Cannabis in acute migraine treatment project:
Response to National Institutes of Health Critique
Ethan Russo, M.D.
I recently received the "formal" critique of our team's
proposal by the NIH Review Committee. Although I would admit to
discouragement, and my doubts as to how to rectify deficiencies that may
not in fact exist, my research partners and I intend to re-submit this
proposal to NIH for the Spring cycle. The critique contained many instances
calling for elements that the protocol in fact already contained. ...of 29
members of the review team... only eight were neurologists, and none appear
to be headache specialists.
AS A NEUROLOGIST with research interests in migraine and ethnobotany, it
was natural that I would be interested in the controversy concerning
medical marijuana. Over the years, I have had numerous patients relate to
me the efficacy of smoked Cannabis in allaying their migraine symptoms.
In 1997, with benefit of some financial support from MAPS, I submitted an
application to the National Institutes of Health (NIH) for a grant to
study the use of smoked marijuana in the treatment of migraine. This
application process has been mandated by the Federal government as
necessary for the approval of any therapeutic clinical Cannabis studies.
To date, the Short-term Effects of Cannabinoids in HIV Patients study of
Dr. Donald Abrams and his team remains the only other application of this
type to NIH. That study was recently approved, while the Cannabis in Acute
Migraine Treatment Project was rejected.
I recently received the "formal" critique of our team's proposal by the
NIH Review Committee. After examining it, I feel that virtually all points
of that criticism can be adequately addressed. My team plans, with
additional support from MAPS, to submit a revised grant application to NIH
for its June 1, 1998 grant cycle. The following is a review of the points
of the critique along with my initial inclinations as to how they might be
It is always a daunting task to defend one's work, particularly when the
effort involved was as intense as for this one, and with so much at stake.
The entire protocol was written so as to incorporate systematically the
approaches and procedures that were outlined by Dr. Robert Temple
(Associate Director for Medical Policy, FDA Center for Drug Evaluation and
Research) at the NIH Workshop on the Medical Utility of Marijuana in
The study is designed to examine migraine sufferers who have either failed
to respond to or tolerate subcutaneous sumatriptan injection, the current
ne plus ultra in acute headache management. Patients selected for the
study would then be initially treated with one of the following: smoked
Cannabis with 4% THC content (the highest potency provided by NIDA), oral
dronabinol 10 mg. (synthetic THC), placebo capsules, or an injected
meperidine/hydroxyzine mixture (a common emergency room fallback
The criticisms leveled at the protocol by the NIH are multiple, and
occasionally contradictory. One reviewer felt the protocol too ambitious,
another not sufficiently rigorous. Finding middle ground acceptable to all
was not the intent of this study. Rather, the guidelines from Dr. Temple
did not call for either a preliminary "open label" study of therapeutic
marijuana use or definitive "Phase 3" studies. They did call for
comparison of smoked marijuana to oral dronabinol, as well as a control.
We arrived at the figure of 30 study subjects through a sophisticated
statistical analysis that indicated that this number would be sufficient
to demonstrate clinically relevant differences between the four study
Precedent of anecdotal accounts
Another criticism revolved around the inclusion in the protocol of
multiple "anecdotal" accounts as evidence of the efficacy of Cannabis in
headache treatment. It bears repeating that this agent has been used
therapeutically and continuously for 4000 years or more, and was
pre-eminent, or nearly so in migraine treatment for eight decades among
American and European physicians (Grinspoon and Bakalar, 1997; Mikuriya,
1973; Russo, 1998). We might still be using it were it not for the
government's prohibition of Cannabis on false pretenses in 1937. During
that previous era, there were no controlled studies, nor were any needed
for this agent. Doctors as prominent as Queen Victoria's personal
physician, J. Russell Reynolds (Russell, 1890), Sir William Gowers
(Gowers, 1888), and Sir William Osler, the father of modern medicine
(Osler and McRae, 1915), preferred Cannabis for migraine patients because
it worked effectively and safely. In more modern times, there have been no
controlled studies of therapeutic use of Cannabis solely because they have
been politically prohibited. This is precisely why studies such as ours
should be allowed to proceed.
The proposal contained many scientific citations as to proposed mechanisms
for Cannabis' analgesic effects and modulation of serotonergic mechanisms,
but apparently these were not sufficiently compelling to the reviewers.
One critique suggested that marijuana might work on headaches due to its
soporific effect, promotion of relaxation, or because of its anti-nausea
properties. I find this unsupported by the facts. Most people who use
Cannabis therapeutically do not fall asleep; rather, many use only enough
to reduce symptoms so that they may return to their prior activities.
Several of my patients use it in this manner. It is well known that
relaxation techniques may modestly reduce migraine pain, but transiently,
and incompletely. As to nausea, the 5-HT3 antagonists ondansetron and
granisetron are powerful agents in its control, but have no effect on
migraine pain (Peroutka, 1990). These criticisms betray a basic lack of
familiarity with migraine pathogenesis.
Elements overlooked by reviewers
The critique contained many instances calling for elements that the
protocol in fact already contained: use of visual analogue scales for
symptom quantification, clear exclusions for pregnancy, drug abuse, etc.
Perhaps the protocol was not carefully read, or the appendices that
contained some of this material were not circulated. In any event, it is
difficult to be criticized for omissions that did not, in fact, occur. A
valid request would be tighter controls for women in childbearing years to
ensure that pregnancy risks are minimized (i.e., contraception, spousal
One reviewer suggested that anyone who ever smoked marijuana be excluded
from participation in the study. I have never seen this as a criterion for
previous studies, and it seems totally unnecessary. We planned a mixture
of experienced and Cannabis-naive subjects to more closely test
"real-world" clinical issues.
Objections to confidentiality procedure
Objections were raised as to confidentiality procedures. We outlined every
reasonable precaution for locked records, limited access, etc. I
personally felt these were adequate. It is true to say they are not
foolproof, but short of draconian police-state tactics, they would be the
best that could be provided. The study would receive the usual intense
monitoring by NIH personnel, and additionally the local Investigational
Review Board, which happens to be located one floor below that where the
study would be performed.
A stinging personal criticism was leveled at me, questioning my ability to
carry out the study due to a perceived lack of experience in "human
trials." This seems to be a variation of the old chestnut that one has to
have a job to get a job. In fact, I have been carrying out "human trials"
for twenty years: it is called the practice of medicine, where every
prescription is an experiment with its failures, side effects and
pitfalls. To say that this study contains elements beyond my expertise is
unfounded, unsubstantiated, and inaccurate. As a faculty member of two
universities at an undergraduate, graduate, and professional level, and
with personal recommendations from two distinguished chairmen of
university departments of pharmacy for this study, I had hoped not to be
disparaged in this manner.
One critic upbraided me for inclusion of an anecdote that suggested
Cannabis was no better than standard pharmaceutical for one patient. Is
that surprising? Nothing works for everyone. That is called clinical
variation, and inclusion of such information is required in a critical
review of the subject in order for it to retain the kind of scientific
objectivity that I am not applying in this document written for readers of
the MAPS Bulletin.
Inclusion criteria questioned
One reviewer questioned selection criteria for patients. How would we know
that they really had migraine, and not some more dire brain disease? It
was even suggested that patients might require MRI scans before entry
(each scan costs $1,200). Actually, established criteria exist, provided
by the IHS (International Headache Society) and were incorporated in our
questionnaires (Headache, 1988). Each subject can be clinically examined
prior to entry, and this has been sufficient for virtually all previous
clinical headache protocols. Imaging studies for migraine patients are not
Another questioned whether 30 study subjects could be recruited. I believe
that I could find them solely from my patient clientele! Many headache
patients are seeking better treatments and are very open to "new ideas,"
for better or worse, even ones that are currently illegal. Let us crunch a
few numbers. Migraine afflicts 14% of females and 8% of males (Linet et
al., 1989), for a composite of 11%. One fourth of those are severe or
2.75% (Stewart et al., 1992). About 70% of people respond to subcutaneous
administration of sumatriptan (Mathew, 1997). About 30% fail, or an even
greater number have sufficient side effects that they prefer not to use
it. Multiplying that by an estimated adult local population of 60,000,
that would be: 60,000 x 0.0275 x 0.30 = 495 potential subjects. I feel
that this is, in fact, a very conservative figure. Obviously, not all
would wish to be part of a study in which they would smoke marijuana, but
this is a university town, and many would not object; some may be doing so
now. I am confident we can recruit sufficient subjects if only allowed to
Question of placebo
Another issue concerned use of placebo. One reviewer mistakenly thought
that certain subjects would be stuck with placebo or other treatment for
their entire course of ten treatments. I believe they failed to understand
the randomization scheme as it was presented. Here I was caught in a bind.
I would prefer not to use placebo: it is inhumane. It was my intention to
eschew "dummy dope" that would require subjects to smoke an inert material
with the attendant risks, but no benefits. It has previously been shown
that even marijuana-naive subjects can detect when they are receiving
placebo as compared to active Cannabis. The placebo was included in the
protocol because it was considered essential by the NIH Committee on the
Medical Utility of Marijuana. Moreover, no subject in our study would
receive placebo more than once.
Another questioned our use of intramuscular meperidine. Once more, I
included it because, for better or worse, it seems to be the drug of
choice in treating migraine in emergency rooms across the United States. I
personally never use it, and do not recommend it. However, it does provide
a recognized point of comparison to a potential alternative treatment such
as smoked Cannabis. Alternatives such as morphine increase nausea, while
butorphanol (Stadol) has been associated with myriad dangers (Fisher and
One reviewer felt a two hour period of observation was insufficient, and
suggested patients be kept overnight. This requirement alone would serve
to more than triple the cost of the study (not that we the taxpayers
should be concerned). Since migraine is primarily an outpatient disease,
this stipulation represents extreme overkill, and would impair subject
recruitment, perhaps prohibitively. One of the main aims of this study is
to ascertain whether people can function better after migraine treatment
with Cannabis. They can not do that wasting time and money in the
hospital. What about that confidentiality anyway? In this small town, your
nurse might be a friend of your cousin, and tell him you were in the
hospital. We plan to treat patients up to ten times in a six month
period. Another fear expressed was that patients might not have 10
headaches during business hours in the 6-month period of time that each
patient will be enrolled in the study. I feel this is unlikely. Most
headaches are generated in AM hours, and our selected study subjects will
have sufficient frequency of attacks to ensure that many will reach this
goal. Our statistical analysis did not require that all study subjects
meet the ten-treatment goal.
Rigor of clinical measures
A difficult issue revolved around whether our clinical measures would be
adequate to answer the questions asked. In fact they are more rigorous
than those employed in the studies that established the efficacy of
sumatriptan in migraine treatment (Cady et al., 1991). Again, I am
confident that useful results will be obtained if the study is ultimately
allowed to proceed.
One reviewer wondered how non-responders to sumatriptan might be
characterized, and why they might be better treated with Cannabis. The
initial issue has been studied (Visser et al., 1996). The answer is that
sumatriptan non-responders may be obese, or take the medicine too early.
Beyond that, the study found no features distinguishing responders from
non-responders. I would add one other observation from my clinical
experience: people with chronic daily headache (a difficult subset of
migraine) respond poorly to subcutaneously sumatriptan. Because this
proposal focuses on migraineurs with episodic attacks, CDH patients would
not be accepted for inclusion.
Inadequacy of review process
Finally, I would level criticism of my own at The National Institutes of
Health. Not unexpectedly, none of the reviewers of my protocol were on the
panel of the Workshop on the Medical Utility of Marijuana that proposed
criteria for clinical Cannabis studies. What is surprising, and
unacceptable is that this group was apparently not informed of NIH's own
expressed suggestions for such studies into the medical use of marijuana.
Unfortunately, government agencies have a longstanding tradition of
ignoring their own commissions' recommendations. Additionally, of 29
members of the review team for the Division of Neurological Diseases and
Stroke, only eight were neurologists, and none appear to be headache
specialists. I do know this much: none are members of the American
Association for the Study of Headache, the premier research organization
devoted to the study of migraine. I am, and would have hoped for
examination by a jury of my peers.
As if that were not enough, this proposal was initially assigned to the
AIDS Division of the National Institute on Drug Abuse, although it
pertained to neither HIV nor "abuse," and was not re-assigned until I
pointed out the inherent contradiction. This indicates that the NIH
bureaucracy has been operating as a "split-brain preparation." That is,
the right hemisphere has no idea what the left hemisphere is doing.
In summary, I am extremely disappointed with the repudiation of this
proposal. It has considerably greater merit and validity than the
criticisms would allow. Although I would admit to discouragement, and my
doubts as to how to rectify deficiencies that may not in fact exist, my
research partners and I intend to re- submit this proposal to NIH for the
Contributions of interested parties to MAPS, earmarked for this purpose,
will be most appreciated.
- Cady, R.K., Wendt, J.K., Krichner, J.R., et al., "Treatment of Acute
Migraine with Subcutaneous Sumatriptan," JAMA 265:2831-2835, 1991.
Fisher, M.A., and Glass, S., "Butorphanol (Stadol): A Study in Problems of
Current Drug Information and Control," Neurology 48:1156-1160. 1997.
- Gowers, W.R., A Manual of Diseases of the Nervous System, Blakiston,
- Grinspoon, L. and Bakalar, J.B., Marihuana: The Forbidden Medicine, Yale
Univ., New Haven, 1993. Headache Classification Committee of the
International Headache Society, "Classification and Diagnostic Criteria
for Headache Disorders, Cranial Neuralgias, and Facial Pain," Cephalalgia
8 (suppl. 7):1-96. 1988.
- Linet, M.S., Stewart, W.F., Celentano, D.D., "An Epidemiological Study of
Headache Among Adolescents and Young Adults," JAMA 261:2211- 2216, 1989.
- Mathew, N.T., "Serotonin 1D (5-HT 1D) Agonists and Other Agents in Acute
Migraine," Neurol. Clinics 15:61-83, 1997.
- Mikuriya, T.H. (ed.), Marijuana: Medical Papers 1839-1972, Medi-Comp
Press, Oakland, CA, 1973.
- Osler, W. and McCrae, T., The Principles and Practice of Medicine,
Appleton and Co., New York and London, 1915.
- Peroutka, S.J., "Developments in 5-Hydroxytryptamine Receptor Pharmacology
in Migraine," Neurol. Clinics 8:829-839, 1990.
- Russell, J.R., "Therapeutical Uses and Toxic Effects of Cannabis Indica,"
Lancet 1:637-638, 1890.
- Russo, E.B., "Cannabis for Migraine Treatment: The Once and Future
Prescription: An Historical and Scientific Review," accepted for
publication in Pain, 1998.
- Stewart, W.F., Lipton, R.B., Celantano, D.D., "Prevalence of Migraine
Headache in the United States," JAMA 267:64-69, 1992.
- Visser, W.H., de Vreind, R.H.M., Jaspers, N.H.W.M., and Ferrari, M.D.,
"Sumatriptan-Nonresponders: A Survey in 366 Migraine Patients," Headache
Ethan Russo, M.D.
Department of Neurosciences Western Montana Clinic
515 West Front Street
Missoula, MT 59807-7609