Practice Points
- Women with multiple sclerosis (MS) may experience greater irritability, sleep disruption, and physical tension during cannabis cessation than men with MS. Clinicians should consider closer follow-up and proactive symptom-management strategies for their patients who are reducing or discontinuing use.
- Because cannabis withdrawal symptoms overlap with common MS-related complaints, clinicians should differentiate among MS-related symptom fluctuations and effects associated with cannabis cessation when evaluating new or worsening symptoms.
- People with heavier cannabis use histories may exhibit greater variability in symptom trajectories during abstinence, underscoring the need for
tailored monitoring plans and individualized
support during cessation.
Individuals with chronic medical or psychiatric conditions may experience heightened or prolonged withdrawal symptoms.1 Multiple sclerosis (MS) is a chronic disease of the central nervous system associated with fatigue, sleep disturbance, mood instability, and cognitive impairment.2 Several of these features, particularly fatigue, sleep disruption, and mood changes, overlap with symptoms reported during cannabis withdrawal,3 potentially intensifying distress and complicating cessation.
In a study of 136 treatment-seeking but otherwise healthy cannabis users (45 women, 91 men), Herrmann et al4 found that women experienced greater and more persistent withdrawal symptoms than men despite comparable use patterns. Neurobiological evidence suggests that sex differences in the endocannabinoid system may contribute to greater withdrawal severity in women.5
Despite consistent findings in the general population, sex-based differences in cannabis withdrawal have not been systematically examined in people with MS. To address this gap, we conducted a secondary analysis of frequent cannabis users with MS from an existing database to evaluate the severity and trajectory of withdrawal symptoms during 28 days of monitored abstinence. We hypothesized that women would report more severe and persistent symptoms than men. This study extends prior work in the same cohort,6,7 in which withdrawal symptoms were not a primary outcome, by examining withdrawal symptom trajectories longitudinally.
Methods
Participants and Design
Nineteen frequent cannabis-using people with MS (11 women) were drawn from a previously published study examining the cognitive and neuroimaging effects of cannabis abstinence in MS.The parent cohort comprised 40 patients recruited from a tertiary MS clinic.7 Demographic and neurological variables included age, education, disease duration, and Expanded Disability Status Scale (EDSS) scores. Inclusion criteria required confirmed MS and frequent cannabis use (≥4 days per week) after disease onset.
The parent study employed a prospective, longitudinal, interventional design in which the users with MS were followed over a 28-day period and allocated to either a cannabis continuation or withdrawal condition using an odd-even allocation procedure. Participants underwent repeated cognitive and MRI assessments at baseline and day 28.7
The present study is a secondary analysis of the withdrawal cohort (n = 19), comprising individuals who successfully completed 28 days of monitored abstinence. Only individuals with objective global cognitive impairment were included, resulting in a clinically selected sample. All participants were able to complete study procedures independently, suggesting sufficient functional capacity for reliable self-report despite the presence of cognitive impairment. This approach was used to enhance sensitivity in detecting meaningful change, as the inclusion of cognitively preserved individuals may introduce ceiling effects.
Cannabis Use and Abstinence Verification
Cannabis-related data included frequency and duration of use, and average monthly consumption in grams. Cannabis use in the parent study was naturalistic and self-directed rather than prescribed or experimentally controlled.7 Cannabis was predominantly inhaled (smoked), consistent with both the parent study and self-reported use patterns in this sample, although formulation and potency were not systematically recorded. Urine analyses indicated predominantly tetrahydrocannabinol (THC) exposure. Creatinine-adjusted urinary cannabidiol (CBD) concentrations were also available but were not included in primary analyses due to their low levels and limited variability in this sample. The source of cannabis (medical vs recreational) was not formally controlled, reflecting real-world use patterns.
Urine samples were collected at baseline and weekly to verify abstinence using creatinine-adjusted THC-COOH ratios. Abstinence was confirmed at the individual participant level, with concentrations declining toward near-zero levels in all participants who completed the protocol. One participant with toxicological evidence of continued cannabis use was excluded. Creatinine normalization accounted for variability in urine concentration.⁸
Withdrawal, Cognitive, and Mood Measures
Baseline Cannabis Withdrawal Questionnaire (CWQ)9 scores were obtained before abstinence and therefore do not reflect cannabis withdrawal, but rather preabstinence symptom reporting. Cannabis withdrawal was assessed using the CWQat baseline and on days 7, 14, 21, and 28, with higher scores indicating greater withdrawal distress. Mood was evaluated using the Hospital Anxiety and Depression Scale (HADS) 10 at baseline and day 28, with the Wechsler Test of Adult Reading (WTAR) 11 administered only at baseline to estimate premorbid intelligence.
Ethics
All participants gave written informed consent. The protocol was approved by the Sunnybrook Health Sciences Centre Research Ethics Board.
Statistical Analysis
Data were screened for outliers and normality. Assumptions for repeated-measures analysis of covariance (ANCOVA) were verified. Group differences in demographic, item-level CWQ, and clinical variables between men and women were assessed with independent-samples ttests.
The primary analysis was a repeated-measures ANCOVA investigating a putative time × sex interaction for CWQ total scores across the 28-day abstinence period, controlling for between-group demographic, clinical, and cannabis-use differences. Baseline CWQ was not included as a covariate because it reflected preabstinence MS-related symptoms rather than withdrawal and could obscure true between-participant differences. The Mauchly test indicated that the assumption of sphericity was violated; therefore, multivariate test statistics (ie, Pillai trace) were used for within-participant effects.
Secondary analyses included 1-way ANCOVAs comparing CWQ total scores between sexes at each time point. All analyses were conducted in IBM SPSS Statistics for Windows, version 29.0.
This study represents a secondary analysis and was not prospectively powered to detect sex differences; therefore, findings should be interpreted as exploratory.
Results
Demographic and Neurological Data
Compared with women, men (n = 8) were older, had higher EDSS scores, and reported greater monthly cannabis use in grams (all P ≤ .050). Men also demonstrated higher WTAR scores (P = .003). Creatinine-adjusted THC-COOH ratios declined steadily across days 7, 14, 21, and 28 for both sexes, confirming adherence to abstinence. HADS anxiety and depression scores remained within normal limits at baseline and day 28 and did not differ between men and women at either time point (all P ≥ .73), indicating overall mood stability (Table 1).
CWQ Symptom Scores by Sex
At baseline (preabstinence), women reported higher CWQ symptom scores than men on several individual items, including difficulty falling asleep, early awakening, physical tension, nausea, restlessness, headaches, and imagining being stoned (all P < .05). Total CWQ scores were higher among women (52.36 ± 22.13) compared with men (24.00 ± 29.55; P = .033). By day 28, sex differences persisted and, in several domains, widened. Women reported significantly greater trouble sleeping, physical tension, irritability, nausea, stomachaches, headaches, and angry outbursts (all P < .05). Total CWQ scores remained statistically higher for women (79.36 ± 32.24) compared with men (40.00 ± 21.09; P = .006) (Table 2).
Repeated-Measures ANCOVA
A repeated-measures ANCOVA examined change in CWQ total scores over the 28-day abstinence period as a function of sex, adjusting for between-group differences that included EDSS, age, monthly cannabis use (grams), WTAR, and THC-COOH/creatinine ratio. A significant main effect of time emerged (F(4,7) = 4.79; P = .035; partial η² = .732). The time × sex interaction was also significant (F(4,7) = 6.34; P = .018; partial η² = .784), indicating that withdrawal trajectories differed significantly between men and women across the abstinence period. A significant time × monthly cannabis use (grams) interaction was observed (F(4,7) = 11.31; P = .004; partial η² = .866), suggesting that heavier cannabis use predicted greater variability and change in withdrawal severity over time. A significant time × THC-COOH/creatinine ratio interaction was also observed (F(4,7) = 5.31; P = .028; partial η² = .752)(Table 3).
Between-participant effects showed that sex did not significantly predict overall CWQ severity when averaged across time points (P = .919), nor did any covariate significantly predict overall levels of withdrawal (all P > .15). This pattern indicates that sex differences were specific to the trajectory of withdrawal rather than absolute symptom levels averaged across all days (Table 3). Although differences were observed at specific time points (eg, baseline and day 28), these do not reflect a significant overall main effect of sex across the entire abstinence period.
Post Hoc Analyses by Sex
Follow-up 1-way ANCOVAs comparing CWQ total scores at each time point revealed significant between-sex differences at both baseline (F(1,17) = 5.43; P = .033) and day 28 (F(1,17) = 8.12; P = .012), with women reporting higher CWQ symptom scores at both assessments. Baseline scores were obtained before abstinence. No significant sex differences were observed at days 7, 14, or 21 (all P > .10; Figure).
Discussion
Our results suggest that sex influenced CWQ symptom trajectories during cannabis abstinence, with women showing a more pronounced and sustained increase in symptom burden than men.
Although men reported heavier cannabis use at baseline, consistent with epidemiological patterns of higher cannabis consumption among men,12 baseline THC/creatinine ratios did not differ significantly, indicating comparable physiological exposure before cessation. Because withdrawal had not yet begun at baseline, group differences in CWQ scores at this time likely reflected sex-related differences in MS symptom burden, such as sleep disturbance, fatigue, and physical tension, rather than withdrawal-specific effects.13 These baseline differences represent a potential confound and limit causal interpretation of sex differences observed during abstinence. This limitation is particularly relevant given the cognitively impaired nature of the sample, which may influence both symptom perception and reporting. The CWQ may have limited specificity in people with MS, given the overlap with MS-related symptoms.
An alternative explanation for the observed trajectories is that cannabis cessation may have unmasked underlying MS-related symptoms rather than reflecting withdrawal alone. Differences from prior analyses of this cohort (Meza et al, 20246) may be explained by the analytic approach. In that study, CWQ was included as a covariate, whereas in the present analysis it was treated as a primary longitudinal outcome, allowing trajectory-based differences to emerge. Minor numerical discrepancies reflect covariate-adjusted vs unadjusted estimates rather than differences in the underlying sample.
During abstinence, women showed higher CWQ scores at specific time points (eg, baseline and day 28) and a more sustained increase over time, which may reflect both withdrawal-related effects and greater underlying MS symptom burden, most notably irritability, sleep disturbance, physical tension, nausea, and headaches. These results parallel previous findings in people without MS, showing a more pronounced withdrawal trajectory among women.4,14 They also align with evidence that endocannabinoid signaling and CB1 receptor regulation are sexually dimorphic.15 Fluctuations in ovarian hormones, particularly estradiol, have been shown to modulate cannabinoid receptor sensitivity and endocannabinoid signaling.15 In people with MS, these hormonal effects may plausibly interact with neuroinflammatory and neuroendocrine alterations inherent to the disease, further shaping withdrawal expression.
Mood remained stable over the withdrawal period. HADS anxiety and depression scores were within normal limits and unchanged across time, and CWQ items related to depression were rarely endorsed. This pattern suggests that the distress items on the CWQ reported by women during abstinence were mild rather than indicative of clinical mood disturbance.
A significant main effect of time indicated that withdrawal severity changed across the 28-day abstinence period. However, the pattern of change was not uniform, as demonstrated by significant interactions between time, sex, and monthly cannabis use. Women exhibited a sustained increase across the 28-day period, whereas men demonstrated a more transient pattern. This contrasts with findings in people without MS, where withdrawal typically peaks early and resolves within 2 to 3 weeks.4 Ongoing neurological and somatic symptoms such as fatigue, sleep disturbance, and physical discomfort may obscure or prolong the subjective withdrawal experience, particularly for women with MS, who demonstrated a more pronounced pattern of withdrawal-related distress across abstinence. Heavier cannabis use was associated with greater variability in symptom change over time. Given the documented use of cannabis among people with MS, this variability may further complicate the interpretation of underlying neurological and somatic symptom burden.16
Laboratory findings offer a mechanistic framework for these sex differences. Cooper and Haney17 reported that women demonstrate heightened subjective and physiological responses to cannabinoids when matched with men for dose and frequency, an effect attributed to hormonal modulation of the endocannabinoid system rather than metabolic factors. This enhanced responsivity may underlie the more pronounced withdrawal symptoms observed among women with MS in the present study.
Limitations
Limitations include a modest sample size, potential selection bias given the predominance of relapsing-remitting MS, the absence of a healthy comparison group, and a follow-up period restricted to 28 days, which may not capture the full course of withdrawal in people with MS. In addition, detailed information regarding cannabis formulation (eg, edibles, tinctures, or potency) was not systematically collected, limiting precise characterization of pharmacological exposure.
Participants were restricted to individuals with objectively defined global cognitive impairment, resulting in a clinically selected sample. This may limit the generalizability of findings to cognitively preserved individuals with MS or to broader cannabis-using populations. Given that all participants had cognitive impairment, deficits in memory and attention may have introduced variability in self-reported withdrawal symptoms, potentially affecting reporting accuracy.
The CWQ may have limited specificity as a measure of cannabis withdrawal in people with MS, given the overlap with MS-related symptoms. This limitation may be further compounded by underlying cognitive impairment, and findings should therefore be interpreted as reflecting overlapping contributions of withdrawal and MS-related symptom expression.
Conclusions
Women with MS showed a more pronounced and sustained increase in CWQ symptom burden across the abstinence period. These findings likely reflect sex-specific differences in the interaction between cannabis cessation and underlying MS-related symptoms, rather than purely greater withdrawal severity. Withdrawal did not decline uniformly, with differences most evident early and late in abstinence. Depression and anxiety remained stable, suggesting cessation is psychologically tolerable despite a more sustained symptom burden in women. Tailored monitoring and sex-specific support may improve cessation success. Further research with larger samples and longer follow-up is needed to clarify mechanisms.
