Skip navigation

Cannabinoids for chemotherapy-induced nausea and vomiting

 

Clinical bottom line

Cannabinoids (oral or intramuscular, not smoked) were more effective antiemetics than some conventional drugs, with NNts of 6 and 8 for complete control of nausea and vomiting compared with conventional treatments. Patients preferred cannabinoids, but adverse event withdrawals and some severe adverse events were more common with cannabinoids.


Reference

MR Tramèr et al. cannabinoids for control of chemotherapy-induced nausea and vomiting: quantitative systematic review. BMJ 2001 323: 16-21. ( http://www.bmj.com/cgi/content/full/323/7303/16 )

Review

Searching several electronic databases (10 August 2000 was date of last search) was for cannabinoids (many names)for full publications of randomised controlled trials of cannabinoids in the setting of chemotherapy-induced nausea and vomiting. Extracted was information closest to complete absence of nausea and vomiting in the first 24 hours of chemotherapy.

Results

Thirty randomised trials were found with 1300 patients. The drugs used were mostly nabilone, dronabilone (tetrahydrocannabinol) given orally, or intramuscular levonantradol. Smoked cannabis had not been tested. Most trials were of acceptable quality.

Complete control of nausea and vomiting was statistically superior with cannabinoids over placebo and active comparators with NNts of 3 to 8 (Table 1). More patients preferred cannabinol than placebo or active comparator.

Table 1: Summary of main results

Endpoint

Control

Number of trials

Number of patients with endpoint/Total number of patients

Number-needed-to-treat

     

Cannabis

Control

(95%CI)

Control of nausea and vomiting
Complete control of nausea

Placebo

4

81/116

66/115

8.0 (4.0 to 775)

Complete control of vomiting

Placebo

4

76/116

41/115

3.3 (2.4 to 5.7)

Complete control of nausea

Active

7

122/207

93/215

6.4 (4.0 to 16)

Complete control of vomiting

Active

6

111/194

90/201

8.0 (4.5 to 38)

Patient's self-rating
Preference

Placebo

4

153/202

27/202

1.6 (1.4 to 1.8)

Preference

Active

14

371/604

156/608

2.8 (2.4 to 3.3)

Active=prochlorperazine, metoclopramide, chlorpromazine, tiethylperazine, haloperidol, domperidone, alizapride

Adverse events

Adverse events occurred significantly more often with cannabinoids. Some could be regarded as beneficial (euphoria, or high). Others were harmful, like dysphoria or depression, hallucination and even paranoia, which occurred in 5% of cannabionoid users. Adverse event withdrawal occurred in 11% of cannabinoid users.

Table 2: Adverse events with cannabinoids

Endpoint

Control

Number of trials

Event rate

Number-needed-to-treat

 

Cannabis

Control

(95%CI)

"Beneficial" central side effects
"High" sensation

active or placebo

8

34.5%

3.0%

3.2 (2.8 to 3.7)

Drowsiness, sedation, somnolence

active or placebo

15

50.3%

30.4%

5.0 (4.0 to 6.8)

Euphoria

active

3

14.3%

0.6%

7.3 (5.2 to 12)

Harmful central side effects
Dizziness

active

9

48.5%

16.6%

3.1 (2.6 to 3.9)

Dysphoria, depression

active or placebo

10

12.5%

0.3%

8.2 (6.3 to 12)

Hallucination

active or placebo

10

6.0%

0.0%

17 (12 to 27)

Paranoia

active or placebo

6

4.9%

0.0%

20 (13 to 42)

Haemodynamic side effects
Hypotension

active or placebo

13

24.9%

10.9%

7.1 (5.3 to 11)

Study withdrawal
Due to side effects

active or placebo

19

10.8%

1.6%

11 (8.9 to 14)

Active=prochlorperazine, metoclopramide, chlorpromazine, tiethylperazine, haloperidol, domperidone, alizapride

Comment

Clearly cannabinoids are effective, but not radically different from other treatments. It could be argued that the level of control, with about 40% of patients vomiting with cannabinoids in these trials, is not as good as the 25% seen with 5-HT3 receptor antagonists combined with dexamethasone and is much the same as the 40% vomiting with 5-HT3 drugs alone (see those results here ). It could be argued, and probably rightly, that if equivalent treatments exist that are without the potentially serious adverse events of paranoia, hallucination, depression and hypotension, that cannabinoids are not a good choice. Others will argue differently. In any event, this report is the current evidence, and makes a good read as well.