Acta psychiat. scand. (1979) 60, 393-404 Department of Psychiatry (Heads: Prof. Y. Lunn, Prof. 0. J . Rafaelsen, T. Vunggaard, M.D.), Rigshospitalet, Copenhagen, Denmark

Delirium tremens SOME CLINICAL FEATURES. PART I P. KRAMPAND R. HEMMINGSEN Twenty patients with delirium tremens and related states were investigated from the time of admission until the acute state was over. Using strict diagnostic criteria the material was divided into two groups according t o the severity of the clinical condition; nine patients had fully developed delirium tremens (grade 3), 11 patients had a less severe clinical state, known as “Predelirium” (grade 2). The material was found t o be representative for the condition under discussion. Patients with grade 2 were admitted during the day and the evening, but not during the night; patients with fully developed delirium tremens (grade 3) were admitted during both the day and the night, and this difference is discussed. The two groups had the same pattern of alcohol abuse, but patients with proper delirium tremens had had a drinking bout preceding the acute state; this was not seen among patients with a less severe clinical state. Patients with grade 2 had had symptoms like sleep disturbances and gastro-intestinal disturbances for 12-24 hours before the admission, whereas grade 3 patients had had such symptoms for about 48 hours. Patients with grade 2 stopped drinking when the first symptoms of the acute state appeared, whereas patients with fully developed delirium tremens continued t o drink in spite of their condition. These anamnestic data were supported by the finding that significantly more patients with proper delirium tremens had alcohol in the blood - several even concentrations about 2 gA - at the time of admission compared t o patients with a less severe clinical state. This lack of “latency period”, which previously has been described as a typical feature in the development of delirium tremens, is discussed. It is concluded that due t o the methodology used, it has been possible t o point out some of the differences between the milder, often harmless, conditions and the potentially serious, fully developed delirium tremens. Key words: Blood alcohol concentration - delirium tremens drinking bout - methodology - preceding symptoms - time of admission.

Delirium tremens (DT) has been known f o r nearly 200 years (Romano (1941)), and the term still covers precisely the same condition as when first described a short lasting psychotic condition related to alcohol abuse; it m a y be dis-

0001-690X/79/100393-12$02.50/0 @ 1979 Munksgaard, Copenhagen

394 tinguished from other delirious states due to the typical clinical picture. It is thus one of the oldest syndromes known in modern psychiatry. Although DT has been recognized as a clinical entity for almost two centuries many aspects of the syndrome are still under discussion (Hemmingsen et al. (1979)). The relation between a prolonged alcohol abuse and DT is well documented (Cutshall (1965), Nielsen (1965), Salum (1972)), but whether DT is a withdrawal psychosis or not is unresolved. The typical features of a fully developed DT have been described by many (Sqjrensen (1959), Lundquist (1961), Nielsen (1965), Salum (1972), Thompson et al. (1975), Keller (1977), Thompson (1978)). After a prodromal state with sleep disturbances, apprehension, restlessness and often nausea and vomiting, the patient becomes increasingly agitated with heavy perspiration, intense tremor, tachycardia and often elevated temperature. He develops vivid hallucinations, predominantly visual and tactile and often of a terrifying nature - the patient sees animals in his surroundings and feels them on his body. Disorientation and suggestibility are outstanding symptoms; the patient does not know the correct date, he does not know where he is, and he can be constantly, intensely occupied with imaginary activities, talking to imaginary persons. In the classical descriptions of DT, the “occupational delirious state”, which could include parts of the patient’s usual job, was often stressed; the patient became completely exhausted, feverish and died in the state of cardio-vascular collapse. In an old Danish description of DT (Pontoppidan (1895)) it was pointed out that this “occupational delirious state” often was less dramatic in Denmark and this was thought to be due to the “indolent, unimaginative, simple-minded Danish national character”. But even without such an “occupational delirious state”, the most severe cases of DT still have to be considered as serious, perhaps life-threatening conditions, if not treated adequately (Thompson (1978)). The duration of the acute psychosis is most often less than 72 hours (Cutshall (1965), Salum (1972), Kramp & Rafaelsen (1978)) and drug treatment does not seem to have changed this course dramatically (Cutshall (1965)). However, the patient’s condition during the acute state, the importance and duration of the various physical and psychic symptoms and possible relations between them are to a large extent unknown, and statements concerning these symptoms more often seem to be based on tradition and clinical experience than on well-documented studies. The purpose of the present study was to investigate patients with DT and less severe clinical states related to alcohol abuse as well clinically as chemically, and hereby to try to clarify some of the many unresolved questions of the syndrome. Using strict inclusion criteria a group of patients were thoroughly investigated from the hour of admission until recovery from the acute state. The investigational program included clinical observation and gradation as well as clinicochemical investigations of blood and cerebrospinal fluid (CSF). In this paper mainly the method used and a description of the material will be presented. A following paper (Kramp et al. (1979)) will present the results of the clinical observations during the acute state. The results of the various laboratory investigations will be presented in subsequent papers.

395

MATERIAL The study comprises all patients admitted from November 1976 to September 1977 to the Department of Psychiatry, Rigshospitalet, with DT or related clinical conditions, who fulfilled the inclusion criteria, who did not meet the exclusion criteria, and who accepted to participate in the study. Inclusion criteria 1) The patient should have a history of alcohol abuse. 2) The actual condition should be related to the abuse. 3) The actual condition should be an acute event; patients with more chronic disabilities related to alcohol abuse, e.g. alcohol hallucinosis, were not included. 4) As a minimum symptoms should comprise intense gross tremor of the extremities, intense perspiration and hallucinations. Exclusion criteria 1) Intake of psychopharmaca ( = psychotherapeutic drugs) (prescribed by physicians or taken spontaneously by the patients themselves), during the last 24 hours before admission. Such patients were excluded because we wanted to investigate the condition “in statu nascendi”. It is a well established fact that even small doses of psychoactive drugs, e.g. sedatives, can change the course considerably (Thompson et al. (1975), Kramp B Rafaelsen (1978)). 2) Somatic illness being a contraindication for the investigational program (see below).

The material thus comprised different degrees of severity of clinical states. In order to obtain well-defined categories of the various clinical states, the patients were evaluated according to a classification system developed by Izikowitz and described by Salum (1972). The system has theoretical advantages (Hemmingsen et al. (1979)), just as it has previously been found suitable in studies of DT and less severe clinical states (Salum (1972), Kramp & Rafaelsen (1978)). The categories are defined as follows: Grade 1: Tremor without hallucinations. Grade 2: Tremor and also hallucinations, but no disorientation. Grade 3: Tremor, hallucinations and disorientation during the acute state. Grade 3 comprises the proper DT, the symptoms mentioned are the general accepted core symptoms of the syndrome (Criteria Committee, National Council of Alcoholism (1972), Keller (1977)). Only patients belonging to grade 2 or 3 were included in the present material. If possible, all the investigational procedures and their purpose were explained to the patients and informed consent was obtained before the patient was included in the study. When the clinical condition made such information meaningless, the patients were informed as soon as the clinical condition had improved sufficiently. METHODS The investigators were on call for alternative days during the study period. When 2.5.

396 a possible case had arrived, the patient was investigated within 30 min after admission by one of the authors to decide whether the patient should be included in the study. In such cases the patient was interviewed about the duration of the alcohol abuse, the daily consumption, the type of alcoholic beverage preferred and possible previous admission with DT. The food intake during the last month before admission was estimated as “sufficient”, “doubtful” or ‘‘clearly insufficient”. Previous serious somatic diseases were recorded. Concerning the actual condition the presence and duration of the symptoms: nausea, vomiting, diarrhoea, sleeplessness, agitation, hallucinations and convulsions were estimated; as was the period between the last drink and the time of investigation. The alcohol consumption during the last 72 hours was compared with the patient’s “habitual” consumption; if exact figures could not be given, the patient was asked to estimate whether he had noticed a marked increased consumption during that period. The patient was also asked to estimate the time period between his last drink and the time of investigation. As one of the blood samples taken immediately after admission - and then repeated daily - was analyzed for alcohol, it was possible to compare the patient’s information with his blood alcohol concentration (BAC). The patient was interviewed immediately after admission, but in some cases the patient’s condition at the time of admission made it impossible to obtain a detailed case history; in such cases the anamnestic data were detailed later on, when the patient’s condition had improved. The patients were classified as belonging to grade 2 or grade 3 (proper DT) according to the earlier described classification system. It should be emphasized that it was the patient’s condition at the time of admission that decided the diagnostic classification. Even if a patient developed some degree of disorientation during the course of the illness, he was still considered a grade 2 patient if he had been fully orientated at the time of admission. In order to check the patient’s information about previous intake of psychoactive drugs, the urine from most of the patients was screened for meprobamate, chlordiazepoxide and diazepam by a thin-layer chromatographic technique. Because treatment with barbital was often initiated before urine could be collected, one of the blood samples, taken at the time of admission, was investigated for barbiturates to see whether the patient had taken such drugs before admission. Statistical methods The Mann-Whitney U-test and Fisher’s exact test were used to test the significance of the differences. A P-value less than 0.05 was considered to indicate statistical significance.

RESULTS Twenty-one patients were included in the study; one was excluded due to incorrect diagnosis.

Table 1. Clinical and anamnestic data concerning 20 patients with DT and less severe clinical states Grade 3 (n = 9)

Grade 2

Variable

(n = 11)

~~

Age (years)

Median

ns. Range

Duration of alcohol abuse (years)

Median

28-55

30-51

20

17 n.s.

Range Daily alcohol consumption (g alcohol) Type of preferred alcohol beverage (no. of patients)

34

40

Beer Liquor Beer and liquor

Previous D T (no. of patients)

5-30

10-25

200-500

ns.

200-500

2 6

ns. n.s. ns.

0 3 6

4

ns.

3

4

3

~~

Food intake before admission (no. of patients)

Sufficient Doubtful Insufficient

4 4 3

ns. ns. ns.

Symptoms before admission (no. of patients)

Nausea Vomiting Diarrhoea Thirst Sleep disturbances Convulsions Drinking bout

6 7

ns. n.s.

8

5

D.S.

5

9 11

n.s. n.s. n.s.

8 9 3 6

Time interval between last drink and admission (hours)

Median Range

0

1 15 1-96

P

< 0.05

P < 0.1

1 4 8

3 1-24

This patient was a man of 60 who was admitted in a confusional state. Eighteen hours after admission he became increasingly delirious and agitated with vivid hallucinations. An alcohol abuse was supposed, and he was consequently included in the study. However, during the next hours he became increasingly febrile, a pneumonia was diagnosed and treated with antibiotics. Thereafter his symptoms disappeared, and - most important - he convincingly denied any alcohol abuse, and this was supported by his relatives.

The material thus consists of 20 patients, all men. Eleven patients were classified as being grade 2, nine as being grade 3. A trend was seen towards more patients with grade 3 being admitted during the night compared to grade 2: f o u r of nine patients with grade 3 were admitted between 0 and 7 a.m., none with grade 2 (P < 0.1). One patient in each group h ad previously been operated f o r peptic ulcer. One

398 Table 2. Blood alcohol concentrations (g/l) in 20 patients with DT and less severe clinical states at the time of admission

BAC

Grade 2 (n = 11) Grade 3 (n = 9)

0

0.010.49

0.500.99

1.001.49

1.5C1.99

2 z .

> 2.50

5

1

3

0

1

1

0

1

2

1

0

0

3

2

patient in grade 3 has had syphilis; seroreaction for syphilis in the blood and spinal fluid was negative at the present admission. Table 1 shows some further anamnestic data. No statistically significant age differences were seen between the two diagnostic groups, even though there was a trend that grade 3 patients were younger (median 34 years) than grade 2 patients (median 40 years). No differences were seen between grades 2 and 3 as to pattern of alcohol abuse or intake of food before admission. No statistically significant differences were seen between grades 2 and 3 in various physical symptoms preceding the admission. All the patients had had sleep disturbances, nearly all had experienced feelings of thirst and agitation; about two thirds suffered from various gastro-intestinal symptoms. None of the patients with grade 2 had had convulsions, which were seen in three cases among patients with grade 3. This difference did not reach a statistically significant level. The duration of the symptoms seemed to be different between the two groups. All patients with grade 3, but only four patients with grade 2, stated that the symptoms had been present for at least 48 hours, whereas the symptoms had been present for only 12-24 hours in the remaining seven patients with grade 2. Statistically significantly more patients with grade 3 had had a “drinking bout” just preceding the admission compared to grade 2. Furthermore it can be mentioned that the only patient with grade 2 having had such an episode developed marked disorientation during the first hours after admission. A trend was seen towards more patients in grade 3 having had a shorter interval between the last drink and admission than had grade 2 patients. BAC at the time of admission is shown in Table 2. It is seen that the majority of grade 3 patients had alcohol in the blood - several even concentrations about 2 gA (44 mmolA) when admitted. About half of grade 2 patients also had alcohol in the blood, but the BAC was lower among grade 2 patients compared to grade 3 patients (P < 0.05). It should be noticed that only two of the 20 patients clinically showed any sign of inebriation, Plasma barbiturate at the time of admission was zero in all cases. The urine screening for drugs was performed in nine patients in grade 2 and in six patients in grade 3. One patient in grade 2 and two patients in grade 3 had traces of chlordiazepoxide in their urine, one patient in grade 2 had traces of meprobamate, no patient had diazepam in the urine.

399 DISCUSSION The present material is supposedly representative for the conditions under discussion. Grade 3 comprised patients with fully developed DT and all the patients suffered to a varying degree from the whole spectrum of mental and phy; sical symptoms previously described as being typical for the syndrome. Grade 2 comprised patients with a less severe clinical state, also known as predelirium, DT incipiens or impending DT, a condition which, unpredictably, may or may not lead to DT (Sulurn (1975), Thompson (1978)). Only patients with severe concomitant somatic diseases were not represented in the material, such patients are usually dealt with by the somatic departments. However, the aim of the present study was to describe some clinical features of DT and this purpose could have been difficult to attain if patients with, for example, hepatic coma, severe head injuries or pancreatitis had been included. It should be emphasized that no patient in the present study suffered from somatic illness and none of the patients investigated by us during the study period were found to have a somatic illness which contraindicated the investigational program. By chance all the patients were men. Nielsen (1965) found the ratio between men and women among patients with “alcoholic psychosis” to be 9:1, Christensen & Sirundbygaard (1968) found the ratio to be 8:l among patients with imNumber of patients

15.

10.

5

12 15 16-19 20-23 0 3 t IMidnight Noon

6-11

f,

-7

Time of admission

Fig. 1. Number of patients admitted during the day and night (n = 71). Hatched columns: Grade 2 (n = 31). White coIumns: Grade 3 (n = 40). Significantly fewer grade 2 patients than grade 3 patients were admitted during the period 0-3 cdmpared to the time periods between 8 and 23. P < 0.05. (Present material and the material by Kramp & Rafaelsen (1978) added together.)

400

pending and fully developed DT, and the same was found by Krump & Rafaelsen (1978). All these studies were performed in the same area as the present, the City of Copenhagen. This sex ratio is similar to that of several other studies (Figurelli (1958), Krystal (1959), Tavel et ul. (1961), Cutshall (1965)). One would thus have expected a few women in the present material, but the fact that none were seen does not represent any statistically significant difference from what was to be expected. We therefore find it justified to conclude that the present material is representative for the conditions investigated. A trend was seen towards more grade 3 patients being admitted during the night, compared to grade 2, even though a statistically significant difference was not obtained. Due to this trend we have looked at a previous material, comprising more patients, where the time of admission also was recorded (Kramp & Rufaelsen (1978)). Exactly the same distribution of time of admission was seen in that material, and we therefore have added together the previous and the present materials. This gives a total of 71 patients, 31 belonging to grade 2, 40 to grade 3. The time of admission is shown in Fig. 1. It is seen that grade 2 patients were admitted during the day and the evening, but none during the night, whereas grade 3 patients were admitted during the night, too. This may be interpreted in two ways: 1) Grade 2 is not such a dramatic and terrifying condition that the patient, or his surroundings, finds it necessary to go to a hospital during the night. The patient may know the nature of his condition and thus know that if he manages to sleep, he will probably get better. On the contrary, a disorientated, agitated and hallucinated grade 3 patient obviously is in need of immediate care, and he is therefore admitted, be it day or night. 2) Darkness and night, with few environmental cues, will per se contribute to the progression of a less severe clinical state into a fully developed DT - in spite of the patient’s attempt to restore himself by alcohol intake, as seen by the high BAC at admission among grade 3 patients. This hypothesis is supported by the clinical experience in the treatment of DT that reduction of environmental cues in the department may cause further deterioration (Gunne (1958), Gross (1967)). No statistically significant age difference was seen between the two diagnostic groups even though there was a trend that patients in grade 3 were younger than grade 2 patients. Cutshall (1965) found the mean age of patients with welldefined DT to be about 46 years; similarly Nielsen (1965) and Sulum (1972) found the majority of patients with DT to be between the ages of 40 and 59; Sulum (1972) did not find any major difference in the number of grade 2 patients below and above the age of 40. Thus there seems to be a trend towards the grade 3 patients in the present material being younger (median 34 years) compared to patients admitted with DT 10-20 years ago. However, larger materials are necessary to substantiate this impression. The pattern of alcohol abuse among the patients in the present material was similar to that found in previous Danish studies of DT (Nielsen (1965), Nielsen (1966), Christensen & Strandbyguard (1968)). The alcohol abuse had lasted for

401 15-20 years, the beverage preference was beer and liquor, and the daily consumption was about 200-500 g alcohol - on the average equivalent to a bottle of hard liquor a day. These findings do not differ from other studies (Lundquist (1961), Cutshall (1965), Salum (1972)). The type of alcoholic beverages abused has been thought to be an important factor, beerdrinkers being assumed not to develop DT, but reviews of the literature have shown this not to be true (Cutshall (1965), Hemmingsen et al. (1979)), in accordance with the h d i n g s of Nielsen (1965). It is not the type of alcoholic beverage, but the amount of alcohol, that is important. According to the exclusion criteria, all patients who had been taking psychoactive drugs during the last 24 hours before admission were excluded from the study. Nevertheless traces of chlordiazepoxide were found in the urine in three cases out of 15 investigated, meprobamate in one case, whereas barbiturate was not detected in the plasma from a single patient. Concomitant drug abuse among alcoholics is not uncommon (Freed (1973)), and our results confirm this but also show, as previously found (Kramp & Rafaelsen (1978)), that barbiturates are not among the drugs used and abused by alcoholics. No patient in the present material had diazepam in the urine at the time of admission, possibly indicating a changed pattern of prescription, as we (Kramp & Rafaelsen (1978)) have found that diazepam was used by a third of a comparable patient material a few years ago. About one third of the patients had previously been admitted with DT or a less severe clinical state. The available data did not permit a more detailed analysis of these preceding admissions, e.g. milder or more severe forms compared to the present episode, but our data are similar to those found by Lundquist (1961) and Salum (1972). About one third of the patients stated that their food intake during the last month before admission had been clearly insufficient, only five found that their appetite and food intake had been as usual. Those figures correspond well with the finding of Lundquist (1961) that some 40 % of patients with DT had had insufficient food intake during the last week before admission. Significantly more patients in grade 3 had had “a drinking bout” preceding the admission and there was a marked trend that the time interval between the last drink and admission was shorter among grade 3 patients (median 3 hours) compared to grade 2 patients (median 15 hours). On the other hand the duration of symptoms, also including hallucinations, before admission was shorter in grade 2 (12-24 hours) than in grade 3 (> 48 hours). Lundquist (1961) found an increased alcohol consumption during the last month before a DT developed, and so did Smith (1953) and Cutshall (1965). Salum (1972) found an increasing severity of the syndrome with an extended duration of the preceding drinking period. However, it was not mentioned whether this drinking period was a more extensive abuse compared to a “habitual abuse” (a drinking bout), or whether it was the duration of the drinking period per se without such an excessive episode. A comparison of our results with the results of Salum is therefore not possible. An increasing alcohol consumption, an excess within the abuse, seems to be one of the factors in that chain of events that leads to a fully developed DT, whereas

402 this phenomenon does not seem to be an important factor in the development of less severe clinical conditions. The time interval between cessation of alcohol consumption and the onset of symptoms, “the latency period”, found in the present material, dif€ers remarkably from the results of several other studies. Salum (1972) found, as we did, that “tremor and certain other symptoms” began during the drinking period. Among grade 2 patients she found that in 75 %, hallucinations occurred within 12 hours after the last alcohol intake; in 10 % the latency period was 12-24 hours. We found that the symptoms among this group of patients started at about the same time as the patients stopped drinking, but due to the inherent ambiguity of this kind of information this difference does not seem crucial. Much more striking is the difference seen in grade 3 patients. Salum (1972) found that almost 80 % of grade 3 patients developed disorientation more than 24 hours after the last drink, the most usual latency period being 1 4 days. It was further stated that disorientation did not follow the occurrence of hallucinations immediately, although the time interval between the appearance of these two phenomena was usually very short. Nielsen (1965) found that only 50 % of patients with fully developed DT had been drinking up to the onset of the disease, the other half was found to have had latency periods between 2 and 8 days. Similarly, Victor & Adams (1953) found that only 4 % of DT patients had a latency period of less than 24 hours. In contrast, we in general did not find any latency period at all. The patients stated that the symptoms had been present for days before admission, but they had been drinking during that period. In most cases it was impossible to decide the duration of the disorientation, but in four cases, where information was available, the duration had been at least 24 hours. Three of these four patients had had the last intake of alcohol a few hours before admission, and this information was supported by their high BAC at admission. The reason for this difference between our results and the results found by others is not clear. The diagnostic criteria used by us were the same as the criteria used by Salum (1972) and the present material does not differ regarding case histories or symptomatology from the above-mentioned studies. It is often stated (e.g. Lundpist (1961), Nielsen (1965), Salum (1972)) that information about alcohol consumption obtained from alcoholics is unreliable and subject to error. However, the validity of the information obtained in the present material could in some way be tested by comparing the information about the time interval between the last drink and admission with the patient’s BAC at the time of admission. These parameters were to a large extent concordant; significantly more (P C 0.05) patients who said they had had the last drink more than 10 hours before admission, had a BAC at nil compared to patients who said they had had the last drink within the last 10 hours before admission. We therefore find it justified to conclude that the various anamnestic data, obtained in the present study, are to a large extent valid. To our knowledge no others have investigated patients with DT immediately after admission, as done in the present material; most often the patients have been investigated within a few days after admission and this methodological difference may partly explain the differences

403 between the present results and the findings of others. Our patients were interviewed at the time of admission independent of time of the day, and they may have been able to communicate more details about their drinking behaviour immediately before admission than would patients interviewed later, e.g. after the acute state had subsided. These patients with fluctuating consciousness often have partial amnesia for the events occurring during the psychosis when they are interviewed afterwards. The importance of an immediate investigation after admission is illustrated by our BAC results: only six patients, five in grade 2 and one in grade 3, had a BAC at nil at the time of admission. The next morning only two of the 20 patients had alcohol in the blood, 0.13 and 0.15 g/l (2.9 and 3.3 mmoVl) respectively. This probably explains why many more patients in the present material were found to have substantial amounts of alcohol in the blood compared to the BAC results in other studies, where the patients have been investigated some time after the hospitalisation, but still, for example, within the first 24 hours after admission (Nielsen (1965), Salum (1972)). CONCLUSION Studying DT and less severe clinical states two important methodological factors must be pointed out: 1) The use of strict diagnostic criteria, which are able to distinguish between proper DT and the less severe clinical states. 2) The investigational procedures must start immediately after the admission of the patient. Due to this methodology it has been possible in the present study to point out some differences between patients with the often harmless less severe clinical states (grade 2) and the fully developed, serious DT (grade 3): grade 3 patients had had a drinking bout just preceding their DT, and they continued to drink in spite of the developing DT. Grade 2 patients had no preceding drinking bout, and the symptoms occurred about the same time as they stopped drinking. Furthermore, the symptoms had lasted longer before admission in patients with grade 3 compared to patients with grade 2. In many other aspects, e.g. the duration and magnitude of alcohol abuse, food intake or symptomatology preceding the admission, no differences were seen between the categories. In the light of the old discussion about the aetiology and pathogenesis of DT - a state of withdrawal or not - the present results seem to be important, as they suggest that DT cannot be looked upon as an absolute withdrawal reaction. For a more thorough discussion of this question, data about the patient’s condition during the acute state also must be taken into consideration. The interpretation of the above mentioned results will therefore be further detailed in the following paper concerning the patients’ condition during DT and less severe clinical states. ACKNOWLEDGEMENT

Part of this study was supported by a grant from the Danish Medical Research Council NO. 512-8570.

404 REFERENCES Christensen, J . K., & N . Strandbygaard (1968): Libriumbehandling af akutte alkoholiske urotilstande. Ugeskr. Laeg. 130, 763-766. Cutshall, B. J . (1965): The Saunders-Sutton Syndrome: An analysis of delirium tremens. Quart. J. Stud. Alcohol 26, 423448. Figurelli, F. A. (1958): Delirium tremens. Reduction of mortality and morbidity with promazine. J. Amer. med. Ass. 166, 747-750. Freed, E. X . (1973): Drug abuse by alcoholics: A review. Int. J. Addict. 8, 451-473. Gross, M . M . (1967): Management of acute alcohol withdrawal states. Quart. J. Stud. Alcohol 28, 655-666. Gunne, L.-M. (1958): Mortaliteten vid delirium tremens. Nord. Med. 60, 1021-1024. Hemmingsen, R., P . Kramp & 0.J . Rafaelsen (1979): Delirium tremens and related clinical states. Etiology, pathophysiology and treatment. Acta psychiat. scand. 59, 337-369. Keller, M . (1977): A lexicon of disablements related to alcohol consumption. In Edwards, G., M . M . Gross, M . Keller, J . Moser & R . Room (eds.): Alcohol-related disabilities. WHO Offset Publ. No. 32, Geneva, pp. 23-60. Kramp, P., R. Hemmingsen & 0.J . Rafaelsen (1979): Delirium tremens: Some clinical features. Part 11. Acta psychiat. scand. 60, 405-422. Kramp, P., & 0 .J . Rafaelsen (1978): Delirium tremens: A double blind comparison of diazepam and barbital treatment. Acta psychiat. scand. 58, 174-190. Krystal, H. (1959): The physiological basis of the treatment of delirium tremens. Amer. J. Psychiat. 116, 137-147. Lundquist, G . (1961): Delirium tremens. Acta psychiat. scand. 36, 443-466. Nielsen, J . (1965): Delirium tremens in Copenhagen. Acta psychiat. scand., Suppl. 187, 1-92. Nielsen, P . V . (1966): Heminevrin (klormetiazol). Ugeskr. Laeg. 128, 973-980. Pontoppidan, K . (1895): Psykiatriske forelaesninger og studier. Tredie raekke. Th. Lind, Copenhagen, pp. 98-114. Romano, J . (1941): Early contributions to the study of delirium tremens. Ann. med. Hist. 3, 128-139. Salum, I . (1972): Delirium tremens and certain other acute sequels of alcohol abuse. Acta psychiat. scand., SuppI. 235, 1-143. Salum, I. (1975): Treatment of delirium tremens. Brit. J. Addict., Suppl. I, 75-80. Smith, J . A . (1953): Methods of treatment of delirium tremens. J. h e r . med. Ass. 152, 384-387. S@rensen,B. F. (1959): Delirium tremens and its treatment. Dan. med. Bull. 6, 261-263. Tavel, E. M., W . Davidson & T . D . Batterton (1961): A critical analysis of mortality associated with delirium tremens. Amer. 3. med. Sci. 242, 18-28. Thompson, W . L. (1978): Management of alcohol withdrawal syndromes. Arch. intern. Med. 138, 278-283. Thompson, W . L., A . D . Johnson, W . L. Maddrey & The Osler Medical Housestaff (1975): Diazepam and paraldehyde for treatment of severe delirium tremens. Ann. intern. Med. 82, 175-180. Victor, M., & R . D . Adurns (1953): The effects of alcohol on the nervous system. Res. Publ. Ass. new. ment. Dis. 32, 526-573. Criteria Committee, National Council on Alcoholism (1972): Criteria for the diagnosis of alcoholism. Amer. J. Psychiat. 129, 127-135. Received March 23, 1979

Peter Kramp, M.D. Ralf Hemmingsen, M.D. Department of Psychiatry Rigshospitalet 9, Blegdamsvej DK-2100 Copenhagen f3 Denmark