Cannabis use among cancer patients and survivors in the United States: a systematic review

Abstract

Background

How cannabis products are being used by cancer patients and survivors in the United States is poorly understood. This study reviewed observational data to understand the modes, patterns, reasons, discontinuation, and adverse experiences of cannabis use.

Methods

PubMed and PsycINFO database searches were conducted between May 2022 and November 2022. Of the 1162 studies identified, 27 studies met the inclusion criteria. The intercoder agreement was strong (0.81).

Results

The majority (74%) of the studies were cross-sectional in design. Study samples were approximately equal proportions of men and women and majority White participants. The prevalence of cannabis use based on national samples ranged between 4.8% and 22%. The most common modes of cannabis intake were topical application (80%), smoking (73%), vaping (12%), and ingestion of edible products (10%). Younger age, male gender, being a current or former smoker, and higher socioeconomic status were associated with greater likelihood of cannabis use. The main motive for cannabis use was management of symptoms due to cancer or cancer treatment such as pain, nausea, lack of sleep, and anxiety. A majority of the participants across studies reported that cannabis helped reduce these symptoms. Lack of symptom improvement, side effects such as fatigue and paranoia, cost, and social stigma were identified as some of the reasons for discontinuing cannabis use.

Conclusions

It appears that cannabis may help cancer patients and survivors manage symptoms. However, more longitudinal studies are needed to determine whether positive experiences of cannabis use outweigh adverse experiences over time in this vulnerable population.

Cannabis use among cancer patients and survivors in the United States has gained increased attention in recent years as medicinal and recreational use of cannabis become more normalized (1). Following 2023 ballot approvals or legislation, 22 states and the District of Columbia allow medical and nonmedical (recreational) cannabis use, an additional 18 states allow medical only, and federal policy changes are in process (2). Cannabis use may be particularly attractive to cancer patients and survivors for its potential in managing symptoms due to cancer and cancer treatment (3). Currently there is a lack of understanding of how cannabis is being used by cancer patients and survivors in the United States.

A recent study found that cancer patients who used cannabis reported statistically lower levels of symptom severity compared with nonusers (4). Other studies have also found cannabis use to be associated with lower levels of opioid use and increased adherence to chemotherapy (5,6). However, the use of cannabis among cancer patients remains controversial, as it also carries risks such as cognitive impairment and respiratory problems (7,8). Thus, understanding the potential benefits and risks of cannabis use is crucial for developing evidence-based recommendations for symptom management and improving the quality of life (QoL) of cancer patients and survivors.

There is also a need to understand sociodemographic and other correlates of cannabis use among cancer patients and survivors. Such knowledge may help health professionals tailor messages regarding cannabis, based on current science, to specific groups of patients. Studies show that there may be sociodemographic differences in the prevalence of cannabis use among cancer patients in the United States (9-11). Male cancer patients appear more likely to report cannabis use compared with female cancer patients (9). Younger patients, those aged younger than 50 years, may be more likely to use cannabis than older patients (10). Men and younger patients are more likely to experience higher cannabis costs when consuming cannabis during cancer treatment, which may lead to substantial out-of-pocket expenses (11). Additionally, cancer survivors who report depressive disorders appear more likely to use cannabis, especially if they are aged younger than 40 years, female, White, single, unemployed, or have low household income (12). Cigarette smoking among cancer patients and survivors may be associated with an increased likelihood of cannabis use (13). Further, there is evidence suggesting that smoking cigarettes may decrease the therapeutic effects of cannabis, potentially by altering the way cannabinoids are metabolized and distributed throughout the body (14).

In sum, recent findings related to cannabis use among cancer patients and survivors in the United States are scattered. There is, thus, a need for synthesizing the findings from population-based, observational studies to understand prevalence and patterns of cannabis use, sociodemographic and other correlates of cannabis use, and positive and adverse experiences of cannabis use in this vulnerable population. This systematic review attempted to address this need. In addition, this review sought to provide an overview of the current state of research in the area in terms of research design and sample characteristics.

Methods

Literature search

A systematic search of the literature was performed using PubMed and PsycINFO, with several combinations of the following and related search words: cannabis, marijuana, cancer patients, cancer survivors, and current treatment. Searches were not restricted by publication date. A manual search was then conducted between May 2022 and November 2022.

Study selection and eligibility criteria

Two authors independently reviewed the studies identified through the search for eligibility. To be included in the review, the studies needed to 1) be observational studies focused on cannabis use among cancer patients or survivors (as opposed to interventional studies where cannabis was assigned to patients or survivors); 2) be empirical, survey-based studies (as opposed to qualitative studies based on interviews and focus groups); 3) involve data collection from human participants; 4) have been published in English, peer-reviewed journals; and 5) be studies conducted in the United States. Studies were excluded if they were 1) intervention studies (eg, clinical trials) where cannabis was provided by researchers to participants; 2) animal studies; 3) review articles (ie, any review of the primary literature); 4) opinion pieces; 5) commentaries; 6) case studies, where data were not provided by participants; 7) studies conducted in countries other than the United States; 8) studies published in non-English languages; and 9) studies published as book chapters or in nonpeer-reviewed journals.

Manual searches of the references of the included articles were also performed, and the Web of Science Cited Reference Search tool, which identifies other articles referencing a specific study, was used for key articles. This review protocol was registered (CRD42023414835) in International Prospective Register of Systematic Reviews database prior to beginning the review. The procedures used in this review followed the Preferred Reporting Items for Systematic reviews and Meta-Analyses protocol (15).

The search identified 1162 studies of which 832 articles were removed as duplicates. All papers from the automated database searches were collated using the Endnote reference management software. After duplicates were deleted, screening was conducted to ensure that studies fulfilled the eligibility criteria. Of the remaining 330 papers, 285 were rejected for not meeting the eligibility criteria based on the review of titles and abstracts. The remaining 45 papers were screened based on the full text. A total of 27 studies were deemed eligible for inclusion in this review. Two reviewers independently assessed the included studies; discrepancies were discussed and resolved by discussion. Intercoder agreement was measured by K statistic. Reasons for exclusion were clearly documented by each reviewer.

Data extraction and reporting method

A data extraction spreadsheet was used to assess whether articles met inclusion criteria and to compare studies. For each study, the first (SA) and second (SC) author collected the following information: 1) general characteristics of the study (author, publication year, study design, recruitment, and other study and sample characteristics); 2) cancer characteristics; 3) cannabis use characteristics; and 4) main findings.

We were unable to perform meta-analysis because of varied research questions, outcomes, and diverse study designs across studies. Instead, a narrative review was conducted to systematically organize and summarize the evidence using narrative synthesis. Narrative review refers to describing the characteristics of review-eligible studies, identifying common themes or patterns across the literature, and providing an overall interpretation of the findings. This narrative approach is particularly useful when the studies are too diverse or methodologically heterogeneous for statistical integration. Thus, each study was described, followed by comparative analysis and synthesis. To evaluate the quality of each included study, we used the Joanna Briggs Institute (JBI) Critical Appraisal assessment checklist (16). Study quality and potential biases were evaluated using the checklist to appraise the information provided regarding each of the following criteria: study participants, sampling methods, validity and reliability of measurement instruments, study follow-up rates, statistical analysis, and consideration of confounding factors. The JBI checklist assesses whether a study addressed each of these characteristics adequately based on a “yes,” “no,” “unclear,” or “not applicable” response for each criterion. In the current analysis the total number of yes responses was counted for each study. A higher number of yes responses indicated a lower risk of bias.

Results

A flow diagram showing the literature search and the results of the study selection process is shown in Figure 1. The K statistic measuring intercoder agreement was strong (0.81).

Figure 1.

PRISMA flow diagram for identification, screening, eligibility, and inclusion of studies in the review.

Quality appraisal

Based on the JBI checklist, 55% (11 of 20) of the cross-sectional studies that have been included in this review may be judged as being of high quality: the number of yes responses ranged from 7 to 8 (of 8) for these studies. Quality appraisal of the longitudinal studies showed that 57% (4 of 7) of the studies scored inadequately on the JBI items (total number of yes responses were 3 of 11), and thus, these studies may not be considered of high quality. Across all included cross-sectional and longitudinal studies, the average number of yes responses was 6.3 and 4.0, respectively.

Study characteristics

Across the 27 studies (17-43), 7 studies were longitudinal (17,26,27,29,34,39,41), 9 reported cross-sectional data from ongoing longitudinal studies (18,21,23-25,33,35-37), and 11 were cross-sectional studies (19,20,22,28,30-32,38,40,42,43). Of the 7 longitudinal studies, 2 studies (17,29) were prospective and 5 were retrospective studies (26,27,34,39,41).

Studies were conducted in different states of the United States (ie, California, Colorado, Florida, Michigan, Minnesota, New York, Oregon, Pennsylvania, Washington) between 2013 and 2020. Of the studies, 6 were conducted nationwide (18,21,23,24,32,43), and 2 of the studies were conducted in 13 states and 2 US territories (ie, Guam and Puerto Rico) (36,37). Only 4 of the 27 studies were conducted in the western United States (ie, California, Oregon, and Washington) (19,20,28,35).

A total of 12 studies (19,20,23,28,31,35-38,40,42,43) reported survey response rates, which ranged from 31.6% to 100% (20,28). Regarding incentives for research participation, 3 studies (25,38,40) indicated having provided monetary compensation, while 1 study (23) indicated not having provided any form of compensation. The rest of the 23 studies did not report information on incentives.

Participant characteristics

Supplementary Table 1 (available online) summarizes the participant characteristics across studies. The 27 studies (17-43) consisted of 477 388 adult (aged 18 years or older) participants, with 22 studies (17-19,21-23,25-35,39-43) focused on cancer patients and 5 studies (20,24,36-38) on cancer survivors. Participant demographics across studies varied in age and ethnicity. The lowest reported mean age was 31.3 (± 5.5), and the highest was 60.9 (± 11.79) years. Samples across studies were predominantly White: 85% (23 of 27) of the studies included more than 70% White participants. Participants across studies represented almost equal proportions of men (52%) and women (48%).

Cancer characteristics

Supplementary Table 2 (available online) summarizes participants’ cancer-related characteristics. Across 27 studies, 9 studies (17,18,20,21,24,30,36-38) did not report cancer-related characteristics of the cancer patients or cancer survivors. The remaining studies reported different types of cancer-related characteristics: 18 studies reported cancer types (ie, gastrointestinal, genitourinary, gynecological, head and neck, hematological, lymphoma, thoracic cancer) (19,20,23,25-29,31-35,39-43), 11 studies reported cancer stage (ie, stage I, stage II, stage III, or stage IV) (19,20,26,27,29,31-34,39,42), 3 studies reported cancer site (ie, colon, larynx, oral cavity, oropharynx, rectum) (20,28,35), and 2 studies (19,31) reported cancer status (ie, active or remission). The most frequently (20 of 27 studies) reported cancer types, for the samples across studies, were gynecological (ie, breast, cervix, endometrium, ovarian, uterine) and gastrointestinal (ie, colorectal, gallbladder, liver, esophageal, pancreatic cancer). Approximately half of the participants across studies had advanced stage (III or IV) status cancer.

Cannabis use characteristics

Samples across studies varied in terms of cannabis use prevalence. Supplementary Table 2 (available online) provides prevalence rates of cannabis use among participants across the 27 studies. The prevalence of cannabis uses among cancer patients and survivors across studies based on national samples ranged between 4.8% (n = 32 320) (18) and 22% (n = 27 645) (23). Of the 27 studies, 20 (17,18,22,23,25-32,34-36,39-43) reported modes of cannabis consumption, which were as follows: 1) smoking, 2) vaping, 3) topical application, and 4) ingestion of edible products. Of the studies, 12 (21,23,28,29,32,35-37,39,40,42,43), with a pooled sample of 193 406 people, showed 73% (n = 141 332) had smoked cannabis. A total of 15 studies (19,21,23,28,30-32,34-37,40-43), with a pooled sample of 205 419, found 12% (22 824) had vaped cannabis. Topical application was examined in 9 studies (19,29,32,34,39,40-43) representing 1671 people of whom 80% (n = 1351) reported using cannabis topically. Finally, 14 studies (19,21,23,28,29,31,32-37,39,42,43) with a pooled sample of 188 690 people revealed 10% (n = 18 965) had ingested edible cannabis products. Studies (17,31) further detailed that those individuals who use cannabis tend to use more than 1 route of administration (ie, dual [enteral ± vaporized ± oromuscosal] or multiple [enteral + vaporized + oromuscosal]). The frequencies of cannabis use (ie, daily, weekly, or monthly) were reported in 13 of the 27 studies (19-24,28,31,32,37,40-42). Both cannabidiol (CBD) and tetrahydrocannabinol (THC) were commonly used, either alone (19) or in combination (25,29,30,34,39,40,42,43) across these 9 studies.

A total of 8 studies (17,19,25,32,33,40,42,43) reported cannabis acquisition history and indicated that cancer patients and survivors acquire cannabis mainly through medical dispensaries (54.6%, 706 of 1291), recreational dispensaries (30.4%, 332 of 1092), and family and friends (64.8%, 549 of 847). Among the 8 studies that reported cannabis acquisition history, participants in 3 studies (25-27) reported having a registration card for medical marijuana. Four studies (32,38,40,43) assessed where participants obtained information on cannabis. The information sources identified were newspaper, television, websites and blogs, social media, friends and family members, physician, nurse, pharmacist in dispensary, movies or documentaries, coworker or colleague, and advertisements. The most frequently accessed sources of information for cannabis were websites and blogs (57.6%, 533 of 925), medical professionals (47.8%, 416 of 879), and friends and family (46.8%, 377 of 804).

Outcome characteristics

Table 1 summarizes the findings of the reviewed studies.

Table 1.

Included study findings on cannabis use among cancer patient and survivors^a

Reference	Total sample	Proportion of cancer or survivor population (%)	Outcomes
Anderson et al. 2019	1120	Cancer (100)	Symptoms (anxiety, lack of appetite, depression, disturbed sleep, fatigue, nausea, pain, vomiting) reduction: 30% at follow-up Adverse effect occurred: 10%
Azagba et al. 2020	32 320	Cancer (4.7)	Being young: higher (OR = 3.95, 95% CI = 2.55 to 6.13) cannabis use Poverty status: lower (OR = 0.37, 95% CI = 0.24 to 0.59) cannabis use Current cigarette use: higher (OR = 1.86, 95% CI = 1.32 to 2.60) cannabis use
Blake et al. 2019	225	Cancer (100)	Reasons for cannabis use: pain control (n = 41, 68.3%); insomnia (n = 33, 55.0%); anxiety (n = 29, 48.3%); nausea (n = 26, 43.3%); appetite stimulation (n = 21, 35.0%) Prescribed opioid use: ↓ 45.0% (after initiation of cannabis use)
Calcaterra et al. 2020	1784	Cancer survivor (100)	QoL score: lower (aOR = 6.14, 95% CI = 8.07 to 4.20) among cannabis user
Cousins et al. 2021	214	Cancer (1.2) + Survivor (3.8)	Being young (aged 18-34 years): higher perceived cannabis acquisition (impossible vs very easy) among past (OR = 0.82, 95% CI = 0.57 to 1.16) and recent (OR = 0.45, 95% CI = 0.20 to 0.99) cancer Being middle age (aged 35-49 years): higher perceived cannabis acquisition (impossible vs very easy) among past (OR = 0.69, 95% CI = 0.48-0.98) but lower among recent (OR = 0.71, 95% CI = 0.43 to 1.16) cancer Being older (aged 50 years and older): lower perceived cannabis acquisition (impossible vs very easy) among past (OR = 1.09, 95% CI = 0.93 to 1.27) and recent (OR = 0.90, 95% CI = 0.68 to 1.20) cancer Being young (aged 18-34 years): lower perceived risk (great vs no risk) among past (OR = 0.71, 95% CI = 0.50 to 1.00) and recent (OR = 0.46, 95% CI = 0.27 to 0.78) cancer Being middle age (aged 35-49 years): lower perceived risk (great vs no risk) among past (OR = 0.68, 95% CI = 0.51 to 0.90) and recent (OR = 0.59, 95% CI = 0.40 to 0.87) cancer Being older (aged 50 years and older): no difference in perceived risk (great vs no risk) among past (OR = 1.08, 95% CI = 0.94 to 1.25) and recent (OR = 1.03, 95% CI = 0.81 to 1.31) cancer
Cousins & Jannausch et al. 2021	1485	Cancer (4.8)	Pain score: lower (mean = 5.5 [± 2.7]) among cancer cannabis user
Dai et al. 2019	169 036	Cancer (18)	Skin cancer: higher current cannabis use Being young (aged 18-34 years): aOR = 1.5, 95% CI = 0.5 to 4.4 Being adult (aged 35-54 years): aOR = 1.3, 95% CI = 0.8 to 1.9 Being adult (aged 55 years and older): aOR = 1.5, 95% CI = 1.1 to 2.1 Other cancer: higher current cannabis use Being young (aged 18-34 years): aOR = 1.5, 95% CI = 0.5 to 4.4 Being adult (aged 35-54 years): aOR = 1.3, 95% CI = 0.8 to 1.9 Being adult (aged 55 years and older): aOR = 1.5, 95% CI = 1.1 to 2.1
Do et al. 2021	20 720	Survivor (4.9)	Age: lower (OR = 0.55, 95% CI = 0.49 to 0.60) cannabis use Pain: higher (OR = 1.14, 95% CI = 1.06 to 1.23) cannabis use
Donovan et al. 2021	85	Cancer (100)	Being young: cannabis users were statistically younger than nonusers (P = .02) Reasons for cannabis use: sleep (n = 14, 53.8%), anxiety (n = 12, 46.2%), nausea (n = 11, 42.3%), pain relief (n = 10, 38.5%), appetite (n = 10, 38.5%)
Donovan & Chang et al. 2019	816	Cancer (100)	Being male: higher (OR = 1.88, 95% CI = 1.32 to 2.67) cannabis use Being single: higher (OR = 2.28, 95% CI = 1.49 to 3.49) cannabis use Symptoms (moderate to severe): higher cannabis use for tiredness (OR = 1.69, 95% CI = 1.03 to 2.77); lack of appetite (OR = 2.25, 95% CI = 1.54 to 3.29); shortness of breath (OR = 1.55, 95% CI = 1.04 to 2.31); anxiety (OR = 1.57, 95% CI = 1.08 to 2.27); depression (OR = 1.56, 95% CI = 1.08 to 2.25); difficulty sleeping (OR = 2.07, 95% CI = 1.35 to 3.15)
Donovan & Oberoi et al. 2019	171	Cancer (100)	Being male: higher (OR = 2.25, 95% CI = 1.16 to 4.38) cannabis use Being ever smoker: higher (OR = 3.13, 95% CI = 1.47 to 6.68) cannabis use Symptoms (moderate to severe): higher cannabis use for pain (OR = 3.86, 95% CI = 1.31 to 11.35); lack of appetite (OR = 2.25, 95% CI = 1.07 to 6.08); nausea (OR = 3.23, 95% CI = 1.11 to 9.39); constipation (OR = 3.34, 95% CI = 1.37 to 8.16); overall well-being (OR = 5.28, 95% CI = 1.65 to 16.87); difficulty sleeping (OR = 4.26, 95% CI = 1.33 to 13.67)
Eliott et al. 2016	15	Cancer (100)	Symptoms: improved among cannabis user for altered sense (73%), pain (67%), appetite (60%), xerostomia (53%), sticky saliva (47%), difficulty chewing (33%), dysphagia (60%), muscle spasm (47%), weight gain or stability (73%), depression (67%), anxiety (33%)
Fehniger et al. 2021	45	Cancer (100)	Symptoms: 71% improved (at least 1 symptom); pain (56%); nausea or vomiting (47%); anorexia (33%); insomnia (27%) Reason for discontinuation: loss of follow-up (41%); death (24%); lack of symptoms improvement (15%); adverse effect (fatigue and paranoia: 2.2%)
Kaufmann et al. 2022	11 590	Cancer (17.2)	Symptoms: severe or chronic pain (83.4%), severe or persistent muscle spasms (24.6%) Reasons for cannabis use: chronic pain (36.5%) Being older (aged 65 years and older): more likely to use cannabis for cancer (aOR = 3.51, 95% CI = 2.84 to 4.33) more likely to use sublingual tincture vs other types (aOR = 0.24, 95% CI = 0.21 to 0.27) use products with a lower THC to CBD ratio (aOR = 1.24, 95% CI = 1.10 to 1.40)
Macari et al. 2020	188	Cancer (100)	Symptoms: pain ↑ (80.9%), appetite ↑ (77.3%), anxiety ↓ (72.7%), nausea ↓ (58.1%), insomnia ↓ (53.3%), tolerance of treatment ↑ (54.6%) Adverse effects: cloudy thinking is the symptom that worsened the most (16.7%), with decreased energy being experienced by (9.8%) of the cannabis users
Mahurin et al. 2022	119	Cancer (100)	Symptoms: improvements (mean = 6.2/10) Antipruritic treatments: moderate improvement of itching (mean = 6.6/10)
Meghani et al. 2021	136	Cancer (100)	Cancer pain: African American participants were estimated to have a mean least pain score of 1.32 (± 0.48) units higher (ie, indicating lower pain relief) than White participants (P = .006)
Nathan et al. 2022	83	Cancer (100)	Symptoms: pain (↓ 0.51), nausea (↓ 0.36), anorexia (↓ 0.4), insomnia (↑ .26), anxiety (↓ 0.08); none statistically significant QoL score: (↓ 0.06) scores; not statistically significant
Newcomb et al. 2021	1433	Cancer (100)	Reasons for cannabis use: stress (54%), sleep (49%), pain (46%), recreation (41%); cancer (4%) Being young (age 18-44 years) at cancer diagnosis: higher (aOR = 3.30, 95% CI = 2.07 to 5.26) current cannabis use QoL score: lower (aOR = 1.52, 95% CI = 1.14 to 2.04) among current cannabis use Being current smoker: higher (aOR = 1.82, 95% CI = 1.15 to 2.88) current cannabis use
Poghosyan et al. 2020	9325	Cancer Survivor (100)	Being male: higher (aOR = 1.97, 95% CI = 1.37 to 2.83) cannabis use Non-Hispanic Black: higher (aOR = 2.31, 95% CI = 1.22 to 4.36) cannabis use Being unmarried: higher (aOR = 1.64, 95% CI = 1.10 to 2.51) cannabis use History of depression: higher (aOR = 1.75, 95% CI = 1.20 to 2.55) cannabis use Being current smoker: higher (aOR = 3.16, 95% CI = 2.04 to 4.98) cannabis use Being former smoker: higher (aOR = 2.42, 95% CI = 1.60 to 3.65) cannabis use Reasons for cannabis use: medical (n = 103, 1.1%), nonmedical (n = 258, 2.7%), both medical and nonmedical (n = 156, 1.6%)
Poghosyan et al. 2021	10 799	Cancer Survivor (100)	Being depressed: higher (aOR = 2.2, 95% CI = 1.3 to 3.7) among daily cannabis user Reasons for cannabis use: medical (n = 278, 47.0%), nonmedical (n = 109, 20.3%), both medical and nonmedical (n = 169, 32.7%)
Potts et al. 2021	194	Cancer Survivor (100)	Reasons for cannabis use: pain (73.3%), stress and/or anxiety (77.9%) Being married: higher (B = −3.37, 95% CI = −4.57 to 2.18; P < .001) potential motives to use among past year cannabis user High income: higher (B = 1.31, 95% CI = 0.56 to 2.07; P = .001) potential motives to use among past year cannabis user Low income: higher (B = −0.79, 95% CI = −1.40 to 0.18; P = .012) actual motives to use among past year cannabis user
Raghunathan et al. 2022	163	Cancer (100)	Symptoms and adverse effects: Among past CBD users: commonly reported benefits were less pain (21%) or anxiety (17%) and improvement in sleep (15%); 92% reported no side effects Among those with past THC use: reported benefits included improvement in appetite (40%), sleep (32%), nausea (28%), and pain (17%); side effects included feeling “high (40%)”
Reblin et al. 2019	73	Cancer (100)	Symptoms vs relief effects: pain (58.3% vs 85.7%), nausea (54.2% vs 92.3%), appetite (47.8% vs 100%), relaxation (52.2% vs 100%), emotional coping (39.1% vs 90%), spasm and/or tremors (21.7% vs 100%), sleep (8.7% vs 100%)
Sura et al. 2022	184	Cancer (100)	Symptoms: (patients who took at least 1 dose of medical marijuana) pain improvement (48.1%), improvement in at least 1 symptom including pain, nausea, and loss of appetite (85.1%) Adverse effects: low (3.72%) Prescribed opioid use: ↓ 44.9% (after initiation of cannabis use)
Tofthagen et al. 2019	162	Cancer (100)	Symptoms: improved nausea and/or vomiting (44.4%), anxiety (50.0%), or sleep (67.9%) Adverse effect: impaired mental functioning (feeling fuzzy or foggy; 5.5%), dry mouth (3.7%), and paranoid and/or thinking or anxiety (3.0%) Reason for cannabis use: pain (59.3%), anxiety (50.6%), sleep (69.1%), depression (27.8%), appetite (44.4%), numbness and/or tingling (24.7%), nausea and/or vomiting (45.1%), diarrhea (8%) Reason for discontinuation: cost (42.8%), societal stigma (17.8%)
Weiss et al. 2021	612	Cancer (100)	Reason for cannabis use: pain (78%), insomnia (70%), anxiety (57%), stress (51%), nausea and/or vomiting (46%)

^aaOR = adjusted odds ration; CBD = cannabidiol; CI = confidence interval; OR = odds ratio; QoL = quality of life; THC = tetrahydrocannabinol; ↑, increase; ↓, decrease.

Demographic factors and cannabis use outcome

Participants’ sociodemographic characteristics assessed across studies included age (18,21,23-25,30,35), gender (26,27,36), race and ethnicity (33), marital status (26,36,38), and income (18,38). Meghani et al. (33) found that in the absence of cannabis use, African American cancer patients tended to report higher levels of pain score compared with White cancer patients; however, African American cancer patients who consumed cannabis did not differ in pain levels from White cancer patients.

Studies found young age to be associated with a more positive and open-minded attitude toward cannabis use (18,23,25,35). Generally, younger individuals were found to show higher prevalence of cannabis use (18,23,25,35), except for 1 study (30) in which older age was associated with increased medical cannabis use. Three studies showed higher cannabis use among men (26,27,36).

Socioeconomic status was also found to impact cannabis use: lower income was associated with lower prevalence of and/or preference for cannabis use (18,38). Three studies also examined the likelihood of cannabis use based on marital status (26,35,38). Their findings were mixed. One study with cancer patients reported greater cannabis use among single (26) or unmarried (36) individuals, whereas another study with cancer survivors reported greater use among married individuals (38).

Reasons for cannabis use

Studies generally indicated that cancer patients and survivors use cannabis for management of symptoms due to cancer or cancer treatment (17,26-32,34,39-42). These symptoms included pain, loss of appetite, nausea, lack of sleep, and increased anxiety.

Positive experiences with cannabis use

Across the 27 studies, 13 studies (17,26-32,34,39-42) reported that cannabis use resulted in symptom reduction. Generally, these studies reported that individuals experienced subjective relief in various domains such as pain, appetite, nausea, sleep, and anxiety. Three studies reported that individuals who use cannabis are more likely to report lower subjective pain (22,24,33). Another study found that cannabis use was associated with moderate mitigation of itching (32). Two studies found that initiation of regular cannabis use reduced the use of prescribed opioids use among cancer patients (19,41). Raghunathan et al. (39) found that among cancer patients who medicated with CBD, the most commonly reported benefits were decreased pain (21%) and anxiety (17%), and improvement in sleep (15%), and among those who medicated with THC, the most commonly reported benefits were improved appetite (40%) and sleep (32%) and reduced nausea (28%) and pain (17%) (39).

Adverse experiences with cannabis use

A total of 6 studies assessed the adverse effects of cannabis use (17,31,37,39,41,42). Of the 6 studies, 3 found minimal adverse effects (17,39,41). Generally observed adverse effects included cloudy thinking (31,41,42), decreased energy (31), impaired mental functioning (42), dry mouth (42), and paranoid thinking (42) and anxiety (42). Three studies linked higher depression with increased cannabis use (26,36,37). Raghunathan et al. (39) showed that a large majority of patients who medicated with CBD (92%) tend to report no side effects, whereas 40% of the patients who medicated with THC tend to report “feeling high” as a side effect (39).

Discontinuation of cannabis use

Two studies identified factors that led patients to discontinue using cannabis (29,42). These included lack of symptom improvement, adverse effects (fatigue and paranoia), and cost and social stigma.

QoL and cannabis use

QoL was evaluated in 3 studies (20,34,35). These studies found that cancer patients and survivors who used cannabis had lower QoL than those who did not use cannabis. The evaluation of QoL in each these 3 studies was conducted by using a different QoL instrument. For example, Calcaterra et al. (20) employed the Quality of Life Questionnaire Core 30; Nathan et al. (34) used the Edmonton Symptom Assessment Score questionnaire covering pain, nausea, anorexia, anxiety, insomnia, and QoL; and Newcomb et al. (35) applied the Functional Assessment of Cancer Therapy-Colon (FACT-C). The FACT-C includes subcategories that focus on physical, social, emotional, functional well-being, and colorectal cancer-related concerns.

Smoking status and cannabis use outcome

Studies found that individuals who are current smokers or who have history of smoking tobacco are more likely to use cannabis (18,27,35,36).

Discussion

The current review set out to synthesize findings of population-based, observational studies on cannabis use among cancer patients and survivors in the United States. In particular, we attempted to examine the current state of the research in terms of research design, regions where the research has been conducted, and sample characteristics. In addition, we attempted to synthesize the findings in terms of cannabis use prevalence and characteristics among cancer patients and survivors, reasons for cannabis use, positive and adverse experiences of cannabis use, and sociodemographic and other correlates of cannabis use.

It was outside the scope of the current review to examine the impact of states’ cannabis-related laws and policies on cannabis use outcomes among cancer patients and survivors. Only 2 studies (32,43) included in the current review appeared to examine the relationship between cannabis legalization status and cancer patients’ likelihood of using cannabis. According to Mahurin et al. (32), patients tend to report increased likelihood of using cannabis if cannabis were legalized in their state. However, they didn’t ask about cannabis laws for each respondent’s state of residence but stated that people in states without legal medical cannabis might get it illegally by smoking or vaping (32). Weiss et al. (43) found that in their sample of cancer patients, those who medicated with cannabis came predominantly from states where medical cannabis was legal. These finding in general agree with findings from a previous study in the general US population that found that cannabis legalization in one’s state is associated with lower perceived risks associated with cannabis (44).

Only 3 among the currently reviewed studies (18,23,38) reported on sampling methods, suggesting a possible methodological limitation. However, 6 studies (18,21,23,24,32,43) analyzed large national datasets, likely utilizing more rigorous sampling approaches, though specific details about sampling were not provided. Moreover, 9 studies (18,21,23-25,33,35-37) collected data as part of parent longitudinal studies; hence, it is possible that sampling strategies for these studies have been described elsewhere. In general, however, as noted by the majority of the studies themselves, studies were mostly based on selective samples, which may have limited their generalizability. Future studies based on rigorous sampling methods may be needed to better understand cannabis use among cancer patients and survivors in the United States.

Types of cannabis reported in the studies also varied in that some studies focused on the mode of consumption (eg, smoking, eating edibles), while others focused on the cannabinoid type (ie, THC and CBD). A recent study reported that oral (ie, sublingual) cannabis products may be associated with short-term improvements in chronic pain but increased risk for dizziness and sedation (45). Future studies should consider reporting modes of cannabis intake and properties of cannabis consumed (eg, THC vs CBD) separately in relation to outcomes and patients’ cancer stages and sites.

Most studies reported age to be negatively associated with cannabis use; however, 1 study (30) noted that older age was associated positively with medicating with cannabis for symptom management. Several of the studies that included a majority of younger participants did not specify the participants’ reasons for using cannabis (eg, medical vs recreational) (18,21,23,25,35). There is a possibility that the association between age and cannabis use among cancer patients and survivors may be partly explained by the purpose of cannabis use (10,46). Future studies need to examine how reasons for using cannabis are associated with cannabis use among cancer patients and survivors.

Although several studies reported race and/or ethnicity of the participants, the breakdown of minority groups (eg, Asian, Pacific Islanders, Native Americans) differed across studies, making comparative analyses across different racial and ethnic groups difficult (15,47). One study examined racial differences between White and African American cancer patients, finding that African American cancer patients who did not use cannabis reported statistically lower pain scores than White patients, whereas no difference was observed among those who use cannabis (33). No other studies examined racial and ethnic differences in this current review, which implies that future studies should consider involving more ethnically diverse samples so that analyses pertaining to racial and ethnic differences are possible. Cannabis-related policy changes have been shown to increase rates of cannabis use across most racial and ethnic populations, but it is currently unclear how cannabis policies affect cannabis use in cancer patients across racial and ethnic groups (47).

The current evidence suggests that specific symptoms of pain, nausea and vomiting, loss of appetite and cachexia, anxiety, sleep disturbance, and medical trauma prompt patients with cancer to medicate with cannabis (1). However, there is also some evidence suggesting that short- and long-term cannabis use may have adverse neuropsychological effects (48). Thus, the oncologists recommending medical cannabis may need to carefully evaluate the potential risks and benefits of cannabis use for individual cancer patients. In specialty services where many patients use cannabis for medical purposes, such as oncology, less than one-third of clinicians report feeling comfortable with their knowledge of medical cannabis, or of recommending its use (49). Cancer patients desired to receive information about cannabis use during their treatment from the oncology providers (50).

Oncologists can play a vital role in reducing the stigma associated with cannabis use and ensuring that patients have access to accurate information to make informed decisions about their health care. Through evidence-based evaluation, communication, patient education, and collaboration with other health professionals, oncologists may guide patients in making informed decisions about incorporating medical cannabis into their overall cancer treatment plan. Oncologists can collaborate with pharmacists and palliative care specialists to ensure comprehensive and coordinated care for their patients. Importantly, however, all health-care providers must consider potential side effects, drug interactions, and individual patient factors when considering medical cannabis as part of a comprehensive treatment plan.

The present study also revealed that cancer patients and survivors who used cannabis had lower QoL than those who did not use cannabis. But the use of diverse measurement tools for assessing QoL across the studies (20,34,35) poses a challenge when synthesizing and comparing the results. For instance, Calcaterra et al. (20) used the Quality-of-Life Questionnaire Core 30, while Nathan et al. (34) relied on the Edmonton Symptom Assessment Score, a commonly used tool particularly in palliative care contexts. Even though both studies had a primary focus on symptom assessment–based QoL, they employed different instruments to measure QoL and were based in clinical settings. In addition, Newcomb et al. (35) used the FACT-C, which encompasses a range of physical, social, emotional, and functional well-being categories specific to colorectal cancer patients. Calcaterra et al. (20) and Newcomb et al. (35) used separate tools to assess QoL scores in their cross-sectional studies, both showing an association between cannabis use and lower QoL. However, Nathan et al. (34) employed a different measurement tool in their longitudinal study where the mean scores of QoL showed no difference before (4.85 [± 2.5]) and after (4.89 [± 2.4]) the initiation of medical cannabis among cancer patients. This methodological diversity underscores the need for care while interpreting and integrating the findings from these studies. Future research should prioritize standardizing QoL measurement tools for cancer patients and survivors using cannabis to ensure more consistent and comparable assessments.

There are some limitations to the current systematic review, which are partly related to the limitations across the studies reviewed. A key limitation across studies has been the unstandardized and inconsistent reporting about cannabis product formulation, route of administration, stages, anatomical site of cancer, sampling method, and purpose of cannabis use. We did not assess any differences in medicating with cannabis and cancer type because of predefined scope and objectives of our current review; however, this highlights a potential avenue for future research. Another limitation is that the present review compared heterogeneous studies that used various study designs (cohort vs cross-sectional studies); therefore, we were unable to assess publication bias because of the diverse outcomes reported in the studies. In addition, data from the studies published in languages other than English were not included. However, the current systematic review summarized studies conducted in the past 10 years on understanding the characteristics and patterns of cannabis use among cancer patients and survivors in the United States. This study is statistically for providing an overview of the current state of population-based, observational studies on cannabis use among cancer patients and survivors in the United States.

Data availability

This systematic review synthesized data exclusively from the published literature. A comprehensive search of PubMed and PsycINFO databases was conducted to identify relevant studies published that met the predetermined eligibility criteria. All data analyzed in this systematic review are available in the public domain.

Author contributions

Samia Amin, PhD (Formal analysis; Methodology; Writing—original draft; Writing—review & editing), Si Woo Chae, MA (Formal analysis; Methodology; Writing—original draft; Writing—review & editing), Crissy T. Kawamoto, BSc (Writing—review & editing), Kristina T. Phillips, PhD (Writing—review & editing) and Pallav Pokhrel, PhD (Conceptualization; Methodology; Supervision; Writing—original draft; Writing—review & editing).

Funding

No funding was used for this study.

Conflicts of interest

None.