Institute of Medicine
National Academy of Science

MEDICAL USE OF MARIJUANA: ASSESSMENT OF THE SCIENCE BASE

Review
By
Lester Grinspoon, M.D.
Harvard Medical School
74 Fenwood Road
Boston, MA 02115

The report has considerable virtues and serious deficiencies. It is clearly written and includes an excellent chapter on cannabis neurophysiology. The committee correctly points out that cannabis constituents have a broad range of potential medical uses, far beyond those considered acceptable today. The authors properly dismiss the thoroughly discredited notions that are still sometimes introduced in opposition to medical marijuana use: the "gateway" theory that marijuana use leads to the use of more dangerous drugs and the view that non-medical use would be encouraged by the availability of cannabis as a medicine. The report's emphasis on the importance of developing a non-pyrolytic device for the inhalation of cannabis vapors is also excellent. The report usefully acknowledges that relief of anxiety and sedation may be therapeutic for some patients. It also acknowledges that any harm from long-term non-medical use should not be a consideration in the evaluation of cannabis for medical purposes.

The main deficiency of the report is insufficient attention to anecdotal evidence—case reports and clinical experience. The omission is justified by the statement (page 3, line 26 of the Executive Summary) that the committee "…emphasizes evidence-based medicine (derived from knowledge and experience and formed by rigorous scientific analysis) as opposed to belief-based medicine (derived from judgment, intuition, and beliefs untested in rigorous science)." Yet anecdotal evidence is by far the richest source of information on the medical utility of cannabis. If it were not for this clinical experience, there would have been no pressure on the Office of National Drug Control Policy to have the Institute of Medicine undertake the task of preparing this report. The almost total absence of any attempt to integrate it into the report will raise questions as to whether the committee has fully addressed all aspects of its charge, which reads: "The study will address what is currently known and not known about the medical use of marijuana."

Anecdotal evidence is the source of much of our knowledge of synthetic medicines as well as plant derivatives. As Louis Lasagna has pointed out, controlled experiments were not needed to recognize the therapeutic potential of chloral hydrate, barbiturates, aspirin, curare, insulin, and penicillin. More recently, the same could be said of lithium. Lasagna asks why regulators are willing to accept the experience of physicians and patients as evidence of adverse effects but not as evidence of therapeutic effects. For example, if the committee had paid more attention to the anecdotal evidence, it would not have dismissed the use of marijuana in glaucoma, epilepsy, and migraine, and might have raised the question of its usefulness in premenstrual syndrome, mood disorders, and alcoholism, to name a few possibilities. One source of this evidence is the series of books on Marijuana, Medicine, and the Law edited by Robert Randall (Washington, D.C.: Galen Press, 1988-1991). It consists largely of testimony presented in support of the petition to reschedule cannabis which was rejected by the Drug Enforcement Administration. Other sources are Lester Grinspoon and James Bakalar, Marijuana, the Forbidden Medicine, Revised and Expanded Edition, Yale University Press, 1997, and several websites.

Most of the recent clinical experience with cannabis involves smoked marijuana. Perhaps because it pays insufficient attention to that experience, the committee suggests a medical role for smoked marijuana (or the inhaled natural plant cannabinoids in any form) that is far too restrictive. The bias is so strong that in several places the report wrongly equates the pharmacological activity of the whole plant with that of THC. This statement is belied by the experience of patients. Many who are familiar with both substances find whole smoked marijuana to be much more useful than THC, and not merely because of the limitations of obligatory oral delivery.

In its attitude toward the cannabis plant, the report echoes the 1997 report of the British Medical Association, which called for research into single compound cannabis derivatives but specifically recommended against pursuing research on whole cannabis. The AMA House of Delegates, as Appendix D indicates, has taken a different position. More recently, so has the House of Lords Science and Technology Committee in its 1998 report on medicinal cannabis. The Committee acknowledges the importance of the fact that patients prefer whole smoked marijuana to dronabinol or nabilone, and concludes that it would be premature to say that the only active substance in cannabis is THC.

Furthermore, while admitting that it may be irrelevant in a medical context, the report proposes further studies of the health risks of smoking marijuana "among populations that use it widely." This issue has been studied extensively over the last 30 years, and any objective overview of the literature indicates that the risks are minimal. I do not believe that further examination of this question is a priority, and the demand for further research on it should not be allowed to delay medical availability. The report contends that "the adverse effects of marijuana are within the range tolerated for other medications." This is somewhat misleading, because the adverse effects of cannabis are generally less than those of other medications used for the same purposes. The remarkable safety of cannabis and the medical significance of that safety are consistently underestimated or understated. Thus, for example, the dangers of chronic non-medical use, including the risk of dependence, are overstated. Despite the concurrent acknowledgment that this should not affect judgments of marijuana's medical value, it is bound to discourage medical development in the present social climate.

I elaborate on these points and others in the following notes on the five chapters and the Executive Summary.

Pages 20-22: Three convincing case histories are presented, but the body of data represented by these accounts is not taken as seriously as it should have been. Although we should of course be cautious about anecdotal evidence, the tone of this very limited discussion is too skeptical.

Page 23: 416-426. The reference to the lack of "stories from people who had tried marijuana but not found it effective" is irrelevant. Since they are unlikely to have suffered any ill effect, they would have no reason to tell their "stories." I do not agree with the conclusion that "careful analysis of data collected under controlled conditions" is needed to determine how many people would find marijuana "more harmful than beneficial." Many millions of dollars have been spent in attempts to demonstrate toxicity which would justify the prohibition of marijuana use for non-medical purposes, and the effort has been entirely unsuccessful. It is difficult to believe that cannabis would be more toxic for medical users. The editors of The Lancet and the New England Journal of Medicine share this view.

Chapter 2

On the whole this chapter is well done, but much of the material is quite technical and perhaps, considering the anticipated audience, should at least in part be presented as an appendix.

Page 6: 10. The statement that CBD is not psychoactive is inconsistent with the work of Carlini and Cunha ("Hypnotic and antiepileptic effects of cannabidiol", Journal of Clinical Pharmacology 21, 1981). These researchers demonstrated the usefulness of cannabidiol as a hypnotic in a controlled study of 15 patients.

Page 48: 870-877. I do not agree that "…it will be important for future studies to firmly establish the abuse potential for marijuana compared to other drugs of abuse…" This should in no way be a priority at this point, especially if it is used as a justification for delaying medical availability. A vast amount of study has been devoted to this subject, and it is overwhelmingly clear that the abuse potential of marijuana "compared to other drugs" is low.

Page 7: Box. Dependence is not an important issue in evaluating medical use, as the committee points out elsewhere in the report.

Page 35: 642-656. It would be more cautious not to give the impression that the adverse effects reported here are at all common, or indeed are anything but rare. For example, the committee might refer to studies by Vera Rubin and Lambros Comitas (Ganja in Jamaica, The Hague: Mouton, 1975) and by C. Stefanis, et al, ("Clinical and Psychophysiological Effects of Cannabis in Long-term Users" In: Brauda and Szara, eds. Pharmacology of Marijuana, Vol. 2, pp. 659-656) which indicate little or not harm from long-term use. My own 31 years of experience with marijuana users and my ongoing review of the scientific literature is inconsistent with reports of serious long-term harm.

Page 60: 1148-1150. I have seen no evidence that the cardiovascular effects are a serious problem for older patients.

Page 62: 1196-1205. Research on Rastafarian and Ethiopian Zion Coptic women suggests that cannabis use is not harmful to pregnant women or the children they are carrying. See Dreher, et al. "Prenatal Marijuana Exposure and Neonatal Outcomes in Jamaica: An Ethnographic Study." Pediatrics 93: 254-260 (February 1994).

Page 63: 1210-1212. In reference to the Hingson study (citation 54), it should be noted that his subjects were using other substances as well.

Page 64: 1283-1288. If the committee members had paid more attention to the abundant anecdotal evidence, they might have discovered that at least some people who are regarded as "dependent" on marijuana are using it in much the same way and for the same reasons that patients use selective serotonin reuptake inhibitors.

Page 66: 1282-1283. The statement that "Smoked marijuana is clearly a poor treatment for any chronic medical condition" is not supported by controlled studies, and the anecdotal evidence clearly refutes it.

Page 67: 1315-1317. It should be noted that the potential dangers of marijuana must be measured against the alternatives.

Page 8: 109-110. Clinical experience makes it seem quite unlikely that "the subjects were tolerant to THC at the time of testing." I have not observed tolerance to marijuana's analgesic properties.

Page 11: 171-172. There is, as far as I know, no evidence that migraine which did not pre-exist the use of cannabis results from cessation of its use.

Page 13: 207-215. This list should include migraine. Cannabis has a long history in the treatment of this disorder; Sir William Osler proclaimed it to be the best available treatment in his 1915 textbook. Despite the development of new medicines, including sumatriptan, many people suffering from migraine still prefer marijuana, partly because it has fewer uncomfortable or dangerous side effects.

Page 14: 226-229. The suggestion that the plant extract should be compared to THC is encouraging (although presumably the authors are not referring to smoked cannabis).

Page 17: 272-274. To assert that emesis caused by cisplatin lasts for "an hour" is a gross understatement. If the definition of emesis includes dry heaves, it can go on for many hours.

Page 20: 328-336. The effectiveness of delta-8-THC in childhood cancer is further proof that delta-9-THC alone is not, as asserted elsewhere in the report, the source of all established marijuana effects.

Page 20: 337-339. When taught how to use marijuana properly, even older patients with little prior experience rarely find it to be dysphoric or anxiogenic.

Page 21: 348-350. In my reading of the Vinciguerra study, 76% of the 56 patients became symptom-free.

Page 21: 356-362 and Page 22: 365-367. The stated finding that "none of these trials showed a clear advantage for smoked marijuana over oral THC" is completely at odds with clinical experience. Nauseous patients often report that they often cannot take THC pills. The onset of relief is too slow, and because of the vagaries of bioavailability, they feel that they have sometimes taken too much or too little. When they smoke marijuana, the effect is immediate, and they learn quickly to titrate the antiemetic dose

Page 22: 369-374. I have never seen an experienced patient in whom marijuana has caused dizziness or orthostatic hypotension. Whether and when "euphoria" should be considered a serious side effect is open to question.

Page 24: 411-424. The surveys referred to here are dismissed as irrelevant because they "deal with perception rather than pharmacological reality based on well-designed outcome studies." Yet earlier, in reference to pain control (page 10), surveys are cited as useful. There seems to be a contradiction here. The surveys clearly reveal that oncologists as well as patients find that cannabis is useful in controlling the nausea and vomiting of cancer chemotherapy. Incidentally, this drug is designated by federal law as "unsafe for use under medical supervision." If the law is correct, many respected oncologists are guilty of malpractice. The committee might find it useful to comment on this anomaly.

Page 25: 442-444. I question the assertion that "The side effects associated with cannabinoids are tolerable, but greater than those seen with other classes of agents." The side effects are different and in my view more benign. I do not believe the word "greater" is appropriate here.

Page 26: 466-468. The passage that reads "…but whether there is a group of patients who might obtain added relief from marijuana or cannabinoid drugs" should be changed to read "…but rather whether there is a group of patients who might obtain added or better relief with fewer side effects from marijuana or cannabinoid drugs."

Page 27: 488-490. Most people who use marijuana for the control of chemotherapy-induced nausea and vomiting find that it improves their mood and reduces anxiety. Chemotherapy patients who use oral THC, especially those who are unfamiliar with it, may experience unpleasant side effects.

Page 28: 502-505. To state that "It is possible that the harmful effects of marijuana for a limited period of time might be outweighed by the antiemetic benefits …" is once again to exaggerate the potential harm. Surely patients are unlikely to do any measurable harm to their pulmonary systems if they receive chemotherapy, say, once every two weeks, and smoke the equivalent of one half to one marijuana cigarette each time.

Page 31: 563-565. Megestrol acetate may be more effective than dronabinol in promoting weight gain, but it is certainly less effective than whole smoked marijuana. Furthermore, in the required doses Megace may precipitate blood clots, heart attacks, and strokes—much more serious side effects than any caused by either smoked marijuana or dronabinol.

Page 32: 579-582. This passage suggests that the Donald Abrams study concerns the medical efficacy of marijuana. That was its original objective, but it is now primarily a study of safety.

Page 35: 639-641. Again, the side effects of megestrol acetate can be more serious than hyperglycemia and hypertension. Dronabinol rarely causes dizziness or lethargy, and its side effects are of far less consequence.

Page 36: 677-683. Again the risks of smoking are exaggerated.

Page 40: 746-747. If "the regular use of smoked marijuana in a chronic condition" presents dangers, the answer would be a device that separates the cannabinoids from other plant material. This point is made elsewhere in the report and should be emphasized here as well.

Page 44: 797-800. The report neglects to mention dantrolene and diazepam, other commonly prescribed antispasticity drugs with far more serious side effects than cannabis.

Page 52: 950-960. The suggestion here is that lack of anecdotal reports indicates that marijuana cannot be useful in the treatment of movement disorders. This is curious, given the committee's skeptical attitude toward anecdotal reports in other contexts. In any case, in my clinical experience some people with dystonias and many with Tourette's syndrome do find marijuana useful.

Page 53: 972-973. The wording of this sentence should also include "some spastic disorders."

Page 53: 973-979. The report states that experiments should include controls for sedation and anxiolytic effects. It should also be noted here that these are correctly described as potentially therapeutic effects elsewhere in the report. (See Executive Summary, page 6, and Chapter 4, page 67: 1318.)

Page 54: 993-996. It is incorrect to say that there are "few anecdotal and individual case reports that marijuana controls seizures in epileptics…" There are many such reports, and they provide a strong scientific basis for further study.

Pages 54-55: 1011-1021. The study by Cunha, Carlini, Pereira, et al, "Chronic administration of cannabidiol to healthy volunteers and epileptic patients", indicating antiepileptic properties of cannabidiol, is not mentioned here.

Pages 59-64: The evidence that cannabis might slow the natural progression of glaucoma or prevent loss of vision is the same as the evidence for any other anti-glaucoma drug: it reduces fluid pressure on the optic nerve. The fact that this evidence is not "direct" does not mean that it is valueless. The need for frequent dosing, emphasized here, is only one factor to be considered, along with efficacy and side effects, in deciding whether to use a medicine. It is not true that "the harmful effects of chronic marijuana smoking clearly outweigh its unproven benefits for the treatment of glaucoma." Therefore I would remove the concluding statement that "clinical studies should not be recommended."

Page 65: Again, the incorrect statement that THC "can produce all the established effects of marijuana" is repeated.

Page 68: 1261-1268. The suggested restrictions require such a high level of bureaucracy that any program based on this recommendation would be no more successful than the now-defunct Compassionate IND program, which never enrolled more than a few dozen patients. People with a medical need for marijuana will simply find other ways to obtain it.

Page 2: 7-11. The statement that "…renewed interest in the therapeutic applications of cannabinoids stems from major scientific discoveries…" tells less than half the story. The main source of interest in their therapeutic properties has been the discoveries of lay people using cannabis for a variety of medicinal purposes. If it were not for that experience, this report would not even exist. The committee leaves present and potential medical cannabis users with little hope when it suggests that the therapeutic potential of cannabis will be realized only by the commercial exploitation of individual cannabinoids. The time and expense required for such a development will simply be too great. Only the federal government might underwrite the effort, as it did in part for dronabinol, but at present this seems doubtful.

Page 20: 6-7. The statement of Unimed's CEO that "… only a small percentage of its market…[is] lost to 'competition from marijuana itself'" should not be allowed to remain without comment. More attention to the anecdotal evidence would show that many if not most patients prefer whole smoked cannabis to dronabinol, and a large number of them exercise this preference, despite the legal risks, by growing marijuana themselves or buying it on the illicit market.

Page 21, lines 14-18. Of the five uses mentioned here, all but the first became known through clinical experience and anecdotal evidence.

Page 22: 18ff. Unimed is quoted as stating in a submission to the DEA that "If the patient smokes two or more cigarettes a day, then Marinol may be cheaper than marijuana; if the patient smokes only one marijuana cigarette, then Marinol may be more expensive." Considering that the cost of Marinol is $8 to $10 per pill, I am not certain that even at present street prices this assertion is correct. It is certainly not accurate for patients who grow their own.

Page 24: 18-19. "The need to satisfy the DEA's regulatory requirements related to drug scheduling" is a major reason why commercial development of any naturally occurring cannabinoid will be extremely difficult. For example, as the report mentions elsewhere, delta-8- THC is not psychoactive and is an extremely effective antinauseant in children with cancer. But pharmaceutical firms regard it as unmarketable because of the scheduling problem. Commercial developers will put great pressure on regulatory agencies to downschedule or not schedule synthetic cannabinoids and analogs. If they are successful, we will have an absurd situation in which natural cannabinoids are restricted as abusable drugs and synthetic cannabinoid receptor agonists, possibly more potent, are not. It is urgent that natural cannabinoids be downscheduled or descheduled, and the report should specifically recommend this.

Page 25: 1-3. The sentence refers to "… competition from other products"; these "other products" will include home-grown and street marijuana, the price of which will depend mainly on how long the American public continues to tolerate the arrest of large numbers of people (more than 600,000 in 1997) on charges of marijuana possession.

Page 31: 16-17. The implication here is that marijuana supplies for clinical research are fairly easy to obtain. The experience of Donald Abrams shows that this is not so.

Page 32: 5-8. This is incorrect. The DEA has never issued an import permit.

Page 36: 6. The report refers to "8 scheduling criteria" here, but only 5 criteria are listed on page 35, beginning at line 21.

Page 37: 16-22. Hortapharm is not the only company developing new marijuana strains. There are also several other groups in the Netherlands and GW Pharmaceuticals in England.

At the cost of some repetition, I will state my reservations and comments with respect to the Executive Summary and its recommendations.

Page 4: 65-67. Again, THC cannot produce "all the established effects of marijuana."

Page 5: Therapeutic effects are not always "modest", and medical use should not necessarily be confined to people who "do not respond to other medications", since marijuana is usually safer and with fewer toxic side effects.

Page 9: 153-154. In 31 years of experience with people who use marijuana medically and non-medically, I have yet to see even a mild withdrawal syndrome.

Page 12: 206-208. The development of non-smoked rapid-onset cannabinoid delivery systems is an urgent need, but it should not be the only goal of clinical trials of smoked marijuana. The report should also emphasize the advantages of developing a non-smoked respiratory delivery system for the natural plant medicine.

Page 14: Recommendation 6. This recommendation refers to "debilitating symptoms", but medicinal use of marijuana should not be so restricted. The recommendation also demands documentation of the failure of all approved medications, without addressing the possibility that both marijuana and the standard medication give equally good results but the side effects of the latter are more uncomfortable or even toxic. Finally, the suggested oversight strategy is far too bureaucratically cumbersome to be workable (see comment on Chapter 4, page 68).