This study reflects experiences with off-label use of CBD during the first years of market access in a cohort of 108 patients with drug-refractory focal-onset, genetic generalised and other epilepsy syndromes, excluding LGS, DS or TSC.

The observed efficacy of CBD was well in line with that observed during RCTs and open-label extension studies for LGS [12, 13], DS [14, 15], and TSC [16], with 50% responder rates of 39% at three months and 42% at one year. The retention rates at three (85%), six (73%), and 12 months (61%) are encouraging, and align well with findings from other post-marketing studies on CBD [32–34]. In addition, the retention is similar to or above that of other ASMs reported in post-marketing studies, including brivaracetam, cenobamate, eslicarbazepine, lacosamide, lamotrigine, levetiracetam, perampanel, topiramate, valproate, and zonisamide in focal or generalised epilepsies [35–47], however, any direct comparisons should be interpreted with caution. In particular, retention data may appear favorable; but cross-study comparisons are inherently limited due to differences in study design and the lack of head-to-head trials [48].

So far, data on the use of CBD in genetic generalised epilepsies has been heterogeneous [49, 50]; however, in our study, patients with this epilepsy type exhibited a 50% responder rate of 52.6% (10/19), suggesting a role for CBD in managing some of these patients. Our results suggest that CBD may offer clinical benefits in certain cases of treatment-resistant epilepsies beyond currently approved indications [51]; however, such use should be considered exploratory, and further prospective studies are needed to establish its efficacy and safety across different epilepsy syndromes.

Mechanism-of-action data support this observation, suggesting that CBD may achieve anti-seizure effects through various pathways, including the modulation of intracellular calcium via GPR55 and TRPV1 receptors and the regulation of adenosine signalling [52]. Although the exact mechanisms remain under investigation, these targets appear to act broadly rather than being limited to specific epilepsy aetiologies or syndromes [52].

Findings on differences in retention rates of CBD between children and adults differ across studies. We observed better retention and higher median daily dosage per kg of body weight for CBD in children, with 22.7 mg per kg compared to 10.9 mg per kg body weight in adults. Higher weight-adjusted dosing in children was also reported in earlier studies on the use of CBD [32, 53]. A retrospective, single-centre study from the University of Wisconsin–Madison [54] reported no differences in retention rates, with 77.2% of paediatric patients (n = 57) and 76.5% of adult patients (n = 51) remaining on CBD over an average of 20 months. In contrast, results from the Italian Expanded Access Program showed higher retention rates in adults than children [55]. Expanded Access Program data from the US revealed no significant differences in seizure frequency reduction between children up to 2 years and adults [56].

The lower-than-recommended median starting dose of 2.2 mg per kg bodyweight per day reflects cautious titration practices in clinical routine, particularly in patients with complex therapy with multiple ASMs or concern for drug interactions. This conservative approach is commonly adopted to improve tolerability, especially in populations outside of the licensed indications. Overall, the TEAE profile observed in our study was in line with the established adverse events profile of CBD. Overall, TEAEs were reported in 37.9% of patients, with discontinuation due to TEAEs occurring in 19.5% of the cohort. The most common TEAEs were gastrointestinal symptoms and sedation. We did not observe any serious TEAEs, and, similar to other post-marketing studies, the overall proportion of patients reporting TEAEs was lower than in RCTs where TEAEs are more rigorously documented. Systematic literature reviews of clinical trials have indicated that CBD is associated with somnolence (particularly when used with CLB), as well as decreased appetite, diarrhoea, and elevated liver enzymes, especially in conjunction with valproate [57–59]. We observed elevated liver enzymes in only one patient; however, data on laboratory tests were not routinely collected. Patients with epilepsy are prone to behavioural and psychiatric TEAEs as well as sleep disorders, with some ASMs reported to worsen these symptoms [60]. In the current study, consistent with findings from other CBD studies, psychobehavioural TEAEs were infrequently observed. In addition, physicians rated 55.6% of patients as showing improvement or no change (32.4%) in behaviour. Overall, CBD appears to have a favourable psychobehavioural profile.

This study is subject to several limitations typical of retrospective chart reviews, including an uncontrolled design, potential gaps or incompleteness in the data, and variability in data collection processes. For example, the absence of detailed records on total seizure counts and TCS may have influenced the accuracy of responder rate calculations. Moreover, seizure frequency data were based on caregiver-reported seizure diaries, which depend on their reliability and accuracy, and may be subject to over- and underreporting [61]. Therapeutic drug monitoring (TDM) was not routinely performed or collected for CBD and other concomitant ASM. Implementing routine TDM for CBD and co-administered ASMs could help minimize TEAEs and support informed dosing adjustments. Recent progress was reported in TDM techniques for precise CBD measurement, that offers significant utility in assessing drug-drug interactions, and might contribute to optimized patient care in complex therapeutic scenarios [62]. The study also lacked a predefined design to statistically evaluate differences between age subgroups, epilepsy types, or the impact of concomitant CLB use. These limitations highlight the need for prospective, controlled studies to provide a more comprehensive understanding of the findings.