THE AMNESIC EFFECT OF DIAZEPAM (VALIUM)

Transcription

1 Brit. J. Anaesth. (1970), 42, 693 THE AMNESIC EFFECT OF DIAZEPAM (VALIUM) BY P. R. F. CLARKE, P. S. ECCERSLEY, J. P. FRISBY AND J. A. THORNTON SUMMARY A double-blind trial of intravenous diazepam (0.24 mg/kg) versus saline was carried out on twelve young intelligent male volunteers to examine the amnesic effect of the drug. Complex psychological testing extended over a 2-hour period. Dense anterograde amnesia was found for approximately 10 minutes with a less severe impairment of memory persisting at least half an hour after injection. It is considered that the memory difficulties were not due to serious reduction of level of consciousness but do focus on the input or "consolidation" processes rather than on retrieval. Further studies are needed to define more clearly the nature of the deficit. The present study arose from the clinical observation of amnesia following the use of intravenous diazepam in oral surgery and dental conservation. Many authors have noted such an effect and several have studied it in clinical series. Recent examples include Brown and Dundee (1968), Fox, Wynands and Bhambhami (1968), Frumin, Herekar and Jarvik (1969), and CNeil and Verrill (1969). In general the amnesic effect has been considered a valuable bonus and exploited to cover a variety of unpleasant and anxiety-provoking procedures. An outline of the pharmacology of diazepam and of its uses in anaesthesia can be found in Knight and Burgess (1968). The phenomenon of postoperative anterograde amnesia has frequently been reported and related variously to premedication technique (Feldman, 1963) or general anaesthesia (Fox, Wynands and Bhambhami, 1968; Gruber and Reed, 1968). Hardy and Wakely (1962) writing on hyoscine and atropine consider that the mechanism involved is analogous to light general anaesthesia causing "central depression" which attacks memory storage as the most vulnerable function. Mazzia and Randt (1966) in an elegant demonstration implicate the arousal system in both anaesthesia and pathological memory disorders. Since Parkes (1968) considers that diazepam acts to suppress stimulation of the hippocampus via the amygdala in the cat, it is tempting to relate the clinical findings to specific disorders of memory of the sort described by Scoville and Milner (1957) and Penfield and Milner (1958) in bilateral hippocampal damage. The complexities of measuring amnesia have been brought out by several authors, e.g. Gruber and Reed (1968), Milner (1968), Whitty and Zangwill (1966), and Deutsch (1969). In order to clarify some of these points it was decided to use diazepam alone in an experimental setting, to ensure that the subject was awake when the information was put in, and to measure a number of different complex cognitive functions. PROCEDURE In order to reduce the individual differences as much as possible, twelve young intelligent male volunteers served as subjects (ten dental students, a research physiologist, ages 20-23, and a senior medical technician, age 34). They were assigned to the experimental or placebo group by the P. R. F. CLARKE,* M.A., DIP.PSYCHOL., A.B.PS.S; P. S. ECCERSLEY,t M.B., CH.B., F.F.A.R.C.S.; J. P. FRISBY^ B.A.; J. A. THORNTON,! MJ)., F.F.A.R.C.S., DA; from the University of Sheffield. Present addresses: * Whiteley Wood Clinic Woofindin Road, Sheffield S10 3TL. t Dept. of Anaesthetics, Royal Infirmary, Sheffield 6. % Dept. of Psychology, University of Sheffield, Western Bank, Sheffield S10 2TN. A brief report of this work was presented orally to the British Psychological Society in London, December 13, 1969.

2 AMNESIC EFFECT OF DIAZEPAM (VALIUM) 691 anaesthetist according to a random number table prepared by one of us who did not see the subjects himself. The anaesthetist alone knew which treatment the subject received but he was not involved in any subjective assessment procedures: the other experimenters were kept in ignorance of the treatment except that the behavioural changes were so marked that there was rarely any doubt about it. The time course of the initial procedure is set out in figure 1. The subjects were brought fasting to the laboratory where the anaesthetist made a brief medical inquiry and examination. As they lay reclining on a couch, a technician gave them a decision-making task, a practice run on a vigilance task and a list of words to remember (on tape). The anaesthetist then administered intravenously either saline (1 ml/ min) or undiluted diazepam (0.24 mg/kg at a rate of 5 mg/min). Thus the injection lasted 3-4 minutes: 2 minutes before it was completed the vigilance test began. This task was interrupted at equally spaced intervals by three more lists of words to remember. At the end of the vigilance task (approximately 15 minutes after the beginning of the injection) the decisionmaking task was presented again while the subject was still reclining on the couch. After this he was taken to a nearby room and tested by the psychologist who gave a triple-associate word-learning test, three intelligence tests, retention tests of the triple-associates, and finally retention tests of the word lists heard on the INJECTION 6O INTELLIGENCE] TRIPLETS AXYZ I2O FIG. 1 Time course of the experimental procedure. D,, D,, D, represent three presentations of the decision-making task. Triplets represents the auditory-verbal tripleassociate learning task. In and Out refer to learning and retrieval aspects respectively. tape and of the events of the session. Thus the triple-associate learning began approximately half an hour after the injection, the retrieval of the associates half an hour after the learning was completed, and the retrieval of the word lists between 1 and 1 hours after they had been heard. The session ended with a third trial of the decision-making task. The subjects were then fed and escorted home when fit. Some weeks later, all the subjects were tested again in clear consciousness by the psychologist. On this occasion one intelligence test was repeated and parallel versions of the other two intelligence tests and of the triple-associate learning and retention tests were used. The tests are described in detail below. The decision-making task. This was a variant of the "expanded-judgement" decision task described by Irwin and Smith (1956). For any one decision, the subject was shown a series of cards drawn from a large pack (N = 300) at 2-second intervals. Each card bore a number in the range and the subject was asked to judge in a forced-choice manner whether or not the mean of the pack was greater or less than 74. He was instructed not to wait until the whole pack had been presented before arriving at his decision but rather to decide "as soon as he felt able to". Each presentation of the task required him to make eight decisions one for each of eight packs whose means varied but whose standard deviations remained constant at 7.5. Three measures were taken on each decision: its accuracy, the number of cards required to make it and the subject's confidence in it (self-rated on a ninepoint scale). These measures were compared over the three presentations of the task. Vigilance task. Random digits were presented on tape at 5-second intervals, with one repetition in every block of 11 digits. The subject had to add 2 to each digit and say the sum aloud except when a digit occurred twice in succession. When that happened he was instructed to say "Repeat" instead. A short practice run was given before the injection. The main test (consisting of 121 digits including 11 repeats) began before the injection was completed and ran for approximately 13 minutes, interrupted at 33-digit intervals by three more memory lists. The scores were the number of omissions of responses, the number

3 692 BRITISH JOURNAL OF ANAESTHESIA of wrong answers, and the number of times the sum was mistakenly given instead of the word "Repeat". Memory lists. lists of 8 words in each of four categories (animals, parts of the body, vegetables, birds) were constructed like those of Bousfield (1953) but selected to be appropriate to a British population. Each list was equated for frequency of occurrence in the English language (Thorndike and Lorge, 1944): only one word occurs times per million, one times, one times, one times, and four between 2 and 9 times. They are thus relatively infrequent words but care was taken not to include subjectively "difficult" words. The 8 words were interleaved and repeated 3, 4 or 5 times to form a complex list of 30 items recorded on tape at approximately 1 every second. The subject's task was to listen to and remember the 8 words used in each list. His scores were the number of lists he thought he had heard, the number of words widiin each list that he recalled and the number of words he could correctly select from recognition sets of 16 words, containing 8 equally frequent "confusion" items which had not been previously presented. Triple-associate learning test. This test was devised by one of the authors for the assessment of various memory difficulties in neurological and psychiatric patients. It was reported orally by Clarke in 1960 and has been used clinically since but as yet has no standardization data. The subject is asked to try to remember 6 sets of 3 words which are read aloud as 6 triplets, the 3 words in a triplet having a phonic similarity (see table I). His learning is tested by prompting him with the first (stimulus) word of each triplet and asking for the other 2 words which went with it. TABLE I Triple associate learning materials (Set I). The stimulus word in each set is underlined; the subject has to learn to respond with the other two words. cur tape thrust win pick courage mate thought worry ridge pad custard bake throttle willing dredge puddle He is corrected at once if he gets them wrong and the whole list of triplets is tried 4 times. This provides a measure of immediate memory (number of response words correct after the first hearing), and 2 measures of learning (number correct summed over 4 trials, and the number which are correct 3 times in a row in these 4 trials). Obviously, in order to measure retention, it is necessary to make sure that material has been adequately learned, and so, where necessary, further trials are given of one or two triplets until at least 4 items have been correct 3 times consecutively. After a "forgetting period" of half an hour, retention is measured in 3 different ways. First the subject is asked to recall as many of the 18 words as he can without regard to order or grouping (free recall). Next he is prompted with the first word of each triplet and is asked to respond with the correct words as in the original learning except that his answers are not corrected (prompt recall). Finally, a list of 21 words is read to him and he is asked to select the 12 which were the response words of the triplets originally learnt (recognition).* Intelligence tests. These were chosen to test a range of different intellectual functions in a brief period. One was a shortened 10-item version* of the synonym-selection form of the Mill Hill Vocabulary (Raven, 1958) to provide a measure of word knowledge relatively insensitive to temporary intellectual inefficiency. This was given on both occasions. The other tests were taken from the Wechsler Adult Intelligence Scale (WAIS) (Wechsler, 1955) on the first occasion and from the Wechsler Bellevue Form II (WB II) (Wechsler, 1946) on the second. These are standard American tests of general intelligence administered individually and are designed to be parallel forms for repeated testing. Each consists of a number of subtests of which only two were given here, the Similarities and the Block Designs. Both of these are relatively sensitive to intellectual impairment even on a temporary basis. The Similarities involves explaining how two named things are alike and might be taken as a measure of verbal skill relevant to word learning. The Block Designs * Fuller details of these tests may be obtained from P. R. F. Clarke.

4 AMNESIC EFFECT OF DIAZEPAM (VALIUM) 693 involves copying patterns with wooden cubes and measures more spatial aspects of intelligence: as the only timed test given it was expected to be specially sensitive to impairment. RESULTS Level of consciousness and general intellectual efficiency. The clinical observation of amnesia after diazepam might be due to a specific disruption of storage and retrieval processes or instead it might reflect general intellectual impairment or simply loss of consciousness. There was clear evidence that the experimental subjects were less efficient on the vigilance task than the controls but it was equally clear that they were not asleep or unconscious (see fig. 2). Errors were few in both groups. All the experimental group but only 2 of the 6 control subjects made errors (Fisher test, P<0.025 one-tail; Siegel, 1956). Omissions were rare: only 2 errors were made, both by experimental subjects. Only 2 experimental subjects gave a substantial number of wrong answers, but the occasional "repeat" task showed more consistent differences suggesting that the bridging of the 5-second gap between digits was difficult for the experimental subjects. Thus while there is evidence that the experimental subjects were clearly awake they nevertheless showed some impairment of intellectual function. This view is supported by the 5 4- n O 1 FORGETTING DIAZEPAM SALINE 2 'REPEAT' i2- \- _ _ O I 2 3 A IO 19 WRONG ANSWERS FIG. 2 Vigilance task errors (raw scores). Vertical axis represents number of subjects in each group obtaining each of the scores along the horizontal axis in response to a total of 121 digits including 11 repeats. finding that 5 of the 6 diazepam subjects showed slurring of speech during the task while none of the controls did. Diazepam marginally affected the decisionmaking task. This was shown by an analysis of variance of "difference scores" calculated for each of the three measures by comparing pre- and postinjection readings, a procedure which enabled each subject to act as his own control. A full discussion of the findings that resulted will be presented elsewhere (Frisby et al., in preparation). For the present it will simply be noted that the two groups of subjects did not differ significantly on either their "accuracy" or on their "number of cards required" difference scores, but that they did so on their "confidence rating" results. Thus, for the most "difficult" packs (these had means actually equal to the criterion of 74), the diazepam group lost confidence in their decisions following injection, whereas the saline group showed a corresponding gain in confidence on the later administrations of the task (P<0.05). No significant differences could be demonstrated between the two groups on the three intelligence tests used. Table II shows that the groups were above average intelligence on the two Wechsler subtests on both occasions although there were some surprisingly low scores for such a sample. Similar figures are not available for the abbreviated version of the vocabulary test used; these scores clustered at a level approximating the top end of the average range for Raven's (1958) norms. In summary then, there is evidence of only slight reduction in level of consciousness with this dose of diazepam and ample evidence that short-term processing of input is accurate even TABLE II Intelligence test scores. WAIS scaled scores at the experimental session; WB II weighted scores at followup. (These scores have a mean=10 and SD = 3.) Similarities Median Range Block designs Median Range Diazepam Test day Saline Follow-up Diazepam Saline

5 694 BRITISH JOURNAL OF ANAESTHESIA during the peak period of the dose. There is also good evidence of normal complex processing within an hour or even within 20 minutes of the injection. Memory. The retention tests for the word lists provide measures of storage of material put in during the first few minutes after the injection. The tripleassociate tests offer evidence of what aspects of memory are most affected approximately half an hour after the injection at a time when the subjects looked and sounded almost normal, just after performing well on the decision-making task and just before intelligence tests showed no general intellectual impairment. Figure 3 presents the frequency distributions of the raw scores for the memory lists, list A (animals) served as an index that both groups could retain this type of information over a period of about 100 minutes because it was presented before the injection. Both groups recalled and recognized the names of the animals well and no significant difference could be demonstrated between them. Incidentally, this finding represents some evidence against the development of L L SL._ O I LIST _D_D_nl_«nA retrograde amnesia in the present study. Lists X, Y and Z, presented after the injection, revealed that diazepam does indeed have an amnesic effect. The experimental group were significantly worse than the controls in recalling all three lists (Mann-Whitney U (Siegel, 1956): list X, P=O.OG1; list Y,- P<0.002; list Z, P<0.002). The differences in recognition were also significant (P=0.001, P=0.001 and P<0.02 respectively). Thus the material appears inaccessible to both recall and recognition although it may be presumed to have been perceived since the vigilance task was not seriously impaired. There is a hint that the memory deficit was already receding after 10 minutes since the experimental group succeeded in retrieving some of list Z. It is worth remarking that while all the controls recalled the categories used in at least three of the lists, only one diazepam subject recalled that he had heard more than the single list of animals (Mann-Whitney U, P<0.004). Figure 4 presents the median learning curves for both groups at test and at follow-up on the triple-associates. The immediate recall scores on this test were not significantly different under saline or diazepam, which supports the evidence _ _ _ _ ] D X I U _D _D D U _ _n _n _p p T WORDS RECALLED! EXPERIMENTAL CONTROL D u b L ^ JJ JD WORDS RECOGNISED FIG. 3 Memory list retrieval (raw scores). Vertical axis represents number of subjects in each group obtaining each of the scores along the horizontal axis. Eight items to be retrieved in each list. _D

6 AMNESIC EFFECT OF DIAZEPAM (VALIUM) 695 I T E M S C O RR E.C T SALINE FOLLOW UP weeks to both groups (table II and fig. 4). These again yielded 13 scores. Four of these had shown significant impairment under diazepam, but now it was found that the two groups did not differ significantly on any of the measures. Moreover, the controls had not changed significantly on any of the 13 measures while the experimental subjects had improved on the four measures which had demonstrated impairment. This improvement was statistically significant (P=0.016) for the two learning scores and approached signifiance (P=0.062) for the two retention scores for poorly learnt items (Walsh tests; Siegel, 1956) TRIALS TRIALS FIG. 4 Triple-associate learning: median learning curves for both groups at the experimental session (set I words) and the follow-up session (set II words). Maximum possible score 12 (vertical axis) on each of the four trials (horizontal axis). from the decision-malting task and the intelligence tests that general intellectual efficiency was not much reduced in the period minutes after the injection. The two measures of learning, however, showed the experimental subjects to be clearly inferior. Total correct responses in four trials showed very little overlap (P =0.002) and the number of items reaching criterion in four trials showed no overlap (P=0.001, Mann- Whitney IT). However, once stored the material appears to be normally accessible. On none of the three retention measures were the diazepam subjects different from the controls for the words which had met the learning criterion. The controls were, however, better at recalling the items which failed to meet that criterion (Free recall, P =0.013; Prompt recall, P = 0.032; Recognition, P = 0.09, Mann-Whitney U). Thus there is good evidence of an impairment which is fairly specific to memory and probably particularly involves the input or storing phase rather than retrieval. Follow-up study. In order to demonstrate that the observed deficit had disappeared and to assure ourselves that our two groups of subjects had similar memory abilities, a parallel set of triple-associates and the intelligence tests were given after several DISCUSSION The present findings demonstrate unequivocally that intravenous diazepam at this dose level produces a clinical picture of dense anterograde amnesia at least for about 10 minutes, followed by a period of better but still impaired memory function. The exact nature of these deficits is less clear. There is no direct evidence that the memory lists were perceived or stored under the influence of diazepam. Although the subjects could carry out the vigilance task fairly accurately it is possible that diey may have gone to sleep during the recitation of the lists. This does not seem very likely in that there was no consistent bunching of errors immediately following each list such as might reflect lowering of conscious level. It is useful to point out here that such an explanation could not account for the learning impairment on the triple-associate test. If it is assumed that the memory lists were perceived and briefly stored, then the deficit might be in any of the phases of memory "consolidation", storage, or retrieval. Since recall is usually a more difficult retrieval task rhan recognition, the fact that both were more or less equally affected offers some evidence to implicate the consolidation or storage phases. This conclusion is supported by the difficulty shown in the triplets test half an hour after the injection. At this stage the subjects were clearly conscious and we can be sure that the material was perceived since they had to recall the response words to reach the learning criterion. Subsequently the retrieval tests demonstrate the expected pattern (recall harder than recognition) and show no impairment relative to the saline

7 696 BRITISH JOURNAL OF ANAESTHESIA group for those items adequately learnt. Nevertheless the learning scores were significantly impaired. Since the measure of learning on each trial was the success of a recall process it might be argued that the deficit was merely in this technique of retrieval over the short term. However, such an explanation would not comfortably account for the subsequent differential retrievability of the "well-learnt" and "poorly-learnt" items which showed up to some extent even in recognition. It should be pointed out that Frumin, Herekar and Jarvik (1969) came to similar conclusions using pictorial material although it must be remembered that they used diazepam in combination with hyoscine. It is interesting that the present diazepam subjects show no evidence of retrograde amnesia and may even benefit from the drug in their recall of the memory list A. Summerneld and Steinberg (1957) concluded that nitrous oxide given after learning may facilitate recall by preventing interference from new input, and Pearlman, Sharpless and Jarvik (1961) consider that a general anaesthetic administered after consolidation is complete will similarly preserve the trace whereas, given beforehand, it may block the consolidation process itself and the trace may be lost. If it is accepted that the defect did occur in the consolidation stage it is still arguable whether this should be seen as a difficulty specific to memory or as a reflection of more general lowering of intellectual efficiency. The subjects may have shown poor concentration or confusion in general or even a lowered enthusiasm consequent to the tranquillizing and relaxant properties of diazepam. The very slight effect on the decisionmaking task and the absence of impairment in the immediate memory and intelligence scores (particularly the Block Designs) contraindicates general intellectual impairment. The clinical impression of the psychologist was that motivation was not reduced. The clinical literature on the amnesic syndrome commonly describes relatively good general intellectual efficiency and immediate memory, although these functions clearly suffer in dementia. On occasion severe specific amnesic states show mainly difficulties in retrieval (Clarke, 1961) although clearly the famous case "H.M." failed to show savings except in a mirrortracing task (which he could not recognize) (Milner, 1968). It remains to be established whether such specific amnesic syndromes are essentially different from the possibly more generalized anterograde amnesia observed following unconsciousness. However, the deficit observed with diazepam may be an experimental model for either, allowing finer investigation of the clinical phenomena. To carry this work further, more sensitive monitoring of consciousness and better experimental definition of the phases of memory are needed. ACKNOWLEDGEMENTS We wish to thank Mr P. Herbert, F.LMX.T., chief technician, and Mrs P. Tatt for their help in carrying out this investigation. REFERENCES Bousfield, W. A. (1953). The occurrence of clustering in the recall of randomly arranged associates. J. gen. Psychol., 49, 229. Brown, S. S., and Dundee, J. W. (1968). Clinical studies of induction agents. XXV: Diazepam. Brit. J. Anaesth., 40, 108. Clarke, P. R. F. (1961). Some problems in the clinical appraisal of memory functions (abstract). Bull. Brit. Psychol. Soc., 43, 54. Deutsch, J. A. (1969). The physiological basis of memory. Arm. Rev. Psychol, 20, 85. Feldman, S. A. (1963). A comparative study of four premedications. Anaesthesia, 18, 169. Fox, G. S., Wynands, J. E., and Bhambhami, M. (1968). A rfinirai comparison of diazepam and thiopentone as induction agents to general anaesthesia. Canad. Anaesth. Soc. J., 15, 281. Frisby, J. P., Thornton, J. A., Barratt, R. F., and Bergman, I. The effects of hypoxia and diazepam on a decision-making task. (In preparation.) Frumin, M. J., Herekar, W., and Jarvik, M. E. (1969). The amnesic effects of diazepam and scopolamine in man. Intern. Congr. Pharmacol., 4th, Basle, p Gruber, R. P., and Reed, D. R. (1968). Postoperative anterograde amnesia. Brit. J. Anaesth., 40, 845. Hardy, T. K., and Wakely, D. (1962). The amnesic properties of hyoscine and atropine in pre-anaesthetic medication. Anaesthesia, 17, 331. Irwin, F. W., and Smith, W. (1956). Further tests of theories of decision in an "expanded-judgement" situation. J. exp. Psychol, 52, 345. Knight, P. F., and Burgess, C. G. (eds.) (1968). Diazepam in Anaesthesia. Bristol: Wright Lambrechts, W., and Parkhouse, J. (1961). Postoperative amnesia. Brit. J. Anaesth, 33, 397. Mazzia, V. D., and Randt, C. (1966). Amnesia and eye-movements in first stage anesthesia. Arch. Neurol. {Chic), 14, 522.

8 AMNESIC EFFECT OF DIAZEPAM (VALIUM) 697 Milner, B. (1968). Alteration of memory in man: reflections on methods; in Weiskrantz, L., Analysis of Behavioral Change, p New York: Harper, Row. CNeU, R., and Verrill, P. J. (1969). Intravenous diazepam in minor oral surgery. Brit. J. oral Surg., 7, 12. Parkes, M. W. (1968). The pharmacology of diazepam; in Knight and Burgess (1968), Diazepam in Anaesthesia. Bristol: Wright. Pearlman, G A. jr., Sharpless, S. K., and Jarvik, M. E. (1961). Retrograde amnesia produced by anaesthetic and convulsant agents. J. comp. physid. Psychol., 54, 109. Penfield, W., and Milner, B. (1958). The memory deficit produced by bilateral lesions in the hippocampal zone. Arch. NeuroL Psychiat. (Chic.), 79, 475. Raven, J. C (1958). Guide to using the MSI HOI Vocabulary Scale. London: Lewis. Scoville, W. B., and Milner, B. (1957). Loss of recent memory after bilateral hippocampal lesions. J. Neurol. Neurosurg. Psychiat., 20, 11. Siegel, S. (1956). Nonparametric Statistics for the Behavioral Sciences. New York: McGraw-Hill. Summerfield, A., and Steinberg, H. (1957). Reducing interference in forgetting. Quart. J. exp. Psychol, 9, 146. Thorndike, E. L., and Lorge, I. (1944). The Teacher's Word-Book of 30,000 Words. New York: Teachers' College, Univ. Columbia. Wechsler, D. (1946). Manual of the Wechsler-BeUevue Intelligence Scale, Form II. New York: Psychol. Corp. (1955). Manual of the Wechsler Adult Intelligence Scale. New York: PsychoL Corp. Whitty, G W. M., and Zangwill, O. L. (eds.) (1966). Amnesia. London: Butterworth. L'EFFET AMNESIQUE DE DIAZEPAM (VALIUM) SOMMAIRE Un essai a double-insu, comparant diazepam intraveineux (0,24 mg/kg) a une solution saline, a iti fait chez douze jeunes volontaires masculins intelligents pour examiner l'effet amnesique du meclicament. Des tests psychologiques complexes occuperent plus de 2 heures. Une forte amnesie anterograde a et observee durant approxirnativement 10 minutes, avec une deterioration moins grave de la memoire durant encore au moins une demi-heure apres rinjection. On considere que les difficultes de memoire n'etaient pas dues a une reduction marquee du nivcau de conscience mais qu'elles concernent plus ks processus de reception ou "consolidation" que de recouvrement, D"autres Etudes sont necessaires pour definir plus nettement la nature du deficit. AMNESIE NACH DIAZEPAM (VALIUM) ZUSAMMENFASSUNG Zwolf junge, intelligente, mannliche Freiwillige beteiligten sich an einem Doppelblind-Versuch mit intravenosen Gaben von 0,24 mg/kg Diazepam im Vergleich mit Kochsalz, um die durch dieses Mittel hervorgerufene Amnesie zu untersuchen. Im Zeitraum von 2 Stunden wurde ein komplexes psychologisches Testprogramrn durchgefuhrt. Schwere anterograde Amnesie trat ungefahr 10 Minuten ein nach der Injektion, wobei die Beeintrachtigung des Gedachtnisses geringer war, jedoch mindestens eine halbe Stunde nach der Injektion anhielt Wir gelangten zu dem SchluO, dao die Gedachtnisschwierigkeitcn nicht aufgrund einer konstanten Verminderung des Bewufltseinsgrad beruhen, sondern mehr den Einlagerungsoder "Konsolidations"prozess beeinflussen und nicht die Wiedererinnerung. Weitere Untersuchungen sind notwendig, um die grundlegenden Auswirkungen zu erhellen. PROLOKGED TRACHEAL INTUBATION: CORRESPONDENCE THE PROBLEM OF CHANGING THE TUBE: AN EASY SOLUTION Sir, Prolonged tracheal intubation is commonly used, as an alternative to tracheostomy, in a variety of conditions coma, respiratory distress and has gained the favour of many anaesthetists, since it is easily performed, and avoids some of the well-known complications of tracheostomy. Intubation may be achieved through the oral or the nasal routes, the nasal being favoured by most, since it is more acceptable to a conscious patient, and less likely to lead to complications such as biting on the tube or displacement. However, whether the tube is introduced through the nose, the mouth or a tracheostomy, it is likely to become obstructed by plugs of mucus, blood or secretions, and therefore has to be removed and replaced by a new one. The replacement is not always easy, since the patient may be unco-operative or in severe respiratory distress, and the procedure has to be carried out as an emergency. Blind nasal intubation is not always successful at the first attempt; laryngoscopy may prove difficult, if not impossible, and may induce trauma. Even in the case of tracheostomy, on a fat, short-necked patient, unless the tube has been in site for some time and the route well established, changing a canmils is likely to be a problem. We have tried successfully a simple technique, which makes the exchange quick and simple. A lubricated stiff ureteral catheter is first inserted within the tracheal tube, whether nasal, oral or tracheostomy; this serves as a guide along which the tube is removed and the new one re-inserted. The whole procedure lasts only a few seconds; it saves time and avoids manoeuvres. We suggest, without having any experience on the subject, that, if secretions are suspected, a Fogarty catheter might be used, the balloon inflated, and the trachea swept, in one stroke, such as is done in case of arterial embolism. G. VOURC'H Paris