Does long-term heavy marijuana use produce toxic effects on the brain?


March 05, 2002

Public Affairs

Belmont, MA - Does long-term heavy marijuana use produce toxic effects on the brain? This is a very critical question, and the study by Solowij and her colleagues in the March 6 issue of the Journal of the American Medical Association tries hard to give us some answers. However, it is extremely difficult to perform reliable studies of heavy marijuana users out of the community, because of what scientists call "confounding variables" - external factors that can distort the results.

Comments from Harrison G. Pope, Jr., director of the Biological Psychiatry Laboratory at McLean Hospital, regarding his editorial in the March 6 issue of JAMA

Suppose we find-as this study did-that long-term heavy marijuana users do more poorly on memory tests than "control" subjects who have used little or no marijuana. How do we know that this difference is attributable to marijuana as opposed to some other factor? Maybe the long-term heavy marijuana users had poorer memories to start with, before they ever began smoking marijuana. Alternatively, the heavy marijuana users may have used substantial amounts of other drugs, such as alcohol or cocaine, and it might be this other drug use, rather than marijuana, that has caused brain toxicity. Another possibility is that heavy marijuana users coming for treatment might be more likely to be suffering from psychiatric problems, like anxiety or depression, as compared to the "control" subjects-and it may be the anxiety or depression that is impairing their memories, rather than their marijuana use. Perhaps some of these heavy marijuana users are taking medications, such as benzodiazepines (drugs in the "Valium" family) to treat their depression or anxiety-and these medications are influencing their performance on memory tests, rather than marijuana use.

Solowij and her colleagues used statistical methods to try to deal with some of these differences, but they did not deal with all of them. The marijuana-using group in their study was composed of people who were coming to clinics seeking psychiatric treatment for their marijuana dependence, whereas the "control" group was composed of ordinary people were not seeking any sort of treatment. But the investigators did not compensate for the possibility that the treatment-seeking marijuana users might have had anxiety disorders, depressive disorders, or use of prescription medications that might have impaired their test performance. Also, 47 percent of the heavy marijuana users had a history of regular use of, dependence on, or clinical treatment for drugs other than marijuana-but many of the analyses in the study did not adjust for this fact, either.

Then there are still the problems of "matching" the marijuana users and the control subjects on other variables. To illustrate, suppose we did a study comparing the average height of 50 people in Seattle with the average height of 50 people in Minneapolis. Our sample in Seattle consists of 30 men and 20 women, whereas the sample in Minneapolis consists of 20 men and 30 women. We find that the average height of people in Minneapolis in shorter than in Seattle. Is this conclusion valid? Of course not: men are slightly taller than women, on average, and the Seattle sample is weighted in favor of men while the Minneapolis sample is weighted in favor of women. Therefore we have to "adjust" for the different sex distribution in the two samples. Once we do this adjustment, will probably find that there is no difference between Seattle and Minneapolis at all. In the same way, investigators comparing marijuana users with controls have to "adjust" for differences in sex distribution, age, and other variables between the two groups. However, Solowij and her colleagues did not adjust for gender, and only partially adjusted for age, in the comparisons between heavy marijuana users and controls, despite the fact that the heavy marijuana users were much older, and included more males, than the control group.

These examples give some idea of how hard it is to test whether long-term marijuana use impairs cognitive functioning. With all of the "background noise" being caused by confounding variables, such as psychiatric disorders, medication use, prior drug use, differences in sex, differences in age, etc., how can we be sure that the differences that we are seeing are attributable to marijuana, rather than something else? For these reasons, the safest thing to say at this point is that the jury is still out on the question of whether long-term marijuana use causes lasting impairment in brain function.

Harrison G. Pope, Jr., director of the Biological Psychiatry Laboratory at McLean Hospital, regarding his editorial in the March 6, 2002, issue of JAMA.

Email this page