Secondary Logo

Journal Logo

Original Article

Oral premedication with fentanyl may be a safe and effective alternative to oral midazolam

Tamura, M.; Nakamura, K.*; Kitamura, R.*; Kitagawa, S.*; Mori, N.; Ueda, Y.

Author Information
European Journal of Anaesthesiology: June 2003 - Volume 20 - Issue 6 - p 482-486


Oral midazolam is commonly administered as a premedication to infants and small children. However, it has a bitter taste [1], so it is necessary to mix it with syrup or juice or to prepare lollipops with it [2]. Even then, the taste is unacceptable to some children. In contrast to midazolam, fentanyl does not have a bitter taste and can be taken orally by children without much difficulty. Oral transmucosal fentanyl citrate has been reported to be an effective and safe premedication for children [3–6], but it is not available in Japan. We therefore administered fentanyl prepared for injection as an oral premedication and tested its efficacy and safety in paediatric anaesthesia compared with oral midazolam.


The protocol was approved by the Ethics Committee of Kyoto City Hospital. Fifty-one children, ASA I, aged 12–107 months, and weighing 10–25 kg, who were scheduled for general anaesthesia between January 1999 and May 2001 in Kyoto City Hospital, were included in the study. The operations performed were for inguinal hernia, strabismus, hydrocoele and undescended testis, and adenotonsillectomy. Patients were randomly assigned (random number table) to fentanyl or midazolam treatment groups in an open-label design. We did not attempt a double-blind design since the two drugs have such different tastes that the observer cannot be adequately blinded to the drug.

Lidocaine tape was affixed to the back of the hand or wrist intended for venepuncture 105 min before transfer to the holding room. Midazolam (5 mg, 1 mL) with syrup (1–2 mL) or fentanyl (0.1 mg, 2mL) with or without syrup (0.5–1 mL) was given orally 30 min before the transfer. The dose of syrup was adjusted according to preoperative information regarding the preference of the patient. The drugs were prepared in small bottles with small orifices originally designed for eye-drops, which the child sucked. If the child was unco-operative, the contents were poured onto the child's buccal mucosa. In a preliminary study, we measured the volume ejected by each push of this bottle containing fentanyl (2 mL) and syrup (0.5 mL) when it was held upside down. The ejected volume averaged 0.84 ± 0.12 mL (n = 17).

Each patient came to the holding room with their parent(s), and an intravenous catheter was placed where the lidocaine tape had been affixed. Patients for whom venepuncture failed at that place were excluded from further study. After intravenous access had been established, incremental doses of thiamylal were given intravenously until the patient could be separated from the parent calmly and transferred to the operating room. General anaesthesia was induced by administration of further intravenous thiamylal and administration of nitrous oxide and sevoflurane by inhalation. Muscle relaxation was induced by intravenous vecuronium and antagonized by neostigmine with atropine if necessary at the end of operation. An acetaminophen (paracetamol) suppository (100–200 mg) was given when the patient complained of pain or in younger children when the anaesthetist believed it necessary.

The excitation-sedation state of the patients was assessed by one of us (M. T., S. K. or R. K.) according to the scores proposed by Hanaoka and colleagues [7] (Table 1; our translation). The scores were recorded preoperatively: (1) upon the patient's arrival at the holding room (T1), (2) during placement of the intravenous catheter (T2), (3) on separation from the parent (T3), and (4) in the operating theatre during the placement of the leads for electrocardiography and pulse oximetry (T4). The postoperative scores were recorded at endotracheal extubation (T5), on transfer to the recovery room (T6) and on transfer from the recovery room to the ward (T7).

Table 1
Table 1:
Excitation-sedation scores proposed by Hanaoka and colleagues [7].

Signs of airway obstruction, bradypnoea (<10 breaths min−1) and hypoxaemia (apparent cyanosis or transcutaneous oxygen saturation <92% during room air breathing) were regarded as respiratory depression. The incidence of postoperative excitement (score 6 in Table 1), somnolence (scores 1 and 2) and vomiting were recorded after transfer to the ward until 6 h after operation. For those patients in whom discharge as a day case was planned, the incidence of the requirement for postponed discharge time was recorded.

Data were analysed by a Spearman's rank correlation test, U-test, Fisher's extract probability test or t-test for unpaired data as indicated. Comparisons with P < 0.05 were considered as significantly different.


Age, body weight, operation times and the duration of anaesthesia did not differ significantly between the midazolam and fentanyl treatment groups (Table 2). The mean dose of thiamylal required was 49 ± 22 mg in the fentanyl group and 34 ± 25 mg in the midazolam group (P < 0.05). Acetaminophen suppositories were given in the recovery room to eight of 24 patients in the midazolam group and to one of 27 patients in the fentanyl group. Excitationsedation scores at any time point did not differ significantly with age (Fig. 1a) or body weight (Fig. 1b), although the dose of midazolam or fentanyl was fixed regardless of differences in bodyweight. The scores did not differ between the fentanyl group and the midazolam group (Fig. 2). Respiratory depression was not observed in either group before induction of anaesthesia, after extubation of the trachea or after transferral to the ward.

Table 2
Table 2:
Patient characteristics.

No patient vomited preoperatively. During the first 6 h postoperatively, vomiting occurred in five patients (18.5%) in the fentanyl group, and in none in the midazolam group. The occurrence of postoperative excitement and somnolence did not differ between the fentanyl and midazolam groups (Table 3).

Table 3
Table 3:
Postoperative complications.
Figure 1.
Figure 1.:
Excitation-sedation scores. (a) Age; (b) body weight. Upper panels (a) ○: T1; □: T2; Δ: T3; ⋄: T4; lower panels (b) ○: T5; □: T6; Δ: T7. At all time points, the scores had no significant relationship to either age or body weight with fentanyl or midazolam.
Figure 2.
Figure 2.:
Scores of children receiving oral premedication. The scores were assessed on arrival at the holding room (T1), during placement of the intravenous catheter (T2), on separation from the parent (T3), during placement of the leads for electrocardiography and pulse oximetry (T4), at tracheal extubation (T5), on transfer to the recovery room (T6), and on transfer to the ward (T7). The box plots show the 25th and 75th percentiles as the tops and bottoms of boxes, and the 10th and 90th percentiles as the whiskers. □: Midazolam; ▪: fentanyl.

Nineteen patients (10 in the fentanyl group and nine in the midazolam group) had been planned for discharge on the day of surgery (6–8 h postoperatively), and all of them did so without alteration.


We administered fentanyl (0.1mg; 4–10 μg kg−1) orally as a premedication to children before minor operations and observed no difference in excitationsedation state compared with children who were premedicated with oral midazolam (5mg; 0.2–0.5 mg kg−1). However, postoperative vomiting was observed in 18.5% children treated with fentanyl.

We did not have a placebo group but used oral midazolam as a control since the present protocol of anaesthetic induction is difficult to perform in nonpremedicated children [8]. Because of the lack of placebo study, one might suppose that both fentanyl and midazolam had no sedative effect in our patients. However, many investigators have already confirmed the sedative effect of oral midazolam [8–14]. The optimal dose of oral midazolam is thought to be 0.5–0.75 mg kg−1 [8–13], but can be less in older children [14]. Therefore, we considered our 5-mg dose of midazolam for all children (10–25 kg body weight) to be within the acceptable range for use as a control for oral fentanyl, and to have exerted a sedative effect in the present study.

We used the same dose of midazolam or fentanyl for all children regardless of body weight, but the preoperative excitation-sedation scores did not correlate with age or body weight in either group. This suggests that a larger dose per body weight is required for preoperative sedation in younger children than in older ones.

The efficacy and safety of different doses of oral transmucosal fentanyl citrate have been studied by several investigators [6, 15–18]. Their results suggest that oral transmucosal fentanyl citrate at doses between 5 and 25 μg kg−1 body weight causes dosedependent sedation of children, while respiratory suppression is common with doses >20 μg kg−1 body weight. In one report [17], postoperative vomiting occurred in 20% of patients treated with 10–15 μg kg−1 body weight fentanyl in oral transmucosal fentanyl citrate and in 47% patients treated with 15–20 μg kg−1. Streis and colleagues [19] demonstrated in healthy volunteers that the peak plasma concentration of fentanyl when swallowed with water (10 mL) was approximately half that of oral transmucosal fentanyl citrate. Moreover, peak plasma concentrations were reached more slowly with fentanyl swallowed (101 min) than with oral transmucosal fentanyl citrate (22 min). In contrast to their study, we administered fentanyl (2 mL with or without syrup up to 1 mL; total volume <3 mL) without the addition of water, from a bottle with a small orifice. In a preliminary test, <1 mL liquid was ejected from this bottle with each push of the bottle. Therefore, we believe that in most of our patients the drug was absorbed through the oral mucosa, not swallowed. This is supported by the fact that although the dose of fentanyl administered was <10μgkg−1, it induced sedation similar to that induced with oral midazolam (0.2–0.5 mg kg−1) only 30 min after administration, and the complication rate of postoperative vomiting was as high as 18.5%.

The doses of preoperative thiamylal and postoperative acetaminophen needed were different in the two groups. This is probably not due to quantitative differences in sedation, but to qualitative differences in the actions of the narcotics and benzodiazepine derivatives. It is well known that benzodiazepine derivatives, including midazolam, and barbiturates act synergistically [20,21]. Midazolam does not have the analgesic action of fentanyl, so acetaminophen was more commonly required in the midazolam treatment group than in the fentanyl group.

We conclude that oral administration of fentanyl without water 30 min before transfer to the holding room before an operation from a bottle with a small orifice is a premedication option for children between 1 and 8 yr of age, although the rate of postoperative vomiting is unfortunately rather high.


1. Anderson BJ, Exarchos H, Lee K, Brown TC. Oral premedication in children: a comparison of chloral hydrate, diazepam, alprazolam, midazolam and placebo for day surgery. Anaesth Intensive Care 1990; 18: 185-193.
2. Shigemi K, Kanbayashi Y, Ohta T, et al. Midazolamatropine lollipop for pediatric premedication. Masui 2000; 49: 496-503.
3. Friesen RH, Carpenter E, Madigan CK, Lockhart CH. Oral transmucosal fentanyl citrate for preanaesthetic medication of paediatric cardiac surgery patients. Paediatr Anaesth 1995; 5: 29-33.
4. Epstein RH, Mendel HG, Witkowski TA, et al. The safety and efficacy of oral transmucosal fentanyl citrate for preoperative sedation in young children. Anesth Analg 1996; 83: 1200-1205.
5. Dsida RM, Wheeler M, Birmingham PK, et al. Premedication of pediatric tonsillectomy patients with oral transmucosal fentanyl citrate. Anesth Analg 1998; 86: 66-70.
6. Esuvaranathan VV, Mukherjee K, Streets C, et al. A comparison of oral transmucosal fentanyl and oral midazolam for premedication in children. Paediatr Anaesth 2000; 10: 697.
7. Hanaoka K, Tachibana N, Tanifuji Y. Application of midazolam for sedation during local anesthesia. Igaku-to-Yakugaku 1985; 14: 840-846.
8. Gillerman RG, Hinkle AJ, Green HM, Cornell L, Dodge CP. Parental presence plus oral midazolam decreases frequency of 5% halothane inductions in children. J Clin Anesth 1996; 8: 480-485.
9. Lyons B, Cregg N, Conway F, Casey W, Doherty P, Moore KP. Premedication for ambulatory surgery in preschool children: a comparison of oral midazolam and rectal thiopentone. Can J Anaesth 1995; 42: 473-478.
10. Cray SH, Dixon JL, Heard CM, Selsby DS. Oral midazolam premedication for paediatric day case patients. Paediatr Anaesth 1996; 6: 265-270.
11. Mitchell V, Grange C, Black A, Train J. A comparison of midazolam with trimeprazine as an oral premedicant for children. Anaesthesia 1997; 52: 416-421.
12. Riva J, Lejbusiewicz G, Papa M, et al. Oral premedication with midazolam in paediatric anaesthesia. Effects on sedation and gastric contents. Paediatr Anaesth 1997; 7: 191-196.
13. Patel D, Meakin G. Oral midazolam compared with diazepam-droperidol and trimeprazine as premedicants in children. Paediatr Anaesth 1997; 7: 287-293.
14. Pywell CA, Hung YJ, Nagelhout J. Oral midazolam versus meperidine, atropine, and diazepam: a comparison of premedicants in pediatric outpatients. AANA J 1995; 63: 124-130.
15. Streisand JB, Stanley TH, Hague B, van Vreeswijk H, Ho GH, Pace NL. Oral transmucosal fentanyl citrate premedication in children. Anesth Analg 1989; 69: 28-34.
16. Ashburn MA, Streisand JB, Tarver SD, et al. Oral transmucosal fentanyl citrate for premedication in paediatric outpatients. Can J Anaesth 1990; 37: 857-866.
17. Schutzman SA, Burg J, Liebelt E, et al. Oral transmucosal fentanyl citrate for premedication of children undergoing laceration repair. Ann Emerg Med 1994; 24: 1059-1064.
18. Egan TD, Sharma A, Ashburn MA, et al. Multiple dose pharmacokinetics of oral transmucosal fentanyl citrate in healthy volunteers. Anesthesiology 2000; 92: 665-673.
19. Streisand JB, Varvel JR, Stanski DR, et al. Absorption and bioavailability of oral transmucosal fentanyl citrate. Anesthesiology 1991; 75: 223-229.
20. DeLorey TM, Kissin I, Brown P, Brown GB. Barbituratebenzodiazepine interactions at the gamma-aminobutyric acid A receptor in rat cerebral cortical synaptoneurosomes. Anesth Analg 1993; 77: 598-605.
21. Wesselman JP, van Wilgenburg H, Long SK. The effects of pentobarbital and benzodiazepines on GABA-responses in the periphery and spinal cord in vitro. Neurosci Lett 1991; 128: 261-264.


© 2003 European Society of Anaesthesiology