Managing Alcohol Withdrawal in the Acutely Ill Hospitalized Adult.
Alcohol withdrawal syndrome (AWS) occurs when regular consumption of alcohol over an extended period of time either ceases or decreases abruptly.[1-3,6] Nursing dissatisfaction with the care of acutely ill patients who also experienced AWS prompted an informal survey of practice. After evidence emerged for unacceptably high incidences of delirium, leather restraint use and intensive care unit (ICU) admissions, an institution wide, multi-disciplinary Alcohol Withdrawal Task Force convened to standardize and guide the approach to manage individuals with acute medical or surgical illness who were also at high risk for alcohol withdrawal. Its primary goal was to avoid cardiorespiratory and neurological morbidities such as seizures, delirium and psychosis associated with severe withdrawal. Secondary goals included reducing symptom severity, treating underlying illness(es), discussing with patients the feasibility and desirability of formal alcoholic rehabilitation and education of patients and significant others. A triphasic symptom-triggered plan was developed and aggressively publicized. It included a variety of pharmacologic protocols to manage autonomic and psychiatric signs and symptoms which clinicians were free to use as proposed or adapt. It also included cues for addressing fluid, electrolyte and nutritional issues. Finally, an interdisciplinary plan of care was developed to anchor the supportive care and structure care delivery.
The purposes of this article are to briefly review the alcohol withdrawal syndrome and report on the protocol's effectiveness in reducing adverse outcomes among acutely ill adults hospitalized for reasons other than alcohol withdrawal. The purposes of the retrospective analysis reported herein are to 1) describe the characteristics of the population identified as being at "high risk" for AWS, and 2) explore the relationships between selected historical, demographic and clinical characteristics and the subsequent clinical course of AWS.
It is estimated that up to 14% of individuals have alcohol dependence at some time during their life. The prevalence of alcohol-related disorders is high in the general population and is higher in general hospitals. Effects of prolonged exposure of the brain to alcohol have been linked strongly to perturbations of central nervous system (CNS) neurotransmitter systems -- notably GABA (major inhibitory) and glutaminergic (major excitatory), their receptors and intracellular signaling systems. The net effect of prolonged exposure is widespread suppression of neurochemical activity; when intake fails precipitously or ceases, removal of alcohol-mediated suppression results in enhanced CNS excitability.[1,6,112,27,35]
Unintentional or involuntary detoxification often occurs as a consequence of unplanned hospitalization and is characterized by a "symptom complex of tremulousness, hallucinations, seizures, confusion and psychomotor and autonomic overactivity" (p.1172). Generalized seizures generally affect fewer than 3% of individuals, but specific high risk populations report much higher incidences -- in inner city hospitals, as many as 20% suffered one or more seizures and Daryanani et al (1994) reported an incidence of 46% on admission. These "rum fits" usually occur after 7-48 hours of abstinence and may precede other signs and symptoms of AWS. Most seizures are tonic-clonic in nature and occur in bursts of 2 to 6. Only 2% develop status epilepticus and 30% are likely to go on to experience alcohol withdrawal delirium.[1,7] The incidence of hallucinations ranges from 5 to 25% with disorganized sensory perceptions occurring in every imaginable combination.[1,22]
The clinical severity of AWS traverses the continuum from brief, non-disabling flu-like symptoms of short duration to critical, sometimes lethal catastrophes (Table 1). At its most severe, substance withdrawal delirium (Table 2), known as delirium tremens (DTs) can occur, characterized by "profound confusion, delusions, vivid hallucinations, tremor, agitation, and sleeplessness -- as well as by the signs of increased autonomic nervous system overactivity." [1, p1176] It is often associated with acute illness. Incidence is estimated at about 5%; but in vulnerable populations, it may approach 25%.[5,9,14,22] Mortality ranges from 5-15% and death is usually due to cardiac or respiratory failure.[3,7]
DSM-IV Diagnostics Criteria for Alcohol Withdrawal
A Cessation of (or reduction in) alcohol use that has been heavy and prolonged. B. Two or more of the following, developing with several hours to a few days after criterion A: (1) autonomic hyperactivity (eg. sweating or pulse > 100) (2) increased hand tremor (3) insomnia (4) nausea or vomiting (5) transient visual, tactile, or auditory hallucinations or illusions (6) psychomotor agitation (7) anxiety (8) grand mal seizures C. The symptoms in Criterion B cause CliniCally significant distress or impairment in social, occupational, or other important areas of functioning. D. The symptoms are not due to a general medical condition and not better accounted for by another mental disorder.
DSM-IV Diagnostic Criteria for Substance Withdrawal Delirium
A Disturbance of consciousness (reduced clarity of awareness of the environment) with reduced ability to focus, sustain or shift attention. B A change in cognition (such as memory deficit, disorientation, language disturbance) or the development of a perceptual disturbance that is not better accounted for by a pre-existing established or evolving dementia. C The disturbance develops over a short period of time (usually hours to days and tends to fluctuate during the course of the day. D There is evidence from the history, physical examination, or laboratory findings that the symptoms in Criteria A and B developed during, or shortly after, a withdrawal syndrome.
The prevalence of alcohol-related problems in community and teaching hospitals ranges from 10-50%;[19,20] up to 20% of admissions may be due to medical complications of excessive use. In a European report, it was estimated that 21% of admissions were alcohol related, affecting a preponderance of men in the 35-55 year range. Australian estimates were slightly lower. Alcohol withdrawal occurs often in association with medical illness[1,2] and may be more severe than the AWS experienced by otherwise healthy individuals who electively cease to ingest alcohol. In the subgroup of acutely ill patients, the incidence of AWS has been reported as high as 82%, DTs 12%, seizures 6.5% and hallucinations 19.6%. Physicians fail to identify more than half of alcoholics who appear for medical care because most patients aren't routinely screened despite recommendations by the National Institute on Alcohol Abuse and Alcoholism, Institute of Medicine and United States Preventive Services Task Force and the availability of valid and reliable screening instruments.[22,30,34]
Individuals who are mildly symptomatic or asymptomatic at 36 hours of abstinence may not require pharmacologic support through their AWS. Reassurance, reality orientation, symptom severity monitoring and supportive nursing care may suffice.[1,4,23,27] However, those who present with moderate or severe symptoms of AWS or serious medical or surgical conditions; or, who by virtue of age or infirmity, are vulnerable to adrenergic hyperactivity and subsequent exhaustion and collapse, ought to be treated pharmacologically.[1,23] There is compelling "evidence that pharmacological management of alcohol withdrawal can treat and prevent symptoms of withdrawal, seizures, and delirium tremens" (p.884). The benzodiazepines (BZD) have been the mainstay of therapy and drugs of choice in the United States (US) since the 1960s, with a wider margin of safety than the barbiturates, especially with respect to respiratory depressant effects, fewer serious drug drug interactions, dependency and abuse. BZD use is based upon their cross tolerance with ethanol and their sedative, anxiolytic and anticonvulsant properties. The recent introduction of flumazenil (Romazicon[C]) has increased their margin of safety.[3,4,11,12,14,17-19,27,32,35]
The ability to predict who might suffer a very symptomatic withdrawal would permit preemptive treatment. Ferguson et al, in a series of logistic regression analyses, identified time abstinent and the presence of serious concurrent medical illness risk factors for delirium tremens; age over 60 years may also be a risk factor for delirium. History of AWS, severe symptoms, maximum number of drinks daily and concurrent serious medical or psychiatric disorders have been identified as risk factors for the development of delirium and fits.[5,4,31] Prior seizures appear to increase the probability of subsequent fits; patients with a positive history actually had more seizures than their counterparts during an episode of care. Finally, patients older than 70 years who suffered one or more seizures on admission were shown to have an increased risk for a complicated AWS.
A retrospective, descriptive, computerized review was conducted for all adult patients identified as being at high risk for alcohol withdrawal. Following retrieval of demographic information, a proportional random sample was drawn from the population and a chart audit was conducted to describe selected historical and clinical parameters, details of treatment over time (96 hours) and complications of management.
The setting is an urban, not-for-profit community hospital of +500 beds located in the Midwest. It lacks an in-house psychiatric or alcohol and other drug addictions (AODA) unit. The population included all patients admitted from August, 1995 through December 31, 1997 who were identified as high risk for alcohol withdrawal. The population included many who received no pharmacologic treatment for their AWS. They were cared for by surgical and medical intern and resident staffs, and over 100 attending staff physicians, usually board certified in family pratice, internal medicine or one of the medical subspecialties.
The CAGE questionnaire was used by nursing staff members to screen all adult admissions. Selection was based upon recommendations of the US Preventive Services Task Force. It is an ideal screening tool for adults and adolescents, as it is easy to administer, reliable, sensitive and specific. Two affirmative CAGE responses, intoxication on admission, positive ethanol toxicology (urine or serum) and/or clinical suspicion were prompts for physicians and nurses to begin clinical monitoring for AWS. That is, the expectation of withdrawal triggered clinical surveillance.
The Symptom Assessment Stabilization Inventory (SASI) was used for clinical monitoring. It is a 14 item scale; items 1 through 9 describe autonomic signs and symptoms; items 10 through 14 generate a "psychosis" score which reflects the intensity of sensory perceptual derangement. Scores range from 0 through 41 (maximum) inclusive. A symptom triggered approach was recommended by our consultant and it mandated drug administration when the threshold to treat scores were exceeded. They included an autonomic score [is greater than] 9 (sedative-hypnotic) and a psychosis score [is greater than] 0 (antipsychotic). This recommendation did not in any way preclude initiating or escalating treatment at lower scores based on clinical judgement. Patient comorbidities and outcomes were tracked on standardized forms.
The data base (Varitrax[TM]) that supports the institution's Outcomes Based Care Management (OBCM) program was used to identify the population for study. We were unable to retrospectively identify study subjects by ICD-9 codes or specific DRGs due to the wide variability in post-discharge principle diagnoses and propensity of some physicians to avoid assigning an alcohol-related diagnosis. Four hundred forty-one episodes of care were identified for study. Demographic variables, concurrent illnesses and historical co-morbidities thought to affect clinical course and data regarding specific outcomes were retrieved for analysis. Varitrax[TM] entries were then matched to specific encounter numbers in order to retrieve length of stay (LOS), cost and secondary and primary diagnoses, admission sources and discharge disposition from the institution's financial data base, TSI[TM]. These files were transferred to Excel[TM] spread sheets. Once 100% accuracy was assured, analysis was done using Statistical Package for the Social Sciences (SPSS[TM]). A proportional (15%) sample was then drawn and a chart review was conducted to allow more detailed analysis. Sixty five charts were selected; seven were missing, incomplete or unavailable. Data was triple checked to assure accuracy of retrieval and entry.
Four hundred forty-one patient episodes of care or encounters were identified. The encounters were sorted based on two primary endpoints: 1) whether or not the subject exceeded one or both threshold to treat scores and 2) whether or not an adverse outcome (eg., over-medication or seizure) occurred.
Group assignment. Group I (n=163) failed to exceed threshold scores during 72 hours of monitoring following their identification of being at risk. At 72 hours, clinical monitoring for AWS ceased per protocol. Group II (n=74) experienced a controlled withdrawal; group III (n=54) suffered uncontrolled withdrawal. Uncontrolled withdrawal was stringently defined by the Task Force as being of sufficient severity to require restraint (soft or leather) use for patient safety, or characterized by one or more seizures. Population data did not allow us to differentiate patients who suffered either, or both complications. Patients who had controlled episodes exceeded threshold to treat score(s) at least once, but did not require restraint use or suffer convulsions. Group IV (n=30) subjects defaulted off the plan prior to completion due to severe intercurrent medical or surgical illness; none suffered uncontrolled withdrawal. Seven patients (Group V) defaulted off and experienced uncontrolled withdrawal. The final group (VI; n=113) was monitored clinically, but lacked documentation on the OBCM data collection forms (nonpermanent chart documents). A hypothesis proposed that since charting is "by exception," no member of this final group exceeded threshold to treat scores.
Following group assignment, and prior to making cross-group comparisons, several categories were collapsed to facilitate analysis. Normal admission source, defined as scheduled admissions, were differentiated from all varieties of emergency department admissions. Patients who went home with and without homecare were combined, as were deaths.
Demographics. Patient and system characteristics are listed in Tables 3 and 4. Individuals enrolled in the plan were predominantly male (63.5%) in their 5th decade (X = 51.6y + 16.01) and most were single (71.2% widowed, separated, divorced, or never married). African Americans and EuroAmericans were roughly equally represented, 42.2% and 49.9% respectively. They overwhelmingly entered the system (78.4%) through the Emergency Department (ED), and approximately the same percentage was discharged home with or without additional support (77.1%). Every patient admitted had one or more serious concurrent medical or surgical illness(es). Twenty-six percent could describe at least one prior episode of AWS, and 10.66% could describe at least one prior AWS convulsion. Sixty percent experienced controlled or no withdrawal (Groups I,II, IV); with inclusion of the indeterminate outcome group, those experiencing no adverse AWS related events reached 86.17%. Length of stays (LOS) and charges were widely variable across groups, reflecting the complexity and severity of the alcohol-related illness as well as the principle diagnosis. Mortality during hospitalization was 1.8%. We were unable to discern whether the deathes were due to a primary illness, the direct result of severe AWS or iatrogenic.
Demographic Characteristics -- Population
Controlled Variables No Withdrawal Withdrawal N 163 74 (36.96%) (16.78%) Age (mean; SD) 50.47 [+ or -] 16.03 50 [+ or -] 13.61 (median; R) 47; 76 47; 51 Gender Male 97 (59.51%) 46 (62.16%) Female 66 (40.50%) 28 (37.83%) Race EuroAmerican 70 (42.94%) 47 (63.51%) African American 81 (49.70%) 20 (27.02%) Hispanic - 1 (1.35%) Unknown 12 (7.36%) 6 (8.11%) Marital Status Married 40 (24.53%) 24 (32.43%) Divorce 21 (12.88%) 11 (14.86%) Separate 4 (2.45%) 2 (2.70%) Widow 17 (10.42%) 4 (5.4%) Single 78 (47.85%) 32 (43.24%) Unknown 3 (1.84%) 1 (1.35%) Uncontrolled Defaulted Off Variables Withdrawal Controlled AW N 54 30 (12.24%) (6.80%) Age (mean; SD) 53 [+ or -] 15.34 55 [+ or -] 17.3 (median; R) 51; 53 56; 59 Gender Male 35 (64.8%) 21 (70%) Female 19 (35.2%) 9 (30%) Race EuroAmerican 31 (57.40%) 22 (73.33%) African American 18 (33.33%) 7 (23.33%) Hispanic - 1 (3.33%) Unknown 5 (9.26%) Marital Status Married 12 (22.22%) 8 (26.66%) Divorce 5 (9.26%) 1 (3.33%) Separate 4 (7.41%) 1 (3.33%) Widow 13 (24.07%) 3 (10%) Single 19 (35.19%) 17 (56.66%) Unknown 1 (1.85%) - Defaulted Off Indeterminate Variables Uncontrolled AW Outcome N 7 113 (1.59%) (25.62%) Age (mean; SD) 58.75 [+ or -] 19.97 51 [+ or -] 17.19 (median; R) 59; 61 47; 67 Gender Male 7 (100%) 74 (65.5%) Female - 39 (34.5%) Race EuroAmerican 3 (42.86%) 47 (41.6%) African American 2 (28.57%) 58 (51.3%) Hispanic - 1 (.9%) Unknown 2 (28.57%) 7 (6.2%) Marital Status Married 3 (42.86%) 31 (27.43%) Divorce 1 (14.29%) 11 (9.73%) Separate - 12 (10.62%) Widow - 15 (13.27%) Single 3 (42.86%) 40 (35.40%) Unknown 4 (3.54%) Variables Totals N 441 Age (mean; SD) 51.6 [+ or -] 16.01 (median; R) 48; 76 Gender Male Female Race EuroAmerican 220 (49.89%) African American 186 (42.18%) Hispanic 3 (.68%) Unknown 32 (7.26%) Marital Status Married 118 (26.76%) Divorce 50 (11.34%) Separate 23 (5.22%) Widow 52 (11.79%) Single 189 (42.986) Unknown 9 (2.04%) Table 4 System Use -- Population Controlled No Withdrawal Withdrawal Variables (n=163) (n=74) Admit Source ER 130 (79.75%) 61 (82.4%) Normal 33 (20.25%) 13 (17.6%) LOS Mean [+ or -] SD 4.963 [+ or -] 3.713 6.0 [+ or -] 5.44 Range (days) 0 - 23 0 - 24 ICU LOS Mean [+ or -] SD .258 [+ or -] 1.21 .851 [+ or -] 2.20 Mode 0 0 Discharge Disposition Home 129 (79.14%) 60 (81.08%) (+/- HHC) AMA 7 (4.29%) 0 TRANSFERS Other acute 3 (1.84%) 2 (2.70%) Rehab 8 (4.91%) 8 (10.81%) Intermediate 5 (3.07%) 1 (1.35%) SNF 7 (4.29%) 1 (1.35%) DIED 1 (.61%) Unknown 3 (1.84%) 2 (2.70%) Uncontrolled Defaulted Off Withdrawal Controlled AW Variables (n=54) (n=30) Admit Source ER 45 (83.33%) 25 (83.33%) Normal 8 (14.81%) 5 (16.67%) ([double dagger]) LOS Mean [+ or -] SD 8.0 [+ or -] 5.924 16.67 [+ or -] 20.86 Range (days) 1 - 26 0 - 108 ICU LOS Mean [+ or -] SD 1.22 [+ or -] 2.08 3.77 [+ or -] 5.83 Mode 0 0 Discharge Disposition Home 39 (72.2%) 14 (46.67%) (+/- HHC) AMA 3 (5.60%) 1 (3.33%) TRANSFERS Other acute 1 (1.85%) 2 (6.67%) Rehab 3 (5.56%) 3 (10%) Intermediate 2 (3.70%) 1 (3.33%) SNF 5 (9.30%) 5 (16.67%) DIED 1 (1.85%) 3 (10%) Unknown 0- 1 (3.33%) Defaulted Off Indeterminate Uncontrolled AW Outcome Variables (n=7) (n=113) Admit Source ER 5 (71.43%) 80 (70 80%) Normal 2 (28.57%) 31 (27.43%) LOS Mean [+ or -] SD 28 [+ or -] 27.48 4.93 [+ or -] 6 32 Range (days) 3 - 76 0 - 57 ICU LOS 3.857 [+ or -] 6.87 .74 [+ or -] 4.69 Mean [+ or -] SD Mode Discharge Disposition Home 3 (42.86%) 95 (84.07%) (+/- HHC) AMA 1 (14.29%) 2 (1.77%) TRANSFERS Other acute 1 (.88%) Rehab 3 (2.65%) Intermediate 2 (1.77%) SNF 3 (42.86%) 4 (3.54%) DIED 3 (2.65%) Unknown 3 (2.65%) Totals Variables (n=441) Admit Source ER 346 (78.46%)([dagger]) Normal 92 (20.86%) LOS Mean [+ or -] SD 6.65 [+ or -] 8.90 Range (days) 0 - 108 ICU LOS .896 [+ or -] 3.35 Mean [+ or -] SD Mode Discharge Disposition Home 340 (77.10%) (+/- HHC) AMA 14 (3.17%) TRANSFERS Other acute 9 (2.04%) Rehab 25 (5.67%) Intermediate 11 (2.50%) SNF 25 (5.67%) DIED 8 (1.81%) Unknown 9 (2.04%)
([dagger]) 3 admissions from acute care facilities
([double dagger]) transfer from acute facility
There were no significant differences across groups on gender (p=.30819), marital status (p=.11022), admission source (p=.40557) or payor (p=.33409). However, there were significant differences in ethnicity. Twice as many EuroAmericans as African Americans experienced withdrawal; and twice as many EuroAmericans as African Americans defaulted off with a significant episode of AWS. Groups also varied in ICU LOS. Patients who defaulted off due to serious intercurrent event, regardless of cause (alcohol-related or not) had significantly longer stays (p=.0000). Discharge destination was similarly different. Approximately 80% of those with no evidence of withdrawal going home and half that number who defaulted off due to serious illness.
Adverse outcomes. Adverse outcomes are summarized on Table 5. Oversedation, as defined by the Task Force, was not uncommon among patients who experienced AWS (controlled or uncontrolled). That is, at some point during their stay, they had a recorded a respiratory rate [is less than] 12 breaths/min, required more than verbal stimulation to respond, were inappropriate or inarticulate in speech or required mechanical ventilation. However, none required mechanical ventilation as a result of treatment; flumazenil (Romazicon[C]) was never administered. In 15.67% of the episodes, patients expressed no interest in stopping their alcohol intake; in some groups, the numbers were much higher. Although participants had positive psychoses scores in 9.3% and 7.0% during acute and stabilization phases of their AWS respectively, the incidence of undermedication was significantly higher in those who underwent uncontrolled withdrawal during both time periods.
Adverse Outcomes: Complicated withdrawal and Complications of Treatment
No Controlled Withdrawal Withdrawal Variables (n=163) (n=74) Acute Phase([Psi]) Positive psychosis score 1 (.61%) 3 (4.05%) Over-sedated([Xi]) 0 7 (9.46%) Stabilization Phase([Psi]) Positive psychosis score 0 3 (4.05%) Over-sedated([Xi]) 0 3 (4.05%) No plans to seek treatment 11 (6.75%) 25 (33.78%) Uncontrolled Defaulted Off Withdrawal Controlled AW Variables (n=54) (n=30) Acute Phase([Psi]) Positive psychosis score 35 (64.81%) 0 Over-sedated([Xi]) 18 (33.33%) 4 (13.3%) Stabilization Phase([Psi]) Positive psychosis score 22 (40.74%) 1 (3.33%) Over-sedated([Xi]) 16 (29.63%) 1 (3.33%) No plans to seek treatment 24 (44.44%) 6 (20.00%) Defaulted Off Indeterminate Uncontrolled AW Outcome Variables (n=7) (n=113) Acute Phase([Psi]) Positive psychosis score 2 (28.57%) Over-sedated([Xi]) 1 (14.29%) Stabilization Phase([Psi]) Positive psychosis score 5 (71.43%) Over-sedated([Xi]) 0 No plans to seek treatment 3 (42.86%) Variables Totals Acute Phase([Psi]) Positive psychosis score 41 (9.3%) Over-sedated([Xi]) 30 (6.8%) Stabilization Phase([Psi]) Positive psychosis score 31 (7.03%) Over-sedated([Xi]) 20 (4.54%) No plans to seek treatment 69 (15.67%)
([Psi]) Protocol defined. Acute phase begins when one or both threshold to treat scores are exceeded, it ends when SASI [is less than] 9 for 12 continuous hours. Stabilization phase follows the acute phase and ends when patient no longer requires medication for symptom management.
([Xi]) Protocol defined. See text for detail.
High risk profile. Using a short list of co-morbidities (Table 6), we began to develop a profile of patients at high risk for severe withdrawal that might trigger a more aggressive therapeutic approach to reduce the incidence of adverse outcomes. Several themes emerged. Those who underwent withdrawal, controlled or uncontrolled, had a significantly higher incidence of prior alcohol withdrawal episodes by history, had AWS seizures (p=.021) during their episode, and tended to suffer hallucinations.
Table 6 Co-morbidities -- Population No Controlled Withdrawal Withdrawal Variables (n=163) (n=74) Concurrent medical/surgical 163 74 illness Prior alcohol withdrawal 31 (19.02%) 37 (50%) Prior AW seizure(s) 5 (3.07%) 21 (28.38%) Uncontrolled Defaulted Off Withdrawal Controlled AW Variables (n=54) (n=30) Concurrent medical/surgical 54 30 illness Prior alcohol withdrawal 25 (46.30%) 5 (16.67%) Prior AW seizure(s) 13 (24.07%) 5 (16.67%) Defaulted Off Indeterminate Uncontrolled AW Outcome Variables (n=7) (n=113) Concurrent medical/surgical 7 113 illness Prior alcohol withdrawal 1 (14.29%) 18 (15.93%) Prior AW seizure(s) 0 3 (2.65%) Variables Totals Concurrent medical/surgical 441 illness Prior alcohol withdrawal 117 (26.53%) Prior AW seizure(s) 47 (10.66%)
(*) NB: Patients frequently reported one than one pre-existing condition.
Population analysis revealed that 13.83% (61/441) suffered a protocol-defined uncontrolled withdrawal. Given our inability to determine who had experienced serious neurological morbidity, we conducted a chart audit. Our goals were to define the nature of the specific event(s) that resulted in their classification as "uncontrolled," broaden our understanding of selected historical (Table 7) and clinical management and systems issues (Tables 8, 9), and describe the "indeterminate outcome" group. Data were collected over the first consecutive 96 hours that followed the patient's identification of being "at risk," which usually coincided with the first 96 hours of hospitalization. The time frame was selected given the temporal course of AWS, in which symptom severity tends to crescendo at 2-3 days, then decreases? We hoped to capture significant events both for those admitted after a short period of abstinence, as well as those admitted intoxicated or with positive toxicology.
Table 7 Subject Characteristics 0-24h -- Sample Controlled No Withdrawal Withdrawal Variables (N=163) (N=74) n 24 11 Age (Mean [+ or -] SD) 48.67 49.82 [+ or -] [+ or -] 15.75 16.20 Clinical Presentation Seizure in the ED/field 1 (4.2%) 1 (9.1%) Intoxicated 3 (12.5%) 0 Positive Etoh Toxicology 3 (12.5%) 3 (27.3%) CAGE performed 13 (54.2%) 7 (63.6%) First SASI (Mean; Mode 3.81; 3 4.91; 2 Range) 0 - 9 1 - 11 First psychosis (Mean; Mode) .06; 0 .333; 0 Ethanol related([Lambda]) Principle Diagnosis 11 (45.8%) 9 (81.8%) Historical Time Abstinent < 2 days 10 (41.7%) 7 (63.6%) [is greater than or equal to] 2 days 3 (12.5%) 0 History Epilepsy 1 (4.2%) 0 Elevated GGT 6 (25%) 7 (63.6%) [Na.sup.+] abnormalities 6 (25%) 4 (36.4%) [K.sup.+] abnormalities 5 (20.8%) 6 (54.5%) Uncontrolled Defaulted Off Withdrawal Controlled AW Variables (N=54) (N=30) n 8 5 Age (Mean [+ or -] SD) 52.13 53.4 [+ or -] [+ or -] 17.69 23.03 Clinical Presentation Seizure in the ED/field 0 0 Intoxicated 2 (25%) 1 (20%) Positive Etoh Toxicology 3 (37.5%) - CAGE performed 5 (62.57%) 4 (80%) First SASI (Mean; Mode 5.86; 4,9 3.6; 3 Range) 2 - 9 2 - 6 First psychosis (Mean; Mode) .57; 0 .4; 0 Ethanol related([Lambda]) Principle Diagnosis 5 (62.5%) 1 (20%) Historical Time Abstinent < 2 days 3 (37.5%) 4 (80%) [is greater than or equal to] 2 days 2 (25%) 0 History Epilepsy 0 0 Elevated GGT 3 (37.5%) 2 (40%) [Na.sup.+] abnormalities 3 (37.5%) 2 (40%) [K.sup.+] abnormalities 1 (12.5%) 2 (40%) Indeterminate Outcome Totals Variables (N=113) N=441 n 16 58 Age (Mean [+ or -] SD) 49.313 50.34 [+ or -] [+ or -] 19.56 17.50 Clinical Presentation Seizure in the ED/field 1 (6.3%) 3 (4.6%) Intoxicated 3 (18.8%) 9 (13.8%) Positive Etoh Toxicology 4 (25%) 13 (20%) CAGE performed 15 (93.8%) 45 (69.2%) First SASI (Mean; Mode 3.62; 2 4.32; 3 Range) First psychosis (Mean; Mode) 0; 0 -196; 0 Ethanol related([Lambda]) Principle Diagnosis 8 (50%) 34 (52.3%) Historical Time Abstinent < 2 days 14 (87.5%) 38 (58.5%) [is greater than or equal to] 2 days 0- 6 (9.2%) History Epilepsy - 1 (1.5%) Elevated GGT 5 (31.3%) 24 (36.9%) [Na.sup.+] abnormalities 4 (25.1%) 19 (29.3%) [K.sup.+] abnormalities 5 (31.3%) 19 (29.2%)
[yen] = No subject was drawn from Group V
([Lambda]) Principle Diagnosis (PD) related to alcoholism include malignancies of esophagus or liver, esophageal varices, esophagitis, Mallory Weiss, macrocytic anaemia, DTs and alcoholic dependency, abuse, intoxications, peripheral neuropathy, gastritis, liver disease and pancreatitis.
From: Gerke P, Hapke U, Rumpf H-J, John U: Alcohol related disease in general hospital patients. Alcohol Alcoholism 1997; 32(2):179-184.
Table 8 System Characteristics -- Sample Controlled No Withdrawal Withdrawal Variables (N=163) (N=74) n 24 11 Admitting Unit Aggregate Unit (GI) 10 (41.7%) 5 (45.5%) Medical unit 3 (12.5%) 3 (27.3%) Surgical unit 2 (8.3%) 2 (18.2%) ICU 1 (4.2%) 2 (18.2%) Other 2 (8.3%) 0 NO ICU LOS 87.5% 7 (63.6%) ICU LOS (range in days) 0 - 2 0 - 5 Admitting Service IM Group A 7 (29.2%) 6 (54.5%) IM/FP 13 (54.3%) 3 (27.3%) 83.5% 81.8% Surgical 4 (16.7%) 0 Medical subspecialty 0 2 (18.2%) Other 0 0 Initial Drug Lorazepam 6 (25%) 4 (36.4%) Chlordiazepoxide 0 3 (27.3%) Phenobarbital 0 0 Combination (BZD/antipsychotic) 2 (8.3%) 4 (36.4%) NO drug therapy 8 (33.3%) 0 Received Loading dose 1 (4.2%) 2 (18.2%) Soft restraints 0 5 (45.5%) Uncontrolled Defaulted Off Withdrawal Controlled AW Variables (N=54) (N=30) n 8 5 Admitting Unit Aggregate Unit (GI) 4 (50%) 2 (40%) Medical unit 3 (37.5%) 1 (20%) Surgical unit 0 2 (40%) ICU 0 0 Other 0 0 NO ICU LOS 50% 0 ICU LOS (range in days) 1 - 7 0 Admitting Service IM Group A 4 (50%) 2 (40%) IM/FP 4 (50%) 3 (60%) 100% 100% Surgical 0 0 Medical subspecialty 0 0 Other 0 0 Initial Drug Lorazepam 5 (62.5%) 2 (40%) Chlordiazepoxide 1 (12.5%) 0 Phenobarbital 1 (12.5%) 0 Combination (BZD/antipsychotic) 0 1 (20%) NO drug therapy 0 2 (40%) Received Loading dose 1 (12.5%) 0 Soft restraints 5 (62.5%) 2 (40%) Indeterminate Totals Outcome Variables (N=113) N=441 n 16 65 Admitting Unit Aggregate Unit (GI) 11 (68.8%) 32 (49.2%) Medical unit 2 (12.5%) 13 (20%) Surgical unit 1 (6.3%) 7 (10.8%) ICU 2 (12.5%) 5 (7.7%) Other 0 5 (7.7%) NO ICU LOS 14 (87.5%) ICU LOS (range in days) 1 - 2 Admitting Service IM Group A 5 (31.3%) 25 (38.5%) IM/FP 8 (50%) 31 (47.7%) 81.3% 86.2% Surgical 1 (6.3%) 5 (7.7%) Medical subspecialty 2 (12.5%) 4 (6.2%) Other 0 0 Initial Drug Lorazepam 8 (50%) 26 (40.%) Chlordiazepoxide 0 4 (6.2) Phenobarbital 0 1 (1.5%) Combination (BZD/antipsychotic) 2 (12.5%) 9 (13.8%) NO drug therapy 5 (31.3%) 15 (23.1%) Received Loading dose 1 (6.3%) 5 (7.7%) Soft restraints 2 (12.5%) Table 9 Symptom Severity -- Sample([Xi]) Controlled No Withdrawal Withdrawal Variables (N=163) (N=74) n 24 11 Day 1 HIGH SASI 4.23 [+ or -] 2.65 6.56 [+ or -] 3.50 (mean [+ or -] SD) (range) HIGH psychosis .08 [+ or -] .289 .5 [+ or -] 1.069 (mean [+ or -] SD) (range) Day 2 HIGH SASI 4.4 [+ or -] 3.22 7.636 [+ or -] 4.45 (mean [+ or -] SD) (range) HIGH psychosis 0 1.7 [+ or -] 3.368 (mean [+ or -] SD) (range) Day 3 HIGH SASI 3.62 5.33 [+ or -] 2.91 (mean [+ or -] SD) (range) HIGH psychosis .5 .89 [+ or -] 2.03 (mean [+ or -] SD) (range) Day 4 HIGH SASI 2.5 4.667 [+ or -] 3.808 (mean [+ or -] SD) (range) HIGH psychosis 0 1 [+ or -] 2.646 (mean [+ or -] SD) (range) Uncontrolled Defaulted Off Withdrawal Controlled AW Variables (N=54) (N=30) n 8 5 Day 1 HIGH SASI 7.8 [+ or -] 3.43 3.5 [+ or -] .71 (mean [+ or -] SD) (range) HIGH psychosis 1.8 [+ or -] 2.49 0 (mean [+ or -] SD) (range) Day 2 HIGH SASI 5.71 [+ or -] 2.81 4.3 [+ or -] 3.22 (mean [+ or -] SD) (range) HIGH psychosis 1 [+ or -] 1.53 0 (mean [+ or -] SD) (range) Day 3 HIGH SASI 5.83 5.25 [+ or -] 5.188 (mean [+ or -] SD) (range) HIGH psychosis 1.5 2.0 [+ or -] 2.828 (mean [+ or -] SD) (range) Day 4 HIGH SASI 4.3 3.0 [+ or -] 1.55 (mean [+ or -] SD) (range) HIGH psychosis 2 .25 [+ or -] .5 (mean [+ or -] SD) (range) Indeterminate Outcome Totals Variables (N=113) 441 n 16 64 Day 1 HIGH SASI 3.5 [+ or -] 2.20 5.08 [+ or -] 3.139 (mean [+ or -] SD) R = 0 - 7 (range) HIGH psychosis 0 .389 [+ or -] 1.153 (mean [+ or -] SD) R = 1 - 5 (range) Day 2 HIGH SASI 4.5 [+ or -] 2.8 5.429 [+ or -] 3.524 (mean [+ or -] SD) R = 0 - 9 R = 0 - 11 (range) HIGH psychosis .08 [+ or -] .29 .56 [+ or -] 1.737 (mean [+ or -] SD) R = 0 - 1 R = 1 - 7 (range) Day 3 HIGH SASI 2.5 4.194 [+ or -] 2.745 (mean [+ or -] SD) R = 1 - 7 R = 0 - 9 (range) HIGH psychosis 0 .67 [+ or -] 1.707 (mean [+ or -] SD) R = 0 - 6 (range) Day 4 HIGH SASI 3.88 3.97 [+ or -] 2.84 (mean [+ or -] SD) R = 0 - 7 R = 0 - 5 (range) HIGH psychosis 0 .83 [+ or -] 2.36 (mean [+ or -] SD) R = 0 - 6 (range)
([Xi]) Up to 60% of charts in one or more group(s) lacked documented autonomic and/or psychosis score(s).
History and clinical presentation -- the first 24 hours. The majority of sample subjects had a CAGE score recorded on admission (70%); over half were assigned an alcohol-related diagnosis, usually gastrointestinal. Fifty-eight percent had been abstinent less than 2 days; 34% were either intoxicated or had a positive ethanol toxicology screen on admission. On admission, groups destined to experience withdrawal tended to have a higher SASI score; the autonomic perturbations may simply have reflected a more severe intercurrent underlying illness. Patients were admitted to nursing units across the institution, and by policy, aggregated by their principle diagnosis. They tended to be primarily cared for by internal medicine and family practice services, and most often situated on medical nursing units. One-third had fluid and electrolyte abnormalities on admission. The majority did not require ICU care (76.9%): 9 had ICU stays of 1 or 2 days (13.9%); 5 spent 3-10 days (7.7%) and 1 patient, 3 weeks.
Management. Symptom severity was generally mild to absent. Those who experienced controlled withdrawal tended to be most symptomatic on day 2 of their episode. The large indeterminate outcome group most closely resembled on all parameters the no withdrawal group. Drug selection overwhelmingly favored the BZD; Lorazepam (Ativan[TM]) was used in 40%, chlordiazepoxide (Librium[TM]) in 6.2% subjects. An additional 13.8% had initial orders for both sedative hypnotics and antipsychotic agents. Phenobarbital and alprazolam (Xanax[TM]) were rarely used. Loading doses were rarely used (7.7%), withdrawal and nonwithdrawal groups were similarly represented. Dosing was conservative; Ativan[TM] dosages ranged from 1-4 mg (day 1), 1-24 mg (day 2), 1-8 mg (day 3) and 1-6 mg (day 4); the maximum daily dose recorded for chlordiazepoxide (Librium[TM]) was 540 mg. Haloperidol (Haldol[TM]) was used exclusively to treat positive psychoses score; the maximum daily dose was 39 mg. Over time, 61.5% (day 1), 38%(day 2), 38.5% (day 3) and 53.8% (day 4) of patients with orders received no lorazepam (AtivanTM). Twenty-three percent never received any medication for their episode; they were predominantly from the first and indeterminate outcome groups. Drug combinations were used on occasion when patients deteriorated and usually included different BZDs, or combinations of BZD and barbiturates. Neither were particularly successful in ameliorating symptoms; care tended to occur in the ICU.
Adverse effects. Attention was turned to the adverse events defined by the Task Force. We were unable to validate that most had experienced either seizures or restraint use during the 4 days of study data retrieval. Three subjects had seizure(s) in the field or upon their arrival in the ED. One patient had a documented history of epilepsy; the other had only one seizure and the final patient was suspected on the basis of history to have had a seizure awaiting transport. No seizures were documented in the 0-24 hour time period, one in the 48-72 hour period. Restraint use, on the other hand, was not uncommon. Twenty-three percent of the sample had orders for soft restraints, 1 (1.5%) for leather restraints and 1 required a sitter. One third of the soft restraints were used in the uncontrolled withdrawal group, as were the leathers. Although subjects may have been restrained after the data collection period, clearly, the incidence of significant adverse events in the "uncontrolled" withdrawal was vastly over-estimated, and our definition overly conservative.
Given the small number of adverse events, and the very small size of the uncontrolled withdrawal group, cross group differences on historical, clinical and system related variables were not statistically significant. However, soft restraint use and time abstinent were significantly different across groups; restraint use was almost nonexistent in the no withdrawal group (p=.00083). Those in the no withdrawal group and indeterminate outcome groups were far more likely to have had their last drink within 48 hours of admission. Those who did experience AWS generally had their last drink greater than 48 hours prior to admission, possibly because the illness which brought them to the hospital precluded habitual intake. That is, they came to our attention at peak risk time for experiencing withdrawal, and did. Significantly, African Americans, who had a lower incident of uncontrolled episodes, were cared for more commonly by one large group of internists who are experienced in the management of this particular clinical problem.
Examination of "indeterminate outcome" data was then undertaken. Only one participant had a threshold to treat SASI (score of 9), and it occurred on day 2. Only 1 participant, also on the second day of data retrieval, had a psychosis score greater than 0 (score of 1). The indeterminate outcome group closely resembled the no withdrawal group, that is, subjects who underwent mild or no withdrawal. Adverse events were almost nonexistent. There was one seizure in the ED and one in the 48-72 hour period. Only two patients in each group were restrained with soft restraints.
Results indicated a paucity of seizures and virtual absence of iatrogenic complications, notably, oversedation requiring pharmacologic reversal or mechanical ventilation. Several patients in the default groups were electively ventilated for reasons unrelated to their AWS. A comparative discussion of our results is limited by the number of published reports of comparable subjects. Most of the experiences described in the literature are of alcoholics admitted for the detoxification or withdrawal.[5,22] And on its face, individuals in residential rehabilitation undergoing supervised AWS and those in the community differ from those with acute medical or surgical conditions that require hospitalization who undergo withdrawal. Recent guidelines noted that "no studies were identified that report on clinical experience in managing AW in patients with specific coexisting medical or substance abuse disorders" (p.148). Recently, several groups published descriptions of protocol development processes for managing AWS.[24,26,37,39] The universal impetus to address the patient issues came from nurses, who had been frustrated with the status quo and had a strong motivation to reduce adverse outcomes for patients and caregivers. Other common experiences included fear of injury and difficulty dealing with this population which was often perceived with fear.
Patch et al and Ryan described subjects similar to ours with respect to age, gender and number of admissions with a principle diagnosis that were alcohol related. Foy et al reported 539 episodes of care; 90 subjects had principle diagnoses of either alcohol intoxication or AWS and two-thirds had secondary diagnoses of organic disease. The majority were adults with primary medical or surgical conditions, similar in age and gender to ours. Different screening and symptom severity instruments, and pharmacologic and adjunctive strategies to rationalize patient management were used, which made cross group comparisons difficult (Table 10).[24,26,37,39]
Table 10 Comparative Reports: Instruments and Outcome Measures Foy et al Screening AWS([Psi]) Symptom Severity AWS Outcome measures seizures hallucinations (in addition to demographics) delirium Patch et al Screening CAGEAID Symptom Severity CIWA-Ar([Psi][Psi]) Outcome measures assaults AMA discharges (in addition to demographics) LOS Ryan Screening CAGE Symptom Severity CIWA-Ar Outcome measures delirium restraints (in addition to demographics) sitters aftercare referrals Segatore, Adams & Lange Screening CAGE Symptom Severity SASI Outcome measures oversedations need for pharmacologic (in addition to demographics) reversal restraint use soft/leather ICU transfer seizure(s)
([Psi]) Alcohol Withdrawal Score
([Psi][Psi]) Clinical Institute Withdrawal Assessment for Alcohol - Revised by authors; established interrater reliability after modification
To the degree that all subjects required inpatient hospitalization and few required intensive care support, groups appeared to be comparable on admission with respect to acuity. Additionally, initial symptom severity scores reported by Patch and our group suggested that the majority of subjects presented with low scores that did not mandate early drug therapy. Sixty two percent of our participants showed no significant signs of AWS (groups I, VII), which may reflect a bias of less severely ill alcoholics. Foy et al reported that 79.04% had no complications from their AWS; Patch et al reported that 31% required no medication while on protocol. Serum gamma glutamyl transferase (GGT) is considered a "state marker of heavy drinking" (p.200). Approximately 40% of our sample had elevated levels on admission. Unfortunately, since we didn't consistently retrieve other liver enzymes, patterns of consumption, or coingestion of P450 inducers which can elevate the enzyme as well, it is of no real value in estimating the number of alcoholics in the sample. Management varied dramatically, from prescriptive, algorithm-driven schemes to ours, which ceded total discretion in medication selection and posology to individual physicians.
Outcomes varied widely across groups. Mean LOS data is comparable across groups at about 5 days[10,24] but reflects the interaction of the AWS, principle and any secondary diagnosis(es). As such, it is not a useful proxy marker for safe and efficient detoxification but supports the assumption of equal acuity across groups. It appears that our soft restraint use was higher than others reported;[24,26,37] 62.5% in our uncontrolled withdrawal group were restrained at some point during the data retrieval period. However, leather restraint use was comparable at under 2%. Fourteen patients (3.17%) left against medical advice (AMA) in our group, perhaps due to undermedication and their desperation for symptom relief; 2% of patients died. Patch et al determined that their deaths were not protocol related. Foy et al reported 46 deaths (8.53%); 2 occurred during AW, both reportedly due to "irreversible respiratory failure following sedation" (p.257). It is unclear whether they were directly protocol related.
Seizures occurred early and rarely in our group, unlike Foy et al who reported that 68 (12.62%) were admitted with seizures; half were alcohol-related. Both contrast sharply with 46% of individuals admitted for alcohol withdrawal who seized? The discrepancy' is due, in part, to our comparably infrequent incidences of prior episodes of AWS (26.53% versus 94%) and AWS seizures (10.66% versus 38%); however, our estimates are limited by our inability to confirm historical reports.
Hallucinations did not occur on admission. During our data collection period of 4 days, approximately 10% of subjects had positive psychosis scores; Foy et al reported late hallucinations in 90 subjects (16.7%). In several of our groups, pychotic symptoms were common -- almost two-thirds of patients who experienced uncontrolled withdrawal had positive psychosis scores. In others, psychoses scores may have been spuriously low because of the liberal use of prochlorperazine (Compazine[TM]) as an anti-emetic, which was liberally prescribed for most patients with pancreatitis and associated disorders. We used haloperidol exclusively to treat psychotic symptoms, an appropriate choice since some evidence of increased mortality has been associated with phenothiazine use in AW.[7,12,35] On first impression, our management of psychotic symptoms appears to have been flawed by widespread undermedication. However, on closer scrutiny, that impression is mitigated by 2 facts. Instrument guidelines identify a score [is greater than] 0 on the psychosis scale as requiring treatment; however, isolated disorientation generates a score that exceeds zero. While the disorientation associated with acute illness requires monitoring and may be a harbinger of delirium, it does not warrant treatment with an antipsychotic medication. A second factor that may have resulted in both underreporting and undermedication of hallucinations suggested itself to us as we supported the care of this aggregate: There was and continues to be some hesitancy or disinclination to treat "quiet" or nondisruptive delirium, even when patients were floridly hallucinating, as well as a free-floating discomfort using haloperidol in this population.
Lorazepam was the drug most commonly selected by physicians; it requires no preparation and use of a short-acting agent simplified titration. It may be the safest empiric choice for the elderly and those who might accumulate drug and suffer from toxic accumulation of metabolite.[3,14,17,18,25,32] Advantages must be weighed against a suggestion that they are more often associated with convulsions late in the detoxification process, call for more frequent assessment and are associated with increased risk of breakthrough agitation, particularly at night.[3,13,27,32] The BZDs differ in pharmacokinetic profiles and cost, not in efficacy, and at this time research does not support the superiority of lorazepam in spite of its prevalent use. Most groups favored BZD monotherapy[10,24,26] in combination with haloperidol if psychosis was present or likely to occur. Others combined BZDs of different half-lives. BZD-barbiturates combinations were rarely used; when used, they were usually unsuccessful in preventing significant adverse outcomes. Drug combinations, though used infrequently, likely reflected clinician discomfort with high dose monotherapy. On many occasions, caregivers overrated deficits in arousal due to lack of familiarity with assessment techniques; this tendency may have resulted in undertreatment and is currently being addressed.
Our Task Force took the decision to support a symptom-triggered approach to management. Although this particular regimen has NOT been studied and is not recommended for patients with seizures or acute comorbid illness, it offers theoretical and practical advantages, including individualization of therapy and reduction of drug amount and duration of exposure.[17,27,28,29] It requires reliable and frequent assessments using a valid and reliable instrument and nursing staff members with the time to do them and who understand the treatment plan.[14,19] The decision to support a symptom-triggered approach was taken in the face of the strong opinion of some that "fixed dosing interval is the gold standard for alcohol withdrawal ... [and] preferable for those patients with acute medical or surgical illness, or patients with a history of delirium tremens" (p.891). Current guidelines endorse the fixed interval approach, accepting the risk of unnecessary drug exposure for the benefit of an increased safety margin; indeed the recommendation is that if patients have a "notable comorbid medical illness, medications should be considered even if withdrawal is mild to moderate [emphasis mine]" (p.148). Endorsements for pharmacologic intervention if the subject presents with a positive past history or with existing complications, regardless of AWS severity on presentation, are becoming more common.[14,21,23,33] Front-load dosing, a variation of fixed dosing which delivers a number of intermediate or longer acting drugs until symptom resolution occurs, may be suitable for symptomatic and asymptomatic patients who are eligible for fixed dose therapy. Advantages of this approach include reduced duration of exposure and clinical monitoring.[27,36]
Based upon consultant recommendation, the SASI was selected to guide pharmacologic management. Drawbacks associated with its use include lack of established reliability and validity; item selection, revisions and cut-off scores were arbitrarily chosen and unsubstantiated. Clearly, prompt identification of deteriorating patients requires a valid instrument. The initial SASI score, which was significantly higher in those who subsequently had an uncontrolled withdrawal (p=.038), underscores the importance of the instrument in the symptom-triggered approach. High admission scores when valid and reliable instruments are used predict development of seizures and delirium.17 The SASI must be subjected to rigorous methodological assessment or its use redefined. Should it possess the psychometric properties that support its use in this population, nursing staff members house-wide will require a standardized orientation to the instrument and acceptable interrater reliability ought to be established.
In our setting, what was particularly interesting was how nurses and physicians monitored and prescribed within and around the symptom-triggered philosophy. As their experience and confidence grew, nursing staff members verbalized a suspicion that regular rather than "as needed" dosing was superior in achieving symptom control. The liberal use of PRN regimens by the physicians made this practice possible. In fact, in many instances, patients received scheduled, tapered doses of anxiolytics. Corroborating these anecdotal reports, and our sense that subjects had better outcomes when treated regularly was evidence from a documentation perspective. On days when patients received multiples doses of anxiolytics or antipsychotics, there rarely was an accompanying identical number of recorded assessments stipulated by protocol. Staffing issues may have precluded recording in some instances; disinclination or lack of familiarity with the instrument or the treatment plan may also have accounted for some deviation. By their own account, nurses relied upon subjective impressions as well as SASI scores when making treatment decisions. The authors suspect that assessment was frequent and informal, especially when patients were acutely changing. Nurses appeared to use the instrument to identify symptoms and assess treatment impact rather than trigger intervention. Routine, 3 or 4 times daily assessments, supplemented by discretionary evaluations represent a far better use of nursing time given the weakness of the instrument.
Uncontrolled withdrawal has been redefined to selectively capture only serious adverse consequences, eg. seizures or personal injury. Similarly, oversedation, which carried a similar cautious bias and, as a result, grossly overestimated the incidence of serious occurrences, was redefined to include only those patients who required pharmacologic reversal of their sedation or mechanical ventilation due to oversedation. We have reconceptualized and are tracking restraint and sitter use as proxy indicators of undermedication.
This report lacks a "before Task Force" standard of comparison. The study population also has a selection bias. Several physicians declined to have their patients enrolled in the data base, but respected management recommendations. Nurses and physicians also had full discretion regarding decisions to default patients off the plan during clinical monitoring or treatment phases. Clearly we were unable to collect and analyze all variables of clinical interest, including possibly important historical facts (eg, consumption amount, pattern of drinking). Many variables retrieved in this retrospective review were irregularly collected and lack the power to ensure the validity of some of the retrieved data. Decisions were also made to limit the chart review, which necessarily excluded collection of such data elements as second line agents used in the treatment of AWS (eg, alpha agonists, nonsedative anti-epileptics) that may have influenced outcome.[4,6] The decision to retrieve sample data for the first four days of treatment has been rationalized, but excluded adverse events that may have occurred before or after the high risk period.
It is clear that approaches of variable flexibility, but universally respectful of the biology of AWS can successfully reduce serious morbidity associated with AWS that occurs in adults with acute illnesses. In our experience, the overwhelming majority of patients identified and treated as high risk for alcohol withdrawal suffered no serious outcomes. None were mechanically ventilated or required pharmacologic reversal due to over-sedation; few had seizures. Physicians used the recommended medications conservatively but without detriment. Our nonprescriptive plan, which offered a buffet style collections of treatment choices was likely its strongest asset. The freedom to use clinical judgment, individualizing particular guidelines, has emerged as a great strength of the Task Force protocol. Its flexibility and scope of options allowed all participants maximum freedom to participate without restricting individual preference. For example, among the sample regimens printed on the order set were fixed and front-loaded dosing protocols, as well as symptom-triggered options. Professional staff willingness to experiment with unfamiliar options and collaborate across disciplines with an open, if sometimes hesitant mind, emerged as probably the greatest strength of the group who care for these patients. For example, many nurses and physicians remain uncomfortable using haloperidol in sufficient doses to produce an effect; however, that hesitancy appears to be decreasing with greater experience. The greatest disappointment arises from the clear dissociation between prescribed and actual use of the instrument and the outcomes. Psychometric testing of the SASI is mandatory; serious, balanced consideration of approaches other than the symptom-triggered approach is as pressing.
In spite of no preprotocol statistics regarding outcomes and iatrogenic complications in AW, the overwhelming impression of all involved is the marked improvement in management of those at risk. We are continuing to screen all admissions and entertain loading and front-loading tapered approaches to treatment. Most exciting is the possibility of building and testing a prediction model that will allow rapid identification and preemptive treatment of high risk individuals for severe withdrawal on admission.
We wish to acknowledge the Task Force members who completed the original work on the protocols, and the professional staff members who continue to care for this group of patients with perseverance and imagination.
[1.] Adams RD, Victor M, Ropper AH: Principles of Neurology, 6th ed. McGraw-Hill, 1997.
[2.] American Psychiatric Association: Diagnostic and Statistical Manual of Mental Disorders, 4th ed. APA, 1994.
[3.] Bird RD, Makela EH: Alcohol withdrawal: What is the benzodiazepine of choice? Ann Pharmacother 1994;28:67-71.
[4.] Brewer C: Second-line and "alternative" treatments for alcohol withdrawal: Alpha-agonists, beta-blockers, anticonvulsants, acupuncture and neuro-electric therapy. Alcohol Alcoholism 1995;30(6):799-803.
[5.] Daryanani HE, Santolaria FJ, Reimers EG et al: Alcohol withdrawal syndrome and seizures. Alcohol Alcoholism 1994; 29(3):323-328.
 Diamond L, Messing RO: Neurologic effects of alcoholism. West J Med 1994;161(3):279-287.
 Etherington JM: Emergency management of acute alcohol problems. Can Fam Physician 1996;42:2423-2431.
 Ewing JA: Detecting alcoholism, the CAGE questionnaire. JAMA 1984;252:1905-1907.
 Ferguson JA, Suelzer CJ, Eckert GJ et al: Risk factors for delirium tremens development. J Gen Int Med 1996; 11(7): 410-414.
[10.] Foy A, Kay J, Taylor A: The course of alcohol withdrawal in a general hospital. Q J Med 1997; 90(4):253-261.
[11.] Gerke P, Hapke U, Rumpf H-J, John U: Alcohol-related diseases in general hospital patients. Alcohol Alcoholism 1997;32(2):179-184.
[12.] Goodwin DW, Gabrielli WF: Alcohol: Clinical aspects. Pages 142-148 in: Substance Abuse, Lowinson JH, Ruiz P, Millman RB, Langrod JG (editors). William & Wilkins, 1997.
[13.] Gorelick DA: Management of alcohol withdrawal [letter]. Mayo Clin Proc 1995;70(12):1224.
[14.] Hall W, Zador D:The alcohol withdrawal syndrome. Lancet 1997;349(9069): 1897-1900.
[15.] Hyman SE: Current trends for alcohol withdrawal [letter]. J Gen Int Med 1995;10(9):523-524.
[16.] Kraemer KL, Mayo-Smith MF, Calkins DR: Impact of age on the severity, course, and complications of alcohol withdrawal. Arch Int Med 1997;157(19): 2234-2241.
[17.] Mayo-Smith MF: Pharmacological management of alcohol withdrawal. JAMA 1997;278(2):144-149.
[18.] Miller N: Pharmacotherapy in alcoholism. J Addictive Dis 1995;14(1):23-45.
[19.] Moberg KD: Alcohol withdrawal in the hospitalized patient. Retrieved from the World Wide Web 4/14,/98. http://www.carle.com/publications/spapers/v40no1/ selectedpfour.html.
[20.] Moore RD, Bone LR, Geller G et al: Prevalence, detection and treatment of alcoholism in hospitalized patients. JAMA 1989;261(3):403-407.
[21.] Morris PR, Mosby EL, Ferguson BL: Alcohol withdrawal syndrome: Current management strategies for the surgery patient. J Oral Maxillofac Surg 1997;55(12):1452-1455.
[22.] Newman JP, Terris DJ, Moore M: Trends in the management of alcohol withdrawal syndrome. Laryngoscope 1995;105(1):1-7.
[23.] Ozdemir V, Bremner KE, Naranjo CA:Treatment of alcohol withdrawal syndrome. Ann Med 1993;26:101-105.
[24.] Patch PB, Phelps GL, Cowan G: Alcohol withdrawal in a medical-surgical setting: The "too little, too late" phenomenon. MEDSURG Nurs 1997;6(2):79-89.
[25.] Peppers MP: Benzodiazepines for alcohol withdrawal in the elderly and in patients with liver disease. Pharmacotherapy 1996;16(1):49-58.
[26.] Ryan LA: Alcohol withdrawal? This protocol works. RN 1997; 60(8): 17-20.
[27.] Saitz R, O'Malley SS: Pharmacotherapies for alcohol abuse. Withdrawal and treatment. Med Clin North Am 1997;81(4):881-907.
[28.] Saitz R, Mayo-Smith MF, Roberts MS et al: Individualized treatment for alcohol withdrawal. JAMA 1994;272(7):519-523.
[29.] Saitz R, Smith MF: Individualizing treatment of alcohol withdrawal [letter]. JAMA 1995;273(3):183-184.
[30.] Schorling JB, Buchsbaum DG: Screening for alcohol and drug abuse. Med Clin North Am 1997;81(4): 845-865.
[31.] Schuckit MA: Alcohol related disorders. Pages 775-791 in: Comprehensive Textbook of Psychiatry. Vol I, 6th ed, Kaplan HI, BJ Sadock BJ (editors). Williams & Wilkins, 1995.
[32.] Shaw GK: Detoxification: The use of benzodiazepines. Alcohol Alcoholism 1995; 30(6):765-770.
[33.] Spies CD, Dubisz N, Funk W et al: Prophylaxis of alcohol withdrawal syndrome in alcohol dependent patients admitted to the intensive care unit after tumour resection. Br J Anaesth 1995;15: 734-739.
[34.] US Preventive Services Task Force: Guide to Clinical Preventive Devices: An Assessment of the Effectiveness of 169 Interventions. Williams & Wilkins, 1989.
[35.] Valenzuela CF, Harris RA: Alcohol: Neurobiology. Pages 119-142 in: Substance Abuse, Lowinson JH, Ruiz P, Millman RB, Langrod JG (editors). William & Wilkins, 1997.
[36.] Wasilewski D, Matsumoto H, Kur E et al: Assessment of diazepam loading dose therapy of delirium tremens. Alcohol Alcoholism 1996; 31(3):273-278.
[37.] Watling SM, Fleming C, Casey P, Yanos J: Nursing-based protocol for treatment of alcohol withdrawal in the intensive care unit. Am J Crit Care 1995;4(1):66-70.
[38.] Yost DA: Alcohol withdrawal syndrome. Am Fam Physician 1996;54(2):657-664, 669.
[39.] Zacharias S, Rodriguez-Garcia A, Honz N, Hooper C: Development of an alcohol withdrawal clinical pathway: An interdisciplinary process. J Nurs Care Qual 1998;12(3):9-18.
Questions or comments about this article may be directed to: Milena Segatore, RN, MScN, CNRN, St. Joseph's Hospital, 5000 W. Chambers St., Milwaukee, WI 53210-1688. She is a neurology clinical specialist. Debra Adams, RN, MSN, is a nurse eductor at St. Joseph's Hospital. Susan Lange, RN, BSN, is a director at St. Joseph's Hospital.3
|Printer friendly Cite/link Email Feedback|
|Author:||Segatore, Milena; Adams, Debra; Lange, Susan|
|Publication:||Journal of Neuroscience Nursing|
|Article Type:||Statistical Data Included|
|Date:||Jun 1, 1999|
|Next Article:||Antihypertensive Treatment for the Neurological Patient: A Nursing Challenge.|