June 13, 2008

From Sander Greenland:

Re: Brain abnormalities seen in heavy pot smokers (June 2, 2008): As a researcher in this topic area, I have sent a letter to the Archives of General Psychiatry regarding the small epidemiologic survey by Yucel and colleagues(1). This study has generated considerable media coverage, including at your own website. I was disappointed by your coverage, which largely ignored many problems that could nullify the validity and generalizability of the study. As I wrote to the Archives: 1) The study failed to control for potentially important differences between the groups. The study did note much more tobacco use among the cannabis users, which along with their heavy cannabis use indicates unhealthy behavior patterns among the users relative to the controls.

The study also failed to note there was also more alcohol use among the cannabis users. They may have failed to note and control for this difference because it was not “statistically significant” -- a fact that does not lessen the potential confounding effects of alcohol use(2). This failure means that some of the difference they report may be due to co-use of alcohol and poorer health behaviors among the users.

2) The users in the study were self selected for a rather extreme behavior (heavy cannabis use for decades), and also for volunteering for the study. It remains possible that pre-existing conditions (including the observed anatomical differences) led to the preference for heavy cannabis use. It is also possible that there was selective nonparticipation by professional and anatomically normal users who would not want to reveal a long-term illegal and socially unaccepted behavior.

Self-selection could thus have acted to create an unusual population of heavy users, and nonparticipation could create a sample further weighted toward the lower functioning and anatomically unusual end of the spectrum.

3) The 15 cannabis users in the study all had used more than 5 joints a day for more than 10 years. Most epidemiologic research has been among occasional users (e.g., Hashibe et al.(3)). Among them, joints have typically been shared or extinguished to be used repeatedly, and thus represent multiple doses, sometimes 2-4 doses. In the latter case the users in the Yucel et al. study may have been consuming over 4x5 = 20 doses per day for over 10 years, with a mean of 20 years. If a single dose were the analogue of (say) a 4-oz. glass of wine, the users in this study were consuming an analogue of over 80 oz. of wine a day, or over 3 standard (750ml) bottles a day for an average of 20 years. Thus the subjects in this study would be the cannabis equivalent of severe very long-term alcoholics, who also show extensive brain abnormalities (as well as liver and other alcohol-related disease).

The above problems are not mutually exclusive and may all have been acting in the Yucel et al. study. Thus, while further study is certainly warranted, the study is at best a very tentative and limited result which may have no relevance for the vast majority of users, and may not even reflect genuine effects in heavy users.


1. Yücel M, Solowij N, Respondek C, Whittle S, Fornito F, Pantelis C, Lubman DI. Regional brain abnormalities associated with long term heavy cannabis use. /Arch Gen Psychiatry. / 2008;65(6):694-701.

2. Rothman KJ, Greenland S, Lash TL. /Modern Epidemiology/, 3rd ed. Philadelphia: Lippincott-Williams Wilkins, p. 261.

3. Hashibe M, Morgenstern H, Cui Y, Tashkin DP, Zhang Z-F, Cozen W, Mack TM, Greenland S (2006). Marijuana use and the risk of lung and upper-aerodigestive tract cancers: results of a population-based case-control study. /Cancer Epidemiology Biomarkers and Prevention/, *15*, 1829-1834.

Sander Greenland, Professor
Departments of Epidemiology and Statistics
UCLA School of Public Health


Post a Comment

<< Home