Volume 56, Issue 10 p. 1243-1253
Editor's Choice: Pharmacokinetics/Pharmacodynamics
Free Access

Pharmacokinetic Properties and Human Use Characteristics of an FDA-Approved Intranasal Naloxone Product for the Treatment of Opioid Overdose

Philip Krieter PhD

Philip Krieter PhD

National Institute on Drug Abuse, National Institutes of Health, Bethesda, MD, USA

Search for more papers by this author
Nora Chiang PhD

Nora Chiang PhD

National Institute on Drug Abuse, National Institutes of Health, Bethesda, MD, USA

Search for more papers by this author
Shwe Gyaw MD

Shwe Gyaw MD

National Institute on Drug Abuse, National Institutes of Health, Bethesda, MD, USA

Search for more papers by this author
Phil Skolnick PhD, DSc (hon)

Corresponding Author

Phil Skolnick PhD, DSc (hon)

National Institute on Drug Abuse, National Institutes of Health, Bethesda, MD, USA

Corresponding Author:

Phil Skolnick, PhD, National Institute on Drug Abuse, 6001 Executive Boulevard, Bethesda, MD 20892-9551

Email: [email protected]

Search for more papers by this author
Roger Crystal MD

Roger Crystal MD

Lightlake Therapeutics, New York, NY, USA

Search for more papers by this author
Fintan Keegan MSc

Fintan Keegan MSc

Adapt Pharma, Ltd, Dublin, Ireland

Search for more papers by this author
Julie Aker MT (ASCP)

Julie Aker MT (ASCP)

Concentrics Research, Indianapolis, IN, USA

Search for more papers by this author
Melissa Beck BA

Melissa Beck BA

Concentrics Research, Indianapolis, IN, USA

Search for more papers by this author
Jennifer Harris BA

Jennifer Harris BA

Concentrics Research, Indianapolis, IN, USA

Search for more papers by this author
First published: 05 May 2016
Citations: 72


Parenteral naloxone has been approved to treat opiate overdose for over 4 decades. Intranasal naloxone, administered “off label” using improvised devices, has been widely used by both first responders and the lay public to treat overdose. However, these improvised devices require training for effective use, and the recommended volumes (2 to 4 mL) exceed those considered optimum for intranasal administration. The present study compared the pharmacokinetic properties of intranasal naloxone (2 to 8 mg) delivered in low volumes (0.1 to 0.2 mL) using an Aptar Unit-Dose device to an approved (0.4 mg) intramuscular dose. A parallel study assessed the ease of use of this device in a simulated overdose situation. All doses of intranasal naloxone resulted in plasma concentrations and areas under the curve greater than those observed following the intramuscular dose; the time to reach maximum plasma concentrations was not different following intranasal and intramuscular administration. Plasma concentrations of naloxone were dose proportional between 2 and 8 mg and independent of whether drug was administered to 1 or both nostrils. In a study using individuals representative of the general population, >90% were able to perform both critical tasks (inserting nozzle into a nostril and pressing plunger) needed to deliver a simulated dose of naloxone without prior training. Based on both pharmacokinetic and human use studies, a 4-mg dose delivered in a single device (0.1 mL) was selected as the final product. This product can be used by first responders and the lay public, providing an important and potentially life-saving intervention for victims of an opioid overdose.

Opioid overdose is a serious and evolving public health problem in the United States.1, 2 Thus, more than 28,000 overdose deaths3 and 750,000 annual emergency department visits4 have been attributed to prescription opioids (eg, oxycodone, methadone) and heroin. Although the introduction of abuse deterrent formulations has apparently stabilized the death rate due to prescription opioids, an unintended consequence has been a dramatic rise in the rate of heroin-induced fatalities.1, 5 As part of a comprehensive effort to limit opioid-induced fatalities, multiple government agencies6 have endorsed wider access to naloxone (17-allyl-4,5α-epoxy-3,14-dihyroxymorphinan-6-one HCl), a high-affinity opiate receptor antagonist that has been used to treat the symptoms of opioid overdose, including respiratory depression, for over 40 years.7, 8

Naloxone has been approved by the US Food and Drug Administration (FDA) for parenteral administration, but in an attempt to reduce the morbidity and mortality associated with opioid overdose, there has been a dramatic increase in its off-label use by the intranasal (IN) route.9-11 Although most morbidity and mortality incidents are the result of accidental overdose involving prescription opioids,12 the distribution of improvised IN naloxone “kits” has largely been confined to individuals with opiate (eg, heroin) use disorders at high risk of overdose, the friends and family of these individuals, and first responders.10 These improvised IN naloxone kits generally consist of 1 or 2 prefilled syringes, each containing 2 mL of naloxone HCl (1 mg/mL) and a mucosal atomizing device. Individuals administering the naloxone are instructed to give half a vial, or 1 mL, in each nostril (for a total dose of 2 mL); a second dose can be administered if the patient has not responded to the first dose.9, 10 Despite multiple reports describing the effectiveness of using these improvised intranasal devices in reversing opiate overdose,9-11, 13, 14 a high error rate has been associated with both kit assembly and proper IN administration, even in individuals receiving training.15 Moreover, it is not known if the pharmacokinetic (PK) properties of naloxone produced by these improvised IN devices are equivalent to the approved dose of parenterally administered naloxone. Here, we describe the PK properties and usability profile of an IN naloxone HCl formulation delivered in low volume (0.1 mL) that was recently approved by the FDA for the treatment of opioid overdose.


Pharmacokinetic Study

Study Participants

The PK study was approved by the MidLands Independent Review Board (Overland Park, Kansas); all subjects gave written informed consent before participation. The study was carried out in accordance with the International Conference on Harmonisation for Good Clinical Practices guidelines.16 This trial was registered as NCT02572089 (www.clinicaltrials.gov).

Male and female volunteers aged 18 to 55 years with body mass index (BMI) 18 to 30 kg/m2 participated in the PK study. Subjects were currently not taking either prescription or over-the-counter medications; nonsmokers and subjects who smoked 20 or fewer cigarettes per day were enrolled. Screening procedures conducted within 21 days of study initiation included the following: medical history, physical examination, evidence of nasal irritation, 12-lead electrocardiogram, complete blood count, clinical chemistry, coagulation markers, hepatitis and human immunodeficiency screening, urinalysis, and urine drug screen. Female subjects were tested for pregnancy at screening and admission to the clinic. Subjects were excluded if they had either abnormal nasal anatomy or symptoms, an upper respiratory tract infection, used opioid analgesics for pain relief within the previous 14 days, or, in the judgment of the investigator, had significant acute or chronic medical conditions. Subjects were required to abstain from grapefruit juice and alcohol from 72 hours prior to admission to the end of the last blood draw of the study and from nicotine- and caffeine-containing products and food for at least 1 hour prior to and 2 hours after dose administration. On days of dosing, a subject's vital signs were required to be within the normal range before administration of naloxone, defined as systolic blood pressure >90 mm Hg and ≤140 mm Hg; diastolic blood pressure >55 mm Hg and ≤90 mm Hg; resting heart rate >40 beats per minute (bpm) and ≤100 bpm; and respiratory rate >8 respirations per minute (rpm) and ≤20 rpm.

Study Design

The PK study was an inpatient, open-label, randomized, 5-period, 5-treatment, 5-sequence, crossover study conducted at Vince & Associates Clinical Research (Overland Park, Kansas). Subjects were randomly assigned to 1 of the 5 sequences to ensure at least 6 subjects in each sequence. On the day after clinic admission, participants were administered the study drug in randomized order with a 4-day washout period between doses. Subjects remained in the clinic for 19 days until all 5 treatments were administered and returned 3 to 5 days after discharge for a follow-up visit. Subjects were fasted overnight before each dosing day and received 1 of the following 5 treatments:
  1. 2 mg naloxone IN (a single 0.1-mL spray of the 20 mg/mL formulation in 1 nostril)

  2. 4 mg naloxone IN (a single 0.1-mL spray of the 20 mg/mL formulation in each nostril)

  3. 4 mg naloxone IN (a single 0.1-mL spray of the 40 mg/mL formulation in 1 nostril)

  4. 8 mg naloxone IN (a single 0.1-mL spray of the 40 mg/mL formulation in each nostril)

  5. 0.4 mg naloxone IM (1.0 mL of the 0.4 mg/mL naloxone HCl for injection in the gluteus maximus)

These doses were chosen based on a pilot study using 2 mg and 4 mg naloxone delivered in low volumes (0.1-0.2 mL) with a different device (data not shown). The IN devices were coded so that neither the staff nor the subjects knew the concentration of naloxone solution administered. IN naloxone was administered in the supine position, and subjects remained in this position for approximately 1 hour after dosing. Subjects were instructed not to breathe when the drug was administered to simulate an opioid overdose with a patient in respiratory arrest. The nasal passage was examined by medical personnel for irritation using a 6-point scale at predose and at 5 minutes and 0.5, 1, 4, and 24 hours postdose. Nasal irritation was scored as follows: 0 (normal-appearing mucosa, no bleeding); 1 (inflamed mucosa, no bleeding); 2 (minor bleeding that stops within 1 minute); 3 (minor bleeding taking 1 to 5 minutes to stop); 4 (substantial bleeding for 4 to 60 minutes, does not require medical intervention); and 5 (ulcerated lesions, bleeding that requires medical intervention). Twelve-lead ECGs were collected predose and at 1 hour and 6 hours postdose. Venous blood samples were collected for the analyses of plasma naloxone concentrations predose and at 2.5, 5, 10, 15, 20, 30, 45, 60, 120, 180, 240, 300, 360, 480, and 720 minutes postdose using Vacutainer® tubes containing sodium heparin. The plasma was stored at –20°C until analyzed.

Study Drugs

Naloxone HCl powder (cGMP grade, 99.8% purity) was purchased from Mallinckrodt (St. Louis, Missouri); naloxone HCl for injection was manufactured by Hospira, Inc (Lake Forest, Illinois). Nasal devices (Aptar Unit-Dose device for liquids, Louveciennes, France) were supplied by Lightlake Therapeutics, Inc (New York, New York). The devices, containing naloxone HCl concentrations of 20 mg/mL and 40 mg/mL, were manufactured by DPT Laboratories, Ltd (Lakewood, New Jersey) and delivered a volume of 0.1 mL.

Analytical Methods

Plasma naloxone concentrations were determined using a validated liquid chromatography-tandem mass spectrometry (LC-MS/MS) assay. Plasma samples (0.2 mL) were added to individual wells of a 96-well plate along with 0.05 mL methanol:water (3:7) containing the internal standard (0.1 ng naloxone-d5) and 0.2 mL 1 M potassium phosphate (pH 7.2). After vortex mixing, the plate was loaded onto a preconditioned 96-well SPE plate and washed sequentially with 0.1% formic acid, acetonitrile, and dichloromethane:isopropanol (8:2). Naloxone was eluted with dichloromethane/isopropanol (8:2) containing 2% ammonium hydroxide to a new 96-well plate. After evaporation, the residue was reconstituted in 0.2 mL methanol:water (9:1) and submitted to LC-MS/MS analysis. Naloxone was analyzed using an AB MDS Sciex API-5000 LC-MS/MS system (Framingham, Massachusetts) with an atmospheric pressure chemical ionization source operated in the positive ion mode. The mobile phase consisted of a gradient increasing from 60% mobile phase A (0.04% ammonium hydroxide):40% mobile phase B (methanol:acetonitrile, 1:1) to 20% A/80% B with a flow rate of 0.4 mL/min through a 50×2.1 mm XBridge C18 column. Naloxone eluted at approximately 2 minutes. Ions monitored were m/z 328.3 and 212.1 for naloxone and m/z 333.3 and 212.1 for the internal standard. The calibration curves (peak area ratios) were linear (r2 > 0.980) over the concentration range of 10.0 pg/mL to 10 ng/mL; the lower limit of quantitation (LLOQ) was 10.0 pg/mL. The interday precision of the calibration curves and quality control samples ranged from 2.1% to 7.9%, and the accuracy ranged between –2.4% and 3.8% during the analysis of the samples.

Data Analyses

The safety population included all subjects who received at least 1 dose of naloxone; the PK population included all subjects who received at least 1 dose of naloxone with sufficient data to calculate meaningful PK parameters. PK parameters were calculated using standard noncompartmental methods and a validated installation of WinNonlin® Phoenix, version 6.3 (Pharsight Corp, St. Louis, Missouri). Values of peak plasma concentrations (Cmax) and the time to reach Cmax (tmax) were the observed values obtained directly from the concentration-time data. The terminal elimination half-life (t½) was estimated by linear regression analysis. The area under the concentration-time curve from time 0 to the last quantifiable concentration (AUC0-t) was determined by the linear trapezoidal method and extrapolated to infinity (AUC0-∞) by adding the value of the last quantifiable concentration divided by the terminal rate constant (λz). The extrapolated percentage of AUC0-∞ was less than 20% for all concentration profiles; therefore, only AUC0-∞ is reported. The apparent total body clearance (CL/F) was calculated as the dose divided by AUC0-∞. PK comparisons were performed using a mixed-effects model in which sequence, period, and treatment were independent factors. Dose proportionality for all IN doses of naloxone was accessed using Cmax and AUC0-∞ parameters. In this analysis the mixed-effects power model urn:x-wiley:00912700:media:jcph759:jcph759-math-0001 was used, and 90% confidence intervals (90%CI) were constructed for the ratio of the dose-normalized geometric mean values (Rdnm) of the parameters.17 All analyses of demographic and safety data were performed using SAS® statistical software, version 9.3 (SAS Institute, Inc, Cary, North Carolina).

Human Factors and Usability Studies

Study Participants

The study was reviewed and approved by Concentrics Institutional Review Board (Indianapolis, Indiana); participants or a guardian reviewed and gave written informed consent before participation. The study was carried out in accordance with the Draft Guidance for Industry and Food and Drug Administration Staff: Applying Human Factors and Usability Engineering to Optimize Medical Device Design, June 22, 201118 and the Guidance for Industry—Label Comprehension Studies for Nonprescription Drug Products, August 2010.19

Adolescents aged 12 to 17 years and adults 18 years of age or older participated in the 2 human use studies. The REALM test20 and REALM-Teen test21 were administered to the adults and adolescents, respectively, in order to screen literacy levels for information only in analyzing the population. The tests are based on a list of 66 words commonly found on medication labels. The subjects needed to be able to read, speak, and understand the nature of the study procedures. They were excluded if they had ever been trained or employed as a healthcare professional or had participated in any clinical trial, product label study, or market research study in the past 12 months.

Study Design

The study was conducted in rooms equipped with 1-way mirrors for observation. To simulate a real-life emergency, study participants were challenged with administering the medication to an unconscious victim, simulated by a full-sized mannequin. No training on the use of the device was provided prior to the usability assessment.

Intranasal devices (Aptar Unit-Dose device, Louveciennes, France) were filled with 0.1 mL of water and packed into a blister card. The blister card, along with a Quick Study Guide (QSG) and patient information section of the package insert were packed into a carton.

Study A (2 devices) was slightly more complex and was conducted prior to study B (1 device) in order to determine if individuals were able to perform critical tasks without reviewing the QSG. The objective of the QSG was to provide clear and concise instructions (combined with pictures) for use in a crisis situation with limited time to interpret the directions (Figure 1). Subjects in study A were randomized to 1 of 2 arms: subjects in arm 1 were given an opportunity to read the QSG in advance of the simulation, whereas subjects in arm 2 did not review the QSG in advance. Subjects in study B (1 device) did not review the QSG in advance of the simulation.

Details are in the caption following the image
Quick Start Guide for use with 1 (left panel) and 2 (right panel) intranasal devices used in study B and study A, respectively.

Subjects were presented with a scenario of an unconscious overdose victim simulated by a life-sized mannequin similar to those used for cardiopulmonary resuscitation training. Subjects were given the product with labeling and asked to proceed as they would in a real-life emergency; no training or coaching was provided either prior to or during the simulation. Background noise, in the form of TV and radio, was introduced into the scenario to simulate voices and noise from onlookers. A trained observer (located behind a 1-way mirror) documented the steps that the subject took during the simulation. Once the subject completed the simulation, an interview was conducted in a separate room to evaluate comprehension of key concepts in the patient information section of the package insert. After the comprehension interview, additional questions were asked about any incorrect actions that were observed during the human factors testing; this information was obtained in order to identify any potentially confusing sections of the labeling.

Data Analyses

The primary endpoints for the critical tasks were (1) inserting the device nozzle into a nostril and (2) pressing the plunger to release a dose into the nose. Secondary endpoints included (3) checking for response, (4) calling 911, and (5) moving to a recovery position after administering dose. Study A also included (6) waiting 2 to 3 minutes to assess the effectiveness of the first dose and (7) readministration using a new unit (if needed). As a subject interfaced with a mannequin rather than a person, the observer was able to make judgments on mitigating circumstances in which the subject was either restricted or confounded by the mannequin. This allows a response that is “not perfect or technically correct” to be considered as “correct” if the subject's intent indicates that he or she either understood the correct action or that the apparent incorrect action has no safety risk. An example would be partial insertion into the mannequin's nose because it lacked flexibility of a human nose. These were added to the final correct performance scores of primary and secondary tasks.

The correct score and lower boundary of the 95%CI were calculated for each of the 2 human factors primary endpoints. The correct score was the point estimate for execution of both critical tasks, defined as the total number of subjects who correctly completed both critical tasks, divided by the number of subjects who performed the tasks, multiplied by 100. Success criteria for the combined primary endpoints had a lower bound threshold of at least 69% for the correct scores in study A (2 devices) or at least 73% for the correct scores in study B (1 device). This was based on sample size and CIs around a mean, a point estimate of the population mean. For all remaining human factors tasks and comprehension objectives, correct scores and 2-sided 95%CIs were computed; however, no thresholds were established. Subgroup analyses were conducted to evaluate any potential differences between subjects with low literacy and adolescent subjects aged 12 to 17 years. An error rate of 6% for each task was estimated based on preliminary qualitative work using untrained users completing each task. This led to a projected probability of completing both critical core tasks correctly for any particular subject to be at least 88%. This estimate is consistent with rates accepted by the FDA for other approved products expected to be used by the lay public, including nasal sprays. With a sample size of 30 subjects per arm, the probability of having the lower bound of the 2-sided 95%CI (of the estimated proportion of subjects correctly completing all core tasks) above the predefined 69% threshold was 87%, assuming the true correct demonstration of the core tasks rate was 88%. With a sample size of 50 subjects, the probability of having the lower bound of the 2-sided 95%CI (of the estimated proportion of subjects correctly completing all core tasks) being above the predefined 73% threshold was 88%, assuming the true correct simulated use demonstration of the core tasks rate was 88%.


Pharmacokinetic Study

Subject Characteristics

Eighteen male and 12 female subjects (Table 1) received at least 1 dose of naloxone; 28 subjects completed the study. One male subject discontinued the study after 1 treatment period (treatment C) due to a systolic blood pressure greater than 140 mm Hg prior to the start of the second period. This predose elevation in blood pressure was judged to be unrelated to drug treatment. One other male subject withdrew for personal reasons after completing 4 treatments; this individual did not receive treatment C.

Table 1. Pharmacokinetics of Naloxone: Subject Demographics
All Male Female
Number 30 18 12
Mean age, years (range) 35.8 (22–55) 36.9 (22–55) 34.2 (24–46)
White 7 3 4
Black/African American 23 15 8
Hispanic or Latino 2 2 0
Not Hispanic or Latino 28 16 12
Mean weight, kg (range) 80.2 (56–102) 86.2 (56–102) 71.3 (57–85)
Mean BMI,a kg/m2 (range) 26.5 (19.6–29.8) 26.8 (19.6–29.8) 26.1 (22.0–28.7)
  • a BMI, body mass index.


Naloxone plasma concentrations were above the lower limit of quantitation (10.0 pg/mL) at 2.5 minutes after IN administration, the first collection time point, in all but 1 (an individual receiving an 8-mg dose) of 114 samples collected; concentrations were measurable in 22 of 29 samples following IM injection (data not shown). The median tmax values after IN and IM doses ranged from 20 minutes to 30 minutes, indicating that naloxone was absorbed rapidly following either route of administration (Table 2; Figure 2). The mean Cmax values increased from 3.1 ng/mL to 10.3 ng/mL as the IN dose increased from 2 mg to 8 mg; the mean Cmax value following IM dosing was 0.9 ng/mL. AUC0-∞ increased from 4.7 ng·h/mL to 15.8 ng·h/mL as the IN dose increased 4-fold from 2 mg to 8 mg. The terminal elimination half-life of naloxone after all 4 IN regimens was approximately 2 hours; it was 1.3 hours after the IM injection. The geometric mean ratios of the dose-corrected Cmax values of the IN doses compared to the IM dose ranged between 55.1% and 70.8%, whereas the ratio was approximately half for AUC0-∞ (Table 3). Based on the actual and dose-corrected values of Cmax and AUC, the IM and IN doses were not equivalent. The pharmacokinetic properties of naloxone following IN administration were similar between male and female subjects for all 4 treatments (Figure 3, Table 4).

Table 2. Pharmacokinetics of IN and IM Naloxone
Parameter (Units)a 2 mg, 1 Spray 20 mg/mL IN (A) 4 mg, 2 Sprays 20 mg/mL IN (B) 4 mg, 1 Spray 40 mg/mL IN (C) 8 mg, 2 Sprays 40 mg/mL IN (D) 0.4 mg IM (E)
Nb 29 29 29 29 29
Cmax (ng/mL) 3.1 (36.1) 6.5 (32.3) 5.3 (44.6) 10.3 (38.1) 0.9 (31.2)
tmax (hours) 0.3 (0.3,1.0) 0.3 (0.2, 0.6) 0.5 (0.2, 1.0) 0.3 (0.2, 1.0) 0.4 (0.1, 2.1)
AUC0-∞(ng·h/mL) 4.7 (29.8) 9.7 (26.7) 8.5 (39.0) 15.8 (23.1) 1.8 (22.7)
t½ (hours) 1.9 (34.6) 2.4 (31.7) 2.2 (29.1) 2.2 (39.0) 1.3 (27.8)
CL/F (L/h) 452.5 (24.4) 435.4 (22.3) 534.7 (35.9) 530.4 (22.0) 232.8 (22.4)
  • a Mean (%CV) for all except median (range) for Tmax. %CV, percent coefficient of variation; AUC0-inf, area under the plasma concentration-time curve from time zero to infinity CL/F, apparent clearance; Cmax, maximum plasma concentration; t½, terminal half-life; tmax, time to Cmax.
  • b Number of subjects receiving treatment.
Details are in the caption following the image
Plasma concentrations of naloxone following intranasal and intramuscular administration of naloxone HCl. Twenty-eight subjects were randomized in a 5-period, 5-treatment, 5-sequence crossover study, receiving 1 or 2 doses (0.1 mL per nostril) of a naloxone HCl formulation (20 and 40 mg/mL) or an intramuscular injection of 0.4 mg.  IN, intranasal; IM, intramuscular.
Table 3. Naloxone Statistical Summary of Treatment Comparisons (Intranasal vs Intramuscular Administration)
Parameter (Units)a IN Administration (Test) Comparison Ratio (Test/Reference) of Adjusted Meansb 90%CI for Ratio
Cmax/Dose (ng/mL/mg) 2 mg IN (Trt A) A vs E 66.7 (59.5-74.7)
4 mg IN (Trt B) B vs E 70.8 (63.2-79.4)
4 mg IN (Trt C) C vs E 55.1 (49.2-61.8)
8 mg IN (Trt D) D vs E 55.3 (49.3-62.0)
AUC0-∞ /Dose (ng·h/[mL·mg]) 2 mg IN (Trt A) A vs E 51.9 (48.3-55.8)
4 mg IN (Trt B) B vs E 53.5 (49.8-57.5)
4 mg IN (Trt C) C vs E 46.2 (43.0-49.8)
8 mg IN (Trt D) D vs E 43.9 (40.8-47.2)
  • a AUC0-∞/Dose, AUC per milligram naloxone administered; Cmax/Dose, Cmax per milligram naloxone administered; IM, intramuscular, IN, intranasal; Trt, treatment.
  • b Geometric least-squares mean ratio between treatments, expressed as a percentage of IM administration (reference, treatment E); see Table 2 for additional detail.
Details are in the caption following the image
Plasma concentrations of naloxone in male and female subjects following intranasal administration of 2, 4, and 8 mg naloxone HCl. Male and female subjects received either a single 0.1 mL intranasal dose (20 and 40 mg/mL) or 2 doses (1 in each nostril) of naloxone HCl (40 mg/mL).
Table 4. Pharmacokinetics of Naloxone in Female and Male Subjects
2 mg, 1 Spray 20 mg/mL IN (A) 4 mg, 2 Sprays 20 mg/mL IN (B) 4 mg, 1 Spray 40 mg/mL IN (C) 8 mg, 2 Sprays 40 mg/mL IN (D) 0.4 mg IM (E)
Parameter (Units)a Female Male Female Male Female Male Female Male Female Male
Nb 12 17 12 17 12 17 12 17 12 17
Cmax (ng/mL) 2.8 (25.6) 3.3 (39.5) 6.4 (25.0) 6.6 (37.0) 5.1 (41.0) 5.4 (47.7) 9.6 (29.4) 10.9 (41.9) 1.1 (28.2) 0.8 (29.2)
Tmax (h) 0.3 (0.3, 0.8) 0.3 (0.3, 1.0) 0.3 (0.2, 0.6) 0.3 (0.2, 0.5) 0.5 (0.2, 0.8) 0.5 (0.2, 1.0) 0.3 (0.2, 0.8) 0.4 (0.2, 1.0) 0.4 (0.1, 0.8) 0.5 (0.2, 2.1)
AUC0-∞ (ng·h/mL) 4.6 (31.3) 4.8 (29.5) 9.6 (19.3) 9.8 (31.4) 7.8 (39.3) 8.9 (38.9) 14.6 (22.6) 16.7 (22.5) 1.8 (20.6) 1.8 (24.7)
t½ (h) 1.7 (33.6) 2.1 (33.7) 2.5 (35.4) 2.3 (28.8) 2.4 (33.1) 2.0 (23.7) 2.4 (49.1) 2.1 (26.5) 1.1 (20.7) 1.4 (27.9)
CL/F (L/h) 463.5 (25.3) 447 (24.3) 432.6 (22.0) 437.4 (23.2) 571.1 (33.1) 509.0 (38.4) 572.2 (21.9) 501.0 (20.6) 226.5 (21.0) 237.3 (23.7)
  • a Mean (%CV) for all except median (range) for Tmax; %CV, percent coefficient of variation; AUC0-∞, area under the plasma concentration-time curve from time zero to infinity; CL/F, apparent clearance; Cmax, maximum plasma concentration; t½, terminal half-life; Tmax, time to Cmax.
  • b Number of subjects receiving treatment.

A power model was used to test for dose proportionality following IN naloxone.17 This was assessed by plotting the slope of the log dose and the log of Cmax and AUC0-∞. A slope equal to 1.0 would indicate dose proportionality. The slope and 90%CI of the line for Cmax were 0.83 and 0.74 to 0.93. For AUC0-∞, the slope and 90%CI of the line were 0.85 and 0.78 to 0.91, indicating exposure to naloxone approached dose proportionality as the dose was increased from 2 mg to 8 mg.


There were no differences in the safety profile of IN naloxone compared to IM dosing. No significant erythema, edema, erosion, or other signs were observed in the nasal cavity prior to or after administration of IN naloxone. Few adverse events (AEs) associated with the nasal mucosa were reported; all were mild (grade 1), transient, and with no clear relationship to dose (Table 5). Vital signs, ECG, and clinical laboratory parameters did not reveal any clinically significant changes or evidence of QTcF prolongation after naloxone administration. Two additional AEs were reported with IN administration: 1 for nasal pain after the 2-mg dose and 1 for headache after the 8-mg dose.

Table 5. Nasal Mucosal Adverse Events Following IN Naloxone
Treatment Erythemaa Edemaa Othera, b Totala
2 mg (20 mg/mL, 1 spray) 3 3 1 7
4 mg (20 mg/mL, 2 sprays) 1 0 0 1
4 mg (40 mg/mL, 1 spray) 1 2 0 3
8 mg (40 mg/mL, 2 sprays) 0 1 0 1
  • a All AEs were grade 1.
  • b Nasal pain.

Human Factors and Usability

Subject Characteristics

The human factors studies were conducted in a population of 116 participants (Table 6). Combined, 33 subjects were adolescents (aged 12 to 17 years), and 83 were adults (aged 18 years or older). A total of 59 (50.9%) had low literacy, and 57 (49.1%) had normal literacy.

Table 6. Human Factors and Usability Study Demographics
Study A (N = 63) Study B (N = 53)
Characteristics n % n %
Male 33 52.4 25 47.2
Female 30 47.6 28 52.8
12 to 17 17 27.0 16 30.2
18 to 24 4 6.3 4 7.5
25 to 34 6 9.5 9 17.0
35 to 44 11 17.5 9 17.0
45 to 54 9 14.3 9 17.0
55 to 64 8 12.7 3 5.7
65 or older 8 12.7 3 5.7
White 38 60.3 30 56.6
African American/Black 14 22.2 11 20.8
Native American 1 1.6 1 1.9
Asian/Pacific Islander 1 1.6 1 1.9
Other 9 14.3 10 18.9
Hispanic Origin
Yes 10 15.9 13 24.5
No 53 84.1 40 75.5
Adult Education
Less than high school 4 8.7 3 8.1
Completed high school/GED 12 26.1 13 35.1
Some college/technical school 18 39.1 11 29.7
Graduated from college/technical school or more 12 26.1 10 27.0
Adolescent education
6th Grade 4 23.5 1 6.3
7th Grade 5 29.4 1 6.3
8th Grade 2 11.8 2 12.5
9th Grade 3 17.6 6 37.5
10th Grade 1 5.9 3 18.8
11th Grade 2 11.8 3 18.8
$0 to $14,999 9 14.3 3 5.7
$15,000 to $24,999 4 6.3 8 15.1
$25,000 to $34,999 7 11.1 8 15.1
$35,000 to $44,999 8 12.7 7 13.2
$45,000 to $64,999 14 22.2 5 9.4
$65,000 to $74,999 7 11.1 3 5.7
$75,000 or more 7 11.1 14 26.4
Refused 7 11.1 5 9.4
Employment Status
Employed full-time 17 27.0 15 28.3
Employed part-time 8 12.7 9 17.0
Unemployed 6 9.5 6 11.3
Student 16 25.4 17 32.1
Retired 8 12.7 3 5.7
Disabled 1 1.6 1 1.9
Self-employed 2 3.2 1 1.9
Homemaker 2 3.2 1 1.9
Other 3 4.8 0 0
REALM Category
Low literacy 43 68.3 16 30.2
Normal literacy 20 31.7 37 69.8

Subjects demonstrated the ability to correctly perform both critical tasks, meeting the success threshold in studies using either 1 or 2 devices (Table 7). Using 1 device, 90.6% of subjects (n = 48 of 53) were able to correctly perform both critical tasks. Similar results were obtained in the study using 2 devices: in arm 1, which included a review of the QSG, 90.6% (n = 29 of 32) of subjects correctly performed both critical tasks, while in arm 2, 90.3% (n = 28 of 31) of subjects correctly performed both critical tasks without reviewing the QSG. In all cases the lower bound of the 95%CI was exceeded. Five secondary human factors tasks (Table 7) were also evaluated with correct performance ranging from 37.7% to 93.8% depending on whether 1 or 2 devices was used and whether or not the QSG was reviewed in advance. Among the secondary tasks evaluated, 70% to 80% of subjects correctly administered a second IN dose, although fewer than 60% were able to correctly administer this second dose within 2 to 3 minutes of the first dose (Table 7). Among other secondary tasks evaluated, the lowest scoring items included moving the victim to a recovery position and waiting 2 to 3 minutes before administering a second dose (Table 7). Comprehension was also evaluated on 9 communication objectives found in the patient information section of the package insert to insure that the general population could understand key messages. Overall, the scores were high, ranging from 70% to 100%, with most scores in the 90% to 100% range (Table 8).

Table 7. Human Factors Study: Primary and Secondary Endpoints
Study A Study B
2 Sprays Arm 1—QSG 2 Sprays Arm 2—QSG 1 Spray QSG Not
Reviewed Prior to Test Not Reviewed Reviewed
(N = 32) Prior to Test (N = 31) Prior to Test (N = 53)
Tasks Endpoint Type % (n) 95%CI % (n) 95%CI % (n) 95%CI
1a) Check for a response Secondary 93.8 (30) 79.2–99.2 77.4 (24) 58.9–90.4 83.0 (44) 70.2–91.9
2a) Insert nozzle into nostril Primary 100.0 (32) 89.1–100.0 100.0 (31) 88.8–100.0 96.2 (51) 87.0–99.5
2b) Press plunger to release dose into nose (Location and Dose Release Combined) Primary 90.6 (29) 75.0–98.0 90.3 (28) 74.3–98.0 90.6 (48) 79.3–96.9
3a) Call 911 Secondary 75.0 (24) 56.6–88.5 80.6 (25) 62.5–92.6 71.7 (38) 57.7–83.2
3b) Move to recovery position after administering dose Secondary 68.8 (22) 50.0–83.9 48.4 (15) 30.2–66.9 37.7 (20) 24.8–52.1
4a) Wait 2–3 minutes Secondary 59.4 (19) 40.6–76.3 54.8 (17) 36.0–72.7 NA NA
4b) Readminister using new unit (if needed) Secondary 80.0 (24) 61.4–92.3 70.0 (21) 50.6–85.3 NA NA
Success threshold, lower bound only 69% 69% 73%
Table 8. Comprehension Objectivesa
Study A (2 Devices) Study B (1 Device)
Total N = 63 Total N = 53
Comprehension Objectives % (n) 95%CI % (n) 95%CI
Primary Objectives
Q.1 What is the drug's purpose? 100.0 (63) 94.3–100.0 96.2 (51) 87.02–99.5
Q.6 After use, is it still necessary to get emergency medical help? 98.4 (62) 91.5–100.0 100.0 (53) 93.3–100.0
Q.8 How should the drug be administered? 95.2 (60) 86.7–99.0 100.0 (53) 93.3–100.0
Q.7 What is an example of a sign of an opioid emergency? 90.5 (57) 80.4–96.4 94.3 (50) 84.3–98.8
Q.4 What is an example of a potential withdrawal symptom after someone receives this product? 87.3 (55) 76.5–94.4 77.4 (41) 63.8–87.7
Q.2 What does the drug do? 85.7 (54) 74.6–93.3 86.8 (46) 74.7–94.5
Exploratory Objectives
Q.9 Who should not use this drug? 92.1 (58) 82.4–97.4 92.5 (49) 81.8–97.9
Q.5 What should you tell your healthcare provider before using this drug? 87.3 (55) 76.5–94.4 79.2 (42) 66.0–89.2
Q.3 Will this have any effect in people who are not taking opioid medicines? 79.4 (50 ) 67.3–88.5 69.8 (37) 55.7–81.7
  • a Values represent the percentage of subjects providing a correct response with the 95% confidence interval (CI). The number of subjects responding to each question is in parentheses.


Here, we demonstrate that IN administration of a naloxone HCl (2 to 8 mg) formulation in low volume (0.1 to 0.2 mL) results in PK parameters that either equal or exceed those observed following the IM dose of naloxone (0.4 mg) that is approved for the treatment of opioid overdose. The time to peak naloxone plasma concentration following IN administration was independent of dose and ranged from 0.3 hour to 0.5 hour compared with 0.4 hour following IM injection (Table 4). The maximum plasma concentration obtained between 2 mg and 8 mg approached dose proportionality, exceeded the maximum concentration produced by the 0.4 mg IM dose of naloxone, and was independent of delivery to either 1 or both nostrils (Table 2, Figure 2). Moreover, at the earliest time sampled (2.5 minutes), plasma naloxone concentrations after IN administration of 2 mg to 8 mg were higher than following the 0.4-mg IM dose; this rapid rise may be advantageous for reversing respiratory depression in opioid overdose.

The high plasma concentrations of naloxone following IN administration reported here (Figure 2, Table 2) may be viewed as surprising based on a study reporting an absolute bioavailability of 4% after IN administration.22 In that study, naloxone (0.8 and 2 mg) was administered in volumes of 1 and 2.5 mL/nostril with a mucosal atomizing device. Optimal delivery volumes for IN drug administration are ≤0.2 mL/nostril.23 This low bioavailability was attributed to the delivery of relatively large volumes to healthy, awake volunteers who swallowed significant amounts of drug that pooled in the nasopharynx.22 These authors hypothesized that unconscious patients with depressed oropharyngeal reflexes may swallow less IN naloxone, thereby increasing absorption and bioavailability. Although there are no peer-reviewed, published studies describing the PK properties of IN naloxone dosed with an improvised device (that is, 2 mg to 4 mg administered in a volume of up to 2 mL in each nostril), based on the Dowling et al study,22 plasma concentrations are likely to fall below the lowest approved (0.4 mg) parenteral dose.24 Nonetheless, multiple reports9-11 have described the reversal of an apparent opioid overdose using these IN improvised devices. These retrospective reports focus on successful rescues but do not provide insight into treatment failures.24-26

Multiple studies have demonstrated sex differences in the PK properties of drugs,27, 28 and an interaction between sex and route of administration has also been documented. For example, the Cmax of cocaine is twice as high in males compared to females when administered IN29 but is similar when given IV.30 In the present study, no sex differences were observed in the PK profile of IN naloxone at doses between 2 mg and 8 mg (Figure 3, Table 4), indicating that both males and females will respond similarly in a rescue situation.

Administration of IN naloxone obviates the potential for needlestick harm and is especially useful for individuals either unable or unwilling to administer an injection. Moreover, the single-use spray is intuitive and rapid: subjects could use the device and deliver a dose of naloxone without either prior training or review of the QSG (Table 7). A single-use device also avoids the potential for confusion around the timing of a second dose, which is associated with an error rate of >40% whether or not the QSG is reviewed (Table 5). Because naloxone has a short half-life (Table 2) relative to many prescription opioids including hydrocodone and methadone,31-33 the plasma concentrations produced by these doses of IN naloxone may provide a more durable reversal of opioid overdose than either improvised IN naloxone kits or a lower approved initial dose (Figures 2 and 3).

Retrospective analyses suggest that improvised IN naloxone kits can be used effectively, even in the absence of training.9, 34 Nonetheless, a prospective comparative usability study15 found that despite training, fewer than 60% of participants were able to successfully administer a simulated IN dose using an improvised nasal naloxone kit; in the absence of training, none of the 42 participants in this study successfully assembled the device and delivered a simulated dose of naloxone.15 By contrast, human factors studies demonstrated that participants were able to correctly perform the 2 critical tasks (insert device in nostril and deliver a dose of medication) necessary to administer a simulated dose of naloxone using an Aptar Unit-Dose device with a success rate >90%, even when no prior training or guidance was provided (Table 7). This high rate of success is not surprising because the device is intuitive, and has been used effectively in other IN products available to the lay public, including sumatriptan (Imitrex®). Moreover, comprehension scores were high for most of the key communication objectives in the patient information section of the package insert (Table 8), indicating that the labeling is clear.

Because naloxone is a competitive opiate receptor antagonist, its efficacy in reversing overdose depends on both the quantity of opioids ingested and their relative affinities for opiate receptors. This information is rarely available to first responders (or bystanders) attempting to resuscitate a suspected overdose. A single dose of 4 mg naloxone was selected as the final formulation for Narcan® Nasal Spray because it delivers approximately the same amount of drug as 2 mg IM (Table 3), the highest recommended initial dose for treating suspected opioid overdose.35 In addition to the simplicity associated with a single administration, the 4-mg IN dose increases the potential for a reversal of opioid overdose compared to either improvised IN devices2426 or the approved autoinjector that delivers 0.4 mg of naloxone.36 This is especially relevant because of the dramatic rise in the abuse of high-potency opioids such as fentanyl37 that require higher concentrations of naloxone to treat overdose.25 Based on the selectivity of naloxone for opiate receptors and the human use characteristics described here, this product can be used safely and effectively with no prior training to provide an important and potentially life-saving intervention for the treatment of opioid overdose.


R.C. is an employee of Lightlake Therapeutics. F.K. is an employee of Adapt Pharma, Ltd. J.A., M.B., and J.H. are employees of Concentrics Research. P.K., N.C., S.G., and P.S. are employees of the National Institutes of Health. The work was supported in part by Lightlake Therapeutics, Adapt Pharma, and NIDA contracts N01DA-12-8905, N01DA-13-8920, and N01DA-14-8914.