Clinical Profiles of Concurrent Cannabis Use in Chronic Pain: A CHOIR Study

Abstract

Objective

Despite evidence of the analgesic benefits of cannabis, there remains a relative scarcity of research on the short- and long-term effects of cannabis use in individuals with chronic pain.

Design

The current study is a secondary analysis of clinical data from the Collaborative Health Outcomes Information Registry (CHOIR).

Setting

Data were drawn from a cohort of patients of a multidisciplinary tertiary care pain clinic.

Subjects

The study sample consisted of data from 7,026 new patient visits from CHOIR; of these, 1,668 patients with a follow-up time point within 180 days were included in a longitudinal analysis.

Methods

Clinical data were analyzed to characterize cross-sectional differences in pain and indicators of psychological and physical function according to self-reported, concurrent cannabis use. Additionally, a propensity score–weighted longitudinal analysis was conducted, examining cannabis use as a predictor of changes in clinical variables across time.

Results

Cross-sectional analyses suggested significantly poorer sleep and significantly higher intensities of pain, emotional distress, and physical and social dysfunction in patients reporting ongoing cannabis use; however, these differences were relatively small in magnitude. However, no differences between cannabis users and nonusers in terms of longitudinal changes in clinical variables were noted.

Discussion

Our results are among the first to examine concurrent cannabis use as a prognostic variable regarding trajectories of pain-related variables in tertiary care. Future studies may benefit from examining the effect of cannabis initiation, concurrent medication use, and specific aspects of cannabis use (dose, duration of use, or cannabis type) on clinical outcomes.

Introduction

Interest in cannabis as a treatment for chronic pain has grown steadily through the past decade. However, the base of evidence supporting the use of cannabis in treating the physical, functional, and psychosocial detriments associated with chronic pain conditions remains limited. Use of cannabis in the management of chronic pain has been associated with reductions in pain intensity and unpleasantness [1], and meta-analytic results suggest significant short-term reduction in chronic neuropathic pain for one of every five to six patients treated [2]. Among patients presenting to pain clinics, roughly one out of three report some level of perceived benefit from cannabis use in terms of their pain [3]. Further, a recent one-year, controlled observational safety study of cannabis use reported generally good tolerability, with mostly mild–moderate and expected adverse events for chronic pain management [4]. Some studies also note improvements in self-reported mood, pain, muscle spasms, sleep, and quality of life, as well as decreases in opioid prescription [3,5]. Improvements in sleep initiation, overall sleep quality, and depressive and anxious symptomatology have also been noted [1,2], though mood improvements are not apparent in all chronic pain populations [6].

However, there may be significant variability in the analgesic benefits of cannabis, and some of this variability may be due to heterogeneity in terminology and chemical composition of the substances under examination. Herbal marijuana (most commonly Cannabis sativa or Cannabis indica) and synthetic substances such as dronabinol or nabilone act on cannabinoid receptors in the central and peripheral nervous systems, and the chemical agents most active in these substances appear to be delta-9-tetrahydrocannabinol (THC) or cannabidiol (CBD), which appear to have distinct psychoactive effects [7]. Although it is beyond the scope of the current paper to discuss these differences in detail, excellent reviews exist elsewhere on this topic [7,8]. It is notable, however, that these factors appear to have implications for differential analgesic benefit: Whereas moderate-dose THC appears to contribute to analgesia in response to a capsaicin-based pain stimulus, high-dose THC has been associated with greater pain intensity in healthy individuals [6]. Additionally, most existing studies (beyond the Ware article [4] reported above) reporting high-quality evidence for analgesia are also experimental or short-term clinical trials lasting no longer than seven weeks [1,2,6,9,10]. In fact, recent reviews have highlighted the need for long-term pragmatic studies of the potential effectiveness of cannabis for patients with chronic pain [11]. Prior clinical trials must also be considered in light of limitations, including small sample size and limited exposure to study drug and follow-up [1,2,6,9,10,12,13].

Further, there are several potential risks related to long-term cannabis use, including short-term memory deficits, increased risks of depression, anxiety, and psychotic symptoms, addiction, impaired functional outcomes (e.g., poor school performance, lower income, increased criminal behavior, unemployment), and impaired brain development in younger people who use cannabis in the long term [7,14]. It should be noted that these reviews may encapsulate research from both recreational and medically driven use of cannabis, which may complicate interpretation of these findings. Taken together, these issues highlight the need for additional research concerning the short-term and long-term effects of cannabis use, particularly in individuals coping with symptoms that are likely to persist in the long term (e.g., chronic pain).

Prior studies have described the patterns and prevalence of cannabis use among chronic pain populations [15]. The current study sought to evaluate the effects of cannabis use on pain characteristics, sleep, functional indicators, and psychosocial function within a sample of self-reported cannabis users utilizing data from the Collaborative Health Outcomes Information Registry (CHOIR) within a tertiary pain clinic setting. Given the paucity of data concerning the clinical characteristics of cannabis use in chronic pain, our analysis was undertaken primarily as a descriptive study for patients presenting for treatment at a tertiary care center for their chronic pain conditions.

Methods

Procedure

The current study is a secondary analysis of data from the CHOIR pain data registry (http://choir.stanford.edu). CHOIR data were collected in a tertiary care pain clinic as part of patients’ initial clinic visits and follow-up visits. In the current study, data from patients who completed an optional cannabis use questionnaire were included in an analysis examining the cross-sectional and longitudinal relationships between self-reported cannabis use and various clinical variables (e.g., pain intensity, measures of mood, function, and sleep). Prior CHOIR publications have utilized both cross-sectional data [16,17] and longitudinal data [18,19]; however, no publications have presented data related to cannabis use from CHOIR. Patients from this tertiary care center may receive a combination of analgesic medications and conservative management approaches, interventional pain therapies (e.g., injections, infusions, or implantable devices), psychological interventions (e.g., cognitive–behavioral therapy or Acceptance and Commitment Therapy), and/or physical modalities (e.g., physical therapy). PROMIS measures were administered using a computerized adaptive testing (CAT) approach, which employs item response theory (IRT)–based estimation that performs comparably to traditional static assessment forms in terms of reliability while reducing respondent burden by decreasing the number of items necessary to establish a reliable estimate [20–22].

Participants

Data from initial clinic visits of 7,025 patients from a tertiary care pain clinic were examined; of these, 899 endorsed current cannabis use (12.8% of the sample). Patients who completed an additional CHOIR assessment within 180 days after their first assessment were included in the longitudinal analysis (mean number of days after initial evaluation = 81.5). Of the patients who completed a follow-up assessment (N = 1668), 223 (13.4%) reported ongoing cannabis use. Patients were asked to endorse the degree to which their use of cannabis was medical and recreational on an 11-point numeric rating scale (NRS), with higher scores reflecting stronger endorsement for that reason for use; as this was an optional question with a low frequency of response (total N at initial visit = 190), we opted not to include this variable in further analyses and present it only to characterize the current sample. Among respondents who answered these questions, there was a strong endorsement of cannabis use for medical reasons (mean [SD] = 8.35 [2.58]) and a low level of endorsed recreational use (mean [SD] = 2.21 [1.86]). The sample was 67.2% female, predominantly married (54.4% of the overall sample), and had a mean age of 49.9 years; 42.8% of the sample reported current employment, and 16.8% reported being on some form of disability. Patients self-reported the etiology of their pain using several broad categories that were not mutually exclusive. According to these categories, 32.7% of the sample reported pain related to nerve issues, 21.7% reported a prior diagnosis of arthritis or fibromyalgia, 18.1% reported pain related to muscle issues, 14.4% reported pain stemming from a disk issue, 8.9% reported issues with pain related to their bones, 2.9% reported pain stemming from the effects of an infection, and 2.1% reported cancer-related pain. Notably, 33.2% of the sample reported that their pain was of an undiagnosed or unknown etiology. Average pain intensity on a 0–10 NRS over the past seven days (SD) in the overall sample was 5.45 (2.21), and average duration of pain (SD) was 8.16 (10.4) years.

Measures

PROMIS Assessments

Patient-Reported Outcome Measurement Information System (PROMIS) instruments assessing depression, anxiety, pain interference, sleep disturbance, social isolation, and fatigue were utilized. Depression items assess primarily affective and cognitive aspects of depression such as feelings of hopelessness, sadness, guilt, loneliness, anhedonia, and decreased positive emotions, while anxiety items assess fear, dread, worry, and indices of physiological hyperarousal (e.g., dizziness, tension) [23]. Pain interference items assess the extent to which pain symptoms interfere with various domains of life function (e.g., social, recreational, occupational, or physical activities) [24]. Sleep disturbance items assess perceptions related to the restorativeness, quality, and depth of sleep [24], while fatigue items were designed to assess a range of feelings of subjective tiredness or exhaustion that may be prominent enough to interfere with daily function [24]. Social isolation items assess primarily the perception of being alone, avoided by others, or detached from others [25]. All PROMIS measures are scored on a t score metric, with a mean score of t = 50 and a standard deviation of 10.

Pain Intensity

Average pain intensity over the previous seven days was rated on an 11-point numerical rating scale from 0 to 10, where 0 refers to no pain at all and 10 refers to the worst pain imaginable. This assessment of pain intensity has been validated for use in chronic pain populations [26].

Pain Catastrophizing

Pain catastrophizing was assessed using the Pain Catastrophizing Scale, a 13-item scale assessing maladaptive cognitive and affective responses to pain. Prior factor analytic studies have suggested that the PCS is comprised of three factors: feelings of helplessness related to pain, an inability to disengage from negative thoughts about pain, and magnification of the intensity or negative consequences of pain [27]. Items are coded on a scale from 0 (“not at all”) to 4 (“all the time”); the PCS is scored as a sum total of these items ranging from 0 to 52, with higher scores corresponding to greater levels of catastrophic appraisal of pain. The internal consistency of this measure in the current sample was high (Cronbach’s α = 0.954).

Cannabis Use

Current cannabis use was assessed at patients’ initial clinic visit using a binary (“yes/no”) response variable. Patients were prompted while completing other CHOIR measures with the following text: “Patients have asked whether the Stanford Pain Management Center prescribes medical cannabis cards. While this is not a service we provide, we would like to understand how our patients as a group use medical cannabis. Below are a few optional questions. Your responses will not impact your individual medical care at the Pain Management Center. Are you using cannabis?” If patients responded affirmatively to the initial cannabis use item, other questions were administered assessing dose, duration, and frequency of cannabis use, as well as respondents’ perceived reasons for use (e.g., medical necessity, social or recreational purposes, or spiritual purposes). However, patient responses to aspects of cannabis use, such as duration and frequency of use, were completed inconsistently and in the form of free-text responses, which substantially reduced the available data for analysis. Further, other aspects of the questionnaire (focused on the quantity of CBD or THC dosage) were filled out very infrequently and raised concerns about the quality of patient responses. As a result, the current analysis focused on self-reported binary use of cannabis only.

Analytic Plan

Cross-sectional analyses were conducted using SPSS, version 25. Demographic differences between cannabis users and nonusers were computed using independent-samples t tests for continuous variables that demonstrated homogeneity of variance between groups, Welch’s t test for continuous variables that demonstrated heterogeneity of variance between groups, and chi-square tests (for categorical data). Due to the nature of the racial status variable, only identifiable racial categories were included; data from patients who declined to answer or who did not know their racial status were not included in this analysis. Similarly, use of independent samples and Welch’s t tests were computed to assess differences in clinical variables between patients who endorsed cannabis use and those who did not during their initial clinic visit. Effect sizes for the cross-sectional t tests were computed using Hedges’ g with a pooled standard deviation, which was utilized due to the unequal sample size between groups [28]. The following variables were specified as outcomes in cross-sectional and longitudinal analysis: average pain intensity, depressive symptoms, anxiety symptoms, sleep disturbance, fatigue, social isolation, pain interference, and pain catastrophizing.

Regarding longitudinal analyses, we assumed that preexisting differences between patients on clinical variables at initial visit might bias the results of the longitudinal analyses. Given that patients were not randomly assigned to cannabis use, propensity weighting was used to control for preexisting differences between cannabis users and nonusers. Logistic regression models were weighted by the inverse of the probability of cannabis use. Covariates included the risk factors listed in the patient sample and baseline differences. The distribution of propensity scores (PS) across exposures was checked for balance and overlap. A stabilization procedure was used to reduce the risk of undue influence from data points with abnormally large weights [29]. Overlap in predicted probabilities suggested that propensity analysis was appropriate (range = 0.002–0.069). Sensitivity analysis was also performed after trimming extreme propensity scores (<0.5% or >99.5%). Stabilized weights were trimmed according to the method proposed by Harder and colleagues [30]. Unadjusted absolute value of standardized differences between cannabis users and nonusers exceeded 25% for nine of 14 covariates. After applying inverse probability weights, all standardized differences were significantly reduced and had an absolute value of <10% (which is less than the recommended upper limit of 0.25, a frequently used criterion for adequacy of covariate balancing) [30,31]. Longitudinal analysis on outcomes was performed used a generalized estimating equation (GEE) analysis from baseline to a follow-up time point within 180 days using the SAS GENMOD command (SAS 7.15; SAS Institute Inc, Cary, NC, USA). Models were estimated examining a cannabis use-by-time interaction in changes in each clinical variable across time. Given the relatively large number of outcome variables on which we were testing group differences, we opted for a conservative adjustment for multiple comparisons in the alpha level (α = 0.05/8 = 0.00625).

Results

Cross-Sectional Analysis

Demographic Differences

The demographic characteristics of the sample were determined using baseline clinical data. A larger proportion of male patients (16.3%) than female patients (11.3%) reported current use (χ²(1, N = 6851) = 32.4, P < 0.001). Racial status was also significantly associated with differential rates of cannabis use (χ²(7, N = 6385) = 51.3, P < 0.001); these results are presented in Table 1. Cannabis use did not vary between Hispanic and non-Hispanic patients (χ²(1, N = 6396) = 2.45, P = 0.12). Patients endorsing current cannabis use were also significantly younger (45.8 vs. 50.5, Welch’s t(6376) = 7.91, P < 0.001). No significant differences in pain duration were noted between cannabis users and nonusers (t(4147) = 0.59, P = 0.56).

Table 1.

Racial composition of patient sample

	No Cannabis, No. (%)	Cannabis, No. (%)	Total No.
American Indian or Alaskan Native	40 (88.9)	5 (11.1)	45
Asian or Asian American	535 (96.4)	20 (3.6)	555
Black or African American	198 (85.7)	33 (14.3)	231
Native Hawaiian or Pacific Islander	26 (89.7)	3 (10.3)	29
Caucasian	3,580 (85.6)	602 (14.4)	4,182
Other	1,174 (87.4)	169 (12.6)	1,343
Unknown	222 (89.2)	27 (10.8)	249
Patient declined to answer	140 (90.9)	14 (9.1)	154

Clinical Variables

At their initial clinic visits, patients reporting current cannabis use reported significantly higher levels of pain (t(1224.6) = 4.88, P < 0.001) and pain-related interference in daily activities (t(1192.4) = 10.42, P < 0.001), compared with those who denied current cannabis use (see Table 2 for group differences and effect sizes for each clinical variable). Similarly, patients reporting cannabis use showed greater levels of psychological and psychosocial distress, with significantly higher levels of depressive symptoms (t(1152.6) = 11.81, P < 0.001), anxious symptoms (t(11.67) = 11.67, P < 0.001), pain catastrophizing (t(6754) = 7.81, P < 0.001), sleep disturbance (t(6561) = 7.37, P < 0.001), fatigue (t(1137.7) = 9.12, P < 0.001), and social isolation (t(6460) = 12.05, P < 0.001). Examination of Hedges’ g scores for relative effect size produced the largest differences between cannabis users and nonusers, found in social isolation, depression, anxiety, fatigue, and pain interference, with scores ranging from 0.32 to 0.41, suggesting small to moderate differences. Small group differences were also noted between cannabis users and nonusers on pain catastrophizing, sleep disturbance, and pain intensity.

Table 2.

Baseline differences between self-reported cannabis users and nonusers in clinical variables

		No.	Mean	SD	Hedges’ g
Average pain intensity	Cannabis	899	5.77	2.08	0.16
	No cannabis	6,125	5.41	2.23
Pain interference	Cannabis	841	66.23	6.55	0.34
	No cannabis	5,818	63.66	7.60
Fatigue	Cannabis	838	61.62	9.46	0.32
	No cannabis	5,760	58.40	10.18
Depression	Cannabis	836	57.43	9.03	0.41
	No cannabis	5,750	53.43	9.98
Anxiety	Cannabis	834	58.51	9.09	0.41
	No cannabis	5,737	54.54	9.90
Sleep disturbance	Cannabis	833	58.60	8.94	0.27
	No cannabis	5,730	56.04	9.40
Social isolation	Cannabis	825	51.24	9.22	0.45
	No cannabis	5,637	46.96	9.56
Pain catastrophizing	Cannabis	856	25.09	12.72	0.29
	No cannabis	5,900	21.42	12.86

All t tests significant at P < 0.001.

Longitudinal Analysis

Using this propensity score–weighted group, no interactions between cannabis use and time were found in predicting any clinical variables (P > 0.05 in all cases; see Table 3). This lack of significant cannabis use-by-time interactions suggested that self-reported cannabis users and nonusers did not show significantly different levels of change in examined clinical variables across time.

Table 3.

Summary table of longitudinal analyses

Outcome	Intercept (B, SE)	Time Effect (B, SE)	Cannabis Group Effect (B, SE)	Interaction (B, SE)
Average pain intensity	5.43 (.035)*	–0.344 (0.054)*	0.037 (0.134)	–0.021 (0.158)
Pain interference	63.92 (0.123)*	–0.924 (0.198)*	0.228 (0.492)	0.862 (0.616)
Fatigue	58.65 (0.164)*	–0.042 (0.220)	0.485 (601)	0.932 (0.636)
Depression	53.89 (0.163)*	–0.060 (0.219)	0.324 (0.565)	0.702 (0.622)
Anxiety	54.96 (0.162)*	0.478 (0.228)**	0.249 (0.592)	1.11 (0.584)***
Sleep disturbance	56.26 (0.152)*	–0.581 (0.216)*	0.601 (0.528)	0.782 (0.686)
Social isolation	47.33 (0.160)*	0.216 (0.209)	0.611 (0.528)	0.366 (0.552)
Pain catastrophizing	22.23 (0.207)	3.39 (0.294)*	0.480 (0.710)	–0.657 (0.852)

All interactions not significant at P < 0.05.

^* P < 0.01;

^** P < 0.05;

^*** P < 0.10.

Discussion

Using a retrospective sample of tertiary care pain clinic patients, we examined self-reported cannabis use as a predictor of concurrent and longitudinal indicators of pain intensity and physical and psychosocial function. Overall, our results were mixed, with cross-sectional analyses associating cannabis use with poorer baseline pain-related measures, but no significant differences were noted between cannabis users and nonusers in terms of longitudinal trajectories of clinical variables.

In our cross-sectional analysis, patients who reported ongoing cannabis use tended toward greater levels of pain intensity, emotional distress, social isolation, fatigue, pain catastrophizing, sleep disturbance, and pain-related interference, compared with those with no current self-reported cannabis use. Cross-sectional differences on clinical variables were small according to statistical standards, but were consistent across all domains. Further, these differences were of a magnitude that met the published thresholds for minimally important differences for pain interference [32,33], depression [34], and anxiety [34]. Notably, the largest differences between patients using and not using cannabis were in psychosocial domains: social isolation, depressive symptoms, and anxious symptoms, whereas the smallest difference was noted in terms of average pain intensity. The results of our longitudinal analysis were less revealing, however. Our analysis indicated no significant differences longitudinally, suggesting that the presence of self-reported cannabis use did not significantly predict differential changes in pain or other indicators of physical or psychosocial health across time.

Although our analytic approach was largely exploratory and we did not initiate our analysis with a directional hypothesis in mind (namely, whether cannabis would predict better or worse levels of pain, sleep, mood, or function), we found no evidence that patients who reported ongoing use of cannabis were better off in terms of any examined clinical variable, either at their initial visits or across time. These findings echo other recent examinations of cannabis use across time, which have indicated no apparent benefit to chronic pain–relevant outcomes such as pain severity [35]. We cannot state, due to the nature of our data collection, whether individuals with chronic pain are at elevated risk for poorer outcomes as a result of cannabis use, or whether individuals who otherwise demonstrate greater levels of pain, distress, and overall health and functional impairments might be more likely to use cannabis as a result of these issues. It is possible (and perhaps likely, given the relatively strong evidence for short-term analgesic benefit for cannabis-naïve patients) that there is a curvilinear effect in terms of the benefits of cannabis in chronic pain. It may be that patients begin to experience more significant pain relief at the outset of cannabis use and then, across time, begin to show increased tolerance and poorer overall adaptation to pain as its benefits decline (perhaps in concert with increasing symptoms of cannabis dependence). Supporting this hypothesis are data from a previous longitudinal study, which reported improvements in individuals with chronic pain one year after initiation of medical cannabis use [4]. However, as we did not begin tracking cannabis use at its outset, and only collected data about current use (which may have been ongoing for months or years), these interpretations are speculative and warrant additional, focused study in the future. We urge prospective studies that may better answer these questions, particularly given the continued and high need for additional nonopioid interventions for chronic pain.

Limitations

Our results should be interpreted with an understanding of the study’s limitations, however. First, the nature of our data set was a convenience sample of treatment-seeking individuals with chronic pain, who presented with relatively high levels of pain intensity, a long duration of pain (roughly eight years on average), and high levels of physical impairment and psychosocial distress; these individuals may be qualitatively different from individuals with more episodic or shorter-term pain complaints. In addition, the nature of our data collection, nested within a multidisciplinary, tertiary care clinical setting, must be acknowledged in terms of interpreting our findings for a few reasons. First, the cannabis use variable was optional, and a large proportion of patients may have opted not to complete this assessment, either due to a concern about how this would affect their care or due to other reasons. Second, cannabis use was assessed using a low-resolution question: a binary “yes/no” indicator of current cannabis use. We were not able to reliably assess duration, frequency, method, or amount of cannabis use in this sample, nor did we have any objective measure (e.g., urine drug screens) that would confirm cannabis use. We also did not distinguish between recreational and medical use in this assessment, leaving the possibility of further heterogeneity in terms of how cannabis may have been used by respondents. Similarly, we did not have available THC or CBD concentrations for ongoing use; although some patients also provided information for these domains, the data were highly variable in terms of both their availability and apparent quality and did not appear suitable for immediate analysis.

It is also notable that our longitudinal analysis was marked by a high degree of data attrition (i.e., having a reduced sample of 1,668 patients from the initial sample of roughly 7,000 who had follow-up data to be included). It is possible that there was some characteristic shared by the patients who followed up regularly for care and completed follow-up CHOIR assessments that obscured the nature of any longitudinal relationships between cannabis use and clinical outcomes. A prospective data collection examining the effectiveness of cannabis in individuals with chronic pain may yield different results; improvements in pain and other outcomes have been shown in longitudinal controlled follow-up studies of cannabis for chronic pain [4], and these findings need to be replicated.

Future Directions

As noted previously, we urge replication of our findings in other samples of individuals with chronic pain, as well as greater depth of assessment (regarding duration, frequency, and dosing) in terms of cannabis use. To date, evidence regarding the clinical implications of medical vs nonmedical use of cannabis has been mixed. Among medical cannabis users, there is some evidence of lower levels of pain among patients with longer-term cannabis use compared with those who are just initiating use of this substance [36], but also evidence of poorer self-rated health among medical cannabis users compared with those endorsing recreational use [37]. There is a relatively clearer indication of increased vulnerability among nonmedical cannabis users to anxiety disorders and substance use disorders, however [38], and licensed medical cannabis users may demonstrate distinct patterns of cannabis use compared with unlicensed and nonmedical users [39]. As our data regarding reasons for use were available only in a subset of our patient sample and were highly skewed toward strictly medical use, we opted to exclude them from our analysis, but these factors are nevertheless key considerations for future research in cannabis use in this population. Further, although more objective indicators of cannabis use (e.g., recent urine drug testing results) were not available in the current patient sample, they may have provided a clearer indication of current use of cannabis in the current study sample and may be beneficial in future studies.

In addition to clearer indicators of current cannabis use, there are other clinical variables that may prove to be valuable for future research. Recent evidence indicates that regular cannabis use has been associated with lower body mass index [40], and differences in BMI may have implications for both indicators of psychosocial function [41] and response to treatment in chronic pain [42]. Another key consideration for future studies may be the specificity of the pain conditions being treated: Some evidence has emerged for the analgesic benefit of cannabis in some forms of neuropathy [1,2], though others have shown relatively less benefit [6]. Another key consideration for future studies concerns other pharmacological interventions for pain that may not have been assessed or appropriately represented in our analysis. Given the highly heterogeneous nature of our pain sample and the high level of variability in terms of ongoing treatment for each patient, we were not able to capture, for example, co-occurring opioid or nonopioid analgesic medication use as a key covariate in our study. Recent evidence is emerging of the potential utility of cannabis in decreasing chronic opioid use in individuals with chronic pain [5,43]; as a result, it may be that the role of cannabis as a potential adjuvant medication (particularly in the context of reducing or replacing opioid use) was not adequately represented in our sample.

Conclusions

The current study examined the cross-sectional and longitudinal relationships of self-reported cannabis use with indicators of pain, mood, sleep, and psychosocial function in a large sample of treatment-seeking individuals with chronic pain. In general, our results suggested worse pain and poorer overall function at baseline among those individuals who endorsed ongoing use of cannabis that persisted across time, though we are not able to definitively state whether worse pain and function are consequences of or predictors of ongoing cannabis use. The current study provides a comparison point against which more systematic, controlled prospective studies may compare efficacy for cannabis use as a treatment for chronic pain.

 

Funding sources: The research reported in this publication was supported by the National Institute on Drug Abuse (NIDA) of the National Institutes of Health (NIH) under Award Number T32DA035165 (JAS), NIDA R01 DA045027 (JMH), NIH K24DA029262 (SCM), NIH T32 DA035165 (SCM), and the Redlich Pain Endowment (SCM).

Disclaimer: The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

Disclosure and conflicts of interest: JAS is on the scientific advisory board of TribeRx. MAW is an employee of Canopy Growth Corporation; Canopy did not provide resources toward or have influence, financially or otherwise, over the conception or development of the manuscript. The authors have no other disclosures of any additional funding or any relationships that might lead to a conflict of interest.