12 December 2002

No evidence for cannabis 'gateway' causation of hard drug use.

Reassessing the marijuana gateway effect. Morral AR, McCaffrey DF, Paddock SM. Addiction (2002) 97;1493-1504

Dear Colleagues,

These authors use a meticulously designed formula for detecting whether household survey figures of drug use can be explained without using the a proposed effect that previous cannabis use makes an individual more likely to use other illicit drugs (the 'gateway' theory). The authors make a simple assumption about drug use behaviour and then look to see if the comprehensive age-group, annual data available can be made to fit without a 'gateway' or 'enhancing' factor involving prior cannabis use. This exercise reminds me of the elegant mathematical demonstration that the square root of two is an irrational number and cannot be represented by a fraction such as p/q where p and q are integers.

For the exercise, the authors assumed that each individual has a given propensity to try drugs to which they are exposed at a given age (this survey only looked at "ever tried" criteria). Thus some were quite vulnerable to drugs they came across while others would be less likely to use them. They assumed further that such individuals would be equally likely to use any drug they came across (ie. that there was no such thing as the 'gateway' theory) and then compared their calculations with the comprehensive drug prevalence figures together for each age 12 to 22 years from the US National Household Survey of Drug Abuse (NHSDA).

After complex statistical computations for both cannabis and 'hard' drugs, the authors conclude that the drug use prevalence figures are consistent with their assumptions, needing no 'priming' or 'gateway' effect for the increases seen in drug use prevalence from the household surveys. Rather than claiming that this disproves the 'gateway' theory (which remains to be supported by any specific evidence to my knowledge), the authors are very modest in their conclusions. They state that a gateway 'priming' effect may still exist but that their figures do not support it. It would appear that if such a gateway effect did in fact exist, that there would have to be an equal and opposite effect keeping the numbers using hard drugs constant. This latter would require some subjects to have a decreased propensity or 'immunity' to drug use. Such an interpretation would be curious or even spurious and thus, the reader is left to seriously question the gateway theory based on these data and the original assumption.

In this study, perhaps for the first time, the 'gateway' theory of enhanced vulnerability has been put to scientific scrutiny and found wanting.

Starting a series of commentaries by selected experts James Anthony writes a long yet rather unconvincing plea that we not abandon the gateway theory, in spite of these results. Rather than giving cogent reasons, he calls on an odd comparison with the shared viral causes of chickenpox and shingles. This seems a weak analogy, as behavioural problems such as drug use do not have a single origin like a virus (although there are occasionally similarities between drug epidemics and those caused by infectious disease, see Frischer M).

Kenkel and Mathios call for economic examinations of the supply and demand situations for drugs, forgetting that this has been done 'de facto' in Holland and South Australia for 20 years. But the trove of interesting evidence from decriminalised areas, including independent statistics is once again ignored, even by serious commentators on drug policy.

Lynskey claims that Morral et al's work is inconsistent with other findings but the two studies he quotes do not seem to address causation of 'progressive' drug use, merely its existence (including alcohol and tobacco). He then states that his own and other work shows that there is a strong genetic component (~50%) to drug use and/or dependence, and that such a common factor may explain Morral's findings. But such a propensity to all drug use is precisely what Morral's group assumed ... and then deduced that this fitted the detailed serial drug prevalence findings without the need for an additional 'priming' or gateway effect. Indeed, a gateway effect would have produced very different prevalence figures from those found by the household surveys.

The study authors reply with the only really robust statements of the otherwise bland 'debate' regarding the existence of the 'gateway' or stepping-stone theory of drug use. Although they say their work does not disprove the gateway theory they are "sceptical about the prospects for any [future] definitive tests [of the validity of the marijuana gateway effect]". They go on to say that US prohibition policies result in 'an enormous black market and over 600,000 arrests annually with associated economic, personal and social consequences. Evaluation of prohibition policies should balance the known and possible harms of legal access to marijuana against the known and possible harms associated with marijuana prohibition." An American president propounded the same sentiment 20 years ago, but without any moves on the official, lignified American stance on strict cannabis prohibition. It is interesting that cannabis for medical treatment is now gaining popular interest again. It was widely used in the first half of the twentieth century (eg. 'cannabis tincture').

There have been claims that a new study by Lynskey et al. in JAMA 'proves' the gateway theory, however this is not the claim of the authors. They state of their study of adolescent drug use in twins who 'used' or 'did not use' cannabis by age 17 that: "While the findings of this study indicate that early cannabis use is associated with increased risks of progression to other illicit drug use and drug abuse/dependence, it is not possible to draw strong causal conclusions solely on the basis of the associations shown in this study." There are serious limitations in this study which was a cross sectional retrospective study using phone interviews from the twins up to 15 years after the exposure.

Regardless of the existance of a gateway theory and other possible serious dangers of cannabis, our current laws seem not to protect young people. None of the many subjects in these trials was prevented from using cannabis under the current system. As in Holland, it would seem more logical for the authorities to regulate the drug as occurs with alcohol and tobacco, two drugs which are also often called 'gateway' drugs (with some justification). Prohibition may indeed be the main common factor linking cannabis, heroin, cocaine and other illicit drug use. Lynskey et al. support this by writing: ".. access to cannabis use may provide individuals with access to other drugs as they come into contact with drug dealers. ... [Dutch decriminalization of cannabis] may have been partially successful as rates of cocaine use among those who have used cannabis are lower in the Netherlands than in the United States."

comments by Andrew Byrne ..