A Large-Scale Naturalistic Examination of the Acute Effects of Cannabis on Pain

Abstract

Introduction: Cannabis use for pain relief is commonly reported, yet laboratory studies and clinical trials suggest that cannabinoids are weak analgesics, and it is unclear whether perceived reductions in pain from before to after cannabis use relate to factors such as dose, method of administration, phytocannabinoid content, or the age or gender of the user. We determined whether inhalation of cannabis decreased self-reported pain ratings as well as whether user gender, age, time, method of administration, tetrahydrocannabinol (THC)/cannabidiol (CBD) content, or dose of cannabis contribute to changes in these ratings. We also examined whether tolerance may develop to the analgesic effects of cannabis over time.

Materials and Methods: Archival data were obtained from Strainprint^®, a medical cannabis app that allows patients to track symptoms before and after using different strains and doses of cannabis. Latent change score models and multilevel models were used to analyze data from 131,582 sessions in which inhaled cannabis was used to treat “muscle pain,” “joint pain,” or “nerve pain.”

Results: For all three pain symptoms, severity ratings decreased significantly after cannabis use. Women reported higher baseline and postcannabis pain severity than did men, and men reported larger decreases in pain than did women. Neither THC nor CBD content nor their interaction predicted reductions in pain ratings. However, vaping was associated with larger reductions in joint pain ratings than was smoking, and lower doses were associated with larger reductions in nerve pain ratings. Additionally, for all three pain symptoms, the dose of cannabis used to manage pain increased significantly over time.

Conclusions: Inhaled cannabis reduces self-reported pain severity by ∼42–49%. However, these reductions appear to diminish across time, and patients use larger doses across time, suggesting that analgesic tolerance develops with continued use.

Introduction

Pain is the most commonly reported reason for using medical cannabis,^1–4 and medical users overwhelmingly report that cannabis substantially reduces pain.^2,5 In contrast, human laboratory studies, and randomized trials conducted with chronic pain patients suggest that cannabis and cannabinoids produce limited decreases in pain.^6–9 However, most of those studies have examined single cannabinoids or a fixed ratio of cannabinoids and relied on either oral or oromucosal administration. Studies focusing on inhaled whole plant cannabis remain rare, despite inhalation being the most common method of administration reported by medical cannabis users.²

Factors that have been proposed to influence cannabinoid analgesia include sex/gender of the user,¹⁰ dose, product type,¹¹ method of administration,¹¹ and cannabinoid content of the cannabis used,^12,13 although studies investigating these variables are rare. In one of the few studies of its kind, data from the Releaf™ app were analyzed to examine the influences of method of administration, and cannabidiol (CBD) and tetrahydrocannabinol (THC) content on changes in pain severity ratings from before to after cannabis use.¹¹ Inhaled cannabis reduced pain ratings by 47%, with flower predicting greater reductions than edibles, pills, or tinctures, although there were no significant differences among the different methods of inhalation. Additionally, high levels of THC predicted larger reductions in pain ratings, whereas cannabis high in CBD yielded smaller reductions in pain ratings. Similarly, in a previous study⁵ we analyzed data from the Strainprint^® app to examine changes in headache and migraine severity ratings from before to after inhaling cannabis and to determine whether these changes varied as a function of gender, cannabinoid content, dose, or time. We found a 47% reduction in headache and 50% reduction in migraine severity ratings from before to after cannabis use that were unrelated to cannabinoid content or dose of the cannabis used. Furthermore, men reported larger reductions in headache than did women and there was evidence for the development of tolerance over time.⁵

In the present study, we used a large archival dataset obtained from the medical cannabis app Strainprint to examine changes in self-reported pain severity from before to after cannabis use in a large sample of medical cannabis patients. The first objective was to determine whether ratings of “muscle pain,” “joint pain,” and “nerve pain” would decrease from before to after cannabis use. The second objective was to explore factors that could predict changes in these ratings, including time, gender, age, CBD and THC content, dose, and method of administration.

There is evidence from rodent studies that acute antinociceptive potency of cannabinoids is greater in females than males.^14,15 However, with repeated administration, either no sex differences are observed, or greater antinociceptive effects are seen in male rats.^15,16 Similarly, in studies examining the analgesic effects of cannabis/cannabinoids in chronic cannabis-using humans, either no gender differences^17,18 or greater effects in men have been observed.^5,19 Therefore, we predicted that the self-reported reductions in pain from before to after cannabis use would be greater in men than women.

The final objective of this study was to investigate whether tolerance would develop over time as a function of cannabis use. Although repeated use of cannabis likely produces tolerance to most of its effects in humans and animals,^15,20,21 few studies have examined the extent to which medical users experience tolerance to the putative analgesic effects of cannabis. We hypothesized that the self-reported reductions in pain from before to after cannabis use would decrease over time, and dose would increase with repeated cannabis use over time.

Materials and Methods

Procedure

Archival data were obtained from Strainprint Technologies, a real-world technology platform with a journaling app that allows medical cannabis users to track changes in symptom severity as a function of cannabis use. Immediately before using cannabis, Strainprint users select the condition(s)/symptom(s) they are using cannabis to manage and rate their severity from 0 (none) to 10 (extreme). Users are further prompted to indicate their method of administration, the strain of cannabis they are about to use, and the producer/distributor of that strain by selecting from >3000 cannabis products sold in Canada. Laboratory-verified cannabinoid content (e.g., %THC, %CBD) is prepopulated in the app using data from the websites of cannabis distributors in Canada. If a product is not prepopulated within the app, users can manually enter the cannabinoid content. They further indicate the “dose” (number of puffs) taken. After an onset period determined by the method of administration (e.g., 20 min after inhalation), a push notification is sent to prompt users to rerate their symptom severity after cannabis use. The app also records the time elapsed between the two symptom ratings.

For the present study, we obtained anonymous data from Strainprint for individuals who used the app to track symptoms of “muscle pain,” “joint pain,” and “nerve pain.” This included anonymous ID codes, cannabis treatment session numbers, gender, age, pain symptom, self-reported symptom severity before and after each tracked session of medical cannabis use, time lag between these two symptom ratings, cannabinoid content (%THC, %CBD) for the cannabis used in each session, the method of obtaining the cannabinoid content data (i.e., from cannabis producer vs. app user), as well as the method of administration, and dose. The Human Research Protection Program determined that this study does not meet the criteria of Human Subjects Research and therefore did not require IRB review.

Inclusion/exclusion criteria

Due to potential differences in onset of effects across different routes of administration, we included only tracked sessions involving inhalation methods of administration. Furthermore, only tracked sessions in which symptoms were rerated within 4 h of cannabis use were included, given that the acute effects of inhaled cannabis dissipate after 3–4 h.^22,23 Finally, to increase the reliability and validity of the data, we only included sessions for which laboratory-verified cannabinoid content data was prepopulated within the app.

Participants

The final sample comprised 3548 medical cannabis users who reported using cannabis for pain (1868 women, 1634 men, 46 “other”). This sample collectively used the Strainprint app 131,582 times over a span of 34 months (from March 2017 to January 2020), to track symptoms of muscle pain, joint pain, and/or nerve pain. Table 1 displays demographic characteristics and information on the number of cannabis use sessions tracked for the entire sample, and for subsamples broken down by symptom. Table 2 displays descriptive statistics pertaining to the cannabis constituents and doses used to manage each pain symptom.

Table 1.

Demographic Characteristics and Tracked Sessions

Symptom	Age			Gender (n)		No. of tracked sessions
	N (sample)	M (SE)	Range	Men	Women	N (sessions)	M (SE)	Range
Muscle pain	2494	38.42 (0.04)	18–97	1100	1368	57,204	208.21 (1.32)	1–2132
Joint pain	2164	39.78 (0.18)	18–116	1032	1102	56,679	319.92 (2.39)	1–3563
Nerve pain	863	42.91 (0.08)	17–73	346	507	17,699	227.51 (2.56)	1–2096
Total sample	3548	39.61 (0.03)	17–116	1634	1868	131,582	258.92 (1.24)	1–3563

Sample sizes reported for the total sample do not equal the sum of the three subsamples because numerous people used Strainprint^® to track multiple pain symptoms.

SE, standard error.

Table 2.

Cannabis Use Characteristics

Symptom	Dose (no. of puffs)		THC (%)		CBD (%)
	M (SE)	Range	M (SE)	Range	M (SE)	Range
Muscle pain	9.86 (0.03)	1–31	15.21 (0.03)	0–83.00	2.59 (0.02)	0–46.00
Joint pain	10.29 (0.03)	1–30	14.58 (0.03)	0–86.50	3.16 (0.02)	0–40.60
Nerve pain	9.72 (0.05)	1–30	14.67 (0.05)	0–67.30	3.24 (0.04)	0–46.00
Total sample	10.02 (0.02)	1–31	14.86 (0.02)	0–86.50	2.92 (0.01)	0–46.00

CBD, cannabidiol; THC, tetrahydrocannabinol.

Data analysis

The percentage of cannabis use sessions involving reductions in pain severity ratings following cannabis use were computed for each pain symptom for the entire sample as well as for the two genders. These percentages were compared between men and women using chi-square analyses.

To examine change in symptom severity ratings from before to after cannabis use, two time points' latent change score (LCS) models were used. This approach allowed us to examine changes in pain ratings from before to after cannabis use within subjects across time²⁴ and as a function of specific predictors of interest (session, gender, age, time lag between pre and postcannabis use ratings, method of administration, dose, THC, CBD). Estimates for each of these predictors are interpreted as standardized beta coefficients, which describe the average influence of the predictor variable on the change factor. Positive coefficients indicate that higher levels of the predictor variable are associated with smaller reductions in symptom severity ratings following cannabis use, and negative coefficients indicate associations with larger decreases in symptom severity. The LCS model is specified using a structural equation modeling approach to model the change between “before” and “after” cannabis use as a latent factor. In the final LCS models, the fixed effects of all predictors of interest were specified, as well as both the fixed and random effects of symptom severity before and after cannabis use. A detailed description of this approach is provided elsewhere.⁵ All LCS models were fit using Mplus version 8.3.²⁵

Longitudinal multilevel models were used to examine changes in dose of cannabis used as a function of cannabis use sessions over time. Both the fixed and random linear effects of time/cannabis use session on dose were estimated. All longitudinal multilevel models were fit using SAS Proc Mixed, with maximum likelihood estimation and incomplete data treated using missing at random assumptions. Standardized beta coefficients are reported with values indicating increases in dose over time in standard deviation units. A Bonferroni corrected alpha of 0.017 (0.05/3 types of pain) was used to determine statistical significance for all analyzes.

Results

Overall change in symptom severity

Muscle pain ratings were reduced in 95.5% of cannabis use sessions, joint pain ratings were reduced in 89.8% of sessions, and nerve pain ratings were reduced in 84.8% of sessions. A common index used for responder analyses in pain clinical trials is the percentage of people who report a 30% reduction in pain. Our results revealed that 72%, 71.4%, and 59% of sessions involved a 30% or greater reduction in muscle, joint, and nerve pain ratings, respectively. As depicted in Figure 1, the average magnitude of self-reported pain reductions was 48.6% for muscle pain, 49.6% for joint pain, and 42.1% for nerve pain.

FIG. 1.

Change in muscle, joint, and nerve pain severity ratings from before to after cannabis use. Error bars represent standard error of the means, *p<0.001.

Gender differences

Table 3 displays the percentage of cannabis use sessions for which pain severity ratings were reduced from before to after use in men and women. The table also shows the mean pain severity ratings before and after use, for each of the pain symptoms in each gender. Men reported a greater percentage of sessions than women, during which pain severity ratings were reduced after cannabis use [muscle: χ²(1)=202.19, p<0.001; joint: χ²(1)=62.05, p<0.001; nerve: χ²(1)=116.52, p<0.001]. Additionally, relative to men, women reported significantly higher pain severity both before [muscle: t(56,694)=24.61, p<0.001; joint: t(56,028)=43.15, p<0.001; nerve: t(17,516)=6.84, p<0.001] and after cannabis use [muscle: t(56,694)=65.75, p<0.001; joint: t(56,028)=64.94, p<0.001; nerve: t(17,516)=46.37, p<0.001].

Table 3.

Percentage of Sessions Involving Pain Reduction and Mean Pain Severity Ratings Before and After Cannabis Use, by Symptom and Gender

	Muscle pain		Joint pain		Nerve pain
	Men	Women	Men	Women	Men	Women
% Sessions pain reduction	92.9%	89.5%	90.5%	88.3%	94.7%	79.1%
Mean rating before use	5.97	6.37	5.56	6.32	6.73	6.94
Mean rating after use	2.65	3.82	2.54	3.71	3.02	4.77

LCS models predicting change in symptom severity

As depicted in Figure 1, the fixed effects of baseline LCS models confirmed that the ratings of all pain symptoms were significantly reduced from before to after cannabis use (muscle: μ_Δ=−2.98, standard error [SE]=0.15, p<0.001; joint: μ_Δ=−2.89, SE=0.14, p<0.001; nerve: μ_Δ=−2.87, SE=0.33, p<0.001). The covariance between symptom severity ratings before cannabis use and the latent change factor was statistically significant and negative for all pain symptoms (muscle: covariance estimate=−1.72, SE=0.29, p<0.001; joint: covariance estimate=−2.01, SE=0.20, p<0.001; nerve: covariance estimate=−1.77, SE=0.80, p=0.03), indicating that more severe symptoms before cannabis use were associated with greater reductions in pain severity ratings following cannabis use. The variance (i.e., random effects tested) of the latent change factors were also statistically significant for all pain symptoms (muscle: variance estimate=4.42, SE=0.36, p<0.001; joint: variance estimate=4.38, SE=0.28, p<0.001; nerve: variance estimate=6.24, SE=0.99, p<0.001). This suggests that there were significant differences among individuals in the rate of change in each symptom following cannabis use.

As shown in Table 4, the fixed effects tested in the conditional LCS models predicting symptom rating changes show that men reported significantly larger reductions than women in ratings of all three pain symptoms from before to after cannabis use. Furthermore, vaping was associated with larger reductions in joint pain ratings than smoking, while lower doses and earlier cannabis use sessions predicted larger reductions in nerve pain ratings. There were no other main effects. Models including the interaction between THC and CBD indicated no significant interactions.

Table 4.

Latent Change Score Models with Predictors of Change in Pain Severity Ratings

Predictor	Muscle pain			Joint pain			Nerve pain
	ß	SE	p	ß	SE	P	ß	SE	p
Time/session	0.07	0.09	0.43	0.06	0.05	0.22	0.20	0.08	0.01
Gender (men=1)	−0.24	0.06	<0.001	−0.19	0.04	<0.001	−0.30	0.08	0.001
Age	0.07	0.04	0.09	0.07	0.04	0.03	0.05	0.10	0.60
Time lag	−0.004	0.02	0.87	0.03	0.02	0.24	0.02	0.03	0.43
Method (Vape=1)	−0.07	0.05	0.15	−0.13	0.08	0.01	−0.08	0.08	0.34
Dose	−0.04	0.04	0.39	0.002	0.06	0.97	0.16	0.06	0.003
THC	0.06	0.04	0.16	0.13	0.08	0.11	0.07	0.06	0.19
CBD	0.06	0.05	0.23	0.11	0.06	0.09	0.09	0.06	0.13

Values corresponding to statistically significant predictors (p<0.016) are displayed in bold font. The table presents three separate LCS models, using time/cannabis use session, gender (women=0, men=1), age, time lag between pre and postcannabis use symptom ratings, method of administration (smoke=0, vape=1), dose (no. of puffs), THC content, and CBD content to predict reductions in pain severity ratings. Positive coefficients indicate that higher values of the predictor are associated with smaller reductions in pain severity ratings, whereas negative values indicate that higher values of the predictor are associated with larger reductions in pain severity ratings. Models containing the interaction between THC×CBD were tested, but no significant interactions were detected, and the inclusion of the interaction term did not alter the pattern of results.

Longitudinal multilevel models predicting changes in dose across cannabis use sessions

The multilevel models in which both random and fixed effects of time on dose of cannabis used were specified revealed that the dose of cannabis used to manage all pain symptoms increased significantly across cannabis use sessions over time (muscle: β=0.02, SE=0.002, p<0.001; joint: β=0.02, SE=0.003, p<0.001; nerve: β=0.06, SE=0.01, p<0.001).

Discussion

The results of this large-scale investigation of real-time changes in pain severity ratings from before to after cannabis use revealed that for each type of pain examined, severity ratings decreased in the vast majority of cannabis use sessions (>84%). The size of these reductions was substantial and significant, with joint pain showing the largest reduction (∼50%) and nerve pain showing the smallest reduction (∼42%) in severity ratings after cannabis use, which is remarkably consistent with two previous studies using app data that showed an average 47% reduction in self-reported pain intensity ratings¹¹ and headache intensity ratings⁵ from before to after cannabis use.

Nevertheless, in the present study, there were significant individual differences in the size of these reductions, suggesting that not all people experience equivalent reductions in pain severity. Indeed, consistent with our hypothesis and previous research,^5,19 men reported greater reductions in pain severity than did women for all three types of pain. Men also recorded a significantly higher percentage of sessions than did women, during which pain severity was reduced. While the larger doses men typically use¹⁸ may explain this difference, gender differences in dose were statistically controlled for in the LCS models and, therefore, dose cannot account for these gender differences. Additional research is needed to understand the mechanisms underlying these gender differences.

Dose was associated with symptom change, but only for nerve pain severity ratings, with higher doses associated with smaller reductions in nerve pain ratings. The intractability of nerve pain may have contributed to this finding. Participants who used cannabis for nerve pain reported the highest pre-session pain severity of the pain types examined; additionally, nerve pain was associated with the lowest reduction in symptom severity, the lowest percentage of sessions in which symptom reduction was reported, and the highest percentage of sessions in which symptoms were reported to remain unchanged after cannabis use (Table 3). Nerve/neuropathic pain is known to be challenging to treat, and current drug therapy guidelines include “…weak recommendations against the use of cannabinoids…” (p. 539).²⁶ Randomized controlled trials examining the efficacy of cannabis/cannabinoids for neuropathic pain show significant but clinically small reductions in pain, averaging less than −1 point on a 0–10 scale.²⁶ Moreover, a placebo-controlled study in neuropathic pain patients indicated that moderate doses are more effective for pain than high or low doses of THC.²⁷

We found no evidence that THC, CBD, nor their interaction influenced change in pain severity ratings, which is consistent with our previous findings pertaining to headache/migraine pain.⁵ Nevertheless, these findings conflict somewhat with one previous study, which found that higher levels of THC were associated with greater reductions in pain ratings, whereas higher levels of CBD were associated with smaller reductions.¹¹ However, effects of THC and CBD concentrations found in that previous study were limited to musculoskeletal, headache-related, gastrointestinal, and “other/nonspecified” pain not examined in the present investigation. As such, the role of THC and CBD may vary as a function of pain type. Furthermore, cannabis can contain over 100 cannabinoids, over 250 terpenes, as well as numerous other molecules,^28,29 and therefore it may be one of these other constituents or an entourage effect that is responsible for the perceived therapeutic properties of cannabis on the types of pain examined in our study. Unfortunately, information on other constituents was missing too frequently to permit meaningful analyses.

Results did reveal that vaping was associated with larger reductions in joint pain than smoking. This finding may reflect slightly different pharmacokinetics of these two methods of administration and stronger effects of vaporized versus smoked cannabis.³⁰ Alternatively, given that concentrates are typically vaped, whereas flower is commonly smoked, this finding may represent an effect of the type of cannabis consumed. Consistent with this interpretation, cannabis concentrates have previously been shown to be associated with larger reductions in headache severity ratings than flower.⁵ Future investigations, particularly clinical trials directly comparing the analgesic effects of concentrates and flower as well as smoked flower and vaped flower are needed.

Later cannabis use sessions were associated with smaller reductions in nerve pain ratings than were earlier cannabis use sessions, which suggests that tolerance may develop with repeated cannabis use. Consistent with these findings, the multilevel models indicated that dose of cannabis used to manage all three types of pain increased slightly across cannabis use sessions over time. While the increase in dose used over time may not be indicative of tolerance per se (it may also be driven by factors such as increased pain intensity over time, growing dissatisfaction with pain relief, etc.), these findings are consistent with substantial evidence of tolerance development to other effects of cannabis and cannabinoids in human studies,²⁰ and to the antinociceptive effects of THC in animal studies.^15,21 Taken together with the finding that use of higher doses of cannabis did not result in greater perceived pain reduction (and may even be associated with less pain reduction), the present findings suggest that limiting cannabis consumption may help to maintain pain relief over time.

Limitations and strengths

Limitations to this study include possible lack of clinician-verified diagnoses of pain, and sampling bias. It is likely that the sample overrepresents individuals who perceive cannabis to be beneficial in managing their pain, given that individuals who do not find it beneficial would likely stop using cannabis and the Strainprint app. Also, only a single item was used to assess each pain symptom and standard definitions of these symptoms were not provided for users. Because the data were obtained from medical cannabis users through an app, it was not possible to obtain a placebo control group, nor was it possible to experimentally manipulate cannabis use. As such, we cannot make causal conclusions nor rule out expectancy effects.

These study limitations are offset by several strengths. First, the data were obtained from a very large sample of individuals who were using a variety of cannabis products in their natural environment, which enhances ecological validity. The ability to track changes in different pain symptoms across an extremely large number of cannabis use sessions (>131,000) over a 34-month time period is an additional strength of the study. Indeed, one drawback of randomized controlled trials has been the relatively short period of time over which patients are tracked.⁹

Conclusions

The present findings suggest that cannabis reduces the perceived severity of pain by ∼40–50%. However, tolerance may develop with repeated use. Collectively, these results suggest that cannabis has short-term benefits for reducing pain that may wane over time.

Abbreviations Used

CBD

cannabidiol

LCS

latent change score

THC

tetrahydrocannabinol

Author Disclosure Statement

No competing financial interests exist.

Funding Information

No funding was received for this article.

Cite this article as: Cuttler C, LaFrance EM, Craft RM (2022) A large-scale naturalistic examination of the acute effects of cannabis on pain, Cannabis and Cannabinoid Research 7:1, 93–99, DOI: 10.1089/can.2020.0068.