Effects of Lixisenatide VersusLiraglutide (Short- and Long-Acting GLP-1 Receptor Agonists)on Esophageal and GastricFunction in Patients WithType 2 DiabetesDiabetes Care 2020;43:2137–2145 | https://doi.org/10.2337/dc20-0720

OBJECTIVE

Short-acting glucagon-like peptide 1 receptor agonists (GLP-1 RAs) decelerategastric emptyingmore than long-actingGLP-1RAs.Delayedgastric emptying is a riskfactor for gastroesophageal reflux disease.Weaimed tomeasure esophageal refluxand function as well as gastric emptying and acid secretion during treatment withshort-acting (lixisenatide) and long-acting (liraglutide) GLP-1 RAs.

RESEARCH DESIGN AND METHODS

A total of 57 subjectswith type2diabeteswere randomized to a10-week treatmentwith lixisenatide or liraglutide. Changes from baseline in the number of refluxepisodes during 24-h pH registration in the lower esophagus, lower esophagussphincter pressure, gastric emptying (13C-sodium octanoate acid breath test), andgastric acid secretion (13C-calcium carbonate breath test) were analyzed.

RESULTS

Gastric emptying half-timewas delayed by 52min (D 95% CI 16, 88) with lixisenatide(P 5 0.0065) and by 25 min (3, 46) with liraglutide (P 5 0.025). There was nodifference in the number of reflux episodes (mean6 SEM33.76 4.1 vs. 40.16 5.3for lixisenatide and liraglutide, respectively, P 5 0.17) or the extent of gastro-esophageal reflux (DeMeester score) (35.1 6 6.7 vs. 39.7 6 7.5, P 5 0.61), withsimilar results for the individualGLP-1RAs. No significant changes frombaseline inother parameters of esophagealmotility and lower esophageal sphincter functionwere observed. Gastric acidity decreased significantly by 220.7% (240.6, 20.8)(P 5 0.042) with the GLP-1 RAs.

CONCLUSIONS

Lixisenatide exerted a more pronounced influence on gastric emptying afterbreakfast than liraglutide. Neither lixisenatide nor liraglutide had significant effectson esophageal reflux or motility. Gastric acid secretion appears to be slightlyreduced by GLP-1 RAs.

1Diabetes Division, Department of Medicine I, StJosef-Hospital Bochum, Ruhr-University Bochum,Bochum, Germany2Profil, Neuss, Germany

Corresponding author: Juris J.Meier, juris.meier@ruhr-uni-bochum.de

Received 2 April 2020 and accepted 9 June 2020

Clinical trial reg. no. NCT02231658, clinicaltrials.gov

This article contains supplementary material on-line at https://doi.org/10.2337/figshare.12490601.

© 2020 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. More infor-mation is availableathttps://www.diabetesjournals.org/content/license.

Daniel R. Quast,1 Nina Schenker,1

Björn A. Menge,1 Michael A. Nauck,1

Christoph Kapitza,2 and Juris J. Meier1

Diabetes Care Volume 43, September 2020 2137

EMER

GINGTH

ERAPIES:

DRUGSANDREG

IMEN

S

https://doi.org/10.2337/dc20-0720http://crossmark.crossref.org/dialog/?doi=10.2337/dc20-0720&domain=pdf&date_stamp=2020-08-06mailto:juris.meier@ruhr-uni-bochum.demailto:juris.meier@ruhr-uni-bochum.dehttp://www.clinicaltrials.govhttp://www.clinicaltrials.govhttps://doi.org/10.2337/figshare.12490601https://www.diabetesjournals.org/content/licensehttps://www.diabetesjournals.org/content/license

Glucagon-like peptide 1 receptor ago-nists (GLP-1 RAs) can be categorized intolong-acting (e.g., liraglutide, exenatideonce weekly, dulaglutide, albiglutide,semaglutide) and short-acting (e.g., ex-enatide b.i.d. [unretarded], lixisenatide)compounds (1). Both subclasses of GLP-1RAs differ markedly from each otherwith respect to their pharmacokineticand pharmacodynamic properties (2,3).Mainly, short-acting GLP-1 RAs lead toshort-lived peaks of drug concentrationsafter subcutaneous injection, the dura-tion of which are several hours, withensuing troughs with near-zero drugconcentrations (4). Long-acting GLP-1RAs provide circulating drug concentra-tions that are constantly elevated, withminor fluctuations over a 24-h or 7-dayperiod (5,6). Consequently, premeal ad-ministration of short-acting GLP-1 RAselicits a significant delay of gastric emp-tying even after prolonged treatmentperiods (weeks/months), thereby pri-marily reducing postprandial glucose el-evations (7). In contrast, long-acting GLP-1RAs initially delay gastric emptying, buttheir pharmacokinetic profile leads todesensitization for this effect over a pe-riod of weeks to months, known astachyphylaxis (8). Therefore, during lon-ger-term treatment, long-acting GLP-1RAs influence postprandial glucose risespredominantly by their effects on insulinand glucagon secretion and have greatereffects on fasting plasma glucose due togreater exposure during the night hours(7). Earlier studies with acute intrave-nous infusion of native human GLP-1have also indicated an inhibition of gas-tric acid secretion, but whether this isalso the case with GLP-1 RAs has not yetbeen determined (9).Delayed gastric emptying and gastric

distension are well-established risk fac-tors in the pathophysiology of gastro-esophageal reflux disease (GERD) (10).Thus, delayed gastric emptyingmay tem-porarily increase the intragastric contentof both solids and acidic liquids, therebyincreasing the likelihoodof reflux into thelower esophagus (11). Several studieshave reported higher frequencies ofGERD symptoms in people with type 2diabetes (12–14). So far, little informa-tion is available on the effects of GLP-1RAs on the incidence of GERD (15).Treatment with GLP-1 RAs may theoret-ically further increase the incidence ofgastroesophageal reflux in thesepatients

(16). However, the effect of GLP-1 RAtreatment on gastroesophageal refluxand esophageal motility has not yetbeen determined.

Therefore, in the current study theeffects of a short- anda long-actingGLP-1RA (lixisenatide and liraglutide, respec-tively) on gastroesophageal reflux, esoph-agealmotility, gastric emptying, and gastricacid secretion were compared over 10weeks of treatment.

RESEARCH DESIGN AND METHODS

Study DesignThe current study was a randomized,bicentric, investigator-blinded, parallel-group study in subjects with type 2 di-abetes. The study was performed at theDiabetes Division, St Josef-Hospital, andat Profil. Informed consent was obtainedfrom all participants. Patients were ran-domized using the program RANCODEProfessional 3.6 (idv Datenanalyse &Versuchsplanung, Gauting, Germany).The randomization and blinding processis described in detail in SupplementaryMaterial. The study was approved by theethics committee of Ruhr-University Bo-chum (registration no. 14-5177 FF), andinvestigations were carried out in accor-dance with the principles of the Decla-ration of Helsinki as revised in 2008.

Study PopulationThe study included male and femalepatients aged 18–70 years with a rangeof BMI of 18–40 kg/m2 (inclusive) whohad been diagnosedwith type 2 diabetesfor at least 3 months. Treatment with astable regimen of metformin and/or asulfonylurea or metformin plus thiazoli-dinedione for at least 1 month or treat-ment with a stable regimen of insulinwith or without additional oral antidia-betes drug treatment (metformin, sulfo-nylureas, thiazolidinediones) for at least1 month was obligatory.

Exclusion criteria included contraindi-cations (including known or suspectedhypersensitivity) to GLP-1 RAs, currentuse of GLP-1 RAs or inhibitors of dipep-tidyl-peptidase 4, intake of medicationthat may influence gastric acid secretionor gastrointestinal motility, decompen-sated glycemic control (HbA1c $10%),preexisting significant concomitant dis-eases (except those typically associatedwith type 2 diabetes, e.g., arterial hyper-tension, dyslipidemia), or prominent dia-betes complications affecting parameters

to be measured in the current study (in-cluding history of GERD or overt gastro-paresis), as judged by the investigator.Furthermore,clinicallyrelevant impairmentof hepatic, renal, or cardiac function asindicated by alanine aminotransferase, bil-irubin, and/or alkaline phosphatase greaterthan threefold the upper limit of normal atscreening, estimated glomerular filtrationrate ,60 mL/min/1.73 m2 (estimate ac-cording to the MDRD study equation) atscreening, clinically relevant electrocardio-gramfindings, or symptomsof heart failure(New York Heart Association class III–IV) asjudged by the investigator were defined asexclusion criteria. Smokers or participantsnot able or willing to refrain from smokingwere not included.

Safety AssessmentMonitoring of adverse events includedsymptomatic and severe hypoglycemicepisodes and gastrointestinal disordersusing a structured questionnaire thatwill be published separately. Laboratorymonitoring included pancreatic, renal,and hepatic parameters. Laboratory anal-ysis was performed at MLM MedicalLaboratories Moenchengladbach GmbHand the laboratory of St Josef-Hospital.

Study Procedures

Screening Visit

At a screening visit, participants had theirmedical history taken and underwent aclinical examination. In addition, the fol-lowing anthropometric variables were as-sessed: age, height, weight, heart rate,blood pressure, and waist and hip circum-ference. Venous blood samples weretaken in the fasting state for themeasure-ment of standard hematological and clin-ical chemistry variables.

Interventions

After screening, participants were ran-domized to receive either lixisenatide orliraglutide. Participants randomized tolixisenatide were administered 10 mgonce daily for 1 week, followed by a20-mg once-daily maintenance dose forthe remainder of the trial. Participantsrandomized to receive liraglutide wereadministered 0.6mg once daily forweekoneafter initiation and1.2mgoncedailyfor week two, followed by 1.8 mg oncedaily for the remaining period of thetrial. In both groups, the GLP-1 RA wasadministered 30 min before breakfast.Subjects were asked to attend the clin-ical unit in the morning in the fasting

2138 Reflux and GI Motility in GLP-1 RAs Diabetes Care Volume 43, September 2020

https://doi.org/10.2337/figshare.12490601https://doi.org/10.2337/figshare.12490601

condition (only still water was allowedafter 8:00 P.M. the evening before thevisit) for screening and all other exper-imental visits.

Experimental Visits

24-h Ambulatory Esophageal pHMeasurement.

Following a high-resolution esophageal ma-nometry (HRM), an esophageal monitoringtube for a 24-h ambulatory manometric pHmeasurement was inserted. This tubecontained a pH electrode that was placed5 cm above the upper border of the loweresophageal sphincter (LES), as previouslydetermined through HRM. Assessment ofgastroesophageal reflux was based on thenumber and duration of episodes withpH,4.0 in the lower third of the esophagusand the DeMeester score (17,18).HRM. HRM was performed using a solid-state catheter with 36 circumferentialpressure sensors spacedat1-cm intervals(Unisensor, Attikon, Switzerland). A se-riesof 10 swallowswith 5mL stillwater atroom temperature was registered permanometry. The topographical plots andmanometric parameters were analyzedusing the software ViMeDat (StandardInstruments GmbH, Karlsruhe, Germany),which analyzes the manometry resultsbased on the Chicago Classification ofesophageal disorders (19).GastricEmptyingofSolidMealComponents:13C-Sodium Octanoate Breath Test. The13C-sodiumoctanoate breath test is a non-invasive test tomeasure gastric emptying(20). Study medication was injected 1 hbefore the meal. After an overnight fastof ;12 h, baseline breath samples weretaken into gas-tight sampling bags. Sub-jectswere then served ameal with a totalcaloric amount of ;1,175.7 kJ with 24%carbohydrate, 16% protein, and 60% fat.This meal was composed of egg, whitebread, margarine, and still water. It in-cluded 91 mg 13C-octanoic acid, mixedinto scrambled eggs to label the meal,and had to be consumed within 10 min.Further breath samples were collectedevery 15 min after the test meal for 3 h,followed by every 30min for another 3 h.Gastric emptying and gastric emptyinghalf-timewereanalyzedusingtheWagner-Nelson method (16,21,22). While thestudy drugswere taken before breakfast,all other morning medication was takenafter completion of the gastric emptyingtest.Noninvasive Determination of Gastric Acid

Secretion: 13C-Calcium Carbonate Breath

Test. The 13C-calcium carbonate breath

test (13C-CC-BT) has been designed toestimate gastric acid secretion in a non-invasive manner (23). Study medicationwas injected 1 h before ingestion of thetracer. After an overnight fast of ;12 h,baseline breath samples were taken.Thereafter, subjects ingested 0.2 g 13C-labeled calcium carbonate dispersed instill water. Breath samples were takentwice at baseline and every 15 min overthe test duration of 90 min. 13C-CC-BTwas analyzed as described by Shinkaiet al. (24). Maximum 13C-CaCO3–derived13CO2 exhalation was chosen for estima-tion of gastric acid secretion, as it isreported to show a good correlationwith the total secretion of gastric acid(24).

End PointsPrimary end point was defined as changefrom baseline (week 21) in the numberof reflux episodes over 24 h after 10weeks of treatment with lixisenatide ver-sus liraglutide. Secondaryendpoints (eachafter 10 weeks of treatment with liraglu-tide versus lixisenatide) were change frombaseline (week 21) in the time charac-terized by a pH,4.0 in the lower third ofthe esophagus, change from baseline ofperistaltic tone in the esophagus andpressure of the LES, change from baselineof gastric emptying, and change frombaseline of the amount of nonstimulated,stimulated, and steady-state gastric acidsecretion.

Statistical Analyses and Sample SizeSubject characteristics are expressed asmeans 6 SD or number (proportion oftotal). Results are expressed as means6SEM or D (95% CI). Data were analyzedusing Student t test for paired samples orStudent t test with Welch correction forunpaired samples. Categorical parame-ters (contingency tables) were assessedwith Fisher exact test. Gastric emptyingwas analyzed using repeated-measuresANOVA with gastric content (as % ofthe initial value) over 360 min as thedependent variable, treatment (afterversus before initiation of GLP-1 RAtreatment) as fixed variable, and subjectas random independent variable. Differ-ences after 10 weeks of treatment withlixisenatide and liraglutide versus base-line are expressed as D (with or without95% CI). Statistical significance was de-fined as P , 0.05. For statistical analysesandfigures, GraphPadPrism, version 8.0.0

for Windows (GraphPad Software, SanDiego, CA) (www.graphpad.com), and Sta-tistica (data analysis software system),version 13.5 for Windows (TIBCO Soft-ware, Palo Alto, CA) (http://tibco.com),were used.

Based on the assumption of a mean of30 reflux episodes per 24 h and an SD of35, n 5 50 subjects would provide 90%power (at an a-level of 0.05) to detect adifference by 15, e.g., a 50% change aftertreatment with a GLP-1 RA. For differ-ences in changes from baseline betweenlixisenatide and liraglutide treatment,n 5 25 per groups would provide 90%power to detect a difference of 15 refluxepisodes per 24 h assuming an SD of 15.

RESULTS

Baseline DataA total of 110 male or female patientswith type 2 diabetes were screened forthe study. Of these, 57 (51.8%) subjects(all Caucasians) were enrolled. Seven(12.3%) participants did not completethe study (two [28.7%] participants with-drew consent before receiving studymedication, two [28.7%] were incompli-ant with the study protocol, and three[42.9%] had unsuccessful esophagealmanometry or pH measurements). Ofthe participants that completed thestudy, 26 (52%) received liraglutide and24 (48%) received lixisenatide. Baselinecharacteristics of completers are de-picted in Table 1. Aside from pulserate, there was no significant differencebetween the groups at baseline (mean6SD 68.7 6 8.6 bpm in lixisenatide and63.56 8.1 bpm in liraglutide, P5 0.035)(Supplementary Table 1). Participantstreated with liraglutide experienced moreweight loss (P 5 0.0078) and a greaterreduction in BMI (P 5 0.020) comparedwith those treated with lixisenatide(Supplementary Table 1). There was nosignificant difference in the concomitantuse of other glucose-lowering agents be-tween the groups (Supplementary Table2). In consideration of the study period of10weeks, changes in HbA1c from baselinewere not measured.

Esophageal pH MeasurementBaseline measurements were similar inboth treatment groups (Table 2). Overall,there were no significant differences inthe parameters of pHmeasurement (Fig.1A–C). The number of reflux episodes didnot change significantly after 10weeks of

care.diabetesjournals.org Quast and Associates 2139

http://www.graphpad.comhttp://tibco.comhttps://doi.org/10.2337/figshare.12490601https://doi.org/10.2337/figshare.12490601https://doi.org/10.2337/figshare.12490601https://doi.org/10.2337/figshare.12490601http://care.diabetesjournals.org

treatment with lixisenatide (mean 6SEM 27.26 6 5.21 to 32.11 6 5.09,P 5 0.23) or liraglutide (39.43 6 6.14to 47.29 6 8.88, P 5 0.34) (Table 2).Therealsowerenot significant changes inDeMeester score and time with pH,4.0during the 24-h measurement periodafter 10 weeks of treatment (Table 2).At baseline, a DeMeester scoring indicat-ing increased esophageal reflux (.14.7)was present in 7 (36.8%) participants inthe lixisenatide and in 14 (66.7%) pa-tients in the liraglutidegroup. Therewerenumerical increases in those with abnor-mal results after 10 weeks of treatmentfor both GLP-1 RAs (from 21 [52.5%] to26 [65.0%], P 5 0.36), with lixisenatidetreatment (from 7 [36.8%] to 11 [57.9%],P 5 0.33) or with liraglutide treatment(from14 [66.7%] to15 [71.4%],P.0.99),all of which were not significant. pHmea-surement and analysis were completedin 19 (79.2%) patients in the lixisenatidegroup and 21 (80.8%) patients of theliraglutide group (Supplementary Table3).

High-Resolution EsophagealManometryAmong the parameters examined, nodifference was found for treatmentwith GLP-1 RAs in general (Fig. 1E–H) orfor each GLP-1 RA studied individually(Table 2). Baseline measurements weresimilar. Neither lixisenatide nor liraglu-tide led to a significant change in restingpressure of the LES (Fig. 1E), mean distal

pressure amplitude (Fig. 1F), maximumdistal pressure amplitude (Fig. 1G), or therelaxation of the LES after the act ofswallowing (Fig. 1H). HRMwas successfulin 23 (88.5%) participants of the liraglu-tide group and 19 (79.2%) of the lixisena-tide group (Supplementary Table 3).

Gastric EmptyingAfter 10 weeks of treatment, gastricemptying was significantly delayed withlixisenatide (P , 0.0001) (Fig. 2A) andliraglutide (P , 0.0002) (Fig. 2B). Gastricemptying half-time was delayed, with amean6SEMincreaseby25610min (P50.025) in the liraglutide group and by526 17 min (P5 0.0065) in the lixisena-tide group. The breath testing for gastricemptying was successful in 21 (87.5%)participants in the lixisenatide group andin 26 (100%) participants in the liraglutidegroup (Supplementary Table 3).

At baseline, three (5.3%) patientsexhibited a prolongation of gastric emp-tyingof.2SD (216.864.6min).Noneofthese patients reported symptoms ofgastroparesis at screening. The resultsof the study were not altered by exclu-sion of these patients from analysis (datanot shown).

Gastric Acid SecretionThere was no significant baseline differ-ence between the lixisenatide group andthe liraglutide group (P 5 0.88). Thepooled analysis showed a significant re-duction of gastric acid production with

the GLP-1 RAs, by –20.7 6 9.9% (P 50.042) (Fig. 1D). For the individual agents,no significant difference was found (Ta-ble 2). In five (10%) participants of thetotal study cohort, no 13C excursion wasregistered at both study visits (one [20%]in the lixisenatide group and four [80%]in the liraglutide group, respectively).These participants were excluded fromthe primary analysis. However, with in-clusion of these participants, the overallchanges between baseline and 10 weekswere still significant (P 5 0.044) andsimilar for lixisenatide (P 5 0.19) andliraglutide (P 5 0.11). The 13C-CC-BTyielded results in 26 (100%) participantsin the liraglutide group and in 24 (100%)participants in the lixisenatide group(Supplementary Table 3).

SafetyNo significant difference in hypoglyce-mic or gastrointestinal adverse eventswas found between the groups (Supple-mentary Table 4). All patients with hy-poglycemia were under treatment withinsulin or sulfonylurea. No severe hypo-glycemic episodes occurred during thetrial.

CONCLUSIONS

The current study was designed to ex-amine the effects of lixisenatide andliraglutide on gastroesophageal reflux,esophageal motility, gastric emptying,and gastric acid secretion in subjectswith type 2 diabetes. We report that

Table 1—Patients’ characteristics at baseline

All participants Treated with lixisenatide Treated with liraglutideSignificance ofdifferences (P)

n female/n male (% female) 18/32 (36) 10/14 (41.7) 8/18 (30.8)

Age, years 60.4 6 7.4 60.7 6 7.6 60.2 6 7.3 0.81

Duration of type 2 diabetes, years 13.5 6 1.0 12.6 6 1.1 14.4 6 1.7 0.38

Arterial hypertension 32 (64) 15 (62.5) 18 (69.2) 0.77

Duration of arterial hypertension, years 10.4 6 1.5 15.6 6 7.5 11.9 6 1.9 0.64

Dyslipidemia 19 (38) 6 (25) 13 (50) 0.087

Diabetic neuropathy* 4 (8) 0 4 (15.4 0.11

Diabetic retinopathy 7 (14) 1 (4.2) 6 (23.1) 0.10

HbA1c, % (mmol/mol) 7.5 6 0.8 (58.8 6 8.3) 7.5 6 0.8 (58.7 6 8.2) 7.5 6 0.8 (58.9 6 8.6) 0.93

Fasting plasma glucose, mmol/L 8.8 6 2.3 9.0 6 1.7 8.7 6 2.7 0.61

Participants smoking 5 (10) 3 (12.5) 2 (7.7) 0.66

Participants consuming alcohol 32 (64) 15 (62.5) 17 (65.4) .0.99

Alcohol consumption of participantsconsuming alcohol, units/week 3.7 6 4.3 3.5 6 4.2 3.9 6 4.5 0.82

Height, cm 172.0 6 10.6 169.5 6 12.0 174.2 6 8.7 0.12

Data are expressed as means6 SD or n (%) of patients unless otherwise indicated. P values were calculated from Student t test withWelch correctionor Fisher exact test. *Based on vibration perception threshold and medical history.

2140 Reflux and GI Motility in GLP-1 RAs Diabetes Care Volume 43, September 2020

https://doi.org/10.2337/figshare.12490601https://doi.org/10.2337/figshare.12490601https://doi.org/10.2337/figshare.12490601https://doi.org/10.2337/figshare.12490601https://doi.org/10.2337/figshare.12490601https://doi.org/10.2337/figshare.12490601https://doi.org/10.2337/figshare.12490601

Table

2—Para

meters

ofuppergastro

intestin

almotility,

gastro

eso

phagealreflux,

gastric

emptying,andgastric

acid

secre

tionatbase

lineandafte

r10

weeksoftre

atm

entwith

lixisenatid

eorlira

glutid

e

Treatmen

twith

lixisenatide

Treatmen

twith

liraglutid

eDifferen

cebetw

eenthegro

ups

Baselin

eAfter

10weeks

oftreatm

ent

Differen

cevs.

baselin

eP

Baselin

eAfter

10weeks

oftreatm

ent

Differen

cevs.

baselin

eP

Differen

cebetw

eentreatm

ents

P

Esophageal

reflux

Episo

des/24

h27.3

65.2

32.16

5.14.8

(23.4,

13.1)0.23

39.46

6.147.3

68.9

7.9(2

9.1,24.8)

0.3423.0

(221.9,

15.9)0.75

DeM

eestersco

re23.2

66.0

27.46

6.44.2

(210.9,

19.2)0.57

45.86

11.150.8

612.8

5.0(2

28.9,38.9)

0.7620.8

(238.1,

36.5)0.97

%tim

epH,4.0

2.96

1.51.9

60.7

21.0

(23.2,

1.2)0.34

5.16

2.08.4

63.1

3.3(2

2.8,9.4)

0.2724.3

(210.8,

2.2)0.19

Esophageal

manometry

Restin

gpressu

reoftheLES

(mmHg)

26.16

2.626.1

62.5

0.0(2

4.1,4.1)

0.9931.1

63.0

32.96

3.51.9

(23.8,

7.6)0.50

21.9

(28.7,

5.0)0.58

Mean

distal

pressu

ream

plitu

de(m

mHg)

64.96

7.964.4

66.1

20.6

(212.7,

11.6)0.93

81.36

7.483.2

67.5

1.9(2

7.1,10.9)

0.6622.5

(217.1,

12.2)0.74

Maxim

um

distal

pressu

ream

plitu

de

(mmHg)

176.46

16.7177.1

620.8

0.7(2

29.9,31.2)

0.96197.0

615.6

180.66

9.2216.4

(243.2,

10.5)0.22

17.0(2

22.4,56.5)

0.39Po

stdeglu

titiveeso

phageal

relaxation(%

)66.3

63.7

61.86

3.924.4

(211.9,

3.1)0.23

63.96

3.060.1

63.6

23.8

(210.2,

2.6)0.23

20.6

(210.2,

8.9)0.89

Gastric

emptyin

ghalf-tim

e(m

in)

88.46

7.7135.9

616.0

52.0(16.3,

87.8)0.0065

97.56

9.0122.1

612.8

24.6(3.4,

45.7)0.025

27.5(2

13.2,68.1)

0.17

Gastric

acidsecretio

n:maxim

um

13C

excretion(D

13C

maxim

um)(&

)138.9

624.4

107.56

11.9231.4

(279.3,

16.4)0.19

145.26

15.7118.0

69.5

227.3

(260.6,

6.0)0.10

24.2

(261.3,

53.0)0.88

Data

arepresen

tedas

mean

s6

SEMor,fo

rdifferen

cevs.b

aseline(D),as

mean

D(95%

CI).D

ifference

betw

eentreatm

ents,as

mean

D(95%

CI),is

disp

layedforchanges

from

baselin

e.Pvalu

eswere

calculated

from

Studen

tttest

forpaired

samples

orStu

den

tttest

with

Welch

correctio

n.Po

stdeglu

titive,after

theact

ofsw

allowing.

care.diabetesjournals.org Quast and Associates 2141

http://care.diabetesjournals.org

neither lixisenatide nor liraglutide treat-ment had significant effects on the num-ber of gastroesophageal reflux episodesper 24-h period. Likewise, esophagealmotility was not affected. After 10weeksof treatment, both GLP-1 RAs elicited asignificant delay in gastric emptying ofsolid components of a test breakfast.Lixisenatide and liraglutide led to aminorsuppression of gastric acid secretion,which only reached statistical signifi-cance when analyses for both GLP-1RAs were combined.The lack of effect of both GLP-1 RAs on

gastroesophageal reflux and on variousmeasures ofesophagealmotility is a noveland somewhat unexpected finding, sinceany persisting delay in gastric emptyinginduced especially with a short-acting

GLP-1 RA might have been expected topromote gastroesophageal reflux events(10). The small increase of participantswith a DeMeester scoring indicating sig-nificant reflux seen in the lixisenatide didnot show statistical significance. In theliraglutide group, no such effectwas seen.The difference may also be due to thehigher percentage of participants with apathological DeMeester scoring at base-line in the liraglutide group comparedwith the lixisenatide group (66.7% vs.36.8%, respectively). Nonetheless, noneof the described results were significant.Thus, even though gastric emptying issignificantly delayed in both agents, nosignificant effects on gastroesophagealreflux were observed. For lixisenatide,the absence of such effects might reflect

the transient nature of GLP-1 receptoractivation with lixisenatide. For liraglu-tide, a less pronounced effect on gastricemptying may further reduce the risk ofreflux.

Another important end point of thecurrent study was the impact of lixisena-tide and liraglutide on esophageal mo-tility. Esophagealmotility is regulated in acomplex manner, and multiple aspectsranging from propulsion and relaxationto the resting tone of the esophagus andthe LES need to be considered (25).Because of the inhibitory effect ofGLP-1 RAs on gastric emptying, a missinginfluence on the resting pressure of theLES was rather unexpected. A prolongedgastric filling could have been expectedto show at least a minor influence on the

Figure1—Variousparametersbeforeandafter treatmentwith lixisenatide (▼) and liraglutide (4).Dataarepresentedas individual valuesandmeans6SEM. P values were calculated from Student t test for paired samples. For differences between lixisenatide and liraglutide treatments, see Table 2.A–C: Parameters derived from24-hpHmeasurements.Measurementswereperformed5 cmabove the LES (identified throughmanometry).A: Numberof refluxepisodesper24h.B: DeMeester score. Thegraybottomsegment indicates thephysiologic range (DeMeester scoreofup to14.7).C: Percentagetime with pH,4.0 during the registration period. D: Gastric acid content in the fasting state, 60 min after injection of the study drug. E–H: Results ofthe high-resolutionmanometry. Visualized are resting pressure of the LES (E), mean distal pressure amplitude (F), maximumdistal pressure amplitude(G), and postdeglutitive esophageal relaxation (H).

2142 Reflux and GI Motility in GLP-1 RAs Diabetes Care Volume 43, September 2020

resting pressure due to a more pro-nounced upper gastric distension. How-ever, neither a short-acting nor a long-acting GLP-1 RA showed a significanteffect on any of the esophageal pressureparameters examined. It should benotedthat in clinical practice, the evaluation ofHRM involves not only the quantitativeanalyses conducted in the current studybut also a visual analysis of the respectivepressure changes over time in morequalitative terms. In the current study,numeric analysis was favored due to itsbetter reproducibility and lower depen-dence on investigators’ experience.Inhibition of gastric acid secretion has

previously been reported during intra-venous GLP-1 administration after pen-tagastrin stimulation (26). In the currentstudy, an indirect measure of gastric acidsecretion based on calcium carbonatemetabolism was employed. Notably, un-like in earlier studies, this assessmentwasperformednot after stimulation (i.e.,by pentagastrin) but, rather, under base-line conditions. This approach was usedsuccessfully to evaluate acid suppressionafter acid blocker treatment, but nouse of this method in evaluation ofGLP-1 RAs has been reported so far(23). Interestingly, a small, but significant(mean6 SEM –20.76 9.9%, P5 0.042),reduction in gastric acid secretion wasfound in the combined analysis ofeffects of either lixisenatide or liraglutide

treatment. A significant effect was notobserved in the individual analysis oflixisenatide’s or liraglutide’s effects, per-haps because of low statistical power.However, compared with the effect ofproton pump inhibitors the effect isminor (23). The minimal inhibition ofgastric acid secretion observed withboth our study drugs and previouslywith native GLP-1 (27) is compatiblewith a reduction in gastrin secretionas recently indicated (28), as well aswith the expression of GLP-1 receptorson parietal cells (29).

Fiveparticipants (four in the liraglutidegroup) showed no significant gastric acidproduction on both visits. Since the in-take of proton pump inhibitors or othermedication that may influence gastricacidity or gastrointestinal motility wasnot allowed in the current study, the lackof gastric acid production suggests anunderlying condition such as atrophicgastritis, which is known to reduce acidproduction and 13C excursion in 13C-CC-BT(24).

Whether there is a clinical impact ofthe reduction of gastric acid secretionwith GLP-1 RAs is unclear. Theoretically,any reduction in gastric acid secretionmight reduce the subsequent risk forreflux events or damage caused thereby,e.g., reflux esophagitis or duodenal ulcers.Recently, the administration of GLP-1was shown to reduce gastrin secretion

in healthy individuals and in individualswith type 2 diabetes, suggesting a pos-sible mechanism for the observed re-duction in acidity (28). However, giventhe small magnitude of this effect, amajor clinical impact appears unlikely.

Lixisenatide and liraglutide signifi-cantly inhibited gastric emptying after10 weeks of treatment. The short-actingGLP-1 RA lixisenatide showed a morepronounced effect, with a delay of thegastric emptying half-time of more thantwice that with the long-acting GLP-1 RAliraglutide. Previous studies reported alarge difference in the inhibition of gas-tric emptying with short-acting versuslong-acting GLP-1 RAs, which cannot bereported in the current study (30,31).Indeed, the delay in nutrient absorptionsubsequent to delayed gastric emptyingis believed to represent the main mech-anism of action reducing postprandialglucose excursions with short-actingGLP-1 RAs but applies mainly to mealsbefore which such medications havebeen injected (32). Nevertheless, withlong-acting GLP-1 RAs, the effects ongastric emptying are largely diminishedduring chronic treatment because oftachyphylaxis (2). In contrast with thisand in line with the present results, arecent study reported a persistent effectof the long-acting formulation of exena-tide on gastric emptying after a treat-ment period of 8 weeks (33). Hence, itremains unclear how much time it takesto reach a steady state in therapy withshort-acting versus long-acting GLP-1RAs and after which period tachyphylaxiscan be observed.

Since recruitment was a major chal-lenge during this study, the study pro-tocolwas amended to allow the inclusionof patients pretreated with dipeptidyl-peptidase 4 inhibitors. In total, fourpatients with sitagliptin and one pa-tient with vildagliptin (lixisenatide group)were recruited (Supplementary Table 2).A washout period of 4 weeks precededthe inclusion. However, previous studieshave shown that both sitagliptin andvildagliptin have no impact on gastricemptying (34,35). Furthermore, all re-ported parameters maintained their sta-tistical significancewhentheseparticipantswere excluded from the analyses.

A couple of limitations need to beconsidered: The assessment of gastricemptying and gastric acid secretion wasperformed using noninvasive breath

Figure 2—Gastric emptying of solid components in lixisenatide (A) and liraglutide (B) before andafter 10 weeks of treatment. Gastric content was calculated with the Wagner-Nelson equation.Data are expressed as means6 SEM. Data were analyzed using repeated-measures ANOVA withgastric content (as % of the initial value) over 360min as the dependent variable, treatment (aftervs. before initiation of GLP-1 RA treatment) as fixed variable, and subject as random independentvariable. The results are presented with P values marked ‘a’ representing the overall effect of theintervention, P values marked ‘b’ representing the effect of time, and P values marked ‘ab’representing the interaction of intervention and time on gastric emptying. If a or ab weresignificant (P, 0.05), post hoc (Duncan) tests were performed to locate significant differences tosingle time points. For gastric emptying half-time, see Table 2.

care.diabetesjournals.org Quast and Associates 2143

https://doi.org/10.2337/figshare.12490601http://care.diabetesjournals.org

tests,which allowed for the estimationofthe respective parameters without ra-dioactive exposure. However, only semi-quantitative estimates can be generatedusing these measurements. While weacknowledge that gastric emptying scin-tigraphy represents the gold standard,the octanoate breath test has previouslybeen validated and used extensively inprevious studies with GLP-1 RAs (36,37).Indeed, theWagner-Nelson approach forderiving gastric emptying from 13CO2exhalation data gives similar resultscompared with scintigraphy (22). Also,aspiration of gastric juice after pentagas-trin stimulation may be considered thegold standard for the determination of(maximum) gastric acid secretion, butsuch an approach appears further awayfrom physiological conditions than thecurrently employed calcium carbonatebreath test. A placebo arm and a thirdvisit after 7 days may have added furtherinformation. Due to the study design, noconclusion about early effects can bedrawn. The study was powered to ex-clude an increase in reflux episodesof .50%, but no firm conclusions re-garding less pronounced changes in thenumber of reflux events can be drawn.The study design was investigatorblinded, but open-label for patients. Fi-nally, improper placement of the pHmeasurement tube caused by using onlypH variation for placement is susceptiblefor artifacts and can influence the De-Meester score. To avoid this, the properpositioning of the pHmeasurement tubewas ascertained by prior manometricdetermination of the LES, thereby pro-viding additional validation. It is alsotheoretically possible that changes inmeal size or composition induced bythe GLP-1 RAs might have reducedthe frequency of reflux episodes,thereby potentially obscuring an in-duction of GERD.In conclusion, the current study has

demonstrated that despite the delay ingastric emptying, neither lixisenatide norliraglutide leads to significant alterationsin gastroesophageal reflux events oresophageal motility. These findings pro-vide additional reassurance as to theoverall gastrointestinal safety of GLP-1RAs. The minor reduction of gastric acidsecretion might protect against the riskof reflux or gastritis.

Acknowledgments. The technical assistanceof B. Baller (St Josef-Hospital) and B. Kronshage(Profil) is greatly acknowledged.Duality of Interest. This study was sponsoredby Novo Nordisk.M.A.N. has been a member of advisory

boards for or has consulted with AstraZeneca,Boehringer Ingelheim, Eli Lilly & Co., Fractyl,GlaxoSmithKline, Hoffman-La Roche, MENARINI/BERLIN-CHEMIE, Merck Sharp & Dohme, NovoNordisk, and Versatis; has received grant supportfrom Eli Lilly & Co., MENARINI/BERLIN-CHEMIE,Merck Sharp & Dohme, and Novartis Pharma;and has also served on the speakers’ bureauof AstraZeneca, Boehringer Ingelheim, Eli Lilly &Co.,GlaxoSmithKline,MENARINI/BERLIN-CHEMIE,Merck Sharp & Dohme, Novo Nordisk, and SunPharma. C.K. has received travel grants from EliLilly & Co. and is a co-owner of Profil, whichreceived research funds from Adocia, Biocon,Boehringer Ingelheim, Dance Biopharm Holdings,Eli Lilly & Co., Gan & Lee Pharmaceuticals, Med-Immune, Mylan, Nordic Bioscience, Nestlé, NovoNordisk A/S, Poxel SA, Sanofi, Wockhardt, XerisPharmaceuticals, and Zealand Pharma A/S. J.J.M.has received lecture honoraria and consultingfees from AstraZeneca, BERLIN-CHEMIE, Boeh-ringer Ingelheim, Eli Lilly, Merck Sharp & Dohme,Novo Nordisk, Novartis, and Sanofi; has receivedreimbursement of congress participation fees andtravelexpensesfromMerckSharp&Dohme,NovoNordisk, and Sanofi; and has initiated projectssupported by Boehringer Ingelheim, Merck Sharp& Dohme, Novo Nordisk, and Sanofi. No otherpotential conflicts of interest relevant to thisarticle were reported.The study sponsor was not involved in the

design of the study; the collection, analysis, andinterpretation of data; drafting the manuscript;or the decision to submit the manuscript forpublication.Author Contributions. D.R.Q. performed theesophageal manometry and the esophageal pHmeasurement, analyzed and discussed the data,and contributed to the manuscript preparation.N.S. and B.A.M. helped to perform the experi-ments and discussed the data. M.A.N. analyzedthe data, discussed the results, and criticallyreviewed the manuscript. J.J.M. and C.K. designedthe study, helped to perform the experiments,analyzed the data, and prepared the manuscript.All authors approved the final version of themanuscript. J.J.M. is the guarantor of this workand, as such, had full access to all the data in thestudy and takes responsibility for the integrity ofthe data and the accuracy of the data analysis.

References1. Nauck MA, Meier JJ. Management of endo-crine disease: are all GLP-1 agonists equal in thetreatment of type 2 diabetes? Eur J Endocrinol2019;181:R211–R2342. Meier JJ. GLP-1 receptor agonists for individ-ualizedtreatmentof type2diabetesmellitus.NatRev Endocrinol 2012;8:728–7423. AbdElAzizMS,KahleM,Meier JJ,NauckMA.Ameta-analysis comparing clinical effects of short-or long-acting GLP-1 receptor agonists versusinsulin treatment from head-to-head studies intype 2 diabetic patients. Diabetes Obes Metab2017;19:216–227

4. Trujillo JM, Goldman J. Lixisenatide, a once-daily prandial glucagon-like peptide-1 receptoragonist for the treatment of adults with type 2diabetes. Pharmacotherapy 2017;37:927–9435. Jacobsen LV, Flint A, Olsen AK, Ingwersen SH.Liraglutide in type 2 diabetes mellitus: clinicalpharmacokinetics and pharmacodynamics. ClinPharmacokinet 2016;55:657–6726. Geiser JS, Heathman MA, Cui X, et al. Clinicalpharmacokinetics of dulaglutide in patients withtype 2 diabetes: analyses of data from clinicaltrials. Clin Pharmacokinet 2016;55:625–6347. Meier JJ, Rosenstock J, Hincelin-Méry A, et al.Contrasting effects of lixisenatide and liraglutideon postprandial glycemic control, gastric emp-tying, and safety parameters in patients withtype2diabetesonoptimized insulinglarginewithor withoutmetformin: a randomized, open-labeltrial. Diabetes Care 2015;38:1263–12738. Jelsing J, Vrang N, Hansen G, Raun K, Tang-Christensen M, Knudsen LB. Liraglutide: short-lived effect on gastric emptying – long lastingeffects on body weight. Diabetes Obes Metab2012;14:531–5389. Wettergren A, Wøjdemann M, Meisner S,Stadil F, Holst JJ. The inhibitory effect of gluca-gon-like peptide-1 (GLP-1) 7-36 amide on gastricacid secretion in humans depends on an intactvagal innervation. Gut 1997;40:597–60110. Bor S. Reflux esophagitis, functional andnon-functional. Best Pract Res ClinGastroenterol2019;40-41:10164911. WangKY, ChenYW,WangTN, et al. Predictorof slower gastric emptying in gastroesophagealreflux disease: survey of an Asian-Pacific cohort.J Gastroenterol Hepatol 2019;34:837–84212. Sun XM, Tan JC, Zhu Y, Lin L. Associationbetweendiabetesmellitusandgastroesophagealreflux disease: a meta-analysis. World J Gastro-enterol 2015;21:3085–309213. RoufMA, KhanM, Sharif JU, et al. Prevalenceof GERD in type II diabetes mellitus patientsadmitted in a tertiary care hospital of Bangla-desh. Mymensingh Med J 2017;26:710–71514. Lluch I, Ascaso JF, Mora F, et al. Gastro-esophageal reflux in diabetes mellitus. Am JGastroenterol 1999;94:919–92415. Noguchi Y, Katsuno H, Ueno A, et al. Signalsof gastroesophageal reflux disease caused byincretin-based drugs: a disproportionality anal-ysisusing the Japaneseadversedrugevent reportdatabase. J Pharm Health Care Sci 2018;4:1516. Phillips LK, Rayner CK, Jones KL, HorowitzM.Measurement of gastric emptying in diabetes. JDiabetes Complications 2014;28:894–90317. Johnson LF, Demeester TR. Twenty-four-hour pH monitoring of the distal esophagus. Aquantitative measure of gastroesophageal re-flux. Am J Gastroenterol 1974;62:325–33218. Gyawali CP, Kahrilas PJ, Savarino E, et al.Modern diagnosis of GERD: the Lyon Consensus.Gut 2018;67:1351–136219. Kahrilas PJ, Bredenoord AJ, Fox M, et al.;International High ResolutionManometryWork-ing Group. The Chicago Classification of esoph-ageal motility disorders, v3.0. NeurogastroenterolMotil 2015;27:160–17420. ChoiMG,CamilleriM, BurtonDD, ZinsmeisterAR, ForstromLA,NairKS. [13C]octanoic acidbreathtest for gastric emptying of solids: accuracy, re-producibility, and comparison with scintigraphy.Gastroenterology 1997;112:1155–1162

2144 Reflux and GI Motility in GLP-1 RAs Diabetes Care Volume 43, September 2020

21. Sanaka M, Nakada K, Nosaka C, Kuyama Y.The Wagner-Nelson method makes the [13C]-breath test comparable to radioscintigraphy inmeasuring gastric emptying of a solid/liquidmixed meal in humans. Clin Exp PharmacolPhysiol 2007;34:641–64422. Trahair L, Nauck M, Wu T, et al. Measure-mentof gastricemptyingusing scintigraphyanda13C-octanoic acid breath test with Wagner-Nelsonanalysis in type 2 diabetes (54(th) EASD AnnualMeeting of the European Association for theStudyof Diabetes: Berlin, Germany, 1 - 5October2018). Diabetologia 2018;61:1–62023. Miyamoto S, Tsuda M, Kato M, et al. Eval-uation of gastric acid suppression with vonopra-zan using calcium carbonate breath test. J ClinBiochem Nutr 2019;64:174–17924. Shinkai H, Iijima K, Koike T, et al. Calciumcarbonate breath test for non-invasive estima-tion of gastric acid secretion. Tohoku J Exp Med2014;232:255–26125. Tolone S, Savarino E, Docimo L. Is there a rolefor high resolutionmanometry inGERDdiagnosis?Minerva Gastroenterol Dietol 2017;63:235–24826. Meier JJ, Nauck MA, Kask B, et al. Influenceof gastric inhibitory polypeptide on pentagastrin-stimulated gastric acid secretion in patients withtype 2 diabetes and healthy controls. World JGastroenterol 2006;12:1874–1880

27. Nauck MA, Bartels E, Orskov C, Ebert R,Creutzfeldt W. Lack of effect of synthetichuman gastric inhibitory polypeptide and glu-cagon-like peptide 1 [7-36 amide] infused atnear-physiological concentrations on penta-gastrin-stimulated gastric acid secretion innormal human subjects. Digestion 1992;52:214–22128. Rehfeld JF, Knop FK, Asmar M. Gastrinsecretion in normal subjects and diabetes pa-tients is inhibited by glucagon-like peptide 1:a role in the gastric side effects of GLP-1-deriveddrugs? Scand J Gastroenterol 2019;54:1448–145129. Pyke C, Heller RS, Kirk RK, et al. GLP-1receptor localization in monkey and humantissue: novel distribution revealed with exten-sively validated monoclonal antibody. Endocri-nology 2014;155:1280–129030. Horowitz M, Flint A, Jones KL, et al. Effect ofthe once-daily human GLP-1 analogue liraglutideon appetite, energy intake, energy expenditureand gastric emptying in type 2 diabetes. DiabetesRes Clin Pract 2012;97:258–26631. Jones KL, Rigda RS, Buttfield MDM, et al.Effects of lixisenatide on postprandial bloodpressure, gastric emptying and glycaemia inhealthy people and people with type 2 diabetes.Diabetes Obes Metab 2019;21:1158–1167

32. Becker RH, Stechl J, Steinstraesser A, GolorG, Pellissier F. Lixisenatide reduces postprandialhyperglycaemia via gastrostatic and insulino-tropic effects. Diabetes Metab Res Rev 2015;31:610–61833. Jones KL, Huynh LQ, Hatzinikolas S, et al.Exenatide onceweekly slows gastric emptying ofsolids and liquids in healthy, overweight peopleat steady-state concentrations. Diabetes ObesMetab 2020;22:788–79734. Stevens JE, ButtfieldM,WuT, et al. Effects ofsitagliptin on gastric emptying of, and the gly-caemic and blood pressure responses to, a car-bohydrate meal in type 2 diabetes. DiabetesObes Metab 2020;22:51–5835. He YL. Clinical pharmacokinetics and phar-macodynamics of vildagliptin. Clin Pharmacoki-net 2012;51:147–16236. Lorenz M, Pfeiffer C, Steinsträsser A, et al.Effects of lixisenatide once daily on gastric emp-tying in type 2 diabetes–relationship to postpran-dial glycemia. Regul Pept 2013;185:1–837. Meier JJ,MengeBA, SchenkerN, et al. Effectsof sequential treatment with lixisenatide, insulinglargine, or their combination on meal-relatedglycaemic excursions, insulin and glucagon se-cretion, and gastric emptying in patients withtype 2 diabetes. Diabetes Obes Metab 2020;22:599–611

care.diabetesjournals.org Quast and Associates 2145

http://care.diabetesjournals.org