Title Page
Contents
Abstract 13
Chapter Ⅰ. Introduction 16
1. Background 16
A. Non-alcoholic fatty liver disease 16
B. NAFLD and type 2 diabetes 18
C. Sodium-glucose cotransporter-2 Inhibitors 20
D. Glucagon-like peptide-1 receptor agonists 23
E. Recent Changes in Clinical Guidelines 26
2. Study Objectives 27
Chapter Ⅱ. Literature Review 28
1. Randomized Clinical Trials of SGLT-2 Inhibitors 28
A. Randomized Clinical Trials on Cardiovascular Outcomes 28
B. Randomized Clinical Trials on Hepatic Outcomes 32
2. Observational Studies on SGLT-2 Inhibitors 34
3. Randomized Clinical Trials of GLP-1RA 37
A. Randomized Clinical Trials on Cardiovascular Outcomes 37
B. Randomized Clinical Trials on Hepatic Outcomes 40
4. Observational Studies on GLP-1RA 43
Chapter Ⅲ. Methods 46
1. Data Source 46
A. National Health Insurance Claims Database in General 46
B. Comparison of Health Examination Precipitant and National Health Insurance Beneficiary 49
2. Study Design 55
3. Study Population 60
4. Exposure Definition 68
5. Study Outcomes and Follow-Up 70
A. Primary Outcome: Cardiovascular Effectiveness 70
B. Secondary Outcome: Safety 70
C. Exploratory Outcomes: Hepatic Effectiveness 71
6. Potential Confounders 77
7. Propensity Score Methods 81
8. Statistical Analyses 83
9. Sensitivity Analyses 85
10. Ethic approval 88
Chapter Ⅳ. Results 89
1. Identification of Study Cohort 89
A. SGLT-2 inhibitor vs DPP-4 inhibitor cohort 89
B. GLP-1RA vs DPP-4 inhibitor cohort 91
2. Baseline Characteristics of Study Cohort 93
A. SGLT-2 inhibitor vs DPP-4 inhibitor cohort 93
B. GLP-1RA vs DPP-4 inhibitor cohort 99
3. Primary Analysis 106
A. Incidence and Hazard of Cardiovascular Effectiveness Outcome 106
B. Incidence and Hazard of Safety Outcome 116
4. Sensitivity Analysis 120
A. Application of Intention-To-Treat Approach 120
B. Application of Alternative NAFLD definition (Hepatic Steatosis Index) 123
C. Restriction of Health Examination Assessment Periods 125
D. Application of Alternative Propensity Score Methodology (PS fine stratification) 127
E. Application of Competing Risk Model 129
F. Varying Grace Periods 131
G. Comparison between NAFLD severity 135
H. Applying laboratory and anthropometric measurements to analysis 137
I. Effect according to underlying cardiovascular risk 139
5. Exploratory Analysis 141
A. Incidence and Hazard of Hepatic Effectiveness Outcome 141
Chapter Ⅱ. Discussion 147
1. Discussion 147
2. Strength and Limitation 152
A. Strength 152
B. Limitation 153
Chapter Ⅵ. Conclusion 155
List of Abbreviations 157
References 159
논문요약 175
Table 1. Results of cardiovascular or hepatic outcome trials with SGLT-2 inhibitors 22
Table 2. Results of cardiovascular or hepatic outcome trials with GLP-1RA 25
Table 3. Summary of SGLT-2 inhibitors randomized clinical trials on cardiovascular outcomes 31
Table 4. Summary of SGLT-2 inhibitors randomized clinical trials on hepatic outcomes 33
Table 5. Summary of observational studies on SGLT-2 inhibitors 36
Table 6. Summary of GLP-1RA randomized clinical trials on cardiovascular outcomes 39
Table 7. Summary of GLP-1RA randomized clinical trials on hepatic outcomes 42
Table 8. Summary of observational studies on GLP-1RA 45
Table 9. Variables included in the NHIS database 48
Table 10. Target trial emulation summary 56
Table 11. Definitions of conditions to identify study population 62
Table 12. Duration between the latest health examination date and cohort entry date 66
Table 13. Baseline characteristics difference between patients with diabetes who has record of health examination record versus patients with diabetes 66
Table 14. Drug codes to identify exposure of interest 69
Table 15. Definition of study outcomes 73
Table 16. List of covariates for generating propensity score 79
Table 17. Baseline characteristics of before 1:1 propensity score matched patients initiating SGLT-2 inhibitors versus DPP-4 inhibitors according to NAFLD status 95
Table 18. Baseline Characteristics of 1:1 Propensity Score Matched Patients initiating SGLT-2 inhibitors versus DPP-4 inhibitors according to NAFLD status. 97
Table 19. Baseline characteristics of before 1:1 propensity score matched patients initiating GLP-1RAs versus DPP-4 inhibitors according to NAFLD status. 102
Table 20. Baseline Characteristics of 1:1 Propensity Score Matched Patients initiating GLP-1RA versus DPP-4 inhibitors according to NAFLD status. 104
Table 21. Results from sensitivity analyses applying intention-to-treat approach 122
Table 22. Results from sensitivity analyses applying alternative NAFLD definition 124
Table 23. Results from sensitivity analyses restricting health examination assessment period 126
Table 24. Results from sensitivity analysis applying propensity score fine stratification 128
Table 25. Results from sensitivity analysis applying competing risk model 130
Table 26. Results from sensitivity analysis varying grace periods to 90 days 133
Table 27. Results from sensitivity analysis varying grace periods to 45 days 134
Table 28. Results from sensitivity analysis comparison between NAFLD severity 136
Table 29. Results from sensitivity analysis applying additional laboratory and anthropometric measurement 138
Table 30. Results from sensitivity analysis according to underlying cardiovascular risk 140
Table 31. Hepatic effectiveness outcomes for in the 1:1 propensity score matched cohort of new-users of SGLT-2 inhibitors vs new-users of DPP-4 inhibitors varied by NAFLD status 142
Table 32. Hepatic effectiveness outcomes for in the 1:1 propensity score matched cohort of new-users of GLP-1RA vs new-users of DPP-4 inhibitors varied by NAFLD status. 145
Figure 1. Conceptual illustration of the distribution of health examinations within health insurance source 50
Figure 2. Absolute standardized mean difference of characteristics between health examination recipients and the general population in 2010 and 2019, among those aged 65 years and above, or with a history of diabetes 53
Figure 3. Absolute standardized mean difference of characteristics between health examination recipients and the general population in 2010 and 2019, among those with a history of cancer, or coronary artery disease 54
Figure 4. Graphical representation of study design 59
Figure 5. Flow chart of study population selection among new-users of SGLT-2 inhibitors versus DPP-4 inhibitors with varying NAFLD status 90
Figure 6. Flow chart of study population selection among new-users of GLP-1RAs versus DPP-4 inhibitors with varying NAFLD status 92
Figure 7. Cardiovascular effectiveness outcomes for in the 1:1 propensity score matched cohort of new-users of SGLT-2 and new-users of DPP-4 inhibitors varied by NAFLD status 107
Figure 8. Cardiovascular effectiveness outcomes for in the 1:1 propensity score matched cohort of new-users of GLP-1RA vs new-users of DPP-4 inhibitors varied by NAFLD status 109
Figure 9. Cumulative incidence of major adverse cardiovascular events among patients with diabetes varied by NAFLD status in 1:1 propensity score matched cohort of new-users of SGLT-2 inhibitor vs new-users of DPP-4 inhibitor 111
Figure 10. Cumulative incidence of hospitalization for heart failure among patients with diabetes varied by NAFLD status in 1:1 propensity score matched cohort of new-users of SGLT-2 inhibitor vs new-users of DPP-4 inhibitor 112
Figure 11. Cumulative incidence of major adverse cardiovascular events among patients with diabetes varied by NAFLD status in 1:1 propensity score matched cohort of new-users of GLP-1RA vs new-users of DPP-4 inhibitor 114
Figure 12. Cumulative incidence of hospitalization for heart failure among patients with diabetes varied by NAFLD status in 1:1 propensity score matched cohort of new-users of GLP-1RA vs new-users of DPP-4 inhibitor 115
Figure 13. Safety outcomes for in the 1:1 propensity score matched cohort of new-users of SGLT-2 inhibitor vs new-users of DPP-4 inhibitors varied by NAFLD status 117
Figure 14. Safety outcomes for in the 1:1 propensity score matched cohort of new-users of GLP-1RA vs new-users of DPP-4 inhibitors varied by NAFLD status 119