Alcohol and Alcoholism Advance Access originally published online on September 16, 2006
Alcohol and Alcoholism 2006 41(6):611-615; doi:10.1093/alcalc/agl068
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
DEVELOPMENT OF A RATING SCALE TO PREDICT THE SEVERITY OF ALCOHOL WITHDRAWAL SYNDROME
1 Department of Psychiatry, Psychosomatics, and Psychotherapy, J. W. Goethe University, Frankfurt/M., Germany and 2 Department of Psychiatry and Psychotherapy, University Medical School of Lübeck, Lübeck, Germany
* Author to whom correspondence should be addressed at: Department of Psychiatry and Psychotherapy, Vivantes Klinikum Hellersdorf, (Teaching hospital of the Charité, Berlin), Myslowitzer Strasse 45, 12621 Berlin, Germany; Tel: +49/30 56803560, Fax: +49/30 56803562; E-mail: tilman.wetterling{at}vivantes.de
(Received 7 March 2006; first review notified 8 May 2006; in revised form 2 August 2006; accepted 2 August 2006)
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Aim: Various factors that may influence the severity of the alcohol withdrawal syndrome (AWS) have been identified. We tested the predictive value of these factors compiled in a newly developed scale, LARS (Luebeck alcohol withdrawal risk scale). Method: A total of 100 individuals (81 males, 19 females, mean age: 47.6 ± 9.9 years) consecutively transferred to inpatient detoxification were included in this prospective study. All fulfilled the ICD-10 criteria for alcohol dependence. The LARS was applied at the time of admission. The course of the AWS was assessed by AWS-scale at least every 4 h. The maximum AWS-score was taken as indicator of the severity of AWS. Results: The mean AWS-scoremax was 6.5 ± 3.3. In all 20% of the patients developed a severe AWS (AWS-scoremax
10). The maximum score usually occurred within 36 h after the last drink. A short version, the LARS11, was developed by statistically grounded item reduction. The optimal cut-off of the LARS11 was calculated as 10. The positive predictive value for severe AWS was 76%, while the negative predictive value was 98.7%. The sensitivity and specificity were high (95 or 92.5%, respectively). Conclusion: LARS11 assessed immediately before detoxification appears to provide a useful estimate of mild/moderate versus severe AWS, and is now ready to be validated in an independent sample. |
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
A great problem in the clinical management of alcoholics is the treatment of the alcohol withdrawal syndrome (AWS). Clinical manifestations of AWS range from mild symptoms (anxiety, mild agitation, tremor, insomnia, mild tachycardia, and hypertension) to more severe mental symptoms (like hallucinations or confusion) or complications such as seizure and cardiac arrhythmia. In view of the high costs of inpatient detoxification there is an ongoing debate about who must be treated as an inpatient and who can be treated as an outpatient. An estimate of the risk of developing a complicated withdrawal is mandatory, especially since a substantial proportion of alcoholics undergoing detoxification develops a potentially lethal delirium tremens, the rate ranging from 7 to 33% depending on setting and investigated study sample (Wetterling et al., 1994
, 1997; Ferguson et al., 1996; Palmstierna, 2001; Lee et al., 2005).
However, thus far there are only a few studies focused on prediction of withdrawal severity. The predictive power of the commonly applied Clinical Institute Withdrawal Assessment-A (CIWA-A) scale has been proven in only a few studies (Shaw et al., 1981; Sellers et al., 1983; Foy et al., 1988; Kraemer et al., 2003). Cases with initial values 15 on the CIWA-A scale (Foy et al., 1988) or CIWA-Ar values 10, respectively (Kraemer et al., 2003) ran a high risk of developing a complicated AWS. In a previous study we showed that the AWS-scale, developed from the CIWA-A scale by a statistically grounded item-reduction, also delivers a good and early prognosis for the course of AWS, if the first score was assessed upon admission (Wetterling et al., 1997).
In order to develop a scale for estimating the risk of a complicated AWS, the possible factors predicting a severe withdrawal or a delirium tremens were identified by literature analysis. There were few publications covering this topic and hardly any prospective studies. The following factors have been shown to be associated with a higher risk of severe AWS:
- — previous delirium tremens or complicated withdrawal (Wetterling et al., 1994; Wojnar et al., 1999b; Palmstierna, 2001; Fiellin et al., 2002; Kraemer et al., 2003; Lee et al., 2005)
- — number of prior detoxifications (Schuckit et al., 1995
; Wojnar et al., 1999a)
- — quantity and frequency of alcohol intake (Shaw et al., 1981
; Schuckit et al., 1995; Wojnar et al., 1999b)
- — serum electrolyte abnormalities, particularly hypokalemia (Tönnesen, 1972
; Wadstein and Skude, 1978; Beckmann, 1990; Laso et al., 1990; Jost et al., 1992; Wetterling et al., 1994)
- — use of sedative-hypnotics (Schuckit et al., 1995
; Wojnar et al., 1999b; Kraemer et al., 2003)
- — alcohol withdrawal symptoms at blood alcohol concentrations
1 g/l (Palmstierna, 2001)
- — severe somatic diseases (infections like pneumonia, coronary heart disease, alcohol liver disease, tachycardia, and anaemia) (Schuckit et al., 1995
; Wojnar et al., 1999b; Ferguson et al., 1996; Palmstierna, 2001; Lee et al., 2005), though the data concerning the type of liver diseases contributing to AWS severity are controversial (Shaw et al., 1981; Barrio et al., 2004)
- — burn injury (Lukan et al., 2002
).
- — number of prior detoxifications (Schuckit et al., 1995
Studies concerning seizure as a predictor of severe AWS gave inconsistent results (Essardas Daryanani et al., 1994; Wetterling et al., 1994, 2001; Palmstierna, 2001; Fiellin et al., 2002). ‘Alcohol markers’, carbohydrate deficient transferrin (CDT), gamma glutamyl transferase (GGT), alanine aminotransferase (ALAT), aspartate aminotransferase (ASAT), or mean red cell volume (MCV) only showed a weak correlation with the severity of AWS, and particularly the sensitivity was too low for a screening test in general hospital (Wetterling et al., 1998). In another study ASAT values 80 U/l were associated with a higher risk of complicated AWS (Kraemer et al., 2003).
The contribution of age on AWS severity is controversial. While some authors (Foy et al., 1997; Kraemer et al., 1999; Wojnar et al., 1999) found an influence of age others suggest only little contribution (Wojnar et al., 2001; Wetterling et al., 2001, 2002). Although a high proportion of alcohol-dependent people suffer from psychiatric disorders like anxiety, depression, or personality disorders (Schneider et al., 2003), there is no evidence that psychiatric comorbidity substantially influences the severity of AWS (Wetterling and Junghanns, 2000). Factors indicating bad social conditions like homelessness may contribute to the development of delirium (Ferguson et al., 1996).
As a significantly higher risk was observed in subjects with more than one of the above mentioned factors (Ferguson et al., 1996; Lee et al., 2005; Kraemer et al., 2003) we decided to construct a new scale LARS (Luebeck Alcohol withdrawal Risk Scale, see Table 3) which is a compilation of those possible risk factors for AWS severity that can easily and routinely be assessed upon admission such as the history of previous detoxifications, the recent drinking behaviour and any complications during the past 4 weeks which patients usually can remember well.
The present study was aimed to evaluate the clinical feasibility of the LARS in an unselected sample of patients with chronic alcohol dependence undergoing inpatient detoxification.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Patients
The sample consisted of patients consecutively referred to a psychiatric ward of the General University Hospital in Frankfurt specialized for the detoxification of all alcoholics without serious concurrent internal or surgical disorders. The hospital has a catchment area of
192 000 people. 100 individuals (81 males, 19 females, mean age: 47.6 ± 9.9 years) were included in this prospective study. All fulfilled the ICD-10 criteria for alcohol dependence (WHO, 1992). Patients gave informed consent to participate. Patients with severe cognitive disorders were excluded. The sample is characterized in Table 1.
|
Procedure
At admission the blood alcohol concentration was assessed by a breathalyzer (Draeger Alcotest 7110, Lübeck, Germany). The electrolyte serum concentrations were measured by routine methods from EDTA blood samples collected upon admission to treatment.
The LARS (Table 3) was completed at the time of admission. Patients then underwent a symptom-triggered detoxification treatment based on the results of a structured assessment with the AWS-scale (Wetterling et al., 1997), rated by trained nurses every 4 h at minimum. The AWS allows a separate assessment of somatic as well as mental withdrawal symptoms. As described in our previous publication the medication (carbamazepine or clomethiazole) was chosen according to the first AWS rating (within 2 h of admission). The maximum AWS sum score during index withdrawal (AWS-scoremax) was taken as measure of withdrawal severity.
Statistics
All statistical tests were calculated using of SPSS-PC program package, version 12.0 (SPSS, Chicago). The internal consistency was assessed by Cronbach's alpha and the corrected item score—total score correlation. The sensitivity, specificity, and the positive and negative predictive value (PPV/NPV) were calculated according to the following formulae (all values are given in %):
- Sensitivity = true positives/(true positives ± false negatives) x 100.
- Specificity = true negatives/(true negatives + false positives) x 100.
- PPV = true positives/all patients with positive test results x 100.
- NPV = true negatives/all patients with negative test results x 100.
- Specificity = true negatives/(true negatives + false positives) x 100.
In all tests P-values > 0.05 were classified as not significant.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Description of the sample
Across the sample the reported mean duration of alcohol dependence was 13.9 ± 8.0 years with a minimum duration of 2 years (Table 1). In comparison to males female smoked less frequently and showed a significantly lower number of previous detoxifications. They drank significantly less alcohol during the drinking period prior to index detoxification (191 ± 82 versus 330 ± 165 g/day) (ANOVA df = 1, F = 12.87, P = 0.001), even if the amount per kg bodyweight was calculated. However, there was no difference with respect to the rate of alcohol intoxications at index admission (
84%), blood alcohol concentration and rate of delirium tremens before admission.
Severity of alcohol withdrawal syndrome
The mean AWS-scoremax was 6.5 ± 3.3. In all 25% developed a mild AWS (AWS-scoremax 
5), 55% a moderate AWS (AWS-scoremax 6–9), and 20% a severe AWS (AWS-scoremax 10), the maximum score usually occurring within 36 h after the last drink. Six subjects developed delirium tremens. Females scored non-significantly higher than males.
Development of a scale (LARS) for the prediction of severe AWS
The initial LARS scale (Table 3) applied to the patients consists of 22 items compiled in 4 subscores:
- LARS I: history of alcohol use (detoxification and withdrawal complications) and abuse of sedatives;
- LARS II. drinking habits and possible alcohol use complications during the past 4 weeks; in this subscale some items assumed to indicate bad physical condition prior to detoxification (malnutrition, sleep disturbances, etc.) were included;
- LARS III: clinical symptoms at clinical investigation;
- LARS IV: laboratory values of serum electrolytes at admission.
- LARS II. drinking habits and possible alcohol use complications during the past 4 weeks; in this subscale some items assumed to indicate bad physical condition prior to detoxification (malnutrition, sleep disturbances, etc.) were included;
All four LARS subscores as well as the total LARS score significantly differentiated patients with severe withdrawal (AWS-scoremax 10) from patients with mild to moderate withdrawal (AWS-scoremax <10) (Table 2). There was no difference between groups in mean age.
|
In order to identify those LARS items providing the best prediction of complicated withdrawal the positive and negative predictive values as well as the sensitivity and specificity were calculated (data not shown in detail). All single items only revealed low PPVs (max. 56.7% for chloride) while some yielded high NPVs (max. 100% for most LARS II items and BAC
1 g/l, and tremor, sweating and pulse rate 100/min when BAC is 1 g/l).
The optimal cut-off of the total LARS score was 16:76 of the 80 subjects developing only mild or moderate AWS showed a lower LARS score, while all 20 alcoholics developing a severe AWS had a LARS score of 17 (
2 = 60.2 P < 0.001).
Thus, the sensitivity was 100%, while the specificity was 88%. The PPV was 69%, and the NPV 100%.
In order to test the reliability of the LARS scale the internal consistency of the items was proven by calculating Cronbach's alpha. A value of Cronbach's alpha >0.75 was considered to demonstrate a good consistency with the whole score. Furthermore, the corrected item score—total score correlation was estimated (Table 3). In the next step items with a corrected item score—total score correlation r < 0.4 were excluded. The remaining 11 items were compiled in new shorter version of the LARS, the LARS11 (Table 3), at a cut-off score of 10, PPV was now 76%, NPV, 98.7%; sensitivity, 95%; and specificity, 92.5%. Since in clinical practice the chloride concentration in serum may not easily be available, we also calculated the values for a scale without chloride LARS10 (cut-off, 9): PPV, 67.9%; NPV, 98.6%; sensitivity, 95%; and specificity, 88.8%.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
In clinical management an early and reliable prognosis of subsequent AWS severity would help to decide what interventions are adequate to ensure a safe detoxification with a good cost–benefit ratio. This prospective study showed that a compilation of easily available data known to be individually associated with an increased risk of severe AWS (the LARS) yields a good prediction of subsequent withdrawal severity. The calculated predictive positive and negative values of the LARS and particularly of the LARS11 and LARS10 that both were developed by statistically grounded item-reduction suggests that these scales could potentially provide a clinical means to estimate the risk of an imminent severe alcohol withdrawal. LARS11 and LARS10 now need to be tested in an independent sample, in outpatient alcohol detoxification and in samples that include younger, and older, patients.
LARS is easy to administer and potentially can improve treatment decisions in emergency rooms and be of economic value as it can help to reserve more intensive therapies for selected cases.
One of the most severe forms of AWS is delirium tremens (Wetterling et al., 1994, 1997; Ferguson et al., 1996; Palmstierna, 2001; Lee et al., 2005). It is commonly regarded as an acute brain syndrome mainly characterized by hyperexcitability of the central nervous system. The factors indicating a high risk of complicated AWS identified in this study were signs of a damage of the central and peripheral nervous system: previous delirium and/or seizures, ataxia, and polyneuropathy and possibly also of hyperexcitability of the CNS (sleep disturbances and nightmares). The other factors found in this study reflect a dysbalanced homeostatis: the occurrence of withdrawal symptoms like tremor, sweating, and pulse rate >100/min, despite of BAC >1 g/l suggest a habituation at higher alcohol levels. The latter risk factors were also found by some other authors (Palmstierna, 2001; Lee et al., 2005). Decreased serum concentrations of chloride the main ‘inhibiting’ ion in the CNS can also be interpreted as an indicator of hyperexcitability.
The number of prior detoxifications showed a good internal consistency as assessed by Cronbach's alpha, but no high item-scale correlation. So although other studies (Schuckit et al., 1995; Wojnar et al., 1999a) found an association with the severity of AWS, this item was not included in the LARS11. Our results may be due to the stratification of this item in the LARS.
There is an important limitation with respect to the use of LARS11: the participants of this study were without serious concurrent medical or surgical disorder and predominately aged between 40 and 55 years. In severely medically ill persons predominantly referred to the internal or surgical department and in young and old alcoholics the clinical feasibility of the LARS has not been appraised yet. Medical conditions like pneumonia, coronary heart disease, alcohol liver disease, and anaemia, and limited physiologic reserve were shown to increase the risk of severe AWS (Schuckit et al., 1995; Ferguson et al., 1996; Wojnar et al., 1999b; Palmstierna, 2001; Lee et al., 2005) and might outweigh the prediction of LARS11. However, our sample also included some cases developing delirium tremens, but no Wernicke–Korsakoff syndrome. Although all cases received a standardized treatment (also including routine parental vitamin B12 administration if AWS-score 10), bias of AWS-score cannot completely be excluded by treatment effects, but treatment was needed for ethical reasons.
To our knowledge there are no comparable scales predicting the severity of AWS. Further studies are required to validate our findings with LARS in other samples.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Barrio, E., Tome, S., Rodriguez, I. et al. (2004) Liver disease in heavy drinkers with and without alcohol withdrawal syndrome. Alcoholism: Clinical and Expertimental Research 28, 131–136.
Beckmann, J. (1990) [Alcohol withdrawal delirium and hypokalemia].[in German] Nervenarzt 61, 444–446.[Web of Science][Medline]
Essardas Daryanani, H., Santolaria, F. J., Gonzalez Reimers, E. et al. (1994) Alcoholic withdrawal syndrome and seizures. Alcohol and Alcoholism 29, 323–328.
Ferguson, J. A., Suelzer, C. J., Eckert, G. J. et al. (1996) Risk factors for delirium tremens development. Journal of General Internal Medicine 11, 410–414.[Web of Science][Medline]
Fiellin, D. A., O'Connor, P. G., Holmboe, E. S. et al. (2002) Risk for delirium tremens in patients with alcohol withdrawal syndrome. Substance Abuse 23, 83–94.
Foy, A., Kay, J. and Taylor, A. (1997) The course of alcohol withdrawal in a general hospital. QJM 90, 253–261.
Foy, A., March, S. and Drinkwater, V. (1988) Use of an objective clinical scale in the assessment and management of alcohol withdrawal in a large general hospital. Alcoholism: Clinical and Expertimental Research 12, 360–364.
Jost, A., Hermle, L., Spitzer, M. et al. (1992) [Clinical and laboratory differentiation of alcohol withdrawal syndrome (‘predelirium’) and alcoholic delirium] [in German] Psychiatrische Praxis 19, 16–22.[Web of Science][Medline]
Kraemer, K. L., Mayo-Smith, M. F. and Calkins, D. R. (1997) Impact of age on the severity, course, and complications of alcohol withdrawal. Archives of Internal Medicine 157, 2234–2241.
Kraemer, K. L., Mayo-Smith, M. F. and Calkins, D. R. (2003) Independent clinical correlates of severe alcohol withdrawal. Substance Abuse 24, 197–209.
Laso, F. J., Gonzalez-Buitrago, J. M., Martin-Ruiz, C. et al. (1990) Inter-relationship between serum potassium and plasma catecholamines and 3':5'cyclic monophosphate in alcohol withdrawal. Drug and Alcohol Dependence 26, 183–188.[CrossRef][Web of Science][Medline]
Lee, J. H., Jang, M. K., Lee, J. Y. et al. (2005) Clinical predictors for delirium tremens in alcohol dependence. Journal of Gastroenterology and Hepatology 20, 1833–1837.[CrossRef][Web of Science][Medline]
Lukan, J. K., Reed, D. N. Jr, Looney, S. W. et al. (2002) Risk factors for delirium tremens in trauma patients. Journal of Trauma 53, 901–906.
Palmstierna, T. (2001) A model for predicting alcohol withdrawal delirium. Psychiatric Services 52, 820–823.
Schneider, U., Altmann, A., Baumann, M. et al. (2001) Comorbid anxiety and affective disorder in alcohol-dependent patients seeking treatment: the first multicentre study in Germany. Alcohol and Alcoholism 36, 219–223.
Schuckit, M. A., Tipp, J. E., Reich, T. et al. (1995) The histories of withdrawal convulsions and delirium tremens in 1648 alcohol dependent subjects. Addiction 90, 1335–1347.[CrossRef][Web of Science][Medline]
Sellers, E. M., Naranjo, C. A., Harrison, M. et al. (1983) Oral diazepam loading: simplified treatment of alcohol withdrawal. Clinical Pharmacology and Therapy 34, 822–826.
Shaw, J. M., Kolesar, G. S., Sellers, E. M. et al. (1981) Development of optimal treatment tactics for alcohol withdrawal. Assessment and effectiveness of supportive care. Journal of Clinical Psychopharmacology 1, 382–388.[Web of Science][Medline]
Tönnesen, E. (1972) Delirium and hypokalaemia. Lancet ii, 97.
Wadstein, J. and Skude, G. (1978) Does hypokalaemia precede delirium tremens? Lancet ii, 549.
Wetterling, T. and Junghanns, K. (2000) Psychopathology of alcoholics during withdrawal and early abstinence. European Psychiatry 15, 483–488.[CrossRef][Web of Science][Medline]
Wetterling, T., Kanitz, R.-D., Veltrup, C. et al. (1994) Clinical predictors of alcohol withdrawal delirium. Alcoholism: Clinical and Experimental Research 18, 1100–1102.[CrossRef][Web of Science][Medline]
Wetterling, T., Kanitz, R.-D., Besters, B. et al. (1997) A new rating scale for the assessment of the alcohol withdrawal syndrome (AWS-Scale)—as a basis for a score—controlled inpatient treatment. Alcohol and Alcoholism 32, 753–760.
Wetterling, T., Kanitz, R.-D., Renner, F. et al. (1998) Does CDT predict the severity of alcohol withdrawal syndrome? Alcoholism: Clinical and Experimental Research 22, 1053–1056.[CrossRef][Web of Science][Medline]
Wetterling, T., Driessen, M., Kanitz, R.-M. et al. (2001) The severity of alcohol withdrawal is not age-dependent. Alcohol and Alcoholism 36, 75–78.
Wetterling, T., Backhaus, J. and Junghanns K. (2002) [Addiction in the elderly—an underestimated diagnosis in clinical practice?] [in German] Nervenarzt 73, 861–866.[CrossRef][Web of Science][Medline]
Wojnar, M., Bizon, Z. and Wasilewski, D. (1999a) Assessment of the role of kindling in the pathogenesis of alcohol withdrawal seizures and delirium tremens. Alcoholism: Clinical and Experimental Research 23, 204–208.[CrossRef][Web of Science][Medline]
Wojnar, M., Bizon, Z. and Wasilewski, D. (1999b) The role of somatic disorders and physical injury in the development and course of alcohol withdrawal delirium. Alcoholism: Clinical and Experimental Research 23, 209–213.[CrossRef][Web of Science][Medline]
Wojnar, M., Wasilewski, D., Zmigrodzka, I. et al. (2001). Age-related differences in the course of alcohol withdrawal in hospitalised patients. Alcohol and Alcoholism 36, 577–583.
World Health Organization (1992) ICD-10 (Chapter F). Geneva, World health organization.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  B. Habermeyer, M. Weiland, R. Mager, G. A. Wiesbeck, and F. M. Wurst Letter to the Editor * A clinical lesson: Glioblastoma multiforme masquerading as depression in a chronic alcoholic Alcohol Alcohol., January 1, 2008; 43(1): 31 - 33. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||