Experience With an Adult Alcohol Withdrawal Syndrome Practice Guideline in Internal Medicine Patients

Karen M. Stanley, MS; Cathy L. Worrall, PharmD, FAPhA; Shayna L. Lunsford, MS; Kit N. Simpson, Dr.PH; Justin G. Miller, MD; Anne P. Spencer, PharmD


Pharmacotherapy. 2005;25(8):1073-1083. 

In This Article


Pilot patients were evaluated based on clusters of symptoms representing the manifestations of AWS. The newly introduced practice guideline provided a plan for management before the emergence of more severe AWS. The plan involved administration of three different classes of drugs—benzodiazepine, central α-adrenergic agonist, and neuroleptic—based on patient-specific symptoms. Retrospective review revealed that the control patients were treated with nonstandardized approaches that did not include assessment with a validated instrument to accurately identify and categorize AWS symptoms. These patients typically were given a scheduled dose of a benzodiazepine regardless of symptoms. This practice contributed to higher benzodiazepine consumption in the control group. Unnecessary benzodiazepine administration can result in suppressed cognition and decreased mental acuity that may be incorrectly attributed to other medical conditions and may mask pain. These patients may also have an increased risk for falls and decreased ability to participate in patient care activities.

The lower benzodiazepine consumption in pilot patients may have been related to regularly scheduled assessments with the AWS type indicator, symptom-triggered drug administration, and adjunctive therapy with clonidine and haloperidol based on symptoms. The physician practice of treating symptoms of delirium in the control patients with benzodiazepine rather than an adjunctive neuroleptic may also have contributed to the control group's higher benzodiazepine intake. Although identified as being at risk for AWS, six (19%) pilot patients and one (2%) control patient required no drug (p=0.01). Experience using the practice guideline with regular symptom assessment has illustrated the difficulty in predicting which patients will develop AWS. Some patients who reported heavy daily alcohol consumption for an extended period, but no previous AWS, did not exhibit symptoms or require drug therapy.

More than 70% of control patients were discharged from the hospital receiving tapered benzodiazepine therapy. This practice may be dangerous because of the potential for this patient population to combine benzodiazepine and alcohol after discharge. The practice was avoided for the pilot patients, most of whose benzodiazepine therapy was tapered off before discharge without a significant increase in LOS.

No significant adverse events were noted in either the pilot or the control group. The possibility of neuroleptic-induced QTc interval prolongation was a concern that prompted an order for a 12-lead ECG before the start of haloperidol in the pilot group. Seventy-two percent of pilot patients and 28% of controls had a prolonged QTc of 450 msec or greater. Because of the possible complication of torsade de pointes, haloperidol therapy was avoided in pilot patients with a prolonged QTc interval. Every effort should be made to perform an ECG at hospital admission for any patient at risk for AWS. This is especially important for internal medicine patients who tend to have numerous medical problems and a complicated drug regimen that may include other drugs known to prolong the QTc interval.[8]

No between-group differences were found regarding the main outcome measurements. Hospital LOS between groups did not differ significantly, although a trend was seen toward an increased LOS in the pilot group (p=0.07). When patients who received no drug therapy were excluded from this analysis, the trend diminished (p=0.13). As stated earlier, there was a tendency to complete tapered drug therapy in pilot patients before discharge. This practice suggests an emphasis on patient care that views AWS as a significant health risk and resists the current trend to discharge patients as soon as their initial admitting problem has been resolved.

Nonrandomized inclusion of patients in each group may have introduced selection bias to this performance improvement project. A few patients who were eligible for the pilot group may not have been included due to physician preference. The small sample size and resultant power limited analysis of the outcome measurements.

Before initiation of this pilot project, which was endorsed by the interdisciplinary staff of the internal medicine unit, extensive education was provided to the attending and resident physicians, pharmacists, and nursing staff about AWS and the adult AWS practice guideline. Even with this careful planning, however, some barriers to implementation were encountered. Monthly rotation of three teams of internal medicine residents made the provision of continuing resident education regarding use of the practice guideline a challenge. The regular nursing staff of the pilot unit supported the project, but the frequent use of float nurses unfamiliar with the practice guideline contributed to implementation inconsistencies and posed a challenge for nursing education.

Results of this project and the previous project with surgery patients[2] have generated interest in symptom-triggered AWS assessment and treatment throughout our hospital. The plan to establish a consistent method for managing AWS in all adult patient care areas has been embraced by the new hospitalist physician group, whose practice was established after completion of the internal medicine pilot project. A new AWS work group was formed to begin revising the adult AWS practice guideline for implementation in all adult patient care areas. This group consisted of four members of the original AWS management performance improvement team, one hospitalist physician, and three physician consultants specializing in addiction.

After reviewing the results of the two projects, the work group decided to continue using the AWS typology method of assessment and treatment, including the AWS type indicator and the adult AWS flow sheet. The work group suggested minimal revision of the adult AWS practice guideline algorithms. The format of the physician order set (Figures 5 and 6) was changed significantly to make it more user-friendly and consistent with the format of other standard order sets used throughout the hospital. The new orders include the option of scheduled lorazepam dosing for patients who have experienced an AWS seizure associated with their current hospitalization. The instruction section (Figure 4) of the revised adult AWS practice guideline contains information about using the anticonvulsive agent carbamazepine[9,10] to treat AWS symptoms in patients with closed head injuries or other conditions in which benzodiazepines may cause confusion or disinhibition.

Adult alcohol withdrawal syndrome orders (page 1).

Adult alcohol withdrawal syndrome orders (page 2).

The revised practice guideline and physician order set were approved by the hospital forms committee and endorsed by the medical executive committee for hospitalwide implementation. To provide for a consistent method of education for the physicians, nursing staff, pharmacists, and social workers in all adult patient care areas, a four-module video program was developed and made available on the hospital intranet. Completion of a pretest and posttest is required of participants to aid in evaluating the effectiveness of this method of educating staff.

In July 2001, the Food and Drug Administration approved the clinical use of the percent carbohydrate-deficient transferrin (%CDT), a biologic test that detects excessive alcohol consumption.[11,12] The %CDT laboratory test is the only one approved in the United States specifically for detection of heavy drinking. A %CDT of 2.6 or greater warns the clinician that the patient probably has been consuming four or more alcoholic drinks/day for 2 weeks or longer and may be at risk for AWS. This test is part of the revised adult AWS orders and is particularly helpful when information about daily alcohol intake cannot be obtained or the accuracy of the patient report is in question. When the %CDT test result is less than 2.6 and other clinical indicators of heavy drinking have been considered, use of the practice guideline may be discontinued. If the test result is 2.6 or greater, the practice guideline is continued. Test results can also be used by staff providing educational counseling to the patient, to emphasize the negative consequences of excessive drinking and stress the importance of accepting assistance with referral for treatment.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.