ADA 2019 poster book. Novo Nordisk. Contents. Presentation schedule. Insulin degludec. Insulin degludec/liraglutide. Liraglutide diabetes

Transkript

1 Novo Nordisk ADA 9 poster book Contents Presentation schedule Insulin degludec Insulin degludec/liraglutide Liraglutide diabetes Semaglutide once weekly Semaglutide oral American Diabetes Association 79 th Scientific Sessions June 7, 9 San Francisco, CA, USA Devices Diabetes research Novo Nordisk A/S, Novo Allé, DK-288 Bagsværd, Denmark Tel: Fax: novonordisk.com Changing Diabetes and the Apis bull logo are registered trademarks of Novo Nordisk A/S

2 Presentation schedule Oral presentations Poster presentations SATURDAY 8 JUNE 9 SATURDAY 8 JUNE 9 SUNDAY 9 JUNE 9 SUNDAY 9 JUNE 9 eposter Theater presentations MONDAY JUNE 9 MONDAY JUNE 9

3 Presentation schedule Oral presentations SATURDAY 8 JUNE 9 8:5 AM 8:3 AM Incretins Marching On S OR Montanya 8:3 AM 8:45 AM Incretins Marching On S-57 Oral semaglutide vs. empagliflozin added-on to metformin monotherapy in uncontrolled type 2 diabetes: PIONEER 2 55-OR Pratley Oral semaglutide vs. liraglutide and placebo in T2D: PIONEER 4 SUNDAY 9 JUNE 9 8:45 AM 9: AM Innovations in Insulin Therapy S OR Aroda 9: AM 9:5 AM Innovations in Insulin Therapy S OR Watada DUAL VIII (NCT256): significantly longer time to treatment intensification with insulin degludec/liraglutide (IDegLira) vs. insulin glargine (IGlar U) in a 4-week (wk) randomized trial mirroring clinical practice IDegLira improves glycemic control in Japanese patients with type 2 diabetes (T2D) uncontrolled on premixed insulin therapy

4 Presentation schedule eposter Theater presentations MONDAY JUNE 9 2: PM : PM The Skinny on Incretins eposter Theater A 983-P Buse Effect and safety of flexible dose adjustment with oral semaglutide vs. sitagliptin in type 2 diabetes: PIONEER P Zinman Oral semaglutide as add-on to insulin in T2D: PIONEER 8

5 Presentation schedule Poster presentations SATURDAY 8 JUNE 9 :3 AM 2:3 PM 2-E Clinical Therapeutics/New Technology Insulins Poster Hall SUNDAY 9 JUNE 9 2: PM : PM 2-C Clinical Therapeutics/New Technology Incretin Based Therapies Poster Hall -P Eliasson Use of IDegLira is associated with high persistency rates up to 2 months (M) and improved AC after 6M in Swedish clinical practice 995-P Brown Semaglutide in patients with type 2 diabetes: real-world analysis in the Canadian LMC Diabetes Registry: the SPARE study SUNDAY 9 JUNE P Viljoen Efficacy and safety of semaglutide by baseline BMI in SUSTAIN 5 and 7 2: PM : PM Complications Hypoglycemia Poster Hall 8-LB Kahkoska Outcomes of type 2 diabetes (T2D) clustering replicated in the DEVOTE trial -LB 5-LB Heller Alexopoulos 374-P de Valk Non-severe hypoglycemia predicts increased risk of subsequent severe events in patients with type 2 diabetes Diabetes duration did not moderate the reduction in severe hypoglycemia seen with degludec vs. glargine U in DEVOTE Reduced rates of hypoglycemia irrespective of the definition used when switching to insulin degludec (degludec) from other basal insulins in routine clinical care: the ReFLeCT study -P Pratley 3-P Aroda 4-P Mosenzon 6-P Visaria 3-P Meier Efficacy and safety of semaglutide.5 mg vs. dulaglutide.5 mg once weekly in type 2 diabetes: a post-hoc analysis of SUSTAIN 7 Clinical response (HbA c % and/or body weight 5% reduction) to semaglutide by baseline HbA c and body weight in the SUSTAIN program Oral semaglutide vs. placebo in patients with type 2 diabetes and moderate renal impairment: PIONEER 5 Real-world effectiveness of semaglutide in early users from a US commercially insured (CI) and Medicare Advantage (MA) population Effect of upper gastrointestinal disease on the pharmacokinetics of oral semaglutide in subjects with type 2 diabetes 375-P de Valk Switching to insulin degludec from other basal insulins reduces rates of hypoglycemia across patient subgroups in routine clinical care: the ReFLeCT study 2: PM : PM 2-B Clinical Therapeutics/New Technology Glucose Monitoring and Sensing Poster Hall 89-LB Hachmann- Nielsen (Bergenstal) Low level of agreement in evaluation of continuous glucose monitoring profiles 2: PM : PM 2-C Clinical Therapeutics/New Technology Incretin Based Therapies Poster Hall -P Uusinarkaus Semaglutide-induced weight loss is associated with improved health-related quality of life and treatment satisfaction 2: PM : PM 2-D Clinical Therapeutics/New Technology Insulin Delivery Systems Poster Hall 26-LB Adolfsson 76-P Adolfsson Improved insulin adherence after introduction of a smart connected insulin pen Increased time in range (TIR) observed after introduction of a connected insulin pen -LB 983-P Buse Rosenstock Oral semaglutide vs. sitagliptin: efficacy by baseline HbA c and background OAD in PIONEER 3 Effect and safety of flexible dose adjustment with oral semaglutide vs. sitagliptin in type 2 diabetes: PIONEER 7 MONDAY JUNE 9 2: PM : PM 4-A Pediatrics Obesity and Type 2 Diabetes Poster Hall 329-P Lynch National differences and similarities in youth-onset type 2 diabetes 985-P Zinman Oral semaglutide as add-on to insulin in T2D: PIONEER P Zinman Efficacy of semaglutide by background sodium glucose co-transporter-2 inhibitor: a post hoc analysis of SUSTAIN 9

6 Insulin degludec Late-breaking poster presentation 8-LB Poster presentation 374-P Outcomes of type 2 diabetes (T2D) clustering replicated in the DEVOTE trial Reduced rates of hypoglycemia irrespective of the definition used when switching to insulin degludec (degludec) from other basal insulins in routine clinical care: the ReFLeCT study Late-breaking poster presentation 5-LB Poster presentation 375-P Diabetes duration did not moderate the reduction in severe hypoglycemia seen with degludec vs. glargine U in DEVOTE Switching to insulin degludec from other basal insulins reduces rates of hypoglycemia across patient subgroups in routine clinical care: the ReFLeCT study

7 Insulin degludec Late-breaking poster presentation 8-LB Outcomes of type 2 diabetes (T2D) clustering replicated in the DEVOTE trial AR KAHKOSKA¹, E HACHMANN-NIELSEN², KR KLEIN¹, KG KONGSBAK², K KVIST², JB BUSE¹ University of North Carolina School of Medicine, Chapel Hill, NC, USA; ²Novo Nordisk A/S, Søborg, Denmark T2D is a heterogeneous disease. Individuals in the Swedish All New Diabetics in Scania (ANDIS) cohort with newly diagnosed T2D were grouped by 6 demographic and clinical variables to show 4 distinct T2D subtypes with differential risk for nephropathy and retinopathy. We tested the predictive validity of this clustering system for patients with advanced T2D in DEVOTE (a large, global, randomized, double-blind, cardiovascular outcomes trial; median observation time:.99 yrs) for major adverse cardiovascular event (MACE)-free survival, severe hypoglycemia (SH)-free survival, and overall survival rates. Subjects (N=7637, mean age=65. yrs, mean T2D duration=6.4 yrs, mean glycated hemoglobin [AC]=8.43%) were assigned to a cluster for which they had the smallest Euclidean distance to the cluster center based on available baseline variables: AC, BMI, age, age at diagnosis. Insulin resistance and sensitivity measures were not available. The 4 DEVOTE clusters showed baseline characteristics consistent with the original ANDIS clusters, with significant differences in MACEincidence and SH-incidence (Table). The results were confirmed using data from the LEADER trial (data not shown). The study suggests that clusters derived from early T2D can be replicated in longstanding T2D. Future work should characterize differences in treatment response across clusters to improve outcomes across the heterogeneous T2D population. Poster

8 Insulin degludec Late-breaking poster presentation 5-LB Diabetes duration did not moderate the reduction in severe hypoglycemia seen with degludec vs. glargine U in DEVOTE A-S ALEXOPOULOS¹, K KVIST², E HACHMANN-NIELSEN², JB BUSE³ Duke University, Durham, NC, USA; ²Novo Nordisk A/S, Søborg, Denmark; ³University of North Carolina, Chapel Hill, NC, USA Hypoglycemia is associated with significant morbidity in type 2 diabetes (T2D). Progression of T2D leads to declining counter-regulatory responses and heightened hypoglycemia risk. In this secondary analysis of data from DEVOTE, a CVOT with 7637 patients with T2D at high risk of cardiovascular events, we investigated diabetes duration as a moderator of severe hypoglycemia (SH) with use of glargine U (IGlar U) versus insulin degludec. Patients in DEVOTE were categorized by diabetes duration in years: <, 5, 5 and. Prevalence of SH was examined by treatment arm (Table) using a negative binomial model with log-observation time as an offset term, and diabetes duration categories as moderators. Both proportion of patients with SH events and rate of SH increased with progression of disease in both arms. Degludec preserved its relative benefit over IGlar U for SH across diabetes duration groups, even in less progressed disease groups, i.e. an observed 52% relative risk reduction for degludec vs. IGlar U for patients within years of disease onset (95%CI:.29;.8). There was no significant difference in treatment ratios across categories (p-value for interaction=.5). Similar findings were confirmed in the SWITCH 2 trial (data not shown). Thus, diabetes duration did not moderate the relative reduction in rate of SH for degludec vs. IGlar U in DEVOTE. Poster

9 Insulin degludec Poster presentation 374-P Reduced rates of hypoglycemia irrespective of the definition used when switching to insulin degludec (degludec) from other basal insulins in routine clinical care: the ReFLeCT study HW DE VALK¹, M FEHER 2,3, TK HANSEN 4, J JENDLE 5, Á MERCHANTE 6,7, MM KOEFOED 8, E PARVARESH RIZI 8, E ZIMMERMANN 8, GP FADINI 9 Department of Internal Medicine, University Medical Center Utrecht, Utrecht, The Netherlands; ²Beta Cell Diabetes Centre, Chelsea and Westminster Hospital, London, UK; 3 University of Surrey, Guildford, UK; 4 Steno Diabetes Center Aarhus, Aarhus University Hospital, Aarhus, Denmark; 5 Faculty of Medicine and Health, School of Medical Sciences, Örebro University, Örebro, Sweden; 6 University General Hospital of Castellón, Castellón de la Plana, Spain; 7 Jaume I University, Castellón de la Plana, Spain; 8 Novo Nordisk A/S, Søborg, Denmark; 9 Department of Medicine, Division of Metabolic Diseases, University of Padova, Padova, Italy ReFLeCT was a multicenter, prospective, observational study designed to investigate the safety and effectiveness of switching to degludec from other basal insulins in patients with type (TD) or type 2 diabetes (T2D). ReFLeCT comprised a 4-week baseline period (pre-switch basal insulin) and 2-month follow-up period (degludec). The primary endpoint of overall hypoglycemia reported in patient diaries was reduced during follow-up vs. baseline in TD and T2D without compromising glycemic control. As different hypoglycemia definitions can impact study outcomes, the previous (prespecified) and updated (post hoc) American Diabetes Association (ADA) hypoglycemia definitions were analyzed and presented here (Fig.). Definitions consisted of, documented asymptomatic and symptomatic, pseudo, probable symptomatic, and Level, 2 and 3 (severe) (Fig.). Hypoglycemic events were analyzed using fully adjusted, negative binomial regression models. In TD (n=556) and T2D (n=6), the estimated rate ratios across the previous and the updated ADA hypoglycemia definitions were significantly lower during the 2-month follow-up vs. baseline, except for asymptomatic hypoglycemia in TD (Fig.). Overall, switching to degludec from other basal insulins is associated with lower rates of hypoglycemia across a range of definitions. Poster

10 Insulin degludec Poster presentation 375-P Switching to insulin degludec from other basal insulins reduces rates of hypoglycemia across patient subgroups in routine clinical care: the ReFLeCT study HW DE VALK¹, M FEHER 2,3, TK HANSEN 4, J JENDLE 5, Á MERCHANTE 6,7, MM KOEFOED 8, E PARVARESH RIZI 8, E ZIMMERMANN 8, GP FADINI 9 Department of Internal Medicine, University Medical Center Utrecht, Utrecht, The Netherlands; ²Beta Cell Diabetes Centre, Chelsea and Westminster Hospital, London, UK; 3 University of Surrey, Guildford, UK; 4 Steno Diabetes Center Aarhus, Aarhus University Hospital, Aarhus, Denmark; 5 Faculty of Medicine and Health, School of Medical Sciences, Örebro University, Örebro, Sweden; 6 University General Hospital of Castellón, Castellón de la Plana, Spain; 7 Jaume I University, Castellón de la Plana, Spain; 8 Novo Nordisk A/S, Søborg, Denmark; 9 Department of Medicine, Division of Metabolic Diseases, University of Padova, Padova, Italy ReFLeCT, a multicenter, prospective, observational study evaluated the safety and effectiveness of switching from other basal insulins to insulin degludec (degludec) in patients with type (TD) or type 2 diabetes (T2D) in routine clinical practice. ReFLeCT comprised a 4-week baseline period (pre -switch basal insulin) and 2-month follow-up period (degludec). The primary endpoint of overall hypoglycemia reported in patient diaries was reduced during the 2-month follow-up period vs. baseline, without compromising glycemic control. In pre-specified subgroup analyses of the primary endpoint, we assessed if the overall result was robust in different subgroups, characterized according to baseline AC (<7.5, 7.5 <8.5, 8.5 <9.5, 9.5%), diabetes duration (quartiles) and physician s reason for initiating degludec (hypoglycemia [Yes/No]). The estimated rate ratios of hypoglycemia were similar within subgroups (no significant interactions), and demonstrated overall lower rates (the majority significantly lower) during the 2-month followup periods vs. baseline in patients with TD or T2D (Figure). Irrespective of baseline characteristics or physician s reason for initiating degludec, switching to degludec from other basal insulins reduced rates of overall hypoglycemia in patients with TD or T2D, in routine clinical practice. Poster

11 Insulin degludec/liraglutide Oral presentation 47-OR Poster presentation -P DUAL VIII (NCT256): significantly longer time to treatment intensification with insulin degludec/liraglutide (IDegLira) vs. insulin glargine (IGlar U) in a 4-week (wk) randomized trial mirroring clinical practice Use of IDegLira is associated with high persistency rates up to 2 months (M) and improved AC after 6M in Swedish clinical practice Oral presentation 48-OR Publication only IDegLira improves glycemic control in Japanese patients with type 2 diabetes (T2D) uncontrolled on premixed insulin therapy Improved glycemic control among patients newly prescribed IDegLira across various background therapies in US real-world practice

12 Insulin degludec/liraglutide Oral presentation 47-OR DUAL VIII (NCT256): significantly longer time to treatment intensification with insulin degludec/liraglutide (IDegLira) vs. insulin glargine (IGlar U) in a 4-week (wk) randomized trial mirroring clinical practice VR ARODA,2, G GONZÁLEZ-GALVEZ³, M HALUZÍK 4, RJ SILVER 5, R GRØN 6, N HALLADIN 6, P ÖRSY 6, G SESTI 7 Brigham and Women s Hospital, Boston, MA, USA; ²MedStar Health Research Institute, Hyattsville, MD, USA; ³Jalisco Institute of Research in Diabetes and Obesity S.C., Guadalajara, Jalisco, Mexico; 4 Institute for Clinical and Experimental Medicine and Institute of Endocrinology, Prague, Czech Republic; 5 Southern New Hampshire Diabetes and Endocrinology, Nashua, NH, USA; 6 Novo Nordisk A/S, Søborg, Copenhagen, Denmark; 7 Department of Medical and Surgical Sciences, University Magna Graecia of Catanzaro, Catanzaro, Italy Patients (pts) (n=2) with T2D (AC 7 %) on oral antidiabetic drugs (OADs) were randomized : to open label IDegLira or IGlar U in a 4-wk trial to assess treatment durability. The primary endpoint was time from randomization to treatment intensification (AC 7.% at 2 consecutive visits including Wk 26); pts who met the primary endpoint discontinued study drug. Baseline characteristics were similar. Over 4 wks, fewer pts with IDegLira required intensification vs. IGlar U (37.4% vs. 66.2%). Pts treated with IDegLira had a significantly longer time to intensification (median: >2 years/~ year for IDegLira/IGlar U; Fig.). There was greater effect with IDegLira vs. IGlar U after 4 wks, had intensification not been needed, in terms of: pts achieving AC <7% (55.7 vs. 28.5%), and AC <7% with no weight gain (.9 vs. 6.3%), lower estimated mean insulin dose (36 vs. 5 U; estimated treatment difference 4.9 U), and 56% lower rate of severe or blood glucose confirmed symptomatic hypoglycemia (.38 vs..86 events/patientyear of exposure, (p<. for all). Safety results were similar. Improved long-term glycemic control, evidenced by significantly longer time to treatment intensification, was achieved with IDegLira vs. IGlar U in pts previously uncontrolled on OADs.

13 Insulin degludec/liraglutide Oral presentation 48-OR IDegLira improves glycemic control in Japanese patients with type 2 diabetes (T2D) uncontrolled on premixed insulin therapy H WATADA¹, BF AGNER², A DOSHI³, R GRØN², MF RANTHE², LK BILLINGS 4,5 Department of Metabolism and Endocrinology, Juntendo University, Tokyo, Japan; ²Novo Nordisk A/S, Søborg, Denmark; ³PrimeCare Medical Group, Houston, TX, USA; 4 NorthShore University Health System, Evanston, IL, USA; 5 University of Chicago Pritzker School of Medicine, Chicago, IL, USA Diabetes management is moving towards choosing therapies that reduce treatment burden, risk of hypoglycemia and weight gain. DUAL VII showed that IDegLira, a fixed-ratio insulin degludec (degludec)/liraglutide combination, had non-inferior AC lowering, lower hypoglycemia rates, no weight gain and less treatment burden vs. basal bolus therapy. In contrast, there are no trial data on the efficacy and safety of switching patients (pts) from premixed insulin therapy (with basal and bolus components) to a fixed-ratio combination such as IDegLira. DUAL II Japan (NCT29948) a 26 week phase 3a multicenter, randomized, double-blind, treat-to-target trial compared the efficacy and safety of IDegLira vs. 5 U degludec in Japanese pts with uncontrolled T2D (AC 7.5.%) on 5 U basal or premixed insulin therapy + metformin ± an oral antidiabetic drug. Primary analysis confirmed superior change in AC with IDegLira vs. degludec 5, and post hoc analyses found the treatment effect was independent of insulin type at baseline (basal or premixed; interaction test, p=.5229). This post hoc evaluation describes novel data in a subset of 39 pts who switched from premixed insulin to IDegLira. Mean (standard deviation [SD]) AC was 8.3% (.73) at baseline and 6.7% (.93) at Week 26. Mean (SD) body weight reduced by.5 (2.9) kg. Daily insulin dose was 4. (3.4) U at baseline and 34.2 (.5) U at Week 26. Rate of hypoglycemic events (defined as severe or plasma glucose <56 mg/dl) was 2.6 events/patient-year of exposure (52 events in 7 pts). No pts withdrew. This post hoc study is the first to evaluate the switch from premixed insulin to IDegLira, in pts with uncontrolled T2D. IDegLira initiation resulted in improved AC and weight loss with most pts not experiencing hypoglycemic events. This study offers insight into the effectiveness and safety of switching pts from premixed insulin therapy to IDegLira, and provides a support for further investigation into this transition.

14 Insulin degludec/liraglutide Poster presentation -P Use of IDegLira is associated with high persistency rates up to 2 months (M) and improved AC after 6m in Swedish clinical practice B ELIASSON, J EKELUND 2, A-M SVENSSON¹, M MIFTARAJ 2, A-C MÅRDBY 3,4, JDDR FERNANDES 5 Department of Molecular and Clinical Medicine, University of Gothenburg, Gothenburg, Sweden; Department of Medicine, Sahlgrenska University Hospital, Gothenburg, Sweden; 2 Centre of Registers Västra Götaland, Gothenburg, Sweden; 3 Novo Nordisk, Malmö, Sweden; 4 Section of Epidemiology and Social Medicine, Institute of Medicine, University of Gothenburg, Gothenburg, Sweden; 5 Novo Nordisk A/S, Søborg, Denmark This retrospective study examines outcomes associated with use of IDegLira a fixed-ratio insulin degludec/liraglutide combination in a large real-world cohort of patients with type 2 diabetes in Sweden. Data were obtained from the Swedish National Diabetes Registry on patients who initiated IDegLira (index date) until the end of 7 (or 3 June/ January 7 for 6M/2M for persistency analyses). Persistency at 6M was defined as more IDegLira withdrawal within 6M of index date, and 2M persistency was more withdrawal in M6 2. Clinical variables were assessed at baseline (closest measurement to index date in previous 2M) and 6M ± 45 days after IDegLira initiation. Mean daily IDegLira dose was calculated from volume withdrawn during defined time periods of 6M. Baseline characteristics for overall cohort (N=3) are in Table. At 6 and 2M, 228/2432 (94%) and 482/76 (84%) patients were persistent, respectively. In persistent patients, least squares mean [95% CI] change in AC and body mass index was -.9% [.;.8] and.3 kg/m² [.5;.2] at 6M, respectively and mean daily IDegLira dose was U at all 8-day periods assessed (Table). After 6M, IDegLira use in Sweden is associated with improved AC and no clinically significant change in BMI at a moderate mean daily insulin dose, and 84% were persistent at 2M. Poster

15 Insulin degludec/liraglutide Publication only Improved glycemic control among patients newly prescribed IDegLira across various background therapies in US real-world practice LE EGEDE¹, A BOGDANOV², L FISCHER², JDDR FERNANDES³, L KALLENBACH² Medical College of Wisconsin, Milwaukee, WI, USA; ²Veradigm Health, San Francisco, CA, USA; ³Novo Nordisk A/S, Søborg, Denmark IDegLira is a fixed-ratio combination of degludec and liraglutide indicated for the treatment of type 2 diabetes (T2D). This is the first real-world study describing change in glycated hemoglobin (HbA c ) among US patients initiating IDegLira. The primary objective was to observe the effect in metabolic control of initiating IDegLira in adult T2D patients previously on basal insulin (Basal), glucagon-like peptide- receptor agonist as only injectable (GLP-), and no injectable therapy (N-INJ) between March 7 and June 8 (index date). Patients with HbA c data 6 months prior to and 6m post index date were included from the Practice Fusion Electronic Health Record database. Of the,384 IDegLira patients, 6 patients had baseline, and follow-up HbA c (8 ± 45 days) data. Changes in clinical outcomes were evaluated by paired t-test. Overall there was significant change in HbA c of.9% (p<.), and there was an HbA c reduction in all three background therapy groups, ranging from.7 in the Basal to. in N-INJ (p<.). There was no significant change in weight in the overall group despite initiating or intensifying insulin therapy with IDegLira, as expected weight in the GLP- subgroup increased (3. lb; p<.5). Consistent with previous real-world studies, IDegLira is effective at lowering HbA c across different background therapies with minimal impact on weight.

16 Liraglutide diabetes Late-breaking poster presentation -LB Non-severe hypoglycemia predicts increased risk of subsequent severe events in patients with type 2 diabetes

17 Liraglutide diabetes Late-breaking poster presentation -LB Non-severe hypoglycemia predicts increased risk of subsequent severe events in patients with type 2 diabetes S HELLER¹, E HACHMANN-NIELSEN², K KVIST² University of Sheffield, Sheffield, UK; ²Novo Nordisk A/S, Søborg, Denmark It is well-known that higher rates of non-severe hypoglycemic episodes (NSHEs) associate with a greater risk of severe hypoglycemic episodes (SHEs) in patients with type diabetes. We investigated whether a similar association existed in patients with type 2 diabetes (T2D). We explored if annual rate of NSHEs was associated with time to first SHE, cardiovascular (CV) death, time to first major adverse CV event (MACE), and all cause death in patients with T2D using data from LEADER, a CV outcomes trial with 934 patients with T2D. We used a Cox proportional hazards model adjusted for randomized treatment arm, and annual rate of NSHE as a time-dependent covariate with three levels; A: <2 NSHEs per year (reference), B: 2 NSHEs, or C: 2 NSHEs. Hazard ratios were used to estimate the association between NSHE and each of the outcomes. Higher rates of NSHE were associated with a higher rate of severe hypoglycemia, MACE, CV death, and all-cause death in patients with T2D (Figure); our results suggest that in this T2D population, a high rate of NSHEs may be associated with more harmful outcomes. Poster

18 Semaglutide once weekly Poster presentation 986-P Poster presentation 3-P Efficacy of semaglutide by background sodium glucose co-transporter-2 inhibitor: a post hoc analysis of SUSTAIN 9 Clinical response (HbA c % and/or body weight 5% reduction) to semaglutide by baseline HbA c and body weight in the SUSTAIN program Poster presentation 995-P Poster presentation 6-P Semaglutide in patients with type 2 diabetes: real-world analysis in the Canadian LMC Diabetes Registry: the SPARE study Real-world effectiveness of semaglutide in early users from a US commercially insured (CI) and Medicare Advantage (MA) population Poster presentation 998-P Poster presentation -P Efficacy and safety of semaglutide by baseline BMI in SUSTAIN 5 and 7 Semaglutide-induced weight loss is associated with improved health-related quality of life and treatment satisfaction Poster presentation -P Efficacy and safety of semaglutide.5 mg vs. dulaglutide.5 mg once weekly in type 2 diabetes: a post hoc analysis of SUSTAIN 7

19 Semaglutide once weekly Poster presentation 986-P Efficacy of semaglutide by background sodium glucose co-transporter-2 inhibitor: a post hoc analysis of SUSTAIN 9 B ZINMAN¹, V BHOSEKAR², RS BUSCH³, I HOLST 4, B LUDVIK 5, J THRASHER 6, VC WOO 7, A PHILIS-TSIMIKAS 8 Lunenfeld-Tanenbaum Research Institute, Mount Sinai Hospital, Toronto, ON, Canada; 2 Novo Nordisk, Bengaluru, India; 3 Albany Medical Center Division of Community Endocrinology, Albany, NY, USA; 4 Division of Endocrinology and Metabolism, Medical University of Vienna, Vienna, Austria; 5 Novo Nordisk A/S, Søborg, Denmark; 6 Arkansas Diabetes and Endocrinology Center, Little Rock, AR, USA; 7 University of Manitoba, Winnipeg, MB, Canada; 8 Scripps Whittier Diabetes Institute, San Diego, CA, USA The SUSTAIN clinical trials demonstrated the efficacy and safety of once-weekly subcutaneous semaglutide, a glucagon-like peptide- analog for the treatment of type 2 diabetes (T2D). SUSTAIN 9 investigated semaglutide. mg vs. placebo as add-on to a stable dose of sodium glucose co transporter-2 inhibitor (SGLT-2i) therapy, with or without metformin or a sulfonylurea. The primary and secondary endpoints were, respectively, change from baseline in HbA c and body weight at week 3. In this post hoc analysis, SUSTAIN 9 data were analyzed by background SGLT- 2i (empagliflozin, canagliflozin, dapagliflozin or other [ipragliflozin, luseogliflozin and tofogliflozin; drugs available only in Japan]). In total, 32 subjects were randomized to semaglutide or placebo. Reductions in HbA c and body weight were greater with semaglutide vs. placebo. There was no significant interaction between background SGLT 2i and treatment effect (interaction p-value >.5 for both endpoints), with a smaller observed weight reduction in the Other group (Table). No safety concerns were identified when adding semaglutide to SGLT-2i therapy. No diabetic ketoacidosis or lower limb amputation events occurred. In conclusion, in subjects with T2D already receiving an SGLT-2i, semaglutide generally resulted in superior HbA c and body weight reductions vs. placebo, regardless of background SGLT-2i therapy. Poster

20 Semaglutide once weekly Poster presentation 995-P Semaglutide in patients with type 2 diabetes: real-world analysis in the Canadian LMC Diabetes Registry: the SPARE study RE BROWN, AR LIU 2, R MAHBUBANI 2, R ARONSON¹ ¹LMC Diabetes & Endocrinology, Toronto, ON, Canada; ²Novo Nordisk Canada Inc., Mississauga, ON, Canada The Canadian LMC Diabetes Registry will evaluate real-world clinical outcomes in GLP- RA naïve patients with T2D who initiate once weekly semaglutide as part of their usual diabetes therapy in the first year of semaglutide availability in Canada. Primary outcome is change in AC at 3 6 months follow-up. In this interim retrospective analysis of the first six months of semaglutide availability, 542 patients who initiated semaglutide were eligible (37% on. mg dose and 63% on.5 mg dose). They had a mean age of 57.3 ±. years, T2D duration of 2. ± 7.4 years, and baseline AC of 8.23 ±.34%. At 3-6 months follow-up, 466 patients who remained on-treatment showed an AC reduction of.88 ±.5% (p<.), with 36% of patients achieving an AC reduction.% and 43% of patients achieving ISAC 7.%. Similar benefit in AC lowering was seen in subgroups based on prior diabetes therapy, concomitant insulin use and prior use of DPP-4i (Figure). Mean weight reduction was 3. ± 3.6 kg and 23.6% of patients achieved a weight loss 5%. Reductions were also seen for systolic BP (5. ± 3. mmhg, p<.), diastolic BP (2.2 ± 9.3 mmhg, p<.) and alanine aminotransferase (.7 ± 2.9 U/L, p=.3). In this real-world analysis of a large Canadian cohort, patients initiating semaglutide had significant reductions in AC and weight at three to six months, regardless of prior and concomitant diabetes therapy. Poster

21 Semaglutide once weekly Poster presentation 998-P Efficacy and safety of semaglutide by baseline BMI in SUSTAIN 5 and 7 A VILJOEN, JP FRIAS 2, T GONDOLF 3, O HANSEN 3, N WIJAYASINGHE 3, J UNGER 4 Borthwick Diabetes Research Centre, Lister Hospital, Stevenage, UK; ²National Research Institute, Los Angeles, CA, USA; ³Novo Nordisk A/S, Søborg, Denmark; 4 Catalina Research Institute, LLC. Monclair, CA, USA Semaglutide, a once-weekly glucagon-like peptide- analog for type 2 diabetes, showed significant, clinically meaningful reductions in HbA c and body weight in the SUSTAIN clinical trial program. Higher body mass index (BMI) at baseline (BL) was associated with greater weight loss during semaglutide therapy. Figure: Change in HbA c (%) by baseline BMI (kg/m2) in SUSTAIN 5 and 7 Values shown This post hoc analysis evaluated change in HbA c by BL BMI (<25, 25 <3, 3 <35 and 35 kg/m 2 ) for semaglutide vs. comparators by trial for SUSTAIN 5 and 7. Safety data were pooled and analyzed stratified by trial. Reductions in mean HbA c (%) from BL were greater in all BMI subgroups with semaglutide vs. comparators (Figure). There were no significant interactions between treatment and BMI, indicating a consistent effect of semaglutide vs. comparator on change in HbA c across BMI subgroups. In all treatment arms, adverse events (AEs) occurred in a similar proportion of subjects across BMI subgroups. Gastrointestinal AEs were higher with semaglutide, but decreased with increasing BL BMI, vs. comparators (semaglutide: <25 kg/m²=48.8%, 25 <3 kg/m²=43.%, 3 <35 kg/m²=39.4% and 35 kg/m 2 =39.3% vs. comparators range: %). Premature treatment discontinuation due to AEs was higher in all BMI subgroups with semaglutide vs. comparators ( % vs %). In conclusion, the efficacy of semaglutide in lowering HbA c does not appear to be influenced by BL BMI. Semaglutide had an acceptable safety profile in all BMI subgroups. Values shown are estimated mean changes from baseline for subjects on treatment without rescue medication. Comparators were volume-matched placebo (SUSTAIN and 5), sitagliptin mg (SUSTAIN 2), exenatide ER 2. mg (SUSTAIN 3), IGlar (SUSTAIN 4) and dulaglutide.75 and.5 mg (SUSTAIN 7). BL, baseline; BMI, body mass index; exenatide ER, exenatide extended release; IGlar, insulin glargine; MET, metformin; OAD, oral antidiabetic drug; SU, sulfonylurea; TZD, thiazolidinedione. Poster

22 Semaglutide once weekly Poster presentation -P Efficacy and safety of semaglutide.5 mg vs. dulaglutide.5 mg once weekly in type 2 diabetes: a post-hoc analysis of SUSTAIN 7 RE PRATLEY, VR ARODA 2, T GONDOLF 3, T HANSEN 3, I LINGVAY 4, J LÜDEMAN 5, TV SKJØTH³, A VILJOEN 6 Florida Hospital, Orlando, FL, USA; 2 Brigham and Women s Hospital, Boston, MA, USA; 3 Novo Nordisk A/S, Søborg, Denmark; 4 University of Texas Southwestern Medical Center, Dallas, TX, USA; 5 diabetes-falkensee, Diabetes Centre and Centre for Clinical Studies, Falkensee, Germany; 6 Lister Hospital, Stevenage, UK Semaglutide and dulaglutide are glucagon-like peptide- receptor agonists for the treatment of type 2 diabetes (T2D). In SUSTAIN 7, an international, open-label, parallel-group trial, adults with inadequately controlled T2D were randomized (:::) to once-weekly subcutaneous semaglutide or dulaglutide at low (.5 vs..75 mg) or high (. vs..5 mg) doses. Semaglutide provided superior glycemic control and reductions in body weight at both low and high doses. To study doses available in clinical practice, we performed a post hoc analysis of the effects of semaglutide low (.5 mg) vs. dulaglutide high (.5 mg) dose at week 4, using statistical methods in published prespecified analyses. Subjects mean age was 56 years, HbA c 8.2%, diabetes duration 7.4 years; 77% were white. Efficacy data (Table) show similar glycemic control and blood pressure, and greater weight loss, for semaglutide.5 mg vs. dulaglutide.5 mg. Table. Efficacy data for semaglutide low (.5 mg) vs. dulaglutide high (.5 mg) dose Frequency and severity of adverse events (AEs) were similar for semaglutide and dulaglutide, including gastrointestinal (GI) AEs (nausea: 23% vs. %, respectively; vomiting: %, each treatment). Premature discontinuation due to GI AEs occurred in 5% of subjects in each arm. In summary, semaglutide.5 mg showed similar improvements in glycemic control and greater weight loss vs. dulaglutide.5 mg at week 4, while showing similar tolerability in subjects with T2D. Values are mean (SE) change from overall baseline mean (SD) or ETD (95% CI), unless otherwise indicated, fom a mixed model for repeated measurements analysis using on-treatment without rescue medication data. CI, confidence interval; ETD, estimated treatment difference; NS, not significant; SD, standard deviation, SE, standard error. Poster

23 Semaglutide once weekly Poster presentation 3-P Clinical response (HbA c % and/or body weight 5% reduction) to semaglutide by baseline HbA c and body weight in the SUSTAIN Program VR ARODA¹, JF LARSEN², SH ØSTOFT², RR REA³, D SUGIMOTO 4, U ZIEGLER², T VILSBØLL 5 Brigham and Women s Hospital, Boston, MA, USA; ²Novo Nordisk A/S, Søborg, Denmark; 3 Hospital de Clínicas da Universidade Federal do Paraná, Paraná, Brazil; 4 Cedar Crosse Research Center, Chicago, IL, USA; 5 Steno Diabetes Center Copenhagen, Gentofte, University of Copenhagen, Hellerup, Denmark The SUSTAIN clinical development program assessed semaglutide once weekly, a glucagon like peptide- (GLP-) analog, across the continuum of type 2 diabetes (T2D) care, including in drugnaïve subjects and those on a background of oral antidiabetic drugs and/or insulin. Semaglutide demonstrated superior reductions in HbA c and body weight (BW) vs. placebo and comparators (sitagliptin, exenatide extended release, insulin glargine,dulaglutide). Figure. Scatter plot of baseline HbA c (%) and baseline body weight (kg) in reductions of HbA c % and body weight 5% (composite and individual component endpoints) at week 3 HbAc.% and body weight loss 5.% HbA c.% Body weight loss 5.% Proportion achieving composite endpoint: 42.% Proportion not achieving composite endpoint: 58.% Proportion achieving HbAc.%: 76.6% Proportion not achieving HbAc.%: 23.4% Proportion achieving body weight loss 5.%: 49.9% Proportion not achieving body weight loss 5.%: 5.% This post hoc analysis assessed the proportion of semaglutide-treated subjects achieving clinically relevant responses (composite endpoint: % HbA c reduction and 5% BW loss; individual components: % HbA c or 5% BW loss) by baseline HbA c and BW across SUSTAIN 5 and 7. A consistently high proportion of semaglutide-treated subjects achieved clinically relevant responses (composite, 42.%; % HbA c reduction, 76.6%; 5% BW loss, 49.9%), regardless of baseline HbA c and BW (Figure). Baseline HbA c and BW did not affect the likelihood of achieving the composite and 5% BW endpoints. A % HbA c reduction was significantly more likely to be achieved with higher baseline HbA c (43% increased odds per % unit higher baseline HbA c ; p<.). Semaglutide provides meaningful reductions in HbA c and BW across a range of HbA c and BW. Individuals with poorly controlled HbA c were more likely to achieve a % HbA c reduction. Not achieving composite endpoint Achieving composite endpoint Not achieving HbAc.% Achieving HbAc.% Not achieving weight loss 5.% Achieving weight loss 5.% Baseline HbA c (%, n=3,66) was 8.2 (.9) and baseline body weight (kg, n=3,64) was 93.3 (22.) across the pooled phase 3 trials (SUSTAIN 5 and 7 trials). Values are mean (standard deviation). In the current analysis subjects achieving clinical response were classified based on: HbA c reduction.% and body weight loss 5.% (far left); HbA c reduction % (middle); and body weight loss 5% (far right) at week 3. Semaglutide.5 mg and. mg on-treatment without rescue medication data' are shown with missing values for HbA c or body weight imputed from a mixed model for repeated measurements. Poster

24 Semaglutide once weekly Poster presentation 6-P Real-world effectiveness of semaglutide in early users from a US commercially insured (CI) and Medicare Advantage (MA) population J VISARIA, TD TAN 2, P PETRARO 2, B NEPAL, V WILLEY HealthCore, Inc., Wilmington, USA; 2 Novo Nordisk, Inc., Plainsboro Township, USA With semaglutide s FDA approval in Dec 7, this study sought to provide real-world evidence on its effectiveness in a cohort of early users. Claims and lab result data from a broad national US CI and MA population were used to identify T2DM patients who initiated semaglutide between 2//7 6/3/8 (first claim date set as index). Of these, patients who had 2 month pre-index health plan eligibility as well as HbA c result within both 3 months pre- and 3 months post-index were selected. Changes in HbA c were assessed in all patients, GLP- naïve patients and GLP- naïve patients with a pre-index HbA c >9%. Of 7 individuals with T2DM initiating semaglutide, 48.6% were female with a median age of 52 years. HbA c was significantly reduced in all patients (.3%), GLP- naïve patients ( 2.%) and HbA c >9% GLP- naïve patients ( 2.9%) (all p<., Figure ). Attainment of HbAc <7% increased from pre- to post-index: % (all),.8 49.% (GLP- naïve) and 32.% (pre index HbA c >9% GLP- naïve) (all p<.). Semaglutide initiation was associated with a significant reduction in HbA c and increase in HbA c goal attainment in real-world practice in this preliminary T2DM cohort. Ongoing research will include a broader group of semaglutide users as well as allow for longer post-initiation follow-up to fully assess its effect. Poster

25 Semaglutide once weekly Poster presentation -P Semaglutide-induced weight loss is associated with improved health-related quality of life and treatment satisfaction KT UUSINARKAUS, HW RODBARD², L VAN GAAL³, JP WILDING 4, T HANSEN 5, A SANDBERG 5, S TADAYON 5, MJ DAVIES 6 DaVita Medical Group, Colorado Springs, CO, USA; ²Endocrine and Metabolic Consultants, Rockville, MD, USA; ³Antwerp University Hospital, Antwerp, Belgium; 4 University of Liverpool, Liverpool, UK; 5 Novo Nordisk A/S, Søborg, Denmark; 6 University of Leicester, Leicester, UK Semaglutide, a glucagon-like peptide- analog for the once-weekly treatment of type 2 diabetes, provided superior glycemic control and weight loss (WL) vs. comparators in the SUSTAIN clinical trial program. This post hoc analysis assessed if WL was associated with patient-reported health-related quality of life (HRQoL) and treatment satisfaction improvements in SUSTAIN 2 5 and 7. Data for both semaglutide doses (.5 and. mg) were pooled across trials (N=2,88). Change in HRQoL (Short Form-36 Health Survey version 2 [SF-36v2 ] Physical Component Summary [PCS] and Mental Component Summary) and treatment satisfaction (Diabetes Treatment Satisfaction Questionnaire status version) scores were evaluated in subjects who achieved 5% and % WL vs. those who did not at end of treatment (3, 4 or 56 weeks). Overall, 5.% and 7.4% of subjects achieved 5% and % WL with semaglutide. Significantly greater improvements in the overall PCS score and most of its components were reported in subjects achieving 5% and % WL vs. those not achieving these responses (Table). WL responses also correlated with overall treatment satisfaction and perception of hyperglycemia (Table). In conclusion, semaglutide-induced WL was associated with improvements in PCS domains of the SF-36v2, overall treatment satisfaction and perception of hyperglycemia across the SUSTAIN 2 5 and 7 trials. Poster

26 Semaglutide oral Oral presentation 54-OR Late-breaking poster presentation -LB Oral semaglutide vs. empagliflozin added on to metformin monotherapy in uncontrolled type 2 diabetes: PIONEER 2 Oral semaglutide vs. sitagliptin: efficacy by baseline HbA c and background OAD in PIONEER 3 Oral presentation 55-OR Poster presentation 4-P Oral semaglutide vs. liraglutide and placebo in T2D: PIONEER 4 Oral semaglutide vs. placebo in patients with type 2 diabetes and moderate renal impairment: PIONEER 5 eposter Theater presentation 983-P Poster presentation 3-P Effect and safety of flexible dose adjustment with oral semaglutide vs. sitagliptin in type 2 diabetes: PIONEER 7 Effect of upper gastrointestinal disease on the pharmacokinetics of oral semaglutide in subjects with type 2 diabetes eposter Theater presentation 985-P Oral semaglutide as add-on to insulin in T2D: PIONEER 8

27 Semaglutide oral Oral presentation 54-OR Oral semaglutide vs. empagliflozin added on to metformin monotherapy in uncontrolled type 2 diabetes: PIONEER 2 E MONTANYA¹, J ROSENSTOCK², LH CANANI³, C DEEROCHANAWONG 4, J GUMPRECHT 5, S LINDBERG 6, I LINGVAY 7, AL SØNDERGAARD 6, M TREPPENDAHL 6, HW RODBARD 8, FOR THE PIONEER 2 INVESTIGATORS Hospital Universitari Bellvitge, CIBERDEM, University of Barcelona, Barcelona, Spain; ²Dallas Diabetes Research Center at Medical City, Dallas, TX, USA; ³Endocrine Division, Hospital de Clínicas de Porto Alegre, Porto Alegre, Brazil; 4 Rajavithi Hospital, Rangsit Medical School, Bangkok, Thailand; 5 Department of Internal Medicine, Diabetology and Nephrology in Zabrze, Medical University of Silesia, Katowice, Poland; 6 Novo Nordisk A/S, Søborg, Denmark; 7 Departments of Internal Medicine and Clinical Sciences, University of Texas Southwestern Medical Center, Dallas, TX, USA; 8 Endocrine and Metabolic Consultants, Rockville, MD, USA The GLP- analog oral semaglutide (sema) 4 mg once daily (QD, N=4) was compared with the SGLT-2i empagliflozin (empa) 25 mg QD (N=4) in patients (pts) with T2D uncontrolled on metformin in a 52-week randomized, open-label trial (NCT ). Endpoints included change from baseline to Week 26 in HbA c (primary) and body weight (confirmatory secondary). Two scientific questions were addressed through defining two estimands: treatment policy (regardless of trial product discontinuation or rescue medication) and trial product (on trial product without rescue medication) in all randomized pts. Oral sema was superior to empa at reducing HbA c at Week 26; the difference was also significant at Week 52 (Table). More patients achieved HbA c <7.% with oral sema. Weight loss with oral sema was not superior at Week 26, but significantly greater at Week 52 (trial product estimand). Proportions of adverse events (AEs) were similar between oral sema (7.5%) and empa (69.2%). The most frequent AEs with oral sema were mild/moderate gastrointestinal events. Premature trial product discontinuations due to AEs were % (oral sema) vs. 4% (empa). In conclusion, oral sema provided superior reductions in HbA c but not body weight at Week 26, and significant reductions in HbA c and body weight (trial product estimand) at Week 52 vs. empa. Oral sema was well tolerated within the established safety profile of GLP-RAs.

28 Semaglutide oral Oral presentation 55-OR Oral semaglutide vs. liraglutide and placebo in T2D: PIONEER 4 R PRATLEY, A AMOD 2, ST HOFF 3, T KADOWAKI 4, I LINGVAY 5, MA NAUCK 6, KB PEDERSEN 3, T SAUGSTRUP 3, JJ MEIER 6, FOR THE PIONEER 4 INVESTIGATORS Florida Hospital Translational Research Institute for Metabolism and Diabetes, Orlando, FL, USA; 2 Life Chatsmed Garden Hospital and Nelson R Mandela School of Medicine, Durban, South Africa; ³Novo Nordisk A/S, Søborg, Denmark; 4 Department of Diabetes and Metabolic Diseases, Graduate School of Medicine, The University of Tokyo, Tokyo, Japan; 5 University of Texas Southwestern Medical Center at Dallas, Dallas, TX; USA 6 Diabetes Center Bochum-Hattingen, St. Josef-Hospital, Ruhr-University Bochum, Bochum, Germany Patients (pts) with T2D uncontrolled on metformin ± SGLT-2i were randomized to oral semaglutide (sema) 4 mg once daily (N=285), liraglutide (lira).8 mg (N=284) or placebo (pbo, N=42) in a phase 3a, 52-week (wk), double-blind, double-dummy trial (NCT286349). Endpoints: Change (baseline to Wk 26) in HbA c (primary) and body weight (BW, confirmatory secondary). Two estimands addressed two efficacy-related questions: Treatment policy (regardless of trial product discontinuation or rescue medication) and trial product (on trial product without rescue medication) in all randomized pts. Treatment policy estimand: Oral sema was non-inferior(margin.4%) to lira and superior to pbo in reducing HbA c, and superior to both in reducing BW at Wk 26 (Table). Differences in both HbA c and BW were significant at Wk 52. Trial product estimand: Oral sema gave significant reductions in HbA c and BW vs. lira and pbo at Wks 26 and 52. Oral sema had comparable tolerability to lira; % (oral sema) vs. 9% (lira) and 4% (pbo) prematurely discontinued trial product due to adverse events (primarily gastrointestinal; 5% [oral sema] vs. 3% [lira] discontinued due to nausea). In conclusion, oral sema was well tolerated in pts with T2D on metformin ± SGLT-2i, was non-inferior vs. lira and superior vs. pbo in reducing HbA c, and was superior in reducing BW vs. both lira and pbo. The reduction in HbA c was significantly better vs. lira when evaluated by the trial product estimand.

29 Semaglutide oral eposter Theater presentation 983-P Effect and safety of flexible dose adjustment with oral semaglutide vs. sitagliptin in type 2 diabetes: PIONEER 7 JB BUSE¹, BW BODE², YM CHO³, E CHRISTIANSEN 4, CL HERTZ 4, A MERTENS 5, SO WALLENSTEIN 4, TR PIEBER 6, FOR THE PIONEER 7 INVESTIGATORS Department of Medicine, University of North Carolina School of Medicine, Chapel Hill, NC, USA; ²Atlanta Diabetes Associates, Emory University School of Medicine, Atlanta, GA, USA; ³Department of Internal Medicine, Seoul National University College of Medicine, Seoul, Korea; 4 Novo Nordisk A/S, Søborg, Denmark; 5 Department of Chronic Diseases, Metabolism and Aging (CHROMETA), KU Leuven, Leuven, Belgium; 6 Division of Endocrinology and Diabetology, Medical University of Graz, Graz, Austria Oral semaglutide (sema) is the first oral GLP- receptor agonist (RA). Flexible dose adjustment of oral sema (3, 7, and 4 mg) once daily (N=253) was tested vs. sitagliptin (sita, mg; N=25) in a 52-week (wk), randomized, open-label phase 3a trial in patients (pts) with T2D uncontrolled on 2 oral glucose-lowering drugs (NCT28498). Oral sema dose could be adjusted every 8 wks based on HbA c (escalated if 7.%) and tolerability. Endpoints at Wk 52: HbA c <7.% (primary); body weight change (secondary). Two scientific questions were addressed by two estimands: treatment policy (regardless of trial product discontinuation or rescue medication) and trial product (on trial product without rescue medication) in all randomized pts. At Wk 52, 9., 3.2 and 59.4% of oral sema pts were on 3, 7 and 4 mg, respectively. Oral sema was superior to sita for HbA c <7.% and weight loss, and gave a significantly better reduction in HbA c at Wk 52 (Table). Fewer pts on oral sema needed rescue medication (3 vs. 6%). Adverse events (AEs) were more common with oral sema (78 vs. 69%), mostly mild/moderate and transient nausea (2 vs. 2%). Trial product discontinuation due to AEs (mostly gastrointestinal) was higher with oral sema (9 vs. 3%). In conclusion, flexible dose adjustment with oral sema provided superior glycemic control and weight loss at Wk 52 vs. sita, and was well tolerated with safety consistent with the GLP-RA class. Poster

30 Semaglutide oral eposter Theater presentation 985-P Oral semaglutide as add-on to insulin in T2D: PIONEER 8 B ZINMAN¹, VR ARODA², JB BUSE³, B CARIOU 4, SB HARRIS 5, ST HOFF 6, KB PEDERSEN 6, MJ TARP-JOHANSEN 6, E ARAKI 7, FOR THE PIONEER 8 INVESTIGATORS Lunenfeld Tanenbaum Research Institute, Mt. Sinai Hospital, University of Toronto, Toronto, ON, Canada; ²Division of Endocrinology, Diabetes & Hypertension, Brigham and Women s Hospital, Harvard Medical School, Boston, MA, USA; MedStar Health Research Institute, Hyattsville, MD, USA; ³Division of Endocrinology & Metabolism, University of North Carolina School of Medicine, Chapel Hill, NC, USA; 4 L institut du Thorax, INSERM CIC 43, CHU Nantes, Nantes, France; 5 Department of Family Medicine, Schulich School of Medicine and Dentistry, Western University, London, ON, Canada; 6Novo Nordisk A/S, Søborg, Denmark; 7 Department of Metabolic Medicine, Faculty of Life Sciences, Kumamoto University, Kumamoto, Japan Patients (pts) with T2D uncontrolled on insulin ± metformin were randomized to oral semaglutide (sema) 3 mg (N=84), 7 mg (N=82) or 4 mg (N=8), or placebo (pbo; N=84) in a 52- week (wk) double-blind trial (NCT3287). Total daily insulin dose was reduced by an optional % and capped at baseline (BL) level for Wk 26 and freely adjustable for Wk Endpoints were change from BL to Wk 26 in HbA c (primary) and body weight (BW; confirmatory secondary). Two scientific questions were addressed by defining two estimands: a treatment policy estimand (regardless of trial product discontinuation or rescue medication use), and a trial product estimand (on trial product without rescue medication use) in all randomized pts. Oral sema was superior to pbo in reducing HbA c and BW at Wk 26 (treatment policy estimand). Significantly greater, dose-dependent HbA c and BW reductions vs. pbo were achieved at Wk 26 and 52 (both estimands; Table). Total daily insulin dose was significantly reduced from BL with oral sema vs. pbo at Wk 26 (except 3 mg) and 52. The rate of hypoglycemia was not significantly different vs. pbo (Table). Most frequent AE with oral sema was nausea ( % of pts vs. 7.% with pbo) and % of pts prematurely discontinued due to AEs (vs. 2.7% with pbo). As add-on to insulin ± metformin, oral sema had superior HbA c and BW reductions vs. pbo at Wk 26. It also reduced insulin use by Wk 52 and was well tolerated without significantly increasing hypoglycemia rate. Poster

31 Semaglutide oral Late-breaking poster presentation -LB Oral semaglutide vs. sitagliptin: efficacy by baseline HbA c and background OAD in PIONEER 3 J ROSENSTOCK¹, D ALLISON², AL BIRKENFELD 3,4, TM BLICHER 5, S DEENADAYALAN 5, A KOUSHOLT 5, MJ DAVIES 6 Dallas Diabetes Research Center at Medical City, Dallas, TX, USA; ²Hillcrest Family Health Center, Waco, TX, USA; ³Department and Outpatient Department of Medicine III, Carl Gustav Carus University Hospital Dresden, Dresden, Germany; 4 Paul Langerhans Institute Dresden, Helmholtz Center Munich, University Hospital, Technical University Dresden, German Center for Diabetes Research (DZD e.v.), Dresden, Germany; 5 Novo Nordisk A/S, Søborg, Denmark; 6 Diabetes Research Centre, University of Leicester, Leicester, UK Exploratory analyses of PIONEER 3 (NCT267865) assessed the efficacy (HbA c change from baseline [BL], achievement of HbA c <7.%) of oral semaglutide (sema;3, 7, or 4 mg), a novel oral GLP- receptor agonist, vs. sitagliptin (sita) mg at Week 26 by BL HbA c and background OAD in pts with T2D uncontrolled onmetformin ± sulfonylurea. HbA c was reduced across all BL HbA c and OAD groups in all treatment arms; reductions were greater with higher BL HbA c. HbA c reductions were significantly greater with oral sema 7 and 4 mg vs. sita in all groups, except for 7 mg in the HbA c 8.% group (Figure ). Achievement of HbA c <7.% was greater with oral sema 7 and 4 mg vs. sita in all groups (Figure 2). In conclusion, oral sema 7 and 4 mg significantly improved glycemic control vs. sita across most HbA c groups, and irrespective of background OAD use. Poster

32 Semaglutide oral Poster presentation 4-P Oral semaglutide vs. placebo in patients with type 2 diabetes and moderate renal impairment: PIONEER 5 O MOSENZON¹, S ROSENLUND², JW ERIKSSON³, SR HELLER 4, RE PRATLEY 5, T SATHYAPALAN 6, TM BLICHER², OH HELS 2, C DESOUZA 7 Diabetes Unit, Department of Internal Medicine, Hadassah Hebrew University Hospital, Jerusalem, Israel; ²Novo Nordisk A/S, Søborg, Denmark; ³Clinical Diabetology and Metabolism, Department of Medical Sciences, Uppsala University, Uppsala, Sweden; 4 Academic Unit of Diabetes, Endocrinology and Metabolism, Department of Oncology & Metabolism, University of Sheffield, Sheffield, UK; 5 Florida Hospital Translational Research Institute for Metabolism and Diabetes, Orlando, FL, USA; 6 Academic Diabetes, Endocrinology and Metabolism Research Group, Hull York Medical School, University of Hull, Hull, UK; 7 Division of Diabetes, Endocrinology, and Metabolism, Department of Internal Medicine, University of Nebraska Medical Center, Omaha, NE, USA Patients (pts) with T2D and renal impairment have limited glucose-lowering treatment options. Oral semaglutide (sema) 4 mg once daily (N=63) was compared with placebo (pbo, N=6) in a 26 week, randomized, phase 3a trial (NCT282778) in pts with T2D and moderate renal impairment (egfr 3 59 ml/min/.73 m²), receiving 2 glucose-lowering agents (could include basal insulin). Endpoints: change in HbA c (primary) and body weight (BW, secondary) to Week 26. Two estimands were defined: treatment policy (regardless of trial product discontinuation or rescue medication) and trial product (on trial product without rescue medication) in all randomized pts. Oral sema was superior to (treatment policy) and significantly better than (trial product) pbo in reducing HbA c and BW (Table). More pts reached HbA c, BW loss, and composite targets with oral sema. At follow-up, egfr was unchanged. Fewer pts on oral sema required rescue medication (4 vs. %). Adverse events were more frequent with oral sema (74 vs. 65%), most often mild/moderate nausea (9 vs. 7%). More patients on oral sema prematurelydiscontinued trial product (5 vs. 5%), mainly due to nausea and vomiting. In conclusion, in pts with T2D and moderate renal impairment, oral sema provided superior glycemic control and BW loss compared to pbo at 26 weeks, did not appear to affect renal function, and was well tolerated in line with the population and GLP-RA class. Poster

33 Semaglutide oral Poster presentation 3-P Effect of upper gastrointestinal disease on the pharmacokinetics of oral semaglutide in subjects with type 2 diabetes JJ MEIER, C GRANHALL 2, U HOEVELMANN 3, A NAVARRIA 2, L PLUM-MOERSCHEL 5, C RAMESH 4, A TANNAPFEL, C KAPITZA 3 St Josef Hospital, Ruhr-University Bochum, Bochum, Germany; ²Novo Nordisk A/S, Søborg, Denmark; ³Profil, Neuss, Germany; 4 Novo Nordisk, Bangalore, India; 5 Profil, Mainz, Germany The GLP- analog semaglutide has been co-formulated with the absorption enhancer sodium N (8 [2 hydroxybenzoyl] amino) caprylate (SNAC) to allow oral administration. Since oral semaglutide is absorbed in the stomach, the effect of upper gastrointestinal disease (UGD) on exposure to oral semaglutide has been investigated in an open-label, parallel-group trial (NCT ). Subjects aged 8 8 years with T2D and with UGD (N=36; chronic gastritis [n=5], gastroesophageal reflux disease [GERD; n=8] or both [n=23]) or without UGD (N=9) received oral semaglutide 3 mg oncedaily for 5 days followed by 7 mg for 5 days. Key endpoints were area under the semaglutide plasma concentration-time curve (AUC) (primary) and maximum semaglutide plasma concentration (Cmax) from 24 h after the th dose. Semaglutide exposure was not statistically significantly different in subjects with vs. without UGD (Figure; estimated group ratios with/without UGD [95% CI] AUC:.8 [.8;.75] and Cmax:.6 [.77;.76]). Time to maximum plasma concentration and half-life of semaglutide were similar in subjects with vs. without UGD. Oral semaglutide was well tolerated. All adverse events (AEs) were mild or moderate with no withdrawals due to AEs. In conclusion, no significant difference in exposure to oral semaglutide in subjects with vs. without UGD was found; hence, no dose adjustment is expected to be required. Poster

34 Devices Late-breaking poster presentation 26-LB Poster presentation 76-P Improved insulin adherence after introduction of a smart connected insulin pen Increased time in range (TIR) observed after introduction of a connected insulin pen

35 Devices Late-breaking poster presentation 26-LB Improved insulin adherence after introduction of a smart connected insulin pen P ADOLFSSON, NV HARTVIG, 2 A KAAS, 3 N NYGÅRD KNUDSEN, 4 A-C MÅRDBY, 5 JB MØLLER, 6 J HELLMAN 7 The Hospital of Halland Kungsbacka, Institution of Clinical Sciences University of Gothenburg, Kungsbacka, Sweden; 2 Global Development, Data Science, Novo Nordisk A/S, Søborg, Denmark; 3 Medical & Science Digital Health, Novo Nordisk A/S, Søborg, Denmark; 4 Epidemiology/Digital Health, Novo Nordisk A/S, Søborg, Denmark; 5 Medical Affairs, Novo Nordisk Sweden, Malmö, Sweden; 6 Digital Health, Partnerships & Commercial Strategy, Novo Nordisk A/S, Søborg, Denmark; 7 Department of Medical Sciences, Uppsala University, Uppsala, Sweden An association between missed insulin injections and its impact on HbA c levels has been established. The smart connected NovoPen 6 captures and allows visualization of insulin injections (date and time of injection and number of units) and thus has the potential to move dialog between patients and healthcare professionals (HCPs) away from guessing about doses taken, missed doses and optimal injection time in relation to meals. This non-interventional study investigated whether the use of NovoPen 6 influenced the behavior of patients with type diabetes (TD) in terms of missed bolus dose (MBD) injections. Patients were recruited from 2 Swedish diabetes clinics. At baseline they received a NovoPen 6 for bolus insulin injections. At each HCP visit, pen data were downloaded at the clinic. Follow-up was after 5 HCP visits. Adults with TD (n=8) using continuous glucose monitoring (CGM) and NovoPen 6 for bolus injections were included in the analyses. The frequency of MBD was analyzed using the GRID algorithm to detect meals from the CGM signal. MBD was defined as meals with no bolus injection within 5 to +6 minutes from the start of the meal, as detected by the algorithm. The change in number of MBD from baseline to 5 HCP visits was analyzed using a mixed model. A significant decrease of 43% in the average daily number of MBD was observed from baseline to after 5 HCP visits (median time in study 6 months) from.74 (95% CI [.62;.88]) to.42 (95% CI [.3;.6]) (p=.2). Based on the assumption that patients have three main meals per day, this corresponded to a decrease from 25% to 4% in MBD. The results indicated that patients achieved more well-dosed meals, even though these are not detected by the GRID algorithm, because of a lower CGM response. These real-world findings confirm that MBD is a problem for patients with TD and that a smart connected pen can support good injection behavior, with fewer missed and more well-dosed mealtime injections. This could subsequently lead to better glycemic control. Poster

36 Devices Poster presentation 76-P Increased time in range (TIR) observed after introduction of a connected insulin pen P ADOLFSSON¹, NV HARTVIG², A KAAS³, N NYGÅRD KNUDSEN 4, A-C MÅRDBY 5, J HELLMAN 6 The Hospital of Halland Kungsbacka, Institution of Clinical Sciences University of Gothenburg, Kungsbacka, Sweden; 2 Global Development, Data Science, Novo Nordisk A/S, Søborg, Denmark; 3 Medical & Science Digital Health, Novo Nordisk A/S, Søborg, Denmark; 4 Epidemiology/Digital Health, Novo Nordisk A/S, Søborg, Denmark; 5 Medical Affairs, Novo Nordisk Sweden, Malmö, Sweden; 6 Department of Medical Sciences, Uppsala University, Uppsala, Sweden The capture and visualization of insulin injection patterns together with blood glucose data has potential to improve dialogue between patients and HCPs about dose, injection time and missed injections. This may improve glycemic control. The objective of this non-interventional study was to investigate how a connected insulin pen (NovoPen 6) influences glycemic control in patients with TDM in a real-world setting. Patients were recruited from 2 Swedish diabetes clinics. At baseline patients received a NovoPen 6, which logs time of injection and number of insulin units. HCP visits were conducted according to clinical practice. At each visit, pen data were downloaded at the clinic to support patient-hcp dialogue. Adults with TDM (n=94) using NovoPen 6 for basal and/or bolus and CGM were included in the analyses. The effect on TIR was based on a 4-day interval after each visit using a linear mixedeffects model. A significant increase in TIR from baseline to after 5 HCP visits (median time in study 6 months) of.9 hrs/day (9.2 to. hrs/day) was found (p=.9). Accordingly significant reduction in time spent in hyper- and L2 hypoglycemia (.8 hrs/day p=.3; -.3 hrs/day p=.5) was observed. There was no change in time in L hypoglycemia (p =.8). These real-world findings in patients with TDM highlight the potential benefit on glycemic control when accurate connected pen data contribute to patient-hcp dialogue. Poster

37 Diabetes research Late-breaking poster presentation 89-LB Poster presentation 329-P Low level of agreement in evaluation of continuous glucose monitoring profiles National differences and similarities in youth-onset type 2 diabetes

38 Diabetes research Late-breaking poster presentation 89-LB Low level of agreement in evaluation of continuous glucose monitoring profiles E HACHMANN-NIELSEN¹, K KVIST¹, RM BERGENSTAL² Novo Nordisk A/S, Søborg, Denmark; 2 International Diabetes Center, Minneapolis, MN, USA Nearly years after the introduction of Continuous Glucose Monitoring (CGM) it is clear that CGM reports can effectively guide clinical decision making but the large amount of information contained in the CGM profile may also complicate the evaluation of the profiles. Figure Distribution of test-rank of 9 Ambulatory Glucose Profiles (AGPs) (A-I) between non-experts and between diabetes experts. We hypothesized that there would be a relatively low level of agreement in evaluations of CGMprofiles performed by a diverse group of HCPs and patients, with higher level of agreement between diabetes experts. We investigated the interobserver variation in the two groups when asked to subjectively prioritize 9 Ambulatory Glucose Profiles from best to worst through an internet-based portal ( A total of 88 responses were collected. The groups were defined as non-experts (N=73) and self-evaluated experts (> years of experience and > CGM evaluation per week) (N=5). The interobserver variation between the groups were calculated using Interclass Correlation Coefficients (ICC) showing moderate agreement among the non-experts group (ICC=.6) and a good degree of agreement among the diabetes experts (ICC=.75) [Figure]. Despite standardized CGM metrics and reports there is still a relatively low level of agreement in interpretation and ranking of CGM-profiles even among diabetes experts. This may allow for undesirable variation in diabetes management. More tools are needed to optimize the use of CGM for clinical decision making. Poster

39 Diabetes research Poster presentation 329-P National differences and similarities in youth-onset type 2 diabetes J LYNCH, M BARRIENTOS-PÉREZ 2, M HAFEZ³, MY JALALUDIN 4, M KOVARENKO 5, R PATURI 6, D WEGHUBER 7 ON BEHALF OF THE NOVO NORDISK PAEDIATRIC TYPE 2 DIABETES GLOBAL EXPERT PANEL University of Texas Health Science Center San Antonio, San Antonio, TX, USA; ²Angeles Hospital of Puebla, Puebla City, México; ³Department of Paediatrics, Cairo University, Cairo, Egypt; 4 Department of Paediatrics, Faculty of Medicine, University of Malaya, Kuala Lumpur, Malaysia; 5 Novosibirsk Medical University, Novosibirsk, Russia; 6 Diabetes Research Society, Hyderabad, India; 7 Department of Paediatrics, Paracelsus Medical School, Salzburg, Austria With the recent increased awareness of youth-onset type 2 diabetes (T2D) cases, an overview of national differences in epidemiology data is needed to obtain a global picture of the disease development. This review examined national epidemiology data of youth-onset T2D published between st January 8 and 5 th January 8, and searched for national guidelines on this topic to understand better national similarities and differences. Of the 5 articles and 4 congress abstracts identified, 42 studies were included. The highest reported prevalence rates of youth-onset T2D were found in the USA and China (38 and 8 cases/, people, respectively), and lowest in Denmark and Ireland (.6 and.2 cases/, people, respectively) (Figure). However, the highest incidence rates were reported in Taiwan and the UK (63 and 33 cases/, people, respectively), with the lowest in Fiji and Austria (.4 and.6 cases/, people, respectively) (Figure). These differences may be partly explained by variations in diagnostic criteria, national screening recommendations, and ethnicity within countries. Our review suggests that published national epidemiology data for youth-onset T2D may underestimate its disease burden. Finding optimal diagnostic criteria and screening strategies for this disease should be of high interest to every country. Poster

40 Insulin degludec Late-breaking poster presentation 8-LB 8-LB Outcomes of Type 2 Diabetes (T2D) Clustering Replicated in the DEVOTE and LEADER Trials Anna R. Kahkoska ; Elise Hachmann-Nielsen 2 ; Klara R. Klein ; Kristine G Kongsbak 2 ; Kajsa Kvist 2 ; John B. Buse University of North Carolina, Chapel Hill, North Carolina U.S.A; 2 Novo Nordisk A/S Søborg, Denmark qrs.ly/329xx28 Aim Type 2 diabetes (T2D) is a heterogeneous disease. Multiple studies have therefore attempted to characterize more precise T2D patient populations.,2 Individuals in the Swedish All New Diabetics in Scania (ANDIS) cohort with newly diagnosed diabetes were grouped by 6 demographic and clinical variables (autoantibodies, sex, age at diagnosis, body mass index (BMI), hemoglobin A c (HbA c ), and HOMA estimates of beta cell function and insulin resistance) to show 4 distinct T2D subtypes with differential risk for nephropathy and retinopathy. 2 (Cluster was essentially identical to type diabetes). Cluster 2 (SIDD- severe insulin-deficit diabetes) Cluster 3 (SIRD- severe insulin-resistant diabetes) Cluster 4 (MOD- mild obesity-related diabetes) Cluster 5 (MARD- mild age-related diabetes). The objective of this study was to test the predictive validity of the same clustering system for advanced T2D in the DEVOTE trial and the LEADER trial for predicting time to first episode of severe hypoglycemia (SH), time to first major adverse cardiovascular event (MACE), time to cardiovascular (CV) death, and time to all-cause mortality. Methods Data came from the DEVOTE and LEADER trials. DEVOTE was a large, geographically diverse cardiovascular outcomes trial in advanced T2D patients at high risk of CV events. 3 For confirmation, all analyses were replicated in the LEADER trial, likewise a cardiovascular outcomes trial in advanced T2D patients. 4 Individuals enrolled in the DEVOTE and LEADER trials were included in the analysis if they had data on three key variables for clustering (BMI, HbA c, and age at diagnosis, calculated as baseline age in years minus years since diagnosis). HOMA and auto-antibodies were not measured in the DEVOTE and LEADER trials. Participants were assigned to one of the four clusters described in the ANDIS cohort based on baseline HbA c, BMI, age, and calculated age at diagnosis. Insulin resistance and sensitivity measures were not available. For each participant, the Euclidean distance to exact cluster centers were calculated and participants were assigned to the cluster for which the distance was the smallest. Time-to-event analysis was used to compare differences in outcomes between the four clusters. Specifically, Kaplan-Meyer curves were depicted and p-values for the log-rank test were determined. HbA c (%) HbAc (%) Results The analysis included: 7,673 participants in the DEVOTE trial with a mean age of 65. years, mean T2D duration of 6.4 years, and mean HbA c of 8.43%. 9,34 participants in the LEADER trial with mean age of 64 years, mean T2D duration of 2.8 years, and mean HbA c of 8.7%. The four T2D clusters from the ANDIS cohort were represented in the DEVOTE trial as: SIDD-like: 6.7% (N =,26), SIRD-like: 24.5% (N =,847), MOD-like: 26.8% (N = 2,632), MARD-like: 34.9% (N = 2,632). Similar distribution was seen in the LEADER trial. The four replicated clusters showed differences in baseline characteristics consistent with the original ANDIS clusters despite clustering on a subset of the variables [Figure ]. For example, patients in the SIDD-like cluster had higher HbA c, and lower BMI. No other cluster had this profile. Figure 2 Mean HbA c and BMI over time in DEVOTE (A & C) and LEADER (B & D) according to T2D cluster Differences in HbA c and BMI were retained over time [Figure 2]. Figure 3 Time to first event for the four T2D clusters: Upper panel DEVOTE data, lower panel LEADER data. A: time to first SH event, B: time to first MACE, C: time to CV-death, D: time to all-cause mortality Figure Comparison of baseline characteristics of T2D Clusters in the DEVOTE trial (top row), the LEADER trial (bottom row). BMI (m/kg 2 ) SIDD-like MOD-like SIRD-like MARD-like BMI (m/kg 2 ) SIDD-like MOD-like SIRD-like MARD-like Age at diagnosis (yr) SIDD-like MOD-like SIRD-like MARD-like Age at diagnosis (yr) SIDD-like MOD-like SIRD-like MARD-like SIDD-like MOD-like SIRD-like MARD-like SIDD-like MOD-like SIRD-like MARD-like HbA c (%) A B C D 6 2 Patients with event (%) % 5% 5% Years since randomization A B C D % % BMI (m/kg 2 ) Key Result p=.6 p<. p=.623 p=.7632 % % % % % % % p=.85 p=.4 p<. p=.7 % % % % % 5% % % Years since randomization % 5% % % SIDD-like MOD-like SIDD-like SIRD-like MOD-like MARD-like SIRD-like MARD-like There was significant difference (p<.5) across the T2D clusters in time to first SH and time to first MACE in the DEVOTE trial, and time to first MACE, time to first CV death and time to all-cause mortality in the LEADER trial [Figure 3]. In the DEVOTE trial, the proportion of MACE incidence and CV death was highest among Cluster 2 SIDD (2.% and 4.8% respectively) and lowest among the Cluster 4 MARD (7.7% and 3.2%, respectively). The LEADER trial showed similar trends. The results were not modified by insulin use at baseline. Discussion The data replicate patterns of previous T2D clusters derived in a predominantly northern European cohort early in T2D, 2 and further replicated in Chinese and US populations. 5 T2D clusters in DEVOTE and LEADER represent a geographically-diverse population with higher HbA c, greater CVD risk, and longer disease duration (mean diabetes duration: 6.4 years). Overall, the SIDD-like cluster appeared to have to highest risk of MACE, CV and all-cause mortality and may benefit from early intervention. The SIDD-cluster has previously shown significantly higher rates of retinopathy compared the other clusters in the ANDIS cohort. 2 To reduce disparity in survival outcomes, more work is needed to understand if and how optimal treatment regimens may differ according to subgroups. T2D clusters that are optimized by incorporation of treatment response variables over longitudinal follow-up could inform future treatment algorithms for clinical practice. Conclusion Clusters derived from early T2D can be replicated in long-standing T2D and may confer information about time to first MACE, CV event, and all-cause mortality. Patients with T2D, a high HbA c and low BMI (i.e. the SIDD-like phenotype) show significantly shorter time to first MACE and CVdeath compared to other subgroups of T2D. The consistency of clusters across different populations adds to the evidence that T2D is not a homogeneous disease, but instead may have different prognoses and require different treatment according to specific disease subtypes. This trial was sponsored by Novo Nordisk. Presented at the American Diabetes Association, 79th Scientific Sessions. June 7, 9, San Francisco, CA, USA. References: () Li L, Cheng W-Y et al. Sci Transl Med 5; 7 (3): 3ra74-3ra74. (2) Ahlqvist E et al. Lancet Diabetes Endocrinol 8; 6 (5): (3) Marso SP et al. Am Heart J 6; 79: (4) Marso SP et al. New Eng J Med. 6; 375 (4): (5) Zou Xet al. Lancet Diabetes Endocrinol 9; 7 (): 9-. Abstract

41 Insulin degludec Late-breaking poster presentation 5-LB 5-LB Diabetes duration did not moderate the reduction in hypoglycemia seen with degludec vs glargine U in DEVOTE and SWITCH 2 Anastasia-Stefania Alexopoulos; Kajsa Kvist 2, Elise Hachmann-Nielsen 2 ; John B. Buse 3 Duke University, Durham, North Carolina U.S.A; 2 Novo Nordisk A/S Søborg, Denmark, ; 3 University of North Carolina, Chapel Hill, North Carolina U.S.A qrs.ly/329xx28 Aim In patients with type 2 diabetes (T2D), hypoglycemia is associated with significant morbidity. Progression of T2D leads to declining counter-regulatory responses and heightened risk of hypoglycemia. 2 DEVOTE was a cardiovascular outcomes trial in 7,637 T2D patients at high risk of cardiovascular events. SWITCH 2 investigated the risk of hypoglycemia in 72 T2D patients at high risk of hypoglycemia, both trials comparing insulin degludec with glargine U. 3,4 In this secondary analysis of DEVOTE and SWITCH 2, we investigated whether diabetes duration moderated the hypoglycemia benefits with insulin degludec compared to insulin glargine U. Clinical question: Since progression of T2D leads to higher risk of hypoglycemia, is the hypoglycemic benefit of degludec over glargine U more prominent in patients with longer versus shorter duration of disease? Methods DEVOTE was a double-blind, treat-to-target, event-driven cardiovascular outcomes trial, with 7,637 patients randomized to either insulin degludec or glargine U. [3] SWITCH 2 was a 64 week, double-blind, treat-to-target cross-over trial consisting of two 32-week treatment periods; patients were randomized : to receive once-daily insulin degludec followed by insulin glargine U (n=36), or to receive insulin glargine U followed by insulin degludec (n=36). [4] Patients in DEVOTE and SWITCH 2 were categorized by patient s duration of diabetes at baseline: <, 5, 5 and years. [Table ] We examined severe hypoglycemia in the DEVOTE trial, and overall symptomatic hypoglycemia in the SWITCH 2 trial. For the analysis of severe hypoglycemia in DEVOTE, severe hypoglycemia was defined (as per the American Diabetes Association), as a hypoglycemic event requiring the assistance of another person for corrective action. It was an externally adjudicated outcome. a [5] For the analysis of overall symptomatic hypoglycemia in SWITCH 2, overall symptomatic hypoglycemia was a composite measure including symptomatic hypoglycemia (any documented glucose 56 mg/dl) and severe hypoglycemia. In SWITCH 2, severe hypoglycemia by diabetes duration was not investigated due to the small number of events. In DEVOTE, events of non-severe hypoglycemia were not systematically collected. Rate of severe hypoglycemia in DEVOTE was investigated with the pre-specified model from the trial stratified by diabetes duration, i.e. a negative binomial model with log-observation time as offset. Rates of overall symptomatic hypoglycemia by diabetes duration in SWITCH 2 was likewise analyzed with the pre-specified confirmatory model from the trial protocol, i.e. a Poisson model with subjects as random effect and log-observation time as off-set. Table Baseline characteristics in DEVOTE (top panel) by diabetes duration category and treatment arm, and in SWITCH 2 (lower panel) by diabetes duration Diabetes Duration (DEVOTE) Age (years, mean) IDeg n=95 < years -5 years 5- years years iglar n=985 IDeg n=92 iglar n=938 IDeg n= 85 iglar n= 769 IDeg n=,95 iglar n=, Gender (% F) 34.6% 36.% 37.4% 36.8% 36.4% 37.3% 39.7% 39.8% BMI (kg/m 2, mean) HbA c (% mean) Established CV disease (%) Insulin naïve (%) 84.% 83.2% 83.8% 84.2% 87.6% 86.% 86.6% 86.4% 26.3% 24.9% 7.5% 8.% 3.3% 2.5% 8.2%.% Basal only (%) 4.9% 42.% 42.3% 4.4% 36.% 36.4% 34.% 32.5% Basal-bolus (%) 32.8% 33.% 4.% 4.6% 5.6% 5.% 57.7% 57.5% egfr (ml/min/.73 m 2, mean) Diabetes Duration (SWITCH 2) < years -5 years 5- years years n=222 n=5 n= 35 n= 58 Age (years, mean) Gender (%, F) 4.9% 47.3% 53.3% 48.% BMI (kg/m 2, mean) HbA c (mean %) egfr (ml/min/.73 m 2, mean) ideg: insulin degludec; iglar: insulin glargine U; CV: cardiovascular; BMI: body mass index (kg/m 2 ); HbA c:haemoglobin A c. Note: SWITCH 2 was a cross-over trial, baseline characteristics are therefore not presented by treatment. a Clinical event definition: an episode requiring assistance of another person to actively administer carbohydrate, glucagon, or take other corrective actions. Plasma glucose values may not be available during an event, but neurological recovery following the return of plasma glucose to normal is considered sufficient evidence that the event was induced by a low plasma glucose concentration Figure DEVOTE Figure Estimated rate ratios of hypoglycemia for insulin degludec vs. insulin glargine U by diabetes duration. Total population SWITCH 2 Rate of hypoglycemia by diabetes duration and treatment (per PYE) DEVOTE SWITCH 2 4 Ideg Iglar U < years -5 years 5- years >= years a) Rates of severe hypoglycemia in the DEVOTE by diabetes duration and treatment, i.e. number of episodes per PYO (patient year of observation). Figure 3 HbA c over time by diabetes duration DEVOTE SWITCH 2 8,5 8, 7,5 7,.6 [.48;.76] 95% CI < years.48 [.29;.8] 95% CI -5 years.76 [.48;.2] 95% CI 5- years.68 [.4;.4] 95% CI years.52 [.35;.77] 95% CI HbA c (%),, Rate ratio Months since randomization < years -5 years 5- years >= years Total population.7 [.6;.8] 95% CI < years.9 [.69;.] 95% CI -5 years.6 [.46;.8] 95% CI 5- years.69 [.52;.9] 95% CI years.62 [.46;.82] 95% CI,, Rate ratio Ideg Iglar U < years -5 years 5- years >= years b) Rates of overall hypoglycemia in SWITCH 2 trial by diabetes duration and treatment HbA c (%) 8, 7,5 7, 6,5 Key Result Weeks since randomization < years -5 years 5- years >= years Results In DEVOTE, baseline characteristics were similar between treatment groups across diabetes duration categories. For both trials renal function declined with greater diabetes duration. The proportion of women and patients on bolus insulin were higher in those with longer diabetes duration in DEVOTE [Table ]. Rates of hypoglycemia tended to increase with diabetes duration in both trials. In DEVOTE, HbA c was similar across categories, while in SWITCH 2, HbA c appeared to increase slightly with diabetes duration. HbA c differences remained throughout the trials [Figure 2 and 3] Rates of hypoglycemia (both severe and overall symptomatic) appeared lower with degludec versus glargine U across all categories of diabetes duration. [Figure ] There was no evidence of a significant interaction between treatment and diabetes duration (p>.5 in both DEVOTE and SWITCH 2), indicating that the benefit of insulin degludec on hypoglycemia is not dependent on a patient s duration of diabetes. Discussion Insulin degludec is an ultralong-acting basal insulin with low day-to-day variability. In DEVOTE and SWITCH 2, degludec was superior to glargine U in terms of lower rates of severe and overall symptomatic hypoglycemia, respectively. 3,4 As expected, rate of hypoglycemia increased with diabetes duration in both trials, likely due to loss of counterregulatory responses over time. 2 The rates of hypoglycemia (both severe and overall) with degludec versus glargine U were consistent with the overall trial results across diabetes duration categories, and hence not showing any signs of a more prominent hypoglycemia benefit for patients with longer diabetes duration. Conclusion In a large cohort of patients with insulin-treated T2D, rates of hypoglycemia increased with diabetes duration. Insulin degludec preserved its relative benefit over insulin glargine U with respect to rate of hypoglycemia across diabetes duration categories. Diabetes duration did not effect the reduction in hypoglycemia seen with degludec versus glargine U in either the DEVOTE or SWITCH 2 trial. The DEVOTE and SWTICH 2 trials were sponsored by Novo Nordisk and registered with ClinicalTrials.gov (NCT and NCT36) Presented at the American Diabetes Association, 79th Scientific Sessions. June 7, 9, San Francisco, CA, USA. References: () International Hypoglycaemia Study Group et al. Hypoglycaemia, cardiovascular disease, and mortality in diabetes: epidemiology, pathogenesis and management. Lancet Diabetes Endocrinol. 9 (2) Sprague and Arbelaez. Glucose counterregulatory responses to hypoglycemia, Pediatr Endocrinol Rev, vol. 9, pp ;. (3) Marso et al. The DEVOTE study group. Efficacy and Safety of Degludec versus Glargine in Type 2 Diabetes. NEJM, 377;8. 7. (4) Wysham et al. Effect of Insulin Degludec vs Insulin Glargine U on Hypoglycemia in Patients With Type 2 Diabetes The SWITCH 2 Randomized Clinical Trial. JAMA ;38(): (5) International Hypoglycaemia Study Group. Glucose concentrations of less than 3. mmol/l (54 mg/dl) should be reported in clinical trials: A joint position statement of the American Diabetes Association and the European Association for the Study of Diabetes. Diabates Care, 4(): Abstract

42 Insulin degludec Poster presentation 374-P 374-P Reduced rates of hypoglycemia irrespective of the definition used when switching to insulin degludec from other basal insulins in routine clinical care: the ReFLeCT study Harold W. de Valk¹, Michael Feher 2,3, Troels Krarup Hansen 4, Johan Jendle 5, Ángel Merchante 6,7, Mette Marie Koefoed 8, Ehsan Parvaresh Rizi 8, Esther Zimmermann 8, Gian Paolo Fadini 9 Department of Internal Medicine, University Medical Center Utrecht, Utrecht, The Netherlands; 2 Beta Cell Diabetes Centre, Chelsea and Westminster Hospital, London, UK; 3 University of Surrey, Guildford, UK; 4 Steno Diabetes Center Aarhus, Aarhus University Hospital, Aarhus, Denmark; 5 Faculty of Medicine and Health, School of Medical Sciences, Örebro University, Örebro, Sweden; 6 University General Hospital of Castellón, Castellón de la Plana, Spain; 7 Jaume I University, Castellón de la Plana, Spain; 8 Novo Nordisk A/S, Søborg, Denmark; 9 Department of Medicine, Division of Metabolic Diseases, University of Padova, Padova, Italy qrs.ly/329xx28 Background Hypoglycemia is a frequent event in patients with diabetes treated with insulin and has been linked to impaired glycemic control.,2 Randomized controlled trials have demonstrated that degludec is associated with less hypoglycemia than with other basal insulins at equivalent glycemic control, across a broad spectrum of patients with diabetes. 3 7 ReFLeCT (Results From Real-World Clinical Treatment with Tresiba ) was a multicenter, prospective, observational study that evaluated the safety and effectiveness of switching from other basal insulins to degludec, as part of routine clinical care, in patients with type (TD) or type 2 diabetes (T2D). 8,9 As different hypoglycemia definitions can impact study outcomes, the present analysis of the ReFLeCT study analyzed previous (pre-specified) and updated (post hoc) American Diabetes Association (ADA) hypoglycemia definitions. Aim The objective of this secondary analysis of the ReFLeCT study was to investigate the change in the rate of hypoglycemia after switching to degludec from other basal insulins, according to different hypoglycemia definitions, in patients with TD or T2D. Methods ReFLeCT study ReFLeCT was a prospective, observational study conducted across seven European countries. 8,9 Patients aged 8 years with TD or T2D who were already on insulin, and whose physician advised that they should switch to degludec treatment, were eligible for inclusion. 8,9 The study comprised a baseline period (4 weeks prior to switching to degludec) and a follow-up period (up to 2 months after switching to degludec). 8,9 Patients attended visits according to routine clinical practice, and could attend up to four visits during the 2-month follow-up period. Patients were instructed to complete 4-week study diaries prior to each visit, collecting day-by-day information on hypoglycemic events. The primary endpoint was the change from the baseline period in the number of overall hypoglycemic events during the 2-month follow-up period. 8,9 In ReFLeCT, switching to degludec from other basal insulins was associated with significantly reduced rates of overall hypoglycemia in combination with improved glycemic control in insulin-treated adults with TD or T2D. 8,9 Hypoglycemia definitions The hypoglycemia definitions included in this analysis consisted of:» Previous (pre-specified) ADA definitions Asymptomatic hypoglycemia: an event not accompanied by typical symptoms of hypoglycemia but with a measured plasma glucose concentration 7 mg/dl (3.9 mmol/l). Documented symptomatic hypoglycemia: an event during which typical symptoms of hypoglycemia are accompanied by a measured plasma glucose concentration 7 mg/dl (3.9 mmol/l). Pseudo-hypoglycemia: an event during which the person with diabetes reports any of the typical symptoms of hypoglycemia with a measured plasma glucose concentration >7 mg/dl (3.9 mmol/l) but approaching that level. Probable symptomatic hypoglycemia: an event during which symptoms typical of hypoglycemia are not accompanied by a plasma glucose determination but that was presumably caused by a plasma glucose concentration 7 mg/dl (3.9 mmol/l). Severe hypoglycemia (Level 3 hypoglycemia): severe hypoglycemia, denoted by severe cognitive impairment that requires external assistance for recovery. 2» Updated (post hoc) ADA definitions Level 2 hypoglycemia: an event with a measured plasma glucose concentration <54 mg/dl (3. mmol/l) indicating serious, clinically important hypoglycemia.,2 Level hypoglycemia: an event with a measured plasma glucose concentration 54 <7 mg/dl ( mmol/l).,2 Statistical analysis The numbers of hypoglycemic events were converted to rates per patient-year of exposure for analysis purposes. Rate ratios for hypoglycemia between the 4-week baseline and 2-month follow-up periods, according to different definitions, were analyzed using negative binomial regression specifying a log-transformed follow-up time offset term adjusted for baseline covariates. Baseline covariates included period (pre/post-switch to degludec), baseline AC, gender, body mass index, duration of diabetes, age and country, in addition to bolus insulin (Yes/No) and sulfonylureas or glinides (Yes/No) for T2D. All statistical tests were two-sided with a significance level of p<.5. Results Baseline characteristics from the overall ReFLeCT study are presented in Table. Seventy (2.6%) patients in the TD group and 67 (.%) in the T2D group withdrew from the study during the follow-up period. In total, 48 TD patients and 56 T2D patients contributed to the present analysis with diary data and complete covariate information. TD Estimated rate ratios (ERRs) demonstrated significantly lower rates of hypoglycemia across the previous ADA hypoglycemia definitions during the 2-month follow-up versus the 4-week baseline period, except for asymptomatic hypoglycemia in patients with TD (Figure a). ERRs also demonstrated significantly lower rates of hypoglycemia for the updated ADA definitions during the 2-month follow-up versus the baseline period (Figure a). T2D ERRs demonstrated significantly lower rates of hypoglycemia across all previous ADA hypoglycemia definitions during the 2-month follow-up versus the 4-week baseline period in patients with T2D (Figure b). The number of Level 3 hypoglycemic events in patients with T2D was insufficient to allow for statistical comparison. ERRs also demonstrated significantly lower rates of hypoglycemia for the updated ADA definitions for hypoglycemia during the 2-month follow-up versus the baseline period (Figure b). Table : Baseline characteristics of patients TD Full analysis set, n T2D Age, years 47.4 (5.7) 65.2 (9.6) Female/male, % 44.2/ /59.6 Duration of diabetes, years 2.4 (3.5) 8. (9.5) BMI, kg/m (4.7) 3. (6.3) Body weight, kg 76.4 (5.6) 87.6 (9.6) AC, % 8. (.3) 8.2 (.4) FPG, mg/dl mmol/l Antidiabetic therapies at baseline, n (%) Proportion on basal insulin Proportion on bolus insulin Proportion on non-insulin antidiabetic therapy 59 (7) 8.8 (3.9) 556 (.) 58 (9.4) 54 (9.7) 62 (56) 9. (3.) 6 (.) 384 (62.8) 379 (62.) Data are mean (SD), unless otherwise specified. BMI, body mass index; FPG, fasting plasma glucose; SD, standard deviation; TD, type diabetes; T2D, type 2 diabetes. Figure : Rate ratios of hypoglycemia according to different hypoglycemia definitions in patients with (a) TD and (b) T2D a) TD ADA asymptomatic hypoglycemia Glucose level 7 mg/dl (3.9 mmol/l) without typical symptoms ADA-documented symptomatic hypoglycemia Glucose level 7 mg/dl (3.9 mmol/l) with typical symptoms ADA pseudo-hypoglycemia Glucose level >7 mg/dl (3.9 mmol/l) with reported symptoms ADA probable symptomatic hypoglycemia No glucose measurement, but assumed glucose level 7 mg/dl (3.9 mmol/l), with reported symptoms Level 3 (ADA severe hypoglycemia) An episode requiring assistance of another person Level 2 hypoglycemia Glucose level <54 mg/dl (3. mmol/l) Level hypoglycemia Glucose level 54 <7 mg/dl ( mmol/l) b) T2D ADA asymptomatic hypoglycemia Glucose level 7 mg/dl (3.9 mmol/l) without typical symptoms ADA-documented symptomatic hypoglycemia Glucose level 7 mg/dl (3.9 mmol/l) with typical symptoms ADA pseudo-hypoglycemia Glucose level >7 mg/dl (3.9 mmol/l) with reported symptoms ADA probable symptomatic hypoglycemia No glucose measurement, but assumed glucose level 7 mg/dl (3.9 mmol/l), with reported symptoms Level 3 (ADA severe hypoglycemia) An episode requiring assistance of another person Level 2 hypoglycemia Glucose level <54 mg/dl (3. mmol/l) Level hypoglycemia Glucose level 54 <7 mg/dl ( mmol/l) Rate ratio [95% CI] 4-week baseline period 2-month follow-up period.88 [.7;.9].83 [.76;.92].44 [.29;.67].53 [.36;.77].28 [.4;.56].8 [.7;.9].9 [.8;.99] Favors degludec.48 [.27;.87].54 [.44;.68].42 [.28;.63].36 [.8;.7] N/A.56 [.4;.79].54 [.43;.67] N % Favors previous basal insulin p<.5; p<.. Severe hypoglycemia, an episode requiring the assistance of another person to actively administer carbohydrate, glucagon or take other corrective actions. Fig a: Models were adjusted for period (pre/post-switch to degludec), baseline AC, gender, BMI, duration of diabetes, age and country. Total follow-up time (patient years) was 38.5 for the 4-week baseline period and 4.5 for the 2-month follow-up period. Fig b: Models were adjusted for period (pre/post-switch to degludec), baseline AC, gender, BMI, duration of diabetes, bolus insulin (Yes/No), sulfonylureas or glinides (Yes/No), age and country. Total follow-up time (patient years) was 4.8 for the 4-week baseline period and 8.8 for the 2-month follow-up period. %, percentage of patients with an event; ADA, American Diabetes Association; BMI, body mass index; CI, confidence interval; E, number of events; R, rate of events per patient-year of exposure; N, number of patients with an event; TD, type diabetes; T2D, type 2 diabetes. Conclusions Switching to degludec from other basal insulins in routine clinical practice was generally associated with lower rates of hypoglycemia when using different hypoglycemia definitions in patients with diabetes. Definitions for Level, 2 and 3 hypoglycemia were well represented in the rate of events and for the change between the baseline and follow-up periods (except for Level 3 hypoglycemia for T2D), strengthening the generalizability of the results from this study. This analysis of ReFLeCT corroborates the findings of the primary study that switching to degludec from other basal insulins is associated with reduced rates of overall hypoglycemia in patients with TD and T2D in routine clinical care. E R N % Key result E R The study was sponsored by Novo Nordisk and is registered with ClinicalTrials.gov (NCT23927). Presenter Harold W. de Valk has received advisory board fees (paid directly to University Medical Center Utrecht) from Novo Nordisk, Eli Lilly and MSD. The authors are grateful to João Diogo Da Rocha Fernandes, Novo Nordisk for review of and input to the poster, and to Jon Griffiths, Watermeadow Medical (supported by Novo Nordisk) for writing assistance. Presented at the American Diabetes Association, 79th Scientific Sessions. June 7, 9, San Francisco, CA, USA. References: () Zekarias & Seaquist. Hypoglycemia in Diabetes: Epidemiology, Impact, Prevention and Treatment (Accessed Mar 9); (2) UK Hypoglycemia Study Group. Diabetologia 7;5:4 7; (3) Ratner et al. Diabetes Obes Metab 3;5:75 84; (4) Davies et al. Diabetes Obes Metab 6;8:96 9; (5) Marso et al. N Engl J Med 7;377:723 32; (6) Lane et al. JAMA 7;38:33 44; (7) Wysham et al. JAMA 7;38:45 56; (8) Fadini et al. Diabetic Medicine 9:36 (Suppl. );6 (Poster 8); (9) Fadini et al. Diabetic Medicine 9:36 (Suppl. );6 (Poster 82); () Seaquist et al. Diabetes Care 3;36:384 95; () American Diabetes Association. Diabetes Care 9;42(Suppl. ):S6 S7; (2) Agiostratidou et al. Diabetes Care 7;4: Abstract

43 Insulin degludec Poster presentation 375-P 375-P Switching to insulin degludec from other basal insulins reduces rates of hypoglycemia across patient subgroups in routine clinical care: the ReFLeCT study Harold W. de Valk, Michael Feher 2,3, Troels Krarup Hansen 4, Johan Jendle 5, Ángel Merchante 6,7, Mette Marie Koefoed 8, Ehsan Parvaresh Rizi 8, Esther Zimmermann 8, Gian Paolo Fadini 9 Department of Internal Medicine, University Medical Center Utrecht, Utrecht, The Netherlands; 2 Beta Cell Diabetes Centre, Chelsea and Westminster Hospital, London, UK; 3 University of Surrey, Guildford, UK; 4 Steno Diabetes Center Aarhus, Aarhus University Hospital, Aarhus, Denmark; 5 Faculty of Medicine and Health, School of Medical Sciences, Örebro University, Örebro, Sweden; 6 University General Hospital of Castellón, Castellón de la Plana, Spain; 7 Jaume I University, Castellón de la Plana, Spain; 8 Novo Nordisk A/S, Søborg, Denmark; 9 Department of Medicine, Division of Metabolic Diseases, University of Padova, Padova, Italy qrs.ly/329xx28 Background Hypoglycemia occurs frequently in patients with type (TD) or type 2 diabetes (T2D) treated with insulin; it can have considerable negative effects on patient quality-of-life and is often a major limiting factor to glycemic control in clinical practice.,2 Insulin degludec (degludec) is a basal insulin analog with an ultra-long duration of action and a stable glucose-lowering profile. 3 Randomized controlled trials have demonstrated that degludec is associated with less hypoglycemia than with other basal insulins at equivalent glycemic control, across a broad spectrum of patients with diabetes. 4 8 ReFLeCT (Results From Real-World Clinical Treatment with Tresiba ) was a multicenter, prospective, observational study that evaluated the safety and effectiveness of switching from other basal insulins to degludec, as part of routine clinical care, in patients with TD or T2D. 9, Aim The aim of the present pre-specified subgroup analyses of the primary endpoint of ReFLeCT was to assess whether the overall result lower overall hypoglycemia rates after switching from other basal insulins to degludec was consistent in different subgroups of patients categorized according to baseline characteristics or physician s reason for switching to degludec. Methods ReFLeCT study ReFLeCT (NCT23927) enrolled patients aged 8 years with TD or T2D who were already on insulin, and whose physician advised that they should switch to degludec treatment. 9, The study comprised a baseline period (4 weeks prior to switching to degludec) and a follow-up period (2 months ±45 days after switching to degludec). 9, Patients attended visits according to routine clinical practice, and could attend up to four visits during the 2-month follow-up period. Patients were instructed to complete 4-week study diaries prior to each visit, collecting day-by-day information on hypoglycemic events. The primary endpoint was the change from the baseline period in the number of overall hypoglycemic events during the 2-month follow-up period. 9, Overall hypoglycemia was defined as any event recorded as hypoglycemia in the diary irrespective of symptoms, blood glucose level or time of day. In ReFLeCT, switching to degludec from other basal insulins was associated with significantly reduced rates of overall hypoglycemia in combination with improved glycemic control, in insulin-treated adults with TD or T2D. 9, Pre-specified subgroup analyses Data were analyzed separately for patients with TD and T2D. Patients were characterized according to baseline AC categories (<7.5, 7.5 <8.5, 8.5 <9.5, 9.5%), diabetes duration quartiles and physician s reason for initiating degludec (concern about hypoglycemia or not hypoglycemia-related). The number of hypoglycemic events was converted to rates per patient-year of exposure (PYE) for analysis purposes. Overall hypoglycemia was analyzed using negative binomial regression models specifying a log-transformed follow-up time offset term. Analyses were adjusted for relevant baseline covariates and included all patients with complete covariate information.» Covariates included period (pre/post-switch to degludec), AC, gender, body mass index, duration of diabetes, age and country.» For T2D, additional covariates included bolus insulin (Yes/No) and sulfonylurea or glinides (Yes/No). All statistical tests were two-sided with a significance level of p<.5. Results The baseline characteristics of patients with TD and T2D are presented in Table. Median (interquartile range [IQR]) AC was 7.9% (7.2; 8.7) in patients with TD and 8.% (7.3; 8.9) in those with T2D. For diabetes duration, median (IQR) was 9.8 years (.6; 3.6) and 6.8 years (.9; 22.8) in patients with TD and T2D, respectively. The most common physician s reasons for switching to degludec were concern about hypoglycemia (64.6% in TD; 36.2% in T2D) and the need to improve glycemic control (63.9% in TD; 73.3% in T2D) (Figure ). There were 8.7% of patients with TD and.8% of patients with T2D who switched to degludec to increase daily dosing flexibility. TD Estimated rate ratios (ERRs) of hypoglycemia were consistently in favor of degludec across subgroups (8 33% lower during the 2-month follow-up period versus the 4-week pre-switch baseline period), with the majority showing a statistically significant difference (Figure 2a). T2D ERRs of hypoglycemia were consistently in favor of degludec across subgroups (27 6% lower during the 2-month follow-up period versus the 4-week pre-switch baseline period), with all showing a statistically significant difference except for patients with baseline AC 8.5 <9.5 (Figure 2b). Table : Baseline characteristics of patients TD T2D Full analysis set, n Age, years 47.4 (5.7) 65.2 (9.6) Female/male, % 44.2/ /59.6 Duration of diabetes, years 2.4 (3.5) 8. (9.5) BMI, kg/m² 26. (4.7) 3. (6.3) AC, % 8. (.3) 8.2 (.4) FPG, mg/dl [mmol/l] 59 (7) [8.8 (3.9)] 62 (56) [9. (3.)] Prior antidiabetic therapy, n (%) Basal insulin before switch Insulin glargine Insulin detemir Other /missing Metformin SGLT-2 inhibitors GLP- receptor agonists DPP-4 inhibitors Sulfonylureas Thiazolidinediones Figure : Physicians reasons for switching basal insulin to degludec Number of patients (64.6%) 354 (63.9%) Concern about hypoglycemia Increase dosing flexibility TD 48 (8.7%) 23 (4.2%) 22 (36.2%) 355 (63.8) 26 (22.7) 75 (3.5) 4 (7.4) 4 (.7) 4 (.7) 2 (.4) () (.2) 448 (73.3%) Improve glycemic control Other Physicians could select multiple reasons, with patients included in counts for each reason recorded. Percentages are based on the number of patients with available data (TD: 554; T2D: 6). Patients were switched to degludec for only one reason (TD: n=3; T2D: n=444), more than one reason (TD: n=224; T2D: n=5) or there was no reason provided (TD: n=; T2D: n=7) TD, type diabetes; T2D, type 2 diabetes T2D 72 (.8%) 36 (59.) 27 (.8) 23 (.) 37 (5.2) 66 (.8) 64 (.5) 49 (8.) 4 (6.7) 8 (.3) Basal insulin dose, U/day 25. (4.) 37.5 (33.9) Bolus insulin dose, U/day 27.3 (6.9) 38.9 (3.7) Proportion on bolus insulin, n (%) 58 (9.4) 384 (62.8) Time of basal insulin injections, n (%) Morning 67 (2.) 82 (3.4) Evening 39 (57.4) 436 (7.4) Morning and evening 7 (3.6) 93 (5.2) Hypoglycemia-prone patients, n (%) 379 (68.2) 346 (56.6) Data are mean (SD), unless otherwise specified The split between insulin glargine units/ml and 3 units/ml was not collected, but, based on the study timeline, the majority of patients might have been using glargine units/ml; Including pre-mix insulins; Defined as patients having any of the following: ) experienced at least one severe hypoglycemic event within the last year; 2) moderate chronic renal failure, defined as egfr 3 59 ml/min/.73 m² per CKD-EPI as specified by physician statement (or according to national reference definitions if they differ from the values stated); 3) hypoglycemic symptom unawareness (history of impaired autonomic responses); 4) diabetes duration >5 years for TD or exposed to insulin >5 years for T2D %, percentage of patients; BMI, body mass index; CKD-EPI, Chronic Kidney Disease Epidemiology Collaboration formula; DPP-4, dipeptidyl peptidase-4; egfr, estimated glomerular filtration rate; FPG, fasting plasma glucose; GLP-, glucagon-like peptide-; n, number of patients; SD, standard deviation; SGLT-2, sodium-glucose co-transporter 2; TD, type diabetes; T2D, type 2 diabetes; U, units 2 (3.4%) Figure 2: Rate ratios of overall hypoglycemia in patients with (a) TD or (b) T2D a) TD Diabetes duration quartiles (years).6 (n=8).92 [.77;.] > (n=2).67 [.56;.79] > (n=2).88 [.75;.3] >3.6 (n=2).79 [.67;.93] Baseline AC (%) <7.5 (n=74).77 [.67;.87] 7.5 <8.5 (n=5).79 [.67;.92] 8.5 <9.5 (n=95).75 [.59;.96] 9.5 (n=62).9 [.67;.23] Reason for switch Hypoglycemia (n=3).77 [.7;.85] Not hypoglycemia (n=7).9 [.74;.9] b) T2D Diabetes duration quartile (years).9 (n=36) > (n=27) > (n=26) >22.8 (n=27) Baseline AC (%) <7.5 (n=58) 7.5 <8.5 (n=82) 8.5 <9.5 (n=) 9.5 (n=75) Reason for switch Hypoglycemia (n=78) Not hypoglycemia (n=338) p<.5; p<.; p values relate to a test for an interaction between period and subgroup for overall hypoglycemia Analyzed using negative binomial regression models. All models were adjusted for period (pre/post-switch to degludec), baseline AC, gender, body mass index, duration of diabetes, age and country. For T2D, there was additional adjustment for baseline bolus insulin (Yes/No) and sulfonylurea or glinides (Yes/No). The rate ratio represents the change in the 2-month follow-up period rate compared with the 4-week baseline period rate, with the significance of a test of the hypothesis that the rate ratio= indicated by asterisks. Based on patient diary periods with 26 3 days. %, percentage of patients with an event; CI, confidence interval; E, number of events; n, number of patients included in fully-adjusted analyses; N, number of patients with an event; R, rate of events per patient-year of exposure; TD, type diabetes; T2D, type 2 diabetes Discussion Overall, switching to degludec from other basal insulins was associated with lower rates of hypoglycemia across different subgroups of patients with TD or T2D. There were no significant interactions between period (pre- or post-switch to degludec) and subgroup for overall hypoglycemia. Conclusion Favors degludec Favors previous basal insulin Key result Rate ratio [95% CI] p value 4-week baseline period 2-month follow-up period Rate ratio [95% CI] p value 4-week baseline period 2-month follow-up period.45 [.3;.66].46 [.33;.66].45 [.27;.74].5 [.36;.7].43 [.32;.56].43 [.29;.62].73 [.47;.4].46 [.27;.79].4 [.3;.5].6 [.46;.8] Favors degludec Favors previous basal insulin These new data support the primary study findings and provide further evidence that the reduced rates of hypoglycemia associated with switching from other basal insulins to degludec are not restricted to specific patient subgroups based on diabetes duration, baseline glycemic control or physician s reason for switching basal insulin. Irrespective of baseline characteristics or physician s reason for initiating degludec, switching to degludec from other basal insulins was associated with reduced rates of overall hypoglycemia in patients with TD or T2D, in real-world clinical practice N N % % E E R R N N % % E E R R The study was sponsored by Novo Nordisk and is registered with ClinicalTrials.gov (NCT23927). Presenter Harold W. de Valk has received advisory board fees (paid directly to University Medical Center Utrecht) from Novo Nordisk, Eli Lilly and MSD. The authors are grateful to João Diogo Da Rocha Fernandes, Novo Nordisk for review of and input to the poster, and to Anna Campbell, Watermeadow Medical (supported by Novo Nordisk) for writing assistance. Presented at the American Diabetes Association, 79th Scientific Sessions. June 7, 9, San Francisco, CA, USA. References: () Khunti et al. Diabetes Res Clin Pract 7;3:2 9; (2) Cryer. Diabetes 8;57:369 76; (3) Haahr & Heise. Clin Pharmacokinet 4;53:787 8; (4) Ratner et al. Diabetes Obes Metab 3;5:75 84; (5) Davies et al. Diabetes Obes Metab 6;8:96 9; (6) Marso et al. N Engl J Med 7;377:723 32; (7) Lane et al. JAMA 7;38:33 44; (8) Wysham et al. JAMA 7;38:45 56; (9) Fadini et al. Diabetic Medicine 9:36 (Suppl. );6 (Poster 8); () Fadini et al. Diabetic Medicine 9:36 (Suppl. );6 (Poster 82) Abstract

44 Insulin degludec/liraglutide Poster presentation -P -P Use of IDegLira is associated with high persistency rates up to 2 months and improved HbA c after 6 months in Swedish clinical practice Björn Eliasson, Jan Ekelund 2, Ann-Marie Svensson,2, Mervete Miftaraj 2, Ann-Charlotte Mårdby 3,4, João Diogo Da Rocha Fernandes 5 Department of Molecular and Clinical Medicine, University of Gothenburg, Gothenburg, Sweden; Department of Medicine, Sahlgrenska University Hospital, Gothenburg, Sweden; 2 Centre of Registers Västra Götaland, Gothenburg, Sweden; 3 Novo Nordisk, Malmö, Sweden; 4 Section of Epidemiology and Social Medicine, Institute of Medicine, University of Gothenburg, Gothenburg, Sweden; 5 Novo Nordisk A/S, Søborg, Denmark Presenting author qrs.ly/ih9xx35 Aim IDegLira is a once-daily, fixed-ratio combination of basal insulin degludec and the glucagon-like peptide- receptor agonist (GLP-RA), liraglutide, used for the treatment of patients with type 2 diabetes (T2D). The maximum daily dose of IDegLira is 5 units (U) (5 U degludec and.8 mg liraglutide). Several randomised clinical trials have confirmed the safety and efficacy of IDegLira, 9 but real-world evidence investigating IDegLira has been limited. IDegLira was launched in Sweden in May 5. This real-world study aimed to assess persistence in the use of IDegLira in patients with T2D and characterize the clinical outcomes associated with IDegLira persistence. Methods This retrospective cohort study included patients initiating IDegLira between 26 May 5 and 3 December 7. Data were obtained from registers that have nationwide coverage: the Swedish National Diabetes Registry (NDR), and the Swedish Prescribed Drug Register (SPDR), and linked on an individual level using unique patient identifiers. Persistence was defined as:» Patients who collected IDegLira from a pharmacy at least one more time within 6 months after the first collection (index date), were defined as having 6-month persistency.» Patients who collected IDegLira from a pharmacy at least one more time within 6 months after the index date, and at least one more time between the 6 and 2 months after initiation, were defined as having 2-month persistency. Dose calculations were only made for patients with at least 8 days between first and last collection and were calculated as follows for the first 8 days:» A,..,A k was the amount of IDegLira that was collected at dates D,..,D k, where K is the smallest integer such that D k D 8 days.» Average dose during the first 8 days was calculated as: ([A +..+A k-2 ]+[8 D k- ]A k- /[D k D k- ])/8.» Average dose for other periods are calculated similarly. Change in outcomes was examined using an analysis of variance (ANOVA), with change in HbA c (glycated hemoglobin)/body mass index (BMI) as the dependent variable, and with persistency as the independent variable, and with pre-index HbA c /BMI as a covariate. Results 365 patients made at least one pharmacy collection of IDegLira during the study period (average number of collections was 3.6/year); however, 65 patients were not found in the Swedish NDR and were excluded from the study (Figure ). The study population included 3 patients with T2D, of whom 2432 met the inclusion criteria for persistency analysis (Figure ). At baseline, mean (standard deviation [SD]) age was 6.3 years (.79), BMI was 33.2 kg/m 2 (5.69), HbA c was 9.% (.5) and duration of diabetes was 2.7 years (7.58) (Table ). Prior to IDegLira initiation, 66% of patients were on insulin and 6% were on insulin ± GLP-RA (Table ). Figure : Patient disposition Inclusion criteria Diagnosed T2D Age 8 years Initiating IDegLira between and months 2 months 5 non-persistent patients 278 non-persistent patients 365 patients 3 patients 2432 patients 76 patients 65 patients Patients were not found in the Swedish NDR or could not be verified to be diagnosed with T2D 568 patients Patients did not initiate IDegLira on or before 3 June patients Patients did not initiate IDegLira on or before January persistent patients 228 persistent patients Patients were defined as non-persistent at 6 months if IDegLira was not collected at the pharmacy within 6 months after the first collection. Patients were defined as persistent at 6 months if IDegLira was collected from a pharmacy at least one more time within 6 months after the first collection (index date). The daily dose of IDegLira was calculated for 988 patients. Patients were defined as non-persistent at 2 months if IDegLira was not collected at the pharmacy in months 2 and months 6 2, or if IDegLira was not collected in the first 6 months but collected in months 6 2. Patients were defined as persistent at 2 months if IDegLira was collected from a pharmacy at least one more time within 6 months after the index date, and at least one more time in the 6 2 months after initiation. IDegLira, insulin degludec/liraglutide; NDR, National Diabetes Register; T2D, type 2 diabetes. Table : Baseline characteristics N Mean SD Age (years) Sex Male Female Duration of diabetes (years) BMI (kg/m 2 ) HbA c (%) Systolic blood pressure (mmhg) Diastolic blood pressure (mmhg) LDL (mmol/l) HDL (mmol/l) Medication prior to initiating IDegLira Proportion of patients (%) Non injectable (including diet-only) Insulin ± OADs GLP-RA ± OADs Insulin + GLP-RA ± OADs Selected concomitant medications and comorbidities Antihypertensive treatment Lipid-lowering treatment Microalbuminuria Macroalbuminuria Retinopathy The number of patients in the cohort differs due to data availability. BMI, body mass index; GLP-RA, glucagon-like peptide- receptor agonist; HDL, high density lipoprotein; IDegLira, insulin degludec/ liraglutide; HbA c, glycated hemoglobin; LDL, low density lipoprotein; N, number of patients in the cohort; OAD, oral antidiabetic drug; SD, standard deviation. At 6 months, persistence was determined in 2432 patients. Of these, 94% were considered persistent (Figure 2). At 2 months, persistence was determined in 76 patients. Of these, 84% were considered persistent (Figure 2). The mean daily dose of IDegLira in patients who were persistent for 6 months, between the first and last collection, was 33.4 U (Table 2).» Patients who had no collection of medication during the 6 months pre-index used the lowest mean daily dose of IDegLira (23.6 U), between the first and last collection of IDegLira (Table 2).» Patients who were treated with a free combination of insulin and GLP-RA preindex used the highest mean daily dose of IDegLira (39.8 U), between the first and last collection of IDegLira (Table 2).» After 6 months, least squares mean [95%] change in HbA c was.9% [.;.8] and.5% [.8;.2] in persistent and non-persistent patients, respectively (Figure 3A). After 6 months, least squares mean [95%] change in BMI was.34 kg/m 2 [.5;.2] and.23 kg/m 2 [.8;.2] in persistent and non-persistent patients, respectively (Figure 3B) Figure 2: Persistency of IDegLira at 6 and 2 months Key result Percentage of patients 6% n=5 6% n= % n=228 84% n= months 2 months Persistent Non-persistent IDegLira, insulin degludec/liraglutide. Table 2: Average daily dose of IDegLira in persistent patients Overall N Mean (U) SD Median (U) Mean daily dose during days Mean daily dose during days Mean daily dose during days Mean daily dose between first and last withdrawal Pre-index treatment Insulin, long-acting GLP-RA Insulin and GLP-RA MDI OAD only No collection of medication during 6 months pre-index Mean daily dose (U) between first and last withdrawal. GLP-RA, glucagon-like peptide- receptor agonist; IDegLira, insulin degludec/liraglutide; HbA c, glycated hemoglobin; MDI, multiple daily insulin injections; N, number of patients; OAD, oral antidiabetic drug; SD, standard deviation; U, units. Figure 3: Change in HbA c (A) and BMI (B) 6 months after initiating treatment with IDegLira A HbA c (%) at 6 months after initiation Persistent Non-persistent 9.% 9.% B BMI (kg/m 2 ) at 6 months after initiation Data are mean values. Treatment period is the observation closest to date of last collection of IDegLira within ± 9 days. BMI, body mass index; IDegLira, insulin degludec/liraglutide; HbA c, glycated hemoglobin; n, number of patients. Discussion 8.% 33.2 The majority of patients who initiated IDegLira persisted with treatment over the -year study period. Reductions of ~.% in HbA c were observed in patients who persisted with IDegLira treatment for 6 months. The mean dose of IDegLira was up to ~ U lower in this real-world setting compared with what has been reported in randomized clinical trials (~33 U in this study vs. ~45 U in DUAL II 2 ). An improvement in glycemic control was observed during this real-world analysis, despite this difference in dose. Outcomes observed are consistent with previous real-world studies. Limitations of this study include:» Risk of reverse causation, as the improvement in HbA c may have resulted in persistence.» Risk of confounding bias due to differences in baseline characteristics between persistent and non-persistent patients. Conclusion (.9) (.5) 8.5% n=99 n=62 Persistent Non-persistent (.34) (.23) 3.9 n=643 n=46 After 2 months, 84% of patients (482/76) were persistent. IDegLira use in Sweden was associated with improved HbA c and no clinically significant change in BMI at a moderate mean daily insulin dose (~33 U), after 6 months. This retrospective analysis was sponsored by Novo Nordisk A/S. Presenter Björn Eliasson has participated in advisory panels for Amgen Inc, AstraZeneca, Eli Lilly and Company, Merck Sharp & Dohme Corp., Novo Nordisk A/S and Sanofi; received research support from Sanofi; received consulting fees from Novo Nordisk A/S and Sanofi; and has participated in speakers bureau for Boehringer Ingelheim International GmbH, Eli Lilly and Company, Novo Nordisk A/S and Sanofi. The authors are grateful to Sejal Varsani, Watermeadow Medical (supported by Novo Nordisk A/S) for writing assistance. Presented at the American Diabetes Association, 79th Scientific Sessions. June 7, 9, San Francisco, CA, USA. References: () Gough et al. Diabetes Obes Metab 5;7():965 73; (2) Buse et al. Diabetes Care 4;37(): ; (3) Linjawi et al. Diabetes Ther 7;8(): 4; (4) Rodbard et al. Diabet Med 7;34(2):89 96; (5) Lingvay et al. JAMA 6;35(9):898 97; (6) Harris et al. Diabetes Obes Metab 7;9(6):858 65; (7) Billings, et al. Diabetes Care 8;4(5):9 6; (8) DUAL VIII, [last accessed May 9]; (9) Philis-Tsimikas et al. Diabetes Obes Metab 9;2(6):399 48; () Price et al. Diabetes Obes Metab 8;: Abstract

45 Liraglutide diabetes Late-breaking poster presentation -LB -LB Non-Severe Hypoglycemia Predicts Increased Risk of Subsequent Severe Events in Patients With Type 2 Diabetes Simon Heller ; Elise Hachmann-Nielsen 2 ; Kajsa Kvist 2 The University of Sheffield, Sheffield, UK (s.heller@sheffield.ac.uk); 2 Novo Nordisk A/S Søborg, Denmark qrs.ly/iy9xx3h Aim It is well-known that higher rates of non-severe hypoglycemic episodes (NSHEs) associate with a greater risk of severe hypoglycemic episodes in patients with type diabetes. We aimed to investigate whether a similar association existed in patients with type 2 diabetes (T2D). We also aimed to investigate the association between non-severe hypoglycemia and other adverse events: time to first Major Adverse Cardiovascular Event (MACE ), time to cardiovascular (CV) death, and time to all-cause mortality. Methods We used data from the LEADER trial; a cardiovascular outcomes trial with patients randomized to either the GLP-RA liraglutide or placebo. The LEADER trial included 934 T2D patients at high risk of cardiovascular events; pre-existing CV-disease (8%) or risk factors for CV-disease (9%). [Table ] The trial information and baseline data has previously been published in details. 2,3 During the total trial period of 35,563 patient years of observation (median follow-up of 3.8 years), a total of 27,933 NSHEs were registered (BS <3. mmol/l). There was 433 severe hypoglycemic episodes,,32 first 3-point-MACEs, 497 cases of CV death and 828 cases of all-cause mortality. In this secondary analysis we explored if the annual rate of NSHEs was associated with time to first severe hypoglycemic episode, time to first MACE, time to CV-death and time to all cause mortality. A Cox proportional hazards model was used, adjusted for randomized treatment arm, and annual rate of NSHE as a timedependentcovariate with threelevels; Group A: <2 NSHEs per year (reference) Group B: 2- NSHEs per year Group C: 2 NSHEs per year The time-dependent covariate was updated at each NSHE event time. The association between NSHE and outcome is estimated with Hazard ratios (HR). The robustness of the results was investigated with three sensitivity analysis:. adjusting the primary analysis for baseline information (sex, baseline HbA c, diabetes duration, age and insulin treatment) 2. Rather than updating the time-dependent covariate at the time of each event it is updated in windows of size days. The NSHE event rate at the closure of each window is used as covariate value for the following window. The HR for each value of the timedependentcovariate is usedto investigate the association. 3. The first year of observation is used to categorize all patients according to group A-C. The subsequent follow-up time beyond the first year is used to investigate the association with a Cox regression model with two covariates with constant values. The first sensitivity analysis investigated if a high annual NSHE event rate can be moderated through selected baseline characteristics. The second sensitivity analysis investigated the dependence of the results toward the method of accounting for the dynamic NSHE rate. The analysis was performed with a range of window sizes. The third sensitivity analysis was performed to avoid the timedependent covariate but instead categorize patients at a given followup time and use this as constant covariate throughout the analysis. In this analysis the number of events (severe hypoglycemia, MACE; CVdeaths) is notably reduced. Results Baseline characteristics according to an exclusive A-C grouping where patients are categorized according to their highest observed annual NSHE rate is similar with regard to age, BMI and gender distribution. Patients with risk time in group C had lower baseline HbA c, longer duration of diabetes and less likely to be insulin naïve. [Table ] Table Baseline characteristics Overall Group A Group B Group C N Age, years 64.3 (7.2) 64.9 (7.26) (7.2) (7.2) BMI, kg/m (6.3) (6.35) 3.9 (6.) 3.54 (6.7) HbAc (%) 8.7 (.5) 8.72 (.55) 8.73 (.52) 8.48 (.38) Female sex (%) 36% 35% 36% 37% Diabetes duration, years 2.8 (8.).9(7.6) 4.5(8.4) 6.2(8.5) Existing CVD/CKD (%) 8% 8% 83% 85% CVD risk factors (%) 9% % 7% 5% Insulin naive (%) 55.5% 6% 44% 36% Table notes: Baseline characteristics for the total trial population and those experiencing a NSHE annual event rate >2 (group C), for those experiencing an annual event rate>2 but not >2 at any time during trial (group B) and for with n annual event rate <=2 during trial (group A). Due to incomplete baseline information 7 subjects were not included in the table. Figure 2 Time to first severe hypoglycemic event Group A (reference) Group B Group C Time to first MACE Group A (reference) Group B Group C Time to first CV death Group A (reference) Group B Group C Time to first all-cause death Group A (reference) Group B Group C Sensitivity analysis : primary analysis adjusted for baseline information Time to first: Severe hypoglycemic event Group A Group B Group C MACE Group A Group B Group C CV death Group A Group B Group C All-cause death Group A Group B Group C,,, Hazard ratio Hazard ratios (95% CI) for severe hypoglycemia, MACE, CV death and all-cause death by NSHE rate groups. N/A 2.4 [.6; 2.86] 95%CI 2.86 [.99; 4.] 95%CI N/A. [.87;.7] 95%CI.35 [.9;.66] 95%CI N/A.88 [.69;.2] 95%CI.55 [.2; 2.4] 95%CI N/A.99 [.83;.9] 95%CI.6 [.24; 2.6] 95%CI Time to first: Severe hypoglycemic event Group A Group B Group C MACE Group A Group B Group C CV death Group A Group B Group C All-cause death Group A Group B Group C,,, Hazard ratio N/A.98 [.42; 2.75] 95%CI 5. [2.84; 8.84] 95%CI N/A.2 [.86;.2] 95%CI.5 [.; 2.23] 95%CI N/A.4 [.88;.49] 95%CI 2.8 [.7; 3.7] 95%CI N/A.9 [.88;.34] 95%CI.8 [.; 2.92] 95%CI N/A 2.3 [.74; 3.7] 95%CI 3. [2.8; 4.44] 95%CI N/A. [.87;.6] 95%CI.32 [.7;.63] 95%CI N/A.88 [.7;.6] 95%CI.32 [.7;.63] 95%CI N/A. [.84;.] 95%CI.56 [.2; 2.] 95%CI,,, Hazard ratio Figure 3 Sensitivity analysis (-3) Sensitivity analysis 2: days of observation windows Time to first: Severe hypoglycemic event Group A Group B Group C MACE Group A Group B Group C CV death Group A Group B Group C All-cause death Group A Group B Group C,,, Hazard ratio Key Result Sensitivity analysis 3: Patients categorized by NSHE annual event rate in first 2 months N/A.83 [.32; 2.55] 95%CI 4.93 [2.84; 8.56] 95%CI N/A.6 [.89;.25] 95%CI.24 [.8;.89] 95%CI N/A.87 [.66;.5] 95%CI.57 [.88; 2.78] 95%CI N/A.8 [.89;.3] 95%CI.35 [.83; 2.9] 95%CI Higher rates of NSHE was associated with a higher rate of severe hypoglycemia, MACE, CV death and all-cause death in patients with T2D. [Figure 2] For MACE, CV-death and all-cause mortality the association was drivenby patientswith an annualevent 2. The rate for severe hypoglycemia was increased - also when the annual NSHE event rate was 2 per year. The sensitivity analyses support the primary findings. In the third sensitivity analysis the total number of events is notably reduced, which affects the CI but the point estimates are consistent. [Figure 3] Discussion There is an increasing amount of evidence pointing to hypoglycaemia as a detrimental factor in development of complications to both type and type 2 diabetes. 4 Previously a number of effect pathways has been demonstrated, one being that even non-severe hypoglycemia is associated with acute and persistent prothrombotic effects illustrating a possible mechanism by which hypoglycemia can increase CV risk. 5 Moreover secondary analysis of a number of large landmark trials has consistently shown the association between hypoglycemia and increased CV risk. 6 As the findings are limited to observational associations it is therefore continuously discussed if hypoglycemia is a marker or mediator of the associated CV risk. Our results supports the findings that there is a strong association between the rate of non-severe hypoglycemia and adverse outcomes, consistent withinmultiplesensitivity analysis. Independent of causality, reducing the risk of any hypoglycemia by lifestyle intervention or pharmacological solutions may be beneficial for any patient including those at high CV risk. Conclusion Higher rates of NSHE was associated with a higher rate of severe hypoglycemia, MACE, CV death and all-cause death in patients with T2D (Figure 2-3); Our results suggest that in this T2DM population, a high rate of NSHEs is associated with serious adverse events and should be avoided. 3-point MACE (CV death, non-fatal MI, non-fatal stroke) This study was sponsored by Novo Nordisk. The LEADER trial was sponsored by Novo Nordisk and was registered with ClinicalTrials.gov (NCT7948) Presented at the American Diabetes Association, 79th Scientific Sessions. June 7, 9, San Francisco, CA, USA. References: () Kovatchev et al, Diabetes Care 2:87-875, 998 (2) Marso et al. Am Heart J. 3 Nov;66(5):823-3 (3) Marso et al. N Engl J Med 6; 375:3-322 (4) International Hypoglycaemia Study Group. Lancet Diabetes Endocrinol 7: , 9 (5) Chow et al. Diabetes Care 4: , 8 (6) Bonds et al. BMJ 34:b499, Abstract

46 Semaglutide once weekly Poster presentation 986-P 986-P Efficacy of semaglutide by background sodium glucose co-transporter-2 inhibitor: a post hoc analysis of SUSTAIN 9 Bernard Zinman ; Vaishali Bhosekar 2 ; Robert S. Busch 3 ; Ingrid Holst 4 ; Bernhard Ludvik 5 ; James Thrasher 6 ; Vincent C. Woo 7 ; Athena Philis-Tsimikas 8 Mount Sinai Hospital, Toronto, ON, Canada ( zinman@lunenfeld.ca; Tel: ); 2 Novo Nordisk, Bengaluru, India; 3 Albany Medical Center Division of Community Endocrinology, Albany, NY, USA; 4 Novo Nordisk A/S, Søborg, Denmark; 5 st Medical Department and Karl Landsteiner Institute for Obesity and Metabolic Disorders, Rudolfstiftung Hospital, Vienna, Austria; 6 Medical Investigations, Inc., Little Rock, AR, USA; 7 University of Manitoba, Winnipeg, MB, Canada; 8 Scripps Whittier Diabetes Institute, San Diego, CA, USA. (qrs.ly/mo9xx3r) Aim Clinical management guidelines recommend glucagon-like peptide- receptor agonists (GLP-RAs) and sodium glucose co-transporter-2 inhibitors (SGLT-2is) for use after metformin in patients with type 2 diabetes (T2D) and established cardiovascular disease.,2 However, clinical trial data on the efficacy and safety of the concomitant use of GLP-RAs and SGLT-2is in the management of T2D are scarce. The SUSTAIN 9 trial showed that, in subjects with T2D inadequately controlled with SGLT-2i therapy with or without metformin or a sulfonylurea, the addition of the GLP-RA semaglutide (Novo Nordisk, Denmark) improved glycemic control, lowered body weight, and was generally well tolerated. 3 This post hoc analysis aimed to determine whether the efficacy and safety of semaglutide vs placebo was consistent in subjects on different background SGLT-2is. Methods SUSTAIN 9 trial design SUSTAIN 9 was a randomized, double-blind, placebo-controlled, multinational trial (Figure ). 3 Subjects were randomized to semaglutide. mg or placebo, both subcutaneous once weekly, as an add-on to SGLT-2i therapy, with or without metformin or a sulfonylurea. 3 Semaglutide dosing began at.25 mg, doubling every 4 weeks until the maintenance dose was reached. 3 The primary and confirmatory secondary endpoints were change from baseline in HbA c and body weight at week 3, respectively. 3 Figure : SUSTAIN 9 trial design 3 Stable treatment with SGLT-2i ± MET or SU for minimum 9 days 32 subjects with T2D Age 8 years HbA c 7..% egfr 6 ml/min/.73 m 2 Randomization (:) + Semaglutide. mg Placebo Dose escalation 8 weeks Treatment maintenance 22 weeks Treatment duration 3 weeks Follow-up 5 weeks Japan: minimum age for enrollment years mmol/mol. Stable treatment with SGLT-2i was defined as having started SGLT-2i treatment 9 days before screening. egfr, estimated glomerular filtration rate; MET, metformin; SGLT-2i, sodium glucose co-transporter-2 inhibitor; SU, sulfonylurea. Statistical analysis In this post hoc analysis, changes in HbA c and body weight from baseline to week 3 were analyzed by background SGLT-2i: canagliflozin; dapagliflozin (includes dapagliflozin and dapagliflozin propanediol monohydrate); empagliflozin; or other (ipragliflozin L-proline, luseogliflozin, and tofogliflozin; drugs available only in Japan) using an analysis of covariance. Proportions of subjects achieving HbA c targets (<7.% and 6.5%), weight-loss responses ( 5% and %), and the triple composite endpoint of HbA c <7.%, no weight gain, and no severe or blood glucose-confirmed hypoglycemia were analyzed by background SGLT-2i using a logistic regression model. A test for interaction was used to evaluate any impact of background SGLT-2i on treatment effect. Results Baseline characteristics and demographics A total of 32 subjects with T2D, on stable treatment with SGLT-2i with or without metformin or a sulfonylurea, were randomized to receive either semaglutide or placebo; one subject was assigned to semaglutide but did not receive treatment. 3 Baseline characteristics for subjects enrolled in SUSTAIN 9 are shown in Table. 3 Baseline characteristics were similar between the SGLT-2i subgroups (data not shown). Table : Baseline characteristics and demographics Semaglutide. mg (n=5) Glycemic control and body weight Placebo (n=5) Total (N=32) Age, years 57.5 (8.9) 56.6 (.) 57. (9.5) Diabetes duration, years 9.8 (6.3) 9.6 (5.9) 9.7 (6.) HbA c, % 8. (.8) 8. (.8) 8. (.8) FPG, mg/dl 64.2 (38.) 6.2 (39.2) 62.7 (38.6) BMI, kg/m2 3. (6.2) 32.7 (6.9) 3.9 (6.6) egfr, ml/min/.73 m (5.3) 96. (5.) 95.2 (5.2) Data presented as mean (standard deviation). BMI, body mass index; egfr, estimated glomerular filtration rate; FPG, fasting plasma glucose. At 3 weeks, there was no significant interaction between background SGLT-2i and treatment effect for any of the endpoints (interaction p>.5 for all; Figures 2 and 3). Across background SGLT-2i groups, reductions in HbA c (Figure 2A) and body weight (Figure 2B) were significantly greater with semaglutide vs placebo. The proportions of subjects achieving American Diabetes Association and American Association of Clinical Endocrinologists HbA c targets of <7.% and 6.5%, respectively, weight-loss responses of 5%, and the composite endpoint of HbA c <7.% without weight gain or blood glucose-confirmed symptomatic hypoglycemia, were also significantly greater for semaglutide vs placebo, with a similar pattern for all SGLT-2is (Figure 3A C, E). Figure 2: Change from baseline in (A) HbA c and (B) body weight at 3 weeks by SGLT-2i therapy (A) Change in HbA c from baseline Baseline HbAc, % Change in HbA c (%-point) from baseline ETD [95% CI]: semaglutide. mg vs placebo p-value for interaction All subjects Canagliflozin Dapagliflozin Empagliflozin n [.6;.24].4..5 [.89;.3].5.34 [.65;.3] [.58;.97] Semaglutide. mg Key results p<. vs placebo. p-value for interaction of background SGLT-2i on treatment effect. 32 subjects were randomized and 3 received trial medication; one subject was randomized but not included in this analysis due to not receiving SGLT-2i at screening. Includes subjects receiving dapagliflozin or dapagliflozin propanediol monohydrate. 25 subjects received other SGLT-2is and were randomized to receive semaglutide. mg (n=5) or placebo (n=); data not presented due to small number of subjects. On-treatment without rescue medication data analyzed using an analysis of covariance with treatment, subgroup and treatment by subgroup interaction, stratification factor, and region as fixed factors, and baseline value as covariate. Mean baseline values and mean changes from baseline are shown. BW, body weight; CI, confidence interval; ETD, estimated treatment difference; SGLT-2i, sodium glucose co-transporter-2 inhibitor. Figure 3: Proportion of subjects achieving, at 3 weeks, (A) HbA c <7.%, (B) HbA c 6.5%, (C) weight loss 5%, (D) weight loss %, (E) triple composite of HbA c <7.% without weight gain and no severe or blood glucose-confirmed symptomatic hypoglycemia (A) HbA c <7.% (ADA) Baseline HbA c, % Subjects achieving HbAc <7.% (%) OR [95% CI]: semaglutide. mg vs placebo p-value for interaction (C) Weight loss 5% Subjects achieving weight loss 5% (%) n All subjects Canagliflozin Dapagliflozin Empagliflozin n Baseline BW, kg OR [95% CI]: semaglutide. mg vs placebo p-value for interaction [2.8;58.3] [8.4;68.49] [9.5;8.93] [4.95;42.] All subjects Canagliflozin Dapagliflozin Empagliflozin [5.77;26.84] (E) Triple composite endpoint n Baseline HbA c, % Baseline BW, kg Subjects achieving triple composite endpoint (%) OR [95% CI]: semaglutide. mg vs placebo p-value for interaction [2.78;5.85] [3.95;4.9] [3.9;37.7] All subjects Canagliflozin Dapagliflozin Empagliflozin [8.39;32.85] [4.4;63.53] 6.55 [5.4;5.7] [3.96;3.] (B) Change in body weight from baseline Baseline BW, kg. Change in BW (kg) from baseline ETD [95% CI]: semaglutide. mg vs placebo p-value for interaction n Placebo Baseline HbAc, % Subjects achieving HbA c 6.5% (%) All subjects Canagliflozin Dapagliflozin Empagliflozin OR [95% CI]: semaglutide. mg vs placebo p-value for interaction [ 4.7; 2.93] (B) HbA c 6.5% (AACE) [5.63;2.52] (D) Weight loss % Subjects achieving weight loss % (%) [ 7.; 3.47] 63.6 p<. vs placebo; p<. vs placebo. p-value for interaction of background SGLT-2i on treatment effect. 32 subjects were randomized and 3 received trial medication; one subject was randomized but not included in this analysis due to not receiving SGLT-2i at screening. Includes subjects receiving dapagliflozin or dapagliflozin propanediol monohydrate. 25 subjects received other SGLT-2is and were randomized to receive semaglutide. mg (n=5) or placebo (n=); data not presented due to small number of subjects. On-treatment without rescue medication data analyzed using an analysis of covariance with treatment, subgroup and treatment by subgroup interaction, stratification factor, and region as fixed factors, and baseline value as covariate. AACE, American Association of Clinical Endocrinologists; ADA, American Diabetes Association; BW, body weight, CI, confidence interval; NE, not estimable; OR, odds ratio; SGLT-2i, sodium glucose co-transporter-2 inhibitor [ 4.9;.97] [5.78;235.7] 6.2 [4.39;59.89] [ 5.37; 2.4] n All subjects Canagliflozin Dapagliflozin Empagliflozin n Baseline BW, kg OR [95% CI]: semaglutide. mg vs placebo p-value for interaction [.5;444.8] All subjects Canagliflozin Dapagliflozin Empagliflozin [2.8;36.59] <. 5.8 [.8;3.75] Semaglutide. mg Placebo 5.5 [.78;32.53].7465 NE The proportion of subjects achieving a weight-loss response of % was significantly greater for semaglutide vs placebo for all subjects, but there was no significant difference between the treatment arms when analyzed by individual SGLT-2i group (Figure 3D). Results for subjects who received other SGLT-2is (n=25; data not shown) were similar to those for subjects receiving canagliflozin, dapagliflozin, or empagliflozin. Safety No new safety concerns were identified when adding semaglutide to SGLT-2i therapy. Reported adverse events were similar between the SGLT-2i subgroups (Table 2). Results for subjects who received other SGLT-2is (n=25; data not shown) were similar to those for subjects receiving canagliflozin, dapagliflozin, or empagliflozin. Table 2: Adverse events by background SGLT-2i Subjects reporting AEs, n (%) Discussion This post hoc analysis of the SUSTAIN 9 trial shows that, when once-weekly semaglutide is used as an add-on to SGLT-2i therapy, no interaction is observed between the individual SGLT-2is and treatment effect. This indicates that the same efficacy can be expected with semaglutide, regardless of background SGLT-2i. Reductions in HbA c and body weight, compared with placebo, were significant and clinically relevant. These findings are consistent with those of previous SUSTAIN trials. 4 Conclusion All subjects Canagliflozin Dapagliflozin Empagliflozin Semaglutide. mg (n=5) Placebo (n=5) Semaglutide. mg (n=39) Placebo (n=29) Semaglutide. mg (n=43) Placebo (n=62) Semaglutide. mg (n=52) Placebo (n=5) All AEs 4 (69.3) 9 (6.3) 34 (87.2) 2 (72.4) 29 (67.4) 33 (53.2) 35 (67.3) 32 (64.) Serious AEs 7 (4.7) 6 (4.) 3 (7.7) (3.4) (2.3) (.6) 3 (5.8) 4 (8.) AEs leading to premature treatment discontinuation 3 (8.7) 3 (2.) 4 (.3) (3.4) 3 (7.) 2 (3.2) 5 (9.6) Gastrointestinal AEs 56 (37.3) (3.2) 4 (35.9) 4 (3.8) 8 (4.9) 8 (2.9) 2 (4.4) 6 (2.) Severe or BG-confirmed symptomatic hypoglycemia 4 (2.7) 2 (5.) (2.3) (.9) ADA classified, including hypoglycemia episodes classified as severe, documented symptomatic, asymptomatic, probable symptomatic, and pseudo-hypoglycemia. 25 subjects who received other SGLT-2is were randomized to receive semaglutide. mg (n=5) or placebo (n=); data not presented due to small number of subjects. Includes subjects receiving dapagliflozin or dapagliflozin propanediol monohydrate. AEs include events that had onset, or increase in severity, from first exposure to the planned follow-up visit scheduled 5 weeks (with a 7-day visit window) after end-of-treatment visit at 3 weeks (on-treatment data). Severe or BG-confirmed hypoglycemia defined as a plasma glucose value below 3. mmol/l (55.8 mg/dl) with symptoms consistent with hypoglycemia. All AEs coded using the most recent version of the Medical Dictionary for Regulatory Activities. n, number of subjects experiencing event; %, percentage of subjects experiencing event; ADA, American Diabetes Association; AE, adverse event; BG, blood glucose; SGLT-2i, sodium glucose co-transporter-2 inhibitor. In this post hoc analysis of SUSTAIN 9, in subjects with T2D already receiving an SGLT-2i, once-weekly semaglutide resulted in superior HbA c and body weight reductions vs placebo; effects were consistent across SGLT-2i subgroups. No new safety concerns were identified. This trial was sponsored by Novo Nordisk and is registered with ClinicalTrials.gov (NCT38633). Presenter Bernard Zinman has received grants from Novo Nordisk, and has received personal fees from AstraZeneca, Boehringer Ingelheim, Eli Lilly, Janssen, Merck, Novo Nordisk, and Sanofi. The authors acknowledge Emre Yildirim, Novo Nordisk, for review of and input to the poster, and the medical writing assistance of David Jenkins, PhD, and Ellen Robertshaw, PhD, AXON Communications (supported by Novo Nordisk). Presented at the 79th Scientific Sessions of the American Diabetes Association, June 7, 9, San Francisco, CA, USA. References: () Davies MJ et al. Diabetes Care 8;4:2669 7; (2) No authors listed. Diabetes Care 9;42 (Suppl ):S S2; (3) Zinman B et al. Lancet Diabetes Endocrinol 9;7:356 67; (4) Sorli C et al. Lancet Diabetes Endocrinol 7;5:25 6; (5) Ahrén B et al. Lancet Diabetes Endocrinol 7;5:34 54; (6) Ahmann AJ et al. Diabetes Care 8;4:258 66; (7) Aroda VR et al. Lancet Diabetes Endocrinol 7;5:355 66; (8) Rodbard HW et al. J Clin Endocrinol Metab 8;3:229 3; (9) Marso SP et al. N Engl J Med 6;375:834 44; () Pratley RE et al. Lancet Diabetes Endocrinol 8;6: Abstract

47 Semaglutide once weekly Poster presentation 995-P Semaglutide in Patients with Type 2 Diabetes: Real-World Analysis in the Canadian LMC Diabetes REgistry: The SPARE Study Ruth E. Brown, PhD; Aiden R. Liu 2, MSc; Reena Mahbubani 2, RPh, MPharm, BSc.; Ronnie Aronson, MD, FRCPC, FACE LMC Diabetes & Endocrinology, Toronto, ON, Canada; 2 Novo Nordisk Canada Inc., Toronto, ON, Canada Objective To evaluate real-world outcomes in adults with type 2 diabetes (T2D) who had initiated the once-weekly glucagon-like peptide receptor agonist (GLP- RA) semaglutide as part of usual clinical therapy Semaglutide has been marketed in Canada since March 8. Methods Retrospective cohort analysis of the LMC Diabetes Registry with Primary Outcome of change in AC at follow-up (3 to 6 months): Key Inclusion criteria: Adult patients with T2D with first prescription of semaglutide in 8 GLP- RA naïve for > 6 months before the index date baseline AC and follow-up AC An on-treatment analysis was used for the main analysis Results Participant Characteristics Baseline Characteristics of the Semaglutide Cohort (N=887) Age 58. ±.2 Male gender, N (%) 52 (58.7) Duration of T2D (years).7 ± 7.4 AC (%) 8.36 ±.4 Weight (kg) 99.4 ± 22.7 BMI (kg/m 2 ) 34.6 ± 7.2 Diabetes therapies during the pre-index period, N (%) Metformin 793 (89.4) Sulfonylurea 36 (4.6) SGLT2 inhibitor 568 (64.) DPP-4 inhibitor 537 (6.5) Insulin 378 (42.6) #oral therapies for diabetes.9 ±.9 Note: 6.5% of the cohort discontinued a DPP-4i on the index date Results AC and Weight Outcomes AC at Baseline and Follow-up in Full Cohort and Pre-defined Subgroups Median follow-up time was 4.4 months 3.6% of the cohort discontinued semaglutide during the three to six month follow-up period statistically significant change from baseline (p<.5); OHA = oral antihyperglycemic agent 44.% and 77.4% of the cohort achieved AC <7.% and <8.%, respectively 66.3% and 43.% of the cohort achieved AC reductions of.5% and.%, respectively Weight Change from Baseline in Full Cohort and Pre-defined Subgroups statistically significant change from baseline (p<.5); OHA = oral antihyperglycemic agent 29.5% and 5.6% of the cohort achieved a weight loss of 5% and %, respectively Results Secondary and Exploratory Outcomes Baseline and 3 to 6 Month Change in Secondary Clinical Outcomes n Baseline Change P Fasting glucose (mmol/l) ± ± 2.9 <. BMI (kg/m2) ± ±.3 <. SBP (mmhg) ± ± 2.8 <. DBP (mmhg) ± ± 8.9 <. ALT (U) ± ± 2.7 <. The semaglutide cohort had significant reductions in fasting glucose, BMI, systolic blood pressure, diastolic blood pressure, and ALT at follow-up Self-Reported Hypoglycemia at Baseline and 3 to 6 Month Follow-up On-treatment cohort (n = 732) Baseline Follow-up P incidence of any weekly hypoglycemia 4.% 5.3% NS incidence of severe hypoglycemia (annualized).4%.2% NS Insulin users (n = 293) incidence of any weekly hypoglycemia 7.9% 9.6% NS incidence of severe hypoglycemia (annualized).7% 2.4% NS Conclusion In a real-world specialist practice setting, GLP- RA naïve patients who initiated semaglutide had significant improvements in AC and body weight at follow-up, regardless of concomitant diabetes therapy Disclosures: This study was funded by Novo Nordisk Canada Author Correspondence: Ruth Brown, ruth.brown@lmc.ca Abstract

48 Semaglutide once weekly Poster presentation 998-P 998-P Efficacy and safety of semaglutide by baseline BMI in SUSTAIN 5 and 7 Adie Viljoen ; Juan P. Frias 2 ; Theis Gondolf 3 ; Jeff Unger 4 Borthwick Diabetes Research Centre, Lister Hospital, Stevenage, UK ( adie.viljoen@nhs.net; Tel: ); 2 National Research Institute, Los Angeles, CA, USA; 3 Novo Nordisk A/S, Søborg, Denmark; 4 Catalina Research Institute LLC, Montclair, CA, USA. (qrs.ly/mo9xx3r) Aim Semaglutide (Novo Nordisk, Denmark) is a glucagon-like peptide- (GLP-) analog for the once-weekly treatment of type 2 diabetes (T2D). The efficacy and safety of semaglutide were evaluated in the SUSTAIN clinical development program, which covered the continuum of care in T2D, including in drug-naïve subjects and those on background medication with oral antidiabetic drugs±insulin. 2 7 Across the SUSTAIN trials, semaglutide showed superior reductions in HbA c and body weight vs placebo and all active comparators (sitagliptin, exenatide extended release, insulin glargine, dulaglutide), and enabled a greater proportion of subjects to achieve clinically meaningful ( 5%) weight-loss responses. 2 7» A higher body mass index (BMI) at baseline was generally associated with greater weight loss during semaglutide therapy. 8,9 As exposure to a drug may be affected by body weight, this post hoc analysis aimed to evaluate if reductions in HbA c were affected by baseline BMI in the SUSTAIN trials. Methods SUSTAIN 5 and 7 trial designs In SUSTAIN 5 and 7, adults with T2D (HbA c 7..% for SUSTAIN, 4, and 5, or 7..5% for SUSTAIN 2, 3, and 7) were randomized to semaglutide.5 mg or. mg, placebo or active comparator (Figure ). 2 7 Figure : SUSTAIN 5 and 7 trials SUSTAIN 2 3 (N=,23; 56 w) vs sitagliptin MET, TZD, MET/TZD All semaglutide-treated subjects followed a fixed dose-escalation regimen: 2 7» The semaglutide.5 mg maintenance dose was reached after 4 weeks of semaglutide.25 mg once weekly; the semaglutide. mg maintenance dose was reached after an additional 4 weeks of semaglutide.5 mg once weekly. Key endpoints were similar across trials: 2 7» The primary endpoint was the change in HbA c from baseline to end of treatment.» The confirmatory secondary endpoint was the change in body weight from baseline to end of treatment. Post hoc analysis For this post hoc analysis of data from the SUSTAIN 5 and 7 trials:» Subjects were grouped by baseline BMI (<25, 25 to <3, 3 to <35, and 35 kg/m 2 ).» Change in HbA c was evaluated for semaglutide vs placebo or active comparator by trial for SUSTAIN 5 and 7 in a mixed model for repeated measurements, with treatment, BMI subgroup, and HbA c at baseline as covariates, and interaction between treatment and BMI subgroups at baseline.» Safety data were pooled and analyzed by a Cochran Mantel Haenszel analysis stratified by trial. Results Baseline characteristics and demographics Baseline characteristics were broadly consistent across SUSTAIN 5 and 7, with mean baseline HbA c and body weight values ranging from % and kg, respectively (Table ). Diabetes duration at baseline ranged from 4.2 to 3.3 years, reflecting the continuum of T2D care covered by the SUSTAIN trials (Table ). Table : Baseline characteristics and demographics by trial SUSTAIN 2 vs placebo SUSTAIN 2 3 vs sitagliptin SUSTAIN 3 4 vs exenatide ER SUSTAIN 4 5 vs IGlar SUSTAIN 5 6 vs placebo SUSTAIN 7 7 vs dulaglutide Figure 2: Estimated change in HbA c by baseline BMI Key result Change in HbA c from baseline (%-point) All subjects <25 25 to <3 3 to <35 35 SUSTAIN vs placebo (drug-naïve) BL HbAc (%) = n= SUSTAIN 2 vs sitagliptin (background: MET, TZD, MET/TZD) BL HbAc (%) = n= SUSTAIN 3 vs exenatide ER (background: 2 OADs) BL HbAc (%) = n= SUSTAIN 4 vs IGlar (background: MET, MET/SU) BL HbAc (%) = n= SUSTAIN 5 vs placebo (background: add-on to basal insulin±met) BL HbAc (%) = n= Safety In all treatment arms, adverse events (AEs) occurred in similar proportions of subjects across BMI subgroups (Table 2). The proportion of subjects with gastrointestinal AEs was higher with semaglutide vs placebo or active comparators; however, these events decreased with increasing baseline BMI. Premature treatment discontinuation due to AEs:» Decreased with increasing baseline BMI, potentially reflecting the trend in gastrointestinal AEs.» Was higher in all BMI subgroups with semaglutide vs placebo or active comparators. Table 2: Adverse events by baseline BMI Discussion Semaglutide.5 mg Semaglutide. mg Comparator Baseline BMI subgroup, kg/m2 <25 25 to <3 3 to <35 35 <25 25 to <3 3 to <35 35 <25 25 to <3 3 to <35 35 AEs 89 (7.8) 249 (69.) 284 (7.6) 33 (7.) 96 (72.2) 337 (69.9) 375 (68.9) 43 (72.) (66.) 365 (67.) 448 (69.9) 462 (68.6) Serious AEs 5 (3.9) 9 (5.2) 3 (7.7) 29 (6.4) 4 (3.6) 29 (6.) 36 (6.6) 58 (.2) 6 (3.4) 26 (5.2) 46 (6.9) 53 (7.6) Gastrointestinal AEs 59 (47.8) 55 (43.2) 55 (38.8) 7 (38.4) 65 (5.5) (43.7) 22 (39.) 228 (39.9) 4 (2.2) 52 (25.8) 9 (28.9) 84 (25.5) AEs leading to premature treatment discontinuation 6 (3.5) 38 (.5) 22 (5.7) 5 (3.4) 23 (8.) 44 (9.2) 42 (7.7) 39 (6.9) 5 (8.3) 29 (4.6) 24 (3.7) 5 (2.3) On-treatment data for number (%) of subjects in the safety analysis set experiencing at least one event. Comparators were: placebo (SUSTAIN and 5), sitagliptin (SUSTAIN 2), exenatide extended release (SUSTAIN 3), insulin glargine (SUSTAIN 4), and dulaglutide (SUSTAIN 7). AE, adverse event; BMI, body mass index. Achieving glycemic control in T2D is challenging, and responsiveness to therapy is important. Background medication SUSTAIN 2 (N=388; 3 w) vs placebo Monotherapy Drug-naïve SUSTAIN 3 4 (N=83; 56 w) vs exenatide ER 2 OADs SUSTAIN 4 5 (N=,89; 3 w) vs insulin glargine MET, MET/SU SUSTAIN 7 7 (N=,; 4 w) vs dulaglutide Add-on to MET Insulin SUSTAIN 6 was a cardiovascular outcomes trial in subjects at high risk of CV disease and, as such, was not included in the present analysis. Exenatide ER, exenatide extended release; MET, metformin; N, number of randomized subjects; OAD, oral antidiabetic drug; SU, sulfonylurea; TZD, thiazolidinedione; w, weeks. OAD SUSTAIN 5 6 (N=397; 3 w) vs placebo Add-on to basal insulin±met Age, years 53.7 (.3) 55. (.) 56.6 (.7) 56.5 (.4) 58.8 (.) 56. (.6) Diabetes duration, years 4.2 (5.5) 6.6 (5.) 9.2 (6.3) 8.6 (6.3) 3.3 (7.8) 7.4 (5.7) % 8. (.9) 8. (.9) 8.3 (.) 8.2 (.9) 8.4 (.8) 8.2 (.9) c mmol/mol 64.5 (9.3) 64.8 (.) 67.7 (.4) 65.8 (9.7) 67.9 (9.2) 66.4 (.) Body weight, kg 9.9 (23.8) 89.5 (.3) 95.8 (2.5) 93.5 (2.8) 9.7 (2.) 95.2 (22.6) BMI, kg/m (7.7) 32.5 (6.2) 33.8 (6.7) 33. (6.5) 32.2 (6.2) 33.5 (6.8) Data are mean (SD). BMI, body mass index; exenatide ER, exenatide extended release; IGlar, insulin glargine; SD, standard deviation. Glycemic control Reductions in mean HbA c (%) from baseline were generally greater with semaglutide vs placebo or active comparator in all BMI subgroups (Figure 2).» The only exception was the <25 kg/m 2 BMI subgroup for semaglutide.5 mg vs insulin glargine (.7% vs.9%, respectively) and semaglutide.5 mg vs dulaglutide.75 mg (.4% vs.6%, respectively). There were no significant interactions between treatment and BMI, with the exception of semaglutide.5 mg in SUSTAIN SUSTAIN 7 vs dulaglutide (background: add-on to MET) BL HbAc (%) = n= Semaglutide.5 mg Semaglutide. mg Placebo Sitagliptin Exenatide ER IGlar Dulaglutide.75 mg Dulaglutide.5 mg Values shown are estimated mean changes from baseline for subjects on treatment without rescue medication. BL, baseline; BMI, body mass index; exenatide ER, exenatide extended release; IGlar, insulin glargine; MET, metformin; OAD, oral antidiabetic drug; SU, sulfonylurea; TZD, thiazolidinedione. In this post hoc analysis of the SUSTAIN 5 and 7 trials, the estimated treatment differences in mean HbA c for semaglutide vs placebo or active comparators did not appear to be influenced by baseline BMI, indicating a consistent effect of semaglutide. Reductions in mean HbA c from baseline were generally greater with semaglutide vs placebo or active comparators in all BMI subgroups. AEs occurred in a similar proportion of subjects in all treatment arms and across BMI subgroups. Gastrointestinal AEs decreased with increasing BMI. Conclusion Semaglutide consistently reduced HbA c vs placebo or active comparators in subjects with T2D regardless of their baseline BMI. Semaglutide had an acceptable safety profile in all BMI subgroups. These trials were sponsored by Novo Nordisk and are registered with ClinicalTrials.gov (NCT54897; NCT9388; NCT8858; NCT228932; NCT23538; NCT26484). Presenter Adie Viljoen has received grants from Sanofi, consulting fees from AstraZeneca, Boehringer, Eli Lilly, MSD, Napp, Novo Nordisk, and Sanofi, and has received research support from AstraZeneca, Amgen, Eli Lilly, Regeneron, Novartis, Novo Nordisk, and Sanofi. The authors are grateful to Claus Dethlefsen, Novo Nordisk, for review of and input to the poster. The authors acknowledge the medical writing assistance of Heather Mittman, AXON Communications (supported by Novo Nordisk). Presented at the 79th Scientific Sessions of the American Diabetes Association, June 7, 9, San Francisco, CA, USA. References: () Lau J et al. J Med Chem 5;58:737 8; (2) Sorli C et al. Lancet Diabetes Endocrinol 7;5:25 6; (3) Ahrén B et al. Lancet Diabetes Endocrinol 7;5:34 54; (4) Ahmann AJ et al. Diabetes Care 8;4:258 66; (5) Aroda VR et al. Lancet Diabetes Endocrinol 7;5:355 66; (6) Rodbard HW et al. J Clin Endocrinol Metab 8;3:229 3; (7) Pratley RE et al. Lancet Diabetes Endocrinol 8;6:275 86; (8) Leiter LA et al. Presented at the 77th Scientific Sessions of the American Diabetes Association, June 9 3, 7, San Diego, CA, USA. Poster 5-P; (9) Viljoen A et al. Presented at the 78th Scientific Sessions of the American Diabetes Association, June 22 26, 8, Orlando, FL, USA. Poster 83-P; () Carlsson Petri KC et al. Diabetes Ther 8;9:533 47; () Fonseca VA et al. Diabetes Care 9;32:S5 6. Abstract

49 Semaglutide once weekly Poster presentation -P -P Efficacy and safety of once-weekly semaglutide low dose.5 mg vs once-weekly dulaglutide high dose.5 mg in type 2 diabetes: a post hoc analysis of SUSTAIN 7 Richard E. Pratley ; Vanita R. Aroda 2 ; Theis Gondolf 3 ; Thomas Hansen 3 ; Ildiko Lingvay 4 ; Jörg Lüdemann 5 ; Trine Vang Skjøth 3 ; Adie Viljoen 6 AdventHealth Orlando, Translational Research Institute for Metabolism and Diabetes, Orlando, FL, USA ( richard.pratley.md@adventhealth.com; Tel: ); 2 Brigham and Women s Hospital, Boston, MA, USA; 3 Novo Nordisk A/S, Søborg, Denmark; 4 University of Texas Southwestern Medical Center, Dallas, TX, USA; 5 diabetes-falkensee, Diabetes Centre and Centre for Clinical Studies, Falkensee, Germany; 6 Lister Hospital, Stevenage, UK. (qrs.ly/mo9xx3r) Aim Semaglutide (Novo Nordisk, Denmark) is a glucagon-like peptide- receptor agonist indicated as an adjunct to diet and exercise to improve glycemic control in adults with type 2 diabetes (T2D). In SUSTAIN 7, an international, open-label, parallel-group trial, adults with inadequately controlled T2D were randomized (:::) to receive subcutaneous semaglutide once weekly or dulaglutide once weekly at low (.5 vs.75 mg, respectively) or high (. vs.5 mg, respectively) doses. 2» Semaglutide provided superior glycemic control and reductions in body weight vs dulaglutide at both low- and high-dose drug comparisons. 2 This post hoc analysis aimed to compare the effects of semaglutide low (.5 mg) vs dulaglutide high (.5 mg) dose at week 4.» This comparison was not prespecified in the primary analysis of SUSTAIN 7. 2» This comparison was implemented to reflect options available in clinical practice and to ensure a thorough assessment of clinical efficacy and safety. Methods SUSTAIN 7 trial design (Figure ) Data were collected from all patients randomized and exposed to treatment (full analysis set), and the data analyzed in this post hoc analysis were prior to use of any rescue medication. Baseline characteristics were broadly consistent between semaglutide and dulaglutide (Table ).» Mean age was 56 years. Baseline HbA c was %; diabetes duration years. Figure : SUSTAIN 7 trial design 2, patients with T2D Age 8 years HbA c 7..5% Stable treatment with metformin (,5 mg/day or MTD) for 9 days prior to screening egfr 6 ml/min/.73 m 2 Randomization (:::) Semaglutide.5 mg Dulaglutide.75 mg Semaglutide. mg Dulaglutide.5 mg Dose escalation 4 8 weeks Treatment maintenance weeks Treatment duration 4 weeks Follow-up 5 weeks Prespecified comparison Prespecified comparison Current post hoc comparison Semaglutide dose escalation from starting dose of.25 mg once weekly, dose doubled every 4 weeks until trial maintenance dose achieved. Dulaglutide.75 mg and.5 mg dosed once weekly without dose escalation. egfr, estimated glomerular filtration rate; MTD, maximum tolerated dose. Results Table : Baseline characteristics and demographics Change in HbA c and body weight from baseline, and proportions of subjects achieving targets Low-dose semaglutide.5 mg resulted in similar improvements in glycemic control and significantly greater weight loss vs high-dose dulaglutide.5 mg at week 4 (Figure 2). Similar proportions of subjects achieved HbA c <7.% and 6.5%, but with low-dose semaglutide.5 mg, significantly larger proportions of subjects achieved weight loss 5% and % compared with high-dose dulaglutide.5 mg (Figure 3). Systolic and diastolic blood pressure There were no significant differences in change from baseline in systolic and diastolic blood pressure between low-dose semaglutide.5 mg and high-dose dulaglutide.5 mg (Table 2). Safety Low-dose semaglutide.5 mg (n=3) High-dose dulaglutide.5 mg (n=299) Male, n 69 (56.%) 7 (57.2%) Female, n 32 (43.9%) 28 (42.8%) Age, years 56 (.9) 56 (.6) Diabetes duration, years 7.7 (5.9) 7.6 (5.6) % 8.3 (.) 8.2 (.9) mmol/mol 67.5 (.5) 66. (9.7) SBP, mmhg 34 (4.8) 32 (3.6) DBP, mmhg 8 (9.) 8 (8.7) Body weight, kg 96.4 (24.4) 93.4 (2.8) BMI, kg/m (7.) 33. (6.6) FPG mg/dl 76.3 (45.7) 72.5 (4.2) mmol/l 9.8 (2.5) 9.6 (2.3) Values are mean (SD) unless otherwise indicated. BMI, body mass index; DBP, diastolic blood pressure; FPG, fasting plasma glucose; SBP, systolic blood pressure; SD, standard deviation. The rate of adverse events (AEs) and serious AEs overall was similar for low-dose semaglutide.5 mg and high-dose dulaglutide.5 mg (Table 3). The rate of gastrointestinal (GI) AEs was similiar for low-dose semaglutide.5 mg and high-dose dulaglutide.5 mg; the GI AEs were mainly mild or moderate for both treatment groups (Table 3). The proportions of subjects discontinuing treatment due to any AE or due to GI AEs were similar for low-dose semaglutide.5 mg and high-dose dulaglutide.5 mg (Table 3). Figure 2: Change from baseline in HbA c and body weight at week 4 Proportion of subjects (%) Change in HbA c from baseline (%) HbA c Body weight Low-dose High-dose Low-dose High-dose Semaglutide semaglutide.5 mg Dulaglutide dulaglutide.5 mg Semaglutide semaglutide.5 mg Dulaglutide dulaglutide.5 mg Low-dose Semaglutide semaglutide.5 mg mg ETD [95% CI]:.4 [.3;.] p=.76 High-dose Dulaglutide dulaglutide.5 mg mg.4 CI, confidence interval; ETD, estimated treatment difference. Figure 3: Proportion of subjects achieving targets at week Low-dose Semaglutide semaglutide.5 mg High-dose Dulaglutide dulaglutide.5 mg mg Change in body weight from baseline (kg) Proportion of subjects (%) Semaglutide Low-dose semaglutide.5.5 mg mg ETD [95% CI]:.58 [ 2.35;.82] p<. HbA c <7.% HbA c 6.5% Weight loss 5% Weight loss % EOR [95% CI]:.6 [.8;.68] p=.4226 CI, confidence interval; EOR, estimated odds ratio. EOR [95% CI]:.7 [.78;.73] p=.459 EOR [95% CI]:.6 [.2;2.29] p=.93 3 Dulaglutide High-dose dulaglutide.5.5 mg mg 3. 4 Semaglutide Low-dose semaglutide.5.5 mg mg Key result EOR [95% CI]: 2. [.5;3.43] p=.38 8 High-dose Dulaglutide dulaglutide.5.5 mg mg Table 2: Change from baseline in systolic and diastolic blood pressure at week 4 SBP DBP Low-dose semaglutide.5 mg High-dose dulaglutide.5 mg Baseline, mmhg 34 (4.8) 32 (3.6) Change from baseline at week 4, mmhg 2.4 (.8) 2.9 (.8) ETD [95% CI].42 [.68;2.52] p-value.697 Baseline, mmhg 8 (9.) 8 (8.7) Change from baseline at week 4, mmhg.6 (.5). (.5) ETD [95% CI].54 [.86;.79] p-value.426 Values are mean (SD) unless otherwise indicated. CI, confidence interval; DBP, diastolic blood pressure; ETD, estimated treatment difference; SBP, systolic blood pressure; SD, standard deviation. Table 3: Adverse events by treatment Discussion In the original SUSTAIN 7 study, semaglutide was superior to dulaglutide at both low- and high-dose drug comparisons in improving glycemic control and reducing body weight, with a similar safety profile in subjects with T2D. In this post hoc analysis, the comparison of low-dose semaglutide.5 mg vs high-dose dulaglutide.5 mg showed similar glycemic control, but with more weight loss and more subjects achieving 5% and % weight loss with semaglutide low dose.5 mg vs dulaglutide high dose.5 mg after 4 weeks, with a similar safety profile in subjects with T2D. These results suggest that the low dose of semaglutide.5 mg can provide greater weight-loss benefit to patients with T2D vs high-dose dulaglutide.5 mg, alongside similar glycemic control and a similar safety profile. Conclusion Low-dose semaglutide.5 mg High-dose dulaglutide.5 mg n % E R n % E R Overall AEs Serious AEs Fatal events < AEs leading to premature treatment discontinuation GI AEs leading to premature treatment discontinuation GI AEs Mild Moderate Severe Most frequent GI AEs Nausea Diarrhea Vomiting AEs include events that had an onset, or increase in severity, from first exposure to the planned follow-up visit scheduled 5 weeks (+7-day visit window) after the end of treatment visit at week 4 (on-treatment data). AEs include events that had an onset, or increase in severity, from randomization to the end of trial regardless of treatment or rescue medication status (in-trial data). One subject receiving dulaglutide.5 mg had four events resulting in a fatal outcome. AE, adverse event; E, events; GI, gastrointestinal; R, rate of events per patient-years. In summary, semaglutide low dose.5 mg showed greater weight loss and similar improvements in glycemic control vs dulaglutide high dose.5 mg at week 4, with a similar safety profile in subjects with T2D, previously uncontrolled on metformin treatment. This trial was sponsored by Novo Nordisk and is registered with ClinicalTrials.gov (NCT26484). Presenter Richard Pratley has received consulting fees from AstraZeneca, GlaxoSmithKline, Glytec LLC, Janssen, Ligand Pharmaceuticals, Lilly, Merck, Mundipharma, Novo Nordisk, Pfizer, and Sanofi; has received grants from Lexicon Pharmaceuticals, Ligand Pharmaceuticals, Lilly, Merck, Novo Nordisk, and Sanofi; and lecture and other fees from AstraZeneca and Novo Nordisk. The authors acknowledge the medical writing assistance of Heather Mittman, AXON Communications (supported by Novo Nordisk). Presented at the 79th Scientific Sessions of the American Diabetes Association, June 7, 9, San Francisco, CA, USA. References: () Novo Nordisk. Ozempic (semaglutide) Prescribing Information 7. Available at: drugsatfda_docs/label/7/9637lbl.pdf. Accessed April 9; (2) Pratley RE et al. Lancet Diabetes Endocrinol 8;6: Abstract

50 Semaglutide once weekly Poster presentation 3-P 3-P Clinical response (HbA c % and/or body weight 5% reduction) to semaglutide by baseline HbA c and body weight in the SUSTAIN program Vanita R. Aroda ; Julie R. Larsen 2 ; Signe Harring Østoft 2 ; Rosangela R. Rea 3 ; Danny Sugimoto 4 ; Uwe Ziegler 2 ; Tina Vilsbøll 5 Brigham and Women s Hospital, Boston, MA, USA ( varoda@bwh.harvard.edu; Tel: ); 2 Novo Nordisk A/S, Søborg, Denmark; 3 Hospital de Clínicas da Universidade Federal do Paraná, Paraná, Brazil; 4 Cedar Crosse Research Center, Chicago, IL, USA; 5 Steno Diabetes Center Copenhagen, Gentofte; University of Copenhagen, Hellerup, Denmark (qrs.ly/mo9xx3r) (qrs.ly/mo9xx3r) Scan for more Novo Nordisk scientific materials Background The SUSTAIN clinical development program assessed subcutaneous semaglutide once weekly, a glucagon-like peptide- (GLP-) analog, across the continuum of type 2 diabetes (T2D) care, including drug-naïve subjects and those on a background of oral glucose-lowering drugs and/ or insulin. 6 In the SUSTAIN 5 and 7 trials:» Semaglutide demonstrated superior reductions in HbA c (Figure A) and body weight (Figure B) vs placebo and comparators. 7» Significantly greater proportions of subjects treated with semaglutide (.5 mg and. mg) vs placebo or comparators achieved clinically meaningful outcomes for the composite outcome of HbA c reduction.% and body weight loss 5.% (Figure C). 8 Figure : Change from baseline in (A) HbA c and (B) body weight, and (C) proportion of subjects achieving composite endpoint of HbA c reduction.% and body weight loss 5.% Background: Treatment duration: (A) (B) (C) Baseline HbA c (%): (mmol/mol): Baseline BW (kg): Change in HbAc from baseline (%) Change in BW from baseline (kg) Proportion of subjects achieving composite endpoint (%) SUSTAIN vs placebo (N=388) Treatment-naïve SUSTAIN 2 vs sitagliptin (N=,23) SUSTAIN 3 vs exenatide ER (N=83) SUSTAIN 4 vs IGlar (N=,89) SUSTAIN 5 vs placebo (N=397) SUSTAIN 7 vs dulaglutide (N=,) MET±TZD 2 OADs MET±SU Add-on to insulin±met MET, MET/SU 3 weeks 56 weeks 56 weeks 3 weeks 3 weeks 4 weeks Semaglutide.5 mg Semaglutide. mg Placebo Sitagliptin mg p<. vs comparator. In SUSTAIN 7, comparisons were semaglutide.5 mg vs dulaglutide.75 mg, and semaglutide. mg vs dulaglutide.5 mg. p<. vs semaglutide.5 mg. SUSTAIN 7 results based on an exploratory analysis. BW, body weight; exenatide ER, exenatide extended release; IGlar, insulin glargine; MET, metformin; OAD, oral antidiabetic drug; SU, sulfonylurea; TZD, thiazolidinedione Exenatide ER 2. mg IGlar Dulaglutide.75 mg Dulaglutide.5 mg Aim The aim of this post hoc analysis was to assess the proportion of semaglutide-treated subjects achieving clinically relevant responses (composite endpoint:.% HbA c reduction and 5.% body weight loss; individual components:.% HbA c or 5.% body weight loss) by baseline HbA c and body weight across the SUSTAIN 5 and 7 trials. Methods This analysis was based on pooled data from the semaglutide arms (.5 mg and. mg) across the SUSTAIN 5 and 7 trials (Table ). Subjects achieving clinically relevant responses by week 3 were classified based on:» HbA c reduction.% and body weight loss 5.% (composite endpoint).» HbA c reduction.%.» Body weight loss 5.%. Logistic regression was used to evaluate whether baseline HbA c and body weight were predictive of achieving the composite endpoint and individual components of clinical response. Semaglutide.5 mg and. mg on-treatment without rescue medication data were used with missing values for HbA c or body weight imputed from a mixed model for repeated measurements. Results Baseline characteristics such as age, baseline HbA c, and body weight were comparable across the SUSTAIN trials (Table ). Differences in diabetes duration (and background medication) reflected the continuum of T2D care, ranging from 4.2 years in treatment-naïve subjects (SUSTAIN ) to 3.3 years in insulin-treated subjects (SUSTAIN 5) (Figure, Table ). Table : Baseline characteristics and demographics by trial in subjects receiving semaglutide (treatment groups pooled) SUSTAIN vs placebo (n=258) SUSTAIN 2 vs sitagliptin (n=88) SUSTAIN 3 vs exenatide ER (n=44) SUSTAIN 4 vs IGlar (n=722) Pooled semaglutide treatment arms SUSTAIN 5 vs placebo (n=263) SUSTAIN 7 vs dulaglutide (n=6) Age, years 53.7 (.5) 55.4 (9.8) 56.4 (.3) 56.6 (.4) 58.8 (9.7) 55.8 (.8) T2D duration, years 4.2 (5.5) 6.6 (5.) 9. (6.) 8.6 (6.3) 3.3 (7.7) 7.5 (5.8) HbA c, % 8. (.9) 8. (.9) 8.4 (.) 8.2 (.9) 8.3 (.8) 8.3 (.9) HbA c, mmol/mol 65. (9.3) 64.2 (.) 67.9 (.4) 66. (9.8) 67.6 (9.) 66.8 (.3) Body weight, kg 93.4 (24.5) 89.6 (.6) 96.2 (22.5) 93.9 (2.9) 92.6 (.9) 95.9 (22.7) BMI, kg/m (8.) 32.5 (6.4) 34. (7.2) 33. (6.5) 32.4 (6.2) 33.6 (6.8) SUSTAIN 3 included only semaglutide. mg. Values are mean (SD). BMI, body mass index; exenatide ER, exenatide extended release; IGlar, insulin glargine; SD, standard deviation. Figure 2: Proportion of semaglutide-treated subjects (A) achieving or (B) not achieving the composite endpoint of.% HbA c reduction and 5.% body weight loss by baseline HbA c and body weight (A) Baseline HbA c (%) Dashed lines indicate mean baseline values for HbA c (8.2%/66 mmol/mol) and body weight (93.3 kg). Primary result Proportion achieving a response: 42.% (B) Proportion not achieving a response: 58.% % 8.9% 3 3.% 23.9% Baseline body weight (kg) Figure 3: Proportion of semaglutide-treated subjects (A) achieving or (B) not achieving.% HbA c reduction by baseline HbA c and body weight Baseline HbA c (%) % Figure 4: Proportion of semaglutide-treated subjects (A) achieving or (B) not achieving 5.% body weight loss by baseline HbA c and body weight For Figures 3 and 4: Dashed lines indicate mean baseline values for HbA c (8.2%/66 mmol/mol) and body weight (93.3 kg) Baseline HbAc (mmol/mol) Baseline HbA c (%) % 22.% % 24.2% Baseline body weight (kg) Achieving composite endpoint Not achieving composite endpoint (A) Proportion achieving a response: 76.6% (B) Proportion not achieving a response: 23.4% Baseline HbA c (%) %.8% % Baseline body weight (kg) 24.% Baseline HbAc (mmol/mol) Baseline HbA c (%) % 3 4.3% 34.7% Baseline body weight (kg) Achieving HbAc.% reduction Not achieving HbAc.% reduction (A) Proportion achieving a response: 49.9% (B) Proportion not achieving a response: 5.% 3 34.% 25.2% Baseline body weight (kg) Achieving 5.% weight loss 7.% Baseline HbA c (mmol/mol) Baseline HbA c (%) % Not achieving 5.% weight loss.2% % 22.9% Baseline body weight (kg) 24.4% Baseline HbA c (mmol/mol) Baseline HbAc (mmol/mol) Baseline HbA c (mmol/mol) Mean (standard deviation) baseline for pooled semaglutide.5 mg and. mg across all trials was 8.2% (±.9)/66 mmol/mol (±) for HbA c and 93.3 kg (22.) for body weight. Overall, 42.% (n=,289/3,66) of subjects achieved the composite endpoint of.% HbA c reduction and 5.% body weight loss (Figure 2).» Baseline HbA c and body weight did not affect the likelihood of achieving the composite endpoint. For individual endpoints:» 76.6% of subjects (n=2,349/3,66) achieved.% HbA c reduction (Figure 3).» 49.9% of subjects (n=,53/3,66) achieved 5.% body weight loss (Figure 4). With every % higher baseline HbA c, the odds of achieving a response of.% HbA c reduction increased by 43% (p<.). Discussion Treatment guidelines recommend an individualized treatment approach for people with T2D. 9» Baseline HbA c is a known predictor of magnitude of HbA c reduction with diabetes treatments, including GLP- receptor agonists.» Understanding treatment responses by baseline characteristics in an inherently heterogeneous population with T2D, spanning a wide range of baseline HbA c and body weight values, is important and may facilitate individualized approaches to treatment. In this post hoc analysis, clinically meaningful responses of HbA c reduction.% and body weight loss 5.% were achieved with semaglutide once weekly, across a range of baseline HbA c and body weight values, in subjects with T2D.» Subjects with a relatively high baseline HbA c, and hence poor glycemic control, were more likely to achieve a.% HbA c reduction than those with lower baseline HbA c, likely reflecting the glucose-dependent action of GLP- in combination with the demonstrated efficacy of semaglutide.» Otherwise, the likelihood of clinically meaningful responses with semaglutide was consistent across baseline HbA c and body weight values. Conclusion In this post hoc analysis across the SUSTAIN 5 and 7 trials, semaglutide once weekly provided clinically meaningful responses in composite and individual endpoints of.% HbA c reduction and/or 5.% body weight loss, across a range of baseline HbA c and body weight values in patients across the continuum of T2D care. These trials were sponsored by Novo Nordisk and are registered with ClinicalTrials.gov (NCT54897; NCT9388; NCT8858; NCT228932; NCT23538; NCT26484). Presenter Vanita R. Aroda has served as consultant for Adocia, AstraZeneca, BD, Novo Nordisk, Sanofi, and Zafgen, and has received research support from AstraZeneca/BMS, Calibra, Eisai, Janssen, Novo Nordisk, Sanofi, and Theracos. The presenter s spouse is an employee of Merck Research Laboratories. The authors acknowledge the medical writing assistance of Daniella Pfeifer, PhD, AXON Communications (supported by Novo Nordisk). Presented at the 79th Scientific Sessions of the American Diabetes Association, June 7, 9, San Francisco, CA, USA. References: () Sorli C et al. Lancet Diabetes Endocrinol 7;5:25 6; (2) Ahrén B et al. Lancet Diabetes Endocrinol 7;5:34 54; (3) Ahmann AJ et al. Diabetes Care 8;4:258 66; (4) Aroda VR et al. Lancet Diabetes Endocrinol 7;5:355 66; (5) Rodbard HW et al. J Clin Endocrinol Metab 8;3:229 3; (6) Pratley RE et al. Lancet Diabetes Endocrinol 8;6:275 86; (7) Aroda VR et al. Diabetes Metab 9 Jan 4. doi:.6/j.diabet [Epub ahead of print]; (8) Rodbard HW et al. Endocr Pract 9 Mar 3. doi:.458/ep [Epub ahead of print]; (9) American Diabetes Association. Diabetes Care 8;4(Suppl ):S73 S85; () DeFronzo RA et al. Diabet Med ;27:39 7. Abstract

51 Semaglutide once weekly Poster presentation 6-P Real-World Effectiveness of Semaglutide in Early Users from a US Commercially Insured and Medicare Advantage Population Jay Visaria, Tam Dang-Tan 2, Paul Petraro 2a, Bal Nepal, Vincent Willey HealthCore, Inc., Wilmington, DE, USA; 2 Novo Nordisk, Inc., Plainsboro Township, NJ, USA; a At the time of the study 6-P qrs.ly/mo9xx3r BACKGROUND Methods Results Figure 3 Change in HbAc Figure 4 HbAc Goal Attainment Glucagon-like peptide receptor agonists (GLP- RA) have been marketed in the US since 5 Although effective in lowering blood glucose with low rates of hypoglycemia,,2 the GLP- RA class of medications is not widely used in the US in diabetes treatment 3 Recently, interest in the GLP- RA class has risen with introduction of once weekly formulations and the publication of multiple cardiovascular outcomes trials demonstrating evidence of cardiovascular benefit 4,5,6,7 Once weekly semaglutide subcutaneous (SC) is the most recent GLP- RA approved in the US (Semaglutide SC, December 7) for adults with type 2 diabetes mellitus (T2DM), and as such, real-world evidence of its use is currently lacking Objective To describe real-world once-weekly semaglutide SC use and its effectiveness in a cohort of early users Methods Study Design and Data Source This retrospective, observational cohort study utilized medical and pharmacy claims integrated with outpatient laboratory data from the HealthCore Integrated Research Database (HIRD ) for the period between 2//6 and 9/3/8 (Figure ) o The HIRD contains integrated, longitudinal claims data on over 7 million individuals across 4 geographically diverse health plans across the US going back to 6 Inclusion Criteria Patients with pharmacy prescription claim for semaglutide SC between 2//7 and 6/3/8 o The service date of the first semaglutide SC claim was defined as the index date Continuous eligibility for medical and pharmacy benefits 2 months prior to the index date Presence of claim with a diagnosis for T2DM during the 2-month pre-index continuous eligibility period Presence of HbAc laboratory test result o o Pre-index: measured within 9 days prior to the index date Post-index: measured during the time period starting at least 9 days following the index date to the end of continuously eligible follow-up Subgroups: o o GLP- RA naïve: patients with no claims for any GLP- RA during the 2-month pre-index period GLP- RA naïve with high baseline HbAc: among GLP- RA naïve semaglutide SC initiators with a HbAc >9% Figure Study schema Study Outcomes Patient demographics and baseline comorbidities Changes in HbAc and HbAc target attainment between the pre-index measurement period and post-index measurement period were assessed in all patients, GLP- RA naïve patients, and GLP- RA naïve patients with a pre-index HbAc >9% The latest available HbAc value within the pre-index and post-index measurement periods were considered o Index date -365 days Change in HbAc was defined as the difference between post-index HbAc and pre-index HbAc value o HbAc targets of <7% and <8% were assessed Statistical Analyses Pre-index HbAc measurement Study Period Index date - 9 days Index date + 9 days 2 months pre-index eligibility (baseline period) diagnosis of type 2 diabetes Index Date (Date of first semaglutide SC claim) Descriptive statistics including means (standard deviation [SD]) and relative frequencies for continuous and categorical data, respectively, were reported The level of statistical significance was set to.5 a priori Post-index HbAc measurement Patient identification period 2//6 2//7 6/3/8 9/3/8 Paired t-test and McNemar s test were used to compare pre-index and post-index HbAc values and HbAc target attainment, respectively Sample selection and creation of analytic variables were performed using the Instant Health Data (IHD) platform (BHE, Boston, MA). Statistical analyses were undertaken with R, version 3.2. (R Foundation for Statistical Computing, Vienna, Austria) Results Patient Selection Study period Medical and pharmacy claims and eligibility data end of continuously eligible post-index follow-up A total of 7 patients were included overall, with 5 GLP- RA naïve patients, inclusive of 25 patients who were GLP- RA naïve with HbAc>9% (Figure 2) Baseline Patient Demographic and Clinical Characteristics Overall, the average age of patients was 52 years, and 49% were female (Table ) Just over half of all patients were prescribed their first claim of semaglutide SC by an endocrinologist The most common baseline comorbid conditions were dyslipidemia (87%), hypertension (72%), and obesity (53%) A third of patients used insulin prior to starting semaglutide SC overall, this proportion increased to more than half in GLP- RA naïve patients with baseline HbAc>9% Table Baseline Patient Demographic and Clinical Characteristics All study patients N=7 GLP- RA naïve subset N=5 High baseline Ac>9% GLP- RA naïve subset N=25 Age at index (years), mean [SD] 52.2 [8.9] 52.9 [8.3] 52.3 [7.2] Age categories (years), n (%) % 3 5.9% 4.% % % % % 2 3.9% 4.% 75.9%.%.% Female, n (%) % % 2 48.% Geographic region of residence of patient 2, n (%) Northeast 3 2.% 7 3.7% 4 6.% Midwest 5 4.7% 3 5.9%.% South % % 9 76.% West 6 5.% 6.8% 2 8.% Health plan type, n (%) HMO 3 28.% % 8 32.% PPO % 2 4.2% 4.% CDHP % % 6 24.% Other 3 2.8% 2 3.9% 4.% Prescribing physician specialty associated with Index semaglutide SC claim, n (%) Endocrinology % 39.2% 2 48.% Non-physician clinician 7 5.9% 2.6% 5.% Primary care physician % % 5.% Others 8 7.5% 4 7.8% 3 2.% Missing/Unknown 2.9% 2 3.9%.% Comorbidities of Interest, n (%) Dyslipidemia % % 2 84.% Hypertension % % 7 68.% Obesity % % 44.% Retinopathy % % 7 28.% Sleep apnea 22.6% 9.6% 4 6.% Neuropathy 8 6.8% % 5.% Ischemic heart disease 5 4.% 7 3.7% 4.% Nephropathy 8 7.5% 4 7.8% 4.% Antidiabetic medications use3 7.% 5.% 25.% Insulin (any insulin) % % 3 52.% Basal insulin (any basal insulin) % % 3 52.% Non-basal insulin (any non-basal insulin) 4 3.% 5 9.8% 4 6.% Non-insulin anti-diabetic (any anti-diabetic medication) 7.% 5.% 25.% Biguanides (metformin) % % 6 64.% Sodium/glucose cotransporter 2 (SGLT2) inhibitors % % 5.% GLP- RA, excluding semaglutide SC %.%.% Sulfonylureas % % 8 32.% Dipeptidyl peptidase (DPP-IV) inhibitors 9.4% 8 5.7% 4 6.% Thiazolidinediones (TZDs) 6 5.6% 2 3.9% 4.% Abbreviations: SD, standard deviation; HbAc, glycosylated hemoglobin; GLP- RA, Glucagon-like peptide receptor agonists; HMO, Health Maintenance Organization; PPO, Preferred Provider Organization; CDHP, Consumer-Driven Health Plan Patients with no claims for any GLP- RA during the 2-month pre-index period 2 The remaining number of patients are in the category Missing/Unknown 3 Use of insulin or non-insulin antidiabetic medications within 9 days prior to the index date Mean HbAc (%) Outcomes Figure 2 Patient Selection Flow Chart Patients with pharmacy claim for semaglutide SC during the patient identification period (2//7 6/3/8) N=,88 Patients with 2 months pre-index continuous eligibility N=,57 claim for type 2 diabetes during the 2 months pre-index continuous eligibility period N=,433 Patients with valid pre-index HbAc result during the three months prior to the first semaglutide SC claim N= Overall (N=7) GLP- RA Naive (N=5) HbAc >9% GLP- RA Naive (N=25) Pre-index HbAc measurement -2. Patients with valid post-index HbAc result during the period starting 9 days after semaglutide SC initiation N=7 (FINAL OVERALL STUDY SAMPLE) Patients with no claims for a GLP- RA during the 2-month pre-index period N=5 (GLP- RA naive subset) Patients with baseline HbAc >9% N=25 (High baseline Ac >9% GLP- RA naive subset).8 Post-index HbAc measurement p<. Semaglutide SC initiation was associated with a significant reduction in HbAc in all patients (-.3%), GLP- RA naïve patients (-2.%), and HbAc >9% GLP- RA naïve patients (-2.9%) (all p<.) (Figure 3) Proportion of patients (%) % 8% 6% 4% % % 22% 47% Discussion/Limitations These results reflect an early after-launch, initial cohort of individuals with T2DM prescribed semaglutide SC; characteristics of future patients prescribed semaglutide SC may differ Due to the need to escalate the dose for semaglutide SC combined with the study s limited follow-up, the full impact of semaglutide SC use may not have been captured The study sample was identified from claims in one large US administrative claims database and consisted of patients with commercial and Medicare Advantage health insurance plans; results may not be generalizable to patients enrolled in other types of health insurance or those residing outside the US HbAc lab values were only available for a subset population in the claims database The study is subject to the inherent limitations associated with pre/post study designs Conclusions 44% 8% 2% This real-world study is the first analysis in the US to describe real-world effectiveness of semaglutide SC in individuals with T2DM 49% Early semaglutide SC initiation and HbAc assessment were associated with clinically and statistically significant reductions in HbAc and increases in HbAc target attainment in this preliminary analysis Additional real-world evidence generation is ongoing that will examine the full breadth of patients prescribed semaglutide SC after this initial, early after-launch cohort as well as allowing for longer follow-up to assess the impact of fully escalated doses and the durability of its effect 29% 77% 32% % % HbAc <7% HbAc <8% HbAc <7% HbAc <8% HbAc <7% HbAc <8% Overall (N=7) GLP- RA Naive (N=5) HbAc >9% GLP- RA Naive (N=25) Pre-index HbAc goal attainment Post-index HbAc goal attainment 6% p<. Attainment of HbAc <7% increased from the pre-index to the post-index period for all patients (22% to 47%), GLP- RA naïve patients (2% to 49%), and HbAc >9% GLP- RA naïve patients (% to 32%) (all p<.) (Figure 4) Sponsorship: Study funding was provided by Novo Nordisk, Inc. Jay Visaria, Vincent Willey, and Bal Nepal are employees of HealthCore, Inc., an independent research organization that received funding from Novo Nordisk, Inc. for the conduct of the study. Tam Dang Tan is employee of Novo Nordisk, Inc., and Paul Petraro was an employee of Novo Nordisk, Inc. when study was conducted. Poster presentation at the 79th Scientific Sessions of American Diabetes Association, June 7-, 9, San Francisco, California Corresponding author: Jay Visaria, HealthCore, Inc., 23 Justison St. Wilmington, DE, 98, USA, jvisaria@healthcore.com, Tel: (32) References: () Victoza [prescribing information]. Plainsboro, NJ: Novo Nordisk, Inc.; revised August 7. (2) Trulicity [prescribing information]. Indianapolis, IN: Eli Lilly and Company; revised January 9. (3) Willey VJ, Kong S, Wu B, et al. Estimating the real-world cost of diabetes mellitus in the United States during an 8-year period using 2 cost methodologies. Am Health Drug Benefits. 8;(6):3-8. (4) Hernandez AF, Green JB, Janmohamed S, et al. Albiglutide and cardiovascular outcomes in patients with type 2 diabetes and cardiovascular disease (Harmony Outcomes): a double-blind, randomised placebo-controlled trial. Lancet. 8;392(57): (5) Holman RR, Bethel MA, Mentz RJ, et al. EXSCEL Study Group. Effects of once-weekly exenatide on cardiovascular outcomes in type 2 diabetes. N Engl J Med 7;377: (6) Marso SP, Daniels GH, Brown-Frandsen K, et al. LEADER Steering Committee; LEADER Trial Investigators. Liraglutide and cardiovascular outcomes in type 2 diabetes. N Engl J Med 6;375: (7) Marso SP, Bain SC, Consoli A, et al. SUSTAIN- 6 Investigators. Semaglutide and cardiovascular outcomes in patients with type 2 diabetes. N Engl J Med 6;375: Abstract

52 Semaglutide once weekly Poster presentation -P -P Semaglutide-induced weight loss is associated with improved health-related quality of life and treatment satisfaction Kari T. Uusinarkaus ; Helena W. Rodbard 2 ; Luc Van Gaal 3 ; John P. Wilding 4 ; Thomas Hansen 5 ; Anna Sandberg 5 ; Sayeh Tadayon 5 ; Melanie J. Davies 6 DaVita Medical Group, Colorado Springs, CO, USA ( kariuusinarkaus@gmail.com; Tel: ); 2 Endocrine and Metabolic Consultants, Rockville, MD, USA; 3 Antwerp University Hospital, Antwerp, Belgium; 4 Institute of Ageing and Chronic Disease, University of Liverpool, Liverpool, UK; 5 Novo Nordisk A/S, Søborg, Denmark; 6 Diabetes Research Centre/Department of Health Sciences, University of Leicester, Leicester, UK. (qrs.ly/mo9xx3r) Aim Semaglutide (Novo Nordisk, Denmark) is a glucagon-like peptide- (GLP-) analog currently approved for the once-weekly subcutaneous treatment of type 2 diabetes (T2D), and has shown reductions in HbA c and body weight across the SUSTAIN clinical trial program. 2 7 Health-related quality of life (HRQoL) and treatment satisfaction were evaluated in the SUSTAIN 2 5 and 7 trials. 3 7 Studies have suggested that weight loss in patients with T2D may be associated with an increase in HRQoL. 8,9 The aim of this post hoc analysis was to assess if weight loss was associated with improvements in patient-reported HRQoL and treatment satisfaction in SUSTAIN 2 5 and 7. Methods Changes in HRQoL (measured by the Physical Component Summary [PCS] and Mental Component Summary [MCS] scores of the Short Form-36 Health Survey version 2 [SF-36v2 ]) and treatment satisfaction scores (measured by the Diabetes Treatment Satisfaction Questionnaire status version [DTSQs]) were evaluated in subjects who achieved 5% and % weight loss ( responders ) vs those who did not ( non-responders ) at end of treatment (3, 4, or 56 weeks) in SUSTAIN 2 5 and 7. The weight-loss responders were chosen to represent meaningful changes at the individual level, as weight losses of 5% and % are known to be clinically meaningful. Estimated responder differences are evaluated in this analysis. Data were pooled across the trials (N=2,88; comparator data not evaluated), and presented by dose (semaglutide.5 mg or. mg) and overall. Patient-reported outcome (PRO) scales Norm-based scoring was used for the SF-36v2, setting the general population mean to 5 for each domain; higher and increasing scores indicate better health. The standard DTSQs scales range from to 6 on a 7-point Likert scale, where 6 indicates the highest treatment satisfaction and the lowest, with the exception of questions on the perception of hyperglycemia and hypoglycemia, where 6 indicates the lowest treatment satisfaction and the highest.» The overall treatment satisfaction is the sum of all scores, excluding the perception of hyperglycemia and hypoglycemia. Statistical analysis Body weight and PROs were analyzed using on-treatment without rescue medication data.» Missing body weight (kg) data were imputed from a mixed model for repeated measurements with treatment, region, and stratum as fixed factors and baseline value as covariate, all nested within visit.» PRO data were analyzed using an analysis of covariance controlled for treatment, strata, and baseline values of body weight and PROs. Safety was assessed using on-treatment data. Results Subject disposition and baseline characteristics for SUSTAIN 2 5 and 7 are shown in Table. Overall, 5.% and 7.4% of subjects achieved 5% and % weight loss with semaglutide, respectively (Figure ). Significantly greater improvements in the overall PCS score and most of its components were reported in responders vs non-responders in the semaglutide. mg and pooled groups (Figure 2A). Overall treatment satisfaction was improved in responders vs non-responders in the semaglutide. mg and pooled groups (Figure 2B). Perception of hyperglycemia, but not hypoglycemia, improved in responders vs non-responders with both doses of semaglutide (Figure 2C). The safety profile of semaglutide by dose and overall is shown in Table 2. Table : Subject disposition and baseline characteristics in SUSTAIN 2 5 and 7 Subject disposition, n (%) Semaglutide.5 mg Semaglutide. mg Semaglutide pooled (.5 mg and. mg) Randomized,5,6 2,85 Exposed,4 (99.9),64 (99.6) 2,88 (99.8) Treatment completers,4 (86.5),339 (83.5) 2,38 (84.8) Subjects who discontinued treatment prematurely 63 (3.5) 265 (6.5) 428 (5.2) Baseline characteristics Male, % 647 (53.7) 845 (52.7),492 (53.) Diabetes duration, years 7.9 (5.8) 8.6 (6.5) 8.3 (6.2) HbA c, % 8.2 (.9) 8.2 (.9) 8.2 (.9) Body weight, kg 93. (2.8) 93.5 (2.9) 93.3 (2.8) BMI, kg/m (6.5) 33. (6.7) 33. (6.6) Denominator for the percentage is the number of subjects randomized. Denominator for the percentage is the full analysis set. Baseline characteristics are calculated using full analysis set data; values are mean (SD) unless otherwise indicated. Values are n (%). BMI, body mass index; n, number of subjects; SD, standard deviation. Figure : Proportions of subjects achieving weight-loss responses in SUSTAIN 2 5 and 7 Proportion of subjects (%) Physical PCS Bodily pain General health Physical functioning Role-physical Mental MCS Vitality 42.3 Semaglutide.5 mg n/n=59/,4 Social functioning Role-emotional Mental health Subjects achieving 5% weight loss 57.6 Semaglutide. mg n/n=924/,64 5. Semaglutide pooled (.5 mg +. mg) n/n=,433/2,88 Semaglutide.5 mg Proportion of subjects (%) Semaglutide. mg Semaglutide pooled (.5 mg and. mg) Favors <5/% WL.2 Semaglutide.5 mg n/n=35/,4 ERD [95% CI] Subjects achieving % weight loss Favors 5/% WL 22. Semaglutide. mg n/n=354/, Semaglutide pooled (.5 mg +. mg) n/n=489/2,88 Observed on-treatment without rescue medication data with missing body weight (kg) values imputed from a mixed model for repeated measurements with treatment, region, and stratum as fixed factors and baseline value as covariate, all nested within visit. All imputed continuous data were dichotomized. N, number of subjects contributing to the analysis; n, number of subjects responding. 5% vs <5% WL % vs <% WL Table 2: Safety profile of semaglutide in SUSTAIN 2 5 and 7 Figure 2: Estimated responder differences for HRQoL and treatment satisfaction in SUSTAIN 2 5 and 7 (A) HRQoL (SF-36v2 ) Overall treatment satisfaction Overall treatment satisfaction Perception of hyperglycemia Perception of hypoglycemia Semaglutide.5 mg Semaglutide.5 mg Semaglutide. mg Semaglutide.5 mg Favors <5/% WL ERD [95% CI] Favors 5/% WL Semaglutide. mg Favors 5/% WL ERD [95% CI] Semaglutide. mg Favors <5/% WL Semaglutide pooled (.5 mg and. mg) n (%) E R n (%) E R n (%) E R Any AEs 854 (7.9) 3, ,5 (7.7) 5, ,4 (7.4) 9, Serious AEs 77 (6.4) (7.5) (7.) Gastrointestinal AEs 528 (43.9), (45.) 2, ,249 (44.5) 3, Nausea 233 (9.4) (.4) (9.9), Vomiting 96 (8.) (9.5) (8.9) Diarrhea 62 (3.5) (3.6) (3.5) Severe or BG-confirmed hypoglycemia 36 (3.) (4.6) (3.9) On-treatment data. An episode that is severe according to the ADA classification or BG-confirmed by a plasma glucose value (<56 mg/dl [3. mmol/l]) with symptoms consistent with hypoglycemia. %, percentage of subjects experiencing at least one event; ADA, American Diabetes Association; AE, adverse event; BG, blood glucose; E, number of events; n, number of subjects in the safety analysis set experiencing at least one event; R, event rate per patient-years of exposure. (B) Treatment satisfaction (DTSQs) (C) Perception of glycemia (DTSQs) Semaglutide pooled (.5 mg and. mg) % vs <5% WL % vs <% WL Semaglutide pooled (.5 mg and. mg) Key results % vs <5% WL % vs <% WL Observed on-treatment without rescue medication and mixed model for repeated measurements imputed data were included. PROs at end of treatment were analyzed by trial using an analysis of covariance with study-specific strata and responder as fixed factors, and baseline PROs and baseline body weight as covariates. Estimates are weighted means of individual trials with weight /SE^2. The individual DTSQs scales are not shown. CI, confidence interval; DTSQs, Diabetes Treatment Satisfaction Questionnaire status version; ERD, estimated responder difference; HRQoL, health-related quality of life; MCS, Mental Component Summary; PCS, Physical Component Summary; PRO, patient-reported outcome; SE, standard error; SF-36v2, Short Form-36 Health Survey version 2 ; WL, weight loss. Discussion This analysis found that weight loss was associated with improvements in the PCS score of SF-36v2, overall treatment satisfaction, and perception of hyperglycemia in subjects achieving weight-loss responses vs those not achieving these responses.» This association appears to be dose dependent, with the changes observed in the pooled semaglutide group being driven by the semaglutide. mg data. In all semaglutide groups, there was a significant association between weight loss and the perception of hyperglycemia. There was no difference in the perception of hypoglycemia, potentially due to the low rate of hypoglycemia observed in the SUSTAIN trials. The safety profile of semaglutide in the SUSTAIN 2 5 and 7 trials was consistent with that of other GLP- receptor agonists (GLP-RAs).,2 GLP-RAs may offer HRQoL and treatment satisfaction benefits, which are often associated with the drugs effects on weight. 3» The SCALE-Diabetes trial in subjects with obesity and T2D showed that weight loss with liraglutide 3. mg may be associated with improved HRQoL. 9» Furthermore, a real-world study showed that a 3% weight loss after 3 6 months of GLP-RA therapy was associated with increased adherence over 8 months vs those without this early weight-loss response. 4 PROs are assessed by patients, and subjective interpretations may confound results; therefore, it can be difficult to infer how changes in HRQoL are influenced by adverse events. The results of this analysis may be also confounded in part by the greater numbers of responders in the semaglutide. mg groups than in the semaglutide.5 mg groups. The focus of this analysis is the effect of weight loss on treatment satisfaction; it does not compare the % vs 5% responder groups. The results in the present analysis are clinically relevant as they suggest that weight loss can be a driver for treatment satisfaction and improved HRQoL. Conclusion Weight loss was associated with improvements in the PCS score of the SF-36v2, overall treatment satisfaction, and the perception of hyperglycemia across the SUSTAIN 2 5 and 7 trials. These data suggest that weight loss may be an important factor affecting HRQoL and treatment satisfaction improvements during T2D treatment with semaglutide. These trials were sponsored by Novo Nordisk and are registered with ClinicalTrials.gov (NCT9388; NCT8858; NCT228932; NCT23538; NCT26484). Presenter Kari Uusinarkaus has received consulting fees from AstraZeneca and Novo Nordisk, speaker fees from Amarin, and consulting and speaker fees from Amgen and Regeneron. The authors are grateful to Ingrid Holst, Novo Nordisk, for review of and input to the poster. The authors acknowledge the medical writing assistance of Flavia Sciota, PhD, AXON Communications (supported by Novo Nordisk). Presented at the 79th Scientific Sessions of the American Diabetes Association, June 7, 9, San Francisco, CA, USA. References: () Novo Nordisk. Ozempic (semaglutide) Prescribing Information 7. Available at: Accessed April 9; (2) Sorli C et al. Lancet Diabetes Endocrinol 7;5:25 6; (3) Ahrén B et al. Lancet Diabetes Endocrinol 7;5:34 54; (4) Ahmann AJ et al. Diabetes Care 8;4:258 66; (5) Aroda VR et al. Lancet Diabetes Endocrinol 7;5:355 66; (6) Rodbard HW et al. J Clin Endocrinol Metab 8;3:229 3; (7) Pratley RE et al. Lancet Diabetes Endocrinol 8;6:275 86; (8) Ridderstråle M et al. Health Qual Life Outcomes 6;4:3; (9) Wharton S et al. Can J Diabetes 5;39(Suppl. ):S35; () Jensen MD et al. Obesity 4;22(Suppl. 2):S5 S39; () Russell-Jones D et al. Diabetologia 9;52:46 55; (2) Diamant M et al. Lancet Diabetes Endocrinol 4;2:464 73; (3) Billings LK et al. J Manag Care Spec Pharm 8;24(Suppl. 9-a):S3 S4; (4) Durden E et al. J Manag Care Spec Pharm 9 March 2. doi:.8553/jmcp [Epub ahead of print]. Abstract

53 Semaglutide oral eposter Theater presentation 983-P 983-P Effect and Safety of Flexible Dose Adjustment with Oral Semaglutide vs Sitagliptin in Type 2 Diabetes: PIONEER 7 John B. Buse, Bruce W. Bode 2, Young Min Cho 3, Erik Christiansen 4, Christin L. Hertz 4, Ann Mertens 5, Signe O. Wallenstein 6, Thomas R. Pieber 7 University of North Carolina School of Medicine, Chapel Hill, NC, USA (john_buse@med.unc.edu); 2 Atlanta Diabetes Associates, Atlanta, GA, USA; 3 Seoul National University College of Medicine, Seoul, Republic of Korea; 4 Novo Nordisk A/S, Søborg, Denmark; 5 KU Leuven, Leuven, Belgium; 6 Novo Nordisk A/S, Aalborg, Denmark; 7 Medical University of Graz, Graz, Austria - For the PIONEER 7 investigators qrs.ly/v69xx44 Background and aims Oral semaglutide is the first glucagon-like peptide- receptor agonist (GLP-RA) in an oral formulation, and it is currently in development for the treatment of type 2 diabetes (T2D). Oral semaglutide is a co-formulation of semaglutide and an absorption enhancer, sodium N-(8-[2-hydroxylbenzoyl] amino) caprylate (SNAC). Three doses (3, 7, and 4 mg) of oral semaglutide have been studied in the phase 3a trials, with superior reductions in HbA c in patients with T2D insufficiently controlled on diet and exercise, and superior reductions in HbA c for the higher doses vs sitagliptin (a dipeptidyl peptidase-4 [DPP-4] inhibitor) in patients with T2D on metformin ± sulfonylurea. 2 PIONEER 7 sought to evaluate the efficacy and safety of oral semaglutide, administered according to an individualized and flexible dose adjustment approach, compared with a fixed dose of sitagliptin. Methods PIONEER 7 was a randomized, open-label, active-controlled, multinational phase 3a trial in patients with T2D uncontrolled on 2 oral glucose-lowering drugs. Patients were randomized to receive once-daily oral semaglutide with flexible dose adjustment or once-daily oral sitagliptin mg for the 52-week treatment period (Figure ). After 52 weeks, patients could complete the trial or continue in a 52-week extension phase (currently ongoing). Figure Trial design. Key inclusion criteria Age 8 years T2D 9 days Stable doses of 2 OADs for 9 days HbA c % (58 8 mmol/mol) N=54 Potential dose adjustment: Week: Dose adjustment criteria: 3 mg mg sitagliptin (N=25) 8 9 years for patients in South Korea. Metformin, a sodium glucose cotransporter 2 inhibitor, a sulfonylurea, and/or a thiazolidinedione. OAD, oral anti-diabetes drug. The primary endpoint was the achievement of HbA c <7.% at week 52. The confirmatory secondary efficacy endpoint was the change from baseline to week 52 in body weight (kg). Two scientific questions were addressed by two estimands (Figure 2). X 3 or 7 mg 6 X Primary evaluation 3, 7, or 4 mg oral semaglutide (N=253) 24 X Efficacy HbA c <7.%: dose continued HbA c 7.%: dose escalated to the next dose level Tolerability Moderate to severe nausea/vomiting for 3 days in the week prior to the visit: dose maintained or reduced 32 X 4 X 48 X 52 5 week follow-up Figure 2 Estimands in the PIONEER 7 trial. Results A total of 54 patients were randomized (Figure ). Baseline characteristics were similar between treatment groups (Table ). Rescue medication was initiated in 5.9% of patients receiving sitagliptin, and 3.2% receiving oral semaglutide flexible dose adjustment. Of patients remaining on treatment by the end of the trial, the majority (59.4%) were receiving oral semaglutide 4 mg, 3.2% were receiving oral semaglutide 7 mg, and 9.% were receiving oral semaglutide 3 mg (dose information was missing for.4%) (Figure 3). Efficacy TREATMENT POLICY ESTIMAND Definition: Effect regardless of trial product discontinuation or rescue medication. Analyzed by: Pattern mixture model with multiple imputation to handle missing data. Primary and confirmatory secondary endpoints were controlled for multiplicity. Patient B Initiates rescue medication Patient C Discontinues trial product Patient D Withdraws TRIAL PRODUCT ESTIMAND Definition: Effect assuming all patients remained on trial product and did not use rescue medication. Analyzed by: Mixed model for repeated measures. A B C D A B C D Patient A Completes treatment Data used Data imputed/predicted Last data point used Oral semaglutide flexible dose adjustment was superior to sitagliptin in achieving HbA c <7.% (Figure 4). Oral semaglutide flexible dose adjustment significantly reduced HbA c and body weight compared with sitagliptin (Figure 5). Oral semaglutide flexible dose adjustment was superior to sitagliptin for reducing body weight (Figure 6). More patients achieved a body weight loss of 5% (27.% vs 2.%) and % (6.4% vs 2.%), and the composite endpoint of HbA c <7.% without hypoglycemia and without body weight gain (45.2% vs 4.7%) with oral semaglutide versus sitagliptin for the treatment policy estimand. The odds of achieving these outcomes were statistically significantly better with oral semaglutide. Results were similar for the trial product estimand. Table Baseline demographics and disease characteristics. Oral semaglutide flex Sitagliptin mg Patients, N Male, n (%) 45 (57.3) 4 (55.8) Age, y 57 ± 58 ± HbAc, % 8.3 ± ±.6 HbAc, mmol/mol 67. ± ± 6.5 Duration of diabetes, y 8.6 ± ± 6.2 FPG, mg/dl 77 ± ± 46 Body weight, kg 88.9 ± ±. BMI, kg/m ± ± 6. Data are mean ± SD unless otherwise specified. BMI, body mass index; flex, flexible dose adjustment; FPG, fasting plasma glucose; SD, standard deviation. Figure 3 Prescribed dose at, 8, 6, and 48 weeks. Proportion of patients (%) Week Week 8 Week 6 Week % 3.2% 9.% Oral semaglutide 3 mg Oral semaglutide 7 mg Oral semaglutide 4 mg Missing information Figure 4 Patients achieving HbA c <7.% at week 52 (primary endpoint). Proportion of patients (%) Treatment policy estimand Oral semaglutide flex Discontinued treatment Final percentages are a proportion of the patients remaining on oral semaglutide at the end of the trial. Trial product estimand 63 Key Result Observed proportions. P<. for odds of achieving the target when comparing oral semaglutide flex with sitagliptin mg. Flex, flexible dose adjustment Sitagliptin mg 28 Figure 5 Change from baseline to week 52 in HbA c. Mean baseline HbA c : 8.3%. Change from baseline (%) Line graph: values are observed means ± standard error of the mean. Bar graphs: estimated mean changes from baseline to week 52. P<. vs sitagliptin. Flex, flexible dose adjustment. Line graph: values are observed means ± standard error of the mean. Bar graphs: estimated mean changes from baseline to week 52. P<. vs sitagliptin. Flex, flexible dose adjustment. For both estimands, the change from baseline to week 52 in Diabetes Treatment Satisfaction Questionnaire (DTSQs) scores appeared similar for oral semaglutide vs sitagliptin for 7 out of 8 domains, including satisfaction with treatment, convenience and flexibility of treatment, as well as total treatment satisfaction. The difference in score for unacceptable feelings of high blood sugar favored oral semaglutide vs sitagliptin. Safety Treatment policy estimand Time since randomization (weeks) Oral semaglutide flex Sitagliptin mg There were no severe 3 hypoglycemic episodes during the trial.. Low and similar proportions of patients experienced blood-glucose confirmed (<56 mg/dl [3. mmol/l]) symptomatic hypoglycemic episodes between treatment groups (oral semaglutide, 5.5% [4 events/ patient years] vs sitagliptin, 5.6% [9 events/ patient years]). Episodes generally occurred in patients on a background of sulfonylurea treatment. -.8 Trial product estimand Figure 6 Change from baseline to week 52 in body weight. Mean baseline body weight: 88.6 kg Change from baseline (kg) Treatment policy estimand Time since randomization (weeks) Oral semaglutide flex Sitagliptin mg -.7 Trial product estimand -.8 Adverse events (AEs) were more common with oral semaglutide, most of which were nausea that was mild/moderate and of short duration (Table 2). Trial product discontinuation due to AEs (mostly gastrointestinal) was higher with oral semaglutide. Event adjudication committee-confirmed neoplasms occurred in 3.2% of patients with oral semaglutide and.8% with sitagliptin, with no clustering to specific organs. Table 2 Safety overview. Patients experiencing at least one AE, n (%) Discussion PIONEER 7 provides information relevant to physicians by using a clinically important primary endpoint (achievement of HbA c <7.%), and a flexible dose adjustment every second month, which more closely replicates the individualized approach of adjusting dose according to efficacy and tolerability. Conclusion Oral semaglutide flex Sitagliptin mg Patients, N All AEs 97 (77.9) 72 (68.8) Most frequent AEs ( 5%) by preferred term Nausea 53 (.9) 6 (2.4) Nasopharyngitis 26 (.3) 3 (5.2) Headache 25 (9.9) 5 (6.) Diarrhea 22 (8.7) 8 (3.2) Abdominal pain upper 6 (6.3) 3 (.2) Vomiting 4 (5.5) 2 (.8) Dyspepsia 3 (5.) 2 (.8) Upper respiratory tract infection 9 (3.6) 5 (6.) Serious AEs 24 (9.5) 24 (9.6) Fatal AEs 2 (.8) AEs by severity Severe 6 (6.3) 8 (7.2) Moderate 4 (4.) 75 (3.) Mild 67 (66.) 44 (57.6) Premature trial product discontinuation due to AEs 22 (8.7) 8 (3.2) One patient on sitagliptin mg was not exposed to trial product and so excluded from safety analyses. In-trial. AE, adverse event; flex, flexible dose adjustment. Flexible dose adjustment with oral semaglutide provided superior glycemic control and weight loss at week 52 vs sitagliptin. These benefits were observed despite less frequent use of additional glucose-lowering therapies with oral semaglutide versus sitagliptin. Oral semaglutide was also well tolerated, with a safety profile consistent with the GLP-RA class. This trial was sponsored by Novo Nordisk A/S and is registered with ClinicalTrials.gov (NCT28498). The authors acknowledge Graham Allcock (Spirit Medical Communications Group Ltd) for medical writing assistance, and Brian Bekker Hansen (Novo Nordisk A/S) for reviewing the poster. Presented at the 79 th Scientific Sessions of the American Diabetes Association, San Francisco, CA, USA, June 7, 9. References: () Aroda, et al. Diabetes 8; 67 (Suppl ): 2-LB; (2) Rosenstock, et al. JAMA 9; 32: 466 8; (3) Seaquist ER, et al. Diabetes Care 3; 36: PIONEER 7 ADA poster_for print_7may9 LLv INDESIGN.indd 2/5/9 4:3:36 Abstract

54 Semaglutide oral eposter Theater presentation 985-P 985-P Oral Semaglutide as Add-on to Insulin in T2D: PIONEER 8 Bernard Zinman, Vanita R. Aroda 2, John B. Buse 3, Bertrand Cariou 4, Stewart B. Harris 5, Søren Tetens Hoff 6, Karen Boje Pedersen 6, Mads Jeppe Tarp-Johansen 6, Eiichi Araki 7 Lunenfeld-Tanenbaum Research Institute, Mount Sinai Hospital, University of Toronto, Toronto, Ontario, Canada (zinman@lunenfeld.ca); 2 Division of Endocrinology, Diabetes, and Hypertension, Brigham and Women s Hospital, Harvard Medical School, Boston, MA, USA; and MedStar Health Research Institute, Hyattsville, MD, USA; 3 Division of Endocrinology & Metabolism, University of North Carolina School of Medicine, Chapel Hill, NC, USA; 4 Department of Endocrinology, L Institut du Thorax, CIC INSERM 43, CHU Nantes, UNIV Nantes, Nantes, France; 5 Department of Family Medicine, Schulich School of Medicine and Dentistry, Western University, London, Ontario, Canada; 6 Novo Nordisk A/S, Søborg, Denmark; 7 Department of Metabolic Medicine, Faculty of Life Sciences, Kumamoto University, Kumamoto, Japan For the PIONEER 8 Investigators qrs.ly/v69xx44 Aim Adding a glucagon-like peptide- receptor agonist (GLP-RA) to insulin can reduce glycated hemoglobin (HbA c ) and body weight without increasing the risk of hypoglycemia. Treatment guidelines recommend this as an effective intensification strategy for patients uncontrolled on insulin. 2 Oral semaglutide is the first GLP-RA in an oral formulation currently in development for the treatment of type 2 diabetes (T2D). PIONEER 8 assessed the efficacy, safety, and tolerability of oral semaglutide in combination with three common insulin regimens. Methods PIONEER 8 was a randomized, double-blind, placebo-controlled, parallel-group phase 3a trial in patients with T2D uncontrolled on insulin ± metformin. Patients were randomized to receive once-daily oral semaglutide 3, 7, or 4 mg, or placebo for the 52-week treatment period (Figure A). Randomization was stratified by patients country of origin (Japanese or non-japanese) and background treatment (± metformin, and basal, basal-bolus, or premixed insulin). A % reduction in total daily insulin dosage was recommended at randomization and was to be maintained to week 8. The treatment period was then split into two insulin dosing stages (Figure B). A capped insulin period covered weeks 8 26, during which total daily insulin dosage should not exceed the dosage at randomization. For weeks 26 52, total daily insulin dosage was freely adjustable. Figure A: Trial design; B: Insulin dosing periods. A B Key inclusion criteria Age 8 years T2D 9 days HbA c % (53 8 mmol/mol) Stable dose of either basal, basal bolus or premixed insulin for 9 days Stable dose of metformin for 9 days (if used) Aim of insulin titration: FPG 7 99 mg/dl; HbA c <7.% (<53 mmol/mol). -% Week: Capped insulin use 3 mg oral semaglutide (N=84) 3 mg Primary evaluation 7 mg oral semaglutide (N=82) 3 mg 7 mg 4 mg oral semaglutide (N=8) Placebo (N=84) Week: Dose escalation 5 week follow-up Insulin dosage at randomization Freely adjustable insulin use Adjustments to total daily insulin dosage were made based on the lowest of three self-measured blood glucose values measured on three consecutive days before each visit. A >% increase relative to baseline in total daily insulin dosage was considered rescue medication. FPG, fasting plasma glucose; HbA c, glycated hemoglobin; T2D, type 2 diabetes. The primary endpoint was change in HbA c, and the confirmatory secondary endpoint was change in body weight (kg), both from baseline to week 26. Two scientific questions were addressed by two estimands (Figure 2). Figure 2 Estimands in the PIONEER 8 trial. Results Of the 73 randomized patients (Figure A), 95.3% completed the trial. At screening, 4.9% of patients were on basal insulin, 38.9% were on basal bolus, and 7.6% were on premixed insulin. Baseline characteristics were similar across treatment groups (Table ). By week 26, 2.7%,.%, 2.2%, and 4.9% of patients had initiated rescue medication in the oral semaglutide 3, 7, and 4 mg, and placebo groups, respectively. This was 29.3%, 8.%, 7.%, and 36.4% by week 52. Rescue medication was primarily a >% increase in total daily insulin dosage relative to baseline. Table Baseline demographics and disease characteristics. Oral semaglutide 3 mg 7 mg 4 mg Placebo Patients, N Male, n (%) 2 (55.4) 3 (56.6) 85 (47.) 5 (57.) Age, y 6 ± 9 6 ± 6 ± 6 ± HbA c, % 8.2 ± ± ± ±.7 Duration of diabetes, y 5. ± ± ± ± 7.9 FPG, mg/dl 58.4 ± ± ± ± 47.4 Body weight, kg 85.9 ± ± ± ± 2.4 BMI, kg/m 2 3. ± ± ± ± 6.5 Total daily insulin dosage, U 6 ± ± ± ± 48 Data are mean ± SD unless otherwise specified. BMI, body mass index; FPG, fasting plasma glucose; HbA c, glycated hemoglobin; SD, standard deviation. Efficacy TREATMENT POLICY ESTIMAND Definition: Effect regardless of trial product discontinuation or rescue medication. Analyzed by: Pattern mixture model with multiple imputation to handle missing data. Primary and confirmatory secondary endpoints were controlled for multiplicity. TRIAL PRODUCT ESTIMAND Definition: Effect assuming all patients remained on trial product and did not use rescue medication. Analyzed by: Mixed model for repeated measures. Oral semaglutide was superior to placebo (P<.5) at reducing HbA c (Figure 3) and body weight (Figure 4) from baseline at week 26 (treatment policy estimand). Reductions in HbA c and body weight from baseline were statistically significantly greater (P<.5) with oral semaglutide vs placebo at week 26 (trial product estimand), and week 52 (both estimands). Figure 3 Change from baseline in HbA c. Mean baseline HbA c : 8.2% Change from baseline in HbA c (%) Treatment policy estimand Line graph: observed means ± standard error of the mean. Bar graphs: estimated mean changes from baseline to weeks 26 and 52. P<. for the test of no difference vs placebo. Figure 4 Change from baseline in body weight. Mean baseline body weight: 85.9 kg Change from baseline in body weight (kg) Time since randomization (weeks) Line graph: observed means ± standard error of the mean. Bar graphs: estimated mean changes from baseline to weeks 26 and 52. P<.5 for the test of no difference vs placebo. Figure 5 shows the observed changes from baseline in the total daily insulin dosage. Oral semaglutide 3 mg Oral semaglutide 7 mg Oral semaglutide 4 mg Placebo Insulin use was statistically significantly lower (P<.5) with oral semaglutide vs placebo at weeks 26 and 52 (both estimands), except 3 mg at week 26 (treatment policy estimand). Figure 5.. Change from baseline in HbA c (%) Treatment policy estimand Time since randomization (weeks) Change from baseline in body weight (kg) Week 26 Week Week 26 Week 52.5 Oral semaglutide 3 mg Oral semaglutide 7 mg Oral semaglutide 4 mg Placebo Change from baseline in total daily insulin dosage. Mean baseline total daily insulin dosage: 58 U Change from baseline in total daily insulin dosage (U) Treatment policy estimand Trial product estimand Time since randomization (weeks) Oral semaglutide 3 mg Oral semaglutide 7 mg Oral semaglutide 4 mg Placebo Observed means ± standard error of the mean. Data collected after total daily insulin dosage had increased by >% relative to baseline were excluded from the presentation and analysis for the trial product estimand, as this was considered rescue medication Trial product estimand Week 26 Week Trial product estimand -2.9 Key Results Week 26 Week 52.6 More patients achieved HbA c <7.% and the composite outcome with oral semaglutide vs placebo at weeks 26 and 52 (P. for the estimated odds ratio vs placebo, both outcomes and estimands; Table 2). Table 2 Patients achieving outcome, % Safety Treatment policy estimand Oral semaglutide Trial product estimand Oral semaglutide 3 mg 7 mg 4 mg Placebo 3 mg 7 mg 4 mg Placebo Patients, N HbA c <7.% Patients achieving secondary efficacy outcomes. Week Week HbA c <7.% without hypoglycemia and without body weight gain Week Week Observed proportions. Severe (according to American Diabetes Association classification 3 ) or blood glucose confirmed (<56 mg/dl [3. mmol/l]) symptomatic episodes. Comparable proportions of patients experienced adverse events (AEs) while on treatment (Table 3). In total, there were more events with oral semaglutide vs placebo (626, 555, and 586 for 3, 7, and 4 mg vs 464). Nausea was most frequently reported with oral semaglutide (Table 3), and nasopharyngitis with placebo (4.7% of patients). Gastrointestinal disorders were the most frequently reported AEs leading to premature discontinuation of oral semaglutide. Similar proportions of patients experienced severe or blood glucose confirmed (<56 mg/dl [3. mmol/l]) symptomatic hypoglycemic episodes (Table 3). The majority occurred in patients on basal-bolus insulin. The observed rates of episodes per years of exposure were 5, 2, 86, and 82 with oral semaglutide 3, 7, and 4 mg, and placebo. Table 3 Safety overview. Patients experiencing at least one AE, n (%) Discussion In PIONEER 8, HbA c and body weight reductions at week 26 were superior with oral semaglutide vs placebo when added to insulin. Oral semaglutide may help reduce concerns over weight gain and hypoglycemia associated with insulin use, and may have an insulin-sparing effect at the 7 and 4 mg doses. The two consecutive insulin dosing periods allowed both the glucose-lowering effect of oral semaglutide to be determined in a controlled setting, and data to be obtained longer term in a setting more reflective of clinical practice. Conclusion Oral semaglutide 3 mg 7 mg 4 mg Placebo Patients, N All AEs 37 (74.5) 42 (78.5) 5 (83.4) 39 (75.5) AEs by severity Mild 23 (66.8) 26 (69.6) 34 (74.) 9 (64.7) Moderate 62 (33.7) 65 (35.9) 62 (34.3) 59 (32.) Severe 7 (9.2) 7 (9.4) 3 (7.2) 9 (4.9) GI disorders Nausea 2 (.4) 3 (6.6) 42 (23.2) 3 (7.) Diarrhea 6 (8.7) 22 (2.2) 27 (4.9) (6.) Vomiting (6.) 4 (7.7) 8 (9.9) 7 (3.8) SAEs 25 (3.6) 9 (.5) 2 (6.6) 7 (9.2) Premature trial product discontinuation due to AEs 3 (7.) 6 (8.8) 24 (3.3) 5 (2.7) Deaths 3 (.7) Severe or BG-confirmed symptomatic hypoglycemia, n/n (%) 52 (28.3) 47 (26.) 48 (26.5) 54 (29.3) Basal 8/77 (.4) 2/76 (5.8) /76 (3.2) 6/8 (.) Basal-bolus 36/7 (5.7) 29/73 (39.7) 3/7 (44.3) 27/72 (37.5) Premixed 8/36 (22.2) 6/32 (8.8) 7/35 (.) /32 (34.4) Events occurred while on treatment unless otherwise specified. Nausea events were mostly mild or moderate in severity and of short duration; In-trial events; Severe (according to American Diabetes Association classification 3 ) or BG-confirmed (<56 mg/dl [3. mmol/l]) symptomatic episodes. AE, adverse event; BG, blood glucose; GI, gastrointestinal; n/n (%), number and proportion of patients experiencing hypoglycemic episodes over number of patients on each insulin regimen; SAE, serious adverse event. Oral semaglutide can effectively improve glycemic control and reduce body weight without increasing the risk of hypoglycemia in patients with T2D on insulin ± metformin. The safety profile was consistent with other GLP-RAs. This trial was sponsored by Novo Nordisk and is registered with ClinicalTrials.gov (NCT3287). The authors acknowledge Sophie Walton (Spirit Medical Communications Group Ltd) for medical writing assistance, and Brian Bekker Hansen (Novo Nordisk) for reviewing the poster. Presented at the 79 th Scientific Sessions of the American Diabetes Association Congress, San Francisco, CA, USA, June 7, 9. References: () Castellana, et al. Diabetes Metab Res Rev 9; 35: e382; (2) ADA. Diabetes Care 9; 42: S9 S2; (3) Seaquist ER, et al. Diabetes Care 3; 36: PIONEER 8_Zinman ADA9 poster_for print_7may9 INDESIGN.indd 7/5/9 :22:8 Abstract

55 Semaglutide oral Late-breaking poster presentation -LB -LB Oral Semaglutide vs Sitagliptin: Efficacy by Baseline HbA c and Background OAD in PIONEER 3 Julio Rosenstock, Dale C. Allison 2, Andreas L. Birkenfeld 3, Thalia Marie Blicher 4, Srikanth Deenadayalan 4, Astrid Kousholt 4, Melanie J. Davies 5 Dallas Diabetes Research Center at Medical City, Dallas, TX, USA (juliorosenstock@dallasdiabetes.com); 2 Hillcrest Family Health Center, Waco, TX, USA; 3 Department and Outpatient Department of Medicine III, Carl Gustav Carus University Hospital Dresden, Dresden, Germany, and Paul Langerhans Institute Dresden, Helmholtz Center Munich, University Hospital, Technical University Dresden, German Center for Diabetes Research (DZD e.v.), Dresden, Germany; 4 Novo Nordisk A/S, Søborg, Denmark; 5 Diabetes Research Centre, University of Leicester, Leicester, UK. qrs.ly/v69xx44 Aim Oral semaglutide is the first glucagon-like peptide- receptor agonist in an oral formulation currently in development for the treatment of type 2 diabetes (T2D). The PIONEER 3 trial compared the efficacy, long-term adverse event (AE) profile, and tolerability of oral semaglutide with sitagliptin as an add-on to metformin ± sulfonylurea in patients with T2D. Oral semaglutide 7 and 4 mg were superior to sitagliptin at improving glycemic control and reducing body weight after 26 weeks. This exploratory analysis of PIONEER 3 investigated whether baseline glycated hemoglobin (HbA c ) or background oral anti-diabetic agent (OAD) had an effect on the glycemic efficacy of oral semaglutide vs sitagliptin. Methods PIONEER 3 was a randomized, 78-week, double-blind, double-dummy, parallel-group, multicenter, multinational phase 3a trial in patients with T2D uncontrolled on metformin ± sulfonylurea. Patients were randomized to receive once-daily oral semaglutide 3, 7, or 4 mg, or sitagliptin mg for the 78-week treatment period (Figure ). Oral semaglutide was initiated at 3 mg, then escalated to 7 mg after 4 weeks, and 4 mg after a further 4 weeks, until the randomized dose was achieved. Figure Trial design Key inclusion criteria Age 8 years T2D 9 days HbA c 7..5% (53 9 mmol/mol) Stable dose of metformin ± SU for 9 days Week: Primary evaluation 3 mg oral semaglutide (N=466) 3 mg mg sitagliptin (N=467) HbA c, glycated hemoglobin; SU, sulfonylurea; T2D, type 2 diabetes. 7 mg oral semaglutide (N=466) 3 mg 7 mg 4 mg oral semaglutide (N=465) week follow-up For this exploratory analysis, patients were categorized into subgroups according to their HbA c at baseline ( 8.%, >8. 9.%, >9.%), and their background OAD (metformin alone, metformin + sulfonylurea). The endpoints used to assess glycemic efficacy were the change from baseline to week 26 in HbA c (primary endpoint), and whether a patient achieved HbA c <7.% at week 26. Efficacy endpoints were analyzed using the full analysis set (all randomized patients). The exploratory analyses were based on the trial product estimand, evaluating the treatment effect for all randomized patients, assuming that all patients continued taking trial product for the entire planned duration of the trial and did not use rescue medication. Data collected after discontinuation of trial product or initiation of rescue medication were excluded. Changes from baseline in HbA c were analyzed using a mixed model for repeated measurements with region, background medication and interaction between treatment and subgroup as categorical fixed effects and baseline HbA c as a covariate, all nested within visit. An unstructured residual covariance matrix was used. For the binary endpoint, missing values of HbA c were imputed from patients randomized to the same trial product using an analysis of covariance-based sequential multiple imputation model. One thousand complete datasets were imputed. Each complete dataset was analyzed using logistic regression with region, background medication, and interaction between treatment and subgroup as categorical fixed effects, and baseline HbA c as a covariate. The results were combined using Rubin s rule. 2 Safety endpoints were summarized by descriptive statistics. These summaries included AEs collected for patients during treatment, and are based on the safety analysis set (all randomized and exposed patients). Results Of 864 patients randomized, 863 were included in efficacy analyses, and 86 were exposed to treatment and included in safety analyses. Baseline characteristics by baseline HbA c and background OAD subgroup are shown in Table. Table Baseline demographics and disease characteristics by subgroup Overall Baseline HbA c subgroups OAD subgroups 8.% >8. 9.% >9.% Metformin Metformin alone + SU Patients, N (FAS) Oral semaglutide 3 mg mg mg Sitagliptin mg Female, n (%) 879 (47.2) 49 (48.) 276 (46.5) 94 (46.2) 454 (46.) 425 (48.5) Age, y 58 ± 59 ± 57 ± 56 ± 57 ± 59 ± 9 HbA c, % 8.3 ± ± ± ± ± ±.9 Duration of diabetes, y 8.6 ± ± ± ± ± ± 6. FPG, mg/dl 7 ± 45 5 ± 3 77 ± 44 4 ± 5 67 ± ± 49 Body weight, kg 9.2 ± ± ± ± ± ±.4 BMI, kg/m ± ± ± ± ± ± 6. Data are mean ± SD unless otherwise specified. BMI, body mass index; FAS, full analysis set; FPG, fasting plasma glucose; HbA c, glycated hemoglobin; OAD, oral anti-diabetic agent; SD, standard deviation; SU, sulfonylurea. Change from baseline in HbA c HbA c reductions at week 26 were greater with higher baseline HbA c across all treatment arms (Figure 2Ai). The reductions were broadly comparable across background OAD subgroups (Figure 2Bi). Overall, oral semaglutide 7 and 4 mg showed greater reductions in HbA c vs sitagliptin across all baseline HbA c and background OAD subgroups (Figure 2Aii and 2Bii). When comparing baseline HbA c subgroups, treatment differences for oral semaglutide 7 and 4 mg vs sitagliptin tended to be greater for subgroups with higher baseline HbA c, but the only statistically significant interaction was for oral semaglutide 4 mg (P=.395; Figure 2Aii). Across background OAD subgroups, the treatment differences were similar (Figure 2Bii). Key Result Figure 2 Change from baseline in HbA c at week 26. A: By baseline HbA c ; B: By background OAD. Estimated change from baseline in HbA c (%) Ai Bi Estimated change from baseline in HbA c (%) Overall 8.% >8. 9.% >9.% Aii ETD [95% CI] vs sitagliptin mg Oral semaglutide 3 mg Overall 8.% >8. 9.% >9.% Oral semaglutide 7 mg Overall 8.% >8. 9.% >9.% Oral semaglutide 4 mg Overall 8.% >8. 9.% >9.% Bii HbA c (%) Favors oral semaglutide Favors sitagliptin ETD [95% CI] vs sitagliptin mg Oral semaglutide 3 mg Overall Metformin alone Metformin + SU Oral semaglutide 7 mg Overall Metformin alone Metformin + SU Oral semaglutide 4 mg Overall Metformin alone Metformin + SU HbA c (%) Favors oral semaglutide Interaction P value Favors sitagliptin Interaction P value P value for test of no interaction between treatment and subgroup. CI, confidence interval; ETD, estimated treatment difference; HbA c, glycated hemoglobin; OAD, oral anti-diabetic agent; SU, sulfonylurea. -.4 Overall Metformin alone Metformin + SU Oral semaglutide 3 mg Oral semaglutide 7 mg Oral semaglutide 4 mg Sitagliptin mg Patients achieving HbA c <7.% The observed proportion of patients achieving HbA c <7.% at week 26 was greater with lower baseline HbA c across all treatment arms (Figure 3Ai). Overall, the odds of achieving HbA c <7.% were greater with oral semaglutide 7 and 4 mg vs sitagliptin across all baseline HbA c subgroups. The odds ratios tended to be greater for the subgroups with higher baseline HbA c, but the only statistically significant interaction was for oral semaglutide 7 mg (P=.; Figure 3Aii). The observed proportion of patients achieving HbA c <7.% was greater with metformin alone than with metformin + sulfonylurea (Figure 3Bi). The odds of achieving HbA c <7.% were greater with oral semaglutide 7 and 4 mg vs sitagliptin across background OAD subgroups (Figure 3Bii). Across background OAD subgroups, the odds ratios were similar. Figure 3 Patients achieving HbA c <7.% at week 26. A: By baseline HbA c ; B: By background OAD. Ai Observed proportion of patients (%) Bi Observed proportion of patients (%) Overall 8.% >8. 9.% >9.% Overall Metformin alone Metformin + SU Aii Bii Estimated odds ratio (log scale) Favors sitagliptin Favors oral semaglutide Estimated odds ratio (log scale) Favors sitagliptin Favors oral semaglutide Oral semaglutide 3 mg Oral semaglutide 7 mg Oral semaglutide 4 mg Sitagliptin mg P value for test of no interaction between treatment and subgroup. CI, confidence interval; HbA c, glycated hemoglobin; OAD, oral anti-diabetic agent; SU, sulfonylurea. 24 Estimated odds ratio [95% CI] vs sitagliptin mg Oral semaglutide 3 mg Overall 8.% >8. 9.% >9.% Oral semaglutide 7 mg Overall 8.% >8. 9.% >9.% Oral semaglutide 4 mg Overall 8.% >8. 9.% >9.% Estimated odds ratio [95% CI] vs sitagliptin mg Oral semaglutide 3 mg Overall Metformin alone Metformin + SU Oral semaglutide 7 mg Overall Metformin alone Metformin + SU Oral semaglutide 4 mg Overall Metformin alone Metformin + SU Interaction P value Interaction P value Safety Across treatment arms, similar proportions of patients experienced AEs or serious AEs (Table 2). The most frequent AEs were nausea for oral semaglutide 7 and 4 mg, and nasopharyngitis for oral semaglutide 3 mg and sitagliptin. Gastrointestinal AEs were the most frequent cause of premature discontinuation of trial product due to AEs ( % of patients with oral semaglutide and 2.6% with sitagliptin). Table 2 Safety overview. Patients experiencing at least one AE, n (%) Oral semaglutide 3 mg 7 mg 4 mg Sitagliptin mg Patients, N All AEs 37 (79.4) 363 (78.2) 37 (79.6) 388 (83.3) AEs by severity Mild 323 (69.3) 38 (68.5) 32 (69.) 34 (73.) Moderate 86 (39.9) 7 (36.9) 99 (42.8) 97 (42.3) Severe 47 (.) 37 (8.) 4 (8.6) 53 (.4) SAEs 64 (3.7) 47 (.) 44 (9.5) 58 (2.4) Premature trial product discontinuation due to AEs Severe or BG-confirmed symptomatic hypoglycemia 26 (5.6) 27 (5.8) 54 (.6) 24 (5.2) 23 (4.9) 24 (5.2) 36 (7.7) 39 (8.4) Severe (according to American Diabetes Association classification 3 ) or BG-confirmed (<56 mg/dl [3. mmol/l]) symptomatic episodes. AE, adverse event; BG, blood glucose; SAE, serious adverse event. Discussion In PIONEER 3, oral semaglutide 7 and 4 mg were superior to sitagliptin in reducing HbA c, and were associated with meaningful weight loss. The odds of patients achieving HbA c <7.% were statistically significantly greater at these doses vs sitagliptin.,4 This exploratory analysis found that: Overall, patients receiving oral semaglutide 7 and 4 mg had greater HbA c reductions and greater odds of achieving HbA c <7.% than those receiving sitagliptin, regardless of baseline HbA c. Comparing baseline HbA c subgroups, treatment differences in the change from baseline in HbA c, and odds of achieving HbA c <7.% for oral semaglutide 7 and 4 mg vs sitagliptin tended to be greater for subgroups with higher baseline HbA c. The efficacy of oral semaglutide vs sitagliptin was consistent with the results from the overall analysis, irrespective of background OAD. Conclusion Overall, oral semaglutide 7 and 4 mg improved glycemic control vs sitagliptin across HbA c subgroups, and irrespective of background OAD use. This trial was sponsored by Novo Nordisk and is registered with ClinicalTrials.gov (NCT267865). The authors acknowledge Sophie Walton (Spirit Medical Communications Group Ltd) for medical writing assistance. Presented at the 79 th Scientific Sessions of the American Diabetes Association Congress, San Francisco, CA, USA, June 7, 9. References: () Rosenstock J, et al. JAMA 9; 32: 466 8; (2) Little RJA, et al. Statistical Analysis With Missing Data. 987, John Wiley & Sons; (3) Seaquist ER, et al. Diabetes Care 3; 36: ; (4) Rosenstock J, et al. Presented at ENDO 9: poster SAT-39. PIONEER 3_Rosenstock ADA9 poster_for print_3may9 INDESIGN.indd 3/5/9 :29:2 Abstract

56 Semaglutide oral Poster presentation 4-P 4-P Oral Semaglutide vs Placebo in Patients with Type 2 Diabetes and Moderate Renal Impairment: PIONEER 5 Ofri Mosenzon, Signe Rosenlund, 2 Jan W. Eriksson, 3 Simon R. Heller, 4 Richard E. Pratley, 5 Thozhukat Sathyapalan, 6 Thalia Marie Blicher, 2 Ole H. Hels, Sr., 2 Cyrus Desouza 7 Diabetes Unit, Division of Internal Medicine, Hadassah Hebrew University Hospital, Jerusalem, Israel (ofrim@hadassah.org.il); 2 Novo Nordisk A/S, Søborg, Denmark; 3 Clinical Diabetology and Metabolism, Department of Medical Sciences, Uppsala University, Uppsala, Sweden; 4 Academic Unit of Diabetes, Endocrinology and Metabolism, Department of Oncology and Metabolism, University of Sheffield, Sheffield, UK; 5 AdventHealth Translational Research Institute for Metabolism and Diabetes, Orlando, FL, USA; 6 Academic Diabetes, Endocrinology and Metabolism Research Group, Hull York Medical School, University of Hull, Hull, UK; 7 Division of Diabetes, Endocrinology, and Metabolism, Department of Internal Medicine, University of Nebraska Medical Center, Omaha, NE, USA For the PIONEER 5 Investigators qrs.ly/v69xx44 Aim Semaglutide is a glucagon-like peptide- (GLP-) analog that reduces glycated hemoglobin (HbA c ) and body weight in patients with type 2 diabetes (T2D) uncontrolled on oral glucose-lowering drugs.,2 Currently, all GLP- receptor agonists (GLP-RAs) are given subcutaneously. An innovative oral semaglutide tablet, co-formulated with the absorption enhancer, sodium N-(8-[2-hydroxybenzoyl] amino) caprylate (SNAC), is in clinical development. 3 T2D is commonly associated with renal impairment. In this phase 3a trial, PIONEER 5, the efficacy and safety of once-daily oral semaglutide were compared with placebo in patients with T2D and moderate renal impairment, added to existing medication. Methods PIONEER 5 was a 26-week, randomized, double-blind, placebo-controlled, parallel-group, multicenter, phase 3a trial. Adult patients with T2D diagnosed 9 days prior to screening, HbA c % (53 8 mmol/mol), and moderate renal impairment (estimated glomerular filtration rate [egfr] 3 59 ml/min/.73 m 2 ) were eligible. Background glucose-lowering medication was required to be given at stable doses for at least 9 days prior to screening, and could be either: Metformin ( 5 mg or maximum tolerated dose) and/or a sulfonylurea (at least half maximum approved dose, or maximum tolerated dose); or Basal insulin, with or without metformin. Patients were randomized to receive once-daily oral semaglutide 4 mg or placebo for the 26-week treatment period (Figure ). Oral semaglutide was initiated at 3 mg, then escalated to 7 mg after 4 weeks, and 4 mg after a further 4 weeks. Randomization was stratified by background medication (metformin alone, sulfonylurea ± metformin, or basal insulin ± metformin) and renal function (egfr or 3 44 ml/min/ 73 m 2 ). Figure Trial design. Patients 8 years with T2D and moderate renal impairment (egfr 3 59 ml/min/.73 m 2 ) Metformin ± SU, or basal insulin ± metformin Week: : N=324 3 mg 7 mg Oral semaglutide 4 mg (n=63) Placebo (n=6) Baseline End of treatment Screening Dose escalation egfr, estimated glomerular filtration rate; SU, sulfonylurea; T2D, type 2 diabetes. Primary endpoint: Change in HbA c from baseline to week 26. Follow-up period Confirmatory secondary endpoint: Change in body weight from baseline to week 26. Further endpoints included the proportion of patients reaching targets for HbA c and weight loss, and composite endpoints. 3 End of trial Adverse events (AEs) were recorded throughout treatment and up to 5 weeks after the last dose of trial product. Laboratory parameters were also assessed, and physical examinations carried out, up to 5 weeks after the last dose of trial product. Two scientific questions related to efficacy were addressed through the definition of two estimands (Figure 2). Binary endpoints were analyzed by a logistic regression model. Figure 2 Estimands in the PIONEER 5 trial. TREATMENT POLICY ESTIMAND Definition: Effect regardless of trial product discontinuation or rescue medication. Analyzed by: Pattern mixture model with multiple imputation to handle missing data. Primary and confirmatory secondary endpoints were controlled for multiplicity. Results Of 324 patients randomized, 34 (96.9%) completed the trial. In total, 33 patients (8.6%) in the oral semaglutide group and 4 (87.6%) in the placebo group completed 26 weeks on treatment. In each group, 27 patients completed treatment without use of rescue medication. Baseline demographics and disease characteristics were balanced (Table ). Table Baseline demographics and disease characteristics. Oral semaglutide 4 mg Placebo Patients, N 63 6 Female, n (%) 8 (49.) 88 (54.7) Age, y 7 ± 8 7 ± 8 HbA c, % 8. ± ±.7 Duration of diabetes, y 4. ± ± 7.4 FPG, mg/dl 63.6 ± ± 5. Body weight, kg 9.3 ± ± 7.5 BMI, kg/m ± ± 5.5 egfr, ml/min/.73 m 2 47 ± 48 ± UACR (mg/g), median (range) 9.2 ( ) 4. ( ) Data are mean ± SD unless otherwise specified. BMI, body mass index; egfr, estimated glomerular filtration rate; FPG, fasting plasma glucose; SD, standard deviation, UACR, urinary albumin creatinine ratio. Efficacy TRIAL PRODUCT ESTIMAND Definition: Effect assuming all patients remained on trial product and did not use rescue medication. Analyzed by: Mixed model for repeated measures. Oral semaglutide was superior to placebo for reduction in HbA c from baseline to week 26, by the treatment policy estimand (Figure 3). Oral semaglutide was superior to placebo for reduction in body weight from baseline to week 26 for the treatment policy estimand (Figure 4). In addition, there were statistically significantly greater reductions in HbA c and body weight for oral semaglutide compared with placebo at week 26 for the trial product estimand (Figures 3 and 4). More patients achieved the targets of HbA c <7.% and 6.5% (P<. for both targets and estimands), and weight loss 5% and % (P<. for both targets and estimands) with oral semaglutide compared with placebo. Compared with placebo, more patients achieved composite endpoints with oral semaglutide (Figure 5). Figure 5 Proportion of patients (%) Figure 3 Change from baseline in HbA c. Mean (SE) HbA c change from baseline (%-points) Patients achieving composite targets. Treatment policy estimand 5.6 HbA c <7.%, no hypoglycemia, no weight gain HbA c reduction.% with weight loss 3% Trial product estimand N= Oral semaglutide 4 mg Treatment policy estimand Placebo Trial product estimand Observed proportions. P<. for the test of no difference vs placebo for the estimated odds ratio; Severe (according to American Diabetes Association classification 4 ) or blood glucose-confirmed (<56 mg/dl [3. mmol/l]) symptomatic episodes. Safety More patients reported an AE with oral semaglutide than placebo; proportions of serious AEs were similar (Table 2). There were more premature trial product discontinuations due to AEs in the oral semaglutide group than the placebo group, mainly due to gastrointestinal events (primarily nausea). There was more confirmed symptomatic hypoglycemia with oral semaglutide (9 vs 3 events per patient-years), but no severe hypoglycemic events. Line graph: observed means ± standard error of the mean. Bar graphs: estimated mean changes from baseline to week 26. CI, confidence interval; ETD: estimated treatment difference; HbA c, glycated hemoglobin; SE, standard error; sema, semaglutide. Figure 4 Change from baseline in body weight. Mean (SE) body weight change from baseline (kg) Oral sema 4 mg Placebo Oral sema 4 mg Placebo Mean baseline HbA c: 8.% Treatment policy estimand Time since randomization (weeks) Mean baseline body weight: 9.8 kg Treatment policy estimand Time since randomization (weeks) Line graph: observed means ± standard error of the mean. Bar graphs: estimated mean changes from baseline to week 26. CI, confidence interval; ETD: estimated treatment difference; SE, standard error; sema, semaglutide. Mean HbA c change from baseline (%-points) Mean body weight change from baseline (kg) Oral sema 4 mg. Oral sema 4 mg 3.4 ETD [95% Cl]:.8 [.,.6] P<. ETD [95% Cl]: 2.5 [ 3.2,.8] P<. Placebo.2 Placebo.9 Mean HbA c change from baseline (%-points) Mean body weight change from baseline (kg) Trial product estimand Oral sema 4 mg. Trial product estimand Oral sema 4 mg ETD [95% Cl]:. [.2,.8] P<. ETD [95% Cl]: 2.7 [ 3.5,.9] P<. Key Result Placebo. Placebo. 7.9 Nausea was more common in patients receiving oral semaglutide with egfr ml/min/.73 m 2 than 3 44 ml/min/.73 m 2. Table 2 Safety overview. Patients with at least one AE, n (%) Oral semaglutide 4 mg Placebo Patients, N 63 6 All AEs (73.6) 5 (65.2) Mild 6 (65.) 89 (55.3) Moderate 6 (37.4) 42 (26.) Severe (6.) 5 (9.3) Gastrointestinal disorders Nausea 3 (9.) 2 (7.5) Constipation 9 (.7) 6 (3.7) Vomiting 9 (.7) 2 (.2) Serious AEs 7 (.4) 7 (.6) AE-related premature trial product discontinuation 24 (4.7) 8 (5.) Deaths (.6) 2 (.2) Severe/blood-glucose confirmed symptomatic hypoglycemia 9 (5.5) 3 (.9) In-trial events; Severe (American Diabetes Association classification 4 ) or blood glucose-confirmed (<56 mg/dl [3. mmol/l]) symptomatic episodes. AE, adverse event. Overall, renal function was unchanged throughout the trial in both treatment groups: median egfr ratios (week 3 to baseline) were.2 (range.27,.96) for oral semaglutide and. (.68, 2.7) for placebo. Geometric mean urinary albumin creatinine ratios (week 26 to baseline):.86 (range.4, 56.7) for oral semaglutide and.9 (., 79.59) for placebo. Two patients in the oral semaglutide group had three adjudicated non-serious events of acute kidney injury (stage ). One patient in the placebo group had a non-serious event of acute kidney injury (stage 2). There were three deaths during the trial: one in each group from cardiovascular causes, and another with placebo from an unconfirmed cause. Discussion GLP-RAs have demonstrated effective glycemic control in patients with T2D and impaired renal function. 5,6 The results of PIONEER 5 indicate that oral semaglutide is efficacious and well tolerated in patients with T2D and moderate renal impairment. In addition, oral semaglutide did not affect renal function. Oral semaglutide could provide the first oral GLP-RA for patients with T2D and renal impairment, who currently lack effective oral treatments. More generally, an oral GLP-RA could be advantageous for patients who require effective glycemic control without injections. Conclusion Oral semaglutide effectively improved glycemic control and reduced body weight vs placebo in patients with T2D and moderate renal impairment. Oral semaglutide tolerability was consistent with the patient population and the known safety profile of GLP-RAs. Oral semaglutide could benefit patients with T2D and renal impairment. This trial was sponsored by Novo Nordisk and is registered with ClinicalTrials.gov (NCT282778). The authors acknowledge Stephen Purver (Spirit Medical Communications Group Ltd) for medical writing assistance, and Brian Bekker Hansen (Novo Nordisk) for reviewing the poster. Presented at the 79 th Scientific Sessions of the American Diabetes Association Congress, San Francisco, CA, USA, June 7, 9. References: () Davies MJ, et al. Diabetologia 8; 6: ; (2) Aroda VR, et al. Diabetes Metab 9; pii: S :3222 2; (3) Buckley ST, et al. Sci Transl Med 8; : pii: eaar747; (4) Seaquist ER, et al. Diabetes Care 3; 36: ; (5) Davies MJ, et al. Diabetes Care 6; 39: 222 3; (6) Tuttle KR, et al. Lancet Diabetes Endocrinol 8; 6: PIONEER 5_Mosenzon ADA9 poster_approved_3.5.9 INDESIGN.indd 3/5/9 :4:57 Abstract

57 Semaglutide oral Poster presentation 3-P 3-P Effect of Upper Gastrointestinal Disease on the Pharmacokinetics of Oral Semaglutide in Subjects with Type 2 Diabetes Juris J Meier, Charlotte Granhall 2, Ulrike Hoevelmann 3, Andrea Navarria 2, Leona Plum-Moerschel 5, Chethana Ramesh 4, Andrea Tannapfel, Christoph Kapitza 3 St Josef Hospital, Ruhr-University Bochum, Bochum, Germany (juris.meier@rub.de); 2 Novo Nordisk A/S, Søborg, Denmark; 3 Profil, Neuss, Germany; 4 Novo Nordisk, Bangalore, India; 5 Profil, Mainz, Germany Aim Oral semaglutide is a human glucagon-like peptide- analogue co-formulated with the absorption enhancer, sodium N-(8-[2- hydroxybenzoyl] amino) caprylate (SNAC). Oral semaglutide has been shown to improve glycemic control and reduce body weight in patients with type 2 diabetes (T2D). 2,3 Upper gastrointestinal (GI) tract comorbidities, such as chronic gastritis and gastroesophageal reflux disease (GERD), are common in patients with T2D. 4,5 Since the stomach is the main site of absorption of oral semaglutide, the effect of upper GI disease on exposure to oral semaglutide was investigated. Methods Population Male or female subjects with T2D aged 8 8 years with body mass index of kg/m 2 and HbA c of 6. 9.% with or without upper GI disease were included. Subjects in the group with upper GI disease had either chronic gastritis, GERD, or both diseases with diagnosis based on the criteria shown in Table. Table Assessment Criteria for disease diagnosis GERD Symptoms (Yes/No) Symptoms present Chronic gastritis Upper GI disease diagnosis criteria Macroscopic (endoscopy) esophagus Findings of mucosal breaks or other Los Angeles classification clinically significant abnormal of esophagitis: findings indicating GERD no mucosal breaks mucosal breaks grade A, B, C or D Macroscopic (endoscopy) stomach (normal/abnormal) Microscopic (mucosal biopsies) Updated Sydney classification: Chronic inflammation (mononuclear cells) Activity (neutrophils) Atrophy Intestinal metaplasia Helicobacter pylori Findings not applicable for diagnosis At least mild chronic inflammation Each assessment graded as absent, mild, moderate, or marked for the corpus and antrum GERD, gastroesophageal reflux disease; GI, gastrointestinal Subjects were identified and diagnosed with chronic gastritis and/or GERD using a stepwise screening process involving a Helicobacter pylori infection urea breath test (Figure ). All subjects underwent an esophagogastroduodenoscopy and histopathological examination of mucosal biopsies to confirm correct group assignment. Figure Screening and group assignment Step (visit A) Step 2 (visit B) Step 3 (visit C) Basic assessments H. Pylori breath test Matching was revoked in case EGD or biopsies indicated presence of chronic gastritis and/or GERD ECG, electrocardiogram; EGD, esophagogastroduodenoscopy; GERD, gastroesophageal reflux disease; GI, gastrointestinal; H. Pylori, Helicobacter Pylori Trial design This was an open-label, parallel group trial (NCT ). All subjects were administered oral semaglutide 3 mg once daily for five days, followed by 7 mg for five days (Figure 2). Subjects received oral semaglutide with ml of water in a fasting state ( 6 hours) before the first meal of the day. Water was not allowed within 2 hours before dosing. Subjects continued fasting for 3 minutes after oral semaglutide administration. Figure 2 Outcome H. Pylori test Negative Positive Statistical analysis Trial design Screening If no symptoms of GERD; match on body weight to a subject with upper GI disease Symptoms of GERD Lab, ECG, vital signs 3 mg oral semaglutide Lab, ECG, vital signs Lab, ECG, vital signs EGD and biopsies EGD and biopsies EGD and biopsies 7 mg oral semaglutide Group assignment No chronic gastritis or GERD Chronic gastritis and/or GERD GERD ± chronic gastritis Chronic gastritis ± GERD Control group GI disease group GI disease group GI disease group Day 6 Days 5 Days 6 Day, 2 7, 22 Day 47 ±2 days PK, pharmacokinetic Endpoints Dosing and PK sampling PK sampling Follow-up The primary endpoint was the area under the plasma concentration time curve (AUC) for semaglutide from time 24 h after the th dose (AUC 24h,Day ). The secondary endpoint was observed maximum plasma concentration for semaglutide after the th dose (C max,day ). Safety and tolerability of oral semaglutide were also assessed. The primary endpoint, AUC 24h,Day was log-transformed and analyzed using an linear normal model. Mean difference in log-transformed AUC 24h,Day between the two groups was estimated and backtransformed to the original scale and presented as a ratio together with the corresponding two-sided 95% confidence intervals. The secondary endpoint, C max,day, was analyzed similarly. Results A total of 36 subjects with upper GI disease and 9 subjects without upper GI disease were included in the trial (Figure 3); 53 subjects completed the trial. Figure 3 There were no major differences in baseline characteristics between the two groups (Table 2). Table 2 With upper GI disease (n=36) Without upper GI disease (n=9) Age, years 62 (8) 58 (2) Body weight, kg 92.3 (7.) 83.5 (6.) Body mass index, kg/m (3.6) 28.4 (4.5) Male, n (%) 25 (69.4) 3 (68.4) White, n (%) 36 (.) 8 (94.7) HbA c, % 7.2 (.7) 7.2 (.9) Duration of diabetes, years.7 (7.5).7 (5.9) Gastric emptying half-life, minutes Baseline demographic and clinical characteristics 62 (8.) 7 (24.4) Gastric emptying coefficient 2.92 (.46) 2.84 (.66) Gastric emptying was assessed by a 3 C-octanoic acid breath test. Data are mean (standard deviation), except gastric emptying half-life which is geometric mean (coefficient of variation), unless specified otherwise. GI, gastrointestinal Pharmacokinetics Subjects in trial GERD and chronic gastritis n=23 (64%) GERD only n=8 (22%) Chronic gastritis only n=5 (4%) GERD, gastroesophageal reflux disease; GI, gastrointestinal Subjects without upper GI disease, N=9 Subjects with upper GI disease. N=36 The geometric mean concentration time profiles of semaglutide over 24 hours after the th dose in subjects with and without upper GI disease are shown in Figure 4. There were no statistically significant differences in AUC 24h,Day or C max,day between subjects with and without upper GI disease (Table 3). Other secondary endpoints were similar between the groups: median t max,day was. h in both groups, and the geometric mean t ½,Day was 4 h in subjects with upper GI disease and 42 h in subjects without upper GI disease. Figure 4 Geometric mean semaglutide concentration time profiles during a 24-hour dosing interval after the th dose GI, gastrointestinal; SEM, standard error of the mean Table 3 AUC 24h (nmolh/l) Pharmacokinetic endpoints of semaglutide after the th dose N in full analysis set N with evaluable profiles Estimated mean 95% CI P value With upper GI disease , Without upper GI disease , Group ratio (with vs. without upper GI disease) C max (nmol/l).8.8, With upper GI disease , 9.9 Without upper GI disease , 9.23 Group ratio (with vs. without upper GI disease) AUC, area under the plasma concentration time curve; CI, confidence interval; C max, maximum plasma concentration; GI, gastrointestinal Safety Semaglutide (nmol/l) A total of 93 adverse events (AEs) were reported (Table 4)..6.77, All of the AEs were mild (n=73) or moderate (n=) in severity. The most frequently reported AEs were GI disorders (43 events in 8 subjects), including diarrhea and upper abdominal pain, followed by infections and infestations ( events in 9 subjects), particularly nasopharyngitis. There were no deaths or AEs leading to withdrawal With upper GI disease Table 4 Conclusion Without upper GI disease Error bar is ±SEM Reference line for lower limit of quantification. Values below lower limit of quantification are imputed. Time since last dosing (hour) Adverse events by system organ class With upper GI disease (N=36) n (%) E Without upper GI disease (N=9) n (%) E Adverse events 24 (66.7) 77 (52.6) 6 Gastrointestinal disorders 5 (4.7) 37 3 (5.8) 6 Infections and infestations 6 (6.7) 8 3 (5.8) 3 Metabolism and nutrition disorders 8 (22.2) 8 (.) Nervous system disorders 3 (8.3) 3 3 (5.8) 3 E, number of adverse events; GI, gastrointestinal; n, number of subjects with adverse event 24 Key Result There was no significant difference in exposure of semaglutide between the groups and no dose adjustment for oral semaglutide is expected to be required in subjects with upper GI disease. Oral semaglutide was well tolerated in subjects with upper GI disease Semaglutide (nmol/l) This trial was sponsored by Novo Nordisk and is registered with ClinicalTrials.gov (NCT ) The authors acknowledge the medical writing assistance of Andy Bond, Spirit Medical Communications Group Limited, supported by Novo Nordisk A/S. Presented at the 79th Scientific Sessions of the American Diabetes Association Congress, San Francisco, CA, USA, June 7-, 9 References: () Buckley et al. Sci Transl Med. 8; : eaar747 (2) Aroda et al. Abstract 2-LB. In: American Diabetes Association - 78th Annual Scientific Sessions; June 8; Orlando, Florida, USA (3) Rosenstock et al. JAMA. 9; 32: (4) Boehme et al. Hepatogastroenterology. 7; 54: (5) Kim et al. World J Gastroenterol. ; 6: Upper GI Disease PK_ADA 9 poster_6may9 INDESIGN.indd 7/5/9 3:: Abstract

58 Devices Late-breaking poster presentation 26-LB 26-LB Improved insulin adherence after introduction of a smart connected insulin pen Peter Adolfsson, Niels V Hartvig 2, Anne Kaas 3, Nikoline Nygård Knudsen 4, Ann-Charlotte Mårdby 5, Jonas B Møller 6, Jarl Hellman 7 The Hospital of Halland Kungsbacka, Institution of Clinical Sciences University of Gothenburg, Kungsbacka, Sweden; 2 Global Development, Data Science, Novo Nordisk A/S, Søborg, Denmark; 3 Medical & Science Digital Health, Novo Nordisk A/S, Søborg, Denmark; 4 Epidemiology/Digital Health, Novo Nordisk A/S, Søborg, Denmark; 5 Medical Affairs, Novo Nordisk Sweden, Malmö, Sweden; 6 Digital Health, Partnerships & Commercial Strategy, Novo Nordisk A/S, Søborg, Denmark; 7 Department of Medical Sciences, Uppsala University, Uppsala, Sweden qrs.ly/g9xx4j Background The association between missed insulin injections and the impact on HbA c levels in insulin-dependent diabetes is well established, with the unwanted effect of increasing the risk of diabetes-related complications. 4 The smart connected NovoPen 6 collects and stores data on the date and time of insulin injections and the number of units administered. These data are then downloaded using near field connectivity to a centralized database. This allows healthcare professionals (HCPs) and patients to look at injection data together when discussing insulin treatment. If the injection data are further combined with glucose/continuous glucose monitoring (CGM) data the potential to improve patient-hcp dialog is thought to be even greater. The possibility to have a combined view of insulin injections and CGM data and the potential for improved dialog between patients and HCPs can eliminate any guessing about doses taken, missed doses and optimal injection time in relation to meals. An engaging and open patient-hcp dialog has been identified as highly important for optimal disease management, and could reduce the number of missed insulin injections to improve treatment adherence. 5,6 It is therefore of interest to assess whether use of the NovoPen 6 can reduce the number of missed injections in everyday clinical use. Figure : Study design (Pre-baseline) Baseline period Follow-up Aim To investigate whether the use of NovoPen 6 can influence the behavior of patients with type diabetes (TD) in terms of change in numbers of missed bolus dose (MBD) injections. Figure 2: Using NovoPen 6 with the Glooko/Diasend system. NovoPen 6 2. Data capture HCP Patient HCP, healthcare professional. In-clinic data box Blood glucose data 3. Data storage Storage of data on cloud 4. HCP review of data Visualization of patient data (insulin dosing and glucose data) on same interface using HCP platform Anonymized data transfer to Novo Nordisk for research and product development purposes Hereafter, the study continued with HCP visits according to clinical practice. At each visit pen data were available for download and use by the patient and HCP during the consultation (Figures and 2). With this study design, it was possible to compare the number of MBD injections between the baseline and follow-up periods. CGM and dosing data from the first 4 days following a clinic visit were used in the analyses. The 4-day time period was chosen to be in line with the international consensus on the use of CGM. 7 Visit 5 was chosen as the earliest point for follow-up, as patients would on average have been in the study for 8 days, allowing for sufficient interaction with HCPs and discussion of available pen data. MBDs were identified using the clinically validated Glucose Rate Increase Detector (GRID) algorithm 8 to detect meals from the CGM signal. A MBD was defined as an occasion where no bolus injection had occurred within 5 to +6 minutes from the start of a meal, as detected by the algorithm (Figure 3). Figure 3: Detection of missed bolus insulin doses by the GRID algorithm Glucose (mmol/l) 5 5 Feb 7 : Feb 7 6: Feb 7 2: Feb 7 8: Feb 8 : Statistical analyses Pen and CGM data for each patient were linked based on patient IDs. Data from days with unacceptable CGM coverage (<7%) or where bolus injections were not available, were excluded. Each day was aggregated to the number of MBD meals, the number of on-time meals and total number of meals. A generalized linear mixed model based on the Poisson distribution was applied with visit number (baseline,, 2, 3, 4, 5+) as fixed effect and patient and visit nested within patient as random effects. The model allows for unbalanced and missing data. The estimated difference between the follow-up period (visits 5) and the baseline period was obtained on the logarithmic-scale. Estimates and 95% confidence intervals were converted to the original scale. Results Eighty-one adults with TD with a mean [min; max] age of 39.2 years [8; 83] were included in these analyses. A total of 892 days were analyzed. A significant decrease of 43.% in the average daily number of MBD injections was observed from the baseline period to the follow-up period, from.74 (95% CI [.62;.88]) to.42 (95% CI [.3;.6]) (p=.2) (Figure 4 and Table ). Based on the assumption that patients have three main meals per day, this corresponded to a decrease from 24.7% (95% CI [.8; 29.4]) to 4.% (95% CI [9.9; 9.9] ) in MBD injections (Table ). A significant increase in the number of daily, undetected meals was observed from the baseline period to the follow-up period, from.54 (95% CI [.37;.7]) to.94 (95% CI [.69; 2.4]) (Figure 4 and Table ). These results indicate that patients achieved more well-dosed meals, as indicated by the slight increase in the number of on-time doses observed in the follow-up period compared with the baseline period (Table ). The increase was not statistically significant, however, because well-dosed meals tend to have a lower CGM response and are as such undetected by the GRID algorithm (Figure 3). Figure 4: Mean number of daily meals and dosing behaviors from the baseline period to the followup period Average number of meals per day p=.2 MBD Baseline NS On-time dose Follow-up p=.3 Undetected Key result 6 4 Percent of three meals CGM Baseline Visit Visit 2 Visit 3 Visit 4 Visit 5 Insulin dose, time and date Pre-baseline was the period before study commencement where patients were already using CGM, but without concurrent use of the NovoPen 6. CGM, continuous glucose monitoring. Methods This pilot study was a prospective, non-interventional study running from May 7 Nov 8. Twelve diabetes clinics from different parts of Sweden participated. Patients with TD using CGM were included if their treating physicians decided to offer them a NovoPen 6. At baseline, patients received a NovoPen 6 for basal and/or bolus insulin injections. Baseline was then followed by a baseline period between pen introduction and visit, during which the patient started to use the NovoPen 6 but without access to downloads of injection data. The first data download occurred at visit, using the Glooko/Diasend in-clinic system to transfer data from the pen to the Glooko/Diasend server. From here the data were accessed via the Glooko/Diasend HCP web portal and the patient and HCP had the first chance to look at the data together. Bolus dose (U) Feb 7 : Feb 7 6: On-time dose MBD Feb 7 2: Feb 7 8: Feb 8 : Example of a day with two meals detected. The solid dark blue line represents the CGM signal and the light blue shaded areas each represent a detected meal. The gray, dashed line represents a glucose level of 7.2 mmol/l and the gray shaded area represents a target glycemic range of 3.9 mmol/l, as previously reported. 6 Meals are detected when the CGM signal is 7.2 mmol/l and increases steeply over 3 45 minutes. A bolus dose within 5 minutes before to 6 minutes after a meal starts is considered on-time, whereas a dose outside of this time window is considered a MBD. Male patient, aged 3 at baseline. CGM, continuous glucose monitoring; GRID, Glucose Rate Increase Detector; MBD, missed bolus dose. Table : Mean number of daily meals and dosing behaviors from the baseline period to the follow-up period Estimated relative change [95% CI] MBD 43.% [ 6.5; 8.] On-time dose Undetected meals 2.7% [ 24.7; 4.2] 25.4% [8.7; 43.7] Baseline level [95% CI] Daily meals (n).74 [.62;.88].57 [.48;.69].54 [.37;.7] Follow-up level [95% CI] Proportion Daily meals Proportion of 3 meals (n) of 3 meals 24.7% [.8; 29.4] 9.% [5.9; 23.] 5.5% [45.6; 56.7].42 [.3;.6].59 [.43;.8].94 [.69; 2.3] 4.% [9.9; 9.9] 9.6% [4.5; 26.7] 64.6% [56.4; 7.] P value Mean and 95% CI based on a mixed Poisson model, with visit number (baseline,, 2, 3, 4, 5+) as fixed effect and patient and visit nested in patient as random effects. Assuming 3 meals per day on average. CGM, continuous glucose monitoring; CI, confidence interval; HCP, healthcare professional; MBD, missed bolus dose; n, number. Estimated mean number of daily meals with 95% confidence intervals. MBD are meals with missed bolus doses. On-time doses are meals where a bolus dose is taken. Undetected are meals that are not detected by the CGM signal, assuming an average of three meals per day. CGM, continuous glucose monitoring; HCP, healthcare professional; MBD, missed bolus dose; NS, not significant. Conclusion These real-world findings confirm that missed bolus dose injections are the reality for patients with TD and that the smart connected NovoPen 6 can support good injection behavior, with fewer missed and more welldosed mealtime injections. This could subsequently lead to better glycemic control and thus lower the risk of diabetes-related complications. The study was sponsored by Novo Nordisk. Presenter Peter Adolfsson has received research support or advisory board fees from Eli Lilly, Novo Nordisk, Roche, and Research and Development, Region Halland, and is an employee of Region Halland. The authors are grateful to Melissa Voigt Hansen, Novo Nordisk for review of and input to the poster and to Jon Griffiths, Watermeadow Medical (supported by Novo Nordisk) for writing assistance. Presented at the American Diabetes Association, 79th Scientific Sessions. June 7, 9, San Francisco, CA, USA. References: () The Diabetes Control and Complications Trial Research Group. N Engl J Med 993;329:977 86; (2) Nathan et al. N Engl J Med 5;353: ; (3) Holman et al. N Engl J Med 8;359:565 76; (4) Munshi et al. Diabetes Care 3;36:543 9; (5) Heisler et al. J Gen Intern Med 2;7:243 52; (6) Ritholz et al. Chronic Illn 4;:33 3; (7) Danne et al. Diabetes Care 7;4:63 4; (8) Harvey et al. J Diabetes Sci Technol 4;8:37. Abstract

59 Devices Poster presentation 76-P 76-P Increased time in range (TIR) observed after introduction of a connected insulin pen Peter Adolfsson, Niels V Hartvig 2, Anne Kaas 3, Nikoline Nygård Knudsen 4, Ann-Charlotte Mårdby 5, Jarl Hellman 6 The Hospital of Halland Kungsbacka, Institution of Clinical Sciences University of Gothenburg, Kungsbacka, Sweden; 2 Global Development, Data Science, Novo Nordisk A/S, Søborg, Denmark; 3 Medical & Science Digital Health, Novo Nordisk A/S, Søborg, Denmark; 4 Epidemiology/Digital Health, Novo Nordisk A/S, Søborg, Denmark; 5 Medical Affairs, Novo Nordisk Sweden, Malmö, Sweden; 6 Department of Medical Sciences, Uppsala University, Uppsala, Sweden qrs.ly/g9xx4j Background Insulin pens have become the most widely used devices for delivering insulin. Despite their convenience, however, there are shortcomings. In particular, poor documentation of insulin therapy can result in inadequate glycemic control for patients with diabetes. Smart insulin pens offer automatic access to insulin injection data, and could help overcome barriers of poor adherence, clinical inertia and incorrect dosing. The smart connected NovoPen 6 collects and stores data on the date and time of insulin injections and the number of units administered. These data are then downloaded using near field connectivity to a centralized database. This allows healthcare professionals (HCPs) and patients to look at injection data together when discussing insulin treatment. If the injection data are further combined with glucose/continuous glucose monitoring (CGM) data the potential to improve patient-hcp dialog is thought to be even greater. The possibility to have a combined view of insulin injections and CGM data gives the HCP and the patient a more complete picture of current glycemic status. Thus, both patient-hcp dialog and treatment approaches can be improved. An engaging and open patient-hcp dialog has been identified as highly important for optimal disease management. Therefore, the NovoPen 6 has the potential to improve glycemic control. 2,3 Figure : Study design (Pre-baseline) CGM Baseline period Baseline Visit Visit 2 Visit 3 Visit 4 Visit 5 Insulin dose, time and date Follow-up Pre-baseline was the period before study commencement where patients were already using CGM, but without concurrent use of the NovoPen 6. CGM, continuous glucose monitoring. Aim The objective of this non-interventional study was to investigate how a smart connected insulin pen (NovoPen 6) influences glycemic control in patients with type diabetes (TD) in a real world setting. Figure 2: Using NovoPen 6 with the Glooko/Diasend system. NovoPen 6 2. Data capture HCP, healthcare professional. Methods HCP Patient In-clinic data box Blood glucose data 3. Data storage Storage of data on cloud 4. HCP review of data Visualization of patient data (insulin dosing and glucose data) on same interface using HCP platform Anonymized data transfer to Novo Nordisk for research and product development purposes This pilot study was a prospective non-interventional study running from May 7 Nov 8. Twelve diabetes clinics from different parts of Sweden participated. Patients with TD using CGM were included if their treating physicians decided to offer them a NovoPen 6. At baseline, patients received a NovoPen 6 for basal and/or bolus insulin injections. Baseline was then followed by a baseline period between pen introduction and visit, during which the patient started to use the NovoPen 6 but without access to downloads of injection data. The first data download occurred at visit, using the Glooko/Diasend in-clinic system to transfer data from the pen to the Glooko/Diasend server. From here the data were accessed via the Glooko/Diasend HCP web portal and the patient and HCP had the first chance to look at the data together. Hereafter, the study continued with HCP visits according to clinical practice. At each visit pen data were available for download and use by the patient and HCP during the consultation (Figures and 2). This study design permitted comparison between the baseline and follow-up periods. CGM and dosing data from the first 4 days following a clinic visit were used in the analyses. The 4-day period was chosen to be in line with the international consensus on the use of CGM. 4 Visit 5 was chosen as the earliest point for follow-up, as patients would on average have been in the study for 8 days, allowing for sufficient interaction with HCPs and discussion of available pen data. Time in range (TIR), time spent in hyperglycemia and time spent in L ( mmol/l) and L2 hypoglycemia (<3. mmol/l) were compared between the baseline and follow-up periods, which was defined as any point after the fifth HCP visit. Table : Baseline levels and estimated changes to follow up of key glycemic parameters Baseline level [95% CI] TIRange (3.9 mmol/l) (hours) 9.9 [8.28;.] TIHyper (> mmol/l) (hours).8 [.8; 2.79] TIHypo L ( mmol/l) (hours) TIHypo L2 (<3. mmol/l) (hours) Mean glucose (mmol/l) Coefficient of variation (%).72 [.52;.92].47 [.32;.6].9 [.53;.64] [34.33; 37.45] Estimated mean change [95% CI].89 [.79; 2.99].78 [ 2.96;.6].5 [.36;.7].33 [.56;.].34 [.96;.28] 3.84 [ 6.2;.56] p-value Estimated mean baseline level and change between the follow-up period (visits 5) and the baseline period with 95% CI. Linear mixed model, with visit number (baseline,, 2, 3, 4, 5+) as fixed effect, patient and visit nested in patient as random effects, and with exponential covariance function. N=94, visits=23, CGM days=2552. CGM, continuous glucose monitoring; CI, confidence interval; TIRange, time in range (3.9 mmol/l); TIHyper, time in hyperglycemia (> mmol/l); TIHypo L, time in L hypoglycemia ( mmol/l); TIHypo L2, time in L2 hypoglycemia (<3. mmol/l). Results Ninety-four adults with TD with a mean [min; max] age of 4. years [8; 83] were included in the analyses. A total of 64 patients used NovoPen 6 for bolus insulin only, 7 for basal and bolus insulin and 5 for basal insulin only. For the majority, insulin degludec was the basal insulin and insulin aspart was the bolus insulin. Seven patients did not have connected pen data in the 4-day periods studied and patient used biphasic insulin aspart 3, neither bolus nor basal insulin (Figure 3). A significant increase of.9 hours per day (~2% of the baseline level) in mean TIR from the baseline period to the follow-up period was observed (p=.9; Figure 4 and Table ). Accordingly, a significant reduction in mean time spent in hyperglycemia (> mmol/l) and L2 hypoglycemia (<3. mmol/l) of.8 hours per day (p=.3) and.3 hours per day (p=.5), respectively, was also observed (Figure 4 and Table ). There was no significant change in mean time spent in L hypoglycemia ( mmol/l; p=.8; Figure 4 and Table ). While the mean glucose level did not change significantly, the coefficient of variation was reduced by 3.8% from the initial level of 35.9% (Table ). This shows that the improved TIR is obtained primarily by more stable glucose levels over the day. In terms of bolus insulin dose (n=8), a significant increase from the baseline period to the follow-up period of 28%, to a dose of 32. U/day was observed. There was no significant change in mean basal insulin dose (n=22). Figure 3: Patient treatment characteristics Glucose monitoring technique CGM (5 mins) FGM (5 mins) 83 Human insulin Bolus Insulin Insulin aspart 79 Basal & Bolus 7 Use of connected pen Mix Basal 5 Bolus 64 Faster-acting insulin aspart Basal Insulin Insulin detemir Insulin degludec 2 Interval between CGM readings. Numbers indicate numbers of patients. Seven patients did not have connected pen data at any of the CGM days studied. One patient used biphasic insulin aspart 3 that is neither considered bolus nor basal insulin in the analysis. CGM, continuous glucose monitoring, FGM, flash glucose monitoring. Figure 4: Mean difference in the time spent in glycemic ranges from the baseline period to the follow-up period Difference from baseline (hours) 2 2 p<.5. Estimated mean difference in time spent in glycemic ranges with 95% CI. The difference is observed between the baseline period and the follow-up period. Baseline is the period after treatment initiation but before the first visit. Analysis is based on CGM data from a 4-day interval after each visit ( 7% coverage). CGM, continuous glucose monitoring; CI, confidence interval; HCP, healthcare professional; TIRange, time in range (3.9 mmol/l); TIHyper, time in hyperglycemia (> mmol/l); TIHypo L, time in L hypoglycemia ( mmol/l); TIHypo L2, time in L2 hypoglycemia (<3. mmol/l). Patients above 8 years (n=94) are included. Conclusion TIRange (3.9 ) TIHyper (>) Glycemic level (mmol/l) Key result TIHypo L (3. 3.9) TIHypo L2 (<3.) These real-world findings in patients with TD highlight the potential benefit to glycemic control when connected pen data contribute to the patient-hcp dialog. Patients with a smart connected pen obtained more stable CGM profiles, with more time in range and less time spent in hyperglycemia and hypoglycemia. The study was sponsored by Novo Nordisk. Presenter Peter Adolfsson has received research support or advisory board fees from Eli Lilly, Novo Nordisk, Roche, and Research and Development, Region Halland, and is an employee of Region Halland. The authors are grateful to Melissa Voigt Hansen, Novo Nordisk, for review of and input to the poster, and to Elizabeth Hilsley, Watermeadow Medical (supported by Novo Nordisk) for writing assistance. Presented at the American Diabetes Association, 79th Scientific Sessions. June 7, 9, San Francisco, CA, USA. References: () Klonoff and Kerr. J Diabetes Sci Technol 8;2:55 3; (2) Heisler et al. J Gen Intern Med 2;7:243 52; (3) Ritholz et al. Chronic Illn 4;:33 3; (4) Danne et al. Diabetes Care 7;4:63 4. Abstract

60 Diabetes research Late-breaking poster presentation 89-LB 89-LB Low level of agreement in evaluation of continuous glucose monitoring profiles Elise Hachmann-Nielsen ; Kajsa Kvist ; Richard M. Bergenstal 2 Novo Nordisk A/S Søborg, Denmark; 2 International Diabetes Center, Minneapolis, U.S.A qrs.ly/ul9z2dc Introduction After the introduction of Continuous Glucose Monitoring (CGM) it has become recognized that CGM gives the opportunity for better and more informed diabetes management compared to using HbA c alone. Nevertheless, the large amount of information inherent to a CGM-profile reading may complicate the evaluation of the profiles and lead to suboptimal clinical application of the information. In 7, the International Consensus on Use of Continuous Glucose Monitoring was published to provide guidance for clinicians, patients, and researchers in utilizing, interpreting, and reporting CGM data in clinical care and research. [] We aimed to investigate whether there is agreement in assessing glucose profiles, with focus on which profiles are better with respect to glucose management. Methods To investigate the level of agreement, we created a simple, publicly available webpage ( where users can rank 9 Ambulatory Glucose Profiles (AGPs) (Figure and 2) along with prioritized argumentation for the choices and submit background information about themselves. AGP has been recommended as a standard for visualization of CGM data since 7. [-3] Figure The webpage Figure notes: The landing page from the crowdsourcing webpage The webpage is continuously active and contributions can still be made to this research by accessing the page. All participant have accepted their responses to be used for research and publication. This webpage was launched online with free access. The webpage was introduced at the Diabetes Technology Meeting (DTM) 8 in Washington and shared via [4] The glucose profiles were randomly selected from the RWE database Cornerstones4Care. [5] We investigated the interobserver variation between responders when asked to subjectively prioritize the 9 AGPs from best to worst in their own opinion. This was analyzed both within the full group of responders and divided the groups into self-evaluated experts (> years of experience and > CGM evaluation per week) and non-experts. The level of agreement was also evaluated from 3 validated experts (all highly involved in patient management and research within the CGM-field) who also ranked the 9 profiles via the same webpage. The interobserver variation in the four populations was determined using the Interclass Correlation Coefficient (ICC) a descriptive statistic that allows for qualitative assessment of agreement. [6] We hypothesized that there would be a relatively low level of agreement in ranking of glucose profiles partly due to the possible individual prioritization of the importance of the key CGM metrics. We also hypothesized that the level of agreement would be higher among the experts than the non-experts and that the highest level of agreement would be among the validated experts. Figure 2 The 9 Ambulatory Glucose Profiles as seen on Results A total of 96 responses were collected. The overall group (N=96) consisted of a broad range of expertise, including physicians (diabetologists and endocrinologists, 36%), nurses (3%), other (6%) e.g. patients, software developers and engineers, who had never/sometimes/often been in contact with CGM-profiles and showed moderate agreement within the all-raters group (ICC =.65). [Figure 3a] The non-experts (N=77) was the subgroup of raters with little or no familiarity but consisted also of both physicians (22%), nurses (3%) and other (75%); they likewise showed moderate agreement similar to the all-raters group (ICC =.62). [Figure 3b] Within the self-evaluated experts (N=9), there were only physicians (%) and here the degree of consensus was good (ICC=.77). [Figure 3c] Within the validated experts (N=3) the consensus was excellent (ICC =. 92). On the webpage, the profiles are not labelled with number or letter but can be freely rearranged using drag-and-drop. The letters inserted here corresponds to the graphs in figure 3. Figure 3 G D A B I H C E F a) b) c) Interclass Correlation Coefficients and distribution of glucose profile ranking (a) all, (b) non-experts, (c) experts Key Result Discussion Crowdsourcing from a publicly available webpage allows for unfiltered information from a diverse population. The overall group included both physicians and nurses deeply involved in CGM management, HCPs with very little familiarity to guidelines and/or practical experience as well as patients, software developers and engineers. The non-experts all had less than years experience and/or evaluated less than CGM per week. The low level of agreement in the overall group and non-expert subgroup is not surprising and merely represent the difficulty for intuitive interpretation of CGM profiles or use of published standard AGP evaluation tools. Within the expert-group, all indicated to have at least years of experience and at least weekly assessment of glucose profiles. They must be expected to have both practical experience and be updated on the International Consensus guideline - however there was still relatively low level of agreement indicating a challenge in the evaluation. This low level of agreement in the self-evaluated experts may be due to 4 key metrics being suggested by the 7 Consensus panel without guidance on how to prioritize metrics and no targets for the metrics being provided. The three validated experts are all highly involved in patient management and research within the CGM-field and showed clearly a higher level of agreement. This must be the aim to ensure standardized evaluation and hereby treatment of patients. To simplify the evaluation, many of the key metrics in a CGM profile may benefit from a combined evaluation in a composite metric, but this has not yet been implemented. Conclusion Despite published guidance and consensus there is still a relatively low level of agreement in the evaluation of CGMprofiles even between diabetes experts. This allows for potentially undesired difference in diabetes management. More data and tools (AI-assisted or a composite metric) may be needed to assist diabetes management with CGM in the future. This study was sponsored by Novo Nordisk A/S Presented at the American Diabetes Association, 79th Scientific Sessions. June 7, 9, San Francisco, CA, USA. 8 more ratings has been collected since abstract submission References: () Danne et al. Diabetes Care 7 Dec; 4(2): 63-64; (2) (3) Bergenstal et al. Diabetes Technol Ther. 3 Mar;5(3):98-2 ; (4) (5) 8 (Diabetes Technology Meeting) November 8-, 8; Rockville, MD; Day #2 Highlights; (6) Koo et al. J Chiropr Med. 6 Jun; 5(2): 55 63; Abstract