Care for hospitalized patients with unhealthy alcohol use: a narrative review

In patients with unhealthy alcohol use, characterization of severity requires an assessment of recent alcohol-related problems, which will often satisfy the criteria for an alcohol use disorder in hospitalized patients [16]. A number of screening instruments for alcohol use disorder have been developed, and some are listed in Table 2[21, 40].

Importantly, as for any ordinal screening instruments, different cutoff scores will result in a tradeoff between sensitivity and specificity. In addition, relatively low cutoff scores on an instrument that assesses both the magnitude of drinking and alcohol-related consequences can be used to screen for unhealthy alcohol use and alcohol use disorders (e.g., as with the AUDIT [41]). Ultimately however, a diagnosis of an alcohol use disorder should be based on an assessment of established criteria.

The American Psychiatric Association (Diagnostic and Statistical Manual/DSM) and World Health Organization (International Classification of Disease/ICD) have developed criteria for a diagnosis of current alcohol dependence that substantially overlap, and require the presence of at least 3 of the diagnostic criteria during the past 12 months (or co-occurring more distantly for alcohol dependence in remission). The current DSM-IV criteria for dependence and abuse are listed in Table 3. Importantly, due to evidence that alcohol problems exist on a continuum rather than sharply defined categories, the upcoming revision of the DSM in 2013 (DSM-V) will controversially do away with the terms “abuse” and “dependence”, and will have only “alcohol use disorder” [42, 43]. Individuals satisfying 2 or 3 of the 11 criteria will be classified as having “moderate” alcohol use disorder, and those meeting 4 or more criteria will be classified with “severe” alcohol use disorder. This DSM-V reclassification may cause some initial confusion, but the concept of a single disorder that varies in severity will likely be a welcome change for many hospital-based clinicians.

After initial screening and clinical assessment, patients with unhealthy alcohol use will have been categorized as having or not having an alcohol use disorder. In the latter, the clinical effort should be aimed at drinking reduction to levels that are unlikely to contribute to health problems. To accomplish this, the evidence, while mixed, generally favors a modest effect of brief interventions during hospitalization. A meta-analysis [44] of 4 randomized controlled trials (one in trauma patients did not exclude dependence, two excluded dependence, and one excluded patients with alcohol-related physical problems) concluded that brief intervention for unhealthy drinkers reduced alcohol consumption by an average of approximately 5 drinks/week 6 months post-intervention. Exclusion of the trauma-based trial due to unblinded outcome assessment reduced the magnitude and statistical significance of this estimate. Reported alcohol use at one year was synthesized for 4 other studies (3 with variable exclusion of patients with alcohol use disorders) and, although there was roughly a 2 drink per week reduction in the intervention groups, this was not statistically significant relative to controls. Importantly, while the evidence base for outcomes other than alcohol use is relatively under-developed, this meta-analysis found evidence that brief intervention was associated with decreased mortality at one year. In addition, a US study of trauma patients reported a 47% reduction in recurrent trauma over the 3 years following receipt of brief intervention [45].

Overall, the evidence tends to favor brief intervention for hospitalized patients with less severe unhealthy alcohol use. A number of different strategies have been studied, but brief intervention in the hospital should include at a minimum feedback about alcohol use (including associations with conditions the patient may have such as hypertension), advice to reduce consumption to safer levels, an explanation of why these limits are recommended, a non-confrontational inquiry to determine a patient’s interest in reducing their drinking, and determination of a plan to achieve drinking goals [46]. This can be delivered by a variety of health care personnel [44, 47], which can lower the cost of brief intervention.

The potential for acute complications renders the detection of alcohol use disorders and associated heavy drinking particularly important at the time of hospitalization. A number of informative reviews are available for guiding care [7, 48‐50], and only a brief summary is provided here (Table 4). In patients with prior severe withdrawal or unstable medical disease, clinicians may wish to institute preventive measures for acute alcohol withdrawal with fixed dose benzodiazepines (e.g., chlordiazepoxide 50 mg every 6 hours for 24 hours followed by 25 mg every 6 hours for 48 hours) [48]. Patients must still be monitored for over-sedation or inadequate dosing. Active withdrawal can be managed with symptom-triggered or scheduled benzodiazepine dosing, with the former resulting in decreased benzodiazepine use and shorter treatment duration [51]. Several potential regimens are listed in Table 5[49, 51‐53], and use of a shorter-acting benzodiazepine (e.g., lorazepam) will require a taper to prevent recurrent symptoms. Monitoring and dosing must be guided by a validated symptom scale such as the Clinical Institute Withdrawal Assessment for Alcohol-revised (CIWA-Ar) or somewhat shorter CIWA-AD [54‐56] administered by trained personnel. Inappropriate selection of patients for symptom-triggered therapy has been reported, with the major reasons being inability to communicate (and thus assess all withdrawal symptoms), lack of heavy drinking in the days preceding hospitalization, and other causes of delirium [57]. Thus an accurate history is required and other causes of delirium must always be considered. For example, chronic heavy drinking suppresses the immune system [58, 59], and clinicians must remain vigilant for infections such as pneumonia or sepsis. Other causes of delirium (e.g., hypoxia, stroke, effects of drugs other than alcohol, etc.) may also need to be ruled out depending on the clinical circumstances.

Initial treatment of severe withdrawal may require intravenous benzodiazepines (e.g., 2 to 4 mg of lorazepam or 5 to 10 mg of diazepam) and additional dosing as needed to control symptoms. Usually benzodiazepines alone are sufficient, and a strong evidence base supports their use [60]. The rare patient with severe withdrawal who fails benzodiazepine treatment will require intensive monitoring and treatment with barbiturates or propofol [49]. Phenobarbital is most frequently used in this circumstance, with 30 mg having roughly equivalent effects on withdrawal symptoms as 2 mg lorazepam, 25 mg chlordiazepoxide, or 10 mg diazepam [53]. Some centers with expertise in alcohol withdrawal treatment use anticonvulsants (e.g., carbamazepine, gabapentin) as a primary prophylaxis or treatment rather than benzodiazepines, but the evidence base for justifying clinical recommendations is currently under-developed. In particular, anticonvulsants have not been shown to reduce the incidence of alcohol withdrawal seizure or delirium tremens [61]. The use of intravenous alcohol infusions continues in some locations [62], mainly in surgical specialties, due to some surgeons’ familiarity and comfort with this treatment. This approach has been discouraged due to potential toxicity (remembering that first-pass metabolism can be substantial in chronic heavy drinkers and is bypassed by this route), and evidence shows that this approach is at best no more efficacious for prophylaxis than benzodiazepines [63], and may have a high failure rate [64, 65].

Chronic heavy drinking (e.g., averaging at least 50 to 60 grams of alcohol per day or usually exceeding daily limits) increases perioperative morbidity due to increased risks for acute withdrawal syndrome, pneumonia and ARDS, wound infections, bleeding, myocardial dysfunction, and enhanced stress responses [59, 66]. Thus pre-operative detection should trigger efforts at detoxification and achieving several weeks of abstinence before elective procedures. For the more likely emergent procedures in hospitalized patients, parenteral thiamine, electrolyte replacement, and close monitoring and treatment for post-operative withdrawal are important components of care, and causes of delirium other than withdrawal should be considered as they would be for patients who are not heavy drinkers. Prevention of acute withdrawal with benzodiazepines is an important consideration in this population [67], and perioperative morphine may reduce post-operative pneumonia by ameliorating neuroendocrine and immune imbalances induced by chronic heavy drinking and peri-operative abstinence [68]. Much of the perioperative management will naturally be directed by the anesthesiologist and surgeon, and any consulting physicians should make certain that these providers are aware of the patient’s heavy drinking. Of note, inappropriate use of symptom-triggered therapy for withdrawal has been observed in surgical patients [57]. It is equally important to communicate to all providers when a known heavy drinker has passed the time when initial manifestations of acute alcohol withdrawal would be expected (within 2 to 3 days of the last drink).

Some research suggests hospitalization has the potential to facilitate treatment for alcohol use disorders by increasing recognition of the need to reduce drinking and intent to do so [69‐72]. Despite this, brief interventions are unlikely to be effective at reducing drinking for hospitalized patients with alcohol dependence [73], but factors including avoidance of heavily drinking friends and engaging in alcohol treatment after hospitalization predict drinking reduction [71]. Thus referral for outpatient treatment should be pursued and integrated into brief intervention (i.e., SBIRT), but, even among research participants, overall compliance with referral is modest [73, 74]. Brief intervention may be more effective in enhancing referral for women and younger adults [75]. Some evidence suggests that peer-involvement from volunteer Alcoholics Anonymous members as a supplement to brief intervention can increase treatment initiation and abstinence following hospitalization [76].

Following hospital discharge, integrated outpatient treatment (i.e., post-discharge care for alcoholism and medical conditions with the same healthcare provider) may be effective under ideal conditions [77, 78]. A related approach involves a chronic disease management strategy in primary care with as-needed addiction specialist support, which may improve the quality of outpatient care and facilitate recovery [79, 80]. Thus it is reasonable to refer patients to this mode of primary care if it exists locally.

Other potential interventions prior to discharge must be extrapolated from studies in treatment-seeking alcohol-dependent populations or treatment strategies implemented for other chronic medical and psychiatric conditions. Taking advantage of findings in the alcoholism pharmacotherapy field that show modest efficacy for several medications coupled with brief counseling strategies [81, 82], an approach for alcohol dependent inpatients might include brief interventions and initiation of pharmacotherapy during the hospitalization. Several potential medications are listed in Table 6, together with a selection of their generally small effect sizes [83‐87]. With the exception of disulfiram, medications for relapse prevention target brain receptors thought to be involved in relapse, including subsets of opioidergic, serotenergic, glutamatergic, and dopaminergic receptors [88], and complete abstinence may not be realized. As in the treatment of other conditions such as hypertension, reduced post-hospitalization drinking due to medication use may be considered a successful albeit partial response, even if the ultimate goal is abstinence. However, with the possible exception of transitional clinics following hospital discharge, continuation of medication will mainly be in the hands of ambulatory clinicians.