SEDATION OF CHILDREN REQUIRING ARTIFICIAL VENTILATION USING AN INFUSION OF MIDAZOLAM

Transcription

1 Br.J. Anaesth. (986), 8, 0-08 SEDATION OF CHILDREN REQUIRING ARTIFICIAL VENTILATION USING AN INFUSION OF MIDAZOLAM P. D. BOOKER, A. BEECHEY AND A. R. LLOYD-THOMAS Long-term intubation of the trachea and intermittent positive pressure ventilation (IPPV) are among the commonest procedures carried out in the Intensive Care Unit (ICU). It is recognized that these can be subjectively distressing (Henschel, 9; Shovelton, 99; Green, 980) and, consequently, sedation is usually required. Since the withdrawal of alphaxalone alphadolone (Althesin, Glaxo), and the withdrawal of etomidate (Hypnomidate, Janssen) from use as sedatives in the ICU, the range of drugs available for this purpose is limited. Although the combination of an opioid with a benzodiazepine is used frequently, the durations of action of certain benzodiazepines, or their metabolites, can cause prolonged sedation (Editorial, 98). In addition, some i.v. formulations are associated with a high incidence of thrombophlebitis (Kawar and Dundee, 982). Moreover, if sedatives are prescribed and given on an intermittent "as required" basis to such patients, there is a danger that sedation may be inadequate (Miller-Jones, 980). Midazolam (Hypnovel, Roche) is an imidazolebenzodiazepine presented in aqueous solution. When administered as a single i.v. bolus, it has a plasma half-life of approximately 20 min (Smith, Eadie and O'Rourke-Brophy, 98). Its two major psychoactive metabolites have even shorter half-lives and do not affect the overall duration of action of the drug (Clarke, 982). Thrombophleb- P. D.BOOKER, M.B., F.F.A.R.CS.; A. BEECHEY,* M.B., F.F.A.R.CS.; A. R. LLOYD-THOMAS,** M.B., F.F.A.R.CS.; Royal Liverpool Children's Hospital, Myrtle Street, Liverpool L DG. Present addresses: Department of Anaesthesia, Queens Medical Centre, Nottingham "Department of Anaesthesia, Saint Bartholomew's Hospital, West Smithfield, London ECA BE. Correspondence to P.D.B. SUMMARY The sedation of 0 children aged 6 months to 9 years who had undergone open heart surgery was studied. During artificial ventilation a midazolam infusion was used in conjunction with the administration of morphine {and tubocurarine). Sedation for patients breathing spontaneously with positive airway pressure was continued with midazolam alone. The duration of the midazolam infusion (2-6 fig kg* min- ) ranged from 2 to 9 h. Forty-seven of the children were sedated uneventfully; the remaining three children needed small doses of other sedative agents. In 0 of the children, blood samples were taken for serum midazolam assay and a short Synacthen test was performed. There was no clinical evidence of accumulation of midazolam, but midazolam concentrations were so variable that no conclusions could be drawn. All patients in whom they were measured (n = 0) had high basal cortisol concentrations, but displayed normal responses to Synacthen. itis has not been a problem (Al-Khudairi, Whitwam and McCloy, 982). These properties suggested that it could be used to provide long-term sedation when given by i.v. infusion. Clinical experience with midazolam used in this way is limited, particularly in paediatric patients. This paper describes the use of an infusion of midazolam in 0 children who required sedation after open heart surgery. PATIENTS AND METHODS Fifty children, aged 6 months to 9 yr, were studied following elective cardiac surgery with cardiopulmonary bypass. The indications for surgery are

2 MIDAZOLAM INFUSION IN CHILDREN 0 TABLE I. Indication! for surgery Atrial septal defect Ventricular septal defect Complex ventricular septal defect Transposition of the great vessels Fallot's tetralogy Aortic stenosis Other valve defects Other shown in table I. Patients who had been prescribed benzodiazepines in the previous 2 weeks or who had liver dysfunction other than as a result of cardiac failure were excluded. The children were premedicated with trimeprazine. mg kg" by mouth 3 h before operation. Morphine 20 ug kg" i.m. and atropine 20 ug kg" i.m. were given 2 h later. Anaesthesia was induced with thiopentone mg kg" and maintained subsequently with 0-66% nitrous oxide in oxygen. Morphine 20 ug kg" i.v. and tubocurarine mg kg" i.v. were given before tracheal intubation was performed. Patients were ventilated manually, using the Jackson-Rees modification of Ayre's T piece, at rates of 0-80 b.p.m., with sufficient fresh gas flow to produce mild hypocarbia (Pa COl. kpa). Monitoring was instituted and venous access secured, after which a bolus of midazolam 200 ug kg" was given. Before surgery an infusion of midazolam was started at a rate of 2 ug kg" min". Infusions were made up by diluting midazolam 3 mg kg" in 0 ml of % glucose in water. The infusion and bolus doses of midazolam were given through an i.v. cannula reserved exclusively for this purpose. On the institution of cardiopulmonary bypass, patients were cooled to 6-26 C and another dose of tubocurarine mg kg" was given. No more neuromuscular blocking agent was given until the patient arrived in the ICU. Rewarming to a temperature greater than 3 C was achieved before bypass was discontinued. The infusion of midazolam was stopped at the time of insertion of the last skin suture. The patients were then transferred to the ICU. Every min thereafter, a standard stimulus an orange stick drawn firmly along the sole of the foot was applied. Sedative drugs were recommenced when the child had recovered sufficiently either to respond to this stimulus by withdrawing the whole leg or to show spontaneous eye-opening. An infusion of morphine was started at the rate of 0. ug kg" min" in those aged 6 months to yr and at 0.3 ug kg" min" in those older than yr. A bolus of midazolam 200 ug kg" was given and the infusion restarted, initially at 2 ug kg" min". Tubocurarine 200 ug kg" was also given; subsequent doses of neuromuscular blocker were limited to a maximum of 00 ug kg" h". Quality of sedation was assessed by the nursing staff, who adjusted the rate of infusion as required. Patients were considered to be adequately sedated if they were asleep, tolerating ventilation and yet able to show a slight response to nursing procedures. If the patient was restless, the rate of the infusion of midazolam was increased in steps of ug kg" min", until satisfactory sedation was achieved. If there was no response to nursing procedures, the rate of infusion was decreased. The duration of postoperative IPPV was determined by cardiovascular and respiratory variables. Pa COt was maintained in the range -. kpa and arterial and central venous pressures were monitored continuously. Urine output was recorded hourly. When the patient's condition had improved such that weaning from IPPV was appropriate, the morphine and tubocurarine were stopped. Spontaneous respiration with continuous positive airway pressure (CPAP) was initiated when at least h had elapsed after the administration of the last dose of neuromuscular blocking drug. The spontaneously breathing patient was then sedated with midazolam alone, until considered fit for extubation. The infusion of midazolam was stopped and extubation was performed when the patient was able to respond to verbal commands or was opening his eyes spontaneously. All assessments and decisions concerning weaning from IPPV and extubation were taken by one person (P.D.B.). Blood samples for assay of midazolam concentrations were taken from 0 patients at the following times: () After the induction of anaesthesia. (2) At the time of last skin suture. (3) When first awake in intensive care. () 2.00 and daily. () On cessation of infusion of midazolam. (6) When the patient awoke and was ready for extubation.

3 06 BRITISH JOURNAL OF ANAESTHESIA TABLE II. Age and sex of the patients studied TABLE IV. Duration of midazolam infusion Age (y) Male Female < < TABLE III. Dose range of midazolam infusions Duration (h) tt > 92 Number 2 0 Dose min" ) < ^ - -6 >6 Number Serum was separated, frozen and assayed (Roche Laboratories PLC). A short Synacthen test was performed on these patients when they were stable, not requiring inotropic support and were sedated by midazolam alone. Within 8 h of their transfer to the paediatric ward, patients older than 6 yr were interviewed about any recollections of their stay in the intensive care unit. RESULTS In this series there were 2 boys and 26 girls whose ages ranged from 6 months to 9 yr (table II). In general, they weighed less than the median weight for their age. Distributions of age and weight were not markedly skewed. The mean duration of the perioperative infusion of midazolam was 208 min (SD 68). After return to the ICU, the mean time to leg withdrawal was 32 min (SD 8). There was no correlation between this time and the duration of the perioperative infusion (r = 0.06). Clinically adequate sedation was obtained in patients (9%). The midazolam infusion rate ranged from.6 to 9. ig kg" min~ l (mean 3.8, SD.3). Only two patients required less than 2 ug kg" min" and one, more than 6 Hg kg" min" (table III). The rate of infusion 0 did not correlate with age (r = 0.0) or weight (r = 0.09). The mean duration of the infusion of midazolam was 6. h (range 3-9, SD 2.9). Eight patients were sedated for longer than 96 h (table IV). At no time was any change in cardiovascular variables, or the need for cardiovascular support, attributed to the infusion of midazolam. The mean time to recovery was 02 min (range 30-32, SD ). Overall, 9 patients (98%) were judged to be ready for removal of the tracheal tube within h of stopping the midazolam and in all the trachea was extubated uneventfully. The remaining patient required 32 min to recover sufficiently to permit extubation. In retrospect, his sedation had exceeded the requirements of the study. The time from stopping the infusion to recovery and extubation did not correlate with age (r = 0.6) or weight (r = 0.8). However, in the one patient who was oversedated, recovery was slow. Cortisol secretion was not inhibited by this sedative regimen. The mean plasma cortisol concentration was 9 nmol litre- (SD 26). Following the administration of Synacthen, the mean concentration increased to 328 nmol litre" (SD 3). One patient (2%) displayed an area of redness around the infusion site. No other local sequelae were seen. Of the patients interviewed after discharge from the ICU, none could recall anything of the period of intubation, whether during the period of artificial ventilation or when breathing spontaneously. DISCUSSION In paediatric intensive care, it is essential that adequate sedation be provided, not only to aid compliance with artificial ventilation, but also to

4 MIDAZOLAM INFUSION IN CHILDREN prevent awareness of a very frightening environment. Ideally, the sedative agent should have no cardiovascular or respiratory effects and its short elimination half-life would be by metabolic pathways independent of hepatic, renal and pulmonary function (Editorial, 98). With the recent loss of etomidate and Althesin, sedation in intensive care units is often provided by narcotic analgesics and benzodiazepines (Buchanan and Cane, 98; Farina, Levati and Tognoni, 98). Indeed, it was our usual practice to administer a long-acting benzodiazepine as part of the anaesthetic techniques described above. On return to the ICU, patients were sedated with morphine by infusion and intermittent doses of diazepam. Compliance with artificial ventilation was aided by the administration of tubocurarine. However, placing a limit of 00 jig kg" h" on the dose of the neuromuscular blocking drug resulted in unacceptable quantities of diazepam being given. This caused a prolongation of the time to extubation because the children were oversedated. It also implies that our previous sedative regimen was inadequate. Accordingly, we decided to continue to limit the prescription of neuromuscular blockers and to maintain existing infusion rates of morphine which had been shown to produce satisfactory analgesia in children (Lynn, Opheim and Tyler, 98). However, we decided to replace the long-acting benzodiazepine given during anaesthesia and the intermittent diazepam given in the intensive care unit by an infusion of midazolam. Midazolam has been shown to be a safe and acceptable drug in paediatric practice when given as a single i.m. or i.v. injection either as premedication or as an induction agent (Cole, 982; Sjovall et al., 98; Gemperle and Rouge, 98). However, there have been no reports of its use as an infusion in paediatric practice. Therefore, we performed a dose ranging pilot study in which, after a bolus dose of 200 ug kg", an initial infusion rate of 2 ug kg" min" was found to be optimal. The formulation of midazolam, in particular its water solubility, has facilitated its use (Fragen, Gahl and Caldwell, 98). The low incidence of thrombophlebitis in our series (2%) confirms its acceptable formulation, and is similar to that found in other series (Al-Khudairi, Whitwam and McCloy, 982). The infusion of midazolam provided smooth and reliable sedation, especially in the avoidance 0 of sleep-wakening cycles often associated with the technique of "as required" sedation. Almost all patients (9%) were adequately sedated with infusion rates of between 2 and 6 ug kg" min". The rapid onset of hypnotic activity (-2 min) after i.v. injection, and the relatively brief half-life of midazolam (Smith, Eadie and O'Rourke- Brophy, 98) allowed the nursing staff to titrate the rate of infusion against the response of the patient, thus producing adequate sedation. Bolus doses of 00 ug kg", given by the nursing staff, were authorized if the child became very restless. However, these were not found necessary in this series of patients: simple adjustment of the infusion rate provided an adequately rapid response to control the minor restlessness that indicated inadequate sedation. Midazolam concentrations measured in 0 patients were so variable as to prevent any conclusion being drawn. Variability of effect is a feature of sedation with benzodiazepines in general and has been noted when midazolam was used in other situations (Gamble et al., 98). This may explain why three patients (6%) were sedated inadequately, despite an infusion rate of 6 ug kg" min". The addition of trimeprazine mg kg", given every 6 h via a nasogastric tube, provided satisfactory control. The design of this study did not permit statistical analysis of any of the cardiovascular effects of the infusion of midazolam. However, continuous monitoring and nursing of these patients, by staff trained to notice and act upon adverse cardiovascular changes, yield confidence that major adverse effects did not occur, and a clinical impression of cardiovascular stability was gained. Benzodiazepines are not regarded as possessing significant cardiovascular effects. The use of midazolam, compared with thiopentone, as an induction agent, has been associated with greater stability of arterial pressure in response to laryngoscopy and tracheal intubation (Boralessa, Senior and Whitwam, 983). Among our requirements for extubation are that the patient must have full protective laryngeal reflexes, respond to commands or spontaneously open the eyes. Thus we wait for most of the sedation to "wear off" before deciding to extubate the trachea. Forty-nine of 0 patients (98%) met these requirements within h, despite infusions which lasted for 92 h in some patients. Reports of a prolongation of the action of midazolam in patients having major surgery and repeated

5 BRITISH JOURNAL OF ANAESTHESIA 08 administration of the drug (Byatt et al., 98; Harper et al., 98), have not been confirmed by our clinical findings. In these reports, the mode of administration of midazolam has not always been completely clear, but it is possible that the emphasis on constant patient assessment and the titration of the infusion rate against patient response prevented accumulation of the drug in this study. The prolongation of action, which appears to be more pronounced in adults, appears to be an alteration in the pharmacokinetics of the drug. Increases in the distribution half-life and the distribution volume have been reported (Byatt et al., 98). A formal pharmacokinetic study has been undertaken and is reported in the following paper (Lloyd-Thomas and Booker, 986). Midazolam is converted to -hydroxy-methyl and -hydroxy metabolites (Clarke, 982). The peak concentrations of these metabolites are seen -20 min after administration and at h have decreased to 0% of peak values (Clarke, 982). The short half-life of these potentially psychoactive metabolites was reflected in the rapid recovery time of the patients and is in contrast to clinical experience with repeated intermittent doses of diazepam. Concern has been expressed recently over the ACTH-resistant suppression of cortisol synthesis by infusions of etomidate (Ledingham and Watt, 983). This drug is thought to cause enzyme specific inhibition (Wagner et al., 983). However, this is not known to occur with benzodiazepines. In patients receiving midazolam infusions, we found high cortisol concentrations. We conclude that this reflects the high degree of stress to which these children were subjected. Indeed, it is opposite to the depression of cortisol concentrations seen in adults sedated with oxazepam, which is thought to be common to all the benzodiazepines (Gram et al., 98). Nevertheless, the patients studied demonstrated a normal response to Synacthen, albeit superimposed on increased basal concentrations of cortisol. We suggest that an infusion of midazolam, when accompanied by an infusion of morphine, provides effective sedation which is superior to that provided by sedative regimens based upon intermittent administration. REFERENCES Al-Khudairi D., Whitwam, J. G., and McCloy, R. G. (982). Midazolam and diazepam for gastroscopy. Anaesthesia, 3, 002. Boralessa, H., Senior D. F., and Whitwam, J. G. (983). Cardiovascular response to intubation. Anaesthesia, 38,623. Buchanan, N., and Cane, R. D. (98). Drug utilization in a general intensive care unit. Intern. Care Med.,,. Byatt, C. M., Lewis, L. D., Dawling, S., and Cochrane, G. M. (98). Accumulation of midazolam in patients receiving mechanical ventilation in an intensive care unit. Br. Med. J., 289, 99. Clarke, R. S. J. (982). Biotransformation of intravenous anaesthetic agents; in Recent Advances in Anaesthesia and Analgesia (eds R. S. Atkinson and C. Langton-Hewer), p.. Edinburgh: Churchill-Livingstone. Cole, W. H. J. (982). Midazolam in paediatric anaesthesia. Anaesth. Intens. Care, 0, 22. Editorial (98). Sedation in the Intensive Care Unit. Lancet,, 388. Farina, M. L., Levati, A., and Tognoni, G. A. (98). A multicentre study of ICU drug utilization. Intens. Care Med.,, 2. Fragen, R. J., Gahl, F., and CaldweU, N. J. (98). A water soluble benzodiazepine, Ro2-398, for induction of anesthesia. Anesthesiology, 9,. Gamble, J. A. S., Kawar, P., Dundee, J. W., Moore, J., and Briggs, L. P. (98). Evaluation of midazolam as an induction agent. Anaesthesia, 36, 868. Gemperle, G. N., and Rouge, J. C. (98). Experience clinique du midazolam en anesthesie pediatriquc. Medicine tt Hygiene, 39, 3. Gram, L. F., Christensen, L., Kristensen, C. B., and Kragh- Sorensen, P. (98). Suppression of plasma cortisol after oral administration of oxazepam in man. Br. J. Clin. Pharmacol.,, 6. Green, D. (980). Paralysis or sedation for controlled ventilation. Lancet,,. Harper, K. W., Collier, P. S., Dundee, J. W., Elliott, P., Halliday, N. J., and Lowry, K. G. (98). Age and nature of operation influence the pharmacokinetics of midazolam. Br. J. Anaesth.,, 866. Henschel, E. O. (9). The Guillain-Barre syndrome. A personal experience. Anesthesiology,, 228. Kawar, P., and Dundee, J. W. (982). Frequency of pain and venous sequelae following the i.v. administration of certain anaesthetics and sedatives. Br. J. Anaesth.,, 93. Ledingham, I. McA., and Watt, I. (983). Influence of sedation on mortality in critically ill multiple trauma patients. Lancet,, 20. Lloyd-Thomas, A. R., and Booker, P. D. (986). Infusion of midazolam in paediatric patients after cardiac surgery. Br. J. Anaesth., 8, 09. Lynn, A. M., Opheim, K. E., and Tyler, D. C. (98). Morphine infusion after pediatric cardiac surgery. Crit. Care Med., 2, 863. Miller-Jones, C. M. H. (980). Sedation for ventilation. A retrospective study of fifty patients. Anaesthesia, 3, 0. Shovelton, D. A. (99). Reflections on an Intensive Therapy Unit. Br. Med. J.,,3. Sjovall, S., Kanto, J., Iisalo, E., Himberg, J. J., and Kangas, L. (98). Midazolam versus atropine plus pethidine as premedi cation in children. Anaesthesia, 39, 22. Smith, M. T., Eadie, M. J., and O'Rourke-Brophy, T. (98). The pharmacokinetics of midazolam in man. Eur. J. Clin. Pharmacol., 9, 2. Wagner, R. L., White, P. F., Kan, P. B., Rosenthal, M. H., and Feldman, D. (98). Inhibition of adrenal steroidogenesis by the anaesthetic etomidate. N. Engl. J. Med., 30,.