leave them?" Van Os himself admits that his study does not eliminate all the confounding factors.
Iversen also points out problems with how psychosis is defined in each study. Van Os's results, for example, seem to show that psychosis is relatively common in cannabis smokers. But he chose a very broad definition - a single psychotic symptom, such as hearing voices or a paranoid belief. That's not the same as schizophrenia. About 20 per cent of the general population live with these symptoms: only 1 in 20 of them ever require treatment.
Iversen isn't the only doubter. Last year John Macleod of the University of Birmingham, UK, carried out a systematic review of the 16 high-quality long-term studies that have looked at cannabis and mental health and concluded that there was "no strong evidence" of a link with psychosis (The Lancet, vol 363, p 1579). "I'm not saying there is no causal association, but at the moment, by the conventional standards of epidemiology, the evidence is not particularly strong," he says.
There are other reasons to doubt the reality of a link. If it were genuine, schizophrenia should be becoming more common as teenage cannabis consumption goes up. But it isn't. In 2003 researchers at the University of New South Wales in Sydney found that, despite a steep rise in cannabis use among Australian teenagers over the past 30 years, there had been no rise in the prevalence of schizophrenia.
In the genes
The only study to find a rise in schizophrenia was carried out in the inner London district of Camberwell, where continual demographic changes make interpreting the result almost impossible.
It seems the question of whether cannabis use can indeed lead to schizophrenia will be unresolved for some time yet. One research avenue that may shed further light on the matter is to look at whether genes are involved. One of the members of Cannon's group, Avsahlom Caspi of King's College London, is an authority on gene-environment interactions. A couple of years ago he made a splash by showing that a genetic predisposition to depression could interact with a traumatic experience such as bereavement to trigger the illness (New Scientist, 26 July 2003, p 15). "We thought, let's put this model to work on cannabis and psychosis," says Cannon.
She and her colleagues reanalysed the data from the New Zealand study, this time adding in another variable - genetic predisposition to schizophrenia. The gene they investigated, called COMT, encodes an enzyme (catechol-O-methyl transferase) that breaks down a signalling chemical in the brain called dopamine. COMT comes in two forms, one of which is marginally more common in people with schizophrenia and is thought to be a risk factor for the disease.
The results were crystal clear. The team found that in people with two copies of the "normal" version of COMT, smoking cannabis had little effect on their mental health. In people with one normal and one "bad" form of the gene, smoking cannabis slightly increased their risk of psychosis. But for people with two copies of the bad gene, cannabis spelled trouble: smoking it as a teenager increased their likelihood of developing psychosis by a factor of 10. The results have not yet been published, and Cannon warns that they need replicating, but even so she says "this is a very large effect, similar to the size of smoking and lung cancer. This is a very significant finding." Case, perhaps, closed.
What should be done about it, however, remains an open question. Van Os advocates that teenagers with a personal or family history of mental illness be urged to steer clear of the drug. He also advocates legal changes: governments should focus on keeping cannabis out of the hands of teenagers and outlawing extra-strong varieties of cannabis, such as skunk and white widow.
For Iversen though, this is still stretching the evidence. "What the data show is that the risk applies to a small minority of young people who start smoking cannabis at a very young age," he says. "Are we going to change the law for the benefit of a vulnerable minority? A small minority of people are vulnerable to liver damage if they drink even a small amount of alcohol, but we haven't changed the law to protect them." "
From issue 2492 of the brilliant "New Scientist" magazine, 26 March 2005, page 44
