Intranasal Glucagon for Treatmentof Insulin-Induced Hypoglycemiain Adults With Type 1 Diabetes: ARandomized CrossoverNoninferiority StudyDiabetes Care 2016;39:264–270 | DOI: 10.2337/dc15-1498

OBJECTIVE

Treatment of severe hypoglycemia with loss of consciousness or seizure outside ofthe hospital setting is presently limited to intramuscular glucagon requiring re-constitution immediately prior to injection, a process prone to error or omission. Aneedle-free intranasal glucagon preparation was compared with intramuscularglucagon for treatment of insulin-induced hypoglycemia.

RESEARCH DESIGN AND METHODS

At eight clinical centers, a randomized crossover noninferiority trial was con-ducted involving 75 adults with type 1 diabetes (mean age, 336 12 years; mediandiabetes duration, 18 years) to compare intranasal (3 mg) versus intramuscular(1 mg) glucagon for treatment of hypoglycemia induced by intravenous insulin.Success was defined as an increase in plasma glucose to ‡70 mg/dL or ‡20 mg/dLfrom the glucose nadir within 30 min after receiving glucagon.

RESULTS

Mean plasma glucose at time of glucagon administrationwas 486 8 and 496 8mg/dLat the intranasal and intramuscular visits, respectively. Success criteria were met atall but one intranasal visit and at all intramuscular visits (98.7% vs. 100%; difference1.3%, upper end of 1-sided 97.5% CI 4.0%). Mean time to success was 16 min forintranasal and 13 min for intramuscular (P < 0.001). Head/facial discomfort wasreported during 25% of intranasal and 9% of intramuscular dosing visits; nausea(with or without vomiting) occurred with 35% and 38% of visits, respectively.

CONCLUSIONS

Intranasal glucagon was highly effective in treating insulin-induced hypoglycemiain adults with type 1 diabetes. Although the trial was conducted in a controlledsetting, the results are applicable to real-world management of severe hypogly-cemia, which occurs owing to excessive therapeutic insulin relative to the im-paired or absent endogenous glucagon response.

Hypoglycemia is a major barrier to the attainment of glycemic control targets inpatients with diabetes (1). Patients with type 1 diabetes bear the greatest burden ofproblematic hypoglycemia. In such patients, mild-to-moderate hypoglycemia(blood glucose 50–70 mg/dL) occurs frequently and responds in most cases to

1University of Pennsylvania Perelman School ofMedicine, Philadelphia, PA2Jaeb Center for Health Research, Tampa, FL3Locemia Solutions ULC, Montreal, QC, Canada4Yale School of Medicine, New Haven, CT5Jacobs School of Medicine and Biomedical Sci-ences, University at Buffalo, The State Universityof New York, Buffalo, NY6Indiana University School of Medicine, Indian-apolis, IN7Barbara Davis Center for Childhood Diabetes,Aurora, CO8Harold Schnitzer Diabetes Health Center,Oregon Health & Science University, Portland,OR9University of Florida, Gainesville, FL10University of Minnesota, Minneapolis, MN11Northwest Lipid Metabolism and DiabetesResearch Laboratories, Seattle, WA12JSS Medical Research, St. Laurent, QC, Canada

Corresponding author: Nicole C. Foster, t1dstats@jaeb.org.

Received 9 July 2015 and accepted 20 October2015.

Clinical trial reg. no. NCT01994746, clinicaltrials.gov.

This article contains Supplementary Data onlineat http://care.diabetesjournals.org/lookup/suppl/doi:10.2337/dc15-1498/-/DC1.

A slide set summarizing this article is availableonline.

*A complete list of participating investigators andcoordinators is provided in the SupplementaryData.

© 2016 by the American Diabetes Association.Readersmayuse this article as longas thework isproperly cited, the use is educational and not forprofit, and the work is not altered.

Michael R. Rickels,1 Katrina J. Ruedy,2

Nicole C. Foster,2 Claude A. Piche,3

Helene Dulude,3 Jennifer L. Sherr,4

William V. Tamborlane,4

Kathleen E. Bethin,5 Linda A. DiMeglio,6

R. Paul Wadwa,7 Andrew J. Ahmann,8

Michael J. Haller,9 Brandon M. Nathan,10

Santica M. Marcovina,11

Emmanouil Rampakakis,12 LinyanMeng,12

and Roy W. Beck,2 for the T1D Exchange

Intranasal Glucagon Investigators*

264 Diabetes Care Volume 39, February 2016

EPIDEM

IOLO

GY/HEA

LTHSERVICES

RESEA

RCH

oral carbohydrate administration (2,3).However, further reductions in bloodglucose can lead to hypoglycemia-induced cognitive dysfunction, whichmay make it impossible for an individualto take appropriate corrective measuresto restore normal blood glucose (4). Inthe most severe hypoglycemic events,administration of oral carbohydrateby a second party is precluded becausethe patient is combative, unconscious,or having a seizure. Such severe hypogly-cemia, which is not uncommon, can beassociated with significant morbidity (5)and mortality (6). In the T1D Exchangeregistry, about 1 in 10 participants report-ed hypoglycemia-related seizure or loss ofconsciousness in the prior year, with aneven higher annual frequency (1 in 5) forthose with .40 years’ diabetes duration(7). Other studies have reported similarfrequencies (8,9).Treatment of severe hypoglycemia

should be administered when possibleby bystanders while awaiting emer-gency medical services and currently in-volves parenteral administration ofglucagon. Glucagon, the principal coun-terregulatory hormone to insulin, is in-hibited by insulin and fails to activate inresponse to hypoglycemia in patientswith type 1 diabetes (10,11). Currentcommercially available glucagon prepa-rations are unstable in solution andavailable only as a lyophilized powderthat must be reconstituted in diluentimmediately prior to injection. Thus,glucagon is often not available whenneeded and, even when available, maybe improperly administered owing tothe complexity of the preparation andadministration processes (12). Thus, al-ternative delivery systems for the ad-ministration of glucagon for severehypoglycemia are needed.More than 20 years ago, several small

studies demonstrated the ability of in-tranasally administered glucagon solu-tion or powder to increase plasmaglucose levels in healthy volunteers(13) in a dose-dependent fashion (14)and to correct insulin-induced hypogly-cemia in patients with type 1 diabetes(14–17). None of these studies led tocommercialization of a nonparenteralglucagon preparation. In the currentstudy, we evaluated the efficacy andsafety of a novel intranasal glucagon de-livery system in comparison with stan-dard intramuscular glucagon for the

treatment of insulin-induced hypoglyce-mia in adults with type 1 diabetes.

RESEARCH DESIGN AND METHODS

The study was conducted at eight diabe-tes clinics within the T1D ExchangeClinic Network (18). Participants in-cluded men and women aged 18–64years with type 1 diabetes of at least 2years’ duration and weighing $50 kgwith BMI 20–35 kg/m2. Exclusion crite-ria included history of a severe hypogly-cemic episode in the month prior toenrollment, pheochromocytoma, or in-sulinoma; history of seizure disorder;cardiovascular, liver, or kidney disease;use of a b-blocker; or consumption ofthree or more alcoholic beverages daily.Additional protocol details are availableat ClinicalTrials.gov (NCT01994746). TheInstitutional Review Board of each par-ticipating institution approved the studyprotocol, and all participants providedwritten informed consent.

Enrolled participants underwent twoglucagon dosing visits, scheduled 1–4weeks apart. Participants were ran-domly assigned to receive either 3 mgintranasal glucagon (Locemia Solutions)or 1 mg intramuscular glucagon (Gluca-Gen HypoKit [Novo Nordisk]) (19) duringthe first dosing visit, with the other glu-cagon preparation administered duringthe second dosing visit in a crossoverfashion using a block design stratifiedby clinic; the randomization sequencewas generated by the coordinating cen-ter and revealed to clinic center staffusing a central study Web site upon en-rollment of each participant. The intra-nasal glucagon formulation consisted of3 mg glucagon, the phospholipid dode-cylphosphocholine as an absorption en-hancer, and b-cyclodextrin as a bulkingagent in a total dosemass of 30mg pow-der, contained within a single-use one-step dispensing device. The tip of thedevice was inserted in one nostril, andthe dose was delivered by simplydepressing a plunger connected to a pis-ton that gently discharges the powderinto the nostril (Supplementary Fig. 1).No inhalation or other cooperative mea-sure is required from the patient, as ab-sorption takes place through the nasalmucosa. An earlier phase I study showednasal congestion,with orwithout concom-itant use of a decongestant, did not ad-versely affect glucagon pharmacokineticsor the glycemic response in otherwise

healthy subjects given the 3-mg dose dur-ing and after recovery from a commoncold (C.A. Piche, written communication).A phase II study established a dose re-sponse for intranasal glucagon thatattained a maximal increase in blood glu-cose with the 3-mg dose, presumably dueto saturable absorption across the nasalmucosa (20).

Eachglucagondosingvisitwasconductedafter an overnight fast of at least 8 h witha starting plasma glucose $90 mg/dL.Hypoglycemia was induced by an infu-sion of insulin initially at 2 mU/kg/mini.v. that could be increased to 3 mU/kg/min or decreased to 1.5 or 1 mU/kg/minas necessary to achieve a controlled de-cline in plasma glucose. For participantsusing an insulin pump, the continuoussubcutaneous insulin infusion was sus-pended during the procedure, while par-ticipants using multiple daily injections ofinsulin took their last long-acting insulindose the day before testing. Plasma glu-cose concentrationsweremeasuredevery5–10 min after the start of the insulin in-fusion using a U.S. Food and Drug Admin-istration–approved glucose analyzer (YSI2300, Yellow Springs Instruments, YellowSprings, OH; Beckman, Beckman Coulter,Brea, CA; Analox GM9, Analox Instru-ments, Lunenburg, MA; or HemoCue,HemoCueWorldwide, Angelholm, Sweden).Once the glucose concentration was,60mg/dL, the insulin infusion was stopped,and 5 min later blood samples for glucoseandglucagonwere collected andglucagonwas administered at t = 0. Glucagon wasdeliveredwith the subject lying in a lateralrecumbent position either in the deltoidmuscle of the nondominant arm for theintramuscular administration or nare ofthe same side for the intranasal adminis-tration. Serial blood samples for glucoseand glucagon were collected at t = 5, 10,15, 20, 25, 30, 40, 50, 60, and 90 min. In-sulin was measured at t = 0, 30, and60min. Hypoglycemia symptomswere as-sessed by the Edinburgh HypoglycemiaScale (21) at baseline, once the glucosewas ,75 and ,60 mg/dL, and at t = 15,30, 45, and 60 min after the administra-tion of glucagon. Nasal and nonnasalsymptoms were ascertained at baselineand at t = 15, 30, 60, and 90 min afterglucagon administration.

Biochemical AnalysisAll samples for analysis of glucose, glu-cagon, and insulin were collected on ice

care.diabetesjournals.org Rickels and Associates 265

into tubes containing EDTAwith the pro-tease inhibitor aprotinin immediatelyadded, centrifuged at 48C, separated,and frozen at 2808C until subsequentanalysis. Plasma glucose concentrationswere measured by the glucose hexoki-nasemethod using an automated glucoseanalyzer (Roche Module P; Roche Diag-nostics, Indianapolis, IN), insulin by twosite immune-enzymometric assay using aTosoh 2000 auto-analyzer (Tosoh Biosci-ences, San Francisco, CA), and glucagonby a commercially available radioimmu-noassay (Millipore, Billerica, MA) at theNorthwest Lipid Research Laboratories(University of Washington, Seattle, WA).

Statistical AnalysisTreatment success was defined as an in-crease in plasma glucose to $70 mg/dLor an increase of $20 mg/dL from theglucose nadir within 30 min after receiv-ing study glucagon, without receivingadditional actions to increase theplasma glucose level. This success crite-ria accounts for the treatment goal toeither normalize glycemia or raise theglucose to a level above which cognitivefunction would allow further treatmentwith oral carbohydrate. The sample sizewas computed to be 75 individuals com-pleting both dosing visits under the fol-lowing assumptions: success rate of 95%in each treatment arm, noninferioritylimit of 10%, one-sided a of 0.025, and80% power. It was preplanned to re-place and not include in efficacy analy-ses participants who did not have twovalid dosing visits, defined as receipt ofstudy glucagon (intramuscular or intra-nasal) and no rescue treatment for se-vere hypoglycemia prior to or within thefirst 10 min after glucagon administra-tion. Safety analyses included all dosingvisits in which glucagon was received.All reported glucose and glucagon con-

centrations are from the central labora-tory, except for 3 instances out of 1,650for which the central lab glucose mea-surement was missing and a local glucosemeasurement was used. Nadir glucoseconcentration was defined as the mini-mum glucose measurement within 10minutes after glucagon administration. Atreatment comparison of nadir glucoseconcentration was completed using a ttest.The proportion of successes in each

treatment arm and the difference inproportions were computed, and a

one-sided 97.5% CI was obtained fromthe one-sample mean of the paired dif-ferences in outcomes (outcome ob-served, 1; outcome not observed, 0)across dosing visits. The difference inproportion of successes between treat-ment arms and one-sided 97.5% CI alsowas calculated using a Poisson regres-sion model incorporating a generalizedestimating equation with adjustmentsfor nadir glucose and treatment period(first treatment for participant versussecond treatment for participant).

Summary statistics for plasma glucoseconcentrations at each time point acrossthe dosing visit were calculated with im-putation for missing glucose values andglucose values after receipt of interven-tion treatment using the Rubin methodbased on available glucose measure-ments and treatment arm. A treatmentcomparison of the blood glucose con-centration over the 90 min after admin-istration of glucagon was performedusing a linear mixed model with re-peated measures.

A treatment arm comparison of thetime from administration of glucagon tooccurrence of success was performed forthe subset of dosing visits with nadir glu-cose concentration ,50 mg/dL by con-structing Kaplan-Meier curves and byusing the marginal Cox proportional haz-ards model for clustered data adjustingfor nadir blood glucose and treatmentperiod. If a participant received additionalintervention treatment to raise the glu-cose concentration prior to success, theremaining time points were consideredfailures. Missing glucose values were im-puted using the Rubin method based onavailable lab glucose measurements andtreatment arm.

Peak glucagon concentrations werecompared between treatment groupsusing an ANOVA model with participantincluded as a random effect. A treat-ment comparison of time to peak gluca-gon concentration was performed usinga Wilcoxon signed rank test. Three glu-cagon values were removed from theanalysis owing to suspected sample col-lection error; these glucagon valueswere imputed using the Rubin methodbased on available glucagon measure-ments and treatment arm.

A treatment comparison of the Edin-burgh Hypoglycemia Scale score at eachtime point after glucagon administra-tion was completed using linear mixed

models with repeated measures adjustingfor the treatment period and score atvisit arrival. Similarly, a treatmentcomparison of the occurrence of nau-sea, with or without vomiting, wascompleted using a generalized linearmixed model with random participanteffect.

Results are expressed as mean 6 SDor median (interquartile range). Dataanalyses were performed using SASsoftware, version 9.4 (SAS Institute,Cary, NC).

RESULTS

Between December 2013 and May2014, 75 participants successfully com-pleted both dosing visits and were in-cluded in the efficacy analyses, and 77completed at least one dosing visit andwere included in the safety analyses(Supplementary Fig. 1). Among the 77included in the analyses, mean 6 SDage was 33 6 12 years and BMI 26 6 4kg/m2 and 45 (58%) were female and74 (96%) non-Hispanic white. Medianduration of type 1 diabetes was 18(interquartile range 9–25) years; 57(74%) were using an insulin pump andthe rest multiple daily injections of in-sulin. Mean HbA1c was 8.3 6 1.8%(Supplementary Table 1). Among the75 who successfully completed bothdosing visits, 72 (96%) were non-Hispanic white.

Median time between the two dosingvisits was 16 (interquartile range 14–22)days. The median total amount of in-sulin given to induce hypoglycemiawas 0.09 (0.07–0.13) units/kg for theintranasal glucagon visits and 0.10(0.08–0.14) units/kg for the intramus-cular glucagon visits. Plasma insulinlevels 0, 30, and 60 min after glucagonadministration were similar betweenarms (Table 1).

At the time of insulin infusion discon-tinuation, mean6 SD local glucose con-centrationwas 556 5 and 556 4mg/dLfor the intranasal and intramusculardosing visits, respectively. Mean centrallaboratory glucose concentrations were48 6 8 mg/dL and 49 6 8 mg/dL, re-spectively, 5 min after stopping insulinimmediately prior to glucagon adminis-tration (t = 0) and 44 6 8 mg/dL and476 8 mg/dL, respectively, at the nadir(Table 1). The nadir glucose concentra-tion occurred at t = 0 for most intramus-cular glucagon visits (49 of 75 visits

266 Intranasal Glucagon and Hypoglycemia Diabetes Care Volume 39, February 2016

[65%]) and at t = 5 for most intranasalvisits (59 of 75 visits [79%]).Success criteria weremet on 74 of the

75 (98.7%) intranasal glucagon visits andall 75 (100%) of the intramuscular glu-cagon visits (unadjusted difference1.3%, one-sided 97.5% CI 4.0%; adjusteddifference 1.5%, one-sided 97.5% CI4.3%). The one intranasal dosing visitnot meeting success criteria had a nadir

glucose concentration of 47 mg/dLwith a rise by 30 min to 65 mg/dL andby 40 min to 72 mg/dL without anyother intervention. As can be seen inFig. 1A, the rise in glucose concentra-tions after intranasal glucagon laggedbehind the rise after intramuscular glu-cagon by ;5 min (P , 0.001).

Within 5 min of administration of bothproducts, glucagon levels had reached

pharmacological levels indicating rapidabsorption via both routes of administra-tion. Initially, glucagon concentrations af-ter intranasal administration were lowerthan concentrations after intramuscularwith the mean6 SD peak 3,1556 1,956and 3,6726 1,726 pg/mL (P = 0.003) andmedian (minimum, maximum) time topeak 20 min (10, 90) and 15 min (5, 60)(P , 0.001), respectively, but were notdifferent after 20 min (Fig. 1B).

Examination of only dosing visitswhere the nadir glucose was ,50 mg/dL(Fig. 2) shows that time to success afterintranasal glucagon lagged slightly be-hind time to success after intramuscularglucagon (mean times to reach 70mg/dLor 20 mg/dL increase were 16 min and13 min with intranasal glucagon andintramuscular glucagon, respectively;P , 0.001).

Symptoms of hypoglycemia, as repre-sented by Edinburgh Hypoglycemia Scalescores, were greater in the intranasalgroup compared with the intramusculargroup for the first 45min after administra-tion but similar thereafter (SupplementaryFig. 2).

Transient head or facial discomfort wasreported after 19 (25%) intranasal gluca-gon administrations and after 7 (9%) in-tramuscular administrations. Vomitingoccurred after 13 (17%) intranasal and 9(12%) intramuscular administrations; nau-sea without vomiting occurred during anadditional 14 (18%) intranasal and20 (26%)intramuscular administrations (P = 0.59)(nausea, with or without vomiting, afterintranasal glucagon administration [36%]vs. intramuscular administration [38%])(Table 2).

CONCLUSIONS

In the study, intranasal glucagon consis-tently corrected insulin-induced hypogly-cemia in adults with type 1 diabetes,meeting the predefined definition of non-inferiority to intramuscular injection ofglucagon. In the one case in which intra-nasal glucagon did not meet the study-defined success criteria, hypoglycemiawas corrected without any additional in-tervention but after the time frame forstudy-defined success. Average glucoseconcentrations and time to meet the pri-mary end point after intranasal glucagonlagged;3 min behind glucose concentra-tions after intramuscular glucagon, consis-tent with the glucagon concentrationsshowing a relative delay in achievement

Table 1—Study protocol characteristics by treatment arm

Intranasal(N = 75)

Intramuscular(N = 75)

Central lab glucose concentration immediately prior toglucagon administration (mg/dL)20 to ,30 1 (1) 1 (1)30 to ,40 9 (12) 7 (9)40 to ,50 28 (37) 31 (41)50 to ,60 34 (45) 28 (37)60 to ,70 3 (4) 8 (11)$70 0 0Mean 6 SD 48 6 8 49 6 8Difference (intramuscular – intranasal),

mean 6 SD 1 6 9

Central lab glucose at nadir (mg/dL)a

20 to ,30 3 (4) 2 (3)30 to ,40 17 (23) 11 (15)40 to ,50 38 (51) 31 (41)50 to ,60 15 (20) 28 (37)60 to ,70 2 (3) 3 (4)$70 0 0Mean 6 SD 44 6 8 47 6 8Difference (intramuscular – intranasal),

mean 6 SD 3 6 9

Insulin infusionLocal glucose concentration when insulin infusion

initiated (mg/dL), mean 6 SD 161 6 48 165 6 54Local glucose concentration when insulin infusiondiscontinued (mg/dL), mean 6 SD 55 6 5 55 6 4

Duration of insulin infusion (min), median(25th–75th percentile) 54 (41–72) 60 (42–80)

Amount of insulin received prior to discontinuation(units/kg), median (25th–75th percentile)b 0.09 (0.07–0.13) 0.10 (0.08–0.14)

Lab insulin concentration (mU/mL), median(25th–75th percentile)

Prior to glucagon administrationc 55.3 (35.6–72.2) 53.5 (36.5–69.6)30-min postglucagon administrationd 11.8 (7.0–22.5) 12.9 (7.5–17.8)60-min postglucagon administratione 6.5 (4.3–14.1) 7.1 (4.9–10.9)

Data are n (%) unless otherwise indicated. aMinimum central lab glucose concentrationmeasured after glucagon administration; nadir occurred at time 0 for 13 (17%) participantsreceiving intranasal glucagon and 49 (65%) participants receiving intramuscular glucagon, attime 5 min for 59 (79%) participants receiving intranasal glucagon and 25 (33%) participantsreceiving intramuscular glucagon, and at time 10 min for 3 (4%) participants receiving intranasalglucagon and 1 (1%) participant receiving intramuscular glucagon; P = 0.002 from a t test for thetreatment comparison of nadir glucose. bTotal amount of insulin received during the insulininfusion, including the priming dose of insulin (if a priming dosewas given); 14 (19%) participantsin the intranasal glucagon arm received a priming dose of insulin (mean 6 SD priming insulindose 0.03 6 0.01 units/kg), and 16 (21%) participants in the intramuscular glucagon armreceived a priming dose of insulin (mean6 SD priming insulin dose 0.046 0.02 units/kg). cSix of160 measurements were removed from the analysis owing to potential sample collection error(3 intranasal glucagon dosing visits and 3 intramuscular glucagon dosing visits). dMissing for 1intramuscular glucagon dosing visit. eFour of 160 values were not included in the analysis owingto potential sample collection error (1 intranasal glucagon dosing visit and 3 intramuscularglucagon dosing visits).

care.diabetesjournals.org Rickels and Associates 267

of peak glucagon levels with intranasalglucagon of ;5 min. Nevertheless, phar-macologic levels of glucagonwere presentby 5 min after administration by eitherapproach. The slight delay in glycemic re-sponse would likely be clinically inconse-quential and in many circumstancesmight be offset by the time required,errors, and failures to deliver amongnonmedical caregivers in preparingand administering the injectable intra-muscular formulation. Indeed, a recenthuman factors study simulating treat-ment of severe hypoglycemia showsnasal delivery of glucagon is muchfaster with a higher success rate than

injection for trained caregivers ofinsulin-using persons (16 s vs. 1.9 minfor time to administer, 94% vs. 50% fordelivery) and for untrained acquain-tances (26 s vs. 2.4 min, 93% vs. 20%for delivery) for intranasal and intra-muscular dosings, respectively (22).

Several small studies have docu-mented the ability of intranasallysprayed glucagon to increase plasmaglucose levels in healthy volunteers(13) in a dose-dependent fashion (14)and to correct insulin-induced hypo-glycemia in patients with type 1 diabe-tes (14–16). In the case of glucagonadministration for the treatment of

severe hypoglycemia, impaired or ab-sent consciousness would prevent pa-tient inhalation of sprayed solution toassist with absorption. Another smallstudy tested the response to insulin-induced hypoglycemia in patients withtype 1 diabetes using a dry powder for-mulation containing either 1 or 2 mgglucagon and found similar pharmaco-dynamics with the 2-mg dose of gluca-gon compared with 1 mg glucagonadministered intramuscularly (17). Inthese prior reports, the glucose-raisingeffect of intranasal glucagon was simi-lar in onset but of shorter durationcompared with intramuscular admin-istration (13,15,17). In our study, 3mg intranasal glucagon exhibitedequivalent durability of the glucose-raising effect as with intramuscularglucagon for the 90-min postdosingobservation period.

In the current study, there was a delayin resolution of hypoglycemia symp-toms at the intranasal compared withintramuscular dosing visit; however,the magnitude of hypoglycemia symp-toms was not different between visits,likely because of the similar nadir glu-cose levels achieved. While we did notmeasure plasma epinephrine, given thecomparable degree of autonomic symp-tom generation and the pharmacologiclevels of glucagon present by 5 min un-der both dosing conditions, we do notbelieve other counterregulatory hor-mones played a significant role in thisstudy. As expected, transient nasalsymptoms and head or facial discomfortoccurred more frequently with the

Figure 1—Plasma glucose (A) and glucagon (B) concentrations over time according to treatment visit. Data are mean6 SD, with the solid black barsand line representing the intranasal visit and the white bars and black dotted line representing the intramuscular visit.

Figure 2—Time to plasma glucose concentration $70 mg/dL or an increase $20 mg/dL fromnadir concentration in participants with nadir glucose ,50 mg/dL. N: intranasal 58 (77%) andintramuscular 44 (59%).

268 Intranasal Glucagon and Hypoglycemia Diabetes Care Volume 39, February 2016

intranasal preparation. However, in thesetting in which intranasal glucagon isgiven to an individual with severe cog-nitive impairment and often lossof consciousness, this will not be rele-vant. Nausea and vomiting, well-knownconsequences of glucagon administra-tion owing to relaxation of gastrointes-tinal smooth muscle (23), occurred atsimilar frequencies with both glucagonpreparations.

This study had certain limitations. Theglucagon dosing administration wasnot blinded. Although this could haveinfluenced investigator handling of theinsulin infusion protocol to inducehypoglycemia, there was no evidenceof bias, since the actual amounts of in-sulin administered, as well as the preglu-cagon dosing insulin and glucose levels,were nearly identical between thestudy visits. The study lacked a sham

treatment condition that would havecontrolled for possible spontaneousrecovery from hypoglycemia. How-ever, hospital policies preclude nottreating hypoglycemia ,40 mg/dL,and with this degree of hypoglycemiaexperienced during 27% and 18% ofthe intranasal and intramuscular glu-cagon dosing visits, respectively, avery high rate of rescue treatmentwould have been required withoutan intervention. It was also not ethicalto reduce glucose levels to the point ofseizure or loss of consciousness or con-tinue insulin administration oncereaching moderate hypoglycemia. Nev-ertheless, the degree of glucose eleva-tion with intranasal was similar to thatof the intramuscular and should bemore than sufficient for the individualto achieve a glucose level where con-sciousness is restored and further oraltreatment of hypoglycemia can be ad-ministered. In the study, glucagon wasadministered by trained health careprofessionals under nonemergencyconditions; whether similar results canbe obtained with either glucagon prep-aration in an outpatient setting remainsto be determined.

In conclusion, this study has demon-strated that intranasal glucagon is effec-tive for the correction of insulin-inducedhypoglycemia in adults with type 1 diabe-tes. Although hypoglycemia was inducedin a controlled setting by administeringinsulin, this should approximate the real-world setting of severe hypoglycemia thatoccurs owing to excessive therapeutic in-sulin availability relative to the impairedor absent endogenous glucagon re-sponse (10) that normally works to limithypoglycemia (11). A needle-free intra-nasal preparation is likely to be pre-ferred due to its relative simplicitycompared with injectable formulations,particularly since individuals who mustadminister rescue glucagon are usuallynot trained medical professionals. Inthe current injectable form, administra-tion of glucagon is a relatively complexprocess and, outside the home, is oftenunavailable. This leads to suboptimaluse of an otherwise effective medica-tion, unnecessary delays in treatment,and costly use of emergency medicalsystems including ambulance services,emergency room visits, and hospitaladmissions. Given the results of thisstudy, intranasal glucagon delivery appears

Table 2—Adverse events by treatment visita

Adverse event group Intranasal (N = 77) Intramuscular (N = 76)

GIAbdominal discomfort 1 (1) 1 (1)Nausea 17 (22) 21 (28)Vomiting 13 (17) 9 (12)Total ($1 GI events)b 28 (36) 29 (38)

Head discomfortEar pain 2 (3) 1 (1)Eye pain 1 (1) 0Facial pain 2 (3) 0Headache 18 (23) 7 (9)Neck pain 1 (1) 0Total ($1 head pain events)b 19 (25) 7 (9)

NasalNasal congestion 6 (8) 1 (1)Nasal discomfort 8 (10) 0Nasal edema 1 (1) 0Rhinorrhoea 2 (3) 1 (1)Total ($1 nasal events)b 14 (18) 1 (1)

Weakness/fatigueFatigue 6 (8) 5 (7)Lethargy 0 1 (1)Muscular weakness 2 (3) 0Total ($1 weakness/fatigue events)b 8 (10) 6 (8)

OcularEye pruritus 2 (3) 1 (1)Lacrimation increased 7 (9) 1 (1)Ocular hyperemia 1 (1) 0Total ($1 ocular events)b 7 (9) 1 (1)

CognitiveConfusional state 1 (1) 0Disturbance in attention 1 (1) 0Somnolence 2 (3) 0Total ($1 cognitive events)b 2 (3) 0

AutonomicHot flush 0 1 (1)Hyperhidrosis 1 (1) 1 (1)Total ($1 autonomic events)b 1 (1) 2 (3)

ThroatCough 1 (1) 0Upper-airway cough syndrome 1 (1) 0Total ($1 throat events)b 2 (3) 0

Skin: pruritus 2 (3) 1 (1)

Hyperglycemia 1 (1) 0

Psychological: anxiety 1 (1) 0

Data are n (%). No serious adverse events were reported. GI, gastrointestinal. aForty-four(57%) of 77 participants receiving intranasal glucagon experienced at least 1 adverse event;35 (46%) of 76 participants receiving intramuscular glucagon experienced at least 1adverse event. bTotal number (%) of participants with at least 1 occurrence of the adverse eventgroup.

care.diabetesjournals.org Rickels and Associates 269

promising, and when commercially avail-able, it can be expected to have a substan-tial beneficial impact on the treatment ofsevere hypoglycemia.

Funding and Duality of Interest. Funding wasprovided by a grant from the Leona M. and HarryB. Helmsley Charitable Trust; by the NationalCenter for Research Resources and the NationalCenter for Advancing Translational Sciences, Na-tional Institutes of Health, through grantsUL1TR000003 (University of Pennsylvania),UL1TR000064 (University of Florida), andUL1TR001108 (Indiana University); and byLocemia Solutions ULC. Studies at the BarbaraDavis Center for Childhood Diabetes were per-formed in their infusion center. Locemia Solutionsprovided the intranasal glucagon product. M.R.R.has received consultancy payments from Longev-ity Biotech, Janssen Research & Development,and Semma Therapeutics and research supportfrom Merck. C.A.P. is an employee and memberof the board for Locemia Solutions, one of theentities that provided financial support for theconduct of this study. H.D. is an employee forLocemia Solutions. J.L.S.’s nonprofit employerhas received grants for support of supplies forinvestigator-initiated studies from Medtronic.W.V.T.’s nonprofit employer has received con-sulting fees from Novo Nordisk and LocemiaSolutions (Locemia consultancy for protocol de-velopment). R.P.W. reports receiving researchsupport from Novo Nordisk and serves as a con-sultant for Novo Nordisk and Medtronic. E.R. andL.M. received payment for pharmacokinetic/pharmacodynamic analysis. R.W.B.’s nonprofitemployer has received consultant payments onhis behalf from Sanofi and Animas and a researchgrant from Novo Nordisk, with no personal com-pensation to R.W.B. No other potential conflictsof interest relevant to this article were reported.Author Contributions. M.R.R. and K.J.R. re-searched data and wrote and edited the man-uscript. N.C.F., E.R., and L.M. researched data,performed statistical analysis, and wrote andedited the manuscript. C.A.P., H.D., J.L.S., W.V.T.,K.E.B.,L.A.D.,R.P.W.,A.J.A.,M.J.H.,B.M.N.,S.M.M.,and R.W.B. researched data and reviewed andedited the manuscript. N.C.F. and R.W.B. are theguarantorsof thisworkand, as such, had full accessto all the data in the study and take responsibilityfor the integrity of the data and the accuracy of thedata analysis.

References1. Cryer PE. The barrier of hypoglycemia in di-abetes. Diabetes 2008;57:3169–3176

2. Brodows RG, Williams C, Amatruda JM.Treatment of insulin reactions in diabetics.JAMA 1984;252:3378–33813. Slama G, Traynard PY, Desplanque N, et al.The search for an optimized treatment of hypo-glycemia. Carbohydrates in tablets, solutin, orgel for the correction of insulin reactions. ArchIntern Med 1990;150:589–5934. Gold AE, MacLeod KM, Deary IJ, Frier BM.Hypoglycemia-induced cognitive dysfunctionin diabetes mellitus: effect of hypoglycemia un-awareness. Physiol Behav 1995;58:501–5115. Whitmer RA, Karter AJ, Yaffe K, QuesenberryCP Jr, Selby JV. Hypoglycemic episodes and riskof dementia in older patients with type 2 diabe-tes mellitus. JAMA 2009;301:1565–15726. McCoy RG, Van Houten HK, Ziegenfuss JY,Shah ND, Wermers RA, Smith SA. Increasedmortality of patients with diabetes reportingsevere hypoglycemia. Diabetes Care 2012;35:1897–19017. Weinstock RS, Xing D, Maahs DM, et al.; T1DExchange Clinic Network. Severe hypoglycemiaand diabetic ketoacidosis in adults with type 1diabetes: results from the T1D Exchange clinicregistry. J Clin Endocrinol Metab 2013;98:3411–34198. Cariou B, Lievre M, Huet D, Charbonnel B,Sert C, Gouet D. Hypoglycemia among 3048insulin-treated patients in real life: frequencyand predictive factors: results from the prospec-tive DIALOG study. Presented at European As-sociation for the Study of Diabetes (Abstract).Barcelona, Spain, EASD, 2013. Available fromhttp://www.easdvirtualmeeting.org/resources/hypoglycaemia-among-3048-insulin-treated-patients-in-real-life-frequency-and-predictive-factors-results-from-the-prospective-dialog-study–2. Accessed 3 December 20159. Khunti K, Alsifri S, Aronson R, et al. Self-reported hypoglycemia: A global study of 24countries with 27,585 insulin-treated patientswith diabetes: the HAT study. Diabetes ResClin Pract 2014;106:S105–S10610. Gerich JE, Langlois M, Noacco C, Karam JH,Forsham PH. Lack of glucagon response to hy-poglycemia in diabetes: evidence for an intrinsicpancreatic alpha cell defect. Science 1973;182:171–17311. Holstein A, Hammer C, Plaschke A, et al.Hormonal counterregulation during severe hy-poglycaemia under everyday conditions in pa-tients with type 1 and insulin-treated type 2diabetes mellitus. Exp Clin Endocrinol Diabetes2004;112:429–43412. HarrismG,Diment A, SulwayM,WilkinsonM.Glucagon administration - underevaluated andundertaught. Pract Diabetes Int 2001;18:22–2513. Pontiroli AE, AlberettoM, Pozza G.Metaboliceffects of intranasally administered glucagon:

comparison with intramuscular and intravenousinjection. Acta Diabetol Lat 1985;22:103–11014. Freychet L, Rizkalla SW, Desplanque N, et al.Effect of intranasal glucagon on blood glucoselevels in healthy subjects and hypoglycaemicpatients with insulin-dependent diabetes. Lan-cet 1988;1:1364–136615. Pontiroli AE, Calderara A, Pajetta E,Alberetto M, Pozza G. Intranasal glucagon asremedy for hypoglycemia. Studies in healthysubjects and type I diabetic patients. DiabetesCare 1989;12:604–60816. Slama G, Alamowitch C, Desplanque N,Letanoux M, Zirinis P. A new non-invasivemethod for treating insulin-reaction: intranasallyophylized glucagon. Diabetologia 1990;33:671–67417. Rosenfalck AM, Bendtson I, Jørgensen S,Binder C. Nasal glucagon in the treatment ofhypoglycaemia in type 1 (insulin-dependent) di-abetic patients. Diabetes Res Clin Pract 1992;17:43–5018. Beck RW, Tamborlane WV, Bergenstal RM,Miller KM, DuBose SN, Hall CA; T1D ExchangeClinic Network. The T1D Exchange clinic regis-try. J Clin Endocrinol Metab 2012;97:4383–438919. Novo Nordisk. GlucaGen prescribing in-formation [Internet]. Available from http://www.novonordiskmedicalinformation.com//file_upload/GlucaGen%20HypoKit%20Prescribing%20Information,April%202014%20.pdf. Accessed9 September 201520. Locemia Solutions ULC. Safety and efficacyof a novel glucagon formulation in type 1 diabeticpatients following insulin-induced hypoglycemia(AMG102) [Internet]. Available from https://clinicaltrials.gov/ct2/show/study/NCT01556594?term=amg+medical&rank=1. Accessed 23 June201521. Hepburn DA, Deary IJ, Frier BM, Patrick AW,Quinn JD, Fisher BM. Symptoms of acute insulin-induced hypoglycemia in humans with andwithout IDDM. Factor-analysis approach. Diabe-tes Care 1991;14:949–95722. Yale JF, Piche C, LafontaineM, et al. Needle-free nasal delivery of glucagon is superior toinjectable delivery in simulated hypoglycaemiarescue. Poster presented at European Associa-tion for the Study of Diabetes (Abstract). Stock-holm, Sweden, September 2015. Available fromhttp://www.easdvirtualmeeting.org/resources/needle-free-nasal-delivery-of-glucagon-is-superior-to-injectable-delivery-in-simulated-hypoglycaemia-rescue–3. Accessed 3 December 201523. Pacchioni M, Orena C, Panizza P, Cucchi E,Del Maschio A, Pontiroli AE. The hypotonic ef-fect of intranasal and intravenous glucagon ingastrointestinal radiology. Abdom Imaging1995;20:44–46

270 Intranasal Glucagon and Hypoglycemia Diabetes Care Volume 39, February 2016