A large new study cuts through the fug.
Everyone who confuses correlation with causation ends up dead.
Everyone knows this line, yet plenty of people – including in the healthcaring professions – never truly absorb the fact that A may not be the reason for B despite being associated with it.
Take cannabis and psychosis, a correlation that makes this otherwise relatively benign drug as suspect to many doctors as garlic is to vampires.
Disclosure: your scribe is not personally a big fan of the weed, having spent too many chunks of her youth paranoid, red-eyed and couchbound courtesy of some vile hydroponic moonshine, and still avoids it despite the influx of pleasanter strains from the US, with their less-than-medical-sounding names like Gorilla Glue and Girl Scout Cookies.
Even back then, given that some people loved and could barely live without the stuff, it seemed obvious that we must have different default settings to have such different experiences from the same substance.
A new study in JAMA Psychiatry pits three hypotheses about the nature of the link between adolescent cannabis use and psychosis: contributing risk, i.e. cannabis causes psychosis by disrupting neurodevelopmental processes; shared vulnerability, i.e. the starting conditions make someone both more liable to have psychosis and to want to consume cannabis; and self-medication, in which cannabis is used to treat the early symptoms of psychosis and associated distress.
Each has some evidence, the authors write; but astoundingly, most of the handful of longitudinal studies in adolescents “did not test whether cannabis use preceded symptom changes or vice versa”.
Using data from the Adolescent Brain Cognitive Development Study, a large longitudinal study of child and adolescent brain development, with a cohort of nearly 12,000 children aged an average 9.5 years at baseline.
To test for contributing risk, they looked at whether psychosis spectrum symptoms increased after cannabis initiation (adjusted for a trend decrease in symptoms); for shared vulnerability, they looked at symptom differences differed between adolescents who did and did not use cannabis; and for self-medication, they tested whether psychosis spectrum symptoms increased leading up to the initiation of cannabis use.
They found that, consistent with the shared vulnerability hypothesis, those who used cannabis at some point during the study period had more psychosis spectrum symptoms and greater distress from symptoms than those who didn’t.
Consistent with the self-medication hypothesis, the number and distress from symptoms did increase leading up to cannabis initiation. After initiation, symptoms did not go up but distress did go down in the short term, before increasing later.
The authors hasten to add that “this finding does not indicate that adolescent cannabis use is an effective means for reducing psychosis spectrum symptoms”.
What they did not find evidence for in the primary analyses was the contributing risk hypothesis, “as psychosis spectrum symptoms generally did not increase after cannabis initiation”, though in sensitivity analyses the results were a bit more mixed.
The authors say their results, while they support the non-causal hypotheses, don’t exclude some causal element and that an integrated hypothesis may be best: “Specifically, shared risk vulnerabilities and attempts to self-medicate symptom-related distress may lead to initial cannabis initiation in adolescence, whereas subsequent increases in the frequency and quantity of cannabis use throughout adolescent development may then contribute to psychosis onset in young adulthood.”
Still, it supports our hunch about starting conditions and suggests if a kid is hitting the bong, it could be a sign something is amiss rather than a crime to be punished.
Send Sour Diesel and story tips to penny@medicalrepublic.com.au.
