No evidence-based guidelines for the management of alcohol withdrawal have been developed. Mayo-Smith, representing the American Society of Addiction Medicine Working Group on Pharmacological Management of Alcohol Withdrawal, reports on a meta-analysis performed to determine which agents are suitable for treating this condition and to develop an evidence-based guideline for the pharmacologic management of alcohol withdrawal.

The following outcomes were examined: severity of alcohol withdrawal, occurrence of delirium, withdrawal seizures and cost. The group used a published system to classify the strength of the recommendations. The group reviewed 134 articles on alcohol withdrawal. Sixty-five of the studies were prospective, controlled clinical trials, and 42 different medications were evaluated. The recommendations developed by the Working Group were graded according to the level of the evidence.

Six prospective trials demonstrated that benzodiazepines are more effective than placebo in decreasing the signs and symptoms associated with alcohol withdrawal. Seizures were significantly reduced in patients who received benzodiazepines. Long-acting benzodiazepines were somewhat more effective than short-acting agents in preventing seizures. The data suggested that long-acting agents could be associated with increased sedation in some patients. The Working Group recommends the use of benzodiazepines for relief of alcohol withdrawal (grade A recommendation).

Two prospective, randomized controlled trials have shown that use of symptom-triggered treatment, based on a symptom score generated by using the revised Clinical Institute Withdrawal Assessment-Alcohol, is as effective as fixed-dose therapy. Moreover, this approach results in significantly less medication being given during the course of the withdrawal.

A review of five studies of the use of beta-adrenergic antagonists revealed evidence that these agents reduce the autonomic manifestations of withdrawal. Whether these agents increase or reduce seizures during withdrawal could not be determined. One study found that the incidence of delirium was increased with use of propranolol. The Working Group does not recommend beta-adrenergic antagonists as monotherapy (grade B recommendation). These agents may be used in conjunction with benzodiazepines.

Studies have consistently demonstrated that centrally acting alpha-adrenergic agonists, such as clonidine, are effective in relieving symptoms of mild to moderate withdrawal. It is not known what effect clonidine may have on the incidence of delirium or seizures. These agents are not recommended as monotherapy (grade B recommendation).

Carbamazepine, used in Europe for alcohol withdrawal, has been shown in clinical studies to be equally as effective as barbital and oxazepam in patients with mild or moderate withdrawal symptoms. Carbamazepine is not recommended as monotherapy (grade B recommendation).

Studies indicate that phenothiazines are less effective than benzodiazepines in preventing seizures and delirium. The Working Group does not recommend neuroleptics as monotherapy because they do not reduce delirium and may increase seizures (grade A recommendation).

Low magnesium levels are known to occur during alcohol withdrawal and then return to normal. A double-blind placebo-controlled trial showed that intramuscular administration of magnesium as an adjunct to benzodiazepine therapy did not decrease the severity of withdrawal symptoms. Routine parenteral administration of magnesium is not recommended (grade B recommendation).

One large trial showed that use of thiamine did not reduce delirium or seizures. But the Working Group states that persons with alcohol dependence frequently have thiamine deficiency. Administration of thiamine is recommended at the time of initial evaluation in all patients with alcohol dependence (grade C recommendation).

The Working Group recommends individualized treatment and use of structured assessment scales for monitoring withdrawal symptoms (grade A recommendation). Such an approach allows titration of doses to meet individual needs. However, the Working Group recognizes that structured assessment and an individualized approach may not always be feasible. A fixed-dose schedule is an acceptable alternative if additional medication is administered when symptoms are not controlled with scheduled doses. The accompanying table gives examples of treatment regimens that meet the recommendations.