Conscious Sedation With Remifentanil and Midazolam During Brief Painful Procedures in Children

Objective  To test the hypothesis that remifentanil, because of its favorable pharmacokinetic properties, would be advantageous to use in combination with midazolam to provide analgesia and sedation during brief painful procedures.

Design  Prospective observation and data collection.

Setting  University hospital.

Patients  Seventeen children aged 2 to 12 years, who underwent 20 brief, painful procedures.

Interventions  Administration of intravenous midazolam hydrochloride, 0.05 mg/kg, and remifentanil hydrochloride, 1 µg/kg, followed by a remifentanil infusion at 0.1 µg·kg⁻¹·min⁻¹. The dose was titrated at 5-minute intervals to levels of sedation and analgesia.

Main Outcome Measures  Successful remifentanil doses, times to discharge readiness, side effects, complications, and requirement for additional medications.

Results  The technique was successful in 17 of 20 procedures. The mean±SD successful dose was 0.4±0.2 µg·kg⁻¹·min⁻¹. Four children developed hypoxemia that abated with mild stimulation; 1 child became unresponsive and required positive-pressure ventilation. The mean±SD time to reach discharge criteria was 9.5±4.3 minutes. Hypoxemia was avoided in 10 of 13 patients by continuous stimulation throughout the procedure.

Conclusion  The use of remifentanil and midazolam during brief, painful procedures results in rapid times to discharge but is complicated by a high incidence of life-threatening respiratory depression at subtherapeutic levels.

THE ROUTINE administration of sedation and analgesia, which often reaches levels of deep sedation¹ during painful medical procedures, has become standard practice in most medical centers in recent years. This improvement in patient care was facilitated by the availability of propofol because of its rapid onset and offset, lack of emetogenic properties, and preservation of spontaneous ventilation at doses that provide adequate hypnosis.^2,3 Despite its safety under usual conditions, propofol administration invariably results in a loss of consciousness at levels required for many painful procedures. In fact, when used during painful procedures, sedative doses of propofol are usually inadequate and result in the administration of doses consistent with a state of deep sedation, which may be indistinguishable from general anesthesia. Thus, the use of propofol is restricted to practitioners who are adequately trained in airway management and maintenance of general anesthesia.

Many agents (eg, propofol) and combinations (eg, midazolam and fentanyl citrate) have been used to provide analgesia and sedation. However, side effects, ranging from minor (eg, nausea, vomiting) to serious (eg, respiratory depression) have been frequently reported.⁴ Remifentanil, an ultra–short-acting, µ-opioid was recently released for use during general anesthesia.⁵ Unlike other natural or synthetic opioids, remifentanil is rapidly hydrolyzed by nonspecific plasma and tissue esterases that impart brevity of action (terminal elimination half-life of 8-10 minutes) and rapid, precise titratability when used as a continuous intravenous infusion. Discontinuation of remifentanil results in an extremely rapid return to the patient's baseline mental and cardiorespiratory status (within 10 minutes in most cases) regardless of the duration of the infusion.⁶ Because of these desirable properties, it was hypothesized that remifentanil would be an ideal opioid to use in combination with midazolam to provide analgesia and anxiolysis to patients undergoing brief, painful procedures.

The research subjects review board of the University of Rochester, Rochester, NY, approved this prospective observational study. Fasted children, aged 2 to 12 years, requiring sedation or general anesthesia during painful medical procedures were eligible to receive remifentanil hydrochloride and midazolam hydrochloride. Verbal and written consent from their parent(s) and assent from the children (when appropriate) were obtained. Exclusion criteria included obesity (>150% ideal body weight) or any systemic disease or craniofacial anomaly that could impair breathing. The majority of the children included in this report had prior placement of an indwelling central venous catheter. Children received intravenous midazolam hydrochloride, 0.05 mg/kg; intravenous ondansetron hydrochloride, 2 mg; and oxygen, administered via nasal cannula at 3 L/min or the "blow-by" technique at 10 L/min. Monitors included continuous electrocardiography and measurement of oxygen saturation by pulse oximetry (SpO₂), and blood pressure determinations at 3-minute intervals. Respiratory rate and depth were continuously assessed by visual inspection throughout the period of sedation. All children had at least one parent in attendance during the procedure. All procedures were performed in an area with standard emergency equipment immediately available in accordance with the recommendations in the American Academy of Pediatrics' guidelines for sedation of pediatric patients.⁷

Following the midazolam premedication, a bolus of remifentanil hydrochloride, 1 µg/kg, was infused over 1 minute, and followed by an infusion dose of 0.1 µg·kg⁻¹·min⁻¹ (via infusion pump). Five minutes after beginning the remifentanil infusion, the children were assessed for level of consciousness and comfort during the injection of skin local anesthesia using a scoring system based on the American Academy of Pediatrics' definitions of "conscious" and "deep" sedation (Table 1).⁷ If the child scored less than 3, indicating he or she had not yet reached a sufficiently sedated condition, the infusion dose was doubled and the child was reassessed in 5 minutes. This process continued until either a score of 3 was attained or respiratory depression, with unresponsiveness to verbal prompting or tactile stimulation, intervened. The 5-minute interval was chosen based on the known pharmacokinetics of remifentanil, which indicate that approximately 5 minutes is required to achieve a steady state plasma concentration following changes in infusion dosage.⁶ Respiratory depression was defined as a respiratory rate below 8/min, or if the SpO₂ decreased below 90%. If a score of 4 was attained, indicating that the child could no longer respond appropriately to verbal command or mild physical stimulation, the infusion dose was halved. Outcome measures included successful remifentanil infusion doses (defined as the dose required for sufficient analgesia regardless of respiratory status), times to discharge readiness⁸ (Table 2), and adverse events. These included hypoxemia (SpO₂ <90%), hypotension (systolic blood pressure <30% of baseline), bradycardia (heart rate <60/min), lack of adequate analgesia or anxiolysis (need for additional agents), or emesis. We defined a treatment failure as the need for additional sedative agents to complete the procedure comfortably when remifentanil administration resulted in respiratory depression despite continuing anxiety and/or lack of adequate analgesia.

Parametric data were compared using t tests and nonparametric data were compared using the Mann-Whitney rank sum test (Sigma-Stat for Windows; Jandel, San Rafael, Calif). Statistical significance was achieved at P<.05. Data are presented as mean±SD.

Seventeen children received remifentanil-midazolam for 20 procedures, which included bone marrow aspirate and/or biopsy (n=13), renal biopsy (n=4), and closed fracture reduction (n=3). There were 11 boys and 9 girls who weighed 26±13 kg and were aged 7.4±3.5 years. This technique was successful in 17 of the 20 procedures. The successful dose was 0.4±0.2 µg·kg⁻¹·min⁻¹. The time to reach discharge criteria for successful sedations was 9.5±4.3 minutes. The duration of all sedations was 17.4±6.8 minutes.

In the 17 successes, 4 children developed hypoxemia (SpO₂, 83%-89%) that rapidly corrected with verbal command or mild physical stimulation. Of the 13 successfully sedated children who did not develop hypoxemia, 10 developed apnea during the procedure and required verbal prompting to breathe. None of the children experienced hemodynamic instability or postprocedure emesis.

The remifentanil-midazolam technique was discontinued in 3 children. Their circumstances are described in Table 3. A 2-year-old who was the ninth child enrolled (failure 2 in Table 3) became unresponsive and hypoxemic despite continuing anxiety, and required positive-pressure ventilation by mask. This prompted a change in the age requirements for enrollment so that only children older than 4 years and able to respond to instructions to breathe deeply were subsequently included.

Remifentanil has been used successfully in combination with midazolam to provide sedation during breast biopsies in adults.⁹ In this study, successful infusion doses of remifentanil hydrochloride ranged from 0.05 to 0.11 µg·kg⁻¹·min⁻¹ when patients received anywhere from 0 to 8 mg of midazolam hydrochloride. At the higher doses of midazolam, these authors found an increased incidence of unresponsiveness and respiratory depression.

Remifentanil does not differ from other opioids with regard to its side effect profile.⁶ Opioids characteristically cause dose-dependent depression of the ventilatory response to hypercarbia and blunting of the ventilatory response to hypoxia.¹⁰ When administered alone, the ventilatory effects of benzodiazepines are variable.¹¹ However, when a benzodiazepine is combined with an opioid, the likelihood of respiratory depression is increased.^4,12 Bailey et al¹² studied the effect of combining midazolam and fentanyl on the slope of the ventilatory response to hypercarbia.¹² Although the addition of fentanyl to midazolam did not change the slope of the ventilatory response, the incidence of apnea was significantly increased. The high incidence of apnea in our patients, therefore, is not surprising. One can only speculate that if remifentanil were used alone, the incidence of respiratory depression would be less, but in children anxiolysis and amnesia are important components of a satisfactory sedation.

Other important side effects of opioids include emesis and difficult ventilation resulting from opioid-induced glottic spasm.¹³ Glottic spasm did not occur in any of our patients. It is possible that pretreatment with midazolam attenuated this effect or perhaps children respond differently than adults, but no data exist to substantiate these hypotheses. In our patients, the addition of ondansetron was effective in preventing narcotic-associated emesis in all patients.

An advantage of using remifentanil is that its effects are terminated rapidly when the infusion is discontinued, regardless of the duration of the infusion.⁶ Our use of remifentanil resulted in rapid recovery and discharge readiness. In an analysis of recovery from propofol administration,³ the mean time to responsiveness was 12 minutes, whereas when remifentanil is used, most children are ready for discharge home by that time. This can be a potential advantage to some families. For example, in our clinic, an 8-year-old child who receives bone marrow aspirations every 2 months continues to receive remifentanil at his and his parents' request because his rapid recovery and desire to remain conscious during the procedure are important priorities. We have not had similar requests from families whose children received remifentanil successfully but propofol subsequently. Although no studies in children have determined the optimal sedative agent, a propofol-based technique has become the preferred method because of its efficacy and safety at therapeutic doses, and the pleasant nausea-free hypnotic state that children and parents have come to expect during painful procedures. Although recovery is longer, a propofol-based technique may save time during the actual sedation because there is no delay waiting to achieve steady state levels as with remifentanil.

In summary, the major finding of this prospective, observational investigation is that the combination of midazolam and remifentanil was not a useful conscious sedation technique for painful procedures due to a high incidence of respiratory depression at subtherapeutic doses and frequent need for additional sedatives to provide patient comfort. Because of the high frequency of apnea, we have since discontinued the use of the remifentanil and midazolam combination for conscious sedation, and all subsequent patients receive a propofol-based technique. Although times to discharge are generally shorter when using remifentanil, the use of propofol rarely results in apnea or hypoxemia at therapeutic doses. If one chooses to use remifentanil, patients should be selected based on their ability to comprehend that they will be frequently prompted to breathe throughout the procedure. Thus, this technique is generally contraindicated in children younger than 4 or 5 years, or in older children with cognitive impairment. Likewise, practitioners who administer this combination of sedatives should expect respiratory depression, be alert to its onset, and be skilled in airway maintenance techniques.

Accepted for publication February 24, 1999.

This work was supported in part by a grant from Glaxo Wellcome Inc, Research Triangle Park, NC.

Presented in part at the 1998 Annual Meeting of the Society for Pediatric Anesthesia, Scottsdale, Ariz, February 13, 1998.

I am grateful to Fritz Rodriquez, MD, for his expert technical assistance and to Denham S. Ward, MD, PhD, for his review of the manuscript and many helpful suggestions.

Editor's Note: As one of the reviewers wrote, this study provides "another milepost on the bumpy road leading to a safe and effective sedative regimen . . . in children." I guess we've miles to go before they sleep.—Catherine D. DeAngelis, MD

Corresponding author: Ronald S. Litman, DO, Department of Anesthesiology, Box 604, Strong Memorial Hospital, 601 Elmwood Ave, Rochester, NY 14642 (e-mail: Ronald_Litman@urmc.rochester.edu).