Emergence of the New Classes of Compounds for Managing Heart Disease in Diabetes Mellitus

Transcription

1 Emergence of the New Classes of Compounds for Managing Heart Disease in Diabetes Mellitus Mikhail Kosiborod, MD Professor of Medicine (Cardiology) Saint Luke s Mid America Heart Institute University of Missouri-Kansas City

2 Research Grants: Disclosures AstraZeneca, Boehringer Ingelheim Consultant/Advisory Board: AstraZeneca, Sanofi, GSK, Amgen, Boehringer-Ingelheim, Novo Nordisk, Merck (Diabetes), Eisai, Janssen, Intarcia, ZS Pharma, Glytec

3 Take Home Points Fundamental paradigm shift in T2D management CVD remains the main cause of death and disability in T2D Intensive glucose control has not convincingly reduced CV events Several classes of glucose-lowering therapies improve CV outcomes seen in RCTs and real world studies These emerging data should shift focus of T2D Rx from HbA1c alone to comprehensive CV risk reduction

4 Causes of Death in Type 2 Diabetes 25 Percent of U.S. deaths in individuals with diabetes Is c hemic Heart Disease CV Disease CHF Vascular Causes of Death Other Heart Disease Cancer Liver Disease Renal Disease Nonvascular Causes of Death P neumonia National Vital Statistics System 2008

5 T2DM and CV Events Over Time Rawshani et al. New Engl J Med. 2017

6 The Puzzles of Diabetes and Its Complications MACROVASCULAR COMPLICATIONS MICROVASCULAR COMPLICATIONS GLUCOSE SMOKING BP GENETICS INFLAMMATION OBESITY GENETICS GLUCOSE BP INSULIN RESISTANCE AGE LIPIDS Courtesy of Silvio Inzucchi MD, Yale University. Fruchart JC. Circulation 2004;109:III15-III19. Yau JW et al. Diabetes Care 2012;35: MJ Fowler. Clin Diabetes 2008;26:77-82.

7 Impact of Intensive Glucose-Lowering Therapy In DM: Summary of Major RCTs Study Microvasc CVD Mortality UKPDS 33 (7.0 vs. 7.9%) DCCT / EDIC* (7.2 vs. 9.1%) ACCORD (6.4% vs. 7.5%) Initial Trial Long Term F/U ADVANCE (6.3% vs. 7.0%) VADT (6.9% vs. 8.4%) *in T1DM. Courtesy of Silvio Inzucchi MD, Yale University. Adapted: Kendall DM, Bergenstal RM. International Diabetes Center 2009, 2015 UKPDS Group. Lancet 1998;352:854; Holman RR. NEJM 2008;359:1577; DCCT Group. NEJM 1993;329;977; Nathan DM. NEJM 2005;353:2643. Gerstein HC. NEJM 2008;358:2545; Patel A. NEJM 2008;358:2560; Duckworth W. NEJM 2009;360:129. (erratum:361:1024); DCCT Group. JAMA 2015;313:45; Zoungas S. NEJM 2014;371:1392; Hayward RA NEJM 2015;372:23.

8 ACCORD HbA 1c and CV Death

9 Intensive Glucose Control and Hospitalization For Heart Failure In Type 2 Diabetes Trials Number of events (annual event rate, %) More intensive Less intensive HbA 1c (%) Favours more intensive Favours less intensive Hazard ratio (95% CI) Hospitalised/fatal heart failure ACCORD ADVANCE UKPDS VADT Overall 152 (0.90) 124 (0.75) (0.83) 231 (0.88) (0.06) 6 (0.11) (1.80) 85 (1.94) ( ) 0.95 ( ) 0.55 ( ) 0.92 ( ) 1.00 ( ) (Q=3.59, p=0.31, I 2 =16.4%) FM Turnbull et al. Diabetologia (2009) 52: Hazard ratio (95% CI)

10 Summary Points Hyperglycemia is associated with higher risk of CV complications Large trials of intensive glucose control have failed to reduce adverse CV events, despite dramatic improvements in HbA 1c

11 Pre-marketing Analyses Upper CL of 95% CI <1.8 For a HR= events Post-marketing Analyses Upper CL of 95% CI <1.3 For a HR= events sponsors should demonstrate that the therapy will not result in an unacceptable increase in CV risk Hazard Ratio 1.8 Courtesy of Silvio Inzucchi MD, Yale University. Courtesy, Darren McGuire, UTSW Medical Ctr, Meta-analysis strategy using Phase 2/3 data Blinded central adjudication of CVD events in Phase 2/3 Inclusion of high-risk subjects: advanced CVD, elderly, CKD Minimum exposure of 2 years in large CVOT Approximately 15,000 pt-yrs

12 Saxagliptin and Cardiovascular Outcomes In Patients With Type 2 Diabetes Mellitus SAVOR-TIMI 53 Investigators, NEJM 2013;369:

13 Alogliptin After ACS In T2DM Patients Time To First Endpoint EXAMINE Investigators, NEJM 2013;369:

14 Sitagliptin Effect On CV Outcomes In T2DM Rates of Primary CV Outcomes Hazard ratio, 0.98 (95% CI, ) P=0.65 TECOS Study Group, NEJM.org June 26, 2015.

15 Outcome No CV Benefit With Lixisenatide In ELIXA: Intention To Treat Analysis Primary outcome (CV death, nonfatal MI, nonfatal stroke, or hospitalization for unstable angina) Lixisenatide (n=3034) Placebo (n=3034) HR (95% CI) 13.4% 13.2% 1.02 ( ) Cardiovascular death 5.1% 5.2% 0.98 ( ) MI (fatal/non-fatal) 8.9% 8.6% 1.03 ( ) Stroke (fatal/nonfatal) 2.2% 2.0% 1.12 ( ) Unstable angina 0.4% 0.3% 1.11 ( ) Primary outcome plus hosp for HF 15% 15.5% 0.97 ( ) Hospitalization for heart failure 4.0% 4.2% 0.96 ( ) Primary outcome plus HF hosp plus coronary revasc. 21.8% 21.7% 1.00 ( ) Medscape. June 8, 2015.

16 Large Non-Insulin CVOTs In T2DM Study SAVOR EXAMINE TECOS CAROLINA CARMELINA DPP4-i saxagliptin alogliptin sitagliptin linagliptin linagliptin Comparator placebo placebo placebo sulfonylurea placebo N 16,500 5,400 14,000 ~6,000 ~7000 Results June Study LEADER ELIXA SUSTAIN 6 EXSCEL REWIND GLP1-RA liraglutide lixisenatide semaglutide exenatide LR dulaglutide Comparator placebo placebo placebo placebo placebo N 16,500 14,000 3,297 14,752 9,622 Results Study EMPA-REG CANVAS DECLARE NCT SGLT-2-i empaglifozin canagliflozin dapagliflozin ertugliflozin Comparator placebo placebo placebo placebo N 7300 ~10,000 17,160 ~8,000 Results Sept Courtesy of Silvio Inzucchi MD, Yale University.

17 LEADER SUSTAIN 6 IRIS

18 Key inclusion criteria Key exclusion criteria ELIXA (N=6068) LEADER (N=9340) SUSTAIN-6 (N=3297) EXSCEL (N=14,752) ACS 180 d of screening A1C <5.5% or >11% PCI within 15 days of screening Planned coronary revasc within 90 d post-screening egfr <30 ml/min/1.73m 2 Preexisting CVD, cerebrovasc disease, vascular disease, renal or HF at 50 y with pre-existing CVD OR 60 y with CVD risk A1C <5.5% or >11% GLP-1RA or DPP-4 inhibitor use within 3 m of screening Calcitonin >50 ng/l Acute coronary/cerebrovasc event in previous 14 days Planned coronary revasc. Chronic HF (NYHA class IV) Preexisting CVD at 50 y OR pre-cvd 60 y GLP-1RA or pramlintide use within 90 before screening DPP-4 inhibitor use within 30 days before screening Calcitonin 50 ng/l Acute coronary or CV event w/in 90 d of randomization Planned coronary revasc. Chronic HF (NYHA class IV) A1C 6.5% to 10% (inclusive), any level of CV risk, up to 3 OADs, or insulin alone or in combo with up to 2 OADs A1C <6.5% or >10% 2 episodes of severe hypo within 12 months Current/prior GLP-1 RA use egfr <30 ml/min/1.73m 2 Prior pancreatitis Personal or family history of MEN-2 Baseline calcitonin >40 mg/l Mean age, y Male, % White race, % Black race, % Asian race, % Other, race, % Hispanic, % BMI, kg/m HbA1c, % DM Duration, y Current smoker, % Previous smoker, % Total Chol, mg/dl 153 (44) (45.3) Not reported Not reported LDL 78 (35) 89.5 (35.5) 82.3 (45.6) ~95 Triglycerides 164 (113) (140.0) Not reported ~200 Pfeffer MA, et al. N Engl J Med 2015;373: ; Marso SP et al. N Engl J Med 2016;375:311-22; Marso SP et al. N Engl J Med 2016;375: ; Holman R et al. N Engl J Med Epub ahead of print

19 ELIXA (N=6068) LEADER (N=9340) SUSTAIN-6 (N=3297) EXSCEL (N=14,752) % with CV disease Coronary artery disease, % Stable angina pectoris, % UA pectoris,% MI, % (lira)/30.0 (pbo) PCI, % 16.5 Revasc: CABG, % 7.9 (lira)/38.6 (pbo) Stroke, % Transient ischemic attack, % (lira)/16.6 (pbo) Peripheral artery disease, % Hypertension, % Heart failure, % Mortality rate (events/100 pt-y) egfr, % <30 ml/min/1.73m ml/min/1.73m ml/min/1.73m 2 90 ml/min/1.73m Qualifying ACS event, n (%) STEMI Non-STEMI UA BL albumin/ creatinine ratio, mg/g, n (%) < (44.0) 2348 (38.7) 1042 (17.2) 4441 (74.3) 1148 (19.2) 389 (6.5) Not reported Not reported Not reported Pfeffer MA, et al. N Engl J Med 2015;373: ; Marso SP et al. N Engl J Med 2016;375:311-22; Marso SP et al. N Engl J Med 2016;375: ; Holman R et al. N Engl J Med Epub ahead of print.

20 LEADER Trial: Primary Outcomes CV Death, Non-fatal MI or Non-fatal Stroke CV Death The cumulative incidences were estimated with the use of the Kaplan Meier method, and the hazard ratios with the use of the Cox proportional-hazard regression model. The data analyses are truncated at 54 months, because less than 10% of the patients had an observation time beyond 54 months. Marso SP et al, NEJM, 2016.

21 Marso SP et al, NEJM, SUSTAIN 6: CV Outcomes

22 Primary composite: MACE-3 (CV death, nonfatal MI, nonfatal stroke) Intention-to-Treat Analysis for Non-inferiority & Superiority HR (95% CI) 0.91 (0.83, 1.00) P value (non-inferiority) <.001 P value (superiority) 0.061

23 Primary composite cardiovascular outcome components Exenatide N=7356 Placebo N=7396 Hazard Ratio 95% CI P value MACE 839 (11.4%) 3.7 per 100 pt-yrs 905 (12.2%) 4.0 per 100 pt-yrs , 1.00 <.001 (non-inferiority) (superiority) CV-death Non-fatal MI Non-fatal stroke 229 (3.1%) 455 (6.2%) 155 (2.1%) 258 (3.5%) 470 (6.4%) 177 (2.4%) , , , (homogeneity among components) Exenatide favoured Placebo favoured

24 Secondary endpoints Exenatide N=7356 Placebo N=7396 Hazard Ratio 95% CI P value All-cause mortality 507 (6.9%) 584 (7.9%) , CV-death 340 (4.6%) 383 (5.2%) , Fatal or non-fatal MI 483 (6.6%) 493 (6.7%) , Fatal or non-fatal stroke 187 (2.5%) 218 (2.9%) , Hospitalisation for ACS 602 (8.2%) 570 (7.7%) , Hospitalisation for heart failure 219 (3.0%) 231 (3.1%) , Exenatide favoured Placebo favoured

25 Meta-anlaysis: MACE-3 endpoint Bethel et al. TLDE 2017; in press

26 Meta-analysis: All-cause mortality Bethel et al. TLDE 2017; in press

27 Number needed to treat to prevent one CV: cardiovascular; MACE: major adverse cardiovascular event. Presented at the American Diabetes Association 76 th Scientific Sessions, Session 3-CT-SY24. June , New Orleans, LA, USA.

28 Antihyperglycaemic medication drop-in Proportion (%) Metformin Sulfonylureas Thiazolidinediones DPP-4i SGLT-2i* GLP-1 Receptor Agonists** Exenatide Placebo Subjects are considered to have taken a new medication if there is no indication of usage at baseline visit as well as indication of usage during at least one post-randomisation visit. Percentages presented are for available data. * Information regarding SGLT-2 inhibitor usage was added to the ecrf on 9th May, Percentages presented are for ITT population. ** Includes GLP-1 receptor agonist usage other than study medication Insulin

29 Censoring Analysis MACE-3 Endpoint Censored medication* # Uncensored Events (rate per 100 pt-yrs) Exenatide / Placebo Exenatide / Placebo Hazard Ratio 95% CI P value None 7356 / (3.7) / 905 (4.0) , Metformin 7039 / (3.7) / 865 (4.0) , Sulfonylurea 6900 / (3.7) / 861 (4.1) , DPP-4i 6886 / (3.7) / 856 (4.0) , Insulin 6793 / (3.7) / 821 (4.0) , SGLT-2i 7138 / (3.7) / 895 (4.0) , GLP-1 RA 7215 / (3.7) / 883 (4.0) , SGLT-2i + GLP-1 RA 7340 / (3.7) / 903 (4.0) , Any 5612 / (3.6) / 705 (4.1) , *Patients starting concomitant medications censored at the time of drop-in Exenatide favoured Placebo favoured

30 Summary: GLP-1 RA class effects For the class, there were significant reductions in MACE-3 CV mortality All cause mortality No differences in Hypoglycemia Pancreatitis Pancreatic cancer A favorable riskbenefit balance

31 REWIND (dulaglutide) Ongoing CVOTs of GLP1 RAs 9622 patients Broad risk profile Estimated completion July 2018 HARMONY Outcomes (albiglutide) 9575 patients Established CVD Completion date March 2018 Several others planned

32 LEADER SUSTAIN 6 IRIS

33 EMPA-REG (N=7034) CANVAS (N=10,143) Key inclusion criteria Established CVD 30 with a history of symptomatic ASCVD, or 50 with 2 or more CV risk factors; antihyperglycemic agent treated or naïve Key exclusion criteria Drug naïve A1C 7% & 9% or on therapy 7% & 10% History of DKA, T1D, pancreas or beta-cell transplantation, or Uncontrolled hyperglycemia with glucose >240 mg/dl diabetes secondary to pancreatitis or pancreatectomy Planned cardiac surgery or angioplasty within 3 m Unstable diabetes regimen (<8 weeks) ACS, stroke, or TIA within 2 months History of 1 severe hypo episode prior to screening egfr <30 ml/min/1.73m 2 Established CVD, % CAD, % Stable angina, % Unstable angina,% MI, % PCI, % CABG, % Stroke, % TIA, % PAD, % Hypertension, % 94 (on anti-hypertensive therapy) 90.0 Heart failure, % egfr, % <30 ml/min/1.73m ml/min/1.73m ml/min/1.73m 2 90 ml/min/1.73m 2 BL alb/cr ratio, mg/g, n (%) < (29) 771 (11) Zinman B et al. N Engl J Med 2015;373: ; Neal B, et al. N Engl J Med. 2017; Epub ahead of print. Mean: 76.5± (<60) Median: 12.3 (IQR 6.65 to 42.1)

34 Primary Outcomes EMPA REG: event rates 10.5 vs. 12.1%, ARR 1.6% EMPA-REG CANVAS Median follow-up=2.4 yrs Median follow-up=3.1 yrs Zinman B et al. N Engl J Med 2015;373: ; Neal B, et al. N Engl J Med. 2017; Epub ahead of print.

35 Total Mortality EMPA REG: 5.7% vs. 8.3%, ARR 2.6% EMPA-REG CANVAS Zinman B et al. N Engl J Med 2015;373: ; Neal B, et al. N Engl J Med. 2017; Epub ahead of print.

36 CV Death EMPA-REG EMPA REG: 3.7% vs. 5.9%, ARR 2.2% CANVAS Zinman B et al. N Engl J Med 2015;373: ; Neal B, et al. N Engl J Med. 2017; Epub ahead of print.

37 Hospitalization for Heart Failure EMPA-REG EMPA REG: ARR 1.4% CANVAS Zinman B et al. N Engl J Med 2015;373: ; Neal B, et al. N Engl J Med. 2017; Epub ahead of print.

38 Number needed to treat (NNT) to prevent one death across landmark trials in patients with high CV risk Simvastatin 1 for 5.4 years Ramipril 2 for 5 years Empagliflozin for 3 years High CV risk 5% diabetes, 26% hypertension Pre statin era High CV risk 38% diabetes, 46% hypertension Pre ACEi/ARB era <29% statin T2DM with high CV risk 92% hypertension >80% ACEi/ARB >75% statin S investigator. Lancet 1994; 344: , 2. HOPE investigator N Engl J Med 2000;342:145 53,

39 David Mathews, ADA

40 DECLARE Dapagliflozin Effects on CardiovascuLAR Events Inclusion Criteria T2DM, 40 yrs Established CVD (secondary prevention) or Multiple Risk Factors (primary prevention N=17,276 1:1 Double-blind Placebo Dapagliflozin (10 mg/d) All other DM agents per treating MD Screening Duration is event-driven: 1,390 events Powered for superiority Primary Endpoints MACE: CV Death, MI, Ischemic Stroke CV death or hospitalization for heart failure Estimated completion

41 How does data from randomized clinical trials compare with real-world clinical practice? 01/05/

42 Data sources Cohort 1 HHF US Truven Health MarketScan Claims and Encounters and linked Medicare Supplemental and Coordination of Benefits (Medicare Supplemental) databases Norway Linked Prescribed Drug, National Patient and Cause of Death Registries Sweden Linked Prescribed Drug, National Patient and Cause of Death Registries Cohort 2 All-cause death and composite HHF/all-cause death Denmark Linked Prescribed Drug, National Patient and Cause of Death Registries UK Clinical Practice Research Datalink (CPRD) dataset The Health Improvement Network (THIN) dataset Germany Diabetes-Patienten-Verlaufsdokumentation (Diabetes Prospective Follow-Up; DPV) Kosiborod M et al. Circulation. 2017; doi.org/ /circulationaha

43 Patient population for all countries/databases combined 1,392,254 new users of SGLT2 inhibitor or other glucose-lowering drug fulfilling the eligibility criteria 166,033 SGLT2 inhibitor 1,226,221 other glucose-lowering drug 11,505 (7%) excluded during 1:1 match process 1:1 propensity match 1,071,693 (87%) excluded during 1:1 match process 154,528 SGLT2 inhibitor 154,528 other glucose-lowering drug Kosiborod M et al. Circulation. 2017; doi.org/ /circulationaha

44 Contribution of SGLT2i Compounds All Countries Combined US Only Europe Only Proportion of exposure time (%) 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% 5.5% 6.7% 5.6% 41.8% 52.7% hhf 51% 49.1% 42.3% 45.3% All-cause Mortality hhf + Allcause Mortality 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% 5.1% 5.3% 5.3% 19% 19.3% 19.3% 75.9% 75.4% 75.4% hhf All-cause Mortality hhf + Allcause Mortality 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% 6.3% 8.3% 6.1% 91.9% 90.1% 92.2% 1.8% 1.5% 1.7% hhf All-cause Mortality hhf + Allcause Mortality Canagliflozin Dapagliflozin Empagliflozin hhf, hospitalization for heart failure; SGLT2i, sodium glucose co-transporter 2 inhibitor Kosiborod M et al. Circulation. 2017;136(3):

45 Key study results Kosiborod M et al. Circulation. 2017; doi.org/ /circulationaha

46 Hospitalization for heart failure P-value for SGLT2 inhibitor vs other glucose-lowering drug: <0.001 Heterogeneity p-value: Data are on treatment, unadjusted. Kosiborod M et al. Circulation. 2017; doi.org/ /circulationaha

47 All-cause death primary analysis P-value for SGLT2i vs other glucose-lowering drug: <0.001 Heterogeneity p-value: Data are on treatment, unadjusted. Kosiborod M et al. Circulation. 2017; doi.org/ /circulationaha

48 Hospitalization for heart failure or all-cause death P-value for SGLT2i vs other glucose-lowering drug: <0.001 Heterogeneity p-value: Data are on treatment, unadjusted. Kosiborod M et al. Circulation. 2017; doi.org/ /circulationaha

49 LOWER RISK OF CARDIOVASCULAR EVENTS AND DEATH ASSOCIATED WITH INITIATION OF SGLT-2 INHIBITORS VERSUS OTHER GLUCOSE LOWERING DRUGS - REAL WORLD DATA ACROSS THREE MAJOR WORLD REGIONS WITH MORE THAN 400,000 PATIENTS: THE CVD-REAL 2 STUDY Mikhail Kosiborod, MD on behalf of the CVD-REAL Investigators and Study Group

50 Background (1) Outcomes trials demonstrated lower risks of cardiovascular events with sodium glucose cotransporter-2 inhibitors (SGLT-2i) in patients with type 2 diabetes (T2D), most with established cardiovascular disease (CVD) 1,2 CVD-REAL a large, international pharmaco-epidemiologic study demonstrated that SGLT-2i were associated with similar CV effects - across compounds, and in a much broader population of patients with T2D seen in clinical practice 3 However, those analyses focused on all-cause death (ACD) and heart failure, and only included patients from the United States and Europe 1. Neal B et al. N Engl J Med 2017;377: ; 2. Zinman B et al. N Engl J Med 2015;373: ; 3. Kosiborod M et al. Circulation 2017;136:

51 Background (2) Most patients with T2D reside outside the US and Europe (primarily in Asia-Pacific and the Middle East) 1 There are important differences in patient characteristics, treatment patterns, and types of adverse CVD events experienced by patients in these world regions (e.g. stroke being much more common in Asia 2 ) Data from large, well-designed comparative effectiveness studies evaluating CV outcomes with various T2D therapies has been limited outside the US and Europe 1. International Diabetes Federation. IDF Diabetes Atlas: 8th edition, 2017; 2. Ueshima H, et al. Circulation 2008;118:

52 Study Objectives To evaluate the relationship between the initiation of SGLT-2i vs. other glucose-lowering drugs (ogld) and a broad range of CV outcomes (allcause death, HHF, MI and stroke) in patients with T2D from three major world regions: Asia-Pacific, Middle East, and North America

53 Countries and Data Sources Australia National Diabetes Services Scheme (NDSS)* Canada Manitoba Population Health Research Data Repository Israel The Maccabi Health Management Organization Japan Medical Data Vision Singapore SingHealth Diabetes Registry South Korea National Health Insurance Service (NHIS) *Included in the ACD analysis only

54 Outcomes All-cause death Hospitalization for heart failure (HHF) All-cause death or HHF Myocardial infarction (MI) Stroke

55 Patient Population 3,917,551 new user episodes of SGLT-2i or ogld fulfilling the eligibility criteria 249,348 SGLT-2i 3,668,203 ogld 1:1 propensity match 14,284 (6%) excluded 3,433,139 (94%) excluded 235,064 SGLT-2i 235,064 ogld

56 Baseline Characteristics Baseline characteristic, n (%) SGLT-2i (N=235,064) ogld (N=235,064) Standardized Difference Age, years, mean (SD) 57 (12) 57 (13) 0.4% Women 105,843 (45) 106,863 (46) 0.9% Established cardiovascular disease * 59,222 (27) 56,576 (26) 2.7% Acute myocardial infarction 7,624 (3) 7,479 (3) 0.4% Unstable angina 12,480 (6) 12,235 (6) 0.5% Heart failure 15,151 (7) 14,741 (7) 0.7% Atrial fibrillation 6,026 (3) 5,843 (3) 0.5% Stroke 20,983 (10) 20,153 (9) 1.3% Peripheral arterial disease 2,446 (1) 2,384 (1) 0.3% Microvascular disease 116,370 (53) 114,630 (52) 1.6% Chronic kidney disease 4,211 (2) 4,021 (2) 0.6% *Myocardial infarction, unstable angina, stroke, heart failure, transient ischemic attack, coronary revascularization or occlusive peripheral artery disease; diabetic mono-/polyneuropathy, diabetic eye complications, diabetic foot/peripheral angiopathy, or diabetic kidney disease

57 Baseline Therapies Baseline therapies, n (%) SGLT-2i (N=235,064) ogld (N=235,064) Standardized Difference Baseline therapies, n (%) SGLT-2i (N=235,064) ogld (N=235,064) Standardized Difference Cardiovascular therapies Antihypertensive therapy 147,166 (63) 145,014 (62) 1.9% Loop diuretics 16,451 (7) 16,100 (9) 0.6% Thiazides 17,608 (8) 17,173 (7) 0.7% ACE inhibitors 20,199 (9) 20,062 (9) 0.2% ARBs 109,620 (47) 109,347 (47) 0.2% Statins 153,694 (65) 153,466 (65) 0.2% Beta-blockers 44,786 (19) 43,947 (19) 0.9% Aldosterone antagonists 6,719 (3) 6,548 (3) 0.4% Includes angiotensin converting enzyme inhibitors, angiotensin receptor blockers, Ca2+ channel blockers, β-blockers, thiazides; ACE, angiotensin-converting enzyme; ARB, angiotensin receptor blockers Glucose-lowering therapies Metformin 173,783 (74) 175,266 (75) 1.4% Sufonylurea 121,209 (52) 119,466 (51) 1.5% DPP-4 inhibitor 130,674 (56) 128,096 (55) 2.2% Thiazolidinedione 30,503 (13) 29,573 (13) 1.2% GLP-1 receptor 6,163 (3) 6,022 (3) 0.4% agonist Insulin 46,486 (20) 44,480 (19) 2.2% DPP-4, Dipeptidyl peptidase-4; GLP-1, glucagon-like peptide-1

58 Use of SGLT-2i: Proportion of Exposure Time *In South Korea and Japan; In Japan only.

59 RESULTS

60 Hospitalization for Heart Failure P-value for SGLT2i vs. ogld: p<0.001 ITT, unadjusted analysis Heterogeneity p-value: p<0.001

61 All-Cause Death P-value for SGLT2i vs. ogld: p<0.001 ITT, unadjusted analysis Heterogeneity p-value: p<0.001

62 Composite of All-Cause Death or HHF P-value for SGLT2i vs. ogld: p<0.001 ITT, unadjusted analysis Heterogeneity p-value: p<0.001

63 Myocardial Infarction P-value for SGLT2i vs. ogld: p=0.001 ITT, unadjusted analysis Heterogeneity p-value: p=0.787

64 Stroke P-value for SGLT2i vs. ogld: p<0.001 ITT, unadjusted analysis Heterogeneity p-value: p=0.029

65 DAPA-HF: A Study to Evaluate the Effect of Dapagliflozin on the Incidence of Worsening Heart Failure or Cardiovascular Death in Patients with Chronic HFrEF Population: Patients with/without T2DM and established heart failure (LVEF 40%) NTproBNP 600 pg/ml egfr 30 ml/min/1.73m 2 Stable SoC heart failure treatment N = ~4600 Randomization 1:1 Placebo + SoC Dapagliflozin 10 mg + SoC Primary endpoint: Time to composite of CV death or hospitalization for heart failure or equivalent event ^ urgent, unplanned, visit for worsening heart failure requiring intravenous therapy for heart failure V1 V2 V3 V4 V5 V6 V7,8,9, etc. PACD SCV Enrollment Randomization 14 ±7d ±3d ±7d ±7d 240 ±14d 360 ±14d * ** *Site visit every 4 th month until SED **SCV within 6 weeks post SED Duration is event-driven: 844 events Powered for superiority Study duration ~33 months Average follow-up ~24 months CV, cardiovascular; egfr, estimated glomerular filtration rate; HFrEF, heart failure with reduced ejection fraction; LVEF, left ventricular ejection fraction; N, number of patients; NTproBNP, N-terminal pro b-type natriuretic peptide; SCV, study closure visit; PACD, primary anticipated closure date; SED, study end date; SoC, standard of care; T2DM, Type 2 diabetes mellitus

66 EMPEROR-Reduced: Empagliflozin Outcome Trial in Patients with Chronic Heart Failure With Reduced Ejection Fraction 2850 patients 18 years, with chronic heart failure with reduced ejection fraction (NYHA class II IV, LVEF 40%) 1:1 Double-blind Placebo Empagliflozin 10 mg All other HF medications per treating physician Screening Estimated completion: June 2020 Primary endpoint: Time to first event of: CV death or hospitalization for heart failure Secondary endpoints: Occurrence of adjudicated hhf Change from baseline in egfr Time to occurrence of sustained reduction in egfr Time to occurrence of all cause mortality Time to onset of T2DM Change from baseline in KCCQ egfr, estimated glomerular filtration rate; HF, heart failure; hhf, hospitalization for heart failure; KCCQ, Kansas City Cardiomyopathy Questionnaire; LVEF, left ventricular ejection fraction; NYHA, New York Heart Association; T2DM, Type 2 diabetes mellitus EMPEROR-Reduced, Accessed March 2017

67 EMPEROR-Preserved: Empagliflozin Outcome Trial in Patients with Chronic Heart Failure With Preserved Ejection Fraction 4126 patients 18 years, with chronic heart failure with preserved ejection fraction (NYHA class II IV, LVEF >40%) 1:1 Double-blind Placebo Empagliflozin 10 mg All other HF medications per treating physician Screening Estimated completion: June 2020 Primary endpoint: Time to first event of: CV death or hospitalization for heart failure Secondary endpoints: Occurrence of adjudicated hhf Change from baseline in egfr Time to occurrence of sustained reduction in egfr Time to occurrence of all-cause mortality Time to onset of T2DM Change from baseline in KCCQ egfr, estimated glomerular filtration rate; HF, heart failure; hhf, hospitalization for heart failure; KCCQ, Kansas City Cardiomyopathy Questionnaire; LVEF, left ventricular ejection fraction; NYHA, New York Heart Association; T2DM, Type 2 diabetes mellitus EMPEROR-Preserved, Accessed March 2017

68 LEADER SUSTAIN 6 IRIS

69 IRIS Primary Outcome Cumulative Event-Free Survival Probability 100% 95% 90% 85% 80% HR % CI, 0.62 to 0.93 P=0.007 Pioglitazone Placebo Months in Trial 9.0%* 11.8%* *cumulative event rates. Kernan WN et al. N Engl J Med, epub Feb 17, 2016 DOI: /NEJMoa

70 Subgroup Analyses Outcomes With and Without CVD at Baseline ITT, adjusted analysis

71 Limitations Possibility of residual, unmeasured confounding cannot be definitively excluded Mortality data were available only from inpatient settings in Japan and Singapore However, most fatal events in these countries occur in the hospital Sensitivity analyses excluding data from Japan and Singapore produced similar results Did not examine safety SGLT-2i experience in real-world practice is still relatively short Longer-term follow up required to examine whether effects are sustained over time

72 Conclusions Large, international study across three major world regions, over 400,000 patients and large number of events for each outcome Initiation of SGLT-2i vs. oglds associated with lower risk of death, HHF, MI and stroke Directionality of associations generally consistent across countries Results stable in multiple sensitivity analyses and across patient subgroups Findings suggest that CV effects of SGLT-2i may extend across patient ethnic and racial backgrounds, geographic regions, as well as CV risk continuum

73 Absolute Rates of CV Events in Patients Treated with SGLT-2i and ogld p<0.001 for all comparisons Event Rate per 100 Patient-Years Established Cardiovascular Disease SGLT-2i ogld SGLT-2i ogld No Known Cardiovascular Disease Heart Failure Death Heart Failure or Heart Failure Death Heart Failure or Cavender M at al.. Presented at ADA Scientific Sessions; June 9 May 13, 2017; San Diego, CA. Death Death

74 Clinical Implications No significant heterogeneity across countries, despite geographic variations in use of SGLT-2i (predominance of canagliflozin in US and dapagliflozin in other countries) The observed cardiovascular benefits are likely class-related Broad population of patients with Type 2 diabetes in general practice, the overwhelming majority (87%) of whom did not have known cardiovascular disease Benefits may extend to those at the lower end of the risk spectrum HHF and death analyses similar to those seen in EMPA-REG OUTCOME Benefits appear to translate to real-world clinical practice Kosiborod M et al. Circulation. 2017; doi.org/ /circulationaha

75 ? Why Is Heart Failure Important In Type 2 Diabetes?

76 5-year Kaplan-Meier Survival Estimates For 115,803 Adults Age 65 Years In Fee-for-service Medicare With Diabetes By Incident Heart Failure Status 1.00 Diabetes, HF Free Proportion Surviving Diabetes with incident HF Years A. Bertoni et al. Diabetes Care 27: , 2004.

77 EMPEROR-Reduced: Empagliflozin Outcome Trial in Patients with Chronic Heart Failure with Reduced Ejection Fraction 2850 patients 18 years, with chronic heart failure with reduced ejection fraction (NYHA class II IV, LVEF 40%) 1:1 Double-blind Placebo Empagliflozin 10 mg All other HF medications per treating physician Screening Estimated completion: June 2020 Primary endpoint: Time to first event of: CV death or hospitalization for heart failure Secondary endpoints: Occurrence of adjudicated hhf Change from baseline in egfr Time to occurrence of sustained reduction in egfr Time to occurrence of all cause mortality Time to onset of T2DM Change from baseline in KCCQ clinicaltrials.gov. NCT

78 EMPEROR-Preserved: Empagliflozin Outcome Trial in Patients with Chronic Heart Failure with Preserved Ejection Fraction 4126 patients 18 years, with chronic heart failure with preserved ejection fraction (NYHA class II IV, LVEF >40%) 1:1 Double-blind Placebo Empagliflozin 10 mg All other HF medications per treating physician Screening Primary endpoint: Time to first event of: CV death or hospitalization for heart failure Estimated completion: June 2020 Secondary endpoints: Occurrence of adjudicated hhf Change from baseline in egfr Time to occurrence of sustained reduction in egfr Time to occurrence of all cause mortality Time to onset of T2DM Change from baseline in KCCQ clinicaltrials.gov. NCT

79 DAPA-HF: A Study to Evaluate the Effect of Dapagliflozin on the Incidence of Worsening Heart Failure or Cardiovascular Death in Patients with Chronic HFrEF Population: Patients with/without T2D and established heart failure (LVEF 40%) NT-proBNP 600 pg/ml egfr 30 ml/min/1.73 m 2 Stable SoC heart failure treatment N = ~4500 Randomization 1:1 Placebo + SoC Dapagliflozin 10 mg + SoC Primary endpoint: Time to composite of CV death or hospitalization for heart failure or equivalent event a Estimated study end date: December 2019 Study duration: ~3 years clinicaltrials.gov. NCT

80 ADA-EASD Position Statement: Managing Hyperglycemia in Type 2 Diabetes Update Monotherapy Efficacy * Hypo risk Weight Side effects Costs Dual therapy Efficacy * Hypo risk Weight Side effects Costs Healthy eating, weight control, increased physical activity & diabetes education Metformin high low risk neutral/loss GI / lactic acidosis low If HbA1c target not achieved after ~3 months of monotherapy, proceed to 2-drug combination (order not meant to denote any specific preference - choice dependent on a variety of patient- & disease-specific factors): Metformin + Sulfonylurea high moderate risk gain hypoglycemia low Metformin + Thiazolidinedione high low risk gain edema, HF, fxs low Metformin + DPP-4 inhibitor intermediate low risk neutral rare high Metformin + SGLT2 inhibitor intermediate low risk loss GU, dehydration high Metformin + GLP-1 receptor agonist high low risk loss GI high Metformin + Insulin (basal) highest high risk gain hypoglycemia variable Triple therapy Metformin + Sulfonylurea + TZD If HbA1c target not achieved after ~3 months of dual therapy, proceed to 3-drug combination (order not meant to denote any specific preference - choice dependent on a variety of patient- & disease-specific factors): Metformin + Thiazolidinedione + SU Metformin + DPP-4 Inhibitor + SU Metformin + SGLT-2 Inhibitor + SU Metformin + GLP-1 receptor agonist + SU Metformin + Insulin (basal) + TZD or DPP-4-i or DPP-4-i or TZD or TZD or TZD or DPP-4-i or SGLT2-i or SGLT2-i or SGLT2-i or DPP-4-i or Insulin or SGLT2-i or GLP-1-RA or GLP-1-RA or Insulin or Insulin or GLP-1-RA or Insulin or Insulin Diabetes Care 2015;38:140. Diabetologia 2015;58:429.

81 CW14 New Paradigm For Managing Blood Glucose In Type 2 Diabetes CVD Risk Factors Only Established CAD History of MI/Stroke Recent ACS Recent Stroke Asymptomatic LV Dysfunction History of HF Symptomatic HF/ HF Hospitalization End Stage HF? GLP-1 Agonists Metformin? Metformin? GLP-1? DPP-4i or SGLT2i SGLT2 Inhibitors Pioglitazone Not adequately studied SGLT2i seem promising Modify based on co-morbidities, contraindications, HbA 1c control preferably with agents that have proven safety Future trials should target combination therapies

82 Slide 81 CW14 Formatting adjusted Carolyn Whiting, 4/24/2017

83 Take Home Points Fundamental paradigm shift in T2D management CVD remains the main cause of death and disability in T2D Intensive glucose control has not convincingly reduced CV events Several classes of glucose-lowering therapies improve CV outcomes seen in RCTs and real world studies These emerging data should shift focus of T2D Rx from HbA1c alone to comprehensive CV risk reduction