`Intranasal Midazolam Formulation in Healthy Volunteers
`
`Daniel P. Wermeling, PharmD,
`FASHP*
`
`Kenneth A. Record, PharmD*
`
`Thomas H. Kelly, PhD†
`
`Sanford M. Archer, MD‡
`
`Thomas Clinch§
`
`Anita C. Rudy, PhD§
`
`We evaluated the pharmacokinetics and pharmacodynamics of single 5-mg doses
`of midazolam after administration of a novel intranasal (IN) formula, IM, and IV
`midazolam in an open-label, randomized, 3-way cross-over study in 12 healthy
`volunteers. IN doses were delivered as 0.1-mL unit-dose sprays of a novel
`formulation into both naris. Blood samples were taken serially from 0 to 12 h after
`each dose. Plasma midazolam concentrations were determined by liquid
`chromatography/mass
`spectrometry/mass
`spectrometry. Noncompartmental
`analysis was used to estimate pharmacokinetic parameters. The mean midazolam
`bioavailabilities and % coefficient of variation were 72.5 (12) and 93.4 (12) after the
`IN and IM doses, respectively. Median time to maximum concentration was 10 min
`for IN doses. Adverse events were minimal with all routes of administration, but
`nasopharyngeal irritation, eyes watering, and a bad taste were reported after IN
`doses. Our results support further development of this novel midazolam nasal
`spray.
`(Anesth Analg 2006;103:344 –9)
`
`Midazolam’s potency and short duration of clini-
`
`cal activity make it an excellent drug for premedica-
`tion, and its advantages have been reviewed (1).
`Although midazolam is marketed only in injectable
`and oral syrup formulations, it is the most extensively
`studied intranasal (IN) benzodiazepine. One survey
`reported that 8% of United States anesthesiologists
`have used IN midazolam, off-label, to premedicate
`pediatric patients preoperatively (2). The first clinical
`investigation of IN midazolam in children was re-
`ported by Wilton et al. (3). Another investigation
`achieved anxiolysis with IN midazolam with a mean
`onset of 15–20 min (4). Estimates of the bioavailability
`of midazolam following the IN route of administra-
`tion have ranged from 50%– 83% in human studies
`(5–11). Most studies have used a dilute aqueous
`midazolam injection solution that is not suitable for
`nasal administration because of low pH and subopti-
`mal concentration causing nasal run-off.
`The formulations used in the aforementioned trials
`
`From the *University of Kentucky College of Pharmacy; †Uni-
`versity of Kentucky College of Medicine; ‡Division of Otolaryngol-
`ogy – Head & Neck Surgery, University of Kentucky A. B. Chandler
`Medical Center; §Intranasal Therapeutics, Inc., Lexington, Ken-
`tucky.
`Accepted for publication April 17, 2006.
`Supported, in part, by Intranasal Technology, Inc., Lexington,
`Kentucky.
`Address correspondence and reprint requests to Daniel P. Wer-
`meling, Pharm.D., University of Kentucky College of Pharmacy, 725
`Rose Street, Lexington, Kentucky 40536. Address e-mail
`to
`dwermel@uky.edu.
`Copyright © 2006 International Anesthesia Research Society
`DOI: 10.1213/01.ane.0000226150.90317.16
`
`344
`
`are generally not appropriate for nasal administration.
`The formulations will be irritating because the aque-
`ous solution is buffered to pH 3. Moreover,
`the
`formulae are too dilute for nasal administration. The
`adult nasal cavity can only receive and retain about
`100 –150 L of liquid, requiring the dose of midazolam
`be solubilized within this volume. A nasal formula
`with a concentration of 2.5 mg per 100 L is necessary
`to give a single spray per naris. Thus, the formulation
`methods must be considered when evaluating any
`estimates of the pharmacokinetics and pharmacody-
`namics in these reports. Given these findings, there
`appears to be an unmet medical need to develop an
`optimal midazolam formulation for nasal delivery.
`The objectives of the study were to evaluate the
`bioavailability of a novel IN midazolam formulation
`and to compare the pharmacodynamic effects on
`psychomotor performance and subjective reporting of
`drug effect after 5 mg doses of midazolam via IN, IM,
`and IV routes of administration. The specific aims
`were to: 1) to obtain an IN bioavailability of more than
`or equal to 70% compared with an IV dose; 2) achieve
`maximum IN concentration within 10 min of admin-
`istration; 3) observe sedative properties from IN ad-
`ministration within 10 min; and, 4) demonstrate a
`nonirritating, well-tolerated formula.
`Absolute bioavailability of alternative drug deliv-
`ery routes is frequently compared with an IV formula
`for a reference. Area under the concentration-time
`curve for each route is calculated. Bioavailability is
`simply the ratio of the IN to IV area under the curve,
`assuming the same dose was administered and clear-
`ance was constant. In this type of study it is generally
`AQUESTIVE EXHIBIT 1130 Page 0001
`
`Vol. 103, No. 2, August 2006
`
`
`
`assumed that clearance remains constant within each
`subject across test arm investigations.
`
`METHODS
`Twelve nonsmoking, healthy subjects (6 men, 6
`women) between the ages of 20 and 29 yr (22.3 ⫾ 2.8
`yr, mean ⫾ sd) and weighing 60 to 92 kg (71.1 ⫾ 10.4
`kg) participated in this inpatient study. All subjects
`gave written informed consent for the study which
`was approved by the Medical IRB of the University of
`Kentucky. Eleven of the volunteers who enrolled in
`the study were Caucasian and one was part-Asian.
`Subjects were within 20% of ideal body weight in
`relation to height and elbow breadth and weighed at
`least 60 kg. The subjects were in good health and had
`no clinically significant previous nasal surgery or
`polyps or other physical abnormalities of the nose or
`any systemic medical illness. They abstained from
`alcohol and caffeine-containing beverages and pre-
`scription and nonprescription drugs that might inter-
`act with midazolam metabolism or nasal physiology
`48 h before the dosing period and during the study.
`Subjects were admitted to the Clinical Research Center
`at University of Kentucky on the evening before each
`study day. The subjects fasted for 8 h before receiving
`the study drug and continued to fast for 2 h after drug
`administration.
`A randomized, open-label, 3-way cross-over design
`was used. On three different occasions, separated by 1
`wk, the subjects received a single dose of each of the
`following three treatments in random order, counter-
`balanced so that an equal number of subjects received
`each treatment first, second, or third:
`• Treatment A: 5 mg IV of midazolam infused over
`15 min
`• Treatment B: 5 mg of IM midazolam
`• Treatment C: 5 mg IN midazolam solution (2.5
`mg/100 L per sprayer/naris)
`
`The 25 mg/mL IN midazolam formulation was
`prepared aseptically, creating a sterile product, under
`Good Manufacturing Practices conditions in the Uni-
`versity of Kentucky College of Pharmacy Center for
`Pharmaceutical Science and Technology. The IN for-
`mulation, a nonaqueous solution containing midazo-
`lam 25 mg/mL, polyethylene glycol 400, butylated
`hydroxytoluene, saccharin, and propylene glycol, pro-
`vided 2.5 mg of midazolam in 0.1 mL spray from a
`modified version of a commercially available unit-
`dose spray pump (Pfeiffer of America, Princeton, NJ,
`unit dose system). Commercially available midazolam
`(Versed® Injection; Roche Laboratories, Nutley, NJ)
`was purchased for comparative IV and IM adminis-
`tration.
`Before study drug administration, subjects gently
`blew their noses. A physician administered a spray to
`each naris and the subjects remained in a semi-
`recumbent position, with the head of the bed elevated
`at a 20°– 40° angle for 30 min. The IV dose, 5 mg
`
`Vol. 103, No. 2, August 2006
`
`midazolam in 10 mL sterile saline solution, was ad-
`ministered by infusion over a period of 15 min in an
`antecubital vein of the contralateral arm for blood
`sampling.
`Serial blood samples were obtained through an
`indwelling venous catheter according to the following
`schedule: 0 (pre-dose), 5, 10, 20, 30, and 45 min, and 1,
`1.5, 2, 3, 4, 8, and 12 h after drug administration.
`Venous blood samples were collected in 10-mL hepa-
`rinized Vacutainer® tubes. After collection, the blood
`was centrifuged at 4°C, and the plasma was trans-
`ferred to polypropylene tubes. The plasma was stored
`at or below ⫺20°C at the study site until shipped to
`AAI Development Services, Inc., Kansas City Facility
`in Shawnee, KS for midazolam assay.
`Plasma samples were analyzed by AAI International,
`a Good Laboratory Practices compliant laboratory, using
`liquid chromatography/mass spectrometry/mass spec-
`trometry (LC/MS/MS) to determine the concentrations
`of midazolam, 1-hydroxymidazolam, and the added
`internal standard triazolam-d4. Analytes were ex-
`tracted from plasma using liquid phase extraction.
`They were separated using reverse phase high per-
`formance liquid chromatography on a 3-micron C18
`column. The analytes were detected using a
`PE/Sciex API III⫹ LC/MS/MS system in multiple
`reaction monitoring (MRM) mode for the following
`precursor/products (m/z): midazolam, 326 and 291;
`1-hydroxymidazolam, 342 and 203; and triazolam-
`d4, 347 and 312; with retention times of 1.75, 1.53,
`and 1.50 min, respectively. The lower limit of detec-
`tion for midazolam and 1-hydroxymidazolam was
`0.50 ng/mL. The method was linear over a concen-
`tration range of 0.50 –500.0 ng/mL using a 0.5-mL
`sample volume. Overall accuracy (inter-batch) was
`89.3%–106.0% for midazolam and 93.3%–108.0% for
`1-hydroxymidazolam. The inter batch overall preci-
`sion (%CV) was 2.9%–9.4% for midazolam and
`3.4%–113.0% for 1-hydroxymidazolam. The specific-
`ity, linearity, sensitivity, accuracy, precision, and
`stability of the data were within the Food and Drug
`Administration guidelines for Bioanalytical Method
`Validation. Freeze/thaw stability (3 cycles) and
`long-term sample storage (13 mo, ⫺20°C) were
`tested and were acceptable.
`A physician was in attendance for at least 4 h after
`each dose, and subjects were observed throughout the
`study session by a research nurse. Vital signs (arterial
`blood pressure, heart rate, respiratory rate) were mea-
`sured before and at 10, 20, 30, 45, 60, 75, 90, 105 min
`and 2, 3, 4, 6, 8, and 12 h after each dose. Continuous
`pulse oximetry monitoring was done for all subjects
`for the first 6 h and as clinically indicated thereafter.
`Any observation of oxygen saturation ⬍90% was
`recorded as an adverse event. In addition to sponta-
`neously reported subjective symptoms, which were
`allowed at any time, subjects were also questioned as
`to their adverse event experience each time that vital
`signs were recorded (12).
`AQUESTIVE EXHIBIT 1130 Page 0002
`© 2006 International Anesthesia Research Society 345
`
`
`
`An otolaryngologist examined the nasal passages to
`evaluate development of local mucosal irritation, in-
`flammation, bleeding, and excoriation or ulceration at
`screening, before dosing on each study day, at 2– 4 h
`after each dose, and within 72 h after treatment.
`Pharmacokinetic parameters were determined us-
`ing standard noncompartmental methods (13) with
`log-linear least square regression analysis (weighting
`factor 1/Y) to determine the elimination rate constants
`(
`Z) (WinNonlin version 3.2; Pharsight Corp., Palo
`Alto, CA). Time to and maximum plasma concentra-
`tion (Tmax and Cmax), elimination half-life (t1/2), area
`under the plasma concentration-time curve from time
`zero to infinity (AUC0 –⬁) were also calculated. The
`absolute bioavailability (F) for the IN and IM dosage
`formula F ⫽
`forms was determined by the
`AUCIN,0 –⬁/AUCIV,0 –⬁ for the IN dose and F ⫽
`AUCIM,0 –⬁/AUCIV,0 –⬁ for the IM dose.
`The subjects completed assessments of drug-
`induced impairment including a Digit-Symbol Substi-
`tution Task (DSST), Visual Analog self-measures and
`the Stanford Sleepiness Scale (SSS) at 0 (1 h before
`dosing as a practice session), 10, 20, 30, and 45 min and
`1, 1.5, 2, 3, 4, 6, 8, and 12 h (14,15).
`The Visual Analog Scale (VAS) consisted of 10
`statements (“Stimulated,” “Sedated,” “High,” “Anx-
`ious,” “Fatigued,” “Hungry,” “Headache,” “Feel a
`drug effect,” “Like the drug effect,” and “Willing to
`take the drug again”) that were presented sequentially
`above a 100-mm line labeled “not at all” on the left end
`and “extremely” on the right end.
`For the SSS, subjects described their current level of
`sleepiness among the following options: “Feeling ac-
`tive and alert, wide awake,” “Functioning at high
`level, but not peak, able to concentrate,” “Relaxed,
`awake, not at full alertness, responsive,” “A little
`foggy, not at peak, let down, ” “Fogginess, beginning
`to lose interest in remaining awake, slowed down,”
`“Sleepiness, prefer to be lying down, fighting sleep,
`woozy,” or “Almost in reverie, sleep onset soon, lost
`struggle to remain awake” (15).
`DSST performance was analyzed according to total
`trial rate, correct trial rate, incorrect trial rate, and
`percentage of trials that were correct. Ratings on the
`VAS were scored based on the number of discrete
`units between the subject’s rating and the left end-
`point on each 100-unit scale. When subjects failed to
`initiate the DSST or the SSS scales because they were
`asleep, they were assigned a “0” for total response rate
`and a “7” for the sleep rating. No other substitutions
`for missing values were possible given the nature of
`the other quantitative measures.
`Dependent variables were analyzed as a function of
`route and time after dose. Analyses of peak effects,
`time to peak effects, and area under the curve for
`response (AUCR), using linear trapezoidal rules, were
`also evaluated. Separate AUCR analyses were com-
`pleted between baseline and 4 h after dose (AUCR0 – 4;
`i.e., over the duration of peak effects) as well as
`
`346
`
`Intranasal Midazolam Kinetics/Dynamics
`
`between baseline and 12 h after dose (AUCR0 –12⬁, i.e.,
`over the full time course).
`Statistical analyses were performed using Proc
`GLM with PC SAS (version 6.12; SAS Institute, Cary,
`NC). An analysis of variance model with factors
`sequence, subject nested within sequence, treatment,
`and period, was performed for peak effects, time to
`peak effects, and AUCR. Gender and route were used
`to analyze peak effects, time to peak effects, and
`AUCR. For VAS analysis, given the number of missing
`values because of subjects being asleep, degrees of
`freedom were inconsistent across variables and condi-
`tions. The least square means for each treatment group
`and pairwise comparisons between treatment groups
`were presented. To assess the gender effect and
`gender-by-route interaction, an analysis of variance
`with factors gender, period, route, and gender-by-
`route interaction was performed.
`Descriptive statistics, mean and standard deviation,
`were calculated for the pharmacokinetic parameters.
`The statistical tests were two-sided with a critical level
`of 0.05. An analysis of variance with factors sequence,
`subject nested within sequence, treatment, and period
`was performed for log-transformed AUC and Cmax.
`The carryover effect for the three treatments was
`analyzed using an analysis of variance of
`log-
`transformed AUC and Cmax. Analysis of variance with
`factors sequence, subject nested within sequence,
`treatment and period for sequence (P ⬎ 0.1) indicated
`that carryover effects were not significant. The differ-
`ence in Tmax values between the IN and IM treatment
`was compared using an analysis of variance of rank-
`transformed Tmax.
`
`RESULTS
`All 12 subjects completed the study without clini-
`cally significant or serious adverse events. There were
`no clinically relevant changes in physical examination,
`nasal evaluations, or laboratory tests. Doses of the
`study drug were well tolerated and events were mild
`to moderate and temporary (2–90 min). The most
`common adverse events seen with nasal administra-
`tion were ones frequently associated with midazolam
`and nasal administration, e.g., eye watering (58% of
`IN doses), dizziness (25%, 25%, and 17% for IV, IM,
`and IN doses, respectively), bad taste (25% after IN
`doses), and nasal congestion/feeling nasopharyngeal
`irritation (100% after IN doses). Three of 12 subjects
`reported a bad taste immediately after IN administra-
`tion of midazolam that lasted 2–20 min. Four of 12
`subjects noted blurred vision, one each in the IV and
`IM groups and 2 in the IN group. No subjects experi-
`enced respiratory depression, apnea, laryngospasm,
`bronchospasm, or wheezing.
`Mean midazolam plasma concentration versus time
`curve profiles (n ⫽ 12) over the first 2 h after IV, IM,
`and IN administration are shown in Figure 1. Mid-
`azolam was rapidly absorbed after IN administration,
`AQUESTIVE EXHIBIT 1130 Page 0003
`ANESTHESIA & ANALGESIA
`
`
`
`Figure 1. Plasma concentrations of midazolam after 5-mg IV,
`IM, and IN midazolam administration. Values are mean
`(⫾ sd) for 12 subjects for each dose.
`
`with concentrations reaching a peak in 2 individuals at
`5 min and in 75% of the individuals in ⱕ10 min
`⫽ 10 min). Mean pharmacokinetic pa-
`(median Tmax
`rameters are presented in Table 1. Cmax values after
`the IN dose were higher than those after the IM dose.
`A significantly shorter Tmax was observed for the IN
`formulation compared with the IM formulation (P ⫽
`0.0001). Levels of 1-hydroxymidazolam were very
`low, and as such, are not reported. The ratios of
`metabolite to parent AUCs were 0.16 to 0.22 for the 3
`routes of administration.
`
`DSST
`Differences in drug-induced psychomotor and cog-
`nitive impairment were observed across routes of
`administration. Subjects were awakened, when pos-
`sible, to initiate performance tasks at all scheduled
`time points but were unable to complete all tasks on
`numerous occasions. Three subjects failed to complete
`performance tasks at 6 time points, as the result of
`sleepiness, after IM dose administration; 7 subjects
`failed to complete performance tasks at 13 time points
`after IN dose administration; and 11 subjects failed to
`complete performance tasks at 29 time points after IV
`dose administration. Figure 2 presents trial rate on the
`DSST as a function of time after dose administration.
`No gender or gender-by-route interactions were ob-
`served, except for AUCR0 – 4 (for gender, P ⫽ 0.0375).
`The carryover effect was not significant (P ⬎ 0.1) for
`all 8 DSST and parameters.
`
`Figure 2. Mean (n ⫽ 12) digit symbol substitution test (DSST)
`trial rating as a function of time over 4 h after 5-mg
`midazolam doses administered via IV, IM, and IN routes of
`administration.
`
`Table 2 presents pharmacodynamic parameters for
`DSST Trial Rate.
`It was obvious that differences
`among routes of administration occurred during the
`first 30 min after drug administration. On all mea-
`sures, the order of magnitude of effects were identical
`with IV producing larger effects with a faster onset of
`action than IN, which in turn produced larger effects
`with a faster onset than IM. Significant effects of route
`were obtained on AUCR0 – 4. Follow-up tests indicated
`a significant difference between the IV and IM routes
`only, with the IV route engendering significantly
`greater AUCR0 – 4 than the IM route (P ⫽ 0.002). No
`significant gender, route, or gender-by-route interac-
`tions were obtained on time to peak effect or
`AUCR0 –12.
`Figure 3 presents the relationship between midazo-
`lam concentrations from each route of administration
`in relation to DSST trial rates. The data parallel
`conclusions from Figure 2 and 4 showing a rapid
`affect of each route with an IV⬎IN⬎IM orientation. A
`shallow clockwise hysteresis is present for each route
`of administration.
`Figure 4 presents SSS ratings as a function of time
`after dose administration. Table 2 presents pharmaco-
`dynamic parameters for the SSS. As with the DSST
`task, the order of magnitude of effects on most phar-
`macodynamic outcome measures were identical, with
`
`Table 1. Pharmacokinetic Parameters of Midazolam After IV Infusion Over 15 min, IM and IN Administration of 5 mg Midazolam in
`Healthy Volunteers
`
`AUC0 –⬁ (ng 䡠 h/mL)
`F (%)
`t1/2 (h)
`Cmax (ng/mL)
`Tmax (min)
`Formulation
`assume 100%
`3.14 (0.7)
`186 (31)
`167 (48)
`12.4 (6.8)
`5 mg IV
`93.4 (12)
`4.17 (2.09)
`175 (39)
`59 (29)
`29.2 (10.9)
`5 mg IM
`72.5 (12)
`3.25 (0.97)
`134 (26)
`80 (17)
`10.3 (5.0)
`5 mg IN
`Tmax ⫽ time to maximum plasma concentration; Cmax ⫽ maximum plasma concentration; t1/2 ⫽ elimination half-life; AUC0 –⬁ ⫽ area under the plasma concentration-time curve from time zero
`to infinity; F ⫽ bioavailability; IV ⫽ intravenous; IM ⫽ intramuscular; IN ⫽ intranasal
`Data are mean (SD), n ⫽ 12.
`Median and range are 10 (5–31), 30 (20 – 60), and 10 (5–20) min for IV, IM, and IN Tmax, respectively.
`Shorter Tmax was observed for IN formulation compared with IM formulation (P ⫽ 0.0001), which was similar to IV formulation.
`
`AQUESTIVE EXHIBIT 1130 Page 0004
`© 2006 International Anesthesia Research Society 347
`
`Vol. 103, No. 2, August 2006
`
`
`
`Table 2. Mean Single-Dose Midazolam Pharmacodynamic (PD) Parameters and Multiple Comparisons After IV Infusion Over 15 min,
`IM and IN Administration of 5 mg Midazolam in Healthy Volunteers
`
`Treatment A
`5 mg IV
`
`Treatment B
`5 mg IM
`
`Treatment C
`5 mg IN
`
`IN vs IV IN vs IM IV vs IM Gender
`
`0.6689
`0.4318
`0.0731
`0.4635
`
`0.0568
`0.2232
`0.8514
`0.4117
`
`0.3233
`0.2129
`0.1116
`0.7692
`
`0.3690
`0.7997
`0.4703
`0.5990
`
`0.1635
`0.0497
`0.0020
`0.6575
`
`0.0081
`0.1456
`0.5908
`0.1858
`
`1.000
`0.8451
`0.0375
`0.0986
`
`0.0037
`0.9035
`0.1658
`0.1878
`
`PD Parameter
`DSST
`Peak effect (trials/s)
`Time to minimum (min)
`AUCR0–4 (rating ⫻ h)
`AUCR0–12 (rating ⫻ h)
`
`0.55 (0.06)
`12.5 (6.2)
`⫺0.41 (0.19)
`⫺0.30 (0.36)
`
`0.53 (0.06)
`133 (23)
`⫺0.20 (0.18)
`⫺0.21 (0.57)
`
`0.55 (0.08)
`59 (133)
`⫺0.30 (0.21)
`⫺0.16 (0.35)
`
`Stanford Sleepiness Scale
`5.0 (1.8)
`6.3 (1.2)
`Peak effect (rating)
`49 (25)
`27 (27)
`Time to peak (min)
`AUCR0–4 (rating ⫻ h)
`6.1 (2.8)
`7.1 (4.2)
`AUCR0–12 (rating ⫻ h)
`9.2 (8.7)
`4.4 (8.8)
`Values in parentheses for parameters are SD. DSST ⫽ digit symbol substitution test
`
`5.4 (1.9)
`45 (47)
`7.4 (3.7)
`7.1 (6.6)
`
`IV producing larger effects with a faster onset than IN,
`which in turn produced larger effects with a faster
`
`Figure 3. Digit symbol substitution test (DSST) trial rating as
`a function of midazolam concentration after 5-mg midazo-
`lam doses administered via IV, IM, and IN routes of
`administration.
`
`onset than IM. An exception was for AUCR0 – 4 in
`which the order was IN, IV, IM, suggesting that IN
`had an overall greater effect than IV in the first 4 h.
`There were no statistical differences in time to peak
`sleep ratings, AUCR0 – 4, or AUCR0 –12 values for the
`different routes. Significantly greater peak sleep rat-
`ings were observed following the IV route, which
`were significantly greater than following the IM route
`(P ⫽ 0.0081) but no different compared to the IN route.
`No significant gender, route or gender-by-route inter-
`actions were obtained on time to peak effect,
`AUCR0 – 4, or AUCR0 –12 values. Gender differences
`were significant for peak effect of sleep ratings. Fe-
`males had significantly higher sleep ratings than
`males for peak effect (P ⫽ 0.0037 in follow-up analy-
`sis). Significant gender-by-route interactions were ob-
`served for AUCR0 –12 in main effects analysis (P ⫽
`0.0145). A possible explanation for this finding is that
`the female subjects had a significantly lower weight
`than their male counterparts, providing a higher dose
`on a mg/kg basis.
`Statistical comparisons for IM and IN to the IV
`route were precluded because of the relatively large
`number of missing values from the IV route. How-
`ever, significant differences in ratings were observed
`across time for all VAS scales (P ⬍ 0.01), with the
`exception of Anxious. Differences in ratings of High
`were observed as a function of route (P ⬍ 0.05), with
`ratings after IN doses significantly larger than after IM
`doses. Route by time interactions were obtained on
`ratings of Sedated (P ⬍ 0.005), High (P ⬍ 0.0001),
`Headache (P ⬍ 0.05), and Feel Drug (P ⬍ 0.01). Simple
`effects analyses of these interactions indicated signifi-
`cant time effects during both the IN and IM routes,
`and significant differences between the IN and IM
`routes 10 (Sedated, High, Feel drug), 20 (High, Head-
`ache), and 60 (Headache) min after dose.
`
`Figure 4. Mean change from baseline (n ⫽ 12) Stanford
`Sleepiness Scale rating as a function of time over 4 h after
`5-mg midazolam doses administered via IV, IM, and IN
`routes of administration.
`
`348
`
`Intranasal Midazolam Kinetics/Dynamics
`
`DISCUSSION
`The earliest clinical studies of midazolam nasal
`delivery used dilute aqueous solutions, approximately
`AQUESTIVE EXHIBIT 1130 Page 0005
`ANESTHESIA & ANALGESIA
`
`
`
`1–5 mg/mL, of the marketed injection product deliv-
`ered via syringes or droppers (3– 8). Although that
`research permitted a proof of concept, the formulation
`approach was far from optimal. Midazolam solubility
`in water is limited, and must be buffered to approxi-
`mately pH 3 to remain in solution. A more concen-
`trated solution of midazolam, approximately 20 –25
`mg/mL, is required so that the desired dose can be
`administered in about 100 L. Most commercial nasal
`spray devices deliver 100 L as the generally accepted
`liquid volume that can be retained in the nasal cavity
`providing sufficient retention, allowing for absorp-
`tion, without anterior or posterior nasopharynx run-
`off. Moreover, the devices must deliver precise and
`accurate volumes for systemic delivery of very potent
`medications, such as midazolam.
`Gudmundsdottir et al. (9) prepared a nonsterile
`aqueous solution, buffered to pH 4.2 and preserved
`with benzalkonium chloride,
`in which midazolam
`solubility was increased to 17 mg/mL by cyclodextrin
`complexation. Multiple sprays were necessary to de-
`liver a 4- to 5-mg dose. The 28 mg/mL formulation
`used by Knoester et al. (11) had water and propylene
`glycol as co-solvents, buffered to pH 4, with the
`addition of 1% benzyl alcohol as an antimicrobial
`preservative (11). These investigators reported excel-
`lent pharmacokinetics, with a maximum concentra-
`tion of 71 ng/mL achieved in 14 minutes on average,
`and an average bioavailability of 83%. These data were
`superior to those of Gudmundsdottir et al., who demon-
`strated a 64% bioavailability, with a maximum concen-
`tration of only 42 ng/mL achieved in 15.5 minutes (9).
`Both products caused nasal irritation, suggesting mida-
`zolam itself may be the primary cause.
`The formulation used in our study was a 25 mg/mL
`solution using a nonaqueous co-solvent system of
`propylene glycol and PEG 400. The product is sterile,
`eliminating the use of irritating antimicrobial preser-
`vatives. A buffer system is not required in this format.
`Despite these formulation changes, the subjects still
`reported mild-to-moderate nasal and throat irritation,
`further suggesting the drug itself is the offending
`agent. Additional strategies are needed to manage
`irritation and bitter taste.
`This formulation produced excellent pharmacoki-
`netic and pharmacodynamic properties. Most notable
`was a median tmax of 10 minutes, faster than previous
`formulation reports, while achieving a peak level of 80
`ng/mL, well above the 40 ng/mL threshold to induce
`sedation. The formulation demonstrated reproducible
`results across the subjects and a very small sd for each
`parameter as compared with previous reports.
`Psychomotor impairment and sedation paralleled
`the pharmacokinetics producing impairment and se-
`dation within 20 minutes. Pharmacodynamic analyses
`indicated clearly that all three routes of administration
`engendered performance impairment and subjective
`ratings of Sleep and VAS ratings of Fatigued, Sedated,
`High, Headache, and Feel Drug Effect. The duration of
`
`Vol. 103, No. 2, August 2006
`
`the effects occurred over a 3-hour interval after drug
`administration. The magnitude of the effects produced
`a ranking of IV⬎IN⬎IM on most parameters. Given
`similarities in the psychomotor recovery after the
`different routes of administration,
`it appears that
`route is not a major influence on recovery times, as the
`pharmacokinetic and pharmacodynamic profiles con-
`verged within approximately 1 hour in this study.
`A limitation of this study was not obtaining a blood
`sample at 15 minutes after IN/IM administration and
`at the conclusion of the IV infusion. The deficiency has
`a potentially modest effect of over-estimating bioavail-
`ability of nasal administration. The data in aggregate
`suggest a good correlation between pharmacokinetic
`variables and pharmacodynamic response. The im-
`pact of this oversight is likely negligible in relation to
`the objectives of this pilot study.
`In conclusion, this study clearly demonstrates that
`this IN midazolam formulation is rapidly and reliably
`absorbed. The formulation is worthy of further inves-
`tigation as a therapeutic alternative for a convenient,
`noninvasive, and rapidly acting sedative.
`
`REFERENCES
`1. Folayan MO, Faponle A, Lamikanra A. Seminars on controver-
`sial issues. A review of the pharmacological approach to the
`management of dental anxiety in children. Int J Paediatr Dent
`2002;12:347–54.
`2. Kain ZN, Mayes LC, Bell C, et al. Premedication in the United
`States: a status report. Anesth Analg 1997;84:427–32.
`3. Wilton NCT, Leigh J, Rosen DR, Pandit UA. Preanesthetic
`sedation of preschool children using intranasal midazolam.
`Anesthesiology 1988;69:972–5.
`4. Zedie N, Amory D, Wagner BHJ, O’Hara DA. Comparison of
`intranasal midazolam and sufentanil premedication in pediatric
`outpatients. Clin Pharmacol Ther 1996;59:341–8.
`5. Rey E, Delaunay L, Pons G, et al. Pharmacokinetics of midazo-
`lam in children: comparative study of intranasal and intrave-
`nous administration. Eur J Clin Pharmacol 1991;41:355–7.
`6. Walbergh EJ, Wills RJ, Eckhert J. Plasma concentration of
`midazolam in children following intranasal administration.
`Anesthesiology 1991;74:233–5.
`7. Bjorkman S, Rigemar G, Idvall J. Pharmacokinetics of midazo-
`lam given as an intranasal spray to adult surgical patients. Br J
`Anaesth 1997;79:575–80.
`8. Burstein AH, Modica R, Hatton M, et al. Pharmacokinetics and
`pharmacodynamics of midazolam after intranasal administra-
`tion. J Clin Pharmacol 1997;37:711–8.
`9. Gudmundsdottir H, Sigurjonsdottir JF, Masson M, et al. Intra-
`nasal administration of midazolam in a cyclodextrin based
`formulation: bioavailability and clinical evaluation in humans.
`Pharmazie 2001;56:963–6.
`10. Loftsson T, Gudmundsdottir H, Sigurjonsdottir JF, et al. Cyclo-
`dextrin solubilization of benzodiazepines: formulation of mida-
`zolam nasal spray. Int J Pharm 2001;212:29–40.
`11. Knoester PD, Jonker DM, Van Der Hoeven RT, et al. Pharma-
`cokinetics and pharmacodynamics of midazolam administered
`as a concentrated intranasal spray: a study in healthy volun-
`teers. Br J Clin Pharmacol 2002;53:501–7.
`12. Coda BA, Rudy AC, Archer SM, Wermeling DP. Pharmacoki-
`netics and bioavailability of single dose intranasal hydromor-
`phone hydrochloride in healthy volunteers. Anesth Analg
`2003;97:117–23.
`13. Gibaldi M, Perrier D. Pharmacokinetics. 2nd ed. New York:
`Marcel Dekker; 1982: 409–17.
`14. McLeod DR, Griffiths RR, Bigelow GE, et al. An automated
`version of the digit symbol substitution test (DSST). Behav Res
`Methods Instrum 1982;14:463–6.
`15. Hoddes E, Zarcone V, Smythe H, et al. Quantification of
`sleepiness: a new approach. Psychophysiology 1973;10:431–6.
`
`AQUESTIVE EXHIBIT 1130 Page 0006
`© 2006 International Anesthesia Research Society 349
`
`