As detailed in the “Special considerations” section, clinical trials for cannabis in the treatment of neurologic disease are quite limited in number and quality due to a combination of legal restrictions placed on cannabis, variable study design, and frequent lack of blinding. With that in mind, the following is a summary of pertinent animal and human trials of cannabis for selected neurologic diseases.
Multiple sclerosis. The evidence of effect and benefit of cannabis in patients with neurologic disease is arguably the most robust in patients with multiple sclerosis. In animal models with experimental autoimmune encephalomyelitis, THC reduced spasticity and tremor in mice models. Similar effects can be seen by inhibiting endocannabinoid reuptake or degradation in the spinal cord, whereas administration of cannabinoid receptor antagonists can worsen spasticity in these mice. Cannabinoid receptor agonists can decrease spasms and pain in mice models. In humans, high concentrations of CB1 are seen on primary afferent nociceptors, whereas CB2 receptors have been implicated in regulating inflammatory pain.
In a survey of 112 patients with multiple sclerosis who self-medicated with cannabis, more than 90% noted improvement in spasticity, pain, and tremor (10). In a separate survey of 225 patients with multiple sclerosis, 54.3% approved of cannabis use, whereas 33.2% were neutral; 19.5% of patients admitted use of it, whereas an additional 50.2% stated they would use it if it were legal (05). The Cannabinoids in Multiple Sclerosis (CAMS) trial was a randomized controlled study of 630 patients with either relapsing-remitting multiple sclerosis, primary progressive multiple sclerosis, or secondary progressive multiple sclerosis receiving THC and cannabidiol versus dronabinol versus placebo. It demonstrated no objective change in primary outcome (spasticity) between groups, but significant subjective improvement in spasticity, pain, sleep, and spasms. Since then, numerous large, randomized, and often double-blind European trials using the THC/CBD combo drug nabiximols (aka oral cannabis extract, OCE) have also shown significant subjective improvement primarily in spasticity and pain, without changes in disability measures.
A 2014 AAN summary of evidence-based guidelines for alternative medicine for the treatment of multiple sclerosis examined 9 studies: 3 class I, 2 class II, and 4 class III, most lasting 6 to 15 weeks (40). The authors concluded that OCE was effective for reducing subjective spasticity symptoms and pain, with benefit maintained for possibly 1 year. Similar effects are possible with use of oral THC or oromucosal cannabinoid spray. Of note, the authors felt it was probably ineffective for objective spasticity and tremors. They suggested clinicians may offer OCE to patients to reduce patient-reported symptoms of spasticity and pain (Level A). These recommendations were based on oral and inhaled cannabis, but inadequate data were found to support smoked cannabis for any indication.
A 2014 AAN systematic review of cannabis for the treatment of select neurologic disorders found similar results as the CAM study and AAN alternative medicine study (23). The authors concluded that for spasticity in multiple sclerosis, OCE is effective in reducing patient-reported scores (2 class I studies), probably ineffective for objective measures of spasticity at 12 to 15 weeks (1 class I study), but possibly objectively effective at 1 year (1 class II study). Similar conclusions were reached for THC and nabiximols. For patients with multiple sclerosis and painful spasms or central pain, OCE is effective (2 class I studies), and THC/nabiximols are probably effective (1 class I study each). Also, THC and OCE are probably ineffective (1 class I study), and nabiximols possibly ineffective (1 class II study), for treating multiple sclerosis–related tremor. It is worth emphasizing that in all studied aspects of the effects of cannabis on multiple sclerosis symptoms, there was no convincing evidence of benefit with inhaled/smoked cannabis.
Epilepsy. Rodent models have shown that there is increased expression of CB1 receptors in the short term after an event of status epilepticus, with similar findings in febrile seizure rodent models. This has been interpreted to suggest that there is a compensatory neuroprotective effect to suppress neuroexcitability after seizures. Treatment of epileptic mice with THC or similar CB1 receptor agonists reduce seizure frequency to a level even better than mice treated with phenytoin and phenobarbital, whereas the cannabinoids 2-arachidonoyl glycerol and arachidonyl ethanolamide reduce the frequency of both spontaneous and miniature excitatory postsynaptic currents in epileptic tissue only. To further support the neuroprotective hypothesis of the cannabinoid system, a number of additional animal studies have shown that cannabinoid receptor antagonists lower seizure threshold and increase seizure frequency and duration in epileptic mice in the short term, ie, days. However, longer term use may result in development of tolerance or loss of protective effect over time. Studies of human chronically epileptic tissue (vs. controls) have shown a decrease in CB1 receptors and endocannabinoid ligands at the hippocampus, suggesting that over time the compensatory neuroprotective response is suppressed as part of the disease state in chronic epilepsy, predisposing to higher likelihood of seizures (19; 39; 26; 20).
A Cochrane review in 2012 gathered all RTCs evaluating the use of cannabis in epilepsy (15). Primary outcome in this review was seizure freedom for at least 12 months, whereas secondary outcomes were seizure relief for 6 months and any potential adverse events. The review found only 4 small (the largest study was 15 patients) “low-quality, high-bias” randomized trials of short duration taking place between 1978 and 1990, none of which explored the review’s desired primary or secondary outcomes. The only data that could be gleaned were that 200 to 300 mg doses of cannabidiol “may” be safe. The 2014 AAN review of cannabis for treatment of epilepsy did not find sufficient data to support or refute its effect on epilepsy (no class I-III studies) (23).
A randomized, double-blind, placebo-controlled trial of 120 children (average age 10) with Dravet syndrome receiving either cannabidiol oral solution (20 mg/kg/day) versus placebo found that the number of convulsive seizures per month in the treatment group went from 12.4 to 5.9 versus 14.9 to 14.1 in the placebo group. Five percent of the treatment group became seizure-free versus 0% in placebo group. There were no differences in the rate of nonconvulsive seizures. Side effects included diarrhea, nausea and vomiting, fatigue, loss of appetite, somnolence, and abnormal liver function tests (LFTs) (on patients concomitantly on valproic acid, which was transient) (12).
A randomized, double-blind, placebo-controlled trial of 171 children (median age 15, 34% over age 18) with Lennox-Gastaut syndrome receiving either cannabidiol oral solution (20 mg/kg/day) versus placebo found that there was a 44% reduction in seizures in the treatment group versus 22% in the placebo group. There was a similar side effect profile to a study by Devinsky and colleagues (12). There was 1 death in the CBD group, which was felt to be unrelated to the treatment.
Epilepsy patients often perceive cannabis as decreasing their seizure frequency. For example, 1 study surveyed 457 adult epilepsy patients, of which 18% were using cannabis; of that, 68% reported decreased frequency of seizures on cannabis (37). The 2016 American Epilepsy Society position statement on cannabis for the treatment of epilepsy notes the following (02):
| || |
The anecdotal reports of positive effects of the marijuana derivative cannabidiol for some individuals with treatment-resistant epilepsy give reason for hope. However, we must remember that these are only anecdotal reports, and robust scientific evidence for the use of marijuana is lacking. The lack of information does not mean that marijuana is ineffective for epilepsy. It merely means that we do not know if marijuana is a safe and effective treatment for epilepsy, which is why it should be studied using the well-founded research methods that all other effective treatments for epilepsy have undergone. Such safety concerns coupled with a lack of evidence of efficacy in controlled studies result in a risk/benefit ratio that does not support use of marijuana for treatment of seizures at this time...research is desperately needed. AES calls on government and private funders to support well-designed clinical research into all promising treatments for epilepsy. To increase clinical research into the effectiveness and safety of marijuana as a possible treatment for resistant epilepsy, the American Epilepsy Society urges that marijuana’s status as a Federal DEA Schedule 1 controlled substance be reviewed… At present, the epilepsy community does not know if marijuana is a safe and effective treatment nor do we know the long-term effects that marijuana will have on learning, memory and behavior, especially in infants and young children…
Parkinson disease. The effect of the endocannabinoid system within the basal ganglia is complicated and not clearly understood. CB1 receptors are found throughout the subcortex, involving both the direct and indirect pathways, and coexpressed with both D1 and D2 receptor locations. They are localized mainly on medium spiny GABAergic neurons of striatum, concentrated in dendrites and presynaptic axon terminals innervating globus pallidus externa and interna, as well as substantia nigra pars reticulata and compacta. Depending on the region in which they are expressed, the endocannabinoid receptors may serve to inhibit glutamate release, as well as either promote or inhibit GABA activity in each of these regions. As such, it is not clear whether the endocannabinoid system would even support a hypokinetic or hyperkinetic state, let alone more nuanced effects on individual regions.
Basic research on cannabis’s effects on Parkinson is limited and conflicting. Rodent and primate Parkinson disease models show significant overactivity of the endocannabinoid system in the striatum, with these changes reversing in the setting of levodopa therapy. Increased CB1 receptors are also seen in the striatum of postmortem Parkinson disease patients. Whether these findings suggest that endocannabinoids are contributing to parkinsonism or are a compensatory reaction to the disease is not clear. Other Parkinson disease animal models show increased cannabinoid ligand 2-AG in the globus pallidus external, suggesting inhibition of a portion of the direct pathway by the endocannabinoid system. Some studies have suggested reduction in CB1 receptor signaling may be associated with dyskinesias, whereas rodent studies have suggested that cannabinoid agonists can induce catalepsy and akinesia (31; 07; 13).
Similarly, the clinical effects are also conflicting. A study examining Unified Parkinson's Disease Rating Scale (UPDRS) scores of 22 Parkinson patients at baseline and 30 minutes after cannabis inhalation (without placebo) showed improvement of mean total UPDRS score (p < 0.004), with subsets showing significant improvement in rigidity (7.6 to 6.5), tremor (7.6 to 3.6), and bradykinesia (13.1 to 8.6) (25). There was also significant improvement of sleep and pain scores. These effects lasted 2 to 3 hours. In an uncontrolled survey of 399 Parkinson patients taking cannabis for months, 46% reported some symptomatic improvement overall (45% in rest tremor, 38% in bradykinesia, and 14% in dyskinesias), though 4% felt there was worsening of symptoms (38). On the other hand, a small randomized, double-blind, placebo-controlled trial of 14 Parkinson disease patients using cannabis showed no benefit in UPDRS, no evidence of a neuroprotective effect, and no definite improvement in quality-of-life measures (09). A small randomized, double-blind, placebo-controlled exploratory trial of a CB1 antagonist did not demonstrate improvement in motor function (29).
With regard to the clinical effect of cannabis on dyskinesias, a randomized double-blind, placebo-controlled crossover study of 17 patients with Parkinson disease receiving oral cannabis versus placebo for 4 weeks did not demonstrate improvement in UPDRS or dyskinesias, both objectively or subjectively, though 65% of patients did not achieve target level of cannabis in their blood (08). In contrast, another randomized double-blind, placebo-controlled crossover trial of 7 patients showed a significant reduction in dyskinesias with a cannabinoid receptors agonist; however, 2 of the patients had a disproportionately higher benefit compared to the mild effect of the other 5 (36). The 2014 AAN review concluded that OCE is probably ineffective for treating levodopa-induced dyskinesias in patients with Parkinson disease (1 class I study) (23).
Headache. A number of studies in animal models have demonstrated that cannabinoid agonists act to potentiate the activity of 5-HT1A receptors (the same effect as a number of migraine abortive medications, including triptans) and antagonize 5-HT2 receptors (the same effect as a number of migraine prophylactics) (34). Midbrain periaqueductal grey matter, a proposed site of migraine generation, has been shown in rodents to be modulated by endocannabinoids (27). Cannabinoids have been shown to inhibit neuronal firing in the trigemino-cervico-vascular complex and inhibit local inflammatory production as well as vasodilatation; these effects are reversed with cannabinoid antagonists (01). Other studies in rodents have shown a dose-dependent suppression of cortical-spreading depression by application of THC (and similar agonists) (22). Analysis of CSF in chronic migraine sufferers showed AEA ligand levels lower than those of normal controls, raising the question of endocannabinoid deficiency in these patients (35).
Unfortunately, there are no rigorous and well-controlled studies exploring cannabis for treatment of headache. Historically, there are similar anecdotal reports dating back hundreds to thousands of years, as noted earlier. A number of small case reports and survey studies suggest a possible benefit in acute migraine/headache, but these studies are weak and few. Conversely, some case reports suggest a worsening of headaches after cannabis intake.
Miscellaneous. The 2014 AAN review on cannabis and selected neurologic diseases did not find sufficient evidence to support or refute the use of cannabis for Huntington disease, tics/Tourette syndrome, or cervical dystonia, subsequent reviews have also not found sufficient evidence to support its use in these conditions (23; 03).
Table 1. Guideline Positions and Author’s Position on Cannabis for Specific Neurologic Diseases
Guideline positions on cannabis for specific neurologic diseases
Author’s position on cannabis for specific neurologic diseases
Multiple sclerosis (MS)
• Spasticity in MS: OCE, THC, and nabiximols are effective in reducing patient-reported scores, probably ineffective for objective measures of spasticity at 12-15 weeks, but possibly effective for objective measures at 1 year.
• Painful spasms or central pain in MS: OCE is effective and THC/nabiximols are probably effective.
• MS-related tremor: THC and OCE are probably ineffective. Nabiximols is possibly ineffective.
Same as guideline position
For bladder dysfunction related to multiple sclerosis, nabiximols is probably effective for void reduction.
OCE is probably ineffective for treating levodopa-induced dyskinesias (23).
Weak and/or conflicting basic science and clinical research does not support cannabis use in Parkinson disease. Furthermore, this patient population is particularly vulnerable to cannabis side effects (psychosis, fatigue, memory problems).
• The current risk-benefit ratio does not support use of cannabis in epilepsy (02).
• Data are insufficient to support or refute the efficacy of cannabinoids for reducing seizure frequency (23).
Compelling basic science research and some well-done clinical trials in children with Dravet syndrome and Lennox–Gastaut syndrome showed benefit but otherwise poor clinical research overall. There are currently inadequate data to support use of cannabis in epilepsy (broadly), but it may be beneficial in specific pediatric epilepsy conditions.
No guidelines exist.
Compelling basic science research, but poor clinical research. There are currently inadequate data to support use of cannabis for headache.
Other selected neurologic diseases
• Huntington disease: Data are insufficient to support or refute efficacy of cannabis for Huntington disease.
• Tourette syndrome: Data are insufficient to support or refute efficacy of THC for reducing tic severity.
• Cervical dystonia: Data are insufficient to support or refute the efficacy of dronabinol (23).
Poor basic and clinical research in other neurologic diseases. There are currently inadequate data to support use of cannabis for other neurologic diseases.