CHALLENGES IN PREDICTION, DIAGNOSIS, AND TREATMENT OF ALCOHOL WITHDRAWAL IN MEDICALLY ILL HOSPITALIZED PATIENTS: A TEACHABLE MOMENT

Story From the Front Lines

An older man was admitted to the hospital for planned bowel resection. He reported drinking 8 to 10 ounces of whiskey daily for many years. During previous hospitalizations, he had no documented alcohol withdrawal symptoms or seizures. His last drink was 3 days prior to admission.

Over the first week postoperatively, he had no symptoms of alcohol withdrawal. On postadmission day 6 he developed an anastomotic leak, requiring urgent reoperation and antibiotics. On postadmission day 9, he became acutely disoriented and inattentive, with new abdominal tenderness. Imaging showed large intra-abdominal abscesses, and percutaneous drains were placed. His heart rate and blood pressure remained normal.

On postadmission day 13 he became more agitated and confused. Suspecting alcohol withdrawal, his clinicians implemented an institutional alcohol withdrawal treatment protocol incorporating the Clinical Institute Withdrawal Assessment for Alcohol (CIWA) scale. Owing to anxiety, agitation, hallucinations, and disorientation, he scored high on CIWA. He was given 10 mg of diazepam, with improvement in his symptoms requiring no further medication. The next day his confusion and agitation worsened, again scoring high on CIWA and triggering another 10-mg dose of diazepam. He became sedated, requiring transfer to a monitored unit.

Teachable Moment

Alcohol withdrawal is common in hospitalized patients and is associated with considerable morbidity and mortality.¹ Alcohol withdrawal comprises a wide spectrum of severity, from mild symptoms of anxiety, tremor, and diaphoresis, to severe manifestations including seizure and delirium tremens. Among patients with alcohol withdrawal at risk for severe manifestations, standard treatment involves benzodiazepines, commonly dosed using symptom-triggered protocols incorporating CIWA.^1,2

Accurately diagnosing alcohol withdrawal is challenging in medically ill hospitalized patients because multiple etiologies cause similar constellations of symptoms. The CIWA scale was developed for the management of alcohol withdrawal without concurrent acute medical illness, and has not been validated in medically ill inpatients.³Six of the 10 CIWA questions rely on patients’ specific responses regarding symptoms, and patients may not be able to answer owing to delirium, language barriers, or intubation. As with this patient, hyperactive delirium alone may lead to high CIWA scores that trigger benzodiazepine dosing.

Accordingly, CIWA protocols are overused in acute care hospitals, causing considerable harm. Unnecessary benzodiazepine use can cause sedation, falls, respiratory depression, aspiration, delirium, and prolonged hospitalization. Risk is higher when combined with other sedatives, including postoperative analgesia, and among older patients.^2,3 In a 2018 study³ of hospitalized patients placed on a CIWA protocol, 20% of patients had no documentation of recent alcohol use, 57% had zero or 1 documented risk factor for alcohol withdrawal, and 14% were unable to communicate when CIWA was ordered. Fifteen percent of these patients had benzodiazepine-associated adverse events.³

In this case, poor understanding of the natural history of alcohol withdrawal led the inpatient team to misdiagnose alcohol withdrawal and inappropriately implement a CIWA protocol. Alcohol withdrawal symptoms generally begin 6 to 12 hours after the last drink and peak at 24 to 72 hours.² Although withdrawal-related seizures can occur at any time during this course, delirium tremens tends to occur 3 to 7 days after the last drink, and only after progressing through more mild symptoms.² For this patient, acute onset of severe alcohol withdrawal more than 1 week after the last drink would be highly unusual, especially without first demonstrating mild withdrawal.

Clinical tools can assist in stratifying patients’ risk of developing severe alcohol withdrawal, potentially allowing low-risk patients to avoid benzodiazepines altogether. The most promising tool is the Prediction of Alcohol Withdrawal Severity Scale (PAWSS), which was developed for medically ill inpatients and incorporates information on history of withdrawal symptoms, addiction treatment, and high-risk drinking practices.¹ A PAWSS score of 4 or higher is reported to have 93.1% sensitivity and 99.5% specificity for predicting severe withdrawal.¹ However, PAWSS should be used with caution. Patients may be unable to answer questions owing to feelings of stigma or if they have not previously abstained from alcohol long enough to have gone through withdrawal. The PAWSS has not yet been independently validated, and the original study outcome for severe alcohol withdrawal (defined as CIWA score of ≥15, physician diagnosis, or administration of benzodiazepines or barbiturates) may not be reliable or patient centered. Despite these limitations, PAWSS can be a helpful tool incorporated into preoperative assessments or at hospital admission.

Patients with heavy alcohol intake should also be offered preoperative interventions to decrease alcohol use, which have been shown to reduce postoperative complications.⁴ Naltrexone and acamprosate are associated with significant reductions in alcohol consumption, and could be considered preoperatively or before hospital discharge.⁵

For this patient, a better understanding of the natural history of alcohol withdrawal and appreciation for the limitations and potential harms of CIWA protocols could have prevented unnecessary benzodiazepine administration, sedation, and prolonged hospitalization. Clinicians should consider a broad differential diagnosis for delirium in hospitalized patients with a history of alcohol use, and carefully consider the appropriateness of CIWA protocols before initiating one.