FROM RESEARCH TO PRACTICE / THE USE OF IN DIABETES MANAGEMENT

Basal Insulin Use With GLP-1 Receptor Agonists Sarah L. Anderson and Jennifer M. Trujillo

■ IN BRIEF The combination of basal insulin and a -like 1 receptor agonist is becoming increasingly common and offers several potential benefits to patients with type 2 diabetes. Clinical studies have demonstrated improved glycemic control and low risks of hypoglycemia and weight gain with the combination, which provides a safe and effective alternative to basal- bolus insulin with less treatment burden. Fixed-ratio combination products that administer both agents in a single injection are in the pipeline and will University of Colorado Skaggs School of Pharmacy and Pharmaceutical Sciences, offer additional options for clinicians and patients. This review focuses on the Aurora, CO rationale for, clinical evidence on, and implications of using this combination of Corresponding author: Jennifer M. Trujillo, therapies in the treatment of type 2 diabetes. [email protected] ype 2 diabetes is a chronic, pro- abetes achieved a goal A1C <7% in DOI: 10.2337/diaspect.29.3.152 gressive disease that remains the 2007–2010 period, indicating the ©2016 by the American Diabetes Association. Tdifficult to control despite a continued need for effective and du- Readers may use this article as long as the work is properly cited, the use is educational and not growing armamentarium of treat- rable treatment approaches (1). for profit, and the work is not altered. See http:// ment options. National data show The typical clinical course of creativecommons.org/licenses/by-nc-nd/3.0 for details. that only 52.5% of people with di- type 2 diabetes involves periods

152 SPECTRUM.DIABETESJOURNALS.ORG a n d e r s o n a n d t r u j i l l o

TABLE 1. Complementary Characteristics of Basal Insulin and GLP-1 Receptor Agonists Basal Insulin GLP-1 Receptor Agonists Mechanism of action • Augments basal insulin • Increases glucose-dependent insulin secretion from the pancreas • Increases glucose disposal • Decreases glucose-dependent secretion of • Decreases hepatic glucose glucagon production • Slows gastric emptying • Increased satiety Glucose profile • Primarily lowers FPG • Short-acting agents primarily lower PPG effects excursions • Longer-acting agents lower PPG and FPG Effect on weight • Increases weight • Decreases weight FROM RESEARCH TO PRACTICE Main adverse effect • Hypoglycemia • GI issues (nausea, vomiting, diarrhea) • Low risk of hypoglycemia Administration • Subcutaneous injection once or • Subcutaneous injection once or twice daily or requirements twice daily once weekly of adequate glycemic control fol- Rationale for Combining Basal (Table 2). The pharmacokinetics, lowed by glycemic deterioration and Insulin with GLP-1 Receptor pharmacodynamics, dosing, and the need for additional treatment. Agonists clinical effects differ among these Ideally, add-on therapies should pro- Several considerations make the agents. The shorter-acting agents vide significant glucose reduction combination of basal insulin and a ( and ) are asso- with a complementary mechanism GLP-1 receptor agonist appealing ciated with more significant delays in of action, without adding signifi- (Table 1). First, the combination gastric emptying and more targeted cant side effects. Basal insulin (e.g., provides complementary effects on reductions in PPG. Compared to detemir, glargine, or degludec) is a the glucose profile. Basal insulin de- short-acting GLP-1 receptor agonists, recommended option at many stages livers sustained insulin throughout the longer-acting agents (albiglu- of type 2 diabetes and is often added the day, and therapy can be individ- tide, , exenatide XR, and to oral antidiabetic agents (OADs) to ualized by titrating the dose to the ) have less of an effect on achieve target fasting plasma glucose patient’s target FPG levels. However, gastric emptying and rely more on the (FPG) levels. However, it increases barriers such as fear of hypoglyce- insulin secretion mechanism, which the risk of hypoglycemia and weight mia and weight gain may limit the leads to reductions in both FPG initiation and adequate titration of gain, and its dosage must be titrated and PPG (4,12). Of note, although effectively to achieve glycemic tar- basal insulin (5,6). In addition, basal liraglutide and lixisenatide are both gets (2,3). Glucagon-like peptide 1 insulin does not target postprandial dosed once daily, lixisenatide has a (GLP-1) receptor agonists are also glucose (PPG); thus, more than half shorter half-life and primarily lowers reasonable options at many stages of of patients treated with basal insulin type 2 diabetes. These agents effec- do not achieve A1C targets (7–9). PPG, whereas liraglutide has a longer tively lower A1C, show favorable GLP-1 receptor agonists stimulate half-life, thus lowering both PPG and effects on weight, and have a low risk glucose-dependent insulin secretion, FPG (13). These distinctions should of hypoglycemia (4). Their use may inhibit glucose-dependent glucagon be considered when selecting a spe- be limited by gastrointestinal (GI) secretion, slow gastric emptying, and cific GLP-1 receptor agonist to use in adverse effects (AEs), administration increase satiety. Some evidence also combination with basal insulin. requirements, and cost. The combi- suggests that GLP-1 receptor ago- Combining therapy may also nation of basal insulin and a GLP-1 nists may preserve β-cell function potentially lower the risk of AEs. receptor agonist offers several poten- (10,11). Through these mechanisms, Basal insulin may cause weight gain tial benefits to patients with type 2 GLP-1 receptor agonists provide im- and hypoglycemia. GLP-1 receptor diabetes. This review focuses on the provements in both PPG and FPG agonists usually cause weight loss and rationale for, clinical evidence on, levels, complementing the effects of have a low risk of hypoglycemia but and implications of the combination basal insulin. do cause GI AEs. Combining the two of basal insulin and GLP-1 receptor There are currently six GLP-1 classes may allow patients to achieve agonists in the treatment of type receptor agonists approved for use glycemic control with lower doses, 2 diabetes. in either the United States or Europe which could potentially result in

VOLUME 29, NUMBER 3, SUMMER 2016 153 FROM RESEARCH TO PRACTICE / THE USE OF INSULIN IN DIABETES MANAGEMENT

TABLE 2. GLP-1 Receptor Agonists Approved for Use in the United States or Europe Short-Acting Long-Acting Generic name Exenatide Lixisenatide Liraglutide Exenatide XR Dulaglutide (trade name) (Byetta) (Lyxumia) (Victoza) (Bydureon) (Tanzeum, (Trulicity) Eperzan) Year of FDA 2005 Pending 2010 2012 2014 2014 approval Year of EMA 2006 2013 2009 2011 2014 2014 approval Dose 5–10 μg 10–20 μg 0.6–1.8 mg 2 mg 30–50 mg 0.75–1.5 mg Administration* Twice daily Once daily Once daily Once weekly Once weekly Once weekly Delivery Multi-use pen Multi-use pen Multi-use pen Single-use Single-use Single-use pen pen# pen# Renal dosing Not recom- None None Not recom- None None mended if mended if CrCl is <30 CrCl is <30 mL/min; mL/min; use caution if use caution if CrCl is 30–50 CrCl is 30–50 mL/min mL/min *Delivered via subcutaneous injection. #Requires reconstitution. CrCl, creatinine clearance; EMA, European Medicines Agency; FDA, U.S. Food and Drug Administration. fewer AEs than when the individual determined that combination thera- receptor agonist–treated patients; the drugs are used independently. py yielded a mean reduction in A1C most common AEs were predictably Finally, adding a GLP-1 receptor of 0.44% (95% CI –0.60 to –0.29, GI related. agonist to basal insulin offers a poten- P <0.0001) and weight of 3.22 kg Diamant et al. randomized 627 tially safer and easier approach to (95% CI –4.90 to –1.54, P <0.0001). patients with insufficient glycemic achieving glycemic control compared Combination therapy did not increase control despite 12 weeks of intensive to prandial insulin. Initiation of the relative risk of hypoglycemia (rel- protocol-driven glargine titration to prandial insulin increases treatment ative risk 0.99, 95% CI 0.76–1.29, add twice-daily exenatide or thrice- burden for patients and management P <0.0001). Although these findings daily lispro to their regimen (15). burden for providers. It requires one were favorable, only 4 of the 15 stud- Patients receiving exenatide reduced to three injections per day in addi- ies in the meta-analysis used an active their glargine dose by 10% (or more tion to the basal insulin injection, comparator. if their A1C was ≤8%), and patients significant patient education, dose The four active comparator stud- receiving lispro reduced their glargine titration, and increased monitoring ies, plus a fifth study published dose by 33–50%. Exenatide was ini- and increases the risk of hypoglyce- after the meta-analysis was com- tiated at 5 μg twice daily for 4 weeks mia and weight gain. Alternatively, pleted, compared the addition of and then up-titrated to 10 μg initi- adding a GLP-1 receptor agonist a GLP-1 receptor agonist or pran- ated at a dose to replace the reduced involves fewer injections, simpli- dial insulin to basal insulin therapy dose of glargine, divided across three fied dose titration, less education, (15–19). Of these, two evaluated meals. After exenatide or lispro ini- reduced weight, and a lower risk of twice-daily exenatide (15,16); two tiation, glargine was titrated as in hypoglycemia. evaluated once-daily liraglutide and the 12-week optimization phase, and Clinical Evidence lixisenatide, respectively (17,18); and lispro was titrated based on premeal Evidence supporting the use of GLP- one evaluated once-weekly albiglutide glucose values. After 30 weeks of 1 receptor agonists in combination (19). Overall, these studies demon- therapy after the optimization phase, with basal insulin is growing. A meta- strated efficacy with regard to A1C patients in each treatment arm expe- analysis from 2014 described 15 and body weight reductions of basal rienced similar reductions in A1C studies (n = 4,348) in which GLP-1 insulin plus a GLP-1 receptor ago- based on the initial A1C at the end receptor agonists were added to basal nist compared to basal insulin plus of the optimization phase (exenatide insulin therapy or vice versa in type prandial insulin (Table 3). There was –1.13%, 95% CI –1.24 to –1.03; lis- 2 diabetes patients (14). The authors minimal hypoglycemia in the GLP-1 pro –1.10%, 95% CI –1.20 to –1.0).

154 SPECTRUM.DIABETESJOURNALS.ORG a n d e r s o n a n d t r u j i l l o each study: Diamant et al., blood glucose < al., glucose et blood Diamant study: each # cose < cose (26 weeks) (19) al. Rosenstock et (18)§ weeks) (26 al. et Roy-Duval (Duration) TrialClinical (12 weeks) (16) al. et Shao (30 weeks) Diamant (15) (17) weeks) (28 al. et Mathieu P < 0.05 versus0.05 baseline. † 60 mg/dL; and Rosenstock et al., blood glucose < al., glucose et blood mg/dL; Rosenstock 60 and

TABLE 3 . Clinical Trials Comparing Basal Insulin Plus a GLP-1 Receptor Agonist to Prandial Background , metformin, Glargine + Glargine metformin metformin metformin Glargine Glargine Glargine Glargine Degludec pioglita Therapy zone, or None both + P < - ± ± 0.05 versus comparator group. ‡Met noninferiority criteria. &Units/kg/day. §Abstract only. *Definition of hypoglycemia was unique to unique was of hypoglycemia &Units/kg/day. criteria. only. *Definition §Abstract noninferiority ‡Met group. comparator versus 0.05 566 890 627 177 60 n Newly di Newly Unknown of Typeof 2 Diabetes Diabetes Duration agnosed (years) Mean Mean 11.5 11.8 FROM RESEARCH TO PRACTICE 11 54 mg/dL; Shao et al., blood glucose < 54 mg/dL; al., glucose et Shao blood - Comparator Exenatide + Exenatide Lixisenatide (each meal) (each (each meal) (each (each meal) meal) (each (each meal) (each Albiglutide Liraglutide (one meal) (one + glargine Exenatide glargine (largest Aspart Aspart Aspart Aspart Aspart Lispro Lispro meal) Arms 70 mg/dL. **Episodes per patient-year of exposure. NA, not applicable; NR, not reported. not NR, applicable; not NA, exposure. of patient-year per 70 mg/dL. **Episodes

Baseline Baseline A1C (%) A1C . 8 7.6 7.59 7.79 7.72 7.76 8.2 8.5 8.3 8.4 7.7 7.7 Baseline Baseline (mg/dL) 168.3 153.0 153.0 165.4 123.3 126.0 109.8 129.6 115.2 118.4 119.7 FPG FPG 70 mg/dL; Mathieu et al., blood glucose < 70 al., glucose mg/dL; et Mathieu blood Baseline Baseline Insulin (units/ 0.66& 0.70& Basal Basal Dose day) 43.4 61.5 47.0 61.1 NA NA 65 65 67 –0.82‡ 0.74† – –1.34‡ –1.42# –1.31# Δ A1C A1C Δ –1.13‡ –0.66 –0.39 –1.28 –1.55 –1.10 (%) Insulin (units/ 0.64& 0.65& Basal Basal Dose 51.5† Final Final 53.2 day) 56.8 50.6 61† NR NR 64 67 –0.73#† Weight Weight 77#† –7.77# +3.27# +0.81 –0.7† –2.8† –2.5† +0.9 +0.9 +2.1 +1.3 (kg) Δ 56 mg/dL; Roy-Duval et al., glu et mg/dL;56 blood Roy-Duval Hypogly- cemia* 1.00** 8.15** 44.6 15.8 31.5 29.9 37.5 (%) 30 10 41 0 Nausea Nausea 20.7 11.2 (%) 2.9 1.4 20 25 0 0 1 2 0 Vomiting (%) 0.6 6.7 6.7 5.7 1.4 0 0 2 2 9 0 Diarrhea Diarrhea 10.3 (%) 3.3 0.3 4.3 13 1 3 7 0 0 0 -

VOLUME 29, NUMBER 3, SUMMER 2016 155 FROM RESEARCH TO PRACTICE / THE USE OF INSULIN IN DIABETES MANAGEMENT

The difference in A1C between aspart group gained weight (+3.27 and the mean degludec dose in the exenatide and lispro was –0.04% kg, P <0.001). Patients in the exen- aspart group was 0.66 units/kg/day (95% CI –0.18 to 0.11) in the per- atide group experienced decreases at study initiation and 0.64 units/kg/ protocol population and –0.03% in liver enzymes compared to the day at week 26 (P not reported). The (95% CI –0.16 to 0.11) in the intent- aspart group (P <0.001). One-third mean aspart dose at week 26 was 0.21 to-treat population, which met of patients in the exenatide group units/kg/day. Patients receiving lira- noninferiority criteria. The glargine reported a mild to moderate GI AE, glutide lost weight, whereas patients dose decreased from a mean of 61.5 and 10% of patients in the aspart receiving aspart gained weight (–2.8 to 56.8 units in the exenatide group group reported an episode of symp- vs. +0.9 kg, P <0.0001 for the com- and from 61.1 to 51.5 units in the tomatic hypoglycemia. The authors parison). Patients randomized to lispro group (P <0.001 for between- concluded that exenatide was effec- liraglutide were less likely to expe- group difference). Patients receiving tive at decreasing A1C, weight, and rience hypoglycemia than those in exenatide lost weight, whereas those elevated liver enzymes in newly diag- the aspart group (estimated rate ratio receiving lispro gained weight. The nosed type 2 diabetes patients with 0.13, 95% CI 0.08–0.21, P <0.0001). incidences of minor and non-noc- obesity and NAFLD (16). The most common AEs in the lira- turnal hypoglycemia were lower in BEGIN: VICTOZA ADD-ON glutide group were GI-related. The the exenatide group compared to the was a 26-week study by Mathieu authors concluded that the addition lispro group (30 vs. 41%, P = 0.004, et al. that evaluated the addition of of liraglutide was more effective at and 15 vs. 34%, P <0.001, respec- once-daily liraglutide or aspart with reducing A1C than aspart; however, tively). However, the incidence of the largest meal in 177 patients with aspart was only given once daily. nocturnal hypoglycemia was similar type 2 diabetes receiving degludec Liraglutide had positive effects on between groups (25 vs. 27%, P not with or without metformin who weight and was associated with a low reported). GI AEs were more prev- had completed the 52-week BEGIN risk of hypoglycemia (17). alent in the exenatide group, but no ONCE-LONG trial plus a 52-week Roy-Duval et al. evaluated the cases of pancreatitis were observed in extension and had an A1C ≥7% (17). addition of the GLP-1 receptor ago- either group. The authors concluded These 177 patients had a mean A1C nist lixisenatide 20 μg daily versus that twice-daily exenatide was a rea- of 7.7% at the end of the 104 total daily glulisine or thrice-daily glulisine sonable alternative to prandial insulin weeks from previous study enroll- in 890 patients with type 2 diabetes for type 2 diabetes patients with sub- ment. Patients receiving liraglutide poorly controlled on glargine ± met- optimal glycemic control despite had their degludec dose decreased by formin after a 12-week optimization glargine therapy (15). 20% at study initiation, and the dose phase (18). Data are currently pub- Shao et al. also studied the addi- could not be increased until week 6. lished only in abstract form. At the tion of twice-daily exenatide or Liraglutide was initiated at 0.6 mg end of a 12-week run-in period, the prandial insulin to glargine (16). daily for 1 week and increased to 1.2 mean daily doses of glargine were This study enrolled 60 patients with mg daily for 5 weeks. At week 5, if similar between the lixisenatide, newly diagnosed type 2 diabetes, the mean prebreakfast FPG value was daily glulisine, and thrice-daily glu- obesity (BMI ≥28 kg/m2), and non- ≥90 mg/dL, liraglutide was increased lisine groups (67 ± 32 units, 65 ± alcoholic fatty liver disease (NAFLD) to 1.8 mg daily. From week 6 onward, 32 units, and 65 ± 27 units, respec- with elevated liver enzymes. Thirty if the mean prebreakfast FPG value tively). Lixisenatide was deemed patients were randomized to twice- was ≥90 mg/dL, liraglutide was noninferior to either once- or thrice- daily exenatide plus glargine, and 30 increased to 1.8 mg daily, or the daily glulisine with respect to A1C patients were randomized to thrice- degludec dose could be increased per change from the end of the 12-week daily aspart plus glargine. Exenatide the study-specified titration schedule. optimization phase to the end of was initiated at 5 μg twice daily Patients receiving aspart started on a the 26-week study period (–0.59 vs. for 4 weeks and up-titrated to 10 dose of 4 units before their largest –0.52 vs. –0.83%) and was superior μg twice daily for 8 weeks. Insulin meal of the day. Aspart was titrated in terms of weight loss, as patients dosing algorithms and mean daily weekly based on the mean of three lost weight with lixisenatide and doses of insulin were not reported. premeal or bedtime blood glucose gained with glulisine (P <0.0001 for At the end of 12 weeks, A1C had values. At 26 weeks, A1C decreased lixisenatide vs. each glulisine arm). decreased by 1.42% in the exen- more in the liraglutide group than in The mean daily dose of glargine in atide group and 1.31% in the aspart the aspart group (–0.74 vs. –0.39%, the lixisenatide, daily glulisine, and group (both P <0.001 vs. baseline P = 0.0024). The mean degludec thrice-daily glulisine groups remained A1C at diagnosis). Patients in the dose in the liraglutide group was the same or decreased (67 ± 36 units, exenatide group lost weight (–7.77 0.70 units/kg/day at study initiation 64 ± 36 units, and 61 ± 29 units, kg, P <0.001), whereas those in the and 0.65 units/kg/day at 26 weeks, respectively; the treatment difference

156 SPECTRUM.DIABETESJOURNALS.ORG a n d e r s o n a n d t r u j i l l o between the lixisenatide and thrice- common in the patients receiving in the degludec group, and 1.21% in daily glulisine groups was statistically albiglutide. The authors concluded the liraglutide group (P <0.0001 for significant). The patients receiving that once-weekly albiglutide was IDegLira vs. either degludec or lira- lixisenatide experienced significantly comparably efficacious to thrice-daily glutide). The daily degludec dose was less hypoglycemia than those receiv- lispro as add-on therapy for type 2 37% lower in the IDegLira group (39 ing thrice-daily glulisine (P <0.0001); diabetes patients not optimized on units) versus the degludec group (62 however, patients in the lixisenatide glargine ± OADs. Albiglutide was units; estimated treatment difference group experienced more GI AEs com- associated with weight loss, less hypo- –23.4 units, 95% CI –26.4 to –20.3, pared to either glulisine group. The glycemia, and fewer injections per P <0.0001). The rate of confirmed authors concluded that once-daily lix- week compared to lispro (20). hypoglycemia was significantly lower isenatide may be a preferred add-on in the IDegLira group compared to Fixed-Ratio Combinations therapy to glargine for improving the degludec group (rate ratio 0.63, glycemic control while minimizing Two fixed-ratio combinations (FRCs) 95% CI 0.50–0.79, P <0.0001) (25). containing a GLP-1 receptor agonist

hypoglycemia and promoting weight The DUAL II study evaluated the FROM RESEARCH TO PRACTICE loss (18). and a basal insulin in a single injection safety and efficacy of degludec versus Finally, the Harmony 6 study by are in the development pipeline. Novo IDegLira, each in combination with Rosenstock et al. evaluated once- Nordisk has submitted materials for metformin in 413 patients with a weekly albiglutide versus thrice-daily U.S. Food and Drug Administration mean baseline A1C of 8.7–8.8 ± 0.7% lispro as add-on therapy in 566 (FDA) approval of an FRC containing already receiving basal insulin and patients with uncontrolled type 2 degludec and liraglutide (IDegLira) OADs (26). Patients were randomized diabetes despite therapy with glargine (21). The combination product has to receive IDegLira plus metformin ± OADs (19). Patients receiving albi- been approved in Europe under the or degludec plus metformin for glutide initiated at 30 mg weekly and trade name Xultophy and is supplied 26 weeks, with titration aimed to up-titrated to 50 mg weekly if their as a 3-mL pen device containing 100 achieve an FPG of 72–90 mg/dL. A1C was >8% between weeks 8 and units/mL degludec and 3.6 mg/mL li- The maximum doses allowed were 50 12. Patients receiving thrice-daily raglutide (22). submitted ma- units for degludec and 50 units for lispro had their dose initiated based terials for FDA approval of an FRC degludec plus 1.8 mg of liraglutide on self-monitoring of blood glucose containing glargine and lixisenatide for IDegLira. At equivalent deglu- data and were titrated to a prepran- (LixiLan or iGlarLixi) in December dec doses (mean dose of 45 units in dial goal of 80–130 mg/dL. Patients 2015 and redeemed a priority review each group), IDegLira decreased A1C in each group had their glargine voucher to shorten the review time to by more than twice that of degludec titrated in increments of 2–8 units to 6 months (23). (–1.9 vs. –0.9%; estimated treatment a FBG goal of 100 mg/dL. The mean The DUAL I study evaluated difference –1.1%, 95% CI –1.3 to daily glargine dose increase was sim- IDegLira versus each of its individ- –0.8%, P <0.0001). Patients random- ilar between groups over the 26-week ual components in insulin-naive ized to IDegLira lost weight (–2.7 study period (albiglutide 47.0–53.2 patients taking metformin with or kg), whereas patients on degludec units vs. lispro 43.4–50.6 units, P without with a base- maintained their weight (P <0.0001), not reported). At the end of 26 weeks, line A1C of 7–10% (24). At the and the incidence of hypoglycemia albiglutide and lispro decreased A1C end of 26 weeks, patients receiving between groups was similar (24 vs. by 0.82 and 0.66%, respectively IDegLira experienced an A1C reduc- 25%). The most common AEs of (P = 0.053). The treatment difference tion of 1.9% compared to a 1.4% IDegLira were GI-related, but the of –0.16% between the two groups reduction with degludec (IDegLira incidence was much less than that met prespecified noninferiority cri- noninferior; treatment difference observed in other liraglutide stud- teria (19). After 52 weeks, patients –0.47%, 95% CI –0.58 to –0.36, ies; incremental titration allowed for in both arms experienced further P <0.0001) and a 1.3% reduction improved GI tolerability. The authors decreases in A1C (–1.01 vs. –0.84%, with liraglutide (IDegLira superior; concluded that the FRC IDegLira was respectively, P = 0.086), but albiglu- treatment difference –0.64%, 95% superior to degludec for optimizing tide no longer met noninferiority CI –0.75 to –0.53, P <0.0001) (24). glycemic control in patients inad- criteria versus lispro. Patients receiv- A 26-week extension to this study equately controlled on OADs and ing albiglutide lost weight (–0.96 confirmed sustained efficacy and basal insulin (26). kg), whereas those receiving lispro safety in 78% (n = 1,311/1,663) of Post-hoc analyses of the DUAL I gained weight (+1.66 kg, P <0.001). the patients enrolled in the original extension and DUAL II studies deter- Hypoglycemia occurred more fre- DUAL I study (25). At the end of 52 mined that IDegLira is efficacious quently in lispro-treated patients (39 weeks, mean A1C had decreased by irrespective of baseline A1C, duration vs. 23%), but overall, AEs were more 1.84% in the IDegLira group, 1.4% of diabetes, and diabetes treatment

VOLUME 29, NUMBER 3, SUMMER 2016 157 FROM RESEARCH TO PRACTICE / THE USE OF INSULIN IN DIABETES MANAGEMENT at screening. In the DUAL I study, their primary endpoints showing vated PPG and A1C levels. This may IDegLira significantly reduced A1C superior A1C reduction by LixiLan also be a reasonable option for pa- irrespective of baseline A1C; this versus the comparator arms (29,30). tients on a large dose of basal insulin trend held true for the DUAL II A phase IIB study evaluated LixiLan or those who are not willing or able study in all baseline A1C categories versus glargine (each in combination to optimize their basal insulin dose. except for those with an A1C <7.5%. with metformin) in 323 type 2 dia- A GLP-1 receptor agonist would be The dose of insulin and rates of hypo- betes patients with a mean baseline a more straightforward approach to glycemia were lower for the IDegLira A1C of 8%. The results of this study, glycemic control with less treatment versus the degludec group, and as currently reported in abstract form, burden and less training, monitoring, expected, the rate of hypoglycemia demonstrated superior glycemic effi- and dose titration necessary than with was higher in the IDegLira versus cacy with LixiLan versus glargine prandial insulin. In most case scenar- the liraglutide group in DUAL I. In without increased risk of hypogly- ios, it would increase the injection DUAL II, insulin doses and rates of cemia. Patients receiving LixiLan burden only modestly. hypoglycemia were similar between experienced a decrease in A1C from Clinicians should consider groups. In both studies, reduction in a mean of 8.1 to 6.3% after 24 patients’ primary glucose profile A1C with IDegLira was independent weeks. Nearly two-thirds of patients defect to determine whether a short- of type 2 diabetes duration and base- achieved an A1C <7% with no doc- or long-acting GLP-1 receptor agonist line insulin dose but varied based on umented hypoglycemia (31). would be preferred. Adherence to baseline OAD use (27). Clinical Implications administration requirements, The DUAL V study evaluated Treatment guidelines from the patients’ preferences, and insurance the safety and efficacy of IDegLira American Diabetes Association in- coverage also will help to determine compared to continued titration of dicate that either basal insulin or which GLP-1 receptor agonist would in type 2 diabetes GLP-1 receptor agonists are reason- be most appropriate. patients not controlled on metformin able agents to add on as second- or When initiating a GLP-1 receptor and insulin glargine (28). A1C reduc- third-line options (2). In 2015, these agonist, the basal insulin dose may tion was greater in patients taking recommendations were updated to need to be decreased as the GLP-1 IDegLira compared to patients include GLP-1 receptor agonists as receptor agonist starts to take effect. continuing with glargine (–1.81 vs. alternatives to prandial insulin to Across clinical studies, there were –1.13%; estimated treatment dif- treat PPG excursions in patients with varying approaches to adjusting basal ference –0.59%, 95% CI –0.74 to uncontrolled A1C despite reaching insulin doses, with some empirically –0.45%, meeting criteria for nonin- FPG goals with basal insulin (3). decreasing the dose by 10–20% when feriority and superiority, P <0.001). These recommendations are well the GLP-1 receptor agonist was ini- Treatment with IDegLira was asso- supported by clinical evidence and tiated. This basal dose adjustment ciated with weight loss compared offer several advantages, making the should depend on baseline glucose to weight gain with glargine (–1.4 use of this combination in clinical and A1C levels, as well as the onset vs. +1.8 kg, P <0.001). IDegLira practice increasingly more common. of action, dose titration schedule, and had lower rates of hypoglycemia but There likely will be several clinical expected impact on glucose profile of higher rates of GI AEs compared to scenarios in which the combination the specific GLP-1 receptor agonist. insulin glargine (28). may be used. Each scenario requires Patients should be educated about The completed but not yet pub- unique considerations to ensure safe the potential for GI AEs, which likely lished LixiLan-O (n = 1,170) and and effective use. In most of these sit- will be transient in nature, and close LixiLan-L (n = 736) phase III uations, patients likely also would be follow-up is warranted. 30-week clinical studies evaluated taking one or more additional OADs the efficacy and safety of LixiLan. Adding Basal Insulin to a GLP-1 such as metformin. LixiLan-O evaluated LixiLan verses Receptor Agonist each of its individual components Adding a GLP-1 Receptor The addition of basal insulin to a in insulin-naive type 2 diabetes Agonist to Basal Insulin GLP-1 receptor agonist is also well patients not controlled on metformin This scenario is well supported by supported by clinical evidence. This (29). LixiLan-L evaluated LixiLan clinical evidence showing decreased is a reasonable option to consider for compared to glargine in type 2 dia- A1C and weight compared to pla- patients with uncontrolled FPG and betes patients not controlled on basal cebo and similar efficacy with less A1C levels despite treatment with one insulin plus one or two OADs (30). hypoglycemia compared to prandial or more OADs and a GLP-1 receptor Patients in each study continued on insulin. This option would be a rea- agonist. Clinicians should educate pa- metformin throughout. Both stud- sonable treatment option in patients tients about how to titrate their dose ies have been reported to have met with controlled FPG levels but ele- to a target FPG range and, for pa-

158 SPECTRUM.DIABETESJOURNALS.ORG a n d e r s o n a n d t r u j i l l o tients taking a , should maximum dose is 50 units of insulin tion statement of the American Diabetes Association (ADA) and the European evaluate the additive risk of hypogly- degludec and 1.8 mg of liraglutide Association for the Study of Diabetes cemia and consider discontinuation (22). When added to a sulfonylurea, (EASD). Diabetes Care 2012;35:1364–1379 of the sulfonylurea. a reduction of the dose of the sulfo- 3. Inzucchi SE, Bergenstal RM, Buse JB, nylurea should be considered. et al. Management of hyperglycemia in Adding Basal Insulin Plus type 2 diabetes, 2015: a patient-centered a GLP-1 Receptor Agonist Switching to an FRC Product approach: update to a position statement of the American Diabetes Association and Simultaneously It may be appropriate for patients to the European Association for the Study of Adding basal insulin plus a GLP-1 switch from basal insulin alone or a Diabetes. Diabetes Care 2015;38:140–149 receptor agonist simultaneously has GLP-1 receptor agonist alone to an 4. Meier JJ. GLP-1 receptor agonists for been shown to be safe and effective in FRC product. In either scenario, the individualized treatment of type 2 diabetes insulin-naive patients (16,24,25,29). prescribing information for IDegLira mellitus. Nat Rev Endocrinol 2012;8:728–742 recommends a starting dose of 16 5. Cryer PE. Hypoglycemia: the limit- In DUAL-1, adding both simultane- ing factor in glycemic management of ously (IDegLira) allowed patients to units of and 0.6 type I and type II diabetes. Diabetologia use lower doses of each component units of liraglutide (22). This regimen 2002;45:937–948 FROM RESEARCH TO PRACTICE to achieve the same glycemic control could lead to challenges in short-term 6. Peyrot M, Rubin RR, Lauritzen T, et glycemic control for patients who al. Resistance to insulin therapy among as with the basal insulin degludec patients and providers: results of the and glycemic control superior to that were already taking higher doses of cross-national Diabetes Attitudes, Wishes, with liraglutide (24,25). Adding the either component individually. To and Needs (DAWN) study. Diabetes Care combination also negated some of date, transitioning from patients tak- 2005;28:2673–2679 the side effects of each component, ing <20 units or >50 units of basal 7. Blak BT, Smith HT, Hards M, Maguire A, Gimeno V. A retrospective database study particularly the GI AEs of the GLP- insulin per day has not been studied. of insulin initiation in patients with type 2 1 receptor agonist, likely because of For patients who are switching from diabetes in UK primary care. Diabet Med the slower titration. This may be a a long-acting GLP-1 receptor ago- 2012;29:e191–198 reasonable option for a wide range nist, there may be overlapping effects 8. Giugliano D, Maiorino MI, Bellastella G, during the transition because of its Chiodini P, Ceriello A, Esposito K. Efficacy of patients; post-hoc analysis of the of insulin analogs in achieving the hemo- DUAL studies demonstrated efficacy longer duration of action. Either sce- globin A1c target of <7% in type 2 diabetes: independent of duration of disease, nario necessitates close follow-up and meta-analysis of randomized controlled trials. Diabetes Care 2011;34:510–517 background therapies, and baseline education during the transition. 9. Raccah D, Bretzel RG, Owens D, Riddle A1C levels (except for patients with Conclusion M. When basal insulin therapy in type 2 A1C levels <7.5%) (27). Clinical studies have demonstrated diabetes mellitus is not enough—what next? FRCs offer the advantage of fewer improved glycemic control and low Diabetes Metab Res Rev 2007;23:257–264 injections and possibly improved 10. Bunck MC, Diamant M, Corner A, et al. risk of hypoglycemia and weight gain One-year treatment with exenatide improves adherence and less transient GI AEs with the combination of basal insulin beta-cell function, compared to insulin resulting from the slower titration of and a GLP-1 receptor agonist. This glargine, in metformin treated type 2 diabe- the GLP-1 receptor agonist compo- tes patients: a randomized, controlled trial. approach provides a safe and effec- Diabetes Care 2009;32:762–768 nent. Challenges are similar to those tive alternative to basal-bolus insu- 11. Bunck MC, Corner A, Eliasson B, et al. with many combination products. lin with less treatment burden. FRC Effects of exenatide on measures of beta-cell Patients’ AEs related to one com- products are in the pipeline and will function after 3 years in metformin-treated ponent may limit their ability to offer additional options for clinicians patients with type 2 diabetes. Diabetes Care 2011;34:2041–2047 titrate the dose or their persistence and patients. 12. Meier JJ, Rosenstock J, Hincelin-Mery with both components. Patients also A, et al. Contrasting effects of lixisenatide may reach the maximum dose of the Duality of Interest and liraglutide on postprandial glycemic combination but still require more control, gastric emptying, and safety param- Dr. Trujillo is an advisory board consultant eters in patients with type 2 diabetes on basal insulin to reach FPG goals. for Sanofi. No other potential conflicts optimized insulin glargine with or without Prescribing information for IDegLira of interest relevant to this article were metformin: a randomized, open-label trial. indicates that the FRC pen is rec- reported. Diabetes Care 2015;38:1263–1273 ommended for patients with basal 13. Kapitza C, Forst T, Coester HV, References Poitiers F, Ruus P, Hincelin-Mery A. insulin needs of 20–50 units (22). 1. Stark Casagrande S, Fradkin JE, Saydah Pharmacodynamic characteristics of lix- The recommended starting dose of SH, Rust KF, Cowie CC. The prevalence isenatide once daily versus liraglutide once IDegLira is 10 dose steps (10 units of meeting A1C, blood pressure, and LDL daily in patients with type 2 diabetes insuf- goals among people with diabetes, 1988– ficiently controlled on metformin. Diabetes insulin degludec and 0.36 mg liraglu- 2010. Diabetes Care 2013;36:2271–2279 Obes Metab 2013;15:642–649 tide). The dose can be titrated by 2 2. Inzucchi SE, Bergenstal RM, Buse JB, et 14. Eng C, Kramer CK, Zinman B, dose steps every 3 days until an FPG al. Management of hyperglycemia in type 2 Retnakaran R. Glucagon-like peptide-1 target of <90 mg/dL is reached. The diabetes: a patient-centered approach: posi- receptor agonist and basal insulin combina-

VOLUME 29, NUMBER 3, SUMMER 2016 159 FROM RESEARCH TO PRACTICE / THE USE OF INSULIN IN DIABETES MANAGEMENT

tion treatment for the management of type 2 aspx?mID=2978&sKey=880b06fe-7bc0- tio combination of insulin degludec and diabetes: a systematic review and meta-anal- 41d3-8a60-eccaaefd88cd&cKey=a4a9d018- liraglutide (IDegLira). Diabetes Care ysis. Lancet 2014;384:2228–2234 2d00-4efe-a431-b34e39d6635c&mKey 2014;37:2926–2933 =2dbfcaf7-1539-42d5-8dda-0a94abb089e8. 15. Diamant M, Nauck MA, Shaginian R, 27. Rodbard HW, Buse JB, Woo V, et al. Accessed 5 April 2016 et al. Glucagon-like peptide-1 receptor ago- Benefits of combination of insulin degludec nist or bolus insulin with optimized basal 21. Novo Nordisk. Xultophy (NN9068) type and liraglutide are independent of baseline insulin in type 2 diabetes. Diabetes Care 2 diabetes: filed for registration. Available glycated hemoglobin level and duration 2014;37:2763–2773 from http://www.novonordisk.com/rnd/ of type 2 diabetes. Diabetes Obes Metab 16. Shao N, Kuang HY, Hao M, Gao XY, pipelinedetails.1428300648863_2.html. 2016;18:40–48 Accessed 23 February 2016 Lin WJ, Zou W. Effects of exenatide on 28. Lingvay I, Manghi FP, Garcia- obesity and NAFLD with elevated liver 22. Novo Nordisk. Xultophy®. Annex Hernandez P, et al. Effect of insulin glargine enzymes in patients with type 2 diabetes. I: Summary of product characteristics. up-titration vs insulin degludec/liraglutide Diabetes Metab Res Rev 2014;30:521–529 Available from http://ec.europa.eu/health/ on glycated hemoglobin levels in patients 17. Mathieu C, Rodbard HW, Cariou B, documents/community-register/2014/ with uncontrolled type 2 diabetes: the et al. A comparison of adding liraglutide 20140918129550/anx_129550_en.pdf. DUAL V randomized controlled trial. versus a single daily dose of Accessed 20 February 2016 JAMA 2016;315:898–907 to insulin degludec in subjects with type 2 23. Zealand Pharmaceuticals. Zealand 29. Sanofi-Aventis. Sanofi announces that diabetes (BEGIN: VICTOZA ADD-ON). announces that Sanofi has submitted first LixiLan phase III study met primary Diabetes Obes Metab 2014;16:636–644 LixiLan for regulatory review in the US, endpoint [Press Release]. Available from 18. Roy-Duval C, Hanefeld M, Gentile S, et triggering a USD 20 million milestone http://www.news.sanofi.us/2015-07-29- al. Advancing basal insulin glargine with payment [Press Release]. Available Sanofi-Announces-that-First-LixiLan- prandial lixisenatide QD vs insulin glulisine from https://newsclient.omxgroup. Phase-III-Study-Met-Primary-Endpoint. QD or TID in type 2 diabetes: the GetGoal- com/cdsPublic/viewDisclosure. Accessed 20 February 2016 Duo2 evidence-based trial (Abstract #78). action?disclosureId=690565&lang=en. Accessed 20 February 2016 30. Sanofi-Aventis. Sanofi reports positive Diabetologia 2015;58(Suppl. 1):S39 top-line results in second pivotal LixiLan 19. Rosenstock J, Fonseca VA, Gross JL, 24. Gough SC, Bode B, Woo V, et al. phase III study [Press Release]. Available et al. Advancing basal insulin replacement Efficacy and safety of a fixed-ratio combi- from http://www.news.sanofi.us/2015-09-14- in type 2 diabetes inadequately controlled nation of insulin degludec and liraglutide Sanofi-Reports-Positive-Top-Line-Results- with insulin glargine plus oral agents: a (IDegLira) compared with its components in-Second-Pivotal-LixiLan-Phase-III-Study. comparison of adding albiglutide, a weekly given alone: results of a phase 3, open-label, Accessed 20 February 2016 randomised, 26-week, treat-to-target trial in GLP-1 receptor agonist, versus thrice-daily 31. Rosenstock J, Guerci B, Paranjape S, prandial . Diabetes Care insulin-naïve patients with type 2 diabetes. Lancet Diabetes Endocrinol 2014;2:885–893 Berria R, Souhami E, Aroda VR. Improved 2014;37:2317–2325 glucose control without increased hypogly- 20. Fonseca VL, Ahrén B, Chow F, 25. Gough SC, Bode BW, Woo VC, et al. cemia risk at any level of HbA1c reduction et al. Once weekly GLP-1 receptor One-year efficacy and safety of a fixed com- with insulin glargine/lixisenatide fixed-ratio agonist albiglutide vs prandial lispro bination of insulin degludec and liraglutide combination (LixiLan) vs. insulin glargine added to basal glargine in type 2 in patients with type 2 diabetes: results of a alone both added on to metformin in diabetes: similar glycaemic control with 26-week extension to a 26 week main trial. type 2 diabetes (T2DM) [Abstract 169- weight loss and less hypoglycaemia Diabetes Obes Metab 2015;17:965–973 OR]. Presented at the American Diabetes [Abstract]. Available from http://www. 26. Buse JB, Visbøll T, Thurman J, et al. Association’s 75th Scientific Sessions, 5–9 abstractsonline.com/plan/ViewAbstract. Contribution of liraglutide in the fixed-ra- June 2015, Boston. Mass.

160 SPECTRUM.DIABETESJOURNALS.ORG