One of the keys to accomplishing the above-mentioned tasks is to understand how alcohol affects the brain leading to dependence and withdrawal. The main effects of alcohol intoxication are euphoria, anesthesia, amnesia, and at higher levels of intoxication respiratory depression and coma. These effects are caused not by a single mechanism, but by several within the cell and cell membrane, thus making treatment of alcoholism with a single medication virtuously impossible. First alcohol activates the n-methyl-d-aspartate (NMDA) receptor, which is a glutamate receptor. This activation leads to intoxication, impaired cognition and learning. Chronic attenuation of NMDA also causes the reported excitotoxic effects on neurons resulting in cell death likely leading to Wernicke-Korsokoff syndrome, blackouts and cerebellar degeneration, and susceptibility to seizures during withdrawal. Second, alcohol results in increase in endogenous opioids in the brain causing the euphoric state, which reinforces further intake of alcohol. Activation of the GABA A-type receptors, which causes inhibitory chloride flux into the cell, resulting in the anxiolytic, motor skill impairment, and sedative effects seen in intoxication with alcohol. Finally, alcohol effects the serotonin and dopamine reward centers in the brain leading to further use. Although drugs such as naltrexone, Disulfram and Fluoxetine have been used in the treatment of alcoholism, they are simply tools in treatment and are not cures for the disease.

Metabolic abnormalities such as hypokalemia, due to alterations in aldosterone level, renal and extra-renal losses, and changes in the distribution of potassium levels across the cell membrane are often seen in withdrawal. Hypomagnesemia is common and may predispose the patient to withdrawal seizures. Hypophosphatemia is common due to malnutrition and is frequently symptomatic, thus predisposing the patient to fatal cardiac failure and rhabdomyolysis. Finally because of hyperthermia, diaphoresis, vomiting, and tachypnea, patients in alcohol withdrawal are often volume depleted leading to severe cardiovascular compromise.

Alcohol withdrawal is a clinical condition that physicians must be knowledgeable about, as it extends across the disciplines of medicine. It is more than just a single entity but rather a spectrum of signs and symptoms ranging from very mild to the severest form, delirium tremens, which is associated with a documented risk of mortality. Sutton first clinically documented the alcohol withdrawal syndrome in the medical literature in 1813 in Tracts on Delirium Tremens. Alcohol withdrawal is defined in the DSM-IV by a history of cessation or reduction in heavy and prolonged use of alcohol, and the presence of two or more of the symptoms of alcohol withdrawal. The physician must be able to distinguish the symptoms of alcohol withdrawal from other health problems. Patients presenting with alcohol withdrawal often have other existing medical problems, so identification is crucial. The mildest form of alcohol withdrawal includes symptoms due to increased central nervous system and sympathetic activity. These usually consist of agitation, increased sweating, tachycardia, increased hand tremor, GI upset, insomnia, palpitations, headache, and anorexia. Alcoholic hallucinosis may also be present in the form of transient tactile visual or auditory hallucinations, with visual being the most common. Alcoholic hallucinosis should not be confused with hallucinations seen with delirium tremens because the latter is usually associated with clouding of the sensorium, where as the former is not. Patients experiencing alcohol withdrawal may also have seizures. Alcoholic withdrawal seizures are usually generalized tonic-clonic convulsions occurring within the first 48 hours after the last drink. Seizures can be attributed to alcohol withdrawal if the patient has a normal EEG, and a history of documented seizure activity in withdrawal, and no other cause for their seizures can be determined.

The most severe form of alcohol withdrawal is delirium tremens (DTs), which carries a mortality risk. The medical literature of the early twentieth century reported this mortality risk to be approximately 37%. Improvements in recognition, pharmacotherapy, and treatment of co-morbid conditions have led to the more modern figure of 1-5%. In-patients experiencing alcohol withdrawal approximately 5% will develop DTs. Death, when it occurs in DTs, is usually due to arrhythmias or complications from the DTs, such as pneumonia. Ferguson et al, describe the following risk factors for the development of DTs:

• History of Sustained Drinking
• Age Greater than 30
• History of previous DTs
• Presence of concurrent illness
• Greater number of days since last drink

The characteristic features of DTs are hallucinations, diaphoresis, agitation, low-grade fever, tachcycardia, hypertension, and disorientation. Other physiologic changes seen in DTs include elevated cardiac indices, oxygen delivery, and oxygen consumption. Because of Hyperventilation and respiratory alkalosis, there is a significant decrease in cerebral blood flow, which has been shown to correlate with the length of the preceding alcohol binge, and the degree of delirium.

Natural History of the Alcohol Withdrawal Process: Foy et al, QJM, 1997

Victor and Adams wrote the fist comprehensive account of the natural history of alcohol withdrawal in 1953. They described four different "states" seen either separately from one another, or in combination during the withdrawal process. Their states were listed as tremulous, hallucinatory, epileptic, and delirious. They determined that the tremulous/transient hallucination symptom complex peaked within the first day and lasted for "a few days"; seizures were most common early on the second day, hallucinations usually began within the first 24 hours and lasted about 3 days. The delirious state began most commonly on the fourth day and was often over in 72 hours. Mortality form delirium was determined to be 15%. No other more modern studies in the natural history of alcohol withdrawal were conducted until 1997, when Foy et al conducted a study in order to determine a more accurate view of the natural history of alcohol withdrawal, incidences of seizures, hallucinations, delirium, and the risk factors for these events.

They found that complications occurred in the following order: seizures, followed by hallucinations and then delirium. Seizures occurred the earliest with 50% occurring on admission and 90% by 9 hours. Of those whom had hallucinations, 50% occurred by 20.5 hours and 90% by 64 hours. The hallucinations were mostly short-lived with a median duration of 6 hours and 90^th percentile of 46 hours.

Of the patients who experienced delirium, 50% occurred within 46 hours and 90% by 85 hours. The duration of delirium was wider and more variable with a median duration of 23 hours and 90^th percentile of 100 hours. Excluding the patients admitted already with complicated withdrawal, the median time of onset of withdrawal was 5 hours and 90% were withdrawing by 24 hours from admission. The median time of onset of withdrawal of those who experienced complications was 7 hours versus 4 hours for those who did not experience complications during the withdrawal. For the patients that had a measured BAL of 0, the median time to onset was 0 hours (i.e. on admission) and 75% were in withdrawal with in 1 hour. The median time of resolution of withdrawal for those with complications was 33 hours versus 22 hours for those without, and 99% of the reactions were over in 120 hours.

Of the 426 patients in the study, 113 experienced complications. There were 46 deaths either during the admission or subsequently, but only 2 died within the period of alcohol withdrawal. Both deaths were due to complications of over sedation during delirious episodes leading to fatal respiratory failure. Of the 113 complications, there were a few patients who experienced more than one complication; therefore, reports of events are higher than the total number of patients with complications. Ten patients developed seizures, 90 hallucinations, and 45 experienced delirium.

From the study they identified the following risk factors for developing complications during withdrawal:

1. Age greater than 70
2. Need for assisted ventilation
3. Pathology of the CNS, hypoxia, and femur fractured conveyed a greater risk of delirium
4. Seizures on admission
5. Delaying initial assessment and subsequent diagnosis for more than 24 hours

As above mentioned, alcohol withdrawal syndrome has a wide variation in presentation and symptoms, thus leading to great subjectivity in the assessment of the severity of withdrawal by healthcare workers. There is a need for objective analysis in assessing severity for use in guiding healthcare workers in administering the proper treatment to the patient experiencing withdrawal from alcohol. This problem was identified by Gross et al in 1973 who developed one of the earliest scales in assessing withdrawal severity, which consisted of a 32 item scoring system called the Total Severity Assessment Scale (TSA), which they then modified to an eleven item scale called the Selective Severity Assessment Scale (SSA). This scale was deficient in its applicability in following the clinical course of the withdrawal reaction and application of appropriate treatment.

Shaw et al identified the need for a scale that could be applied more frequently and used to adequately quantify and direct patient treatment. They called their scale the Clinical Institute Withdrawal Assessment for Alcohol Withdrawal or CIWA-A. This scale incorporated the original SSA items of observed tremor, sweating, clouding of the sensorium, hallucinations, quality of contact, agitation, and seizures. They substituted nausea and vomiting for sleeping and eating disturbances, to allow the scale to be scored every 30 minutes rather than daily. Also they substituted tactile, visual, and auditory disturbances for hallucinations, since these usually precede hallucinations in alcohol withdrawal. Headache and flushing of face were also added. The scale was scored on a basis of 1-7 points of severity in each interaction, with the total score being tallied to determine severity of withdrawal as defined in their study results as:

• Mild-20.4 (+/-2.6)
• Moderate 24.2 (+/- 5.4)
• Severe 29 (+/- 7.6)

1. The systematic evaluation of patients using the CIWA-A scale to score the severity of their withdrawal coupled with "supportive" nursing care was effective in the treatment of patients suffering from alcohol withdrawal.
2. 75% of the patients in their study required no medication at all in their treatment.
3. Nurses could be trained to administer the scale in a timely fashion, and use it to objectively evaluate the effectiveness of their care without direct physician supervision.
4. The scale was also valid when compared to physician assessments.
5. Using the scale was also important because frequent nursing intervention was beneficial to the patient. The interaction reoriented the patient and helped calm them, thus preventing the need for medication in many circumstances.

Although Shaw developed and proved the validity of the CIWA-A scale, their study did not include patients with more than just alcohol withdrawal as their primary diagnosis. It was not clear if the use of such a scale would be beneficial in a general medical facility, where patients often had other existing co-morbidities. In 1988 Foy et al set out to determine its usefulness in such a situation. Their study was performed in the Royal Newcastle Hospital in Australia on 203 adult general medical and surgical patients age 20-75, who in addition to various other conditions, met criteria for alcohol withdrawal in hopes to determine that CIWA testing was valid in this population.

Patients were assessed by nursing staff using the CIWA-A every 4 hours for the first 24-48 hours. Scoring then ceased when it was less than or equal to 10. When the score was greater than 10 it was increased to every 2 hours. If >15, scores were done hourly. Treatment was advised when scores were >15 on 2 separate occasions, or >20 once. The recommended treatment was the loading technique of diazepam20 mg at 2-hour intervals until the scores had fallen to <10.

The group assessed the following outcomes: Occurrence of severe withdrawal (confusion, hallucination, or seizures after admission), highest score prior to developing complications, or if no complications, prior to discharge, and use of benzodiazepines for symptoms. One hundred ten had scores>15 and received at least 1 dose of medication. Fifteen of the remaining 93 who had scores <15 received prophylactic treatment with at least 20 mg of diazepam. They had 37 complications (4 seizures, 33 confusion). These patients had mean highest scores of 21.8 compared to those who had no complications having a mean highest score of 15.6 (statistically significant p<0.001). Of the patients who had scores >15, 75 were treated, with only 11 developing severe withdrawal versus 35 who were not treated of which 21 developed severe withdrawal (RR 3.72). Of the 93 with CIWA-A <15, 78 were not treated and only 5 developed severe withdrawal (RR 1.92). Of the 15 with scores of <15 who received prophylactic diazepam none developed severe withdrawal.

Eleven patients who had high scores and were adequately treated or had low scores went on to have severe withdrawal. These patients were suffering from femur fractures (3), shock (2), and septicemia (1).

The key findings from this study were as follows:

1. Patients who developed severe alcohol withdrawal had higher scores on assessment scales than other patients even before the development of seizures, hallucinations or confusion; therefore, a severity scale used in a general hospital does predict who is at greater risk for severe withdrawal.
2. The higher the score the greater the risk that an untreated patient would develop severe withdrawal.
3. The scale can be used as a guide to treatment.
4. Benzodiazepines do appear to prevent complications when given early as determined by a high CIWA-A score.
5. The only limitations to using the CIWA-A in a general hospital were in dealing with patients who were critically ill (hypoxia, shock or septicemia) or had femur fractures, and thus require special attention.

In 1987 Sullivan et al identified that the original CIWA-A had several items that were redundant and by eliminating these, it could be more efficient to administer without losing accuracy of assessing the withdrawal severity. They reviewed data from 137 subjects involved in previous studies in alcohol withdrawal who underwent CIWA-A scoring as part of the protocol. After analyzing the cases and their scores, he found the items of convulsions, quality of contact, general hallucinations, flushing, and thought disturbances to be inadequate or redundant, and reassessed the cases using his revised scale the CIWA-Ar and determined that improved efficacy was achieved without any significant loss in accuracy. Trained nursing staff could reliably administer this scale in <1 minute. His scale consists of vomiting, sweats, tremor, anxiety, agitation, tactile disturbances, auditory hallucinations, visual hallucinations, headache, and clouding of the sensorium, and is the one which is preferred in most hospitals that use assessment scales in the evaluation of alcohol withdrawal. (See attachment)

The study consisted of retrospectively analyzing the medical record of 117 patients who had been treated for alcohol withdrawal in the previous 24 months without the use of the CIWA-Ar. This S- (control) group and they were treated with fixed-dosing of benzodiazepines without criteria for drug administration. The S+ group consisted of 133 patients treated after the initiation of the scale over the next 24 months, and examined prospectively. They were similar in age, race, sex, co-morbidities, drinking history, and admitting BAL to the S- group. These patients were evaluated by trained nurses and the decision on weather or not to treat the patient was made by house staff at Roger Williams General Hospital (a subsidiary of Brown University), based on their CIWA-Ar scores.

Patients were evaluated hourly during the initial phases of the withdrawal period, and then as needed with the CIWA-Ar. If they had a score>10 they received Diazepam 20 mg, or Librium 100 mg orally.

Total dosage was calculated as diazepam equivalents (DZE). They found that the S+ group required a mean dose of 50 mg, versus the S- group requiring a mean of 75mg. There was a greater proportion of people in the S+ group receiving low-dose (<20 mg DZE) treatment (44/133) compared with the S- group (25/117). There was no difference in the rate of complications between the groups, and each had one patient enter the DTs, both of whom had addiction to medium range benzodiazepines prior to admission.

From the study it can be concluded that:

1. The CIWA-Ar was an effective guide in directing medication administration.
2. Using the CIWA-Ar led to an improvement in the appropriateness of pharmacotherapy without a difference in morbidity.
3. When the scale was used, patients with greater dependence, and hence worse withdrawal received greater amounts of medicine and vice versa. Thus, there was titration of drug administration to therapeutic requirement in a more appropriate manner when the CIWA-Ar was used.
4. A lower average of medication used in the CIWA-Ar would lead to financial savings without increasing the rate of complications.
5. The use of the CIWA-Ar scale can also help in writing the appropriate amount of PRN medication.

The treatment of acute alcohol withdrawal is one that in the past has undergone many studies in the medical literature, but still the treatments used varied widely among physicians. In1997 Mayo-Smith formulated a working group on pharmacological management of alcohol withdrawal in order to produce guidelines for the treatment of alcohol withdrawal. The performed a meta-analysis on 134 articles, which included 65 prospective controlled trials and 42 different medications, addressing the outcomes they were interested in studying. These outcomes were: (1) severity of alcohol withdrawal, (2) withdrawal delirium, and (3) seizures from withdrawal, (4) completion of withdrawal, (5) entry into rehab, (6) adverse effects, and (7) cost.

They first looked at the agents used in treating alcohol withdrawal. It was determined that the benzodiazepines were the agent of choice in treating alcohol withdrawal. It appeared that all of the benzodiazepines were equally efficacious in reducing the signs and symptoms of alcohol withdrawal, but some were preferred over others. They found that the longer acting agents such as Chlordiazepoxide (Librium) may be more effective in preventing seizures, and also lead to a smoother withdrawal with less breakthrough or rebound symptoms. These agents also could cause excess sedation in special populations such as the elderly or those with marked liver disease. The longer-acting agents also have less potential for abuse than do shorter acting agents such as alprazolam, diazepam, and lorazepam. Several other agents that have been used in the past in treating withdrawal were also examined:

b -Blockers: There was some evidence that they may reduce the autonomic manifestations of withdrawal, but no studies showed they conferred any benefit in reducing seizure activity. Delirium is a known side effect of the more centrally penetrating b -Bs such as propanolol, and in one case lead to improper diagnosis of delirium from alcohol withdrawal. There were no studies that either proved or disproved this as a class effect among all b -Bs.

Clonidine: Studies did show that Clonidine helped ameliorate the symptoms in mild to moderate withdrawal. There was no data on the ability to reduce or increase the incidence of delirium or seizures. There is a well-documented rebound hypertension effect with Clonidine.

Carbamazepine: This is widely used in Europe, and equal in efficacy to barbital and oxazepam in patients with mild to moderate withdrawal. There was limited data comparing its efficacy in reducing seizures (except in animal studies) and preventing delirium. There was no abuse potential.

Neuroleptic Agents: They have some effectiveness in reducing signs and symptoms, but less effective than benzodiazepines in preventing delirium. Furthermore they increase the incidence of seizures No controlled studies showed their effectiveness in calming patients.

Magnesium: No data showed benefit in reducing signs and symptoms, or reducing seizures and delirium.

Ethyl Alcohol: There have been small case reports of ETOH given orally or IV in preventing withdrawal symptoms exist, but these are very small and uncontrolled without objective or quantitative assessment of withdrawal severity. There are also no trials comparing it to benzodiazepines or placebo. IV infusions require very close monitoring and are very expensive to administer. They are associated with tissue damage at the infusion site, and has well known and severe side effects and toxicity. They furthermore are counterintuitive in the treatment of alcohol abuse for the hospitalized patient.

Thiamin: It is useful in preventing Wernike-Korsakoff syndrome, but does not reduce delirium or seizures.

The working group also examined the studies already mentioned in this paper and others, and found the use of the CIWA-Ar to be of great importance in treating alcohol withdrawal. They compared the long-time standard of fixed-dosing of medication to symptom-triggered therapy. They found that symptom triggered therapy, which consists of giving medication to patients based on symptomatic need, as determined by their CIWA-Ar score, was equally effective as fixed-dosing.

In 1994 Saitz et al compare symptom triggered therapy with fixed dosing of medications. They randomized 111 patients in a double blinded placebo controlled study to receive either Chlordiazepoxide 50 mg every 6 hours for 4 doses then 25 mg for 8 doses, or placebo on the same schedule (the patients also received additional 25-100 mg if CIWA-Ar was >8). The placebo group received only the 25-100 mg when CIWA-Ar scores were >8. This group was also given placebo medication at the above listed fixed schedule. They concluded that patients receiving symptom-triggered therapy required less medication (mean 100mg vs. 425 mg), and had a shorter duration of medication treatment (9 hours vs. 68 hours), with no significant increase in complications (seizures, hallucination, or delirium) during treatment of withdrawal. They also concluded that symptom-triggered therapy did not lead to a worsened withdrawal severity; therefore, they concluded that fixed-dosing of medication leads to increased intensity of treatment and prolongs hospital course. They did refer to similar critically ill patients, as previously mentioned in this paper, that may need special consideration in determining scheduling of therapy due to their condition possibly confusing their CIWA-Ar scores.

They compared the average cost of Chlordiazepoxide 25 mg ($0.033), Diazepam 5 mg ($.071), and lorazepam 1 mg ($0.115). The cost of the longer acting agent is cheaper and is equally efficacious. Also they found that some practitioners routinely use continuous infusions of short acting agents such as lorazepam, which can result in large hospital costs. In one study by Hoey, hospital guidelines were established that led practitioners to use Chlordiazepoxide whenever it was appropriate in the treatment of alcohol withdrawal. This led to an average cost of $59.79 where the average cost of treatment before implementing the guidelines was $1008.72 due to physicians regularly using short acting agents often in infusion form. They had equivalent outcomes and no increase in adverse effects in their study after the guidelines were implemented. There is no evidence that infusion of short acting agents provides better outcomes than oral or IV bolus therapy with longer acting agents.

Final Recommendations:

In their paper they put forth the final specific treatment recommendation:

Monitoring:

-Monitor patient every 4-8 hours by means of CIWA-Ar until score has been < 8-10 for 24 hours; use additional assessments as needed.

-For patients with CIWA-Ar <8-10, supportive non-pharmacologic monitoring is acceptable.

-Patients with CIWA-Ar scores 8-15 benefit from medication thus reducing risk of complications.

-CIWA-Ar scores ≥ 15 have a significant risk of major complications if left untreated.

Symptom Triggered Regimens:

Administer one of the following every hour when the CIWA-Ar ≥ 8-10:

Librium 50-100 mg

Diazepam 10-20 mg

Lorazepam 2-4 mg

Repeat CIWA-Ar 1 hour after every dose to assess need for further medication

Fixed Schedule Regimens:

If necessary to give medication on a fixed schedule then administer one of the following:

Librium 50 mg every 6 hours for 4 doses then 25 mg every 6 hours for 8 doses

Diazepam 10 mg every 6 hours for 4 doses then 5 mg every 6 ours four 8 doses

Lorazepam 2 mg every 6 hours for 4 doses then 1 mg every 6 hours for 8 doses

Provide additional medication as needed CIWA-Ar ≥ 8-10 with above.

Other benzodiazepines may be used at equivalent doses.

The group put forth the following caveats:

1. For patients with a history of withdrawal seizures it is reasonable to provide one of the listed medications at presentation. It is also reasonable to monitor the patient the patient and provide symptom-triggered therapy.
2. For patients with severe co-morbidities medication should be considered even if symptoms are mild to moderate. (This is the group that fixed-dosing may be more appropriate.)
3. b -Blockers, Clonidine, and carbamazepine are not recommended as monotherapy because they do not reduce delirium or seizures. Neuroleptics should not be used as monotherapy because they do not reduce delirium and may increase seizures. They may be used in conjunction with benzodiazepines for marked agitation or severe hallucinations.
4. Routine administration of magnesium is not beneficial.
5. Ethyl alcohol is not recommended due to lack of controlled studies and known adverse events.
6. It is recommended that thiamine be administered to all alcoholics at initial administration.