Long‐term outcome of alcohol withdrawal seizures

Abstract

Background and purpose

Alcohol withdrawal seizures (AWS) are a well‐known complication of chronic alcohol abuse, but there is currently little knowledge of their long‐term relapse rate and prognosis. The aims of this study were to identify risk factors for AWS recurrence and to study the overall outcome of patients after AWS.

Methods

In this retrospective single‐center study, we included patients who were admitted to the Emergency Department after an AWS between January 1, 2013 and August 10, 2021 and for whom an electroencephalogram (EEG) was requested. AWS relapses up until April 29, 2022 were researched. We compared history, treatment with benzodiazepines or antiseizure medications (ASMs), laboratory, EEG and computed tomography findings between patients with AWS relapse (r‐AWS) and patients with no AWS relapse (nr‐AWS).

Results

A total of 199 patients were enrolled (mean age 53 ± 12 years; 78.9% men). AWS relapses occurred in 11% of patients, after a median time of 470.5 days. Brain computed tomography (n = 182) showed pathological findings in 35.7%. Risk factors for relapses were history of previous AWS (p = 0.013), skull fractures (p = 0.004) at the index AWS, and possibly epileptiform EEG abnormalities (p = 0.07). Benzodiazepines or other ASMs, taken before or after the index event, did not differ between the r‐AWS and the nr‐AWS group. The mortality rate was 2.9%/year of follow‐up, which was 13 times higher compared to the general population. Risk factors for death were history of AWS (p < 0.001) and encephalopathic EEG (p = 0.043).

Conclusions

Delayed AWS relapses occur in 11% of patients and are associated with risk factors (previous AWS >24 h apart, skull fractures, and pathological EEG findings) that also increase the epilepsy risk, that is, predisposition for seizures, if not treated. Future prospective studies are mandatory to determine appropriate long‐term diagnostic and therapeutic strategies, in order to reduce the risk of relapse and mortality associated with AWS.

INTRODUCTION

Although alcohol abuse is the seventh leading cause of mortality and morbidity [1], little is known about alcohol withdrawal seizures (AWS), a well‐known complication of alcohol withdrawal. AWS are considered acute symptomatic seizures (formerly known as provoked seizures), which usually occur 6–48 h after a significant reduction or discontinuation of chronic alcohol intake, with a peak between 13 and 24 h of abstinence [2]. In extreme cases, alcohol withdrawal leads to status epilepticus [3], which is often pharmacoresistant, and persistent epilepsy. Subacute encephalopathy with seizures in alcoholics (SESA), a subtype of focal nonconvulsive status epilepticus (NCSE), was first described in 1981 by Niedermeyer et al. [4], and is characterized by encephalopathy, lateralized periodic discharges on electroencephalogram (EEG), chronic microvascular ischemia on neuroimaging studies and a high recurrence rate when chronic antiseizure medication (ASM) is stopped.

The pathophysiological mechanisms underlying AWS presumably include chronic receptor changes. Ethanol stimulates inhibitory GABA‐A and GABA‐B receptors, at the same time suppressing excitatory glutamatergic N‐methyl‐D‐aspartate (NMDA) receptor activity. Chronic alcohol consumption leads to a reduction in GABA‐A receptors and an increase in NMDA receptor activity, respectively. Hence, during withdrawal periods, an overactivity of NMDA‐mediated glutamatergic signaling, together with the reduction of the GABAergic inhibitory inputs, causes neuronal hyper‐excitability, ultimately leading to generalized seizures or NCSE [5, 6, 7].

According to European guidelines [8], the diagnostic work‐up should include history taking (including drinking history), neuroimaging and laboratory screening, in particular gamma‐glutamyl‐transpeptidase (gGT) levels. In order to exclude the use of other substances that might contribute to the onset of seizures, a toxicology screening can be added. Routine EEG is not recommended but helps to rule out an underlying epileptic disorder, including NCSE or focal dysfunction. Typically, post‐ictal EEGs in AWS are normal, or show non‐specific alterations, such as reduction of background activity [2, 9, 10].

In contrast to AWS, unprovoked seizures are associated with brain lesions or typical epileptiform discharges in the EEG, for which ASM is usually offered. This distinction is of major importance: in the latter case, an ASM with the optimal profile for the needs of the patient is chosen by the neurologist or other medical doctor in charge, whereas in the case of acute symptomatic seizures the underlying cause should be removed (in case of AWS, avoidance of abrupt withdrawal, anti‐addiction programs in dedicated facilities, etc.). However, like all acute symptomatic seizures, AWS are not necessarily a single event, and in some patients they may recur, in particular, because the removal of the underlying factors is not easy [11]. Despite the high frequency of AWS as a neurological emergency, not much is known about the risk factors for delayed AWS relapses and the long‐term evolution of patients with AWS. In this retrospective study, we aimed to identify risk factors for AWS recurrence and to study the overall outcome of patients after AWS.

METHODS

This retrospective single‐center study was conducted in patients admitted to the Emergency Department (ED) of the University Hospital of Geneva with probable AWS and for whom an EEG was performed between January 1, 2013 and August 10, 2021. In the ED of our university hospital, EEG is offered to the majority of patients with AWS, to look for AWS complications or NCSE. Inclusion criteria were isolated seizures after alcohol withdrawal, either measured with blood alcohol level or determined from the history and examination.

For each patient, the following anamnestic data were collected: age, gender, history of AWS, history of cirrhosis, and pre‐existing treatment with benzodiazepines and with ASMs. We also monitored whether benzodiazepines or ASMs were prescribed after the index event for those who never received these treatments (“de novo”). Because only a fraction of patients had a neurological control visit after the index event, the presence or change of long‐term medical treatment could not be reliably determined. However, the vast majority of seizure emergencies in the canton of Geneva are referred to the ED of the University Hospital, where EEGs can be carried out 365/365 days of the year, so we believe that we have a reliable estimation of recurrences of generalized seizures.

In addition to the elements of the patient's personal history, the following laboratory findings were obtained upon admission and included in the study: mean corpuscular value, gGT level and plasma sodium concentration. Blood alcohol is not regularly studied in our ED, in particular if the patient reports no alcohol consumption for the last 48 h and withdrawal symptoms are noted. EEG findings were classified into the following categories: (i) normal, with or without the presence of beta‐rhythms, vigilance changes or non‐specific focal slowing; (ii) encephalopathic EEG (diffuse continuous, intermittent slowing or background slowing unrelated to vigilance); and (iii) presence of epileptiform EEG abnormalities, such as spikes, spike‐waves, intermittent rhythmic (focal) delta activity, which are together referred to as interictal epileptiform discharges (IEDs), and seizures. Computed tomography (CT) reports were also reviewed and the following information was extracted: (i) acute intracranial hemorrhages, including intraparenchymal hemorrhages, subarachnoid hemorrhages, subdural hemorrhages, and epidural hemorrhages; (ii) cortical–subcortical atrophy not attributable to aging; (iii) traumatic cerebral sequelae; and (iv) cranial fractures.

Patients’ ED files were reviewed until April 29, 2022, to determine the occurrence of AWS relapses and possible relevant factors surrounding these events. Patients were subdivided into those with (r‐AWS group) and without relapses (no relapse AWS [nr‐AWS] group) that occurred during the follow‐up period.

Statistical analyses

For dichotomous (yes/no) variables, relative frequencies of events were calculated and comparisons were carried out with chi‐squared tests (χ ²) and, when more than 20% of cells had an expected count lower than 5, Fisher's exact test. For quantitative variables, descriptive statistical parameters included mean, median, and standard deviation, comparisons between groups were performed through Mann–Whitney U‐tests. Moreover, we also performed logistic regression, including either “AWS relapse during follow‐up” or “death during follow‐up” as the dependent variables. Survival statistics included Kaplan–Meier curves, their comparison through log‐rank tests, and Cox proportional hazards regression.

RESULTS

Relapse rate

In total, we identified 199 patients with AWS, of whom 22 (11%) presented another seizure during follow‐up. Their mean (± SD) age was 53 (± 12) years, and patients were predominantly male (78.9%; p < 0.001). There were no differences in age (p = 0.22) or sex (p = 0.36) between the r‐AWS and nr‐AWS groups. Follow‐up ranged from 61 to 3383 days (2 months to 9.2 years), with a median of 1766 days (4.8 years). Out of 22 patients with at least one AWS relapse, 14 had one relapse, six had two relapses and two had three relapses. The median time between the index event and the first AWS relapse was 470.5 days.

Clinical variables

Among all patients, 55 (27.6%) had a history of at least one previous AWS, 11 in the r‐AWS (50%) and 44 (24.8%) in the nr‐AWS group (χ ² = 6.2, p = 0.013; log‐rank χ ² for first AWS relapse = 7.2, p = 0.007; Figure 1). Cirrhosis was found in 22 patients, yet no difference between the two groups was noted (p = 0.75). Table 1 provides a summary of the results.

FIGURE 1.

Cumulative freedom from alcohol withdrawal seizures (AWS) relapse. Light blue curve: patients without history of previous AWS; dark blue curve: patients with a history of previous AWS. Log‐rank test between the two groups (event = first AWS relapse): χ ² = 7.2, p = 0.007.

TABLE 1.

Patient characteristics

	All (N = 199)	Patients with relapse (r‐AWS; N = 22)	Patients without relapse (nr‐AWS; N = 177)	p value
Gender	42 women, 157 men	3 women, 19 men	39 women, 138 men	0.36
Age, years Mean ± SD Median	53.47 ± 12.24 54	51.14 ± 9.9 52.5	53.76 ± 12.5 54	0.22
History of AWS, n (%)	55 (27.6)	11 (50)	44 (24.85)	0.013
History of BZDs or ASMs or both, n (%)	88 (44.2)	6 (27.3)	82 (46.3)	0.09
History of BZDs, n (%)	86 (46.7)	6 (27.3)	80 (45.2)	0.1
History of ASMs, n (%)	7 (3.5)	0	7 (3.9)	0.34
BZDs and/or ASMs after the index event, n (%)	72 (36.5); N = 197	9 (42); N = 21	63 (35.8); N = 176	0.52
BZDs after the index event, n (%)	48 (24.3); N = 197	8 (38); N = 21	40 (22.7); N = 176	0.12
ASMs after the index event, n (%)	29 (14.7); N = 197	1 (4.7); N = 21	28 (15.9); N = 176	0.17
Severe hyponatremia, n (%)	5 (2.6); N = 193	2 (9.1); N = 21	3 (1.7); N = 172	0.09 a
Cirrhosis, n (%)	22 (11.1)	2 (9.1)	20 (11.3)	0.75
Pathological CT findings, n (%)	65 (35.7); N = 182	7 (35); N = 20	58 (35.8); N = 162	0.94
Brain atrophy, n (%)	48 (26.3); N = 182	4 (20); N = 20	44 (27.1); N = 162	0.49
Intracranial hemorrhages, n (%)	19 (10.4); N = 182	3 (15); N = 20	16 (9); N = 162	0.48
Traumatic sequelae, n (%)	6 (3.3); N = 182	0; N = 20	6 (3.7); N = 162	0.38
Cranial bone fractures, n (%)	4 (2); N = 182	3 (15); N = 20	1 (0.6); N = 162	0.004 a
EEG: Normal, focal slowing, n (%)	175 (88.8); N = 197	18 (85.7); N = 21	157 (89.2); N = 176	0.63
EEG: Encephalopathy, n (%)	13 (6.6); N = 197	1 (4.8); N = 21	12 (6.8); N = 176	0.72
EEG: Epileptiform discharges, n (%)	10 (5.1); N = 197	3 (14.3); N = 21	7 (4); N = 176	0.077 a

Note: Level of significance = 0.05; p values in bold are significant.

Abbreviations: AWS, alcohol withdrawal seizures; BZDs, benzodiazepines; CT, computed tomography; EEG, electroencephalogram; N, number of patients for which a given variable was known (when not indicated, all patients were concerned); nr‐AWS, patients without alcohol withdrawal seizure relapses during follow‐up; r‐AWS, patients with alcohol withdrawal seizure relapses during follow‐up, SD, standard deviation.

^a Fisher's exact test.

Laboratory

The majority of patients had elevated gGT values, without differences between the r‐AWS and nr‐AWS group. Sodium levels at the index event were available in 193 patients and were lower in the r‐AWS group (133.5 vs. 136.8 mEq/L; p = 0.002). Five patients (2.6%) had severe hyponatremia (defined as <125 mEq/L) at the moment of the index event and two of these had subsequent relapses (comparison between r‐AWS and nr‐AWS groups: Fisher's exact test p = 0.09), although they had no severe hyponatremia at the moment of the relapse. None of the patients with severe hyponatremia had IEDs in the EEG.

Computed tomography findings

Brain CT scans were performed in 182 patients, of whom 65 (35.7%) had abnormal CT scans and 12 (6.6%) had several pathological findings. Intracranial hemorrhages were found in 19 patients (10.4%), significant brain atrophy in 48 patients (26.3%) and traumatic cerebral sequelae in six patients (3.3%). Four patients had skull fractures (2.2%). Only skull fractures were significantly associated with AWS relapse during follow‐up (three out of four had AWS relapses, χ ² = 17.1, p < 0.001; Fisher's exact test p = 0.004).

Electroencephalogram findings

An EEG report was available for 197/199 patients at the time of the index AWS (21/22 in r‐AWS group; 176/177 in nr‐AWS group) and for all patients at the time of AWS relapse. EEG was normal or showed either diffuse high frequency patterns or vigilance changes in 175 patients (88.8%). Encephalopathic EEG findings, that is, diffuse continuous or intermittent slowing or background slowing unrelated to vigilance, were described in 13 patients (6.6%). Epileptiform EEG abnormalities including IEDs and seizures were identified in 10 patients (5.1%), and in two of these 10 patients in the context of intracranial hemorrhage. None of the EEG features was different between the groups, except for epileptiform abnormalities, which were found much more frequently in the r‐AWS group (3/21 patients) than in the nr‐AWS group (7/176 patients; p = 0.042, although Fisher's exact test p = 0.07). Epileptiform abnormalities were also significantly more often noted in patients with brain atrophy than those without (12.5% vs. 2.2%; Fisher's exact p = 0.012).

Drug treatment effects

Benzodiazepine treatments included oxazepam, lorazepam and clonazepam, while ASMs included valproic acid, lamotrigine, levetiracetam, phenobarbital, and gabapentin. Prior drug therapy (benzodiazepines and/or ASM) was more frequently identified in the medical history of patients without relapse, although this did not reach the level of significance (46.3% vs. 27.3%, Fisher's exact test p = 0.09). When we considered them separately, pre‐existing benzodiazepines and ASM treatments were present in 86 (46.7%) and seven patients (3.5%), respectively, at the time of the index AWS (no significant difference between the groups; p = 0.1 and p = 0.34, respectively).

Regarding drug treatments following the index event (“de novo”) and before EEG execution, 72 patients (36.5%) received either ASMs or benzodiazepines. Again, no difference was noted between the groups (p = 0.52).

Mortality

A total of 25 patients (12.5%) died, after a mean (±SD; median) of 811 (± 482; 848) days from the index event. We calculated a mortality rate of 2.9% per year of follow‐up for survival and 3.2 per 100 person‐years of follow‐up for survival. The mean (±SD; median) age at death was 60 (±12; 62) years. With the log‐rank test, survival rate did not differ significantly between the r‐AWS group and the nr‐AWS group, but was significantly lower in patients with a history of previous AWS (Figure 2; χ ² = 14.9, p < 0.001; log‐rank χ ² = 14, p < 0.001) and encephalopathic EEG (χ ² = 4.1, p = 0.043). In particular, the mortality rate was 1.6% per year and 1.7 per 100 person‐years among patients without a history of previous AWS, whereas it was 6.2% per year and 7.3 per 100 person‐years among those without a history of previous AWS.

FIGURE 2.

Cumulative survival rate. Light blue curve: patients without history of previous alcohol withdrawal seizures (AWS); dark blue curve: patients with a history of previous AWS. Log‐rank test between the two groups (event = death): χ ² = 14, p < 0.001.

Prediction of relapse and mortality

The logistic regression model included “AWS relapse during follow‐up” as the dependent variable and the following independent variables: “history of previous AWS”, “epileptiform EEG abnormalities”, “encephalopathic EEG findings” and “cranial bone fractures”. The model was significant (χ ² = 16.8; p = 0.002, Nagelkerke R ² = 0.17). This was mainly due to “epileptiform EEG abnormalities” (regression coefficient β1 = 1.6; odds ratio = 5, 95% confidence interval [CI] 1.1–22.1; p = 0.03) and “cranial bone fractures” (regression coefficient β2 = 3.4; odds ratio = 30.3 with 95% CI 2.8–325.8; p = 0.005), while “history of previous AWS” was marginally significant (regression coefficient β3 = 1; odds ratio = 2.6, 95% CI 0.98–7; p = 0.054).

Concerning mortality prediction, the relative risk of death related to the risk factor “history of previous AWS” was 3.9 (95% CI 1.9–8.2; p < 0.001), while Cox regression showed a hazard ratio of 4 (95% CI 1.8–9.1; p = 0.001).

DISCUSSION

Our retrospective study provided the following main findings: (i) patients with AWS experience relapse more often if they already have a history of AWS or (ii) in case of skull fractures; (iii) AWS are associated with increased mortality; and (iv) history of previous AWS is a predictor for death.

In a previous study, seizure relapse within 6 h after alcohol withdrawal was noted in 24% of patients treated with placebo [12]. None of our patients experienced an early seizure relapse in the ED, which may have been due to the transient, but regular use of benzodiazepines in the acute phase to avoid withdrawal symptoms while still in the ED. However, 11% experienced delayed AWS recurrences, with a median time of approximately 1.3 years between the index event and the first AWS relapse.

A history of AWS predisposes to relapse, consistent with earlier observations [11]. Whether this association integrates into the definition of established epilepsy (i.e., two recurrent seizures in >24 h, increasing the risk to >60% within the next 10 years if not treated) [13] remains to be studied. We identified several other risk factors that are well‐established markers also for post‐traumatic epilepsy [14, 15]: the presence of IEDs in the EEG was associated with the development of an epileptic disorder [9], and skull fractures. A CT scan is often performed after AWS, in particular, if head trauma is suspected [10, 16]. Previous studies found intracranial lesions in 6%–15% (e.g., subdural hematomas, hygromas and skull fractures with subarachnoid hemorrhages) [17, 18, 19]. In our study, head CT disclosed at least one structural abnormality in more than one third of patients. If approximately one out of 10 patients with AWS will have another seizure, it might be worthwhile to consider CT and EEG regularly to identify patients at risk for recurrence.

Regarding metabolic causes, we noted significantly lower sodium levels in r‐AWS patients. However, severe hyponatremia was found only in a minority of patients and equally often in both groups. Hyponatremia is a well‐known side effect of alcohol abuse, and based on our findings, we hypothesize that lower sodium levels may facilitate seizures, together with other clinical, EEG‐ and CT‐based risk factors mentioned above. Other studies found that acute symptomatic seizures caused by metabolic disturbances are associated with a higher risk of relapse, perhaps due to an underlying tendency of such an abnormality to recur [20, 21]. However, in our sample, the two r‐AWS patients with severe hyponatremia did not show the same metabolic disturbance at the moment of relapse, suggesting that the recurrence of metabolic dysfunctions is rare and not relevant to the relapse.

Interestingly, patients with brain atrophy had significantly more IEDs than their counterparts. Although previous studies did not find differences in the prevalence of atrophy between alcoholic patients with and without epilepsy, a more generalized pattern of brain atrophy was found in the former group, suggesting its potential epileptogenic role [16, 22]. Also in other epileptogenic disorders, such as temporal lobe epilepsy or dementia, cortical thinning and atrophy (of the hippocampus or cortex) are associated with increased epileptogenicity, probably due to loss of inhibitory neurons or altered neuronal circuits [23, 24]. Nonetheless, additional evidence is necessary to corroborate these observations and to determine if brain atrophy in patients with alcohol abuse share the same mechanisms as in patients with other epilepsy syndromes.

Consistent with the existing body of literature, AWS affect mainly middle‐aged men, which is quite consistent across time and geographical regions [12, 25, 26, 27]. The reason for this association is not clear. One previously proposed explanation was the higher preponderance of alcohol abuse in men, which, however, does not explain the high percentage of AWS. Men are more often involved in binge drinking with higher quantities of alcohol, leading to AWS and frequent delirium tremens. Women with alcohol use disorder tend to have more severe cognitive and motor impairment related to significantly lower but also more stable alcohol exposure compared to men, which may prevent major blood level changes [27]. Finally, a genetic origin could be hypothesized. Patients with AWS more often present the 9‐repeat allele of Dopamine transporter gene 1 (DAT1) compared to alcoholics without AWS [28]. However, a sex preponderance of this allele could not yet be confirmed.

Mortality is increased in patients with substance abuse, including alcohol abuse. In a population‐based retrospective cohort from Canada, mortality was examined in patients who presented at least twice a year to the ED for alcohol‐related reasons [26]. While AWS were not identified as a subgroup, the all‐cause 1‐year mortality rate was 5.4%. In a retrospective observational study, the mortality rate in patients with substance abuse was 3.3% within 2 years following their index event [29]. Another study reported a fourfold increased mortality risk for patients presenting with alcohol‐related seizures, compared with the general population [25]. In our study, we found a mortality rate of 2.9% per year (or 3.2 per 100 person‐years) of follow‐up in patients with AWS, which is approximately 13 times higher than the overall mortality (0.22%) in the canton of Geneva in the age group 40–64 years [30, 31]. The average age at death (59.2 years) was significantly lower than the average life expectancy for men and women in Switzerland in 2020, which is 81 and 81.5 years, respectively. Moreover, a history of previous AWS increases the risk of death by approximately four times, compared to patients without a history of previous AWS. It is not clear if repeated AWS correspond to an untreated epilepsy disorder or if they indicate the final stages of chronic alcoholism and major cerebral and physical injury associated with a high risk for death.

The role of benzodiazepines, phenytoin, phenobarbital and pregabalin in preventing further AWS has already been studied, but most of the studies looked at short‐term outcomes (e.g., length of hospital, ED or intensive care unit stay, seizure recurrence within the next 6–12 h) and found beneficial effects of some molecules [32, 33]. Since published studies and meta‐analyses mainly concern short‐term treatments (days or few weeks) [34, 35], the role of long‐term ASM is still unclear. In our study, previously prescribed as well as de novo‐introduced benzodiazepines or new ASM after AWS did not show long‐term protection against AWS recurrence, but we do not know the effective length of the treatment, nor the patient compliance level. Prospective studies could determine the benefit of ASMs, the optimal molecule, as well as the need for regular follow‐up, to prevent AWS relapses and impact on mortality.

Alcohol withdrawal seizures are related to an acute decrease in alcohol intake. The abrupt interruption of drinking could be due to a lack of supply but also linked to psychiatric factors. In a previous observation, depression, bipolar disorders, personality and anxiety disorders were described in 50% of all patients with alcohol use disorders, some of which may predispose patients to change (repeatedly) their alcohol consumption pattern [36]. Since our AWS patients did not see a psychiatrist as part of their work‐up in the ED, we do not possess any information on relevant psychiatric comorbidities. To the best of our knowledge, there is no prospective clinical trial taking into account such comorbidities.

Our study has several limitations, including its retrospective nature. Our findings may apply only to generalized tonic–clonic seizures. Focal seizures with alteration of consciousness only are most likely incorrectly assigned to alcoholization and therefore, we might underestimate the presence of epilepsy in patients with alcohol abuse if they are not associated with generalized seizures. Another limitation is the low number of pathological EEGs recorded, which might have influenced our results. This should be re‐addressed in future studies with larger sample sizes. Also, we included only patients for whom an EEG was requested, so we do not know the outcome for those patients with a diagnosis of AWS who did not undergo EEG recording. Most patients are not referred to neurologists or psychiatrists for follow‐up after discharge from the ED, due to the label of “acute symptomatic seizures”, or because patients do not desire to undergo behavioral and medical treatment. Thus, our study, like others, is limited by the absence of a structured follow‐up, including a detailed history of duration and type of ASM and benzodiazepine treatment. However, we were able to obtain a reliable estimate of the incidence of recurrent generalized tonic–clonic AWS, since almost all inaugural generalized seizures (and their relapses) are seen in the ED of our hospital.

In conclusion, a substantial number of AWS patients (11%) present an AWS relapse, and risk factors include mainly those that are known for the development of an epilepsy disorder. Patients with AWS, or patients with substance abuse in general, require particular medical attention but are less likely to receive specialist care, despite repeated visits in the ED and a high mortality rate. Increased mortality is a consistent finding in most studies on alcohol abuse, but few management options are provided to these patients in the ED. The classification of AWS as acute symptomatic seizure with very little risk of recurrence needs to be revisited, and true relapse rates, together with the potential benefits of ASM, should be addressed in prospective studies.

CONFLICT OF INTEREST STATEMENT

Serge Vulliémoz and Margitta Seeck have shares at Epilog(R).

ETHICAL APPROVAL

The study was conducted according to 1964 Declaration of Helsinki and its later amendments. Approval from ethical committee was waived because of the nature of the study (retrospective).

Sansone G, Megevand P, Vulliémoz S, Corbetta M, Picard F, Seeck M. Long‐term outcome of alcohol withdrawal seizures. Eur J Neurol. 2024;31:e16075. doi: 10.1111/ene.16075

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are available on reasonable request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.