Sallan, Stephen, et al, "Antiemetics in Patients Receiving Chemotherapy for Cancer." The New England Journal of Medicine. 1980; 302(3): pp. 135-138.
Delta-9-tetrahydrocannabinol (THC) is an effective
antiemetic as compared with placebos in patients receiving chemotherapy
for cancer. In this study we compared THC with prochlorperazine
(Compazine) in a randomized, double-blind, crossover trial with
patients who had failed to benefit from standard antiemetic therapy.
Regardless of the emetic activity of the chemotherapeutic agents, there
were more complete responses to THC courses (in 36 of 79 courses) than
to prochlorperazine (in 16 of 78 courses). Of 25 patients who were
treated with both drugs and who expressed a preference, 20 preferred
THC (P = 0.005). Among patients under 20 years of age there was a
higher proportion of complete responses to THC courses (15 of 20) than
among older patients (21 of 59 courses; P = 0.004).
Increased food intake occurred more frequently with THC (P =
0.008) and was associated with the presence of a "high." Of 36 THC
courses resulting in complete antiemetic responses, 32 were associated
with a high.
We conclude that THC is an effective antiemetic in many patients who
receive chemotherapy for cancer and for whom other antiemetics are
Nausea and vomiting of central origin occur after the administration
of a variety of chemotherapeutic agents for cancer and frequently
constitute the major illness resulting from such treatment. Control
with classic antiemetics has been incomplete and variable (1).
Phenothiazines have traditionally been the drugs of choice in both
the control and prevention of vomiting. Controlled studies of the
antiemetic effects of phenothiazines in patients receiving chemotherapy
for cancer have demonstrated that the phenothiazines are more effective
than placebos but generally do not adequately control nausea and
vomiting (2). Another oral agent, delta-9- tetrahydrocannabinol (THC,
the active ingredient of marihuana), has been shown to have antiemetic
properties superior to those of placebos in reducing vomiting caused by
chemotherapy (3). In addition, THC has been proved to stimulate
appetite and food consumption (4).
The purpose of this study was to compare the effects of orally
administered THC with the effects of prochlorperazine (Compazine), one
of the most commonly prescribed phenothiazine antiemetics, on nausea
and vomiting in patients receiving chemotherapy for cancer.
Eighty-four patients known to have neoplasms were enrolled in the
study. Fifty-one were male and 33 female, and their ages ranged from
nine to 70 years (average age, 32.5 years). All but two patients had
received previous chemotherapy (including doxorubicin,
cyclophosphamide, high-dose methotrexate, cisplatin, bleomycin, and
vinblastine), their nausea and vomiting were inadequately controlled by
conventional antiemetics, including the phenothiazines. Pregnant women
and patients with a history of emotional instability or untoward
reactions to psychoactive drugs were not eligible.
The study was thoroughly explained to the patients. They were told
that they would receive either a conventional antiemetic or a
marihuana-like drug to control nausea and vomiting. Five subjects known
to use marihuana agreed not to smoke it during the course of the study.
THC was supplied by the National Institute on Drug Abuse. The drug
was suspended in 0.12 ml of sesame oil and supplied in gelatin capsules
to blind the study. The THC dosage was 10 mg per square meter of
body-surface area, with 15 mg the amount most commonly administered. On
the basis of our previous experience with variability of response, the
five patients with body-surface areas of less than 1 square meter each
received 10 mg of THC (3). Prochlorperazine tablets were crushed, and
10 mg of the drug was placed with a filler of powdered lactose into
opaque capsules that were identical in appearance to those containing
The experimental plan was a randomized, double-blind design in which
each patient was to receive three one-day courses of a study drug (two
courses with one drug and one course with the other). Neither the
person administering the drug nor the one recording the patient's
response knew which drug the patient received. The study days were
arranged so that chemotherapy regimens were the same for each of the
three courses. If the chemotherapy changed, the patient was not
evaluated for comparison. Each drug course consisted of three doses
given every four hours; the first dose was administered one hour before
chemotherapy, and the other two doses three and seven hours after
chemotherapy,. Each patient was randomized to one of the six distinct
ways of arranging two drugs in a sequence of three trials: TTP, PPT,
TPT, PTP, TPP, or PTT, where T refers to THC and P to prochlorperazine.
In this way any effects of order were balanced.
Nausea, vomiting, food intake, and the development of a "high" were
assessed by the patient on the day after treatment through the use of a
self-administered questionnaire (the questionnaire is available to
readers on request). In addition, the patient, nurses, and other ward
and clinic personnel in contact with the patient were interviewed by
one of the authors, who also reviewed the questionnaire and nurses'
On the basis of responses to the questionnaires, the antiemetic
effect of each study drug was rated as one of the following: complete
response, defined as no nausea or vomiting after chemotherapy; partial
response, a reduction in the severity of nausea and vomiting; or no
response, no reduction in the severity of nausea and vomiting. Food
intake during the 24 hours after chemotherapy was rated by the patient
as no food intake, less than usual, average, or more than usual.
Patients indicated whether or not they had become high and were asked
to describe the condition if they so desired. For the purpose of this
study, a high was characterized as a mood change that consisted of a
tendency to laugh, elation, heightened awareness, mild aberrations of
fine motor coordination, or minimal distortion of activities and
interactions with others.
The significance tests reported in this paper are all two-sided
tests. They are either tests associated with 2-by-2 contingency tables
or the sign test. All P values were computed with exact tests.
To assess the antiemetic activity of the two study drugs, the
analysis was carried out in two parts. Patients who received both
antiemetics were analyzed according to their antiemetic preference,
whereas patients who received only one antiemetic were analyzed by
means of comparison with other patients. The analysis of preference is
more precise because it supplies information on the individual patient,
thus eliminating many variables. The results were further assessed on
the basis of emetic potential of the chemotherapeutic agents
administered and the development of a high.
The study randomized 84 patients; however, not all patients completed the assigned three courses of drug treatment (Table 1).
Twenty-seven patients received only one course and were removed from
the study for various reasons: two died of cancer; four had THC
toxicity, characterized as being too high; one refused to accept the
risk of vomiting with subsequent courses of other antiemetics after
having a complete response to THC; seven had changes in chemotherapy
regimens and thus became ineligible for comparison according to study
criteria; and the remaining 13 patients vomited during the first course
and chose to quit the study. Of these 27 who were removed from the
study, 15 had received THC, and 12 prochlorperazine. Six patients had a
complete response, and nine had no response to THC; with
prochlorperazine, one patient had a complete response, and 11 had no
response. The rate of complete response to THC (six patients of 15) was
significantly higher than the rate for prochlorperazine (one patient of
12; P = 0.05). All six patients who had complete responses to THC
became high, whereas only two of the nine patients who had no response
to THC became high.
Response and preference may depend on the emetic activity of the
chemotherapeutic agents that the patient is taking. Various cancer
drugs and drug combinations were separated into classes A, B, and C;
class A includes the strongest emetics, class B the moderate emetics,
and class C the weakest emetics (Table 2).
The response to THC and prochlorperazine as a function of the total
number of antiemetic courses and class of emetic activity is summarized
in Table 3. This
analysis compares responses between patients and not within individual
patients and thus reflects the antiemetic effect regardless of the
number of study courses completed. Complete responses resulted from 36
of 79 THC courses and 16 of 78 prochlorperazine courses.
The antiemetic preferences of patients who received two or three drug courses are summarized in Table 4.
Of 25 patients who expressed a preference, 20 preferred THC to
prochlorperazine (P = 0.005). The degree of preference for either
antiemetic did not depend on the class of emetic activity of the
patient's chemotherapy for cancer. However, eight patients without an
antiemetic preference took class A treatments, as compared with 13 who
took class B and C treatments. This trend was not statistically
significant (P = 0.14)
The antiemetic response to THC correlated with the development of a high. As summarized in Table 5,
32 of 41 courses associated with a high also resulted in a complete
response, whereas only four of 38 courses without a high resulted in a
complete response (P<0.0001). <p /> The ages of patients in
each antiemetic group having more than one course are the same by
virtue of the experimental design. The average ages of the single-drug
groups are 34.2 years for THC and 33.4 years for prochlorperazine. An
analysis of response by age showed that THC was superior to
prochlorperazine in patients of every age. The rate of complete
response to prochlorperazine, 16 courses of 78, did not change markedly
with age. On the other hand, complete responses to THC numbered 15 of
20 for patients under 20 and 21 courses of 59 for patients 20 or older
(P = 0.004).
Food intake after chemotherapy was evaluated in 28 patients who
completed three courses of antiemetic therapy. Responses were obtained
from the questionnaire and were categorized into one of three outcomes:
more food intake with THC, more food intake with prochlorperazine, or
the same food intake with both drugs. Table 6
summarizes the preferences of the patients according to occurrence of a
high and of increased food intake. Eight of the 28 patients ate more
during the prochlorperazine courses, and 19 ate approximately the same
amount during all courses. We found a very strong association between
increased food intake during the THC course and the presence of a high.
Eight of 16 patients who became high reported an increased food intake
that was conditional on the high; none of the 12 who failed to become
high reported a greater intake (P = 0.008).
The results of this randomized, double-blind study demonstrate that
THC is superior to prochlorperazine as an antiemetic in patients
receiving chemotherapy for cancer. However, our conclusions are
conditional in that the patients in this study were presumed refractory
to conventional antiemetics including the control drugs, the
phenothiazines. Of these presumed refractory patients, 23 had a
complete or partial response to prochlorperazine. This variability of
response is consistent with experimental evidence in the prevention of
apomorphine-induced vomiting in normal human volunteers. In those
volunteers the same doses of apomorphine had different emetic effects,
and the same doses of antiemetics had variable antiemetic effects (5).
Failure to attain a high after ingesting oral THC and the
concomitant absence of an antiemetic effect may be explained by either
pharmacologic or psychologic phenomena. Relatively little is known
about human gastrointestinal absorption of this highly water-insoluble
cannabinoid. Oral administration may result in erratic absorption,
thereby precluding standardization of the THC dose. Plasma levels would
provide important information in determining THC activity. An attempt
to correlate the antiemetic activity of THC with its plasma
concentrations has demonstrated the existence of a relatively steep
dose- response curve (6). Variability in achieving a high has been
reported in subjects who have never used marihuana. Many persons do not
become high on their first exposure to marihuana, as has been reported
anecdotally by regular users over the years (7). It has been
hypothesized that prior pharmacologic sensitization or repeated
exposure to marihuana reduces psychologic inhibition and is necessary
to experience a high state.
We attempted to evaluate the influence of the study day on the
antiemetic effect by examining patients who had different responses to
the same antiemetic drug on different study days. We found that there
was no influence of drug sequences on response, which supports the
findings of Herman et al (8). As in our previous study, the time of
onset and duration of the high were unrelated to previous marihuana use
(3). However, one patient who had used marihuana routinely before the
study found the intensity of the THC high to be much greater than that
previously induced by marihuana, possibly because of the development of
a reverse tolerance (9). He found the THC high unpleasant and refused
further THC treatment after the first study day. It is important to
note that except in the previously mentioned four patients who had THC
toxicity, no side effects of either study drug were reported by any
patient. The absence of reports of somnolence differs from our initial
experience (3). Although we observed somnolence in some patients, we
cannot explain their failure to report it.
Synthetic analogues of THC, such as the cannabinoid nabilone, are
also potent antiemetics. A recent study comparing prochlorperazine with
nabilone showed that nabilone was superior to prochlorperazine in
controlling nausea and vomiting induced by chemotherapy (8). A limited
number of our patients who responded favorably to THC have continued
using the drug during subsequent chemotherapy. Several of these
patients have reported a diminution in the antiemetic effect of THC
with repeated doses. This anecdotal finding contradicts a previous
report (9) of the development of reverse tolerance in persons who use
marihuana often. THC analogues may prove advantageous if they appear
effective with repeated courses of chemotherapy.
We attempted to correlate the degree of emetic activity of the
chemotherapeutic agents with the extent of response to the antiemetic
agents. Although THC was apparently a better antiemetic than
prochlorperazine, no distinction could be drawn between the degree of
antiemetic activity and the degree of emetic stimulation. This finding
agrees with the results from studies of nabilone, (8) indicating that
both THC and nabilone provide antiemetic activity against many
Experimental evidence demonstrating the antiemetic effects of both
chlorpromazine and THC has been reported by ten Ham and de Jong (10).
Using animal models of stereotypic behavior after apomorphine
administration, they compared the activity of THC with that of
chlorpromazine in Wistar rats and homing pigeons. Whereas THC was shown
to antagonize low doses of apomorphine but not high doses,
chlorpromazine appears to have a very strong antiapomorphine effect,
even at high doses of apomorphine. Further studies of the pharmacology
of THC and other cannabinoids in monkeys, dogs, and other laboratory
animals are ongoing. However, it is important to consider species
differences in vomiting behavior. Some laboratory animals, such as
monkeys, are known to be poor models for vomiting, (11) and the
so-called crop vomiting of pigeons may have an entirely different
mechanism from that of the vomiting activity of mammals (12).
In summary, these data demonstrate that THC is a safe and effective
antiemetic for patients receiving chemotherapy for cancer. In many
patients, THC was comparable to the most commonly used phenothiazine,
and THC was preferred by most patients who expressed preferences.
Failure to respond to THC may be due to pharmacologic factors such as
failure of absorption. THC can be used safely at a dosage of 10 mg per
square meter every four hours for at least three doses. Further
investigation is indicated for THC and other cannabinoid derivatives,
alone and in combination with other approaches to the control of
vomiting in patients receiving chemotherapy.
- Browne DC, Sparks R. Vomiting mechanisms: a clinical study of thiethylperazine. South Med J. 1961; 54:953-61.
- Moertel CG, Reitemeier RJ. Controlled clinical studies of orally
administered antiemetic drugs. Gastroenterology. 1969; 57:262-8.
- Sallan SE, Zinberg NE, Frei E III. Antiemetic effect of
delta-9-tetrahydrocannabinol in patients receiving cancer chemotherapy.
N Engl J Med. 1975; 293:795-7.
- Hollister LE. Hunger and appetite after single doses of marihuana,
alcohol, and dextroamphetamine. Clin Pharmacol Ther. 197 1; 12:44-9.
- Shields KG, Ballinger CM, Hathaway BN. Antiemetic effectiveness of
haloperidol in human volunteers challenged with apomorphine. Anesth
Analg (Cleve). 1971; 50:1017-27.
- Chang AE, Shiling DJ, Stillman RC, et al. A prospective evaluation
of delta-9-THC as an antiemetic in cancer patients receiving high dose
methotrexate. Ann Intern Med. (in press).
- Weil AT, Zinberg NE, Nelsen JM. Clinical and psychological effects of marijuana in man. Science. 1968; 162:1234-42.
- Herman TS, Einhorn LH, Jones SE, et al. Superiority of nabilone
over prochlorperazine as an antiemetic in patients receiving cancer
chemotherapy. N Engl J Med. 1979; 300:1295-7.
- Lemberger L, Tamarkin NR, Axelrod J, Kopin IJ.
Delta-9-tetrahydrocannabinol: metabolism and distribution in long-term
marihuana smokers. Science. 1971; 173:72-4.
- ten Ham M, de Jong Y. "Anti-emetic" activity of tetrahydrocannabinol in rats and pigeons. N Engl J Med. 1978; 298:798-9.
- Borison HL. Animal techniques for evaluating antiemetic drugs. In:
Nodin JH, Siegler PE, eds. Animal and clinical pharmacologic techniques
in drug evaluation. Chicago: Year Book, 1964: 418-22.
- Borison HL, McCarthy LE, London SW. Cannabinoids and emesis. N Engl J Med. 1978; 298:1480- 1.
Summary of Patients According to Number of Treatment Courses Completed
|Classification of Patients
||No. of Patients
|Ineligible for study
||4 vomited before chemotherapy; insufficient data for 7
|Completed 1 course
||15 received THC; 12 received prochlorperazine
|Completed 2 courses
||Each patient received THC and prochlorperazine
|Completed 3 courses
||56 courses of THC; 58 courses of prochlorperazine
Classification of Chemotherapeutic Agents According to Emetic Activity
||Greatest emetic activity
Combinations of agents inclucing cisplastin, dacarbazine, doxorubicin, and cyclophosphamide
||Moderate emetic activity
Combinations of agents including high-dose methotrexate, cyclophosphamide, doxorubicin, and actinomycin D.
Cisplatin and high-dose actinomycin D as single agents
||Least emetic activity
Single agents including high-dose methotrexate, cyclophosphamide, and doxorubicin
Response of Patients to Antiemetics According to Emetic Activity of Patients' Chemotherapy
No. of Courses
Summary of Antiemetic Preferences of Patients Receiving Two or Three Courses of Antiemetics
|Emetic Activity of Chemotherapy
No. of Patients
Response to THC Related to Development of High
No. of Courses
Antiemetic Preference Related to Food Intake and Development of a High
|Increased Food Intake
||Same Food Intake
||Increased Food Intake
||Same Food Intake
no. of patients
Copyrighted material. Reprinted by permission.