Alcohol Withdrawal

Nicole Bouchard

BACKGROUND

Chronic alcoholism and alcohol withdrawal syndrome (AWS) are serious disorders that affect millions of people worldwide. AWS is frequently encountered in hospitalized patients and contributes significantly to patient morbidity and mortality. AWS also imposes a considerable financial burden on hospitals, as these patient visits are often prolonged and not fully reimbursed. In spite of the high volume of medical admissions for moderate to severe AWS, relatively little evidence-based literature addresses its management. This is particularly true in the intensive care unit (ICU), where management varies significantly among institutions.

PATHOPHYSIOLOGY

When the central nervous system is exposed to long-term ethanol use, compensatory changes occur to counter ethanol’s depressant effects on the inhibitory centers of the cerebral cortex. The net effect of these compensatory changes is to restore cerebral homeostasis despite the near-constant presence of ethanol. Tolerance to ethanol is an example of such a compensatory changes change. Alcohol potentiates γ-aminobutyric acid (GABA_A) signaling by increasing the GABA_A chloride channel opening. With chronic ethanol use, the persistently stimulated inhibitory GABA_A receptors become down-regulated and less sensitive to ethanol.^1–5 Ethanol also inhibits glutamate, the excitatory neurotransmitter that binds to and activates the excitatory N-methyl- D-aspartate (NMDA) receptor. In chronic ethanol use, NMDA receptor systems are up-regulated and become more sensitive to glutamate.^6–9 Ethanol use is also associated with increased brain dopamine; this is thought to be a contributing factor to some of its acutely pleasurable effects and addictive qualities.^9–11

When ethanol exposure is abruptly terminated, there is a loss of homeostasis, and these neurotransmitter systems become imbalanced. This absence of ethanol’s depressant effects on the already desensitized, down-regulated GABA system combined with the enhanced NMDA excitatory system and dysregulation of the dopaminergic system is primarily responsible for the development of unopposed CNS excitation and the hyperexcited state associated with AWS.^12–14

Chronic ethanol use is also thought to desensitize α₂ receptors. The increased dopamine seen with ethanol use is metabolized to norepinephrine (NE) by dopamine-β-hydroxylase, resulting in an increase in available NE. Chronically impaired α₂ receptor activity in the face of increased NE results in adrenergic receptor up-regulation and adrenergic hypersensitivity. These effects may explain the increased sympathetic nervous system activity observed in alcohol withdrawal.^15–17

HISTORY AND PHYSICAL EXAM

Initial steps in the successful treatment of adult patients at risk of or actively experiencing AWS include early recognition of AWS, appropriate patient disposition, early initiation of symptom-triggered therapy (STT),^18–25 and/or front-loading with benzodiazepines.²⁶ Early recognition of AWS requires detailed history taking and an awareness of the stages of withdrawal.

Clinical manifestations of alcohol withdrawal occur along a spectrum and often coexist with other pathophysiologic states. Psychological symptoms range from mild anxiety, insomnia, craving, irritability, and labile emotions to significant agitation, trouble thinking clearly, and altered mental status (AMS) or frank delirium. Physical symptoms include headache, diaphoresis, nausea and vomiting, tachycardia, hypertension, tremor, tongue fasciculations, hyperthermia, and seizures.

The typical withdrawal timeline after a patient’s last drink is as follows:

• 6 to 12 hours: acute tremulousness (“the shakes”), insomnia, headache
  
• 12 to 24 hours: visual and/or auditory hallucinations, also known as alcohol hallucinosis
  
• 6 to 48 hours (or earlier with rapidly declining blood alcohol level [BAL]): seizures (“rum fits”) (typically several, usually short and self-terminating)
  
• 72 to 96 hours: delirium tremens (DT or “DTs”) characterized by AMS, delirium, hyperdynamic circulation, hyperventilation, and hyperthermia. DTs may persist for up to 2 weeks (1 week is more common)

Although AWS tends to occur in a temporal progression, there is no fixed sequence. Completion of alcohol withdrawal typically lasts 4 to 7 days; however, significant withdrawal can last up to 2 weeks.²⁷ Historically, approximately 5% of all patients with AWS will progress to DTs²⁸; this number jumps to over 30% for patients who experience withdrawal seizures.²⁹ In modern times, advances in medical therapy and ICU care have decreased the mortality from AWS with DT from 37% to 5%.^30–33

DIFFERENTIAL DIAGNOSIS

It is important to consider the complete clinical picture when caring for a patient with AWS. Not only does AWS frequently exist as a result of or in parallel with another pathophysiologic process, but its signs and symptoms can also be masked or complicated by other processes (Table 52.1). Occasionally, symptoms such as isolated tachycardia are wrongly attributed to AWS in the ethanol-dependent patient. Similarly, confounding AWS with other diagnoses (encephalopathy, psychosis, delirium from another cause, CNS injury, infection) can lead to wrong diagnoses, inappropriate sedation, and withheld treatments.³⁴

DIAGNOSTIC EVALUATION

A history of physiologic tolerance to ethanol from years of heavy use, coupled with recent cessation or reduction in ethanol intake, places patients at risk for developing AWS, severe AWS, and/or DT. Risk factors for the development of severe AWS and/or DT include duration of ethanol abuse; quantity of ethanol consumed; history of repeated episodes of AWS, DTs, or seizures; withdrawal symptoms with a positive BAL; mild intoxication with BAL >300 to 400 mg/dL; and comorbid infections.^27,35–38 Hypokalemia, thrombocytopenia, and presence of structural brain lesions are additional independent predictors of more severe withdrawal.³⁶ Certain populations may also be predisposed to severe AWS, and there may be a genetic or racial component (Whites appear more at risk than Blacks for severe AWS).^33,39 When admitted to the hospital, high-risk patients should undergo careful risk assessment, followed by management using an established clinical care pathway. Unrecognized or undertreated withdrawal may progress to more severe withdrawal and exacerbate future episodes of AWS, underscoring the importance of early recognition and treatment.

The most commonly used and highly validated scoring systems for the assessment and STT of AWS are the Clinical Institute Withdrawal Assessment—Alcohol (CIWA-Ar, Ar: Alcohol-revised, and, where available, a new version CIWA-Ad) (Table 52.2).^{22,24,26,27,34,40–44} Care must be taken to apply the CIWA score correctly to at-risk patients to avoid erroneous attribution of symptoms to patients not in AWS.³⁴ Some centers use abridged, institution-specific assessment tools; these may be convenient at a local hospital level but have not been extensively validated.

The strength of the CIWA tools is the ability to detect AWS at an early stage, when treatment will be maximally beneficial. The CIWA-Ar/Ad score becomes more difficult to evaluate in patients with very high scores, DTs, or benzodiazepine-resistant AWS. In these cases, the goal of STT is to maintain light sedation, and an agitation/sedation score such as the Richmond Agitation Sedation Score (RASS) may be a more useful assessment tool (Table 52.3).

Initial patient classification into mild, moderate, or severe AWD should be based on CIWA-Ar score and clinical picture. At initiation of treatment, CIWA-Ar/Ad/RASS (the latter for ICU patients) should again be assessed. The result of these scores, along with clinical picture, should then guide medication dosing; scoring should be repeated as indicated by the degree of AWS and frequency of medication dosing. No single guideline exists for managing this process; a sample protocol is provided in Figures 52.1–52.3.

MANAGEMENT GUIDELINES

Moderate to severe AWS is most commonly treated with benzodiazepines. Ideal pharmacologic management is tailored to the individual patient and controls hyperadrenergic symptoms, anxiety, agitation, and delirium while minimizing adverse effects. Drug selection and dosing strategy should be informed by the desired pharmacodynamic and pharmacokinetic properties of the drugs, as well as by individual patient characteristics (existing symptoms, underlying illnesses, and/or past episodes of AWS). Ongoing subjective assessments using CIWA-Ar/Ad and RASS scores are essential.

When compared to standing doses or continuous infusions of benzodiazepines, well-executed STT with benzodiazepines plus adjunctive agents can decrease total benzodiazepine dose, incidence of intubation and admission to the ICU, and hospital and ICU length of stay (LOS).^{18,20–25,42,45–47} In one study, well-executed STT was shown to be comparable to front-loading, an approach in which patients are typically titrated to lid lag, a low CIWA score (<8), or a state of calm.²⁶ In front-loading, the use of long-acting benzodiazepines with active metabolites can lead to heavier sedation but confers the advantage of autotapering. In the emergency department (ED), front-loading followed by SST is often used in the initial control phase for patients in AWS.⁴⁸ Generally speaking, institutional clinical guidelines and individual patient parameters should guide the practitioner’s choice of approach.

Benzodiazepines

The five most commonly used benzodiazepines are diazepam, chlordiazepoxide, lorazepam, clonazepam, and oxazepam. There are several important considerations when choosing a pharmacologic agent for a patient in AWD (Table 52.4).

Chlordiazepoxide and diazepam have a longer duration of action due to the presence of active metabolites. This may decrease the rate of breakthrough symptoms and have an autotapering effect.^49,50 Parenteral diazepam has a short onset and is preferred for rapid titration in severe cases; doses of 5 to 20 mg can be given every 5 to 10 minutes. Higher bolus doses can be used if the patient demonstrates tolerance. Oral chlordiazepoxide can be used effectively in moderate withdrawal and can be rapidly titrated at doses of 50 to 100 mg/hour. Patients may also be managed with aggressive oral dosing of chlordiazepoxide. The experience of recent national drug shortages established the efficacy of this approach in patients previously thought to require parenteral medications.^51–53

Lorazepam, which has a longer time to peak effect (15 to 20 minutes), may lead to iatrogenic oversedation if titrated too rapidly. Lorazepam can be titrated in doses of 1 to 4 mg every 15 to 30 minutes. If higher doses are used, a longer dosing interval is preferred. Due to lorazepam’s lack of active metabolites, nondependence on renal or hepatic mechanisms for clearance and predictable t_1/2, it is preferred over diazepam and chlordiazepoxide for use in patients with COPD, hepatic dysfunction (INR > 1.6), renal dysfunction (Cr_Cl < 30 mL/min, S_Cr > 2 mg/dL), or age > 65 years.⁵⁴ Prolonged lorazepam infusions carry a risk of toxicity from propylene glycol (the diluent in lorazepam infusions) with associated metabolic acidosis and renal failure. Lorazepam is also incompatible with numerous other infusions.

In general, it is recommended to avoid using different benzodiazepines together (e.g., oral chlordiazepoxide and intravenous benzodiazepines), except during the initial control phase when a patient is progressing from mild to more severe symptoms, or to provide a basis for autotapering (when chlordiazepoxide is added). The deliberate and measured use of different classes of medications based on mechanism of action and synergy can yield good results. Haphazard polypharmacy with benzodiazepines and other sedatives in difficult-to-control patients increases the chances of unexpected synergy and oversedation.

Patients with resistant alcohol withdrawal (defined below) require ICU-level monitoring and a more aggressive pharmacotherapeutic approach. Such patients often require adjunctive therapies like phenobarbital or propofol (see below).

Resistant Alcohol Withdrawal (RAW) is defined as the following:

Barbiturates

Barbiturates are GABA_A agonists and have been employed as monotherapy and, more commonly, as adjuncts in severe DT or RAW.^{20,51,52,55–59} The most studied barbiturate in the setting of AWS is phenobarbital, which is typically initiated in a monitored setting when high-dose benzodiazepines have failed to control symptoms. A standard starting dose of phenobarbital of 30 mg can be used, titrated to effect every 30 minutes. As the onset of barbiturates’ effect can be delayed, it is prudent to wait approximately 30 minutes before adding an additional agent. The long half-life of barbiturates also confers a beneficial self-tapering effect. One study showed decreased ICU admissions with phenobarbital given as a single dose of 19 mg/kg (vs. placebo) early in the course of treatment.⁵⁸ It is unclear, however, whether this outcome demonstrates superiority of loading with a barbiturate, or the benefit of a front-loaded approach in general paired with a less aggressive institutional protocol. In times of nationwide benzodiazepine shortages, the use of barbiturates in the treatment of AWS should be considered a reasonable therapeutic option.^51,52

Propofol

Propofol is a GABA_A agonist and an NMDA antagonist. It is highly effective at controlling severe AWS symptoms and is generally reserved for cases refractory to high-dose benzodiazepines and barbiturates. Intubation is generally recommended because of propofol’s potency and the narrow window between an effective and toxic dose. Propofol’s short half-life makes it unsuitable as a primary or exclusive agent. Intubated patients on propofol infusions can be started on a benzodiazepine or chlordiazepoxide via nasogastric tube, in an effort to wean from propofol and establish a base of long-acting benzodiazepines that permit autotapering (see Tapering section).^60,61 Propofol infusions are susceptible to tachyphylaxis and should only be used in a highly monitored setting.⁶²

Antipsychotics

In historical studies of AWS, benzodiazepines have been shown far superior to monotherapy with antipsychotics in controlling symptoms and preventing seizures or DT’s.^41,63,64 Bolus therapy with the antipsychotic haloperidol, given with clonidine and benzodiazepines, has been described as a successful combination when compared to continuous monotherapy infusions.^24,41,42 It is not clear whether the strength of the treatment reported was the result of attentive STT/bolus therapy, or the combination of the medications used, or both. As a general rule, while there may be a role for antipsychotics in patients in AWS with comorbid psychiatric disorders or pronounced hallucinations, monotherapy with these agents should be avoided.

Anticonvulsants

Some studies have suggested that benzodiazepine use in the treatment of AWS can lead to more craving—of both ethanol and benzodiazepines—and that alternative, non–benzodiazepine-based regimens—such as anticonvulsants—may decrease craving and relapse after detoxification.^65–69 Since these studies considered only short-term outcomes, it is difficult to know whether the effect on craving continues beyond the initial period following detoxification.

The anticonvulsant topiramate is an antiglutaminergic and GABA-enhancing drug, and studies have shown it to attenuate AWS in rodents.⁷⁰ Human studies have shown a benefit over placebo for preventing AWS and suggest a possible role for treating ethanol dependence.^9,10,70 Valproic acid,^71–73 gabapentin,^67,74 and carbamazepine^66,71,75 have also been evaluated in the treatment of AWS, primarily in mild to moderate cases in the settings of inpatient and outpatient detoxification units. Studies suggest them to be inadequate as sole agents but potentially more useful in combination therapy.^33,71,76

A 2010 Cochrane review of anticonvulsants for AWS concluded that there was insufficient evidence in their favor, excepting the use of carbamazepine, which may outperform benzodiazepines in treating some aspects of AWS.⁷⁷ Compared to benzodiazepines, carbamazepine appears to have more benefit in symptom control and craving in less medically severe cases and in the outpatient setting, but less benefit in preventing seizures and DTs.⁷⁵

Gamma-hydroxybutyric acid (GHB), also a GABA_A agonist, was found to be comparable, but not superior, to benzodiazepines.⁷⁸ GHB has been used more widely in Europe, but concerns for GHB addiction have tempered its use in the United States.⁷⁹ Phenytoin has not been shown to be beneficial in seizures related to AWS.^80–82

Of note, in 2013, an unpublished case series was presented at a national meeting regarding the use of valproic acid or gabapentin with clonidine patches. The investigator described successful detoxification of medically moderate to severe AWS without the use of a benzodiazepine.^53,83 Since the findings remain unpublished, it is difficult to comment on the results, except to say that alternative regimens to benzodiazepines are being explored and discussed.

Clonidine and Dexmedetomidine

Clonidine and dexmedetomidine are centrally acting α₂ agonists. α₂ agonists decrease NE release and reduce symptoms of sympathetic overdrive. These agents have been studied in rodent models for the treatment of AWS^84–87 as well as in a limited number of human trials. The surgical literature has reported the successful use of clonidine with haloperidol and benzodiazepines in the treatment of surgical intensive care unit (SICU) and trauma patients with AWS;^{24,41,42,88,89} one randomized controlled trial found transdermal clonidine to be as effective as chlordiazepoxide in the management of mild AWS;⁹⁰ and a case series reported using high-dose clonidine patches in combination with valproic acid or gabapentin in the treatment of moderate to severe AWS.⁸³ Most studies (case reports and uncontrolled case series) of dexmedetomidine in AWS have focused on its use as an adjunctive agent.^91–101 Both clonidine and dexmedetomidine can produce sedation as well as hypotension and bradycardia, and acute withdrawal from dexmedetomodine has been described in patients on prolonged infusions.¹⁰² There is likely a role for clonidine and dexmedetomidine as adjuncts in the treatment of AWS, but more studies are required to fully comment on appropriate case selection, outcomes, and patient safety.¹⁰³ At this time, neither agent is recommended for use as monotherapy in AWS.

Baclofen

Baclofen is a stereo-selective agonist of GABA_B. It has been shown to be effective in suppressing AWS in rodents^104,105 and, in a small number of human trials, to be comparable, but not superior, to benzodiazepines.^106–108 It is notable that the benzodiazepine doses were low in the human trial comparison groups, suggesting that the study population was experiencing less severe withdrawal. A recent Cochrane review concluded that there were insufficient data to make outcome or safety conclusions regarding baclofen’s use for AWS.¹⁰⁹ Baclofen has shown some effect in supporting efforts at alcohol abstinence in alcohol-dependent patients with liver cirrhosis. This application has gained popularity, particularly in Europe, and warrants further study.^{106,107,110–113}

Ethanol

Ethanol therapy for AWS (IV or PO) is not recommended due to a high failure rate and potential for complications.^114–117 In both PO and IV forms, alcohol has large free water content, an association with electrolyte and behavioral disturbances, and a tendency to cause hypoglycemia. These factors, as well as difficulties in titration to adequate blood levels, make it an unsuitable choice.¹¹⁵ In a randomized trial of IV ethanol versus scheduled-dose diazepam in a trauma ICU, IV ethanol was found to be inferior to diazepam and was associated with greater treatment failures.¹¹⁷ Prophylaxis with ethanol in patients, particularly for elective surgical admissions, is practiced in some centers.¹¹⁵ A discussion of this is beyond the scope of this chapter, and it remains a controversial practice.

Beta-blockers

Beta-adrenergic agents should not be used to control the hypertension and tachycardia associated with AWS until the underlying cause (i.e., the hyperadrenergic AWS state) is treated. Once fluid status and appropriate sedation are administered, vital sign abnormalities usually normalize. If hypertension coexists, beta-blockers and other antihypertensive agents can be used to control blood pressure. Beta-blockers, by masking abnormal vital signs, can also obscure the diagnosis of delirium tremens.¹¹⁸

Tapering

Once initial control of withdrawal symptoms is achieved with benzodiazepines and/or phenobarbital/propofol and a stable clinical trend is established for 24 to 28 hours, a plan for tapering must be implemented. A recommended approach is to taper by approximately 20% per day of total daily benzodiazepine equivalent dose. Taper with chlordiazepoxide when possible (anticipate starting with ≥100 mg chlordiazepoxide PO q2–8h in severe cases). For example, if a patient had a total of 700 mg chlordiazepoxide PO and 8 mg lorazepam IV over 24 hours (equivalent to ~900 mg chlordiazepoxide), then the next day’s dose would be 720 mg/day (divided q6–8h). Patients treated with long-acting agents with active metabolites may exhibit varying degrees of an “autotaper” effect while still receiving STT. If during tapering, a patient’s CIWA-Ar scores increase to >10, give supplemental medication for breakthrough symptoms and consider a slower taper by increasing the daily dose, that is, tapering by 10% per day. Propofol infusions should be tapered as early as possible (after 2 to 3 days) because of the associated risk of infection and hypertriglyceridemia. Tapering from nonbenzodiazepines is not well described in moderate to severe AWS.

General Supportive Care

The following additional supportive care measures for the patient with AWS should be considered, particularly if the patient is heavily sedated or bedbound.

Related posts:

Extracorporeal Membrane Oxygenation Sickle Cell Disease Transfusion Therapy The Central Venous and Pulmonary Artery Catheter Pancreatitis Subarachnoid and Intracerebral Hemorrhage