For more than a decade, oral midazolam has been the most commonly ordered anesthetic premedication for pediatric patients in the United States (US) (1). Before 1998, however, there was no commercially prepared oral formulation available on the US market. Before the release of Versed® syrup by Roche Laboratories (Nutley, NJ), oral preparations of midazolam consisted of a mixture of the IV solution and a palatable liquid diluent. A wide variety of additives have been used, with the specific choice being a matter of local practice and preference. This resulted in formulations that differed from practice to practice in chemical composition, active drug concentration, and pH. Since the introduction of the commercially prepared syrup, both it and mixtures based on the IV formulation remain in clinical use.
A recent publication suggests that Versed Syrup is effective as a sedative premedication at a dose of 0.25 mg/kg, a dose that has not been demonstrated to be effective in studies using extemporaneous preparations of the IV formulation (2). The purpose of this study was to directly compare an equivalent dose of Versed Syrup with a standardized mixture of the IV drug in Syrpalta® (Humco Lab, Texarkana, TX) with regard to achieved sedative effect and plasma midazolam concentration at three specified times after oral administration. On the basis of pilot data obtained in 11 nonrandomized patients, we hypothesized that use of the IV/Syrpalta mixture would result in more effective sedation with correspondingly higher plasma midazolam levels.
With IRB approval and informed parental consent, ASA status I and II pediatric patients 2–10 yr old, weighing <40 kg, and scheduled for surgical procedures with minimal anticipated blood loss were enrolled in this prospective, randomized, and double-blinded study. Exclusion criteria included 1) diagnoses that may influence drug absorption (e.g., gastroesophageal reflux treated with antacid or gastrokinetic medication, history of small-bowel resection, malabsorption syndromes) and 2) the use of substances known to either induce or inhibit the cytochrome P-450 3A4 system.
Patients were randomized to receive either the commercially available Versed Syrup (Group 1) or a mixture of the IV midazolam formulation (midazolam HCl injection; Baxter Healthcare Corp., Deerfield, IL) in Syrpalta (Group 2) at a dose of 0.5 mg/kg body weight. The IV/Syrpalta mixture was prepared to yield a final drug concentration of 2 mg/mL, equivalent to the commercial syrup. Syrpalta is a flavored syrup consisting of 83% sucrose, purified water, glycerin, 0.1% sodium benzoate, 0.001% benzalkonium chloride, 0.2% alcohol, and artificial coloring and flavors (pH 5). All patients were observed by one of the two principal investigators at the time of medication administration to verify the ingestion of the entire dose.
Exactly 15 and 30 min after drug administration, each patient was assessed for sedative effect by one of the principal investigators, who was blinded to group. Assessments were made by using the 20-point Observer’s Assessment of Alertness/Sedation Scale (OAA/S) (3) (Appendix 1). By design, equal numbers of patients in each group (12 and 13) were evaluated by each of the two investigators to minimize the effect of interobserver variability.
Immediately after the 30-min OAA/S assessment, all patients received an inhaled induction consisting of sevoflurane in nitrous oxide (N2O). After establishment of IV access, patients underwent endotracheal intubation facilitated by neuromuscular blockade. Exactly 45, 60, and 90 min after drug administration, venous blood was obtained for the purpose of determining plasma midazolam level. Levels were obtained by National Medical Services, Inc. (Willow Grove, PA) by using standard gas chromatography/mass spectrometry analysis. Total IV fluid administration was recorded for each patient at the time of the 90-min specimen acquisition.
Statistical analysis was performed with MINITAB statistical software, Release 12 (Minitab, Inc., State College, PA). Demographic data and plasma midazolam levels were analyzed with a two-tailed Student’s t-test. Results are reported as the mean ± sd. Sedation scores were compared by using the Mann-Whitney U-test. Results are reported as the median and range. The number of patients with a measurable sedative response (defined as an OAA/S score of ≤17) (4) in each group was compared by using Fisher’s exact test. A P value of <0.05 was considered statistically significant.
There were 50 patients in the final analysis: 25 in Group 1 and 25 in Group 2. The study groups were comparable with respect to age, weight, sex distribution, and intraoperative fluid administration (Table 1). No significant blood loss occurred in any surgical procedure.
Median OAA/S scores at 15 min after administration were not significantly different (P = 0.095) between groups, with a median value of 18 (range, 16–20; 95% confidence interval, 17–19) in Group 1 and 17 (range, 11–20; 95% confidence interval, 16–18) in Group 2. However, at 30 min, OAA/S scores were significantly different (P = 0.023), with a median score of 17 (range, 11–20; 95% confidence interval, 15–18) in Group 1 and 15 (range, 9–19; 95% confidence interval, 14–16) in Group 2. Twenty-three of 25 Group 2 patients and 16 of 25 Group 1 patients had a 30-min OAA/S score of ≤17 (P = 0.017).
At least 2 plasma midazolam levels were obtained in all 50 patients. Three of the 150 specimens (one 60-min specimen in Group 1 and two 90-min specimens in Group 2) were either lost or mishandled. One excessively high 45-min value in Group 2 was determined to be an outlier, as defined by the extreme studentized deviate statistic, and was thereby omitted from analysis. However, inclusion of this patient in the analysis would not significantly alter the results or overall conclusions of the study. This left 146 of the intended 150 samples available for statistical analysis. Plasma midazolam levels were significantly higher in Group 2 at all three acquisition times. At 45 min, the mean level in Group 1 was 43.4 ± 19.6 ng/mL, and in Group 2, it was 63.1 ± 23.9 ng/mL (P = 0.003). At 60 min, the levels were 30.8 ± 17.9 ng/mL and 45.8 ± 18.2 ng/mL in Groups 1 and 2, respectively (P = 0.005). At 90 min, the levels were 21.0 ± 8.9 ng/mL and 28.9 ± 12.6 ng/mL, respectively (P = 0.017). The results are displayed in Figure 1. The relationship between plasma midazolam level and sedative response (OAA/S score of ≤17) for all patients, regardless of treatment group, is displayed in Figure 2.
Generally, two different preparations of oral midazolam are in clinical use. One is a commercially prepared and marketed formulation (Versed Syrup). The other consists of the IV formulation combined with a variety of palatable liquids. We compared sedation scores and plasma midazolam levels in children 2–10 years old receiving Versed Syrup (Group 1) with those in children receiving an IV formulation mixed in Syrpalta (Group 2), a commercially available palatable diluent. In Group 2 patients, the median sedation score was significantly lower 30 minutes after oral administration, and the mean plasma midazolam level was significantly higher at each of 3 specified acquisition times (45, 60, and 90 minutes after drug administration). At 45 minutes after administration, a time occurring within the recommended premedication to induction interval (5), the plasma midazolam level was >45% higher in Group 2 patients. In addition, more Group 2 patients (92%) than Group 1 patients (64%) had a detectable sedative effect at 30 minutes, as defined by an OAA/S score of ≤17.
The safety of oral midazolam in pediatric patients at a dose up to 1.0 mg/kg has been reported by McMillan et al. (6). Specifically, they reported no episodes of apnea, bradypnea, or arterial oxygen desaturation at any time during the conduct of their investigation. No patient failed to respond to stimulation or was unarousable. Five of the patients in our study attained 30-minute OAA/S scores of ≤12. Four of the five were in Group 2, and one was in Group 1. Peak plasma levels in this group of five patients ranged from 71 to 100 ng/mL. In contrast to the findings of McMillan et al., one Group 2 patient (with an OAA/S score of 9) was unresponsive to either loud verbal command or mild prodding. Although it occurred in only a small minority of patients in our series (10%), there is the potential for deep sedation at an administered dose of 0.5 mg/kg.
The overall distribution of the sedation response and the corresponding plasma midazolam concentration (Fig. 2) is similar to that seen in adolescent patients (4). At less than approximately 70 ng/mL, the sedative response is highly variable. At more than 70 ng/mL, detectable sedation is the expected effect. These results suggest a pharmacokinetic and pharmacodynamic variability in this pediatric age group that was also characteristic of the adolescents in the previous report.
Structurally, benzodiazepines consist of a benzene ring fused to a seven-membered diazepine ring. The midazolam molecule is an equilibrium mixture of two structural forms, the relative amounts of which are pH dependent. The first form contains an open hydrolyzed diazepine ring that renders the molecule polar and water soluble. The second form contains a closed diazepine ring and is lipophilic and more readily bioavailable. At a pH of >4.5, the molecule exists almost entirely in the lipophilic closed-ring form. As the pH decreases, the relative proportion of the open-ring form increases, such that at a pH of <2.5, the open-ring form predominates. Versed Syrup is pH-adjusted between 2.8 and 3.6 to ensure water solubility. In contrast, the IV formulation in Syrpalta has a pH range of 4.5 to 5.0. Because of the pH of this mixture, the predominant form of the drug contacting the oral mucosa is the lipophilic and more readily absorbable form, probably permitting greater oral mucosal uptake. The bioavailability of sublingually administered midazolam is substantially greater than that given orogastrically (7). Because of extensive first-pass hepatic metabolism (8), only 40%–50% of an orogastric dose reaches the systemic circulation. Karl et al. (9) demonstrated that sublingual midazolam premedication in pediatric patients resulted in satisfactory sedation and cooperation with anesthetic induction. This effect was evident at a dose less than half the recommended oral dose and was effective despite imperfect compliance with the method of administration. Lammers et al. (10) demonstrated a more rapid onset of sedation in pediatric patients premedicated with midazolam when the drug was mixed with the antacid sodium citrate before oral administration. Thus, the lower sedation score and higher plasma levels in our Group 2 patients may reflect enhanced pH-dependent oral absorption of the IV/Syrpalta mixture, with more drug bypassing first-pass metabolism.
In a recent multicenter study examining efficacy and safety for pediatric product labeling of Versed Syrup, Cote et al. (2) offered evidence supporting a claim of efficacy for a 0.25 mg/kg dose of Versed Syrup for preoperative sedation/anxiolysis in the pediatric population. Using an abbreviated but similar assessment scale to the OAA/S scale, 45 (94%) of 48 patients in the multicenter study receiving a 0.25 mg/kg dose of Versed Syrup demonstrated satisfactory sedation at 30 minutes after administration. This is substantially more than the 64% of our patients exhibiting detectable sedation who received twice this dose of Versed Syrup. This 0.25 mg/kg dose was compared only with larger doses of the same product, without either active controls or placebo controls for comparison. Because the efficacy of this dose has not been convincingly demonstrated for preparations using the IV drug in study protocols that did incorporate controls, a claim is made for the superior efficacy and consistency of effect of Versed Syrup. Cote et al. (2) quite legitimately state that these various extemporaneous formulations often suffer from a lack of standardization and consistency in preparation, particularly with regard to pH and concentration. We standardized both by using a readily available and consistently formulated commercial syrup (Syrpalta) as the diluent. In a 60:40 mixture with the IV product, this yields a final drug concentration of 2 mg/mL, which is equivalent to the commercial preparation. In so doing, we have demonstrated that this specific and reproducible extemporaneous preparation produces more reliable sedation and correspondingly higher plasma midazolam levels than an equivalent dose of Versed Syrup. Our findings also present evidence for a consistent pharmacokinetic handling of the two preparations. The coefficient of variation for achieved blood levels was no greater with our preparation than for levels obtained with Versed Syrup, and the decline in blood levels over time directly paralleled that of the commercial preparation. Although it is reasonable to assume a cost reduction associated with the use of generic preparations, at the time of this publication the average wholesale price per milligram of midazolam was no different between the two products used in this study (approximately US$0.63/mg). Actual costs to the hospital pharmacy are determined by negotiated contract and may differ from institution to institution.
1. Kain ZN, Mayes LC, Bell C, et al. Premedication in the United States: a status report. Anesth Analg 1997; 84: 427–32.
2. Cote CJ, Cohen IT, Suresh S, et al. A comparison of three doses of a commercially prepared oral midazolam syrup in children. Anesth Analg 2002; 94: 37–43.
3. Chernik DA, Ghouri D, Laine H, et al. Validity and reliability of the observer’s assessment of alertness/sedation scale: study with intravenous midazolam. J Clin Psychopharmacol 1990; 10: 244–51.
4. Brosius KK, Bannister CF. Oral midazolam premedication in preadolescents and adolescents. Anesth Analg 2002; 94: 31–6.
5. Weldon BC, Watcha MF, White PF. Oral midazolam in children: effect of time and adjunctive therapy. Anesth Analg 1992; 75: 51–5.
6. McMillan CO, Spahr-Schopfer IA, Sikich N, et al. Premedication of children with oral midazolam. Can J Anaesth 1992; 39: 545–50.
7. Fujii J, Inotsume N, Nakano M. Relative bioavailability of midazolam following sublingual versus oral administration in healthy volunteers. J Pharmacobiodyn 1988; 11: 206–9.
8. Reves JG, Fragan RJ, Vinik HR, Greenblatt DJ. Midazolam: pharmacology and uses. Anesthesiology 1985; 62: 310–24.
9. Karl HW, Rosenberger JL, Larach MG, Ruffle JM. Transmucosal administration of midazolam for premedication of pediatric patients: comparison of the nasal and sublingual routes. Anesthesiology 1993; 78: 885–91.
© 2003 International Anesthesia Research Society
10. Lammers CR, Rosner JL, Crockett DE, et al. Oral midazolam with antacid may increase the speed of onset of sedation in children prior to general anaesthesia. Paediatr Anaesth 2002; 12: 26–8.