Diabetes Care Volume 37, October 2014 2723

Lawrence A. Leiter,1 Molly C. Carr,2 Efficacy and Safety of the Once- Murray Stewart,2 Angela Jones-Leone,2 CARE/EDUCATION/NUTRITION/PSYCHOSOCIAL CLIN Rhona Scott,3 Fred Yang,2 and Weekly GLP-1 Receptor Agonist Yehuda Handelsman4 Versus in Patients With Type 2 and Renal Impairment: A Randomized Phase III Study Diabetes Care 2014;37:2723–2730 | DOI: 10.2337/dc13-2855

OBJECTIVE To evaluate weekly subcutaneous albiglutide versus daily sitagliptin in renally impaired patients with and inadequately controlled glycemia on a regimen of diet and exercise and/or oral antihyperglycemic .

RESEARCH DESIGN AND METHODS In this phase III, randomized, double-blind, multicenter, 52-week study, the pri- mary study end point was HbA1c change from baseline at week 26 in patients with renal impairment, as assessed with estimated glomerular filtration rate and cat- egorized as mild, moderate, or severe (‡60 to £89, ‡30 to £59, and ‡15 to £29 mL/min/1.73 m2, respectively). Secondary end points included fasting plasma glucose (FPG), weight, achievement of treatment targets, hyperglycemic rescue, and safety. 1Division of Endocrinology and Metabolism, Li Ka RESULTS Shing Knowledge Institute and Keenan Research Centre for Biomedical Science, St. Michael’s Hos- Baseline demographics were similar across treatment and renal impairment pital, University of Toronto, Toronto, Ontario, groups with overall mean age of 63.3 years, BMI of 30.4 kg/m2,HbA of 8.2% Canada 1c 2 (66 mmol/mol), and diabetes disease duration of 11.2 years. HbA change from GlaxoSmithKline, King of Prussia, PA 1c 3GlaxoSmithKline, Uxbridge, Middlesex, U.K. baseline at week 26 was significantly greater for albiglutide than sitagliptin 4Metabolic Institute of America, Tarzana, CA 2 2 P ( 0.83% vs. 0.52%, = 0.0003). Decreases in HbA1c, FPG, and weight were seen Corresponding author: Lawrence A. Leiter, through week 52. Time to hyperglycemic rescue through week 52 was significantly [email protected]. longer for albiglutide than sitagliptin (P = 0.0017). Results of safety assessments Received 5 December 2013 and accepted 23 June were similar between groups, and most adverse events (AEs) were mild or mod- 2014. erate. The incidences of gastrointestinal AEs for albiglutide and sitagliptin were as Clinical trial reg. no. NCT01098539, clinicaltrials follows: overall, 31.7%, 25.2%; diarrhea, 10.0%, 6.5%; nausea, 4.8%, 3.3%; and .gov. vomiting, 1.6%, 1.2%, respectively. This article contains Supplementary Data online at http://care.diabetesjournals.org/lookup/ CONCLUSIONS suppl/doi:10.2337/dc13-2855/-/DC1. Once-weekly albiglutide therapy in renally impaired patients with type 2 diabetes © 2014 by the American Diabetes Association. Readers may use this article as long as the work provided statistically superior glycemic improvement with almost similar tolera- is properly cited, the use is educational and not bility compared with daily sitagliptin therapy. for profit, and the work is not altered. 2724 Albiglutide in Type 2 Diabetes Renal Impairment Diabetes Care Volume 37, October 2014

The steady rise in the prevalence of type in patients with type 2 diabetes and re- related to pancreatitis, and other events 2 diabetes worldwide has become a ma- nal impairment whose glycemia was in- identified by the pharmacovigilance jor health care issue due to associated adequately controlled on their current group as related to pancreatitis, as well increases in patient morbidity and mor- regimen of either diet and exercise or as amylase and/or lipase measurements tality secondary to cardiovascular, renal, specific antihyperglycemic therapy. (three or more times the upper limit of and neurological disease complications normal) regardless of suspected etiology, and the rising monetary and resource RESEARCH DESIGN AND METHODS to determine whether adjudication for costsneededtomanagethesecompli- Study Design pancreatitis was warranted. cations. Renal insufficiency is a com- This phase III, randomized, double- After randomization, patients who mon comorbidity in patients with type blind, active-controlled, two parallel- experienced persistent hyperglycemia 2 diabetes, and the prevalence of group, multicenter, 52-week study (Har- qualified, on the basis of prespecified chronic kidney disease (CKD) in pa- mony 8) (GlaxoSmithKline study number HbA1c and/or fasting plasma glucose tients with type 2 diabetes has been GLP114130) compared the efficacy and (FPG) values, to undergo masked dose increasing (1), with the reported over- safety of once-weekly subcutaneous in- titration followed by hyperglycemia res- all incidence of renal impairment in pa- jections of albiglutide with that of daily cue as needed (Supplementary Table 2). tients with type 2 diabetes ranging oral sitagliptin in renally impaired pa- Given the blinded nature of the study, from15to35%(basedonestimated tients with type 2 diabetes. Renal impair- uptitration also occurred in the sitaglip- glomerular filtration rate [eGFR] ,60 ment was defined as mild (eGFR $60 to tin group, but the actual dose received mL/min/1.73 m2)(2–4). #89 mL/min/1.73 m2), moderate (eGFR did not change. Patients continued to Current treatment options for glyce- $30 to #59 mL/min/1.73 m2), or severe receive study in a blinded mic control are significantly limited for (eGFR $15 to #29 mL/min/1.73 m2) fashion after rescue. patients with type 2 diabetes and CKD based on the MDRD formula (22). Albiglu- The primary end point of the study was (5,8–11), which often leads to the in- tide (30 mg) was given subcutaneously to evaluate the efficacy of albiglutide troduction of sulfonylurea and once weekly (with treatment-masked compared with that of sitagliptin on the therapy to maintain glycemic control. uptitration, if needed, to 50 mg weekly), change in HbA1c from baseline at week Further, therapeutic options for glyce- and sitagliptin was dosed based on the 26. Secondary end points over time in- mic control can be associated with sig- eGFR value at randomization per the cluded HbA1c, FPG, body weight, propor- nificant risks of , weight sitagliptin package insert. All patients tion of patients who met prespecified gain, and fluid retention, which may continued to receive their prescribed HbA1c treatment targets, time to hy- add to the complexity of maintaining oral antihyperglycemic medication reg- perglycemic rescue, and population blood glucose, body weight, and blood imen (, , pharmacokinetics of albiglutide. Pharma- pressure in this population (12–15). sulfonylurea, or any combination of cokinetics are not reported here, as a Albiglutide, a novel, long-acting glu- these oral antihyperglycemic medica- population model that combined data cagon-like peptide 1 receptor (GLP-1R) tions) for the duration of the study from this and three other phase III studies agonist, was designed to retain the ther- with the exception of patients with was reported separately (23). Safety end apeutic actions of GLP-1 while having a GFR ,60 mL/min/1.73 m2,whowere points over time included evaluations greatly extended duration of action. Al- washed off their background metfor- of AEs and SAEs, safety events of special biglutide was synthesized through ge- min. Instructions for downtitration of interest, clinical laboratory parameters, netic modification that resulted in the sulfonylureas were also provided to vital sign measurements, electrocardio- attachment of two modified recombi- avoid hypoglycemia. gram readings, and physical examinations. nant human GLP-1 fragments linked in There were four specific study peri- Immunogenicity was also evaluated. tandem to the amino terminus of the ods: prescreening and screening (~2 There were no changes to the statistical coding sequence for human albumin, weeks), run-in (4 weeks), treatment pe- plan, which was finalized before the da- and it retains GLP-1 glucose-dependent riod (52 weeks, including 26 weeks of tabase was frozen and treatment codes insulinotropic activities both in vitro and initial treatment and evaluation for pri- were unblinded. in vivo (16). With a half-life of approxi- mary efficacy and safety followed by an This study was conducted in accor- mately 5 days, the pharmacokinetic additional 26 weeks of treatment for dance with International Conference profile of albiglutide allows for once- secondary efficacy and safety assess- on Harmonisation Good Clinical Practi- weekly subcutaneous injections (16–20). ments), and posttreatment follow-up ces and the Declaration of Helsinki, and To date, reported clinical studies, in- (8 weeks). Patient assessments during all patients provided written informed cluding a phase I study in patients with the treatment period varied between consent before they participated in this varying degrees of renal impairment weekly and quarterly. Independent, study. (GLP108370), have shown albiglutide to blinded adjudication and review of all be an effective, safe, and tolerable ther- suspected cardiovascular events took Patients apy when administered to patients with place over the entire treatment period. Male and nonpregnant, nonlactating fe- type 2 diabetes (16–21). An independent, blinded pancreatitis male patients $18 years of age with his- This study was conducted to compare adjudication committee reviewed all re- torical diagnoses of type 2 diabetes were the efficacy, safety, and tolerability of ported adverse events (AEs) of pancreati- enrolled. Patients were required to have albiglutide given subcutaneously once tis, serious AEs (SAEs) identified through baseline HbA1c between 7.0 and 10.0% weekly with sitagliptin given orally daily broad and narrow SMQ searches as (53–86 mmol/mol), BMI between 20 and care.diabetesjournals.org Leiter and Associates 2725

45 kg/m2, fasting C-peptide level of $0.8 albiglutide versus sitagliptin had at least rescue and the proportion of patients ng/mL (0.26 nmol/L), GFR of $15 to ,90 91% power according to a one-sided, who received hyperglycemia rescue. mL/min/1.73 m2, hemoglobin of $10 two-sample t test and a test-wise signif- The safety analyses included compara- g/dL for male patients and $9g/dL icance level of 0.025, with 200 com- tive summaries of vital sign measure- for female patients, and normal levels pleted patients per treatment group ments, laboratory values, physical of thyroid-stimulating hormone or clin- for a noninferiority margin of 0.4. With examination assessments, electrocar- ically euthyroid. the significance on the noninferiority diogram readings, and on-therapy AE in- Patients with malignant disease (except hypothesis, the superiority hypothesis cidence rates, with events “on therapy” squamous cell or basal cell carcinoma) had at least 90% power to reject the defined as events that occurred within were excluded from the study. Further null hypothesis if the actual albiglutide 56 days of treatment regardless of exclusion criteria included a history of di- superiority was 0.35% with a two-sided, rescue. abetic gastroparesis, current ongoing two-sample t test and a test-wise signif- All analyses were conducted with SAS symptomatic biliary disease or history of icance level of 0.05. Allowing for as great software (SAS Institute, Cary, NC), pancreatitis, significant gastrointestinal as 20% early withdrawal, loss to glyce- version 9.1 or higher. (GI) surgery or surgeries thought to signif- mic follow-up, and rescue, we random- icantly affect upper GI function, recent ized 250 patients to each treatment RESULTS (within predefined time scales) clinically group. Patients significant cardiovascular and/or cere- The primary analysis of the change in This study was conducted at 134 centers brovascular disease, a history of human HbA1c from baseline response at week in 15 countries between 7 May 2010 and immunodeficiency virus infection, and 26 was applied to the intent-to-treat 30 May 2012. A total of 771 patients acute symptomatic hepatitis B or C in- population using an ANCOVA model were assessed for eligibility, and 507 pa- fection. Patients who met additional ex- with main effects for treatment group, tients were randomly assigned to re- clusion criteria, including requirements region, renal impairment severity, his- ceive albiglutide (n = 254) or sitagliptin for levels of total bilirubin, alanine tory of prior myocardial infarction, and (n = 253) (Supplementary Fig. 1). At least aminotransferase, aspartate amino- age category as factors and baseline 97% of patients in each group received transferase, amylase, lipase, or fasting HbA1c as a continuous covariate. Treat- one or more doses of study medication. triglycerides, were also excluded. ment-effect estimates (and associated Overall dropout rates by week 52 were hypothesis tests) of albiglutide were ~20% in the albiglutide group and 25% Randomization and Blinding evaluated within this ANCOVA model in the sitagliptin group. Approximately 500 patients were planned as least squares (LS) means contrasts Patient demographic and baseline to be randomly assigned in a ratio of 1:1 relative to sitagliptin. Patients who qual- characteristics were similar between to receive albiglutide plus sitagliptin ified for hyperglycemia rescue had their the albiglutide and sitagliptin treatment matching placebo or albiglutide matching primary end point value for change in groups. Of the assessed characteristics, placebo plus sitagliptin. Eligible patients HbA from baseline recorded at the only HbA category showed a statisti- were stratified according to severity of 1c 1c time of rescue. Follow-up assessments cally significant difference between renal impairment (mild, moderate, or continued beyond rescue. For patients groups (P , 0.05); however, this differ- severe), prior history of myocardial infarc- who withdrew consent, were lost to ence was not considered to be clinically tion (yes or no), and age (,65 or $65 follow-up, or otherwise discontinued relevant and was likely due to the higher years of age). An interactive voice re- participation, the last observation car- mean HbA at baseline in the sitagliptin sponse system was used for the blinded 1c ried forward (LOCF) method was used group (Table 1). Overall, ~52%, 41%, and randomization, which was based on a se- for the primary analysis of HbA .The 7% of patients had mild, moderate, or questered fixed randomization schedule. 1c ANCOVA analyses were also performed severe renal impairment, respectively. Albiglutide and matching placebo was on HbA change from baseline by visit Approximately 57% of patients (141 supplied as a fixed-dose (30 or 50 mg) 1c up to week 52 with the observed case of 249) on albiglutide required dose up- pen injector system, which was injected (OC) algorithm, which only included pa- titration to 50 mg, with 35% undergoing subcutaneously into the abdomen. Sita- tients with data at the specified analysis uptitration by week 26 and an addi- gliptin (100 mg for patients with eGFR time point and excluded data carried for- tional22.1%byweek48.Themeanal- $50–89 mL/min/1.73 m2, 50 mg for those ward from earlier observations as was biglutide dose was 40.2 mg at week 26 with eGFR $30 to ,50 mL/min/1.73 m2, done under the LOCF procedure. For and 42.4 mg at week 52. The propor- and 25 mg for those with eGFR ,30 this report, LOCF analyses that excluded tion of albiglutide-treated patients mL/min/1.73 m2) and matching placebo posthyperglycemic rescue values were who required uptitration from 30 mg were provided as overcoated tablets or cap- used for the week 26 data and OC anal- weekly to 50 mg weekly was similar sules. No dose adjustments were made for yses that excluded posthyperglycemic among the mild (53% [69 of 128 pa- degrees of renal impairment after random- rescue values were used for the week tients]), moderate (60% [61 of 102 pa- ization. Treatment compliance was as- 52 data unless otherwise stated. tients]), and severe (58% [11 of 19 sessed for both treatment groups through The other continuous secondary effi- patients]) renal impairment subgroups. the return of unused study medication. cacy end points were analyzed in a fash- Treatment compliance was 98% for Sample Size and Statistical Analyses ion similar to that used for the primary albiglutide and 93% for sitagliptin. Com- Based on an expected treatment effect end point. Fifty-two week efficacy was pliance rates ,80% were lower in the of 0% and an SD of 1.2%, the test of evaluated using time to hyperglycemia albiglutide (1.6%) and matching placebo 2726 Albiglutide in Type 2 Diabetes Renal Impairment Diabetes Care Volume 37, October 2014

Table 1—Patient demographics and baseline characteristics by treatment group maintained through week 52 (Supple- (safety population) mentary Fig. 2A). Albiglutide Sitagliptin Total P* FPG N 249 246 495 The change in FPG from baseline at Severity of renal impairment, n (%) 0.9522 week 26 was 21.42 mmol/L in the albi- Mild 128 (51.4) 128 (52.0) 256 (51.7) glutide group and 20.22 mmol/L in the Moderate 102 (41.0) 101 (41.1) 203 (41.0) Severe 19 (7.6) 17 (6.9) 36 (7.3) sitagliptin group (Fig. 1B). At week 26, the difference in LS means (albiglutide Age (years), mean (SD) 63.2 (8.37) 63.5 (9.02) 63.3 (8.69) 0.733 2 , ,65 141 (56.6) 138 (56.1) 279 (56.4) 0.906 vs. sitagliptin) was 1.20 mmol/L (P $65 108 (43.4) 108 (43.9) 216 (43.6) 0.0001). Both treatment groups main- Sex, n (%) 0.692 tained decreases in FPG through week Female 113 (45.4) 116 (47.2) 229 (46.3) 52, with the mean change from baseline Male 136 (54.6) 130 (52.8) 266 (53.7) in FPG showing greater reductions for Race, n (%)† albiglutide than for sitagliptin at each Black 36 (14.5) 42 (17.1) 78 (15.8) time point through week 52 (Supple- Asian 84 (33.7) 76 (30.9) 160 (32.3) mentary Fig 2B). At week 52, both albi- White 113 (45.4) 114 (46.3) 227 (45.8) glutide and sitagliptin demonstrated Other 17 (6.8) 16 (6.5) 33 (6.7) sustained glycemic effects (Supplemen- Diabetes duration (years), mean (SD) 10.83 (7.403) 11.62 (8.476) 11.23 (7.956) 0.271 tary Fig. 2B). Baseline HbA1c, mean (SD) 8.13 (1.036) 8.23 (0.942) 8.18 (0.991) 0.247 ,8.0% (,63.9 mmol/mol) 131 (52.6) 107 (43.5) 238 (48.1) 0.042 Clinically Meaningful HbA1c Response $8.0% ($63.9 mmol/mol) 118 (47.4) 139 (56.5) 257 (51.9) A higher percentage of patients in the Prior myocardial infarction, n (%) 0.841 albiglutide treatment group achieved Yes 21 (8.4) 22 (8.9) 43 (8.7) the treatment targets of HbA ,6.5% No 228 (91.6) 224 (91.1) 452 (91.3) 1c (,48 mmol/mol) and ,7.0% (,53 Body weight (kg), mean (SD) 83.25 (19.902) 82.84 (20.649) 83.04 (20.275) 0.821 2 mmol/mol) at week 26 (albiglutide BMI (kg/m ), mean (SD) 30.35 (5.466) 30.43 (5.828) 30.39 (5.644) 0.873 15.3% and 42.6%, respectively, com- Baseline vital signs, mean (SD) pared with sitagliptin 12.3% and Systolic blood pressure (mmHg) 132.9 (14.78) 132.8 (14.94) 132.9 (14.87) 0.940 Diastolic blood pressure (mmHg) 78.5 (10.34) 78.8 (9.17) 78.6 (9.79) 0.733 30.5%, respectively). The treatment dif- ference between albiglutide and sita- Baseline lipids, mean (SD) fi Total cholesterol (mmol/L) 4.7 (1.21) 4.5 (1.08) 4.6 (1.15) 0.053 gliptin was statistically signi cant (P = HDL (mmol/L) 1.2 (0.32) 1.2 (0.32) 1.2 (0.32) 1.000 0.0077) for the treatment target of LDL (mmol/L) 2.5 (0.95) 2.4 (0.95) 2.5 (0.95) 0.245 HbA1c ,7.0% (,53 mmol/mol) at Triglycerides (mmol/L) 2.3 (1.64) 2.0 (0.97) 2.2 (1.36) 0.014 week 26. The proportion of patients Patients with any diabetes who reached the treatment targets of condition^, n (%) improvement in HbA1c by at least Dyslipidemia 177 (71.1) 172 (69.9) 349 (70.5) 0.776 1.0% (10.9 mmol/mol), 1.5% (16.4 Nephropathy 66 (26.5) 61 (24.8) 127 (25.7) 0.663 mmol/mol), or 2.0% (21.9 mmol/mol) Peripheral neuropathy 62 (24.9) 58 (23.6) 120 (24.2) 0.731 Diabetic retinopathy 40 (16.1) 49 (19.9) 89 (18.0) 0.264 by week 26 was higher in the albiglutide group than in the sitagliptin group. The *The P value was for testing the null hypothesis that the population values (mean or proportion) fi were equal between the two treatment groups (albiglutide and sitagliptin). All tests were two- differences were not signi cant at sided. †Patients could be counted in more than one race category. A P value was not calculated week 52. for race. ^Diabetes conditions were not predefined in the protocol but were investigator assessed. Time to Hyperglycemia Rescue There was a statistically significant dif- ference between albiglutide and sita- (1.6%) groups compared with the sita- groups (data not shown). The treatment gliptin (P =0.0017)inthemeantimeto gliptin (11.4%) and matching placebo difference (albiglutide vs. sitagliptin) hyperglycemia rescue through week 52. (14.8%) groups. was 20.32% (95% CI 20.49, 20.15) The proportion of patients who had re- (23.5 mmol/mol [95% CI 25.4, 21.6]). quired hyperglycemia rescue was lower Efficacy The upper bound of the CI was below in the albiglutide group than in the sita- fi HbA1c Change From Baseline at Primary the prespeci ed noninferiority margin gliptin group at week 26 (6.1% [15 pa- End Point and Over Time of 0.4% (4.4 mmol/mol), indicating tients] vs. 12.1% [29 patients]) and at The model-adjusted LS mean for the pri- noninferiority of albiglutide to sitaglip- week 52 (17.9% [44 patients] vs. 28.3% mary end point of change from baseline tin. A superiority test conducted in ac- [68 patients]). The probability of rescue in HbA1c at week 26 was 20.83% (29.1 cordance with a prespecified, step-wise at the end of 52 weeks was 20% for al- mmol/mol) in the albiglutide group and procedure indicated that albiglutide biglutide and 34% for sitagliptin. Met- 20.52% (25.7 mmol/mol) in the sita- was statistically superior to sitagliptin formin was the most commonly used gliptin group (Fig. 1A), with similar re- (P = 0.0003). The treatment effect of rescue medication in both treatment sults across all three baseline eGFR albiglutide seen at week 26 was groups. care.diabetesjournals.org Leiter and Associates 2727

Figure 1—Summary of efficacy data through week 26 in the intent-to-treat population using an LOCF analysis. Note: Efficacy data through week 52 can be found in Supplementary Fig. 2.

Change in Body Weight occurred in 31.7% of albiglutide-treated SAEs Patients in both treatment groups patients (52.8 AEs/100 person-years) Fatal and nonfatal SAEs were similar be- showed a modest mean loss in body and in 25.2% of sitagliptin-treated pa- tween the albiglutide group (32 patients weight through week 26, with a model- tients (47.4 AEs/100 person-years). [12.9%]) and the sitagliptin group (36 adjusted LS mean weight change from Diarrhea was the most commonly re- patients [14.6%]). With the exception baseline of 20.79 kg for albiglutide ported GI AE in both the albiglutide of atrial fibrillation (reported for four and 20.19 kg for sitagliptin (P , 0.05). (10.0%) and sitagliptin (6.5%) groups. patients on albiglutide and one patient Weight loss from baseline was main- Other AEs prospectively defined as on sitagliptin), the majority of SAEs were tained through week 52 in the albiglu- being of special interest included hypo- reported for no more than two patients in tide group, whereas a small gain in mean glycemic events, injection-site reactions, each treatment group. In both groups, body weight was observed in the sita- potential systemic allergic reactions, SAEs in the “cardiac disorders and infec- gliptin group over this period (20.82 pancreatitis, and thyroid cancer (Table 2). tions and infestations” system organ class kg and 0.32 kg, respectively; P , 0.05 A higher proportion of patients in the were the most commonly reported. [excluding postrescue values]). albiglutide group (24.1%) than in the si- A total of eight deaths occurred during tagliptin group (15.9%) experienced a hy- this study: four on albiglutide (pleural me- Safety poglycemic event, with few events sothelioma, pancreatic pseudocyst, sud- AEs occurring in patients who were not den cardiac death, and cardiac disorder) The incidence of any AE and the event receiving sulfonylurea. One case (0.4%) and four on sitagliptin (ischemic stroke, rates of on-therapy AEs over the course of severe hypoglycemia was reported subarachnoid hemorrhage, malignant of the study were similar between the with albiglutide, and four cases (1.6%) melanoma, and gastroenteritis). One pa- two treatment groups (83.5% and 347 were reported with sitagliptin. Higher tient in the albiglutide group was found AEs/100 person-years with albiglutide proportions of patients in the albiglutide incidentally (on computed tomography and 83.3% and 331 AEs/100 person- group than in the sitagliptin group expe- scan during work-up of hematuria) to years with sitagliptin). A summary of rienced injection-site reactions (8.0% vs. have two pancreatic cysts and developed study AEs is presented in Table 2. Nota- 3.7%, respectively) and AEs that the in- pancreatitis after a pancreatic fine needle bly, there was no marked difference be- vestigator assessed as related to study aspirate. The patient went on to develop tween the albiglutide and sitagliptin medication (21.7% vs. 13.8%, respec- pancreatic pseudocyst and subsequently treatment groups in nausea events tively). The proportion of patients who died; the blinded adjudication commit- (4.8% vs. 3.3%) or vomiting events withdrew due to AEs was 10.4% in the tee considered the event not related to (1.6% vs. 1.2%) (Fig. 2). In both groups, albiglutide group and 10.6% in the sita- treatment. most AEs were mild or moderate in in- gliptin group. Renal failure, renal impair- Other Safety Parameters tensity, and a similar proportion of pa- ment, and hyperglycemia were the only Measurements of hematology, serum tients in each group experienced severe events that led to withdrawal reported chemistry, vital signs, electrocardio- AEs (13.7% with albiglutide and 14.6% by more than more patient in either gram readings, and physical examina- with sitagliptin). On-therapy GI AEs treatment group. tions were generally unremarkable. 2728 Albiglutide in Type 2 Diabetes Renal Impairment Diabetes Care Volume 37, October 2014

Table 2—Patient safety through week 52 (safety population) patients with type 2 diabetes and varying Albiglutide Sitagliptin degrees of renal impairment. The pri- mary study end point of change from N 249 246 baseline in HbA1c demonstrated the non- Overall AE incidence, n (%) inferiority of albiglutide to sitagliptin. Any AE 208 (83.5) 205 (83.3) Furthermore, albiglutide was shown to Fatal AEs 4 (1.6) 4 (1.6) SAEs 32 (12.9) 36 (14.6) be statistically superior to sitagliptin in Drug-related AEs 54 (21.7) 34 (13.8) lowering HbA1c at week 26. Nearly 60% AEs leading to withdrawal of active treatment 26 (10.4) 26 (10.6) of the patients treated with albiglutide Most common AEs, n (%)* achieved HbA1c levels ,7% by week 26. Urinary tract infection 23 (9.2) 20 (8.1) However, it should be noted that more Diarrhea 25 (10.0) 16 (6.5) patients in the albiglutide group than – Upper respiratory tract infection 14 (5.6) 23 (9.3) in the sitagliptin group had HbA levels Nasopharyngitis 14 (5.6) 20 (8.1) 1c ,8% at baseline. Similarly, a rapid Hypertension 14 (5.6) 19 (7.7) Anemia 16 (6.4) 10 (4.1) albiglutide treatment effect was also Peripheral edema 14 (5.6) 8 (3.3) seen with FPG, where there was a steep Constipation 15 (6.0) 6 (2.4) decline from baseline through week 4 GI disorders system organ class, n (%) that was maintained through week 26. Any event 79 (31.7) 62 (25.2) These data are further supported by Nausea 12 (4.8) 8 (3.3) the sustained effect, as evidenced by Vomiting 4 (1.6) 3 (1.2) the lower proportion of patients who Hypoglycemic events, n (%) required hyperglycemia rescue in the Any hypoglycemic event 60 (24.1) 39 (15.9) albiglutide group compared with the With background sulfonylurea 56 (22.5) 35 (14.2) Without background sulfonylurea 4 (1.6) 4 (1.6) sitagliptin group. Documented symptomatic 29 (11.6) 15 (6.1) Recent treatment guidelines have po- Severe 1 (0.4) 4 (1.6) sitioned incretin-based therapies, in- ISRs, n (%) cluding GLP-1R agonists, as prominent Any 20 (8.0) 9 (3.7) therapy options due to their substantial 1–2 15 (75.0) 9 (100.0) effectiveness in improving glycemic con- Total number of events 88 11 trol and other positive attributes such as Deemed mild 70 (79.5) 10 (90.9) modest weight loss, low rates of hypo- ISR-related study withdrawals 0 0 † glycemia, and potential for improved Potential systemic allergic reactions, n (%) b Angioedema 1 (0.4) 1 (0.4) -cell mass/function and possible car- Anaphylaxis 0 0 diovascular protective actions (24). Fur- Exfoliative rash 0 1 (0.4) ther, these guidelines indicate that Lip swelling 1 (0.4) 0 GLP-1R agonists that do not depend Face edema 1 (0.4) 0 on renal function for clearance may be an Pancreatitis, n (%)^ 1 (0.4) 0 appropriate alternative for some patients Thyroid cancer, n (%) 0 0 when metformin use is not an option. ISR, injection-site reaction. *Occurred in $5% of patients in either treatment group. †Potential With few exceptions, albiglutide and allergic reactions were identified through a standard MedDRA query. ^Preferred term of sitagliptin in this study had similar safety pancreatic pseudocyst. profiles, including for events of special interest. Although the relatively limited number of patients with severe renal Shifts from baseline in renal impairment for anti-albumin (anti–human serum al- impairment did experience higher fre- category, as assessed with eGFR, were bumin) antibodies. No reactivity with glu- quencies of GI events compared with similar between groups, with no clearly cagon was observed. One patient in the patients with mild and moderate renal identifiable treatment-associated trends albiglutide arm experienced a nonserious impairment, there were no new safety evident in either serum creatinine or the event of angioedema (local, nonsystemic, concerns identified for albiglutide in this ratio of urine albumin to creatinine. lower lip), and one patient in the sitaglip- study than were included in similar pre- tin arm experienced a serious event of vious clinical trials (16–21). Diarrhea Immunogenicity angioedema (thought to be attributable The incidence of anti-albiglutide antibod- (the most common GI event in both to iodine). Both patients remained in the ies was 3.0% (7 of 231 patients) in the treatment groups) and constipation study. No SAEs or potential systemic aller- albiglutide group, including 1 patient were seen at modestly higher inci- gic reactions were observed in the patient (0.4%) with preexisting antibodies at dences in patients treated with albiglu- with anti-albiglutide antibodies. baseline. All antibodies detected were of tide than in those treated with sitagliptin; low titer and nonneutralizing. Of the CONCLUSIONS however, there was no marked differ- seven anti-albiglutide antibody–positive This study demonstrated that once-weekly ence in nausea or vomiting events in patients, five tested positive for anti– albiglutide therapy effectively and safely either treatment group. Furthermore, up- GLP-1 antibodies and one tested positive lowers blood glucose concentrations in titration of albiglutide did not increase care.diabetesjournals.org Leiter and Associates 2729

Figure 2—Nausea (A) and vomiting (B) through week 52 in the safety population.

the number of reported GI events, and the 30-mg and 50-mg albiglutide doses. (198 on albiglutide and 178 on sitaglip- very few GI events led to withdrawal. In contrast, sitagliptin dosing was deter- tin) than completed week 26 (220 on al- These data should be considered in mined according to each patient’srenal biglutide and 206 on sitagliptin), which the context of three unique design fea- impairment status at baseline, as out- limits the interpretation of the longitudi- tures in this study. First, patients with lined in the sitagliptin prescribing in- nal sustainability response. Finally, varying degrees of renal impairment formation. Finally, the use of rescue HbA1c may be a less reliable measure of were enrolled in this study. Given that medication was embedded in the treat- efficacy in patients with renal impair- many medications used to treat type 2 ment program to facilitate patients re- ment; however, both treatment groups diabetes require dose reduction based maining in the study on blinded study had similar degrees of renal impairment, on renal dysfunction for optimal thera- medication through week 52, which en- so the effect should not bias the inter- peutic benefit, it is important to evalu- abled the collection of additional safety pretation of the results. ate albiglutide in the context of this and efficacy data. While the primary ef- These study data from patients with patient population. However, as a large ficacy assessments were independent CKD suggest a positive benefit-to-risk therapeutic protein that is degraded by of rescue medication, the main safety evaluation for the use of 30 or 50 mg enzymatic catabolism to its constituent assessments included all available pa- albiglutide to reduce blood glucose lev- amino acids, albiglutide should not re- tient information regardless of rescue els in patients with type 2 diabetes and quire renal elimination or any dose ad- status. any degree of renal impairment, thus justment that is based on degree of This study supports the potential ad- potentially filling a current unmet need. renal impairment, as supported by a vantages of albiglutide treatment in pa- Results from additional phase III studies previous albiglutide study in patients tients with type 2 diabetes who have that are currently underway as part of with varying degrees of renal impair- inadequate glycemic control and any the albiglutide clinical development pro- ment (GLP108370). This unique aspect degree of renal insufficiency without gram are consistent with the present ef- of albiglutide could prove more conve- the need for dose adjustment. ficacy and long-term safety findings from nient for this challenging population, There were several potential limita- Harmony 8 (25–29). potentially improving patient medica- tions in this study. First, there were In conclusion, albiglutide had a favor- tion compliance and treatment efficacy. only 19 patients with severe CKD who able benefit-to-risk profile and was gen- A second unique feature of the study were treated with albiglutide (and 17 erally well tolerated in patients with design was that uptitration of the sitagliptin patients). Owing to the small type 2 diabetes and mild, moderate, or blinded study drug was need based sample size and the potential for vari- severe renal impairment. In contrast to when prespecified glycemic thresholds able drug responses in this population, reports on postmarketing experience were met, with no preestablished or these results for patients with severe with currently approved GLP-1R ago- standardized time point for dose upti- CKD cannot be considered definitive. A nists, albiglutide therapy was not associ- tration. Thus, the efficacy and safety second limitation was that fewer pa- ated with worse tolerability than that data for albiglutide were combined for tients completed 52 weeks of treatment observed in other populations. In the 2730 Albiglutide in Type 2 Diabetes Renal Impairment Diabetes Care Volume 37, October 2014

present comparison with sitagliptin, 2. Dreyer G, Hull S, Aitken Z, Chesser A, Yaqoob 17. Madsbad S, Kielgast U, Asmar M, Deacon albiglutide produced statistically supe- MM. The effect of ethnicity on the prevalence of CF, Torekov SS, Holst JJ. An overview of once- rior glycemic improvement and did not diabetes and associated chronic kidney disease. weekly glucagon-like peptide-1 receptor ago- QJM 2009;102:261–269 nists–available efficacy and safety data and require dose adjustment specifictore- 3. Middleton RJ, Foley RN, Hegarty J, et al. The perspectives for the future. Diabetes Obes nal impairment. Given the various restric- unrecognized prevalence of chronic kidney dis- Metab 2011;13:394–407 tions imposed on other antihyperglycemic ease in diabetes. Nephrol Dial Transplant 2006; 18. Rosenstock J, Stewart MW. Albiglutide. therapies in this population, albiglutide 21:88–92 Glucagon-like peptide GLP-1 receptor agonist, 4. New JP, Middleton RJ, Klebe B, et al. Assessing treatment of type 2 diabetes. Drugs Future may be another treatment option for the prevalence, monitoring and management of 2010;35:701–712 patients with type 2 diabetes and renal chronic kidney disease in patients with diabetes 19.RosenstockJ,ReuschJ,BushM,YangF, impairment. compared with those without diabetes in general Stewart M; Albiglutide Study Group. Potential of practice. Diabet Med 2007;24:364–369 albiglutide, a long-acting GLP-1 receptor agonist, 5. Garber AJ, Abrahamson MJ, Barzilay JI, et al. in type 2 diabetes: a randomized controlled trial American Association of Clinical Endocrinolo- exploring weekly, biweekly, and monthly dosing. Duality of Interest. This study was sponsored gists’ comprehensive diabetes management al- Diabetes Care 2009;32:1880–1886 by GlaxoSmithKline. The authors thank the gorithm 2013 consensus statementdexecutive 20. Tomkin GH. Albiglutide, an albumin-based following employees of GlaxoSmithKline for summary. Endocr Pract 2013;19:536–557 fusion of glucagon-like peptide 1 for the poten- their specific contributions: Caroline Perry, PhD, 6. Phung OJ, Scholle JM, Talwar M, Coleman CI. tial treatment of type 2 diabetes. Curr Opin Mol for study management support and Douglas Effect of noninsulin antidiabetic drugs added to Ther 2009;11:579–588 L. Wicks, MPH, CMPP, for management of metformin therapy on glycemic control, weight 21. Seino Y, Nakajima H, Miyahara H, et al. manuscript development. The authors acknowl- gain, and hypoglycemia in type 2 diabetes. Safety, tolerability, pharmacokinetics and phar- edge the editorial support of PPD, Inc., provided JAMA 2010;303:1410–1418 macodynamics of albiglutide, a long-acting GLP- by Amy Myers, MPH (production of draft out- 7. package insert [Internet], 2011. 1-receptor agonist, in Japanese subjects with line, production of manuscript, assembly of Deerfield, IL, Takeda Pharmaceuticals America, Inc. type 2 diabetes mellitus. Curr Med Res Opin tables and figures) and Christopher Barnes, Available from http://www.accessdata.fda.gov/ 2009;25:3049–3057 PhD (production of manuscript), which scripts/cder/drugsatfda/index.cfm?fuseaction= 22. National Kidney Foundation. K/DOQI clini- was funded by GlaxoSmithKline. L.A.L. has search.drugdetails. Accessed 15 August 2013 cal practice guidelines for chronic kidney disease: received research funding from, has provided 8. Kahn SE, Haffner SM, Heise MA, et al.; ADOPT evaluation, classification, and stratification. Am J continuing medical education on behalf of, and/ Study Group. Glycemic durability of rosiglita- Kidney Dis 2002;39(Suppl. 1):S1–S266 or has acted as a consultant to AstraZeneca, zone, metformin, or glyburide monotherapy. N 23. Young MA, Wald JA, Matthews JE, Yang F, Boehringer Ingelheim, Bristol-Myers Squibb, Eli Engl J Med 2006;355:2427–2443 Reinhardt RR. Effect of renal impairment on the Lilly, GlaxoSmithKline, Janssen, Merck, Novo 9. Byetta US package insert [Internet], 2011. pharmocokinetics, efficacy, and safety of albi- Nordisk, Sanofi, Servier, and Takeda. M.C.C., M.S., San Diego, CA, Pharmaceuticals. Avail- glutide. Postgrad Med 2014;126:35–36 A.J.-L., R.S., and F.Y. are employed by and able from http://www.accessdata.fda.gov/ 24. Inzucchi SE, Bergenstal RM, Buse JB, et al.; shareholders of GlaxoSmithKline. M.S. and F.Y. drugsatfda_docs/label/2011/021773s029s030lbl. American Diabetes Association (ADA); Euro- have been issued a patent for GLP-1s in renal pdf. Accessed 21 October 2013 pean Association for the Study of Diabetes disease. Y.H. has received research funding from, 10. Victoza US package insert [Internet], 2012. (EASD). Management of hyperglycemia in type has provided continuing medical education on Princeton, NJ, Novo Nordisk. Available from 2 diabetes: a patient-centered approach: posi- behalf of, and/or has acted as a consultant to http://www.accessdata.fda.gov/drugsatfda_ tion statement of the American Diabetes Asso- Amarin, Amgen Amylin, Boehringer Ingelheim, docs/label/2012/022341s007s009s013lbl. ciation (ADA) and the European Association for Bristol-Myers Squibb, diaDeux, DSI, Eisai, Gilead, pdf. Accessed 21 October 2013 the Study of Diabetes (EASD). Diabetes Care Intarcia, GlaxoSmithKline, Halozyme, Janssen, 11. Januvia US package insert [Internet], 2012. 2012;35:1364–1379 Lexicon, LipoScience, Merck, Novo Nordisk, Whitehouse Station, NJ, Merck & Co. Available 25. Ahren´ B, Stewart M, Cirkel D, et al. Weekly Sanofi, Santarus, Takeda, and Vivus. No other from http://www.accessdata.fda.gov/drugsatfda_ GLP-1 receptor agonist albiglutide is superior to potential conflicts of interest relevant to this docs/label/2012/021995s023lbl.pdf. Accessed 21 sitagliptin or in patients with type 2 article were reported. October 2013 diabetes mellitus on metformin at the 104 week Author Contributions. L.A.L., M.C.C., M.S., 12. Ampudia-Blasco FJ, Rossetti P, Ascaso JF. primary endpoint. Diabetes Care 2014;37: A.J.-L., R.S., F.Y., and Y.H. contributed to study Basal plus basal-bolus approach in type 2 diabe- 2141–2148 design, conduct, or data collection; performed tes. Diabetes Technol Ther 2011;13(Suppl. 1): 26. Pratley R, Stewart M, Cirkel D, Ye J, Perry C, data analysis or interpretation; and contributed S75–S83 Carr MC. Harmony 4: 52-week efficacy of albi- to writing, reviewing, revising, and approving 13. Seaquist ER, Anderson J, Childs B, et al. Hy- glutide vs in patients with type 2 the manuscript for submission. L.A.L. and Y.H. poglycemia and diabetes: a report of a work- diabetes mellitus. Diabetologia 2013;56(Suppl. are the guarantors of this work and, as such, had group of the American Diabetes Association 1):S360–S361 full access to all the data in the study and take and the Endocrine Society. Diabetes Care 27. Nauck M, Stewart M, Perkins C, et al. Har- responsibility for the integrity of the data and 2013;36:1384–1395 mony 2 results at week 52 primary endpoint: the accuracy of the data analysis. 14. Gross JL, Kramer CK, Leitão CB, et al.; Di- once-weekly albiglutide monotherapy for pa- Prior Presentation. Parts of this study were abetes and Endocrinology Meta-analysis Group tients with type 2 diabetes mellitus inade- presented in abstract form at the 73rd Scientific (DEMA). Effect of antihyperglycemic agents quately controlled with diet and exercise. Sessions of the American Diabetes Association, added to metformin and a sulfonylurea on gly- Diabetologia 2013;56(Suppl. 1):S360 Chicago, IL, 21–25 June 2013, and at the 49th cemic control and weight gain in type 2 diabe- 28. Reusch J, Stewart M, Perkins C, et al. Har- Annual Meeting of the European Association for tes: a network meta-analysis. Ann Intern Med mony 1 results at week 52 primary endpoint: the Study of Diabetes, Barcelona, Spain, 23–27 2011;154:672–679 once-weekly albiglutide vs placebo in patients September 2013. 15. Carver C. Insulin treatment and the prob- with type 2 diabetes mellitus not controlled on lem of weight gain in type 2 diabetes. Diabetes pioglitazone 6 metformin. Diabetologia 2013; Educ 2006;32:910–917 56(Suppl. 1):S359–S360 References 16. Bush MA, Matthews JE, De Boever EH, et al. 29. Stewart M, Yang F, Perry C, Carr MC, Home 1. Koro CE, Lee BH, Bowlin SJ. Antidiabetic med- Safety, tolerability, pharmacodynamics and PD. 52-week efficacy of albiglutide vs placebo vs ication use and prevalence of chronic kidney pharmacokinetics of albiglutide, a long-acting pioglitazone in triple therapy (background met- disease among patients with type 2 diabetes glucagon-like peptide-1 mimetic, in healthy formin and glimepiride) in patients with type 2 mellitus in the United States. Clin Ther 2009; subjects. Diabetes Obes Metab 2009;11:498– diabetes: Harmony 5 study. Diabetologia 2013; 31:2608–2617 505 56(Suppl. 1):S361