Graphical abstract
Keywords: Epilepsy, Seizure, Cannabis, Length of stay, Event capturability, Epilepsy risk factors
Highlights
-
•
Cannabis users had a decreased EMU LOS by average of 0.9 days compared to non-users.
-
•
Cannabis users had a 18.1 % increase in event capturability while admitted to the EMU.
-
•
Cannabis users had an increased correlation of specific epilepsy risk factors.
-
•
Cannabis use correlates with lower LOS and higher event capturability in the EMU.
-
•
When planning EMU admission a patient’s social history should be considered.
Abstract
Purpose
Epilepsy affects approximately 50 million people worldwide, many of whom undergo evaluation in Epilepsy Monitoring Units (EMUs) for seizure classification, medication management, or presurgical assessment. With increasing numbers of states legalizing marijuana use, a growing number of EMU admissions involve individuals who use cannabis-based products. If cannabis users are unable to use cannabis products during their EMU admission, this may lead patients to be discharged before recording typical events. This study evaluates the impact of cannabis use on EMU outcomes, specifically length of stay (LOS) and likelihood of capturing a clinical event.
Methods
A retrospective chart review was conducted using a REDCap database including demographics, epilepsy risk factors, anti-seizure medication use, social history, urine drug screen results, and final diagnoses. Patients were categorized as users (self-reported use, positive urine test, or both) or non-users (negative test and no self-report). Primary outcomes in the analysis were LOS and event capture rates.
Results
Cannabis use was associated with a 0.9 day LOS reduction compared to non-users. Among patients admitted for spell classification, event capturability was 18.1 % higher in cannabis users. Additionally, cannabis users had increased prevalence of psychosocial comorbidities: 12.6 % reported physical abuse, 11.1 % sexual abuse, and 10.2 % mental abuse. Psychiatric diagnoses were also more prevalent; users were 18.9 % more likely to have Major Depressive Disorder and 22.1 % more likely to have Generalized Anxiety Disorder.
Conclusions
Cannabis use significantly affects EMU evaluation outcomes and correlates with distinct psychosocial and psychiatric profiles. These findings support a holistic approach in epilepsy patient management.
1. Introduction
According to the World Health Organization (WHO), approximately 50 million people worldwide reported having active epilepsy in 2024. Of these 50 million the WHO estimates that up to 70 % could live seizure free if properly diagnosed and treated [1]. This results in numerous patients seeking out neurologists and epilepsy specialists with 67 % of the United States’ 2.9 million adults diagnosed with epilepsy consulting a neurologist or epilepsy specialist between 2021 and 2022 [2]. This contributes to over 90,000 admissions to Epilepsy Monitoring Units (EMU) across the nation every year. Current literature identifies five reasons for EMU admission: spell classification, presurgical evaluation, medication management, seizure quantification, and intracranial monitoring [3,4]. Admission to the EMU allows physicians to gain a deeper understanding of a patient's specific disease, including seizure types and medication responses, ultimately improving patient care. Several factors influence the duration of a patient’s stay in the EMU, which is essential for maximizing the benefits of admission. The average length of stay (LOS) in the EMU is reported to be 4.8 days. However, specific patient subgroups experience longer LOS, including presurgical patients (average LOS: 7.1 days), patients with symptomatic generalized epilepsy (average LOS: 6.9 days), those with a history of invasive epilepsy management (average LOS: 6.2 days vs. 4.7 days for non-invasive management), and patients taking three or more antiseizure medications (ASMs) (average LOS: 6.3 days vs. 4.3 days for those on fewer than three ASMs) [3]. These findings highlight the need for a comprehensive approach to understanding the factors influencing LOS, including non-medical factors such as substance use.
Globally, it is estimated that there is an increase in the use of marijuana products. Latest reports suggest that marijuana is the most common illicit substance utilized, with its use increasing from about 180 million people in 2011 to 219 million in 2021 (approximately 22 % increase) [5]. As of 2024, marijuana use has been legalized in nearly all U.S. states and territories for medical purposes. Additionally, 24 states (including Arizona in which this study was conducted), the District of Columbia, Guam, and the Northern Mariana Islands, have legalized marijuana for recreational use in addition to medical use [6]. Although marijuana regulations vary by state, the widespread legalization has led to an increase in patients reporting marijuana use or testing positive in drug screenings. In 2021, 19 % (52.5 million) of Americans reported using cannabis at least once, with 30 % (15.75 million) of these users meeting the criteria for cannabis use disorder (CUD) [7]. CUDs are more prevalent among epilepsy patients, with prevalence rates having doubled since 2006, with literature from 2014 suggesting that 4.41 % of epilepsy patients have a CUD [8].
A 2019 study conducted in Oregon surveyed epilepsy patients about their marijuana use, with the majority (87.5 %) reporting that they used marijuana to manage their epilepsy. The most common source of marijuana was local medical dispensaries, licensed for both medical and recreational sales [9]. Although some studies suggest potential efficacy of cannabis-based products for epilepsy treatment [10], cannabis use and CUDs may act as barriers to effective management. Importantly, cannabis-based epilepsy medications and “street cannabis” differ significantly. The Food and Drug Administration (FDA) has approved cannabidiol (CBD) derived from Cannabis sativa for use in medical treatments, requiring specific cultivation and extraction processes that ensure high purity and dosage control [11]. In contrast, “street cannabis” is unregulated, leading to potential inconsistencies in purity and dosage. A 2019 survey found that only 2 out of 39 epilepsy patients could provide an exact dosage of cannabis used in milligrams [9]. The lack of standardization in street cannabis use could hinder treatment efficacy and potentially worsen epileptic conditions due to impurities and unknown drug-drug interactions.
Cannabis use is not only a barrier to effective treatment due to uncontrolled purity and dosage, but it is also associated with a range of risk factors and comorbidities in epilepsy patients. CUDs are linked to depression, bipolar disorder, tobacco use disorder and increased post-traumatic stress disorder (PTSD) severity in the general population [[12], [13], [14], [15]]. Conversely, epilepsy patients with these mental health conditions are more likely to have a CUD [8]. Despite the rising prevalence of CUDs, cannabis use has been shown to exacerbate depression symptoms [16], potentially creating a vicious cycle of worsening mental health and epilepsy. As the prevalence of marijuana/cannabinoid product use increases with its legalization, patients may develop additional epilepsy-related risk factors and comorbidities, which could affect their EMU admission and treatment options.
While much research has focused on the factors influencing EMU LOS, as well as the impact of marijuana and cannabis products on epilepsy and its comorbidities, no studies to date have compared the effects of cannabinoid positivity on EMU admissions. It is postulated that patients who use cannabis, and especially those with cannabis use disorder, may not tolerate EMU admissions given that they are not allowed to use cannabis products as they otherwise would during the hospital stay. This may lead to shorter admissions and inability to capture events, decreasing the diagnostic yield. Given the widespread legalization and increased cannabis use, this paper aims to explore how cannabinoid positivity impacts EMU LOS and subsequently the ability to capture events during the admission.
2. Methods
2.1. Study design
In order to evaluate the association between cannabinoid positivity (along with associated risk factors and comorbidities) and patients' LOS in the EMU, a retrospective clinical chart review was conducted. A query of the electronic medical record (EMR) system was completed with a REDCap database created for this study. This database was designed to collect comprehensive information on pre-admission, during-admission, and post-admission variables.
The database included 6 key fields: demographics, epilepsy risk factors, anti-seizure medication use, social history, urine drug screen results, and final diagnoses. For the urine drug screen, all patients are tested on admission per protocol. If a cannabinoid was detected in the urine, the result was reported in the medical record as cannabinoid positive. Data was extracted directly from the clinical charts of patients admitted to the EMU between January and December 2023. All admissions during this period were reviewed, with exclusions made for patients prescribed cannabidiol (Epidiolex) or those who had a subsequent EMU admission within 2023 (only the first admission was included). Additionally, it should be noted that use of marijuana and non-FDA approved cannabis containing products is not allowed during hospital admission per hospital policy.
2.2. Data classification & comparison
Patients were classified into two groups: those currently using cannabis and those not using cannabis. A patient was classified as a “current cannabis user” if they either self-reported current marijuana or cannabinoid product use (regardless of urine test result), had a positive urine test for cannabinoids without self-reporting use, or both. This approach was taken in order to be inclusive of any patient who may be using cannabis products. Of note, none of the patients were advised by their treating provider to utilize cannabis obtained from a dispensary for medical purposes. Conversely, patients were classified as “not currently using cannabis” if they self-reported no marijuana or cannabinoid product use and had a negative urine test.
LOS was analyzed for each subgroup to determine statistical difference between groups (t-test). The diagnostic yield of an EMU admission was evaluated based on the ability to capture a neurological event during the stay (i.e. event capturability). Event capturability was compared for each subgroup, and the association between changes in LOS and event capturability was analyzed using t-tests.
3. Results
A total of 191 patient clinical charts were analyzed in this retrospective review. Of these, 130 patients (68.06 %) were not currently using cannabis, while 61 patients (31.94 %) were using cannabis at the time of their admission. No patients were excluded from the analysis due to the use of Epidiolex.
Among the 191 patients, 76 (39.8 %) identified as male, 114 (59.7 %) as female, and 1 (0.5 %) identified as other. Within the group of non-users (130 patients), 37.7 % (49) were male, 61.5 % (80) were female, and 0.8 % (1) identified as other. Among those currently using cannabis (61 patients), 44.3 % (27) were male, 55.7 % (34) were female, and none identified as other.
3.1. Baseline characteristics
The mean age of all patients was 44.4 years (SD: 17.64), with a range from 18 to 87 years and a median of 44 years. Non-users had a mean age of 46.6 years (SD: 18.47) with a median of 48 years, while cannabis users had a mean age of 39.8 years (SD: 14.84) and a median of 37 years. Demographic characteristics are shown in Table 1..
Table 1.
Baseline characteristics.
| Current Cannabis User |
||||
|---|---|---|---|---|
| No (N = 130) |
Yes (N = 61) |
Total (N = 191) |
P-value | |
| Sex, n (%) | 0.5621 | |||
| Male | 49 (37.7 %) | 27 (44.3 %) | 76 (39.8 %) | |
| Female | 80 (61.5 %) | 34 (55.7 %) | 114 (59.7 %) | |
| Other | 1 (0.8 %) | 0 (0.0 %) | 1 (0.5 %) | |
| Age at Admission, years | 0.0072 | |||
| Mean (SD) | 46.6 (18.47) | 39.8 (14.84) | 44.4 (17.64) | |
| Median | 48 | 37 | 44 | |
| Range | 18.0, 87.0 | 18.0, 72.0 | 18.0, 87.0 | |
Table 1. showing the base demographics for epilepsy patients included in analysis.
3.2. Marijuana history
Upon admission, during standardized questioning, all patients were asked specifically about a history of marijuana use. Regarding marijuana use history, 59.7 % (114) of all patients reported no history of marijuana use previously in their life, while 40.3 % (77) had a history of use. Among the cannabis non-users (130 patients), 86.2 % (112) had no history of previous marijuana use, and 13.8 % (18) had a history of use, explained by prior use that did not overlap with the admission period. Among the cannabis user group (61 patients), 96.7 % (59) had a prior history of marijuana use, and 3.3 % (2) did not report any history of marijuana use, suggesting possible underreporting or false positive urine drug screening. These data are summarized in Table 2..
Table 2.
Marijuana history- all patients.
| Current Cannabis User |
||||
|---|---|---|---|---|
| No (N = 130) |
Yes (N = 61) |
Total (N = 191) |
P-value | |
| Reported Marijuana History n (%) | 0<.0011 | |||
| No | 112 (86.2 %) | 2 (3.3 %) | 114 (59.7 %) | |
| Yes | 18 (13.8 %) | 59 (96.7 %) | 77 (40.3 %) | |
Table 2. showing the distribution of self-reported previous marijuana use compared to cannabis use when currently admitted to the Epilepsy Monitoring Unit.
3.3. Length of stay – all patients
The mean length of stay (LOS) for all patients was 3.8 days (SD: 1.94), with a median of 4 days and a range from 0 to 14 days. Non-users had a mean LOS of 4.1 days (SD: 2.11), with a median of 4 days, and a range of 0 to 14 days. Cannabis users had a mean LOS of 3.2 days (SD: 1.36), with a median of 3 days and a range of 1 to 8 days. The p-value for this data was 0.007, indicating a statistically significant association between cannabis use and shorter LOS in the EMU. Of note, there was no specific protocol for the decrease of antiseizure medications during EMU admission, therefore data on the influence of changes to ASMs cannot be investigated in this study.
3.4. Event capturability- all patients
Regarding event capturability, 26.2 % (50) of all patients were unable to have an event captured, while 73.8 % (141) had an event captured during their admission. In the non-user group, 30.0 % (39) were unable to have an event captured, and 70.0 % (91) had an event captured. In the cannabis user group, 18.0 % (11) were unable to have an event captured, while 82.0 % (50) had an event captured. The p-value for this data was 0.0792, indicating no statistically significant difference, and no conclusion can be drawn regarding cannabis use and event capturability when looking at all admissions to the EMU for any reason. LOS and event capturability data for all patients is shown in Table 3..
Table 3.
Total length of stay & event capturability- all patients.
| Current Cannabis User |
||||
|---|---|---|---|---|
| No (N = 130) |
Yes (N = 61) |
Total (N = 191) |
P-value | |
| Total Length of Stay, days | 0.0011 | |||
| Mean (SD) | 4.1 (2.11) | 3.2 (1.36) | 3.8 (1.94) | |
| Median | 4 | 3 | 4 | |
| Range | 0.0, 14.0 | 1.0, 8.0 | 0.0, 14.0 | |
| Event Captured, n (%) | 0.0792 | |||
| No | 39 (30.0 %) | 11 (18.0 %) | 50 (26.2 %) | |
| Yes | 91 (70.0 %) | 50 (82.0 %) | 141 (73.8 %) | |
Table 3. showing the total length of stay and events captured during admission to the Epilepsy Monitoring Unit for all patients based on their “Current Cannabis User” classification.
3.5. Subcategory distribution
Further analysis sorted patients by admission type: spell/seizure classification, surgical candidacy, medication management, and intracranial monitoring. Of the 191 patients, 79.1 % (151) were admitted for spell/seizure classification, 18.3 % (35) for surgical candidacy, 12.0 % (23) for medication management, and 11.5 % (22) for intracranial monitoring. Note that some patients were admitted for multiple reasons, and a single patient may be classified into multiple groups. Cannabis use by admission type is detailed in Table 4..
Table 4.
Admission type.
| Current Cannabis User |
||||
|---|---|---|---|---|
| No (N = 130) |
Yes (N = 61) |
Total (N = 191) |
P-value | |
| Spell/Seizure Classification, n (%) | 0.0691 | |||
| No | 32 (24.6 %) | 8 (13.1 %) | 40 (20.9 %) | |
| Yes | 98 (75.4 %) | 53 (86.9 %) | 151 (79.1 %) | |
| Surgical Candidacy, n (%) | 0.9431 | |||
| No | 106 (81.5 %) | 50 (82.0 %) | 156 (81.7 %) | |
| Yes | 24 (18.5 %) | 11 (18.0 %) | 35 (18.3 %) | |
| Medication Management, n (%) | 0.5211 | |||
| No | 113 (86.9 %) | 55 (90.2 %) | 168 (88.0 %) | |
| Yes | 17 (13.1 %) | 6 (9.8 %) | 23 (12.0 %) | |
| Intracranial Monitoring, n (%) | 0.3251 | |||
| No | 113 (86.9 %) | 56 (91.8 %) | 169 (88.5 %) | |
| Yes | 17 (13.1 %) | 5 (8.2 %) | 22 (11.5 %) | |
Table 4. showing how patients were further subcategorized into groups based on their admission type and the distribution of patients classified as currently using cannabis in each.
3.6. Spell/seizure classification patients
In the spell/seizure classification group (151 patients), cannabis users had a mean LOS of 3.2 days (SD: 1.40), with a median of 3 days and a range of 1 to 8 days, while non-users had a mean LOS of 3.6 days, with a median of 4 days and a range of 0 to 11 days. The p-value for this data was 0.0091, indicating that cannabinoid positivity is associated with a shorter LOS in patients admitted for spell/seizure classification.
Additionally, in the spell/seizure classification subgroup, event capture rate showed that 83.0 % (44) of cannabis users had an event captured, compared to 65.3 % (64) of non-users. The p-value for this comparison was 0.0212, indicating that cannabis use is associated with higher event capturability in this patient group. This remained consistent with the overall analysis as noted above in Section 3.3. LOS and the event capturability data for spell/seizure classification patients is shown in Table 5..
Table 5.
Total length of stay & event capturability- spell/seizure classification patients.
| Current Cannabis User |
||||
|---|---|---|---|---|
| No (N = 98) |
Yes (N = 53) |
Total (N = 151) |
P-value | |
| Total Length of Stay, days | 0.0091 | |||
| Mean (SD) | 3.6 (1.47) | 3.2 (1.40) | 3.5 (1.46) | |
| Median | 4 | 3 | 3 | |
| Range | 0.0, 11.0 | 1.0, 8.0 | 0.0, 11.0 | |
| Event Captured, n (%) | 0.0212 | |||
| No | 34 (34.7 %) | 9 (17.0 %) | 43 (28.5 %) | |
| Yes | 64 (65.3 %) | 44 (83.0 %) | 108 (71.5 %) | |
Table 5. showing the total length of stay and events captured during admission to the Epilepsy Monitoring Unit for spell/seizure classification patients based on their “Cannabis User” classification.
3.7. Risk factors & comorbidities
Finally, an analysis of risk factors and comorbidities revealed that 83.2 % (159) of patients had at least 1 risk factor (for either epilepsy or functional/dissociative seizures) or comorbidity, while 16.8 % (32) had none. Among non-users, 80.0 % (104) had at least 1 risk factor/comorbidity, and 20.0 % (26) had none. Among cannabis users, 90.2 % (55) had at least 1 risk factor/comorbidity, and 9.8 % (6) had at none (p = 0.0791), suggesting no significant association between cannabis use and the presence of any risk factor or comorbidity. Of note, these risk factors were inquired for all patients in the study during a structured interview by a neuropsychologist on admission. On further subgroup analysis, specific risk factors of physical abuse, sexual abuse, psychological abuse, Major Depressive Disorder, and Generalized Anxiety Disorder demonstrated an association with cannabis use in patients admitted to the EMU. Other risk factors listed did not demonstrate a statistically significant association. Data is summarized in Table 6.. Unfortunately, numbers in this study were not large enough to investigate the association between risk factors and comorbidities, LOS, or event capturability.
Table 6.
Risk factors & comorbidities all patients.
| Current Cannabis User | ||||
|---|---|---|---|---|
| No (N = 130) |
Yes (N = 61) |
Total (N = 191) |
P-value | |
| Risk Factor/Comorbidity, n (%) | 0.0791 | |||
| At least 1 risk factor/comorbidity present | 104 (80.0 %) | 55 (90.2 %) | 159 (83.2 %) | |
| No risk factors/comorbidities present | 26 (20.0 %) | 6 (9.8 %) | 32 (16.8 %) | |
| Head Trauma, n (%) | 0.6271 | |||
| No | 96 (73.8 %) | 43 (70.5 %) | 139 (72.8 %) | |
| Yes | 34 (26.2 %) | 18 (29.5 %) | 52 (27.2 %) | |
| Physical Abuse, n (%) | 0.0031 | |||
| No | 125 (96.2 %) | 51 (83.6 %) | 176 (92.1 %) | |
| Yes | 5 (3.8 %) | 10 (16.4 %) | 15 (7.9 %) | |
| Sexual Abuse, n (%) | 0.0191 | |||
| No | 121 (93.1 %) | 50 (82.0 %) | 171 (89.5 %) | |
| Yes | 9 (6.9 %) | 11 (18.0 %) | 20 (10.5 %) | |
| Psychological Abuse, n (%) | 0.0151 | |||
| No | 124 (95.4 %) | 52 (85.2 %) | 176 (92.1 %) | |
| Yes | 6 (4.6 %) | 9 (14.8 %) | 15 (7.9 %) | |
| Major Depressive Disorder, n (%) | 0.0141 | |||
| No | 80 (61.5 %) | 26 (42.6 %) | 106 (55.5 %) | |
| Yes | 50 (38.5 %) | 35 (57.4 %) | 85 (44.5 %) | |
| Generalized Anxiety Disorder, n (%) | 0.0041 | |||
| No | 65 (50.0 %) | 17 (27.9 %) | 82 (42.9 %) | |
| Yes | 65 (50.0 %) | 44 (72.1 %) | 109 (57.1 %) | |
| PTSD, n (%) | 0.6481 | |||
| No | 119 (91.5 %) | 57 (93.4 %) | 176 (92.1 %) | |
| Yes | 11 (8.5 %) | 4 (6.6 %) | 15 (7.9 %) | |
| Developmental Delay, n (%) | 0.1261 | |||
| No | 121 (93.1 %) | 60 (98.4 %) | 181 (94.8 %) | |
| Yes | 9 (6.9 %) | 1 (1.6 %) | 10 (5.2 %) | |
| Bipolar Disorder, n (%) | 0.1431 | |||
| No | 130 (100.0 %) | 60 (98.4 %) | 190 (99.5 %) | |
| Yes | 0 (0.0 %) | 1 (1.6 %) | 1 (0.5 %) | |
| History of Meningitis or Encephalitis, n (%) | 0.5821 | |||
| No | 129 (99.2 %) | 60 (98.4 %) | 189 (99.0 %) | |
| Yes | 1 (0.8 %) | 1 (1.6 %) | 2 (1.0 %) | |
| History of Febrile Seizure, n (%) | 0.1661 | |||
| No | 124 (95.4 %) | 55 (90.2 %) | 179 (93.7 %) | |
| Yes | 6 (4.6 %) | 6 (9.8 %) | 12 (6.3 %) | |
| History of Stroke, n (%) | 0.2281 | |||
| No | 123 (94.6 %) | 60 (98.4 %) | 183 (95.8 %) | |
| Yes | 7 (5.4 %) | 1 (1.6 %) | 8 (4.2 %) | |
| History of Birth Trauma, n (%) | 0.3301 | |||
| No | 128 (98.5 %) | 61 (100.0 %) | 189 (99.0 %) | |
| Yes | 2 (1.5 %) | 0 (0.0 %) | 2 (1.0 %) | |
| OCD, n (%) | N/A | |||
| No | 130 (100.0 %) | 61 (100.0 %) | 191 (100.0 %) | |
Table 6. shows the distribution of epilepsy-related risk factors and comorbidities, and cannabis use in Epilepsy Monitoring Unit patients.
4. Discussion
4.1. Marijuana prevalence
Given increasing legalization and accessibility, cannabis use among patients being evaluated for epilepsy is likely to rise. This anticipated increase is due to current literature suggesting that marijuana consumption can decrease epilepsy symptoms. As marijuana becomes more widely available, many patients pursue self-directed cannabis use outside medical supervision. However, most studies investigating symptom relief have focused on controlled, purified doses that are regulated and prescribed by physicians. In contrast, marijuana obtained from local dispensaries often has varying compositions, and current literature indicates that many patients are unaware of both the purity and dosage of their marijuana [9]. Therefore, while marijuana may offer some symptom relief, it does not completely resolve symptoms in patients with epilepsy or related conditions, and an increasing number of patients are being admitted to the EMU with the exact doses and specific contents of their marijuana/cannabis products unclear.
The findings of this study reveal that patients being evaluated for epilepsy are unequivocally influenced by cannabis use. In our sample of 191 patients, 40.3 % were classified as cannabis users, which is 21.3 % higher than the U.S. national average [17]. This is an important insight into the prevalence of cannabis use within the epilepsy patient population.
4.2. Reduced length of stay with cannabis use
One of the novel findings of this study is the association between cannabis use and a shorter LOS in the EMU, with a reduction of about one day. One possibility is that more patients may be suffering from cannabis use disorder than previously identified. The patient's desire, either physiologic or psychologic, to use cannabis may influence their desire to continue with their evaluation in the EMU. This finding aligns with current literature suggesting that a holistic approach is necessary to predict LOS. Our data support the notion that not only a patient's epilepsy history and comorbidities, but also their history with social or addictive substances, may play a significant role in determining LOS. An alternative hypothesis may be that events occur earlier during admissions in the setting of cannabis use. It is standard practice that patients are discharged once events are recorded and a diagnosis is made. As noted below, events are more likely to be captured in patients who are currently using cannabis, and it is likely that they occur earlier in the admission as seen by a shorter mean LOS. However, due to the relatively small sample size, additional research is needed to confirm the validity of this association and explore underlying mechanisms.
4.3. Increased event capturability with cannabis use
A second key finding was the association between cannabis use and increased event capture rates in patients admitted for spell/seizure classification. This finding is counterintuitive, as one would typically expect that patients with shorter hospital stays would have fewer opportunities to capture an event, being connected to the EEG for a shorter time. However, we observed a statistically significant increase (83 % vs 65.3 %) in event capture rate for cannabis users, indicating a potential increase in epilepsy symptoms during hospitalization. One possible explanation is that this phenomenon may be attributed to withdrawal effects. Withdrawal may cause patients to feel anxious, restless, and irritable, in addition to physical symptoms such as gastrointestinal upset and tremors. Unlike other addictive substances like tobacco, which have alternative forms for use during hospitalization, cannabinoid products lack a comparable substitute. Thus, patients who regularly use cannabis may experience these withdrawal symptoms, which could lead to a higher event capture rate – whether epileptic or non-epileptic events. Another possible explanation is the lack of cannabis may uncover underlying anxiety that was being mitigated by cannabinoids. The increased mood symptoms may help trigger events as well. A final explanation may be that the cannabis used by patients is simply producing an antiseizure effect, and its lack of use during admission leads to seizures, just as with any other antiseizure medication. Although these hypotheses are plausible, further research is needed to substantiate the role of cannabis withdrawal in increasing event capturability.
Interestingly, although not statistically significant, a trend toward increased event capture rates in cannabis users were also observed across all reasons for admission, not just those admitted for spell/seizure classification.
4.4. Increased risk factors & comorbidities with cannabis use
Another significant finding was the association between cannabis use and an increased likelihood of certain epilepsy and functional/dissociative seizure risk factors and comorbidities, including abuse (physical, sexual, and psychological), Major Depressive Disorder, and Generalized Anxiety Disorder. These findings align with existing literature that suggests cannabis use is associated with higher rates of these conditions. However, it is important to note that this association does not imply causality. While cannabis use may act as a marker for patients with these comorbidities or risk factors, we cannot conclude that cannabis use directly results in the development of these conditions. It is equally possible that these conditions may contribute to the initiation or continuation of cannabis use. Further study is needed to understand these associations.
5. Conclusion
Based on these results, there is a statistically significant association between cannabis use, LOS, event capturability, and epilepsy risk factors. Specifically, there is a negative association between cannabis use and LOS in patients admitted to the EMU. On average, patients currently using cannabis have a LOS that is 0.9 days shorter compared to patients not using cannabis products. Consequently, admitted patients with cannabinoids in their urine drug screen and/or those who verbally report using non-prescribed cannabis products may have shorter admissions.
Additionally, a positive association exists between cannabis use and event capturability among patients admitted for spell classification. On average, spell classification patients who are currently using cannabis are 18.1 % more likely to have an event captured compared to those not using cannabis. Thus, despite their shorter LOS, the diagnostic yield of the admission for spell classification patients using cannabis remained diagnostically useful, as the event capture rate was not reduced.
Furthermore, these results are in line with current literature regarding epilepsy and functional/dissociative seizure risk factors. Patients admitted to the EMU using cannabis were more likely to have a history of physical, sexual, and psychological abuse, as well as Major Depressive Disorder and Generalized Anxiety Disorder.
Overall, it can be concluded that cannabis use significantly impacts a patient's admission to the EMU. Therefore, a holistic approach must be taken when considering whether an EMU admission would be beneficial for a patient. Despite the shorter LOS observed among patients using cannabis, the increased event capturability indicates that the admission remains effective. At this time, no modifications to EMU admission protocols for patients using cannabinoid products can be recommended. Furthermore, no recommendations are being made to suggest that patient should utilize cannabis products prior to admission, as it is unknown if these products may cause other adverse effects.
5.1. Future research
Additional research is warranted to further elucidate how marijuana and cannabis products affect patients being evaluated for epilepsy. Primarily, future studies should investigate whether cannabis withdrawal contributes to the observed shorter LOS and increased event capturability in spell classification patients. Such research could yield valuable information to improve the effectiveness of patient admissions. Furthermore, additional studies should analyze the impact of cannabinoids on specific seizure types (focal onset, generalized onset) and clarify the causal relationship between cannabis use with epilepsy and functional/dissociative seizure risk factors—determining whether cannabis use leads to an increase in these risk factors or whether the presence of such risk factors predisposes patients to cannabis use. One of the limitations of this study is the retrospective nature of the review. Future research could include prospectively looking at this patient population and attempting to gain some quantification to the degree of cannabis product use to determine a dose effect on LOS and event capture rate. Additionally, presuming that the sudden removal of cannabis from a patient’s system will negatively impact a patient’s mood all by itself, the investigation into methods to mitigate any mood issues on the day of admission for cannabis users could potentially prolong LOS for higher diagnostic yield. Finally, further investigation into the neuropsychological impact of cannabis, including its effects on cognitive function, mood disorders, and overall brain health, is needed to fully understand its influence on the management of patients with epilepsy.
CRediT authorship contribution statement
Oliver Hoerth: Writing – review & editing, Writing – original draft, Visualization, Validation, Software, Project administration, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. Ejerzain Aniles-Renova: Writing – review & editing, Supervision, Project administration, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. Nan Zhang: Writing – review & editing, Validation, Formal analysis. Emily Thompson: Writing – review & editing, Validation, Formal analysis. Kristin A. Kirlin: Writing – review & editing, Supervision, Methodology, Formal analysis. Joseph Drazkowski: Writing – review & editing, Validation, Supervision, Resources, Project administration, Methodology, Formal analysis, Conceptualization.
Funding
This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.
Declaration of competing interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Contributor Information
Oliver Hoerth, Email: ohoerth8@gmail.com.
Joseph Drazkowski, Email: drazkowski.joseph@mayo.edu.
References
- 1.World Health Organization (WHO), Epilepsy. https://www.who.int/news-room/fact-sheets/detail/epilepsy , 2024 (accessed 11 August 2025).
- 2.Centers for Disease Control and Prevention. Epilepsy Facts and Stats / Epilepsy. https://www.cdc.gov/epilepsy/data-research/facts-stats/index.html, 2024 (accessed 23 February 2025).
- 3.Bagić A., Ahrens S., Chapman K., Bai S., Clarke D., Eisner M., et al. Epilepsy monitoring unit practices and safety among NAEC epilepsy centers: A census survey. Epilepsy Behav. 2023;150 doi: 10.1016/j.yebeh.2023.109571. [DOI] [PubMed] [Google Scholar]
- 4.Gazzola D.M., Thawani S., Agbe-Davies O., Carlson C. Epilepsy monitoring unit length of stay. Epilepsy Behav. 2016;58:102–105. doi: 10.1016/j.yebeh.2016.02.031. [DOI] [PubMed] [Google Scholar]
- 5.United Nations Office on Drugs and Crime, Special Points of Interest World Drug Report. https://www.unodc.org/res/WDR-2023/Special_Points_WDR2023_web_DP.pdf , 2023 (accessed 11 August 2025).
- 6.Centers for Disease and Control and Prevention, State Medical Cannabis Laws, Cannabis and Public Health. https://www.cdc.gov/cannabis/about/state-medical-cannabis-laws.html , 2024 (accessed 23 February 2025).
- 7.Centers for Disease Control and Prevention, Cannabis Facts and Stats, Cannabis and Public Health. https://www.cdc.gov/cannabis/data-research/facts-stats/index.html , 2024 (accessed 23 February 2025).
- 8.Lekoubou A., Fox J., Bishu K.G., Ovbiagele B. Trends in documented cannabis use disorder among hospitalized adult epilepsy patients in the United States. Epilepsy Res. 2020;163 doi: 10.1016/j.eplepsyres.2020.106341. [DOI] [PubMed] [Google Scholar]
- 9.Kerr A., Walston V., Wong V.S.S., Kellogg M., Ernst L. Marijuana use among patients with epilepsy at a tertiary care center. Epilepsy Behav. 2019;97:144–148. doi: 10.1016/j.yebeh.2019.05.037. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.O. Devinsky, N. A. Jones, M. O. Cunningham, A. P. Jayasekera, B. J. Whalley, Cannabinoid treatments in epilepsy and seizure disorders, J. American Physiological Society 104 (2024) Issue 2 10.1152/physrev.00049.2021. [DOI] [PubMed]
- 11.Food and Drug Administration (FDA), FDA and Cannabis: Research and Drug Approval Process. https://www.fda.gov/news-events/public-health-focus/fda-and-cannabis-research-and-drug-approval-process , 2023 (accessed 11 August 2025).
- 12.E. L. Dent, P. G. Johnstad, M. Sorkhou, Cannabis use and mood disorders: a systematic review, J. Frontiers Public Health 12, (2024) Sec. Substance Use Disorders and Behavioral Addictions. https://www.frontiersin.org/journals/public-health/articles/10.3389/fpubh.2024.1346207/full. [DOI] [PMC free article] [PubMed]
- 13.Maggu G., Choudhary S., Jaishy R., Chaudhury S., Saldanha D., Borasi M. Cannabis use and its relationship with bipolar disorder: A systematic review and meta-analysis. Ind Psychiatry J. 2023;32:202–214. doi: 10.4103/ipj.ipj_43_23. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC10756590/ [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Rabin R.A., George T.P. A review of co-morbid tobacco and cannabis use disorders: possible mechanisms to explain high rates of co-use. Am J Addict. 2015;24(2):105–116. doi: 10.1111/ajad.12186. [DOI] [PubMed] [Google Scholar]
- 15.Metrik J., Stevens A.K., Gunn R.L., Borsari B., Jackson K.M. Cannabis use and posttraumatic stress disorder: prospective evidence from a longitudinal study of veterans. Psychol Med. 2022;52(3):446–456. doi: 10.1017/S003329172000197X. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Murkar A., Kendzerska T., Robillard R., Quilty L., Saad M. Increased cannabis intake during the COVID-19 pandemic is associated with worsening of depression symptoms in people with PTSD. BMC Psychiatry. 2022;22:554. doi: 10.1186/s12888-022-04185-7. https://bmcpsychiatry.biomedcentral.com/articles/10.1186/s12888-022-04185-7 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.National Center for Drug Abuse Statistics, Marijuana Addiction Statistics [2023]: Usage & Abuse Rates. https://drugabusestatistics.org/marijuana-addiction/ , 2023 (accessed 23 February 2025).