Efficacy and Safety of Dulaglutide Added Onto Pioglitazone And

Diabetes Care Volume 37, August 2014 2159

Carol Wysham,1 Thomas Blevins,2

ﬁ CARE/EDUCATION/NUTRITION/PSYCHOSOCIAL CLIN Ef cacy and Safety of Dulaglutide Richard Arakaki,3 Gildred Colon,4 Pedro Garcia,5 Charles Atisso,6 Added Onto Pioglitazone and Debra Kuhstoss,6 and Mark Lakshmanan6 Metformin Versus Exenatide in Type 2 Diabetes in a Randomized Controlled Trial (AWARD-1) Diabetes Care 2014;37:2159–2167 | DOI: 10.2337/dc13-2760

OBJECTIVE To compare the efﬁcacy and safety of dulaglutide, a once-weekly GLP-1 receptor agonist, with placebo and exenatide in type 2 diabetic patients. The primary objective was to determine superiority of dulaglutide 1.5 mg versus placebo in

HbA1c change at 26 weeks.

RESEARCH DESIGN AND METHODS This 52-week, multicenter, parallel-arm study (primary end point: 26 weeks) ran- 1Redwood Clinic, Spokane, WA domized patients (2:2:2:1) to dulaglutide 1.5 mg, dulaglutide 0.75 mg, exenatide 2 m Texas Diabetes and Endocrinology, Austin, TX 10 g, or placebo (placebo-controlled period: 26 weeks). Patients were treated 3University of Hawaii at Manoa, Honolulu, HI with metformin (1,500–3,000 mg) and pioglitazone (30–45 mg). Mean baseline 4American Telemedicine Center of Puerto Rico, HbA1c was 8.1% (65 mmol/mol). San Juan, PR 5University Hospital Autonomous University of RESULTS Nuevo Leon,´ Monterrey, Mexico 6 6 Lilly Diabetes, Eli Lilly and Company, Indianap- Least squares mean SE HbA1c change from baseline to the primary end point olis, IN 2 6 2 6 2 6 was 1.51 0.06% ( 16.5 0.7 mmol/mol) for dulaglutide 1.5 mg, 1.30 Corresponding author: Mark Lakshmanan, 0.06% (214.2 6 0.7 mmol/mol) for dulaglutide 0.75 mg, 20.99 6 0.06% [email protected]. (210.8 6 0.7 mmol/mol) for exenatide, and 20.46 6 0.08% (25.0 6 0.9 Received 25 November 2013 and accepted 8 mmol/mol) for placebo. Both dulaglutide doses were superior to placebo at May 2014. 26 weeks (both adjusted one-sided P < 0.001) and exenatide at 26 and 52 weeks Clinical trial reg. no. NCT01064687, clinicaltrials (both adjusted one-sided P < 0.001). Greater percentages of patients reached .gov. This article contains Supplementary Data online HbA1c targets with dulaglutide 1.5 mg and 0.75 mg than with placebo and exenatide (all P < 0.001). At 26 and 52 weeks, total hypoglycemia incidence was lower in at http://care.diabetesjournals.org/lookup/ suppl/doi:10.2337/dc13-2760/-/DC1. patients receiving dulaglutide 1.5 mg than in those receiving exenatide; no dula- A slide set summarizing this article is available glutide-treated patients reported severe hypoglycemia. The most common gas- online. trointestinal adverse events for dulaglutide were nausea, vomiting, and diarrhea. © 2014 by the American Diabetes Association. Events were mostly mild to moderate and transient. Readers may use this article as long as the work is properly cited, the use is educational and not CONCLUSIONS for profit, and the work is not altered. Both once-weekly dulaglutide doses demonstrated superior glycemic control ver- See accompanying articles, pp. 2149 sus placebo and exenatide with an acceptable tolerability and safety profile. and 2168. 2160 Efficacy/Safety of Dulaglutide Versus Exenatide Diabetes Care Volume 37, August 2014

Type 2 diabetes is characterized by pro- one with sustained GLP-1 activation. for severe, persistent hyperglycemia; a gressive b-cell failure and insulin resis- This information should be useful in detailed description of the criteria for tance, and intensification of treatment making treatment decisions for individ- rescue therapy is provided in the Sup- is usually required over time. The Amer- ual patients with type 2 diabetes. plementary Data. In addition, patients ican Diabetes Association and European who discontinued the study drug Association for the Study of Diabetes RESEARCH DESIGN AND METHODS because of an adverse event were al- recommend metformin for initial drug Eligible patients at screening were $18 lowed to remain in the study for safety therapy (1). If alternative or combina- years of age with a BMI between 23 and follow-up. tion therapy is necessary, other oral 45 kg/m2 and glycosylated hemoglobin The primary outcome measure was agents such as a sulfonylurea, thiazolidi- A1c (HbA1c) between 7.0% and 11.0% change in HbA1c from baseline to 26 nedione, and dipeptidyl peptidase-4 (53–97 mmol/mol) on oral antihypergly- weeks. Secondary efficacy measures (DPP-4) inhibitor may be used (1). Al- cemic medication (OAM) monotherapy were change in HbA1c from baseline to though optimal second- and third-line or between 7.0% and 10.0% (53–86 52 weeks, percentage of patients with agents have not been firmly established, mmol/mol) on combination OAM ther- HbA1c ,7.0% (53 mmol/mol) or #6.5% when oral agents alone do not allow a apy. Patients were excluded from the (48 mmol/mol), changes in central lab- patient to achieve glycemic control, in- study if they were taking GLP-1 receptor oratory fasting serum glucose (FSG), jectable agents such as a GLP-1 receptor agonists during the 3 months before 8-point self-monitored plasma glucose agonist may be used. Among the avail- screening or were on long-term in- (SMPG) profiles, change in body weight, able GLP-1 receptor agonists, there are sulin therapy. The protocol was ap- and b-cell function and insulin sensitiv- differences in duration of action, fre- proved by local institutional review ity indices (updated HOMA2). quency of dosing, and efficacy and boards, and all patients provided written Safety assessments included adverse safety profiles (2–5). informed consent before participation events, hypoglycemic episodes, vital Dulaglutide is a long-acting human in the trial. The study was conducted signs, electrocardiograms, serial collec- GLP-1 receptor agonist in development in accordance with the Declaration of tion of laboratory parameters (hematol- as a once-weekly treatment for type 2 Helsinki guidelines on good clinical ogy, urinalysis, hepatobiliary analytes, diabetes (6,7). The molecule comprises practices (8). renal analytes, pancreatic enzymes, two identical disulfide-linked chains, Eligible patients entered a lead-in pe- and calcitonin), injection site reactions, each containing an N-terminal GLP-1 an- riod that lasted up to 12 weeks (Fig. 1A). and dulaglutide antidrug antibodies alog sequence covalently linked to a During this period, previous OAMs other (ADAs). Adjudication of pancreatic modified human IgG4 Fc fragment by a than metformin and pioglitazone were events was performed by an indepen- small peptide (6). In contrast to native discontinued, and patients were upti- dent clinical event classification group. GLP-1, dulaglutide is resistant to degra- trated on a dual-OAM regimen of maxi- The following events were adjudicated dation by DPP-4 and is large in size, mally tolerated metformin (1,500–3,000 to assess for possible development of which slows absorption and reduces re- mg/day; .2,550 mg/day allowed only in pancreatitis: investigator-reported pan- nal clearance. The molecular features certain countries outside the U.S. per creatitis, adverse events of serious result in a soluble formulation and a pro- country label) and pioglitazone (30–45 or severe abdominal pain without longed half-life of ;5days,makingit mg/day). Patients were then stabilized known cause, and confirmed cases of suitable for once-weekly subcutaneous for ;8 weeks before randomization, at asymptomatic elevations (three or more administration. Dulaglutide exhibits which time a qualifying HbA1c .6.5% times the upper limit of normal) in pan- GLP-1–mediated effects, including was required for ongoing eligibility. creatic enzymes. Laboratory analyses glucose-dependent potentiation of in- Patients were then randomized to were performed at a central laboratory sulin secretion, inhibition of glucagon one of four arms (2:2:2:1) of subcutane- (Quintiles). Immunogenicity testing was secretion, delay of gastric emptying, ous injections of once-weekly dulaglu- performed by BioAgilytix (Durham, NC) and weight loss. tide 1.5 mg or dulaglutide 0.75 mg, and Millipore (St. Louis, MO). It is important to understand the exenatide, or once-weekly placebo Hypoglycemia was defined as plasma benefits and risks of dulaglutide relative (Fig. 1A) according to a computer- glucose (PG) #70 mg/dL (#3.9 mmol/L) to other GLP-1 receptor agonists with generated random sequence using and/or symptoms or signs attributable differing pharmacological and clini- an interactive voice response system. to hypoglycemia. Severe hypoglycemia cal profiles. The purpose of AWARD-1 Exenatide-treated patients received was defined as an episode requiring (Assessment of Weekly AdministRation 5 mgBIDforthefirst 4 weeks and the assistance of another person to ac- of LY2189265 [dulaglutide] in Diabetes-1) 10 mg BID for the remainder of the tively administer therapy (9). was to compare once-weekly dulaglutide study. After 26 weeks, placebo-treated to placebo and exenatide twice daily patients were switched in a blinded Statistical Analyses (referred to hereafter as exenatide) fashion (1:1) to dulaglutide 1.5 mg or The study was designed with 90% power in patients with type 2 diabetes treated dulaglutide 0.75 mg (52-week data for to show superiority of dulaglutide with maximally tolerated doses of these patients are included in separate versus placebo and 93% power for non- metformin and pioglitazone. This study analyses not reported here). Randomi- inferiority versus exenatide on the allows a direct comparison of the effi- zation was stratified by country. An change from baseline in HbA1c at the cacy and safety profiles of a short-acting add-on rescue therapy was allowed for 26-week primary end point with an SD GLP-1 receptor agonist and a long-acting patients who met prespecified criteria of 1.3%, a one-sided a of 0.025, and a care.diabetesjournals.org Wysham and Associates 2161

patients achieving HbA1c targets (LOCF) was analyzed by using a logistic regression model, with treatment, country, and baseline as covariates. Total hypoglycemia included events that were documented symptomatic, documented asymptomatic, probable, and/ or severe (9). The percentage of patients experiencing adverse events was analyzed by using a x2 test unless there were no sufﬁcient data to meet the as- sumptions of the analysis, in which case a Fisher exact test was conducted. The two-sided signiﬁcance level was 0.05 for secondary end points and 0.10 for interactions.

RESULTS A total of 978 patients were randomized to dulaglutide 1.5 mg, dulaglutide 0.75 mg, exenatide, and placebo. Two patients assigned to exenatide did not re- ceive the drug; thus, the intention-to-treat population comprised 976 patients. De- mographic and baseline characteristics were balanced across all arms (Table 1). At randomization, 86% of patients were receiving $2,500 mg/day of metformin and 45 mg/day of pioglitazone, and the mean doses were similar across arms. A total of 77 (7.9%) patients discontinued the study at 26 weeks; the distribution of patients across treatment arms was Figure 1—Study design (A) and patient disposition (B). All patients underwent a metformin as follows: dulaglutide 1.5 mg, 19 – – (1,500 3,000 mg/day) and pioglitazone (30 45 mg/day) lead-in period that lasted up to 12 (6.8%); dulaglutide 0.75 mg, 17 (6.1%); weeks and was continued for the duration of the study; other OAMs were discontinued. Two doses of dulaglutide (1.5 mg and 0.75 mg) were evaluated along with exenatide and placebo. exenatide, 24 (8.7%); and placebo, 17 Placebo patients continued until week 26 and were then randomized to dulaglutide 1.5 mg or (12.1%). The most common reasons dulaglutide 0.75 mg. aNumber of patients rescued at week 26: dulaglutide 1.5 mg, 4 (1.4%); for study discontinuation at 26 weeks b dulaglutide 0.75 mg, 12 (4.3%); exenatide, 11 (4.0%); placebo, 22 (15.6%). Number of patients were adverse events and patient deci- rescued at week 52: dulaglutide 1.5 mg, 9 (3.2%); dulaglutide 0.75 mg, 25 (8.9%); exenatide, 24 (8.7%); placebo to dulaglutide 1.5 mg, 1 (1.6%); placebo to dulaglutide 0.75 mg, 3 (4.8%). sion (Fig. 1B). Disposition, including patients receiving rescue therapy, throughout the 52-week study period is noninferiority margin of 0.40%. This cor- only data collected before the initiation shown in Fig. 1B. responds to 280 patients per active of rescue medication were used. treatment arm and 140 for placebo, The change from baseline in HbA1c Efficacy with an assumed dropout rate of 11%. and weight at 26 and 52 weeks were The LS mean 6 SE HbA1c change from The type I error rate across all treatment analyzed by ANCOVA, with factors for baseline to the 26-week primary end comparisons for change from baseline in treatment, country, and baseline value point was 21.51 6 0.06% (216.5 6 HbA1c at 26 weeks was controlled at as covariates. The last observation car- 0.7 mmol/mol) for dulaglutide 1.5 mg, 0.025 (one-sided) by tree gatekeeping ried forward (LOCF) was used in the case 21.30 6 0.06% (214.2 6 0.7 mmol/mol) (10). P values were adjusted so that of missing data. Secondary analysis for dulaglutide 0.75 mg, 20.99 6 0.06% each could be compared with 0.025 to methods for HbA1c and weight and (210.8 6 0.7 mmol/mol) for exenatide, assess significance while accounting for methods for other continuous second- and 20.46 6 0.08% (25.0 6 0.9 multiplicity adjustments (11). ary end points over time included a mmol/mol) for placebo (Fig. 2A). Dula- The analyses of efficacy and safety mixed-effects, repeated-measures glutide 1.5 mg and dulaglutide 0.75 mg were based on the intent-to-treat pop- (MMRM) analysis, with additional fac- were superior to placebo (LS mean ulation comprising all randomized pa- tors for visit and treatment-by-visit in- difference; nominal 95% CI: 21.05% tients who received at least one dose teraction and the patient as a random [211.5 mmol/mol]; 21.22 to 20.88% of study treatment. For the assessment effect. Least squares (LS) means and [213.3 to 29.6 mmol/mol] vs. of efficacy and hypoglycemia events, SEs are reported. The percentage of 20.84% [29.2 mmol/mol]; 21.01 to 2162 Efficacy/Safety of Dulaglutide Versus Exenatide Diabetes Care Volume 37, August 2014

Table 1—Baseline characteristics and demographics of patients differences between dulaglutide 1.5 mg Dulaglutide 1.5 mg Dulaglutide 0.75 mg Exenatide Placebo and dulaglutide 0.75 mg versus exena- Variable (n =279) (n =280) (n =276) (n =141) tide were 218 mg/dL and 210 mg/dL, respectively (P , 0.001, both compari- Sex Male 163 (58) 168 (60) 156 (57) 83 (59) sons). At 52 weeks, both dulaglutide Female 116 (42) 112 (40) 120 (44) 58 (41) arms continued with significantly greater Age (years) 56 6 10 56 6 9556 10 55 6 10 changes from baseline in FSG com- , Race pared with exenatide (P 0.001, dula- Hispanic or Latino 93 (33) 102 (36) 91 (33) 45 (32) glutide 1.5 mg; P = 0.005, dulaglutide Not Hispanic or Latino 186 (67) 178 (64) 184 (67) 96 (68) 0.75 mg) (Fig. 2D). American Indian 40 (14) 37 (13) 38 (14) 20 (14) Figure 2E shows the mean of each PG Asian 6 (2) 8 (3) 4 (1) 6 (4) value from the 8-point SMPG profile at Black 24 (9) 24 (9) 18 (7) 10 (7) baseline and 26 weeks. The analysis of Multiple 3 (1) 3 (1) 3 (1) 2 (1) changes in the individual components of Native Hawaiian 1 (,1) 1 (,1) 1 (,1) 0 (0) fi White 205 (74) 207 (74) 211 (76) 103 (73) the daily blood glucose pro le demon- BMI (kg/m2)336 5336 6346 5336 6 strated that dulaglutide 1.5 mg and du- Weight 96 6 20 96 6 21 97 6 19 94 6 19 laglutide 0.75 mg were associated with a greater reduction in the mean of all pre- Diabetes duration (years) 9 6 696 596 696 6 meal PG compared with placebo and ex- HbA (%) 8.1 6 1.3 8.1 6 1.2 8.1 6 1.3 8.1 6 1.3 1c enatide (P , 0.001, both comparisons). HbA (mmol/mol) 65 6 14 65 6 13 65 6 14 65 6 14 1c All active treatment arms had signifi- FPG (mg/dL) 162 6 56 159 6 50 164 6 55 166 6 54 a cantly greater LS mean reductions in OAM treatment postprandial PG compared with placebo 1 OAM 55 (20) 67 (24) 76 (27) 44 (31) , 2 OAMs 155 (56) 142 (51) 135 (49) 62 (44) (P 0.001, all comparisons). Patients .2 OAMs 63 (23) 67 (24) 66 (24) 33 (23) receiving dulaglutide 1.5 mg had a sig- SBP (mmHg) 127 6 15 127 6 15 127 6 15 125 6 14 nificantly greater reduction in the mean DBP (mmHg) 81 6 9776 10 77 6 10 77 6 11 of all postprandial PG values compared with exenatide (P = 0.047). Patients re- Data are mean 6 SD or n (%) unless otherwise indicated. DBP, diastolic blood pressure; SBP, systolic blood pressure. aAt screening. ceiving dulaglutide 1.5 mg and exenatide demonstrated greater reductions in the mean of all 2-h postprandial PG excursions compared with placebo (P = 20.67% [211.0 to 27.3 mmol/mol], 0.75-mg arms (78% and 66%, respectively) 0.003, dulaglutide 1.5 mg; P , 0.001, respectively). Compared with exenatide, compared with exenatide (52%) (P , exenatide), with changes in the exenatide the mean changes from baseline were 0.001, both comparisons) and placebo group significantlygreaterthaninthe superior with dulaglutide 1.5 mg (43%) (P , 0.001, both comparisons) dulaglutide groups (P , 0.001, both (20.52% [25.7 mmol/mol]; 20.66 to (Fig. 2C). At the same time point, 63% comparisons). All three active treatment 20.39% [27.2 to 24.3 mmol/mol]) and 53% of patients receiving dulaglutide arms exhibited similar reductions in the and with dulaglutide 0.75 mg (20.31% 1.5 mg and dulaglutide 0.75 mg, respec- morning meal postprandial PG. Com- [23.4 mmol/mol]; 20.44 to 20.18% tively, achieved an HbA1c target of #6.5% pared with exenatide at the midday [24.8 to 22.0 mmol/mol]). (48 mmol/mol) compared with 38% in the meal, LS mean reductions in postpran- The LS mean HbA1c changes from exenatide arm and 24% in the placebo dial PG were significantly greater for du- baseline to 52 weeks were 21.36 6 arm (P , 0.001, all comparisons). At 52 laglutide 1.5 mg and dulaglutide 0.75 mg 0.08% (214.9 6 0.9 mmol/mol) for du- weeks, 57% and 48% of the dulaglutide (P , 0.001 and P = 0.049, respectively). laglutide 1.5 mg, 21.07 6 0.08% 1.5-mg and dulaglutide 0.75-mg patients, At the evening meal, LS mean reduction (211.7 6 0.9 mmol/mol) for dulaglu- respectively, achieved this target, com- in postprandial PG was significantly tide 0.75 mg, and 20.80 6 0.08% pared to 35% in the exenatide arm (Fig. 2C). greater for dulaglutide 1.5 mg than for (28.8 6 0.9 mmol/mol) for exenatide The majority of effects on FSG (mea- exenatide (P = 0.044). Results were sim- (Fig. 2A). Compared with exenatide, sured by the central laboratory) were ilar for active treatment arms at 52 weeks the LS mean changes from baseline observed within 2 weeks after random- (data not shown). were superior for dulaglutide 1.5 mg ization for all active treatment arms and The LS mean change in body weight (20.56% [26.1 mmol/mol]) and dulaglu- remained steady thereafter (Fig. 2D). (ANCOVA LOCF) from baseline to 26 tide 0.75 mg (20.27% [23.0 mmol/mol]; The LS mean FSG changes from baseline weeks was 21.30 6 0.29 kg for dulaglu- adjusted P , 0.001, both comparisons). to 26 weeks were 243 6 2mg/dLfor tide 1.5 mg, 0.20 6 0.29 kg for dulaglu- Figure 2B shows HbA1c values at baseline dulaglutide 1.5 mg, 234 6 2mg/dLfor tide 0.75 mg, 21.07 6 0.29 kg for and over time up to 52 weeks. dulaglutide 0.75 mg, 224 6 2mg/dL exenatide, and 1.24 6 0.37 kg for pla- At 26 weeks, the percentage of pa- for exenatide, and 25 6 3mg/dLfor cebo (Fig. 2F). Compared with placebo, tients attaining the HbA1c goal of ,7.0% placebo. All active treatment arms change in weight with dulaglutide 1.5 mg, (53 mmol/mol) was significantly higher in were associated with a greater decrease dulaglutide 0.75 mg, and exenatide the dulaglutide 1.5-mg and dulaglutide in FSG compared with placebo. LS mean was significantly different (P , 0.001, care.diabetesjournals.org Wysham and Associates 2163

Figure 2—Efficacy and safety measures through the treatment period. A: Change in HbA1c from baseline at 26 and 52 weeks, ANCOVA LOCF. B:HbA1c over time, MMRM. C: Percentage of patients achieving HbA1c targets, logistic regression. D: Change in FSG over time, MMRM. E: Baseline and 26-week 8-point SMPG profiles, MMRM (solid lines are baseline, and dashed lines are at 26 weeks). F: Change in weight over time, MMRM. G: Incidence of nausea up to 26 weeks. Data are LS mean 6 SE. ††P , 0.001, superiority vs. exenatide; ‡‡P , 0.001, superiority vs. placebo; #P , 0.05 vs. exenatide; *P , 0.05 vs. placebo; ##P , 0.001 vs. exenatide; **P , 0.001 vs. placebo. 2164 Efficacy/Safety of Dulaglutide Versus Exenatide Diabetes Care Volume 37, August 2014

P = 0.010, and P , 0.001, respectively). among the arms with respect to insulin natural causes in the dulaglutide 0.75 Compared with exenatide, the decrease sensitivity estimated by HOMA2-%S mg arm [patient had a history of cardio- in body weight was similar for dulaglu- (Supplementary Table 1). Patients vascular risk factors]) at 90 and 102 days tide 1.5 mg, and there was significantly treated with dulaglutide 1.5 mg demon- postrandomization, respectively. One greater weight gain for dulaglutide strated a significant mean reduction from patient who received dulaglutide 1.5 0.75 mg (LS mean difference: 20.24 kg baseline in total and LDL cholesterol lev- mg for 6 months died of pancreatic can- [P = 0.474] for dulaglutide 1.5 mg, els compared with placebo at 26 weeks cer 9 months after discontinuation from 1.27 kg [P , 0.001] for dulaglutide (Supplementary Table 1). These patients the study. 0.75 mg). The observed differences in also demonstrated a significant reduc- The incidence of adverse events weight between the dulaglutide groups tion in mean triglyceride levels com- was similar across arms (Table 2). and the exenatide group were main- paredwithexenatideat26and52 Gastrointestinal adverse events, includ- tained at 52 weeks (Fig. 2F). weeks and placebo at 26 weeks. No dif- ing nausea, vomiting, and diarrhea, Pancreatic b-cell function, as mea- ferences were observed among arms for were the most commonly reported in sured by HOMA2-%B at 26 weeks, in- change from baseline in mean HDL cho- dulaglutide- and exenatide-treated creased with all active treatment arms lesterol levels. patients; nausea and vomiting events compared with placebo and increased were significantly (P , 0.05, all com- more with dulaglutide 1.5 mg than Safety parisons) higher in dulaglutide- and with exenatide (P , 0.001, all compar- The incidence of serious adverse exenatide-treated patients than in isons). At 52 weeks, both dulaglutide events was similar across treatment placebo-treated patients at 26 weeks. arms had higher HOMA2-%B values ver- arms (Table 2). Two patients died during The incidence of these events was sus exenatide (P , 0.001, both compar- the study (one of myocardial infarction similar among patients receiving dula- isons). No differences were observed in the dulaglutide 1.5 mg arm and one of glutide 1.5 mg and exenatide and

Table 2—Safety assessments, change from baseline in vital signs, and TE dulaglutide ADAs 26 weeks 52 weeks Dulaglutide Dulaglutide Dulaglutide Dulaglutide 1.5 mg 0.75 mg Exenatide Placebo 1.5 mg 0.75 mg Exenatide Variable (n =279) (n =280) (n =276) (n =141) (n =279) (n =280) (n =276) Death 1 (0.4) 1 (0.4) 0 (0) 0 (0) 1 (0.4) 1 (0.4) 0 (0) Serious AEs, n (%) 12 (4) 15 (5) 15 (5) 12 (9) 18 (7) 22 (8) 27 (10) AEs (patients with $1 event) 215 (77) 199 (71) 198 (72) 104 (74) 226 (81) 220 (79) 221 (80) TE AEs ($5% patients) GI events 131 (47)** 83 (30)#* 117 (42)** 26 (18) 142 (51) 94 (34)# 128 (46 Nausea 78 (28)** 45 (16)#* 71 (26)** 8 (6) 81 (29) 47 (17)# 77 (28) Vomiting 47 (17)#** 17 (6)#* 30 (11)** 2 (1) 47 (17) 17 (6)# 33 (12) Diarrhea 31 (11) 22 (8) 16 (6) 8 (6) 36 (13) 26 (9) 21 (8) Dyspepsia 22 (8)* 5 (2)# 19 (7) 4 (3) 23 (8) 6 (2)# 20 (7) Constipation 12 (4) 5 (2) 5 (2) 2 (1) 16 (6)# 5 (2) 5 (2) Flatulence 14 (5) 3 (1) 6 (2) 3 (2) 16 (6) 3 (1) 7 (3) Infections and infestations 74 (27) 74 (26) 78 (28) 43 (31) 110 (39) 101 (36) 107 (39) Nasopharyngitis 18 (7) 23 (8) 12 (4) 8 (6) 27 (10) 26 (9) 16 (6) URI 12 (4) 14 (5) 12 (4) 6 (4) 15 (5) 23 (8) 19 (7) UTI 12 (4) 8 (3) 7 (3) 4 (3) 17 (6) 13 (5) 13 (5) Headache 20 (7) 9 (3) 24 (9) 8 (6) 26 (9) 14 (5) 24 (9) Fatigue 10 (4)# 12 (4) 21 (8)* 2 (1) 13 (5) 13 (5) 22 (8) Decreased appetite 22 (8)#* 14 (5) 8 (3) 3 (2) 23 (8)# 15 (5) 9 (3) Peripheral edema 3 (1) 13 (5) 11 (4) 7 (5) 8 (3) 15 (5) 17 (6) Back pain 11 (4) 9 (3) 8 (3) 9 (6) 15 (5) 13 (5) 12 (4) Dizziness 15 (5) 8 (3)# 18 (7)* 2 (1) 18 (7) 9 (3) 21 (8) Arthralgia 8 (3) 10 (4) 9 (3) 3 (2) 10 (4) 16 (6) 13 (5) Pain in extremity 6 (2) 8 (3) 8 (3) 6 (4) 11 (4) 13 (5) 13 (5) Discontinuation due to AE 8 (3) 4 (1) 9 (3) 3 (2) 9 (3) 4 (1) 10 (4) Safety parameters SBP (mmHg) 0.11 6 0.83* 20.36 6 0.82* 0.06 6 0.83* 3.40 6 1.13 0.83 6 0.87 1.62 6 0.85 0.02 6 0.88 DBP (mmHg) 0.76 6 0.55 0.56 6 0.54 20.11 6 0.55 1.25 6 0.75 0.89 6 0.57 0.76 6 0.57 0.02 6 0.58 Heart rate (beats/min) 2.80 6 0.52#* 2.80 6 0.51#* 1.18 6 0.52 0.61 6 0.70 1.68 6 0.56 1.56 6 0.55 1.15 6 0.56 TE ADAs Dulaglutide ADAs 4 (1.4) 2 (0.7) 12 (4.3) 2 (1.4) 5 (1.8) 3 (1.1) 14 (5.1) Exenatide ADAs dd75 (27.2) ddd58 (21.0) Data are n (%) and LS mean 6 SE. AE, adverse event; DBP, diastolic blood pressure; GI, gastrointestinal; SBP, systolic blood pressure; TE, treatment emergent; URI, upper respiratory infection; UTI, urinary tract infection. **P , 0.001 vs. placebo. #P , 0.05 vs. exenatide. *P , 0.05 vs. placebo. care.diabetesjournals.org Wysham and Associates 2165

significantly (P , 0.05, all comparisons) above the upper limit of normal for pan- and the exenatide arms of ;0.3%–0.5% lower in those receiving dulaglutide 0.75 creatic enzymes were similar for the ac- was achieved with a similar or lower risk mg after 52 weeks. The majority of the tive treatments compared with placebo of hypoglycemia, indicating an accept- events were mild to moderate in sever- at 26 weeks and among the active treat- able benefit and hypoglycemic risk pro- ity. Nausea was primarily transient, with ments at 52 weeks (Supplementary Table file for this new, once-weekly GLP-1 new-onset cases occurring primarily in 1). Calcitonin values remained stable receptor agonist. the first 2 weeks of treatment with throughout the study in all treatment At randomization, the majority (86%) both doses of dulaglutide (Fig. 2G). arms. of patients tolerated the uptitration Discontinuations due to adverse There were no clinically relevant to maximum approved doses of both events were similar across treatment changes in LS mean systolic blood pres- metformin and pioglitazone and had a arms at 26 and 52 weeks (Fig. 1B). The sure among the three active treatment study-qualifying HbA1c .6.5% (48 most common adverse event leading to arms at 26 or 52 weeks (Table 2); in the mmol/mol) after 8 weeks of stabiliza- discontinuation was nausea (three for placebo group, there was an increase in tion, demonstrating that the study was dulaglutide 1.5 mg, one for dulaglutide systolic blood pressure of 3.40 mmHg. conducted in a patient population that 0.75 mg, and four for exenatide). One There were no differences observed was appropriate for the addition of a patient from the dulaglutide 1.5 mg among arms for change in diastolic third antihyperglycemic agent. In this arm was given a diagnosis of chronic blood pressure at 26 or 52 weeks (Table population, patients receiving dulaglu- pancreatitis ;7monthsafterstudy 2). Dulaglutide 1.5 mg and dulaglutide tide 1.5 mg achieved a further 1.5% drug initiation. This patient had no signs 0.75 mg were associated with signifi- HbA1c reduction from the baseline or symptoms of pancreatitis before cantly (P , 0.05, all comparisons) mean HbA1c of 8.1% (65 mmol/mol), study initiation but had transient eleva- greater LS mean increases in heart rate with 78% achieving the goal of ,7.0% tions in pancreatic enzymes starting at at 26 weeks compared with exenatide (53 mmol/mol); both results are superior baseline and continuing throughout and placebo; no differences between to exenatide. The previously reported the study, including during the 6 months dulaglutide and exenatide were noted similar glycemic effect of dulaglutide after study drug discontinuation when at 52 weeks (Table 2). (12) and exenatide (13,14) compared the patient was allowed to remain in Ten (1.8%) patients randomized with placebo provides additional sup- the study but off the study drug. to dulaglutide developed treatment- port for the current results. A total of 108 patients (dulaglutide emergent dulaglutide ADAs at least The impact of dulaglutide and exena- 1.5 mg, 10.4%; dulaglutide 0.75 mg, once postbaseline during the 52-week tide on glucose control was evident 10.7%; exenatide, 15.9%; placebo, study (Table 2). Another three patients early in the course of therapy, with a 3.5%) experienced hypoglycemia during who were randomized to placebo devel- near-maximal decrease in FSG observed the first 26 weeks, with significantly oped treatment-emergent dulaglutide as early as 2 weeks after the initiation of fewer patients in the dulaglutide 1.5 ADAs after switching to dulaglutide therapy and a significantly greater mag- mg arm compared with the exenatide at 26 weeks. One patient treated with nitude of effect with dulaglutide com- arm (P =0.007).Themean1-yearad- dulaglutide 1.5 mg with a treatment- pared with exenatide. Although both justed rates of total hypoglycemia emergent dulaglutide ADA experienced GLP-1 receptor agonists improved pre- were 0.45, 1.10, 1.47, and 0.37 events/ local injection site erythema and swelling prandial and postprandial blood glucose patient/year for dulaglutide 1.5 mg, du- for 1 day. In exenatide-treated pa- control as measured on 8-point SMPG laglutide 0.75 mg, exenatide, and pla- tients at 26 weeks, 48% were noted profiles,thereweresomenotabledif- cebo, respectively, at 26 weeks. The to have treatment-emergent exenatide ferences. Changes from baseline in the incidences and rates of total hypoglyce- ADAs (Table 2) of whom one experienced mean of all preprandial and postpran- mia remained lower for dulaglutide 1.5 an injection site reaction. No patients dial PG values were greater with dula- mg than for exenatide at 52 weeks. reported systemic hypersensitivity glutide relative to exenatide. Exenatide There were no events of severe hypogly- reactions. wasassociatedwithasmallermean cemia among dulaglutide-treated pa- postprandial excursion compared with tients, and two events were reported CONCLUSIONS dulaglutide likely because of the higher for exenatide-treated patients. The results of the AWARD-1 study dem- absolute premeal glycemic level with ex- Small median increases in serum li- onstrate that once-weekly dulaglutide enatide. Glycemia is the main factor pase, total amylase, and pancreatic am- in combination with maximally tolerated influencing insulin secretion rates in ylase (p-amylase) that remained within doses of metformin and pioglitazone re- b-cells exposed to a GLP-1 receptor ag- normal range were observed for dula- sults in significantly larger improvements onist; therefore, patients with near- glutide and exenatide; these changes in HbA1c and percentages of patients normal glycemia require less insulin to were significant (P , 0.05, all com- achieving target HbA1c goals compared be secreted after the meal to maintain parisons) compared with placebo with placebo and the active comparator blood glucose levels within the physio- (Supplementary Table 1). Increases in exenatide at 26 weeks. Additionally, du- logic range, which is believed to result in pancreatic enzymes were greater for laglutide 1.5 mg was associated with sig- the observed difference in glucose ex- dulaglutide 1.5 mg than for exenatide nificant weight reduction compared with cursions in the current study. at 26 weeks and were greater for total placebo. Importantly, the clinically rele- Weight loss was similar for dulaglu- amylase and p-amylase at 52 weeks. vant mean difference in HbA1c change tide 1.5 mg and exenatide, despite Incidences of treatment-emergent values from baseline between the dulaglutide the greater reduction of HbA1c with 2166 Efficacy/Safety of Dulaglutide Versus Exenatide Diabetes Care Volume 37, August 2014

dulaglutide 1.5 mg and known weight in clinical practice. The study was per- Sessions of the American Diabetes Association, effects of background thiazolidinedione formed in Mexico, Argentina, and the Chicago, IL, 21–25 June 2013; the Brazilian Di- therapy over time (15,16). Dulaglutide U.S. in a population that was primarily abetes Society 19th Annual Congress 2013, Flo- rianopolis,´ Brazil, 9–11 October 2013; and the 0.75 mg did not have the same weight white and Hispanic. 49th Annual Meeting of the European Associa- loss effect as dulaglutide 1.5 mg and ex- Overall, once-weekly dulaglutide tion for the Study of Diabetes, Barcelona, Spain, enatide, which may indicate a greater therapy is efficacious and safe in com- 23–27 September 2013. GLP-1 receptor agonist concentration bination with metformin and pioglita- requirement to achieve the weight loss zone. Dulaglutide is superior to placebo References 1. Inzucchi SE, Bergenstal RM, Buse JB, et al.; observed with dulaglutide 1.5 mg. These and exenatide with respect to HbA1c results are consistent with the magni- change from baseline and the percent- American Diabetes Association (ADA); Euro- pean Association for the Study of Diabetes tude of weight loss observed in other age of patients achieving glycemic (EASD). Management of hyperglycemia in type studies evaluating GLP-1 receptor ago- targets. Additionally, the observed rapid 2 diabetes: a patient-centered approach: posi- nists on a background therapy of met- improvement in fasting glucose and tion statement of the American Diabetes Asso- formin and thiazolidinedione (17,18). SMPG values, with an attendant low ciation (ADA) and the European Association for The safety profile of dulaglutide in risk of hypoglycemia, represents an im- the Study of Diabetes (EASD). Diabetes Care 2012;35:1364–1379 this trial is generally consistent with portant treatment profile in the man- 2. Fineman MS, Bicsak TA, Shen LZ, et al. Effect the known effects of the GLP-1 receptor agement of patients with type 2 on glycemic control of exenatide (synthetic ex- agonist class. Patients treated with du- diabetes. endin-4) additive to existing metformin and/or laglutide 1.5 mg and exenatide reported sulfonylurea treatment in patients with type 2 diabetes. Diabetes Care 2003;26:2370–2377 similar incidences of gastrointestinal ad- 3. Knudsen LB, Nielsen PF, Huusfeldt PO, et al. verse events, with less frequent report- Acknowledgments. The authors thank the Potent derivatives of glucagon-like peptide-1 ing by patients treated with dulaglutide AWARD-1 team and staff for the conduct of with pharmacokinetic properties suitable for 0.75 mg. The incidence of nausea and this study and Oralee Johnson Varnado (Eli Lilly once daily administration. J Med Chem 2000; – vomiting also appeared to be similar to and Company) for editing assistance. 43:1664 1669 Duality of Interest. This work is sponsored by 4. Rosenstock J, Reusch J, Bush M, Yang F, the incidence observed with liraglutide Eli Lilly and Company. C.W. is an advisory board Stewart M; Albiglutide Study Group. Potential when used with similar background member for Boehringer Ingelheim, Eli Lilly and of albiglutide, a long-acting GLP-1 receptor ag- therapy (18), with nausea reported by Company, Janssen, and Sanofi U.S.; has received onist, in type 2 diabetes: a randomized con- 29%and40%ofpatientstreatedwith institutional grants for research from AstraZeneca, trolled trial exploring weekly, biweekly, and – liraglutide 1.2 mg and 1.8 mg and vomit- Bristol-Myers Squibb, Boehringer Ingelheim, Eli monthly dosing. Diabetes Care 2009;32:1880 Lilly and Company, Intarcia, Jaeb Center for 1886 ing by 7% and 17%, respectively. Overall, Health Research, Janssen, Merck, Novo Nordisk, 5. Werner U, Haschke G, Herling AW, Kramer fewer dulaglutide-treated patients ex- and Sanofi U.S.; and has been a speakers’ bureau W. Pharmacological profile of lixisenatide: perienced hypoglycemia compared member for Bristol-Myers Squibb, Boehringer a new GLP-1 receptor agonist for the treatment with exenatide-treated patients. Cardio- Ingelheim, Eli Lilly and Company, Janssen, Novo of type 2 diabetes. Regul Pept 2010;164:58–64 fi vascular assessments showed an in- Nordisk, and Sano U.S. T.B. has received 6. Barrington P, Chien JY, Tibaldi F, Showalter research grants from Amylin Pharmaceuticals, HD, Schneck K, Ellis B. LY2189265, a long-acting crease in systolic blood pressure with Inc., and served as an advisor and speakers’ glucagon-like peptide-1 analogue, showed a placebo compared with the dulaglutide bureau member for Amylin Pharmaceuticals, dose-dependent effect on insulin secretion in and exenatide arms, which may be Inc., and Eli Lilly and Company. R.A. has re- healthy subjects. Diabetes Obes Metab 2011; partly a result of the increase in weight ceived institutional grants for research from 13:434–438 GlaxoSmithKline, Novartis Pharmaceuticals, Eli 7. Glaesner W, Vick AM, Millican R, et al. Engi- observedintheplaceboarm.Anin- Lilly and Company, Reata Pharmaceuticals, and neering and characterization of the long-acting crease in heart rate was observed with Sanofi; has received consulting fees from Novo glucagon-like peptide-1 analogue LY2189265, dulaglutide and exenatide and was Nordisk; and has received speaker’sfees an Fc fusion protein. Diabetes Metab Res Rev similar to changes observed within the from Amylin Pharmaceuticals, Inc., AstraZeneca, 2010;26:287–296 GLP-1 receptor agonist class (19,20). and Bristol-Myers Squibb. G.C. has received con- 8. World Medical Association declaration of sulting fees from Eli Lilly and Company, Sanofi, Helsinki. Recommendations guiding physicians There were no clinical adverse events Quintiles, and ICON. P.G. has served as a board in biomedical research involving human sub- identified based on serial evaluations member for Novo Nordisk, Merck Sharp & jects. JAMA 1997;277:925–926 of thyroid and pancreatic laboratory pa- Dohme, and Eli Lilly and Company and has re- 9. Workgroup on Hypoglycemia, American Di- rameters. The immunogenicity of dula- ceived speaker’s fees from Bristol-Myers Squibb, abetes Association. Defining and reporting hy- glutide appeared to be low, with ,2% Eli Lilly and Company, Novo Nordisk, Amgen Inc., poglycemia in diabetes: a report from the and GlaxoSmithKline. C.A., D.K., and M.L. are American Diabetes Association Workgroup on of patients developing treatment- employees of Eli Lilly and Company. No other Hypoglycemia. Diabetes Care 2005;28:1245– emergent dulaglutide ADAs in contrast potential conflicts of interest relevant to this 1249 to a 47% incidence in patients treated article were reported. 10. Dmitrienko A, Tamhane AC, Wiens BL. Gen- with exenatide. Author Contributions. C.W., T.B., R.A., G.C., eral multistage gatekeeping procedures. Biom J – Limitations of the clinical application and P.G. researched data and contributed to the 2008;50:667 677 discussion and review and editing of the man- 11. Westfall PHYS. Resampling-Based Multiple of these results include the unforeseen uscript. C.A., D.K., and M.L. researched data and Testing: Examples and Methods for P-Value Ad- decrease in the use of high-dose thiazo- contributed to the writing of the manuscript. justment. New York, Wiley, 1993 lidinedione therapy during the course of M.L. is the guarantor of this work and, as such, 12. Umpierrez GE, Blevins T, Rosenstock J, the study. During the lead-in period, had full access to all the data in the study and Cheng C, Anderson JH, Bastyr EJ 3rd; EGO takes responsibility for the integrity of the data Study Group. The effects of LY2189265, a long- there was forced titration of metformin and the accuracy of the data analysis. acting glucagon-like peptide-1 analogue, in a and pioglitazone to maximally tolerated Prior Presentation. Parts of this study were randomized, placebo-controlled, double-blind doses, which may not always be routine presented in abstract form at the 73rd Scientific study of overweight/obese patients with type care.diabetesjournals.org Wysham and Associates 2167

2 diabetes: the EGO study. Diabetes Obes with type 2 diabetes in the PROactive study thiazolidinedione in patients with type 2 diabe- Metab 2011;13:418–425 (PROspective pioglitAzone Clinical Trial In tes (LEAD-4 Met+TZD). Diabetes Care 2009;32: 13. Buse JB, Henry RR, Han J, Kim DD, Fineman macroVascular Events): a randomised controlled 1224–1230 MS, Baron AD; Exenatide-113 Clinical Study trial. Lancet 2005;366:1279–1289 19. Pratley RE, Nauck M, Bailey T, et al.; 1860- Group. Effects of exenatide (exendin-4) on gly- 16.KahnSE,HaffnerSM,HeiseMA,etal.; LIRA-DPP-4 Study Group. Liraglutide versus si- cemic control over 30 weeks in sulfonylurea- ADOPT Study Group. Glycemic durability of ro- tagliptin for patients with type 2 diabetes who treated patients with type 2 diabetes. Diabetes siglitazone, metformin, or glyburide monother- did not have adequate glycaemic control with Care 2004;27:2628–2635 apy. N Engl J Med 2006;355:2427–2443 metformin: a 26-week, randomised, parallel- 14. DeFronzo RA, Ratner RE, Han J, Kim DD, 17. Drucker DJ, Buse JB, Taylor K, et al.; group, open-label trial. Lancet 2010;375: Fineman MS, Baron AD. Effects of exenatide DURATION-1 Study Group. Exenatide once 1447–1456 (exendin-4) on glycemic control and weight weekly versus twice daily for the treatment of 20. Pratley RE, Nauck MA, Bailey T, et al.; 1860- over 30 weeks in metformin-treated patients type 2 diabetes: a randomised, open-label, non- LIRA-DPP-4 Study Group. Efﬁcacy and safety of with type 2 diabetes. Diabetes Care 2005;28: inferiority study. Lancet 2008;372:1240–1250 switching from the DPP-4 inhibitor sitagliptin to 1092–1100 18. Zinman B, Gerich J, Buse JB, et al.; LEAD-4 the human GLP-1 analog liraglutide after 52 weeks 15. Dormandy JA, Charbonnel B, Eckland DJ, Study Investigators. Efﬁcacy and safety of the in metformin-treated patients with type 2 diabe- et al.; PROactive Investigators. Secondary pre- human glucagon-like peptide-1 analog liraglu- tes: a randomized, open-label trial. Diabetes Care vention of macrovascular events in patients tide in combination with metformin and 2012;35:1986–1993