Phenobarbital and Alcohol Withdrawal Syndrome: A Systematic Review and Meta-Analysis

Alcohol withdrawal syndrome (AWS) is a complication frequently encountered among patients who are chronic alcohol abusers. It is considered to have a significant impact on the United States healthcare system. It not only has a toll on the healthcare spending but also contributes to significant morbidity and mortality. Benzodiazepines are considered first line in the treatment of AWS. Since patients with alcohol use disorder have downregulated gamma aminobutyric acid (GABA) receptors, this often leads to benzodiazepine resistance. Phenobarbital is also used in the management of alcohol withdrawal syndrome. Here we present a systematic review and meta-analysis of the efficacy and safety of the drug. We conducted an electronic database search for relevant studies published between the inception of the project and November 20, 2022, in three databases, including Medline/PubMed, Embase, and Cochrane Library. Our study included all original studies with prime focus on the baseline characteristics of patients admitted to the intensive care unit (ICU) for alcohol withdrawal syndrome and management/monitoring protocol implemented for its treatment. The primary outcomes that were the focus of our study consisted of changes in the length of hospital stay, length of ICU stay, and changes in scoring systems (for alcohol withdrawal assessment and monitoring) following the implementation of phenobarbital. The secondary outcomes included complications such as intubation and mortality. Based on our analysis, the mean difference in hospital stay was statistically significant at -2.6 (95% CI, -4.48, -0.72, P=0.007) for phenobarbital compared to the benzodiazepine group. We were unable to comment on the heterogeneity in our meta-analysis due to the standard deviation not being reported in one study. There was no statistically significant difference regarding the length of stay in the intensive care unit compared to the control/comparative arm, with a mean difference of -1.17 (95% CI, -1.17, 0.09, P=0.07), with considerable heterogeneity (I2=77%, P=0.002). Our meta-analysis also investigated the risk of intubation between the phenobarbital and the control/comparative group. There was statistically significant difference in the incidence of intubation, relative risk (RR) 0.52 (95% CI, 0.25, 1.08, P=0.08), with considerable heterogeneity (I2=80%, P=0.0001). Our study concludes that phenobarbital is an effective tool in the management of AWS in an ICU setting. However, various studies have reported contradictory results, and vital information appears to be lacking. Moreover, there is a lack of uniformity in terms of phenobarbital dosing. Drug administration should be adapted according to the severity of the symptoms. Further studies need to be conducted discussing the safety profile and adverse effects of the drug when it comes to the management of alcohol withdrawal syndrome.


Introduction And Background
In the United States, alcohol abuse represents a significant healthcare burden, with over 14 million Americans suffering from alcohol use disorder. In 2010, a total of $249 billion was spent in the United States on alcohol-related disorders. Despite warnings from the World Health Organization, alcohol consumption is expected to increase until at least 2025. Hospital admissions for chronic alcohol abusers are particularly dangerous since these patients are at an increased risk of developing life-threatening conditions such as alcohol withdrawal syndrome (AWS) [1].
Chronic alcoholism leads to a decrease in gamma aminobutyric acid (GABA) inhibitory receptors as part of a feedback mechanism to counteract the depressive effects of alcohol. Additionally, excitatory glutamate receptors are also upregulated, decreasing the influence of alcohol. When alcohol intake is abruptly ceased, the body does not have enough time to adjust GABA and glutamate regulation, resulting in alcohol 1 2 3, 4 5 1 withdrawal syndrome. Symptoms of AWS begin to manifest eight hours after the last drink and peak 24 to 72 hours later. Patients initially experience nausea, sweating, headaches, shakiness, tachycardia and hypertension. Symptoms become more severe over time, leading to seizures and delirium tremens [2,3].
Pharmacological agents such as benzodiazepines, barbiturates, beta-blockers, butyrophenones, gabapentin, propofol, dexmedetomidine, and valproic acid have been used to treat alcohol withdrawal syndrome to reduce its intensity and prevent life-threatening complications. The ideal agent for treating AWS should have a rapid onset, a strong safety profile and properties such as anxiolysis and sedation. The American Society of Addiction Medicine's 2020 Clinical Practice Guidelines recommend benzodiazepines as the firstline treatment for addiction [4]. The binding of benzodiazepines to GABA receptors increases the frequency of opening of these channels. Benzodiazepines are fully effective when GABA is available at the receptor site. A patient with alcohol use disorder has low levels of GABA, causing benzodiazepines to fail to produce the desired response, resulting in resistant alcohol withdrawal syndrome, characterized by hallucinations, seizures, and delirium tremens [3][4][5]. Therefore, other pharmacological agents can be used to control the debilitating symptoms of alcohol withdrawal syndrome. Phenobarbital (a barbiturate) is being increasingly recommended for the treatment of patients with a contraindication to benzodiazepines [4]. Phenobarbital also acts on glutamate in addition to GABA. GABA and glutamate receptors are both responsible for the symptoms of AWS. The effect of phenobarbital is robust because, unlike benzodiazepines, it does not require endogenous GABA [3,6,7]. Several observational studies have been published on the effectiveness of phenobarbital in AWS, with mixed results. As a result, we conducted our meta-analysis in order to assess phenobarbital's effectiveness in AWS.

Review Search strategy and study selection
The study methodology was based on the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines, in which protocol registration is not required [8]. We conducted an electronic database search for relevant studies published between the inception of the project and November 20, 2022, in three databases, including Medline/PubMed, Embase, and Cochrane Library. In our search, we used keywords, Medical Subject Headings (MeSH) terms, and publication type based on the participants, comparison, intervention, and outcomes (PICO) framework. The following terms were used according to each database: "Barbiturate" OR "Barbi" OR "Barb" OR "Phenobarbital" OR "Phenobarb" OR "Pheno" OR "Phenobarbital" [Mesh] OR "Barbiturates"[Mesh] AND "Alcohol withdrawal syndrome" OR "Alcohol withdrawal" OR "Delirium tremens" OR "Delirium" OR "Alcohol Withdrawal Delirium" [Mesh] OR "Alcohol-Induced Disorders, Nervous System"[Mesh] OR "Substance Withdrawal Syndrome" [Mesh] OR "Alcoholism" [Mesh] AND "Intensive care" OR "Intensive care unit" OR "Critical care" OR "Intensive therapy unit" OR "Intensive Care Units" [Mesh].
The study included all original studies, including cohort, cross-sectional, and case-control studies, which provided information on baseline characteristics, comorbidities, initial assessment of patients with alcohol withdrawal syndrome, management using phenobarbital, and a monitoring protocol implemented in the intensive care unit (ICU) for the treatment of AWS, outcomes, and complications observed. In addition, we included commentary and case series that included at least 10 patients. Exclusion criteria included nonoriginal reports, such as reviews, letters to editors, and commentaries, which did not include original patient data. In addition, case reports and case series involving fewer than 10 patients were excluded. This study excluded the following: full texts with limited data/data not relevant to the study, unextractable or irrelevant data, articles not published in English, duplicate records, animal studies, and overlapped data.
The primary outcomes were the mean length of hospital stay, mean length of ICU stay, and changes in Clinical Institute Withdrawal Assessment for Alcohol (CIWA) scale/additional scoring system scores following the implementation of phenobarbital. The secondary outcomes included complications (such as intubation) and mortality.
We conducted a manual search of the references of our included papers in order to ensure that we did not miss any relevant studies. All original studies that reported the management of AWS in the ICU setting with phenobarbital were included in the analysis. Two independent reviewers screened the titles and abstracts of the papers, followed by a full-text screening to ensure that relevant papers were included in the systematic review. Disagreements were resolved by discussion and by referring to the senior author whenever necessary.

Data extraction
The data extraction sheet was developed by one author using Microsoft Excel (Microsoft, Redmond, WA). Two independent reviewers extracted data using an Excel spreadsheet. The author who developed the data extraction sheet checked the extracted data for accuracy. When necessary, the senior author was consulted for the resolution of disagreements and discrepancies.

Quality assessment and statistical analysis
An independent reviewer evaluated the risk of bias in the included studies. We assessed the quality of the included studies using the risk-of-bias assessment tool developed by the National Institutes of Health (NIH) [9]. An analysis of the descriptive data was conducted using SPSS Statistics, version 26 (IBM Corp., Armonk, NY). RevMan 5.4 (Cochrane Collaboration, Copenhagen) was used to develop the forest plot and funnel plot.

Search Results
We identified 90 records using EndNote 20 (Clarivate, London). After removing duplicate records and manually screening abstracts and titles, 21 records were identified for further full-text screening. A total of 11 studies were assessed for inclusion into our systematic review. Two studies were excluded due to the data being individualized for every patient and lack of mean/median values of the variables examined for the entire study population. Ultimately, nine studies were included in this systematic review and meta-analysis after 12 papers were excluded from the full-text screening phase ( Figure 1).

Study Characteristics and Quality of the Included Studies
The baseline characteristics of the included studies are summarized in Table 1. A total of nine studies were included: one retrospective case series with more than 11 participants and including original data [10], seven retrospective studies [11][12][13][14][15][16][17], and one abstract with data relevant to the study design [18]. The sample sizes of the included studies ranged from 31 to 135 individuals. Patients aged 18 and older were included in the study.  Patients with a high risk for AWS were defined as patients with a previous history of DTs with or without history of alcohol withdrawal seizures and recent alcohol use (≤2 weeks) with positive blood alcohol levels (BALs) >0.1 g/dL, or elevated mean corpuscular volume (MCV), or elevated aspartate aminotransferase (AST) to alanine aminotransferase (ALT) ratio ≥1.5:1. Patients with a medium risk for AWS were defined as patients with active alcohol dependence per patient's reporting plus two of the following criteria: two or more days since last drink, or an elevated BAL >0.1 g/dL on admission, or symptoms of autonomic dysfunction with the BAL >0.1 g/dL, or elevated MCV, or AST:ALT ratio ≥1.5:1.28-31.
<> denotes not mentioned/not available.

Baseline Characteristics and Comorbidities of Patients with AWS
According to Ammar et al., phenobarbital was administered to patients admitted to the surgical intensive care unit for the prevention or worsening of AWS [10]. In a study conducted by Oks et al., patients were treated with phenobarbital according to the pharmacological protocol of the medical intensive care unit to manage AWS [11]. Duby et al. conducted a retrospective study where patients in the preintervention group (PRE) were treated in a non-protocolized fashion and typically received continuous infusions or scheduled doses of benzodiazepines according to physician preferences. The AWS protocol was used to administer escalating diazepam and phenobarbital to patients in the postintervention group (POST) [16]. As part of the retrospective cohort study conducted by Gold et al., patients who required escalating doses of benzodiazepines during the pre-guideline era (i.e., the control group) were admitted to intensive care units. While in the intensive care unit, they received benzodiazepine and phenobarbital as a non-protocolized treatment. In the post-guideline era, patients admitted to the ICU for the management of AWS were administered benzodiazepine and phenobarbital in a protocolized manner [17]. We found that the majority of the studies included in our study included patients with a prior history of AWS, polysubstance abuse, and alcohol withdrawal seizures [10,[12][13][14][15][16][17].

Management and Monitoring Protocols Implemented in the ICU for the Management of AWS
The management protocols implemented for the treatment of AWS in an ICU setting in both control and experimental groups have been described in depth in Table 2. A variety of scoring systems were used for the assessment and monitoring of alcohol withdrawal. Ammar [16]. Apart from benzodiazepines and phenobarbital, a few other medications, particularly sedatives, were used for managing alcohol withdrawal patients both in the control and experimental arms of the studies included in our systematic review and meta-analysis. A few participants in the case series conducted by Ammar et al. received quetiapine, haloperidol, or a combination of the two for the management of alcohol withdrawal syndrome [10]. Patients requiring mechanical ventilation in the study conducted by Gold et al. received propofol [17]. Goodberlet et al. conducted a study where a few of the participants in both control and experimental arms received dexmedetomidine, clonidine, propofol, and antipsychotics [13]. A total of 11 patients in the experimental group and 26 patients in the control group received propofol and dexmedetomidine for the management of alcohol withdrawal syndrome in the study conducted by Saukkonen et al. [18]. A retrospective study conducted by Shah et al. mentions a few subjects, both in the control and experimental arm, requiring sedatives and other adjunctive medications (both ventilated and non-ventilated patients) such as dexmedetomidine, ketamine, propofol, antipsychotics and gabapentin [15]. Participants in the study conducted by Duby et al., required dexmedetomidine (4 patients in the experimental arm vs. 17 patients in the control arm), olanzapine (five patients in the experimental arm vs. seven patients in the control arm), haloperidol (4 patients in the experimental arm vs. 10 patients in the control arm) and quetiapine (two patients in the experimental arm vs. five patients in the control arm) for further management [16].

Results of the Meta-Analysis, Outcomes, and Complications
Our meta-analysis included seven studies. The purpose of this study was to determine the length of hospital stay and ICU stay (days) as well as the rate of intubation (Tables 3, 4). Pooled analysis of the length of hospital stay included two studies with a total of 175 patients. The mean difference was statistically significant at -2.6 (95% CI, -4.48, -0.72, P=0.007). We were unable to comment on the heterogeneity in our meta-analysis due to the SD not being reported in one study ( Figure 2) [14,18]   History of severe withdrawal, withdrawal seizures, or delirium tremens: median APACHE II score of 10 (5:13) in the POST guideline group as compared to 4 (2.8:25) in the PRE guideline group.
Presentation with withdrawal seizures or delirium tremens: median APACHE II score of 10 (5:14.5) in the POST guideline group as compared to 5 (3.25:6.75) in the PRE guideline group.
MICU patients: median APACHE II score of 10 (5:16) in the POST guideline group as compared to 6 (2:10) in the PRE guideline group.
Phenobarbital load (mg/kg ideal body weight): median APACHE II score of 8 (4:13) when patients received a phenobarbital loading dose ≥9 mg/kg ideal body weight as compared to 10 (7:16) in patients receiving a loading dose of <9 mg/kg ideal body weight.
b The number of patients receiving dexmedetomidine was much lower in the phenobarbital group than in the CIWA group: 4 (7%) vs. 17    Source: [14,18] FIGURE 3: Forest plot of the length of ICU stay for the phenobarbital vs.

control group
The green box represents individual study effects and the black diamond represents the combined result of the studies.

control group
The blue box represents individual study effects and the black diamond represents the combined result of the studies.

Risk of Bias of Included Studies
On the NIH risk-of-bias assessment tool, all of the studies included scored 10 or above, except for the abstract (a retrospective study) by Saukkonen et al., which received a score of 5 [11][12][13][14][15][16][17][18]. On the NIH tool, the case series included in our study received an overall score of 8 out of 9 [10].

Discussion
The purpose of our meta-analysis was to evaluate the effectiveness of phenobarbital in the treatment of alcohol withdrawal syndrome. Our study examined the patient's length of hospital stay, length of stay in the intensive care unit, and interventions such as intubations during hospitalization for alcohol withdrawal syndrome. Based on our analysis, the mean difference in hospital stay was statistically significant at -2.6 (95% CI, -4.48, -0.72, P=0.007) compared to the benzodiazepine group. We were unable to comment on the heterogeneity in our meta-analysis due to the SD not being reported in one study [14][15][16][17][18]. Based on the results of our meta-analysis, there was no statistically significant difference in the length of stay in the intensive care unit compared to the control/comparative arm, with a mean difference of - investigated the risk of intubation between the phenobarbital and the control/comparative group. There was a statistically significant difference in the incidence of intubation, RR 0.52 (95% CI, 0.25, 1.08, P=0.08), with considerable heterogeneity (I 2 =80%, P=0.0001) [13][14][15][16][17][18].
Only two studies with a total population of 175 participants compared the mean length of hospital stay between the intervention and the control arm whereas only one study compared the median length of hospital stay between the intervention and the control arm [13,14,18]. Unfortunately, due to the lack of SD values, the heterogeneity of the data could not be evaluated. Both Tidwell et al. and Saukkonen et al. found that phenobarbital treatment reduced hospital stay in both arms. We also found the same result in our analysis. The small population of the subgroup is a limitation of the analysis [14,18]. In contrast, Goodberlet et al. reported the opposite findings showing more prolonged hospitalization in the phenobarbital group compared to benzodiazepine [13].
ICU stay was reported in all the studies included in the systematic review [10][11][12][13][14][15][16][17][18] Tidwell et al. also reported a significant reduction in the length of the ICU stay in the phenobarbital arm [14]. According to Duby et al., the ICU stay was also reduced in the phenobarbital group [16]. According to Goodberlet et al., patients who received phenobarbital spent longer in the intensive care unit. The extended ICU stay can be attributed to the study limitations since patients in the phenobarbital arm had severe symptoms, which could explain the prolonged stay in the ICU. In addition, patients were selected only when dexmedetomidine administration was limited in the institute, thereby influencing the choice of sedatives and other pharmacological agents [13].
We also examined interventions such as intubation. According to the study by Tidwell et al., the phenobarbital arm showed a decreased incidence of intubation [14]. Conversely, none of the five other studies were able to demonstrate such a link. We also found that there was no significant correlation between phenobarbital administration and lower intubation rates when compared to the control group in our meta-analysis. Reduced sedation needs may be responsible for the lower incidence of intubations, as previously reported [13][14][15][16][17][18].
Additionally, adverse effects such as bradycardia, hypotension, and mortality were examined. The only study to record bradycardia and hypotension was that of Nguyen et al., which found no incidence in the parameters of either arm [12]. The remaining studies did not include these parameters. Monitoring and documentation were not as effective as they could have been because the included studies were retrospective. The limitations of these studies include the inability to control the acquisition of the desired data, which may lead to the loss of essential details. The study by Shah et al. showed that front-loaded phenobarbital participants experienced significantly fewer episodes of hypotension than those receiving low intermittent dosing of phenobarbital [15]. As reported by Goodberlet et al., there was no difference in mortality between the two arms [13]. Contrary to this, Duby et al. reported improved mortality outcomes in patients treated with phenobarbital [16]. Additionally, phenobarbital used for epilepsy is easily measurable, giving it an advantage over benzodiazepines. Therefore, phenobarbital is unlikely to cause adverse effects [19].
Some studies have been evaluated in the previous meta-analyses in recent years. Hammond et al. reviewed nine studies in 2016 and found that the phenobarbital group had lower hospital stays, fewer ICU admissions, and decreased mechanical ventilation as compared to the benzodiazepine group [20]. However, the study was limited to 294 patients, posing a serious limitation. A systematic review by Mo et al. found barbiturates to be superior to benzodiazepines in cases of severe AWS or benzodiazepine resistance [21].

Limitations
We have identified a major limitation of our analysis as the inclusion of retrospective studies. This increases the possibility of selection bias and limits the observer's control over documentation and vital signs' monitoring. The relatively small size of our study is another limitation of our analysis. The analysis was limited by the fact that a few critical parameters, such as hospital stay, were reported in only three studies, making it difficult to reach meaningful conclusions. Additionally, no study has focused on the safety profile of phenobarbital. Also, some studies did not provide data on adverse effects. A final problem was the lack of uniform rules or guidelines for administering phenobarbital, which led to disparities between the results of different studies.

Conclusions
In conclusion, phenobarbital can help reduce the length of hospital stay. However, it does not seem to have an impact when it comes to the length of ICU stay and the risk for intubation. Various studies have reported contradictory results. Since some vital details are lacking in the included studies, it is important to mention the need for large-scale, multi-center studies to be conducted before solid conclusions and recommendations can be made. Another important chapter that is missing from the previous studies is the safety of phenobarbital and the uniformity of dosing. Drug administration should also be adapted according to the severity of the symptoms. Additionally, there is a lack of detailed research regarding the adverse effects of phenobarbital administration in the context of alcohol withdrawal syndrome. Accordingly, the current evidence supports recommending phenobarbital for the treatment of alcohol withdrawal syndrome at a broad level.

Conflicts of interest:
In compliance with the ICMJE uniform disclosure form, all authors declare the following: Payment/services info: All authors have declared that no financial support was received from any organization for the submitted work. Financial relationships: All authors have declared that they have no financial relationships at present or within the previous three years with any organizations that might have an interest in the submitted work. Other relationships: All authors have declared that there are no other relationships or activities that could appear to have influenced the submitted work.