Weeding through the haze: a survey on cannabis use among people living with Parkinson’s disease in the US

Symptomatic management of Parkinson’s disease (PD) is complex and many symptoms, especially non-motor symptoms, are not effectively addressed with current medications. In the US, cannabis has become more widely available for medical and recreational use, permitting those in the PD community to try alternative means of symptom control. However, little is known about the attitudes towards, and experiences with, cannabis use among those living with PD. To address this shortcoming, we distributed an anonymous survey to 7,607 people with PD in January 2020 and received 1339 responses (17.6%). 1064 complete responses were available for analysis. Respondents represented 49 states with a mean age of 71.2 years (±8.3) and mean PD duration of 7.4 years (±6.2). About a quarter of respondents (24.5%) reported cannabis use within the previous six months. Age and gender were found to be predictors of cannabis use in this sample (Age OR = 0.95, 95% CI 0.93 to 0.97; Male OR = 1.44, 95% CI 1.03 to 2.03). Users reported learning about cannabis use from the internet/news (30.5%) and friends or other people with PD (26.0%). Cannabis users were more likely to report insufficient control of their non-motor symptoms with prescription medications than non-users (p = 0.03). Cannabis was primarily used for PD (63.6%) and was most often used to treat nonmotor symptoms of anxiety (45.5%), pain (44.0%), and sleep disorders (44.0%). However, nearly a quarter of users (23.0%) also reported they had stopped cannabis use in the previous six months, primarily due to a lack of symptom improvement (35.5%). Three quarters of respondents (75.5%) did not use cannabis, primarily because there was a lack of scientific evidence supporting efficacy (59.9%). Our results suggest that the lack of formal guidance or research evidence about cannabis for PD may in part underlie inconsistencies in both use and reported effectiveness.


Introduction
Parkinson's disease (PD) is the second most common neurodegenerative disorder, affecting more than 1 million Americans at a cost to society of more than $50 Billion dollars. 1,2 While PD is typically defined clinically by four cardinal motor symptoms that include rest tremor, bradykinesia, rigidity and postural instability, 3 the symptoms of PD are broader . Non-motor manifestations of PD are varied and include sleep disturbances such as insomnia, sleep fragmentation and REM sleep behavior disorder (RBD), cognitive changes, depression, anxiety, hallucinations, hyposmia, pain and autonomic dysfunction (constipation, orthostatic hypotension and urinary incontinence). 4,5 Additionally, treatment of PD with levodopa or other dopaminergic agents is commonly associated with involuntary movements (levodopa-induced dyskinesias, LID) that may be troublesome for some patients. The efficacious control of LID and of non-motor features represents a major therapeutic challenge that is often unmet for many of those living with PD. 5 Therefore, many of those living with PD in the US are interested in complementary and alternative or integrative therapies (CAIM) that may bring relief to their troubling non-motor symptoms. 6,7 The medicinal use of cannabis represents a novel, alternative approach toward PD symptom control. Preclinical evidence suggests that cannabinoids could be widely beneficial to neurodegenerative diseases, including PD. 8,9 A small survey in Colorado of individuals with PD found that the handful of identified cannabis users reported it to be among the most effective of complementary therapies. 10,11 However, there remains a lack of evidence of both efficacy and safety to support the recommendation of the use of cannabis as a therapy for PD. 12,13 Nevertheless, there is an increasing interest in the use of medical cannabis. In one recent survey, 95% of movement disorder specialist neurologists reported being asked to prescribe medical cannabis to their patients. 14 Currently, 47 of 50 states in the US permit the sale use, or consumption of hemp-sourced cannabidiol (CBD) products containing less than 0.3% cannabis, the two most common reasons selected were for medical reasons related to their PD only (63.6%, 166/261), or for the combination of both medical and nonmedical (e.g., recreational) reasons (21.5%, 56/261). For those who used cannabis explicitly for their PD, about half (46.8%,104/222) indicated they used cannabis for PD symptoms in general and a similar number (43.2%, 96/222) indicated they used cannabis for specific PD symptoms (see Table 2). Comparable findings were identified through thematic analysis of open-text responses. Among the 246 users (85.1%) who reported their cannabis use was for medical or for both medical and non-medical reasons, the most commonly identified reasons for trying cannabis included curiosity (22.8%, 56/246), the idea that cannabis was a natural substance (16.7%, 41/246) and word of mouth (11.4%, 28/246).
We next sought to understand how cannabis users acquired their knowledge about cannabis as a potential therapy (see Table 2). For the majority of users who used cannabis for any medical or for both medical and non-medical reasons (excluding recreational users and users who did not know their reason for use), the most common sources of information were from the internet/news (30.5%, 75/246) and from friends or other people with PD (26.0%, 64/246). Most of these users (64.4%, 168/246) had not received a prior recommendation to use cannabis from a licensed doctor or provider.
Following up our question on knowledge, we asked these medical and medical and non-medical users how they learned how to use cannabis, such as dosage, type and frequency of use. We found that a majority of users (56.1%, 138/246) reported that they were not provided any information about recommendations on how to use cannabis (such as dosage, type and frequency of use). Of the users who were provided information (43.9%, 108/246), the most common sources of information were reported to be a non-PD doctor (25.0%, 27/108), staff at a dispensary (22.2%, 24/108) and friends or other people with PD (22.2%, 24/108).
Among all users, we also sought to better understand the cannabis formulations, strengths and routes of administration they most often used. Two-thirds (64.4%, 168/261) considered themselves PRN or "as needed" users and the remainder (35.6%, 93/261) considered themselves "regular" users. However, men were more likely to report identifying as a "regular" user than women (OR = 1.75, 95% CI 1.04 to 2.94). The daily routine of cannabis use varied and a summary can be found in Table 3. The majority of users reported their time of use at the end of the day, either in the evening (25.3%, 66/261) or at bedtime (26.4%, 69/261, see Cannabis comes in a number of forms, which can impact how rapidly it is absorbed and for how long it persists in the body. 18 Users of cannabis most frequently reported spraying or dropping (for example, sublingual drops) (29.1%, 76/261), smoking (27.2%, 71/261) and eating or swallowing (19.2%, 50/261) as their primary method of cannabis use (Fig. 3B). Users acquired cannabis most commonly through a medical dispensary (38.7%, 101/261) or a family member or friend (24.5%, 64/261, see Fig. 3C). When asked details about the type of cannabis used, such as strain of cannabis and concentration of CBD and THC, including high THC, low THC, high CBD, etc., about a quarter of all users did not know (22.2%, 58/261, see Fig. 3D). Of those who did (77.8%, 203/261), almost half did not know the specific type (48.3%, 98/203) or dosage (47.0%, 95/203) they used. Chi-square tests revealed significant differences between the selected reasons for cannabis use (PD medical, other medical, non-medical or both medical and non-medical) and the time of day when it was used (p ≪ 0.001), method of use ( p < 0.001), where the cannabis was acquired (p ≪ 0.001), and the type of cannabis used (p < 0.003). Among users, there was also a significant difference between men's and women's reasons for cannabis use (p < 0.02) and the type of cannabis used (p < 0.04).

Symptom management by cannabis
We asked users who reported use for symptom management whether they felt that cannabis addressed their motor and non-motor symptoms. Of these users, most (72.0%, 144/200) reported that its use did not address or only partially addressed their motor symptoms. A nearly equally large majority (70.0%, 140/200) reported that cannabis use did not address or only somewhat addressed their non-motor symptoms.
Respondents who reported using cannabis for general or for specific PD symptoms (90.1%, 200/261) were also asked about perceived symptom improvement. The most common nonmotor symptoms that cannabis users were trying to treat with their cannabis use were anxiety (45.5%, 91/200), pain (44.0%, 88/200) and sleep disorders (44.0%, 88/200). The most common motor symptoms that cannabis users were trying to treat with their cannabis use included stiffness (43.0%, 86/200) and tremor (42.0%, 84/200). Additional symptom information can be found in Fig. 4. There were no differences found between type of cannabis used and the top selected symptoms for anxiety, pain, sleep disturbances, stiffness and tremor. However, there were differences found between type of cannabis used and the selected symptoms, appetite, dystonia and urinary symptoms, but due to the low selection rates, further analysis was not conducted.

Cannabis user experiences
A small portion of users (12.6% 33/261) reported negative side effects from cannabis use. We also polled non-users of cannabis who had indicated a previous negative experience as a . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted June 26, 2020.
About a quarter of cannabis users (23.0%, 60/261) had stopped using cannabis in the previous six months. Among these users as well as the subset of non-users who had indicated a previous negative experience (4.1%, 33/803) a lack of symptom improvement (35.5%, 33/93) was the most commonly selected reason for discontinued use (Fig. 5). Among users who had not discontinued use in the previous six months, the most commonly identified items that cannabis use allowed for included general symptom improvement (40.3%, 81/201), relaxation (14.4%, 29/201) and improvement in daily activities (9.0%, 18/201).

Medication satisfaction
All respondents (both users and non-users) reported on their prescription medication use. A few reported not taking medications for PD motor (6.6%, 70/1064) or for non-motor (5.2%, 55/1064) symptoms. Very few (4.6%, 49/1064) were taking no medication at all, and this group had a significantly shorter disease duration than those taking medication (p << 0.001). However, those not taking PD prescription medications were no more or less likely to report cannabis usage. Among all respondents, men were more likely to be taking prescription medications for PD motor symptoms (OR = 2.11, 95% CI 1.28 to 3.50) or for PD non-motor symptoms (OR = 2.03, 95% CI 1.15 to 3.56) than women.
We asked those taking PD medications whether their prescription medications addressed their motor and non-motor symptoms. For motor symptom management, about half (45.3%, 482/1064) reported that medications mostly addressed their symptoms. For non-motor symptom management, a little less than half (41.5%, 441/1064) reported that medications somewhat addressed their symptoms. There were no reported differences between users and non-users regarding mean levels to which medication addressed motor symptoms; however, there were significant differences between users and non-users for non-motor symptoms, with 67.4 % (165/245) of cannabis users reporting that medications were less effective at addressing their non-motor symptoms than non-users (p < 0.005).
Despite users being less satisfied with prescription medications, they still rated their prescription medications as more efficacious than cannabis for control of both motor and non-motor symptoms. Among users, the majority (84.5%, 202/239) reported that their prescription medication addressed motor symptoms better than or equal to cannabis (p ≪ 0.001), and 74.9% (179/239) of users reported that their prescription medication addressed non-motor symptoms better than or equal to cannabis, but this difference was not significant (p = 0.052).
Among all users, the majority (84.7%, 221/261) reported that cannabis use had no impact on their prescription medication usage, and of the users who had indicated cannabis use for both . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted June 26, 2020. ; https://doi.org/10.1101/2020.06.24.20139162 doi: medRxiv preprint medical and medical and non-medical reasons, most (89.0%, 219/246) reported that they had not thought that cannabis would be a replacement for their PD prescription medications when they began using.

Clinical Trials
We queried all respondents (both users and non-users) about their level of interest in learning about and enrolling in a clinical trial exploring the impact of cannabis on PD symptoms. Most (82.3%, 876/1064) were interested in learning more information about a clinical trial, and more than half (62.3%, 663/1064) were interested in enrolling in a clinical trial. Upon further analysis, users were found to be more likely to be interested in learning more information (OR = 3.63, 95% CI 2.18 to 6.02) and more likely to be interested in clinical trial enrollment (OR = 3.18, 95% CI 2.27 to 4.45) than non-users.

Discussion
We examined cannabis use, knowledge, motivations, and routines among people living with PD. This study is unique in that we captured both medicinal and recreational cannabis use among a large number of people living with PD from nearly all fifty states. From our population sample, nearly a quarter of those surveyed (24.5%) reported the use of cannabis. This is within the wide range of cannabis usage reported among people living with PD in the literature (between 4% and 80%). 10,11,14,19 However, our reported number of users may be higher because we assessed both the recreational and medicinal use of cannabis while other studies cited assessed medicinal use only.
Our results suggest that although there are many people with PD using cannabis as a CAIM treatment for their motor and non-motor symptoms, the lack of formal guidance about cannabis usage for PD may underlie inconsistencies in use and reported effectiveness. The survey results also indicate a knowledge gap among people living with PD, which may be impacting their decisions about cannabis use, as a large portion of users reported not knowing the type, brand or dosage of cannabis they primarily use. This knowledge gap, combined with the receipt of information and recommendations primarily through the internet, friends or other people living with PD or word of mouth, as opposed to a licensed health provider, could be partially responsible for the perceived confusion about expectations and efficacy. We would expect that specific cannabis formulations, ratio of CBD to THC, route of administration, and strain (sativa vs indica) might have a differential impact of specific symptoms that patients are trying to address. 12,13,18 With these factors ignored or unknown and without an individualized and educated approach, the potential for success may be limited and the potential for adverse reactions may be amplified. However, in this study, users who did not know the type of cannabis they used compared to those who did know did not report significantly different levels to which cannabis addressed motor and non-motor symptoms.
Among some users, reported cannabis use may have been beneficial for particular motor and non-motor PD symptoms including anxiety, pain, sleep disorders, stiffness and tremor. These symptoms are consistent with other research studies examining motor and non-motor effects of cannabis on PD symptoms. 11, 14,[20][21][22][23][24] In this study, users of cannabis reported it most improved . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted June 26, 2020. ; the severity of anxiety, pain and sleep disorders. However, overall use and the level to which cannabis was reported to address motor and non-motor symptoms varied widely. This was particularly evident by both the reported low levels of motor and non-motor symptom relief by cannabis compared to prescription medications and the 23.0% discontinuance rate in the previous six months, which was greater than the 14% reported in in a survey of MS patients. 25 Here, discontinued use was primarily reported due to a lack of symptom improvement. These findings are in line with the inconsistent efficacy observed in the limited clinical trials conducted on cannabis and PD. 12,26,27 We found that a majority of people living with PD and using cannabis, were doing so without having been provided information or provider recommendations. This was not entirely surprising; a previous study found a significant cannabis knowledge gap among specialized Parkinson's disease clinicians that paralleled our observed knowledge gap in people with PD. In that survey; the majority of clinicians reported they lacked the formal training and knowledge of cannabis as a treatment alternative for PD motor and non-motor symptoms. 14 It is important that people living with PD are provided information from reputable, evidencebased sources. Considering the growing legality of cannabis access and customization of cannabis formulations marketed to people living with PD, additional research is needed for both clinicians and people with PD to better understand the efficacy of cannabis for PD motor and non-motor symptoms. If cannabis is intended to be used as a medicinal or CAIM treatment for PD, formal recommendations are needed to fill the current knowledge gap around cannabis type, dosage and efficacy. Any future research or clinical trials targeting the efficacy of cannabis should clearly define these recommendations for use and prioritize the communication of such recommendations to enrolled participants.
The large number of non-users reported here suggest that cannabis use may not be viewed by people living with PD as a primary medicinal treatment option for PD. Although 33 states have legalized cannabis use for medical conditions, PD is recognized as a qualifying medical condition in only 17 of these states. 12 In this study, 36.4% of cannabis users had other medical, recreational or a combination of reasons for cannabis use, and the identified reasons for use were associated with variability in responses to use routines and the reported level to which cannabis addressed PD symptoms.
Cannabis use did not have an impact on PD prescription medications. Indeed, nearly all respondents (89%) reported that they knew cannabis would not be a replacement for their current medications. However this finding differs from other studies, in which prescription medication use decreased among cannabis users. 19 Despite the perceived hype around cannabis for PD, 17 a healthy skepticism seems apparent among non-users. Nearly threequarters (76%) would consider cannabis use only if there were scientific evidence to support the claimed benefits. Our results would suggest that people living with PD expect and could benefit from more evidence about use and outcomes.
. CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted June 26, 2020. ; https://doi.org/10.1101/2020.06. 24.20139162 doi: medRxiv preprint It is notable that pain is one of the symptoms best described in the literature that has evidence supporting use of cannabis in other populations such as HIV neuropathy and cancer related pain. 28,29 However, those results have been recently called into question. 30 Supporting the point that patients look towards evidence, here, one of the main uses reported for cannabis among those with PD was for the amelioration of pain symptoms. As evidence is generated around the efficacy of cannabis for PD, such evidence is expected to influence how cannabis use is viewed as a PD treatment option in the future.
There are several limitations to this study. The study response rate of 14% may not have been representative of the PD population. Users may have been more inclined to respond to the survey and report favorable results than non-users, which could have influenced and inflated the number of users and benefits leading to a possible response bias. However, the large number of non-users also participating may have mitigated this potential response bias by users. All respondents were required to complete the survey electronically, and this survey invitation was sent only to those who had an active email address. Given that PD predominantly affects the older population, it is possible that the technology and email requirements excluded several who did not have an active email address or access to an electronic or mobile device. It is similarly noteworthy that the average educational level (and presumably socioeconomic status) of the respondents was high. Thus, our respondents may be more motivated to participate in such surveys, more comfortable with the use of electronic technology, and potentially less concerned about legal ramifications of their participation. Although cannabis is recreationally legal in 11 states and medicinally legal in 33 states, 12 it remains a United States Drug Enforcement Administration Schedule I substance. For this reason, people living with PD may have been reluctant to report their cannabis use, leading to an undercount of user prevalence. Policy around cannabis varies globally and use likely does so as well. This study was limited to the United States and is therefore not reflective of other countries' cannabis use trends or perceptions. The demographics reflected in this study primarily consisted of white, married and highly educated respondents, which likely do not reflect the entire PD population.
In summary, a large survey of individuals with PD regarding their attitudes and knowledge of cannabis revealed that 1 in 5 of people with PD are current cannabis users. Reassuringly, most of these users recognize that cannabis is not a substitute for their current medications and also see the limited efficacy of cannabis for symptom management. Many people with PD justifiably are not using cannabis because of a lack of evidence to support its use with PD. Despite the lack of evidentiary support, it is a concern that PD cannabis users lack resources to guide them on the potential use of cannabis for PD. PD cannabis users identify several non-motor symptoms that may be responsive to cannabis and should be the target of future clinical experiments to see if this reported experience bares out in a larger population.

Standard Protocol Approvals, Registrations, and Patient Consents
The Western Institutional Review Board (WIRB) approved this study. Given the low-risk nature of the protocol and anonymous data collection, the Board found that this research meets the . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted June 26, 2020.

Patient Surveys
A review of publicly available surveys about cannabis use or perceptions of use was conducted, and relevant questions were used or modified to fit this survey. 14,31-35 The survey was then reviewed by three people living with PD to ensure relevance and readability.
An invitation to an anonymous, electronic survey was emailed on January 17, 2020 to 7,607 people with Parkinson's disease. These individuals were identified as having attended an online or in-person Parkinson's Foundation educational event or having called the Parkinson's Foundation Helpline in the 2019 calendar year. Reminder emails, two in total, were sent a week after the initial invitation and a week before the survey closed. The survey remained open for three weeks and closed on February 7, 2020.
The online questionnaire consisted of seven sections: 1. A screening question to confirm that the respondent was a person with Parkinson's disease (number of questions, n = 1). The full questionnaire can be viewed in the Appendix.

A question
Of the 1,339 surveys returned, 261 were incomplete and were discarded from further analysis. There were 1,078 completed responses, however, five were discarded due to incomplete or incorrect geographic information (e.g. one, four or eight digit zip codes) and an additional nine were discarded due to suspected inaccuracy of birth year and diagnosis year (e.g. reported year of birth after 2010 or reported year of diagnosis equal to reported year of birth). 1,064 completed responses were available for analysis. However, when results were further analyzed by gender, five respondents who had selected other or prefer not to answer as their gender were removed from the gender analysis portion of the reported results.

Statistical Analysis
The anonymous survey was collected and managed using Research Electronic Data Capture (REDCap) tools hosted at the Parkinson's Foundation. 36,37 Statistical analyses were performed using the programming language R. 38 Determination of associations between demographic and user status and reported outcomes was performed using Student's t-test for continuous variables and Pearson's Chi-squared test (including the Monte Carlo procedure for p-value approximation where cell counts were below 5) for categorical variables. Variables rating demographic characteristics, user stats and reported outcomes that were dichotomous or could be dichotomized were examined using Pearson's Chi-squared test and unadjusted odds ratios. P < 0.05 was considered statistically significant. Open text collected in the survey were evaluated through thematic analysis.

Data Availability
The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request. . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted June 26, 2020. . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted June 26, 2020. ;  . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted June 26, 2020. ; https://doi.org/10.1101/2020.06.24.20139162 doi: medRxiv preprint Figure 1: A Geographic representation of survey participants. Blue shading within a state depicts the percentage of respondents who were cannabis users. Montana and Maine all had 100% use from the survey participants. Rhode Island was the only state without survey respondents.

Figure 2:
Primary reasons for not using cannabis in the past 6 months (n=803). Responses were either pre-identified in the survey as listed in the category labels or individuals selected "other" and their open text response was analyzed thematically.
. CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted June 26, 2020. ; https://doi.org/10.1101/2020.06.24.20139162 doi: medRxiv preprint Figure 3: Cannabis use by time of day, method, source, and type. Each cannabis user was asked (A) when they used cannabis during the day, (B) the primary method by which they used cannabis, (C) where they most often acquired their supply of cannabis, and (D) the type of cannabis they used based on the concentration of the two main constituents: THC (tetrahydrocannabinol) and CBD (cannabinol). The results were further categorized by type of user (n=261 for all graphs).
. CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted June 26, 2020. ; https://doi.org/10.1101/2020.06.24.20139162 doi: medRxiv preprint Figure 4: Cannabis use for selected PD symptoms. Each cannabis user for PD medical reasons (just PD and both PD and non-medical, n = 200) was asked, "For which of the following symptoms do you use marijuana or cannabis to improve?" . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted June 26, 2020. ; https://doi.org/10.1101/2020.06.24.20139162 doi: medRxiv preprint Figure 5: Reported reason for cannabis cessation. Users (n=93) responded to the survey question "For what reason did you stop using marijuana or cannabis?" Responses were either pre-identified as one of the survey options or individuals selected "other" and their open text response was analyzed thematically. Lack of symptom improvement was a key reason for halting cannabis use.
. CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted June 26, 2020. ; https://doi.org/10.1101/2020.06.24.20139162 doi: medRxiv preprint