Acta psych. scand. (1979)59, 263-275 Department of Psychiatry (Heads: Prof. V.Lunn, Prof. 0. J. Rafaelsen, T. Vanggaard, M.D.), Rigshospitalet, Copenhagen, and the Dumex Company, Copenhagen, Denmark

Barbital and diazepam plasma levels during treatment of delirium tremens P. KRAMP,P. R~NSTED AND T.HANSEN Plasma concentrations of barbital and diazepam were measured daily during a double-blind study of the efficacy of the two drugs in the treatment of delirium tremens and less severe clinical states. Treatment was estimated as satisfactory in the majority of cases; the present study deals with the satisfactory groups only. Both in the barbital group and in the diazepam group the same plasma level was seen in different clinical states. This result is discussed in relation to the theories about the aetiology of delirium tremens, and it is concluded that the data fits best with the assumption that delirium tremens is released from a withdrawal state, but once established, the delirious state is not interrupted by the drugs. The barbital concentrations were rather high, many at a level where non-alcoholics would show pronounced intoxication symptoms not seen in the present material. The diazepam concentrations on the other hand were low, often below a level where a cerebral effect is measurable in normal subjects. On this basis it is concluded, that the two drugs have different modes of action. Barbital may act by its cross-dependence properties with alcohol and thus diminish the withdrawal reaction, whereas diazepam may act by its anti-anxiety effect, but not in the doses here applied, by crossdependence properties with alcohol. Finally, this hypothesis is discussed in relation to clinical experience in the treatment of delirium tremens.

-

Key words: Delirium tremens - alcohol withdrawal alcohol intoxication - barbital - diazepam - plasma concentrations - mode of action aetiology.

-

Barbital, a long-acting barbiturate, has been used in Denmark in the treatment of delirium tremens (DT) and related clinical states for decades with good results. The mortality rate for DT in the 1950s was undoubtedly lower in Denmark compared with many other countries (Nielsen 1965). Barbital is seldom used in other countries, probably due to the risk of drug intoxication with respiratory depression and coma (GreenbZutt & Greenblatt (1972)), but given under careful clinical observation it has been found to be a safe and effective drug in the treatment of DT (S@rensen (1959), Nielsen (1965), Krump & Rufaelsen

(1978)). The use of the two benzodiazepines, chlordiazepoxide and diazepam, in the treatment of DT and related clinical states has increased during recent years 0001490Xf79/030263-13$02.50/0@ 1979Munksgaard, Copenhagen

264 (Fuvuzzu & Martin (1974)). There do not seem to be any major differences between the two drugs (Greenblutt & Shuder (1974)), but diazepam may have some advantages (Brown et ul. (1972), Krump & Rufuelsen (1978)). Benzodiazepines are undoubtedly of value in less severe cases (Rosenfeld & Bizzoco (1961), Cutshall (1965), Kuim er ul. (1969)). They diminish the patient’s anxiety and restlessness and to some extent prevent progression. On the other hand, they do not shorten the episode, and in the treatment of more severe cases, including DT, they may not be ideal (Curshull (1965), Golbert et ul. (1967)). Whichever drug is used, the treatment must be ruled by careful clinical judgment, and the doses needed will thus vary considerably (Thompson et ul. (1975), Krump & Rufuelsen (1978)). The reason for this variation is unknown. It may be due to differences in the severity of the clinical state, route of administration, frequency of drug administration or concomitant drug therapy. Most cases improve within 72 h (Cutshull (1965), Sulum (1975)). Due to the natural course of the disease, the time at which treatment is started will be an important factor for the amount of drug found necessary; this may partly explain the clinical experience, that patients who seem to be in the same clinical state need widely varying drug doses. However, pharmacokinetic explanations may also be relevant. To our knowledge determination of the plasma concentration of drugs, when used in the treatment of DT, has not been performed before. It is thus unknown whether the difference in drug doses is paralleled by differences in plasma values, e.g. whether higher plasma concentrations are necessary in more severe clinical states. In the following, data concerning plasma concentrations of diazepam and barbital obtained during treatment of ongoing DT and related states will be presented. MATERIAL AND METHODS The material has been presented in detail elsewhere (Krump & Rufuelsen (1978)). In summary it comprises patients who participated in a double-blind study concerning the efficacy of diazepam and barbital in the treatment of DT and less severe clinical states. Intake of psychoactive drugs within the 24 h prior to admission was an exclusion criterion; as was alcohol in the organism (measured indirectly in the expired air) at the time when drug treatment was found necessary. Ninety-one patients took part in the study, 47 were treated with barbital, 44 with diazepam. The material was divided into three groups according to the severity of the clinical condition (Sulum (1972)): Grade 1: Tremor. Grade 2: Tremor and hallucinations. Grade 3: Tremor, hallucinations and disorientation (proper DT). Medication with barbital or diazepam was given according to the clinical condition; no other treatment with psychoactive drugs was given during the investigational period. Blood samples, to determine the plasma concentration of the drug given, were taken every morning, or at least 3 h after the latest medication, from 33 patients treated with diazepam and 46 patients treated with barbital. The last blood sample was taken the morning after the last study medication dose was given. Blood samples were analyzed by laboratory technicians without knowledge of

Table 1. Barbital plasma concentrations (pmol/l). Treatment estimated as satisfactory Diagnosis Grade 1 median range 25 % quartile 75 % quartile n

Grade 2 median range 25 % quartile 75 % quartile n

Grade 3 median range 25 % quartile 75% quartile n

Sample 1

Sample 2

Last sample

255 103-505 206 407 15

385 223472 292

320 168-505 223 440 15

310 168-586 233 427 6

342 212-570 238 520 4

334 212-402 244 382 6

201 119-510 144 435 9

397 195-1 140 225 556 8

413 119-1 107 266 467 11

440

6

the research protocol or the treatment of the individual patient. Blood samples for barbital were analysed within 24 h, diazepam samples were centrifuged and the separated plasma stored at -2OOC until analysed. In a few cases a single blood sample was missed, these patients are still included in the material. Only if more than one sample, or the last one, was missing, was the patient excluded. An unexpected result, in spite of the exclusion criteria, was that about onethird of the patients treated with diazepam had a clearly elevated pre-study plasma level of diazepam or its main metabolite, N-desmethyldiazepam (values > 0.18 pmoM of one of the two compounds). (In the following all plasma concentrations will be given in pmol/l, other results will be converted to this). Four of these patients had a pre-study level that was more than half of the maximum plasma concentration obtained during the treatment period. These four patients were excluded from the present study. Only two patients treated with barbital had a modestly elevated plasma pre-study level of barbital, and this did not indicate exclusion from the present study. Thus the material comprises 40 patients treated with barbital and 25 patients treated with diazepam. Analysis The plasma concentration of diazepam and its main metabolite, N-desmethyldiazepam, was determined by gas-chromatography (Arnold (1975)), and the concentration of barbital by U.V. spectrophotometry.

266 Table 2. Diazepam plasma concentrations (pmoU1). Treatment estimated as satisfactory Diagnosis

Sample 1

Sample 2

Last sample

1.15 0.18-2.17 0.53 1.82 9

1.74 1.25-3.43 1.43 2.46 6

1.55 0.18-3.36 0.83 2.14

1.44 0.75-2.88 0.81 2.24

5

1.34 0.87-2.23 0.96 1.84 5

1.32 1.05-2.23 1.10 1.64 6

0.76 0.37-2.31

1.07 0.39-1.23

1.07 0.39-1.23

3

3

-~

~~

Grade 1 median range 25 % quartile 75% quartile n

Grade 2 median range 25 % quartile 75% quartile n

Grade 3 median range 25 % quartile 75 % quartile n

-

10

-

-

3

RESULTS The clinical part of the study has been reported earlier (Kramp & Rafaelsen (1978)). Briefly, the results were evaluated both with respect to the duration and course of the condition, and to a global evaluation of the efficacy and administration of the treatment. In grades 1 and 2 no major differences were seen between the two drugs. In grade 3, proper DT barbital was found superior to diazepam in the global evaluation of the efficacy. Barbital was given orally, and diazepam by intramuscular injections, but it was concluded that this difference in route of administration probably did not iduence the results. The plasma concentrations obtained during the treatment are shown in Table 1 (barbital), Table 2 (diazepam), and Table 3 (diazepam N-desmethyldiazepam) for the patients where the efficacy of the treatment was estimated as satisfactory. The material has been divided into grades 1, 2 and 3 (DT) according to the earlier described classification system. In a few cases the condition lasted longer than 48 h. The plasma concentrations obtained from these patients on day 3 and eventually day 4 were at the same level as the concentrations on day 2, both in the diazepam and the barbital group. Due to the short duration of the condition - most cases improved within 48 h - the last plasma value will often be the second one - or the first value, if the patient improved very quickly. In the different barbital groups (grades 1, 2 and 3, and first, second and last samples) medians at about 350 pmol/l are seen, but with a wide variation as seen from the 25 % and 75 % quartiles. The same is seen in the diazepam groups, where the medians for diazepam

+

261

Table 3. Diazepam Diagnosis Grade 1 median range 25 % quartile 75 % quartile n

Grade 2 median range 25 % quartile 75 % quartile n

Grade 3 median range 25 % quartile 75 % quartile n

+ N-desmethyldiazepam plasma

concentrations (pmoN1). Treatment estimated as satisfactory

Sample 1

Sample 2

Last sample

1.49 0.24-3.76 0.97 2.47 9

3.37 2.35-4.93 2.42 4.78 6

2.76 0.24-5.57 1.21 5.08 10

2.50 1.02-4.13 1.14 3.37 5

2.17 1.22-4.06 1.46 3.31 5

2.32 1.70-4.06 1.80 2.97 6

1.25 1.03-3.70

2.04 1.81-2.55

2.04 1.81-2.55

-

-

-

3

3

3

+

are about 1.40 pmoM, for diazepam N-desmethyldiazepam, 2.30 pmoM. In the barbital group as a whole the efficacy was estimated as non-satisfactory in eight cases, in the diazepam group in six cases. Due to the limited number of cases in each group it is not possible to make similar Tables for the non-satisfactory groups. The single values are shown in Table 4. As to be expected, there is a trend to higher values in the non-satisfactory groups, as more doses were given if the patient did not improve. It should be emphasized, however, that medication was given according to the clinical condition and continued until the patient became calm or fell asleep. Most often the efficacy was estimated as non-satisfactory because the symptoms reappeared after a period with less pronounced symptoms. This course was impossible to decide beforehand, and it was thus impossible to predict whether larger amounts of drugs were necessary in individual patients, as the patient initially responded to the treatment. The code was broken three times due to unsatisfactory effect of the treatment, the drug given each time was diazepam. In two of the patients treatment continued with diazepam but now given intravenously. One patient needed 100 mg during the next few hours before he improved, the other one needed 700 mg during the following 48 h. The plasma concentrations were not determined in these two patients, but the plasma concentrations have probably been higher than the above-mentioned values due to the much larger amount of drug given. In the following only the satisfactory groups will be discussed.

268 Table 4. Barbital, diazepam, and diazepam

+

N-desmethyldiazepam plasma concenrrations, single values (pnol/l). Treatment estimated as non-satisfactory

Grade

Sample 1

Sample 2

472 320 521 309 494 266 163 326

624 646 635 407 510

Last sample

Barbital 1

2 3

635 494

570 624 662 787 434 613 603 461

D iazepam

1 2

3

0.26 0.46 0.58 1.77 0.69

0.30 1.33 2.29 3.08 2.10

0.46 0.30 1.30 3.00 3.08 2.10

Diazepam -?N-desmethyldiazepam

1 2

3

0.59 0.97 0.73 2.75 0.73

0.69 2.07 4.24 4.82 3.41

0.97 0.69 2.90 6.04 4.82 3.41

DISCUSSION Barbital has a long half-life, at least 48 h (Lous (1954), Parke (1971)), but probably longer. The plasma values will therefore to a great extent reflect the levels which were necessary to sedate the patients to such a degree that they fell asleep, this being the aim of the treatment. The results are fairly uniform, the same plasma level is seen in different grades of severity and during the course of the treatment. The values are in general rather high; the 75 % quartiles are thus about the level where patients with barbital poisoning, who are not habituated to the drug, begin to wake up (Myschetzky & Lussen (1971)). HOWever, none of the patients in the present study showed more pronounced intoxication symptoms; a few had at most some dizziness and/or mild speech disturbances. All the patients could, for example eat and drink by themselves after the acute state, in spite of the high plasma barbital. Cross-dependence between alcohol and barbital has been shown both in mice and humans. Goldstein (1972) found that administration of long-acting barbiturates (and diazepam) could remove alcohol withdrawal symptoms in mice. Ftuser et al. (1957) investigated the effect of alcohol administration on the barbiturate withdrawal syndrome. Ten barbiturate abusers were chronically intoxicated with long-acting barbiturates for 3 5 4 4 days. The barbiturate was then suddenly

269 substituted by large doses of alcohol (310-460 ml 95 % alcohol per day). After this, the patients were in the same degree of intoxication and only developed weak barbiturate withdrawal symptoms. Lieber (1976) mentions that larger doses of sedatives are required to achieve a given effect in alcoholics than in nonalcoholics. This may be due both to CNS adaptation to sedatives (pharmacodynamic properties), and to a metabolic adaptation - an increased capacity of the liver in alcoholics to inactivate sedatives (pharmacokinetic properties). Barbital is not, or at most to a very limited degree (< 5 %), metabolized in the liver (Parke (1971)), and the relatively high plasma levels of barbital found in the present study probably reflect CNS adaptation to what is called ,the general sedatives