Primary Results 8 th June 2015 TECOS

Primary Results 8 th June 2015

TECOS • Initiated in advance of FDA requirements, but consistent with that guidance • Large, international trial designed to assess the impact of sitagliptin versus placebo on cardiovascular event rates – When added to usual diabetes care – Minimize difference in glycemia between groups • Randomized, double-blind, placebo-controlled • Academically led in collaboration with industry sponsorship Green JB et al. NEJM 2015; DOI: 10. 1056/NEJMoa 1501352 2

Pragmatic Study Design • Integration with usual diabetes care – Management of diabetes, hypertension and dyslipidaemia remain responsibility of usual care provider – All laboratory results collected locally as available, except Hb. A 1 c, from the usual care setting – TECOS investigator responsible for dose adjustment of blinded study drug according to e. GFR • Streamlined data collection • Streamlined safety reporting Green JB et al. NEJM 2015; DOI: 10. 1056/NEJMoa 1501352 3

Primary Objective To demonstrate that the risk of cardiovascular events in patients treated with sitagliptin in addition to usual care was non-inferior to that in patients treated without sitagliptin in addition to usual care. Green JB et al. NEJM 2015; DOI: 10. 1056/NEJMoa 1501352 4

Primary Composite Cardiovascular Outcome Time to first occurrence of: – Cardiovascular-related death – Nonfatal myocardial infarction – Nonfatal stroke – Hospitalization for unstable angina A Clinical Endpoints Committee, blinded to therapy allocation, reviewed all potential CVD endpoints independently. 5 Green JB et al. NEJM 2015; DOI: 10. 1056/NEJMoa 1501352 5

Secondary Cardiovascular Outcomes Time to — – Secondary composite CV outcome (nonfatal MI, nonfatal stroke, or CV-related death) – First confirmed component event in the primary outcome (Cardiovascular-related death, nonfatal myocardial infarction, nonfatal stroke, hospitalization for unstable angina) – First fatal or nonfatal MI – First fatal or nonfatal stroke – All-cause mortality – Hospitalization for heart failure or CV-related death Green JB et al. NEJM 2015; DOI: 10. 1056/NEJMoa 1501352 6

Other Secondary Outcomes • Change in Hb. A 1 c over time • Change in renal function over time – e. GFR – urinary albumin/creatinine ratio • In patients not receiving insulin at baseline, time to initiation of chronic insulin therapy • In all patients, time to next co-interventional agent (next AHA or chronic insulin) • Medical resource utilization (e. g. , hospitalizations, outpatient visits) — data not shown today Green JB et al. NEJM 2015; DOI: 10. 1056/NEJMoa 1501352 7

Major Inclusion Criteria • Type 2 diabetes (A 1 c ≥ 6. 5% and ≤ 8. 0%) – Stable monotherapy OR dual combination therapy with metformin, pioglitazone, or sulfonylurea or *stable dose of insulin with or without metformin • ≥ 50 years old • Preexisting vascular disease defined as having: – History of myocardial infarction – Prior coronary revascularization – Coronary angiography with at least one ≥ 50% stenosis – History of ischemic stroke – Carotid arterial disease with ≥ 50% carotid stenosis – Peripheral arterial disease with objective evidence • Able to see usual care provider at least twice yearly *Amended 13 Sept 2010 Green JB et al. NEJM 2015; DOI: 10. 1056/NEJMoa 1501352 8

Major Exclusion Criteria • Type 1 diabetes or history of ketoacidosis • History of ≥ 2 episodes of severe hypoglycemia during the 12 months prior to enrollment • Estimated glomerular filtration rate (e. GFR) <30 m. L/min/1. 73 m 2 • Use of another DPP-4 inhibitor, GLP-1 analogue, or thiazolidinedione other than pioglitazone in previous three months • Cirrhosis of the liver • Planned revascularization procedure • Pregnancy or planned pregnancy Green JB et al. NEJM 2015; DOI: 10. 1056/NEJMoa 1501352 9

Statistical Analysis Plan Primary analysis for non-inferiority performed in the per protocol (PP) population – Consistent with ICH E-9/CONSORT statement – Supporting intention-to-treat (ITT) analysis also performed • Prespecified conditional hierarchical analyses, each at =0. 025 one-sided: – Non-inferiority for the secondary composite CV outcome (PP population) – Superiority for primary CV outcome (ITT population) – Superiority for the secondary composite CV outcome (ITT population) Green JB et al. NEJM 2015; DOI: 10. 1056/NEJMoa 1501352 10

Prohibited Medications • Use of open-label DPP-4 inhibitors or GLP-1 receptor agonists • Use of rosiglitazone as concomitant therapy subsequent to enrollment was discouraged, but not prohibited Green JB et al. NEJM 2015; DOI: 10. 1056/NEJMoa 1501352 11

Recruitment: = country involved in TECOS December 2008 – July 2012 ITT population North America 2594, 17. 7% Western Europe 2076, 14. 2% Eastern Europe 3965, 27. 0% Asia Pacific 4565, 31. 1% Latin America 1471, 10. 0% Total 14, 671 Green JB et al. NEJM 2015; DOI: 10. 1056/NEJMoa 1501352

Consort Diagram 14, 735 randomized 64 excluded from all analyses • 11 did not consent • 53 at one site excluded for GCP deviations 14, 671 included in ITT analysis 7332 sitagliptin ITT 7339 placebo ITT 7180 (97. 9%) VS known 7123 (97. 0%) VS known 6972 (95. 1%) completed 6905 (94. 1%) completed 61 (0. 8%) LTFU 71 (1. 0%) LTFU 29 (48%) VS known 33 (46%) VS known 299 (4. 1%) Withdrawn 363 (4. 9%) Withdrawn 179 (60%) VS known 185 (51%) VS known ITT = intention-to-treat; LTFU = lost to follow-up; VS = vital status, GCP = Good Clinical Practice Green JB et al. NEJM 2015; DOI: 10. 1056/NEJMoa 1501352

Per Protocol Population Sitagliptin Placebo 7332 7339 75 (1. 0%) 73 (1. 0%) 7257 7266 Censored early for primary composite 1494 (20. 4%) 1654 (22. 5%) Early study drug discontinuation 1378 (18. 8%) 1505 (20. 5%) Initiation of prohibited medication 70 (1. 1%) 102 (1. 4%) Other 46 (0. 6%) 47 (0. 6%) # Primary events censored 136 (1. 9) 151 (2. 1%) ITT Population (n) Completely excluded for major protocol violation PP Population (n) Green JB et al. NEJM 2015; DOI: 10. 1056/NEJMoa 1501352 14

Baseline Characteristics Characteristic Sitagliptin n=7332 Placebo n=7339 Age (years) 65. 4 ± 7. 9 65. 5 ± 8. 0 2134 (29. 1%) 2163 (29. 5%) White 4955 (67. 6%) 5002 (68. 2%) Black 206 (2. 8%) 241 (3. 3%) Asian 1654 (22. 6%) 1611 (22. 0%) Other 517 (7. 1%) 485 (6. 6%) 886 (12. 1%) 912 (12. 4%) BMI (kg/m 2) 30. 2 ± 5. 6 30. 2 ± 5. 7 e. GFR (m. L/min/1. 73 m 2)* 74. 9 ± 21. 3 74. 9 ± 20. 9 Women Race Hispanic or Latino Values are mean ±SD for continuous variables or n, % for categorical variables. *MDRD formula used to calculate e. GFR. Site-reported values are presented. Green JB et al. NEJM 2015; DOI: 10. 1056/NEJMoa 1501352 15

Baseline Characteristics— CV Risk Management Characteristic Sitagliptin n=7332 Placebo n=7339 Systolic blood pressure (mm. Hg) 135 ± 16. 9 135 ± 17. 1 Diastolic blood pressure (mm. Hg) 77. 1 ± 10. 3 77. 2 ± 10. 6 Total cholesterol (mg/d. L) 166. 1 ± 44. 8 165. 4 ± 45. 9 LDL-C (mg/d. L) 91. 2 ± 63. 8 90. 7 ± 51. 2 HDL-C (mg/d. L) 43. 5 ± 12. 0 43. 4 ± 13. 0 166. 0 ± 101. 0 164. 8 ± 98. 8 Aspirin use 5764 (78. 6%) 5754 (78. 4%) Statin use 5851 (79. 8%) 5868 (80. 0%) Triglycerides (mg/d. L) Medication Values are mean ±SD for continuous variables or n, % for categorical variables. Green JB et al. NEJM 2015; DOI: 10. 1056/NEJMoa 1501352 16

Baseline Characteristics— CV Disease Sitagliptin n=7332 Placebo n=7339 Prior cardiovascular disease 5397 (73. 6%) 5466 (74. 5%) Myocardial infarction 3133 (42. 7%) 3122 (42. 5%) PCI 2814 (38. 9%) 2900 (40. 1%) CABG 1845 (25. 2%) 1819 (24. 8%) ≥ 50% coronary stenosis 3804 (51. 9%) 3883 (52. 9%) Prior cerebrovascular disease 1806 (24. 6%) 1782 (24. 3%) 1297 (17. 7%) 1258 (17. 1%) TIA 280 (3. 8%) 286 (3. 9%) ≥ 50% carotid stenosis 431 (5. 9%) 429 (5. 8%) Peripheral arterial disease 1217 (16. 6%) 1216 (16. 6%) History of heart failure 1303 (17. 8%) 1340 (18. 3%) Characteristic Stroke Green JB et al. NEJM 2015; DOI: 10. 1056/NEJMoa 1501352 17

Baseline Characteristics— Diabetes Characteristic Sitagliptin n=7332 Placebo n=7339 Duration of diabetes (years) 11. 6 ± 8. 1 7. 2 ± 0. 5 Metformin 5936 (81. 0%) 6030 (82. 2%) Sulfonylurea 3346 (45. 6%) 3299 (45. 0%) 196 (2. 7%) 200 (2. 7%) 1724 (23. 5%) 1684 (22. 9%) 50 (33, 80) 50 (32, 80) Monotherapy 3496 (47. 7%) 3498 (47. 7%) Dual combination therapy 3766 (51. 4%) 3768 (51. 3%) Hb. A 1 c (%) Medication taken alone or in combination Thiazolidinedione Insulin Median daily dose (units) Values are mean ±SD or median (IQR) for continuous variables or n, % for categorical variables. Green JB et al. NEJM 2015; DOI: 10. 1056/NEJMoa 1501352 18

Participant Follow-up, years Primary outcome Median (IQR) 2. 84 (2. 19, 3. 61) Secondary outcome Median (IQR) 2. 87 (2. 21, 3. 63) All-cause mortality Median (IQR) Green JB et al. NEJM 2015; DOI: 10. 1056/NEJMoa 1501352 19 3. 02 (2. 31, 3. 76)

Cumulative Proportion of Participants Lost or Withdrawn from Study At the end of Year 1 At the end of Year 2 At the end of Year 3 Sitagliptin: 2. 7% Sitagliptin: 4. 0% Sitagliptin: 5. 0% Placebo: 3. 3% Placebo: 5. 1% Placebo: 6. 0% Green JB et al. NEJM 2015; DOI: 10. 1056/NEJMoa 1501352

Cumulative Proportion of Participants Discontinuing Study Drug At the end of Year 2 At the end of Year 1 Sitagliptin: 11. 1% Sitagliptin: 17. 9% Placebo: 19. 9% Placebo: 12. 2% At the end of Year 3 Sitagliptin: 24. 5% Placebo: 26. 6% Green JB et al. NEJM 2015; DOI: 10. 1056/NEJMoa 1501352

Glycemic Control Least Squares Mean Hb. A 1 c ± 1 SD Overall LS Mean difference -0. 29% (-0. 32, -0. 27), p<0. 0001 Green JB et al. NEJM 2015; DOI: 10. 1056/NEJMoa 1501352

Initiation of Additional Antihyperglycemic Agents Sitagliptin n=7332 Placebo n=7339 Initiation of next antidiabetic medication ITT HR 0. 72 (95% CI 0. 68, 0. 77), p<0. 001 # Patients 1591 (21. 7%) 2046 (27. 9%) 8. 5 11. 6 1 year, % (95% CI) 6. 7 (6. 1, 7. 3) 9. 3 (8. 6, 10. 0) 2 years, % (95% CI) 14. 9 (14. 1, 15. 7) 20. 3 (19. 4, 21. 3) 3 years, % (95% CI) 23. 4 (22. 2, 24. 5) 31. 3 (30. 1, 32. 6) 4 years, % (95% CI) 33. 1 (31. 4, 34. 9) 41. 5 (39. 6, 43. 3) Event rate per 100 pyrs Cumulative incidence (%) of event Green JB et al. NEJM 2015; DOI: 10. 1056/NEJMoa 1501352 23

Initiation of Chronic Insulin Therapy Sitagliptin* n=5608 Placebo* n=5655 Initiation of insulin ITT HR 0. 70 (95% CI 0. 63, 0. 79) p<0. 001 # Patients 542 (9. 7%) 744 (13. 2%) 3. 44 4. 85 1 year, % (95% CI) 3. 2 (2. 8, 3. 7) 4. 8 (4. 3, 5. 4) 2 years, % (95% CI) 6. 4 (5. 8, 7. 1) 9. 7 (8. 9, 10. 5) 3 years, % (95% CI) 9. 8 (9. 0, 10. 7) 14. 1 (13. 1, 15. 1) 4 years, % (95% CI) 13. 2 (12. 1, 14. 5) 17. 5 (16. 3, 18. 9) Event rate per 100 pyrs Cumulative incidence (%) of event *Patients not on insulin at baseline Green JB et al. NEJM 2015; DOI: 10. 1056/NEJMoa 1501352

Estimated Glomerular Filtration Rate (e. GFR) Mean ± 1 SD Estimated overall mean difference*: -1. 34 ml/min/1. 73 m 2 (95%CI -1. 76, -0. 91), p<0. 0001 *Mixed model with random intercept & slope, adjusted for baseline value and region Green JB et al. NEJM 2015; DOI: 10. 1056/NEJMoa 1501352

Severe Hypoglycemia* ITT HR (95% CI): 1. 12 (0. 89– 1. 40), p=0. 33 Sitagliptin Participants with event n (%) 160 (2. 2%) Events per 100 patient-years Placebo 143 (1. 9%) 0. 78 0. 70 *Hypoglycemia requiring assistance Green JB et al. NEJM 2015; DOI: 10. 1056/NEJMoa 1501352 26

Confirmed Acute Pancreatitis and Pancreatic Cancer ITT HR 1. 93 (0. 96, 3. 88), p=0. 065 Patients Events Sitagliptin (n=7332) Placebo (n=7339) Sitagliptin Placebo 23 (0. 3%) 12 (0. 2%) 25 17 Severe 4 0 Mild 19 11 21 16 Unknown severity 0 1 Acute pancreatitis ITT HR 0. 66 (0. 28, 1. 51), p=0. 32 Sitagliptin n=7332 Pancreatic cancer Placebo n=7339 9 (0. 1%) 14 (0. 2%) Green JB et al. NEJM 2015; DOI: 10. 1056/NEJMoa 1501352 27

Confirmed Charter Defined Malignancy* ITT HR 0. 91 (0. 77, 1. 08), p=0. 27 Event Sitagliptin n=7332 Placebo n=7339 278 308 268 (3. 7%) 290 (4. 0%) 41 (0. 6%) 43 (0. 6%) 21 (0. 3%) 28 (0. 4%) 15 (0. 2%) 49 (0. 7%) 35 (0. 5%) 34 (0. 5%) 25 (0. 3%) 11 (0. 1%) 9 (0. 1%) 14 (0. 2%) Total number of malignancies Number of patients with malignancy 5 most common types of malignancy Prostate Lung (bronchus) Colon (large intestine, cecum, appendix) Bladder Melanoma Pancreatic cancer *Newly diagnosed malignancy or 1 st recurrence of previously diagnosed malignancy during the study period Green JB et al. NEJM 2015; DOI: 10. 1056/NEJMoa 1501352

Serious Adverse Events* Serious Adverse Events Sitagliptin N=7332 Placebo N=7339 Patients n (%) Events Neoplasms benign, malignant and unspecified 341 (4. 7%) 405 371 (5. 1%) 470 Injury, poisoning and procedural complications 146 (2. 0%) 165 133 (1. 8%) 153 Gastrointestinal disorders 130 (1. 8%) 143 102 (1. 4%) 121 Musculoskeletal and connective tissue disorders 118 (1. 6%) 136 93 (1. 3%) 102 Respiratory, thoracic and mediastinal disorders 66 (0. 9%) 81 77 (1. 0%) 95 *System Organ Classes with incidence ≥ 1% in either group Green JB et al. NEJM 2015; DOI: 10. 1056/NEJMoa 1501352 29

Primary Composite Cardiovascular Outcome* PP Analysis for Non-inferiority * CV death, nonfatal MI, nonfatal stroke, hospitalization for unstable angina Green JB et al. NEJM 2015; DOI: 10. 1056/NEJMoa 1501352

Primary Composite Cardiovascular Outcome* ITT Analysis for Superiority * CV death, nonfatal MI, nonfatal stroke, hospitalization for unstable angina Green JB et al. NEJM 2015; DOI: 10. 1056/NEJMoa 1501352

Primary Composite Cardiovascular Outcome ITT Population Numbers of patients with events Sitagliptin n=7332 Placebo n=7339 Primary composite CV Outcome 839 (11. 4%) 851 (11. 6%) 4. 06 per 100 pyrs 4. 17 per 100 pyrs ITT HR=0. 98 (0. 89, 1. 08), p=0. 65 Individual components • CV death 311 (4. 2%) 291 (4. 0%) • Nonfatal MI 275 (3. 8%) 286 (3. 9%) • Nonfatal stroke 145 (2. 0%) 157 (2. 1%) • Hospitalization for unstable angina 108 (1. 5%) 117 (1. 6%) Green JB et al. NEJM 2015; DOI: 10. 1056/NEJMoa 1501352

Secondary Composite Cardiovascular Outcome* ITT Analysis for Superiority * CV death, nonfatal MI, nonfatal stroke Green JB et al. NEJM 2015; DOI: 10. 1056/NEJMoa 1501352

Secondary Composite Cardiovascular Outcome ITT Population Numbers of patients with events Sitagliptin n=7332 Placebo n=7339 Secondary composite CV outcome 745 (10. 2%) 746 (10. 2%) 3. 58 per 100 pyrs 3. 62 per 100 pyrs ITT HR=0. 99 (0. 89, 1. 10), p=0. 84 Individual components • CV death 313 (4. 3%) 293 (4. 0%) • Nonfatal MI 285 (3. 9%) 294 (4. 0%) • Nonfatal stroke 147 (2. 0%) 159 (2. 2%) Green JB et al. NEJM 2015; DOI: 10. 1056/NEJMoa 1501352

All-cause Mortality (ITT Population) Sitagliptin n=7332 547 (7. 5%) 537 (7. 3%) 2. 48 per 100 pyrs All-cause mortality Placebo n=7339 2. 45 per 100 pyrs ITT HR=1. 01 (0. 90, 1. 14), p=0. 88 Non-cardiovascular 167 (2. 3%) 171 (2. 3%) Unknown* 109 (1. 5%) 107 (1. 5%) Cardiovascular • Sudden cardiac death 72 (1. 0%) 73 (1. 0%) • Acute myocardial infarction 21 (0. 3%) 27(0. 4%) • Heart failure 28 (0. 4%) 35 (0. 5%) • Stroke 29 (0. 4%) 36 (0. 5%) • Other cardiovascular 8 (0. 1%) 5 (0. 1%) • Presumed cardiovascular 113 (1. 5%) 83 (1. 1%) * Counted as cardiovascular for the primary analysis Green JB et al. NEJM 2015; DOI: 10. 1056/NEJMoa 1501352

Hospitalization for Heart Failure* ITT Analysis * Adjusted for history of heart failure at baseline Green JB et al. NEJM 2015; DOI: 10. 1056/NEJMoa 1501352

Hospitalization for Heart Failure* ITT Population Numbers of patients with events Sitagliptin n=7332 Placebo n=7339 Hospitalization for heart failure† 228 (3. 1%) 229 (3. 1%) 1. 07 per 100 pyrs 1. 09 per 100 pyrs ITT HR=1. 00 (0. 83, 1. 20), p=0. 98 Hospitalization for heart failure or cardiovascular death† 538 (7. 3%) 525 (7. 2%) 2. 54 per 100 pyrs 2. 50 per 100 pyrs ITT HR=1. 02, (0. 90, 1. 15), p=0. 74 * Adjusted for history of heart failure at baseline † Prespecified analyses Green JB et al. NEJM 2015; DOI: 10. 1056/NEJMoa 1501352

Primary Composite Cardiovascular Outcome Prespecified Subgroup Analyses* (1) * ITT Population Green JB et al. NEJM 2015; DOI: 10. 1056/NEJMoa 1501352

Primary Composite Cardiovascular Outcome Prespecified Subgroup Analyses* (2) * ITT Population Green JB et al. NEJM 2015; DOI: 10. 1056/NEJMoa 1501352

Primary Composite Cardiovascular Outcome Prespecified Subgroup Analyses* (3) * ITT Population Green JB et al. NEJM 2015; DOI: 10. 1056/NEJMoa 1501352

Primary Composite Cardiovascular Outcome Prespecified Subgroup Analyses* (4) * ITT Population Green JB et al. NEJM 2015; DOI: 10. 1056/NEJMoa 1501352

Primary Composite Cardiovascular Outcome Prespecified Subgroup Analyses* (5) * ITT Population Green JB et al. NEJM 2015; DOI: 10. 1056/NEJMoa 1501352

Summary of Results (1) • For the primary composite cardiovascular outcome (CV death, nonfatal MI, nonfatal stroke, or hospitalization for unstable angina) sitagliptin, compared with placebo, was noninferior, and not superior • For the secondary composite cardiovascular outcome (CV death, nonfatal MI, or nonfatal stroke) sitagliptin, compared with placebo, was noninferior, and not superior • The rate of hospitalization for heart failure did not differ between sitagliptin and placebo treatment groups • The incidence of severe hypoglycemia did not differ between sitagliptin and placebo treatment groups Green JB et al. NEJM 2015; DOI: 10. 1056/NEJMoa 1501352

Summary of Results (2) • The rates of infections, and deaths from infection, did not differ between sitagliptin and placebo treatment groups • The incidence of overall malignancies did not differ between sitagliptin and placebo treatment groups • Overall, confirmed events of acute pancreatitis were uncommon, but numerically more frequent in the sitagliptin group • Overall, confirmed events of pancreatic cancer were uncommon, but numerically more frequent in the placebo group Green JB et al. NEJM 2015; DOI: 10. 1056/NEJMoa 1501352

Summary of Results (3) • TECOS was a cardiovascular safety study • The study aimed for glycemic equipoise to minimize possible glycemic confounding effects on the outcomes of interest, with the result that there was only a small difference in the Hb. A 1 c levels between the sitagliptin and placebo groups • The utility of sitagliptin as a glucose-lowering agent was confirmed by the more frequent initiation of insulin therapy and the greater need for additional antihyperglycemic agents in the placebo group compared with the sitagliptin group Green JB et al. NEJM 2015; DOI: 10. 1056/NEJMoa 1501352

Acknowledgements TECOS Executive Committee* Rury Holman, Joint Chair Eric Peterson, Joint Chair Paul Armstrong John Buse Robert Josse Keith Kaufman Joerg Koglin Scott Korn John Lachin Darren Mc. Guire Eberhard Standl Peter Stein Shailaja Suryawanshi Frans Van de Werf * Robert M. Califf served as Joint Chair until taking up the post of deputy FDA commissioner on March 1, 2015 TECOS was conducted jointly by …in an academic collaboration with

Data and Safety Monitoring Board Marc Pfeffer, Chair Stuart Pocock John Mc. Murray Hertzel Gerstein Leif Groop Independent statistician Tim Clayton

Operations Committee Argentina: Isaac Sinay, Australia: David Brieger, Steve Stranks, Belgium: Andre Scheen, Brazil: Antonio Chacra, Renato Lopes Bulgaria: Tsvetalina Tankova Canada: Irene Hramiak Chile: Carlos Raffo Grado China: Junbo Ge, Yang Wen Ying Colombia: Pablo Aschner Czech Republic: Jan Skrha Estonia: Anu Ambos Finland: Timo Strandberg France: Michel Marre, Florence Travert Germany: Markof Hanefeld, Axel Riefflin Hong Kong: Juliana Chan Hungary: Peter Ofner India: Priyadarshini Arambam, Johann Christopher, Mukul Manchanda, Atul Mathur, Sanjay Mittal, N. K. Reddy, Naresh Trehan Israel: Julio Wainstein Italy: Giuseppe Ambrosio Latvia: Valdis Pirags Lithuania: Neli Jakuboniene Malaysia: Mafauzy Mohamed Netherlands: Jan H. Cornel New Zealand: Russell Scott, Harvey White Norway: Sigrun Halvorsen Poland: Andrzej Tykarski Romania: Ioan Andrei Veresiu Russia: Alexander V. Dreval, Inna Misinkova Singapore: E. Shyong Tai Slovakia: Boris Krahulec South Africa: Larry Distiller South Korea: Yong Soo Park Spain: Adela Rovira Sweden: Michael Alvarsson Taiwan: Lee-Ming Chuang Turkey: Tuncay Delibasi United Kingdom: Amanda Adler United States: Helena Rodbard

Patients and Sites • We thank the patients, without whom this study and these analyses would not have been possible • We also thank the many investigators and their staff from 673 sites in 38 countries who worked diligently to help ensure TECOS was run to the highest possible standards