The Effect of Small Dose Fentanyl on the Emergence Characteristics of Pediatric Patients After Sevoflurane Anesthesia Without Surgery

Cravero, Joseph P. MD; Beach, Michael MD; Thyr, Brian MD; Whalen, Kate RN

Anesthesia & Analgesia:
doi: 10.1213/01.ANE.0000070227.78670.43

We designed this study to measure the effect of a small dose of IV fentanyl on the emergence characteristics of pediatric patients undergoing sevoflurane anesthesia without any surgical intervention. Thirty-two ASA physical status I or II pediatric outpatients receiving sevoflurane anesthesia for magnetic resonance imaging scans were enrolled and assigned in a random and double-blinded manner to receive either placebo (saline) or 1 μg/kg IV fentanyl 10 min before discontinuation of their anesthetic. The primary outcome measure was the percentage of patients with emergence agitation. We also evaluated the duration of agitation and time to meet hospital discharge criteria. Patients who received fentanyl had a decreased incidence of agitation (12% versus 56%) when compared with placebo. There was no significant difference in time to meet hospital discharge criteria. We conclude that the addition of a small dose of fentanyl to inhaled sevoflurane anesthesia decreases the incidence of emergence agitation independent of pain control effects.

In Brief

IMPLICATIONS: The addition of a small dose of fentanyl given to patients undergoing nonsurgical sevoflurane anesthesia resulted in a significant decrease in emergence agitation in a prospective, randomized, and controlled trial involving pediatric patients.

Author Information

Department of Anesthesiology, Dartmouth-Hitchcock Medical Center, Lebanon, New Hampshire

Accepted for publication March 20, 2003.

Address correspondence and reprint requests to Joseph P. Cravero, MD, Department of Anesthesiology, One Medical Center Dr., Lebanon, NH 03756. Address e-mail to

Article Outline

Emergence agitation (EA) phenomena (also called emergence delirium) in children have been the subject of a number of investigations. Several studies have documented an increase in the incidence of EA when drugs with low solubility are used for anesthesia in children (1–17). Although this problem has not been associated with significant morbidity, marked EA can negate any advantage rapid emergence from general anesthesia offers. Sevoflurane in particular has been associated with an increased amount of agitation on emergence from anesthesia in children when compared with a more soluble anesthetic (halothane) (1–3,5,6,8,10,12–16) even in the absence of any surgical intervention (15).

This study was performed to evaluate whether a small dose of IV fentanyl would improve the emergence characteristics of a group of patients receiving sevoflurane anesthesia and no surgery—thus eliminating the possibility that the primary effect of fentanyl on emergence behavior is strictly related to its pain control. We hypothesized that the addition of fentanyl to the anesthetic technique would decrease agitation independent of any effect on postsurgical pain.

Back to Top | Article Outline


After IRB approval, informed consent was obtained for 32 ASA physical status I or II patients aged 18 mo to 10 yr scheduled to undergo magnetic resonance imaging (MRI) scanning. Any patient with a defined psychological/emotional disorder or cognitive delay was excluded from the study. Any neurological condition that would limit a patient’s ability to communicate with, or understand nursing personnel, was also grounds for exclusion. For preparation, patients and their families viewed a video of the induction process before the day of the study. In addition, parents were counseled on behaviors to expect in their child on induction of anesthesia and they were allowed to be present for induction. Preinduction sedation was left to the discretion of the attending anesthesiologist as is the routine practice at our institution for patients undergoing MRI scans. (Unpublished data from our institution reveal a 10% rate of sedation in this setting.) If the attending anesthesiologist deemed the patient extremely anxious, midazolam was administered orally. For the purposes of this study, any patient who received midazolam was considered ineligible.

If sedation was not required, patients were randomly assigned to receive either placebo (Group P) or fentanyl (Group F) using a computer-generated random number assignment. Patients in Group F received 1 μg/kg fentanyl IV 10 min before the end of the anesthetic. Patients in Group P received an equal volume of saline IV 10 min before the end of the anesthetic. The MRI technicians were asked to evaluate the remaining time in each scan and to inform the anesthesia team when the total time left in a scan (including remaining sequences and interval time) would be 10 min. There were no instances in the study group in which the scan had to be repeated or when the anesthesia was not terminated as expected given the estimate from the MRI technician. Anesthesia providers, MRI technicians, patients, recovery nurses, and the emergence grader (KW) were all blinded to group assignment.

Anesthesia was provided by one of six anesthesiologists. A circle breathing circuit was used in all cases. The induction of anesthesia was initiated with fresh gas flows of nitrous oxide (5 L/min) and oxygen (2 L/min) by mask. Sevoflurane was then added to the circuit at maximal vaporizer output (8%) after 1 min. Upon successful induction of general anesthesia, the flow rates were decreased to 2 L/min oxygen and nitrous oxide was discontinued—this was maintained for the balance of the procedure. All patients had an IV catheter placed after the induction of anesthesia. A laryngeal mask airway (LMA) of appropriate size for the age and weight of the child was placed after the IV was secured. Spontaneous respirations were maintained. Ventilation was assisted if the end-tidal carbon dioxide increased to ≥55 cm H2O and inhaled anesthetic was decreased by 10%. The maintenance anesthetic was delivered at a concentration that maintained a stable heart rate, blood pressure, and breathing rate (baseline ± 20%). At the conclusion of the procedure, the inhaled anesthetic was discontinued, and the fresh gas flow was increased to 10 L/min oxygen. The LMA was removed while the patient was still nonresponsive to stimulation as is routine practice in our department.

Noninvasive blood pressure measurements, electrocardiogram, heart rate, hemoglobin oxygen saturation, and end-tidal carbon dioxide were continuously monitored and recorded every 5 min from the time of induction until leaving the MRI scanner. Upon completion of the MRI scan and removal of the LMA, patients were taken directly to the postanesthesia care unit (PACU). Parents were allowed to be at the child’s bedside immediately upon admission (while the child was still asleep) to the PACU as is routine at our institution.

One trained observer, blinded to patient group assignment, was present at the end of the anesthetic and upon admission to the PACU. The level of agitation was recorded continuously beginning at the discontinuation of anesthesia time using a simple EA scale. This scale rates agitation from 1 to 5—with 1 representing the obtunded patient with no response to stimulation, 2 designates asleep but responsive to movement or stimulation, 3 is awake and appropriately responsive, 4 crying and difficult to console, and 5 describes wild thrashing behavior that requires restraint. This scoring system is very similar or identical to scales used in most pediatric emergence studies, and is the same as that used by the authors in previously published investigations of this phenomenon (15,16). The emergence observer recorded the time (in continuous minutes) at which any change in the agitation level occurred. Agitation levels were recorded until the patient was awake, alert, calm, and responsive to his/her parents or guardians.

In the PACU, the IV catheter was removed as soon as the child was responsive. Observation/recording ended when the child was awake, calm, and cooperative (grading level 3). No changes were made in the care of the study patients in the PACU or the secondary recovery unit (SRU). Bedside nurses determined the time when each child met standard criteria for discharge from the PACU. Patients were then taken to the SRU. There they were observed until the time at which they met all discharge criteria for the SRU. This time (which is the same as time to meet hospital discharge criteria) was recorded. All times were measured using hospital clocks which are synchronized.

For the purposes of this study, EA was defined as an EA score of ≥4 for ≥5 min duration despite all calming efforts by the child’s parents/guardians and nursing personnel.

Data were also collected concerning the duration of the anesthesia and any itching or vomiting that might have occurred in the postanesthesia time period.

Fisher’s exact or Wilcoxon’s ranked sum test was used as appropriate for univariate comparisons. Logistic regression with adjustment for over-dispersion was used to model the presence of EA on an inhaled anesthetic adjusting for patient covariates. A P value of 0.05 was used for statistical significance. Assuming an incidence of EA of ≥55% in the placebo group [based on previously published data from our institution (15)] and 10% in the fentanyl group (based on a small pilot sample), 16 patients per group provided a power of 0.8 using an unadjusted χ2 test.

Back to Top | Article Outline


The reason for performing an MRI was recorded for each patient. Ten patients in Group P had MRIs of the head for evaluation of a first-time seizure or cranial abnormality compared with eight similar patients in Group F. Four patients in each group were having MRIs of the spine. Three patients in each group had extremity MRIs for evaluation of deformity or mass lesion. All participants were carefully evaluated to assure no evidence for neurological issues or developmental delays according to parental report and a review of the medical record. None of the patients were on medication for seizures or attention deficit difficulties. Of the patients who were scanned for “seizure,” all had experienced only one seizure and none were thought to have any sequelae as determined by the treating neurologist. In the course of performing this study, 33 patients were considered but not offered entry into the study because of their neurological status. In addition, three patients were not evaluated because the treating anesthesiologist thought that the child’s emotional status demanded pre-procedural sedation.

The results of the comparisons among the patients are outlined in Table 1. There was no statistically significant difference between the groups in terms of their age, weight, sex, or duration of anesthesia. There was a trend toward longer cases in Group P.

Analysis of the data generated by evaluating the nature of the recovery period in our patients revealed that there was a much larger percentage of patients who had agitation in Group P relative to Group F (56% versus 12%, P = 0.02;Table 2). There was no significant difference in the time to reach hospital discharge criteria as judged by the SRU nurses who were unaware of patient group assignment.

None of the children had significant itching or vomiting as noted by the blinded observer or the recovery nurses who were caring for these children.

Adjustment for age, weight, or sex using logistic regression models did not change the statistical significance or the magnitude of the effect of fentanyl on EA. Although increasing age and weight were associated with decreased odds of agitation, the result was not statistically significant. Boys had increased odds of becoming agitated, but this was also not statistically significant (odds ratio: 2.6, P = 0.29).

Back to Top | Article Outline


Many investigators have described the incidence and characteristics of EA in pediatric patients after sevoflurane anesthesia (1–3,5,6,8,10,12–16). In addition, several studies have appeared that evaluated methods for improving the quality of sevoflurane emergence. Davis et al. (18) investigated the incidence of EA after bilateral myringotomy and tube placement. The incidence of excitement and agitation was less in patients receiving halothane or sevoflurane when ketorolac was given IV after the induction of anesthesia (14% versus 38%, P < 0.05 for sevoflurane patients). The authors argued that EA could be curtailed with adequate pain management. In a similar study, Johannesson et al. (2) found that acetaminophen given after the induction of anesthesia decreased agitation in a group of patients receiving sevoflurane anesthesia.

Clonidine has also been shown to decrease agitation on emergence from anesthesia. Kulka et al. (19) documented a significant decrease in agitation (10% versus 72%, P < 0.001) in a clonidine-treated group undergoing circumcision. Likewise, oral midazolam premedication was shown by Lapin et al. (1) to decrease agitation after ear tube insertion (39% versus 67%).

With respect to the use of opioids and their effect on agitation, Galinkin et al. (20) found that the administration of 2 μg/kg intranasal fentanyl after induction with sevoflurane resulted in therapeutic serum levels of fentanyl and decreased agitation after ear tube placement. Murray et al. (21) evaluated the effect of 0.1 mg/kg oxycodone elixir (premedication) on emergence in children undergoing otolaryngology surgery using halothane or sevoflurane anesthesia. EA was decreased in patients who received halothane anesthesia and oxycodone premedication (15% versus 45%) but not for patients receiving sevoflurane. Finally, it is notable that the first study to describe postoperative agitation (22) found that patients who received an opioid-based anesthetic had a less frequent incidence of postoperative disturbance (wild thrashing) behavior when compared with those who received an insoluble drug—cyclopropane (0.4% versus 8%).

All of the above studies are limited by the difficulty in separating “excitement” behaviors caused by specific anesthetics from “pain” behaviors related to the procedure itself. Various investigators have attempted to separate these two entities. Beskow and Westrin (14) used a combination of behavioral scales whereas Aono et al. (10) used caudal blocks in an effort to differentiate these problems. We believe that EA behaviors specific to anesthetic technique (and treatment) can be more precisely investigated in the absence of surgical pain. In this study, our only interventions involved the placement of an LMA and a peripheral IV catheter after anesthesia was induced. Although we recognize that studies in adults have reported a small incidence of oral pain after LMA use, there are no such reports in children and we have not noted this complaint in our practice. Similarly, IV catheters were removed immediately upon awakening so we do not believe the annoyance of an IV catheter added significantly to agitation.

Our study involved a relatively small number of patients. This is attributed to strict inclusion criteria. In addition, three patients were not included because the attending anesthesiologist believed that the preanesthetic anxiety warranted sedation before the induction of anesthesia. We believe there is clear evidence that midazolam premedication alters emergence behavior (1,23) and should be avoided in a study such as ours. However, despite our small numbers, the large difference between our groups allowed us to find significant differences between the treatment arms.

Like all investigations of EA, our study is limited by the lack of a widely used and validated tool for measuring agitation. With this limitation in mind, we used a simple graded instrument that we have used in previous investigations and is very similar to those used by other investigators. All scoring was done by one blinded observer to eliminate issues of inter-rater variability.

We believe that the time required to meet standard discharge criteria is a more clinically relevant measure than other alertness scoring systems for the purposes of this comparison. Because our nurses were blinded to treatment, we believe the bias in our system was equally distributed. Our data revealed no significant difference in the time required to reach hospital discharge criteria when fentanyl was given.

In summary, we found that the incidence and duration of EA in patients receiving sevoflurane without surgery was significantly decreased by the addition of 1 μg/kg fentanyl 10 minutes before the end of anesthesia. We demonstrated that time to reach hospital discharge criteria was unchanged by the addition of this small dose of fentanyl. The addition of a small dose of fentanyl to an anesthetic using sevoflurane should be considered, even when expected postoperative pain is minimal, to decrease EA.

Back to Top | Article Outline


1. Lapin SL, Auden SM, Goldsmith LJ, Reynolds AM. Effects of sevoflurane anaesthesia on recovery in children: a comparison with halothane. Paediatr Anaesth 1999; 9: 299–304.
2. Johannesson GP, Floren M, Lindahl SG. Sevoflurane for ENT-surgery in children: a comparison with halothane. Acta Anaesthesiol Scand 1995; 39: 546–50.
3. Sarner JB, Levine M, Davis PJ, et al. Clinical characteristics of sevoflurane in children: a comparison with halothane. Anesthesiology 1995; 82: 38–46.
4. Lerman J, Sikich N, Kleinman S, Yentis S. The pharmacology of sevoflurane in infants and children. Anesthesiology 1994; 80: 814–24.
5. Naito Y, Tamai S, Shingu K, et al. Comparison between sevoflurane and halothane for paediatric ambulatory anaesthesia. Br J Anaesth 1991; 67: 387–9.
6. Greenspur JC, Hannallah RS, Wellborn LG, Norden JM. Comparison of sevoflurane and halothane anesthesia in children undergoing outpatient ear, nose, and throat surgery. J Clin Anesth 1995; 7: 398–402.
7. Patel SS, Goa KL. Desflurane: a review of its pharmacodynamic and pharmacokinetic properties and its efficacy in general anaesthesia. Drugs 1995; 50: 742–67.
8. Epstein RH, Mendel HG, Guarnieri KM, et al. Sevoflurane versus halothane for general anesthesia in pediatric patients: a comparative study of vital signs, induction, and emergence. J Clin Anesth 1995; 7: 237–44.
9. Lerman J, Davis PJ, Welborn LG, et al. Induction, recovery, and safety characteristics of sevoflurane in children undergoing ambulatory surgery. Anesthesiology 1996; 84: 1332–40.
10. Aono J, Ueda W, Mamiya K, et al. Greater incidence of delirium during recovery from sevoflurane anesthesia in preschool boys. Anesthesiology 1997; 87: 1298–300.
11. Nathanson MH, Freedman B, Smith I, White PF. Sevoflurane versus desflurane for outpatient anesthesia: a comparison of maintenance and recovery profiles. Anesth Analg 1995; 81: 1186–90.
12. Welborn LG, Hannallah RS, Norden JM, et al. Comparison of emergence and recovery characteristics of sevoflurane, desflurane, and halothane in pediatric ambulatory patients. Anesth Analg 1996; 83: 917–20.
13. Sury MR, Black A, Himinton L, et al. A comparison of the recovery characteristics of sevoflurane and halothane in children. Anaesthesia 1996; 51: 543–6.
14. Beskow A, Westrin P. Sevoflurane causes more postoperative agitation in children than does halothane. Acta Anaesthesiol Scand 1999; 43: 536–41.
15. Cravero J, Surgenor S, Whalen K. Emergence agitation in paediatric patients after sevoflurane anaesthesia and no surgery: a comparison with halothane. Paediatr Anaesth 2000; 10: 419–24.
16. Cravero JP, Beach M, Dodge CP, Whalen K. Emergence characteristics of sevoflurane compared to halothane in pediatric patients undergoing bilateral pressure equalization tube insertion. J Clin Anesth 2000; 12: 897–401.
17. van Vlymen JM, White PF. Outpatient anesthesia. In: Miller RD, ed. Anesthesia. New York: Churchill Livingstone, 2000: 2226–7.
18. Davis PJ, Greenberg JA, Gendelman M, Fertal K. Recovery characteristics of sevoflurane and halothane in preschool-aged children undergoing bilateral myringotomy and pressure equalization tube insertion. Anesth Analg 1999; 88: 34–8.
19. Kulka PJ, Bressem M, Tryba M. Clonidine prevents sevoflurane-induced agitation in children. Anesth Analg 2001; 93: 335–8.
20. Galinkin JL, Fazi LM, Cuy RM, et al. Use of intranasal fentanyl in children undergoing myringotomy and tube placement during halothane and sevoflurane anesthesia. Anesthesiology. 2000; 93: 1378–83.
21. Murray DJ, Cole JW, Shrock CD, et al. Sevoflurane versus halothane: effect of oxycodone premedication on emergence behaviour in children. Paediatr Anaesth 2002; 12: 308–12.
22. Eckenhoff JE, Kneale DH, Drpps RD. The incidence and etiology of postanesthetic excitement: a clinical survey. Anesthesiology 1961; 22: 667–73
23. Kain ZN, Mayes LC, Wang SM, et al. Postoperative behavioral outcomes in children: effects of sedative pre-medication. Anesthesiology 1999; 90: 758–65.
© 2003 International Anesthesia Research Society