sign in sign up
<<
Home , Insulin glargine

Diabetes Care 1

Julio Rosenstock,1 Vivian A. Fonseca,2 Advancing Basal Jorge L. Gross,3 Robert E. Ratner,4 Bo Ahren,´ 5 Francis C.C. Chow,6 Fred Yang,7 Replacement in Diane Miller,7 Susan L. Johnson,8 Murray W. Stewart,7 and Inadequately Controlled With Lawrence A. Leiter,9 for the Harmony-6 Plus Oral Agents: Study Group AComparisonofAdding , a Weekly GLP-1 Receptor , Versus Thrice- Daily Prandial DOI: 10.2337/dc14-0001 MRIGTCNLGE N THERAPEUTICS AND TECHNOLOGIES EMERGING

OBJECTIVE GLP-1 receptor may provide an alternative to prandial insulin for advanc- 1Dallas Diabetes and Endocrine Center, Medical ing basal insulin therapy. Harmony 6 was a randomized, open-label, active- City Dallas Hospital, Dallas, TX controlled trial testing once-weekly albiglutide vs thrice-daily prandial insulin 2Tulane University Health Sciences Center, New lispro as an add-on to titrated once-daily insulin glargine. Orleans, LA 3Endocrine Division, Hospital de Cl´ınicas de Porto RESEARCH DESIGN AND METHODS Alegre and Universidade Federal do Sul, Porto Alegre, Brazil – – 4 Patients taking basal insulin (with or without oral agents) with HbA1c 7 10.5% (53 Medstar Health Research Institute, Hyattsville, 91 mmol/mol) entered a glargine standardization period, followed by randomi- MD zation to albiglutide, 30 mg weekly (n = 282), subsequently uptitrated to 50 mg, if 5Department of Medicine, Lund University, Lund, necessary, or thrice-daily prandial lispro (n = 281) while continuing Sweden 6The Chinese University of Hong Kong, Shatin, and/or . Glargine was titrated to fasting plasma of <5.6 New Territories, Hong Kong mmol/L, and lispro was adjusted based on glucose monitoring. The primary end 7GlaxoSmithKline, Upper Merion, PA 8GlaxoSmithKline, Research Triangle Park, NC point was the difference in the HbA1c change from baseline at week 26. 9Keenan Research Center, Li Ka Shing Knowledge RESULTS Institute, St. Michael’s Hospital, and Division of Endocrinology and Metabolism, University of At week 26, HbA1c decreased from baseline by 20.82 6 SE 0.06% (9.0 mmol/mol) Toronto, Toronto, Ontario, Canada with albiglutide and 20.66 6 0.06% (7.2 mmol/mol) with lispro; treatment dif- Corresponding author: Julio Rosenstock, ference, 20.16%(95%CI20.32 to 0.00; 1.8 mmol/mol; P < 0.0001), meeting the [email protected]. noninferiority end point (margin, 0.4%). Weight decreased with albiglutide but Received 1 January 2014 and accepted 24 April increased with lispro (20.73 6 0.19 kg vs. +0.81 6 0.19 kg). The mean glargine 2014. dose increased from 47 to 53 IU (albiglutide) and from 44 to 51 IU (lispro). Adverse reg. no. NCT00976391, clinicaltrials events for albiglutide versus lispro included severe (0 vs. 2 events), .gov. documented symptomatic hypoglycemia (15.8% vs. 29.9%), nausea (11.2% vs. This article contains Supplementary Data online at http://care.diabetesjournals.org/lookup/ 1.4%), vomiting (6.7% vs. 1.4%), and injection site reactions (9.5% vs. 5.3%). suppl/doi:10.2337/dc14-0001/-/DC1 CONCLUSIONS R.E.R. is currently affiliated with the American Diabetes Association, Alexandria, VA. Weekly albiglutide is a simpler therapeutic option than thrice-daily lispro for © 2014 by the American Diabetes Association. advancing basal insulin glargine therapy, resulting in comparable HbA1c reduction See http://creativecommons.org/licenses/by- with weight loss and lower hypoglycemia risk. nc-nd/3.0/ for details. Diabetes Care Publish Ahead of Print, published online June 4, 2014 2 Albiglutide vs. Lispro in Patients on Glargine Diabetes Care

Because b-cell function progressively 2b dose-ranging study, once-weekly al- run-in/stabilization to insulin glargine declines in type 2 diabetes, most people biglutide was associated with numeri- (4–8 weeks), 52-week treatment, with will ultimately require exogenous insu- cally greater improvement in HbA1c the primary outcome at 26 weeks re- lin therapy to meet glycemic goals. A versus twice-daily and sub- ported here, and an 8-week follow-up common approach for initiating insulin stantially less nausea and vomiting (22). (Supplementary Fig. 1). Patients taking therapy is the addition of basal insulin to Adding once-weekly albiglutide to other intermediate- or long-acting insu- oral agent therapy (1). Clinical trials type 2 diabetes insufficiently controlled lins were switched to glargine, and their have demonstrated that basal insulin on insulin glargine may provide glycemic dose was titrated/stabilized over ;8 analogs can effectively treat glycemic control to a difficult-to-treat population weeks. Patients continued taking their targets when structured insulin adjust- in a more convenient and simpler man- current regimen of metformin (MET), ment algorithms are systematically used ner than advancing to the conventional pioglitazone (PIO), and a-glucosidase (2). However, only 50–60% of patients basal/bolus regimen. The present trial inhibitors for the duration of the study. reach glycemic goals, and the number evaluated the efficacy and safety of Use of , glinides, or DPP-4 may be lower in clinical practice if basal once-weekly albiglutide versus thrice- inhibitors was discontinued at week 21 insulin replacement is not properly op- daily prandial lispro added to titrated per protocol. timized (3–5). When the addition of basal glargine in patients with type 2 di- Patients were stratified by HbA1c basal insulin to oral agents is insufficient abetes uncontrolled with glargine with (#8.5% [69 mmol/mol] or .8.5% to achieve adequate glycemic control, or without oral antihyperglycemic [69 mmol/mol]), history of myocardial injections of prandial insulin are often agents. To our knowledge, this is the infarction (yes or no), and current oral added (6). The addition of prandial in- first trial to directly compare these two therapy (MET without PIO, PIO without sulin to basal insulin has been shown to intensification approaches in patients MET, both, or neither). The two treat- further reduce HbA1c levels compared with type 2 diabetes taking basal insulin ment arms were 1) albiglutide, 30 mg with basal insulin alone (7,8). However, with inadequate glycemic control. The weekly, uptitrated to 50 mg if necessary, + this approach is generally limited by hy- full study was 52 weeks in duration. Re- once-daily titrated glargine; or 2)thrice- poglycemia and weight gain (9,10). sults to the primary end point of 26 daily titrated preprandial lispro + once- Recent studies have shown that GLP-1 weeks are reported here. daily titrated glargine. receptor agonists (RAs) used in con- Before week 8, albiglutide uptitration junction with basal insulin have comple- RESEARCH DESIGN AND METHODS was not allowed. If HbA1c was .8% be- mentary effects resulting in meaningful Patients tween weeks 8 and 12, albiglutide could glucose-lowering benefits. Basal insulin This study was conducted in accordance be uptitrated to 50 mg once weekly. analogs provide diurnal and especially with Good Clinical Practice standards, all Similarly, if HbA1c was $7.5% between nocturnal coverage of postabsorptive applicable privacy requirements, and weeks 12 and 26, uptitration of albiglu- periods, reducing hepatic glucose pro- the guiding principles of the Declaration tide was allowed. duction and resulting in improvements of Helsinki. Institutional review board ap- Glargine was titrated similarly to in nocturnal and fasting plasma glucose proval was received, and written in- achieve FPG of ,5.6 mmol/L in both (FPG) levels (2). GLP-1 RAs stimulate in- formed consent was obtained from each groups according to an insulin titration sulin secretion and suppress patient before participation in the study. scheme provided to investigators, which secretion, both in a glucose-dependent Key inclusion criteria were men or they followed at their discretion. Titra- manner, with marked reductions in women aged 18–75 years, type 2 diabe- tion was based on FPG measurements postprandial glucose levels (11–13) tes inadequately controlled on glargine, from the preceding 2 days (based on thereby providing a complementary detemir, or NPH insulin, with or without self-monitoring of blood glucose), and mechanism of action to basal insulin. oral antidiabetes drugs, for $6months recommended insulin increases ranged In addition, weight loss often exper- and ,5years;HbA1c $7.0% (53 from 2 IU to 8 IU, depending on the de- ienced with GLP-1 RAs can counteract mmol/mol) and #10.5% (91 mmol/mol), gree of hyperglycemia. Insulin dose weight gain with exogenous insulin ther- and BMI $20 kg/m2 and #45 kg/m2. could also be decreased in the case of apy. Randomized, placebo-controlled Major exclusion criteria were ongoing hypoglycemia, and was to be decreased trials evaluating use of GLP-1 RAs in con- symptomatic biliary disease or history by 10–15% in the case of severe hypo- junction with basal insulin have demon- of pancreatitis, lipase level above upper glycemia (Supplementary Table 1). strated meaningful reductions in limit of normal (ULN), recent clinically For patients assigned to the lispro glycated hemoglobin A1c (HbA1c), significant cardiovascular or cerebrovas- arm, preprandial insulin was started weight loss, and low risk of hypoglyce- cular disease, and history or family his- based on self-monitored blood glucose mia (14–18). tory of medullary carcinoma or multiple data and distributed among the pa- Albiglutide is a GLP-1 RA composed endocrine neoplasia type 2 (full criteria tient’s meal times at the investigator’s of a GLP-1 dimer fused to recombinant are in the Supplementary Methods). discretion. The lispro dose was titrated human albumin. An amino acid substi- at the investigator’s discretion to tution at position eight of the GLP-1 di- Study Design achieve a preprandial glucose level of mer creates resistance to dipeptidyl Harmony 6 was a randomized, open- 4.4–7.2 mmol/L and peak (1–2 h) post- peptidase-4 (DPP-4) degradation. Albi- label, active-controlled, parallel-group, prandial glucose of ,10 mmol/L based glutide has a half-life of ;5 days, allow- multicenter phase 3 study. The study on average of the two previous days’ ing for weekly dosing (19–23). In a phase comprised four periods: screening, home glucose monitoring results (before care.diabetesjournals.org Rosenstock and Associates 3

and 2 h after each meal and at bedtime; not patients had signs or symptoms sug- both continued beyond rescue. Supplementary Methods). gestive of pancreatitis. Criteria for assess- HbA1c, FPG, and body weight changes Hyperglycemia rescue criteria were ing probability of pancreatitis appear in were analyzed analogous to the primary specified in the protocol based on labo- Supplementary Table 2. The committee end point. ratory values and titration history. Be- also assessed the likelihood that pancre- The proportions of patients meeting fore week 4, patients were not allowed atitis cases were attributable to a study laboratory-based rescue criteria and to receive additional rescue medication drug. subsequently rescued or not, through to lower blood glucose levels. After Anti-albiglutide antibodies were mea- 26 weeks, are summarized. Time to hy- week 4, patients were eligible to receive sured by indirect ELISA. Samples perglycemia rescue up to the 6-month rescue medication if they were not that were confirmed positive for anti- primary end point was evaluated using meeting prespecified HbA1c goals albiglutide antibodies were further Kaplan-Meier curves. The proportion of (weeks 4–12: 9.0% and ,0.5% change tested by ELISA for GLP-1, glucagon, patients achieving meaningful response from baseline; weeks 12–16: 8.5%; and human albumin cross-reactivity, as levels were analyzed by treatment weeks 16–26: 8.0%) and had not well as albiglutide neutralizing activity, comparisons using nonparametric, co- received a recent titration. After notifi- using a cell-based reporter gene assay. variance-adjusted, extended Mantel- cation that a patient in either arm met Haenszel tests with logistic regression predefined laboratory criteria for hyper- Statistical Analysis models as supportive analyses. glycemia rescue, the investigator re- With 250 patients planned in each treat- viewed titration history to determine if ment group, assuming 10% loss to RESULTS full rescue criteria were met. Choice of follow-up, the study would have 94% Overall, 586 patients were enrolled in rescue medication, which could include power to reject the null hypothesis of this study, and 566 received treatment: more intensive insulin titration or the inferiority for HbA1c change from base- 285 with albiglutide and 281 with lispro addition of other glucose-lowering med- line, assuming an expected treatment (Fig. 1). More than 90% of patients in ications, was determined by the investi- group difference of 0.0% and SD of each group completed active treatment gator. Addition of other GLP-1 analogs 1.2%, using a one-sided, two-sample t through week 26. AEs, loss to follow-up, was prohibited, and use of a DPP-4 in- test and a test-wise significance level of and withdrawal of consent were the hibitor as rescue was discouraged. 0.025 for a noninferiority margin of most common reasons for discontinuing The primary objective was to evaluate 0.4%. albiglutide; withdrawal of consent and the efficacy of albiglutide + glargine vs Primary analysis of the HbA1c change loss to follow-up were the most com- thrice-daily lispro + glargine on the from baseline response was analyzed mon reasons for discontinuation among HbA1c change from baseline at week using an ANCOVA model that incorpo- the lispro group. Three albiglutide-treated 26. Secondary efficacy objectives in- rated treatment group, region, age cat- patients did not have both baseline cluded the change from baseline over egory, and prior myocardial infarction as and postbaseline HbA1c values and time in HbA1c andFPG,proportionof covariates, with baseline HbA1c as a con- were excluded from efficacy analyses. patients meeting HbA1c treatment goals tinuous covariate. The treatment effect Most demographic and baseline charac- of ,7.0% (53 mmol/mol) and ,6.5% estimate for the albiglutide + glargine teristics were similar between groups (48 mmol/mol), change from baseline group was evaluated with this ANCOVA (Table 1). Of note, 30 patients (15 per in weight, and time to meeting hyper- model as least-squares means contrast arm) continued treatment glycemia rescue criteria. Safety events relative to the preprandial lispro + glar- at study entry and during the study. of special interest included pancreatitis, gine group. With significance for the There was a reduction in HbA1c from a thyroid tumors, systemic allergic reac- noninferiority test, the superiority test mean (6 SD) baseline of 8.5 6 0.9% tions, immunogenicity, other adverse for the combination of albiglutide and (69 mmol/mol) to 7.7 6 1.1% (61 events (AEs), and abnormalities in labo- glargine relative to combination of glar- mmol/mol) at week 26 in the albiglu- ratory parameters, vital sign measure- gine and preprandial lispro would use a tide group and from 8.4 6 0.9% (68 ments, electrocardiograms, and one-sided superiority test at a signifi- mmol/mol) to 7.8 6 1.1% (62 mmol/mol) physical examinations. Hypoglycemia se- cance level of 0.025. A multiple- in the lispro group. Model-adjusted verity was defined according to Ameri- comparisons adjustment strategy was change from baseline (6 SE) was can Diabetes Association criteria (24). implemented for the multiple inferential 20.82 6 0.06% (9.0 mmol/mol) for al- Potential cardiovascular events were tests among the secondary objectives to biglutide and 20.66 6 0.06% (7.2 adjudicated by a cardiovascular end preserve the study’s nominal criterion mmol/mol) for lispro. The treatment dif- point masked committee, and a separate significance level of 0.05. Missing values ference for albiglutide minus lispro of pancreatitis masked committee for primary end point analysis were im- 20.16% (95% CI 20.32 to 0.00) (1.8 adjudicated potential events of pancrea- puted using the last observation carried mmol/mol) met the prespecified pri- titis. The remit of the pancreatitis adju- forward. For patients who met labora- mary end point of noninferiority (0.4%) dication committee included review of tory-based hyperglycemia rescue criteria to lispro (P , 0.0001) and was close to all reported AEs of pancreatitis, review before week 26 or discontinued from ac- meeting the prespecified primary end of serious AEs to determine if adjudica- tive study participation, the HbA1c at the point of statistical superiority (P = tion for pancreatitis was warranted, and time of rescue was carried forward for 0.0533). Decreases from baseline over review of amylase and/or lipase mea- the primary end point analysis. Assess- time in HbA1c were observed through surements ($3 times ULN), whether or ments and subsequent adjustments of week 26 in both groups, with a steep decline 4 Albiglutide vs. Lispro in Patients on Glargine Diabetes Care

Figure 1—Patient disposition. ITT, intent to treat.

from baseline to week 8 (Fig. 2A). At (20.99 vs. –0.71 mmol/L; P = 0.2366). included in the analysis (least-squares each visit, the magnitude of change from The glargine dose increased similarly in mean difference, 20.06%; 95% CI baseline was numerically greater in the both the albiglutide (baseline, 47.0 IU; 20.22 to 0.11 [0.7 mmol/mol]). Time albiglutide group versus the lispro group. 26 weeks, 53.2 IU) and lispro (baseline, to hyperglycemia rescue evaluated A sensitivity analysis that used observed 43.4 IU; 26 weeks, 50.6 IU) arms. The with a Kaplan-Meier estimate of the HbA1c valueswithnomissingdataimputa- lispro dose increased from 15.5 IU (ini- probability of hyperglycemia rescue at tion showed findings consistent with tial average dose on day of randomiza- 6 months was comparable for the two the intent-to-treat population. An explor- tion) to 30.6 IU. In the albiglutide arm, treatment groups (albiglutide, 22.9%; atory post hoc analysis was also done to 145 patients (51%) uptitrated to 50 mg lispro, 24.4%). examinechangefrombaselineinHbA1c in before week 26. Weight decreased from baseline over patients as a function of whether they At 26 weeks, fewer patients in the time in the albiglutide group and in- continued sulfonylurea therapy. Results albiglutide arm (79 [28%]) than in creased in the lispro group (Fig. 2C). At from patients who discontinued sulfonyl- the lispro arm (107 [38%]) met the week 26, the model-adjusted least- urea were comparable with overall results. laboratory-based criteria for rescue, in- squares mean (SE) weight change from Descriptive statistics excluding patients dicating that they were not achieving baseline for albiglutide was 20.7 6 0.2 who continued on sulfonylurea appear in prespecified glycemic parameters. kg, and the change for lispro was +0.8 6 Supplementary Table 3. Among these patients, numerically 0.2 kg (P , 0.0001). The between-group The HbA1c treatment goal of ,7.0% more patients received rescued therapy difference in weight change from base- (53 mmol/mol) was reached by 30% of in the albiglutide group relative to the line to week 26, 21.5 kg (95% CI 22.1 to patients receiving albiglutide and 25% of insulin lispro group (61 of 79 vs. 62 of 21.0), was significantly different (P , those receiving lispro; HbA1c ,6.5% 107). Of note, patients meeting labora- 0.0001) in favor of albiglutide at each (48 mmol/mol) was reached by 11% of tory criteria for rescue (indicating that time point from week 4 through week 26. albiglutide-treated and 8% of lispro- glycemic parameters were not met at a Overall, AEs and serious AEs occurred treated patients at week 26. Changes specific time point) may not have actu- in a similar proportion in the two study in FPG were consistent with HbA1c ally received rescue therapy in the con- arms (Table 2). Gastrointestinal events changes over time, decreasing in both ventional sense. Rather, they may have (most commonly diarrhea, nausea, and groups at all on-treatment visits (Fig. recently had more intensive insulin titra- vomiting) were relatively low but oc- 2B). Mean FPG values were lower for tion or may have received rescue ther- curred more frequently in the albiglutide- the albiglutide group than for the lispro apy after week 26. Importantly, there treatedgroupandwerethemost group at each visit. The difference did was no difference between treatment common AEs leading to withdrawal not meet statistical significance at the arms in HbA1c change from baseline at (1.4% in the albiglutide arm). However, time of the 26-week primary end point 26 weeks when postrescue values were the incidence of nausea and vomiting care.diabetesjournals.org Rosenstock and Associates 5

Table 1—Demographics and baseline characteristics occurred during this study. Of note, ; Insulin glargine in combination with 8% of the patients were excluded at the screening visit because of a lipase Albiglutide Lispro level greater than the ULN. (n =285) (n =281) P value Age, mean 6 SD, years 54.8 6 9.1 56.3 6 8.9 0.0437 CONCLUSIONS Male, n (%) 132 (46) 136 (48) NS Harmony 6 is the first study to use an Race, n (%) Not donea active comparator, such as prandial lis- Black 39 (14) 34 (12) pro, when a GLP-1 RA was added to Asian 48 (17) 50 (18) basal insulin glargine. Results demon- White 176 (62) 174 (62) Other 31 (11) 24 (9) strated that once-weekly albiglutide Ethnicity, n (%) NS plus basal insulin resulted in clinically Hispanic/Latino 75 (26) 70 (25) meaningful HbA1c reduction (20.82% 2 Weight, mean 6 SD, kg 92.5 6 21.5 91.6 6 21.0 NS [9.0 mmol/mol] vs. 0.66% [7.2 Diabetes duration, mean 6 SD, years 11 6 7116 6NSmmol/mol] with thrice-daily lispro), which met the prespecified noninferior- HbA1c, mean 6 SD, % 8.5 6 0.9 8.4 6 0.9 NS FPG, mean 6 SD, mmol/L 8.5 6 3.0 8.5 6 3.1 NS ity margin after 26 weeks. Moreover, albiglutide was associated with weight OAD therapy, n (%)b NS MET 196 (69) 191 (68) loss and lower rates of hypoglycemia. TZD 7 (3) 4 (1) As expected, more gastrointestinal AEs Both 18 (6) 19 (7) occurred with albiglutide than with lis- Neither 64 (23) 67 (24) pro, although rates of nausea and vom- Prior myocardial infarction 22 (7.7) 27 (9.6) NS iting were low, at ,5% with albiglutide, OAD, oral antidiabetes drug; TZD, . aThe test for race was not done because at all time points. Mild injection site re- patients could check more than one race category; hence, it was not proper to perform test of actions were also more frequent with b proportions between two treatment groups. During the study, 30 patients (15 per arm) stayed albiglutide than with lispro, and low on background sulfonylurea therapy (with or without MET/TZD). rates of anti-albiglutide antibodies were observed. Trials with both twice-daily exenatide was ,5% of patients at all time points in be driven by patients who remained on and and once-daily lixisena- the albiglutide arm, and incidence of sulfonylurea therapy (Supplementary tide in conjunction with basal insulin rel- vomiting decreased to nearly 0 at 26 Table 4). ative to placebo resulted in glycemic weeks (Supplementary Fig. 2). The pro- There were six potential systemic al- improvements, weight loss, and low portion of patients who had events in lergic reactions in the albiglutide group risk of hypoglycemia (14–18). The cur- the prerescue period was similar to and three in the lispro group through rent study adds to this body of knowl- that of the overall population. week 26; none led to withdrawal and edge by demonstrating that add-on Injection site reactions occurred more most were cutaneous in nature. One pa- therapy of once-weekly albiglutide to frequently among those in the albiglu- tient in the albiglutide arm had a serious basal insulin glargine is comparable tide group versus the lispro group (9.5% AE of angioedema attributed to lisinopril. to adding thrice-daily prandial insulin vs. 5.3%). All events of injection site re- The patient discontinued lisinopril and to basal insulin therapy in improving action were mild or moderate in inten- remained on the study drug without fur- glucose control with the added value of sity; 1.1% of patients in the albiglutide ther events. Anti-albiglutide antibody in- weight loss, less hypoglycemia, and the group and 0% in the lispro group with- cidence was 3.3% overall (9 of 269 obvious simplicity advantage of fewer in- drew from active treatment due to events patients, including 2 with preexisting jections. This finding is especially im- captured as injection site reactions. antibodies); all were nonneutralizing. portant in view of the patient-centered Documented prerescue hypoglyce- Among those patients with potential approach to hyperglycemia manage- mia occurred nearly twice as frequently systemic allergic reactions, one patient ment advocated by the American in the lispro arm than in the albiglutide with an event of injection site erythema Diabetes Association/European Associ- arm through week 26 (15.8% of patients tested positive for anti-albiglutide ation for the Study of Diabetes position and 0.9 events/patient/year vs. 29.9% of antibodies. statement, and the alternative of using patients and 2.3 events/patient/year; In one albiglutide-treated patient, a 1 weekly injection versus 21 injections Table 2). Severe hypoglycemia occurred thyroid nodule was discovered after a over a 1-week period would be quite in two lispro-treated (0.7%) and in no baseline elevated level of appealing to patients (25). In addition albiglutide-treated patients. An explor- 140.16 pmol/L (ULN: men, 3.23 pmol/L; to fewer injections, requiring less fre- atory post hoc analysis found that rates women, 1.75 pmol/L), and albiglutide was quent glucose monitoring may be an of hypoglycemia in the small proportion discontinued after one dose. The patient additional cost-savings benefitfor of subjects continuing on sulfonylurea was subsequently diagnosed with medul- many people with type 2 diabetes who was similar to the population overall. lary thyroid with confirmed MEN- need to advance basal insulin replace- Therefore, rates of hypoglycemia in the 2b genotype. No events of potential ment therapy with a basal/bolus overall population did not appear to pancreatitis requiring adjudication regimen. 6 Albiglutide vs. Lispro in Patients on Glargine Diabetes Care

Figure 2—Change over time in mean HbA1c (A), mean FPG (B), and weight (C). (A high-quality color representation of this figure is available in the online issue.)

When compared within the GLP-1 RA because nausea and vomiting early in glargine was not optimized, as shown by class, longer-acting agents are associ- the course of treatment may increase the small increase in insulin glargine dose ated with less gastric emptying and ap- the likelihood of poor adherence, and in both groups after randomization re- parently lower rates of gastrointestinal consequently, poor clinical outcomes. sulting in some further improvements in AEs. Albiglutide has demonstrated Yearly persistence (defined as medica- FPG, but mean values remained above lower rates of nausea and vomiting tion possession ratio .80%) for twice- target. Thus, it is conceivable that greater than liraglutide (26) and like weekly ex- daily exenatide and liraglutide has been reductions in HbA1c would have been enatide, which was also associated with estimated at only 20% and 31%, respec- achieved in both arms had the insulin lower rates versus twice-daily exenatide tively (29). glargine been systematically titrated in a or versus once-daily liraglutide (27,28). The current study has a number of more structured fashion to target FPG Importantly, rates of nausea and vomit- potential limitations. First, it was not ,100 mg/dL. This approach would have ing in the current study were low strictly designed as a treat-to-target tri- allowed better appreciation of what fur- throughout the course of treatment. al, and no glucose-monitoring committee ther improvements could have been This finding may be important for avoid- was involved to enforce the execution of achieved with thrice-daily prandial insulin ing the early discontinuation of therapy prandial insulin titrations. Second, and for versus weekly albiglutide as well the ef- oftenseenwithmostGLP-1RAs, the same reason, titration of basal insulin fect on hypoglycemia and body weight care.diabetesjournals.org Rosenstock and Associates 7

Table 2—AEs occurring in 5% of patients in either arm through week 26 (all events would likely have been achieved if the and events occurring prerescue) and hypoglycemia events (prerescue) basal glargine dose had been systemat- $5% in any treatment group (26 weeks) ically titrated, as shown previously in the twice-daily exenatide added-on to Albiglutide + insulin Lispro + insulin Event glargine (n =285) glargine (n =281) insulin glargine study (14). Although glargine may not have been optimally Overall safety, n (%) titrated in this study, the albiglutide- Any AE 209 (73.3) 199 (70.8) Any serious AE 21 (7.4) 19 (6.8) treated patients showed a numerically Any AE leading to withdrawal 15 (5.3) 1 (0.4) greater decrease in FPG than did the Any related event 75 (26.3) 30 (10.7) lispro-treated patients, despite a similar Most common AEs (%) increase in glargine in both groups. This Diarrhea 13.0 4.3 finding is consistent with the fact that Nausea 11.2 1.4 long-acting GLP-1R agonists have been Urinary tract infection 8.4 6.0 showntohaveagreaterimpactonfast- Vomiting 6.7 1.4 ing than on postprandial glucose levels. Headache 6.7 3.6 Nasopharyngitis 6.3 7.8 One additional study limitation is that Upper respiratory tract infection 6.3 3.2 use of insulin necessitated an open-label Diabetic retinopathy 4.9 6.4 approach, which may have biased the Back pain 3.9 5.0 reporting of events but was unavoidable Edema peripheral 3.2 5.7 given the nature of the compounds Hypertension 2.5 5.0 tested. Most common AEs prerescue period In conclusion, once-weekly albiglu- Any AE, n (%) 200 (70.2) 183 (65.1) tide added-on to basal insulin glargine Diarrhea (%) 12.6 4.3 fi Nausea (%) 10.9 1.4 in type 2 diabetes insuf ciently treated Urinary tract infection (%) 8.1 5.7 with basal insulin and oral antidiabetes Vomiting (%) 6.7 1.4 drugs resulted in comparable glycemic Headache (%) 6.3 3.2 control but with weight loss, lower risk Nasopharyngitis (%) 5.6 7.5 of hypoglycemia, and fewer injections Upper respiratory tract infection (%) 5.3 3.2 per week versus thrice-daily lispro and dd Diabetic retinopathy (%) with a relatively low frequency of gas- Back pain (%) dd Edema peripheral (%) ddtrointestinal AEs for the GLP-1 RA class. Hypertension (%) ddThis result suggests that once-weekly al- Hypoglycemia events (prerescue)a n (%) Events (n)/rateb,c n (%) Events (n)/rateb,c biglutide may represent a reasonable option and a simpler approach to ther- d Any hypoglycemic event 70 (24.6) 188/1.31 107 (38.1) 430/3.09 apy intensification in basal insulin-treated Severe 0 0 2 (0.7) 2/0.01 patients, where health care support and/ Documented symptomatic 45 (15.8) 134/0.94 84 (29.9) 325/2.33 Asymptomatic 19 (6.7) 31/0.22 22 (7.8) 43/0.31 or patient willingness to perform the nec- Probable symptomatic 7 (2.5) 8/0.06 8 (2.8) 16/0.11 essary self-management tasks for basal- Relative 11 (3.9) 15/0.10 11 (3.9) 31/0.22 bolus insulin therapy are limited. Not applicable 0 0 1 (0.4) 1/0.01 Missing 0 0 2 (0.7) 12/0.09 aDefinitions according to Workgroup on Hypoglycemia, American Diabetes Association (ADA), Acknowledgments. The authors thank the 2005. Severity as derived using the ADA guidelines for categorization of hypoglycemic events: following employees of GlaxoSmithKline for severe, requires assistance; documented symptomatic, symptoms, glucose of #3.9 mmol/L; their specific contributions: Caroline Perry, asymptomatic, no symptoms, glucose #3.9 mmol/L; probable symptomatic, symptoms, PhD, for study management support, and Sergio glucose not measured; relative, symptoms, glucose .3.9 mmol/L. bEvent rate per patient-year. Forero-Schwanhaeuser, MD, and Douglas L. cRates obtained post hoc. dPatients with more than one hypoglycemic event were counted in all Wicks, MPH, CMPP, for management of manu- severity categories reported. script development. Editorial support was pro- vided by Beth Burke, PhD (assistance with the production of draft outline, production of manuscript first draft, assembling Tables and when used closer to glycemic targets. A A1chieve trial, the GINGER (Glulisine Figures, and collating author comments), and further treat-to-target study would be re- in Combination with Insulin Glargine Joelle Suchy, PhD (assistance with the produc- quired to understand whether the non- in an Intensified Insulin Regimen) trial, tion of draft outline, production of manuscript fi inferiority finding in glucose control and the AT-LANTUS (A Trial Comparing rst draft, assembling Tables and Figures, collating author comments, fact checking, and would remain under these stricter trial Lantus Algorithms to Achieve Normal referencing), at MediTech Media, Hamilton, New conditions. Blood Glucose Targets in Subjects With Jersey, which was funded by GlaxoSmithKline. Trials in which insulin is more aggres- Uncontrolled Blood Sugar With Type 2 Funding. Funding for this study was provided sively titrated can achieve lower HbA Diabetes Mellitus) trial, HbA values at by GlaxoSmithKline NCT00976391. 1c 1c fi ; Duality of Interest. J.R. has served on scienti c at study end, typically 7% (3,8,30). study end are consistent with what was advisory boards and received honorarium, con- However, in studies where less rigorous seen in the current study (;7.5%) sulting fees, or grants/research support from titration of insulin is done, such as the (7,31–33). Greater reductions in HbA1c GLP-1 RA manufacturers GlaxoSmithKline, Novo 8 Albiglutide vs. Lispro in Patients on Glargine Diabetes Care

Nordisk, Sanofi, Eli Lilly, Roche, , and Annual Meeting of the European Association for 15. Lind M, Jendle J, Torffvit O, Lager I. Glucagon- Intarcia Therapeutics. V.A.F. has received re- the Study of Diabetes, Berlin, Germany, 1–5 like 1 (GLP-1) analogue combined with search support (to Tulane University) in the October 2012 (Fonseca V et al; 473-P). insulin reduces HbA1c and weight with low risk form of grants from Novo Nordisk, Sanofi,Eli of hypoglycemia and high treatment satisfac- Lilly, Daiichi-Sankyo, Pamlabs, Reata Pharma- References tion. Prim Care Diabetes 2012;6:41–46 ceuticals, and Halozyme Therapeutics, and has 1. Unger J. Insulin initiation and intensification 16. DeVries JH, Bain SC, Rodbard HW, et al.; received honoraria for consulting and lectures in patients with T2DM for the primary care phy- Liraglutide-Detemir Study Group. Sequential in- fi from GlaxoSmithKline, Takeda, Novo Nordisk, sician. Diabetes Metab Syndr Obes 2011;4:253– tensi cation of metformin treatment in type 2 fi Sano , Eli Lilly, Daiichi-Sankyo, Pamlabs, XOMA, 261 diabetes with liraglutide followed by random- and AstraZeneca. J.L.G. has received honorar- 2. Rosenstock J, Davies M, Home PD, Larsen J, ized addition of basal insulin prompted by A1C – ium and/or grants/research support for serving Koenen C, Schernthaner G. A randomised, 52- targets. Diabetes Care 2012;35:1446 1454 fi on scienti c advisory boards from GLP-1 RA week, treat-to-target trial comparing insulin de- 17. Riddle MC, Forst T, Aronson R, et al. Adding manufacturers Bristol-Myers Squibb, Eli Lilly, temir with insulin glargine when administered once-daily for type 2 diabetes inad- GlaxoSmithKline, Novo Nordisk, and Roche. R.E.R. as add-on to glucose-lowering drugs in insulin- equately controlled with newly initiated and received research support from GlaxoSmithKline naive people with type 2 diabetes. Diabetologia continuously titrated basal insulin glargine: for the performance of this study. B.A. has received 2008;51:408–416 a 24-week, randomized, placebo-controlled honoraria for lecturing and membership on advi- 3. Rosenstock J, Ahmann AJ, Colon G, Scism- study (GetGoal-Duo 1). Diabetes Care 2013;36: sory boards for GLP-1 RA or DPP-4 inhibitor manu- Bacon J, Jiang H, Martin S. Advancing insulin 2497–2503 facturers Bristol-Myers Squibb, GlaxoSmithKline, therapy in type 2 diabetes previously treated 18. Rosenstock J, Rodbard HW, Bain SC, et al.; Eli Lilly, Novartis, Novo Nordisk, Merck, and with glargine plus oral agents: prandial pre- Liraglutide-Detemir Study Group. One-year sus- Sanofi. F.C.C.C. has received research support mixed (insulin lispro protamine suspension/lis- tained glycemic control and weight reduction in from Boehringer Ingelheim, GlaxoSmithKline, pro) versus basal/bolus (glargine/lispro) type 2 diabetes after addition of liraglutide to Merck, Novartis, Novo Nordisk, Sanofi, and Takeda, therapy. Diabetes Care 2008;31:20–25 metformin followed by accord- and has served on advisory boards for Boehringer 4. Heintjes EM, Thomsen TL, Penning-van Beest ing to HbA1c target. J Diabetes Complications Ingelheim, Eli Lilly, Merck, Novartis, and Novo FJ, Christensen TE, Herings RM. Glycemic con- 2013;27:492–500 Nordisk. F.Y., D.M., S.L.J., and M.W.S. are em- trol and long-acting utilization in 19. Madsbad S, Kielgast U, Asmar M, Deacon C, ployed by GlaxoSmithKline and are eligible or own patients with type 2 diabetes. Adv Ther 2010; Torekov SS, Holst JJ. An overview of once- GlaxoSmithKline stock options or individual corpo- 27:211–222 weekly GLP-1 receptor agonists - available effi- rate shares. F.Y. and M.W.S., as employees of 5. Hermansen K, Davies M, Derezinski T, cacy and safety data and perspectives for the GlaxoSmithKline, have patent rights to GLP-1 RA in Martinez Ravn G, Clauson P, Home P. A 26- future. Diabetes Obes Metab 2011;13:394–407 renal disease. L.A.L. has received research funding week, randomized, parallel, treat-to-target trial 20. Rosenstock J, Stewart MW. Albiglutide. from, has provided continuing medical education comparing insulin detemir with NPH insulin as Drugs Future 2010;35:701 on behalf of, and/or has acted as a consultant to add-on therapy to oral glucose-lowering drugs 21. Bush MA, Matthews JE, De Boever EH, et al. AstraZeneca, Boehringer Ingelheim, Bristol-Myers in insulin-naive people with type 2 diabetes. Di- Safety, tolerability, and Squibb, Eli Lilly, GlaxoSmithKline, Janssen, Merck, abetes Care 2006;29:1269–1274 of albiglutide, a long-acting Novo Nordisk, Sanofi, Servier, and Takeda. No 6. Ampudia-Blasco FJ, Rossetti P, Ascaso JF. glucagon-like peptide-1 mimetic, in healthy other potential conflicts of interest relevant to this Basal plus basal-bolus approach in type 2 diabe- subjects. Diabetes Obes Metab 2009;11:498– article were reported. tes. Diabetes Technol Ther 2011;13(Suppl. 1): 505 Author Contributions. All authors contrib- S75–S83 22. Rosenstock J, Reusch J, Bush M, Yang F, uted to the interpretation of the data and the 7. Davies M, Sinnassamy P, Storms F, Gomis R; Stewart M; Albiglutide Study Group. Potential development and approval of the manuscript. AT.LANTUS Study Group. Insulin glargine-based of albiglutide, a long-acting GLP-1 receptor ag- All authors were fully involved in the manuscript therapy improves glycemic control in patients onist, in type 2 diabetes: a randomized con- development and assume responsibility for the with type 2 diabetes sub-optimally controlled trolled trial exploring weekly, biweekly, and direction and content. J.R. was involved in the on premixed insulin therapies. Diabetes Res monthly dosing. Diabetes Care 2009;32:1880– study design, collection, and interpretation of Clin Pract 2008;79:368–375 1886 the data, the drafting, reviewing, and editing of 8. Holman RR, Farmer AJ, Davies MJ, et al.; 4-T 23. Tomkin GH. Albiglutide, an albumin-based the manuscript, and contributing to the discus- Study Group. Three-year efficacy of complex in- fusion of glucagon-like peptide 1 for the poten- sion. V.A.F., L.A.L., B.A., and S.L.J. were involved sulin regimens in type 2 diabetes. N Engl J Med tial treatment of type 2 diabetes. Curr Opin Mol in interpretation of the data, reviewing and 2009;361:1736–1747 Ther 2009;11:579–588 editing the manuscript, and contributing to the 9. Cryer PE. Hypoglycemia: still the limiting fac- 24. Workgroup on Hypoglycemia, American Di- discussion. J.L.G., R.E.R., and F.C.C.C. were tor in the glycemic management of diabetes. abetes Association. Defining and reporting hy- involved in collection and interpretation of the Endocr Pract 2008;14:750–756 poglycemia in diabetes: a report from the data, reviewing and editing the manuscript, and 10. Gross JL, Kramer CK, Leitão CB, et al.; Di- American Diabetes Association Workgroup on contributing to the discussion. F.Y. was involved abetes and Endocrinology Meta-analysis Group Hypoglycemia. Diabetes Care 2005;28:1245– in study design, conducted statistical analyses, (DEMA). Effect of antihyperglycemic agents 1249 was involved in interpretation of the data, and added to metformin and a sulfonylurea on gly- 25. Inzucchi SE, Bergenstal RM, Buse JB, et al. reviewed and edited the manuscript. D.M. cemic control and weight gain in type 2 diabe- Management of hyperglycaemia in type 2 dia- conducted statistical analysis, was involved in tes: a network meta-analysis. Ann Intern Med betes: a patient-centered approach. Position interpretation of the data, and reviewed and 2011;154:672–679 statement of the American Diabetes Associa- edited the manuscript. M.W.S. was involved in 11. Drucker DJ. The biology of incretin hor- tion (ADA) and the European Association for study design, interpretation of the data, review- mones. Cell Metab 2006;3:153–165 the Study of Diabetes (EASD). Diabetologia ing and editing the manuscript, and contribut- 12. Wajcberg E, Amarah A. Liraglutide in the 2012;55:1577–1596 ing to the discussion. J.R. is the guarantor of this management of type 2 diabetes. Drug Des Devel 26.PratleyRE,NauckMA,BarnettAH,etal. work and, as such, had full access to all the data Ther 2010;4:279–290 Once-weekly albiglutide versus once-daily lira- in the study and takes full responsibility for the 13. Wajcberg E, Tavaria A. Exenatide: clinical glutide in patients with type 2 diabetes inade- integrity of the data and the accuracy of the aspects of the first incretin-mimetic for the quately controlled on oral drugs (HARMONY 7): data analysis. treatment of type 2 diabetes mellitus. Expert a randomised, open-label, multicentre, non- Prior Presentation. Portions of data from this Opin Pharmacother 2009;10:135–142 inferiority phase 3 study. Lancet Diabetes En- study were presented at the 72nd Scientific 14. Buse JB, Bergenstal RM, Glass LC, et al. Use of docrinol 2014;2:289–297 Sessions of the American Diabetes Association, twice-daily exenatide in basal insulin-treated pa- 27. Buse JB, Nauck M, Forst T, et al. Exenatide Philadelphia, Pennsylvania, 8-12 June 2012 tients with type 2 diabetes: a randomized, con- once weekly versus liraglutide once daily in pa- (Rosenstock J et al; 55-OR), and at the 48th trolled trial. Ann Intern Med 2011;154:103–112 tients with type 2 diabetes (DURATION-6): care.diabetesjournals.org Rosenstock and Associates 9

a randomised, open-label study. Lancet 2013; 30. Raskin P, Gylvin T, Weng W, Chaykin L. Com- 32. Fritsche A, Larbig M, Owens D, Haring¨ HU; 381:117–124 parison of insulin detemir and insulin glargine GINGER study group. Comparison between a 28. Buse JB, Drucker DJ, Taylor KL, et al.; using a basal-bolus regimen in a randomized, con- basal-bolus and a premixed insulin regimen in DURATION-1 Study Group. DURATION-1: exe- trolled clinical study in patients with type 2 dia- individuals with type 2 diabetes-results of the natide once weekly produces sustained glycemic betes. Diabetes Metab Res Rev 2009;25:542–548 GINGER study. Diabetes Obes Metab 2010;12: control and weight loss over 52 weeks. Diabetes 31. Home P, Naggar NE, Khamseh M, et al. An 115–123 Care 2010;33:1255–1261 observational non-interventional study of peo- 33. Oguz A, Benroubi M, Brismar K, et al. Clin- 29. Chou E, Wang H, Miao W, Germe M, Fournier ple with diabetes beginning or changed to in- ical outcomes after 24 months of insulin therapy M, Schwartz S. Adherence to GLP-1 agonist ther- sulin analogue therapy in non-Western in patients with type 2 diabetes in five coun- apy in U.S. managed care. Diabetes 2012;61 countries: the A1chieve study. Diabetes Res tries: results from the TREAT study. Curr Med (Suppl. 1):1219-P. Clin Pract 2011;94:352–363 Res Opin 2013;29:911–920