To realize the full potential of cannabis as a medicine in the setting of the present prohibition system, we would have to address all these problems and more. A delivery system that successfully navigated this minefield would be cumbersome, inefficient, and bureaucratically top-heavy. Government and medical licensing boards would insist on tight restrictions, challenging physicians as though cannabis were a dangerous drug every time it was used for any new patient or purpose. There would be constant conflict with one of two outcomes: patients would not get all the benefits they should, or they would get the benefits by abandoning the legal system for the black market or their own gardens and closets.

A solution now being proposed, notably in the Institute of Medicine (IOM) Report, is what might be called the "pharmaceuticalization" of cannabis: prescription of isolated individual cannabinoids, synthetic cannabinoids, and cannabinoid analogs. The IOM Report states that "...if there is any future for marijuana as a medicine, it lies in its isolated components, the cannabinoids, and their synthetic derivatives." It goes on: "Therefore, the purpose of clinical trials of smoked marijuana would not be to develop marijuana as a licensed drug, but such trials could be a first step towards the development of rapid-onset, non-smoked cannabinoid delivery systems." This position was recently echoed by Antonio Maria Costa, Executive Director, Office on Drugs and Crime, the United Nations at the International Symposium on Cannabis in Stockholm on March 7th, 2003: "I am not sure I understand the controversy about the medical virtues of cannabis: First, if and when they are ascertained, society should definitely make use of them. Who would oppose the advances of medicine? Who would stand in the way of reducing suffering? My concern is to prevent that, by proclaiming the (medical) virtues of cannabis, we open a back door to its wider (recreational) consumption. Society would end up regretting such abuse, just as we now regret tobacco addiction. If proven to be medically useful:, and this is my second point, cannabis should be treated like any other medicine, namely as a pharmaceutical preparation to be prescribed for specific symptoms in accordance with properly determined dosages and standards. In other words, either we are serious about the medical properties of cannabis (and we, in this hall, take the question very seriously) or it is just a matter of using such properties as a Trojan horse to reach other goals – namely, the de facto decriminalization of its production and trafficking. In this case I would be strongly negative." Some cannabinoid analogs may indeed have advantages over whole smoked or ingested marijuana in limited circumstances. For example, cannabidiol may be more effective as an anti-anxiety medicine and an anticonvulsant when it is not taken along with THC, which sometimes generates anxiety. Other cannabinoids and analogs may prove more useful than marijuana in some circumstances because they can be administered intravenously. For example, 15 to 20 percent of patients lose consciousness after suffering a thrombotic or embolic stroke, and some people who suffer brain syndrome after a severe blow to the head become unconscious. The new analog dexanabinol (HU-211) has been shown to protect brain cells from damage when given immediately after the stroke or trauma; in these circumstances, it will be possible to give it intravenously to an unconscious person. Presumably other analogs may offer related advantages. Some of these commercial products may also lack the psychoactive effects which make marijuana useful to some for non-medical purposes. Therefore, they will not be defined as “abusable” drugs subject to the constraints of the Comprehensive Drug Abuse and Control Act. Nasal sprays, vaporizers, nebulizers, skin patches, pills, and suppositories can be used to avoid exposure of the lungs to the particulate matter in marijuana smoke.

The question is whether these developments will make marijuana itself medically obsolete. Surely many of these new products would be useful and safe enough for commercial development. It is uncertain, however, whether pharmaceutical companies will find them worth the enormous development costs. Some may be (for example, a cannabinoid inverse agonist that reduces appetite might be highly lucrative), but for most specific symptoms, analogs or combinations of analogs are unlikely to be more useful than natural cannabis. Nor are they likely to have a significantly wider spectrum of therapeutic uses, since the natural product contains the compounds (and synergistic combinations of compounds) from which they are derived. For example, the naturally occurring THC and cannabidiol of marijuana, as well as dexanabinol, protect brain cells after a stroke or traumatic injury.

The cannabinoids in whole marijuana can be separated from the burnt plant products (which comprise the smoke) by vaporization devices that will be inexpensive when manufactured in large numbers. These devices take advantage of the fact that finely chopped marijuana releases the cannabinoids by vaporization when air flowing through the marijuana is held within a fairly large temperature window below the ignition temperature of the plant material. Inhalation is a highly effective means of delivery, and faster means will not be available for analogs (except in a few situations such as parenteral injection in a patient who is unconscious or suffering from pulmonary impairment). It is the rapidity of the response to inhaled marijuana which makes it possible for patients to titrate the dose so precisely. Furthermore, any new analog will have to have an acceptable therapeutic ratio. The therapeutic ratio (an index of the drug's safety) of marijuana is not known because it has never caused an overdose death, but it is estimated, on the basis of extrapolation from animal data, to be an almost unheard of 20,000 to 40,000. The therapeutic ratio of a new analog is unlikely to be higher than that; in fact, new analogs may be much less safe than smoked marijuana because it will be physically possible to ingest more of them. And there is the problem of classification under the Comprehensive Drug Abuse and Control Act for analogs with psychoactive effects. The more restrictive the classification of a drug, the less likely drug companies are to develop it and physicians to prescribe it. Recognizing this economic fact of life, Unimed Pharmaceuticals Inc. has fairly recently succeeding in getting Marinol (dronabinol) reclassified from Schedule 2 to Schedule 3. Nevertheless, many physicians will continue to avoid prescribing it for fear of the drug enforcement authorities.

A somewhat different approach to the pharmaceuticalization of cannabis is being taken by a British company, G. W. Pharmaceuticals. It is attempting to develop products and delivery systems which will skirt the two primary popular concerns about the use of marijuana as a medicine: the smoke and the psychoactive effects (the "high"). To avoid the need for smoking, G. W. Pharmaceuticals has developed an electronically controlled dispenser to deliver cannabis extracts sublingually in carefully controlled doses. The company expects its products (extracts of marijuana) to be effective therapeutically at doses too low to produce the psychoactive effects sought by recreational and other users. My clinical experience leads me to question whether this is possible in many or even most cases. The issue is complicated by tolerance to the psychoactive effects. Recreational users soon discover that the more often they use marijuana, the less “high” they experience. A patient who smokes cannabis frequently for the relief of, say, chronic pain or elevated intraocular pressure will experience little or no "high". Furthermore, as a clinician who has considerable experience with medical cannabis use, I have to question whether the psychoactive effect is always separable from the therapeutic. And I strongly question whether the psychoactive effects are necessarily undesirable. Many patients suffering from serious chronic illnesses report that cannabis generally improves their spirits. If they note psychoactive effects at all, they speak of a slight mood elevation — certainly nothing unwanted or incapacitating.

The great advantage of the administration of cannabis through the pulmonary system is the rapidity with which its effects are experienced. This in turn allows for the self-titration of dosage, the best way of adjusting individual dosage. With other routes of delivery the response time is longer and self-titration becomes more difficult. Thus, self-titration is not possible with oral ingestion of cannabis. While the response time for sublingual or oral mucosal administration of cannabis is shorter than it is with oral ingestion, it is significantly longer than that from absorption through the lungs and therefore a considerably less useful route of administration for self-titration.

Furthermore, the design of the G. W. Pharmaceuticals dispenser negates whatever self-titration capacity sublingual administration may have. The device has electronic controls that monitor the dose and prevent delivery if the patient tries to take more than the physician or pharmacist has set it to deliver during predetermined time windows. The proposal to use this cumbersome and expensive device apparently reflects a concern that patients cannot accurately titrate the therapeutic amount or a fear that they might take more than they need and experience some degree of "high" (always assuming, doubtfully, that the two can easily be separated, especially when cannabis is used infrequently). Because these products will be considerably more expensive than natural marijuana, they will succeed only if patients are intimidated by the legal risks, and patients and physicians consider the health risks of smoking marijuana (with and without a vaporizer) much more compelling than is justified by either the medical or epidemiological literature