Commentary on "Gulati P, Chavan BS, Sidana A. Comparative efficacy of baclofen and lorazepam in the treatment of alcohol withdrawal syndrome. Indian J Psychiatry. 2019;61:60-4".

Sir,

The study by Gulati et al., 2019[1] comparing baclofen and lorazepam in alcohol withdrawal syndrome (AWS) in line of finding safer alternatives for managing AWS is well desired. However, we would like to highlight a few methodological and ethical issues.

One specific concern is the use of zolpidem by 15 (45.4%) patients for sleep in the baclofen group. As zolpidem improves sleep and anxiety[2](both symptoms of alcohol withdrawal), results in the baclofen group may have been exaggerated comparable to lorazepam.

It is unclear how many patients currently presenting with confusion/disorientation (suggesting complicated withdrawal) were included. On the other hand, the mean Clinical Institute Withdrawal Assessment for Alcohol-revised (CIWA-Ar) scores of both the groups were 14–15, suggesting mild–moderate withdrawal. Patients with CIWA-Ar ≤10 are considered having no or minimal withdrawal and usually do not need medication. Thus inclusion of patients at both ends of AWS severity induces considerable heterogeneity and limits generalization of results.

Another issue is the inclusion of patients with caffeine and nicotine dependence. As some symptoms of their withdrawal, e.g., headache, agitation, difficulty concentrating, nausea/vomiting, anxiety, and insomnia, may overlap with alcohol withdrawal, such patients could have been analyzed separately. It is also not mentioned if such patients got any additional treatment as this would affect the interpretation of results.

Baclofen has been used in dose of 30–270 mg/day in various studies,[3] but the rationale of using 30 mg/day for the current study has not been described. It is also not apparent if subjects were recruited from the emergency or outpatient department setting as this would also affect the results.

For the comparison of results, time × treatment interaction effects could have been more informative. Even though the authors have used ANOVA, its statistics have not been provided.

There are a few ethical concerns. It was not clear whether consent was taken from patients or caregivers. Patients under intoxication or acute withdrawal may be confused and may not understand implications of consent. How was the capacity of consent assessed (Mini-Mental State Examination or other tool) is not mentioned. As all patients would not have consented to participate in the study where they might receive a less established drug, refusal rate should have been provided.

Regarding safety, baclofen is known to affect renal functions,[4] but no testing at baseline or during the study was reported. Name of adverse effects' checklist was not provided. Further, investigations were reported for 45 (baclofen = 22 and lorazepam = 23) instead of 64 subjects.

Authors report no financial support, but baclofen is an expensive drug and it appears patients paid for it; however, it is customary that patients are provided the drug from the researchers if a new treatment is being tested.

Another aspect is the design of the study which appears to be parallel-group randomized controlled trial (RCT) involving a drug being tested for new indication; therefore, approval of agencies, e.g., the Directorate General of Health Services and registration with Clinical Trials Registry-India, may be needed.

Even though the study recommends using baclofen in AWS, recent reviews of RCTs of baclofen suggest caution in view of poor-quality evidence.[35]

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.