Refracting Green

Kevin Sabet appeared at 2pm today before the House/Senate joint judiciary committee to present his anti-marijuana legalization argument (his trip and presentation were paid for by the Oregon Narcotics Enforcement Association–highly ironic considering his persistent railing against the “big marijuana lobby”).  I tried to talk to him after the presentation, but he essentially flipped me the bird and ran out of the capital.  Very professional.  During this presentation, he claimed (among other things) that:

(1)  adolescent marijuana use rates have increased in states where medical marijuana is legal (he used data from a single state–Colorado–to support this claim),

(2)  marijuana use rates will increase in states that legalize adult recreational use of the drug

(3)  1 in 6 adolescents who try marijuana will become addicted

(4)  marijuana use is associated with declines in IQ

All of the above points are demonstrably invalid; furthermore, they are the result of an incredibly narrow and pernicious reading of the collected scientific evidence.  My claim—that Dr. Sabet is willfully misleading his audiences while presenting these claims across the nation—is not new or novel, but my presentation of evidence demonstrating his inaccuracies is.

First, adolescent marijuana use in the United States has traditionally followed what is known as a “cohort effect” over time; what this means is that rates of use by youths has gone both up and down at various times over the past 35 years, but the rates of use between states at any given time are fairly similar.  Most recently, youth use declined in the US from 1999 until 2007, when it started to trend back up (see Figure 1).

Figure 1. 

Source:  Anderson, Hansen, and Rees 2012

What is interesting about this national trend is that certain states have experienced slower rates of growth in youth usage—contrary to Dr. Sabet’s hypothesis, these states have medical use provisions (Anderson, Hansen, and Rees 2012–link here).  This deserves restating:  states with medical marijuana laws have slowed the national trend of increasing adolescent marijuana use.   Dr. Sabet presented data from Colorado—which has, like all other states in the US, seen an increase in adolescent marijuana use in recent years—without acknowledging this nationwide trend or the research demonstrating that his claim is false and unfounded.

Second, marijuana use is likely to increase in states that legalize recreational use, but only for a few years and, in all likelihood, not in significant proportions.  Policy experts often forget that marijuana is a niche drug and that only a small percentage of any particular population seems to enjoy its effects—it’s legal status is not likely to impact this phenomenon.  The Netherlands and Portugal offer the most illuminating example for states considering legalization of recreational use.  In the Netherlands, marijuana use was legalized (de facto) in 1972; Portugal decriminalized possession up to 28g in 2001.  In both nations, use of the drug increased slightly for 2-3 years following legalization, then plummeted.  Today, the lifetime use rates (i.e. percent of population over the age of 12 who have used the drug at least once) of both the Netherlands (19.8%) and Portugal (10%) are significantly lower than the US (42.4%).  No other nation in the world has higher rates of lifetime use than the US—this should be an indication that our approach to this drug has failed.  If the outcome is similar in the US (and we have no reason to believe that it would not be), legalization will result in lower lifetime use rates.  An important side note:  41% of Dutch teens report that marijuana is easy to obtain in their schools, while 75% of US teens report the same (MacCoun 2011).

Third, do 1 in 6 adolescents (16%) who try marijuana become addicted?  This claim requires some serious twisting of the data.  First, marijuana is not physically addictive in a classical sense; to receive the label of “physically addicting,” the substance must cause physical withdrawal symptoms upon cessation of use.  Many addiction experts now agree that a substance does not need to produce physical symptoms to qualify as “addictive”; instead, users who cannot stop using a drug while it is deleteriously affecting aspects of their lives are considered “dependent”.  Using this definition, approximately 8% of regular users can be categorized as “dependent” on the drug.  It takes a wild stretch of the imagination (and no supporting evidence) to suggest that 16% of all kids who try marijuana become addicted to it.

Finally, does marijuana negatively affect a person’s intelligence?  Dr. Sabet claims lower IQ scores result from using marijuana; however, experts on intelligence testing long ago threw out the archaic notion that this complex concept can be captured in a number (Gould 1981).  Anyone who makes a claim about “IQ scores” today is about 35 years behind the advances of modern science.  The most recent study cited by Sabet is Meier et al.’s (2012) longitudinal research on marijuana use and IQ scores; despite statistically significant correlations between IQ score declines and long term use originating during early adolescence a, the authors of that study are very careful to note that “adult-onset cannabis users did not appear to experience IQ decline as a function of persistent cannabis use” (5).  If IQ scores were a valid metric for intelligence or cognitive functioning, this would tell us that we need laws in place that reduce adolescent exposure to marijuana (which the current prohibition has definitively not accomplished).

More sophisticated versions of this argument examine brain activity in mice; the most recent and widely cited version of this approach (Raver, Haughwout, and Keller 2013) attempted to simulate chronic exposure to marijuana in adolescents by subjecting mice to an intense regime of drug use.  The mice were developmentally equivalent to 13 year olds at the time of exposure and the experiment ended when they were equivalent to 15.5 years old.  The results suggested that chronic marijuana use leads to a modest but statistically significant 30% decrease in cortical oscillation activity.  This study’s methods are more than a bit controversial though:  to achieve this 30% decline in cortical oscillation activity, researchers had to expose the mice to the human equivalent of 100 joints per day for 2.5 years.  The marijuana equivalent consisted of pure THC (one of 483 active chemical constituents in marijuana) and WIN55-212, an incredibly potent, synthetic version of THC (both of these are problematic, as other active cannabinoids present in marijuana flowers are THC agonists and neuroprotectants).  Even more problematic was the delivery of the marijuana equivalent, where researchers had to sedate the mice and inject THC or WIN55-212—to do so, they exposed the animals to a mix of ethanol and ketamine.  Both of these drugs have been shown to decrease cortical oscillations on their own and severely call into question the face validity of the study.  A number of other studies have demonstrated that marijuana use leads to minor, but reversible short-term cognitive impairment (even with heavy chronic use).

Closing note:  the science on marijuana is woefully inadequate, but we do know some things–and what we know directly contradicts what Kevin Sabet presented today.