Background information: Alcohol has potentiating effects on the inhibitory GABA neurotransmission system and inhibitory effects on the excitatory glutamate neurotransmission system. Chronic alcohol use causes changes to preserve homeostasis, and when the stimulus is removed, alcohol withdrawal results due to decreased inhibition via the GABA system and increased excitation via the glutamate system. Treatment options for alcohol withdrawal include benzodiazepines such as lorazepam, diazepam, or chlordiazepoxide, gabapentin, and phenobarbital (Buell et al.). Tolerance to alcohol can also result in cross-tolerance to benzodiazepines and to a lesser degree barbiturates. Benzodiazepines and phenobarbital are positive allosteric modulators of GABA, and phenobarbital also blocks glutamate signaling. However, the main benefit of phenobarbital is that in down-regulation of GABA-A receptors alpha subunits that occurs in chronic alcohol use, benzodiazepines cannot bind, whereas phenobarbital binds directly to the ion channel and can function without the GABA-A receptors alpha subunit. Phenobarbital can be used as a monotherapy or in combination with benzodiazepines to treat alcohol withdrawal (Hawa et al.).
REBEL Cast Ep115 – Phenobarbital vs Lorazepam in Alcohol Withdrawal
Clinical question: Is a phenobarbital-based or lorazepam-based alcohol detoxification protocol associated with a lower hospital length of stay compared to Ativan for alcohol withdrawal?
Paper: Hawa F et al. Phenobarbital Versus Lorazepam for Management of Alcohol Withdrawal Syndrome: A Retrospective Cohort Study. Cureus. 2021. PMID: 33728215
What They Did:
- Retrospective cohort study (n=606) of patients admitted to three hospitals in Michigan for alcohol withdrawal between 2016 and 2018
- Aged 18-100 years
- Admitted for alcohol intoxication or withdrawal treated with a CIWA-AR driven protocol
- CIWA-AR = Clinical Institute Withdrawal Assessment for Alcohol Scale, Revised
- Admitted directly to the ICU
- Pregnant patients
- Patients who were transferred
- Patients who received an addiction medicine consultation
- Hospital length of stay
- Alcohol-related 30-day readmission
- All-cause 30-day readmission
- 30-day emergency department visits
- ICU transfers
|Lorazepam (n=543)||Phenobarbital (n=63)||Unadjusted p-value||Propensity-score adjusted p-value|
|Initial CIWA Score (+/- SD)||9.05 (5.87%)||11.02 (7.54%)||0.049||0.859|
|Maximum CIWA Score (+/- SD)||15.90 (7.1%)||26.29 (6.57%)||0.666||0.290|
|Initial BAL (n, %):||0.865||0.906|
|<10||133 (26.55%)||15 (25.86%)|
|10-100||72 (14.37%)||8 (13.79%)|
|101-200||63 (12.57%)||6 (10.34%)|
|201-300||87 (17.37%)||14 (24.14%)|
|301-400||95 (18.96%)||9 (15.52%)|
|>400||51 (10.18%)||6 (10.34%)|
Table 1: Selected clinical characteristics with and without propensity scoring for demographic and clinical variables.
Overall, the group that received phenobarbital had a higher initial CIWA score than the group that received lorazepam. The group that received phenobarbital had a higher maximum CIWA score, but the difference was not significantly significant. The difference in initial blood-alcohol levels were not statistically significant between the groups.
The primary outcome, hospital length of stay, was significantly shorter both before and after propensity score adjustment for the phenobarbital group. The 30-day emergency department readmission rate and all-cause 30-day readmission were significantly lower only when adjusted for propensity scores.
|Lorazepam (n=543)||Phenobarbital (n=63)||Unadjusted p-value||Propensity-score adjusted p-value|
|Alcohol-related 30-day readmission (n, %)||65 (11.97%)||6 (9.52%)||0.715||0.045|
|All-cause 30-day readmission (n, %)||77 (14.18%)||7 (11.11%)||0.635||0.020|
|30-day emergency department visits (n, %)||101 (18.6%)||7 (11.11%)||0.195||0.015|
|ICU transfers (n, %)||38 (7.04%)||2 (3.23%)||0.416||0.114|
|LOS (Days +/-SD)||3.66 (2.32)||2.805 (1.26)||<0.001||<0.001|
Table 2: Primary and secondary outcomes with and without propensity scoring for demographic and clinical variables.
- Although this was not a randomized controlled trial, propensity scoring was used to adjust for variables that may have influenced physician selection of a lorazepam or phenobarbital protocol. The authors appropriately intended for this study to be a pilot observational study and did not complete a prospective randomized controlled trial to test for non-inferiority due to sample size limitations.
- A broad range of variables were included that potentially influences treatment decisions and/or outcomes.
- The protocols used in the study were using symptom-driven medications. Although symptom-triggered protocols require experienced staff and close monitoring, in previous studies the outcomes are associated with lower benzodiazepine doses and shorter lengths of stay in the hospital (Sachdeva et al.).
- More information would have been helpful regarding the statistical methods regarding propensity scores. The study only mentioned that propensity-score weighting was used instead of propensity-score matching (i.e., 1:1 matching using propensity score, which typically discards some patients who do not have matches) (Olmos et al.).
- The group that received phenobarbital had a higher maximum CIWA score than the group that received lorazepam, which could mean that phenobarbital was not as effective as lorazepam, although this is potentially confounded by patients who were transferred to the ICU were excluded from the study. Although CIWA has subjective components, when a patient has a CIWA in the 20s ICU transfer should be considered.
- The study only examined symptom-triggered phenobarbital based on CIWA-Ar scoring, and it is unclear if there was front-loading, as phenobarbital (and benzodiazepines) sometimes are administered. This is beyond the scope of the paper but has been examined in other literature (Rosensen et al.)
- The protocol at this hospital system was that all patients admitted to the ICU were started on a phenobarbital protocol, so the unadjusted score could reflect that sicker patients were started on the phenobarbital protocol based on variables that were not measured in the analysis.
- The study does not address crossover, i.e., did some patients fail benzodiazepine treatment while inpatient and then get started on phenobarbital? Phenobarbital is also used in practice as a benzodiazepine-sparing adjunct in treating alcohol withdrawal.
- The study did not include emergency department treatment in categorizing the patients, e.g., was there any crossover between treatment in the emergency department and treatment after admission?
- Hospital C had 0 patients in the phenobarbital group (58 in the lorazepam group), which leads to concern about selection bias, i.e., were all the patients who the physicians would have started on phenobarbital sent to the ICU and thus excluded from the study?
- The protocols in the hospitals in the study included phenobarbital and lorazepam IV, p.o., and IM, and the physician was able to choose the method of administration. One of the main benefits of phenobarbital is the 30-minute peak serum concertation when administered IV.
- Two randomized clinical trials exist for emergency department patients at this time:
- A randomized, double-blind, placebo-controlled trial of emergency department patients (n=102) with alcohol withdrawal in the emergency department were treated with a loading dose of 10mg/kg phenobarbital versus placebo (Rosenson et al.) resulted in decreased ICU admissions (8% in the phenobarbital loading group versus 25% in the placebo group). Both groups received symptom-triggered lorazepam per institutional protocol. This study differed in that the phenobarbital administration was administered in a symptom-triggered protocol as opposed to a bolus.
- A randomized, blinded clinical trial (Hendey et al.) of symptom-triggered phenobarbital versus symptom-triggered lorazepam in the emergency department (n=44). The phenobarbital arm received 260mg IV followed by 130mg IV PRN with no medications if discharged, and the lorazepam arm received 2mg doses followed by a chlordiazepoxide (Librium) taper if discharged. There was found to be no significant difference in CIWA scores at 48 hours between the two groups (p=0.6). A potential benefit in the phenobarbital group is that no medications were prescribed at discharge.
- This study evaluated phenobarbital in the setting of a symptom-triggered protocol. A symptom-triggered protocol has been demonstrated to have improved outcomes when compared to a fixed taper when using benzodiazepines for alcohol withdrawal. However, a potential downside of symptom-triggered phenobarbital is that phenobarbital is sometimes underdosed. Large randomized clinical trials are warranted to provide further information regarding phenobarbital as an adjunct to benzodiazepine symptom-triggered management, phenobarbital using a loading dose, and phenobarbital monotherapy, among other topics.
“Phenobarbital may be a reasonable alternative to lorazepam in management of AWS patients admitted to general medical units. When taking into account the BZDs-related adverse events including oversedation, encephalopathy, agitation, and increased risk of rebound withdrawal symptoms, phenobarbital may represent a reasonable alternative with a potential for improved outcomes.”
Although not the subject of this study, across multiple medical conditions, lower hospital lengths of stay are associated with lower morbidity and mortality as well as lower healthcare costs. Hospital length of stay is an important metric to take into consideration when starting someone on a protocol for alcohol detoxification in the emergency department. This topic will need to be revisited pending an RCT being completed currently (Filewood et al.) on a loading dose of phenobarbital versus placebo in addition to symptom-triggered benzodiazepine treatment in both study arms.
Clinical Bottom Line: Consider phenobarbital when starting treatment for alcohol withdrawal for patients who meet admission criteria.
- Hawa F, Gilbert L, Gilbert B, Hereford V, Hawa A, Al Hillan A, Weiner M, Albright J, Scheidel C, Al-Sous O. Phenobarbital Versus Lorazepam for Management of Alcohol Withdrawal Syndrome: A Retrospective Cohort Study. Cureus. 2021 Feb 11;13(2):e13282. doi: 10.7759/cureus.13282. PMID: 33728215.
- Buell D, Filewod N, Ailon J, Burns KEA. Practice patterns in treating severe alcohol withdrawal: a multidisciplinary, survey. J Intensive Care Med. 2019;23: 885066619847119.31122170
- Mo Y, Thomas MC, Karras GE Jr. Barbiturates for the treatment of alcohol withdrawal syndrome: A systematic review of clinical trials. J Crit Care. 2016 Apr;32:101-7. doi: 10.1016/j.jcrc.2015.11.022. Epub 2015 Dec 8. PMID: 26795441.
- Filewod N, Hwang S, Turner CJ, Rizvi L, Gray S, Klaiman M, Buell D, Ailon J, Caudarella A, Ginocchio GF, Santos M, Sandhu G, Dewhurst N, Sequeira K, Burns KEA. Phenobarbital for the management of severe acute alcohol withdrawal (the PHENOMANAL trial): a pilot randomized controlled trial. Pilot Feasibility Stud. 2022 Jan 22;8(1):14. doi: 10.1186/s40814-021-00963-4. PMID: 35065662.
- Sachdeva A, Chandra M, Deshpande SN. A comparative study of fixed tapering dose regimen versus symptom-triggered regimen of lorazepam for alcohol detoxification. Alcohol Alcohol. 2014 May-Jun;49(3):287-91. doi: 10.1093/alcalc/agt181. Epub 2014 Jan 8. PMID: 24407777.
- Olmos, Antonio and Govindasamy, Priyalatha (2015) “A Practical Guide for Using Propensity Score Weighting in R,” Practical Assessment, Research, and Evaluation: Vol. 20, Article 13. https://doi.org/10.7275/jjtm-r398.
- Rosenson J et al. Phenobarbital for Acute Alcohol Withdrawal: A Prospective Randomized Double-Blind Placebo-Controlled Study. J Emerg Med 2013. PMID: 22999778.
- Hendey GW, Dery RA, Barnes RL, Snowden B, Mentler P. A prospective, randomized, trial of phenobarbital versus benzodiazepines for acute alcohol withdrawal. Am J Emerg Med. 2011 May;29(4):382-5. doi: 10.1016/j.ajem.2009.10.010. Epub 2010 Mar 25. PMID: 20825805.
For More on this Topic Checkout:
- EmCrit: “Alcohol withdrawal,” Josh Farkas
- Tox and Hound: “Fear and loathing in alcohol withdrawal,” Jeff Lapoint
Post Peer Reviewed By: Salim R. Rezaie, MD (Twitter: @srrezaie)