After about a decade’s hiatus in the research into smoked marijuana’s medical uses, the FDA has approved a study comparing smoked marijuana to the oral THC pill in the treatment of weight loss caused by the HIV-related wasting syndrome. The principle investigator in the study is Dr. Donald Abrams of the San Francisco Community Consortium. Previous editions of the MAPS newsletter have reported on this project, and on the $50,000 pledge that MAPS received for the study from an individual in the Netherlands.
As a result of my work assisting Dr. Abrams with protocol development, I was asked by officials at the FDA to prepare a Clinical Plan outlining the entire research agenda for the development of the medical uses of marijuana (this plan will be more fully discussed in the next issue of the MAPS newsletter). This Clinical Plan formed the basis for further discussions with the FDA. These discussions focused on the sequence, size and design of the studies that the FDA will require to be successfully carried out prior to any approval of the prescription availability of marijuana for one or more medical indications.
Protocol design issues
The discussions with the FDA led to some very helpful suggestions about how to design the marijuana research protocol. Unfortunately, there are still some major obstacles in the road ahead. As a result of the work on the Clinical Plan, a major protocol design issue has arisen which involves the clash of differing sets of concerns, all of which have some validity. The FDA requires at least two "adequate and well-controlled trials" demonstrating safety and efficacy before it will approve a drug for prescription use. Traditionally, these studies are designed in the standard double-blind placebo-controlled manner. If a study is not double-blind, the FDA is not inclined to consider it "well-controlled." Given that a single medical marijuana study will probably take over a year to conduct and cost over $150,000, it is essential that the FDA will seriously consider the data it generates.
Though the FDA has approved Dr. Abrams’ study comparing smoked marijuana to Marinol, it considered it a pilot study that was not "well-controlled." The rationale was that since the study design compares smoked marijuana to oral THC capsules, all the patients could easily determine which experimental treatment they were receiving simply by noticing if their medicine is smoked or swallowed. This knowledge in turn compromises the double-blind. As a result, bias can be consciously or unconsciously introduced into the experiment by either the subjects, the researchers, or both. Limiting the potential for bias to influence the results of the study is the fact that patients suffering from the wasting syndrome are by definition not able to control their weight at will ( if they could they wouldn’t be in the study). Nevertheless, subjects would be able to alter their reports concerning their appetite, an outcome variable that was going to be measured through patient self-reports.
Despite everyone’s best intentions, there is no known way to conduct effectively a double-blind study with two psychoactive drugs like marijuana and the oral THC capsule. The best way to attempt to create a double-blind experiment with these two drugs is to give patients either active marijuana and placebo THC, or placebo marijuana and an active THC capsule. Since the pill takes about 45 minutes to take effect, the subjects would be asked to first take the pill, wait about 45 minutes, and then smoke some marijuana.
Even with the attempt to confuse the patient about which drug was active, patients would still be able to tell which active drug they received. The rapid, almost instantaneous onset of the psychoactive effects of smoked marijuana differs dramatically from the slow onset of the THC capsule. Even the nature of the subjective experience is different. Furthermore, the oral THC pill’s absorption into the body is quite variable from day to day depending on diet and other factors. The pill’s actual time of onset could never be perfectly correlated with any standardized time suggested to the patient about when to smoke the marijuana. After several days in the study, at the longest, all the patients would know which drug they received, the active marijuana or the active pill.
After some discussion, the FDA officials and I agreed that a design involving high dose marijuana (10% THC), medium dose marijuana (3-5% THC) and placebo marijuana ( less than 1% THC) would come the closest to being a true double-blind study. Such a study would be considered as one of the two required "adequate and well-controlled" studies. I therefore suggested to Dr. Donald Abrams that we redesign the study by dropping the use of Marinol and adding a medium dose group and a placebo marijuana group. Dr. Abrams agreed to present the new study design to the Scientific Advisory Board (SAC) of the Community Consortium at its December 2, 1993 meeting.
No to placebo marijuana
To my surprise, the SAC rejected the proposed protocol design. The reasons for its decision are valid, as is the FDA’s competing need for a "well-controlled" study. Some mutual accommodation will need to be developed. While the SAC approved of the use of high- and medium-dose marijuana, it refused to administer placebo marijuana to its patients. The rationale was that the subjects would be exposed to the potentially harmful effects of smoking without receiving any benefits. Furthermore, after a few days in the study, they would certainly be able to determine whether they were smoking active marijuana or not, thereby compromising the double-blind. Since the double-blind would not be effective anyway, the SAC saw no reason to go through with an empty formality that carried even a slight degree of risk.
I had also thought that patients in the study would indeed be able to tell if they had active or placebo marijuana. However, I thought a good-faith effort at the double-blind would be worth it in order for the study as a whole to be considered "well-controlled". Furthermore, I considered the risk of harm to the subjects from the smoke from the placebo marijuana to be minimal, especially since there are simple ways to reduce the risk of smoking, for example with a water pipe. In deciding to recommend this design, I tried to weigh the risk/benefit ratio for the subjects who would be randomly assigned to the placebo marijuana group. While I recognized that they would not benefit from their three months in the study, and would also be exposed to smoke without therapeutic value, I thought that their opportunity to obtain legal access to smoked marijuana after the conclusion of the three month study period might adequately compensate them for participating in the study in the placebo marijuana group.
The SAC had one additional concern that even I could see was compelling. The SAC pointed out the ethical problem of not providing patients suffering from the life-threatening wasting syndrome with some active treatment, especially since there already are two drugs approved by the FDA for this indication – Marinol and a new drug called Megace. When there are no FDA-approved drugs for an illness, having a placebo group is easily justified since there are no proven alternatives that have been demonstrated to be safe and effective. Indeed, given that all drugs have some side-effects, administering a drug of unknown safety and efficacy is probably harder to justify than giving a placebo.
Though there are FDA-approved drugs for the wasting syndrome, the SAC was not enthusiastic about the therapeutic value of either Marinol or Megace. Marinol has only been demonstrated to increase appetite but not weight. While Megace has a small effect on weight gain, this may primarily be water weight of little consequence to overall health. Despite their shortcomings, the SAC preferred to administer one of these drugs to the control group instead of placebo marijuana. Using either of these drugs, however, would still not result in a double-blind study. Furthermore, using an active placebo with some therapeutic potential creates another problem. When the data from standard double-blind studies are analyzed, the important variable is the difference between the effect of the treatment and the effect of the placebo. The statistical probability that the observed effect of the treatment is due to the treatment and not to chance depends in part on how large the difference is between the two groups. However, when the drug used as a placebo has some therapeutic effectiveness of its own, the true effect of the treatment is more difficult to determine.
One possible solution
Dr. Abrams, the SAC, the FDA and I are all still trying to develop an appropriate protocol design. My preference is to use low-THC marijuana that still has some therapeutic effect, perhaps 2% marijuana. The study design would then be patterned on a standard dose-response study, though without an inactive placebo. Using three groups with high-, medium- or low-THC content at least maximizes the chance that patients and researchers will be blind as to which treatment group the patients are in. The problem of comparing the results of the high and medium doses to an active placebo remains to be addressed, but.there seems to be no way to avoid this. In any case, isn’t this why some people become statisticians?
Some progress has resulted from all this struggle to craft the best design for the experiment. The SAC decided that it would be a good idea to conduct a small pilot study with whatever design we come up with, rather than enter into a big trial without first testing out the methodology with a smaller group of patients. Since we have already been pledged $50,000, we do have enough funding for a pilot study. As soon as the protocol design is finalized and all the regulatory permissions are in hand, we can begin the experiment instead of waiting until larger sums are available. It should be much easier to raise the remaining funds once the pilot study is underway. At that time we would have actually demonstrated that this project was not just another pipe dream but instead was an actual study that could have dramatic implications for the medical treatment of the wasting syndrome, and for medical marijuana research in general.
FDA grant program
In a good-faith attempt to help the medical marijuana debate be resolved on the scientific merits of research data, an FDA official who did not have to do so nevertheless informed me about a grant program that I did not know existed. This grant program had the potential to provide substantial funding for the wasting syndrome study. This gesture convinced me more than ever that working with the FDA to conduct scientific research is an essential part of breaking the logjam concerning the medical use of marijuana.
The grant I was shown was from FDA’s Office of Orphan Products Development. These grants are for research into the therapeutic potential of drugs intended for diseases which afflict 200,000 people per year or less, such as the AIDS-related wasting syndrome. I had erroneously assumed that marijuana would first need to be formally approved by the FDA as an Orphan Drug before any funding proposals would be considered. I was told instead that researchers need only have an approved protocol , which Dr. Abrams already had. The grant application deadline was December 27, 1993, about a month and a half after I was informed about the program. If there were no unexpected problems, it seemed that there just might be enough lead time for Dr. Abrams and the Community Consortium to prepare a grant application for the research study, and possibly obtain funding support for the study directly from the FDA.
The SAC’s decision to reject the proposed protocol design with the placebo marijuana group had one major unfortunate side-effect. In the absence of a protocol design that the SAC was willing to implement and the FDA was willing to consider well-controlled, Dr. Abrams was not able to prepare an application for FDA funding for the study in time for the December 27, 1993 application date. Since this grant program is on an annual cycle, the opportunity for FDA funding for medical marijuana research slipped through our fingers for at least a year. While other sources of funding will probably become available eventually, FDA funding would have been particularly appropriate for such a controversial drug at this stage in its development.
National Medical Marijuana Day
Another idea to help raise funds for medical marijuana research is to initiate a National Medical Marijuana Day, perhaps in April or May of 1994. The idea would be to catalyze hundreds of small, local fundraising events all synchronized to take place on the same day. The events would ideally be co-sponsored by local medical marijuana groups, NORML chapters, and the Cannabis Action Network, some of the same groups that helped with the symposium commemorating the 50th anniversary of LSD. In addition, we would advise the local organizers to seek co-sponsorship of the events with their local AIDS groups. If each local event raised $500 and there were 100 events, $50,000 would be generated.
The central organizers would need to assemble a "Local Organizers Starter Packet" composed of a video about the issue, some informational material, and organizational tips. This kit could be sold at a price that covers costs, perhaps $25. It should also be possible to arrange a benefit concert to take place in San Francisco on the same day with some very popular bands. More on this idea later. If you would be interested in sponsoring a local event, please contact MAPS.