Current issues of ACP Journal Club are published in Annals of Internal Medicine


Symptom-triggered chlordiazepoxide was effective for alcohol withdrawal

ACP J Club. 1995 Jan-Feb;122:8. doi:10.7326/ACPJC-1995-122-1-008

Related Content in the Archives
Review: Benzodiazepines are more effective than neuroleptics in reducing delirium and seizures in alcohol withdrawal

Source Citation

Saitz R, Mayo-Smith MF, Roberts MS, et al. Individualized treatment for alcohol withdrawal. A randomized double-blind controlled trial. JAMA. 1994 Aug 17;272:519-23.



To compare the efficacy and efficiency of a symptom-triggered regimen with that of a fixed-schedule approach for administering chlordiazepoxide in patients with the alcohol withdrawal syndrome.


Randomized, double-blind, placebo-controlled trial with 24-hour follow-up.


Detoxification unit in a Veterans Affairs medical center.


101 patients (mean age 47 y, 99% men) who abused or were dependent on alcohol according to the definition of the Diagnostic and Statistical Manual of Mental Disorders, Revised 3d Edition, and were admitted for treatment of alcohol withdrawal. Exclusion criteria were medical or psychiatric illness requiring hospitalization; history of seizures; inability to take oral medication; current use of or withdrawal from opiates, benzodiazepines, barbiturates, clonidine, or β-blockers; and lack of consent to participate. 96 patients (95%) completed the study.


Patients were allocated in blocks of 10 to chlordiazepoxide according to a fixed schedule (50 mg for 4 doses, then 25 mg for 8 doses every 6 hours [ n = 50]) or to chlordiazepoxide as needed on a symptom-triggered basis (25 mg/h to 100 mg/h when a score of 8 was achieved on the Clinical Institute Withdrawal Assessment for Alcohol, revised scale [ n = 51]). Patients in the fixed-schedule group also received the as-needed medication, and patients in the "as-needed" group received placebo on a fixed schedule.

Main outcome measures

Duration of medication from time of admission to last dose and total amount of chlordiazepoxide administered.

Main results

Analysis was done on an intention-to-treat basis. Median medication duration was shorter and total amount of chlordiazepoxide administered was less in the symptom-triggered group than in the fixed-dose group (median medication duration 9 h [interquartile range 0 to 43 h] vs 68 h [interquartile range 64 to 73 h]; median total dosages 100 mg [interquartile range 0 to 400 mg] vs 425 mg [interquartile range 350 to 750 mg]; for both comparisons P < 0.001). Groups did not differ in number and amount of as-needed doses administered, incidence of delirium tremens, hallucinations, seizures, lethargy, discharge from the hospital against medical advice, readmission within 30 days, or rate of entry into a rehabilitation program.


Patients who were treated for alcohol withdrawal with symptom-triggered medication required less chlordiazepoxide and recovered faster than did patients who received fixed-schedule medication.

Source of funding: Roche Laboratories.

For article reprint: Dr. R. Saitz, Section of General Internal Medicine, 91 East Concord Street, #200, Boston Medical Center & Boston University School of Medicine, Boston, MA 02118, USA. FAX 617-414-4676.


Overworked interns love the "Librium protocol" for alcohol withdrawal; they simply order a tapering dose of chlordiazepoxide and everybody gets better. This study suggests a different approach: Tailor the dosage to the symptoms and everybody still gets better, only quicker and with less drug. The study by Saitz and colleagues has no serious design flaws (1), but the study raises several questions about symptom-triggered treatment.

Are the findings generalizable to all alcoholic persons? Most study participants were men in their late 40s who had no concomitant medical illness. Would similar treatment be equally effective in other groups, such as younger alcoholic women or patients with concomitant medical illness?

How much does it cost? Patients required intensive monitoring by nurses with special training in the use of the Clinical Institute Withdrawal Assessment for Alcohol, revised scale. Successful emulation could require a large investment in a trained and committed nursing staff.

Could it reduce mortality? Probably not; death caused by alcohol withdrawal is now rare. Could it reduce morbidity? Probably not; alcoholics are cross-tolerant to benzodiazepines and are not dangerously sedated by high doses. Could it reduce the subjective discomfort of alcohol withdrawal? Probably not; either fixed or symptom-triggered doses of chlordiazepoxide would control symptoms well.

So why bother with symptom-triggered chlordiazepoxide? Because it could probably reduce the cost of medical care and expose patients to less potential risk by administering less medication. A reduction of 59 hours in mean duration of treatment could translate into major savings.

Symptom-triggered chlordiazepoxide is an elegant refinement which likely will require additional nursing and physician training to implement. A care-path approach to monitoring and management should be evaluated. Until then, general medical and surgical units could benefit from education that: a) Librium (chlordiazepoxide) is not the best benzodiazepine, and b) high doses given early in moderate-to-severe withdrawal present drawn-out treatment course and may present sedation-induced like complications (2, 3).

Michael Phillips, MD
St. Vincent's Medical Center of RichmondStaten Island, New York, USA


1. Fuller RK, Gordis E. Refining the treatment of alcohol withdrawal [Editorial]. JAMA. 1994;272:557-8.

2. Mayo-Smith MF, for the American Society of Addiction Medicine Working Group on Pharmacological Management of Alcohol Withdrawal. Pharmacological management of alcohol withdrawal. A meta-analysis and evidence-based practice guideline. JAMA. 1997;278:144-51.

3. Holbrook AM, Crowther R, Lotter A, Cheng C, King D. Diagnosis and management of acute alcohol withdrawal. CMAJ. 1999;160:675-80.