<<

1318 Care Volume 39, August 2016

Julio Rosenstock,1 Bruno Guerci,2 Prandial Options to Advance Basal Markolf Hanefeld,3 Sandro Gentile,4 Ronnie Aronson,5 Francisco J. Tinahones,6 Glargine Therapy: Testing Christine Roy-Duval,7 Elisabeth Souhami,7 Marek Wardecki,8 Jenny Ye,9 Plus Basal Insulin Riccardo Perfetti,9 and Simon Heller,10 on behalf of the GetGoal Duo-2 Trial Versus Either as Investigators Basal-Plus or Basal- in : The GetGoal Duo-2 Trial Diabetes Care 2016;39:1318–1328 | DOI: 10.2337/dc16-0014 CLIN CARE/EDUCATION/NUTRITION/PSYCHOSOCIAL

OBJECTIVE To provide evidence-based options on how to intensify basal insulin, we explored head-to-head prandial interventions in overweight patients with type 2 diabetes inadequately controlled on basal with or without 1–3 oral antidi- 1Dallas Diabetes and Endocrine Center at Medi- abetic agents (OADs). cal City, Dallas, TX 2University of Lorraine and the Department of RESEARCH DESIGN AND METHODS Diabetology, Metabolic Diseases and Nutrition, Brabois Adult Hospital, Vandœuvre-les-Nancy,` Patients were randomized to lixisenatide once daily or insulin glulisine given once France 3 or thrice daily, added to glargine, with or without , if HbA1c remained GWT-TUD, Study Centre Prof. Hanefeld, Dres- ‡7to£9% (‡53 to £75 mmol/mol) after 12 weeks of glargine optimization with den Technical University, Dresden, Germany 4Department of Clinical and Experimental Med- OADs other than metformin stopped at the start of optimization. Coprimary end points icine, Second University of Naples, Naples, Italy at 26 weeks were 1) noninferiority (95% CI upper bound <0.4% [<4.4 mmol/mol]) 5LMC Diabetes & Endocrinology, Toronto, On- in HbA reduction with lixisenatide versus glulisine once daily, and either 2a) non- tario, Canada 1c 6 inferiority in HbA reduction for lixisenatide versus glulisine thrice daily or 2b)supe- Centre in Physiopathology of Obesity and Nutri- 1c tion (CIBEROBN), Carlos III Institute of Health riority in body weight change for lixisenatide versus glulisine thrice daily. Fasting and Hospital Virgen de la Victoria, Malaga, Spain and postprandial plasma , composite efficacy/safety end points, and ad- 7Sanofi, Paris, France verse events were also assessed. 8Sanofi, Warsaw, Poland 9Sanofi, Bridgewater, NJ 10 RESULTS Academic Unit of Diabetes, Endocrinology and Metabolism, University of Sheffield, Sheffield, Baseline characteristics were similar between arms (n = 298, diabetes and basal U.K. 2 insulin duration of 12.2 and 3.2 years, respectively; BMI 32.2 kg/m ). HbA1c im- Corresponding author: Julio Rosenstock, proved from 8.5% to 7.9% (69 to 63 mmol/mol) with glargine optimization and [email protected]. further to 7.2%, 7.2%, and 7.0% (55, 55, and 53 mmol/mol) with lixisenatide and Received 4 January 2016 and accepted 29 April glulisine once daily and thrice daily, respectively; all coprimary end points were 2016. met. Symptomatic and body weight were lower in lixisenatide reg. no. NCT01768559, clinicaltrials versus glulisine patients. More gastrointestinal events occurred with lixisenatide. .gov. This article contains Supplementary Data online CONCLUSIONS at http://care.diabetesjournals.org/lookup/ Short-acting -like -1 receptor as add-on to basal insulin suppl/doi:10.2337/dc16-0014/-/DC1. may become a preferred treatment intensification option, attaining meaningful © 2016 by the American Diabetes Association. Readers may use this article as long as the work is glycemic targets with fewer hypoglycemic events without weight gain versus properly cited, the use is educational and not for basal-plus or basal-bolus in uncontrolled basal insulin-treated type 2 diabetes. profit, and the work is not altered. care.diabetesjournals.org Rosenstock and Associates 1319

The conventional recommendation in The present 26-week trial (GetGoal were eligible for screening. Patients type 2 diabetes inadequately controlled Duo-2) took the above findings further were on basal insulin for at least 6 months with basal insulin plus oral antidiabetic and evaluated head-to-head in a ran- at screening (stable dose $20 units/day agents (OADs) has been the addition of a domized fashion the efficacy and safety for $2 months before screening), alone prandial insulin as a single injection with of lixisenatide 20 mg once daily versus or combined with stable doses of 1–3 the main meal (basal-plus) or progres- insulin glulisine once daily (basal-plus) OADs (metformin [$1.5 mg/day or max- sively covering all meals (basal-bolus) (1). or thrice daily (basal-bolus) for intensi- imum tolerated dose], a DPP-4 inhibitor, Evidence from a randomized head-to- fication of optimized insulin glargine, an SU, or a glinide). Patients receiving head trial comparing initiation of insulin with or without metformin, in uncon- basal insulin alone or with metformin glargine plus insulin glulisine as basal-plus trolled patients on basal insulin with or hadtohaveHbA1c 7.5–10.0% (58– or basal-bolus versus twice-daily pre- without OADs. It is hoped that this trial 86 mmol/mol) at screening. Patients re- mixed insulin demonstrated that these will provide evidence-based data to aid ceiving basal insulin plus an SU and/or regimens can achieve substantial and in the clinical decision-making process a DPP-4 inhibitor and/or a glinide comparable reductions in HbA1c to almost for individualization of basal insulin- hadtohaveHbA1c 7.0–10.0% (53– 7% (53 mmol/mol) over 60 weeks (2). based therapy in patients with type 2 86 mmol/mol) at screening. Because all However, hypoglycemia and weight gain diabetes when intensification of treat- OADs other than metformin were dis- occurred in each prandial insulin injection ment is required. continued at the start of the run-in group. Indeed, unacceptable weight gain phase, the two different HbA1c thresh- and hypoglycemia are pressing issues in RESEARCH DESIGN AND METHODS olds were necessary to compensate for the management of type 2 diabetes, and the subsequent expected increase in Trial Design hypoglycemic risk increases as type 2 di- HbA in the latter group. GetGoal Duo-2 was a randomized, open- 1c abetes advances, impeding attainment of Exclusion criteria included a clinically label, active comparator trial with three glycemic targets (3). relevant history of gastrointestinal dis- parallel arms conducted at 199 centers in The recently updated American Diabe- ease or a history of unexplained/chronic 18 countries. The study was initiated in tes Association/European Association for . Patients were excluded if January 2013 and completed in Decem- the Study of Diabetes treatment guide- they had /aspartate aminotrans- ber 2014. Adults with type 2 diabetes un- lines have largely reiterated this prandial ferase, amylase, or lipase levels more controlled on $6 months’ basal insulin, insulin recommendation but also pro- than three times the upper limit of normal with or without 1–3 OADs, were eligible posed the option of adding a glucagon- or levels .20 pg/mL. to enter a 12-week run-in period. On run- like peptide-1 receptor (GLP-1 in entry, OADs other than metformin (di- RA) to basal insulin (4). Although these peptidyl peptidase-4 [DPP-4] inhibitors, Run-in Phase updated guidelines emphasize the impor- During run-in, insulin glargine was intro- [SU], and glinides) were dis- tance of individualizing treatment and duced (for patients previously on a differ- continued, and insulin glargine was opti- go some way toward revising the exist- ent basal insulin) or continued as part of a mally titrated. After the run-in phase, if ing treatment paradigm, they are not once-daily regimen, at breakfast or at the HbA remained between $7and#9% evidence-based, because the three in- 1c evening meal, as determined at the start ($53 and #75 mmol/mol) and mean jectable options (GLP-1 RA, basal insu- of the run-in period, and titrated every fasting plasma glucose (FPG) was #140 lin, and prandial insulin) have not been 3 days to achieve fasting self-monitored mg/dL (#7.8 mmol/L), patients were ran- compared directly in a single properly plasma glucose (SMPG) between 80 and domized 1:1:1 (by interactive voice or controlled study to address the multiple 100 mg/dL (4.4 and 5.6 mmol/L) while Web response system), and stratified by issues affecting outcomes (5). avoiding hypoglycemia (11). baseline HbA (,8or$8% [,64 or .64 GLP-1 RAs improve glycemia without 1c mmol/mol]) and metformin use, to re- substantially increasing hypoglycemic ceive subcutaneous lixisenatide once Randomized Treatment Phase risk and without weight gain. Lixisenatide Insulin glargine doses were adjusted daily or insulin glulisine once or thrice (Lyxumia; Sanofi, Paris, France) is a once- weekly to maintain fasting daily SMPG daily as add-on therapy to insulin glargine daily prandial GLP-1 RA with robust post- between 80 and 100 mg/dL (4.4 and 5.6 with or without metformin. The trial com- prandial glucose (PPG)-lowering effect, mmol/L) except during the 4 weeks after prised three periods (Supplementary predominantly via delayed gastric empty- randomization when a stable insulin Fig. 1): 1) a screening phase of up to ing and glucose-dependent reductions in dose was maintained. Timing of insulin 2 weeks, followed by a 12-week run-in glucagon release (6–9). glargine injections was established dur- phase; 2) 26 weeks of open-label, random- The GetGoal Duo-1 trial was the first ing the run-in, whereas timing of the ized treatment; and 3) 3 days of follow-up. study to assess the efficacy of lixisenatide once-daily prandial intervention was All patients signed an informed consent in combination with optimally titrated based on the main meal during the week form. The trial protocol complied with basal insulin in patients with type 2 diabe- before randomization. Patients random- the recommendations of the Declaration tes uncontrolled on OADs who were newly ized to receive lixisenatide were adminis- of Helsinki and was approved by indepen- initiating insulin glargine. Significant reduc- tered 10 mg once daily for 2 weeks, dent ethics committees at each center. tions in HbA1c to 7% (53 mmol/mol) and followed by lixisenatide 20 mg once daily marked PPG reductions were achieved Study Population for the remainder of the study, injected with the addition of lixisenatide plus basal Adults with type 2 diabetes for at least 30–60 min before the main meal, as de- insulin versus basal insulin alone (10). 1yearandaBMI.20.0–40.0 kg/m2 fined above. Further details on study 1320 Insulin Plus Lixisenatide or Glulisine QD/TID Diabetes Care Volume 39, August 2016

administration can be found in that a sample size of 285 patients per treat- and 2 were both met at a = 0.025 (one- the Supplementary Appendix. ment arm (855 patients in total) would sided), the study was considered positive. ensure that the upper limit of the two- Coprimary end points were analyzed Assessments sided 95% CI for the adjusted mean differ- using an ANCOVA model with treat- Coprimary objectives at week 26 were: 1) ence in change from baseline at week 26 in ment, week – 1strataofHbA1c (,8or noninferiority of lixisenatide versus insu- HbA1c between the lixisenatide and the in- $8% [,64 or $64 mmol/mol]), ran- lin glulisine once daily in HbA1c reduction; sulin glulisine once-daily arms would not domization strata of metformin use and for lixisenatide versus insulin glulisine exceed 0.4% (4.4 mmol/mol) with at least (yes or no), and country as fixed effects, thrice daily, either 2a) noninferiority in 94% power. This sample size also ensured and using the corresponding baseline HbA1c reduction or 2b) superiority of that the upper limit of the two-sided 97.5% value as a covariate. Missing data in lixisenatide versus insulin glulisine CI for the adjusted mean difference in the analysis of the coprimary end points thrice daily in body weight change. change from baseline at week 26 in were imputed using the last observation Secondary objectives assessed at HbA1c between the lixisenatide and the carried forward method. week 26 included change from baseline insulin glulisine thrice-daily arms would All continuous secondary efficacy end in FPG, change from baseline in PPG dur- not exceed 0.4% (4.4 mmol/mol) with at points, except insulin glulisine and total ing standardized meal tests in patients least 90% power; and at least 90% power daily insulin dose, were analyzed using who received an injection of study med- to detect a difference of 1 kg in change the same primary ANCOVA model as de- ication before breakfast (meal test using from baseline in body weight at week 26 scribed above. All categorical secondary Ensure Plus [Abbott Laboratories]; between the lixisenatide and the insulin efficacy end points were analyzed using a – 600 kcal, carbohydrate 50 55%, glulisine thrice-daily arms, assuming a Cochran–Mantel–Haenszel method strat- – – 15 20%, and fat 25 30%), percentage common SD of 2.75 kg at the 2.5% signif- ified on week – 1HbA (,8or$8% , 1c of patients who achieved HbA1c 7or icance level (two-sided). [,64 or $64 mmol/mol]) and randomi- # , # 6.5% ( 53 or 48 mmol/mol), change The modified intent-to-treat popula- zation strata of metformin use (yes or no). from baseline in body weight, 7-point tion (all randomized patients with at Multiplicity adjustments were not per- fi SMPG pro le, and daily insulin dose. Com- least one dose of study medication formed for secondary efficacy end points. posite end points assessed at week 26 and a baseline assessment and at least In the efficacy analysis, patients were were the percentage of patients achieving one assessment after baseline of any analyzed in the treatment group to which , , HbA1c 7% ( 53 mmol/mol) with no primary or secondary efficacy end point) they were randomized, irrespective of , , weight gain; HbA1c 7% ( 53 mmol/mol) was used for all efficacy analyses, other compliance with the study protocol. without documented symptomatic hy- than assessment of PPG (assessed in pa- , , poglycemia (glucose 60 mg/dL [ 3.3 tientswhoreceivedaninjectionofstudy RESULTS , , mmol/L]; and HbA1c 7% [ 53 mmol/mol]), medication before breakfast). The safety Between 8 January 2013 and 3 Decem- no weight gain, and no documented symp- analysis was conducted on the safety ber 2014, 2,159 patients were screened tomatic hypoglycemia. population (all randomized patients who and 894 patients were randomized Adverse events (AEs), serious AEs, vi- received at least one dose of study medica- (Supplementary Fig. 2). The most fre- tal signs, and laboratory values, includ- tion regardless of the amount of treat- quent reason for run-in failure was ing amylase and lipase, were monitored ment administered). that HbA1c criteria were not met at the throughout the trial. Symptomatic hy- Overall, statistical analyses were as- end of the run-in period. Treatment dis- poglycemia AEs included any event ac- sessed at a significance level of 0.025 continuations occurred in 30 (10.1%), 17 companied by typical signs and/or (one-sided). End points 1 and 2 (2a or (5.7%), and 12 (4.0%) of the 298 patients symptoms of hypoglycemia reported 2b) were assessed separately at a signifi- in the lixisenatide and insulin glulisine fi by the study investigator, and as de ned cance level of 0.025 (one-sided). For co- once-daily and thrice-daily arms, re- per protocol, included symptomatic primary end point 1, lixisenatide was spectively (Supplementary Fig. 2). Insu- , events accompanied by glucose 60 declared noninferior to insulin glulisine lin glargine was administered by 26% of , mg/dL ( 3.3 mmol/L) or, if no glucose once daily if the upper bound of the two- patients in the morning and by 74% in measurement was available, associated sided 95% CI for treatment difference was the evening. Both lixisenatide and insu- with prompt recovery with oral carbohy- ,0.4% (,4.4 mmol/mol) as determined lin glulisine once daily were adminis- fi drate. Severe hypoglycemia was de ned based on regulatory recommendations at tered by 30% of patients at breakfast as the patient requiring assistance along the time of the protocol preparation and and by 70% at the evening meal. with a glucose measurement 36 mg/dL on other studies with similar compounds Baseline characteristics were similar , ( 2.0 mmol/L) or, if no glucose measure- (12,13). For coprimary end points 2a and across treatment groups (Table 1). Ap- ment was available, prompt recovery after 2b, a Hochberg procedure was used to proximately 90% of patients were re- oral carbohydrate, intravenous glucose, control for type I error as follows: if both ceiving (and continued) treatment with or glucagon administration. end points 2a and 2b were met at a = metformin. At screening, the overall 0.025 (one-sided), then the entire end mean durations of type 2 diabetes and Statistical Methods point 2 was met at a = 0.025 (one-sided); basal insulin treatment were 12.2 and Assuming a dropout rate of 20%, a com- if only end point 2a or 2b was achieved at 3.2 years, respectively. Mean overall mon standard deviation of 1.2%, and a a = 0.025 (one-sided), then the end point body weight was 89.1 kg and mean 2 true difference in HbA1c between the treat- that was met was tested at a =0.0125 BMI was 32.2 kg/m . At screening, mean 6 ment groups of zero, it was determined (one-sided). If coprimary end points 1 SD HbA1c for the total randomized care.diabetesjournals.org Rosenstock and Associates 1321

Table 1—Screening demographic data and patient characteristics: randomized population Insulin glargine with or without metformin plus Lixisenatide 20 mg once daily Insulin glulisine once daily Insulin glulisine thrice daily Parameters (n =298) (n = 298) (n =298) Age at screening, years 59.8 6 8.6 60.2 6 8.6 59.4 6 9.5 Male, n (%) 138 (46.3) 135 (45.3) 132 (44.3) White, n (%) 276 (92.6) 280 (94.0) 272 (91.3) BMI, kg/m2 32.3 6 4.6 31.9 6 4.4 32.5 6 4.6 Body weight, kg Start of run-in 89.8 6 17.4 87.9 6 15.8 89.7 6 17.3 End of run-in 90.2 6 17.5 88.4 6 15.8 90.1 6 17.3

HbA1c, % (mmol/mol) Screening 8.5 6 0.7 (69 6 7.7) 8.5 6 0.7 (69 6 7.7) 8.5 6 0.8 (69 6 8.7) End of run-in 7.9 6 0.5 (63 6 5.5) 7.8 6 0.5 (62 6 5.5) 7.9 6 0.5 (63 6 5.5) FPG, mmol/L (mg/dL) Start of run-in 9.2 6 2.9 (165 6 53) 9.3 6 2.9 (167 6 52) 9.5 6 3.0 (171 6 53) End of run-in 6.9 6 2.1 (125 6 37) 6.8 6 1.8 (122 6 32) 6.7 6 1.9 (119 6 34) Variables at screening Duration of type 2 diabetes, years 11.9 6 6.4 12.3 6 6.8 12.4 6 6.8 Duration of basal insulin treatment, years 3.1 6 2.6 3.3 6 3.5 3.2 6 3.1 Daily basal insulin dose, units/day NPH insulin 41 6 20 39 6 18 41 6 20 Insulin glargine 42 6 23 41 6 23 40 6 23 41 6 30 40 6 25 39 6 22 OAD use, n (%) Metformin 262 (87.9) 260 (87.2) 259 (86.9) SU 141 (47.3) 129 (43.3) 142 (47.7) DPP-4 inhibitor 37 (12.4) 29 (9.7) 42 (14.1) Insulin glargine dose, units/day Start of run-in 41 6 22 40 6 22 39 6 21 End of run-in 68 6 32 65 6 32 65 6 27 Patients with evening insulin glargine dosing,* % 73.5 76.8 73.0 Patients with study drug administration at: Breakfast, % 30.2 29.5 – Evening meal, % 69.5 69.8 – Missing, % 0.3 0.7 – Data are presented as the mean 6 SD or as indicated. All data are for screening except where indicated otherwise. Start of run-in, week – 12; end of run-in, week – 1. *Data available for 283 lixisenatide-treated patients and 289 patients in each insulin glulisine arm; evening dosing defined as any dose administered between 1600 and 0400 h. study population was 8.5 6 0.7% (69 6 6.6 mmol/mol) and 7.8 6 0.6% (62 6 6.6 lixisenatide and insulin glulisine once 7.7 mmol/mol), and at the start of run- mmol/mol) and was reduced to 7.2 6 0.8% daily and thrice daily, respectively (Table in, mean 6 SD FPG was 168 6 53 mg/dL (55 6 8.7 mmol/mol) and 7.0 6 0.7% 2). Lixisenatide demonstrated statistical (9.3 6 2.9 mmol/L). Overall mean 6 (53 6 7.7 mmol/mol), respectively, at superiorityinchangefrombaselineat SD HbA1c and FPG at the end of the week 26 (Table 2). The least squares (LS) week 26 in body weight compared week-12 glargine optimization pe- mean treatment difference (95% CI) for with insulin glulisine thrice daily (copri- 6 6 riod decreased to 7.9 0.5% (63 change in HbA1c from baseline to end mary end point LS mean treatment dif- 5.5 mmol/mol) and 122 6 34 mg/dL of study for lixisenatide versus insulin ference –2.0 kg [95% CI –2.59, –1.40]; (6.8 6 1.9 mmol/L), respectively. During glulisine once daily was –0.1% (–0.17, P , 0.0001). the optimization period, the overall insu- 0.06) (–0.5 mmol/mol [–1.9, 0.7]) and lin glargine dose increased from 40 to Secondary End Points versus insulin glulisine thrice daily was Owing to optimization of insulin glargine 66 units/day. 0.2% (0.10, 0.33) (2.3 mmol/mol [1.0, during the run-in period, the change fi Coprimary End Points 3.5]), meeting the prede ned nonin- from baseline in FPG over 26 weeks All coprimary end points were met. feriority criteria (upper bound of the was minimal across the three treatment Mean 6 SD HbA1c with lixisenatide declined two-sided 95% CI for treatment differ- arms, as was the change in the daily in- from 7.8 6 0.6% (62 6 6.6 mmol/mol) ence ,0.4%). sulinglarginedose(Table2).LSmean6 at baseline to 7.2 6 0.8% (55 6 8.7 At week 26, the LS mean 6 SE change SE reductions from baseline in 2-h PPG mmol/mol) at week 26. Mean 6 SD from baseline in body weight in the after a standardized breakfast at week 26 baseline HbA1c in the insulin glulisine once- three treatment groups was –0.6 6 were markedly greater in the lixisenatide and thrice-daily arms was 7.7 6 0.6% (61 6 0.3, +1.0 6 0.3, and +1.4 6 0.3 kg, for arm compared with the insulin glulisine 1322 Insulin Plus Lixisenatide or Glulisine QD/TID Diabetes Care Volume 39, August 2016

Table 2—Response to therapy: modified intent-to-treat population Insulin glargine with or without metformin plus Lixisenatide 20 mg once daily Insulin glulisine once daily Insulin glulisine thrice daily Parameters (n = 297) (n = 298) (n = 295) FPG, mmol/L (mg/dL); n Baseline, mean 6 SD 6.6 6 1.8 (119 6 33); 295 6.9 6 2.0 (123 6 36); 295 6.7 6 1.9 (120 6 34); 294 Week 26 LOCF, mean 6 SD 6.6 6 2.0 (119 6 35); 295 6.7 6 1.9 (120 6 35); 295 6.7 6 2.0 (121 6 36); 294 LS mean change 6 SE 20.2 6 0.1 (–4 6 3); 295 20.2 6 0.1 (–4 6 3); 295 20.1 6 0.1 (–1 6 3); 294 LS mean (95% CI) treatment difference – 20.0 (–0.32, 0.30) 20.2 (–0.48, 0.14) (–0.2 [–5.7, 5.4]) (–3.0 [–8.6, 2.6]) 2-h PPG, mmol/L (mg/dL);* n/N Baseline, mean 6 SD 14.1 6 3.6 (254 6 65); 69/90 13.8 6 3.5 (249 6 63); 55/88 14.6 6 3.5 (262 6 63); 68/295 Week 26 LOCF, mean 6 SD 10.2 6 3.9 (184 6 70); 69/90 12.2 6 3.4 (220 6 60); 55/88 12.7 6 3.9 (229 6 69); 68/295 LS mean change 6 SE 23.6 6 0.6 21.6 6 0.6 21.4 6 0.6 (–66 6 11); 69/90 (–28 6 11); 55/88 (–25 6 11); 68/295 LS mean (95% CI) treatment difference – 22.1 (–3.3, –0.8) 22.2 (–3.4, –1.1) (–37.3 [–59.2, –15.3]) (–40.2 [–61.1, –19.2])

HbA1c, % (mmol/mol); n Baseline after 12-week run-in optimization, mean 6 SD 7.8 6 0.6 (62 6 6.6); 292 7.7 6 0.6 (61 6 6.6); 292 7.8 6 0.6 (62 6 6.6); 295 Week 26 LOCF, mean 6 SD 7.2 6 0.8 (55 6 8.7); 292 7.2 6 0.8 (55 6 8.7); 292 7.0 6 0.7 (53 6 7.7); 295 LS mean change 6 SE 20.6 6 0.1 (–6.6 6 1.1); 292 20.6 6 0.1 (–6.6 6 1.1); 292 20.8 6 0.1 (–8.7 6 1.1); 295 LS mean (95% CI) treatment difference – 20.1 (–0.17, 0.06) 0.2 (0.10, 0.33) (–0.5 [–1.9, 0.7]) (2.3 [1.0, 3.5])

HbA1c ,7% at week 26† n/N (%) 123/292 (42.1) 112/292 (38.4) 145/295 (49.2) Treatment difference, % (95% CI) – 3.7 (–4.03, 11.49) 27.3 (–15.07, 0.56)

HbA1c #6.5% at week 26† n/N (%) 60/292 (20.5) 52/292 (17.8) 91/295 (30.8) Treatment difference, % (95% CI) – 2.7 (–3.59, 9.01) 210.5 (–17.33, –3.59) Body weight, kg; n Baseline, mean 6 SD 90.1 6 17.4; 295 88.4 6 15.9; 295 90.0 6 17.2; 295 Week 26 LOCF, mean 6 SD 89.4 6 18.1; 295 89.3 6 16.3; 295 91.3 6 17.3; 295 LS mean change 6 SE 20.6 6 0.3; 295 1.0 6 0.3; 295 1.4 6 0.3; 295 LS mean (95% CI) treatment difference – 21.7 (–2.26, –1.06) 22.0 (–2.59, –1.40)§ Patients with no weight gain;†‡ n/N (%) 191/295 (64.7) 108/295 (36.6) 90/295 (30.5) Weighted average response rate (95% CI) treatment difference (%) – 28.1 (20.5, 35.8) 34.2 (26.7, 41.7) Insulin glargine dose, U/day; n Baseline, mean 6 SD 67 6 32; 292 65 6 32; 294 65 6 27; 294 Week 26 LOCF, mean 6 SD 67 6 36; 292 64 6 36; 294 61 6 29; 294 LS mean change 6 SE from baseline to week 26 0.7 6 1.0; 292 20.1 6 1.0; 294 23.1 6 1.0; 294 LS mean (95% CI) treatment difference – 0.8 (–1.41, 2.92) 3.8 (1.66, 6.00) Daily insulin glulisine dose, units; n Week 26 LOCF, mean 6 SD– 10 6 8; 295 20 6 13; 293 Total daily insulin dose (glargine + glulisine), units; n Week 26 LOCF, mean 6 SD 67 6 32; 292 74 6 39; 295 81 6 34; 294 Exploratory analyses Patients with $2% weight loss;† n (%) 97(32.9) 33(11.2) 32(10.8) Weighted average response rate treatment difference, % (95% CI) – 21.7 (15.27, 28.11) 22.0 (15.61, 28.43) Patients with $3% weight loss;† n (%) 69 (23.4) 21 (7.1) 18 (6.1) Weighted average response rate treatment difference, % (95% CI) – 16.3 (10.69, 21.97) 17.3 (11.75, 22.85) Patients with $5% weight loss;† n (%) 36 (12.2) 11 (3.7) 7 (2.4) Weighted average response rate treatment difference, % (95% CI) – 8.5 (4.13, 12.86) 9.8 (5.69, 13.98)

Summary statistics (mean 6 SD) for HbA1c, FPG, 2-h PPG, and body weight were based on modified intent-to-treat population; all available data were included in the analysis. An ANCOVA model was used for calculation of LS mean treatment difference for HbA1c, FPG, 2-h PPG, and body weight, with treatment groups, week – 1 strata of HbA1c (,8or$8% [,64 or $64 mmol/mol]), metformin use (yes or no) strata at randomization, and country as fixed effects, and the corresponding baseline value as a covariate. Patients with both baseline and week 26 (LOCF) measurements are included. LOCF, last observation carried forward. *After a standardized meal in patients administered treatment before breakfast. †Weighted average of proportion difference between treatment groups from each strata (week – 1 strata of HbA1c [,8.0 or $8.0%], randomization strata of metformin use) using Cochran–Mantel–Haenszel weights. ‡Prespecified end point. §P , 0.0001. care.diabetesjournals.org Rosenstock and Associates 1323

once- and thrice-daily arms (Table 2). Ro- glulisine once daily and thrice daily, re- GLP-1 RA plus insulin glargine with bust reductions in HbA1c,asseenduring spectively) (Table 3 and Fig. 1E). Serious both a basal-plus and a basal-bolus reg- the run-in period, continued in each of symptomatic hypoglycemia was report- imen. All three injectable options to ad- the treatment arms from the first study ed for two patients in the insulin glulisine vance basal insulin therapy provided on-treatment measurement up to week once-daily arm versus none in the lixise- substantial HbA1c reductions but with 12andremainedrelativelystableuntil natide and insulin glulisine thrice-daily meaningful clinical differences in treat- the end of treatment at week 26 (Fig. arms. ment complexity and intensity of blood 1A). Proportions of patients achieving glucose monitoring, body weight changes, glycemic targets are reported in Table Overall Safety and, most notably, hypoglycemic risk. The 2. LS mean treatment difference in body A greater proportion of patients in the findings of this study, therefore, provide weight change for lixisenatide versus lixisenatide arm experienced gastroin- an evidence-based framework for the insulin glulisine once daily was –1.7 kg testinal AEs compared with in the insulin decision-making process to select the ap- (95% CI –2.26, –1.06). Contrary to treat- glulisine arms (Table 3). and/or propriate treatment strategy on a patient- ment with lixisenatide, treatment with vomiting were the most common AEs in by-patient basis. The outcomes of this the lixisenatide arm (Table 3), resulting basal-plus and basal-bolus resulted in study were positive despite the testing in treatment discontinuation in six pa- increased body weight from week 2 on a very challenging, predominantly tients. Gastrointestinal AEs resulted in through week 26 (Fig. 1B). The percent- overweight and obese population with 11 patients (4%) permanently discontin- age of patients with no weight gain over long-standing type 2 diabetes inade- uing treatment in the lixisenatide arm the course of study treatment was sub- quately controlled on relatively high compared with none in the insulin glulisine stantially higher with lixisenatide com- doses of basal insulin and receiving mul- arms. Serious AEs were observed in sim- pared with the insulin glulisine once- and tiple OADs at study entry. Patients in all ilarnumbersofpatientsinalltreatment thrice-daily arms (Table 2). three treatment arms experienced sub- groups (lixisenatide, 11 of 298 [3.7%]; At week 26, 7-point SMPG profiles stantial improvements in glucose control, insulin glulisine once daily, 11 of 301 were comparable for lixisenatide and in- but the lixisenatide arm also had reduc- [3.7%]; insulin glulisine thrice daily, 14 sulin glulisine once daily, with glucose re- tions in body weight, fewer hypoglycemic of 294 [4.8%]). ductions observed after the evening meal events, and was a simpler regimen com- Three patients died during study pared with the prandial insulin groups. and at bedtime (Fig. 1C). Reduction in glu- treatment; none of these events was cose was also observed after lunch with We propose that with so many effec- adjudicated to be related to the investi- tive glucose-lowering tools available, insulin glulisine thrice-daily (Fig. 1C). gational product. One patient in the HbA should no longer be viewed as A larger proportion of patients treated 1c lixisenatide arm was diagnosed with the primary criterion for the assessment with lixisenatide achieved the composite metastatic pancreatic 35 days af- of treatment success or be the main ba- end points compared with insulin gluli- ter the start of treatment and died after sis upon which to make therapeutic de- sine once daily and thrice daily. In partic- palliative care. Two patients in the insu- cisions. Robust reductions in HbA with ular, patients receiving lixisenatide 1c lin glulisine thrice-daily arm died: one of prandial insulin can occur at the cost of were twice as likely to achieve the triple severe bleeding from a ruptured ankle substantial weight gain, increased oc- composite outcome of HbA ,7% 1c ulcer and the other from exacerbation currence of hypoglycemia, and greater (,53 mmol/mol) without weight gain of chronic heart failure. treatment complexity with the addi- or documented symptomatic hypogly- An investigator reported suspected tional burden and expense from the cemia (Fig. 1D). pancreatitis in one patient in the need for frequent insulin adjustments Safety lixisenatide arm on day 89. As a result, and . Hypoglycemia. Study investigators re- lixisenatide was temporarily discontinued. In contrast to the earlier GetGoal Duo-1 ported the percentage of patients with The symptoms subsequently disappeared, trial (10), patients included in GetGoal symptomatic hypoglycemia was higher with no findings on imaging, and pancre- Duo-2 were not initiating basal insulin; in patients treated with insulin once atic enzyme levels returned to normal. rather, they had been receiving insulin daily or thrice daily versus lixisenatide Treatment with lixisenatide was re- treatment for several years at baseline, (P 5 0.01 and P 5 0.0001, respectively) sumed on day 102 and continued until had more advanced disease, and most (Table 3). Compared with insulin glulisine the planned end-of-treatment visit, and were overweight or obese. This challeng- once daily and thrice daily, a post hoc anal- routine assessment of pancreatic en- ing patient population and the use of two ysis showed that the rate of protocol- zymes showed levels within the normal active comparator regimens in this trial defined symptomatic hypoglycemia events ranges. An independent review commit- highlight the scientific rigor of the study was 25% and 51% lower, respectively, tee adjudicated this case as mild acute design and serve to underscore the with lixisenatide (Table 3). Moreover, pancreatitis. robust nature of our results. an exploratory analysis demonstrated All of the coprimary end points were that protocol-defined nocturnal hypo- CONCLUSIONS met: lixisenatide plus insulin glargine , glycemia (from 2300 to 0600 h) was The GetGoal Duo-2 trial is the first trial was noninferior for HbA1c reductions ver- more common in both insulin glulisine to directly compare prandial lixisenatide sus basal-plus and basal-bolus and was arms versus lixisenatide (estimated with prandial insulin combined with statistically superior for changes in body rateratios[95%CI]of0.6[0.4,0.9]vs. basal insulin glargine and is also the first weight compared with basal-bolus. Symp- 0.5 [0.3, 0.7] for lixisenatide vs. insulin single head-to-head trial to compare a tomatic hypoglycemia occurred in a 1324 Insulin Plus Lixisenatide or Glulisine QD/TID Diabetes Care Volume 39, August 2016

Figure 1—A: Change over time from baseline to week 26 (and last observation carried forward [LOCF]) for mean 6 SE HbA1c in the modified intent-to- treat (mITT) population. B: Change over time from baseline to week 26 (and LOCF) for mean 6 SE body weight (mITT population). All treatment arms are with or without (6) metformin. C:Mean6 SE 7-point SMPG profiles at baseline and week 26 (mITT population). D: Patients achieving composite end points at week 26. E: Hypoglycemia events by hour of the day at week 26 (safety population). Data for symptomatic hypoglycemia per protocol (glucose ,3.3 mmol/L [,60 mg/dL]/recovery with oral carbohydrate if no glucose measurement available). All treatment arms with or without metformin; 70% of patients receiving lixisenatide or insulin glulisine once daily (QD) administered their dose in the evening. TID, thrice daily. care.diabetesjournals.org Rosenstock and Associates 1325

Figure 1—Continued. higher proportion of patients in the basal- and particularly in patients of longer lixisenatide plus insulin glargine, basal- plus and basal-bolus arms than in the type 2 diabetes duration or older age plus, and basal-bolus groups, as expected, lixisenatide plus insulin glargine arm, (14), underscores the necessity of treat- GLP-1 RA class-associated gastrointesti- and nocturnal hypoglycemic events were ment individualization and highlights a nal AEs, predominantly nausea but more frequent with insulin glulisine once need to consider the benefit-to-risk ratio also vomiting and diarrhea, were more daily or thrice daily than with lixisenatide. of any new regimen. common in the lixisenatide plus insulin The negative effect of hypoglycemia in Although the rate of AEs and seri- glargine group. This resulted in a higher patients with type 2 diabetes in general, ous AEs was comparable across the but still relatively small number of 1326 Insulin Plus Lixisenatide or Glulisine QD/TID Diabetes Care Volume 39, August 2016

Table 3—Summary of patients experiencing AEs: safety population Insulin glargine with or without metformin plus Lixisenatide 20 mg once daily Insulin glulisine once daily Insulin glulisine thrice daily Safety parameter (n =298) (n =301) (n =294) Any treatment-emergent AE 221 (74.2) 222 (73.8) 236 (80.3) Any serious treatment-emergent AE 11 (3.7) 11 (3.7) 14 (4.8) Treatment-emergent AE Leading to death 1 (0.3) 0 2 (0.7) Leading to permanent discontinuation 15 (5.0) 2 (0.7) 3 (1.0) Gastrointestinal AEs (system class and preferred term) 105 (35.2) 26 (8.6) 22 (7.5) Nausea 75 (25.2) 5 (1.7) 3 (1.0) Vomiting 26 (8.7) 5 (1.7) 6 (2.0) Diarrhea 20 (6.7) 10 (3.3) 4 (1.4) Pancreatic enzyme increase** Amylase 0 0 0 Lipase 2 (0.7) 1 (0.3) 3 (1.0) Hypoglycemia Symptomatic hypoglycemia* 107 (35.9) 140 (46.5)† 154 (52.4)‡ Symptomatic hypoglycemia per protocol§ 98 (32.9) 117 (38.9) 132 (44.9)| Symptomatic hypoglycemia events per protocol, N¶ 332 395 600 Events by hour of the day 2300 to ,0600 89 136 151 0600 to ,1000 128 111 149 1000 to ,1400 21 44 102 1400 to ,1800 31 54 90 1800 to ,2300 63 49 105 Missing 0 1 3 Severe symptomatic hypoglycemia# 0 2 (0.7) 0 Exploratory analysis Estimated rate ratio lixisenatide-to-insulin glulisine for Symptomatic hypoglycemia events (95% CI)¶ – 0.75 (0.53, 1.06) 0.49 (0.34, 0.69) Nocturnal symptomatic hypoglycemia (95% CI)¶ – 0.58 (0.37, 0.90) 0.47 (0.30, 0.73) All data are n (%) patients with events, unless otherwise stated. **Defined as more than twice the upper limit of normal confirmed by repeat measurement. *Any symptomatic hypoglycemia reported as clinically meaningful by the study investigator regardless of plasma glucose. †P =0.01 vs. lixisenatide (Fisher exact test; post hoc analysis). ‡P = 0.0001 vs. lixisenatide (Fisher exact test; post hoc analysis). §Symptomatic hypoglycemia accompanied by glucose ,3.3 mmol/L (,60 mg/dL) or prompt recovery with oral carbohydrate. |P = 0.0031 vs. lixisenatide (Fisher exact test; post hoc analysis). ¶Number of hypoglycemia events was analyzed using negative binomial regression with a log-link function and the logarithm of the time period for which a hypoglycemic episode was considered treatment-emergent as offset, with treatment, randomization strata of HbA1c, randomization strata of metformin use, and country as fixed effects. #Plasma glucose ,2.0 mmol/L (,36 mg/dL) or with prompt recovery after oral carbohydrate, intravenous glucose, or glucagon administration; 70% of patients receiving lixisenatide or insulin glulisine once daily administered their dose in the evening.

discontinuations overall with lixisenatide A number of studies have directly both studies was the same (7.2% [55 versus the prandial . However, compared the efficacy and safety of mmol/mol]), albeit with lixisenatide 90% completion was observed in the other GLP-1 RAs with basal-plus or once-daily dosing versus lixisenatide group, which is better than basal-bolus, but none have compared twice-daily dosing. the completion rates seen in other GLP-1 both insulin tactics versus a GLP-1 RA The 26-week Harmony 6 study of 566 RA trials (15–17). inthesametrial.Inthe30-week4B patients inadequately controlled on insu- Patients in the lixisenatide arm were study, exenatide twice daily plus insulin lin glargine, with or without metformin twice as likely to achieve the prespeci- glargine demonstrated noninferior and/or , demonstrated non- fied triple-composite outcome of HbA1c HbA1c reductions with benefits in body inferiority of weekly plus insu- ,7% (,53 mmol/mol) without weight weight and significant reductions in non- lin glargine versus and gain or documented symptomatic hypo- nocturnal hypoglycemia versus insulin insulin glargine as basal-bolus in HbA1c glycemia versus the insulin glulisine arms, lispro basal-bolus in patients inade- reductions. Insulin glargine was not opti- confirming a previous meta-analysis (18) quately controlled on basal insulin and mally titrated before randomization in that used propensity score matching to metformin with or without an SU (19). this study, and final HbA1c levels were indirectly compare the efficacy and Baseline HbA1c was slightly higher in the higher than in our study at 7.7 and 7.8% safety of insulin glulisine once daily and 4B study (;8.2% [;66 mmol/mol]) ow- (61 and 62 mmol/mol) in the albiglutide lixisenatide once daily both as add-on to ing to less vigorous titration during the and basal-bolus arms, respectively (20). insulin glargine in 24-week randomized run-in period versus values in GetGoal The 26-week BEGIN: VICTOZA ADD-ON controlled trials. Duo-2; however, the final HbA1c level in study (an addendum study of 413 patients care.diabetesjournals.org Rosenstock and Associates 1327

in whom degludec treatment vs. insu- injectable strategies. Lixisenatide as GlaxoSmithKline, Hoffmann-La Roche, Merck, lin glargine failed [HbA .7% (.53 add-on to basal insulin represents a Merck Sharp & Dohme, Novartis, , 1c fi mmol/mol)]) compared once valuable alternative to treatment inten- and Sano . C.R.-D., E.S., M.W., J.Y., and R.P. are employees of and own stock/shareholders in daily at the highest dose of 1.8 mg as add- sification with basal-plus or basal-bolus, Sanofi. S.H. has served on advisory boards for on to titrated versus in- with a lower risk of hypoglycemia and AstraZeneca; served as a consultant for , sulin aspart basal-plus (21). Baseline without body weight gain, and may Novo Nordisk, and Takeda; and served on speak- HbA was ;7.7% (;61 mmol/mol) in become a preferred therapeutic option. ers bureaus for Boehringer Ingelheim, Eli Lilly, Novo 1c fl fi Nordisk, and Takeda. No other potential con icts BEGIN ADD-ON, and a nal HbA1c level of interest relevant to this article were reported. of ;7.2% (;55 mmol/mol) was reported, Author Contributions. All authors were in- similar to GetGoal Duo-2. volved in the writing, discussion, and review of Acknowledgments. The authors thank the this manuscript. J.R. contributed to the analysis Emerging evidence suggests that GLP-1 study participants, trial staff, and investigators and interpretation of the data and was involved RAs seem to be a more valuable alterna- for their participation. Principal investigators at in the writing of the manuscript. B.G. was a the clinical sites are listed in the Supplementary tive to prandial insulin on top of basal principal investigator and coordinator in France Appendix. Editorial assistance for this publica- insulin. Because GLP-1 RA agents are and contributed to analyzing and discussing the tion was provided by Jane Bryant, PhD (Caudex, found to have differentiating features, results and revising the manuscript. M.H. was an Macclesfield, U.K.), and Sarah Addison, PhD investigator in Germany and contributed to such as frequency of administration (Caudex, London, U.K.), funded by Sanofi. analyzing and revising the manuscript. S.G. and differential effects on FPG and Funding. The GetGoal Duo-2 trial was spon- was involved in the writing, discussion, and sored by Sanofi. PPG, their use in combination therapy review of the manuscript. R.A. contributed to Duality of Interest. J.R. has served on scientific can be individualized further. The pro- the designofthestudy,conductedthe study,and advisory boards and received honoraria or assisted in the writing of the manuscript. F.J.T. nounced PPG reductions associated consulting fees from AstraZeneca, Boehringer was a principal investigator in Spain and con- with lixisenatide make it particularly Ingelheim, Daiichi Sankyo, Eli Lilly, Intarcia, tributed to analyzing and discussing the results suitable for treatment intensification af- Janssen, Lexicon, Merck, Novo Nordisk, and and revising the manuscript. C.R.-D. was the Sanofi and has received grants/research support ter basal insulin because this combina- medical supervisor for the study sponsor, con- from Asahi, AstraZeneca, Boehringer Ingelheim, tion permits improvements in both PPG tributed to the analysis and interpretation of the Bristol-Myers Squibb, Daiichi Sankyo, Eli Lilly, data, and was involved with critical revision and and FPG. Prandial GLP-1 RAs suppress GlaxoSmithKline, Hanmi, Intarcia, Janssen, Lexicon, correction of the manuscript. E.S. contributed to glucagon and also delay gastric empty- MannKind, Merck, Novartis, Novo Nordisk, Pfizer, thedesignofthestudy, wrotethestudyprotocol, Sanofi,andTakeda.B.G.hasservedonadvisory ing, which reduces the rate of postmeal supervised study setup, and contributed to boards for Abbott, Boehringer Ingelheim, Eli Lilly, glucose absorption, resulting in robust analyzing the results and revising the manu- and Novartis; served as a board member for PPG reductions associated with the script. M.W. provided medical oversight of the AstraZeneca, Bristol-Myers Squibb, GlaxoSmithKline, study for the study sponsor, contributed to the meal after administration (7,22). Hence, Novo Nordisk, Roche Diagnostic, and Sanofi; and analysis and interpretation of the data, and was future head-to-head studies comparing has received research support from Dinno Sante,´ involved with critical revision and correction of Eli Lilly, Janssen, Johnson & Johnson, Medtronic, other GLP-1 RAs added to basal insulin the manuscript. J.Y. produced the statistical Merck Sharp & Dohme, Novartis, Novo Nordisk, would be of interest. analyses. R.P. contributed to the study design Sanofi, and Vitalaire. M.H. has served on advisory Limitations of the current study in- and study data discussion; critically reviewed boards for Bristol-Myers Squibb, GlaxoSmithKline, the manuscript, including input into every clude the relatively short 26-week dura- Sanofi, and Takeda and on speakers bureaus for stage of the development of the manuscript; tion and the open-label design; however, Bayer HealthCare, Eli Lilly, GlaxoSmithKline, Roche and approved the finalversion.S.H.wasa Pharmaceuticals, Sanofi, and Takeda. S.G. has blinding would have required additional principal investigator and coordinator in the served as a board member for Sanofi and as a con- injections to mimic the basal-bolus reg- U.K. and contributed to analyzing and discus- sultant (spouse/partner) for Roche Diagnostics. sing the results and revising the manuscript. imen, adding additional complexity to R.A. has served on advisory boards for Janssen, J.R. is the guarantor of this work and, as such, a study that already required ongoing Novo Nordisk, and Sanofi; served as a consultant had full accesstoall ofthedata in thestudyand titration of basal insulin. for AstraZeneca, Bristol-Myers Squibb, Medtronic, takesresponsibilityfortheintegrityofthedata Sanofi, and Takeda; has received research support The findings of the current study and the accuracy of the data analysis. from Abbott, Andromeda Biotech, AstraZeneca, support a change to the current treat- Prior Presentation. This study was presented BD Medical Diabetes Care, Boehringer Ingelheim, orally (OP 13:78) at the 51st Annual Meeting of ment paradigm (4). Historically, prandial Bristol-Myers Squibb, ConjuChem, Diartis the European Association for the Study of Di- insulins have been recommended as Pharmaceuticals, Eli Lilly, Essem Research, abetes, Stockholm, Sweden, 14–18 September fi fi GlaxoSmithKline, ICON, Janssen, Medpace, Med- rst-line intensi cation regimens on 2015, and in poster form (107-LB) at the 75th tronic, Merck, Novartis, Novo Nordisk, Piramal, top of basal insulin (1). We suggest Scientific Sessions of the American Diabetes Asso- Quintiles, Regeneron, Roche, Sanofi, Takeda, that the spectrum of potential benefits ciation, Boston, MA, 5–9 June 2015. fi and Tolerx; and has other relationships with (consistent ef cacy, weight reduction, AstraZeneca, BD Medical Diabetes Care, lower risk of hypoglycemia, and less Boehringer Ingelheim, Bristol-Myers Squibb, References treatment complexity) associated with Novo Nordisk, Sanofi,andTakeda.F.J.T.has 1. Inzucchi SE, Bergenstal RM, Buse JB, et al.; the use of GLP-1 RAs indicate that they served on advisory boards for AstraZeneca, American Diabetes Association (ADA); Euro- should be considered as the first-line Boehringer Ingelheim, Bristol-Myers Squibb, pean Association for the Study of Diabetes Eli Lilly, GlaxoSmithKline, Hoffmann-La Roche, (EASD). Management of in option to advance basal insulin therapy, Merck, Merck Sharp & Dohme, Novartis, Novo type 2 diabetes: a patient-centered approach: with basal plus prandial insulin being re- Nordisk, and Sanofi; served as a consultant for position statement of the American Diabetes served for patients who cannot tolerate AstraZeneca, Boehringer Ingelheim, Bristol-Myers Association (ADA) and the European Associa- GLP-1 RAs due to gastrointestinal AEs Squibb, Eli Lilly, GlaxoSmithKline, Hoffmann-La tion for the Study of Diabetes (EASD). Diabetes Roche, Merck, Merck Sharp & Dohme, Novartis, Care 2012;35:1364–1379 (4). Basal-bolus insulin therapy should Novo Nordisk, and Sanofi; and has received 2. Riddle MC, Rosenstock J, Vlajnic A, Gao L. be relegated to the minority of patients research support from AstraZeneca, Boehringer Randomized, 1-year comparison of three ways who do not respond to the other simpler Ingelheim, Bristol-Myers Squibb, Eli Lilly, to initiate and advance insulin for type 2 1328 Insulin Plus Lixisenatide or Glulisine QD/TID Diabetes Care Volume 39, August 2016

diabetes: twice-daily premixed insulin versus inadequately controlled by established basal insulin: in patients with type 2 diabetes. Di- basal insulin with either basal-plus one prandial a 24-week, randomized, placebo-controlled com- abetes Obes Metab 2011;13:348–356 insulin or basal-bolus up to three prandial injec- parison (GetGoal-L). Diabetes Care 2013;36:2489– 17. Rosenstock J, Reusch J, Bush M, Yang F, tions. Diabetes Obes Metab 2014;16:396–402 2496 Stewart M; Albiglutide Study Group. Potential 3. UK Prospective Diabetes Study (UKPDS) 11. Riddle MC, Rosenstock J, Gerich J; Insulin of albiglutide, a long-acting GLP-1 receptor ag- Group. Intensive blood-glucose control with sul- Glargine 4002 Study Investigators. The treat- onist, in type 2 diabetes: a randomized con- phonylureas or insulin compared with conven- to-target trial: randomized addition of glargine trolled trial exploring weekly, biweekly, and tional treatment and risk of complications in or human NPH insulin to oral therapy of type 2 monthly dosing. Diabetes Care 2009;32:1880– patients with type 2 diabetes (UKPDS 33). diabetic patients. Diabetes Care 2003;26:3080– 1886 Lancet 1998;352:837–853 3086 18. Raccah D, Lin J, Wang E, et al. Once-daily 4. Inzucchi SE, Bergenstal RM, Buse JB, et al. Man- 12. Monnier L, Colette C. Addition of rapid- prandial lixisenatide versus once-daily rapid- agement of hyperglycemia in type 2 diabetes, acting insulin to basal insulin therapy in acting insulin in patients with type 2 diabetes 2015: a patient-centered approach: update to a type 2 diabetes: indications and modalities. mellitus insufficiently controlled with basal in- position statement of the American Diabetes As- Diabetes Metab 2006;32:7–13 sulin: analysis of data from five randomized, sociation and the European Association for the 13. Nathan DM, Zinman B, Cleary PA, et al.; controlled trials. J Diabetes Complications Study of Diabetes. Diabetes Care 2015;38:140–149 Diabetes Control and Complications Trial/ 2014;28:40–44 5. Peyrot M, Rubin RR, Kruger DF, Travis LB. Epidemiology of Diabetes Interventions and 19. Diamant M, Nauck MA, Shaginian R, et al.; Correlates of insulin injection omission. Diabe- Complications (DCCT/EDIC) Research Group. 4B Study Group. Glucagon-like peptide 1 recep- tes Care 2010;33:240–245 Modern-day clinical course of tor agonist or bolus insulin with optimized basal 6. Ahren´ B, Gautier JF, Berria R, Stager W, mellitus after 30 years’ duration: the diabetes insulin in type 2 diabetes. Diabetes Care 2014; Aronson R, Bailey CJ. Pronounced reduction of control and complications trial/epidemiology of 37:2763–2773 postprandial glucagon by lixisenatide: a meta- diabetes interventions and complications and 20. Rosenstock J, Fonseca VA, Gross JL, et al.; analysis of randomized clinical trials. Diabetes Pittsburgh epidemiology of diabetes complica- Harmony 6 Study Group. Advancing basal insu- Obes Metab 2014;16:861–868 tions experience (1983-2005). Arch Intern Med lin replacement in type 2 diabetes inadequately 7. Lorenz M, Pfeiffer C, Steinstrasser¨ A, et al. 2009;169:1307–1316 controlled with insulin glargine plus oral agents: Effects of lixisenatide once daily on gastric emp- 14. Zammitt NN, Frier BM. Hypoglycemia in a comparison of adding albiglutide, a weekly tying in type 2 diabetes–relationship to post- type 2 diabetes: pathophysiology, frequency, GLP-1 receptor agonist, versus thrice-daily prandial glycemia. Regul Pept 2013;185:1–8 and effects of different treatment modalities. prandial insulin lispro. Diabetes Care 2014;37: 8. Meier JJ. GLP-1 receptor agonists for individ- Diabetes Care 2005;28:2948–2961 2317–2325 ualized treatment of type 2 diabetes mellitus. 15. Gallwitz B, BohmerM,SegietT,etal.¨ 21. Mathieu C, Rodbard HW, Cariou B, et al.; Nat Rev Endocrinol 2012;8:728–742 Exenatide twice daily versus premixed insulin BEGIN: VICTOZA ADD-ON (NN1250-3948) study 9. Meier JJ, Rosenstock J, Hincelin-Mery´ A, et al. aspart 70/30 in metformin-treated patients group. A comparison of adding liraglutide Contrasting effects of lixisenatide and liraglu- with type 2 diabetes: a randomized 26-week versus a single daily dose of to tide on post-prandial glycemic control, gastric study on glycemic control and hypoglycemia. insulin degludec in subjects with type 2 diabetes emptying, and safety parameters in patients Diabetes Care 2011;34:604–606 (BEGIN: VICTOZA ADD-ON). Diabetes Obes with type 2 diabetes on optimized insulin 16. Garber A, Henry RR, Ratner R, Hale P, Chang Metab 2014;16:636–644 glargine with or without metformin: a random- CT, Bode B; LEAD-3 (Mono) Study Group. 22. Edwards CM, Stanley SA, Davis R, et al. ized, open-label trial. Diabetes Care 2015;38: Liraglutide, a once-daily human glucagon-like Exendin-4 reduces fasting and postprandial 1263–1273 peptide 1 analogue, provides sustained im- glucose and decreases energy intake in healthy 10. Riddle MC, Aronson R, Home P, et al. Adding provements in glycaemic control and weight volunteers. Am J Physiol Endocrinol Metab once-daily lixisenatide for type 2 diabetes for 2 years as monotherapy compared with 2001;281:E155–E161