What cannabis / psychosis link?

I was planning on busting out a quick response to the recent news of an apparent cannabis / psychosis link, noting that meta-analysis is fraught with peril - in fact, the last round of Ecstasy toxicity scare stories were based on a flawed meta-analysis. But our own Omgoleus noted in a comment to the last post:

Note that the only thing new here is a meta-analysis, where researchers (possibly feeling lazy) go over the results of all the previous studies in a field to see what they say collectively, rather than doing any new work.

This has the effect of improving statistical power (because you effectively have a big new sample composed of all the people in all the studies) but most definitely does NOT solve any methodological problems or limitations of the original study.

So, if there were any biases in the original studies, the biases are still there in the meta-analysis. Caveat emptor, as the somewhat-misapplied saying goes.

It's not just that they round up old studies and attempt to come up with new conclusions. It's that they can also pick and choose which studies to include, thus skewing the conclusion toward whatever bias they might have, which clearly happened in the past. Now I'm not saying that's what happened this time, although the article does note the meta-analysis did exclude subjects who were already showing signs of psychosis at the start of the studies in question. But without examining the entirety of the literature, it's hard to say. Luckily for us, Dominic Holden over at the SLOG has already dug a little deeper, and noted:

By reading these headlines, you would believe recent findings show conclusive evidence that marijuana use leads to increased risk of psychoses. You would be wrong. The report is actually based on a meta-analysis of previously debunked studies, and the findings, as the article’s text immediately volunteers, are completely ambiguous.

Indeed, take a look at this very recent study, via a link that Holden provided:

Marijuana use is not associated with heightened symptoms of schizophrenia, according to data to be published in the journal Schizophrenia Research.

Investigators at London’s Institute of Psychiatry assessed whether the prior use of cannabis in patients with schizophrenia was associated with appreciable changes in schizophrenic symptoms compared with patients who had no history of marijuana use. Researchers performed logistic regression analysis on 757 volunteers with cases of first onset schizophrenia. Of these, 182 (24 percent) had reportedly used cannabis in the year prior to diagnosis, while 552 (73 percent) had not. (The remaining three percent had no data available.)

Investigators reported no statistically significant "differences in syptomatology between schizophrenic patients who were or were not cannabis users" after controlling for patients’ age, sex, and ethnicity.

Researchers also failed to find "any evidence that cannabis users with schizophrenia were more likely to have a family member with the disorder."

These findings "argue against a distinct schizophrenic-like psychosis caused by cannabis," authors concluded.

The study is the largest trial ever conducted to compare cannabis using and non-using schizophrenic patients, investigators said.

What seems to be clear is that:

Although investigators did not assess whether cannabis consumers had greater odds of contracting schizophrenia compared to those who did not have a history of smoking pot, prior reviews have downplayed such an association. Most recently, Britain's Advisory Council on the Misuse of Drugs (ACMD) concluded in 2006, "For individuals, the current evidence suggests, at worst, that using cannabis increases lifetime risk of developing schizophrenia by one percent."

A separate 2006 report by Britain’s Beckley Foundation speculated that cannabis may "precipitate schizophrenia in people who are already vulnerable" to the disease, but it also acknowledged that the "increased rates of cannabis use in the last thirty years have not been accompanied by a corresponding increase in the rate of psychosis in the population."

Even the authors of the meta-analysis under critique here are clear that:

There could be something else about marijuana users, “like their tendency to use other drugs or certain personality traits, that could be causing the psychoses,” Zammit said.

And finally, as Holden points out:

Is this not obvious? If this study shows anything, it’s that psychotics are smoking pot to chill the fuck out.

But it's not clear that this does study does show anything, without carefully examining all "35 studies on the long-term effects of cannabis use in Europe, the US and Australasia" to determine if, as Omgoleus describes above, any of these studies have "methodological problems or limitations," the most common of which in studies like this is the inability to control for any other kind of drug use or environmental issue that may have an impact on a person's mental state. Given that subjects in these studies are typically self-reporting users of illicit substances (i.e., no one is taking a pool of subjects off into a private facility, and then dosing them with medical-grade marijuana and nothing else), these studies are very difficult to take at face value.

That said, if anyone out there actually has read these same 35 studies - while at the same time examining the literature to make sure any studies that don't support the meta-analysis (like, oh, gee, the one that Holden pointed us to) weren't excluded - feel free to ping us with your thoughts.