Alcohol withdrawal syndrome (AWS) is characterized by varied symptoms that range from mild to severe intensity depending on several factors including the quantity, frequency and duration of alcohol intake, and the number of prior withdrawal episodes, as well as individual differences in the vulnerability [
Subtyping of the AWS has been attempted in the past, as Gross [
Several rating instruments have been used to measure severity of alcohol withdrawal [
In our setup, we use CIWA-Ar as part of the measure for the management of alcohol withdrawal symptoms. It is generally observed that alcohol withdrawal symptoms fluctuate in presentation and severity across time. The present study was carried out to explore the dimensionality of this scale in an attempt to identify a set of underlying factors that exist and can explain the interrelationships among various manifestations of acute alcohol withdrawal symptoms. The knowledge of these underlying factors may enhance our understanding of AWS and better prediction of complications thus management plans.
This was a cross-sectional hospital-based study, conducted at Centre for Addiction Psychiatry, Central Institute of Psychiatry, Ranchi, India, a tertiary care referral centre during May 2005 to June 2006. The study was approved by the institutional review board. Study sample included 201, only male fulfilling ICD-10 DCR (World Health Organization) [
Patients admitted with alcohol dependence syndrome with acute withdrawal were evaluated with CIWA-Ar immediately after admission then every six hours, as a routine protocol of the ward. Written informed consent to participate for the study was obtained from all the patients. The sociodemographic details were obtained from patient and their relatives. Detailed physical examination, mental status examination, and planned screening laboratory investigation were done to ensure conformity of study criteria. The patients get admitted with varying lengths of abstinence, ranging from hours to days, so we initially took all the ratings of all the patients and arranged them as rating at first 6 hours of abstinence and at every six hours like at 6, 12, 18, 24, 30, 36, 42, 48, 52, 58, and 64 hours till CIWA-Ar scoring reaches below 10. The averages of each rating of CIWA-Ar scores were computed to see the severity of withdrawal symptoms across time span of abstinence. Many patients were admitted after overnight, 12 to 18 hours of abstinence, so we included the patients who were admitted with at least 24 hours of abstinence. Hence rating at the 24 hours of abstinence was considered as first rating. Meanwhile management and medications were continued as per ward protocol and no adjustment for study purpose was done. The standard detoxification protocol included thiamine supplementation, benzodiazepines either lorazepam or diazepam, and correction of fluids and electrolytes if any and other symptomatic treatment of associated conditions like dyspepsia or concurrent injury, wound, and infections.
The collected data on 201 patients was statistically analyzed, using statistical package for social sciences (SPSS, Inc., Chicago, Illinois) version 10.0 for Windows. Exploratory factor analysis (maximum likelihood method) was carried out to identify factor structure on all items of CIWA-Ar for day three. Kaiser-Meyer-Olkin measure of sample adequacy and Bartlett’s test of sphericity were also done to assess appropriateness of conducting factor analysis. Two criteria for retaining the number of components were considered: Kaiser’s criterion [
Table
Sample characteristics (
Variables | Mean | SD |
---|---|---|
Age | 37.18 | 9.35 |
Age of onset of drinking (in years) | 21.63 | 4.99 |
Duration of dependence (in years) | 6.49 | 5.06 |
Average amount (in mL) | 1266.59 | 870.85 |
Maximum amount (in mL) | 1802.36 | 1191.72 |
Last intake (hours before admission) | 13.64 | 9.27 |
|
||
Variables |
|
% |
|
||
Gender | ||
Male | 201 | 100 |
Marital status | ||
Married | 170 | 84.6 |
Single | 31 | 15.4 |
Education | ||
Illiterate | 18 | 9.0 |
Up to class 10 | 71 | 35.3 |
Above class 10 | 112 | 55.7 |
Habitat | ||
Rural | 25 | 12.4 |
Urban | 132 | 65.6 |
Semiurban | 44 | 22.0 |
Occupation | ||
Unemployed | 34 | 17.0 |
Employed | 167 | 83.0 |
Socioeconomic status (monthly income in Rs) | ||
Lower (upto 5000) | 68 | 34.0 |
Middle (5000–20000) | 123 | 61.0 |
Higher (above 20000) | 10 | 5.0 |
Past history of detoxification | 94 | 46.8 |
Past history of withdrawal seizure | 46 | 22.9 |
Past history of delirium tremens | 4 | 2.0 |
Family psychiatric illness | 24 | 12.0 |
Family substance dependence | 133 | 66.2 |
The item frequency and mean of all six hourly CIWA-Ar ratings were calculated; the mean scores of CIWA-Ar at 24 hours and at 36 hours are shown in Table
CIWA-Ar items frequency, mean, and SD at 24 hours and 36 hours (
CIWA-Ar items | 24 hours | 36 hours | ||||
---|---|---|---|---|---|---|
Present% | Mean | SD | Present% | Mean | SD | |
(1) Anxiety | 91 | 3.67 | 2.06 | 91 | 2.96 | 1.92 |
(2) Nausea | 25.9 | .51 | 1.09 | 58.7 | .92 | .95 |
(3) Paroxysmal sweats | 69.2 | 1.31 | 1.54 | 75 | 1.04 | .86 |
(4) Headache and fullness in head | 28.9 | .44 | .86 | 12.9 | .17 | .49 |
(5) Tremor | 94.5 | 3.82 | 1.99 | 94.5 | 4.62 | 1.78 |
(6) Visual disturbances | 30.3 | 1.01 | 1.75 | 53.2 | .72 | .90 |
(7) Auditory disturbances | 9.5 | .29 | 1.38 | 55.1 | 3.35 | 2.54 |
(8) Tactile disturbances | 10 | .30 | 1.33 | 65.7 | 2.25 | 2.20 |
(9) Orientation and clouding of sensorium | 11 | .29 | .95 | 49.3 | 1.05 | 1.30 |
(10) Agitation | 65.7 | 1.63 | 1.99 | 80.6 | 3.30 | 2.06 |
|
||||||
Total CIWA-Ar score | 13.32 ± 9.27 | 20.4 ± 9.09 |
Factor analysis (extraction method-maximum likelihood) with the 10 items of CIWA-Ar for day three, resulted in initial three factors with eigenvalues greater than unity. The scree plot was also showing clear inflexion, supporting three factors (Figure
Factor analysis (maximum likehood) with varimax rotation showing factor structure of CIWA-Ar (
CIWA-Ar items | Factor loading | % positive |
Item scores |
|||
---|---|---|---|---|---|---|
Delirious | Autonomic | Nonspecific | Mean | SD | ||
Tactile disturbances |
|
65.7 | 2.26 | 2.20 | ||
Auditory disturbances |
|
.150 | .163 | 75.1 | 3.35 | 2.54 |
Orientation and clouding of sensorium |
|
−.137 | 49.3 | 1.06 | 1.30 | |
Agitation |
|
.120 | 80 | 3.30 | 2.06 | |
Anxiety |
|
91 | 2.96 | 1.92 | ||
Paroxysmal sweats |
|
.277 | 75 | 1.05 | .86 | |
Tremor |
|
.211 | 94.5 | 4.63 | 1.78 | |
Headache and fullness in head | .172 |
|
12.9 | .18 | .49 | |
Nausea and vomiting | −.156 | .375 |
|
58.7 | .93 | .95 |
Visual disturbances | .233 |
|
53.2 | .72 | .90 | |
Eigenvalue | 3.4 | 2.4 | 1 | |||
Variance (%) | 34.34 | 24.25 | 10.04 | Total = |
||
Factor mean (SD) | 9.97 (7.4) | 8.81 (3.96) | 1.64 (1.41) | |||
Cronbach’s alpha | .91 | .66 | .26 |
Scree plot, showing three factors above eigenvalue of one and showing clear inflexion of the graph.
The first factor named as “delirious factor,” which had highest loading from tactile disturbances (.999), followed by auditory disturbances, orientation and clouding of sensorium, and agitation. It explained 34.34% of variance and showed good internal consistency (Cronbach’s alpha = .91). The second factor named as “autonomic factor” reflected four-item loading, highest from anxiety, followed by paroxysmal sweats, tremor, and headache or fullness in head. It explained 24.25% of variance and showed moderate internal consistency (Cronbach’s alpha = .66). The third factor named as “nonspecific factor” reflected two-item loadings, nausea, and visual disturbances that explained 10.04% of variance and a Cronbach’s alpha of .26.
We examined the factor structure of the CIWA-Ar in a population of adult men hospitalized to a tertiary psychiatric institute for treatment of alcohol dependence. The ideal study of AWS could have been the assessment before starting medication, but this was practically not possible for many reasons. Firstly, many of the patients come for admission after 12 to 18 hours of abstinence and severe withdrawal, so keeping them drug free was ethically not possible. Thus natural AWS presentation and its severity were masked by routine benzodiazepam administration and thiamine supplement. Secondly, many other patients were referred from primary care centers with initial management, including long acting benzodiazepines like diazepam that masks the AWS. Thirdly, many other patients came even before onset of withdrawal and in a state of intoxication; the AWS was not fully evolved in terms of range of symptoms and severity. For all these reasons, the mean CIWA-Ar score for the initial 24 hours of abstinence (first day) of admission was only 13.32 (SD 9.27). Later the sequential rating found more prominent withdrawal symptoms reaching highest mean score of 20.4 at thirty-six hours then gradually decreased. As drug free AWS was not possible, we consider that the higher mean score of CIWA-Ar represents the AWS better than low score. Also as both AWS medications and alcohol itself are CNS depressant and act in a similar way, either medication or alcohol intake should not make much difference in clinical picture. So we decided to proceed for factor analysis with highest mean CIWA-Ar scoring at the 36th hour.
The severity of withdrawal symptoms and appearance of complete sets of withdrawal symptoms at the 36th hour may have been influenced by plasma half-life of benzodiazepams being used for detoxification. In this study, choice of medication was with treating team of the hospital; however, only either intermediate acting lorazepam or long acting diazepam was used for this purpose. There was no use of short acting benzodiazepines, which causes varying and rebound withdrawal symptoms across different time frame with its dosing and changing plasma concentration. Another more important influencing reason for varying withdrawal symptoms across different time frame could have been the dose of benzodiazepines, but we did not interfered with any medications or dosing and it was continued as per ward protocol to ensure naturalistic conditions. However, the equivalent benzodiazepines mean dose at 36 hour was 30 mg of diazepam per day. But this equivalent dose may not be accurate as many patients were on oral medications and others were on parenteral benzodiazepines.
We excluded any other comorbid substance use disorders or polysubstance dependence but a few patients may also have undisclosed benzodiazepine or organic inhalant abuse or addiction. These medications with CNS depressant effect do mask and modify the withdrawal symptoms. We also excluded any comorbid general medical condition especially epilepsy for that reason, patients on antiepileptic either taking it regularly or skipping will modify the alcohol withdrawal symptoms. For that matter any psychotropic drugs causing CNS depression or any effecting stimulants will alter the withdrawal symptoms. Most of the patients needed proton pump inhibitor drugs like pantoprazole or omeprazole for the alcohol induced dyspepsia, peptic ulcer disease, or gastroesophageal reflux disease, but these medications do not impose any effect on alcohol withdrawal symptoms.
In a previous study by Pittman et al. [
We found a three-factor solution based on rotated eigenvalues and scree plot analysis. The first factor explained 34.34% variance and consisted of four items, namely, tactile disturbances, auditory disturbances, orientation and clouding of sensorium, and agitation. This factor appears to represent
The second “autonomic” factor explained 24.25% of variance and consisted of four items: anxiety, paroxysmal sweats, tremor, and headache and fullness in head with factor scores of .998, .660, .528, and .245 respectively. Cronbach’s alpha for this factor was 0.66 showing adequate internal consistency. Within this factor the highest loadings was with anxiety. It probably represents mixed mechanism of CNS rebound hyperactivity along with adrenal hyperactivity. This factor may be a result of the practice of not using adrenergic medication on routine basis at our institute. This factor was in accordance to study of Pittman et al. [
There are three proposed physiologic bases for the symptom manifestation of alcohol withdrawal symptoms: CNS excitation, adrenergic hyperactivity, and delirium, which may be attributed to different neurotransmitters and they respond selectively to different pharmacotherapy [
Strength of our study includes large sample size and not interfering with any medications or management strategies thus providing setting of naturalistic conditions. The use CIWA-Ar is the most widely accepted alcohol withdrawal assessment scale and selection of abstinent hours was important to allow time for full appearance of symptoms, even though under cover of detoxification medication. There was for better coverage and inclusion of withdrawal symptoms at 36 hours as indicated by total CIWA-Ar score and item frequency.
These results have wider implications for the recognition and management of AWS, particularly, for better understanding and identification of the symptom profiles for and differential management plans across subtypes of AWS. The use of adrenergic antagonists may have a valuable role in addition to benzodiazepines, in a set of patients with autonomic features. One of the limitations in our study is that it includes male only patients; however, gender can be an important issue in AWS presentation and its severity. Even studies found that sex hormone affects the AWS by modulating the function of the GABA-A receptor [
It is known that autonomic arousal is an important mechanism in AWS; thus, other physiological measures and biological markers for objective assessment may be included in future studies. Further studies may also be carried out including cases of mild to moderate severity and in both sex to uncover the differences. Also, factor analysis depends heavily on the population studied; therefore, studies on different population may be required to generalize our findings. For the clinical practice, it is advisable not to overdepend on rating scales and it must not replace a thorough clinical evaluation of the patient’s medical status in prediction of those at risk of severe alcohol withdrawal.
The acute alcohol withdrawal symptoms was most severe at 36 hours of abstinence in our sample. This study finds multidimensionality of alcohol withdrawal symptoms as measured with CIWA-Ar; we found three factors explaining 68.74 percentage of variance and named as delirious, autonomic and nonspecific. These factors of the CIWA-Ar represent high internal consistency among the items.
The authors report no conflict of interests. The authors alone are responsible for the content and writing of this paper.