PSYCHOMOTOR RECOVERY AFTER THREE METHODS OF SEDATION DURING SPINAL ANAESTHESIA

Transcription

1 British Journal of Anaesthesia 1990; 64: SYCHOMOTOR RECOVERY AFTER THREE METHODS OF SEDATION DURING SINAL ANAESTHESIA I. G. KESTIN,. B. HARVEY AND C. NIXON SUMMARY In a double-blinded study, we have investigated psychomotor recovery after three sedation schemes during spinal anaesthesia for transurethral resection of prostate. Thirty-nine patients were allocated randomly to receive i.v. midazolam only, i.v. midazolam with antagonism by f/umazenil or an infusion of propofol. The psychomotor tests were of long and short term memory, critical flicker fusion threshold, attention and concentration, cognition, mental arithmetic, and sedation and anxiety levels, and were performed before operation and at 30 min, 60 min, 2 h and after operation. Free recall of a picture learning task was the most sensitive test, showing a significant impairment in all groups, persisting for more than 2 h. ropofol was associated with the most rapid recovery and least impairment compared with preoperative baseline. erformance improved in all the tests after antagonism with f/umazenil, but was still significantly impaired compared with baseline in tests of picture recall and letter deletion. Significant decreases in performance in all tests were detected 1 h after flumazenil. No increase in anxiety was recorded after flumazenil. We conclude that infusion of propofol is the technique with the most rapid recovery, while only brief and incomplete antagonism of the effects of midazolam may be expected with flumazenil. KEY WORDS Recovery: psychomotor function. Sedation: midazolam. propofol. flumazenil. Some of the advantages of regional anaesthesia in the elderly include diminution in postoperative confusion- and mental deterioration [1,2]. An additional advantage of regional anaesthesia for transurethral resection of prostate (TUR) is that mental confusion is one of the signs of TUR syndrome, and this sign is masked by general anaesthesia. The advantages of regional anaesthesia may be lost if the patient remains heavily sedated at the end of a procedure, and it has been suggested also that heavy sedation may have been associated with cardiac arrest during spinal anaesthesia [3]. It is possible that propofol or midazolam with flumazenil might provide perioperative sedation, but with rapid postoperative psychomotor recovery. We have therefore compared psychomotor recovery after three sedation schemes: midazolam alone, midazolam with flumazenil at the end of surgery, and an infusion of propofol. ATIENTS AND METHODS After District Ethics Committee approval and informed patient consent, we studied 39 patients undergoing TUR under spinal anaesthesia using a randomized, double-blind method. atient exclusions included those with CNS disorders or concurrent benzodiazepine therapy. The psychomotor tests were administered to the patients in the same order by the same investigator on the day before surgery and at 30 min, 60 min, 2 h and after operation. Memory. One minute to memorize a card showing pictures of nine familiar objects. Critical flicker fusion threshold (CFFT). This was measured three times with ascending frequency and three times in descending frequency and the mean of the six readings taken as the CFFT. I. G. KESTTN*, F.F./LR.C.S. ;.B.HARVEY, F.F.A.R.C.S.; C. NIXON, F.F.A.R.C.S.; Department of Anaesthesia, lymouth General Hospital, lymouth L4 8QQ. Accepted for ublication: June 16, resent address, for correspondence: Sir Humphry Davy Department of Anaesthesia, Bristol Royal Infirmary, Bristol BS2 8HW.

2 676 BRITISH JOURNAL OF ANAESTHESIA Serial 7s. Reverse counting aloud from 100 to 44 in steps of 7. The time taken and number of errors were scored. Letter deletion. One minute to delete as many of the letters p and b as possible. Correct deletions and the errors (incorrect deletions or omissions) were scored. Forward and backward digit span. Reverse counting. The number of errors and time taken to write all the numbers from 1000 down to 976. Backward spelling of a common five-letter word. atients then performed a simple unscored task so that the interval between Memory and icture recall was always 15 min. icture recall. The patient was asked to recall the objects memorized in the first test (1 min allowed), and the number recalled was taken as the free recall score. Verbal clues were given for the remaining objects, and any additional objects remembered were noted as the verbal cued recall score. Then the patient was given another picture card showing 20 objects, including the nine original objects, and asked to point out those that had appeared on the original card. The total number recognized was taken as the visual recognition score (the patient was told that incorrect identifications were subtracted from the score, to discourage guessing). There were five different cards used for the picture recall test, together with a corresponding pair for the visual recognition test. No object on any of the cards appeared in more than one test sequence. Before the patient performed the postoperative test, two 10-cm visual analogue scales (VAS) were used to measure the patient's own subjective assessment of sedation and anxiety. One end of the sedation scale was randomly labelled "asleep", and the other "mentally fully alert", and the distance from the alert end was measured as the sedation score. One end of the anxiety scale was labelled "worst anxiety possible", the other "completely calm" and the central 1-cm area was shaded and labelled "usual"; the distance from the calm end was taken as the anxiety score. Anaesthetic management No premedication was used and the patients were allocated randomly to one of three groups: group i.v. midazolam only; group i.v. midazolam followed by antagonism with flumazenil; group i.v. propofol. In groups and, increments of midazolam 1 mg were given i.v. until the patient reached sedation level 3, and further midazolam was given i.v. to maintain this level of sedation during surgery. The anaesthetist was unaware of the group to which the patient belonged. In group, propofol was given using a loading dose and three-stage infusion scheme. The suggested dose in the study programme was: loading dose of 0.5 mg kg" 1 over 20 s, followed immediately by an infusion of 5 mg kg" 1 h" 1 for 10 min, 4 mg kg" 1 h" 1 for 10 min and 3 mg kg" 1 h" 1 thereafter. The anaesthetist was free to choose a multiple of this whole scheme on clinical grounds, and could vary the maintenance infusion rate to maintain the patient at sedation level 3. In all three groups, 0.5% bupivacaine in 8% glucose was used for the spinal block and the patient breathed 40 % oxygen during surgery. The following sedation scale was used: 1 = fully alert and orientated; 2 = drowsy; 3 = eyes closed, rousable to command; 4 = eyes closed, reusable with mild physical stimuli, for example earlobe tug; 5 = asleep, unrousable with mild physical stimuli. The sedation level was recorded every 5 min during surgery. In the Recovery Unit, the first psychomotor tests were administered 30 min after the end of surgery. At 50 min after surgery, 5 ml of either flumazenil 500 ug (group ) or saline (groups and ) was given in a double-blinded method: 2 ml initially, followed at 1-min intervals by 1-ml increments. The psychomotor tests were repeated at 60 min, 2 h and after surgery. Data analysis The continuously variable data were analysed for normality by the Schapiro-Wilk test. Skewed data were transformed logarithmically and tested for normality. These data subsets were analysed by ANOVA for repeated measures, with pairwise comparisons made using the Fisher protected least significant difference method. Ordinal data were analysed by the Kruskal-Wallis test, with pairwise comparisons made using the method suggested by Conover [4]. The overall statistical significance level was 5 %. RESULTS There were 13 patients in each group, and no significant differences between the groups (table I). In the propofol group, the mean initial loading dose was 0.54 mg kg" 1 (range mg kg" 1 ),

3 SYCHOMOTOR RECOVERY 677 TABLE I. atient and anaesthetic data (mean (SD)) Age (yr) Weight (kg) Duration of sedation (min) Midazolam dose (mg kg" 1 ) (n = 13) (n= 13) (n = 13) 68(7) 73(5) 71(7) 72 (10) 77(7) 71 (10) 46(12) 45(11) 40(12) 0.08 (0.02) 0.07 (0.03) o Before op. 30 min 60 min Time after op. FIG. 1. Serial 7s test (mean, SEM) in the midazolam-saline () (A)> midazolam-flumazenil () ( ) and propofol () (x) groups. < 0.05 compared with: * group ; $ group ; f preoperative baseline ' 33 ^ 32 t 31 & t* Before op. 30 min 60 min Time after op. t* 24h FIG. 2. Critical flicker fusion threshold (CFFT) (mean, SEM) in the midazolam-saline () (A), midazolam-flumazeni] () ( ) and propofol () (x) groups. < 0.05 compared with: * group ; $ group ; t preoperative baseline. followed by the two short-duration infusions with the appropriate doses. All patients achieved sedation level 2 or 3 within 5 min, and the mean rate of the maintenance infusion of propofol was 3.7 mg kg" 1 h" 1 (95 % confidence limits mg kg" 1 h" 1 ). Three patients received additional i.v. boluses of propofol mg to achieve a brief increase in sedation level. There

4 678 BRITISH JOURNAL OF ANAESTHESIA i= Before op. 30min 60min Time after op. FIG. 3. Reverse counting test (mean, SEM) in the midazolam-saline () (A), midazoum-flumazeni] () ( ) and propofol () (x) groups. < 0.05 compared with: *group ; \ preoperative baseline. TABLE II. ostoperative anxiety scores (mean (SD)) 30min 6.1(1.2) 5.8(1.2) 6.5(1.1) ostoperative anxiety scores (cm) 60 min 6.2 (0.7) 6.3(1.1) 5.9(1.0) 5.7 (0.8) 5.9 (0.9) 5.6 (0.7) 5.2 (0.5) 5.5(0.5) 5.1 (0.4) TABLE III. ostoperative sedation scores (mean (95% confidence limits)). < 0.05 compared with: * group ; $ group ; -\30-min score TABLE IV. Free recall: number of objects recalled (median (range)). < 0.05 compared with: * group ; %group ; t preoperative baseline baseline 6 (4-8) 6 (3-9) 5 (4-9) 30 min It* (0-4) 0+* (0-3) 3t (1-7) ostoperative score 60 min It* (0-4) 3tS (0-5) 4t (1-5) 2f* (1-4) 2f* (1-3) 4t (2-5) 4t (3-6) 5+ (3-6) 5 (4-7) ostoperative sedation score (cm) 30 min 60 min 7.3* ( ) 7.5* ( ) 5.1 ( ) 6.4* ( ) 4.0$t ( ) 3.7 ( ) 3.8 ( ) 3.9 ( ) 2.6 ( ) 0.9 ( ) 0.7 ( ) 0.7 ( ) were no significant differences between groups in duration or depth of sedation. Four patients received i.m. pethidine for postoperative analgesia within the first, and were excluded from the 24-h psychomotor testing. There were no significant differences between the three groups in baseline preoperative performance in any of the tests. The results of the psychomotor tests are shown in figures 1-3 and tables II-V. There were no significant changes between any of the groups or from baseline performance within each group in the following tests: verbal cued recall, visual recognition recall, forward digit span or error rate in serial 7s and reverse counting. 30 min after anaesthesia There were no significant differences between groups and and both groups were significantly different from group and their own baseline performance in the following tests: free recall, CFFT, serial 7s, letter deletion, reverse counting, and backward spelling. There was a significant difference from baseline performance in groups and in reverse digit span. The VAS score for sedation arousal in these groups was significantly different from group, in which free recall, CFFT and serial 7s tests were significantly different from their own baseline.

5 SYCHOMOTOR RECOVERY 679 TABLE V. Letter deletion test: number of letters deleted (median (range)). < 0-OS compared with preoperative baseline ostoperative scores baseline 30min 60min 26 (17-39) 24 (13-38) 21.5 (11-35) 12f (7-24) 16.5f (4-31) 19 (7-33) 15t (4-28) 15t (8-34) 17 (9-31) 19+ (9-42) 18 (8-30) 19 (8-39) 26.5 (7-38) 31.5 (17-45) 24 (17-^3) 60 min after anaesthesia After administration of flumazenil, performance was improved significantly in group compared with group in tests of free recall, CFFT, serial 7s, backward spelling, backward digit span and the sedation score, but was significantly below baseline in free recall and letter deletion (tables IV, V). There were no significant differences between groups and, or between group and their baseline performance except for free recall. Flumazenil was associated with a decrease in sedation, but no increase in anxiety (tables II, III). 2 h after anaesthesia There were no significant differences between groups and, and both groups were significantly inferior to group and their own baseline performance in the following tests: free recall, CFFT and serial 7s. after anaesthesia There were no differences between the three groups or from their own baseline performance in any of the tests except for free recall, where groups and were below baseline performance. DISCUSSION The psychomotor tests used in this study tested the patients for memory, both "long term" (picture recall) and "short term" (digit span), attention and concentration (serial 7s, reverse counting and spelling), central integration (CFFT), cognition (mental arithmetic, backward digit span and backward spelling), and subjective assessment of sedation and anxiety. The obvious effect of sedation with midazolam was profound memory deficit for free recall at the picture learning task, while performance of the visual recognition test was unimpaired. This difference between performance in tests of free recall compared with visual recognition recall has been shown to be associated with low levels of sedation with benzodiazepines [5]. In that study [5], free recall was impaired initially, with relative preservation of recognition recall, and as the dose of benzodiazepine was increased, recognition recall was affected progressively. Both our study and that of Sage, Close and Boas [5] would suggest, therefore, that, the memory deficit with low doses of benzodiazepines involves access to memory trace, rather than formation (as memory is available for recognition but not for free recall). Other workers have suggested otherwise : that benzodiazepines affect principally consolidation and storage memory processes, rather than access [6, 7]. Using an auditory learning task, Clarke and colleagues [7] found that both free recall and verbal recognition were affected by benzodiazepines, and concluded that the memory deficit was in the consolidation process. However, as shown by Sage's group, this pattern is characteristic of larger doses of benzodiazepines. Clarke and colleagues found also that visual recognition was not affected by benzodiazepines, which would support the hypothesis that the memory deficit involves access, rather than formation (at least for visual material). Forward digit span was found to be unaffected by sedation with midazolam, whereas there was significant impairment at 30 min in the test of backward digit span in the midazolam groups. Other studies have shown that short term memory is not affected by benzodiazepines [6]. Backward digit span, therefore, must involve additional cognitive mechanisms sensitive to midazolam, in addition to short term memory, which is not

6 680 BRITISH JOURNAL OF ANAESTHESIA affected by midazolam. This has been found also by Adam [8]. Backward spelling probably tests the same mechanisms as backward digit span, and both tests in this study showed a similar pattern of results. Flumazenil was found to antagonize only partially most of the psychomotor effects of midazolam. erformance after flumazenil had been given did not consistently achieve preoperative levels. Some studies in volunteers have shown full antagonism of the cognitive and amnesic effects of midazolam with large doses of flumazenil either 200 mg orally [6,9] or 10 mg i.v. [10]. All clinical studies have documented excellent antagonism of sedation after larger doses of benzodiazepines than used in this study [5, 11]. Flumazenil was associated with greater alertness, but no increase in anxiety; some studies have reported increases in anxiety after large doses of flumazenil [12, 13]. The duration of the flumazenil antagonism was brief: 1 h after flumazenil, patients were indistinguishable from the group who received midazolam alone, and were below baseline performance in several tests. Resedation has been reported also in other studies [5, 6, 14, 15]. The implications of these results are that only brief and incomplete antagonism of the amnesic and cognitive effects of midazolam may be expected from i.v. flumazenil. Oral flumazenil has been shown to have a longer duration of action, but there was still evidence of resedation at 6h [6]. This is likely to limit its routine use in clinical anaesthesia. Even though the doses of midazolam used in this study were relatively low (mean mg kg" 1 ), diminished cognition, memory attention and concentration and central integration persisted for more than 2 h in the midazolam group. Sedation with an infusion of propofol was found to be associated with the most rapid psychomotor recovery. The propofol scheme used in this study was derived from the study by Roberts and colleagues [16], who reported a regimen designed to achieve and maintain a blood concentration of 3 ug ml" 1 within 2 min. This concentration is sufficient for surgical anaesthesia; as the pharmacokinetics of propofol are linear [17], multiples of this regimen should alter blood concentration accordingly. The concentration that produces sedation rather than anaesthesia is not known, so it was chosen initially in this study to use 50 % of the dose regimen used by Roberts [16], giving an estimated blood concentration of 1.5 ug ml" 1. The advantage of using this loading dose and three-stage infusion technique is that all the patients were sedated within 5 min. Another study of an infusion of propofol for sedation used a two-stage infusion technique without a loading dose, and the patients required 26 min to achieve the maintenance sedation level [18]. We conclude that an infusion of propofol is the technique with the most rapid recovery after sedation during spinal anaesthesia. The cognitive and amnesic effects of low dose midazolam sedation may persist for more than 2 h after operation, and flumazenil may antagonize only briefly and incompletely these effects of midazolam. REFERENCES 1. Chung F, Meier R, Lautenschlager E, Cannichael FJ, Chung A. General or spinal anesthesia: which is better in the elderly? Anesthtsiology 1987; 67: Hole A, Terjesen T, Breivik H. Epidural versus general anaesthesia for total hip arthroplasty in elderly patients. Acta Anaesthesiologica Scandinavica 1980; 24: Caplan RA, Ward RJ, osner K, Cheney FW. Unexpected cardiac arrest during spinal anesthesia: a closed claims analysis of predisposing factors. Antsthtsiobgy 1988; 68: Conover WJ. ractical Nonparametric Statistics. New York: John Wiley and Sons, 1980; Sage DJ, Close A, Boas RA. Reversal of midazolam sedation with Ancxate. British Journal of Anaesthesia 1987; 59: O'Boyle C, Lambe R, Darragh A, Taffe W, Brick I, Kenny M. Ro antagonizes the effects of diazepam in man without affecting its bioavailability. British Journal of Anaesthesia 1983; 55: Clarke RF, Eccersley S, Frisby J, Thornton JA. The amnesic effects of diazepam (Valium). British Journal of Anaesthesia 1970; 42: Adam N. Disruption of memory functions associated with general anesthetics. In: Kilstrom JF, Evans FJ, eds. Functional Disorders of Memory. Hillsdale, New Jersey: Lawrence Erlbaum Association, 1979; Darragh A, Lambe R, Scully M, Brick I, O'Boyle C, Wilson Downie W. Investigation in man of the efficacy of a benzodiazepine antagonist, Ro Lancet 1981; 2: Lauven M, Schwilden H, Stoeckel H, Greenblatt DJ. The effects of a benzodiazepine antagonist Ro in the presence of stable concentrations of midazolam. Anesthesiology 1985; 63: Kirkegaard L, Knudsen L, Jensen S, Kruse A. Benzodiazepine antagonist Ro Antagonism of diazepam sedation in outpatients undergoing gastroscopy. Anaesthesia 1986; 41: Louis M, Forster A, Suter M, Gemperle M. Clinical and hemodynamic effects of a specific benzodiazepine antagonist (Ro ) after open heart surgery. Anesthesiology 1984; 61: A Ricou R, Forster A, Bruckner A, Chastonay, Gemperle M. Clinical evaluation of a specific benzodiazepine

7 SYCHOMOTOR RECOVERY 681 antagonist (Ro ). Studies in elderly patients after regional anaesthesia under benzodiazepine sedation. British Journal of Anaesthesia 1986; 58: Klotz U, Kanto J. harmacokinetics and clinical use of flumazenil (Ro ). Clinical harmacokinetics 1988; 14: Duvaldestein, Lebrault C, Guirimand F, Raimbault E. Efficacy of flumazenil reversal after midazolam induced anesthesia. Anesthesiology 1988; 63: A Roberts FL, Dixon J, Lewis GTR, Tackley RM, rys- Roberts C. Induction and maintenance of propofol anaesthesia. A manual infusion scheme. Anaesthesia 1988; 43 (Suppl.): Spelina KR, Coates D, Monk CR, rys-roberts C, Norley I, Turtle MJ. Dose requirements of propofol by infusion during nitrous oxide anaesthesia in man. I: atients premedicated with morphine sulphate. British Journal of Anaesthesia 1986; 58: Wilson E, Mackenzie N, Grant IS. A comparison of propofol and midazolam by infusion to provide sedation in patients who receive spinal anaesthesia. Anaesthesia 1988; 43 (Suppl.):