Sleep and Anxiety in Cancer and CBD Oil: Research and Doses

What does the research say?

In a study of 144 people with advanced cancer, pure CBD oil was compared with placebo. There was no clear difference between the two groups in total symptom burden, quality of life, anxiety or depression. In a similar study using an oil with equal amounts of THC and CBD, total symptom burden, anxiety, depression and quality of life were also no different from placebo (on the quality-of-life scale, only the pain domain favored the THC oil); general well-being was better in the placebo group.

In a study of 50 women with advanced breast cancer and significant anxiety, a single dose of CBD was given before an imaging scan. On the main measure there was no difference from placebo; on a secondary measure, anxiety 2–4 hours later was lower in the CBD group. This is a small, single finding.

In a study of 20 people with a brain tumor and significant anxiety, patients took 600 mg of pure CBD a day for 3 weeks and placebo for 3 weeks. Anxiety and depression symptoms went down more during the placebo period; there was no sign that CBD worked. The study was stopped early because too few patients could take part.

No pure CBD study targeting sleep in cancer as its main goal was found. In an open-label study of 39 women with breast cancer who took CBD for 15 weeks for joint pain, sleep and anxiety did not change when all participants were analyzed together; improvement was seen only in those who completed treatment. In analyses pooling cancer pain studies of THC-containing products, sleep either did not change at all or improved very slightly. In two small studies using CBD-dominant capsules with a little THC in patients with chemotherapy-related nerve damage, sleep was no better than with a dummy treatment (placebo); in one of them, sleep went worse in those taking cannabis. In a small 1974 study of 12 anxious patients receiving radiation therapy for lung cancer, THC improved night-time sleep but also caused tiredness and confusion; the study's abstract does not say whether anxiety changed.

Studies that used quality of life as their main measure show a similar picture. In 56 patients receiving radiation therapy for head and neck cancer, the THC-like drug nabilone did not delay the worsening of quality of life; mood and sleep were also no different from a dummy treatment (placebo). In 243 patients with advanced cancer and loss of appetite and weight, cannabis extract or THC did not improve quality of life more than placebo. An analysis pooling studies in cancer and nervous system diseases found no effect on quality of life; in a pooled analysis of cachexia studies, quality of life improved slightly less in those taking cannabinoids than in those taking a dummy treatment or an appetite-stimulating drug (megestrol). In a brain tumor study that compared two different THC:CBD ratios and had no placebo group, the total quality-of-life score did not change, and sleep improved in both groups.

Registries following cancer patients who use medical cannabis in Israel, Canada, the US, the UK and Thailand reported that anxiety, sleep and quality of life improved over time. In a US program of about 2,000 patients, those taking higher CBD doses had slightly greater improvement in sleep and anxiety (most of these patients were using THC-dominant products); in a Canadian registry, on the other hand, CBD-dominant products did not bring clear relief. In about 2,900 cancer patients in Thailand, quality of life and anxiety scores also improved over 3 months with an oil providing only about 0.2 mg of THC a day; this dose was far below those used in controlled studies, and sleep was not measured separately in this study. These studies had no comparison group, patients knew what they were taking, and those who stopped treatment or died were left out of the results; so it is unknown how much of the improvement came from cannabis. Some were run by researchers linked to organizations that make or prescribe cannabis products.

In their guidelines, MASCC, the international association working on supportive care for people with cancer (2023), and the American Society of Clinical Oncology (ASCO, 2024) stated that there is not enough evidence to recommend cannabis or cannabinoids for anxiety, depression or sleep problems in cancer. ASCO also said that, outside a clinical trial, oral CBD at doses of 300 mg a day or higher should not be recommended to cancer patients for symptom burden; its reasons were that no benefit has been shown and that liver enzymes can rise reversibly (low-quality evidence, weak recommendation). This does not mean that lower doses work: no CBD dose has been shown to help with symptom burden in cancer.

Strength of evidence: ●●○○ Weak evidence. Finding: No effect found. What does this mean?

Ongoing research on this topic

Trials testing CBD for this topic on ClinicalTrials.gov: 3 under way (1 recruiting); results are awaited from 1 that ended in the last two years. A trial being under way does not show that CBD works.

NCT06266611: Phase 2, recruiting, target 185 participants
NCT03948074: Phase 2, completed, results awaited, 91 participants
NCT07565974: Phase 2, not yet recruiting, target 69 participants

All trials on this topic and their history

Possible side effects and drug interactions

In the pure CBD oil study, side effects were similar in the two groups; shortness of breath was reported slightly more often in the CBD group. In a joint review of two studies, no significant rise in liver enzymes was seen in cancer patients over 28 days of use. With the THC-containing oil, confusion, feeling “high” and an exaggerated sense of well-being were more common. In analyses pooling cancer studies, psychiatric side effects (e.g., hallucinations, nightmares, suspiciousness) were more common with THC-dominant products; anxiety events were more frequent with high-dose synthetic THC than with low doses.

CBD can slow the breakdown of some medicines in the body. Blood levels of citalopram, used for anxiety and depression, rose in patients who added CBD. Drowsiness may increase when it is taken with medicines that cause sleepiness. Levels of some cancer drugs, such as everolimus, may also rise.

Interactions with medicines used for this condition

Interactions reported for these medicines in peer-reviewed publications and official prescribing information. A medicine that is not listed can still interact with CBD; tell your doctor about every medicine you take.

Sleep and Anxiety in Cancer: medicines used and their interactions with CBD
Medicine	Importance	Warning	Source
Everolimus	High importance	Everolimus, sirolimus: In a controlled study in healthy volunteers taking approved CBD, the blood level of everolimus rose about 2.5-fold; in patients with TSC (tuberous sclerosis complex), everolimus and sirolimus levels rose in most patients when CBD was added. The prescribing information for both medicines recommends close monitoring of blood levels with CBD, and lowering the dose if needed. Do not start without talking to your doctor; your drug level should be monitored.	EMA Epidyolex SmPC 4.5, PMID 37132402, PMID 31924480, Everolimus (Afinitor) prescribing information §2.11 and §7.1, dailymed.nlm.nih.gov, Everolimus (Zortress) prescribing information §5.22 and §7.1, dailymed.nlm.nih.gov, Sirolimus prescribing information §5.21 and §7.5, dailymed.nlm.nih.gov
Warfarin	High importance	Warfarin: If you take warfarin, tell your doctor before starting CBD. Case reports have described INR rising and the warfarin dose needing to be lowered; more frequent INR monitoring is recommended.	PMID 29387536, PMID 33016308, PMID 30326170
Immunotherapy (checkpoint inhibitors, e.g., nivolumab)	Caution	Immunotherapy (checkpoint inhibitors): In some observational studies, cannabis users had a lower response to immunotherapy or shorter survival; other studies found no difference, and the results conflict. Cause and effect has not been proven, most studies were done with cannabis containing THC, and there are no data specific to pure CBD. Guidelines advise caution; do not start without talking to your oncologist.	PMID 30670598, PMID 32872248, PMID 35454955, PMID 38610922, PMID 39737331, PMID 36535195, PMID 40184324, PMID 38478773, PMID 36872397
Tamoxifen	Caution	Tamoxifen: In one case, the active breakdown product of tamoxifen (endoxifen) was found to be lower while the person was taking 40 mg of CBD a day. In a small study of 15 patients taking tamoxifen, about 50 mg a day of THC-free CBD oil lowered endoxifen by 12.6% on average; the researchers did not consider this difference important at this dose, and the effect of higher doses is not known. Do not start without talking to your oncologist.	PMID 33191836, PMID 37543688
Citalopram (SSRI) (Cipram, Citol, Relaxol), Escitalopram (SSRI) (Cipralex, Citoles, Eslorex)	Caution	Citalopram, escitalopram: CBD raised citalopram blood levels: a significant rise in a series of 6 patients; in healthy volunteers, citalopram exposure (AUC) rose by 43% after 12 days of approved CBD (2.5 mg/kg twice a day). If you take these medicines, talk to your doctor.	PMID 34121064, PMID 41622700
Chemotherapy / cancer medicines (general)	Information	Chemotherapy and cancer medicines: Data on interactions between CBD and cancer medicines are very limited. Talk to your oncologist before using it.	PMID 30776990, PMID 38478773
Apixaban (DOAC), Rivaroxaban (DOAC)	Information	DOAC blood thinners: If you take a blood thinner, tell your doctor about CBD. There are no clinical data on interactions between these medicines and CBD; there is only a general warning in the package leaflet.	dailymed.nlm.nih.gov

To check interactions with the medicines you take: CBD drug interaction checker.

How was CBD used in studies?

In the pure CBD oil study, the dose ranged from 50 mg to 600 mg a day (study dose; not a personal recommendation) and was increased every few days; half of the patients chose a dose of 400 mg a day (study dose; not a personal recommendation) or less. In the pre-imaging anxiety study, a single dose of 400 mg (study dose; not a personal recommendation) was used.

In the oil containing THC and CBD, the dose was adjusted over two weeks from 2.5 mg each of THC and CBD a day up to 30 mg each (study dose; not a personal recommendation). There is no information on pure CBD doses for sleep in cancer.

Study details (for health professionals)

In advanced cancer, CBD oil did not improve total symptom burden, quality of life or anxiety more than placebo (MedCan1). A 1:1 THC:CBD oil also made no difference in total symptom burden. In studies of a THC+CBD spray for cancer pain, the Cochrane 2023 meta-analysis found no difference from placebo in sleep problems (4 RCTs, 1,314 people; SMD −0.06; moderate certainty); a 2022 meta-analysis calculated a 'very small' improvement (−0.19 cm on a 10 cm scale). In two small RCTs of CBD-dominant, low-THC capsules for chemotherapy-induced nerve damage, sleep was no better than placebo; in one, it went worse with cannabis. For anxiety before a scan, a single 400 mg dose of CBD was ineffective on the primary outcome. In patients with brain tumors, 600 mg/day of pure CBD for 3 weeks did not reduce anxiety and depression symptoms more than placebo (a small crossover RCT stopped early). There is no randomized THC study in cancer that targets psychological symptoms as its primary goal; there is only a small placebo-controlled pilot from 1974 in 12 anxious lung cancer patients, with randomization not stated (THC improved night-time sleep but caused tiredness and confusion; its effect on anxiety was not reported in the abstract, and there are no numerical results). In two THC/nabilone RCTs that used quality of life as their main measure (243 and 56 patients), there was no difference from placebo; in a pooled analysis of cachexia RCTs, quality of life improved slightly less with cannabinoids than in those taking placebo or an appetite-stimulating drug (megestrol) (SMD −0.25). For psychological symptoms in cancer, the MASCC 2023 guideline reviewed 15 RCTs and wrote that none of them used these symptoms as the primary outcome and that there is no high-quality evidence sufficient to recommend cannabis; a 2024 meta-analysis also found no clinically meaningful effect on depression, anxiety or stress. Patient registries in Israel, Canada, the US, the UK and Thailand reported improvement from baseline in anxiety, sleep and quality of life; but these studies had no comparison group, in most of them patients mainly used THC products, patients who died or stopped were not included in the analysis, and some were run by authors linked to organizations that make or prescribe cannabis products. In the Thai cohort (2,873 cancer patients; quality of life +0.13, anxiety −1.46 points; sleep not measured separately), the daily amount of THC in the oil used was only about 0.2 mg, and the public agency that makes the product funded the study; the authors also point to the possible share of the placebo effect. October 2026 search (PubMed, Cochrane, ClinicalTrials.gov): 13 studies with cannabinoids that had a comparison group (11 placebo-controlled, 1 waitlist-controlled, 1 comparing two THC:CBD ratios) and 8 systematic reviews were found on this topic; all are in the table below. The table also includes 1 open-label study, 8 observational studies and 2 safety sub-analyses of the same trials. Four cancer pain RCTs that examined sleep only as a secondary measure were not listed one by one; their sleep results are included in pooled form in the Cochrane review in the table.

Evidence assessment: Weak (pure CBD: anxiety and quality of life); there is no pure CBD study that targets sleep as its main goal; all 3 placebo-controlled studies with pure CBD found no difference from placebo on their main measure

CBD alone

Pure CBD oil was no different from placebo for anxiety, quality of life or total symptom burden in advanced cancer; in brain tumors, 600 mg/day also did not reduce anxiety; in breast cancer, a single 400 mg dose before imaging was no different from placebo for anxiety on the main measure (in an exploratory comparison, the anxiety level after the dose was lower in the CBD group); there is no pure CBD study that targets sleep as its main goal (in a small open-label study done for joint pain, sleep and anxiety did not change when all participants were analyzed together). In real-world data, sleep and anxiety improved slightly more in those taking higher CBD doses, but most of these patients were using THC-dominant products and there was no comparison group; in one patient registry, CBD-dominant products did not bring clear relief.

Products containing THC

A 1:1 THC:CBD oil did not change total symptom burden; it was also no different from placebo for anxiety, depression or quality of life (on the quality-of-life scale, only the pain domain favored the THC oil), and general well-being improved more with placebo on day 14. In analyses pooling placebo-controlled studies for cancer pain, there was either no difference in sleep (Cochrane 2023: 4 studies, 1,314 people) or a very small improvement (2022 meta-analysis: −0.19 cm on a 10 cm scale); in two small studies of CBD-dominant, low-THC capsules for nerve damage, sleep was no better than placebo, and in one it was worse. In studies that used quality of life as their main measure, THC, cannabis extract or the THC-like drug nabilone was no different from placebo; in a pooled analysis of cachexia studies, quality of life improved slightly less in those taking cannabinoids than in those taking placebo or an appetite-stimulating drug (megestrol). Real-world registries reported improvement in anxiety, sleep and quality of life, but there was no comparison group.

Evidence: Moderate for sleep (effect very small or absent; from cancer pain studies); very low for anxiety and depression (no randomized study used these symptoms as its main measure; studies that looked at them as a secondary measure found no significant difference from placebo); low for quality of life (no benefit or a small negative effect)

Dose range in studies

Pure CBD: 50–600 mg/day (MedCan1, median chosen dose 400 mg/day), 600 mg/day in brain tumors (3 weeks), single dose 400 mg (Nayak 2024), 200 mg/day for joint pain in breast cancer (15 weeks, open-label; Fleege 2025); THC-containing 1:1 oil: from 2.5+2.5 mg up to 30+30 mg/day; THC or cannabis extract (2.5 mg THC, + 1 mg CBD in the extract) 2 times a day (Strasser 2006, Brisbois 2011); nabilone starting at 0.5 mg, up to 2 mg/day (Côté 2016)

Starting dose and titration

MedCan1: increased every 3 days; MedCan2: increased over 14 days according to tolerance/effect

These are doses researchers gave under medical supervision; they are not a recommended dose for you. "mg/kg/day" means milligrams per kilogram of body weight per day.

Sleep and Anxiety in Cancer and CBD: studies reviewed
Study	Participants	Product	Dose and titration	Duration	Result	Source
Hardy 2023 (MedCan1-CBD) Phase IIb, double-blind, placebo-controlled, dose-escalation RCT	144 randomized (58 CBD / 63 placebo in the primary analysis) Advanced cancer, palliative care, ESAS ≥10/90	Synthetic CBD oil 100 mg/mL (CBD made by synthesis, not from the plant; as described in the MedCan2 publication) and placebo oil	Starting at 0.5 mL once a day (50 mg/day) → 2 mL 3 times a day (600 mg/day); median chosen by patients 400 mg/day Increased every 3 days (ASCO 2024 text)	28 days (primary: day 14)	On day 14, the change in total symptom distress score was −3.0 with CBD and −6.2 with placebo (p=0.24, no difference). Responder rate 44.8% vs 58.7% (p=0.13). No effect on quality of life, depression or anxiety (on day 28, change in anxiety CBD −1.10, placebo −0.79; depression −0.50 and −0.63; as reported by Cochrane 2023); sleep was not measured separately. The share of patients who felt better or much better on day 14 was 53% in the CBD group and 65% in the placebo group; on day 28, 70% and 64% Side effects: No difference between arms; dyspnea more frequent with CBD	PMID 36409969, doi:10.1200/JCO.22.01632, ANZCTR ACTRN12618001220257, protocol: Good 2019, PMID 31810437, PMID 38307703, disease progression and survival)
Hardy 2025 (MedCan2) Double-blind, placebo-controlled RCT	144 randomized; 121 in the day-14 analysis (56 THC:CBD, 65 placebo) Advanced cancer, palliative care	Oil: THC 10 mg + CBD 10 mg/mL (1:1)	Starting at 0.25 mL/day (2.5 mg THC + 2.5 mg CBD) → 3 mL/day (30 mg + 30 mg); median dose chosen by patients 1.5 mL (15 mg THC + 15 mg CBD/day) Increased over 14 days according to tolerance/effect	28 days	No difference in total symptom distress (p=0.76). Small benefit in ESAS pain (−1.42 vs −0.46; p=0.04), at the cost of more psychotomimetic toxicity. No difference in anxiety (−1.46 vs −1.17; p = 0.48) or depression (p = 0.21) scores; general well-being improved more in the placebo group on day 14 (baseline-adjusted p = 0.02; no difference left on day 28). On the quality-of-life scale (EORTC QLQ-C15-PAL), only the pain domain favored THC:CBD; no difference on the depression-anxiety-stress scale (DASS). On day 28, the share who felt better was higher in the THC:CBD group (88% vs 66%; p = 0.05), but those who stopped early were not included in this analysis. No opioid-sparing effect Side effects: Confusion, feeling 'high', exaggerated sense of well-being more frequent; most grade 1–2	PMID 40705150, doi:10.1007/s00520-025-09763-5, ANZCTR ACTRN12619000037101, funding from the Australian government (MRFF), the manufacturer supplied the drug and placebo and had no role in the study
Nayak 2024 Phase 2, double-blind, placebo-controlled RCT	50 randomized (25 CBD, 25 placebo) Women with advanced breast cancer and clinical anxiety	FDA-approved CBD (oral, pure)	400 mg single dose None	Single dose before a scan (imaging)	The primary outcome (change on the fear subscale from before to 2–4 hours after the dose) did not differ between groups (−19.1 vs −15.0; p = 0.37). In a secondary (exploratory) comparison, the anxiety score 2–4 hours after the dose was lower in the CBD group (51.5 vs 58.0; p = 0.02); this was not the prespecified main outcome Side effects: No grade 3–4 toxicity. Mild or moderate side effects in 2 people in the CBD group: grade 1 nausea in 1 person, grade 2 nausea and pain in 1 person (possibly drug-related according to the authors); in the registry's 1-week side-effect table, 1 nausea and 1 pain in the CBD group and no side effects in the placebo group	PMID 39680411, doi:10.1001/jamanetworkopen.2024.50391, NCT04482244
Belgers 2026 Double-blind, placebo-controlled crossover RCT; stopped early because of low enrollment (research letter)	20 randomized (15 completed) Adults with a primary brain tumor and stable disease who had significant anxiety	Pure CBD tablet (>99% CBD, <0.1% THC)	600 mg/day	2 × 3 weeks (at least 2 weeks of washout)	Anxiety and depression symptoms fell more with placebo; the probability that CBD was beneficial was 19% for anxiety and 11% for depression (Bayesian analysis); at the end of 3 weeks, significant anxiety was present in 50% after placebo and 59% after CBD Side effects: Side effects mild and similar (CBD 3 events, one of them grade 3; placebo 6 events); 1 rash during CBD	PMID 42750778, doi:10.1093/nop/npag025, Netherlands Trial Register NL9623 (GRASS study). Funding: Anita Veldman Foundation, the authors reported no conflicts of interest
Davies 1974 Placebo-controlled crossover pilot study (randomization and blinding not stated in the abstract)	12 (two groups of 6) Anxious patients receiving radiation therapy in hospital for inoperable lung (bronchial) cancer (United Kingdom)	Oral THC (named Δ1-trans-THC in the publication, Δ9-THC in today's naming) and placebo	10 mg/day, 7 days None (fixed dose)	2 × 7 days (with a 7-day gap between them)	THC caused drowsiness, improved night-time sleep and reduced pain; but on mood scales, tiredness and confusion increased, and cheerfulness and vigor decreased. Whether anxiety changed is not reported in the abstract. According to the authors, caring for stressed patients became easier. The abstract has no numerical results or statistics; a very small study from 1974 Side effects: Drowsiness, tiredness, confusion; mild increase in heart rate and drop in blood pressure (also seen during the placebo period)	PMID 22454883, doi:10.1111/j.1365-2125.1974.tb00257.x
Strasser 2006 (Cannabis-In-Cachexia-Study-Group) Multicenter, double-blind, placebo-controlled phase 3 RCT (3 groups; stopped early on the recommendation of an independent committee)	243 randomized (95 cannabis extract, 100 THC, 48 placebo); 164 completed Patients with advanced cancer and cancer-related anorexia-cachexia syndrome (at least 5% weight loss in the past 6 months)	Oral cannabis extract (2.5 mg THC + 1 mg CBD) or THC (2.5 mg) and placebo	2 times a day, 6 weeks None (fixed dose)	6 weeks	No difference among the three groups in the co-primary outcomes of appetite and quality of life (EORTC QLQ-C30 global health and quality-of-life score). In mood measured daily, improvement was reported by 60% with cannabis extract, 46% with THC and 64% with placebo (p = 0.46; as reported by Mücke 2018). Because it became clear that no meaningful difference between groups would emerge, the independent committee recommended stopping enrollment Side effects: No difference between groups in cannabinoid-related side effects	PMID 16849753, doi:10.1200/JCO.2005.05.1847
Brisbois 2011 Double-blind, placebo-controlled, randomized pilot study (phase 2, two centers)	46 randomized (24 dronabinol, 22 placebo); 21 completed Patients with advanced cancer with reduced appetite and taste and smell disturbances (Canada)	Dronabinol (synthetic THC, oral capsule) and placebo	2.5 mg; once a day for the first 3 days, then 2 times a day; up to 20 mg a day if needed 2 times a day from day 4; increase up to 20 mg/day at the patient's request	18 days	Quality of life, one of the study's aims, improved in both groups; no difference between groups. In the dronabinol group, sleep quality (p = 0.025) and relaxation (p = 0.045) were better than with placebo; taste and smell perception and the enjoyment of food were also rated better. Very small and short; 21 of 46 people completed Side effects: No details in the abstract	PMID 21343383, doi:10.1093/annonc/mdq727, NCT00316563, no results in the registry)
Côté 2016 Double-blind, placebo-controlled RCT (single center)	56 randomized (28 nabilone, 28 placebo); 24 left before the end of the study Patients with squamous cell cancer of the head and neck (mouth, pharynx, larynx) receiving radiation therapy (Canada)	Nabilone (THC-like synthetic cannabinoid, by mouth) and placebo	Week 1: 0.5 mg at bedtime; week 2: 0.5 mg 2 times a day; from week 3: up to 2 mg/day Weekly increase; from week 3, dose adjusted by the radiation oncologist	Throughout radiation therapy (about 7 weeks) + 4 weeks of follow-up	Primary outcome: the time until a 15% worsening of quality of life was not extended by nabilone (p = 0.43). No difference from placebo in mood (p = 0.32), sleep (p = 0.44), pain (p = 0.60), nausea (p = 0.71), appetite (p = 0.33) or weight (p = 0.15). Anxiety was not measured as an efficacy outcome Side effects: No difference from placebo in the frequency of drowsiness (p = 0.32) and anxiety (p = 0.91) as side effects (as reported by Cochrane 2023; the full text could not be read)	PMID 26503964, doi:10.1177/0003489415612801, Funding: Canadian public research funds, the manufacturer supplied the nabilone and placebo (as reported by Cochrane 2023)
Turcott 2018 Double-blind, placebo-controlled RCT (single center)	47 randomized; 33 analyzed (14 nabilone, 19 placebo) Patients with advanced non-small cell lung cancer and loss of appetite (Mexico)	Nabilone (THC-like synthetic cannabinoid, oral capsule) and placebo	0.5 mg for 2 weeks, then 1 mg for 6 weeks Dose increase after week 2	8 weeks	Quality of life (one of the study's three aims) was compared only within each group: in the nabilone group, role (p = 0.030), emotional (p = 0.018) and social functioning (p = 0.036) and insomnia (p = 0.020) improved, while in the placebo group they did not change; a direct comparison of the two groups is not in the abstract. The insomnia score (0–100) changed by −40.7 with nabilone and −9.9 with placebo (as reported by Cochrane 2023). No difference from placebo in appetite or weight; the full text could not be read Side effects: The abstract reports that treatment was safe; no details on side effects in the sources read (the full text could not be read)	PMID 29550881, doi:10.1007/s00520-018-4154-9, NCT02802540, quality-of-life domains were secondary outcomes, no results in the registry). The manufacturer donated the nabilone and placebo and had no role in the study (as reported by Cochrane 2023)
Schloss 2021 Double-blind, randomized phase 2 study comparing two THC:CBD ratios (no placebo group; control group from past patient records for secondary outcomes)	88 enrolled; 83 completed at least 4 weeks (41 on 1:1, 42 on 4:1); 61 at week 12 Adults with high-grade glioma (glioblastoma or grade 3 astrocytoma) taking cannabis in addition to standard treatment (Australia)	Oral whole-plant cannabis oil: 1:1 (4.6 mg THC + 4.8 mg CBD per mL) or 4:1 (15 mg THC + 3.8 mg CBD per mL); single dose at night	Mean tolerated dose 2.25 mL/day of the 1:1 oil (10.4 mg THC + 10.8 mg CBD), 1.8 mL/day of the 4:1 oil (27 mg THC + 6.8 mg CBD) Night-time dose adjusted according to side effects	12 weeks	The primary outcomes were tolerability and quality of life (FACT-Br). Both ratios were well tolerated; the total quality-of-life score did not change clearly with either ratio (1:1: 92.8 → 91.2; 4:1: 93.6 → 87.9; between groups p = 0.90); the sleep item improved in both groups (p = 0.009), and the physical and functional domains were better with the 1:1 ratio. Because there was no placebo group, it cannot be told whether the improvement came from cannabis or from other factors; the abstract's statement that it "significantly improved quality of life" is not supported by the total score Side effects: No serious side effects; dry mouth 13%, night-time tiredness 11%, dizziness 10%, drowsiness 7%; at week 8, mild night-time hallucinations, suspiciousness or excessive cheerfulness in 4 people (6%); dose reduced in 3 people	PMID 34094937, doi:10.3389/fonc.2021.649555, ANZCTR ACTRN12617001287325. Funding: the company that supplied the product (Bioceuticals), the authors reported no conflicts of interest
Zylla 2026 (CanPan) Waitlist-controlled, open-label pilot RCT	34 enrolled; 32 started (16 early start, 16 delayed start) Newly diagnosed locally advanced or metastatic pancreatic cancer, at least one symptom; moderate-to-severe insomnia in 85% (US, Minnesota)	Products chosen for each patient from the state medical cannabis program (THC-dominant, balanced THC:CBD or CBD-dominant; by mouth, spray or inhaled)	Individual; median THC at week 8: 7.3 mg/day Start at a low dose and increase slowly on the advice of the pharmacist and oncology team	8-week comparison period (16 weeks in total)	Feasibility goals, the primary outcome, were met. At week 8, improvement in insomnia was 67% in the early-start group and 30% in the waiting group (p = 0.18; not significant); numerical but non-significant differences also in pain and appetite. No placebo; patients knew what they were taking. In a second, 16-week analysis (16 patients with complete data), those with higher daily THC doses had greater improvement in anxiety (r = 0.52; p = 0.038) and insomnia (r = 0.53; p = 0.033); this is an association without a comparison Side effects: Worsening of cannabis-related complaints such as dry mouth was similar in the two groups (11% vs 20%; p = 0.99)	PMID 41510814, doi:10.1200/OP-25-01165, NCT06605430, 16-week dose–response analysis: Gupta 2026, PMID 42767994, funding was a donation, no conflicts of interest reported
Weiss 2025 Double-blind, placebo-controlled RCT (the target of 100 people was not reached)	46 randomized; 43 completed (23 active, 20 placebo) Patients with non-metastatic breast, colorectal, endometrial or ovarian cancer who had finished chemotherapy and had grade 2–3 chemotherapy-related nerve damage (neuropathy) (88% women; US)	Oral hemp extract gel capsule, CBD-dominant with a little THC; placebo	After week 1, 125–136 mg CBD + 6–11 mg THC a day Days 1–3: 1 capsule 3 times a day; days 4–6: 2 capsules 3 times a day; from day 7: 3 capsules 3 times a day	12 weeks	The primary outcome was sensory symptoms of neuropathy (no difference in the primary analysis). Of this topic's secondary measures, there was no significant difference between groups in sleep disturbance (PROMIS): at 12 weeks, the placebo group improved slightly more (p = 0.13); in a mixed model adjusted for the baseline difference, the difference favored the active group but was again not significant (−2.6; 95% CI −6.0 to 0.9; p = 0.14). Quality of life (QLQ-C30 summary score p = 0.46; global health p = 0.33) did not differ. Half the targeted number of people was reached Side effects: No difference in side effects between groups; no serious side effects	PMID 41211445, doi:10.3389/fonc.2025.1590168, NCT04398446, the authors reported no conflicts of interest
Haney 2025 Double-blind, placebo-controlled pilot RCT (feasibility)	12 (6 active, 6 placebo); all completed Women with breast cancer and painful taxane-related nerve damage (neuropathy) (US)	Oral cannabis capsule (100 mg CBD + 5 mg THC per capsule) and placebo	1 capsule 3 times a day (300 mg CBD + 15 mg THC/day); patients in both groups asked to reduce the dose Dose reduction at the patient's request	8 weeks	Sleep, pain and functional well-being improved over time in both groups; but sleep scores went worse in the cannabis group than with placebo (p < 0.05), and functional well-being was lower in the last 3 weeks (p < 0.02); neuropathy scores were worse than placebo every week. No sign of benefit was seen; a very small feasibility study Side effects: Well tolerated according to the authors; requests for dose reduction in both groups	PMID 40611810, doi:10.1089/can.2025.0028, NCT03782402, results in the registry)
Fleege 2025 Open-label, single-arm phase 2 pilot study (no comparison group)	39 started; 28 completed (5 stopped because of side effects, 6 because they saw no benefit) Women with stage 0–3 hormone receptor-positive breast cancer and aromatase inhibitor-related joint and muscle pain (US)	FDA-approved pure CBD medicine (oral solution)	Started at 25 mg 2 times a day and increased over 4 weeks to 100 mg 2 times a day (200 mg/day) If tolerated, increased by 25 mg 2 times a day once a week	15 weeks	The primary outcome was pain (worst pain fell by at least 2 points in 17 of 39 people). In this topic's secondary measures, anxiety and sleep disturbance did not change in the analysis of all participants (38 people); improvement in anxiety, sleep disturbance and fatigue was seen only in the 28 people who completed 15 weeks (completers are a selected group). Participants' average baseline anxiety score was close to that of the general population (PROMIS T score 49.5). No comparison group Side effects: 40 side effects in 18 patients (46%), all grade 1–2; 5 people stopped because of side effects	PMID 40751295, doi:10.1002/cam4.71117, NCT04754399, PROMIS anxiety T score −3.4, depression −1.7). Funding: public and foundation funds
De Feo 2023 (MASCC guideline) Systematic review and expert guideline (MASCC; RCTs and systematic reviews; search up to November 12, 2021)	15 RCTs and 2 systematic reviews (of the RCTs, 4 on sleep, 5 on mood, 6 on both) Anxiety, depression and insomnia in cancer patients	Cannabis and cannabinoid products (product breakdown not in the abstract)			No study in cancer patients used psychological symptoms as the primary outcome. 6 of the 15 RCTs showed a sign of benefit (5 for sleep, 1 for mood); the studies differed greatly in product, comparison, duration and measurement tools. The guideline's conclusion: there is no high-quality evidence sufficient to recommend cannabis for psychological symptoms in cancer	PMID 36809575, doi:10.1007/s00520-023-07628-3
Crichton 2024 Systematic review and meta-analysis of intervention studies (5 databases; up to May 2023; GRADE)	15 studies (11 randomized, 4 non-randomized), 18 interventions, 1,898 participants Depression, anxiety and stress in cancer patients	THC (70% synthetic), synthetic CBD (with or without THC) and whole-plant extracts			No clinically meaningful effect on the primary outcomes (depression, anxiety, stress); also no effect on emotional functioning, mood changes or quality of life. Anxiety events were more frequent with high-dose synthetic THC than with low doses (OR 2.0; 95% CI 1.4–2.9; very low certainty). The authors concluded that the evidence for efficacy and safety is insufficient Side effects: Side effects, including anxiety, more frequent with high-dose synthetic THC (very low certainty)	PMID 38430618, doi:10.1016/j.maturitas.2024.107941
Creangă-Murariu 2025 Systematic review and meta-analysis (interventional and observational studies; search up to November 4, 2023)	98 publications (all cancer symptoms); 1,039 patients from single-arm studies in the anxiety analysis Symptom management in cancer patients	All types of cannabinoids (THC-dominant, CBD-dominant, balanced)			Anxiety and insomnia were pooled only from single-arm studies, against baseline: anxiety fell (mean change −1.30; 95% CI −2.22 to −0.39; very high inconsistency between studies, I² 100%), while the improvement in insomnia was not significant (−1.08; −2.48 to 0.33); these are not placebo comparisons. In comparisons with placebo, there was no difference in depression (MD 0.60; −0.65 to 1.86) or overall quality of life (MD 0.16; −0.02 to 0.35). Certainty of evidence was low or very low for all outcomes; risk of publication bias was high for anxiety, insomnia and quality of life Side effects: With THC-dominant products, the risk of psychiatric (OR 10.62; 1.35–83.57) and neurological (OR 2.24; 1.15–4.35) side effects increased	PMID 40748522, doi:10.1007/s11912-025-01695-x, PROSPERO CRD42023479375
Häuser 2023 (Cochrane) Cochrane systematic review and meta-analysis (double-blind RCTs; search January 26, 2023)	14 RCTs, 1,823 participants; sleep analysis included 4 RCTs, 1,314 people Adults with cancer pain (sleep problems, depression and anxiety as secondary outcomes)	Nabiximols (THC+CBD mouth spray), THC, nabilone and other synthetic THC analogs, CBD oil			In cancer pain not responding to opioids, THC+CBD spray or THC gave no benefit over placebo for sleep problems (SMD −0.06; 95% CI −0.19 to 0.06; 4 studies, 1,314 people; moderate certainty); a study with a withdrawal design also found no difference. In one study included in the pooled analysis (Portenoy 2012), sleep disruption fell more than with placebo only in the lowest-dose arm (p = 0.003), with no difference at medium and high doses. No difference in depression (single study, 360 people) (low certainty); anxiety was not measured in these spray studies. Of two small studies with nabilone, one found sleep and mood no different from placebo, while in the other the insomnia score improved more with nabilone (−40.7 vs −9.9 on a 0–100 scale; a direct comparison of the two groups was not reported); anxiety was not measured in these two studies (low certainty). In the pure CBD study (MedCan1), changes in anxiety and depression were similar in the two groups; sleep was not measured Side effects: No clinically meaningful difference from placebo in withdrawals due to side effects or in serious side effects (moderate certainty)	PMID 37283486, doi:10.1002/14651858.CD014915.pub2
AminiLari 2022 Systematic review and meta-analysis of RCTs (up to January 2021; GRADE)	39 RCTs, 5,100 patients (33 of the studies in patients with cancer pain or chronic non-cancer pain); 5 RCTs in the cancer pain subgroup (as reported by ASCO 2024) Patients whose sleep was measured; most with chronic pain (including cancer pain)	Oral cannabinoids (38 studies) and inhaled cannabis in 1 study			A very small improvement in sleep disturbance in cancer pain (weighted mean difference −0.19 cm on a 10 cm scale; 95% CI −0.36 to −0.03; moderate-to-high certainty); improvement was larger in chronic non-cancer pain (interaction p = 0.03). According to the authors, any benefit is small Side effects: Risk of dizziness clearly increased (risk difference 29% in studies with at least 3 months of follow-up); small increases in drowsiness, dry mouth, tiredness and nausea	PMID 34546363, doi:10.1093/sleep/zsab234
Belgers 2023 Meta-analysis of RCTs (search August 2–3, 2021)	17 RCTs (4 in cancer, 13 in central nervous system disease); in the cancer subgroup, quality-of-life analysis included 4 studies, 747 patients Adults with cancer or chronic neurological disease	THC (dronabinol, nabilone), THC+CBD (mouth spray, cannabis extract) and CBD			No effect on overall health-related quality of life (g = −0.02; 95% CI −0.11 to 0.06) or mental well-being (g = −0.02; −0.16 to 0.13); no difference between the cancer and neurological disease subgroups. In 65% of the studies, a drug company supplied the drug or sponsored the study	PMID 35861789, doi:10.1089/can.2021.0187, Brain tumor study by the same research group: Belgers 2026 (in this table)
Simon 2022 Systematic review and meta-analysis (RCTs and non-randomized studies; search up to May 2020)	10 studies (4 RCTs, 6 non-randomized); quality-of-life analysis included 4 RCTs, 545 patients Cancer-related cachexia and loss of appetite	Dronabinol, nabilone, THC, cannabis extract and medical cannabis			In the pooled analysis of 4 RCTs, overall quality of life improved slightly less in the cannabinoid groups than in control groups taking placebo or megestrol (an appetite-stimulating drug) (SMD −0.25; 95% CI −0.43 to −0.07; moderate certainty). No benefit for appetite or weight Side effects: No significant difference in the number and severity of side effects in the RCTs	PMID 34881518, doi:10.1002/jcsm.12861
Mücke 2018 Systematic review and meta-analysis (palliative medicine)	9 studies, 1561 participants Palliative care (cancer subgroup)	Cannabinoids			In cancer patients, no difference from placebo in sleep problems (2 studies, 203 patients; SMD −0.09; 95% CI −0.62 to 0.43) or health-related quality of life (2 studies, 420 patients; SMD 0.10; −0.12 to 0.33); the share of patients developing mental health complaints was similar (3.2% vs 4.1%). No difference in calories, appetite, nausea and vomiting, or pain either; quality of evidence low or very low Side effects: No difference in dizziness/mental health in cancer	PMID 29400010, doi:10.1002/jcsm.12273, First German version: Mücke 2016, Schmerz, PMID 26809975, SMD 0.00)
Maida 2008 Prospective observational study: patients taking and not taking nabilone were compared after balancing with propensity scores (not randomized or blinded)	112 (47 nabilone, 65 not taking nabilone) Patients with advanced cancer with pain and other symptoms (Canada)	Nabilone (THC-like synthetic cannabinoid, by mouth), in addition to existing treatment	Not reported (in the abstract)	30 days	The primary outcomes were pain and opioid use (both lower in the nabilone group; p < 0.0001). Of this topic's measures, anxiety (p = 0.028) and total symptom distress (ESAS total; p = 0.021) were better in the nabilone group; the scale has no sleep item. Nabilone was not given at random and the study was not blinded; groups were balanced with propensity scores Side effects: Not reported (in the abstract)	PMID 18402303
Bar-Lev Schleider 2018 Prospective observational study (data routinely collected in a cannabis company's treatment program; no comparison group)	2,970 cancer patients Cancer patients granted a medical cannabis license (Israel); 51% stage 4	The company's cannabis products (product types and THC/CBD ratios not in the abstract)	Not reported (in the abstract)	6 months (2015–2017)	The most common reasons for treatment were sleep problems (78%) and pain (78%). At month 6, 902 patients had died and 682 had stopped treatment; of the 1,211 remaining who responded, 95.9% reported that their condition had improved, 3.7% that it was unchanged and 0.3% that it had worsened. No separate result for sleep and quality of life in the abstract; the full text could not be read. No control group; those who died or stopped were not included in the evaluation. 4 of the 8 authors are from the research department of the company that supplied the product Side effects: No details in the abstract (the authors report it was well tolerated)	PMID 29482741, doi:10.1016/j.ejim.2018.01.023
Aviram 2022 Prospective observational cohort (multicenter; no comparison group)	324 started; 212, 158 and 126 responded at months 1, 3 and 6 Oncology patients receiving a medical cannabis license for the first time for cancer symptoms (Israel)	Licensed medical cannabis: dried flower (smoked or vaporized) or oil extract; at month 6, of those who reported their product, 61% used THC-dominant, 25% mixed and 15% CBD-dominant products (median monthly THC 3 g)	Individual; median 20 g of product a month	6 months	At 6 months, compared with baseline, anxiety (GAD-7: 9 → 6; −2.35; 95% CI −3.31 to −1.40), depression (Beck: 19 → 15; −1.97; −3.26 to −0.68), sleep disturbance (Pittsburgh: 12 → 8; −3.07; −3.95 to −2.18) and quality-of-life score (−0.55; −0.83 to −0.27) improved; total symptom burden fell by a median of 18%. In the same group's first 1-month analysis (108 patients), sleep and distress had improved, while anxiety, depression and quality of life had not changed significantly. No control group; at month 6, only 39% of those who started responded Side effects: Side effects were common (20–30%), but most were not serious and stayed stable during follow-up	PMID 35669038, doi:10.3389/fpain.2022.861037, 1-month study from the same research group: Aviram 2020, PMID 33265945, THC-dominant products were better than CBD-dominant ones only for sleep duration). Funding: a charitable foundation, the authors reported no conflicts of interest
Reddy 2026 (Minnesota cannabis program) Retrospective longitudinal analysis (state medical cannabis program records and sales data; no comparison group)	1,962 cancer patients Cancer patients enrolled in the Minnesota Medical Cannabis Program (US); anxiety and depression were not the reason for enrolling	Products obtained from the program (75% taken by mouth; median daily THC 17 mg, CBD 3 mg; median THC:CBD ratio 7.2)	Individual; mean daily THC 32 mg, CBD 14 mg	2015–2023; measurements after the first 30 days in the program	The sleep disturbance score (0–10) was 6.72 at baseline and then fell by a mean of 1.63 points (24%). Compared with the highest CBD band (>14.3 mg/day), the likelihood of meaningful (at least 30%) improvement was 29–35% lower in the second and third bands; there was no consistent link with THC dose or THC:CBD ratio. In the same patients, the anxiety score improved by a mean of 25% and was linked to CBD dose (0.05 points per 5 mg of CBD); improvement in depression was linked only to route of use (Reddy 2025). No control group; dose was estimated from sales records Side effects: Not reported (in these analyses)	PMID 41120744, doi:10.1007/s12094-025-04069-8, Anxiety and depression analysis in the same 1, 962 patients: Reddy 2025, PMID 41163433, analysis of cancer patients' symptoms over the first 4 months in the same program: Anderson 2019, PMID 30860938, the authors reported no conflicts of interest
Kasvis 2022 (Quebec Cannabis Registry) Patient registry analysis (prospective registry; no comparison group)	358 (171 men, 187 women) Patients starting medical cannabis for cancer symptoms (Canada, Quebec)	Medical cannabis: THC-dominant, balanced THC:CBD or CBD-dominant products	Not reported (in the abstract)	3 months	At 3 months, anxiety (ESAS-r, 0–10; men −1.1, women −1.2) and well-being (men −1.2, women −1.4) improved from baseline in both sexes; the abstract reports no significant improvement for depression (the scale has no sleep item). Balanced THC:CBD products improved anxiety (−1.83) and well-being (−2.01) in men; CBD-dominant products did not give significant relief in either sex. No control group; results are self-reported by patients Side effects: Not reported (in the abstract)	PMID 35723729, doi:10.1007/s00520-022-07193-1, Pain analysis of the same 358 patients: Aprikian 2024, PMID 37130724
Bar-Sela 2013 Prospective observational study (interviews on the day of licensing and 6–8 weeks later; no comparison group)	211 started; 131 attended the second interview (25 had stopped in the first week); 106 evaluated Cancer patients receiving supportive or palliative care who were granted a medical cannabis license (Israel, single center)	Licensed cannabis plant (THC 16–27%, CBD 0.5–2%)	Up to 30 g a month; starting with strains under 20% THC was recommended for the first 2 months Start with a small amount and increase slowly according to need and side effects	6–8 weeks	In the 106 patients who continued treatment, all symptoms improved significantly from baseline (p < 0.001): the share of patients without sleep disturbance rose from 20% to 56%, and the share without mood disturbance from 6% to 46%; the distress thermometer score also fell. 50 of those who started (24%) died before the second interview; only those who continued were evaluated (selection bias). No control group; results are patient-reported Side effects: Memory weakening with long-term use was the only significant side effect (p = 0.002)	PMID 23956774, doi:10.1155/2013/510392
Varadpande 2025 (UK Medical Cannabis Registry) Case series (prospective patient registry; no comparison group)	168 Patients prescribed cannabis-based medicines for cancer pain (United Kingdom)	Cannabis-based medicines (product breakdown not in the abstract)	Not reported (in the abstract)	6 months (measurements at months 1, 3 and 6)	In addition to pain measures, anxiety (GAD-7), sleep quality (single-item scale) and quality of life (EQ-5D-5L) improved significantly from baseline (p < 0.05; sizes not in the abstract). No control group; the authors state that cause and effect cannot be established. 8 of the 12 authors are linked to a private clinic that prescribes cannabis medicines Side effects: 29 side effects in 5 patients (3%); most mild or moderate	PMID 39921589, doi:10.1080/15360288.2025.2457101, Preliminary 16-patient palliative care report from the same registry: Nimalan 2022, PMID 34980282
Stienrut 2024 (Thai national cohort) Prospective observational cohort (30 centers; no comparison group)	21,284 patients, 2,873 of them with cancer; by week 12, 32% of participants had dropped out of follow-up Patients prescribed cannabis oil by traditional Thai medicine practitioners for symptoms such as pain, loss of appetite and insomnia (Thailand); patients with cancer were not selected for anxiety or sleep problems (mean anxiety score at baseline 1.96/10)	Cannabis extract in coconut oil made by the Ministry of Public Health (about 0.07 mg THC per drop; almost no CBD)	Most often 3 drops a day at bedtime (about 0.2 mg THC/day)	3 months	In patients with cancer, quality of life (EQ-5D-5L; 1 = best health state) rose from 0.79 by a mean of 0.13 points over 3 months (95% CI 0.12–0.14); the anxiety score (ESAS, 0–10) fell by 1.46 points (1.35–1.58), the not-feeling-well score by 1.97 points and the depression score by 0.59 points. Sleep was not measured separately: the authors used ESAS's drowsiness item in place of insomnia. No control group; patients were not selected for these symptoms and baseline scores were low; the daily amount of THC was about 0.2 mg (the authors write that the placebo effect may have influenced the results), and reasons for dropping out were not recorded Side effects: Side effects at 0.42% of 67,681 visits (most often dry mouth, dizziness, drowsiness)	PMID 39144529, doi:10.1159/000540153, Thai Clinical Trials Registry TCTR20191231001. Funding: the Department of Thai Traditional and Alternative Medicine of the Ministry of Public Health, which also makes the product, 9 of the 13 authors work in this department, the authors reported no conflicts of interest
Scarborough 2026 (MedCan1+2 substudy) Liver enzyme substudy of two RCTs	287 Advanced cancer	CBD alone or THC/CBD	Escalating dose, ≤600 mg CBD/day Escalating	28 days	No significant rise in ALT/AST; no patient had ALT/AST >3× ULN (>5× in liver metastases)	PMID 41429435, doi:10.1136/spcare-2025-005837, Main studies: Hardy 2023 (MedCan1) and Hardy 2025 (MedCan2), in this table
Downs 2026 (MedCan1 secondary analysis) Surrogate markers of drug interactions in RCT data	MedCan1 participants (number in the analysis not in the abstract; the full text could not be read) Advanced cancer	CBD oil (pure)	MedCan1 dose: 50–600 mg/day, chosen by the patient	28 days (MedCan1 duration)	The chosen CBD dose was unrelated to opioid, benzodiazepine or antipsychotic use; the risk of new side effects did not increase with concomitant medicines	PMID 39539239, doi:10.1089/can.2024.0041, Main study: Hardy 2023 (MedCan1), in this table

Reported side effects

Pure CBD: no difference between arms, dyspnea more frequent (Hardy 2023)
THC-containing: confusion, feeling 'high', exaggerated sense of well-being (Hardy 2025)
In the short term (≤600 mg/day, 28 days), no rise in liver enzymes was seen in cancer patients (Scarborough 2026)
THC-dominant products: increased risk of psychiatric side effects (meta-analysis, OR 10.62; wide confidence interval); anxiety events more frequent with high-dose synthetic THC (OR 2.0; very low certainty) (Creangă-Murariu 2025, Crichton 2024)

Monitoring done in studies

Prescription CBD label: liver tests at baseline, at months 1, 3 and 6, then periodically; stop if ALT/AST >3× ULN + bilirubin >2× ULN, or if transaminases stay >5× ULN
INR in people taking warfarin
Drug blood levels in people taking everolimus/tacrolimus/sirolimus

Who should be careful

Liver disease / liver metastases / hepatotoxic treatment: with prescription CBD, ALT >3× ULN in 12–13% (placebo 1%), dose-related; in moderate/severe hepatic impairment, the label specifies a lower starting dose
People taking warfarin or clopidogrel and those at high risk of bleeding: there is an interaction mechanism (CYP2C9, CYP2C19); no clinical data have been reported showing that CBD directly increases bleeding risk
Before surgery/procedures: ASRA 2023 recommends asking about cannabinoid use and including it in the anesthesia plan; the anesthesia team must be told
People receiving immunotherapy (checkpoint inhibitors): there is an observational signal and a guideline warning
People taking everolimus, tacrolimus, sirolimus or oral targeted drugs: should not be added without consulting the oncologist/pharmacist
Older adults, people taking many medicines, people on sedatives/opioids: risk of somnolence, dizziness and falls; with THC-containing products, confusion and psychotomimetic effects
Use to 'treat' tumors: uncontrolled case series are not evidence; according to ASCO, stopping standard treatment is strongly advised against
Pregnancy/breastfeeding, history of psychiatric illness (for THC-containing products): cancer studies excluded these groups

Drug interactions

Everolimus: exposure increased ~2.5-fold; label: start everolimus at a lower dose, monitor levels (high; human PK data)
Tacrolimus, sirolimus, digoxin (P-gp substrates): the label states that their levels may rise (moderate)
Oral chemotherapy/targeted drugs that are P-gp substrates (e.g., many tyrosine kinase inhibitors): theoretical; no human data
CYP2C19 substrates and prodrugs (clopidogrel): substrate levels rise, and the active metabolite of prodrugs may fall; no human data for cancer drugs
Warfarin: case report of INR elevation; INR should be monitored
CYP3A4 substrates (taxanes, irinotecan, docetaxel, many TKIs): on the label, midazolam was unchanged (licensed CBD 750 mg twice daily); a single dose of a CBD-dominant extract (640 mg CBD + 20 mg THC) increased midazolam AUC by 56% (Bansal 2023, PMID 37313955); herbal cannabis tea did not affect irinotecan and docetaxel PK (n=24); nabiximols did not affect temozolomide PK; these results cannot be generalized
Immunotherapy: in people using cannabis (mostly THC-containing), the nivolumab response rate was 15.9% versus 37.5% (Taha 2019, PMID 30670598, n=140); in a prospective observational study, time to progression and survival were shorter (Bar-Sela 2020, PMID 32872248, n=102); causation not proven, no data specific to pure CBD
Opioids: in MedCan1, CBD dose was unrelated to opioid use; no opioid-sparing effect was shown with THC-containing products; somnolence increases with CNS depressants
Citalopram/escitalopram: in 6 patients who added CBD (200–800 mg/day), citalopram levels rose significantly (Anderson 2021, PMID 34121064)
Valproate and liver-toxic drugs (some chemotherapies, methotrexate): no data on use together with hepatotoxic chemotherapies
Tamoxifen: in one case, endoxifen was found to be low with 40 mg/day CBD (Parihar 2020, PMID 33191836)
No clinical studies have been reported with direct oral anticoagulants (apixaban, rivaroxaban)

Abbreviations: RCT, randomized controlled trial (a study in which participants are assigned to groups by chance); placebo, an inactive substance; SMD, OR and RR, effect measures used in analyses that pool studies; CYP, liver enzymes that break down medicines.

If you are considering CBD

Tell your care team about your sleep problems and anxiety; you can ask about options that suit you, including emotional support.
If you take an antidepressant, an anti-anxiety medicine or a sleep medicine, talk with your doctor and pharmacist about how it may interact with CBD.
Go over your cancer medicines with your oncologist, especially immunotherapy and targeted drugs taken by mouth.
If a product is being considered, look at its label and its independent certificate of analysis (COA); find out whether it contains THC.

Choose your age group, medicines and situation to prepare a short summary you can take to your doctor. Nothing you enter is sent anywhere.

Open the personal warnings tool Convert drops ↔ mg

Bottom line

CBD oil has not been shown to offer a clear benefit over placebo for sleep, anxiety or quality of life in cancer, and as of October 2026 there is no CBD study that targets sleep as its main goal. These symptoms are real and important; talking about them with your care team is the first step toward finding support that will help you.

Frequently asked questions

Does CBD oil work for sleep and anxiety in cancer?

In a well-designed study in advanced cancer, pure CBD oil was not clearly different from a dummy treatment (placebo) for anxiety, quality of life or overall symptom burden. Two small studies, one on anxiety before imaging in breast cancer and one on anxiety in brain tumors, also found no difference from placebo on the main measure. There is no pure CBD study that targets sleep in cancer as its main goal. For these reasons, the evidence is weak.

Does CBD help people with cancer sleep?

We don't know; there is no pure CBD study that targets sleep in cancer as its main goal. In a small study in which CBD was given for joint pain, sleep did not change when all participants were analyzed together. In cancer pain studies of THC-containing products, sleep either did not change at all or improved very slightly; in two small studies with CBD-dominant capsules containing a little THC, sleep was also no better than with a dummy treatment. Patient registries report that sleep improved, but these studies had no comparison group.

Can CBD be used together with my antidepressant or anti-anxiety medicines?

CBD can slow the breakdown of some medicines in the body; for example, blood levels of citalopram rose in patients who added CBD. Drowsiness may increase when it is taken with medicines that cause sleepiness. So talk to your doctor and pharmacist before using them together.

Can CBD harm my liver?

In patients with advanced cancer, no significant rise in liver enzymes was seen over 28 days of use. However, it is known that with CBD used as a medicine, liver enzymes can rise depending on the dose. If you have liver disease or cancer that has spread to the liver, it is important to talk with your doctor.

Can CBD oil be obtained with a prescription in Turkey for sleep and anxiety in cancer?

In Turkey, products derived from hemp are regulated by the Kenevirden Elde Edilen Ürünlere Dair Yönetmelik (Regulation on Products Derived from Hemp) dated January 31, 2026. These products, including CBD, may be sold only in pharmacies. Those classed as medicines and health products are dispensed with a prescription; for supplement products and cosmetics with a THC content of at most 0.3%, the regulation does not state a prescription requirement. As of October 2026, there is no approved (licensed) CBD medicine in Turkey. If you are thinking of using it, talk to your doctor first.

References

Show 32 references

Hardy 2023 (MedCan1-CBD). Phase IIb, double-blind, placebo-controlled, dose-escalation RCT. PMID 36409969, doi:10.1200/JCO.22.01632, ANZCTR ACTRN12618001220257, protocol: Good 2019, PMID 31810437, PMID 38307703, disease progression and survival)
Hardy 2025 (MedCan2). Double-blind, placebo-controlled RCT. PMID 40705150, doi:10.1007/s00520-025-09763-5, ANZCTR ACTRN12619000037101, funding from the Australian government (MRFF), the manufacturer supplied the drug and placebo and had no role in the study
Nayak 2024. Phase 2, double-blind, placebo-controlled RCT. PMID 39680411, doi:10.1001/jamanetworkopen.2024.50391, NCT04482244
Belgers 2026. Double-blind, placebo-controlled crossover RCT; stopped early because of low enrollment (research letter). PMID 42750778, doi:10.1093/nop/npag025, Netherlands Trial Register NL9623 (GRASS study). Funding: Anita Veldman Foundation, the authors reported no conflicts of interest
Davies 1974. Placebo-controlled crossover pilot study (randomization and blinding not stated in the abstract). PMID 22454883, doi:10.1111/j.1365-2125.1974.tb00257.x
Strasser 2006 (Cannabis-In-Cachexia-Study-Group). Multicenter, double-blind, placebo-controlled phase 3 RCT (3 groups; stopped early on the recommendation of an independent committee). PMID 16849753, doi:10.1200/JCO.2005.05.1847
Brisbois 2011. Double-blind, placebo-controlled, randomized pilot study (phase 2, two centers). PMID 21343383, doi:10.1093/annonc/mdq727, NCT00316563, no results in the registry)
Côté 2016. Double-blind, placebo-controlled RCT (single center). PMID 26503964, doi:10.1177/0003489415612801, Funding: Canadian public research funds, the manufacturer supplied the nabilone and placebo (as reported by Cochrane 2023)
Turcott 2018. Double-blind, placebo-controlled RCT (single center). PMID 29550881, doi:10.1007/s00520-018-4154-9, NCT02802540, quality-of-life domains were secondary outcomes, no results in the registry). The manufacturer donated the nabilone and placebo and had no role in the study (as reported by Cochrane 2023)
Schloss 2021. Double-blind, randomized phase 2 study comparing two THC:CBD ratios (no placebo group; control group from past patient records for secondary outcomes). PMID 34094937, doi:10.3389/fonc.2021.649555, ANZCTR ACTRN12617001287325. Funding: the company that supplied the product (Bioceuticals), the authors reported no conflicts of interest
Zylla 2026 (CanPan). Waitlist-controlled, open-label pilot RCT. PMID 41510814, doi:10.1200/OP-25-01165, NCT06605430, 16-week dose–response analysis: Gupta 2026, PMID 42767994, funding was a donation, no conflicts of interest reported
Weiss 2025. Double-blind, placebo-controlled RCT (the target of 100 people was not reached). PMID 41211445, doi:10.3389/fonc.2025.1590168, NCT04398446, the authors reported no conflicts of interest
Haney 2025. Double-blind, placebo-controlled pilot RCT (feasibility). PMID 40611810, doi:10.1089/can.2025.0028, NCT03782402, results in the registry)
Fleege 2025. Open-label, single-arm phase 2 pilot study (no comparison group). PMID 40751295, doi:10.1002/cam4.71117, NCT04754399, PROMIS anxiety T score −3.4, depression −1.7). Funding: public and foundation funds
De Feo 2023 (MASCC guideline). Systematic review and expert guideline (MASCC; RCTs and systematic reviews; search up to November 12, 2021). PMID 36809575, doi:10.1007/s00520-023-07628-3
Crichton 2024. Systematic review and meta-analysis of intervention studies (5 databases; up to May 2023; GRADE). PMID 38430618, doi:10.1016/j.maturitas.2024.107941
Creangă-Murariu 2025. Systematic review and meta-analysis (interventional and observational studies; search up to November 4, 2023). PMID 40748522, doi:10.1007/s11912-025-01695-x, PROSPERO CRD42023479375
Häuser 2023 (Cochrane). Cochrane systematic review and meta-analysis (double-blind RCTs; search January 26, 2023). PMID 37283486, doi:10.1002/14651858.CD014915.pub2
AminiLari 2022. Systematic review and meta-analysis of RCTs (up to January 2021; GRADE). PMID 34546363, doi:10.1093/sleep/zsab234
Belgers 2023. Meta-analysis of RCTs (search August 2–3, 2021). PMID 35861789, doi:10.1089/can.2021.0187, Brain tumor study by the same research group: Belgers 2026 (in this table)
Simon 2022. Systematic review and meta-analysis (RCTs and non-randomized studies; search up to May 2020). PMID 34881518, doi:10.1002/jcsm.12861
Mücke 2018. Systematic review and meta-analysis (palliative medicine). PMID 29400010, doi:10.1002/jcsm.12273, First German version: Mücke 2016, Schmerz, PMID 26809975, SMD 0.00)
Maida 2008. Prospective observational study: patients taking and not taking nabilone were compared after balancing with propensity scores (not randomized or blinded). PMID 18402303
Bar-Lev Schleider 2018. Prospective observational study (data routinely collected in a cannabis company's treatment program; no comparison group). PMID 29482741, doi:10.1016/j.ejim.2018.01.023
Aviram 2022. Prospective observational cohort (multicenter; no comparison group). PMID 35669038, doi:10.3389/fpain.2022.861037, 1-month study from the same research group: Aviram 2020, PMID 33265945, THC-dominant products were better than CBD-dominant ones only for sleep duration). Funding: a charitable foundation, the authors reported no conflicts of interest
Reddy 2026 (Minnesota cannabis program). Retrospective longitudinal analysis (state medical cannabis program records and sales data; no comparison group). PMID 41120744, doi:10.1007/s12094-025-04069-8, Anxiety and depression analysis in the same 1, 962 patients: Reddy 2025, PMID 41163433, analysis of cancer patients' symptoms over the first 4 months in the same program: Anderson 2019, PMID 30860938, the authors reported no conflicts of interest
Kasvis 2022 (Quebec Cannabis Registry). Patient registry analysis (prospective registry; no comparison group). PMID 35723729, doi:10.1007/s00520-022-07193-1, Pain analysis of the same 358 patients: Aprikian 2024, PMID 37130724
Bar-Sela 2013. Prospective observational study (interviews on the day of licensing and 6–8 weeks later; no comparison group). PMID 23956774, doi:10.1155/2013/510392
Varadpande 2025 (UK Medical Cannabis Registry). Case series (prospective patient registry; no comparison group). PMID 39921589, doi:10.1080/15360288.2025.2457101, Preliminary 16-patient palliative care report from the same registry: Nimalan 2022, PMID 34980282
Stienrut 2024 (Thai national cohort). Prospective observational cohort (30 centers; no comparison group). PMID 39144529, doi:10.1159/000540153, Thai Clinical Trials Registry TCTR20191231001. Funding: the Department of Thai Traditional and Alternative Medicine of the Ministry of Public Health, which also makes the product, 9 of the 13 authors work in this department, the authors reported no conflicts of interest
Scarborough 2026 (MedCan1+2 substudy). Liver enzyme substudy of two RCTs. PMID 41429435, doi:10.1136/spcare-2025-005837, Main studies: Hardy 2023 (MedCan1) and Hardy 2025 (MedCan2), in this table
Downs 2026 (MedCan1 secondary analysis). Surrogate markers of drug interactions in RCT data. PMID 39539239, doi:10.1089/can.2024.0041, Main study: Hardy 2023 (MedCan1), in this table

Points that could not be verified from the full text

MedCan3 (THC/CBD 1:20) is ongoing: 128 of the target 150 patients have been enrolled (ANZCTR, September 23, 2026), and the last patient is expected to be enrolled in June 2027; no results (protocol PMID 38693590)
Related studies on ClinicalTrials.gov that were completed or ended early but whose results had not been published as of October 2026: a 91-person N-of-1 study of cannabis oil for cancer symptoms (NCT03948074; a symptom scale that also includes sleep and anxiety; finished in July 2025), a phase 3 dronabinol–placebo study in pancreatic cancer (NCT03984214; 109 patients; main measure the symptom score of a quality-of-life scale; finished in September 2024), a quality-of-life study with cannabinoids in Germany (NCT06097533; 79 patients; ended early in February 2026) and a CBD study for anxiety in glioblastoma (NCT05753007; ended with 2 participants). Their results could not be included in the table.
The full text of some studies is not open access: for MedCan1 (Hardy 2023), Strasser 2006, Brisbois 2011, Côté 2016, Turcott 2018, Zylla 2026, Haney 2025, Maida 2008, Kasvis 2022, Bar-Lev Schleider 2018, Varadpande 2025, the MASCC 2023 guideline, Crichton 2024 and AminiLari 2022, only the abstract, the registry entry or reports in reviews were read. For Davies 1974, only the abstract could be read (the full-text file could not be opened because it required human verification). A 1976 publication on the psychological effects of THC in patients with advanced cancer (Noyes 1976) has no abstract on PubMed; because it could not be read, it was not included in the table.

Research history

Here we record what changed on this page, when and why, with dates. The newest change is at the top.

October 31, 2026
The evidence level text now uses the same level name as the top of the page; the level and the result did not change.
October 30, 2026
The last sentence of the page's short summary was rewritten in a calmer, more guiding tone: instead of "Talk to your doctor before you start.", it now says whom you can talk with, and about what, if you are considering CBD. The conclusion and the strength of evidence did not change.
October 28, 2026
October 2026 supplementary review: sources found for this page during the searches for other topics were read and worked into the page. A cohort without a comparison group that followed 2,873 cancer patients in Thailand for 3 months (Stienrut 2024; about 0.2 mg of THC a day) was added: quality of life and anxiety scores improved, but an improvement seen at such a low dose and without controls cannot be attributed to the product; sleep was not measured separately in this study. The side-effect information for the single-dose pre-imaging CBD study in breast cancer (Nayak 2024) was completed from the full text (mild or moderate nausea in 2 people in the CBD group, pain in one of them; no side effects with placebo). The conclusion and the strength of evidence did not change.
Related studies: Stienrut 2024 (Thai national cohort), published July 17, 2024 (PMID 39144529, doi:10.1159/000540153, Thai Clinical Trials Registry TCTR20191231001. Funding: the Department of Thai Traditional and Alternative Medicine of the Ministry of Public Health, which also makes the product, 9 of the 13 authors work in this department, the authors reported no conflicts of interest)

Older changes (5)

October 28, 2026
In the October 2026 search (PubMed, Cochrane, ClinicalTrials.gov), 24 studies were added to this topic: 9 studies with a comparison group (7 placebo-controlled: Davies 1974, Strasser 2006, Brisbois 2011, Côté 2016, Turcott 2018, Weiss 2025, Haney 2025; waitlist-controlled Zylla 2026; Schloss 2021, comparing two THC:CBD ratios), 1 open-label study (Fleege 2025), 7 systematic reviews including the MASCC 2023 guideline and Cochrane 2023, and 7 observational studies. The THC result for sleep was updated with the Cochrane analysis of cancer pain studies (no difference or a very small improvement); the negative sleep result of two small RCTs with CBD-dominant products was added. The statement 'there is no THC RCT for anxiety' was corrected: there are RCTs that measured anxiety as a secondary measure, and they found no difference; a small 1974 THC pilot in anxious patients was also added. Quality of life and real-world data were described together with their limits; ASCO 2024's weak recommendation against high-dose CBD for symptom burden was added. Anxiety, depression and quality-of-life results, registration numbers and funding were added to the MedCan rows; an internal file note visible to readers in two places was removed. The conclusion and the strength of evidence did not change.
Related studies: Davies 1974, published August 1974 (PMID 22454883, doi:10.1111/j.1365-2125.1974.tb00257.x); Strasser 2006 (Cannabis-In-Cachexia-Study-Group), published July 20, 2006 (PMID 16849753, doi:10.1200/JCO.2005.05.1847); Brisbois 2011, published September 2011 (PMID 21343383, doi:10.1093/annonc/mdq727, NCT00316563, no results in the registry)); Côté 2016, published April 2016 (PMID 26503964, doi:10.1177/0003489415612801, Funding: Canadian public research funds, the manufacturer supplied the nabilone and placebo (as reported by Cochrane 2023)); Turcott 2018, published September 2018 (PMID 29550881, doi:10.1007/s00520-018-4154-9, NCT02802540, quality-of-life domains were secondary outcomes, no results in the registry). The manufacturer donated the nabilone and placebo and had no role in the study (as reported by Cochrane 2023)); Schloss 2021, published May 2021 (PMID 34094937, doi:10.3389/fonc.2021.649555, ANZCTR ACTRN12617001287325. Funding: the company that supplied the product (Bioceuticals), the authors reported no conflicts of interest); Zylla 2026 (CanPan), published January 9, 2026 (PMID 41510814, doi:10.1200/OP-25-01165, NCT06605430, 16-week dose–response analysis: Gupta 2026, PMID 42767994, funding was a donation, no conflicts of interest reported); Weiss 2025, published October 2025 (PMID 41211445, doi:10.3389/fonc.2025.1590168, NCT04398446, the authors reported no conflicts of interest); Haney 2025, published October 2025 (PMID 40611810, doi:10.1089/can.2025.0028, NCT03782402, results in the registry)); Fleege 2025, published August 2025 (PMID 40751295, doi:10.1002/cam4.71117, NCT04754399, PROMIS anxiety T score −3.4, depression −1.7). Funding: public and foundation funds); De Feo 2023 (MASCC guideline), published February 21, 2023 (PMID 36809575, doi:10.1007/s00520-023-07628-3); Crichton 2024, published June 2024 (PMID 38430618, doi:10.1016/j.maturitas.2024.107941); Creangă-Murariu 2025, published September 2025 (PMID 40748522, doi:10.1007/s11912-025-01695-x, PROSPERO CRD42023479375); Häuser 2023 (Cochrane), published June 5, 2023 (PMID 37283486, doi:10.1002/14651858.CD014915.pub2); AminiLari 2022, published February 14, 2022 (PMID 34546363, doi:10.1093/sleep/zsab234); Belgers 2023, published February 2023 (PMID 35861789, doi:10.1089/can.2021.0187, Brain tumor study by the same research group: Belgers 2026 (in this table)); Simon 2022, published February 2022 (PMID 34881518, doi:10.1002/jcsm.12861); Maida 2008, published March 2008 (PMID 18402303); Bar-Lev Schleider 2018, published March 2018 (PMID 29482741, doi:10.1016/j.ejim.2018.01.023); Aviram 2022, published May 2022 (PMID 35669038, doi:10.3389/fpain.2022.861037, 1-month study from the same research group: Aviram 2020, PMID 33265945, THC-dominant products were better than CBD-dominant ones only for sleep duration). Funding: a charitable foundation, the authors reported no conflicts of interest); Reddy 2026 (Minnesota cannabis program), published April 2026 (PMID 41120744, doi:10.1007/s12094-025-04069-8, Anxiety and depression analysis in the same 1, 962 patients: Reddy 2025, PMID 41163433, analysis of cancer patients' symptoms over the first 4 months in the same program: Anderson 2019, PMID 30860938, the authors reported no conflicts of interest); Kasvis 2022 (Quebec Cannabis Registry), published October 2022 (PMID 35723729, doi:10.1007/s00520-022-07193-1, Pain analysis of the same 358 patients: Aprikian 2024, PMID 37130724); Bar-Sela 2013, published July 2013 (PMID 23956774, doi:10.1155/2013/510392); Varadpande 2025 (UK Medical Cannabis Registry), published June 2025 (PMID 39921589, doi:10.1080/15360288.2025.2457101, Preliminary 16-patient palliative care report from the same registry: Nimalan 2022, PMID 34980282)
October 22, 2026
The names of the finding labels were simplified: this page's "No benefit seen" label became "No effect found". The assessment did not change.
October 21, 2026
PubMed and ClinicalTrials.gov were searched for newly published clinical studies and ones missed earlier. Study added: Belgers 2026. The conclusion and the strength of evidence did not change.
Related studies: Belgers 2026, published October 2026 (PMID 42750778, doi:10.1093/nop/npag025, Netherlands Trial Register NL9623 (GRASS study). Funding: Anita Veldman Foundation, the authors reported no conflicts of interest)
October 21, 2026
The full texts of the studies were reread, and the information on the page was checked against its sources. Corrected: drug interaction notes. Points that could not be verified from the full text went down from 4 items to 1. The conclusion and the strength of evidence did not change.
October 19, 2026
Page published. This first review examined 6 studies.

Evidence changes across all topics