Introduction Bernard Zinman CM MD FRCP FACP Director

Introduction Bernard Zinman CM, MD, FRCP, FACP Director, Leadership Sinai Centre for Diabetes Professor of Medicine, University of Toronto 1

Disclosure • Consultations and Honoraria – Astra. Zeneca, Boehringer Ingelheim, Eli Lilly, Janssen, Merck, Novo Nordisk, Sanofi, Takeda • Grant Support – Boehringer Ingelheim, Novo Nordisk, Merck 2

Type 2 diabetes is increasingly prevalent Hazard ratio (95% CI) (diabetes vs no diabetes) • Globally, 387 million people are • At least 68% of people >65 years with living with diabetes 1 diabetes die of heart disease 2 This will rise to 592 million by 20351 3 Mortality risk associated with diabetes (n=820, 900)3 2 1 0 CV death All-cause mortality 1. IDF Diabetes Atlas 6 th Edition 2014 http: //www. idf. org/diabetesatlas; 2. Centers for Disease Control and Prevention 2011; 3. Seshasai et al. N Engl J Med 2011; 364: 829 -41 3

Diabetes is associated with significant loss of life years Men 7 Non-vascular deaths Years of life lost 6 Vascular deaths 6 5 4 4 3 3 2 2 1 1 0 0 50 60 70 Age (years) . 7 5 0 40 80 90 Women 0 40 50 60 70 80 90 Age (years) On average, a 50 -year-old individual with diabetes and no history of vascular disease will die 6 years earlier compared to someone without diabetes Seshasai et al. N Engl J Med 2011; 364: 829 -41 4

Meta-analysis of intensive glucose control in T 2 DM: major CV events including heart failure Number of events More Less intensive Difference in Hb. A 1 c (%) HR (95% CI) Stroke 378 370 -0. 88 0. 96 (0. 83, 1. 10) Myocardial infarction 730 745 -0. 88 0. 85 (0. 76, 0. 94) Hospitalisation for or death from heart failure 459 446 -0. 88 1. 00 (0. 86, 1. 16) 0. 50 Favours more intensive 1. 00 2. 00 Favours less intensive • Meta-analysis of 27, 049 participants and 2370 major vascular events from: – ADVANCE – UKPDS – ACCORD – VADT HR, hazard ratio; CV, cardiovascular Turnbull FM et al. Diabetologia 2009; 52: 2288– 2298 5

Meta-analysis of intensive glucose control in T 2 DM: mortality Number of events More Less intensive Difference in Hb. A 1 c (%) HR (95% CI) All-cause mortality 980 884 -0. 88 1. 04 (0. 90, 1. 20) CV death 497 441 -0. 88 1. 10 (0. 84, 1. 42) Non-CV death 476 432 -0. 88 1. 02 (0. 89, 1. 18) 0. 50 Favours more intensive 1. 00 2. 00 Favours less intensive • Meta-analysis of 27, 049 participants and 2370 major vascular events from – ADVANCE – UKPDS – ACCORD – VADT HR, hazard ratio; CV, cardiovascular Turnbull FM et al. Diabetologia 2009; 52: 2288– 2298 6

Recent trials of newer glucose-lowering agents have been neutral on the primary CV outcome HR: 1. 0 (95% CI: 0. 89, 1. 12) SAVOR-TIMI 53 HR: 0. 96 (95% CI: UL ≤ 1. 16) EXAMINE 2013 HR: 0. 98 (95% CI: 0. 88, 1. 09) TECOS 2014 HR: 1. 02 (95% CI: 0. 89, 1. 17) 2015 ELIXA DPP-4 inhibitors* Lixisenatide EMPA-REG OUTCOME® Empagliflozin CV, cardiovascular; HR, hazard ratio; DPP-4, dipeptidyl peptidase-4 *Saxagliptin, alogliptin, sitagliptin Adapted from Johansen OE. World J Diabetes 2015; 6: 1092 -96 7

Empagliflozin • Empagliflozin is a highly selective inhibitor of the sodium glucose cotransporter 2 (SGLT 2) in the kidney • Glucose reduction occurs by reducing renal glucose reabsorption and thus increasing urinary glucose excretion • In patients with type 2 diabetes, empagliflozin leads to 1: – Significant reductions in Hb. A 1 c – Weight loss – Reductions in blood pressure without increases in heart rate 1. Liakos A et al. Diabetes Obes Metab 2014; 16: 984 -93 8

Empagliflozin modulates several factors related to CV risk Other BP Arterial stiffness Albuminuria Sympathetic nervous system activity Uric acid Glucose Insulin ↑LDL-C ↑HDL-C Triglycerides Weight Visceral adiposity Oxidative stress Adapted from Inzucchi SE, Zinman, B, Wanner, C et al. Diab Vasc Dis Res 2015; 12: 90 -100 9

EMPA-REG OUTCOME® • Randomised, double-blind, placebo-controlled CV outcomes trial • Objective To examine the long-term effects of empagliflozin versus placebo, in addition to standard of care, on CV morbidity and mortality in patients with type 2 diabetes and high risk of CV events CV, cardiovascular 10

Trial design John M Lachin, Sc. D Professor of Biostatistics and Epidemiology, and Statistics, The George Washington University, Rockville, USA 11

Disclosure • Consultations – Boehringer Ingelheim, Merck and Co. , Gilead, Janssen, Novartis, Astra. Zeneca 12

Participating countries 590 sites in 42 countries Asia North America, Australia, New Zealand Latin America Europe Africa 13

Trial design Screening (n=11531) Placebo (n=2333) Randomised and treated (n=7020) Empagliflozin 10 mg (n=2345) Empagliflozin 25 mg (n=2342) • Study medication was given in addition to standard of care – Glucose-lowering therapy was to remain unchanged for first 12 weeks • Treatment assignment double masked • The trial was to continue until at least 691 patients experienced an adjudicated primary outcome event 14

Timeline • September 15 th 2010: First patient entered • April 13 th 2013: Last patient entered • December 15 th 2014: Closeout (final visits) started • April 13 th 2015: Last patient out • Efforts were made to track outcomes and vital status for all patients, including those who discontinued trial medication 15

Patient disposition Placebo Empagliflozin 10 mg Empagliflozin 25 mg N (%) Intent-to-treat population 2333 (100) 2345 (100) 2342 (100) Discontinued study drug prematurely 683 (29. 3) 555 (23. 7) 542 (23. 1) Completed study or died 2266 (97. 1) 2264 (96. 5) Vital status available 2316 (99. 3) 2324 (99. 1) 2327 (99. 4) 16

Key inclusion and exclusion criteria • Key inclusion criteria – – Adults with type 2 diabetes BMI ≤ 45 kg/m 2 Hb. A 1 c 7– 10%* Established cardiovascular disease • Prior myocardial infarction, coronary artery disease, stroke, unstable angina or occlusive peripheral arterial disease • Key exclusion criteria – e. GFR <30 m. L/min/1. 73 m 2 (MDRD) BMI, body mass index; e. GFR, estimated glomerular filtration rate; MDRD, Modification of Diet in Renal Disease *No glucose-lowering therapy for ≥ 12 weeks prior to randomisation or no change in dose for ≥ 12 weeks prior to randomisation or, in the case of insulin, unchanged by >10% compared to the dose at randomisation 17

Pre-specified primary and key secondary outcomes • Primary outcome – 3 -point MACE: Time to first occurrence of CV death, non-fatal MI or non-fatal stroke • Key secondary outcome – 4 -point MACE: Time to first occurrence of CV death, non-fatal MI, non-fatal stroke or hospitalisation for unstable angina CV, cardiovascular; MI, myocardial infarction; MACE, Major Adverse Cardiovascular Event 18

Further pre-specified outcomes • • • CV death Non-fatal MI Non-fatal stroke Hospitalisation for heart failure All-cause mortality • All CV and neurological events were adjudicated by independent, masked, clinical event committees CV, cardiovascular; MI, myocardial infarction 19

Additional analyses • Changes from baseline in: – Hb. A 1 c – Weight – Waist circumference – Systolic and diastolic blood pressure – Heart rate – LDL cholesterol – HDL cholesterol • Safety and tolerability – Adverse events HDL, high density lipoprotein; LDL, low density lipoprotein 20

Statistical testing strategy for MACE • Analysis compared empagliflozin 10 mg and 25 mg (pooled) versus placebo • Hierarchy to be used: 1. Test of non-inferiority for 3 -point MACE 2. Test of non-inferiority for 4 -point MACE 3. Test of superiority for 3 -point MACE 4. Test of superiority for 4 -point MACE • Each tested at =0. 0249, allowing for 0. 0001 penalty for inclusion of interim data in NDA to FDA • Non-inferiority was concluded if two-sided upper bound of 95. 02% CI was <1. 3 • Superiority was concluded if two-sided p≤ 0. 0498 MACE; Major Adverse Cardiovascular Event; NDA, New Drug Application; FDA, Food and Drug Administration 21

Statistical analysis • Analyses of CV outcomes were based on a Cox proportional hazards model • Patients who did not have an event were censored on the last day they were known to be free of the outcome • Cumulative incidence functions were corrected for mortality as a competing risk (except for all-cause mortality) • The primary analysis was conducted in patients treated with ≥ 1 dose of study drug (intent-to-treat population) • The CV outcome analyses were independently validated by statisticians at the University of Freiburg, Germany 22

Further pre-defined analyses of the primary outcome • Secondary analyses: – Comparisons of empagliflozin 10 mg versus placebo and empagliflozin 25 mg versus placebo • Sensitivity analyses: – To assess the robustness of the outcomes, we used three subsets of the data set (two on-treatment sets and one perprotocol set) • Subgroup analyses based on baseline characteristics 23

Baseline characteristics and effectiveness results Christoph Wanner, MD Professor of Medicine, Division of Nephrology, Würzburg University Clinic, Würzburg, Germany 24

Disclosures • Grants from European Foundation of Studies in Diabetes – EFSD/Boehringer Ingelheim European Diabetes Research Programme 25

Baseline characteristics Placebo (n=2333) Empagliflozin 10 mg (n=2345) Empagliflozin 25 mg (n=2342) 63. 2 (8. 8) 63. 0 (8. 6) 63. 2 (8. 6) 1680 (72. 0) 1653 (70. 5) 1683 (71. 9) Europe 959 (41. 1) 966 (41. 2) 960 (41. 0) North America* 462 (19. 8) 466 (19. 9) Asia 450 (19. 3) 447 (19. 1) 450 (19. 2) Latin America 360 (15. 4) 359 (15. 3) 362 (15. 5) Africa 102 (4. 4) 107 (4. 6) 104 (4. 4) Age, years Male Region Data are n (%) or mean (SD) in patients treated with ≥ 1 dose of study drug *Includes Australia and New Zealand 26

Baseline characteristics: type 2 diabetes Hb. A 1 c, % Placebo (n=2333) Empagliflozin 10 mg (n=2345) Empagliflozin 25 mg (n=2342) 8. 08 (0. 84) 8. 07 (0. 86) 8. 06 (0. 84) Time since diagnosis of type 2 diabetes, years ≤ 5 423 (18. 1) 406 (17. 3) 434 (18. 6) >5 to 10 571 (24. 5) 585 (24. 9) 590 (25. 2) >10 1339 (57. 4) 1354 (57. 7) 1318 (56. 3) Metformin 1734 (74. 3) 1729 (73. 7) 1730 (73. 9) Sulphonylurea 992 (42. 5) 985 (42. 0) 1029 (43. 9) Thiazolidinedione 101 (4. 3) 96 (4. 1) 102 (4. 4) 1135 (48. 6) 1132 (48. 3) 1120 (47. 8) 65 (50. 6) 65 (47. 9) 66 (48. 9) Glucose-lowering medication* Insulin Mean daily dose, U** Data are n (%) or mean (SD) in patients treated with ≥ 1 dose of study drug *Medication taken alone or in combination **Placebo, n=1135; empagliflozin 10 mg, n=1132; empagliflozin 25 mg, n=1120 27

Baseline characteristics: CV risk factors Placebo (n=2333) Empagliflozin 10 mg (n=2345) Empagliflozin 25 mg (n=2342) Body mass index, kg/m 2 30. 7 (5. 2) 30. 6 (5. 3) Weight, kg 86. 6 (19. 1) 85. 9 (18. 8) 86. 5 (19. 0) Waist circumference, cm 105. 0 (14. 0) 104. 7 (13. 7) 104. 8 (13. 7) Systolic blood pressure, mm. Hg 135. 8 (17. 2) 134. 9 (16. 8) 135. 6 (17. 0) Diastolic blood pressure, mm. Hg 76. 8 (10. 1) 76. 6 (9. 8) 76. 6 (9. 7) Heart rate, bpm* 70. 7 (0. 2) 71. 0 (0. 2) 70. 5 (0. 2) LDL cholesterol, mg/d. L 84. 9 (35. 3) 86. 3 (36. 7) 85. 5 (35. 2) HDL cholesterol, mg/d. L 44. 0 (11. 3) 44. 7 (12. 0) 44. 5 (11. 8) e. GFR, m. L/min/1. 73 m 2 (MDRD) 73. 8 (21. 1) 74. 3 (21. 8) 74. 0 (21. 4) ≥ 90 m. L/min/1. 73 m 2 488 (20. 9%) 519 (22. 1%) 531 (22. 7%) 60 to <90 m. L/min/1. 73 m 2 1238 (53. 1%) 1221 (52. 1%) 1204 (51. 4%) <60 m. L/min/1. 73 m 2 607 (26. 0%) 605 (25. 8%) 607 (25. 9%) Data are n (%) or mean (SD) in patients treated with ≥ 1 dose of study drug *Mean (SE). LDL, low density lipoprotein; HDL, high density lipoprotein; e. GFR, estimated glomerular filtration rate; MDRD, Modification of Diet in Renal Disease equation 28

Baseline characteristics: CV complications Placebo (n=2333) Empagliflozin 10 mg (n=2345) Empagliflozin 25 mg (n=2342) 2307 (98. 9%) 2333 (99. 5%) 2324 (99. 2%) Coronary artery disease 1763 (75. 6%) 1782 (76. 0%) 1763 (75. 3%) Multi-vessel coronary artery disease 1100 (47. 1%) 1078 (46. 0%) 1101 (47. 0%) History of MI 1083 (46. 4%) 1107 (47. 2%) 1083 (46. 2%) Coronary artery bypass graft 563 (24. 1%) 594 (25. 3%) 581 (24. 8%) History of stroke 553 (23. 7%) 535 (22. 8%) 549 (23. 4%) Peripheral artery disease 479 (20. 5%) 465 (19. 8%) 517 (22. 1%) Single vessel coronary artery disease 238 (10. 2%) 258 (11. 0%) 240 (10. 2%) 244 (10. 5%) 240 (10. 2%) 222 (9. 5%) Any CV risk factor Cardiac failure* Data are n (%) in patients treated with ≥ 1 dose of study drug *Based on narrow standardised Med. DRA query “cardiac failure” 29

Baseline characteristics: CV medication (1) Placebo (n=2333) Empagliflozin 10 mg (n=2345) Empagliflozin 25 mg (n=2342) Anti-hypertensive therapy 2221 (95. 2%) 2227 (95. 0%) 2219 (94. 7%) ACE inhibitors/ARBs 1868 (80. 1%) 1896 (80. 9%) 1902 (81. 2%) Beta-blockers 1498 (64. 2%) 1530 (65. 2%) 1526 (65. 2%) Diuretics 988 (42. 3%) 1036 (44. 2%) 1011 (43. 2%) Calcium channel blockers 788 (33. 8%) 781 (33. 3%) 748 (31. 9%) Mineralocorticoid receptor antagonists 136 (5. 8%) 157 (6. 7%) 148 (6. 3%) Renin inhibitors 19 (0. 8%) 16 (0. 7%) 11 (0. 5%) Other 191 (8. 2%) 193 (8. 2%) 190 (8. 1%) Data are n (%) in patients treated with ≥ 1 dose of study drug ACE, angiotensin-converting enzyme; ARB, angiotensin receptor blockers 30

Baseline characteristics: CV medication (2) Placebo (n=2333) Empagliflozin 10 mg (n=2345) Empagliflozin 25 mg (n=2342) 1864 (79. 9%) 1926 (82. 1%) 1894 (80. 9%) Statins 1773 (76. 0%) 1827 (77. 9%) 1803 (77. 0%) Fibrates 199 (8. 5%) 214 (9. 1%) 217 (9. 3%) Ezetimibe Niacin 81 (3. 5%) 35 (1. 5%) 95 (4. 1%) 56 (2. 4%) 94 (4. 0%) 35 (1. 5%) Other 175 (7. 5%) 172 (7. 3%) 193 (8. 2%) 2090 (89. 6%) 2098 (89. 5%) 2064 (88. 1%) Acetylsalicylic acid 1927 (82. 6%) 1939 (82. 7%) 1937 (82. 7%) Clopidogrel 249 (10. 7%) 253 (10. 8%) 241 (10. 3%) Vitamin K antagonists 156 (6. 7%) 141 (6. 0%) 125 (5. 3%) Lipid-lowering drugs Anti-coagulants and anti-platelets Data are n (%) in patients treated with ≥ 1 dose of study drug 31

Exposure Placebo (n=2333) Empagliflozin 10 mg (n=2345) Empagliflozin 25 mg (n=2342) Treatment duration, years 2. 6 (1. 8 -3. 4) 2. 6 (1. 9 -3. 4) 2. 6 (2. 0 -3. 4) Observation time, years 3. 1 (2. 2 -3. 5) 3. 2 (2. 2 -3. 6) Data are median (interquartile range) in patients treated with ≥ 1 dose of study drug 32

Hb. A 1 c Adjusted mean (SE) Hb. A 1 c (%) 9. 0 8. 5 Placebo 8. 0 Empagliflozin 10 mg Empagliflozin 25 mg 7. 5 7. 0 6. 5 6. 0 0 12 28 40 52 66 80 94 108 122 136 150 164 178 192 206 Week Placebo Empagliflozin 10 mg 2294 2272 2296 2272 Empagliflozin 25 mg 2296 2280 2188 2133 2113 2063 2008 1967 1741 1456 1241 1109 962 2218 2150 2155 2108 2072 2058 1805 1520 1297 1164 1006 2212 2150 2115 2080 2044 1842 1540 1327 1190 1043 705 420 151 749 488 170 795 498 195 All patients (including those who discontinued study drug or initiated new therapies) were included in this mixed model repeated measures analysis (intent-to-treat) X-axis: timepoints with reasonable amount of data available for pre-scheduled measurements 33

Weight Adjusted mean (SE) weight (kg) 90 88 86 Placebo Empagliflozin 10 mg 84 Empagliflozin 25 mg 82 80 0 12 28 52 108 164 220 Week Placebo 2215 2138 1598 1239 425 Empagliflozin 10 mg 2285 1915 2290 1893 2238 2174 1673 1298 483 Empagliflozin 25 mg 2283 1891 2226 2178 1678 1335 489 All patients (including those who discontinued study drug or initiated new therapies) were included in this mixed model repeated measures analysis (intent-to-treat) X-axis: timepoints with reasonable amount of data available for pre-scheduled measurements 34

Waist circumference Adjusted mean (SE) waist circumference (cm) 107 106 Placebo 105 104 Empagliflozin 10 mg 103 Empagliflozin 25 mg 102 101 Placebo 0 12 28 52 108 164 220 Week 2183 2110 1562 1220 418 Empagliflozin 10 mg 2259 1869 2272 1836 2219 2155 1644 1285 475 Empagliflozin 25 mg 2273 1857 2209 2157 1648 1329 486 All patients (including those who discontinued study drug or initiated new therapies) were included in this mixed model repeated measures analysis (intent-to-treat) X-axis: timepoints with reasonable amount of data available for pre-scheduled measurements 35

Systolic blood pressure Adjusted mean (SE) systolic blood pressure (mm. Hg) 145 143 141 139 137 Placebo 135 Empagliflozin 25 mg Empagliflozin 10 mg 133 131 129 127 125 0 16 28 40 52 66 80 94 108 122 136 150 164 178 192 206 Week Placebo 2322 Empagliflozin 10 mg 2322 Empagliflozin 25 mg 2323 2235 2203 2161 2133 2073 2024 1974 1771 1492 1274 1126 981 2250 2235 2193 2174 2125 2095 2072 1853 1556 1327 1189 1034 2247 2221 2197 2169 2129 2102 2066 1878 1571 1351 1212 1070 735 450 171 790 518 199 842 528 216 All patients (including those who discontinued study drug or initiated new therapies) were included in this mixed model repeated measures analysis (intent-to-treat) X-axis: timepoints with reasonable amount of data available for pre-scheduled measurements 36

Diastolic blood pressure Adjusted mean (SE) diastolic blood pressure (mm. Hg) 80 79 78 77 76 Placebo 75 Empagliflozin 25 mg Empagliflozin 10 mg 74 73 72 71 70 0 16 28 40 52 66 80 94 108 122 136 150 164 178 192 206 Week Placebo 2322 Empagliflozin 10 mg 2322 Empagliflozin 25 mg 2323 2235 2203 2161 2133 2073 2024 1974 1771 1492 1274 1126 981 2250 2235 2193 2174 2125 2095 2072 1853 1556 1327 1189 1034 2247 2221 2197 2169 2129 2102 2066 1878 1571 1351 1212 1070 735 450 171 790 518 199 842 528 216 All patients (including those who discontinued study drug or initiated new therapies) were included in this mixed model repeated measures analysis (intent-to-treat) X-axis: timepoints with reasonable amount of data available for pre-scheduled measurements 37

Heart rate (ECG) 75 Adjusted mean (SE) heart rate (bpm) 74 73 72 71 70 Empagliflozin 10 mg Placebo Empagliflozin 25 mg 69 68 67 66 65 0 28 52 80 108 136 164 192 Week Placebo 2174 2127 2032 1928 1796 1300 1002 552 Empagliflozin 10 mg 2205 2137 2064 2006 1877 1366 1045 597 Empagliflozin 25 mg 2192 2127 2066 2006 1907 1383 1086 633 All patients (including those who discontinued study drug or initiated new therapies) were included in this mixed model repeated measures analysis (intent-to-treat) X-axis: timepoints with reasonable amount of data available for pre-scheduled measurements 38

Low-density lipoprotein cholesterol 100 Adjusted mean (SE) LDL cholesterol (mg/d. L) 98 96 94 92 90 Placebo Empagliflozin 25 mg Empagliflozin 10 mg 88 86 84 82 80 04 28 52 80 108 136 164 192 Week Placebo 2297 2273 Empagliflozin 10 mg 2294 2269 2179 2104 2006 1932 1419 1086 694 2205 2143 2072 1998 1474 1133 740 Empagliflozin 25 mg 2287 2256 2188 2132 2060 2020 1503 1169 779 All patients (including those who discontinued study drug or initiated new therapies) were included in this mixed model repeated measures analysis (intent-to-treat) X-axis: timepoints with reasonable amount of data available for pre-scheduled measurements 39

High-density lipoprotein cholesterol 50 Adjusted mean (SE) HDL cholesterol (mg/d. L) 49 48 47 46 Empagliflozin 25 mg Empagliflozin 10 mg 45 Placebo 44 43 42 41 40 04 28 52 80 108 136 164 192 Week Placebo 2297 2273 Empagliflozin 10 mg 2295 2270 2181 2104 2007 1932 1419 1087 694 2209 2144 2074 2001 1475 1134 741 Empagliflozin 25 mg 2289 2259 2191 2135 2064 2022 1507 1170 779 All patients (including those who discontinued study drug or initiated new therapies) were included in this mixed model repeated measures analysis (intent-to-treat) X-axis: timepoints with reasonable amount of data available for pre-scheduled measurements 40

Cardiovascular outcomes Silvio E Inzucchi Professor of Medicine, Yale University School of Medicine, New Haven, CT, USA 41

Disclosure • Consultations and non-financial support – Boehringer Ingelheim, Merck, Janssen, Novo Nordisk, Sanofi/Regeron, Intarcia, Lexicon, Poxel, Takeda, Eli Lilly • CME funding to Yale University – Boehringer Ingelheim, Eli Lilly, Novo Nordisk, Abbott, Merck and Sanofi 42

Primary outcome: 3 -point MACE HR 0. 86 (95. 02% CI 0. 74, 0. 99) p=0. 0382* Cumulative incidence function. MACE, Major Adverse Cardiovascular Event; HR, hazard ratio. * Two-sided tests for superiority were conducted (statistical significance was indicated if p≤ 0. 0498) 43

3 -point MACE Empagliflozin 10 mg HR 0. 85 (95% CI 0. 72, 1. 01) p=0. 0668 Empagliflozin 25 mg HR 0. 86 (95% CI 0. 73, 1. 02) p=0. 0865 Cumulative incidence function. MACE, Major Adverse Cardiovascular Event; HR, hazard ratio 44

3 -point MACE: sensitivity analyses Patients with event/ analysed Empagliflozin Placebo HR (95% CI) p-value Intent-to-treat population 490/4687 282/2333 0. 86 (0. 74, 0. 99)* 0. 0382 227/2308 0. 87 (0. 74, 1. 02) 0. 0839 278/2316 0. 86 (0. 75, 1. 00) 0. 0519 On-treatment analysis** 407/4607 Per protocol analysis*** 487/4654 0. 5 Favours empagliflozin 1. 0 Favours placebo Cox regression analysis. MACE, Major Adverse Cardiovascular Event; HR, hazard ratio. *95. 02% CI. **Excluding events >30 days after last intake of study drug and patients who received study drug for <30 days (cumulative). ***Patients treated with ≥ 1 dose of study drug who did not have important protocol violations. 45

CV death, MI and stroke Patients with event/ analysed Empagliflozin Placebo 3 -point MACE 490/4687 282/2333 HR (95% CI) p-value 0. 86 (0. 74, 0. 99)* 0. 0382 CV death 0. 25 0. 50 Favours empagliflozin 1. 00 2. 00 Favours placebo Cox regression analysis. MACE, Major Adverse Cardiovascular Event; HR, hazard ratio; CV, cardiovascular; MI, myocardial infarction *95. 02% CI 46

CV death HR 0. 62 (95% CI 0. 49, 0. 77) p<0. 0001 Cumulative incidence function. HR, hazard ratio 47

CV death Empagliflozin 10 mg HR 0. 65 (95% CI 0. 50, 0. 85) p=0. 0016 Empagliflozin 25 mg HR 0. 59 (95% CI 0. 45, 0. 77) p=0. 0001 Cumulative incidence function. HR, hazard ratio 48

CV death, MI and stroke Patients with event/analysed Empagliflozin Placebo 3 -point MACE CV death HR (95% CI) p-value 490/4687 282/2333 0. 86 (0. 74, 0. 99)* 0. 0382 172/4687 137/2333 0. 62 (0. 49, 0. 77) <0. 0001 Non-fatal MI Non-fatal stroke 0. 25 0. 50 Favours empagliflozin 1. 00 2. 00 Favours placebo Cox regression analysis. MACE, Major Adverse Cardiovascular Event; HR, hazard ratio; CV, cardiovascular; MI, myocardial infarction *95. 02% CI 49

CV death, MI and stroke Patients with event/analysed Empagliflozin Placebo 3 -point MACE HR (95% CI) p-value 490/4687 282/2333 0. 86 (0. 74, 0. 99)* 0. 0382 CV death 172/4687 137/2333 0. 62 (0. 49, 0. 77) <0. 0001 Non-fatal MI 213/4687 121/2333 0. 87 (0. 70, 1. 09) 0. 2189 Non-fatal stroke 150/4687 60/2333 1. 24 (0. 92, 1. 67) 0. 1638 0. 25 0. 50 Favours empagliflozin 1. 00 2. 00 Favours placebo Cox regression analysis. MACE, Major Adverse Cardiovascular Event; HR, hazard ratio; CV, cardiovascular; MI, myocardial infarction *95. 02% CI 50

Fatal and non-fatal stroke Patients with event/analysed Empagliflozin Placebo HR (95% CI) p-value (0. 89, 1. 56) 0. 2567 Intent-to-treat population 164/4687 69/2333 1. 18 Numerical difference largely driven by events occurring >30 days after treatment stop 0. 5 1. 0 Favours empagliflozin 2. 0 Favours placebo On-treatment analysis* 141/4607 66/2308 1. 04 (0. 78, 1. 40) 0. 7849 0. 5 1. 0 Favours empagliflozin Cox regression analysis. MACE, Major Adverse Cardiovascular Event; HR, hazard ratio; *Excluding events >30 days after last intake of study drug and patients who received study drug for <30 days (cumulative) 2. 0 Favours placebo

3 -point MACE and 4 -point MACE Patients with event/analysed Empagliflozin Placebo 3 -point MACE HR (95% CI) p-value 490/4687 282/2333 0. 86 (0. 74, 0. 99)* 0. 0382 CV death 172/4687 137/2333 0. 62 (0. 49, 0. 77) <0. 0001 Non-fatal MI 213/4687 121/2333 0. 87 (0. 70, 1. 09) 0. 2189 Non-fatal stroke 150/4687 60/2333 1. 24 (0. 92, 1. 67) 0. 1638 4 -point MACE 0. 25 0. 50 Favours empagliflozin 1. 00 2. 00 Favours placebo Cox regression analysis. MACE, Major Adverse Cardiovascular Event; HR, hazard ratio; CV, cardiovascular; MI, myocardial infarction *95. 02% CI 52

3 -point MACE and 4 -point MACE Patients with event/analysed Empagliflozin Placebo 3 -point MACE HR (95% CI) p-value 490/4687 282/2333 0. 86 (0. 74, 0. 99)* 0. 0382 CV death 172/4687 137/2333 0. 62 (0. 49, 0. 77) <0. 0001 Non-fatal MI 213/4687 121/2333 0. 87 (0. 70, 1. 09) 0. 2189 Non-fatal stroke 150/4687 60/2333 1. 24 (0. 92, 1. 67) 0. 1638 599/4687 333/2333 0. 89 (0. 78, 1. 01)* 0. 0795 4 -point MACE 0. 25 0. 50 Favours empagliflozin 1. 00 2. 00 Favours placebo Cox regression analysis. MACE, Major Adverse Cardiovascular Event; HR, hazard ratio; CV, cardiovascular; MI, myocardial infarction *95. 02% CI 53

3 -point MACE: subgroup analysis HR (95% CI) Empagliflozin Placebo All patients Age, years <65 ≥ 65 Sex Male Female Race White Asian Black/African-American Hb. A 1 c, % <8. 5 ≥ 8. 5 Body mass index, kg/m 2 <30 ≥ 30 e. GFR, m. L/min/1. 73 m 2 ≥ 90 60 to <90 <60 4687 p-value for interaction 2333 0. 01 2596 2091 1297 1036 0. 81 3336 1351 1680 653 0. 09 3403 1006 237 1678 511 120 0. 01 3212 1475 1607 726 2279 2408 1120 1213 1050 2425 1212 488 1238 607 0. 06 0. 20 0. 25 0. 50 1. 00 Favours empagliflozin 2. 00 4. 00 Favours placebo For the test of homogeneity of the treatment group difference among subgroups with no adjustment for multiple tests. e. GFR, estimated glomerular filtration rate (according to Modification of Diet in Renal Disease equation) 54

CV death: subgroup analyses HR (95% CI) Empagliflozin Placebo All patients Age, years <65 ≥ 65 Sex Male Female Race White Asian Black/African-American Hb. A 1 c, % <8. 5 ≥ 8. 5 Body mass index, kg/m 2 <30 ≥ 30 e. GFR, m. L/min/1. 73 m 2 ≥ 90 60 to <90 <60 4687 p-value for interaction 2333 0. 21 2596 2091 1297 1036 0. 32 3336 1351 1680 653 0. 43 3403 1006 237 1678 511 120 0. 51 3212 1475 1607 726 2279 2408 1120 1213 1050 2425 1212 488 1238 607 0. 05 0. 15 0. 25 0. 50 1. 00 Favours empagliflozin 2. 00 4. 00 Favours placebo For the test of homogeneity of the treatment group difference among subgroups with no adjustment for multiple tests. e. GFR, estimated glomerular filtration rate (according to Modification of Diet in Renal Disease equation) 55

Heart failure 56

Hospitalisation for heart failure HR 0. 65 (95% CI 0. 50, 0. 85) p=0. 0017 Cumulative incidence function. HR, hazard ratio 57

Hospitalisation for heart failure Empagliflozin 10 mg HR 0. 62 (95% CI 0. 45, 0. 86) p=0. 0044 Empagliflozin 25 mg HR 0. 68 (95% CI 0. 50, 0. 93) p=0. 0166 Cumulative incidence function. HR, hazard ratio 58

All-cause mortality 59

All-cause mortality HR 0. 68 (95% CI 0. 57, 0. 82) p<0. 0001 Kaplan-Meier estimate. HR, hazard ratio 60

All-cause mortality Empagliflozin 10 mg HR 0. 70 (95% CI 0. 56, 0. 87) p=0. 0013 Empagliflozin 25 mg HR 0. 67 (95% CI 0. 54, 0. 83) HR 0. 68 p=0. 0003 (95% CI 0. 57, 0. 82) p<0. 0001 Kaplan-Meier estimate. HR, hazard ratio 61

All-cause mortality, CV death and non-CV death Patients with event/analysed Empagliflozin Placebo HR 95% CI p-value All-cause mortality 269/4687 194/2333 0. 68 (0. 57, 0. 82) <0. 0001 CV death 172/4687 137/2333 0. 62 (0. 49, 0. 77) <0. 0001 Non-CV death 97/4687 57/2333 0. 84 (0. 60, 1. 16) 0. 2852 0. 25 0. 50 Favours empagliflozin 1. 00 2. 00 Favours placebo Cox regression analysis. CV, cardiovascular; HR, hazard ratio 62

Safety and tolerability David Fitchett, MD Cardiologist, St Michael’s Hospital Associate Professor of Medicine, University of Toronto, Canada 63

Disclosures • Consultations – Boehringer Ingelheim, Novo Nordisk, Astra. Zeneca, Sanofi, Merck 64

Adverse events Placebo (n=2333) Empagliflozin 10 mg (n=2345) Empagliflozin 25 mg (n=2342) n (%) Rate One or more AEs 2139 (91. 7%) 178. 67 2112 (90. 1%) 150. 34 2118 (90. 4%) 148. 36 One or more drug-related* AEs 549 (23. 5%) 11. 33 666 (28. 4%) 14. 15 643 (27. 5%) 13. 38 One or more AEs leading to discontinuation 453 (19. 4%) 8. 26 416 (17. 7%) 7. 28 397 (17. 0%) 6. 89 One or more serious AEs 988 (42. 3%) 22. 34 876 (37. 4%) 18. 20 913 (39. 0%) 19. 39 Rate = per 100 patient-years *As reported by the investigator Patients treated with ≥ 1 dose of study drug 65

Adverse events consistent with urinary tract infection Placebo (n=2333) Empagliflozin 10 mg (n=2345) Empagliflozin 25 mg (n=2342) n (%) Rate 423 (18. 1%) 8. 21 426 (18. 2%) 8. 02 416 (17. 8%) 7. 75 10 (0. 4%) 0. 17 22 (0. 9%) 0. 37 19 (0. 8%) 0. 31 Male 158 (9. 4%) 3. 96 180 (10. 9%) 4. 49 170 (10. 1%) 4. 09 Female 265 (40. 6%) 22. 81 246 (35. 5%) 18. 83 246 (37. 3%) 20. 38 Events consistent with UTI Events leading to discontinuation By sex Rate = per 100 patient-years Patients treated with ≥ 1 dose of study drug Based on 79 Med. DRA preferred terms 66

Complicated urinary tract infection Placebo (n=2333) Empagliflozin 10 mg (n=2345) Empagliflozin 25 mg (n=2342) n (%) Rate 41 (1. 8%) 0. 71 34 (1. 4%) 0. 57 48 (2. 0%) 0. 80 16 (0. 7%) 0. 28 13 (0. 6%) 0. 22 16 (0. 7%) 0. 27 Pyelonephritis† 22 (0. 9%) 0. 38 15 (0. 6%) 0. 25 20 (0. 9%) 0. 33 Urosepsis 3 (0. 1%) 0. 05 6 (0. 3%) 0. 10 11 (0. 5%) 0. 18 Complicated urinary tract infection* Urinary tract infection Rate = per 100 patient-years Patients treated with ≥ 1 dose of study drug Events reported in >0. 1% of patients in any group are shown *Pyelonephritis, urosepsis or serious adverse event consistent with urinary tract infection †Based on 15 Med. DRA preferred terms 67

Adverse events consistent with genital infection Placebo (n=2333) Empagliflozin 10 mg (n=2345) Empagliflozin 25 mg (n=2342) n (%) Rate 42 (1. 8%) 0. 73 153 (6. 5%) 2. 66 148 (6. 3%) 2. 55 Serious events 3 (0. 1%) 0. 05 5 (0. 2%) 0. 08 4 (0. 2%) 0. 07 Events leading to discontinuation 2 (0. 1%) 0. 03 19 (0. 8%) 0. 32 14 (0. 6%) 0. 23 Male 25 (1. 5%) 0. 60 89 (5. 4%) 2. 16 77 (4. 6%) 1. 78 Female 17 (2. 6%) 1. 09 64 (9. 2%) 3. 93 71 (10. 8%) 4. 81 Events consistent with genital infection By sex Rate = per 100 patient-years Patients treated with ≥ 1 dose of study drug Based on 88 Med. DRA preferred terms 68

Confirmed hypoglycaemic adverse events Placebo (n=2333) Empagliflozin 10 mg (n=2345) Empagliflozin 25 mg (n=2342) n (%) Confirmed hypoglycaemic adverse events Events requiring assistance 650 (27. 9%) 656 (28. 0%) 647 (27. 6%) 36 (1. 5%) 33 (1. 4%) 30 (1. 3%) 483 (42. 6%) 494 (43. 6%) 464 (41. 4%) 28 (2. 5%) 27 (2. 4%) 25 (2. 2%) Patients taking insulin at baseline Total Events requiring assistance Patients treated with ≥ 1 dose of study drug Plasma glucose <3. 9 mmol/L (70 mg/d. L) and/or requiring assistance 69

Other adverse events (1) Placebo (n=2333) Empagliflozin 10 mg (n=2345) Empagliflozin 25 mg (n=2342) n (%) Rate Diabetic ketoacidosis* 1 (<0. 1%) 0. 02 3 (0. 1%) 0. 05 1 (<0. 1%) 0. 02 Acute kidney injury† 155 (6. 6%) 2. 77 121 (5. 2%) 2. 07 125 (5. 3%) 2. 12 Events consistent with volume depletion§ 115 (4. 9%) 2. 04 115 (4. 9%) 1. 97 124 (5. 3%) 2. 11 Serious events 24 (1. 0%) 0. 42 19 (0. 8%) 0. 32 26 (1. 1%) 0. 43 Events leading to discontinuation 7 (0. 3%) 0. 12 1 (<0. 1%) 0. 02 4 (0. 2%) 0. 07 20 (0. 9%) 0. 35 9 (0. 4%) 0. 15 21 (0. 9%) 0. 35 Venous thrombotic events** Rate = per 100 patient-years Patients treated with ≥ 1 dose of study drug *Based on 4 Med. DRA preferred terms. †Based on 1 standardised Med. DRA query §Based on 8 Med. DRA preferred terms. **Based on 1 standardised Med. DRA query 70

Other adverse events (2) Placebo (n=2333) Empagliflozin 10 mg (n=2345) Empagliflozin 25 mg (n=2342) n (%) Rate Hepatic injury* 108 (4. 6%) 1. 91 80 (3. 4%) 1. 35 88 (3. 8%) 1. 48 Hypersensitivity* 197 (8. 4%) 3. 59 158 (6. 7%) 2. 75 181 (7. 7%) 3. 14 Bone fractures† 91 (3. 9%) 1. 61 92 (3. 9%) 1. 57 87 (3. 7%) 1. 46 Rate = per 100 patient-years Patients treated with ≥ 1 dose of study drug *Based on standardised Med. DRA queries †Based on 62 Med. DRA preferred terms 71

Changes in clinical laboratory parameters Placebo (n=2333) Empagliflozin 10 mg (n=2345) Empagliflozin 25 mg (n=2342) Baseline Change from baseline Haematocrit, % 41. 1 (5. 7) 0. 9 (4. 7) 41. 2 (5. 6) 4. 8 (5. 5) 41. 3 (5. 7) 5. 0 (5. 3) Haemoglobin, g/d. L 13. 4 (1. 5) -0. 1 (1. 2) 13. 4 (1. 5) 0. 8 (1. 3) 13. 5 (1. 5) 0. 8 (1. 3) Serum creatinine, mg/d. L 1. 04 (0. 24) 0. 07 (0. 25) 1. 03 (0. 23) 0. 04 (0. 2) 1. 04 (0. 25) 0. 04 (0. 19) e. GFR m. L/min/1. 73 m 2 74. 8 (20. 6) -4. 5 (12. 9) 75. 2 (21. 1) -2. 5 (13. 1) 75. 0 (21. 4) -2. 8 (13. 4) Sodium, m. Eq/L 141 (2) 0 (2) Potassium, m. Eq/L 4. 3 (0. 4) 0. 0 (0. 4) Calcium, mg/d. L 9. 7 (0. 5) 0. 0 (0. 5) 9. 7 (0. 4) 0. 0 (0. 5) Magnesium, m. Eq/L 1. 7 (0. 2) 0. 0 (0. 2) 1. 7 (0. 2) 0. 1 (0. 2) Phosphate, mg/d. L 3. 7 (0. 3) 0. 0 (0. 3) 3. 7 (0. 3) 0. 1 (0. 3) Electrolytes Data are mean (SD) in patients treated with ≥ 1 dose of study drug Changes from baseline are at last value on treatment, defined as the last measurement ≤ 3 days after the last intake of study drug 72

Implications for practice and conclusions Bernard Zinman CM, MD, FRCP, FACP Director, Leadership Sinai Centre for Diabetes Professor of Medicine, University of Toronto 73

EMPA-REG OUTCOME®: Summary • Empagliflozin reduced risk for 3 -point MACE by 14% • Empagliflozin was associated with a reduction in Hb. A 1 c without an increase in hypoglycaemia, reductions in weight and blood pressure, and small increases in LDL cholesterol and HDL cholesterol • Empagliflozin was associated with an increase in genital infections but was otherwise well tolerated MACE, Major Adverse Cardiovascular Event; HDL, high density lipoprotein; LDL, low density lipoprotein 74

EMPA-REG OUTCOME®: Summary • Empagliflozin reduced hospitalisation for heart failure by 35% • Empagliflozin reduced CV death by 38% • Empagliflozin improved survival by reducing all-cause mortality by 32% CV, cardiovascular 75

EMPA-REG OUTCOME®: Important features • Population studied – A high CV risk population with modest hyperglycaemia on standard glucose-lowering and CV therapy • Follow-up and retention – 97. 0% of patients completed the study and vital status was available for 99. 2% of patients • Two doses of empagliflozin (10 mg and 25 mg) studied – Similar magnitude of reduction with both doses for CV death, allcause mortality and hospitalisation for heart failure CV, cardiovascular 76

Number needed to treat (NNT) to prevent one death across landmark trials in patients with high CV risk Simvastatin 1 Ramipril 2 Empagliflozin for 5. 4 years for 5 years for 3 years 30 56 39 High CV risk T 2 DM with high CV risk 5% diabetes, 26% hypertension 38% diabetes, 46% hypertension 92% hypertension Pre-ACEi/ARB era Pre-statin era 1994 >80% ACEi/ARB <29% statin 2000 >75% statin 2015 1. 4 S investigator. Lancet 1994; 344: 1383 -89, http: //www. trialresultscenter. org/study 2590 -4 S. htm; 2. HOPE investigator N Engl J Med 2000; 342: 145 -53, http: //www. trialresultscenter. org/study 2606 -HOPE. htm 77

EMPA-REG OUTCOME®: Therapeutic considerations • Empagliflozin, as used in this trial, for 3 years in 1, 000 patients with type 2 diabetes at high CV risk: – 25 lives saved (82 vs 57 deaths) • 22 fewer CV deaths (59 vs 37) – 14 fewer hospitalisations for heart failure (42 vs 28) – 53 additional genital infections (22 vs 75) 78

EMPA-REG OUTCOME®: What effect will these results have on clinical practice guidelines? 79

Acknowledgements • We are indebted to the study participants for their commitment to following the trial protocol including adherence to study medication, clinic visits and assessments • We thank the physician investigators, coordinators and their staff from 590 sites in 42 countries who conscientiously enrolled participants and maintained excellent follow-up throughout the study 80

Acknowledgements EMPA-REG OUTCOME® Steering Committee Bernard Zinman [Chair], Lunenfeld-Tanenbaum Research Institute, Toronto, Canada Christoph Wanner, Würzburg University Clinic, Würzburg, Germany John M. Lachin, The George Washington University, Rockville, MD, USA David Fitchett, University of Toronto, Canada Erich Bluhmki, Boehringer Ingelheim, Biberach, Germany Odd Erik Johansen, Boehringer Ingelheim KS, Asker, Norway Hans J. Woerle, Boehringer Ingelheim, Germany Uli C. Broedl, Boehringer Ingelheim, Germany Silvio E. Inzucchi, Yale University School of Medicine, CT, USA 81

Acknowledgements Data Safety Monitoring Board Francine K. Welty, Beth Israel Deaconess Medical Center, Boston, USA Klaus G. Parhofer, University of Munich, Germany Terje R. Pedersen, Oslo University Hospital, Oslo, Norway Kennedy R. Lees, University of Glasgow, UK Tim Clayton, London School of Hygiene and Tropical Medicine, UK Stuart Pocock, London School of Hygiene and Tropical Medicine, UK Mike Palmer, N Zero 1 Ltd, Wilmslow, UK 82

Further reading • The slides from this presentation are available at: www. empa-reg-outcome. com www. easd. org • www. nejm. org 83