BMJ

Confidential: For Review Only Sodium-Glucose Transport Protein 2 (SGLT-2) inhibitors and -Like Peptide-1 (GLP-1) receptor agonists for : A systematic review and network meta-analysis of randomised controlled trials

Journal: BMJ

Manuscript ID BMJ-2020-058523

Article Type: Research

BMJ Journal: BMJ

Date Submitted by the 17-May-2020 Author:

Complete List of Authors: Palmer, Suetonia; University of Otago Christchurch, Medicine Tendal, Britta; Monash University, School of Public Health and Preventive Medicine Mustafa, Reem; University of Kansas School of Medicine, Internal Medicine, Division of Nephrology and Hypertension; McMaster University, Department of Health Research Methods, Evidence and Impact Vandvik, Per; Innlandet Hospital Trust-divisjon Gjøvik, Department of Medicine Li, Sheyu; Sichuan University, Department of Endocrinology and Metabolism, West China Hospital Hao, Qiukui; Sichuan University West China Hospital; McMaster University Department of Medicine, Department of Health Research Methods, Evidence and Impact Tunnicliffe, David; The University of Sydney, School of Public Health Ruospo, Marinella; University of Bari, Department of Emergency and Transplantation Natale, Patrizia; The University of Sydney, School of Public Health Saglimbene, Valeria; University of Bari, Department of Emergency and Organ Transplantation Nicolucci, Antonio; Center for Outcomes Research and clinical Epidemiology (CORESEARCH) Johnson, David; Translational Research Institute Tonelli, Marcello; University of Calgary, Cumming School of Medicine Rossi, Maria; Center for Outcomes Research and clinical Epidemiology Badve, Sunil; The George Institute for Global Health, UNSW Sydney; Department of Renal Medicine, St George Hospital Nadeau-Fredette, Annie-Claire; Universite de Montreal Faculte de medecine Burke, Michael; University of Queensland; Princess Alexandra Hospital, Faruque, Labib; University of Alberta Lloyd, Anita; University of Alberta Ahmad, Nasreen; University of Alberta Liu, Yuanchen; University of Alberta, Department of Medicine Tiv, Sophanny; University of Alberta Millard, Tanya; Monash University, School of Public Health and

https://mc.manuscriptcentral.com/bmj Page 1 of 244 BMJ

1 2 3 Preventive Medicine 4 Gagliardi, Lucia; The Queen Elizabeth Hospital, Endocrine and Diabetes 5 Unit 6 Kolanu, Nithin; Garvan Institute of Medical Research 7 Barmanray, Rahul; The University of Melbourne, Department of Medicine 8 McMorrow, Rita; The University of Melbourne, Department of General 9 Practice Raygoza Cortez, Ana Karina; Universidad Autonoma de Nuevo Leon 10 Confidential:White, Heath; For Monash University Review Only 11 Chen, Xiangyang; Sichuan University 12 Zhou, Xu; Chinese evidence-based medicine center, West China hospital, 13 Sichuan University 14 Liu, Jiali; West China Hospital, Sichuan University, Chinese Evidence- 15 based Medicine Center 16 Flores Rodriguez, Andrea; Universidad Autonoma de Nuevo Leon Díaz Gonzalez-Colmenero , Alejandro; Universidad Autonoma de Nuevo 17 Leon 18 Wang, Yang; Sichuan University 19 Li, Ling; West China Hospital, Sichuan University, Chinese Evidence- 20 based Medicine Center 21 Suntanto, Surya; The University of Sydney, Charles Perkins Centre 22 Research and Education Hub Cesar Solis, Ricardo; Universidad Autonoma de Nuevo Leon 23 Díaz-González , Fernando; Universidad Autonoma de Nuevo Leon 24 Walsh, Michael; St. Joseph's Hospital, Division of Nephrology 25 Guyatt, Gordon; McMaster University 26 Strippoli, Giovanni; Universita degli Studi di Bari Dipartimento 27 Emergenza e Trapianti di Organi 28 diabetes, network meta-analysis, patient outcomes, mortality, SGLT-2 Keywords: 29 inhibitors, GLP-1 receptor agonists 30 31 32 33 Note: The following files were submitted by the author for peer review, but cannot be converted to PDF. 34 You must view these files (e.g. movies) online. 35 Appendix 2 SGLT2-I and GLP1RA NMA_2020.05.11.xlsx 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 https://mc.manuscriptcentral.com/bmj BMJ Page 2 of 244

1 2 3 4 Sodium-Glucose Transport Protein 2 (SGLT-2) inhibitors and Glucagon-Like Peptide- 5 6 1 (GLP-1) receptor agonists for type 2 diabetes: A systematic review and network meta- 7 8 analysis of randomised controlled trials 9 10 11 Authors 12 Confidential: For Review Only 13 Suetonia C. Palmer, nephrologist and professor1, Britta Tendal, living guidelines program 14 15 manager2 Reem A. Mustafa, nephrologist and methodologist3,4, Per Olav Vandvik, 16 17 18 physician and methodologist5, Sheyu Li, diabetologist and research fellow6,7, Qiukui Hao, 19 20 physician and visiting scholar8, David Tunnicliffe, research fellow9, Marinella Ruospo, 21 22 research fellow10, Patrizia Natale, PhD candidate9, Valeria Saglimbene, research fellow10, 23 24 25 Antonio Nicolucci, diabetologist and professor11, David W. Johnson, nephrologist12, 26 27 Marcello Tonelli, nephrologist13, Maria Chiara Rossi, epidemiologist11, Sunil V. Badve, 28 29 nephrologist14 Yeoungjee Cho, nephrologist12, Annie-Claire Nadeau-Fredette, 30 31 32 nephrologist15, Michael Burke, nephrologist16, Labib I Faruque, clinical assistant17, Anita 33 34 Lloyd, researcher17, Nasreen Ahmad, study coordinator17, Yuanchen Liu, researcher17, 35 36 Sophanny Tiv, health informatics programmer17, Tanya Millard, research fellow2, Lucia 37 38 Gagliardi, endocrinology specialist18, Nithin Kolanu, endocrinology trainee19, Rahul D 39 40 41 Barmanray, endocrinologist and PhD candidate20, Rita McMorrow, general practitioner and 42 43 PhD candidate21, Ana Karina Raygoza Cortez, research trainee22, Heath White, senior 44 45 research officer2, Xiangyang Chen, research fellow6, Xu Zhou, lecturer23, Jiali Liu, 46 47 48 researcher24, Andrea Flores Rodríguez, research trainee22, Alejandro Díaz González- 49 50 Colmenero, research trainee22 Yang Wang, researcher25, Ling Li, assistant professor24, 51 52 Surya Sutanto, research scientist26, Ricardo Solis, researcher22 Fernando Díaz González, 53 54 55 research trainee22 Michael T Walsh, nephrologist27 Gordon Guyatt,professor4, Giovanni 56 57 F.M. Strippoli, professor10 58 59 60 1

https://mc.manuscriptcentral.com/bmj Page 3 of 244 BMJ

1 2 3 Affiliations 4 5 6 1 Department of Medicine, University of Otago, Christchurch, New Zealand 7 2 School of Public Health and Preventive Medicine, Monash University, Melbourne, Australia 8 3 Department of Internal Medicine, Division of Nephrology and Hypertension, University of 9 Kansas, Kansas City, USA 10 4 Department of Health Research Methods, Evidence and Impact, McMaster University, 11 Hamilton, Canada 12 5 InstituteConfidential: of Health and Society, University For ofReview Oslo, Norway Only 13 14 6 Department of Endocrinology and Metabolism, West China Hospital, Sichuan University, 15 China 16 7 Ninewells Hospital, University of Dundee 17 8 The Center for Gerontology and Geriatrics, West China Hospital, Sichuan University, China 18 9 Sydney School of Public Health, The University of Sydney, Sydney, Australia 19 10 Department of Emergency and Transplantation, University of Bari, Bari, Italy 20 11 Center for Outcomes Research and Clinical Epidemiology (CORESEARCH), Pescara, Italy 21 22 12 Department of Nephrology, Division of Medicine, University of Queensland at Princess 23 Alexandra Hospital, Woolloongabba, Australia 24 13 Cumming School of Medicine, University of Calgary, Calgary, Canada 25 14 The George Institute for Global Health, Sydney Australia 26 15 Department of Nephrology, Hôpital Maisonneuve-Rosemont, Montréal, Canada 27 16 Mater Private Clinic, Brisbane, Australia 28 29 17 Department of Nephrology, Faculty of Medicine and Dentistry, University of Alberta, 30 Edmonton, Alberta, Canada 31 18 Endocrine and Diabetes Unit, The Queen Elizabeth Hospital, Woodville, Australia 32 19 Garvan Institute of Medical Research, Sydney, Australia 33 20 Department of Medicine, University of Melbourne, Melbourne, Australia 34 21 Department of General Practice and Primary Health Care, University of Melbourne 35 22 Plataforma INVEST Medicina UANL-KER Unit Mayo Clinic (KER Unit Mexico), 36 37 Universidad Autonoma de Nuevo Leon, Monterrey 64460, Mexico 38 23 Evidence-based Medicine Research Center, Jiangxi University of Traditional Chinese 39 Medicine, Nanchang, China 40 24 Chinese Evidence-based Medicine Center and Cochrane China Center, West China Hospital, 41 Sichuan University, Chengdu, Sichuan, China 42 25 West China School of Medicine, Sichuan University, Chengdu, China 43 26 Faculty of Medicine and Health, Charles Perkins Centre, The University of Sydney, Sydney, 44 45 Australia 46 27 Department of Nephrology, Division of Medicine, McMaster University, Hamilton, Canada 47 48 49 50 Corresponding author: 51 52 Professor Giovanni FM Strippoli, MD PhD 53 Department of Emergency and Organ Transplantation 54 University of Bari 55 Piazza Giulio CESARE, 70124 56 57 Bari 58 Italy 59 [email protected] 60 2

https://mc.manuscriptcentral.com/bmj BMJ Page 4 of 244

1 2 3 Word count for text: 3684 4 5 Word count for abstract: 466 6 7 8 9 10 11 12 Confidential: For Review Only 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 3

https://mc.manuscriptcentral.com/bmj Page 5 of 244 BMJ

1 2 3 What is already known on this topic 4 5 6  Although trial results conflict, sodium glucose transporter 2 (SGLT-2) inhibitors and 7 8 glucagon-like peptide-1 (GLP-1) receptor agonists likely reduce cardiovascular and 9 10 kidney disease when added to other glucose-lowering therapy in adults with type 2 11 12 diabetes.Confidential: For Review Only 13 14 15  Uncertainty remains concerning the relative and absolute benefits and harms of these 16 17 drugs across all important outcomes in patients with type 2 diabetes at varying levels 18 19 of cardiovascular and kidney disease risk. 20 21 22 23 What this study adds 24 25  Benefits shared by SGLT-2 inhibitors and GLP-1 receptor agonists include reductions 26 27 in all-cause and cardiovascular mortality, myocardial infarction, kidney failure and 28 29 30 serious hyperglycaemia (high certainty evidence) and lower body weight (low 31 32 certainty) without incurring severe hypoglycaemia (high certainty). 33 34  The two drug classes appeared to have similar effects on cardiovascular mortality, 35 36 myocardial infarction, kidney failure, health-related quality of life and serious 37 38 39 hyperglycaemia. We found high certainty evidence for potentially important benefits 40 41 of SGLT-2 inhibitors over GLP-1 receptor agonists on all-cause mortality and 42 43 hospitalisation for heart failure, and of GLP-1 receptor agonists over SGLT-2 44 45 46 inhibitors on nonfatal stroke. 47 48  Harms also differed, with SGLT-2 inhibitors causing diabetic ketoacidosis and genital 49 50 infection (high certainty) while GLP-1 receptor agonists may increase severe 51 52 53 gastrointestinal events (low certainty). 54 55 56 57 58 59 60 4

https://mc.manuscriptcentral.com/bmj BMJ Page 6 of 244

1 2 3 Abstract 4 5 Objectives 6 7 8 To evaluate the impact of sodium glucose transporter 2 inhibitors (SGLT2Is) and glucagon- 9 10 like peptide-1 receptor agonists (GLP1RAs) on important outcomes in patients with type 2 11 12 diabetes.Confidential: For Review Only 13 14 15 Design 16 17 Systematic review and network meta-analysis. 18 19 20 Setting 21 22 Medline, Embase and Cochrane Central Register of Controlled Trials up to 1 May 2020. 23 24 25 Eligibility criteria for selecting studies and methods 26 27 We included randomised controlled trials that compared SGLT2Is or GLP1RAs to placebo, 28 29 standard care, or other glucose lowering therapy in adults with type 2 diabetes with follow up 30 31 of 24 weeks or longer. Two reviewers independently screened studies for eligibility, extracted 32 33 34 data, and assessed risk of bias. A parallel guideline committee (BMJ Rapid Recommendation) 35 36 provided critical oversight of the systematic review. We performed frequentist random- 37 38 effects pairwise and network meta-analysis and used GRADE to assess the certainty of 39 40 41 evidence. Results include estimates of absolute benefits and harms of treatment per 1000 42 43 patients treated for 5 years for patients at very low risk (no cardiovascular risk factors), low 44 45 risk (3 or more cardiovascular risk factors), moderate risk (cardiovascular disease), high risk 46 47 (chronic kidney disease), and very high risk (cardiovascular disease and kidney disease). 48 49 50 Results 51 52 730 trials including 402,030 patients proved eligible. All results refer to the addition of 53 54 55 SGLT2Is and GLP1RAs to existing anti-diabetic treatment. SGLT2Is and GLP1RAs lowered 56 57 all-cause mortality. For 1000 patients over 5 years, best estimates of absolute mortality 58 59 reductions with SGLT2Is ranged from 5 fewer for very low risk (moderate certainty) to 61 60 5

https://mc.manuscriptcentral.com/bmj Page 7 of 244 BMJ

1 2 3 fewer for very high risk patients (high certainty) and for GLP1RAs from 2 fewer for very low 4 5 6 risk (moderate certainty) to 32 fewer for very high risk patients (high certainty). SGLT2Is 7 8 and GLP1RAs resulted in reductions in cardiovascular mortality, nonfatal myocardial 9 10 infarction, and kidney failure. Potentially important differences between the two agents 11 12 Confidential: For Review Only included the following: SGLT2Is reduced mortality and hospitalisation for heart failure more 13 14 15 than GLP1RAs and GLP1RAs reduce nonfatal stroke more than SGLT2Is (which appeared to 16 17 have no effect). SGLT2Is caused diabetic ketoacidosis and genital infection (high certainty) 18 19 while GLP1RAs may cause severe gastrointestinal events (low certainty). Neither class of 20 21 22 drugs increased serious hypoglycaemia, and both prevented serious hyperglycaemia. Low 23 24 certainty evidence suggested SGLT2Is and GLP1RAs may lower body weight. There was 25 26 little or no evidence about SGLT2Is or GLP1RAs on limb amputation, blindness, eye disease 27 28 29 requiring intervention, or neuropathic pain and evidence regarding health-related quality of 30 31 life was limited. 32 33 34 Conclusions 35 36 In patients with type 2 diabetes, SGLT2Is and GLP1RAs result in similar reductions in 37 38 cardiovascular mortality, myocardial infarction, and kidney failure. Potential advantages of 39 40 SGLT2Is include larger reductions in all-cause mortality and hospitalisation for heart failure. 41 42 43 Reduction in nonfatal stroke appears to be an advantage of GLP1RAs. SGLT2Is increased 44 45 diabetic ketoacidosis and genital infection and GLP1RAs may increase severe gastrointestinal 46 47 events. Limited evidence is available for other important outcomes. 48 49 50 Systematic review registration 51 52 PROSPERO CRD42019153180 53 54 55 56 57 58 59 60 6

https://mc.manuscriptcentral.com/bmj BMJ Page 8 of 244

1 2 3 Introduction 4 5 6 Diabetes affects half a billion people worldwide and accounted for 1.5 million deaths in 7 8 2016.1 Glucose lowering therapy is a mainstay of treatment.2 In people with type 2 diabetes 9 10 and higher risks of cardiovascular disease, several large-scale randomised controlled trials 11 12 Confidential: For Review Only (RCTs) have investigated sodium-glucose cotransporter (SGLT)-2 inhibitors and glucagon- 13 14 15 like peptide-1 (GLP-1) receptor agonists and reported reductions in cardiovascular mortality 16 17 and nonfatal cardiovascular complications with both drug classes.3-5 The mortality reductions 18 19 have not, however, proved consistent across available trials, leaving clinicians and patients 20 21 3 6 7 22 uncertain of the magnitude of benefits. 23 24 Recommendations released in 2019 by the American Diabetes Association included using 25 26 27 SGLT-2 inhibitors and GLP-1 receptor agonists in the management of diabetes for people 28 29 with cardiovascular disease or kidney disease who have not reached their glycaemic target 30 31 goals.2 European Society of Cardiology guidelines in 2019 recommended SGLT-2 inhibitors 32 33 34 and GLP-1 receptor agonists in people with type 2 diabetes and cardiovascular disease or 35 36 high risk of cardiovascular disease.8 National Institute for Health and Care Excellence 37 38 (NICE) guidance updated in 2019 suggested a stepped approach to treatment intensification 39 40 41 of antidiabetic drugs, considering as the first treatment with additions of dual and 42 43 triple therapy from several drug classes, including SGLT-2 inhibitors.9 44 45 46 Several published systematic reviews and meta-analyses have summarised glucose-lowering 47 48 therapies for people with type 2 diabetes, including a network meta-analysis of stepped 49 50 intensification of therapy added to metformin and .10 Results from more recent 51 52 53 meta-analyses report reductions in mortality and selected cardiovascular and renal events 54 55 with both SGLT-2 inhibitors and GLP-1 receptor agonists.11-14 However, a network meta- 56 57 analysis including all available glucose-lowering therapies, the full range of important 58 59 outcomes, GRADE certainty ratings and estimates of absolute benefits and harms in patients 60 7

https://mc.manuscriptcentral.com/bmj Page 9 of 244 BMJ

1 2 3 with varying risks of cardiovascular and kidney disease remains unavailable. Moreover, 4 5 3 7 15 16 6 several large-scale trials published recently necessitate updated evidence synthesis. 7 8 We therefore conducted a systematic review and network meta-analysis evaluating the 9 10 11 benefits and harms of SGLT-2 inhibitors and GLP-1 receptor agonists in adults with type 2 12 Confidential: For Review Only 13 diabetes. This review is conducted as part of the BMJ Rapid Recommendations project, a 14 15 collaborative initiative from the MAGIC Evidence Ecosystem Foundation 16 17 18 (www.magicproject.org) and The BMJ. The aim of the initiative is to provide trustworthy 19 20 practice guidelines within months of newly released evidence and underpinned by rigorous 21 22 evidence summaries. This systematic review informs a BMJ Rapid Recommendation 23 24 addressing SGLT-2 inhibitors and GLP-1 receptor agonists in type 2 diabetes (box 1), and 25 26 27 although the network meta-analysis includes all anti-diabetic drug classes, focuses on these 28 29 two classes added to other anti-diabetic , and in relation to one another.17 30 31 32 Methods 33 34 Protocol registration 35 36 We registered the protocol for this systematic review with PROSPERO (CRD42019153180). 37 38 39 Guideline panel and patient involvement 40 41 As per the BMJ Rapid Recommendations process, a guideline panel that included content 42 43 experts, diabetologists, nephrologists, internal medicine physicians, primary care physicians, 44 45 46 methodologists and patients provided critical oversight of the review. The panel reviewed the 47 48 protocol, identified the population, selected and ranked important patient outcomes, and 49 50 recommended baseline risks on which absolute treatment effects were calculated. 51 52 53 Search strategy 54 55 The literature search designed by an information specialist was conducted in MEDLINE, 56 57 Embase and the Cochrane Central Register of Controlled Trials from March 2016 to 1 May 58 59 60 8

https://mc.manuscriptcentral.com/bmj BMJ Page 10 of 244

1 2 3 2020 without language restriction (appendix 1). The search from a previous network meta- 4 5 10 6 analysis provided records from database inception to March 2016. 7 8 Study selection 9 10 A team of reviewers, working independently, screened citations and evaluated full-text 11 12 Confidential: For Review Only 13 records for eligible studies using Covidence (Covidence systematic review software, Veritas 14 15 Health Innovation, Melbourne, Australia. Available at www.covidence.org). A third reviewer 16 17 (BT or SP) resolved disagreements by consensus. 18 19 20 Parallel group RCTs were eligible if they compared SGLT-2 inhibitors or GLP-1 receptor 21 22 agonists relative to one another or to other glucose lowering treatments, placebo, or standard 23 24 25 care in adults with type 2 diabetes. We included studies reporting outcomes at 24 weeks or 26 27 longer. Treatment was either given as monotherapy or added to non-randomised background 28 29 glucose lowering management and other therapies. 30 31 32 Using definitions of outcomes specified in individual RCTs, the review addressed all patient- 33 34 important outcomes defined by the guideline panel: all-cause mortality, cardiovascular 35 36 37 mortality, nonfatal myocardial infarction, nonfatal stroke, kidney failure, hospitalisation for 38 39 heart failure, severe hypoglycaemia, blindness, eye disease requiring intervention, health- 40 41 related quality of life, body weight, amputation, neuropathic pain, diabetic ketoacidosis, 42 43 44 serious hyperglycaemia, genital infection, Fournier gangrene, severe gastrointestinal events, 45 46 pancreatic cancer, pancreatitis, and glycated haemoglobin A1C. 47 48 49 Data extraction 50 51 For each eligible study, two reviewers independently extracted the following: study 52 53 characteristics (year of publication, country or countries, funding, duration); population 54 55 (setting, sample size, patient demographics and coexisting illnesses); description of 56 57 58 interventions (drug class, name, dose); and outcomes. Reviewers resolved disagreements 59 60 through discussion or if necessary, consultation with a third reviewer. 9

https://mc.manuscriptcentral.com/bmj Page 11 of 244 BMJ

1 2 3 Risk of bias assessment 4 5 Two reviewers independently assessed risk of bias with adjudication by a third reviewer (SP) 6 7 8 using the Cochrane tool for assessing risk of bias in randomised trials that includes random 9 10 sequence generation, allocation concealment, blinding, missing outcome data and selective 11 12 reportingConfidential: of outcomes.18 Each domain was For judged Review as low, unclear or Onlyhigh risk of bias. 13 14 15 Reviewers considered allocation concealment as low risk if there were no reported methods 16 17 for allocation concealment and participants and investigators were blind to treatment 18 19 allocation. 20 21 22 Data synthesis and analysis 23 24 For each direct comparison of two treatments, we conducted a frequentist pairwise meta- 25 26 analysis and reported with corresponding 95% confidence intervals, odds ratios for 27 28 29 dichotomous outcomes, mean differences for continuous outcomes (body weight and 30 31 glycated haemoglobin A1C) and standardised mean difference for health-related quality of 32 33 life. We assessed statistical heterogeneity using the I2 statistic and funnel plots for evidence 34 35 36 of small study effects in analyses including 10 or more studies. 37 38 We conducted network meta-analysis using frequentist methods with restricted maximum 39 40 41 likelihood methods to estimate network heterogeneity, assuming a common heterogeneity 42 43 estimate within a network. Comparisons of the magnitude of common heterogeneity variance 44 45 for the network (tau [τ]) were made against an empirical distribution of heterogeneity 46 47 48 variances in which values of τ<0.5 were low, 0.5 to 1.0 were high, and above 1.0 extreme 49 50 provided assessments of extent of heterogeneity.19 We assessed agreement between direct and 51 52 indirect estimates in every closed loop of evidence using node-splitting approaches and for 53 54 the entire network using design-by-treatment interaction model.20. We imputed missing 55 56 57 standard deviations for continuous variables when absent using standard deviations borrowed 58 59 from other similar RCTs.21 22 60 10

https://mc.manuscriptcentral.com/bmj BMJ Page 12 of 244

1 2 3 The guideline panel recommended five baseline risk categories to estimate absolute effects of 4 5 6 treatment on cardiovascular and kidney outcomes in different categories of patients, 7 8 reflecting typical clinical scenarios in practice. Patient categories were defined as very low 9 10 risk (no or few (<3) cardiovascular risk factors), low risk (3 or more cardiovascular risk 11 12 Confidential: For Review Only factors), moderate risk (cardiovascular disease), high risk (chronic kidney disease [estimated 13 14 15 glomerular filtration rate 45 to 75 ml/min per 1.73 m2 with albuminuria >300 mg/g (30 16 17 mg/mmol) or estimated glomerular filtration rate 15 to 45 ml/min per 1.73 m2]), and very 18 19 high risk (cardiovascular disease and chronic kidney disease)). The panel calculated absolute 20 21 22 treatment effects of the network estimates based on the event rates from the best sources of 23 24 evidence including cohort studies, risk prediction equations or the placebo arm in available 25 26 trials. We estimated baseline absolute risk per 1000 patients treated for 5 years.23-28 27 28 29 All pairwise and network meta-analysis were performed using STATA 13 MP (StataCorp. 30 31 Stata Statistical Software: Release 13. College Station, TX: StataCorp LP) using published 32 33 29 34 routines. 35 36 Assessment of evidence certainty 37 38 39 The GRADE approach provided the methods to assess the certainty of the evidence for direct 40 41 and network comparisons, using a non-contextualised approach.30 31 Ratings of evidence 42 43 certainty for direct estimates included considerations of risk of bias, inconsistency, 44 45 indirectness, imprecision and publication bias. 46 47 48 The lower of the ratings from the two direct estimates forming the dominant first order loop 49 50 51 provided the starting point for certainty ratings for each indirect estimate with further 52 53 downward rating for intransitivity if present. The estimate that provided the most information 54 55 – direct or indirect – provided the basis for the certainty of the network estimates. If these 56 57 estimates contributed similar amounts of information, we chose the higher of the two 58 59 60 certainty judgments. If evidence of incoherence between direct and indirect estimates existed, 11

https://mc.manuscriptcentral.com/bmj Page 13 of 244 BMJ

1 2 3 the certainty of the network estimate was rated down, and we used the estimate with the 4 5 6 highest certainty – direct or indirect – as the best estimate of the treatment effect. MAGICapp 7 8 provided the platform to develop the GRADE summary of findings tables 9 10 (www.magicproject.org). We also provide the absolute estimates of effect and associated 11 12 Confidential: For Review Only uncertainties across risk groups, derived from best relative estimates of effect from the 13 14 15 network meta-analysis, in evidence summaries and decision aids developed by MAGIC 16 17 (insert link). 18 19 20 Results 21 22 Description of included studies 23 24 The electronic search yielded 18,248 unique records. Screening and full-text article analysis 25 26 identified 730 trials including 402,030 patients (figure 1, appendix 2) comparing 11 glucose- 27 28 29 lowering drugs, placebo, or standard care. Figure 2 shows the network of treatment 30 31 comparisons in all available trials. The trial sample sizes ranged from 16 to 17,160. The 32 33 median trial mean age was 57.0 years and median proportion of men was 55.6%. At baseline, 34 35 36 the median trial mean glycated haemoglobin A1C was 8.1% and body mass index was 30.1 37 38 kg/m2. Eligibility criteria included coronary artery disease or macrovascular disease in 33 39 40 trials,4 6 32-62 atrial fibrillation in 1 trial,37 heart failure in 6 trials,63-68 chronic kidney disease 41 42 3 69-97 98-101 43 and/or albuminuria in 30 trials, dialysis for kidney failure in 4 trials and high risk of 44 45 cardiovascular or kidney disease in 9 trials.5 7 15 16 102-106 Treatment with SGLT-2 inhibitors or 46 47 GLP-1 receptor agonists was generally added to background glucose-lowering therapy. 48 49 50 Risk of bias 51 52 Appendix 3 presents risk of bias for each trial. 289 (39.6%) trials were at low risk of bias in 53 54 random sequence generation and 509 (69.7%) were at low risk of bias in allocation 55 56 57 concealment. 449 trials (61.5%) reported blinding for participants and investigators and 93 58 59 trials (12.7%) reported blinding for outcome assessment. 317 trials (43.4%) were adjudicated 60 12

https://mc.manuscriptcentral.com/bmj BMJ Page 14 of 244

1 2 3 as being at low risk of attrition bias. 336 trials (46.0%) were at low risk of bias from selective 4 5 6 outcome reporting. 7 8 Outcomes 9 10 Appendix 4 presents the network plot for each outcome. There was no evidence of global 11 12 Confidential: For Review Only 13 network inconsistency except for health-related quality of life (appendix 5) and no serious 14 15 concerns of incoherence between direct and indirect evidence (appendix 6). Appendix 7 16 17 presents network estimates for each drug comparison for all outcomes. The anticipated 18 19 20 absolute differences in treatment with SGLT-2 inhibitors or GLP-1 receptor agonists 21 22 compared with placebo are shown in tables 1, 2 and 3 and appendix 8. 23 24 25 All-cause and cardiovascular mortality 26 27 228 trials including 283,997 patients reported all-cause mortality. SGLT-2 inhibitors lowered 28 29 30 all-cause mortality compared to placebo (odds ratio (OR) 0.77 (95% CI 0.71 to 0.84); 5 fewer 31 32 per 1000 in 5 years for very low risk patients (moderate certainty); 16 fewer for low risk 33 34 patients (high certainty); 28 fewer for moderate risk patients (high certainty); 39 fewer for 35 36 37 high risk patients (high certainty); and 61 fewer for very high risk patients (high certainty)) 38 39 (table 3). GLP-1 receptor agonists lowered all-cause mortality compared to placebo (OR 0.88 40 41 (95% CI 0.82 to 0.94); 2, 8, 14, 20, and 32 fewer per 1000 in 5 years for very low, low, 42 43 moderate, high and very high risk patients respectively, moderate to high certainty) (table 3). 44 45 46 SGLT-2 inhibitors reduced all-cause mortality compared to GLP-1 receptor agonists (OR 47 48 49 0.88 (95% CI 0.79 to 0.99); 2,7, 13, 18, and 28 fewer per 1000 in 5 years for very low, low, 50 51 moderate, high and very high risk patients respectively, moderate to high certainty) (table 3). 52 53 54 128 trials including 223,679 patients reported cardiovascular mortality. SGLT-2 inhibitors 55 56 lowered cardiovascular mortality compared to placebo (OR 0.83 (95% CI 0.73 to 0.93); 2, 8, 57 58 13, 19 and 30 fewer per 1000 in 5 years for very low, low, moderate, high and very high risk 59 60 13

https://mc.manuscriptcentral.com/bmj Page 15 of 244 BMJ

1 2 3 patients respectively, moderate to high certainty) (table 2), as did GLP-1 receptor agonists 4 5 6 (OR 0.88 (95% CI 0.80 to 0.96); 2, 6, 9, 13, and 21 fewer per 1000 in 5 years for very low, 7 8 low, moderate, high and very high risk patients respectively, moderate to high certainty, table 9 10 2). SGLT-2 inhibitors and GLP-1 receptor agonists did not have different effects on 11 12 Confidential: For Review Only cardiovascular mortality (OR 0.94 (95% CI 0.81 to 1.10), high certainty, appendix 8). 13 14 15 16 Nonfatal myocardial infarction 17 18 199 trials including 262,253 patients reported nonfatal myocardial infarction. SGLT-2 19 20 inhibitors lowered odds of nonfatal myocardial infarction compared to placebo (OR 0.87 21 22 23 (95% CI 0.79 to 0.97); 4, 8, 14, 16 and 25 fewer per 1000 in 5 years for very low, low, 24 25 moderate, high and very high risk patients respectively, moderate to high certainty, appendix 26 27 8), as did GLP-1 receptor agonists (OR 0.92 (95% 0.85 to 0.99); 2, 5, 9, 10, 15 fewer per 28 29 1000 in 5 years for very low, low, moderate, high and very high risk patients respectively, 30 31 32 moderate to high certainty, appendix 8). SGLT-2 inhibitors and GLP-1 receptor agonists did 33 34 not have different effects on nonfatal myocardial infarction (OR 0.95 (95% CI 0.84 to 1.08); 35 36 high certainty, appendix 8). 37 38 39 40 Nonfatal stroke 41 42 174 trials including 259,619 patients reported nonfatal stroke. SGLT-2 inhibitors had little or 43 44 no effect on nonfatal stroke (OR 1.01 (95% CI 0.89 to 1.14); high certainty, appendix 8). 45 46 GLP-1 receptor agonists reduced nonfatal stroke (OR 0.84 (95% CI 0.76 to 0.93); 5, 9, 17, 19 47 48 49 and 30 fewer per 1000 in 5 years for very low, low, moderate, high and very high risk 50 51 patients respectively, moderate to high certainty, appendix 8). SGLT-2 inhibitors had higher 52 53 odds of nonfatal stroke than GLP-1 receptor agonists (OR 1.20 (95% CI 1.03 to 1.41); high 54 55 56 certainty, appendix 8). 57 58 59 60 14

https://mc.manuscriptcentral.com/bmj BMJ Page 16 of 244

1 2 3 Kidney failure 4 5 32 trials including 90,038 patients reported kidney failure, defined generally as estimated 6 7 2 8 glomerular filtration rate below 15 ml/min per 1.73 m or commencement of kidney 9 10 replacement therapy. SGLT-2 inhibitors reduced kidney failure (OR 0.70 (95% CI 0.56 to 11 12 0.89); 1,3,Confidential: 6, 28 and 44 fewer per 1000 inFor 5 years forReview very low, low, moderate,Only high and very 13 14 15 high risk patients respectively, moderate to high certainty, appendix 8). GLP-1 receptor 16 17 agonists also reduced kidney failure (OR 0.78 (95% CI 0.67 to 0.92); 0, 2, 4, 20, 33 per 1000 18 19 in 5 years fewer for very low, low, moderate, high and very high risk patients respectively, 20 21 22 moderate to high certainty, appendix 8). SGLT-2 inhibitors and GLP-1 receptor agonists 23 24 probably did not have different effects on kidney failure (OR 0.90 (95% CI 0.68 to 1.19); 25 26 moderate certainty). 27 28 29 Hospitalisation for heart failure 30 31 32 143 trials including 230,158 patients reported hospitalisation for heart failure. SGLT-2 33 34 inhibitors reduced hospitalisation for heart failure (OR 0.69 (95% CI 0.61 to 0.78); 2, 9, 25, 35 36 33 and 73 fewer per 1000 in 5 years for very low, low, moderate, high and very high risk 37 38 patients respectively, moderate to high certainty, appendix 8). GLP-1 receptor agonists had 39 40 41 little or no effect on hospitalisation for heart failure (OR 0.93 (95% CI 0.84 to 1.03), high 42 43 certainty, appendix 8). SGLT-2 inhibitors reduced hospitalisation for heart failure compared 44 45 with GLP-1 receptor agonists (OR 0.74 (95% CI 0.63 to 0.87), high certainty). 46 47 48 49 Severe hypoglycaemia 50 51 193 trials including 248,896 patients reported severe hypoglycaemia. There was no difference 52 53 in the odds of serious hypoglycaemia comparing SGLT-2 inhibitor or GLP-1 receptor agonist 54 55 treatment to placebo or each other in high or moderate certainty evidence (appendix 8). 56 57 58 59 60 15

https://mc.manuscriptcentral.com/bmj Page 17 of 244 BMJ

1 2 3 Blindness and eye disease requiring intervention 4 5 7 trials including 68,221 patients reported blindness. The results of the network meta-analysis 6 7 8 were uninformative as blindness was rarely reported and the confidence intervals for 9 10 estimated treatment effects were very wide (appendix 7). The outcome of eye disease 11 12 requiringConfidential: intervention was not specifically For reported Review in any trial. Only 13 14 15 16 Health-related quality of life 17 18 24 trials including 11,451 patients reported health-related quality of life during a median 19 20 follow-up of 6 months. The reported instruments included the Diabetes Treatment 21 22 Satisfaction Questionnaire, the Short Form 12 and 36, the EuroQoL 5 Dimension (EQ-5D), 23 24 25 the Diabetes Quality of Life (DQOL), Treatment Satisfaction Status Scale and the Impact of 26 27 Weight on Quality of Life (IWQOL). SGLT-2 inhibitors had uncertain effects on health- 28 29 related quality of life (standardised mean difference (SMD) 0.13 (95% CI -0.11 to 0.36); low 30 31 32 certainty; appendix 8). GLP-1 receptor agonists may increase health-related quality of life 33 34 (SMD 0.12 (95% CI 0.01 to 0.24); low certainty; appendix 8). There was no evidence that 35 36 SGLT-2 inhibitors and GLP-1 receptor agonists had different effects on health-related quality 37 38 of life (SMD 0.01 (95% CI -0.20 to 0.21); low certainty; appendix 8). 39 40 41 42 Body weight 43 44 454 trials including 221,192 patients reported body weight during a median follow-up of 6 45 46 months. SGLT-2 inhibitor therapy and GLP-1 receptor agonist therapy may lower body 47 48 49 weight (mean difference (MD) for SGLT-2 inhibitors -1.94 kg (95% CI -2.24 to -1.63); low 50 51 certainty; MD for GLP-1 receptor agonists -1.37 kg (95% CI -1.64 to -1.11); low certainty; 52 53 appendix 8). SGLT-2 inhibitors appeared to lower body weight to a greater extent than GLP- 54 55 1 receptor agonists (mean difference -0.56 kg (95% CI -0.94 to -0.18); moderate certainty). 56 57 58 59 60 16

https://mc.manuscriptcentral.com/bmj BMJ Page 18 of 244

1 2 3 Glycated haemoglobin A1C 4 5 587 trials including 236,330 patients reported glycated haemoglobin A1C during a median 6 7 8 follow-up of 6 months. SGLT-2 inhibitors (mean difference -0.61% (95% CI -0.68 to -0.54); 9 10 low certainty) and GLP-1 receptor agonists (mean difference -0.90% (95% CI -0.96 to -0.83); 11 12 low certainty)Confidential: may lower glycated haemoglobin For A1C Review levels more than Only placebo (low 13 14 15 certainty). GLP-1 receptor agonists reduced glycated haemoglobin A1C levels to a greater 16 17 extent than SGLT-2 inhibitors (mean difference -0.28% (95% CI -0.38 to -0.19); high 18 19 certainty). 20 21 22 23 Other outcomes 24 25 Results showed the effects of treatment on amputation or neuropathic pain were very 26 27 uncertain (appendix 8). SGLT-2 inhibitor therapy increased diabetic ketoacidosis with high 28 29 certainty. SGLT-2 inhibitors and GLP-1 receptor agonists reduced serious hyperglycaemia 30 31 32 with high certainty. SGLT-2 inhibitors increased genital infection compared with placebo and 33 34 GLP-1 receptor agonists (high certainty). The effects of treatments on Fournier gangrene 35 36 were uncertain. GLP-1 receptor agonists may incur severe gastrointestinal events (low 37 38 certainty). There is probably no increase in pancreatic cancer or pancreatitis with SGLT-2 39 40 41 inhibitor or GLP-1 receptor agonist therapy (low to moderate certainty). 42 43 44 Discussion 45 46 We found high certainty evidence that SGLT-2 inhibitors and GLP-1 receptor antagonists 47 48 when added to other antidiabetic therapy reduced all-cause mortality, cardiovascular 49 50 mortality, nonfatal myocardial infarction, kidney failure and serious hyperglycaemia, without 51 52 53 increasing severe hypoglycaemia (appendix 8). We also observed notable differences 54 55 between SGLT-2 inhibitor and GLP-1 receptor agonists in high certainty evidence; SGLT-2 56 57 inhibitors reduced all-cause mortality and hospitalisation for heart failure more than GLP-1 58 59 60 receptor agonists while GLP-1 receptor agonists reduced nonfatal stroke more than SGLT-2 17

https://mc.manuscriptcentral.com/bmj Page 19 of 244 BMJ

1 2 3 inhibitors. SGLT-2 inhibitors caused diabetic ketoacidosis and genital infection while GLP-1 4 5 6 receptor agonists may increase severe gastrointestinal events. SGLT-2 inhibitors and GLP-1 7 8 receptor agonists may lower body weight and only low quality evidence supported effects of 9 10 treatment on health-related quality of life. Effects of treatment were uncertain for other 11 12 Confidential: For Review Only important outcomes including limb amputation, neuropathic pain, blindness and eye disease 13 14 15 requiring intervention. 16 17 18 Strength and limitations of study 19 20 The strengths of review include a sensitive search to identify eligible trials and independent 21 22 study identification, selection, data extraction and risk of bias adjudication by two reviewers. 23 24 The review used the GRADE approach to report the certainty of available evidence together 25 26 27 with reporting of estimated absolute risks for all outcomes including patients with varying 28 29 risks of cardiovascular and kidney disease. 30 31 32 Limitations include the heterogeneity in clinical settings in the included trials, although the 33 34 consistency of results across studies diminishes this concern. Some outcomes resulted in 35 36 imprecise estimates of effects and low certainty evidence. SGLT-2 inhibitors failed to reduce 37 38 39 nonfatal stroke in the same way as other cardiovascular endpoints, a finding not scientifically 40 41 intuitive. Trials generally did not include patients at lowest cardiovascular risk. Accordingly, 42 43 evidence certainty was graded down for the lowest risk patients due to indirectness. The 44 45 46 panel agreed that a non-contextualised approach was appropriate to assess imprecision in 47 48 estimates of effects to inform judgements about evidence certainty in the network meta- 49 50 analysis. We therefore did not adjudicate whether the range of treatment effects for each 51 52 53 outcome was trivial, small, moderate or large, assess critical outcomes simultaneously or 54 55 make inferences about their relative value to each other.31 The related BMJ Rapid 56 57 Recommendation will adjudicate the importance of the absolute effects of treatment and rate 58 59 certainty accordingly. This approach will likely lower the adjudicated evidence certainty for 60 18

https://mc.manuscriptcentral.com/bmj BMJ Page 20 of 244

1 2 3 several outcomes that inform the Rapid Recommendations. Whether SGLT-2 inhibitors 4 5 6 combined with GLP-1 receptor agonists given together provide additional benefits compared 7 8 to each treatment alone is not answered by this review. 9 10 11 Relation to prior work 12 Confidential: For Review Only 13 This systematic review has included substantially more trials – including large and recently 14 15 published trials – and patients than previously reported reviews and is based on priority 16 17 18 setting and received oversight by a panel with range of clinical and lived experiences. We 19 20 found similar treatment effects for SGLT-2 inhibitors and GLP-1 receptor agonists on 21 22 mortality as a previous network meta-analysis published in 2016, although the precision in 23 24 the treatment estimates has increased substantially.10 Our results provide an empirical basis to 25 26 27 support guideline recommendations made in 2019 that patients at highest risk of 28 29 cardiovascular disease and kidney disease are likely to experience important benefits on risks 30 31 of cardiovascular events and heart failure with SGLT-2 inhibitors.2 8 This network meta- 32 33 34 analysis aims to provide a broader representation of both relative and absolute estimates of a 35 36 wide range of important clinical outcomes. 37 38 39 Implications of study 40 41 This systematic review of 730 clinical trials provides detailed information for decision- 42 43 makers about the benefits and harms of SGLT-2 inhibitors and GLP-1 receptor agonists on 44 45 46 important outcomes in adults with type 2 diabetes. A core finding indicates there is high 47 48 certainty that SGLT-2 inhibitors reduce all-cause and cardiovascular mortality, nonfatal 49 50 myocardial infarction, kidney failure and hospitalisation for heart failure. GLP-1 receptor 51 52 agonist therapy reduces all-cause and cardiovascular mortality, nonfatal myocardial 53 54 55 infarction, nonfatal stroke and kidney failure. SGLT-2 inhibitor therapy reduces all-cause 56 57 mortality and hospitalisation for heart failure to a greater extent than GLP-1 receptor agonist 58 59 therapy while GLP-1 receptor agonist therapy reduces nonfatal stroke more than SGLT-2 60 19

https://mc.manuscriptcentral.com/bmj Page 21 of 244 BMJ

1 2 3 inhibitors. The impact of SGLT-2 inhibitors and GLP-1 receptor agonists on many other 4 5 6 important patient outcomes is uncertain. This analysis provides evidence regarding absolute 7 8 benefits and harms for the associated BMJ Rapid Recommendation, that has used additional 9 10 evidence sources and methodologies to generate guideline recommendations for decision- 11 12 Confidential: For Review Only making in clinical practice. 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 20

https://mc.manuscriptcentral.com/bmj BMJ Page 22 of 244

1 2 3 References 4 5 6 1. Global Health Observatory (GHO) data. 7 https://www.who.int/gho/mortality_burden_disease/causes_death/top_10/en/ World 8 Health Organization. Geneva, Switzerland. Accessed on 8 March, 2020. 9 10 2. American Diabetes Association. 9. Pharmacologic approaches to glycemic treatment: 11 Standards of Medical Care in Diabetes - 2019. Diabetes Care 2019;42(Supplement 12 1):S90-S102.Confidential: doi: 10.2337/dc19-S009 For Review Only 13 14 3. Perkovic V, Jardine MJ, Neal B, et al. and renal outcomes in type 2 diabetes 15 16 and nephropathy. N Engl J Med 2019;380(24):2295‐306. 17 doi:10.1056/NEJMoa1811744 18 19 4. Zinman B, Wanner C, Lachin JM, et al. , cardiovascular outcomes, and 20 mortality in type 2 diabetes. N Engl J Med 2015;373:2117-28. 21 doi:10.1056/NEJMoa1504720 22 23 5. Marso SP, Daniels GH, Brown-Frandsen K, et al. and cardiovascular outcomes 24 in type 2 diabetes. N Engl J Med 2016;375:311-22. doi:10.1056/NEJMoa1603827 25 26 27 6. Rosenberg AE, Sigmon KN, Somerville MC, et al. and cardiovascular 28 outcomes in patients with type 2 diabetes and cardiovascular disease (Harmony 29 Outcomes): a double-blind, randomised placebo-controlled trial. Lancet 30 2018;392(10157):1519-29. doi:10.1016/S0140-6736(18)32261-X 31 32 7. Wiviott SD, Raz I, Bonaca MP, et al. and cardiovascular outcomes in type 2 33 diabetes. N Engl J Med 2019;380(4):347‐57. doi:10.1056/NEJMoa1812389 34 35 36 8. Cosentino F, Grant PJ, Aboyans V, et al. 2019 ESC Guidelines on diabetes, pre-diabetes, 37 and cardiovascular diseases developed in collaboration with the EASD: The Task 38 Force for diabetes, pre-diabetes, and cardiovascular diseases of the European Society 39 of Cardiology (ESC) and the European Association for the Study of Diabetes 40 (EASD). Eur Heart J 2019;41(2):255-323. doi:10.1093/eurheartj/ehz486 41 42 9. National Institute for Health and Care Excellence. Type 2 diabetes in adults: management 43 (NICE Guideline No. 28). Retrieved from 44 45 https://www.nice.org.uk/guidance/ng28/chapter/1-Recommendations#drug-treatment- 46 2. 2019. 47 48 10. Palmer SC, Mavridis D, Nicolucci A, et al. Comparison of clinical outcomes and adverse 49 events associated with glucose lowering drugs in patients with type 2 diabetes: a 50 meta-analysis. JAMA 2016;316(3):313-24. doi:10.1001/jama.2016.9400 51 52 11. Zheng S, Roddick A, Aghar-Jaffar R, et al. Association between use of sodium-glucose 53 cotransporter 2 inhibitors, glucagon-like peptide 1 agonists, and dipeptidyl peptidase 4 54 55 inhibitors with all-cause mortality in patients with type 2 diabetes: A systematic 56 review and meta-analysis. JAMA 2018;319(15):1580-91. 57 doi:10.1001/jama.2018.3024 58 59 60 21

https://mc.manuscriptcentral.com/bmj Page 23 of 244 BMJ

1 2 3 12. Neuen B, Young TK, Heerspink Hiddo JL, et al. SGLT2 inhibitors for the prevention of 4 kidney failure in patients with type 2 diabetes: A systematic review and meta- 5 6 analysis. Lancet Diabetes Endocrinol 2019;7(11):845-54. doi:10.1016/S2213- 7 8587(19)30256-6 8 9 13. Hussein H, Zaccardi F, Khunti K, et al. Cardiovascular efficacy and safety of sodium- 10 glucose co-transporter-2 inhibitors and glucagon-like peptide-1 receptor agonists: A 11 systematic review and network meta-analysis. Diabet Med 2019;36(4):444-53. 12 doi:10.111/dme.13898Confidential: For Review Only 13 14 14. Arnott C, Li Q, Kang AJ, et al. Sodium glucose cotransporter 2 inhibition for the 15 ‐ 16 prevention of cardiovascular events in patients with type 2 diabetes mellitus: A 17 systematic review and meta‐analysis. J Am Heart Assoc 2020;9(3):e014908. 18 doi:10.1161/JAHA.119.014908 19 20 15. Rosenstock J, Perkovic V, Johansen OE, et al. Effect of vs placebo on major 21 cardiovascular events in adults with type 2 diabetes and high cardiovascular and renal 22 risk: the CARMELINA randomized clinical trial. JAMA 2019;321(1):69‐79. 23 24 doi:10.1001/jama.2018.18269 25 26 16. Rosenstock J, Kahn Steven E, Johansen Odd E, et al. Effect of linagliptin vs 27 on major adverse cardiovascular outcomes in patients with type 2 diabetes: the 28 CAROLINA randomized clinical trial. JAMA 2019;322(12):1155-66. 29 doi:10.1001/jama.2019.13772 30 31 17. Siemeniuk R, Agoritsas T, Macdonald H, et al. Introduction to BMJ Rapid 32 Recommendations. BMJ 2016;354:i5191. doi:10.1136/bmj.i5191 33 34 35 18. Higgins JP, Altman DG, Gotzsche PC, et al. The Cochrane Collaboration's tool for 36 assessing risk of bias in randomised trials. BMJ 2011;343:d5928. 37 doi:10.1136/bmj.d5928 38 39 19. Spiegelhalter D, Abram K, Myles J. Bayesian approaches to clinical trials and health- 40 care evaluation: John Wiley & Sons, Ltd, 2004. 41 42 20. Veroniki AA, Vasiliadis HS, Higgins JP, et al. Evaluation of inconsistency in networks of 43 interventions. Int J Epidemiol 2013;42(1):332-45. doi:10.1093/ije/dys222 44 45 46 21. Dias S, Welton N, Caldwell D, et al. Checking consistency in mixed treatment 47 comparison meta-analysis. Stat Med 2020;29(7-8):932-44. doi:10.1002/sim.3767 48 49 22. Furukawa T, Barbui C, Cipriani A, et al. Imputing missing standard deviations in meta- 50 analyses can provide accurate results. J Clin Epidemiol 2006;59(1):7-10. 51 doi:10.1016/j.clinepi.2005.06.006 52 53 23. Hirji I, Andersson S, Guo Z, et al. Incidence of genital infection among patients with type 54 55 2 diabetes in the UK General Practice Research Database. J Diabetes Compl 56 2012;26:501-05. doi:10.1016/j.jdiacomp.2012.06.012 57 58 59 60 22

https://mc.manuscriptcentral.com/bmj BMJ Page 24 of 244

1 2 3 24. Dieleman J, Kerklaan J, Huygen F, et al. Incidence rates and treatment of neuropathic 4 pain conditions in the general population. Pain 2008;137(3):681-8. 5 6 doi:10.1016/.pain.2008.03.002 7 8 25. Takeuchi M, Kawamura T, Sato I, et al. Population-based incidence of diabetic 9 ketoacidosis in type 2 diabetes: medical claims data analysis in Japan. 10 Pharmacoepidemiol Drug Saf 2018;27(1):123-26. doi:10.1002/pds.4271 11 12 26. Hippisley-CoxConfidential: J, Coupland C. Development For and Review validation of risk Only prediction equations to 13 estimate future risk of blindness and lower limb amputation in patients with diabetes: 14 cohort study. BMJ 2015;351:h5441. doi:10.1136/bmj.h5441 15 16 17 27. Basu S, Sussman J, Berkowitz S, et al. Development and validation of risk equations for 18 complications of type 2 diabetes (RECODe) using individual participant data from 19 randomised trials. Lancet Diabetes Endocrinol 2017;5(10):788-98. 20 doi:10.1016/S2213-8587(17)30221-8 21 22 28. Elley C, Robinson T, Moyes S, et al. Derivation and validation of a renal risk score for 23 people with type 2 diabetes. Diabetes Care 2013;36(10):3113-20. doi:10.2337/dc13- 24 25 0190 26 27 29. Chaimani A, Higgins JP, Mavridis D, et al. Graphical tools for network meta-analysis in 28 STATA. PLoS ONE 2013;8(10):e76654. doi:10.1371/journal.pone.0076654 29 30 30. Brignardello-Petersen R, Bonner A, Alexander P, et al. Advances in the GRADE 31 approach to rate the certainty in estimates from network meta-analysis. J Clin 32 Epidemiol 2018;93:36-44. doi:10.1016/j.jclinepi.2017.10.005 33 34 31. Hultcrantz M, Rind D, Akl E, et al. The GRADE Working Group clarifies the construct 35 36 of certainty of evidence. J Clin Epidemiol 2017;87:4-13. 37 doi:10.1016/j.jclinepi.2017.05.006 38 39 32. Gerstein HC, Ratner RE, Cannon CP, et al. Effect of on progression of 40 coronary atherosclerosis in patients with type 2 diabetes mellitus and coronary artery 41 disease: the assessment on the prevention of progression by rosiglitazone on 42 atherosclerosis in diabetes patients with cardiovascular history trial. Circulation 43 2010;121(10):1176-87. doi:10.1161/CIRCULATIONAHA.109.881003 44 45 46 33. Hedblad B, Zambanini A, Nilsson P, et al. Rosiglitazone and carotid IMT progression 47 rate in a mixed cohort of patients with type 2 diabetes and the resistance 48 syndrome: main results from the Rosiglitazone Atherosclerosis Study. J Intern Med 49 2007;261:293-305. doi:10.1111/j.1365-2796.2007.01767.x 50 51 34. Choi D, Kim SK, Choi SH, et al. Preventative effects of rosiglitazone on restenosis after 52 coronary stent implantation in patients with type 2 diabetes. Diabetes Care 53 54 2004;27(11):2654-60. doi:10.2337/diacare.27.11.2654 55 56 35. Pfeffer MA, Claggett B, Diaz R, et al. in patients with type 2 diabetes and 57 acute coronary syndrome. N Engl J Med 2015;373(23):2247-57. 58 doi:10.1056/NEJMoa1509225 59 60 23

https://mc.manuscriptcentral.com/bmj Page 25 of 244 BMJ

1 2 3 36. Tanaka A, Shimabukuro M, Machii N, et al. Effect of empagliflozin on endothelial 4 function in patients with type 2 diabetes and cardiovascular disease: Results from the 5 6 multicenter, randomized, placebo-controlled, double-blind EMBLEM Trial. Diabetes 7 Care 2019;42(10):e159-e61. doi:10.2337/dc19-1177 8 9 37. Kuramitsu S, Miyauchi K, Yokoi H, et al. Effect of on plaque changes in 10 coronary artery following acute coronary syndrome in diabetic patients: The 11 ESPECIAL-ACS study. J Cardiol 2017;69(1):369-76. doi:10.1016/j.jjcc.2016.08.011 12 Confidential: For Review Only 13 38. White WB, Cannon CP, Heller SR, et al. after acute coronary syndrome in 14 patients with type 2 diabetes. N Engl J Med 2013;369(14):1327-35. 15 16 doi:10.1056/NEJMoa1305889 17 18 39. Finn AV, Oh JS, Hendricks M, et al. Predictive factors for in-stent late loss and coronary 19 lesion progression in patients with type 2 diabetes mellitus randomized to 20 rosiglitazone or placebo. Am Heart J 2009;157(2):383.e1-8. 21 doi:10.1016/j.ahj.2008.11.013 22 23 40. Gantz I, Chen M, Suryawanshi S, et al. A randomized, placebo-controlled study of the 24 25 cardiovascular safety of the once-weekly DPP-4 inhibitor omarigliptin in patients with 26 type 2 diabetes mellitus. Cardiovasc Diabetol 2017;16(1):112. doi:10.1186/s12933- 27 017-0593-8 28 29 41. Hirano M, Nakamura T, Kitta Y, et al. Rapid improvement of carotid plaque echogenicity 30 within 1 month of treatment in patients with acute coronary syndrome. 31 Atherosclerosis 2009;203(2):483-8. doi:10.1016/j.atherosclerosis.2008.07.023 32 33 42. Hirano M, Nakamura T, Obata JE, et al. Early improvement in carotid plaque 34 35 echogenicity by in patients with acute coronary syndromes. Circ J 36 2012;76(6):1452-60. doi:10.1253/circj.CJ-11-1524 37 38 43. Hong SJ, Choi SC, Cho JY, et al. Pioglitazone increases circulating microRNA-24 with 39 decrease in coronary neointimal hyperplasia in type 2 diabetic patients- optical 40 coherence tomography analysis. Circ J 2015;79(4):880-8. doi:10.1253/circj.CJ-14- 41 0964 42 43 44. Kaku K, Daida H, Kashiwagi A, et al. Long-term effects of pioglitazone in Japanese 44 45 patients with type 2 diabetes without a recent history of macrovascular morbidity. 46 Curr Med Res Opin 2009;25(12):2925-32. doi:10.1185/0307990903328124 47 48 45. Laberge A, Brassard P, Arsenault B, et al. Positive effect of the PPAR-gamma agonist 49 rosiglitazone on hemodynamic response to exercise in type 2 diabetic men after 50 coronary artery bypass graft surgery: A 1-yr randomized study. Can J Cardiol 51 2016;32(10 Supplement 1):S223-S24. doi:10.1016/j.cjca.2016.07.356 52 53 54 46. Lee HW, Lee HC, Kim BW, et al. Effects of low dose pioglitazone on restenosis and 55 coronary atherosclerosis in diabetic patients undergoing drug eluting stent 56 implantation. Yonsei Med J 2013;54(6):1313-20. doi:10.3349/ymj.2013.54.6.1313 57 58 59 60 24

https://mc.manuscriptcentral.com/bmj BMJ Page 26 of 244

1 2 3 47. Moriwaki K, Takeuchi T, Fujimoto N, et al. Effect of sitagliptin on coronary flow reserve 4 assessed by magnetic resonance imaging in type 2 diabetic patients with coronary 5 6 artery disease. Circ J 2018;82(8):2119‐27. doi:10.1253/circj.CJ-18-0083 7 8 48. Nishio K, Sakurai M, Kusuyama T, et al. A randomized comparison of pioglitazone to 9 inhibit restenosis after coronary stenting in patients with type 2 diabetes. Diabetes 10 Care 2006;29(1):101-6. doi:10.2337/diacare.29.01.06.dc05-1170 11 12 49. OgasawaraConfidential: D, Shite J, Shinke T, et al. For Pioglitazone Review reduces the necrotic-core Only component 13 in coronary plaque in association with enhanced plasma adiponectin in patients with 14 type 2 diabetes mellitus. Circ J 2009;73(2):343-51. doi:10.1253/circj.cj-08-0699 15 16 17 50. Osman A, Otero J, Brizolara A, et al. Effect of rosiglitazone on restenosis after coronary 18 stenting in patients with type 2 diabetes. Am Heart J 2004;147(5):e23. 19 doi:10.1016/j.ahj.2003.12.006 20 21 51. Nissen SE, Nicholls SJ, Wolski K, et al. Comparison of pioglitazone vs glimepiride on 22 progression of coronary atherosclerosis in patients with type 2 diabetes: the 23 PERISCOPE randomized controlled trial. JAMA 2008;299(13):1561-73. 24 25 doi:10.1001/jama.299.13.1561 26 27 52. Phrommintikul A, Wongcharoen W, Kumfu S, et al. Effects of dapagliflozin vs 28 on cardiometabolic parameters in diabetic patients with coronary artery 29 disease: a randomised study. Br J Clin Pharmacol 2019;85(6):1337‐47. 30 doi:10.1111/bcp.13903 31 32 53. Takagi T, Okura H, Kobayashi Y, et al. A prospective, multicenter, randomized trial to 33 assess efficacy of pioglitazone on in-stent neointimal suppression in type 2 diabetes: 34 35 POPPS (Prevention of In-Stent Neointimal Proliferation by Pioglitazone Study). 36 JACC Cardiol Intv 2009;2(6):524-31. doi:10.1016/j.jcin.2009.04.007 37 38 54. Tanaka A, Komukai S, Shibata Y, et al. Effect of pioglitazone on cardiometabolic profiles 39 and safety in patients with type 2 diabetes undergoing percutaneous coronary artery 40 intervention: a prospective, multicenter, randomized trial. Heart Vessel 41 2018;33(9):965‐77. doi:10.1007/s00380-018-1143-3 42 43 55. Dormandy JA, Charbonnel B, Eckland DJ, et al. Secondary prevention of macrovascular 44 45 events in patients with type 2 diabetes in the PROactive Study (PROspective 46 pioglitAzone Clinical Trial In macroVascular Events): a randomised controlled trial. 47 Lancet 2005;366(9493):1279-89. doi:10.1016/S0140-6736(05)67528-9 48 49 56. Hong J, Zhang Y, Lai S, et al. Effects of metformin versus on cardiovascular 50 outcomes in patients with type 2 diabetes and coronary artery disease. Diabetes Care 51 2013;36(5):1304-11. doi:10.2337/dc12-0719 52 53 54 57. Suzuki K, Tanaka S, Aoki C, et al. Greater efficacy and improved endothelial dysfunction 55 in untreated type 2 diabetes with liraglutide versus sitagliptin. Dokkyo Journal of 56 Medical Sciences 2014;41(3):211-20. 57 58 58. Kato Y, Iwata A, Zhang B, et al. Effects of dipeptidyl peptidase-4 inhibitor sitagliptin on 59 coronary atherosclerosis as assessed by intravascular ultrasound in type 2 diabetes 60 25

https://mc.manuscriptcentral.com/bmj Page 27 of 244 BMJ

1 2 3 mellitus with coronary artery disease. IJC Metab Endocr 2017;16:1-9. 4 doi:10.1016/j.ijcme.2017.06.005 5 6 7 59. Varghese A, Yee MS, Chan CF, et al. Effect of rosiglitazone on progression of 8 atherosclerosis: insights using 3D carotid cardiovascular magnetic resonance. J 9 Cardiovasc Magn Reson 2009;11:24. doi:10.1186/1532-429X-11-24 10 11 60. Wang G, Wei J, Guan Y, et al. Peroxisome proliferator-activated receptor-gamma agonist 12 rosiglitazoneConfidential: reduces clinical inflammatory For Review responses in type Only2 diabetes with coronary 13 artery disease after coronary angioplasty. Metabolism 2005;54(5):590-7. 14 doi:10.1016/j.metabol.2004.11.017 15 16 17 61. Cefalu WT, Leiter LA, de Bruin TW, et al. Dapagliflozin's effects on glycemia and 18 cardiovascular risk factors in high-risk patients with type 2 diabetes: A 24-week, 19 multicenter, randomized, double-blind, placebo-controlled study with a 28-week 20 extension. Diabetes Care 2015;38(7):1218-27. doi:10.2337/dc14-0315 21 22 62. Verma S, Mazer CD, Yan AT, et al. Effect of empagliflozin on left ventricular mass in 23 patients with type 2 diabetes and coronary artery disease: the EMPA-HEART 24 25 CardioLink-6 randomized clinical trial. Circulation 2019. 26 doi:10.1161/CIRCULATIONAHA.119.042375 27 28 63. Oe H, Nakamura K, Kihara H, et al. Comparison of effects of sitagliptin and on 29 left ventricular diastolic dysfunction in patients with type 2 diabetes: Results of the 30 3D trial. Cardiovasc Diabetol 2015;14(1):83. doi:10.1186/s12933-015-0242-z 31 32 64. Arturi F, Succurro E, Miceli S, et al. Liraglutide improves cardiac function in patients 33 with type 2 diabetes and chronic heart failure. Endocrine 2017;57(3):464‐73. 34 35 doi:10.1007/s12020-016-1166-4 36 37 65. Tanaka A, Hisauchi I, Taguchi I, et al. Effects of canagliflozin in patients with type 2 38 diabetes and chronic heart failure: a randomized trial (CANDLE). ESC Heart Failure 39 2020. doi:10.1002/ehf2.12707 40 41 66. Dargie HJ, Hildebrandt PR, Riegger GA, et al. A randomized, placebo-controlled trial 42 assessing the effects of rosiglitazone on echocardiographic function and cardiac status 43 in type 2 diabetic patients with New York Heart Association Functional Class I or II 44 45 Heart Failure. J Am Coll Cardiol 2007;49(16):1696-704. 46 doi:10.1016/j.jacc.2006.10.077 47 48 67. Giles TD, Miller AB, Elkayam U, et al. Pioglitazone and heart failure: results from a 49 controlled study in patients with type 2 diabetes mellitus and systolic dysfunction. J 50 Card Fail 2008;14(6):445-52. doi:10.1016/j.cardfail.2008.02.007 51 52 68. McMurray JJV, Ponikowski P, Bolli GB, et al. Effects of vildagliptin on ventricular 53 54 function in patients with type 2 diabetes mellitus and heart failure: A randomized 55 placebo-controlled trial. JACC: heart failure 2018;6(1):8‐17. 56 doi:10.1016/j.jchf.2017.08.004 57 58 59 60 26

https://mc.manuscriptcentral.com/bmj BMJ Page 28 of 244

1 2 3 69. Allegretti AS, Zhang W, Zhou W, et al. Safety and effectiveness of bexagliflozin in 4 patients with type 2 diabetes mellitus and stage 3a/3b CKD. Am J Kidney Dis 5 6 2019;74(3):328-37. doi:10.1053/j.ajkd.2019.03.417 7 8 70. Morikawa A, Ishizeki K, Iwashima Y, et al. Pioglitazone reduces urinary albumin 9 excretion in renin-angiotensin system inhibitor-treated type 2 diabetic patients with 10 hypertension and microalbuminuria: the APRIME study. Clin Exp Nephrol 11 2011;15(6):848-53. doi:10.1007/s10157-011-0512-3 12 Confidential: For Review Only 13 71. Arjona Ferreira JC, Marre M, Barzilai N, et al. Efficacy and safety of sitagliptin versus 14 glipizide in patients with type 2 diabetes and moderate-to-severe chronic renal 15 16 insufficiency. Diabetes Care 2013;36(5):1067-73. doi:10.2337/dc12-1365/-/DC1 17 18 72. Tuttle KR, Lakshmanan MC, Rayner B, et al. versus in 19 patients with type 2 diabetes and moderate-to-severe chronic kidney disease 20 (AWARD-7): a multicentre, open-label, randomised trial. Lancet Diabetes Endocrinol 21 2018;6(8):605. doi:10.1016/S2213-8587(18)30104-9 22 23 73. Bakris GL, Ruilope LM, McMorn SO, et al. Rosiglitazone reduces microalbuminuria and 24 25 blood pressure independently of glycemia in type 2 diabetes patients with 26 microalbuminuria. J Hypertens 2006;24(10):2047-55. 27 doi:10.1097/1.hjh.0000244955.39491.88 28 29 74. Chacra A, Gantz I, Mendizabal G, et al. A randomised, double-blind, trial of the safety 30 and efficacy of omarigliptin (a once-weekly DPP-4 inhibitor) in subjects with type 2 31 diabetes and renal impairment. Int J Clin Pract 2017;71(6):e12955. 32 doi:10.1111/ijcp.12955 33 34 35 75. Pollock C, Stefansson B, Reyner D, et al. Albuminuria-lowering effect of dapagliflozin 36 alone and in combination with and effect of dapagliflozin and saxagliptin 37 on glycaemic control in patients with type 2 diabetes and chronic kidney disease 38 (DELIGHT): a randomised, double-blind, placebo-controlled trial. Lancet Diabetes 39 Endocrinol 2019;7(6):429‐41. doi:10.1016/S2213-8587(19)30086-5 40 41 76. Fioretto P, Del Prato S, Buse JB, et al. Efficacy and safety of dapagliflozin in patients 42 with type 2 diabetes and moderate renal impairment (chronic kidney disease stage 43 44 3A): the DERIVE Study. Diab Obes Metab 2018;20(11):2532‐40. 45 doi:10.1111/dom.13413 46 47 77. Yale JF, Bakris G, Cariou B, et al. Efficacy and safety of canagliflozin over 52 weeks in 48 patients with type 2 diabetes mellitus and chronic kidney disease. Diabetes Obes 49 Metab 2014;16(10):1016-27. doi:10.1111/dom.12348 50 51 78. Barnett AH, Mithal A, Manassie J, et al. Efficacy and safety of empagliflozin added to 52 53 existing antidiabetes treatment in patients with type 2 diabetes and chronic kidney 54 disease: A randomised, double-blind, placebo-controlled trial. Lancet Diabetes 55 Endocrinol 2014;2(5):369-84. doi:10.1016/S2213-8587(13)0208-0 56 57 79. Hiramatsu T, Asano Y, Mabuchi M, et al. Liraglutide relieves cardiac dilated function 58 than DPP-4 inhibitors. Eur J Clin Invest 2018;48(10):e13007. doi:10.1111/eci.13007 59 60 27

https://mc.manuscriptcentral.com/bmj Page 29 of 244 BMJ

1 2 3 80. Hu ZY, Zhao JW. Effect of rosiglitazone on serum CRP and plasma PAI-1 in patients 4 with early type 2 diabetic nephropathy. Central Plains Medical Journal 5 6 2007;34(20):27-8. 7 8 81. Jerums G, Murray RM, Seeman E, et al. Lack of effect of on early diabetic 9 nephropathy and retinopathy: a two-year controlled study. Diabetes Res Clin Pract 10 1987;3(2):71-80. doi:10.1016/s0168-8227(87)80010-4 11 12 82. KohanConfidential: DE, Fioretto P, Tang W, et al. For Long-term Review study of patients Only with type 2 diabetes 13 and moderate renal impairment shows that dapagliflozin reduces weight and blood 14 pressure but does not improve glycemic control. Kidney Int 2014;85(4):962-71. 15 16 doi:10.1038/ki.2013.356 17 18 83. Kashiwagi A, Takahashi H, Ishikawa H, et al. A randomized, double-blind, placebo- 19 controlled study on long-term efficacy and safety of treatment in patients 20 with type 2 diabetes mellitus and renal impairment: Results of the Long-Term 21 ASP1941 Safety Evaluation in Patients with Type 2 Diabetes with Renal Impairment 22 (LANTERN) study. Diab Obes Metab 2015;17(2):152-60. doi:10.1111/dom.12403 23 24 25 84. Davies MJ, Bain SC, Atkin SL, et al. Efficacy and safety of liraglutide versus placebo as 26 add-on to glucose-lowering therapy in patients with type 2 diabetes and moderate 27 renal impairment (LIRA-RENAL): A randomized clinical trial. Diabetes Care 28 2016;39(2):222-30. doi:10.2337/dc14-2883 29 30 85. Lukashevich V, Schweizer A, Shao Q, et al. Safety and efficacy of vildagliptin versus 31 placebo in patients with type 2 diabetes and moderate or severe renal impairment: a 32 prospective 24-week randomized placebo-controlled trial. Diabetes Obes Metab 33 34 2011;13(10):947-54. doi:10.1111/j.1463-1326.2011.01467.x 35 36 86. Groop PH, Cooper ME, Perkovic V, et al. Dipeptidyl peptidase-4 inhibition with 37 linagliptin and effects on hyperglycaemia and albuminuria in patients with type 2 38 diabetes and renal dysfunction: Rationale and design of the MARLINA-T2DTM trial. 39 Diabetes Vasc Dis Res 2015;12(6):455-62. doi:10.1177/1479164115579002 40 41 87. McGill JB, Sloan L, Newman J, et al. Long-term efficacy and safety of linagliptin in 42 patients with type 2 diabetes and severe renal impairment: a 1-year, randomized, 43 44 double-blind, placebo-controlled study. Diabetes Care 2013;36(2):237-44. 45 doi:10.2337/dc12-0706 46 47 88. Nakamura T, Ushiyama C, Osada S, et al. Pioglitazone reduces urinary podocyte 48 excretion in type 2 diabetes patients with microalbuminuria. Metabolism 49 2001;50(10):1193-6. doi:10.1053/meta.2001.26703 50 51 89. Nakamura T, Matsuda T, Kawagoe Y, et al. Effect of pioglitazone on carotid intima- 52 53 media thickness and arterial stiffness in type 2 diabetic nephropathy patients. 54 Metabolism 2004;53(10):1382-6. doi:10.1016/j.metabol.2004.05.013 55 56 90. Nakamura T, Sugaya T, Kawagoe Y, et al. Effect of pioglitazone on urinary liver-type 57 fatty acid-binding protein concentrations in diabetes patients with microalbuminuria. 58 Diabetes Metab Res Rev 2006;22(5):385-9. doi:10.1002/dmrr.633 59 60 28

https://mc.manuscriptcentral.com/bmj BMJ Page 30 of 244

1 2 3 91. Neff KJ, Tobin LM, Hogan AE, et al. The effect of low dose liraglutide on renal 4 inflammation in type 2 diabetic kidney disease: A randomised controlled study. 5 6 Diabet Med 2016;33(Suppl 1):64. doi:10.1111/dme.17_13048 7 8 92. Nowicki M, Rychlik I, Haller H, et al. Long-term treatment with the dipeptidyl peptidase- 9 4 inhibitor saxagliptin in patients with type 2 diabetes mellitus and renal impairment: 10 a randomised controlled 52-week efficacy and safety study. Int J Clin Pract 11 2011;65(12):1230-9. doi:10.1111/j.1742-1241.2011.02812.x 12 Confidential: For Review Only 13 93. Pistrosch F, Passauer J, Herbrig K, et al. Effect of treatment on 14 proteinuria and renal hemodynamic in type 2 diabetic patients with overt nephropathy. 15 16 Horm Metab Res 2012;44(12):914-8. doi:10.1055/s-0032-1314836 17 18 94. Strom Halden TA, Kvitne KE, Midtvedt K, et al. Efficacy and safety of empagliflozin in 19 renal transplant recipients with posttransplant diabetes mellitus. Diabetes Care 20 2019;42(6):1067‐74. doi:10.2337/dc19-0093 21 22 95. Takashima H, Yoshida Y, Nagura C, et al. Renoprotective effects of canagliflozin, a 23 sodium glucose cotransporter 2 inhibitor, in type 2 diabetes patients with chronic 24 25 kidney disease: a randomized open-label prospective trial. Diabetes Vasc Dis Res 26 2018;15(5):469‐72. doi:10.1177/1479164118782872 27 28 96. Grunberger G, Camp S, Johnson J, et al. in patients with Stage 3 chronic 29 kidney disease and type 2 diabetes mellitus: the VERTIS RENAL randomized study. 30 Diabetes Ther 2018;9(1):49‐66. doi:10.1007/s13300-017-0337-5 31 32 97. Yale JF, Bakris G, Cariou B, et al. Efficacy and safety of canagliflozin in subjects with 33 type 2 diabetes and chronic kidney disease. Diabetes Obes Metab 2013;15(5):463-73. 34 35 doi:10.1111/dom.12090 36 37 98. Abe M, Okada K, Maruyama T, et al. Clinical effectiveness and safety evaluation of long- 38 term pioglitazone treatment for erythropoietin responsiveness and insulin resistance in 39 type 2 diabetic patients on hemodialysis. Expert Opin Pharmacother 40 2010;11(10):1611-20. doi:10.1517/1456566.2010.495119 41 42 99. Abe M, Higuchi T, Moriuchi M, et al. Efficacy and safety of saxagliptin, a dipeptidyl 43 peptidase-4 inhibitor, in hemodialysis patients with diabetic nephropathy: A 44 45 randomized open-label prospective trial. Diabetes Res Clin Pract 2016;116:244-52. 46 doi:10.1016/j.diabres.2016.04.034 47 48 100. Arjona Ferreira JC, Corry D, Mogensen CE, et al. Efficacy and safety of sitagliptin in 49 patients with type 2 diabetes and ESRD receiving dialysis: a 54-week randomized 50 trial. Am J Kidney Dis 2013;61(4):579-87. doi:10.1053/j.ajkd.2012.11.043 51 52 101. Ito M, Abe M, Okada K, et al. The dipeptidyl peptidase-4 (DPP-4) inhibitor vildagliptin 53 54 improves glycemic control in type 2 diabetic patients undergoing hemodialysis. 55 Endocrine J 2011;58(11):979-87. doi:10.1507/encrj.EJ11-0025 56 57 102. Neal B, Perkovic V, Mahaffey KW, et al. Canagliflozin and cardiovascular and renal 58 events in type 2 diabetes. N Engl J Med 2017;377(7):644-57. 59 doi:10.1056/NEJMoa1611925 60 29

https://mc.manuscriptcentral.com/bmj Page 31 of 244 BMJ

1 2 3 103. Li JL, Feng ZP, Li QF, et al. Insulin glargine effectively achieves glycemic control and 4 improves insulin resistance in patients with early type 2 diabetes that exhibit a high 5 6 risk for cardiovascular disease. Exp Ther Med 2014;8(1):147-52. 7 doi:10.3892/etm.2014.1688 8 9 104. McGuire DK, Abdullah SM, See R, et al. Randomized comparison of the effects of 10 rosiglitazone vs. placebo on peak integrated cardiovascular performance, cardiac 11 structure, and function. Eur Heart J 2010;31(18):2262-70. 12 doi:10.1093/eurheartj/ehq228Confidential: For Review Only 13 14 105. Scirica BM, Bhatt DL, Braunwald E, et al. Saxagliptin and cardiovascular outcomes in 15 16 patients with type 2 diabetes mellitus. N Engl J Med 2013;369(14):1317-26. 17 doi:10.1056/NEJMoa1307684 18 19 106. Yee MS, Pavitt DV, Dhanjil S, et al. The effects of rosiglitazone on atherosclerotic 20 progression in patients with Type 2 diabetes at high cardiovascular risk. Diabet Med 21 2010;27(12):1392-400. doi:10.1111/j.1464-5491.2010.03089.x 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 30

https://mc.manuscriptcentral.com/bmj BMJ Page 32 of 244

1 2 3 Acknowledgements 4 5 We thank members of the Rapid Recommendations panel for critical feedback on the review 6 7 8 question and outcome selection, GRADE judgments and manuscript. We thank Ruth Mitchell 9 10 for developing the search strategy. 11 12 Confidential: For Review Only 13 Contributors 14 15 SCP, AN, GFMS conceived the study, SCP and GFMS designed the search strategy, SCP 16 17 performed the literature search, SCP, BT, MR, PN, VS, DWJ, GDB, SVB, YC, ACNF, MB, 18 19 LB, AL, NA, YL, ST, TM, LG, NK, RB, RM, AKRC, HW, XYC, XZ, JL, AFR, ADGC, 20 21 22 YW, LL, SS, RS, FDG, MTW screened studies for eligibility, SCP, MR, PN assessed the risk 23 24 of bias, SCP, BT, MR, PN, VS, DWJ, GDB, SVB, YC, ACNF, MB, LB, AL, NA, YL, ST, 25 26 TM, LG, NK, RB, RM, AKRC, HW, XYC, XZ, JL, AFR, ADGC, YW, LL, SS, RS, FDG 27 28 29 performed data extraction, SCP, BT, RM, POV, TA, DT, MTW, GG, GFMS interpreted the 30 31 data analysis, SCP, MR, PN, GFMS assessed the certainty of the evidence, SCP wrote the 32 33 first draft of the manuscript and all other authors revised the manuscript. SCP and GFMS are 34 35 36 guarantors. The corresponding author attests that all listed authors meet authorship criteria 37 38 and that no others meeting the criteria have been omitted. 39 40 41 Funding 42 43 This systematic review did not receive any funding. 44 45 Competing interests 46 47 48 All authors have completed the ICMJE uniform disclosure form at 49 50 www.icmje.org/coi_disclosure.pdf and declare: SCP, BT, RAM, POV, SL, QH, DT, MR, PN, 51 52 VS, SVB, YC, ACNF, MB, LIF, AL, NA, YL, ST, TM, AL, NK, RDB, RM, AKRC, HW, 53 54 CX, XZ, JL, JW, AFR, ADGC, YW, LL, SS, RS, FG, MTW, GG, GFMS: no support from 55 56 57 any organisation for the submitted work; no financial relationships with any organisations 58 59 that might have an interest in the submitted work in the previous three years. AN reports 60 31

https://mc.manuscriptcentral.com/bmj Page 33 of 244 BMJ

1 2 3 funding to institution from AstraZeneca, Sanofi, NovoNordisk, Shionogi, SOBI, AlfaSigma, 4 5 6 Medtronic, Artsana, Theras; board membership for AstraZeneca, NovoNordisk and Eli Lilly 7 8 (consultancy). MCR reports funding to institution from AstraZeneca, Sanofi, NovoNordisk, 9 10 Shionogi, SOBI, AlfaSigma, Medtronic, Artsana, Theras. MT reports grant to institution by 11 12 Confidential: For Review Only Daichi Sankyo in lieu of a personal honorarium. SB reports advisory board membership to 13 14 15 Bayer Australia, speaking honoraria from Bayer Australia and Pfizer Australia; nonfinancial 16 17 research support from Bayer AG. 18 19 20 Patient consent 21 22 No required 23 24 25 Ethical approval 26 27 Not required 28 29 30 Data sharing 31 32 No additional data available 33 34 Transparency 35 36 37 The manuscript’s guarantors (SCP and GFMS) affirm that the manuscript is an honest, 38 39 accurate and transparent account of the study being reported; that no important aspects of the 40 41 study have been omitted; and that any discrepancies from the study as planned and registered 42 43 have been explained. 44 45 46 This manuscript has not been deposited as a preprint. 47 48 49 50 51 52 53 54 55 56 57 58 59 60 32

https://mc.manuscriptcentral.com/bmj BMJ Page 34 of 244

1 2 3 Box 1: Linked resources in the BMJ Rapid Recommendations cluster 4 5  Reference to guideline paper 6 7 8  Reference to network meta-analysis 9 10 11  Reference to prognostic systematic review 12 Confidential: For Review Only 13  Reference to patient perspectives meta-synthesis 14 15  Reference to patient focus group study 16 17 18  Reference to MAGICapp public guideline 19 20 21  Reference to MAGIC multiple comparisons evidence summaries and decision aids 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 33

https://mc.manuscriptcentral.com/bmj Page 35 of 244 BMJ

1 2 3 Table 1 Summary of anticipated absolute differences comparing SGLT-2 inhibitor therapy with placebo treatment per 1000 patients with diabetes 4 type 2 from very low to very high cardiovascular risk, treated for 5 years 5 6 Nonfatal Hospitalisation Diabetic Genital 7 All-cause Cardiovascular myocardial Nonfatal for heart 8 ketoacidosis infection Body weight Risk mortalityConfidential:mortality infarction strokeForKidney Review failure failure Only 9 0 more 10 Very low (no 5 fewer 2 fewer 4 fewer 1 fewer 2 fewer (3 fewer to 4 11 cardiovascular (3 to 6 fewer) (1 to 4 fewer) (1 to 6 fewer) (0 to 1 fewer) (1 to 2 fewer) more) 12 risk factors)      13  Low (3 or 16 fewer 1 more 14 8 fewer 8 fewer 3 fewer 9 fewer 15 fewer (11 to 20 (6 fewer to 8 (3 to 12 fewer) (2 to 12 fewer) (1 to 4 fewer) (7 to 12 fewer) 16 cardiovascular fewer) more)     17 risk factors)   18 28 fewer 1 more -1.94 kg Moderate 13 fewer 14 fewer 6 fewer 25 fewer 2 more 177 more 19 (19 to 35 (12 fewer to 15 (-2.24 to - (cardiovascular (6 to 21 fewer) (3 to 23 fewer) (2 to 9 fewer) (18 to 31 fewer) (0 to 12 (131 to 232 20 fewer) more) 1.63) during 21 disease)     more) more)   6 months 22   39 fewer 1 more 28 fewer  23 High (chronic 19 fewer 16 fewer 33 fewer (27 to 49 (13 fewer to 17 (10 to 40 24 kidney (8 to 30 fewer) (4 to 25 fewer) (23 to 41 fewer) fewer) more) fewer) 25 disease)    26    Very high 27 61 fewer 2 more 44 fewer (cardiovascular 30 fewer 25 fewer 73 fewer 28 (42 to 77 (21 fewer to 27 (16 to 65 29 and chronic (12 to 47 fewer) (6 to 40 fewer) (52 to 92 fewer) fewer) more) fewer) 30 kidney       31 disease) 32 33 Risk categories represent the following patient populations: Very low; No cardiovascular risk factors; Low; three or more cardiovascular risk factors: 34 Moderate; cardiovascular disease: High; chronic kidney disease (reduced glomerular filtration rate or macroalbuminuria): Very high; Cardiovascular disease 35 and chronic kidney disease. Certainty of the evidence for each estimate is shown: High certainty ; Moderate certainty ; Low certainty ; Very 36 low certainty . 37 38 39 40 41 34 42 43 https://mc.manuscriptcentral.com/bmj 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Page 36 of 244

1 2 3 Table 2 Summary of anticipated absolute differences comparing GLP-1 receptor agonist therapy with placebo treatment per 1000 patients with 4 diabetes type 2 from very low to very high cardiovascular risk, treated for 5 years 5 6 Severe 7 Nonfatal Hospitalisation gastro- 8 All-causeConfidential: Cardiovascular myocardial ForNonfatal Review forOnly heart intestinal 9 10 Risk mortality mortality infarction stroke Kidney failure failure events Body weight 11 12 Very low (No 2 fewer 2 fewer 2 fewer 5 fewer 0 fewer 0 fewer 13 cardiovascular (1 to 4 fewer) (1 to 3 fewer) (0 to 5 fewer) (2 to 7 fewer) (0 to 1 fewer) (0 to 1 fewer) 14 risk factors)       15 Low (3 or 2 fewer 16 8 fewer 6 fewer 5 fewer 9 fewer 2 fewer fewer (5 fewer to 1 17 (4 to 13 fewer) (2 to 9 fewer) (1 to 9 fewer) (4 to 14 fewer) (1 to 3 fewer) cardiovascular more) 18      19 risk factors)  6 fewer -1.37 kg 20 Moderate 14 fewer 9 fewer 9 fewer 17 fewer 4 fewer (13 fewer to 2 58 more (9 to (-1.64 to - 21 (cardiovascular (7 to 22 fewer) (3 to 16 fewer) (1 to 16 fewer) (8 to 26 fewer) (2 to 7 fewer) 22 more) 153 more) 1.11) during disease)      23   6 months 24 7 fewer  High (chronic 20 fewer 13 fewer 10 fewer 19 fewer 20 fewer 25 (17 fewer to 3 kidney (10 to 31 (4 to 22 fewer) (1 to 18 fewer) (8 to 29 fewer) (7 to 30 fewer) 26 more) disease) fewer)      27  28 Very high 29 32 fewer 30 fewer 33 fewer 16 fewer (cardiovascular 21 fewer 15 fewer 30 (16 to 38 (13 to 46 (12 to 49 (38 fewer to 7 and chronic (7 to 35 fewer) (2 to 29 fewer) 31 fewer) fewer) fewer) more) kidney   32     33 disease) 34 Risk categories represent the following patient populations: Very low; No cardiovascular risk factors; Low; three or more cardiovascular risk factors: 35 Moderate; cardiovascular disease: High; chronic kidney disease (reduced glomerular filtration rate or macroalbuminuria): Very high; Cardiovascular disease 36 and chronic kidney disease. Certainty of the evidence for each estimate is shown: High certainty ; Moderate certainty ; Low certainty ; Very 37 low certainty . 38 39 40 41 35 42 43 https://mc.manuscriptcentral.com/bmj 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 37 of 244 BMJ

1 2 3 Table 3 GRADE summary of findings to illustrate absolute effects based on cardiovascular and renal risk, for all-cause mortality for SGLT-2 4 inhibitors and GLP-1 receptor agonists compared to placebo or each other 5 6 7 Anticipated absolute effects over 5 years 8 Confidential: For Review Only 9 Anticipated Certainty 10 absolute in 11 effects (95% treatment 12 Relative Risk with Risk with CI) over 5 effects 13 Comparison effect Baseline risk control intervention years (GRADE) Plain text summary 14 15 SGLT 2 OR 0.77 Very low risk Placebo: 20 SGLT-2 5 fewer per Moderate SGLT-2 inhibitor therapy probably 16 inhibitor v (0.71 to 0.84) per 1000 inhibitor: 15 1000 due to reduces all-cause mortality in people 17 placebo per 1000 (from 3 fewer indirectness with diabetes and few or no 18 to 6 fewer) cardiovascular risk factors. 19 20 Low risk Placebo: 70 SGLT-2 16 fewer per High SGLT-2 inhibitor therapy reduces all- 21 per 1000 inhibitor: 54 1000 cause mortality in people with diabetes 22 per 1000 (from 11 fewer and cardiovascular risk factors. 23 to 20 fewer) 24 Moderate risk Placebo: 120 SGLT-2 28 fewer per High SGLT-2 inhibitor therapy reduces all- 25 per 1000 inhibitor: 92 1000 cause mortality in people with diabetes 26 per 1000 (from 19 fewer and established cardiovascular disease 27 to 35 fewer) 28 High risk Placebo: 170 SGLT-2 39 fewer per High SGLT-2 inhibitor therapy reduces all- 29 per 1000 inhibitor: 131 1000 cause mortality in people with diabetes 30 per 1000 (from 27 fewer and chronic kidney disease. 31 32 to 49 fewer) 33 Very high Placebo: 265 SGLT-2 61 fewer per High SGLT-2 inhibitor therapy reduces all- 34 risk per 1000 inhibitor: 204 1000 cause mortality in people with diabetes 35 per 1000 (from 42 fewer and established cardiovascular disease 36 to 77 fewer) and chronic kidney disease. 37 38 39 40 41 36 42 43 https://mc.manuscriptcentral.com/bmj 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Page 38 of 244

1 2 3 4 Anticipated absolute effects over 5 years 5 6 Anticipated Certainty 7 absolute in 8 Confidential: Foreffects Review (95% treatment Only 9 Relative Risk with Risk with CI) over 5 effects 10 Comparison effect Baseline risk control intervention years (GRADE) Plain text summary 11 GLP-1 OR 0.88 Placebo: 20 GLP-1 2 fewer per Moderate GLP-1 receptor agonist therapy 12 Very low risk receptor (0.82 to 0.94) per 1000 receptor 1000 (from 1 due to probably reduces all-cause mortality in 13 14 agonist v agonist: 18 fewer to 4 indirectness people with diabetes and few or no 15 placebo per 1000 fewer) cardiovascular risk factors. 16 Low risk Placebo: 70 GLP-1 8 fewer per High GLP-1 receptor agonist therapy reduces 17 per 1000 receptor 1000 all-cause mortality in people with 18 agonist: 62 (from 4 fewer diabetes and cardiovascular risk factors. 19 per 1000 to 13 fewer) 20 Moderate risk Placebo: 120 GLP-1 14 fewer per High GLP-1 receptor agonist therapy reduces 21 per 1000 receptor 1000 all-cause mortality in people with 22 agonist: 106 (from 7 fewer diabetes and established cardiovascular 23 per 1000 to 22 fewer) disease. 24 High risk Placebo: 170 GLP-1 20 fewer per High GLP-1 receptor agonist therapy reduces 25 per 1000 receptor 1000 all-cause mortality in people with 26 agonist: 150 (from 10 fewer diabetes and chronic kidney disease. 27 per 1000 to 31 fewer) 28 Placebo: 265 GLP-1 32 fewer per High GLP-1 receptor agonist therapy reduces 29 Very high 30 risk per 1000 receptor 1000 all-cause mortality in people with 31 agonist: 233 (from 16 fewer diabetes and established cardiovascular 32 per 1000 to 48 fewer) disease and chronic kidney disease. 33 34 35 36 37 38 39 40 41 37 42 43 https://mc.manuscriptcentral.com/bmj 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 39 of 244 BMJ

1 2 3 4 Anticipated absolute effects over 5 years 5 6 Anticipated Certainty 7 absolute in 8 Confidential: Foreffects Review (95% treatment Only 9 Relative Risk with Risk with CI) over 5 effects 10 Comparison effect Baseline risk control intervention years (GRADE) Plain text summary 11 SGLT 2 OR 0.88 GLP-1 SGLT-2 2 fewer per Moderate SGLT-2 inhibitor therapy probably 12 Very low risk inhibitor v (0.79 to 0.98) receptor inhibitor: 16 1000 due to reduces all-cause mortality compared 13 14 GLP-1 agonist: 18 per 1000 (from 0 fewer indirectness with GLP-1 receptor agonist therapy in 15 receptor per 1000 to 4 fewer) people with diabetes and few or no 16 agonist cardiovascular risk factors. 17 Low risk GLP-1 SGLT-2 7 fewer per High SGLT-2 inhibitor therapy reduces all- 18 receptor inhibitor: 55 1000 cause mortality compared with GLP-1 19 agonist: 62 per 1000 (from 1 fewer receptor agonist therapy in people with 20 per 1000 to 13 fewer) diabetes and cardiovascular risk factors. 21 Moderate risk GLP-1 SGLT-2 13 fewer per High SGLT-2 inhibitor therapy reduces all- 22 receptor inhibitor: 93 1000 cause mortality compared with GLP-1 23 agonist: 106 per 1000 (from 2 fewer receptor agonist therapy in people with 24 per 1000 to 22 fewer) diabetes and established cardiovascular 25 disease. 26 High risk GLP-1 SGLT-2 18 fewer per High SGLT-2 inhibitor therapy reduces all- 27 receptor inhibitor: 132 1000 cause mortality compared with GLP-1 28 agonist: 150 per 1000 (from 3 fewer receptor agonist therapy in people with 29 per 1000 to 32 fewer) diabetes and chronic kidney disease. 30 31 Very high GLP-1 SGLT-2 28 fewer per High SGLT-2 inhibitor therapy reduces all- 32 risk receptor inhibitor: 205 1000 cause mortality compared with GLP-1 33 agonist: 233 per 1000 (from 5 fewer receptor agonist therapy in people with 34 per 1000 to 49 fewer) diabetes and established cardiovascular 35 disease and chronic kidney disease. 36 CI= confidence interval. We used the point estimate of absolute effect for GLP-1 receptor agonist therapy, obtained from GLP-1 receptor agonist therapy 37 versus placebo, to calculate absolute effect for SGLT-2 inhibitors v GLP-1 receptor agonists. 38 39 40 41 38 42 43 https://mc.manuscriptcentral.com/bmj 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Page 40 of 244

1 2 3 Figures 4 5 Figure 1 PRISMA flow diagram of studies included in the review of glucose lowering therapies 6 for type 2 diabetes 7 8 9 RCT=randomised controlled trial. 10 11 Figure 2 Network plot of trials evaluating glucose lowering therapies in type 2 diabetes 12 Confidential: For Review Only 13 The network shows the number of participants assigned to each glucose-lowering 14 15 16 class with the size of each circle proportional to the number of randomly assigned 17 18 participants in the treatment comparisons (sample size for the specific treatment 19 20 shown in brackets). The width of the lines is proportional to the number of trials 21 22 23 comparing the corresponding pair of treatments. The most frequent drug comparison 24 25 was DPP-4 inhibitors compared to placebo. 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 39

https://mc.manuscriptcentral.com/bmj Page 41 of 244 BMJ

1 2 3 Appendices 4 5 6 Appendix 1 Search strategies 7 8 Appendix 2 Included studies 9 10 Appendix 3 Study risk of bias 11 12 AppendixConfidential: 4 Network plots for each outcome For Review Only 13 14 Appendix 5 Evaluations of network inconsistency and heterogeneity 15 16 17 Appendix 6 Direct, indirect and network treatment estimates 18 19 Appendix 7 Network meta-analysis treatment estimates 20 21 Appendix 8 GRADE summary of findings for SGLT-2 inhibitors and GLP-1 receptor 22 23 agonists compared to placebo or each other 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 40

https://mc.manuscriptcentral.com/bmj BMJ Page 42 of 244

1 2 3 4 5 6 Records identified through 7 Duplicate records excluded Confidential:database searching For Review Only 8 (n=3301) 9 (n=21,549) 10 11 12 13 Identification Records removed during 14 Records after duplicates removed automated RCT classifier 15 (n=18,248) process 16 (n=947) 17 18 19 20 21 Records screened on title Records excluded during 22 and abstract screening

23 Screening (n=17,301) (n=14,939) 24 25 26 27 Full-text articles excluded, with 28 29 reasons 30 Full-text articles assessed for (n=1636) • 31 eligibility Not primary research (n=92) • 32 (n=2362) Incorrect patient population 33 (n=57) 34 • Incorrect study design (n=392) Eligibility 35 • Incorrect intervention or 36 comparator (n=278) 37 • Secondary publication of 38 39 included trial (n=772) 40 • Retracted or terminated (n=5) • Ongoing without reported data 41 Randomised controlled trials (n=40) 42 included in review 43 (n=730) 44 45

46 Included 47 48 49 50 51 52 53 54 https://mc.manuscriptcentral.com/bmj 55 56 57 58 59 60 Page 43 of 244 BMJ

1 2 3 DPP-4 inhibitor 4 5 SGLT-2 inhibitor (76,146) 6 (45,203) 7 8 Thiazolidinedione 9 GLP1-receptor (30,953) 10 agonist 11 (64,096) 12 Confidential: For Review Only 13 Sulfonylurea 14 (29,274) 15 16 Basal insulin 17 (7803) 18 19 Metformin 20 (11,751) 21 22 Basal bolus 23 Insulin 24 (1327) Standard therapy 25 (9375) 26 27 Bolus insulin 28 (1280) 29 30 31 Alpha glucosidase Placebo 32 Inhibitor Glitinide (113,542) 33 (6887) (4393) 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

https://mc.manuscriptcentral.com/bmj BMJ Page 44 of 244

1 2 3 Appendix 1 Search strategies 4 5 Table 1 6 7 MEDLINE 8 9 1. exp Diabetes Mellitus Type 2/ 10 2. Diabetes Mellitus/ 11 12 3.Confidential: ((diabetes or diabetes mellitus For or diabetic*) Review adj1 (type 2 orOnly type II or type ii or 13 non-insulin dependent or noninsulin dependent or adult onset or mature onset or 14 late onset)).tw. 15 16 4. NIDDM.tw. 17 5. Diabetic Nephropathies/ 18 6. (diabetic nephropath* or diabetic kidney disease).tw. 19 7. or/1-6 20 21 8. exp Insulin, Long Acting/ 22 9. ((long acting or longer acting or intermediate acting) adj insulin*).tw. 23 10. (insulin adj1 degludec).tw. 24 25 11. (insulin adj1 glargine).tw. 26 12. (insulin adj1 detemir).tw. 27 13. (insulin adj1 zinc).tw. 28 29 14. (insulin adj1 isophane).tw. 30 15. (insulin adj1 aspart).tw. 31 16. (insulin adj1 lispro).tw. 32 33 17. (insulin adj1 lente).tw. 34 18. (insulin adj1 ultralente).tw. 35 19. Sulfonylurea Compounds/ 36 37 20. / 38 21. / 39 22. / 40 41 23. Gliclazide/ 42 24. Glipizide/ 43 25. Glyburide/ 44 45 26. / 46 27. (sulphonylurea or sulphonylureas).tw. 47 28. acetohexamide.tw. 48 49 29. carbutamide.tw. 50 30. chlorpropamide.tw. 51 31. .tw. 52 53 32. gliclazide.tw. 54 33. glimepiride.tw. 55 34. .tw. 56 57 35. glipizide.tw. 58 36. .tw. 59 37. glyburide.tw. 60

https://mc.manuscriptcentral.com/bmj Page 45 of 244 BMJ

1 2 3 38. glycopyramide.tw. 4 5 39. tolazamide.tw. 6 40. / 7 41. biguanides.tw. 8 9 42. Metformin/ 10 43. metformin.tw. 11 44. alpha-Glucosidases/ai [Antagonists & Inhibitors] 12 Confidential: For Review Only 13 45. alpha glucosidase inhibitor*.tw. 14 46. Acarbose/ 15 47. acarbose.tw. 16 17 48. .tw. 18 49. voglibose.tw. 19 50. / 20 51. thiazolidinedione*.tw. 21 22 52. pioglitazone.tw. 23 53. .tw. 24 54. rosiglitazone.tw. 25 26 55. *.tw. 27 56. .tw. 28 57. .tw. 29 30 58. .tw. 31 59. Receptors, Glucagon/ag 32 60. Glucagon-like Peptide 1/ag 33 34 61. (glucagon-like peptide 1 receptor inhibitor* or glucagon-like peptide 1 receptor 35 agonist* or glucagon-like peptide 1 inhibitor* or glucagon-like peptide 1 agonist* 36 or GLP-1 receptor inhibitor* or GLP-1 receptor agonist* or GLP-1 inhibitor* or 37 38 GLP-1 agonist*).tw. 39 62. dulaglutide.tw. 40 63. ( or exendin 4).tw. 41 42 64. liraglutide.tw. 43 65. .tw. 44 66. .tw. 45 46 67. lixisenatide.tw. 47 68. albiglutide.tw. 48 69. Dipeptidyl-peptidase IV Inhibitors/ 49 50 70. (dipeptidyl-peptidase IV Inhibitor* or dipeptidyl-peptidase 4 Inhibitor* or 51 ((DPP4 or DPP 4 or DPP IV) adj inhibitor*)).tw. 52 71. omarigliptin.tw. 53 54 72. vildagliptin.tw. 55 73. sitagliptin.tw. 56 74. saxagliptin.tw. 57 58 75. linagliptin.tw. 59 76. alogliptin.tw. 60 77. septagliptin.tw.

https://mc.manuscriptcentral.com/bmj BMJ Page 46 of 244

1 2 3 78. .tw. 4 5 79. .tw. 6 80. .tw. 7 81. .tw. 8 9 82. dutogliptin.tw. 10 83. retagliptin.tw. 11 84. Sodium-Glucose Transporter 2/ai 12 Confidential: For Review Only 13 85. (sodium glucose transporter 2 inhibitor* or sodium glucose transporter ii 14 inhibitor* or SGLT 2 inhibitor*).tw. 15 86. (sodium glucose cotransporter adj3 inhibitor*).tw. 16 17 87. (sodium glucose co transporter adj3 inhibitor*).tw. 18 88. canagliflozin.tw. 19 89. dapagliflozin.tw. 20 90. empagliflozin.tw. 21 22 91. ertugliflozin.tw. 23 92. .tw. 24 93. bexagliflozin.tw. 25 26 94. henagliflozin.tw. 27 95. ipragliflozin.tw. 28 96. licogliflozin.tw. 29 30 97. .tw. 31 98. remogliflozin.tw. 32 99. sergliflozin.tw. 33 34 100. .tw. 35 101. ( adj1 (analogue* or derivative*)).tw. 36 102. .tw. 37 38 103. / 39 104. Hypoglycemic Agents/ 40 105. or/8-104 41 42 106. and/7,105 43 107. randomized controlled trial.pt. 44 108. controlled clinical trial.pt. 45 46 109. pragmatic clinical trial.pt. 47 110. randomized.ab. 48 111. randomised.ab 49 50 112. placebo.ab. 51 113. clinical trials as topic/ 52 114. randomly.ab. 53 54 115. trial.ti. 55 116. or/107-115 56 117. animals/ not (humans/ and animals/) 57 58 118. 116 not 117 59 119. and/106,118 60

https://mc.manuscriptcentral.com/bmj Page 47 of 244 BMJ

1 2 3 Table 2 4 5 Embase (1974 to 1 May 2020) 6 7 1. diabetes mellitus/ 8 2. non insulin dependent diabetes mellitus/ 9 3. ((diabetes or diabetes mellitus or diabetic*) adj1 (type 2 or type II or type ii or 10 11 non-insulin dependent or noninsulin dependent or adult onset or mature onset or 12 Confidential:late onset)).tw. For Review Only 13 4. NIDDM.tw. 14 15 5. Diabetic Nephropathy/ 16 6. (diabetic nephropath* or diabetic kidney disease).tw. 17 7. Or/1-6 18 19 8. long acting insulin/ 20 9. ((long acting or longer acting or intermediate acting) adj insulin*).tw. 21 10. / 22 23 11. (insulin adj1 degludec).tw. 24 12. / 25 13. (insulin adj1 detemir).tw. 26 27 14. insulin glargine/ 28 15. (insulin adj1 glargine).tw. 29 16. insulin zinc suspension/ 30 17. (insulin adj1 zinc).tw. 31 32 18. / 33 19. (insulin adj1 aspart).tw. 34 20. / 35 36 21. (insulin adj1 lispro).tw. 37 22. isophane insulin/ 38 23. (insulin adj1 isophane).tw. 39 40 24. (insulin adj1 lente).tw. 41 25. (insulin adj1 ultralente).tw. 42 26. meglitinide/ 43 44 27. meglitinide*.tw. 45 28. mitiglinide/ 46 29. mitiglinide.tw. 47 48 30. nateglinide/ 49 31. nateglinide.tw. 50 32. repaglinide/ 51 52 33. repaglinide.tw. 53 34. amylin derivative/ 54 35. (amylin analogue* or amylin derivative*).tw. 55 56 36. pramlintide/ 57 37. pramlintide.tw. 58 38. derivative/ 59 60 39. biguanide*.tw.

https://mc.manuscriptcentral.com/bmj BMJ Page 48 of 244

1 2 3 40. metformin.tw,sh. 4 5 41. sulphonylurea derivative/ 6 42. sulphonylurea*.tw. 7 43. acetohexamide/ 8 9 44. carbutamide/ 10 45. chlorpropamide/ 11 46. glibenclamide/ 12 Confidential: For Review Only 13 47. glibornuride/ 14 48. gliclazide/ 15 49. glimepiride.sh,tw. 16 17 50. glipizide/ 18 51. gliquidone/ 19 52. tolazamide/ 20 53. acetohexamide.tw. 21 22 54. carbutamide.tw. 23 55. chlorpropamide.tw. 24 56. glibenclamide.tw. 25 26 57. gliclazide.tw. 27 58. glimepiride.tw. 28 59. glibornuride.tw. 29 30 60. glipizide.tw. 31 61. gliquidone.tw. 32 62. glyburide.tw. 33 34 63. glycopyramide.tw. 35 64. tolazamide.tw. 36 65. alpha glucosidase inhibitor/ 37 38 66. alpha glucosidase inhibitor*.tw. 39 67. acarbose/ 40 68. miglitol/ 41 42 69. voglibose/ 43 70. acarbose.tw. 44 71. miglitol.tw. 45 46 72. voglibose.tw. 47 73. glitazone derivative/ 48 74. (glitazone adj1 (derivative* or analogue*)).tw. 49 50 75. thiazolidinedione*.tw. 51 76. pioglitazone/ 52 77. rivoglitazone/ 53 54 78. rosiglitazone/ 55 79. pioglitazone.tw. 56 80. rivoglitazone.tw. 57 58 81. rosiglitazone.tw. 59 82. glucagon like peptide 1 receptor agonist/ 60

https://mc.manuscriptcentral.com/bmj Page 49 of 244 BMJ

1 2 3 83. (glucagon-like peptide 1 receptor inhibitor* or glucagon-like peptide 1 receptor 4 5 agonist* or glucagon-like peptide 1 inhibitor* or glucagon-like peptide 1 agonist* 6 or GLP-1 receptor inhibitor* or GLP-1 receptor agonist* or GLP-1 inhibitor* or 7 GLP-1 agonist*).tw. 8 9 84. albiglutide/ 10 85. dulaglutide/ 11 86. exendin 4/ 12 Confidential: For Review Only 13 87. liraglutide/ 14 88. lixisenatide/ 15 89. semaglutide/ 16 17 90. taspoglutide/ 18 91. albiglutide.tw. 19 92. dulaglutide.tw. 20 93. (exenatide or exendin 4).tw. 21 22 94. liraglutide.tw. 23 95. lixisenatide.tw. 24 96. semaglutide.tw. 25 26 97. taspoglutide.tw. 27 98. dipeptidyl peptidase IV inhibitor/ 28 99. (dipeptidyl-peptidase IV Inhibitor* or dipeptidyl-peptidase 4 Inhibitor* or ((DPP4 29 30 or DPP 4 or DPP IV) adj inhibitor*)).tw. 31 100. alogliptin/ 32 101. anagliptin/ 33 34 102. gemigliptin/ 35 103. linagliptin/ 36 104. omarigliptin/ 37 38 105. saxagliptin/ 39 106. sitagliptin/ 40 107. teneligliptin/ 41 42 108. vildagliptin/ 43 109. alogliptin.tw. 44 110. anagliptin.tw. 45 46 111. gemigliptin.tw. 47 112. linagliptin.tw. 48 113. omarigliptin.tw. 49 50 114. saxagliptin.tw. 51 115. sitagliptin.tw. 52 116. teneligliptin.tw. 53 54 117. vildagliptin.tw. 55 118. evogliptin.tw. 56 119. dutogliptin.tw. 57 58 120. retagliptin.tw. 59 121. sodium glucose cotransporter 2 inhibitor/ 60

https://mc.manuscriptcentral.com/bmj BMJ Page 50 of 244

1 2 3 122. (sodium glucose transporter 2 inhibitor* or sodium glucose transporter ii 4 5 inhibitor* or SGLT 2 inhibitor*).tw. 6 123. (sodium glucose cotransporter adj3 inhibitor*).tw. 7 124. (sodium glucose co transporter adj3 inhibitor*).tw. 8 9 125. canagliflozin/ 10 126. dapagliflozin.tw. 11 127. empagliflozin/ 12 Confidential: For Review Only 13 128. ertugliflozin/ 14 129. tofogliflozin/ 15 130. canagliflozin.tw. 16 17 131. dapagliflozin.tw. 18 132. empagliflozin.tw. 19 133. ertugliflozin.tw. 20 134. tofogliflozin.tw. 21 22 135. bexagliflozin.tw. 23 136. henagliflozin.tw. 24 137. ipragliflozin.tw. 25 26 138. licogliflozin.tw. 27 139. luseogliflozin.tw. 28 140. remogliflozin.tw. 29 30 141. sergliflozin.tw. 31 142. sotagliflozin.tw. 32 143. antidiabetic agent/ 33 34 144. oral antidiabetic agent/ 35 145. or/8-144 36 146. and/7,145 37 38 147. randomized controlled trial/ 39 148. double-blind procedure/ 40 149. single-blind procedure/ 41 42 150. random$.tw. 43 151. factorial$.tw. 44 152. placebo$.tw. 45 46 153. (double$ adj blind$).tw. 47 154. (singl$ adj blind$).tw. 48 155. assign$.tw. 49 50 156. allocat$.tw. 51 157. or/147-156 52 158. and/146,157 53 54 159. MEDLINE.cr. 55 160. 158 not 159 56 161. (mouse or mice or murine or rat or rats or dog or dogs or animal*).ti. 57 58 162. 160 not 161 59 60

https://mc.manuscriptcentral.com/bmj Page 51 of 244 BMJ

1 2 3 Table 3 4 5 Cochrane Central Register of Controlled Trials (CENTRAL) 6 7 #1. MeSH descriptor: [Diabetes Mellitus] this term only 8 #2. ((diabetes or diabetic* or "diabetes mellitus") near/1 ("type 2" or "type 9 ii" or "non-insulin dependent" or "non insulin dependent" or 10 11 "noninsulin dependent" or "adult onset" or "mature onset" or "late 12 Confidential:onset")):ti,ab,kw For Review Only 13 #3. ("diabetes mellitus" not "insulin dependent diabetes mellitus"):kw 14 15 #4. #3 and EMBASE 16 #5. NIDDM:ti,ab,kw 17 #6. ((diabetic next nephropath*) or "diabetic kidney disease"):ti,ab,kw 18 19 #7. or #1-#2, #4-#6 20 #8. ((long acting or longer acting or intermediate acting) near/1 21 insulin*):ti,ab,kw 22 23 #9. insulin near/1 degludec:ti,ab,kw 24 #10. insulin near/1 detemir:ti,ab,kw 25 #11. insulin near/1 glargine:ti,ab,kw 26 27 #12. insulin near/1 zinc 28 #13. insulin near/1 aspart:ti,ab,kw 29 #14. insulin near/1 lispro:ti,ab,kw 30 #15. insulin near/1 isophane:ti,ab,kw 31 32 #16. insulin near/1 ultralente:ti,ab,kw 33 #17. insulin near/1 lente 34 #18. meglitinide*:ti,ab,kw 35 36 #19. mitiglinide:ti,ab,kw 37 #20. nateglinide:ti,ab,kw 38 #21. repaglinide:ti,ab,kw 39 40 #22. (amylin next analogue* or amylin next derivative*):ti,ab,kw 41 #23. pramlintide:ti,ab,kw 42 #24. biguanide*:ti,ab,kw 43 44 #25. metformin:ti,ab,kw 45 #26. (sulfonylurea* or sulphonylurea*):ti,ab,kw 46 #27. acetohexamide:ti,ab,kw 47 48 #28. carbutamide:ti,ab,kw 49 #29. chlopropramide:ti,ab,kw 50 #30. glibenclamide:ti,ab,kw 51 52 #31. gliclazide:ti,ab,kw 53 #32. glimepiride:ti,ab,kw 54 #33. glibornuride:ti,ab,kw 55 56 #34. glipizide:ti,ab,kw 57 #35. gliquidone:ti,ab,kw 58 #36. glyburide:ti,ab,kw 59 60 #37. glycopyramide:ti,ab,kw

https://mc.manuscriptcentral.com/bmj BMJ Page 52 of 244

1 2 3 #38. tolazimide:ti,ab,kw 4 5 #39. MeSH descriptior: [alpha-Glucosidases] this term only and with 6 qualifier(s): [Antagonists & inhibitors] 7 #40. alpha next glucosidase next inhibitor*:ti,ab,kw 8 9 #41. acarbose:ti,ab,kw 10 #42. miglitol:ti,ab,kw 11 #43. voglibose:ti,ab,kw 12 Confidential: For Review Only 13 #44. (glitazone near/1 (derivative* or analogue*)):ti,ab,kw 14 #45. thiazolidinedione:ti,ab,kw 15 #46. pioglitazone:ti,ab,kw 16 17 #47. rivoglitazone:ti,ab,kw 18 #48. rosiglitazone:ti,ab,kw 19 #49. MeSH descriptor: [Receptors, Glucagon] explode all trees and with 20 qualifier(s): [Agonists - AG] 21 22 #50. MeSH descriptor: [Glucagon-Like Peptide 1] explode all trees and with 23 qualifier(s): [Antagonists & inhibitors - AI] 24 #51. ("glucagon-like peptide 1 receptor inhibitor" or "glucagon-like peptide 25 26 1 receptor agonist" or "glucagon-like peptide 1 inhibitor" or "glucagon- 27 like peptide 1 agonist" or "GLP-1 receptor inhibitor" or "GLP-1 28 receptor agonist" or "GLP-1 inhibitor" or "GLP-1 agonist"):ti,ab,kw 29 30 #52. ("glucagon-like peptide 1 receptor inhibitors" or "glucagon-like peptide 31 1 receptor agonists" or "glucagon-like peptide 1 inhibitors" or 32 "glucagon-like peptide 1 agonists" or "GLP-1 receptor inhibitors" or 33 34 "GLP-1 receptor agonists" or "GLP-1 inhibitors" or "GLP-1 35 agonists"):ti,ab,kw 36 #53. albiglutide:ti,ab,kw 37 38 #54. dulaglutide:ti,ab,kw 39 #55 (exenatide or "exendin 4"):ti,ab,kw 40 #56 liraglutide:ti,ab,kw 41 42 #57 lixisenatide:ti,ab,kw 43 #58 semaglutide:ti,ab,kw 44 #59 taspoglutide:ti,ab,kw 45 46 #60 ("dipeptidyl-peptidase IV Inhibitor" or "dipeptidylpeptidase 4 Inhibitor" 47 or "dipeptidyl-peptidase IV Inhibitors" or "dipeptidyl-peptidase 4 48 Inhibitors"):ti,ab,kw 49 50 #61 ((DPP4 or DPP 4 or DPP IV) next inhibitor*):ti,ab,kw 51 #62 alogliptin:ti,ab,kw 52 #63 anagliptin:ti,ab,kw 53 54 #64 gemigliptin:ti,ab,kw 55 #65 linagliptin:ti,ab,kw 56 #66 omarigliptin:ti,ab,kw 57 58 #67 saxagliptin:ti,ab,kw 59 #68 sitagliptin:ti,ab,kw 60

https://mc.manuscriptcentral.com/bmj Page 53 of 244 BMJ

1 2 3 #69 teneligliptin:ti,ab,kw 4 5 #70 vildagliptin:ti,ab,kw 6 #71 evogliptin:ti,ab,kw 7 #72 dutogliptin:ti,ab,kw 8 9 #73 retagliptin:ti,ab,kw 10 #74 MeSH descriptor: [Sodium-Glucose Transporter 2] explode all trees 11 and with qualifier(s): [Antagonists & inhibitors - AI] 12 Confidential: For Review Only 13 #75 ("sodium glucose transporter 2 inhibitor" or "sodium glucose 14 transporter ii inhibitor" or "SGLT 2 inhibitor" or "sodium glucose 15 transporter 2 inhibitors" or "sodium glucose transporter ii inhibitors" or 16 17 "SGLT 2 inhibitors"):ti,ab,kw 18 #76 ("sodium glucose cotransporter" near/3 inhibitor*):ti,ab,kw 19 #77 ("sodium glucose co-transporter" near/3 inhibitor*):ti,ab,kw 20 #78 canagliflozin:ti,ab,kw 21 22 #79 dapagliflozin:ti,ab,kw 23 #80 empagliflozin:ti,ab,kw 24 #81 ertugliflozin:ti,ab,kw 25 26 #82 tofogliflozin:ti,ab,kw 27 #83 bexagliflozin:ti,ab,kw 28 #84 henagliflozin:ti,ab,kw 29 30 #85 ipragliflozin:ti,ab,kw 31 #86 licogliflozin:ti,ab,kw 32 #87 luseogliflozin:ti,ab,kw 33 34 #88 remogliflozin:ti,ab,kw 35 #89 sergliflozin:ti,ab,kw 36 #90 sotagliflozin:ti,ab,kw 37 38 #91 MeSH descriptor: [Insulins] this term only 39 #92 MeSH descriptor: [Hypoglycemic Agents] this term only 40 #93 (antidiabetic next agent*):kw 41 42 #94 (oral next antidiabetic next agent*):kw 43 #95 or #8-#94 44 #96 and #7, #95 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

https://mc.manuscriptcentral.com/bmj BMJ Page 54 of 244

1 2 3 Appendix 2 4 Reference list for included studies1-250251-483484-730 5 6 7 3D 8 Oe H, Nakamura K, Kihara H, et al. Comparison of effects of sitagliptin and voglibose on left 9 ventricular diastolic dysfunction in patients with type 2 diabetes: Results of the 3D trial. Cardiovasc 10 Diabetol 2015;14(1):83 11 12 Confidential: For Review Only 4B Study 13 14 Diamant M, Nauck MA, Shaginian R, et al. Glucagon-like peptide 1 receptor agonist or bolus insulin 15 with optimized basal insulin in type 2 diabetes. Diabetes Care 2014;37(10):2763-73 16 17 4T 18 Holman RR, Thorne KI, Farmer AJ, et al. Addition of biphasic, prandial, or basal insulin to oral therapy 19 in type 2 diabetes. N Engl J Med 2007;357(17):1716-30 20 21 1860-LIRA-DPP-4 22 Pratley RE, Nauck M, Bailey T, et al. Liraglutide versus sitagliptin for patients with type 2 diabetes 23 24 who did not have adequate glycaemic control with metformin: a 26-week, randomised, parallel- 25 group, open-label trial. Lancet 2010;375(9724):1447-56 26 27 Abe 2008 28 Abe M, Okada K, Maruyama T, et al. Clinical effectiveness and safety evaluation of long-term 29 pioglitazone treatment for erythropoietin responsiveness and insulin resistance in type 2 diabetic 30 patients on hemodialysis. Expert Opin Pharmacother 2010;11(10):1611-20 31 32 33 Abe 2016 34 Abe M, Higuchi T, Moriuchi M, et al. Efficacy and safety of saxagliptin, a dipeptidyl peptidase-4 35 inhibitor, in hemodialysis patients with diabetic nephropathy: A randomized open-label prospective 36 trial. Diabetes Res Clin Pract 2016;116:244-52 37 38 ACTION-J 39 Yasunari E, Takeno K, Funayama H, et al. Efficacy of pioglitazone on glycemic control and carotid 40 intima-media thickness in type 2 diabetes patients with inadequate insulin therapy. J Diabetes Invest 41 2011;2(1):56-62 42 43 44 ADOPT 45 Kahn SE, Haffner SM, Heise MA, et al. Glycemic durability of rosiglitazone, metformin, or glyburide 46 monotherapy. N Engl J Med 2006;355(23):2427-43 47 48 Ahmann 2015 49 Ahmann A, Rodbard HW, Rosenstock J, et al. Efficacy and safety of liraglutide versus placebo added 50 to basal insulin analogues (with or without metformin) in patients with type 2 diabetes: A 51 52 randomized, placebo-controlled trial. Diab Obes Metab 2015; 17(11):1056-64 53 54 Ahrén 2004 55 Ahrén B, Gomis R, Standl E, et al. Twelve- and 52-week efficacy of the dipeptidyl peptidase IV 56 inhibitor LAF237 in metformin-treated patients with type 2 diabetes. Diabetes Care 57 2004;27(12):2874-80 58 59 Allegretti 2019 60

https://mc.manuscriptcentral.com/bmj Page 55 of 244 BMJ

1 2 3 Allegretti AS, Zhang W, Zhou W, et al. Safety and effectiveness of bexagliflozin in patients with type 2 4 diabetes mellitus and stage 3a/3b CKD. Am J Kidney Dis 2019;74(3):328-37 5 6 7 Alogliptin Study 007 8 Pratley RE, Kipnes MS, Fleck PR, et al. Efficacy and safety of the dipeptidyl peptidase-4 inhibitor 9 alogliptin in patients with type 2 diabetes inadequately controlled by glyburide monotherapy. 10 Diabetes Obes Metab 2009;11(2):167-76 11 12 AlogliptinConfidential: Study 009 For Review Only 13 Pratley RE, Reusch JE, Fleck PR, et al. Efficacy and safety of the dipeptidyl peptidase-4 inhibitor 14 alogliptin added to pioglitazone in patients with type 2 diabetes: a randomized, double-blind, 15 16 placebo-controlled study. Curr Med Res Opin 2009;25(10):2361-71 17 18 Alogliptin Study 010 19 DeFronzo RA, Fleck PR, Wilson CA, et al. Efficacy and safety of the dipeptidyl peptidase-4 inhibitor 20 alogliptin in patients with type 2 diabetes and inadequate glycemic control: a randomized, double- 21 blind, placebo-controlled study. Diabetes Care 2008;31(12):2315-7 22 23 Alvarsson 2003 24 25 Alvarsson M, Sundkvist G, Lager I, et al. Beneficial effects of insulin versus sulphonylurea on insulin 26 secretion and metabolic control in recently diagnosed type 2 diabetic patients. Diabetes Care 27 2003;26(8):2231-7 28 29 APOLLO 30 Bretzel RG, Nuber U, Landgraf W, et al. Once-daily basal insulin glargine versus thrice-daily prandial 31 insulin lispro in people with type 2 diabetes on oral hypoglycaemic agents (APOLLO): an open 32 randomised controlled trial. Lancet 2008;371(9618):1073-84 33 34 35 Apovian 2010 36 Apovian CM, Bergenstal RM, Cuddihy RM, et al. Effects of exenatide combined with lifestyle 37 modification in patients with type 2 diabetes. Am J Med 2010;123(5):468.e9-17 38 39 APPROACH 40 Gerstein HC, Ratner RE, Cannon CP, et al. Effect of rosiglitazone on progression of coronary 41 atherosclerosis in patients with type 2 diabetes mellitus and coronary artery disease: the assessment 42 on the prevention of progression by rosiglitazone on atherosclerosis in diabetes patients with 43 44 cardiovascular history trial. Circulation 2010;121(10):1176-87 45 46 APRIME 47 Morikawa A, Ishizeki K, Iwashima Y, et al. Pioglitazone reduces urinary albumin excretion in renin- 48 angiotensin system inhibitor-treated type 2 diabetic patients with hypertension and 49 microalbuminuria: the APRIME study. Clin Exp Nephrol 2011;15(6):848-53 50 51 Araki 2015 52 53 Araki E, Tanizawa Y, Tanaka Y, et al. Long-term treatment with empagliflozin as add-on to oral 54 antidiabetes therapy in Japanese patients with type 2 diabetes mellitus. Diabetes Obes Metab 55 2015;17(7):665-74 56 57 Araki 2015a 58 Araki E, Inagaki N, Tanizawa Y, et al. Efficacy and safety of once-weekly dulaglutide in combination 59 with sulphonylurea and/or biguanide compared with once-daily insulin glargine in Japanese patients 60

https://mc.manuscriptcentral.com/bmj BMJ Page 56 of 244

1 2 3 with type 2 diabetes: a randomized, open-label, phase III, non-inferiority study. Diabetes Obes 4 Metab 2015;17(10):994-1002 5 6 7 Arechavaleta 2011 8 Arechavaleta R, Seck T, Chen Y, et al. Efficacy and safety of treatment with sitagliptin or glimepiride 9 in patients with type 2 diabetes inadequately controlled on metformin monotherapy: a randomized, 10 double-blind, non-inferiority trial. Diabetes Obes Metab 2011;13(2):160-8 11 12 Arjona FerreiraConfidential: 2013 For Review Only 13 Arjona Ferreira JC, Corry D, Mogensen CE, et al. Efficacy and safety of sitagliptin in patients with type 14 2 diabetes and ESRD receiving dialysis: a 54-week randomized trial. Am J Kidney Dis 2013;61(4):579- 15 16 87 17 18 Arjona Ferreira 2013a 19 Arjona Ferreira JC, Marre M, Barzilai N, et al. Efficacy and safety of sitagliptin versus glipizide in 20 patients with type 2 diabetes and moderate-to-severe chronic renal insufficiency. Diabetes Care 21 2013;36(5):1067-73 22 23 Arturi 2017 24 25 Arturi F, Succurro E, Miceli S, et al. Liraglutide improves cardiac function in patients with type 2 26 diabetes and chronic heart failure. Endocrine 2017;57(3):464-73 27 28 Asian Acarbose Study 29 Chan JC, Chan KW, Ho LL, et al. An Asian multicenter clinical trial to assess the efficacy and 30 tolerability of acarbose compared with placebo in type 2 diabetic patients previously treated with 31 diet. Asian Acarbose Study Group. Diabetes Care 1998;21(7):1058-61 32 33 Aso 2019 34 35 Aso Y, Jojima T, Iijima T, et al. Effects of dapagliflozin, an SGLT2 inhibitor, on hepatic steatosis and 36 fibrosis evaluated by transient elastography in patients with type 2 diabetes and nonalcoholic fatty 37 liver disease. Diabetes 2019;68(Supplement 1) 38 39 Avilés-Santa 1999 40 Avilés-Santa L, Sinding J, Raskin P. Effects of metformin in patients with poorly controlled, insulin- 41 treated type 2 diabetes mellitus. A randomized, double-blind, placebo-controlled trial. Ann Intern 42 Med 1999;131(3):182-8 43 44 45 AWARD-1 46 Wysham C, Blevins T, Arakaki R, et al. Efficacy and safety of dulaglutide added onto pioglitazone and 47 metformin versus exenatide in type 2 diabetes in a randomized controlled trial (AWARD-1). Diabetes 48 Care 2014;37(8):2159-67 49 50 AWARD-2 51 Giorgino F, Benroubi M, Sun JH, et al. Efficacy and safety of once-weekly dulaglutide versus insulin 52 53 glargine in patients with type 2 diabetes on metformin and glimepiride (AWARD-2). Diabetes Care 54 2015;38(12):2241-9 55 56 AWARD-3 57 Umpierrez G, Povedano ST, Manghi FP, et al. Efficacy and safety of dulaglutide monotherapy versus 58 metformin in type 2 diabetes in a randomized controlled trial (AWARD-3). Diabetes Care 59 2014;37(8):2168-76 60

https://mc.manuscriptcentral.com/bmj Page 57 of 244 BMJ

1 2 3 4 AWARD-4 5 6 Blonde L, Jendle J, Gross J, et al. Once-weekly dulaglutide versus bedtime insulin glargine, both in 7 combination with prandial insulin lispro, in patients with type 2 diabetes (AWARD-4): a randomised, 8 open-label, phase 3, non-inferiority study. Lancet 2015;385(9982):2057-66 9 10 AWARD-5 11 Nauck M, Weinstock RS, Umpierrez GE, et al. Efficacy and safety of dulaglutide versus sitagliptin 12 after 52 weeksConfidential: in type 2 diabetes in a randomized For controlled Review trial (AWARD-5). Only Diabetes Care 13 2014;37(8):2149-58 14 15 16 AWARD-7 17 Tuttle KR, Lakshmanan MC, Rayner B, et al. Dulaglutide versus insulin glargine in patients with type 2 18 diabetes and moderate-to-severe chronic kidney disease (AWARD-7): a multicentre, open-label, 19 randomised trial. Lancet Diabetes Endocrinol 2018;6(8):605 20 21 AWARD-8 22 Dungan KM, Weitgasser R, Perez Manghi F, et al. A 24-week study to evaluate the efficacy and safety 23 of once-weekly dulaglutide added on to glimepiride in type 2 diabetes (AWARD-8). Diabetes Obes 24 25 Metab 2016;18(5):475-82 26 27 AWARD-9 28 Pozzilli P, Norwood P, Jodar E, et al. Placebo-controlled, randomized trial of the addition of once- 29 weekly glucagon-like peptide-1 receptor agonist dulaglutide to titrated daily insulin glargine in 30 patients with type 2 diabetes (AWARD-9). Diabetes Obes Metab 2017;19(7):1024-31 31 32 AWARD-10 33 Ludvik B, Frías JP, Tinahones FJ, et al. Dulaglutide as add-on therapy to SGLT2 inhibitors in patients 34 35 with inadequately controlled type 2 diabetes (AWARD-10): a 24-week, randomised, double-blind, 36 placebo-controlled trial. Lancet Diabetes Endocrinol 2018;6(5):370-81 37 38 AWARD-CHN1 39 Chen YH, Huang CN, Cho YM, et al. Efficacy and safety of dulaglutide monotherapy compared with 40 glimepiride in East-Asian patients with type 2 diabetes in a multicentre, double-blind, randomized, 41 parallel-arm, active comparator, phase III trial. Diabetes Obes Metab 2018;20(9):2121-30 42 43 44 Ba 2017 45 Ba J, Han P, Yuan G, et al. Randomized trial assessing the safety and efficacy of sitagliptin in Chinese 46 patients with type 2 diabetes mellitus inadequately controlled on sulfonylurea alone or combined 47 with metformin. J Diabetes 2017;9:667-76 48 49 Bachmann 2003 50 Bachmann W, Petzinna D, Raptis SA, et al. Long-term improvement of metabolic control by acarbose 51 in type 2 diabetes patients poorly controlled with maximum sulfonylurea therapy. Clin Drug Invest 52 53 2003;23(10):679-86 54 55 Bailey 2010 56 Bailey CJ, Gross JL, Pieters A, et al. Effect of dapagliflozin in patients with type 2 diabetes who have 57 inadequate glycaemic control with metformin: a randomised, double-blind, placebo-controlled trial. 58 Lancet 2010;375(9733):2223-33 59 60

https://mc.manuscriptcentral.com/bmj BMJ Page 58 of 244

1 2 3 Bailey 2012 4 Bailey CJ, Iqbal N, T'Joen C, et al. Dapagliflozin monotherapy in drug-naïve patients with diabetes: a 5 6 randomized-controlled trial of low-dose range. Diabetes Obes Metab 2012;14(10):951-9 7 8 Bajaj 2014 9 Bajaj M, Gilman R, Patel S, et al. Linagliptin improved glycaemic control without weight gain or 10 hypoglycaemia in patients with Type 2 diabetes inadequately controlled by a combination of 11 metformin and pioglitazone: A 24-week randomized, double-blind study. Diabet Med 12 2014;31(12):1505-14Confidential: For Review Only 13 14 Bakris 2006 15 16 Bakris GL, Ruilope LM, McMorn SO, et al. Rosiglitazone reduces microalbuminuria and blood 17 pressure independently of glycemia in type 2 diabetes patients with microalbuminuria. J Hypertens 18 2006;24(10):2047-55 19 20 BALANCE 21 Bae JC, Min KW, Kim YH, et al. Efficacy and safety of fixed-dose combination therapy with gemigliptin 22 (50 mg) and rosuvastatin compared with monotherapy in patients with type 2 diabetes and 23 dyslipidaemia (BALANCE): a multicentre, randomized, double-blind, controlled, phase 3 trial. 24 25 Diabetes Obes Metab 2019;21(1):103-11 26 27 Banerji 1995 28 Banerji MA, Chaiken RL, Lebovitz HE. Prolongation of near-normoglycemic remission in black NIDDM 29 subjects with chronic low-dose sulfonylurea treatment. Diabetes 1995;44(4):466-70 30 31 Barnett 2003 32 Barnett AH, Grant PJ, Hitman GA, et al. Rosiglitazone in Type 2 diabetes mellitus: an evaluation in 33 British Indo-Asian patients. Diabetic Med 2003;20(5):387-93 34 35 36 Barnett 2012 37 Barnett AH, Charbonnel B, Donovan M, et al. Effect of saxagliptin as add-on therapy in patients with 38 poorly controlled type 2 diabetes on insulin alone or insulin combined with metformin. Curr Med Res 39 Opin 2012;28(4):513-23 40 41 Barnett 2013 42 Barnett AH, Huisman H, Jones R, et al. Linagliptin for patients aged 70 years or older with type 2 43 44 diabetes inadequately controlled with common antidiabetes treatments: a randomised, double- 45 blind, placebo-controlled trial. Lancet 2013;382(9902):1413-23 46 47 Barzilai 2011 48 Barzilai N, Guo H, Mahoney EM, et al. Efficacy and tolerability of sitagliptin monotherapy in elderly 49 patients with type 2 diabetes: a randomized, double-blind, placebo-controlled trial. Curr Med Res 50 Opin 2011;27(5):1049-58 51 52 53 BEGIN: ADD TO GLP-1 54 Aroda VR, Bailey TS, Cariou B, et al. Effect of adding insulin degludec to treatment in patients with 55 type 2 diabetes inadequately controlled with metformin and liraglutide: a double-blind randomized 56 controlled trial (BEGIN: ADD TO GLP-1 Study). Diab Obes Metab 2016;18(7):663-70 57 58 BEGIN: VICTOZA ADD-ON 59 60

https://mc.manuscriptcentral.com/bmj Page 59 of 244 BMJ

1 2 3 Mathieu C, Rodbard HW, Cariou B, et al. A comparison of adding liraglutide versus a single daily dose 4 of insulin aspart to insulin degludec in subjects with type 2 diabetes (BEGIN: VICTOZA ADD-ON). 5 6 Diabetes Obes Metab 2014;16(7):636-44 7 8 Berberoglu 2010 9 Berberoglu Z, Yazici AC, Demirag NG. Effects of rosiglitazone on bone mineral density and 10 remodelling parameters in postmenopausal diabetic women: a 2-year follow-up study. Clin 11 Endocrinol 2010;73(3):305-12 12 Confidential: For Review Only 13 Bergenstal 2009 14 Bergenstal R, Lewin A, Bailey T, et al. Efficacy and safety of biphasic insulin aspart 70/30 versus 15 16 exenatide in subjects with type 2 diabetes failing to achieve glycemic control with metformin and a 17 sulfonylurea. Curr Med Res Opin 2009;25(1):65-75 18 19 Berndt-Zipfel 2013 20 Berndt-Zipfel C, Michelson G, Dworak M, et al. Vildagliptin in addition to metformin improves retinal 21 blood flow and erythrocyte deformability in patients with type 2 diabetes mellitus - results from an 22 exploratory study. Cardiovasc Diabetol 2013;12(1):59 23 24 25 BEST 26 Retnakaran R, Qi Y, Opsteen C, et al. Initial short-term intensive insulin therapy as a strategy for 27 evaluating the preservation of beta-cell function with oral antidiabetic : a pilot study 28 with sitagliptin. Diabetes Obes Metab 2010;12(10):909-15 29 30 BETA 31 Moon JS, Ha KS, Yoon JS, et al. The effect of glargine versus glimepiride on pancreatic beta-cell 32 function in patients with type 2 diabetes uncontrolled on metformin monotherapy: Open-label, 33 randomized, controlled study. Acta Diabetol 2014;51(2):277-85 34 35 36 Bi 2013 37 Bi Y, Tong GY, Yang HJ, et al. The beneficial effect of metformin on beta-cell function in non-obese 38 Chinese subjects with newly diagnosed type 2 diabetes. Diabetes Metab Res Rev 2013;29(8):664-72 39 40 Bilezikian 2013 41 Rubin MR, Manavalan JS, Agarwal S, et al. Effects of rosiglitazone vs metformin on circulating 42 osteoclast and osteogenic precursor cells in postmenopausal women with type 2 diabetes mellitus. J 43 44 Clin Endocrinol Metab 2014;99(10):E1933-42 45 46 Birkeland 1994a 47 Birkeland KI, Hanssen KF, Urdal P, et al. A long-term, randomized, comparative study of insulin 48 versus sulfonylurea therapy in type 2 diabetes. J Intern Med 1994;236(3):305-13 49 50 Bode 2013 51 Bode B, Stenlöf K, Sullivan D, et al. Efficacy and safety of canagliflozin treatment in older subjects 52 53 with type 2 diabetes mellitus: a randomized trial. Hosp Pract (1995) 2013;41(2):72-84 54 55 Bolinder 2012 56 Bolinder J, Ljunggren Ö, Kullberg J, et al. Effects of dapagliflozin on body weight, total fat mass, and 57 regional adipose tissue distribution in patients with type 2 diabetes mellitus with inadequate 58 glycemic control on metformin. J Clin Endocrinol Metab 2012;97(3):1020-31 59 60

https://mc.manuscriptcentral.com/bmj BMJ Page 60 of 244

1 2 3 Bolli 2009 4 Bolli G, Dotta F, Colin L, et al. Comparison of vildagliptin and pioglitazone in patients with type 2 5 6 diabetes inadequately controlled with metformin. Diabetes Obes Metab 2009;11(6):589-95 7 8 Borges 2011 9 Borges JL, Bilezikian JP, Jones-Leone AR, et al. A randomized, parallel group, double-blind, 10 multicentre study comparing the efficacy and safety of Avandamet (rosiglitazone/metformin) and 11 metformin on long-term glycaemic control and bone mineral density after 80 weeks of treatment in 12 drug-naïveConfidential: type 2 diabetes mellitus patients. For Diabetes Review Obes Metab 2011; 13Only(11):1036-46 13 14 Bosi 2007 15 16 Bosi E, Camisasca RP, Collober C, et al. Effects of vildagliptin on glucose control over 24 weeks in 17 patients with type 2 diabetes inadequately controlled with metformin. Diabetes Care 18 2007;30(4):890-5 19 20 Bosi 2009 21 Bosi E, Dotta F, Jia Y, et al. Vildagliptin plus metformin combination therapy provides superior 22 glycaemic control to individual monotherapy in treatment-naive patients with type 2 diabetes 23 mellitus. Diabetes Obes Metab 2009;11(5):506-15 24 25 26 Bosi 2011 27 Bosi E, Ellis GC, Wilson CA, et al. Alogliptin as a third oral antidiabetic drug in patients with type 2 28 diabetes and inadequate glycaemic control on metformin and pioglitazone: a 52-week, randomized, 29 double-blind, active-controlled, parallel-group study. Diabetes Obes Metab 2011;13(12):1088-96 30 31 Bouchi 2017 32 Bouchi R, Nakano Y, Fukuda T, et al. Reduction of visceral fat by liraglutide is associated with 33 ameliorations of hepatic steatosis, albuminuria, and micro-inflammation in type 2 diabetic patients 34 35 with insulin treatment: a randomized control trial. Endocrine J 2017;64(3):269-81 36 37 Braun 1996 38 Braun D, Schonherr U, Mitzkat HJ. Efficacy of acarbose monotherapy in patients with type 2 39 diabetes: A double-blind study conducted in general practice. Endocrinol Metab 1996;3(4):275-80 40 41 BRL-049653/334 42 Hedblad B, Zambanini A, Nilsson P, et al. Rosiglitazone and carotid IMT progression rate in a mixed 43 44 cohort of patients with type 2 diabetes and the insulin resistance syndrome: main results from the 45 Rosiglitazone Atherosclerosis Study. J Intern Med 2007;261:293-305 46 47 Bryson 2016 48 Bryson A, Jennings PE, Deak L, et al. The efficacy and safety of teneligliptin added to ongoing 49 metformin monotherapy in patients with type 2 diabetes: a randomized study with open label 50 extension. Expert Opin Pharmacother 2016;17(10):1309-16 51 52 53 Bunck 2009 54 Bunck MC, Diamant M, Cornér A, et al. One-year treatment with exenatide improves beta-cell 55 function, compared with insulin glargine, in metformin-treated type 2 diabetic patients: a 56 randomized, controlled trial. Diabetes Care 2009;32(5):762-8 57 58 Buse 2011 59 60

https://mc.manuscriptcentral.com/bmj Page 61 of 244 BMJ

1 2 3 Buse JB, Bergenstal RM, Glass LC, et al. Use of twice-daily exenatide in Basal insulin-treated patients 4 with type 2 diabetes: a randomized, controlled trial. Ann Intern Med 2011;154(2):103-12 5 6 7 Camerini-Davalos 1988 8 Camerini-Davalos RA, Velasco CA, Reddi AS, et al. Delay of progression of diabetic microangiopathy. 9 Metabolism 1988;37(2 Suppl 1):10-8 10 11 Camerini-Davalos 1994 12 Camerini-DavalosConfidential: RA, Velasco CA, Reddi AS. ForEffect of insulin-glipizideReview combination Only on skeletal muscle 13 capillary basement membrane width in diabetic patients. Clin Ther 1994;16(6):952-61 14 15 16 Campbell 1994 17 Campbell IW, Menzies DG, Chalmers J, et al. One year comparative trial of metformin and glipizide in 18 type 2 diabetes mellitus. Diabete Metab 1994;20(4):394-400 19 20 CANDLE 21 Tanaka A, Hisauchi I, Taguchi I, et al. Effects of canagliflozin in patients with type 2 diabetes and 22 chronic heart failure: a randomized trial (CANDLE). ESC Heart Failure 2020 23 24 25 CANTATA-D 26 Lavalle-Gonzalez FJ, Januszewicz A, Davidson J, et al. Efficacy and safety of canagliflozin compared 27 with placebo and sitagliptin in patients with type 2 diabetes on background metformin 28 monotherapy: A randomised trial. Diabetologia 2013;56(12):2582-92 29 30 CANTATA-D2 31 Schernthaner G, Gross JL, Rosenstock J, et al. Canagliflozin compared with sitagliptin for patients 32 with type 2 diabetes who do not have adequate glycemic control with metformin plus sulfonylurea: 33 A 52-week randomized trial. Diabetes Care 2013;36(9):2508-15 34 35 36 CANTATA-M 37 Stenlöf K, Cefalu WT, Kim KA, et al. Efficacy and safety of canagliflozin monotherapy in subjects with 38 type 2 diabetes mellitus inadequately controlled with diet and exercise. Diabetes Obes Metab 39 2013;15(4):372-82 40 41 CANTATA-MSU 42 Wilding JPH, Charpentier G, Hollander P, et al. Efficacy and safety of canagliflozin in patients with 43 44 type 2 diabetes mellitus inadequately controlled with metformin and sulphonylurea: A randomised 45 trial. Int J Clin Pract 2013;67(12):1267-82 46 47 CANTATA-SU 48 Cefalu WT, Leiter LA, Yoon KH, et al. Efficacy and safety of canagliflozin versus glimepiride in patients 49 with type 2 diabetes inadequately controlled with metformin (CANTATA-SU): 52 week results from a 50 randomised, double-blind, phase 3 non-inferiority trial. Lancet 2013;382(9896):941-50 51 52 53 CANVAS 54 Neal B, Perkovic V, Mahaffey KW, et al. Canagliflozin and cardiovascular and renal events in type 2 55 diabetes. N Engl J Med 2017;377(7):644-57 56 57 CARMELINA 58 59 60

https://mc.manuscriptcentral.com/bmj BMJ Page 62 of 244

1 2 3 Rosenstock J, Perkovic V, Johansen OE, et al. Effect of linagliptin vs placebo on major cardiovascular 4 events in adults with type 2 diabetes and high cardiovascular and renal risk: the CARMELINA 5 6 randomized clinical trial. JAMA 2019;321(1):69-79 7 8 CAROLINA 9 Rosenstock J, Kahn Steven E, Johansen Odd E, et al. Effect of linagliptin vs glimepiride on major 10 adverse cardiovascular outcomes in patients with type 2 diabetes: the CAROLINA randomized clinical 11 trial. JAMA 2019;322(12):1155-66 12 Confidential: For Review Only 13 Casner 1988 14 Casner PR. Insulin-glyburide combination therapy for non-insulin-dependent diabetes mellitus: a 15 16 long-term double-blind, placebo-controlled trial. Clin Pharmacol Ther 1988;44(5):594-603 17 18 Cefalu 2015 19 Cefalu WT, Leiter LA, de Bruin TW, et al. Dapagliflozin's effects on glycemia and cardiovascular risk 20 factors in high-risk patients with type 2 diabetes: A 24-week, multicenter, randomized, double-blind, 21 placebo-controlled study with a 28-week extension. Diabetes Care 2015;38(7):1218-27 22 23 Chacra 2017 24 25 Chacra A, Gantz I, Mendizabal G, et al. A randomised, double-blind, trial of the safety and efficacy of 26 omarigliptin (a once-weekly DPP-4 inhibitor) in subjects with type 2 diabetes and renal impairment. 27 Int J Clin Pract 2017;71(6):e12955 28 29 Chakraborty 2011 30 Chakraborty A, Chowdhury S, Bhattacharyya M. Effect of metformin on oxidative stress, nitrosative 31 stress and inflammatory biomarkers in type 2 diabetes patients. Diabetes Res Clin Pract 32 2011;93(1):56-62 33 34 35 Charbonnel 2013 36 Charbonnel B, Steinberg H, Eymard E, et al. Efficacy and safety over 26 weeks of an oral treatment 37 strategy including sitagliptin compared with an injectable treatment strategy with liraglutide in 38 patients with type 2 diabetes mellitus inadequately controlled on metformin: a randomised clinical 39 trial. Diabetologia 2013;56(7):1503-11 40 41 Charpentier 2009 42 Charpentier G, Halimi S. Earlier triple therapy with pioglitazone in patients with type 2 diabetes. 43 44 Diabetes Obes Metab 2009;11(9):844-54 45 46 Chavez 2015 47 Chavez AO, Guardado-Mendoza R, Varvel S, et al. Differential effect of pioglitazone, exenatide, and 48 combination of pioglitazone and exenatide on plasma metabolites in type 2 diabetes. Diabetes 2015; 49 64:A529 50 51 Chen 2015 52 53 Chen Y, Ning G, Wang C, et al. Efficacy and safety of linagliptin monotherapy in Asian patients with 54 inadequately controlled type 2 diabetes mellitus: A multinational, 24-week, randomized, clinical trial. 55 J Diabetes Invest 2015;6(6):692-8 56 57 Chen 2016 58 Chen YH, Tarng DC, Chen HS. Renal outcomes of pioglitazone compared with acarbose in diabetic 59 patients: A randomized controlled study. PLoS ONE 2016;11(11):e0165750 60

https://mc.manuscriptcentral.com/bmj Page 63 of 244 BMJ

1 2 3 4 Chiasson 1994 5 6 Chiasson JL, Josse RG, Hunt JA, et al. The efficacy of acarbose in the treatment of patients with non- 7 insulin-dependent diabetes mellitus. A multicenter controlled clinical trial. Ann Intern Med 8 1994;121(12):928-35 9 10 Chiasson 2001 11 Chiasson JL, Naditch L. The synergistic effect of miglitol plus metformin combination therapy in the 12 treatmentConfidential: of type 2 diabetes. Diabetes Care For 2001;24 (6):989-94Review Only 13 14 CHICAGO 15 16 Mazzone T, Meyer PM, Feinstein SB, et al. Effect of pioglitazone compared with glimepiride on 17 carotid intima-media thickness in type 2 diabetes: a randomized trial. JAMA 2006;296(21):2572-81 18 19 Chien 2011 20 Chien MN, Lee CC, Chen WC, et al. Effect of sitagliptin as add-on therapy in elderly type 2 diabetes 21 patients with inadequate glycemic control in Taiwan. Int J Gerontol 2011;5(2):103-6 22 23 Cho 2019 24 25 Cho K, Omori K, Takase T, et al. Effect of switching from pioglitazone to the SGLT2 inhibitor 26 dapagliflozin on body weight and metabolism-related factors in patients with type 2 diabetes 27 mellitus. Diabetes Obes Metab 2019;21:710-14 28 29 Choi 2004 30 Choi D, Kim SK, Choi SH, et al. Preventative effects of rosiglitazone on restenosis after coronary stent 31 implantation in patients with type 2 diabetes. Diabetes Care 2004;27(11):2654-60 32 33 Chou 2008 34 35 Chou HS, Palmer JP, Jones AR, et al. Initial treatment with fixed-dose combination 36 rosiglitazone/glimepiride in patients with previously untreated type 2 diabetes. Diabetes Obes 37 Metab 2008;10(8):626-37 38 39 Chou 2012 40 Chou HS, Truitt KE, Moberly JB, et al. A 26-week, placebo- and pioglitazone-controlled monotherapy 41 study of rivoglitazone in subjects with type 2 diabetes mellitus. Diabetes Obes Metab 42 2012;14(11):1000-9 43 44 45 CIMT 46 Lundby-Christensen L, Tarnow L, Boesgaard TW, et al. Metformin versus placebo in combination 47 with insulin analogues in patients with type 2 diabetes mellitus-the randomised, blinded 48 Copenhagen Insulin and Metformin Therapy (CIMT) trial. BMJ Open 2016;6(2):e008376 49 50 Civera 2008 51 Civera M, Merchante A, Salvador M, et al. Safety and efficacy of repaglinide in combination with 52 53 metformin and bedtime NPH insulin as an insulin treatment regimen in type 2 diabetes. Diabetes Res 54 Clin Pract 2008;79(1):42-7 55 56 COMPASS 57 Takihata M, Nakamura A, Tajima K, et al. Comparative study of sitagliptin with pioglitazone in 58 Japanese type 2 diabetic patients: the COMPASS randomized controlled trial. Diabetes Obes Metab 59 2013;15(5):455-62 60

https://mc.manuscriptcentral.com/bmj BMJ Page 64 of 244

1 2 3 4 COMPOSIT-I 5 6 Roussel R, Duran-Garcia S, Zhang Y, et al. Double-blind, randomized clinical trial comparing the 7 efficacy and safety of continuing or discontinuing the dipeptidyl peptidase-4 inhibitor sitagliptin 8 when initiating insulin glargine therapy in patients with type 2 diabetes: the CompoSIT-I Study. Diab 9 Obes Metab 2019;21(4):781-90 10 11 COMPOSIT-R 12 Scott R, MorganConfidential: J, Zimmer Z, et al. A randomized For clinical Review trial of the efficacy Only and safety of sitagliptin 13 compared with dapagliflozin in patients with type 2 diabetes mellitus and mild renal insufficiency: 14 the CompoSIT-R study. Diab Obes Metab 2018;20(12):2876-84 15 16 17 CONFIDENCE 18 Xu W, Bi Y, Sun Z, et al. Comparison of the effects on glycaemic control and beta-cell function in 19 newly diagnosed type 2 diabetes patients of treatment with exenatide, insulin or pioglitazone: a 20 multicentre randomized parallel-group trial (the CONFIDENCE study). J Intern Med 2015;277(1):137- 21 50 22 23 Coniff 1994 24 25 Coniff RF, Shapiro JA, Seaton TB. Long-term efficacy and safety of acarbose in the treatment of obese 26 subjects with non-insulin-dependent diabetes mellitus. Arch Intern Med 1994;154(21):2442-8 27 28 Coniff 1995 29 Coniff RF, Shapiro JA, Seaton TB, et al. Multicenter, placebo-controlled trial comparing acarbose (BAY 30 g 5421) with placebo, , and tolbutamide-plus-acarbose in non-insulin-dependent 31 diabetes mellitus. Am J Med 1995;98(5):443-51 32 33 Coniff 1995a 34 35 Coniff RF, Shapiro JA, Seaton TB, et al. A double-blind placebo-controlled trial evaluating the safety 36 and efficacy of acarbose for the treatment of patients with insulin-requiring type II diabetes. 37 Diabetes Care 1995;18(7):928-32 38 39 Costa 1997 40 Costa B, Piñol C. Acarbose in ambulatory treatment of non-insulin-dependent diabetes mellitus 41 associated to imminent sulfonylurea failure: a randomised-multicentric trial in primary health-care. 42 Diabetes and Acarbose Research Group. Diabetes Res Clin Pract 1997;38(1):33-40 43 44 45 CREDENCE 46 Perkovic V, Jardine MJ, Neal B, et al. Canagliflozin and renal outcomes in type 2 diabetes and 47 nephropathy. N Engl J Med 2019;380(24):2295-306 48 49 Cusi 2019 50 Cusi K, Bril F, Barb D, et al. Effect of canagliflozin treatment on hepatic triglyceride content and 51 glucose metabolism in patients with type 2 diabetes. Diabetes Obes Metab 2019;21(4):812-21 52 53 54 CV181-011 55 Rosenstock J, Aguilar-Salinas C, Klein E, et al. Effect of saxagliptin monotherapy in treatment-naïve 56 patients with type 2 diabetes. Curr Med Res Opin 2009;25(10):2401-11 57 58 CV181-013 59 60

https://mc.manuscriptcentral.com/bmj Page 65 of 244 BMJ

1 2 3 Hollander P, Li J, Allen E, et al. Saxagliptin added to a thiazolidinedione improves glycemic control in 4 patients with type 2 diabetes and inadequate control on thiazolidinedione alone. J Clin Endocrinol 5 6 Metab 2009;94(12):4810-9 7 8 CV181-039 9 Pfützner A, Paz-Pacheco E, Allen E, et al. Initial combination therapy with saxagliptin and metformin 10 provides sustained glycaemic control and is well tolerated for up to 76 weeks. Diabetes Obes Metab 11 2011;13(6):567-76 12 Confidential: For Review Only 13 da Silva 2016 14 da Silva G, Nogueira K, Fukui R, et al. Short and long term effects of a DPP-4 inhibitor versus bedtime 15 16 NPH insulin as ADD-ON therapy in patients with type 2 diabetes. Curr Pharm Design 2016; 17 22(44):6716-21 18 19 Dailey 2004 20 Dailey GE, Noor MA, Park JS, et al. Glycemic control with glyburide/metformin tablets in combination 21 with rosiglitazone in patients with type 2 diabetes: a randomized, double-blind trial. Am J Med 22 2004;116(4):223-9 23 24 25 Dapagliflozin 006 26 Wilding JP, Woo V, Soler NG, et al. Long-term efficacy of dapagliflozin in patients with type 2 27 diabetes mellitus receiving high doses of insulin: a randomized trial. Ann Intern Med 28 2012;156(6):405-15 29 30 Dargie 2007 31 Dargie HJ, Hildebrandt PR, Riegger GA, et al. A randomized, placebo-controlled trial assessing the 32 effects of rosiglitazone on echocardiographic function and cardiac status in type 2 diabetic patients 33 with New York Heart Association Functional Class I or II Heart Failure. J Am Coll Cardiol 34 35 2007;49(16):1696-704 36 37 Davidson 2007 38 Davidson JA, McMorn SO, Waterhouse BR, et al. A 24-week, multicenter, randomized, double-blind, 39 placebo-controlled, parallel-group study of the efficacy and tolerability of combination therapy with 40 rosiglitazone and sulfonylurea in African American and Hispanic American patients with type 2 41 diabetes inadequately controlled with sulfonylurea monotherapy. Clin Ther 2007;29(9):1900-14 42 43 44 Davies 2013 45 Davies M, Heller S, Sreenan S, et al. Once-weekly exenatide versus once- or twice-daily insulin 46 detemir: randomized, open-label, clinical trial of efficacy and safety in patients with type 2 diabetes 47 treated with metformin alone or in combination with . Diabetes Care 2013;36(5):1368- 48 76 49 50 Davies 2017 51 Davies M, Pieber TR, Hartoft-Nielsen ML, et al. Effect of Oral Semaglutide Compared With Placebo 52 53 and Subcutaneous Semaglutide on Glycemic Control in Patients With Type 2 Diabetes: a Randomized 54 Clinical Trial. JAMA 2017;318(15):1460-70 55 56 DECLARE-TIMI 58 57 Wiviott SD, Raz I, Bonaca MP, et al. Dapagliflozin and cardiovascular outcomes in type 2 diabetes. N 58 Engl J Med 2019;380(4):347-57 59 60

https://mc.manuscriptcentral.com/bmj BMJ Page 66 of 244

1 2 3 DeFronzo 1995 4 DeFronzo RA, Goodman AM. Efficacy of metformin in patients with non-insulin-dependent diabetes 5 6 mellitus. The Multicenter Metformin Study Group. N Engl J Med 1995;333(9):541-9 7 8 DeFronzo 2005 9 DeFronzo RA, Ratner RE, Han J, et al. Effects of exenatide (exendin-4) on glycemic control and weight 10 over 30 weeks in metformin-treated patients with type 2. Diabetes Care 2005;28(5):1092-100 11 12 DeFronzoConfidential: 2012 For Review Only 13 DeFronzo RA, Burant CF, Fleck P, et al. Efficacy and tolerability of the DPP-4 inhibitor alogliptin 14 combined with pioglitazone, in metformin-treated patients with type 2 diabetes. J Clin Endocrinol 15 16 Metab 2012;97(5):1615-22 17 18 DeFronzo 2015 19 DeFronzo RA, Lewin A, Patel S, et al. Combination of empagliflozin and linagliptin as second-line 20 therapy in subjects with type 2 diabetes inadequately controlled on metformin. Diabetes Care 21 2015;38(3):384-93 22 23 Dei Cas 2017 24 25 Dei Cas A, Spigoni V, Cito M, et al. Vildagliptin, but not glibenclamide, increases circulating 26 endothelial progenitor cell number: A 12-month randomized controlled trial in patients with type 2 27 diabetes. Cardiovasc Diabetol 2017;16(1):27 28 29 Dejager 2007 30 Dejager S, Razac S, Foley JE, et al. Vildagliptin in drug-naïve patients with type 2 diabetes: a 24-week, 31 double-blind, randomized, placebo-controlled, multiple-dose study. Horm Metab Res 32 2007;39(3):218-23 33 34 35 del Prato 2003 36 del Prato S, Erkelens DW, Leutenegger M. Six-month efficacy of vs. placebo or metformin 37 in diet-failed type 2 diabetic patients. Acta Diabetol 2003;40(1):20-7 38 39 Del Prato 2011 40 Del Prato S, Barnett AH, Huisman H, et al. Effect of linagliptin monotherapy on glycaemic control and 41 markers of beta-cell function in patients with inadequately controlled type 2 diabetes: A randomized 42 controlled trial. Diabetes Obes Metab 2011;13(3):258-67 43 44 45 Del Prato 2014 46 Del Prato S, Camisasca R, Wilson C, et al. Durability of the efficacy and safety of alogliptin compared 47 with glipizide in type 2 diabetes mellitus: a 2-year study. Diabetes Obes Metab 2014;16(12):1239-46 48 49 DELIGHT 50 Pollock C, Stefansson B, Reyner D, et al. Albuminuria-lowering effect of dapagliflozin alone and in 51 combination with saxagliptin and effect of dapagliflozin and saxagliptin on glycaemic control in 52 53 patients with type 2 diabetes and chronic kidney disease (DELIGHT): a randomised, double-blind, 54 placebo-controlled trial. Lancet Diabetes Endocrinol 2019;7(6):429-41 55 56 Deng 2017 57 Deng XL, Ma R, Zhu HX, et al. Short article: a randomized-controlled study of sitagliptin for treating 58 diabetes mellitus complicated by nonalcoholic fatty liver disease. Eur J Gastroenterol Hepatol 59 2017;29(3):297-301 60

https://mc.manuscriptcentral.com/bmj Page 67 of 244 BMJ

1 2 3 4 DERIVE 5 6 Fioretto P, Del Prato S, Buse JB, et al. Efficacy and safety of dapagliflozin in patients with type 2 7 diabetes and moderate renal impairment (chronic kidney disease stage 3A): the DERIVE Study. Diab 8 Obes Metab 2018;20(11):2532-40 9 10 Derosa 2003 11 Derosa G, Mugellini A, Ciccarelli L, et al. Comparison between repaglinide and glimepiride in patients 12 with typeConfidential: 2 diabetes mellitus: a one-year, randomized, For Review double-blind assessment Only of metabolic 13 parameters and cardiovascular risk factors. Clin Ther 2003;25(2):472-84 14 15 16 Derosa 2003a 17 Derosa G, Mugellini A, Ciccarelli L, et al. Comparison of glycaemic control and cardiovascular risk 18 profile in patients with type 2 diabetes during treatment with either repaglinide or metformin. 19 Diabetes Res Clin Pract 2003;60(3):161-9 20 21 Derosa 2004 22 Derosa G, Franzetti I, Gadaleta G, et al. Metabolic variations with oral antidiabetic drugs in patients 23 with Type 2 diabetes: comparison between glimepiride and metformin. Diabetes Nutr Metab 24 25 2004;17(3):143-50 26 27 Derosa 2005 28 Derosa G, Cicero AF, Gaddi AV, et al. Long-term effects of glimepiride or rosiglitazone in combination 29 with metformin on blood pressure control in type 2 diabetic patients affected by the metabolic 30 syndrome: a 12-month, double-blind, randomized clinical trial. Clin Ther 2005;27(9):1383-91 31 32 Derosa 2007 33 Derosa G, D'Angelo A, Fogari E, et al. Effects of nateglinide and glibenclamide on prothrombotic 34 35 factors in naive type 2 diabetic patients treated with metformin: a 1-year, double-blind, randomized 36 clinical trial. Intern Med 2007;46(22):1837-46 37 38 Derosa 2009 39 Derosa G, D'Angelo A, Salvadeo SA, et al. Modulation of adipokines and vascular remodelling 40 markers during OGTT with acarbose or pioglitazone treatment. Biomed Pharmacother 41 2009;63(10):723-33 42 43 44 Derosa 2009a 45 Derosa G, Maffioli P, Salvadeo SA, et al. Direct comparison among oral hypoglycemic agents and 46 their association with insulin resistance evaluated by euglycemic hyperinsulinemic clamp: the 60's 47 study. Metabolism 2009;58(8):1059-66 48 49 Derosa 2010 50 Derosa G, Maffioli P, Salvadeo SA, et al. Effects of sitagliptin or metformin added to pioglitazone 51 monotherapy in poorly controlled type 2 diabetes mellitus patients. Metabolism 2010;59(6):887-95 52 53 54 Derosa 2010a 55 Derosa G, Maffioli P, Salvadeo SA, et al. Exenatide versus glibenclamide in patients with diabetes. 56 Diabetes Technol Ther 2010;12(3):233-40 57 58 Derosa 2011 59 60

https://mc.manuscriptcentral.com/bmj BMJ Page 68 of 244

1 2 3 Derosa G, Maffioli P, D'Angelo A, et al. Acarbose on insulin resistance after an oral fat load: a double- 4 blind, placebo controlled study. J Diabetes Complications 2011;25(4):258-66 5 6 7 Derosa 2011a 8 Derosa G, Putignano P, Bossi AC, et al. Exenatide or glimepiride added to metformin on metabolic 9 control and on insulin resistance in type 2 diabetic patients. Eur J Pharmacol 2011;666(1-3):251-6 10 11 Derosa 2012 12 Derosa G,Confidential: Carbone A, D'Angelo A, et al. A randomized, For Review double-blind, placebo-controlled Only trial 13 evaluating sitagliptin action on insulin resistance parameters and beta-cell function. Expert Opin 14 Pharmacother 2012;13(17):2433-42 15 16 17 Derosa 2012a 18 Derosa G, Ragonesi PD, Carbone A, et al. Vildagliptin added to metformin on beta-cell function after 19 a euglycemic hyperinsulinemic and hyperglycemic clamp in type 2 diabetes patients. Diabetes 20 Technol Ther 2012;14(6):475-84 21 22 Derosa 2013 23 Derosa G, Cicero AFG, Franzetti IG, et al. Effects of exenatide and metformin in combination on some 24 25 adipocytokine levels: A comparison with metformin monotherapy. Can J Physiol Pharmacol 26 2013;91(9):724-32 27 28 Derosa 2014 29 Derosa G, Bonaventura A, Bianchi L, et al. Vildagliptin compared to glimepiride on post-prandial 30 lipemia and on insulin resistance in type 2 diabetic patients. Metabolism 2014;63(7):957-67 31 32 Derosa 2014a 33 Derosa G, Ragonesi PD, Fogari E, et al. Sitagliptin added to previously taken antidiabetic agents on 34 35 insulin resistance and lipid profile: A 2-year study evaluation. Fundam Clin Pharmacol 36 2014;28(2):221-29 37 38 DIA3004 39 Yale JF, Bakris G, Cariou B, et al. Efficacy and safety of canagliflozin over 52 weeks in patients with 40 type 2 diabetes mellitus and chronic kidney disease. Diabetes Obes Metab 2014;16(10):1016-27 41 42 Distiller 2014 43 44 Distiller LA, Nortje H, Wellmann H, et al. A 24-week, prospective, randomized, open-label, treat-to- 45 target pilot study of obese type 2 diabetes patients with severe insulin resistance to assess the 46 addition of exenatide on the efficacy of U-500 plus metformin. Endocr Pract 47 2014;20(11):1143-50 48 49 DIVERSITY-CVR 50 Fuchigami A, Shigiyama F, Kitazawa T, et al. Efficacy of dapagliflozin versus sitagliptin on 51 cardiometabolic risk factors in Japanese patients with type 2 diabetes: a prospective, randomized 52 53 study (DIVERSITY-CVR). Cardiovasc Diabetol 2020;19(1) 54 55 Dobs 2013 56 Dobs AS, Goldstein BJ, Aschner P, et al. Efficacy and safety of sitagliptin added to ongoing metformin 57 and rosiglitazone combination therapy in a randomized placebo-controlled 54-week trial in patients 58 with type 2 diabetes. J Diabetes 2013;5(1):68-79 59 60

https://mc.manuscriptcentral.com/bmj Page 69 of 244 BMJ

1 2 3 Dorkhan 2009 4 Dorkhan M, Dencker M, Stagmo M, et al. Effect of pioglitazone versus insulin glargine on cardiac size, 5 6 function, and measures of fluid retention in patients with type 2 diabetes. Cardiovasc Diabetol 7 2009;8:15 8 9 Douek 2005 10 Douek IF, Allen SE, Ewings P, et al. Continuing metformin when starting insulin in patients with Type 11 2 diabetes: a double-blind randomized placebo-controlled trial. Diabetic Med 2005;22(5):634-40 12 Confidential: For Review Only 13 Drent 2002 14 Drent ML, Tollefsen AT, Heusden FH, et al. Dose-dependent efficacy of miglitol, an alpha-glucosidase 15 16 inhibitor, in type 2 diabetic patients on diet alone: results of a 24-week double-blind placebo- 17 controlled study. Diab Nutr Metab 2002;15(3):152-9 18 19 DUAL-I 20 Gough SC, Bode B, Woo V, et al. Efficacy and safety of a fixed-ratio combination of insulin degludec 21 and liraglutide (IDegLira) compared with its components given alone: results of a phase 3, open- 22 label, randomised, 26-week, treat-to-target trial in insulin-naive patients with type 2 diabetes. 23 Lancet Diabetes Endocrinol 2014;2(11):885-93 24 25 26 DUAL-II 27 Buse JB, Vilsboll T, Thurman J, et al. Contribution of liraglutide in the fixed-ratio combination of 28 insulin degludec and liraglutide (IDegLira). Diabetes Care 2014;37(11):2926-33 29 30 DUAL-III 31 Linjawi S, Bode BW, Chaykin LB, et al. The efficacy of IDegLira (Insulin Degludec/Liraglutide 32 Combination) in adults with type 2 diabetes inadequately controlled with a GLP-1 receptor agonist 33 and oral therapy: DUAL III randomized clinical trial. Diabetes Ther 2017;8(1):101-14 34 35 36 DUALTM IX 37 Philis-Tsimikas A, Billings LK, Busch R, et al. Superior efficacy of insulin degludec/liraglutide versus 38 insulin glargine U100 as add-on to sodium-glucose co-transporter-2 inhibitor therapy: a randomized 39 clinical trial in people with uncontrolled type 2 diabetes. Diab Obes Metab 2019;21(6):1399-408 40 41 DURATION-2 42 Bergenstal RM, Wysham C, Macconell L, et al. Efficacy and safety of exenatide once weekly versus 43 44 sitagliptin or pioglitazone as an adjunct to metformin for treatment of type 2 diabetes (DURATION- 45 2): a randomised trial. Lancet 2010;376(9739):431-9 46 47 DURATION-3 48 Diamant M, Gaal L, Stranks S, et al. Once weekly exenatide compared with insulin glargine titrated to 49 target in patients with type 2 diabetes (DURATION-3): an open-label randomised trial. Lancet 50 2010;375(9733):2234-43 51 52 53 DURATION-4 54 Russell-Jones D, Cuddihy RM, Hanefeld M, et al. Efficacy and safety of exenatide once weekly versus 55 metformin, pioglitazone, and sitagliptin used as monotherapy in drug-naive patients with type 2 56 diabetes (DURATION-4): a 26-week double-blind study. Diabetes Care 2012;35(2):252-8 57 58 DURATION-8 59 60

https://mc.manuscriptcentral.com/bmj BMJ Page 70 of 244

1 2 3 Frias JP, Guja C, Hardy E, et al. Exenatide once weekly plus dapagliflozin once daily versus exenatide 4 or dapagliflozin alone in patients with type 2 diabetes inadequately controlled with metformin 5 6 monotherapy (DURATION-8): a 28 week, multicentre, double-blind, phase 3, randomised controlled 7 trial. Lancet Diabetes Endocrinol 2016;4(12):1004-16 8 9 DURATION-NEO-2 10 Gadde KM, Vetter ML, Iqbal N, et al. Efficacy and safety of autoinjected exenatide once-weekly 11 suspension versus sitagliptin or placebo with metformin in patients with type 2 diabetes: The 12 DURATION-NEO-2Confidential: randomized clinical study. For Diabetes Review Obes Metab 2017; 19Only(7):979-88 13 14 EAGLE 15 16 D'Alessio D, Haring HU, Charbonnel B, et al. Comparison of insulin glargine and liraglutide added to 17 oral agents in patients with poorly controlled type 2 diabetes. Diabetes Obes Metab 2015;17(2):170- 18 78 19 20 EASIE 21 Aschner P, Chan J, Owens DR, et al. Insulin glargine versus sitagliptin in insulin-naive patients with 22 type 2 diabetes mellitus uncontrolled on metformin (EASIE): a multicentre, randomised open-label 23 trial. Lancet 2012;379(9833):2262-9 24 25 26 Ebato 2009 27 Ebato C, Shimizu T, Arakawa M, et al. Effect of sulfonylureas on switching to insulin therapy (twice- 28 daily biphasic insulin aspart 30): comparison of twice-daily biphasic insulin aspart 30 with or without 29 glimepiride in type 2 diabetic patients poorly controlled with sub-maximal glimepiride. Diabetes Res 30 Clin Pract 2009;86(1):31-6 31 32 EDIT 33 Sato S, Saisho Y, Kou K, et al. Efficacy and safety of sitagliptin added to insulin in Japanese patients 34 35 with type 2 Diabetes: The EDIT randomized trial. PLoS ONE 2015;10(3) 36 37 Efstathiou 2015 38 Efstathiou S, Skeva I, Mountokalakis T. Dapagliflozin may attenuate adipose tissue inflammation and 39 arterial stiffness in type 2 diabetes. Diabetes 2015;64:A322 40 41 ELEGANT 42 de Wit HM, Vervoort GM, Jansen HJ, et al. Liraglutide reverses pronounced insulin-associated weight 43 44 gain, improves glycaemic control and decreases insulin dose in patients with type 2 diabetes: a 26 45 week, randomised clinical trial (ELEGANT). Diabetologia 2014;57(9):1812-9 46 47 ELIXA 48 Pfeffer MA, Claggett B, Diaz R, et al. Lixisenatide in patients with type 2 diabetes and acute coronary 49 syndrome. N Engl J Med 2015;373(23):2247-57 50 51 ELLENA-IT 52 53 Nakaguchi H, Kondo Y, Kyohara M, et al. Effects of liraglutide and empagliflozin added to insulin 54 therapy in patients with type 2 diabetes: a randomized controlled study ELLENA-IT study. J Diabetes 55 Invest 2020 56 57 EMBLEM 58 59 60

https://mc.manuscriptcentral.com/bmj Page 71 of 244 BMJ

1 2 3 Tanaka A, Shimabukuro M, Machii N, et al. Effect of empagliflozin on endothelial function in patients 4 with type 2 diabetes and cardiovascular disease: Results from the multicenter, randomized, placebo- 5 6 controlled, double-blind EMBLEM Trial. Diabetes Care 2019;42(10):e159-e61 7 8 EMIT 9 Kashiwagi A, Akiyama N, Shiga T, et al. Efficacy and safety of ipragliflozin as an add-on to a 10 sulfonylurea in Japanese patients with inadequately controlled type 2 diabetes: results of the 11 randomized, placebo-controlled, double-blind, phase III EMIT study. Diabetol Int 2015;6(2):125-38 12 Confidential: For Review Only 13 EMLIFA001 14 Kahl S, Gancheva S, Strassburger K, et al. Empagliflozin effectively lowers liver fat content in well- 15 16 controlled type 2 diabetes: A randomized, double-blind, phase 4, placebo-controlled trial. Diabetes 17 Care 2019 18 19 EMPA-HEART CardioLink-6 20 Verma S, Mazer CD, Yan AT, et al. Effect of empagliflozin on left ventricular mass in patients with 21 type 2 diabetes and coronary artery disease: the EMPA-HEART CardioLink-6 randomized clinical trial. 22 Circulation 2019 23 24 25 EMPA-REG BASAL 26 Rosenstock J, Jelaska A, Zeller C, et al. Impact of empagliflozin added on to basal insulin in type 2 27 diabetes inadequately controlled on basal insulin: a 78-week randomized, double-blind, placebo- 28 controlled trial. Diabetes Obes Metab 2015;17(10):936-48 29 30 EMPA-REG H2H-SU 31 Ridderstrale M, Andersen KR, Zeller C, et al. Comparison of empagliflozin and glimepiride as add-on 32 to metformin in patients with type 2 diabetes: A 104-week randomised, active-controlled, double- 33 blind, phase 3 trial. Lancet Diabetes Endocrinol 2014;2(9):691-700 34 35 36 EMPA-REG MDI 37 Rosenstock J, Jelaska A, Frappin G, et al. Improved glucose control with weight loss, lower insulin 38 doses, and no increased hypoglycemia with empagliflozin added to titrated multiple daily injections 39 of insulin in obese inadequately controlled type 2 diabetes. Diabetes Care 2014;37(7):1815-23 40 41 EMPA-REG MET 42 Haring HU, Merker L, Seewaldt-Becker E, et al. Empaglif lozin as add-on to metformin in patients 43 44 with type 2 diabetes: A 24-week, randomized, double-blind, placebo-controlled trial. Diabetes Care 45 2014;37(6):1650-59 46 47 EMPA-REG METSU 48 Haring HU, Merker L, Seewaldt-Becker E, et al. Empagliflozin as add-on to metformin plus 49 sulfonylurea in patients with type 2 diabetes: A 24-week, randomized, double-blind, placebo- 50 controlled trial. Diabetes Care 2013;36(11):3396-404 51 52 53 EMPA-REG MONO 54 Roden M, Weng J, Eilbracht J, et al. Empagliflozin monotherapy with sitagliptin as an active 55 comparator in patients with type 2 diabetes: A randomised, double-blind, placebo-controlled, phase 56 3 trial. Lancet Diabetes Endocrinol 2013;1(3):208-19 57 58 EMPA-REG OUTCOME 59 60

https://mc.manuscriptcentral.com/bmj BMJ Page 72 of 244

1 2 3 Zinman B, Wanner C, Lachin JM, et al. Empagliflozin, cardiovascular outcomes, and mortality in type 4 2 diabetes. N Engl J Med 2015;373:2117-28 5 6 7 EMPA-REG PIO 8 Kovacs CS, Seshiah V, Swallow R, et al. Empagliflozin improves glycaemic and weight control as add- 9 on therapy to pioglitazone or pioglitazone plus metformin in patients with type 2 diabetes: A 24- 10 week, randomized, placebo-controlled trial. Diabetes Obes Metab 2014;16(2):147-58 11 12 EMPA-REGConfidential: RENAL For Review Only 13 Barnett AH, Mithal A, Manassie J, et al. Efficacy and safety of empagliflozin added to existing 14 antidiabetes treatment in patients with type 2 diabetes and chronic kidney disease: A randomised, 15 16 double-blind, placebo-controlled trial. Lancet Diabetes Endocrinol 2014;2(5):369-84 17 18 Engelbrechtsen 2016 19 Engelbrechtsen L, Iepsen EPW, Andersson EA, et al. Weight loss and weight maintenance obtained 20 with or without GLP-1 analogue treatment decrease branched chain amino acid levels. 21 Metabolomics 2016;12(12):181 22 23 Erem 2014 24 25 Erem C, Ozbas HM, Nuhoglu I, et al. Comparison of effects of gliclazide, metformin and pioglitazone 26 monotherapies on glycemic control and cardiovascular risk factors in patients with newly diagnosed 27 uncontrolled type 2 diabetes mellitus. Exp Clin Endocrinol Diabetes 2014;122(5):295-302 28 29 ESPECIAL-ACS 30 Kuramitsu S, Miyauchi K, Yokoi H, et al. Effect of sitagliptin on plaque changes in coronary artery 31 following acute coronary syndrome in diabetic patients: The ESPECIAL-ACS study. J Cardiol 32 2017;69(1):369-76 33 34 35 Esposito 2004 36 Esposito K, Giugliano D, Nappo F, et al. Regression of carotid atherosclerosis by control of 37 postprandial hyperglycemia in type 2 diabetes mellitus. Circulation 2004;110(2):214-9 38 39 Esposito 2011 40 Esposito K, Maiorino MI, Di Palo C, et al. Effects of pioglitazone versus metformin on circulating 41 endothelial microparticles and progenitor cells in patients with newly diagnosed type 2 diabetes-a 42 randomized controlled trial. Diabetes Obes Metab 2011;13(5):439-45 43 44 45 Essen Study 46 Hoffmann J, Spengler M. Efficacy of 24-week monotherapy with acarbose, glibenclamide, or placebo 47 in NIDDM patients. The Essen Study. Diabetes Care 1994;17(6):561-6 48 49 EUREXA 50 Gallwitz B, Guzman J, Dotta F, et al. Exenatide twice daily versus glimepiride for prevention of 51 glycaemic deterioration in patients with type 2 diabetes with metformin failure (EUREXA): an open- 52 53 label, randomised controlled trial. Lancet 2012;379(9833):2270-8 54 55 EUREXA extension 56 Schernthaner G, Rosas-Guzman J, Dotta F, et al. Treatment escalation options for patients with type 57 2 diabetes after failure of exenatide twice daily or glimepiride added to metformin: Results from the 58 prospective European Exenatide (EUREXA) study. Diabetes Obes Metab 2015;17(7):689-98 59 60

https://mc.manuscriptcentral.com/bmj Page 73 of 244 BMJ

1 2 3 EXAMINE 4 White WB, Cannon CP, Heller SR, et al. Alogliptin after acute coronary syndrome in patients with 5 6 type 2 diabetes. N Engl J Med 2013;369(14):1327-35 7 8 Exenatide-113 9 Buse JB, Henry RR, Han J, et al. Effects of exenatide (Exendin-4) on glycemic control over 30 weeks in 10 sulfonylurea-treated patients with type 2 diabetes. Diabetes Care 2004;27(11):2628-35 11 12 EXSCEL Confidential: For Review Only 13 Holman RR, Bethel MA, Mentz RJ, et al. Effects of Once-Weekly Exenatide on Cardiovascular 14 Outcomes in Type 2 Diabetes. N Engl J Med 2017;377(13):1228-39 15 16 17 Feng 2017 18 Feng W, Gao C, Bi Y, et al. Randomized trial comparing the effects of gliclazide, liraglutide, and 19 metformin on diabetes with non-alcoholic fatty liver disease. J Diabetes 2017 20 21 Ferdinand 2019 22 Ferdinand KC, Izzo JL, Lee J, et al. Antihyperglycemic and Blood Pressure Effects of Empagliflozin in 23 Black Patients With Type 2 Diabetes Mellitus and Hypertension. Circulation 2019;139(18):2098-109 24 25 26 Fernandez 2008 27 Fernandez M, Triplitt C, Wajcberg E, et al. Addition of pioglitazone and ramipril to intensive insulin 28 therapy in type 2 diabetic patients improves vascular dysfunction by different mechanisms. Diabetes 29 Care 2008;31(1):121-7 30 31 Ferrannini 2010 32 Ferrannini E, Ramos SJ, Salsali A, et al. Dapagliflozin monotherapy in type 2 diabetic patients with 33 inadequate glycemic control by diet and exercise: a randomized, double-blind, placebo-controlled, 34 35 phase 3 trial. Diabetes Care 2010;33(10):2217-24 36 37 Filozof 2010 38 Filozof C, Gautier JF. A comparison of efficacy and safety of vildagliptin and gliclazide in combination 39 with metformin in patients with Type 2 diabetes inadequately controlled with metformin alone: a 40 52-week, randomized study. Diabetic Med 2010;27(3):318-26 41 42 Finn 2009 43 44 Finn AV, Oh JS, Hendricks M, et al. Predictive factors for in-stent late loss and coronary lesion 45 progression in patients with type 2 diabetes mellitus randomized to rosiglitazone or placebo. Am 46 Heart J 2009;157(2):383.e1-8 47 48 Fischer 1998 49 Fischer S, Hanefeld M, Spengler M, et al. European study on dose-response relationship of acarbose 50 as a first-line drug in non-insulin-dependent diabetes mellitus: efficacy and safety of low and high 51 doses. Acta Diabetologica 1998;35(1):34-40 52 53 54 Foley 2009 55 Foley JE, Sreenan S. Efficacy and safety comparison between the DPP-4 inhibitor vildagliptin and the 56 sulfonylurea gliclazide after two years of monotherapy in drug-naïve patients with type 2 diabetes. 57 Horm Metab Res 2009;41(12):905-9 58 59 Foley 2011 60

https://mc.manuscriptcentral.com/bmj BMJ Page 74 of 244

1 2 3 Foley JE, Bunck MC, Möller-Goede DL, et al. Beta cell function following 1 year vildagliptin or placebo 4 treatment and after 12 week washout in drug-naive patients with type 2 diabetes and mild 5 6 hyperglycaemia: a randomised controlled trial. Diabetologia 2011;54(8):1985-91 7 8 Fonseca 2000 9 Fonseca V, Rosenstock J, Patwardhan R, et al. Effect of metformin and rosiglitazone combination 10 therapy in patients with type 2 diabetes mellitus: a randomized controlled trial. JAMA 11 2000;283(13):1695-702 12 Confidential: For Review Only 13 Fonseca 2003 14 Fonseca V, Grunberger G, Gupta S, et al. Addition of nateglinide to rosiglitazone monotherapy 15 16 suppresses mealtime hyperglycemia and improves overall glycemic control. Diabetes Care 17 2003;26(6):1685-90 18 19 Fonseca 2007 20 Fonseca V, Schweizer A, Albrecht D, et al. Addition of vildagliptin to insulin improves glycaemic 21 control in type 2 diabetes. Diabetologia 2007;50(6):1148-55 22 23 Fonseca 2013 24 25 Fonseca V, Staels B, Morgan JD, et al. Efficacy and safety of sitagliptin added to ongoing metformin 26 and pioglitazone combination therapy in a randomized, placebo-controlled, 26-week trial in patients 27 with type 2 diabetes. J Diabetes Complications 2013;27(2):177-83 28 29 Forst 2003 30 Forst T, Eriksson JW, Strotmann HJ, et al. Metabolic effects of mealtime insulin lispro in comparison 31 to glibenclamide in early type 2 diabetes. Exp Clin Endocrinol Diabetes 2003;111(2):97-103 32 33 Forst 2005 34 35 Forst T, Hohberg C, Fuellert SD, et al. Pharmacological PPARgamma stimulation in contrast to beta 36 cell stimulation results in an improvement in adiponectin and proinsulin intact levels and reduces 37 intima media thickness in patients with type 2 diabetes. Horm Metab Res 2005;37(8):521-7 38 39 Forst 2014 40 Forst T, Guthrie R, Goldenberg R, et al. Efficacy and safety of canagliflozin over 52 weeks in patients 41 with type 2 diabetes on background metformin and pioglitazone. Diabetes Obes Metab 42 2014;16(5):467-77 43 44 45 Forst 2015 46 Forst T, Koch C, Dworak M. Vildagliptin versus insulin in patients with type 2 diabetes mellitus 47 inadequately controlled with sulfonylurea: Results from a randomized, 24 week study. Curr Med Res 48 Opin 2015;31(6):1079-84 49 50 Frederich 2012 51 Frederich R, McNeill R, Berglind N, et al. The efficacy and safety of the dipeptidyl peptidase-4 52 53 inhibitor saxagliptin in treatment-naive patients with type 2 diabetes mellitus: A randomized 54 controlled trial. Diabetol Metab Syndr 2012;4(1) 55 56 FREEDOM-1 57 Rosenstock J, Buse JB, Azeem R, et al. Efficacy and Safety of ITCA 650, a Novel Drug-Device GLP-1 58 Receptor Agonist, in Type 2 Diabetes Uncontrolled With Oral Antidiabetes Drugs: the FREEDOM-1 59 Trial. Diabetes Care 2018;41(2):333-40 60

https://mc.manuscriptcentral.com/bmj Page 75 of 244 BMJ

1 2 3 4 Gaal 2001 5 6 Gaal L, Maislos M, Schernthaner G, et al. Miglitol combined with metformin improves glycaemic 7 control in type 2 diabetes. Diabetes Obes Metab 2001;3(5):326-31 8 9 Gallwitz 2011 10 Gallwitz B, Böhmer M, Segiet T, et al. Exenatide twice daily versus premixed insulin aspart 70/30 in 11 metformin-treated patients with type 2 diabetes: a randomized 26-week study on glycemic control 12 and hypoglycemia.Confidential: Diabetes Care 2011;34(3):604-6 For Review Only 13 14 Gallwitz 2012 15 16 Gallwitz B, Rosenstock J, Rauch T, et al. 2-year efficacy and safety of linagliptin compared with 17 glimepiride in patients with type 2 diabetes inadequately controlled on metformin: a randomised, 18 double-blind, non-inferiority trial. Lancet 2012;380(9840):475-83 19 20 Gantz 2017 21 Gantz I, Chen M, Suryawanshi S, et al. A randomized, placebo-controlled study of the cardiovascular 22 safety of the once-weekly DPP-4 inhibitor omarigliptin in patients with type 2 diabetes mellitus. 23 Cardiovasc Diabetol 2017;16(1):112 24 25 26 Gantz 2017a 27 Gantz I, Okamoto T, Ito Y, et al. A Randomized, Placebo-Controlled Trial Evaluating the Safety and 28 Efficacy of Adding Omarigliptin to Antihyperglycemic Therapies in Japanese Patients with Type 2 29 Diabetes and Inadequate Glycemic Control. Diabetes Ther 2017;8(4):793-810 30 31 Gantz 2017b 32 Gantz I, Sokolova L, Jain L, et al. Use of Prohibited Medication, a Potentially Overlooked Confounder 33 in Clinical Trials: omarigliptin (Once-weekly DPP-4 Inhibitor) Monotherapy Trial in 18- to 45-year- 34 35 olds. Clin Ther 2017;39(10):2024-37 36 37 Garber 2006 38 Garber A, Klein E, Bruce S, et al. Metformin-glibenclamide versus metformin plus rosiglitazone in 39 patients with type 2 diabetes inadequately controlled on metformin monotherapy. Diabetes Obes 40 Metab 2006;8(2):156-63 41 42 Garber 2007 43 44 Garber AJ, Schweizer A, Baron MA, et al. Vildagliptin in combination with pioglitazone improves 45 glycaemic control in patients with type 2 diabetes failing thiazolidinedione monotherapy: a 46 randomized, placebo-controlled study. Diabetes Obes Metab 2007;9(2):166-74 47 48 Garber 2008 49 Garber AJ, Foley JE, Banerji MA, et al. Effects of vildagliptin on glucose control in patients with type 2 50 diabetes inadequately controlled with a sulphonylurea. Diabetes Obes Metab 2008;10(11):1047-56 51 52 53 Gastaldelli 2014 54 Gastaldelli A, Brodows RG, D'Alessio D. The effect of chronic twice daily exenatide treatment on 55 beta-cell function in new onset type 2 diabetes. Clin Endocrinol 2014;80(4):545-53 56 57 GENERATION 58 59 60

https://mc.manuscriptcentral.com/bmj BMJ Page 76 of 244

1 2 3 Schernthaner G, Duran-Garcia S, Hanefeld M, et al. Efficacy and tolerability of saxagliptin compared 4 with glimepiride in elderly patients with type 2 diabetes: A randomized, controlled study 5 6 (GENERATION). Diabetes Obes Metab 2015;17(7):630-38 7 8 Gentile 2001 9 Gentile S, Turco S, Guarino G, et al. Effect of treatment with acarbose and insulin in patients with 10 non-insulin-dependent diabetes mellitus associated with non-alcoholic liver cirrhosis. Diabetes Obes 11 Metab 2001;3(1):33-40 12 Confidential: For Review Only 13 GetGoal Duo-2 14 Rosenstock J, Guerci B, Hanefeld M, et al. Prandial options to advance basal insulin glargine therapy: 15 16 Testing lixisenatide plus basal insulin versus either as basal-plus or basal-bolus in 17 type 2 diabetes: The GetGoal Duo-2 Trial. Diabetes Care 2016;39(8):1318-28 18 19 GetGoal-Duo-1 20 Riddle MC, Forst T, Aronson R, et al. Adding once-daily lixisenatide for type 2 diabetes inadequately 21 controlled with newly initiated and continuously titrated basal insulin glargine: a 24-week, 22 randomized, placebo-controlled study (GetGoal-Duo 1). Diabetes Care 2013;36(9):2497-503 23 24 25 GetGoal-F1 26 Bolli GB, Munteanu M, Dotsenko S, et al. Efficacy and safety of lixisenatide once daily vs. placebo in 27 people with Type 2 diabetes insufficiently controlled on metformin (GetGoal-F1). Diabetic Med 28 2014;31(2):176-84 29 30 GetGoal-L 31 Riddle MC, Aronson R, Home P, et al. Adding once-daily lixisenatide for type 2 diabetes inadequately 32 controlled by established basal insulin: a 24-week, randomized, placebo-controlled comparison 33 (GetGoal-L). Diabetes Care 2013;36(9):2489-96 34 35 36 GetGoal-L-Asia 37 Seino Y, Min KW, Niemoeller E, et al. Randomized, double-blind, placebo-controlled trial of the once- 38 daily GLP-1 receptor agonist lixisenatide in Asian patients with type 2 diabetes insufficiently 39 controlled on basal insulin with or without a sulfonylurea (GetGoal-L-Asia). Diabetes Obes Metab 40 2012;14(10):910-7 41 42 GetGoal-L-C 43 44 Yang W, Min K, Zhou Z, et al. Efficacy and safety of lixisenatide in a predominantly Asian population 45 with type 2 diabetes insufficiently controlled with basal insulin: the GetGoal-L-C randomized trial. 46 Diab Obes Metab 2018;20(2):335-43 47 48 GetGoal-M 49 Ahrén B, Leguizamo Dimas A, Miossec P, et al. Efficacy and safety of lixisenatide once-daily morning 50 or evening injections in type 2 diabetes inadequately controlled on metformin (GetGoal-M). 51 Diabetes Care 2013;36(9):2543-50 52 53 54 GetGoal-M-Asia 55 Yu Pan C, Han P, Liu X, et al. Lixisenatide treatment improves glycaemic control in Asian patients with 56 type 2 diabetes mellitus inadequately controlled on metformin with or without sulfonylurea: a 57 randomized, double-blind, placebo-controlled, 24-week trial (GetGoal-M-Asia). Diabetes Metab Res 58 Rev 2014;30(8):726-35 59 60

https://mc.manuscriptcentral.com/bmj Page 77 of 244 BMJ

1 2 3 GetGoal-O 4 Meneilly GS, Roy-Duval C, Alawi H, et al. Lixisenatide therapy in older patients with type 2 diabetes 5 6 inadequately controlled on their current antidiabetic treatment: The GetGoal-O randomized trial. 7 Diabetes Care 2017;40(4):485-93 8 9 GetGoal-P 10 Pinget M, Goldenberg R, Niemoeller E, et al. Efficacy and safety of lixisenatide once daily versus 11 placebo in type 2 diabetes insufficiently controlled on pioglitazone (GetGoal-P). Diabetes Obes 12 Metab 2013;Confidential:15(11):1000-07 For Review Only 13 14 GetGoal-S 15 16 Rosenstock J, Hanefeld M, Shamanna P, et al. Beneficial effects of once-daily lixisenatide on overall 17 and postprandial glycemic levels without significant excess of hypoglycemia in Type 2 diabetes 18 inadequately controlled on a sulfonylurea with or without metformin (GetGoal-S). J Diabetes 19 Complications 2014;28(3):386-92 20 21 Giles 2008 22 Giles TD, Miller AB, Elkayam U, et al. Pioglitazone and heart failure: results from a controlled study in 23 patients with type 2 diabetes mellitus and systolic dysfunction. J Card Fail 2008;14(6):445-52 24 25 26 Giugliano 1993 27 Giugliano D, Quatraro A, Consoli G, et al. Metformin for obese, insulin-treated diabetic patients: 28 improvement in glycaemic control and reduction of metabolic risk factors. Eur J Clin Pharmacol 29 1993;44(2):107-12 30 31 GLAC 32 Tan MH, Johns D, Strand J, et al. Sustained effects of pioglitazone vs. glibenclamide on insulin 33 sensitivity, glycaemic control, and lipid profiles in patients with Type 2 diabetes. Diabet Med 34 35 2004;21(8):859-66 36 37 GLAD 38 Tan M, Johns D, González Gálvez G, et al. Effects of pioglitazone and glimepiride on glycemic control 39 and insulin sensitivity in Mexican patients with type 2 diabetes mellitus: A multicenter, randomized, 40 double-blind, parallel-group trial. Clin Ther 2004;26(5):680-93 41 42 GLAL 43 44 Tan MH, Baksi A, Krahulec B, et al. Comparison of pioglitazone and gliclazide in sustaining glycemic 45 control over 2 years in patients with type 2 diabetes. Diabetes Care 2005;28(3):544-50 46 47 Glimepiride Combination Group study 1998 48 Riddle MC, Schneider J. Beginning insulin treatment of obese patients with evening 70/30 insulin 49 plus glimepiride versus insulin alone. Glimepiride Combination Group. Diabetes Care 50 1998;21(7):1052-7 51 52 53 Göke 2002 54 Göke B. Improved glycemic control and lipid profile in a randomized study of pioglitazone compared 55 with acarbose in patients with type 2 diabetes mellitus. Treat Endocrinol 2002;1(5):329-36 56 57 Göke 2010 58 59 60

https://mc.manuscriptcentral.com/bmj BMJ Page 78 of 244

1 2 3 Göke B, Gallwitz B, Eriksson J, et al. Saxagliptin is non-inferior to glipizide in patients with type 2 4 diabetes mellitus inadequately controlled on metformin alone: a 52-week randomised controlled 5 6 trial. Int J Clin Pract 2010;64(12):1619-31 7 8 Gómez-Perez 2002 9 Gómez-Perez FJ, Fanghänel-Salmón G, Antonio Barbosa J, et al. Efficacy and safety of rosiglitazone 10 plus metformin in Mexicans with type 2 diabetes. Diabetes Metab Res Rev 2002;18(2):127-34 11 12 Gomis 2011Confidential: For Review Only 13 Gomis R, Espadero RM, Jones R, et al. Efficacy and safety of initial combination therapy with 14 linagliptin and pioglitazone in patients with inadequately controlled type 2 diabetes: a randomized, 15 16 double-blind, placebo-controlled study. Diabetes Obes Metab 2011;13(7):653-61 17 18 Goodman 2009 19 Goodman M, Thurston H, Penman J. Efficacy and tolerability of vildagliptin in patients with type 2 20 diabetes inadequately controlled with metformin monotherapy. Horm Metab Res 2009;41(5):368-73 21 22 Grant 1996 23 Grant PJ. The effects of high- and medium-dose metformin therapy on cardiovascular risk factors in 24 25 patients with type II diabetes. Diabetes Care 1996;19(1):64-6 26 27 GRAVITAS 28 Miras AD, Perez-Pevida B, Aldhwayan M, et al. Adjunctive liraglutide treatment in patients with 29 persistent or recurrent type 2 diabetes after metabolic surgery (GRAVITAS): a randomised, double- 30 blind, placebo-controlled trial. Lancet Diabetes Endocrinol 2019;7(7):549-59 31 32 Grey 2014 33 Grey A, Bolland M, Fenwick S, et al. The skeletal effects of pioglitazone in type 2 diabetes or 34 35 impaired glucose tolerance: A randomized controlled trial. Eur J Endocrinol 2014;170(2):255-62 36 37 Gudipaty 2014 38 Gudipaty L, Rosenfeld NK, Fuller CS, et al. Effect of exenatide, sitagliptin, or glimepiride on beta-cell 39 secretory capacity in early type 2 diabetes. Diabetes Care 2014;37(9):2451-58 40 41 Gurkan 2014 42 Gurkan E, Tarkun I, Sahin T, et al. Evaluation of exenatide versus insulin glargine for the impact on 43 44 endothelial functions and cardiovascular risk markers. Diabetes Res Clin Pract 2014;106(3):567-75 45 46 Gutniak 1987 47 Gutniak M, Karlander SG, Efendi S. Glyburide decreases insulin requirement, increases beta-cell 48 response to mixed meal, and does not affect insulin sensitivity: effects of short- and long-term 49 combined treatment in secondary failure to sulfonylurea. Diabetes Care 1987;10(5):545-54 50 51 Güvener 1999 52 53 Güvener N, Gedik O. Effects of combination of insulin and acarbose compared with insulin and 54 gliclazide in type 2 diabetic patients. Acta Diabetol 1999;36(1-2):93-7 55 56 Guzman 2017 57 Guzman CB, Zhang XM, Liu R, et al. Treatment with LY2409021, a glucagon receptor antagonist, 58 increases liver fat in patients with type 2 diabetes. Diabetes Obes Metab 2017;19(11):1521-28 59 60

https://mc.manuscriptcentral.com/bmj Page 79 of 244 BMJ

1 2 3 Haak 2012 4 Haak T, Meinicke T, Jones R, et al. Initial combination of linagliptin and metformin improves 5 6 glycaemic control in type 2 diabetes: a randomized, double-blind, placebo-controlled study. Diabetes 7 Obes Metab 2012;14(6):565-74 8 9 Hadjadj 2016 10 Hadjadj S, Rosenstock J, Meinicke T, et al. Initial combination of empagliflozin and metformin in 11 patients with type 2 diabetes. Diabetes Care 2016;39(10):1718-28 12 Confidential: For Review Only 13 Halimi 2000 14 Halimi S, Berre MA, Grangé V. Efficacy and safety of acarbose add-on therapy in the treatment of 15 16 overweight patients with Type 2 diabetes inadequately controlled with metformin: a double-blind, 17 placebo-controlled study. Diabetes Res Clin Pract 2000;50(1):49-56 18 19 Hällsten 2002 20 Hällsten K, Virtanen KA, Lönnqvist F, et al. Rosiglitazone but not metformin enhances insulin- and 21 exercise-stimulated skeletal muscle glucose uptake in patients with newly diagnosed type 2 22 diabetes. Diabetes 2002;51(12):3479-85 23 24 25 Halvorsen 2019 26 Halvorsen YC, Walford G, Massaro J, et al. A 96-week, multinational, randomized, double-blind, 27 parallel-group, clinical trial evaluating the safety and effectiveness of bexagliflozin as a monotherapy 28 for adults with type 2 diabetes. Diab Obes Metab 2019;21(11):2496-504 29 30 Han 2018 31 Han KA, Chon S, Chung CH, et al. Efficacy and safety of ipragliflozin as an add-on therapy to 32 sitagliptin and metformin in Korean patients with inadequately controlled type 2 diabetes mellitus: a 33 randomized controlled trial. Diab Obes Metab 2018;20(10):2408-15 34 35 36 Handelsman 2017 37 Handelsman Y, Lauring B, Gantz I, et al. A randomized, double-blind, non-inferiority trial evaluating 38 the efficacy and safety of omarigliptin, a once-weekly DPP-4 inhibitor, or glimepiride in patients with 39 type 2 diabetes inadequately controlled on metformin monotherapy. Curr Med Res Opin 2017:1-8 40 41 Handelsman 2019 42 Handelsman Y, Mathieu C, Del Prato S, et al. Sustained 52 week efficacy and safety of triple therapy 43 44 with dapagliflozin plus saxagliptin versus dual therapy with sitagliptin as addon to metformin in 45 uncontrolled type 2 diabetes. Diab Obes Metab 2017;21(4):883-92 46 47 Haneda 2016 48 Haneda M, Seino Y, Inagaki N, et al. Influence of renal function on the 52-week efficacy and safety of 49 the socium glucose cotransporter 2 inhibitor luseogliflozin in Japanese patients with type 2 diabetes 50 mellitus. Clin Ther 2016;38(1):88e66-88e120 51 52 53 Hanefeld 1991 54 Hanefeld M, Fischer S, Schulze J, et al. Therapeutic potentials of acarbose as first-line drug in NIDDM 55 insufficiently treated with diet alone. Diabetes Care 1991;14(8):732-7 56 57 Hanefeld 2004 58 59 60

https://mc.manuscriptcentral.com/bmj BMJ Page 80 of 244

1 2 3 Hanefeld M, Brunetti P, Schernthaner GH, et al. One-year glycemic control with a sulfonyurea plus 4 pioglitazone versus a sulfonylurea plus metformin in patients with type 2 diabetes. Diabetes Care 5 6 2004;27(1):141-7 7 8 Hanefeld 2007 9 Hanefeld M, Patwardhan R, Jones NP. A one-year study comparing the efficacy and safety of 10 rosiglitazone and glibenclamide in the treatment of type 2 diabetes. Nutr Metab Cardiovasc Dis 11 2007;17(1):13-23 12 Confidential: For Review Only 13 HARMONY 1 14 Reusch J, Stewart MW, Perkins CM, et al. Efficacy and safety of once-weekly glucagon-like peptide 1 15 16 receptor agonist albiglutide (HARMONY 1 trial): 52-week primary endpoint results from a 17 randomized, double-blind, placebo-controlled trial in patients with type 2 diabetes mellitus not 18 controlled on pioglitazone, with or without metformin. Diabetes Obes Metab 2014;16(12):1257-64 19 20 HARMONY 2 21 Nauck MA, Stewart MW, Perkins C, et al. Efficacy and safety of once-weekly GLP-1 receptor agonist 22 albiglutide (HARMONY 2): 52 week primary endpoint results from a randomised, placebo-controlled 23 trial in patients with type 2 diabetes mellitus inadequately controlled with diet and exercise. 24 25 Diabetologia 2016;59(2):266-74 26 27 HARMONY 3 28 Ahren B, Johnson SL, Stewart M, et al. HARMONY 3: 104-week randomized, double-blind, placebo- 29 and active-controlled trial assessing the efficacy and safety of albiglutide compared with placebo, 30 sitagliptin, and glimepiride in patients with type 2 diabetes taking metformin. Diabetes Care 31 2014;37(8):2141-8 32 33 HARMONY 4 34 35 Weissman PN, Carr MC, Ye J, et al. HARMONY 4: randomised clinical trial comparing once-weekly 36 albiglutide and insulin glargine in patients with type 2 diabetes inadequately controlled with 37 metformin with or without sulfonylurea. Diabetologia 2014;57(12):2475-84 38 39 HARMONY 5 40 Home PD, Shamanna P, Stewart M, et al. Efficacy and tolerability of albiglutide versus placebo or 41 pioglitazone over 1year in people with type 2 diabetes currently taking metformin and glimepiride: 42 HARMONY 5. Diabetes Obes Metab 2015;17(2):179-87 43 44 45 HARMONY 6 46 Rosenstock J, Fonseca VA, Gross JL, et al. Advancing basal insulin replacement in type 2 diabetes 47 inadequately controlled with insulin glargine plus oral agents: A comparison of adding albiglutide, a 48 weekly GLP-1 receptor agonist, versus thrice-daily prandial insulin lispro. Diabetes Care 49 2014;37(8):2317-25 50 51 HARMONY OUTCOMES 52 53 Rosenberg AE, Sigmon KN, Somerville MC, et al. Albiglutide and cardiovascular outcomes in patients 54 with type 2 diabetes and cardiovascular disease (Harmony Outcomes): a double-blind, randomised 55 placebo-controlled trial. Lancet 2018;392(10157):1519-29 56 57 Hartemann-Heurtier 2009 58 59 60

https://mc.manuscriptcentral.com/bmj Page 81 of 244 BMJ

1 2 3 Hartemann-Heurtier A, Halbron M, Golmard JL, et al. Effects of bed-time insulin versus pioglitazone 4 on abdominal fat accumulation, inflammation and gene expression in adipose tissue in patients with 5 6 type 2 diabetes. Diabetes Res Clin Pract 2009;86(1):37-43 7 8 Hartley 2015 9 Hartley P, Shentu Y, Betz-Schiff P, et al. Efficacy and tolerability of sitagliptin compared with 10 glimepiride in elderly patients with type 2 diabetes mellitus and inadequate glycemic control: a 11 randomized, double-blind, non-inferiority trial. Drugs and Aging 2015;32(6):469-76 12 Confidential: For Review Only 13 Hasche 1999 14 Hasche H, Mertes G, Bruns C, et al. Effects of acarbose treatment in Type 2 diabetic patients under 15 16 dietary training: a multicentre, double-blind, placebo-controlled, 2-year study. Diabetes Nutr Metab 17 1999;12(4):277-85 18 19 Hattori 2017 20 Hattori A, Takemoto M, Tokuyama H, et al. Sitagliptin but not alpha glucosidase inhibitor reduced 21 the serum soluble CD163, a marker for activated macrophage, in individuals with type 2 diabetes 22 mellitus. Diabetes Res Clin Pract 2017;126:138-43 23 24 25 Hattori 2018 26 Hattori S. Anti-inflammatory effects of empagliflozin in patients with type 2 diabetes and insulin 27 resistance. Diabetol Metab Syndr 2018;10(1):93 28 29 HEELA study 30 Davies MJ, Donnelly R, Barnett AH, et al. Exenatide compared with long-acting insulin to achieve 31 glycaemic control with minimal weight gain in patients with type 2 diabetes: results of the Helping 32 Evaluate Exenatide in patients with diabetes compared with Long-Acting insulin (HEELA) study. 33 Diabetes Obes Metab 2009;11(12):1153-62 34 35 36 Hegele 1995 37 Hegele RA, Connelly PW, Palmason C, et al. Differential response of plasma lipoprotein(a) and 38 apolipoprotein B in NIDDM subjects treated with acarbose. Diabetes Care 1995;18(2):272-3 39 40 Heine 2005 41 Heine RJ, Gaal LF, Johns D, et al. Exenatide versus insulin glargine in patients with suboptimally 42 controlled type 2 diabetes: a randomized trial. Ann Intern Med 2005;143(8):559-69 43 44 45 Heliövaara 2007 46 Heliövaara MK, Herz M, Teppo AM, et al. Pioglitazone has anti-inflammatory effects in patients with 47 Type 2 diabetes. J Endocrinol Invest 2007;30(4):292-7 48 49 Henry 2012 50 Henry RR, Murray AV, Marmolejo MH, et al. Dapagliflozin, metformin XR, or both: initial 51 pharmacotherapy for type 2 diabetes, a randomised controlled trial. Int J Clin Pract 2012;66(5):446- 52 53 56 54 55 Henry 2014 56 Henry RR, Staels B, Fonseca VA, et al. Efficacy and safety of initial combination treatment with 57 sitagliptin and pioglitazone-a factorial study. Diabetes Obes Metab 2014;16(3):223-30 58 59 Hermann 1999 60

https://mc.manuscriptcentral.com/bmj BMJ Page 82 of 244

1 2 3 Hermann LS, Ranstam J, Vaaler S, et al. Effects of antihyperglycaemic therapies on proinsulin and 4 relation between proinsulin and cardiovascular risk factors in type 2 diabetes. Diabetes Obes Metab 5 6 1999;1(4):227-32 7 8 Hiramatsu 2018 9 Hiramatsu T, Asano Y, Mabuchi M, et al. Liraglutide relieves cardiac dilated function than DPP-4 10 inhibitors. Eur J Clin Invest 2018;48(10):e13007 11 12 Hirano 2009Confidential: For Review Only 13 Hirano M, Nakamura T, Kitta Y, et al. Rapid improvement of carotid plaque echogenicity within 1 14 month of pioglitazone treatment in patients with acute coronary syndrome. Atherosclerosis 15 16 2009;203(2):483-8 17 18 Hirano 2012 19 Hirano M, Nakamura T, Obata JE, et al. Early improvement in carotid plaque echogenicity by 20 acarbose in patients with acute coronary syndromes. Circ J 2012;76(6):1452-60 21 22 Hollander 2007 23 Hollander P, Yu D, Chou HS. Low-dose rosiglitazone in patients with insulin-requiring type 2 diabetes. 24 25 Arch Intern Med 2007;167(12):1284-90 26 27 HOME 28 Kooy A, Jager J, Lehert P, et al. Long-term effects of metformin on metabolism and microvascular 29 and macrovascular disease in patients with type 2 diabetes mellitus. Arch Intern Med 30 2009;169(6):616-25 31 32 Home 2018 33 Home P, Shankar RR, Gantz I, et al. A randomized, double-blind trial evaluating the efficacy and 34 35 safety of monotherapy with the once-weekly dipeptidyl peptidase-4 inhibitor omarigliptin in people 36 with type 2 diabetes. Diabetes Res Clin Pract 2018;138:253-61 37 38 Hong 2013 39 Hong SJ, Hyoung PJ, Lim DS. Effects of pioglitazone on brachial artery flow-mediated dilation and 40 circulating levels of microrna-21 in hypertensive type II diabetic patients. J Am Coll Cardiol 41 2013;61(10 Suppl 1):E2069 42 43 44 Hong 2015 45 Hong SJ, Choi SC, Cho JY, et al. Pioglitazone increases circulating microRNA-24 with decrease in 46 coronary neointimal hyperplasia in type 2 diabetic patients- optical coherence tomography analysis. 47 Circ J 2015;79(4):880-8 48 49 Hong 2016 50 Hong S, Park CY, Han KA, et al. Efficacy and safety of teneligliptin, a novel dipeptidyl peptidase-4 51 inhibitor, in Korean patients with type 2 diabetes mellitus: A 24-week multicentre, randomized, 52 53 double-blind, placebo-controlled phase III trial. Diabetes Obes Metab 2016;18(5):528-32 54 55 Horton 2000 56 Horton ES, Clinkingbeard C, Gatlin M, et al. Nateglinide alone and in combination with metformin 57 improves glycemic control by reducing mealtime glucose levels in type 2 diabetes. Diabetes Care 58 2000;23(11):1660-5 59 60

https://mc.manuscriptcentral.com/bmj Page 83 of 244 BMJ

1 2 3 Hotta 1993 4 Hotta N, Kakuta H, Sano T, et al. Long-term effect of acarbose on glycaemic control in non-insulin- 5 6 dependent diabetes mellitus: a placebo-controlled double-blind study. Diabetic Med 1993;10(2):134- 7 8 8 9 Hsieh 2011 10 Hsieh SH, Shih KC, Chou CW, et al. Evaluation of the efficacy and tolerability of miglitol in Chinese 11 patients with type 2 diabetes mellitus inadequately controlled by diet and sulfonylureas. Acta 12 DiabetologicaConfidential: 2011;48(1):71-7 For Review Only 13 14 Hu 2007 15 16 Hu ZY, Zhao JW. Effect of rosiglitazone on serum CRP and plasma PAI-1 in patients with early type 2 17 diabetic nephropathy. Central Plains Medical Journal 2007;34(20):27-8 18 19 Hwang 2008 20 Hwang YC, Lee EY, Lee WJ, et al. Effects of rosiglitazone on body fat distribution and insulin 21 sensitivity in Korean type 2 diabetes mellitus patients. Metab Clin Experiment 2008;57(4):479-87 22 23 Hygum 2020 24 25 Hygum K, Harslof T, Jorgensen NR, et al. Bone resorption is unchanged by liraglutide in type 2 26 diabetes patients: a randomised controlled trial. Bone 2020;132 27 28 Iacobellis 2017 29 Iacobellis G, Mohseni M, Bianco SD, et al. Liraglutide causes large and rapid epicardial fat reduction. 30 Obesity 2017;25(2):311-16 31 32 ILLUMINATE 33 Kashiwagi A, Kazuta K, Goto K, et al. Ipragliflozin in combination with metformin for the treatment of 34 35 Japanese patients with type 2 diabetes: ILLUMINATE, a randomized, double-blind, placebo- 36 controlled study. Diabetes Obes Metab 2015;17(3):304-08 37 38 Inagaki 2012 39 Inagaki N, Atsumi Y, Oura T, et al. Efficacy and safety profile of exenatide once weekly compared 40 with insulin once daily in Japanese patients with type 2 diabetes treated with oral antidiabetes 41 drug(s): results from a 26-week, randomized, open-label, parallel-group, multicenter, noninferiority 42 study. Clin Ther 2012;34(9):1892-908 43 44 45 Inagaki 2013 46 Inagaki N, Watada H, Murai M, et al. Linagliptin provides effective, well-tolerated add-on therapy to 47 pre-existing oral antidiabetic therapy over 1year in Japanese patients with type 2 diabetes. Diabetes 48 Obes Metab 2013;15(9):833-43 49 50 Inagaki 2014 51 Inagaki N, Kondo K, Yoshinari T, et al. Efficacy and safety of canagliflozin monotherapy in Japanese 52 53 patients with type 2 diabetes inadequately controlled with diet and exercise: A 24-week, 54 randomized, double-blind, placebo-controlled, Phase III study. Expert Opin Pharmacother 55 2014;15(11):1501-15 56 57 Inagaki 2015 58 59 60

https://mc.manuscriptcentral.com/bmj BMJ Page 84 of 244

1 2 3 Inagaki N, Onouchi H, Maezawa H, et al. Once-weekly versus daily alogliptin in Japanese 4 patients with type 2 diabetes: A randomised, double-blind, phase 3, non-inferiority study. Lancet 5 6 Diabetes Endocrinol 2015;3(3):191-97 7 8 INICOM 9 Lim S, Han KA, Yu J, et al. Efficacy and safety of initial combination therapy with gemigliptin and 10 metformin compared with monotherapy with either drug in patients with type 2 diabetes: A double- 11 blind randomized controlled trial (INICOM study). Diabetes Obes Metab 2017;19(1):87-97 12 Confidential: For Review Only 13 Inoue 2019 14 Inoue H, Morino K, Ugi S, et al. Ipragliflozin, a sodium-glucose cotransporter 2 inhibitor, reduces 15 16 bodyweight and fat mass, but not muscle mass, in Japanese type 2 diabetes patients treated with 17 insulin: a randomized clinical trial. J Diabetes Invest 2019;10(4):1012-21 18 19 Inoue 2020 20 Onoue T, Goto M, Wada E, et al. Dipeptidyl peptidase-4 inhibitor anagliptin reduces fasting 21 apolipoprotein B-48 levels in patients with type 2 diabetes: A randomized controlled trial. PLoS ONE 22 2020;15(1):e0228004 23 24 25 Insulin Glargine 4014 Study 26 Rosenstock J, Sugimoto D, Strange P, et al. Triple therapy in type 2 diabetes: insulin glargine or 27 rosiglitazone added to combination therapy of sulfonylurea plus metformin in insulin-naive patients. 28 Diabetes Care 2006;29(3):554-9 29 30 INTERVAL 31 Strain WD, Lukashevich V, Kothny W, et al. Individualised treatment targets for elderly patients with 32 type 2 diabetes using vildagliptin add-on or lone therapy (INTERVAL): a 24 week, randomised, 33 double-blind, placebo-controlled study. Lancet 2013;382(9890):409-16 34 35 36 Ito 2011 37 Ito M, Abe M, Okada K, et al. The dipeptidyl peptidase-4 (DPP-4) inhibitor vildagliptin improves 38 glycemic control in type 2 diabetic patients undergoing hemodialysis. Endocrine J 2011;58(11):979- 39 87 40 41 Ito 2017 42 Ito D, Shimizu S, Inoue K, et al. Comparison of ipragliflozin and pioglitazone effects on nonalcoholic 43 44 fatty liver disease in patients with type 2 diabetes: A randomized, 24-week, open-label, active- 45 controlled trial. Diabetes Care 2017;40(10):1364-72 46 47 Jabbour 2014 48 Jabbour SA, Hardy E, Sugg J, et al. Dapagliflozin is effective as add-on therapy to sitagliptin with or 49 withoutmetformin: A 24-Week, multicenter, randomized, double-blind, placebo-controlled study. 50 Diabetes Care 2014;37(3):740-50 51 52 53 Jacob 2007 54 Jacob AN, Salinas K, Adams-Huet B, et al. Weight gain in type 2 diabetes mellitus. Diabetes Obes 55 Metab 2007;9(3):386-93 56 57 Jain 2006 58 Jain R, Osei K, Kupfer S, et al. Long-term safety of pioglitazone versus glyburide in patients with 59 recently diagnosed type 2 diabetes mellitus. Pharmacotherapy 2006;26(10):1388-95 60

https://mc.manuscriptcentral.com/bmj Page 85 of 244 BMJ

1 2 3 4 JEDIS-1 5 6 Kawazu S, Kanazawa Y, Iwamoto Y, et al. Effect of antihyperglycemic drug monotherapy to prevent 7 the progression of mild hyperglycemia in early type 2 diabetic patients: the Japan Early Diabetes 8 Intervention Study (JEDIS). Diabetol Int 2017;8(4):350-65 9 10 Jeon 2011 11 Jeon HJ, Oh TK. Comparison of vildagliptin-metformin and glimepiride-metformin treatments in type 12 2 diabeticConfidential: patients. Diabetes Metab J 2011; 35For(5):529-35 Review Only 13 14 Jerums 1987 15 16 Jerums G, Murray RM, Seeman E, et al. Lack of effect of gliclazide on early diabetic nephropathy and 17 retinopathy: a two-year controlled study. Diabetes Res Clin Pract 1987;3(2):71-80 18 19 Ji 2014 20 Ji L, Ma J, Li H, et al. Dapagliflozin as monotherapy in drug-naive asian patients with type 2 diabetes 21 mellitus: A randomized, blinded, prospective phase III study. Clin Ther 2014;36(1):84-100.e9 22 23 Ji 2016 24 25 Ji L, Han P, Wang X, et al. Randomized clinical trial of the safety and efficacy of sitagliptin and 26 metformin co-administered to Chinese patients with type 2 diabetes mellitus. Journal of Diabetes 27 Investigation 2016;7(5):727-36 28 29 Ji 2017 30 Ji L, Li L, Kuang J, et al. Efficacy and safety of fixed-dose combination therapy, alogliptin plus 31 metformin, in Asian patients with type 2 diabetes: A phase 3 trial. Diabetes Obes Metab 32 2017;19(5):754-58 33 34 35 Jian 2018 36 Jian X, Yang QL, Xiao S, et al. The effects of a sodium-glucose cotransporter 2 inhibitor on diabetic 37 nephropathy and serum oxidized low-density lipoprotein levels. Eur Rev Med Pharmacol Sci 38 2018;22(12):3994-99 39 40 Jibran 2006 41 Jibran R, Suliman MI, Qureshi F, et al. Safety and efficacy of repaglinide compared with 42 glibenclamide in the management of type 2 diabetic Pakistani patients. Pak J Med Sci 43 44 2006;22(4):385-90 45 46 Johnston 1998 47 Johnston PS, Feig PU, Coniff RF, et al. Long-term titrated-dose alpha-glucosidase inhibition in non- 48 insulin-requiring Hispanic NIDDM patients. Diabetes Care 1998;21(3):409-15 49 50 Johnston 1998a 51 Johnston PS, Feig PU, Coniff RF, et al. Chronic treatment of African-American type 2 diabetic patients 52 53 with alpha-glucosidase inhibition. Diabetes Care 1998;21(3):416-22 54 55 Johnston 1998b 56 Johnston PS, Lebovitz HE, Coniff RF, et al. Advantages of alpha-glucosidase inhibition as 57 monotherapy in elderly type 2 diabetic patients. J Clin Endocrinol Metab 1998;83(5):1515-22 58 59 Josse 2003 60

https://mc.manuscriptcentral.com/bmj BMJ Page 86 of 244

1 2 3 Josse RG, Chiasson JL, Ryan EA, et al. Acarbose in the treatment of elderly patients with type 2 4 diabetes. Diabetes Res Clin Pract 2003;59(1):37-42 5 6 7 Jovanovic 2000 8 Jovanovic L, Dailey G, Huang WC, et al. Repaglinide in type 2 diabetes: a 24-week, fixed-dose efficacy 9 and safety study. J Clin Pharm 2000;40(1):49-57 10 11 Jovanovic 2004 12 JovanovicConfidential: L, Hassman DR, Gooch B, et al. Treatment For of Review type 2 diabetes with Only a combination regimen 13 of repaglinide plus pioglitazone. Diabetes Res Clin Pract 2004;63(2):127-34 14 15 16 Jung 2005 17 Jung HS, Youn BS, Cho YM, et al. The effects of rosiglitazone and metformin on the plasma 18 concentrations of resistin in patients with type 2 diabetes mellitus. Metabol Clin Exper 19 2005;54(3):314-20 20 21 Juurinen 2009 22 Juurinen L, Tiikkainen M, Saltevo J, et al. Nateglinide combination therapy with basal insulin and 23 metformin in patients with Type 2 diabetes. Diabetic Med 2009;26(4):409-15 24 25 26 Kadoglou 2007 27 Kadoglou NP, Iliadis F, Liapis CD, et al. Beneficial effects of combined treatment with rosiglitazone 28 and exercise on cardiovascular risk factors in patients with type 2 diabetes. Diabetes Care 29 2007;30(9):2242-4 30 31 Kadoglou 2008 32 Kadoglou NP, Iliadis F, Angelopoulou N, et al. Beneficial effects of rosiglitazone on novel 33 cardiovascular risk factors in patients with Type 2 diabetes mellitus. Diabet Med 2008;25(3):333-40 34 35 36 Kadoglou 2011 37 Kadoglou NP, Kapelouzou A, Tsanikidis H, et al. Effects of rosiglitazone/metformin fixed-dose 38 combination therapy and metformin monotherapy on serum vaspin, adiponectin and IL-6 levels in 39 drug-naïve patients with type 2 diabetes. Exp Clin Endocrinol Diabetes 2011;119(2):63-8 40 41 Kadowaki 2011 42 Kadowaki T, Namba M, Imaoka T, et al. Improved glycemic control and reduced bodyweight with 43 44 exenatide: A double-blind, randomized, phase 3 study in Japanese patients with suboptimally 45 controlled type 2 diabetes over 24 weeks. J Diabetes Invest 2011;2(3):210-7 46 47 Kadowaki 2017 48 Kadowaki T, Inagaki N, Kondo K, et al. Efficacy and safety of canagliflozin as add-on therapy to 49 teneligliptin in Japanese patients with type 2 diabetes mellitus: Results of a 24-week, randomized, 50 double-blind, placebo-controlled trial. Diabetes Obes Metab 2017;19(6):874-82 51 52 53 Kadowaki 2018 54 Kadowaki T, Inagaki N, Kondo K, et al. Efficacy and safety of teneligliptin added to canagliflozin 55 monotherapy in Japanese patients with type 2 diabetes mellitus: a multicentre, randomized, double- 56 blind, placebo-controlled, parallel-group comparative study. Diab Obes Metab 2018;20(2):453-57 57 58 Kaku 2009 59 60

https://mc.manuscriptcentral.com/bmj Page 87 of 244 BMJ

1 2 3 Kaku K. Efficacy and safety of therapy with metformin plus pioglitazone in the treatment of patients 4 with type 2 diabetes: a double-blind, placebo-controlled, clinical trial. Curr Med Res Opin 5 6 2009;25(5):1111-9 7 8 Kaku 2009a 9 Kaku K, Daida H, Kashiwagi A, et al. Long-term effects of pioglitazone in Japanese patients with type 10 2 diabetes without a recent history of macrovascular morbidity. Curr Med Res Opin 11 2009;25(12):2925-32 12 Confidential: For Review Only 13 Kaku 2011 14 Kaku K, Rasmussen MF, Nishida T, et al. Fifty-two-week, randomized, multicenter trial to compare 15 16 the safety and efficacy of the novel glucagon-like peptide-1 analog liraglutide vs glibenclamide in 17 patients with type2 diabetes. J Diabetes Invest 2011;2(6):441-7 18 19 Kaku 2014 20 Kaku K, Kiyosue A, Inoue S, et al. Efficacy and safety of dapagliflozin monotherapy in Japanese 21 patients with type 2 diabetes inadequately controlled by diet and exercise. Diabetes Obes Metab 22 2014;16(11):1102-10 23 24 25 Kaku 2017 26 Kaku K, Sumino S, Katou M, et al. Randomized, double-blind, phase III study to evaluate the efficacy 27 and safety of once-daily treatment with alogliptin and metformin hydrochloride in Japanese patients 28 with type 2 diabetes. Diabetes Obes Metab 2017;19(3):463-67 29 30 Kaku 2019 31 Kaku K, Araki E, Tanizawa Y, et al. Superior efficacy with a fixed-ratio combination of insulin degludec 32 and liraglutide (IDegLira) compared with insulin degludec and liraglutide in insulin-naive Japanese 33 patients with type 2 diabetes in a phase 3, open-label, randomized trial. Diab Obes Metab 34 35 2019;21(12):2674-83 36 37 Kaku 2019a 38 Kaku K, Haneda M, Tanaka Y, et al. Linagliptin as add-on to empagliflozin in a fixed-dose combination 39 in Japanese patients with type 2 diabetes: glycaemic efficacy and safety profile in a two-part, 40 randomized, placebo-controlled trial. Diab Obes Metab 2018;21(1):136-45 41 42 Kanazawa 2010 43 44 Kanazawa I, Yamaguchi T, Yano S, et al. Baseline atherosclerosis parameter could assess the risk of 45 bone loss during pioglitazone treatment in type 2 diabetes mellitus. Osteoporos Int 46 2010;21(12):2013-8 47 48 Kanazawa 2011 49 Kanazawa I, Yamamoto M, Yamaguchi T, et al. Effects of metformin and pioglitazone on serum 50 pentosidine levels in type 2 diabetes mellitus. Exp Clin Endocrinol Diabetes 2011;119(6):362-5 51 52 53 Kaneto 2020 54 Kaneto H, Takami A, Spranger R, et al. Efficacy and safety of insulin glargine/lixisenatide fixed-ratio 55 combination (iGlarLixi) in Japanese patients with type 2 diabetes mellitus inadequately controlled on 56 basal insulin and oral antidiabetic drugs: the LixiLan JP-L Randomized Clinical Trial. Diab Obes Metab 57 2020 58 59 Kato 2015 60

https://mc.manuscriptcentral.com/bmj BMJ Page 88 of 244

1 2 3 Kato H, Nagai Y, Ohta A, et al. Effect of sitagliptin on intrahepatic lipid content and body fat in 4 patients with type 2 diabetes. Diabetes Res Clin Pract 2015;109(1):199-205 5 6 7 Kawamori 2018 8 Kawamori R, Haneda M, Suzaki K, et al. Empagliflozin as add-on to linagliptin in a fixed-dose 9 combination in Japanese patients with type 2 diabetes: glycaemic efficacy and safety profile in a 52- 10 week, randomized, placebo-controlled trial. Diab Obes Metab 2018;20(9):2200-09 11 12 Kelly 2007Confidential: For Review Only 13 Kelly AS, Thelen AM, Kaiser DR, et al. Rosiglitazone improves endothelial function and inflammation 14 but not asymmetric dimethylarginine or oxidative stress in patients with type 2 diabetes mellitus. 15 16 Vasc Med 2007;12(4):311-8 17 18 Kendall 2005 19 Kendall DM, Riddle MC, Rosenstock J, et al. Effects of exenatide (exendin-4) on glycemic control over 20 30 weeks in patients with type 2 diabetes treated with metformin and a sulfonylurea. Diabetes Care 21 2005;28(5):1083-91 22 23 Khaloo 2019 24 25 Khaloo P, Asadi Komeleh S, Alemi H, et al. Sitagliptin vs. pioglitazone as add-on treatments in 26 patients with uncontrolled type 2 diabetes on the maximal dose of metformin plus sulfonylurea. J 27 Endocrinol Invest 2019;42(7):851-57 28 29 Khanolkar 2008 30 Khanolkar MP, Morris RH, Thomas AW, et al. Rosiglitazone produces a greater reduction in 31 circulating platelet activity compared with gliclazide in patients with type 2 diabetes mellitus--an 32 effect probably mediated by direct platelet PPARgamma activation. Atherosclerosis 33 2008;197(2):718-24 34 35 36 Kikuchi 2012 37 Kikuchi M, Kaku K, Odawara M, et al. Efficacy and tolerability of rosiglitazone and pioglitazone in 38 drug-naïve Japanese patients with type 2 diabetes mellitus: a double-blind, 28 weeks' treatment, 39 comparative study. Curr Med Res Opin 2012;28(6):1007-16 40 41 Kim 2014 42 Kim SG, Kim DM, Woo JT, et al. Efficacy and safety of monotherapy in patients with 43 44 type 2 diabetes mellitus over 24-weeks: a multicenter, randomized, double-blind, parallel-group, 45 placebo controlled trial. PLoS ONE 2014;9(4):e92843 46 47 Kim Jeong 2019 48 Kim Jeong M, Kim Sang S, Kim Jong H, et al. Efficacy and safety of pioglitazone versus glimepiride 49 after metformin and alogliptin combination therapy: A randomized, open-label, multicenter, 50 parallel-controlled study. Diabetes Metab J 2019 51 52 53 KIND-KM 54 Tanaka K, Saisho Y, Kawai T, et al. Efficacy and safety of liraglutide monotherapy compared with 55 metformin in Japanese overweight/obese patients with type 2 diabetes. Endocrine J 2015;62(5):399- 56 409 57 58 Kiyici 2009 59 60

https://mc.manuscriptcentral.com/bmj Page 89 of 244 BMJ

1 2 3 Kiyici S, Ersoy C, Kaderli A, et al. Effect of rosiglitazone, metformin and medical nutrition treatment 4 on arterial stiffness, serum MMP-9 and MCP-1 levels in drug naive type 2 diabetic patients. Diabetes 5 6 Res Clin Pract 2009;86(1):44-50 7 8 Ko 2006 9 Ko GT, Tsang PC, Wai HP, et al. Rosiglitazone versus bedtime insulin in the treatment of patients with 10 conventional oral antidiabetic drug failure: a 1-year randomized clinical trial. Adv Ther 11 2006;23(5):799-808 12 Confidential: For Review Only 13 Koffert 2017 14 Koffert JP, Mikkola K, Virtanen KA, et al. Metformin treatment significantly enhances intestinal 15 16 glucose uptake in patients with type 2 diabetes: results from a randomized clinical trial. Diabetes Res 17 Clin Pract 2017;131:208-16 18 19 Kohan 2014 20 Kohan DE, Fioretto P, Tang W, et al. Long-term study of patients with type 2 diabetes and moderate 21 renal impairment shows that dapagliflozin reduces weight and blood pressure but does not improve 22 glycemic control. Kidney Int 2014;85(4):962-71 23 24 25 Kondo 2016 26 Kondo Y, Harada N, Hamasaki A, et al. Sitagliptin monotherapy has better effect on insulinogenic 27 index than glimepiride monotherapy in Japanese patients with type 2 diabetes mellitus: A 52-week, 28 multicenter, parallel-group randomized controlled trial. Diabetol Metab Syndr 2016;8:15 29 30 Kothny 2013 31 Kothny W, Foley J, Kozlovski P, et al. Improved glycaemic control with vildagliptin added to insulin, 32 with or without metformin, in patients with type 2 diabetes mellitus. Diabetes Obes Metab 33 2013;15(3):252-7 34 35 36 Krawcyk 2005 37 Krawczyk A, Gajer G, Grzeszczak W, et al. The impact of metformin on metabolic control in obese 38 type 2 diabetic patients treated with insulin. Preliminary observation. Polish. Diabetologia 39 Doswiadczalna i Kliniczna 2005;5(1):47-51 40 41 Kudo-Fujimaki 2014 42 Kudo-Fujimaki K, Hirose T, Yoshihara T, et al. Efficacy and safety of nateglinide plus vildagliptin 43 44 combination therapy compared with switching to vildagliptin in type 2 diabetes patients 45 inadequately controlled with nateglinide. J Diabetes Invest 2014;5(4):400-09 46 47 Laberge 2016 48 Laberge A, Brassard P, Arsenault B, et al. Positive effect of the PPAR-gamma agonist rosiglitazone on 49 hemodynamic response to exercise in type 2 diabetic men after coronary artery bypass graft 50 surgery: A 1-yr randomized study. Can J Cardiol 2016;32(10 Supplement 1):S223-S24 51 52 53 Lam 1998 54 Lam KS, Tiu SC, Tsang MW, et al. Acarbose in NIDDM patients with poor control on conventional oral 55 agents. A 24-week placebo-controlled study. Diabetes Care 1998;21(7):1154-8 56 57 Lambadiari 2018 58 59 60

https://mc.manuscriptcentral.com/bmj BMJ Page 90 of 244

1 2 3 Lambadiari V, Pavlidis G, Kousathana F, et al. Effects of 6-month treatment with the glucagon like 4 peptide-1 analogue liraglutide on arterial stiffness, left ventricular myocardial deformation and 5 6 oxidative stress in subjects with newly diagnosed type 2 diabetes. Cardiovasc Diabetol 2018;17(1):8 7 8 Langenfeld 2005 9 Langenfeld MR, Forst T, Hohberg C, et al. Pioglitazone decreases carotid intima-media thickness 10 independently of glycaemic control in patients with type 2 diabetes mellitus. Results from a 11 controlled randomized study. Circulation 2005;111:2525-31 12 Confidential: For Review Only 13 LANTERN 14 Kashiwagi A, Takahashi H, Ishikawa H, et al. A randomized, double-blind, placebo-controlled study on 15 16 long-term efficacy and safety of ipragliflozin treatment in patients with type 2 diabetes mellitus and 17 renal impairment: Results of the Long-Term ASP1941 Safety Evaluation in Patients with Type 2 18 Diabetes with Renal Impairment (LANTERN) study. Diab Obes Metab 2015;17(2):152-60 19 20 Lawrence 2004 21 Lawrence JM, Reid J, Taylor GJ, et al. Favorable effects of pioglitazone and metformin compared with 22 gliclazide on lipoprotein subfractions in overweight patients with early type 2 diabetes. Diabetes 23 Care 2004;27(1):41-6 24 25 26 LEAD-1 27 Marre M, Shaw J, Brändle M, et al. Liraglutide, a once-daily human GLP-1 analogue, added to a 28 sulphonylurea over 26 weeks produces greater improvements in glycaemic and weight control 29 compared with adding rosiglitazone or placebo in subjects with Type 2 diabetes (LEAD-1 SU). Diabet 30 Med 2009;26(3):268-78 31 32 LEAD-2 33 Nauck M, Frid A, Hermansen K, et al. Efficacy and safety comparison of liraglutide, glimepiride, and 34 35 placebo, all in combination with metformin, in type 2 diabetes. Diabetes Care 2009;32(1):84-90 36 37 LEAD-3 38 Garber A, Henry R, Ratner R, et al. Liraglutide versus glimepiride monotherapy for type 2 diabetes 39 (LEAD-3 Mono): a randomised, 52-week, phase III, double-blind, parallel-treatment trial. Lancet 40 2009;373(9662):473-81 41 42 LEAD-4 43 44 Zinman B, Gerich J, Buse JB, et al. Efficacy and safety of the human glucagon-like peptide-1 analog 45 liraglutide in combination with metformin and thiazolidinedione in patients with type 2 diabetes 46 (LEAD-4 Met+TZD). Diabetes Care 2009;32(7):1224-30 47 48 LEAD-5 49 Russell-Jones D, Vaag A, Schmitz O, et al. Liraglutide vs insulin glargine and placebo in combination 50 with metformin and sulfonylurea therapy in type 2 diabetes mellitus (LEAD-5 met+SU): a randomised 51 controlled trial. Diabetologia 2009;52(10):2046-55 52 53 54 LEADER 55 Marso SP, Daniels GH, Brown-Frandsen K, et al. Liraglutide and cardiovascular outcomes in type 2 56 diabetes. N Engl J Med 2016;375:311-22 57 58 Lebovitz 2001 59 60

https://mc.manuscriptcentral.com/bmj Page 91 of 244 BMJ

1 2 3 Lebovitz HE, Dole JF, Patwardhan R, et al. Rosiglitazone monotherapy is effective in patients with 4 type 2 diabetes. J Clin Endocrinol Metab 2001;86(1):280-8 5 6 7 Ledesma 2019 8 Ledesma G, Umpierrez GE, Morley JE, et al. Efficacy and safety of linagliptin to improve glucose 9 control in older people with type 2 diabetes on stable insulin therapy: a randomized trial. Diabetes 10 Obes Metab 2019 11 12 Lee 2013Confidential: For Review Only 13 Lee HW, Lee HC, Kim BW, et al. Effects of low dose pioglitazone on restenosis and coronary 14 atherosclerosis in diabetic patients undergoing drug eluting stent implantation. Yonsei Med J 15 16 2013;54(6):1313-20 17 18 Lee 2017 19 Lee SH, Gantz I, Round E, et al. A randomized, placebo-controlled clinical trial evaluating the safety 20 and efficacy of the once-weekly DPP-4 inhibitor omarigliptin in patients with type 2 diabetes mellitus 21 inadequately controlled by glimepiride and metformin. BMC Endocr Disord 2017;17(1):70 22 23 Leiter 2014 24 25 Leiter LA, Carr MC, Stewart M, et al. Efficacy and safety of the once-weekly GLP-1 receptor agonist 26 albiglutide versus sitagliptin in patients with type 2 diabetes and renal impairment: A randomized 27 phase III study. Diabetes Care 2014;37(10):2723-30 28 29 Leiter 2014a 30 Leiter LA, Cefalu WT, De Bruin TWA, et al. Dapagliflozin added to usual care in individuals with type 2 31 diabetes mellitus with preexisting cardiovascular disease: A 24-week, multicenter, randomized, 32 double-blind, placebo-controlled study with a 28-week extension. J Am Geriatr Soc 2014;62(7):1252- 33 62 34 35 36 Leonhardt 1991 37 Leonhardt W, Hanefeld M, Fischer S, et al. Beneficial effects on serum lipids in noninsulin dependent 38 diabetics by acarbose treatment. Arzneimittel-Forschung 1991;41(7):735-8 39 40 Lewin 2007 41 Lewin A, Lipetz R, Wu J, et al. Comparison of extended-release metformin in combination with a 42 sulfonylurea (glyburide) to sulfonylurea monotherapy in adult patients with type 2 diabetes: a 43 44 multicenter, double-blind, randomized, controlled, phase III study. Clin Ther 2007;29(5):844-55 45 46 Lewin 2015 47 Lewin A, DeFronzo RA, Patel S, et al. Initial combination of empagliflozin and linagliptin in subjects 48 with type 2 diabetes. Diabetes Care 2015;38(3):394-402 49 50 Li 2014 51 Li CJ, Yu Q, Yu P, et al. Efficacy and safety comparison of add-on therapy with liraglutide, saxagliptin 52 53 and vildagliptin, all in combination with current conventional oral hypoglycemic agents therapy in 54 poorly controlled Chinese type 2 diabetes. Exp Clin Endocrinol Diabetes 2014;122(8):469-76 55 56 Li 2014b 57 Li JL, Feng ZP, Li QF, et al. Insulin glargine effectively achieves glycemic control and improves insulin 58 resistance in patients with early type 2 diabetes that exhibit a high risk for cardiovascular disease. 59 Exp Ther Med 2014;8(1):147-52 60

https://mc.manuscriptcentral.com/bmj BMJ Page 92 of 244

1 2 3 4 Li 2017 5 6 Li F, Shen Y, Sumn R, et al. Effects of vildagliptin add-on insulin therapy on nocturnal glycemic 7 variations in uncontrolled type 2 diabetes. Diabetes Ther 2017;8(5):1111-22 8 9 Li 2019 10 Li J, Zhang P, Fan B, et al. The efficacy of saxagliptin in T2DM patients with non-alcoholic fatty liver 11 disease: preliminary data. Rev Assoc Med Bras 2019;65(1):33-37 12 Confidential: For Review Only 13 Li 2019a 14 Li H, Xu X, Wang J, et al. A randomized study to compared the effects of once-weekly dulaglutide 15 16 injection and once-daily glimepiride on glucose fluctuation of type 2 diabetes mellitus patients: a 26- 17 week follow-up. J Diabetes Res 2019;2019:6423987- 18 19 LIBRA 20 Kramer CK, Zinman B, Choi H, et al. The impact of chronic liraglutide therapy on glucagon secretion 21 in type 2 diabetes: Insight from the LIBRA trial. J Clin Endocrinol Metabol 2015; 100(10):3702-9 22 23 Lim 2017 24 25 Lim S, Kim KM, Kim SG, et al. Effects of lobeglitazone, a novel thiazolidinedione, on bone mineral 26 density in patients with type 2 diabetes mellitus over 52 weeks. Diabetes Metab J 2017;41(5):377-85 27 28 Lin 2003 29 Lin BJ, Wu HP, Huang HS, et al. Efficacy and tolerability of acarbose in Asian patients with type 2 30 diabetes inadequately controlled with diet and sulfonylureas. J Diabetes Complications 31 2003;17(4):179-85 32 33 Lindstrom 2000 34 35 Lindstrom J, Tuomilehto J, Spengler M. Acarbose treatment does not change the habitual diet of 36 patients with Type 2 diabetes mellitus. Diabetic Med 2000;17(1):20-5 37 38 Lingvay 2018 39 Lingvay I, Desouza CV, Lalic KS, et al. A 26-week randomized controlled trial of semaglutide once 40 daily versus liraglutide and placebo in patients with type 2 diabetes suboptimally controlled on diet 41 and exercise with or without metformin. Diabetes Care 2018;41(9):1926-37 42 43 44 LIPER2 45 Wägner AM, Miranda-Calderin G, Ugarte-Lopetegui MA, et al. Effect of liraglutide on physical 46 performance in type 2 diabetes: results of a randomized, double-blind, controlled trial (LIPER2). 47 Diabetes Metab 2019;45(3):268-75 48 49 LIRA-ADD2SGLT2i 50 Blonde L, Belousova L, Fainberg U, et al. Liraglutide as add-on to sodium-glucose co-transporter-2 51 inhibitors in patients with inadequately controlled type 2 diabetes: LIRA-ADD2SGLT2i, a 26-week, 52 53 randomized, double-blind, placebo-controlled trial. Diab Obes Metab 2020 54 55 Liraglutide-Detemir Study 56 DeVries JH, Bain SC, Rodbard HW, et al. Sequential intensification of metformin treatment in type 2 57 diabetes with liraglutide followed by randomized addition of basal insulin prompted by A1C targets. 58 Diabetes Care 2012;35(7):1446-54 59 60

https://mc.manuscriptcentral.com/bmj Page 93 of 244 BMJ

1 2 3 LIRA-RENAL 4 Davies MJ, Bain SC, Atkin SL, et al. Efficacy and safety of liraglutide versus placebo as add-on to 5 6 glucose-lowering therapy in patients with type 2 diabetes and moderate renal impairment (LIRA- 7 RENAL): A randomized clinical trial. Diabetes Care 2016;39(2):222-30 8 9 LIRA-SWITCH 10 Bailey TS, Takacs R, Tinahones FJ, et al. Efficacy and safety of switching from sitagliptin to liraglutide 11 in subjects with type 2 diabetes (LIRA-SWITCH): a randomized, double-blind, double-dummy, active- 12 controlledConfidential: 26-week trial. Diabetes Obes Metab For 2016; 18Review(12):1191-98 Only 13 14 Liu 2013 15 16 Liu SC, Chien KL, Wang CH, et al. Efficacy and safety of adding pioglitazone or sitagliptin to patients 17 with type 2 diabetes insufficiently controlled with metformin and a sulfonylurea. Endocr Pract 18 2013;19(6):980-88 19 20 Liu 2014 21 Liu B, Wang J, Wang G. Beneficial effects of pioglitazone on retardation of persistent atrial fibrillation 22 progression in diabetes mellitus patients. Int Heart J 2014; 55(6):499-505 23 24 25 Liu 2020 26 Liu L, Yan H, Xia M, et al. Efficacy of exenatide and insulin glargine on nonalcoholic fatty liver disease 27 in patients with type 2 diabetes mellitus. Diabetes Metabol Res Rev 2020 28 29 Liutkus 2010 30 Liutkus J, Rosas Guzman J, Norwood P, et al. A placebo-controlled trial of exenatide twice-daily 31 added to thiazolidinediones alone or in combination with metformin. Diabetes Obes Metab 32 2010;12(12):1058-65 33 34 35 LixiLan-G 36 Blonde L, Rosenstock J, Del Prato S, et al. Switching to iGlarLixi versus continuing daily or weekly 37 GLP-1 RA in type 2 diabetes inadequately controlled by GLP-1 RA and oral antihyperglycemic 38 therapy: the LixiLan-G randomized clinical trial. Diabetes Care 2019;42(11):2108-16 39 40 Lou 2020 41 Lou N, Wu X, Wang Y, et al. Liraglutide improves postprandial blood glucose, blood glucose levels, 42 insulin resistance and physical distribution of t2dm patients with ms. Int J Clin Experimental Med 43 44 2020;13(2):518-24 45 46 Lu 2016 47 Lu CH, Min KW, Chuang LM, et al. Efficacy, safety, and tolerability of ipragliflozin in Asian patients 48 with type 2 diabetes mellitus and inadequate glycemic control with metformin: Results of a phase 3 49 randomized, placebo-controlled, double-blind, multicenter trial. J Diabetes Invest 2016;7(3):366-73 50 51 Lukashevich 2011 52 53 Lukashevich V, Schweizer A, Shao Q, et al. Safety and efficacy of vildagliptin versus placebo in 54 patients with type 2 diabetes and moderate or severe renal impairment: a prospective 24-week 55 randomized placebo-controlled trial. Diabetes Obes Metab 2011;13(10):947-54 56 57 Lund 2009 58 59 60

https://mc.manuscriptcentral.com/bmj BMJ Page 94 of 244

1 2 3 Lund SS, Tarnow L, Frandsen M, et al. Combining insulin with metformin or an insulin secretagogue 4 in non-obese patients with type 2 diabetes: 12 month, randomised, double blind trial. BMJ 5 6 2009;339:b4324 7 8 Ma 2015 9 Ma L, Li J, Zhang L, et al. The effect of metformin and pioglitazone on glucagon of patients with 10 diabetes and metabolic syndrome. [Chinese]. Chinese Journal of Clinical Nutrition 2015;23(2):65-72 11 12 MacauleyConfidential: 2015 For Review Only 13 Macauley M, Hollingsworth KG, Smith FE, et al. Effect of vildagliptin on hepatic steatosis. J Clin 14 Endocrinol Metabol 2015; 100(4):1578-85 15 16 17 Madsbad 2011 18 Madsbad S, Kilhovd B, Lager I, et al. Comparison between repaglinide and glipizide in Type 2 diabetes 19 mellitus: a 1-year multicentre study. Diabetic Med 2001;18(5):395-401 20 21 Maffioli 2013 22 Maffioli P, Fogari E, D'Angelo A, et al. Ultrasonography modifications of visceral and subcutaneous 23 adipose tissue after pioglitazone or glibenclamide therapy combined with rosuvastatin in type 2 24 25 diabetic patients not well controlled by metformin. Eur J Gastroenterol Hepatol 2013;25(9):1113-22 26 27 MAGNA VICTORIA 28 Bizino MB, Jazet IM, Westenberg JJM, et al. Effect of liraglutide on cardiac function in patients with 29 type 2 diabetes mellitus: randomized placebo-controlled trial. Cardiovasc Diabetol 2019;18(1):55 30 31 Marbury 1999 32 Marbury T, Huang WC, Strange P, et al. Repaglinide versus glyburide: a one-year comparison trial. 33 Diabetes Res Clin Pract 1999;43(3):155-66 34 35 36 MARCH 37 Yang W, Shan Z, Tian H, et al. Acarbose compared with metformin as initial therapy in patients with 38 newly diagnosed type 2 diabetes: An open-label, non-inferiority randomised trial. Lancet Diabetes 39 Endocrinol 2014;2(1):46-55 40 41 Marena 1993 42 Marena S, Pagani A, Montegrosso G, et al. Comparison of miglitol and glibenclamide in non insulin- 43 44 dependent diabetic patients. G Ital Diabetol 1993;13(4):383-8 45 46 Mari 2008 47 Mari A, Scherbaum WA, Nilsson PM, et al. Characterization of the influence of vildagliptin on model- 48 assessed -cell function in patients with type 2 diabetes and mild hyperglycemia. J Clin Endocrinol 49 Metab 2008;93(1):103-9 50 51 MARLINA-T2DTM 52 53 Groop PH, Cooper ME, Perkovic V, et al. Dipeptidyl peptidase-4 inhibition with linagliptin and effects 54 on hyperglycaemia and albuminuria in patients with type 2 diabetes and renal dysfunction: Rationale 55 and design of the MARLINA-T2DTM trial. Diabetes Vasc Dis Res 2015;12(6):455-62 56 57 Marre 2002 58 59 60

https://mc.manuscriptcentral.com/bmj Page 95 of 244 BMJ

1 2 3 Marre M, Gaal L, Usadel KH, et al. Nateglinide improves glycaemic control when added to metformin 4 monotherapy: results of a randomized trial with type 2 diabetes patients. Diabetes Obes Metab 5 6 2002;4(3):177-86 7 8 MASTER 9 Mikada A, Narita T, Yokoyama H, et al. Effects of miglitol, sitagliptin, and initial combination therapy 10 with both on plasma responses to a mixed meal and visceral fat in over-weight Japanese 11 patients with type 2 diabetes. "the MASTER randomized, controlled trial". Diabetes Res Clin Pract 12 2014;106Confidential:(3):538-47 For Review Only 13 14 Mathieu 2015 15 16 Mathieu C, Shankar RR, Lorber D, et al. A randomized clinical trial to evaluate the efficacy and safety 17 of co-administration of sitagliptin with intensively titrated insulin glargine. Diabetes Ther 18 2015;6(2):127-42 19 20 Mathieu 2016 21 Mathieu C, Herrera Marmolejo M, Gonzalez Gonzalez JG, et al. Efficacy and safety of triple therapy 22 with dapagliflozin add-on to saxagliptin plus metformin over 52 weeks in patients with type 2 23 diabetes. Diabetes Obes Metab 2016;18(11):1134-37 24 25 26 Matthaei 2015 27 Matthaei S, Catrinoiu D, Celinski A, et al. Randomized, double-blind trial of triple therapy with 28 saxagliptin add-on to dapagliflozin plus metformin in patients with type 2 diabetes. Diabetes Care 29 2015;38(11):2018-24 30 31 Matthaei 2015a 32 Matthaei S, Bowering K, Rohwedder K, et al. Dapagliflozin improves glycemic control and reduces 33 body weight as add-on therapy to metformin plus sulfonylurea: a 24-week randomized, double-blind 34 35 clinical trial. Diabetes Care 2015;38(3):365-72 36 37 Matthews 2005 38 Matthews DR, Charbonnel BH, Hanefeld M, et al. Long-term therapy with addition of pioglitazone to 39 metformin compared with the addition of gliclazide to metformin in patients with type 2 diabetes: a 40 randomized, comparative study. Diabetes Metab Res Rev 2005;21(2):167-74 41 42 Matthews 2010 43 44 Matthews DR, Dejager S, Ahren B, et al. Vildagliptin add-on to metformin produces similar efficacy 45 and reduced hypoglycaemic risk compared with glimepiride, with no weight gain: results from a 2- 46 year study. Diabetes Obes Metab 2010;12(9):780-9 47 48 Mattoo 2005 49 Mattoo V, Eckland D, Widel M, et al. Metabolic effects of pioglitazone in combination with insulin in 50 patients with type 2 diabetes mellitus whose disease is not adequately controlled with insulin 51 therapy: results of a six-month, randomized, double-blind, prospective, multicenter, parallel-group 52 53 study. Clin Ther 2005;27(5):554-67 54 55 McCluskey 2004 56 McCluskey D, Touger MS, Melis R, et al. Results of a randomized, double-blind, placebo-controlled 57 study administering glimepiride to patients with type 2 diabetes mellitus inadequately controlled 58 with rosiglitazone monotherapy. Clin Ther 2004;26(11):1783-90 59 60

https://mc.manuscriptcentral.com/bmj BMJ Page 96 of 244

1 2 3 McGill 2013 4 McGill JB, Sloan L, Newman J, et al. Long-term efficacy and safety of linagliptin in patients with type 5 6 2 diabetes and severe renal impairment: a 1-year, randomized, double-blind, placebo-controlled 7 study. Diabetes Care 2013;36(2):237-44 8 9 McGuire 2010 10 McGuire DK, Abdullah SM, See R, et al. Randomized comparison of the effects of rosiglitazone vs. 11 placebo on peak integrated cardiovascular performance, cardiac structure, and function. Eur Heart J 12 2010;31(18):2262-70Confidential: For Review Only 13 14 MDI Liraglutide 15 16 Lind M, Hirsch IB, Tuomilehto J, et al. Liraglutide in people treated for type 2 diabetes with multiple 17 daily insulin injections: randomised clinical trial (MDI Liraglutide trial). BMJ 2015;351:h5364 18 19 Meneghini 2010 20 Meneghini LF, Traylor L, Schwartz SL. Improved glycemic control with insulin glargine versus 21 pioglitazone as add-on therapy to sulfonylurea or metformin in patients with uncontrolled type 2 22 diabetes mellitus. Endocr Pract 2010;16(4):588-99 23 24 25 Meneilly 2000 26 Meneilly GS, Ryan EA, Radziuk J, et al. Effect of acarbose on insulin sensitivity in elderly patients with 27 diabetes. Diabetes Care 2000;23(8):1162-7 28 29 Mita 2007 30 Mita T, Watada H, Shimizu T, et al. Nateglinide reduces carotid intima-media thickening in type 2 31 diabetic patients under good glycemic control. Arterioscler Thromb Vasc Biol 2007;27(11):2456-62 32 33 Mita 2019 34 35 Mita T, Hiyoshi T, Yoshii H, et al. The effect of linagliptin versus metformin treatment-related quality 36 of life in patients with type 2 diabetes mellitus. Diabetes Ther 2019;10(1):119-34 37 38 Mitrakou 1998 39 Mitrakou A, Tountas N, Raptis AE, et al. Long-term effectiveness of a new alpha-glucosidase inhibitor 40 (BAY m1099-miglitol) in insulin-treated type 2 diabetes mellitus. Diabet Med 1998;15(8):657-60 41 42 Miyagawa 2015 43 44 Miyagawa J, Odawara M, Takamura T, et al. Once-weekly glucagon-like peptide-1 receptor agonist 45 dulaglutide is non-inferior to once-daily liraglutide and superior to placebo in Japanese patients with 46 type 2 diabetes: a 26-week randomized phase III study. Diabetes Obes Metab 2015;17(10):974-83 47 48 Miyazaki 2002 49 Miyazaki Y, Matsuda M, DeFronzo RA. Dose-response effect of pioglitazone on insulin sensitivity and 50 insulin secretion in type 2 diabetes. Diabetes Care 2002;25(3):517-23 51 52 53 Mokta 2018 54 Mokta JK, Ramesh n, Sahai AK, et al. Comparison of safety and efficacy of glimepiride-metformin and 55 vildagliptin- metformin treatment in newly diagnosed type 2 diabetic patients. J Assoc Physicians 56 India 2018;66(8):30-35 57 58 Moretto 2008 59 60

https://mc.manuscriptcentral.com/bmj Page 97 of 244 BMJ

1 2 3 Moretto TJ, Milton DR, Ridge TD, et al. Efficacy and tolerability of exenatide monotherapy over 24 4 weeks in antidiabetic drug-naive patients with type 2 diabetes: a randomized, double-blind, placebo- 5 6 controlled, parallel-group study. Clin Ther 2008;30(8):1448-60 7 8 Moriwaki 2018 9 Moriwaki K, Takeuchi T, Fujimoto N, et al. Effect of sitagliptin on coronary flow reserve assessed by 10 magnetic resonance imaging in type 2 diabetic patients with coronary artery disease. Circ J 11 2018;82(8):2119-27 12 Confidential: For Review Only 13 Moses 2014 14 Moses RG, Kalra S, Brook D, et al. A randomized controlled trial of the efficacy and safety of 15 16 saxagliptin as add-on therapy in patients with type 2 diabetes and inadequate glycaemic control on 17 metformin plus a sulphonylurea. Diabetes Obes Metab 2014;16(5):443-50 18 19 Moses 2016 20 Moses RG, Round E, Shentu Y, et al. A randomized clinical trial evaluating the safety and efficacy of 21 sitagliptin added to the combination of sulfonylurea and metformin in patients with type 2 diabetes 22 mellitus and inadequate glycemic control. J Diabetes 2016;8(5):701-11 23 24 25 Mu 2017 26 Mu Y, Pan C, Fan B, et al. Efficacy and safety of linagliptin/metformin single-pill combination as initial 27 therapy in drug-naive Asian patients with type 2 diabetes. Diabetes Res Clin Pract 2017;124:48-56 28 29 Muller-Wieland 2018 30 Muller-Wieland D, Kellerer M, Cypryk K, et al. Efficacy and safety of dapagliflozin or dapagliflozin plus 31 saxagliptin versus glimepiride as add-on to metformin in patients with type 2 diabetes. Diabetes 32 Obes Metab 2018;20(11):2598-607 33 34 35 Nakamura 2001 36 Nakamura T, Ushiyama C, Osada S, et al. Pioglitazone reduces urinary podocyte excretion in type 2 37 diabetes patients with microalbuminuria. Metabolism 2001;50(10):1193-6 38 39 Nakamura 2004 40 Nakamura T, Matsuda T, Kawagoe Y, et al. Effect of pioglitazone on carotid intima-media thickness 41 and arterial stiffness in type 2 diabetic nephropathy patients. Metabolism 2004;53(10):1382-6 42 43 44 Nakamura 2006 45 Nakamura T, Sugaya T, Kawagoe Y, et al. Effect of pioglitazone on urinary liver-type fatty acid- 46 binding protein concentrations in diabetes patients with microalbuminuria. Diabetes Metab Res Rev 47 2006;22(5):385-9 48 49 Nar 2009 50 Nar A, Gedik O. The effect of metformin on leptin in obese patients with type 2 diabetes mellitus and 51 nonalcoholic fatty liver disease. Acta Diabetol 2009;46(2):113-8 52 53 54 Nathan 1988 55 Nathan DM, Roussell A, Godine JE. Glyburide or insulin for metabolic control in non-insulin- 56 dependent diabetes mellitus. A randomized, double-blind study. Ann Intern Med 1988;108(3):334- 57 40 58 59 Nauck 2007a 60

https://mc.manuscriptcentral.com/bmj BMJ Page 98 of 244

1 2 3 Nauck MA, Duran S, Kim D, et al. A comparison of twice-daily exenatide and biphasic insulin aspart in 4 patients with type 2 diabetes who were suboptimally controlled with sulfonylurea and metformin: a 5 6 non-inferiority study. Diabetologia 2007;50(2):259-67 7 8 Nauck 2007b 9 Nauck MA, Meininger G, Sheng D, et al. Efficacy and safety of the dipeptidyl peptidase-4 inhibitor, 10 sitagliptin, compared with the sulfonylurea, glipizide, in patients with type 2 diabetes inadequately 11 controlled on metformin alone: a randomized, double-blind, non-inferiority trial. Diabetes Obes 12 Metab 2007;Confidential:9(2):194-205 For Review Only 13 14 Nauck 2009 15 16 Nauck MA, Ellis GC, Fleck PR, et al. Efficacy and safety of adding the dipeptidyl peptidase-4 inhibitor 17 alogliptin to metformin therapy in patients with type 2 diabetes inadequately controlled with 18 metformin monotherapy: a multicentre, randomised, double-blind, placebo-controlled study. Int J 19 Clin Pract 2009;63(1):46-55 20 21 Nauck 2011 22 Nauck MA, Prato S, Meier JJ, et al. Dapagliflozin versus glipizide as add-on therapy in patients with 23 type 2 diabetes who have inadequate glycemic control with metformin: a randomized, 52-week, 24 25 double-blind, active-controlled noninferiority trial. Diabetes Care 2011;34(9):2015-22 26 27 Nauck 2016 28 Nauck MA, Di Domenico M, Patel S, et al. Linagliptin and pioglitazone combination therapy versus 29 monotherapy with linagliptin or pioglitazone: A randomised, double-blind, parallel-group, 30 multinational clinical trial. Diabetes Vasc Dis Res 2016;13(4):286-98 31 32 Neff 2016 33 Neff KJ, Tobin LM, Hogan AE, et al. The effect of low dose liraglutide on renal inflammation in type 2 34 35 diabetic kidney disease: A randomised controlled study. Diabet Med 2016;33(Suppl 1):64 36 37 Negro 2005 38 Negro R, Mangieri T, Dazzi D, et al. Rosiglitazone effects on blood pressure and metabolic 39 parameters in nondipper diabetic patients. Diabetes Res Clin Pract 2005;70(1):20-5 40 41 Ning 2016 42 Ning G, Li L, Ma J, et al. Vildagliptin as add-on therapy to insulin improves glycemic control without 43 44 increasing risk of hypoglycemia in Asian, predominantly Chinese, patients with type 2 diabetes 45 mellitus. J Diabetes 2016;8(3):345-53 46 47 Nino 2017 48 Nino A, Okuda I, Wilson TH, et al. Weekly glucagon-like peptide-1 receptor agonist albiglutide as 49 monotherapy improves glycemic parameters in Japanese patients with type 2 diabetes mellitus: a 50 randomized, double-blind, placebo-controlled study. J Diabetes Invest 2017;9(3):558-66 51 52 53 Nishimura 2016 54 Nishimura A, Usui S, Kumashiro N, et al. Efficacy and safety of repaglinide added to sitagliptin in 55 Japanese patients with type 2 diabetes: A randomized 24-week open-label clinical trial. Endocrine J 56 2016;63(12):1087-98 57 58 Nishio 2006 59 60

https://mc.manuscriptcentral.com/bmj Page 99 of 244 BMJ

1 2 3 Nishio K, Sakurai M, Kusuyama T, et al. A randomized comparison of pioglitazone to inhibit 4 restenosis after coronary stenting in patients with type 2 diabetes. Diabetes Care 2006;29(1):101-6 5 6 7 Nogueira 2014 8 Nogueira KC, Furtado M, Fukui RT, et al. Left ventricular diastolic function in patients with type 2 9 diabetes treated with a dipeptidyl peptidase-4 inhibitor- a pilot study. Diabetol Metab Syndr 10 2014;6(1) 11 12 Nowicki 2011Confidential: For Review Only 13 Nowicki M, Rychlik I, Haller H, et al. Long-term treatment with the dipeptidyl peptidase-4 inhibitor 14 saxagliptin in patients with type 2 diabetes mellitus and renal impairment: a randomised controlled 15 16 52-week efficacy and safety study. Int J Clin Pract 2011;65(12):1230-9 17 18 Ogasawara 2009 19 Ogasawara D, Shite J, Shinke T, et al. Pioglitazone reduces the necrotic-core component in coronary 20 plaque in association with enhanced plasma adiponectin in patients with type 2 diabetes mellitus. 21 Circ J 2009;73(2):343-51 22 23 Ohira 2014 24 25 Ohira M, Yamaguchi T, Saiki A, et al. Metformin reduces circulating malondialdehyde-modified low- 26 density lipoprotein in type 2 diabetes mellitus. Clin Invest Med 2014;37(4):E243-51 27 28 Ohira 2014a 29 Ohira M, Yamaguchi T, Saiki A, et al. Pioglitazone improves the cardio-ankle vascular index in 30 patients with type 2 diabetes mellitus treated with metformin. Diabetes Metab Syndr Obes 31 2014;7:313-19 32 33 Omarigliptin Protocol 020 34 35 Gantz I, Okamoto T, Ito Y, et al. A randomized, placebo- and sitagliptin-controlled trial of the safety 36 and efficacy of omarigliptin, a once-weekly dipeptidyl peptidase-4 inhibitor, in Japanese patients 37 with type 2 diabetes. Diabetes Obes Metab 2017;19(11):1602-09 38 39 Onuchin 2010 40 Onuchin SG, Elsukova OS, Solov'ev OV, et al. Capabilities of hypoglycemic therapy in women with 41 decompensated type 2 diabetes mellitus. Terapevticheski Arkhiv 2010;82(8):34-41 42 43 44 Osman 2004 45 Osman A, Otero J, Brizolara A, et al. Effect of rosiglitazone on restenosis after coronary stenting in 46 patients with type 2 diabetes. Am Heart J 2004;147(5):e23 47 48 Ovalle 2004 49 Ovalle F, Bell DS. Effect of rosiglitazone versus insulin on the pancreatic beta-cell function of subjects 50 with type 2 diabetes. Diabetes Care 2004;27(11):2585-9 51 52 53 Owens 2011 54 Owens DR, Swallow R, Dugi KA, et al. Efficacy and safety of linagliptin in persons with type 2 diabetes 55 inadequately controlled by a combination of metformin and sulphonylurea: a 24-week randomized 56 study. Diabetic Med 2011;28(11):1352-61 57 58 Oyama 2008 59 60

https://mc.manuscriptcentral.com/bmj BMJ Page 100 of 244

1 2 3 Oyama T, Saiki A, Endoh K, et al. Effect of acarbose, an alpha-glucosidase inhibitor, on serum 4 lipoprotein lipase mass levels and common carotid artery intima-media thickness in type 2 diabetes 5 6 mellitus treated by sulfonylurea. J Atheroscler Thromb 2008;15(3):154-9 7 8 Pagano 1995 9 Pagano G, Marena S, Corgiat-Mansin L, et al. Comparison of miglitol and glibenclamide in diet- 10 treated type 2 diabetic patients. Diabete Metab 1995;21(3):162-7 11 12 Pan 2008Confidential: For Review Only 13 Pan C, Yang W, Barona JP, et al. Comparison of vildagliptin and acarbose monotherapy in patients 14 with Type 2 diabetes: a 24-week, double-blind, randomized trial. Diabetic Med 2008;25(4):435-41 15 16 17 Pan 2012 18 Pan C, Xing X, Han P, et al. Efficacy and tolerability of vildagliptin as add-on therapy to metformin in 19 Chinese patients with type 2 diabetes mellitus. Diabetes Obes Metab 2012;14(8):737-44 20 21 Pan 2012a 22 Pan CY, Yang W, Tou C, et al. Efficacy and safety of saxagliptin in drug-naïve Asian patients with type 23 2 diabetes mellitus: a randomized controlled trial. Diabetes Metab Res Rev 2012;28(3):268-75 24 25 26 Pan 2015 27 Pan C, Lu J, Li X, et al. Efficacy and safety of retagliptin in Chinese patients with type 2 diabetes 28 mellitus. Diabetes 2015; 64:A307 29 30 Park 2011 31 Park JS, Cho MH, Nam JS, et al. Effect of pioglitazone on serum concentrations of osteoprotegerin in 32 patients with type 2 diabetes mellitus. Eur J Endocrinol 2011;164(1):69-74 33 34 35 Park 2014 36 Park CY, Kang JG, Chon S, et al. Comparison between the therapeutic effect of metformin, 37 glimepiride and their combination as an add-on treatment to insulin glargine in uncontrolled 38 patients with type 2 diabetes. PLoS ONE 2014;9(3):e88779 39 40 Park 2017 41 Park J, Park SW, Yoon KH, et al. Efficacy and safety of evogliptin monotherapy in patients with type 2 42 diabetes and moderately elevated glycated haemoglobin levels after diet and exercise. Diabetes 43 44 Obes Metab 2017;19(12):1681-87 45 46 Parmar Vinendra 2019 47 Parmar Vinendra M, Goswami Sunita S. Efficacy and safety of teneligliptin as add-on therapy to 48 conventional therapy in Indian patients with type 2 diabetes mellitus. Asian J Pharm Clin Res 49 2019;12(12):116-20 50 51 Parthan 2018 52 53 Parthan G, Bhansali S, Kurpad AV, et al. Effect of Linagliptin and Voglibose on metabolic profile in 54 patients with Type 2 Diabetes: a randomized, double-blind, placebo-controlled trial. BMC Pharmacol 55 Toxicol 2018;19:38 56 57 Patel 2013 58 59 60

https://mc.manuscriptcentral.com/bmj Page 101 of 244 BMJ

1 2 3 Kirkman MS, Shankar RR, Shankar S, et al. Treating postprandial hyperglycemia does not appear to 4 delay progression of early type 2 diabetes: the Early Diabetes Intervention Program. Diabetes Care 5 6 2006;29(9):2095-101 7 8 Pavithra 2019 9 Pavithra D, Praveen D, Chowdary PR, et al. A prospective randomized control study on effect of 10 liraglutide on cardiovascular outcomes in type II diabetes mellitus. Drug Invention Today 11 2019;12(11):2549-52 12 Confidential: For Review Only 13 Pavo 2003 14 Pavo I, Jermendy G, Varkonyi TT, et al. Effect of pioglitazone compared with metformin on glycemic 15 16 control and indicators of insulin sensitivity in recently diagnosed patients with type 2 diabetes. J Clin 17 Endocrinol Metab 2003;88(4):1637-45 18 19 Perez 2009 20 Perez A, Zhao Z, Jacks R, et al. Efficacy and safety of pioglitazone/metformin fixed-dose combination 21 therapy compared with pioglitazone and metformin monotherapy in treating patients with T2DM. 22 Curr Med Res Opin 2009;25(12):2915-23 23 24 25 PERISCOPE 26 Nissen SE, Nicholls SJ, Wolski K, et al. Comparison of pioglitazone vs glimepiride on progression of 27 coronary atherosclerosis in patients with type 2 diabetes: the PERISCOPE randomized controlled 28 trial. JAMA 2008;299(13):1561-73 29 30 Perriello 2006 31 Perriello G, Pampanelli S, Pietro C, et al. Comparison of glycaemic control over 1 year with 32 pioglitazone or gliclazide in patients with Type 2 diabetes. Diabetic Med 2006;23(3):246-52 33 34 35 Petrica 2009 36 Petrica L, Petrica M, Vlad A, et al. Nephro- and neuroprotective effects of rosiglitazone versus 37 glimepiride in normoalbuminuric patients with type 2 diabetes mellitus: a randomized controlled 38 trial. Wiener Klinische Wochenschrift 2009;121(23-24):765-75 39 40 Petrica 2011 41 Petrica L, Vlad A, Petrica M, et al. Pioglitazone delays proximal tubule dysfunction and improves 42 cerebral vessel endothelial dysfunction in normoalbuminuric people with type 2 diabetes mellitus. 43 44 Diabetes Res Clin Pract 2011;94(1):22-32 45 46 Philis-Tsimikas 2013 47 Philis-Tsimikas A, Del Prato S, Satman I, et al. Effect of insulin degludec versus sitagliptin in patients 48 with type 2 diabetes uncontrolled on oral antidiabetic agents. Diabetes Obes Metab 2013;15(8):760- 49 66 50 51 Phillips 2001 52 53 Phillips LS, Grunberger G, Miller E, et al. Once- and twice-daily dosing with rosiglitazone improves 54 glycemic control in patients with type 2 diabetes. Diabetes Care 2001;24(2):308-15 55 56 Phillips 2003 57 Phillips P, Karrasch J, Scott R, et al. Acarbose improves glycemic control in overweight type 2 diabetic 58 patients insufficiently treated with metformin. Diabetes Care 2003;26(2):269-73 59 60

https://mc.manuscriptcentral.com/bmj BMJ Page 102 of 244

1 2 3 Phrommintikul 2019 4 Phrommintikul A, Wongcharoen W, Kumfu S, et al. Effects of dapagliflozin vs vildagliptin on 5 6 cardiometabolic parameters in diabetic patients with coronary artery disease: a randomised study. 7 Br J Clin Pharmacol 2019;85(6):1337-47 8 9 PIOCOMB 10 Hanefeld M, Pfützner A, Forst T, et al. Double-blind, randomized, multicentre, and active comparator 11 controlled investigation of the effect of pioglitazone, metformin, and the combination of both on 12 cardiovascularConfidential: risk in patients with type 2 diabetes For receiving Review stable basal insulinOnly therapy: the 13 PIOCOMB study. Cardiovasc Diabetol 2011;10:65 14 15 16 PIOfix 17 Pfützner A, Schöndorf T, Tschöpe D, et al. PIOfix-study: effects of pioglitazone/metformin fixed 18 combination in comparison with a combination of metformin with glimepiride on diabetic 19 dyslipidemia. Diabetes Technol Ther 2011;13(6):637-43 20 21 Pioglitazone 001 study 22 Aronoff S, Rosenblatt S, Braithwaite S, et al. Pioglitazone hydrochloride monotherapy improves 23 glycemic control in the treatment of patients with type 2 diabetes: a 6-month randomized placebo- 24 25 controlled dose-response study. The Pioglitazone 001 Study Group. Diabetes Care 2000;23(11):1605- 26 11 27 28 PIONEER 29 Pfützner A, Marx N, Lübben G, et al. Improvement of cardiovascular risk markers by pioglitazone is 30 independent from glycemic control: results from the pioneer study. J Am Coll Cardiol 31 2005;45(12):1925-31 32 33 PIONEER 1 34 35 Aroda VR, Rosenstock J, Terauchi Y, et al. PIONEER 1: randomized clinical trial comparing the efficacy 36 and safety of oral semaglutide monotherapy with placebo in patients with type 2 diabetes. Diabetes 37 Care 2019;42(9):1724-32 38 39 PIONEER 2 40 Rodbard H, W., Rosenstock J, Canani LH, et al. Oral semaglutide versus empagliflozin in patients with 41 type 2 diabetes uncontrolled on metformin: the PIONEER 2 trial. Diabetes Care 2019;42(12):2272-81 42 43 44 PIONEER 3 45 Rosenstock J, Allison D, Birkenfeld AL, et al. Effect of additional oral semaglutide vs sitagliptin on 46 glycated hemoglobin in adults with type 2 diabetes uncontrolled with metformin alone or with 47 sulfonylurea: the PIONEER 3 randomized clinical trial. JAMA 2019;321(15):1466-80 48 49 PIONEER 4 50 Pratley R, Amod A, Hoff ST, et al. Oral semaglutide versus subcutaneous liraglutide and placebo in 51 type 2 diabetes (PIONEER 4): a randomised, double-blind, phase 3a trial. Lancet 52 53 2019;394(10192):39-50 54 55 PIONEER 5 56 Mosenzon O, Blicher TM, Rosenlund S, et al. Efficacy and safety of oral semaglutide in patients with 57 type 2 diabetes and moderate renal impairment (PIONEER 5): a placebo-controlled, randomised, 58 phase 3a trial. Lancet Diabetes Endocrinol 2019;7(7):515-27 59 60

https://mc.manuscriptcentral.com/bmj Page 103 of 244 BMJ

1 2 3 PIONEER 6 4 Husain M, Birkenfeld AL, Donsmark M, et al. Oral semaglutide and cardiovascular outcomes in 5 6 patients with type 2 diabetes. N Engl J Med 2019;381(9):841-51 7 8 PIONEER 7 9 Pieber TR, Bode B, Mertens A, et al. Efficacy and safety of oral semaglutide with flexible dose 10 adjustment versus sitagliptin in type 2 diabetes (PIONEER 7): a multicentre, open-label, randomised, 11 phase 3a trial. Lancet Diabetes Endocrinol 2019;7(7):528-39 12 Confidential: For Review Only 13 PIONEER 8 14 Zinman B, Aroda Vanita R, Buse John B, et al. Efficacy, safety and tolerability of oral semaglutide 15 16 versus placebo added to insulin with or without metformin in patients with type 2 diabetes: the 17 PIONEER 8 Trial. Diabetes Care 2019;42(12):2262-71 18 19 PIONEER 9 20 Katagiri H, Deenadayalan S, Navarria A, et al. Dose-response, efficacy, and safety of oral semaglutide 21 monotherapy in Japanese patients with type 2 diabetes (PIONEER 9): a 52-week, phase 2/3a, 22 randomised, controlled trial. Lancet Diabetes Endocrinol 2020;8(5):377-91 23 24 25 PioRAGE 26 Koyama H, Tanaka S, Monden M, et al. Comparison of effects of pioglitazone and glimepiride on 27 plasma soluble RAGE and RAGE expression in peripheral mononuclear cells in type 2 diabetes: 28 Randomized controlled trial (PioRAGE). Atherosclerosis 2014;234(2):329-34 29 30 PIRAMID 31 van der Meer RW, Rijzewijk LJ, Jong HW, et al. Pioglitazone improves cardiac function and alters 32 myocardial substrate metabolism without affecting cardiac triglyceride accumulation and high- 33 energy phosphate metabolism in patients with well-controlled type 2 diabetes mellitus. Circulation 34 35 2009;119(15):2069-77 36 37 Pistrosch 2012 38 Pistrosch F, Passauer J, Herbrig K, et al. Effect of thiazolidinedione treatment on proteinuria and 39 renal hemodynamic in type 2 diabetic patients with overt nephropathy. Horm Metab Res 40 2012;44(12):914-8 41 42 Pistrosch 2013 43 44 Pistrosch F, Köhler C, Schaper F, et al. Effects of insulin glargine versus metformin on glycemic 45 variability, microvascular and beta-cell function in early type 2 diabetes. Acta Diabetol 46 2013;50(4):587-95 47 48 Pi-Sunyer 2007 49 Pi-Sunyer FX, Schweizer A, Mills D, et al. Efficacy and tolerability of vildagliptin monotherapy in drug- 50 naïve patients with type 2 diabetes. Diabetes Res Clin Pract 2007;76(1):132-8 51 52 53 Pop-Busui 2009 54 Pop-Busui R, Oral E, Raffel D, et al. Impact of rosiglitazone and glyburide on nitrosative stress and 55 myocardial blood flow regulation in type 2 diabetes mellitus. Metabolism 2009;58(7):989-94 56 57 POPPS 58 59 60

https://mc.manuscriptcentral.com/bmj BMJ Page 104 of 244

1 2 3 Takagi T, Okura H, Kobayashi Y, et al. A prospective, multicenter, randomized trial to assess efficacy 4 of pioglitazone on in-stent neointimal suppression in type 2 diabetes: POPPS (Prevention of In-Stent 5 6 Neointimal Proliferation by Pioglitazone Study). JACC Cardiol Intv 2009;2(6):524-31 7 8 Pratley 2014 9 Pratley RE, Fleck P, Wilson C. Efficacy and safety of initial combination therapy with alogliptin plus 10 metformin versus either as monotherapy in drug-naive patients with type 2 diabetes: A randomized, 11 double-blind, 6-month study. Diabetes Obes Metab 2014;16(7):613-21 12 Confidential: For Review Only 13 PRESERVE-BETA 14 Gerich J, Raskin P, Jean-Louis L, et al. PRESERVE-beta: two-year efficacy and safety of initial 15 16 combination therapy with nateglinide or glyburide plus metformin. Diabetes Care 2005;28(9):2093-9 17 18 PREVENT-J 19 Hirukawa H, Hashiramoto M, Tanizawa Y, et al. Remission of hyperglycemia after withdrawal of oral 20 antidiabetic drugs in Japanese patients with early-stage type 2 diabetes. J Diabetes Invest 21 2018;9(5):1119-27 22 23 PRIDE 24 25 Tanaka A, Komukai S, Shibata Y, et al. Effect of pioglitazone on cardiometabolic profiles and safety in 26 patients with type 2 diabetes undergoing percutaneous coronary artery intervention: a prospective, 27 multicenter, randomized trial. Heart Vessel 2018;33(9):965-77 28 29 PRIME V 30 Koshizaka M, Ishikawa K, Ishibashi R, et al. Comparing the effects of ipragliflozin versus metformin 31 on visceral fat reduction and metabolic dysfunction in Japanese patients with type 2 diabetes 32 treated with sitagliptin: a prospective, multicentre, open-label, blinded-endpoint, randomized 33 controlled study (PRIME-V study). Diabetes Obes Metab 2019;21(8):1990-95 34 35 36 PRISMA 37 Genovese S, Passaro A, Brunetti P, et al. Pioglitazone randomised Italian study on metabolic 38 syndrome (PRISMA): Effect of pioglitazone withmetformin on HDL-C levels in type 2 diabetic 39 patients. J Endocrinol Invest 2013;36(8):606-16 40 41 PROActive 42 Dormandy JA, Charbonnel B, Eckland DJ, et al. Secondary prevention of macrovascular events in 43 44 patients with type 2 diabetes in the PROactive Study (PROspective pioglitAzone Clinical Trial In 45 macroVascular Events): a randomised controlled trial. Lancet 2005;366(9493):1279-89 46 47 PROLOGUE 48 Oyama J, Murohara T, Kitakaze M, et al. The effect of sitagliptin on carotid artery atherosclerosis in 49 type 2 diabetes: The PROLOGUE Randomized Controlled Trial. PLoS Med 2016;13(6):e1002051 50 51 QUARTER 52 53 Schernthaner G, Matthews DR, Charbonnel B, et al. Efficacy and safety of pioglitazone versus 54 metformin in patients with type 2 diabetes mellitus: a double-blind, randomized trial. J Clin 55 Endocrinol Metab 2004;89(12):6068-76 56 57 Quatraro 1986 58 59 60

https://mc.manuscriptcentral.com/bmj Page 105 of 244 BMJ

1 2 3 Quatraro A, Consoli G, Ceriello A, et al. Combined insulin and sulfonylurea therapy in non-insulin- 4 dependent diabetics with secondary failure to oral drugs: a one year follow-up. Diabetes Metab 5 6 1986;12(6):315-8 7 8 Rahman 2010 9 Rahman S, Ismail A-S, Ismail SB, et al. Effect of rosiglitazone and ramipril on macrovasculopathy in 10 patients with type 2 diabetes: Needs longer treatment and/or higher doses? Clin Pharmacol Adv App 11 2010;2(1):83-7 12 Confidential: For Review Only 13 Rahman 2011 14 Rahman IU, Malik SA, Bashir M, et al. Monotherapy with metformin or glimepiride and changes in 15 16 serum sialic acid in type 2 diabetes mellitus. Br J Diabetes Vasc Dis 2011;11(3):137-40 17 18 Raskin 2001 19 Raskin P, Rendell M, Riddle MC, et al. A randomized trial of rosiglitazone therapy in patients with 20 inadequately controlled insulin-treated type 2 diabetes. Diabetes Care 2001;24(7):1226-32 21 22 Raskin 2004 23 Raskin P, McGill J, Saad MF, et al. Combination therapy for type 2 diabetes: repaglinide plus 24 25 rosiglitazone. Diabetic Med 2004;21(4):329-35 26 27 Raskin 2009 28 Raskin P, Lewin A, Reinhardt R, et al. Twice-daily dosing of a repaglinide/metformin fixed-dose 29 combination tablet provides glycaemic control comparable to rosiglitazone/metformin tablet. 30 Diabetes Obes Metab 2009;11(9):865-73 31 32 Raz 2008 33 Raz I, Chen Y, Wu M, et al. Efficacy and safety of sitagliptin added to ongoing metformin therapy in 34 35 patients with type 2 diabetes. Curr Med Res Opin 2008;24(2):537-50 36 37 RECORD 38 Home PD, Pocock SJ, Beck-Nielsen H, et al. Rosiglitazone evaluated for cardiovascular outcomes in 39 oral agent combination therapy for type 2 diabetes (RECORD): a multicentre, randomised, open- 40 label trial. Lancet 2009;373(9681):2125-35 41 42 REGO-F 43 44 Terawaki Y, Iwaya C, Nomiyama T, et al. Efficacy and safety of a combination of an insulin 45 secretagogue and a dipeptidyl peptidase-4 inhibitor in Japanese patients with type 2 diabetes 46 mellitus; the repaglinide glucose oscillation study in Fukuoka (REGO-F). Diabetol Int 2020 47 48 RELEASE 49 de Boer Stefanie A, Heerspink Hiddo JL, Juarez Orozco Luis E, et al. Effect of linagliptin on pulse wave 50 velocity in early type 2 diabetes (RELEASE): a randomized, double-blind, controlled 26-week trial. 51 Diabetes Obes Metab 2017;19(8):1147-54 52 53 54 RESULT 55 Rosenstock J, Goldstein BJ, Vinik AI, et al. Effect of early addition of rosiglitazone to sulphonylurea 56 therapy in older type 2 diabetes patients (>60 years): the Rosiglitazone Early vs. SULphonylurea 57 Titration (RESULT) study. Diabetes Obes Metab 2006;8(1):49-57 58 59 REWIND 60

https://mc.manuscriptcentral.com/bmj BMJ Page 106 of 244

1 2 3 Gerstein HC, Colhoun HM, Dagenais GR, et al. Dulaglutide and cardiovascular outcomes in type 2 4 diabetes (REWIND): a double-blind, randomised placebo-controlled trial. Lancet 5 6 2019;394(10193):121-30 7 8 Reynolds 2002 9 Reynolds LR, Konz EC, Frederich RC, et al. Rosiglitazone amplifies the benefits of lifestyle 10 intervention measures in long-standing type 2 diabetes mellitus. Diabetes Obes Metab 11 2002;4(4):270-5 12 Confidential: For Review Only 13 Reynolds 2007 14 Reynolds LR, Kingsley FJ, Karounos DG, et al. Differential effects of rosiglitazone and insulin glargine 15 16 on inflammatory markers, glycemic control, and lipids in type 2 diabetes. Diabetes Res Clin Pract 17 2007;77(2):180-7 18 19 Ristic 2006 20 Ristic S, Collober-Maugeais C, Pecher E, et al. Comparison of nateglinide and gliclazide in 21 combination with metformin, for treatment of patients with Type 2 diabetes mellitus inadequately 22 controlled on maximum doses of metformin alone. Diabetic Med 2006;23(7):757-62 23 24 25 Robbins 2007 26 Robbins DC, Beisswenger PJ, Ceriello A, et al. Mealtime 50/50 basal + prandial insulin analogue 27 mixture with a basal insulin analogue, both plus metformin, in the achievement of target HbA1c and 28 pre- and postprandial blood glucose levels in patients with type 2 diabetes: a multinational, 24- 29 week, randomized, open-label, parallel-group comparison. Clin Ther 2007;29(11):2349-64 30 31 Roberts 2005 32 Roberts VL, Stewart J, Issa M, et al. Triple therapy with glimepiride in patients with type 2 diabetes 33 mellitus inadequately controlled by metformin and a thiazolidinedione: results of a 30-week, 34 35 randomized, double-blind, placebo-controlled, parallel-group study. Clin Ther 2005;27(10):1535-47 36 37 Rodbard 2016 38 Rodbard HW, Seufert J, Aggarwal N, et al. Efficacy and safety of titrated canagliflozin in patients with 39 type 2 diabetes mellitus inadequately controlled on metformin and sitagliptin. Diabetes Obes Metab 40 2016;18(8):812-19 41 42 Rosenstock 1998 43 44 Rosenstock J, Brown A, Fischer J, et al. Efficacy and safety of acarbose in metformin-treated patients 45 with type 2 diabetes. Diabetes Care 1998;21(12):2050-5 46 47 Rosenstock 2006 48 Rosenstock J, Rood J, Cobitz A, et al. Initial treatment with rosiglitazone/metformin fixed-dose 49 combination therapy compared with monotherapy with either rosiglitazone or metformin in patients 50 with uncontrolled type 2 diabetes. Diabetes Obes Metab 2006;8(6):650-60 51 52 53 Rosenstock 2007 54 Rosenstock J, Baron MA, Dejager S, et al. Comparison of vildagliptin and rosiglitazone monotherapy 55 in patients with type 2 diabetes: a 24-week, double-blind, randomized trial. Diabetes Care 56 2007;30(2):217-23 57 58 Rosenstock 2007a 59 60

https://mc.manuscriptcentral.com/bmj Page 107 of 244 BMJ

1 2 3 Rosenstock J, Kim SW, Baron MA, et al. Efficacy and tolerability of initial combination therapy with 4 vildagliptin and pioglitazone compared with component monotherapy in patients with type 2 5 6 diabetes. Diabetes Obes Metab 2007;9(2):175-85 7 8 Rosenstock 2009 9 Rosenstock J, Rendell MS, Gross JL, et al. Alogliptin added to insulin therapy in patients with type 2 10 diabetes reduces HbA(1C) without causing weight gain or increased hypoglycaemia. Diabetes Obes 11 Metab 2009;11(12):1145-52 12 Confidential: For Review Only 13 Rosenstock 2010 14 Rosenstock J, Inzucchi SE, Seufert J, et al. Initial combination therapy with alogliptin and pioglitazone 15 16 in drug-naïve patients with type 2 diabetes. Diabetes Care 2010;33(11):2406-8 17 18 Rosenstock 2012 19 Rosenstock J, Vico M, Wei L, et al. Effects of dapagliflozin, an SGLT2 inhibitor, on HbA(1c), body 20 weight, and hypoglycemia risk in patients with type 2 diabetes inadequately controlled on 21 pioglitazone monotherapy. Diabetes Care 2012;35(7):1473-8 22 23 Rosenstock 2013 24 25 Rosenstock J, Wilson C, Fleck P. Alogliptin versus glipizide monotherapy in elderly type 2 diabetes 26 mellitus patients with mild hyperglycaemia: A prospective, double-blind, randomized, 1-year study. 27 Diabetes Obes Metab 2013;15(10):906-14 28 29 Rosenstock 2015 30 Rosenstock J, Hansen L, Zee P, et al. Dual add-on therapy in type 2 diabetes poorly controlled with 31 metformin monotherapy: a randomized double-blind trial of saxagliptin plus dapagliflozin addition 32 versus single addition of saxagliptin or dapagliflozin to metformin. Diabetes Care 2015;38(3):376-83 33 34 35 Rosenstock 2016 36 Rosenstock J, Chuck L, Gonzalez-Ortiz M, et al. Initial combination therapy with canagliflozin plus 37 metformin versus each component as monotherapy for drug-naive type 2 diabetes. Diabetes Care 38 2016;39(3):353-62 39 40 Rosenstock 2019 41 Rosenstock J, Perl S, Johnsson E, et al. Triple therapy with low-dose dapagliflozin plus saxagliptin 42 versus dual therapy with each monocomponent, all added to metformin, in uncontrolled type 2 43 44 diabetes. Diabetes Obes Metab 2019;21(9):2152-62 45 46 Ross 2015 47 Ross SA, Caballero AE, Del Prato S, et al. Initial combination of linagliptin and metformin compared 48 with linagliptin monotherapy in patients with newly diagnosed type 2 diabetes and marked 49 hyperglycaemia: A randomized, double-blind, active-controlled, parallel group, multinational clinical 50 trial. Diabetes Obes Metab 2015;17(2):136-44 51 52 53 SAIS1 54 Furumoto T, Oba K, Tsutsui H, et al. A randomized controlled trial comparing the effects of sitagliptin 55 and glimepiride on endothelial function and metabolic parameters: Sapporo athero-incretin study 1 56 (SAIS1). PLoS ONE 2016;11(10):e0164255 57 58 Saleem 2011 59 60

https://mc.manuscriptcentral.com/bmj BMJ Page 108 of 244

1 2 3 Saleem K, Yasin MA, Asrar A, et al. Comparison of repaglinide with glibenclamide in the reduction of 4 HbA1C of type 2 diabetic patients. Pakistan J Med Health Sci 2011;5(1):23-6 5 6 7 Salman 2001 8 Salman S, Salman F, Satman I, et al. Comparison of acarbose and gliclazide as first-line agents in 9 patients with type 2 diabetes. Curr Med Res Opin 2001;16(4):296-306 10 11 Saloranta 2002 12 SalorantaConfidential: C, Hershon K, Ball M, et al. Efficacy For and safety Review of nateglinide in typeOnly 2 diabetic patients 13 with modest fasting hyperglycemia. J Clin Endocrinol Metab 2002;87(9):4171-6 14 15 16 Samson 2011 17 Samson SL, Sathyanarayana P, Jogi M, et al. Exenatide decreases hepatic fibroblast growth factor 21 18 resistance in non-alcoholic fatty liver disease in a mouse model of obesity and in a randomised 19 controlled trial. Diabetologia 2011;54(12):3093-100 20 21 Sato 2019 22 Sato A, Takei M, Hiramatsu K, et al. Effects of sitagliptin on pancreatic beta-cells in type 2 diabetes 23 with sulfonylurea treatment: A prospective randomized study. J Clin Med Res 2019;11(1):15-20 24 25 26 SAVOR-TIMI 53 27 Scirica BM, Bhatt DL, Braunwald E, et al. Saxagliptin and cardiovascular outcomes in patients with 28 type 2 diabetes mellitus. N Engl J Med 2013;369(14):1317-26 29 30 Saxagliptin 014 Study 31 DeFronzo RA, Hissa MN, Garber AJ, et al. The efficacy and safety of saxagliptin when added to 32 metformin therapy in patients with inadequately controlled type 2 diabetes with metformin alone. 33 Diabetes Care 2009;32(9):1649-55 34 35 36 SCALE Diabetes Randomized Clinical Trial 37 Davies MJ, Bergenstal R, Bode B, et al. Efficacy of liraglutide for weight loss among patients with type 38 2 diabetes: The SCALE diabetes randomized clinical trial. JAMA 2015;314(7):687-99 39 40 Scherbaum 2002 41 Scherbaum WA, Göke B. Metabolic efficacy and safety of once-daily pioglitazone monotherapy in 42 patients with type 2 diabetes: a double-blind, placebo-controlled study. Horm Metab Res 43 44 2002;34(10):589-95 45 46 Scherbaum 2008 47 Scherbaum WA, Schweizer A, Mari A, et al. Efficacy and tolerability of vildagliptin in drug-naive 48 patients with type 2 diabetes and mild hyperglycaemia. Diabetes Obes Metab 2008;10(8):675-82 49 50 Schweizer 2007 51 Schweizer A, Couturier A, Foley JE, et al. Comparison between vildagliptin and metformin to sustain 52 53 reductions in HbA1c over 1 year in drug-naive patients with Type 2 diabetes. Diabet Med 54 2007;24(9):955-61 55 56 Schweizer 2009 57 Schweizer A, Dejager S, Bosi E. Comparison of vildagliptin and metformin monotherapy in elderly 58 patients with type 2 diabetes: a 24-week, double-blind, randomized trial. Diabetes Obes Metab 59 2009;11(8):804-12 60

https://mc.manuscriptcentral.com/bmj Page 109 of 244 BMJ

1 2 3 4 Segal 1997 5 6 Segal P, Feig PU, Schernthaner G, et al. The efficacy and safety of miglitol therapy compared with 7 glibenclamide in patients with NIDDM inadequately controlled by diet alone. Diabetes Care 8 1997;20(5):687-91 9 10 Segal 2005 11 Segal P, Eliahou HE, Petzinna D, et al. Long-term efficacy and tolerability of acarbose treatment in 12 patients withConfidential: type 2 diabetes mellitus. Clin DrugFor Invest Review 2005;25(9):589-95 Only 13 14 Seino 2011 15 16 Seino Y, Rasmussen MF, Nishida T, et al. Glucagon-like peptide-1 analog liraglutide in combination 17 with sulfonylurea safely improves blood glucose measures vs sulfonylurea monotherapy in japanese 18 patients with type 2 diabetes: Results of a 52-week, randomized, multicenter trial. J Diabetes Invest 19 2011;2(4):280-86 20 21 Seino 2015 22 Seino Y, Inagaki N, Haneda M, et al. Efficacy and safety of luseogliflozin added to various oral 23 antidiabetic drugs in Japanese patients with type 2 diabetes mellitus. J Diabetes Invest 24 25 2015;6(4):443-53 26 27 Seino 2016 28 Seino Y, Kaneko S, Fukuda S, et al. Combination therapy with liraglutide and insulin in Japanese 29 patients with type 2 diabetes: A 36-week, randomized, double-blind, parallel-group trial. J Diabetes 30 Invest 2016;7(4):565-73 31 32 Seino 2018 33 Seino Y, Terauchi Y, Osonoi T, et al. Safety and efficacy of semaglutide once weekly vs sitagliptin 34 35 once daily, both as monotherapy in Japanese people with type 2 diabetes. Diabetes Obes Metab 36 2018;20(2):378-88 37 38 Shaddinger 2019 39 Shaddinger BC, Soffer J, Vlasakakis G, et al. Efficacy and safety of an albiglutide liquid formulation 40 compared with the lyophilized formulation: a 26-week randomized, double-blind, repeat-dose study 41 in patients with type 2 diabetes mellitus. Diabetes Res Clin Pract 2019;152:125-34 42 43 44 Shah 2011 45 Shah ZH, Saleem K, Mahboob F, et al. A comparative study of repaglinide and glibenclamide in type 2 46 diabetic patients. Pakistan Journal of Medical and Health Sciences 2011;5(3):476-9 47 48 Shalenko 2018 49 Shaienko ZO, Bobyreva LY. Combination of metformin and pioglitazone and its effect in treatment of 50 comorbid pathology. Wiad Lek 2018;71(2 pt 2):278-80 51 52 53 Shankar 2017 54 Shankar RR, Bao Y, Han P, et al. Sitagliptin added to stable insulin therapy with or without metformin 55 in Chinese patients with type 2 diabetes. J Diabetes Invest 2017;8(3):321-29 56 57 Shankar 2017a 58 Shankar RR, Inzucchi SE, Scarabello V, et al. A randomized clinical trial evaluating the efficacy and 59 safety of the once-weekly dipeptidyl peptidase-4 inhibitor omarigliptin in patients with type 2 60

https://mc.manuscriptcentral.com/bmj BMJ Page 110 of 244

1 2 3 diabetes inadequately controlled on metformin monotherapy. Curr Med Res Opin 2017;33(10):853- 4 1860 5 6 7 Shestakova 2018 8 Shestakova MV, Wilding JPH, Wilpshaar W, et al. A phase 3 randomized placebo-controlled trial to 9 assess the efficacy and safety of ipragliflozin as an add-on therapy to metformin in Russian patients 10 with inadequately controlled type 2 diabetes mellitus. Diabetes Res Clin Pract 2018;146:240-50 11 12 Shibuya 2018Confidential: For Review Only 13 Shibuya T, Fushimi N, Kawai M, et al. Luseogliflozin improves liver fat deposition compared to 14 metformin in type 2 diabetes patients with non-alcoholic fatty liver disease: a prospective 15 16 randomized controlled pilot study. Diabetes Obes Metab 2018;20(2):438-42 17 18 SIMPLE 19 Abreu M, Tumyan A, Elhassan A, et al. A randomized trial comparing the efficacy and safety of 20 treating patients with type 2 diabetes and highly elevated HbA1c levels with basal-bolus insulin or a 21 glucagon-like peptide-1 receptor agonist plus basal-bolus insulin: the SIMPLE study. Diabetes Obes 22 Metab 2019;21(9):2133-41 23 24 25 Sit2Mix Trial 26 Linjawi S, Sothiratnam R, Sari R, et al. The study of once- and twice-daily biphasic insulin aspart 30 27 (BIAsp 30) with sitagliptin, and twice-daily BIAsp 30 without sitagliptin, in patients with type 2 28 diabetes uncontrolled on sitagliptin and metformin - The Sit2Mix trial. Primary Care Diabetes 29 2015;9(5):370-76 30 31 Sitagliptin Study 019 32 Rosenstock J, Brazg R, Andryuk PJ, et al. Efficacy and safety of the dipeptidyl peptidase-4 inhibitor 33 sitagliptin added to ongoing pioglitazone therapy in patients with type 2 diabetes: a 24-week, 34 35 multicenter, randomized, double-blind, placebo-controlled, parallel-group study. Clin Ther 36 2006;28(10):1556-68 37 38 Sitagliptin Study 020 39 Charbonnel B, Karasik A, Liu J, et al. Efficacy and safety of the dipeptidyl peptidase-4 inhibitor 40 sitagliptin added to ongoing metformin therapy in patients with type 2 diabetes inadequately 41 controlled with metformin alone. Diabetes Care 2006;29(12):2638-43 42 43 44 Sitagliptin Study 021 45 Aschner P, Kipnes MS, Lunceford JK, et al. Effect of the dipeptidyl peptidase-4 inhibitor sitagliptin as 46 monotherapy on glycemic control in patients with type 2 diabetes. Diabetes Care 2006;29(12):2632- 47 7 48 49 Sitagliptin study 024 50 Seck T, Nauck M, Sheng D, et al. Safety and efficacy of treatment with sitagliptin or glipizide in 51 patients with type 2 diabetes inadequately controlled on metformin: a 2-year study. Int J Clin Pract 52 53 2010;64(5):562-76 54 55 Sitagliptin Study 035 56 Hermansen K, Kipnes M, Luo E, et al. Efficacy and safety of the dipeptidyl peptidase-4 inhibitor, 57 sitagliptin, in patients with type 2 diabetes mellitus inadequately controlled on glimepiride alone or 58 on glimepiride and metformin. Diabetes Obes Metab 2007;9(5):733-45 59 60

https://mc.manuscriptcentral.com/bmj Page 111 of 244 BMJ

1 2 3 Sitagliptin Study 036 4 Goldstein BJ, Feinglos MN, Lunceford JK, et al. Effect of initial combination therapy with sitagliptin, a 5 6 dipeptidyl peptidase-4 inhibitor, and metformin on glycemic control in patients with type 2 diabetes. 7 Diabetes Care 2007;30(8):1979-87 8 9 Sitagliptin Study 049 10 Aschner P, Katzeff HL, Guo H, et al. Efficacy and safety of monotherapy of sitagliptin compared with 11 metformin in patients with type 2 diabetes. Diabetes Obes Metab 2010;12(3):252-61 12 Confidential: For Review Only 13 Sitagliptin Study 051 14 Vilsbøll T, Rosenstock J, Yki-Järvinen H, et al. Efficacy and safety of sitagliptin when added to insulin 15 16 therapy in patients with type 2 diabetes. Diabetes Obes Metab 2010;12(2):167-77 17 18 SMART 19 Du J, Liang L, Fang H, et al. Efficacy and safety of saxagliptin compared with acarbose in Chinese 20 patients with type 2 diabetes mellitus uncontrolled on metformin monotherapy: results of a Phase 21 IV open-label randomized controlled study (the SMART study). Diabetes Obes Metab 22 2017;19(11):1513-20 23 24 25 Smith 2005 26 Smith SR, Jonge L, Volaufova J, et al. Effect of pioglitazone on body composition and energy 27 expenditure: a randomized controlled trial. Metabolism 2005;54(1):24-32 28 29 Softeland 2017 30 Softeland E, Meier JJ, Vangen B, et al. Empagliflozin as add-on therapy in patients with type 2 31 diabetes inadequately controlled with linagliptin and metformin: A 24-week randomized, double- 32 blind, parallel-group trial. Diabetes Care 2017;40(2):201-09 33 34 35 Sohn 2008 36 Sohn TS, Lee JI, Kim IJ, et al. The effect of rosiglitazone and metformin therapy, as an initial therapy, 37 in patients with type 2 diabetes mellitus. Korean Diabetes J 2008;32(5):445-52 38 39 Sone 2019 40 Sone H, Kaneko T, Shiki K, et al. Efficacy and safety of empagliflozin as add-on to insulin in Japanese 41 patients with type 2 diabetes: a randomised, double-blind, placebo-controlled trial. Diabetes Obes 42 Metab 2019 43 44 45 South Danish Diabetes Study 46 Gram J, Henriksen JE, Grodum E, et al. Pharmacological treatment of the pathogenetic defects in 47 type 2 diabetes: the randomized multicenter South Danish Diabetes Study. Diabetes Care 48 2011;34(1):27-33 49 50 SPEAD-A 51 Mita T, Katakami N, Yoshii H, et al. Alogliptin, a Dipeptidyl Peptidase 4 Inhibitor, Prevents the 52 53 Progression of Carotid Atherosclerosis in Patients with Type 2 Diabetes: The Study of Preventive 54 Effects of Alogliptin on Diabetic Atherosclerosis (SPEAD-A). Diabetes Care 2016;39(1):139-48 55 56 SPECIFY 57 Gu T, Ma J, Zhang Q, et al. Comparative effect of saxagliptin and glimepiride with a composite 58 endpoint of adequate glycaemic control without hypoglycaemia and without weight gain in patients 59 60

https://mc.manuscriptcentral.com/bmj BMJ Page 112 of 244

1 2 3 uncontrolled with metformin therapy: results from the SPECIFY study, a 48-week, multi-centre, 4 randomized, controlled trial. Diab Obes Metab 2019;21(4):939-48 5 6 7 Spengler 1989 8 Spengler N, Hansel G. Efficacy of 6 months monotherapy with glucosidase inhibitor acarbose versus 9 sulphonylurea glibenclamid on metabolic control of dietary treated type II diabetics abstract. Eur J 10 Clin Invest 1989;19((2 Pt II)):A71 11 12 SPIKE Confidential: For Review Only 13 Mita T, Katakami N, Shiraiwa T, et al. Sitagliptin attenuates the progression of carotid intima-media 14 thickening in insulin-treated patients with type 2 diabetes: The sitagliptin preventive study of intima- 15 16 media thickness evaluation (SPIKE): A randomized controlled trial. Diabetes Care 2016;39(3):455-64 17 18 SPOTLIGHT 19 Kashiwagi A, Shiga T, Akiyama N, et al. Efficacy and safety of ipragliflozin as an add-on to pioglitazone 20 in Japanese patients with inadequately controlled type 2 diabetes: a randomized, double-blind, 21 placebo-controlled study (the SPOTLIGHT study). Diabetol Int 2015;6(2):104-16 22 23 SPREAD-DIMCAD 24 25 Hong J, Zhang Y, Lai S, et al. Effects of metformin versus glipizide on cardiovascular outcomes in 26 patients with type 2 diabetes and coronary artery disease. Diabetes Care 2013;36(5):1304-11 27 28 Sridhar 2013 29 Sridhar S, Walia R, Sachdeva N, et al. Effect of pioglitazone on testosterone in eugonadal men with 30 type 2 diabetes mellitus: a randomized double-blind placebo-controlled study. Clin Endocrinol 31 2013;78(3):454-9 32 33 St. John Sutton 2002 34 35 St. John Sutton M, Rendell M, Dandona P, et al. A comparison of the effects of rosiglitazone and 36 glyburide on cardiovascular function and glycemic control in patients with type 2 diabetes. Diabetes 37 Care 2002;25(11):2058-64 38 39 Standl 1999 40 Standl E, Baumgartl HJ, Füchtenbusch M, et al. Effect of acarbose on additional insulin therapy in 41 type 2 diabetic patients with late failure of sulphonylurea therapy. Diabetes Obes Metab 42 1999;1(4):215-20 43 44 45 Standl 2001 46 Standl E, Schernthaner G, Rybka J, et al. Improved glycaemic control with miglitol in inadequately- 47 controlled type 2 diabetics. Diabetes Res Clin Pract 2001;51(3):205-13 48 49 START 50 Dou J, Ma J, Liu J, et al. Efficacy and safety of saxagliptin in combination with metformin as initial 51 therapy in Chinese patients with type 2 diabetes: results from the START study, a multicentre, 52 53 randomized, double-blind, active-controlled, phase 3 trial. Diab Obes Metab 2018;20(3):590-98 54 55 START-J 56 Terauchi Y, Yamada Y, Ishida H, et al. Efficacy and safety of sitagliptin as compared with glimepiride 57 in Japanese patients with type 2 diabetes mellitus aged>60years (START-J trial). Diabetes Obes 58 Metab 2017;19:1188-92 59 60

https://mc.manuscriptcentral.com/bmj Page 113 of 244 BMJ

1 2 3 Stewart 2006 4 Stewart MW, Cirkel DT, Furuseth K, et al. Effect of metformin plus roziglitazone compared with 5 6 metformin alone on glycaemic control in well-controlled Type 2 diabetes. Diabet Med 7 2006;23(10):1069-78 8 9 Stocker 2007 10 Stocker DJ, Taylor AJ, Langley RW, et al. A randomized trial of the effects of rosiglitazone and 11 metformin on inflammation and subclinical atherosclerosis in patients with type 2 diabetes. Am 12 Heart J 2007;Confidential:153(3):445.e1-6 For Review Only 13 14 Strojek 2009 15 16 Strojek K, Bebakar WM, Khutsoane DT, et al. Once-daily initiation with biphasic insulin aspart 30 17 versus insulin glargine in patients with type 2 diabetes inadequately controlled with oral drugs: an 18 open-label, multinational RCT. Curr Med Res Opin 2009;25(12):2887-94 19 20 Strojek 2011 21 Strojek K, Yoon KH, Hruba V, et al. Effect of dapagliflozin in patients with type 2 diabetes who have 22 inadequate glycaemic control with glimepiride: a randomized, 24-week, double-blind, placebo- 23 controlled trial. Diabetes Obes Metab 2011;13(10):928-38 24 25 26 Strom Halden 2019 27 Strom Halden TA, Kvitne KE, Midtvedt K, et al. Efficacy and safety of empagliflozin in renal transplant 28 recipients with posttransplant diabetes mellitus. Diabetes Care 2019;42(6):1067-74 29 30 Su 2014 31 Su Y, Su YL, Lv LF, et al. A randomized controlled clinical trial of vildagliptin plus metformin 32 combination therapy in patients with type II diabetes mellitus. Exp Ther Med 2014;7(4):799-803 33 34 35 SUCCESS 36 Yokoh H, Kobayashi K, Sato Y, et al. Efficacy and safety of the dipeptidyl peptidase-4 inhibitor 37 sitagliptin compared with alpha-glucosidase inhibitor in Japanese patients with type 2 diabetes 38 inadequately controlled on metformin or pioglitazone alone (Study for an Ultimate Combination 39 Therapy to Control Diabetes with Sitagliptin-1): A multicenter, randomized, open-label, non- 40 inferiority trial. Journal of Diabetes Investigation 2015;6(2):182-91 41 42 SUMER 43 44 Tamez-Perez HE, Grupo de Estudio Sumer. [Efficacy and safety of initial treatment with glimpeiride 45 versus sitagliptin in type 2 diabetes]. Revista Medica del Instituto Mexicano del Seguro Social 46 2015;53(2):142-8 47 48 Sun 2006 49 Sun Z, Ren X, Jin H, et al. Effect of pioglitazone on serum advanced glycosylation end product peptide 50 and monocyte chemoattractant protein-1 in type 2 diabetic patients. Jiangsu Medical Journal 51 2006;32(2):101-3 52 53 54 Sun 2016 55 Sun W, Zeng C, Liao L, et al. Comparison of acarbose and metformin therapy in newly diagnosed type 56 2 diabetic patients with overweight and/or obesity. Current Medical Research and Opinion 57 2016;32(8):1389-96 58 59 SUPER 60

https://mc.manuscriptcentral.com/bmj BMJ Page 114 of 244

1 2 3 Chen Y, Liu X, Li Q, et al. Saxagliptin add-on therapy in Chinese patients with type 2 diabetes 4 inadequately controlled by insulin with or without metformin: results from the SUPER study, a 5 6 randomized, double-blind, placebo-controlled trial. Diabetes Obes Metab 2018;20(4):1044-49 7 8 SUSTAIN 1 9 Sorli C, Harashima SI, Tsoukas GM, et al. Efficacy and safety of once-weekly semaglutide 10 monotherapy versus placebo in patients with type 2 diabetes (SUSTAIN 1): a double-blind, 11 randomised, placebo-controlled, parallel-group, multinational, multicentre phase 3a trial. Lancet 12 Diabetes Confidential:Endocrinol 2017;5(4):251-60 For Review Only 13 14 SUSTAIN 2 15 16 Ahrén B, Masmiquel L, Kumar H, et al. Efficacy and safety of once-weekly semaglutide versus once- 17 daily sitagliptin as an add-on to metformin, thiazolidinediones, or both, in patients with type 2 18 diabetes (SUSTAIN 2): A 56-week, double-blind, phase 3a, randomised trial. Lancet Diabetes 19 Endocrinol 2017;5(5):341-54 20 21 SUSTAIN 4 22 Aroda VR, Bain SC, Cariou B, et al. Efficacy and safety of once-weekly semaglutide versus once-daily 23 insulin glargine as add-on to metformin (with or without sulfonylureas) in insulin-naive patients with 24 25 type 2 diabetes (SUSTAIN 4): A randomised, open-label, parallel-group, multicentre, multinational, 26 phase 3a trial. Lancet Diabetes Endocrinol 2017;5(5):355-66 27 28 SUSTAIN 5 29 Rodbard HW, Lingvay I, Reed J, et al. Semaglutide Added to Basal Insulin in Type 2 Diabetes (SUSTAIN 30 5): a Randomized, Controlled Trial. J Clin Endocrinol Metabol 2018;103(6):2291-301 31 32 SUSTAIN 6 33 Marso SP, Bain SC, Consoli A, et al. Semaglutide and cardiovascular outcomes in patients with type 2 34 35 diabetes. N Engl J Med 2016;375(19):1834-44 36 37 SUSTAIN 8 38 Lingvay I, Catarig AM, Frias JP, et al. Efficacy and safety of once-weekly semaglutide versus daily 39 canagliflozin as add-on to metformin in patients with type 2 diabetes (SUSTAIN 8): a double-blind, 40 phase 3b, randomised controlled trial. Lancet Diabetes Endocrinol 2019;7(11):834-44 41 42 SUSTAIN 9 43 44 Zinman B, Bhosekar V, Busch R, et al. Semaglutide once weekly as add-on to SGLT-2 inhibitor therapy 45 in type 2 diabetes (SUSTAIN 9): a randomised, placebo-controlled trial. Lancet Diabetes Endocrinol 46 2019;7(5):356-67 47 48 Suzuki 2014 49 Suzuki K, Tanaka S, Aoki C, et al. Greater efficacy and improved endothelial dysfunction in untreated 50 type 2 diabetes with liraglutide versus sitagliptin. Dokkyo Journal of Medical Sciences 51 2014;41(3):211-20 52 53 54 SWIM 55 Kesavadev J, Pillai PBS, Shankar A, et al. Sitagliptin 100 mg vs glimepiride 1-3 mg as an add-on to 56 insulin and metformin in type 2 diabetes (SWIM). Endocrine connections 2017;6(8):748-57 57 58 Takagi 2003 59 60

https://mc.manuscriptcentral.com/bmj Page 115 of 244 BMJ

1 2 3 Takagi T, Yamamuro A, Tamita K, et al. Pioglitazone reduces neointimal tissue proliferation after 4 coronary stent implantation in patients with type 2 diabetes mellitus: an intravascular ultrasound 5 6 scanning study. Am Heart J 2003;146(2):E5 7 8 Takase 2007 9 Takase H, Nakazawa A, Yamashita S, et al. Pioglitazone produces rapid and persistent reduction of 10 vascular inflammation in patients with hypertension and type 2 diabetes mellitus who are receiving 11 angiotensin II receptor blockers. Metabolism 2007;56(4):559-64 12 Confidential: For Review Only 13 Takashima 2018 14 Takashima H, Yoshida Y, Nagura C, et al. Renoprotective effects of canagliflozin, a sodium glucose 15 16 cotransporter 2 inhibitor, in type 2 diabetes patients with chronic kidney disease: a randomized 17 open-label prospective trial. Diabetes Vasc Dis Res 2018;15(5):469-72 18 19 Tao 2018 20 Tao T, Wu P, Wang Y, et al. Comparison of glycemic control and β-cell function in new onset T2DM 21 patients with PCOS of metformin and saxagliptin monotherapy or combination treatment. BMC 22 Endocr Disord 2018;18(1):14 23 24 25 Taskinen 2011 26 Taskinen MR, Rosenstock J, Tamminen I, et al. Safety and efficacy of linagliptin as add-on therapy to 27 metformin in patients with type 2 diabetes: a randomized, double-blind, placebo-controlled study. 28 Diabetes Obes Metab 2011;13(1):65-74 29 30 TECOS 2015 31 Green JB, Bethel MA, Armstrong PW, et al. Effect of sitagliptin on cardiovascular outcomes in type 2 32 diabetes. N Engl J Med 2015;373(3):232-42 33 34 35 T-Emerge 1 36 Raz I, Fonseca V, Kipnes M, et al. Efficacy and safety of taspoglutide monotherapy in drug-naive type 37 2 diabetic patients after 24 weeks of treatment: results of a randomized, double-blind, placebo- 38 controlled phase 3 study (T-Emerge 1). Diabetes Care 2012;35(3):485-7 39 40 T-Emerge 3 41 Henry RR, Mudaliar S, Kanitra L, et al. Efficacy and safety of taspoglutide in patients with type 2 42 diabetes inadequately controlled with metformin plus pioglitazone over 24 weeks: T-Emerge 3 trial. J 43 44 Clin Endocrinol Metab 2012;97(7):2370-9 45 46 T-Emerge 4 47 Bergenstal RM, Forti A, Chiasson JL, et al. Efficacy and safety of taspoglutide versus sitagliptin for 48 type 2 diabetes mellitus (T-Emerge 4 Trial). Diabetes Ther 2012;3(1):1-19 49 50 T-Emerge 5 51 Nauck M, Horton E, Andjelkovic M, et al. Taspoglutide, a once-weekly glucagon-like peptide 1 52 53 analogue, vs. insulin glargine titrated to target in patients with Type 2 diabetes: an open-label 54 randomized trial. Diabetic Med 2013;30(1):109-13 55 56 T-Emerge 6 57 Pratley RE, Urosevic D, Boldrin M, et al. Efficacy and tolerability of taspoglutide versus pioglitazone in 58 subjects with type 2 diabetes uncontrolled with sulphonylurea or sulphonylurea-metformin therapy: 59 a randomized, double-blind study (T-emerge 6). Diabetes Obes Metab 2013;15(3):234-40 60

https://mc.manuscriptcentral.com/bmj BMJ Page 116 of 244

1 2 3 4 T-Emerge 7 5 6 Hollander P, Lasko B, Barnett AH, et al. Effects of taspoglutide on glycemic control and body weight 7 in obese patients with type 2 diabetes (T-Emerge 7 study). Obesity 2013;21(2):238-47 8 9 Teramoto 2007 10 Teramoto T, Yamada N, Shirai K, et al. Effects of pioglitazone hydrochloride on Japanese patients 11 with type 2 diabetes mellitus. J Atheroscler Thromb 2007;14(2):86-93 12 Confidential: For Review Only 13 Teupe 1991 14 Teupe B, Bergis K. Prospective randomized two-years clinical study comparing additional metformin 15 16 treatment with reducing diet in type 2 diabetes. Diabète Metab 1991;17(1 Pt 2):213-7 17 18 Thrasher 2014 19 Thrasher J, Daniels K, Patel S, et al. Efficacy and safety of linagliptin in black/African American 20 patients with type 2 diabetes: A 6-month, randomized, double-blind, placebo-controlled study. 21 Endocrine Pract 2014;20(5):412-20 22 23 TIDE 24 25 Punthakee Z, Bosch J, Dagenais G, et al. Design, history and results of the Thiazolidinedione 26 Intervention with vitamin D Evaluation (TIDE) randomised controlled trial. Diabetologia 27 2012;55(1):36-45 28 29 Tinahones 2017/Tinahones 2017a 30 Tinahones FJ, Gallwitz B, Nordaby M, et al. Linagliptin as add-on to empagliflozin and metformin in 31 patients with type 2 diabetes: two 24-week randomized, double-blind, double-dummy, parallel- 32 group trials. Diabetes Obes Metab 2017;19(2):266-74 33 34 35 Tofogliflozin 003 36 Kaku K, Watada H, Iwamoto Y, et al. Efficacy and safety of monotherapy with the novel 37 sodium/glucose cotransporter-2 inhibitor tofogliflozin in Japanese patients with type 2 diabetes 38 mellitus: A combined Phase 2 and 3 randomized, placebo-controlled, double-blind, parallel-group 39 comparative study. Cardiovasc Diabetol 2014;13(1):65 40 41 Tolman 2009 42 Tolman KG, Freston JW, Kupfer S, et al. Liver safety in patients with type 2 diabetes treated with 43 44 pioglitazone: results from a 3-year, randomized, comparator-controlled study in the US. Drug Safety 45 2009;32(9):787-800 46 47 TOPSCORE 48 Kato Y, Iwata A, Zhang B, et al. Effects of dipeptidyl peptidase-4 inhibitor sitagliptin on coronary 49 atherosclerosis as assessed by intravascular ultrasound in type 2 diabetes mellitus with coronary 50 artery disease. IJC Metab Endocr 2017;16:1-9 51 52 53 TOSCA.IT 54 Vaccaro O, Masulli M, Nicolucci A, et al. Effects on the incidence of cardiovascular events of the 55 addition of pioglitazone versus sulfonylureas in patients with type 2 diabetes inadequately 56 controlled with metformin (TOSCA.IT): a randomised, multicentre trial. Lancet Diabetes Endocrinol 57 2017;5(11):887-97 58 59 Tripathy 2013 60

https://mc.manuscriptcentral.com/bmj Page 117 of 244 BMJ

1 2 3 Tripathy D, Daniele G, Fiorentino TV, et al. Pioglitazone improves glucose metabolism and modulates 4 skeletal muscle TIMP-3-TACE dyad in type 2 diabetes mellitus: A randomised, double-blind, placebo- 5 6 controlled, mechanistic study. Diabetologia 2013;56(10):2153-63 7 8 TROICA 9 Ahn CH, Han KA, Yu JM, et al. Efficacy and safety of gemigliptin, a dipeptidyl peptidase-4 inhibitor, in 10 patients with type 2 diabetes mellitus inadequately controlled with combination treatment of 11 metformin and sulphonylurea: A 24-week, multicentre, randomized, double-blind, placebo- 12 controlledConfidential: study (TROICA study). Diabetes Obes For Metab Review 2017;19(5):635-43 Only 13 14 Truitt 2010 15 16 Truitt KE, Goldberg RB, Rosenstock J, et al. A 26-week, placebo- and pioglitazone-controlled, dose- 17 ranging study of rivoglitazone, a novel thiazolidinedione for the treatment of type 2 diabetes. Curr 18 Med Res Opin 2010;26(6):1321-31 19 20 Türkmen Kemal 2007 21 Türkmen Kemal Y, Güvener Demirag N, Yildirir A, et al. Effects of rosiglitazone on plasma brain 22 natriuretic peptide levels and myocardial performance index in patients with type 2 diabetes 23 mellitus. Acta Diabetologica 2007;44(3):149-56 24 25 26 UKPDS 44 27 Holman RR, Cull CA, Turner RC. A randomized double-blind trial of acarbose in type 2 diabetes shows 28 improved glycemic control over 3 years (U.K. Prospective Diabetes Study 44). Diabetes Care 29 1999;22(6):960-4 30 31 Umpierrez 2006 32 Umpierrez G, Issa M, Vlajnic A. Glimepiride versus pioglitazone combination therapy in subjects with 33 type 2 diabetes inadequately controlled on metformin monotherapy: results of a randomized clinical 34 35 trial. Curr Med Res Opin 2006;22(4):751-9 36 37 Umpierrez 2018 38 Umpierrez GE, Cardona S, Chachkhiani D, et al. A randomized controlled study comparing a DPP4 39 inhibitor (Linagliptin) and basal insulin (Glargine) in patients with type 2 diabetes in long-term care 40 and skilled nursing facilities: Linagliptin-LTC Trial. J Am Med Dir Assoc 2018;19(5):399-404 41 42 Vähätalo 2007 43 44 Vähätalo M, Rönnemaa T, Viikari J. Recognition of fasting or overall hyperglycaemia when starting 45 insulin treatment in patients with type 2 diabetes in general practice. Scand J Prim Health Care 46 2007;25(3):147-53 47 48 Van Eyk 2019 49 Van Eyk HJ, Paiman EHM, Bizino MB, et al. A double-blind, placebo-controlled, randomised trial to 50 assess the effect of liraglutide on ectopic fat accumulation in South Asian type 2 diabetes patients. 51 Cardiovasc Diabetol 2019;18(1):87 52 53 54 Van Gaal 2014 55 Van Gaal L, Souhami E, Zhou T, et al. Efficacy and safety of the glucagon-like peptide-1 receptor 56 agonist lixisenatide versus the dipeptidyl peptidase-4 inhibitor sitagliptin in young (<50 years) obese 57 patients with type 2 diabetes mellitus. J Clin Transl Endocrinol 2014;1(2):31-37 58 59 Vanderheiden 2016 60

https://mc.manuscriptcentral.com/bmj BMJ Page 118 of 244

1 2 3 Vanderheiden A, Harrison L, Warshauer J, et al. Effect of adding liraglutide vs placebo to a high-dose 4 lnsulin regimen in patients with type 2 diabetes a randomized clinical trial. JAMA Int Med 5 6 2016;176(7):939-47 7 8 Varghese 2009 9 Varghese A, Yee MS, Chan CF, et al. Effect of rosiglitazone on progression of atherosclerosis: insights 10 using 3D carotid cardiovascular magnetic resonance. J Cardiovasc Magn Reson 2009;11:24 11 12 Veleba 2015Confidential: For Review Only 13 Veleba J, Janovska P, Kuda O, et al. Combined intervention with pioglitazone and n-3 fatty acids in 14 metformin-treated type 2 diabetic patients: Improvement of lipid metabolism. Nutrition and 15 16 Metabolism 2015;12(1):52 17 18 VERTIS Asia 19 Ji L, Liu Y, Miao H, et al. Safety and efficacy of ertugliflozin in Asian patients with type 2 diabetes 20 mellitus inadequately controlled with metformin monotherapy: VERTIS Asia. Diabetes Obes Metab 21 2019;21(6):1474-82 22 23 VERTIS FACTORIAL 24 25 Pratley RE, Eldor R, Raji A, et al. Ertugliflozin plus sitagliptin versus either individual agent over 52 26 weeks in patients with type 2 diabetes mellitus inadequately controlled with metformin: the VERTIS 27 FACTORIAL randomized trial. Diabetes Obes Metab 2018;20(5):1111-20 28 29 VERTIS MET 30 Rosenstock J, Frias J, Páll D, et al. Effect of ertugliflozin on glucose control, body weight, blood 31 pressure and bone density in type 2 diabetes mellitus inadequately controlled on metformin 32 monotherapy (VERTIS MET). Diabetes Obes Metab 2018;20(3):520-29 33 34 35 VERTIS MONO 36 Terra SG, Focht K, Davies M, et al. Phase III, efficacy and safety study of ertugliflozin monotherapy in 37 people with type 2 diabetes mellitus inadequately controlled with diet and exercise alone. Diabetes 38 Obes Metab 2017;19(5):721-28 39 40 VERTIS RENAL 41 Grunberger G, Camp S, Johnson J, et al. Ertugliflozin in patients with Stage 3 chronic kidney disease 42 and type 2 diabetes mellitus: the VERTIS RENAL randomized study. Diabetes Ther 2018;9(1):49-66 43 44 45 VERTIS SITA2 46 Dagogo-Jack S, Liu J, Eldor R, et al. Efficacy and safety of the addition of ertugliflozin in patients with 47 type 2 diabetes mellitus inadequately controlled with metformin and sitagliptin: the VERTIS SITA2 48 placebo-controlled randomized study. Diabetes Obes Metab 2018;20(3):530-40 49 50 VERTIS SU 51 Hollander P, Liu J, Hill J, et al. Ertugliflozin Compared with Glimepiride in Patients with Type 2 52 53 Diabetes Mellitus Inadequately Controlled on Metformin: the VERTIS SU Randomized Study. 54 Diabetes Ther 2018;9(1):193-207 55 56 Vianna 2017 57 Vianna AGD, De Lacerda CS, Pechmann LM, et al. Vildagliptin has the same safety profile as a 58 sulfonylurea on bone metabolism and bone mineral density in post-menopausal women with type 2 59 diabetes: A randomized controlled trial NCT01679899 NCT. Diabetol Metab Syndr 2017;9(1):35 60

https://mc.manuscriptcentral.com/bmj Page 119 of 244 BMJ

1 2 3 4 VICTORY 5 6 Bertrand OF, Poirier P, Rodés-Cabau J, et al. Cardiometabolic effects of rosiglitazone in patients with 7 type 2 diabetes and coronary artery bypass grafts: A randomized placebo-controlled clinical trial. 8 Atherosclerosis 2010;211(2):565-73 9 10 VISION 11 Ji LN, Pan CY, Lu JM, et al. Efficacy and safety of combination therapy with vildagliptin and 12 metforminConfidential: versus metformin uptitration in ChineseFor patients Review with type 2 diabetes Only inadequately 13 controlled with metformin monotherapy: a randomized, open-label, prospective study (VISION). 14 Diabetes Obes Metab 2016;18(8):775-82 15 16 17 VIVIDD 18 McMurray JJV, Ponikowski P, Bolli GB, et al. Effects of vildagliptin on ventricular function in patients 19 with type 2 diabetes mellitus and heart failure: A randomized placebo-controlled trial. JACC: heart 20 failure 2018;6(1):8-17 21 22 Vongthavaravat 2002 23 Vongthavaravat V, Wajchenberg BL, Waitman JN, et al. An international study of the effects of 24 25 rosiglitazone plus sulphonylurea in patients with type 2 diabetes. Curr Med Res Opin 26 2002;18(8):456-61 27 28 Wajcberg 2007 29 Wajcberg E, Sriwijitkamol A, Musi N, et al. Relationship between vascular reactivity and lipids in 30 Mexican-Americans with type 2 diabetes treated with pioglitazone. J Clin Endocrinol Metab 31 2007;92(4):1256-62 32 33 Wang 2005 34 35 Wang G, Wei J, Guan Y, et al. Peroxisome proliferator-activated receptor-gamma agonist 36 rosiglitazone reduces clinical inflammatory responses in type 2 diabetes with coronary artery disease 37 after coronary angioplasty. Metabolism 2005;54(5):590-7 38 39 Wang 2013 40 Wang H, Ni Y, Yang S, et al. The effects of gliclazide, metformin, and acarbose on body composition 41 in patients with newly diagnosed type 2 diabetes mellitus. Curr Ther Res Clin Exp 2013;75:88-92 42 43 44 Wang 2015 45 Wang MM, Lin S, Chen YM, et al. Saxagliptin is similar in glycaemic variability more effective in 46 metabolic control than acarbose in aged type 2 diabetes inadequately controlled with metformin. 47 Diabetes Res Clin Pract 2015;108(3):e67-e70 48 49 Wang 2016 50 Wang W, Yang J, Yang G, et al. Efficacy and safety of linagliptin in Asian patients with type 2 diabetes 51 mellitus inadequately controlled by metformin: A multinational 24-week, randomized clinical trial. J 52 53 Diabetes 2016;8(2):229-37 54 55 Wang 2016a 56 Wang Y, Xu L, Yuan L, et al. Sodium-glucose co-transporter-2 inhibitors suppress atrial natriuretic 57 peptide secretion in patients with newly diagnosed Type 2 diabetes. Diabetic Med 58 2016;33(12):1732-36 59 60

https://mc.manuscriptcentral.com/bmj BMJ Page 120 of 244

1 2 3 Wang 2017 4 Wang W, Ning G, Ma J, et al. A randomized clinical trial of the safety and efficacy of sitagliptin in 5 6 patients with type 2 diabetes mellitus inadequately controlled by acarbose alone. Curr Med Res Opin 7 2017;33(4):693-99 8 9 Wang 2019 10 Wang W, Nevarez L, Filippova E, et al. Efficacy and safety of once-weekly dulaglutide versus insulin 11 glargine in mainly Asian patients with type 2 diabetes mellitus on metformin and/or a sulphonylurea: 12 a 52-weekConfidential: open-label, randomized phase III Fortrial. Diabetes Review Obes Metab 2019; Only21(2):234-43 13 14 Watanabe 2005 15 16 Watanabe I, Tani S, Anazawa T, et al. Effect of pioglitazone on arteriosclerosis in comparison with 17 that of glibenclamide. Diabetes Res Clin Pract 2005;68(2):104-10 18 19 Webb 2020 20 Webb DR, Htike ZZ, Swarbrick DJ, et al. A randomized, open-label, active comparator trial assessing 21 the effects of 26 weeks of liraglutide or sitagliptin on cardiovascular function in young obese adults 22 with type 2 diabetes. Diabetes Obes Metab 2020 23 24 25 Wolever 2000 26 Wolever TMS, Assiff L, Basu T, et al. Miglitol, an alpha-glucosidase inhibitor, prevents the metformin- 27 induced fall in serum folate and vitamin B12 in subjects with type 2 diabetes. Nutr Res 28 2000;20(10):1447-56 29 30 Wolffenbuttel 1999 31 Wolffenbuttel BH, Landgraf R. A 1-year multicenter randomized double-blind comparison of 32 repaglinide and glyburide for the treatment of type 2 diabetes. Dutch and German Repaglinide Study 33 Group. Diabetes Care 1999;22(3):463-7 34 35 36 Wolffenbuttel 2000 37 Wolffenbuttel BH, Gomis R, Squatrito S, et al. Addition of low-dose rosiglitazone to sulphonylurea 38 therapy improves glycaemic control in Type 2 diabetic patients. Diabetic Med 2000;17(1):40-7 39 40 Wong 2005 41 Wong TY, Szeto CC, Chow KM, et al. Rosiglitazone reduces insulin requirement and C-reactive 42 protein levels in type 2 diabetic patients receiving peritoneal dialysis. Am J Kidney Dis 43 44 2005;46(4):713-9 45 46 Wu 2014 47 Wu S, Li X, Zhang H. Effects of metformin on endothelial function in type 2 diabetes. Exp Ther Med 48 2014;7(5):1349-53 49 50 Wu 2015 51 Wu W, Li Y, Chen X, et al. Effect of linagliptin on glycemic control in Chinese patients with newly- 52 53 diagnosed, drug-naive type 2 diabetes mellitus: A randomized controlled trial. Med Sci Monit 54 2015;21:2678-84 55 56 Xiao 2015 57 Xiao CC, Ren A, Yang J, et al. Effects of pioglitazone and glipizide on platelet function in patients with 58 type 2 diabetes. Eur Rev Med Pharmacol Sci 2015;19(6):963-70 59 60

https://mc.manuscriptcentral.com/bmj Page 121 of 244 BMJ

1 2 3 Xiao 2016 4 Xiao X, Cui X, Zhang J, et al. Effects of sitagliptin as initial therapy in newly diagnosed elderly type 2 5 6 diabetics: A randomized controlled study. Experimental and Therapeutic Medicine 2016;12(5):3002- 7 08 8 9 Xu 2017 10 Xu W, Mu Y, Zhao J, et al. Efficacy and safety of metformin and sitagliptin based triple 11 antihyperglycemic therapy (STRATEGY): a multicenter, randomized, controlled, non-inferiority 12 clinical trial.Confidential: Sci China Life Sci 2017;60(3):225-38 For Review Only 13 14 Yabiku 2017 15 16 Yabiku K, Mutoh A, Miyagi K, et al. Effects of oral antidiabetic drugs on changes in the liver-to-spleen 17 ratio on computed tomography and inflammatory biomarkers in patients with type 2 diabetes and 18 nonalcoholic fatty liver disease. Clin Ther 2017;39(3):558-66 19 20 Yale 2013 21 Yale JF, Bakris G, Cariou B, et al. Efficacy and safety of canagliflozin in subjects with type 2 diabetes 22 and chronic kidney disease. Diabetes Obes Metab 2013;15(5):463-73 23 24 25 Yamakage 2019 26 Yamakage H, Tanaka M, Inoue T, et al. Effects of dapagliflozin on the serum levels of fibroblast 27 growth factor 21 and myokines and muscle mass in Japanese patients with type 2 diabetes: A 28 randomized, controlled trial. J Diabetes Investig 2019 29 30 Yamamoto 2018 31 Yamamoto S, Hayashi T, Ohara M, et al. Comparison of liraglutide plus basal insulin and basal-bolus 32 insulin therapy (BBIT) for glycemic control, body weight stability, and treatment satisfaction in 33 patients treated using BBIT for type 2 diabetes without severe insulin deficiency: a randomized 34 35 prospective pilot study. Diabetes Res Clin Pract 2018;140:339-46 36 37 Yamanouchi 2005 38 Yamanouchi T, Sakai T, Igarashi K, et al. Comparison of metabolic effects of pioglitazone, metformin, 39 and glimepiride over 1 year in Japanese patients with newly diagnosed Type 2 diabetes. Diabet Med 40 2005;22(8):980-5 41 42 Yamasaki 2005 43 44 Yamasaki Y, Katakami N, Hayaishi-Okano R, et al. a-glucosidase inhibitor reduces the progression of 45 carotid intima-media thickness. Diabetes Res Clin Pract 2005;67(3):204-10 46 47 Yan 2019 48 Yan J, Yao B, Kuang H, et al. Liraglutide, sitagliptin, and insulin glargine added to metformin: The 49 effect on body weight and intrahepatic lipid in patients with type 2 diabetes mellitus and 50 nonalcoholic fatty liver disease. Hepatology 2019;69(6):2414-26 51 52 53 Yang 2002 54 Yang WS, Jeng CY, Wu TJ, et al. Synthetic peroxisome proliferator-activated receptor-gamma agonist, 55 rosiglitazone, increases plasma levels of adiponectin in type 2 diabetic patients. Diabetes Care 56 2002;25(2):376-80 57 58 Yang 2011 59 60

https://mc.manuscriptcentral.com/bmj BMJ Page 122 of 244

1 2 3 Yang W, Pan CY, Tou C, et al. Efficacy and safety of saxagliptin added to metformin in Asian people 4 with type 2 diabetes mellitus: a randomized controlled trial. Diabetes Res Clin Pract 2011;94(2):217- 5 6 24 7 8 Yang 2012 9 Yang W, Guan Y, Shentu Y, et al. The addition of sitagliptin to ongoing metformin therapy 10 significantly improves glycemic control in Chinese patients with type 2 diabetes. J Diabetes 11 2012;4(3):227-37 12 Confidential: For Review Only 13 Yang 2013 14 Yang SJ, Min KW, Gupta SK, et al. A multicentre, multinational, randomized, placebo-controlled, 15 16 double-blind, phase 3 trial to evaluate the efficacy and safety of gemigliptin (LC15-0444) in patients 17 with type 2 diabetes. Diabetes Obes Metab 2013;15(5):410-16 18 19 Yang 2015 20 Yang HK, Min KW, Park SW, et al. A randomized, placebo-controlled, double-blind, phase 3 trial to 21 evaluate the efficacy and safety of anagliptin in drug-naive patients with type 2 diabetes. Endocrine J 22 2015;62(5):449-62 23 24 25 Yang 2015a 26 Yang W, Xing X, Lv X, et al. Vildagliptin added to sulfonylurea improves glycemic control without 27 hypoglycemia and weight gain in Chinese patients with type 2 diabetes mellitus. J Diabetes 28 2015;7(2):174-81 29 30 Yang 2016 31 Yang W, Han P, Min KW, et al. Efficacy and safety of dapagliflozin in Asian patients with type 2 32 diabetes after metformin failure: A randomized controlled trial. J Diabetes 2016;8(6):796-808 33 34 35 Yang 2018 36 Yang W, Ma J, Li Y, et al. Dapagliflozin as add-on therapy in Asian patients with type 2 diabetes 37 inadequately controlled on insulin with or without oral antihyperglycemic drugs: a randomized 38 controlled trial. J Diabetes 2018;10(7):589-99 39 40 Yee 2010 41 Yee MS, Pavitt DV, Dhanjil S, et al. The effects of rosiglitazone on atherosclerotic progression in 42 patients with Type 2 diabetes at high cardiovascular risk. Diabet Med 2010;27(12):1392-400 43 44 45 Yki-Jarvinen 1999 46 Yki-Järvinen H, Ryysy L, Nikkilä K, et al. Comparison of bedtime insulin regimens in patients with type 47 2 diabetes mellitus. A randomized, controlled trial. Ann Intern Med 1999;130(5):389-96 48 49 Yki-Jarvinen 2013 50 Yki-Jarvinen H, Rosenstock J, Duran-Garcia S, et al. Effects of adding linagliptin to basal insulin 51 regimen for inadequately controlled type 2 diabetes: A >52-week randomized, double-blind study. 52 53 Diabetes Care 2013;36(12):3875-81 54 55 Yoon 2011 56 Yoon KH, Shockey GR, Teng R, et al. Effect of initial combination therapy with sitagliptin, a dipeptidyl 57 peptidase-4 inhibitor, and pioglitazone on glycemic control and measures of beta-cell function in 58 patients with type 2 diabetes. Int J Clin Pract 2011;65(2):154-64 59 60

https://mc.manuscriptcentral.com/bmj Page 123 of 244 BMJ

1 2 3 Yoon 2011a 4 Yoon KH, Shin JA, Kwon HS, et al. Comparison of the efficacy of glimepiride, metformin, and 5 6 rosiglitazone monotherapy in korean drug-naive type 2 diabetic patients: The practical evidence of 7 antidiabetic monotherapy study. Diabetes Metabol J 2011; 35(1):26-33 8 9 Yuan 2012 10 Yuan GH, Song WL, Huang YY, et al. Efficacy and tolerability of exenatide monotherapy in obese 11 patients with newly diagnosed type 2 diabetes: a randomized, 26 weeks metformin-controlled, 12 parallel-groupConfidential: study. Chinese Med J 2012;125 For(15):2677-81 Review Only 13 14 Zang 2016 15 16 Zang L, Liu Y, Geng J, et al. Efficacy and safety of liraglutide versus sitagliptin, both in combination 17 with metformin, in Chinese patients with type 2 diabetes: a 26-week, open-label, randomized, active 18 comparator clinical trial. Diabetes Obes Metab 2016;18(8):803-11 19 20 ZEUS II 21 Cho Young M, Deerochanawong C, Seekaew S, et al. Efficacy and safety of gemigliptin as add-on 22 therapy to insulin with or without metformin in patients with type 2 diabetes mellitus (ZEUS II 23 study). Diabetes Obes Metab 2020;22:123-27 24 25 26 Zhang 2020 27 Zhang LY, Qu XN, Sun ZY, et al. Effect of liraglutide therapy on serum fetuin A in patients with type 2 28 diabetes and non-alcoholic fatty liver disease. Clin Res Hepatol Gastroenterol 2020 29 30 Zheng 2019 31 Zheng HH, Lei Y, Chen XH, et al. Retinal neuroprotective effect of GLP-1 analogs liraglutide in early 32 diabetic retinopathy. Int Eye Sci 2019;19(2):275-79 33 34 35 Zhu 2003 36 Zhu XX, Pan CY, Li GW, et al. Addition of rosiglitazone to existing sulfonylurea treatment in Chinese 37 patients with type 2 diabetes and exposure to hepatitis B or C. Diabetes Technol Ther 2003;5(1):33- 38 42 39 40 Zib 2007 41 Zib I, Jacob AN, Lingvay I, et al. Effect of pioglitazone therapy on myocardial and hepatic steatosis in 42 insulin-treated patients with type 2 diabetes. J Invest Med 2007;55(5):230-6 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

https://mc.manuscriptcentral.com/bmj BMJ Page 124 of 244

1 2 3 Appendix 3 Risk of bias in included studies 4 5 6 7 8 9 10 11 12 Confidential: For Review Only

13 Study Random sequence generation (selection bias) Allocation concealment (selection bias) Blinding of participants and investigators (performance and detection bias) Blinding of outcome assessment (performance and detection bias) Incomplete data (attrition bias) Selective reporting (reporting bias) 14 3D Low Low High Unclear Low High 15 4B Low Unclear High Unclear Low Low 16 17 4T Low Low High Unclear Low Low 18 1860-LIRA-DPP-4 Low Low High Unclear High Low 19 Abe 2008 Unclear Unclear High Unclear Low High 20 Abe 2016 Unclear Unclear High Unclear Low High 21 22 ACTION-J Low Unclear High Unclear Low High 23 ADOPT Unclear Low Low Low High Low 24 Ahmann 2015 Unclear Low Low Unclear Low High 25 Ahrén 2004 Unclear Low Low Unclear Low High 26 27 Allegretti 2019 Low Low Low Low Low High 28 Alogliptin Study 007 Unclear Low Low Unclear High Low 29 Alogliptin Study 009 Low Low Low Unclear Low Low 30 31 Alogliptin Study 010 Unclear Unclear High Unclear Unclear High 32 Alvarsson 2003 Unclear Unclear High Unclear High Low 33 APOLLO Low Low High Unclear Low High 34 Apovian 2010 Unclear Low Low Unclear High High 35 36 APPROACH Low Low Low Low High Low 37 APRIME Unclear Unclear High Unclear High Low 38 Araki 2015 Low Low High Unclear Low Low 39 Araki 2015a Low Low High Low Low Low 40 41 Arechavaleta 2011 Low Low Low Unclear High Low 42 Arjona Ferreira 2013 Low Low Low Unclear High Low 43 Arjona Ferreira 2013a Low Low Low Unclear High Low 44 Arturi 2017 Unclear Unclear High Unclear Low High 45 46 Asian Acarbose Study Unclear Low Low Unclear Low High 47 Aso 2019 Unclear Unclear High Unclear Unclear High 48 Avilés-Santa 1999 Unclear Low Low Unclear Low High 49 50 AWARD-1 Low Low Low Unclear High Low 51 AWARD-2 Low Low High Low Low Low 52 AWARD-3 Low Low Low Unclear High Low 53 AWARD-4 Low Low High Low High Low 54 55 AWARD-5 Unclear Low Low Unclear Low Low 56 AWARD-7 Low Low High Low Low Low 57 AWARD-8 Unclear Low Low Unclear Low Low 58 AWARD-9 Low Low Low Low Low Low 59 60 AWARD-10 Low Low Low Low Low Low AWARD-CHN1 Low Low Low Low Low Low

https://mc.manuscriptcentral.com/bmj Page 125 of 244 BMJ

1 2 Ba 2017 Low Low Low Low Low Low 3 Bachmann 2003 Unclear Low Low Unclear Unclear Low 4 Bailey 2010 Low Low Low Unclear Low Low 5 6 Bailey 2012 Low Low Low Unclear Low High 7 Bajaj 2014 Low Low Low Low High Low 8 Bakris 2006 Unclear Low Low Unclear High High 9 BALANCE Unclear Unclear High Unclear Unclear High 10 11 Banerji 1995 Unclear Low Low Unclear High High 12 Barnett 2003 Confidential:Unclear Low For ReviewLow Unclear OnlyLow High 13 Barnett 2012 Low Low Low Unclear High Low 14 Barnett 2013 Low Low Low Low Low Low 15 16 Barzilai 2011 Unclear Low Low Unclear High High 17 BEGIN: ADD TO GLP-1 Unclear Low Low Unclear High Low 18 BEGIN: VICTOZA ADD-ON Unclear Unclear High Unclear Low Low 19 Berberoglu 2010 Unclear Unclear High Unclear High High 20 21 Bergenstal 2009 Low Low High Unclear High Low 22 Berndt-Zipfel 2013 Unclear Unclear High Unclear Unclear High 23 BEST Unclear Low Low Unclear High High 24 25 BETA Unclear Unclear High Unclear High High 26 Bi 2013 Low Unclear High Unclear Unclear High 27 Bilezikian 2013 Low Low Low Unclear High Low 28 Birkeland 1994a Unclear Unclear High Unclear Low High 29 30 Bode 2013 Low Low Low Unclear High Low 31 Bolinder 2012 Low Low Low Unclear High Low 32 Bolli 2009 Unclear Low Low Unclear Unclear Low 33 Borges 2011 Unclear Low Low Unclear High Low 34 35 Bosi 2007 Unclear Low Low Unclear High Low 36 Bosi 2009 Unclear Low Low Unclear Low Low 37 Bosi 2011 Unclear Low Low Unclear Low Low 38 Bouchi 2017 Unclear Unclear High Unclear Low High 39 40 Braun 1996 Unclear Low Low Unclear High High 41 BRL-049653/334 Unclear Low Low Unclear High High 42 Bryson 2016 Unclear Low Low Unclear Unclear Low 43 44 Bunck 2009 Unclear Unclear High Unclear High High 45 Buse 2011 Low Low Low Unclear Low Low 46 Camerini-Davalos 1988 Low Low Low Low High High 47 Camerini-Davalos 1984 Low Low Low Unclear Low High 48 49 Campbell 1994 Unclear Unclear High Unclear Low High 50 CANDLE Unclear Unclear High Unclear High Low 51 CANATATA-D Low Low Low Unclear Low Low 52 CANTATA-D2 Low Low Low Unclear High Low 53 54 CANTATA-M Unclear Low Low Unclear High Low 55 CANTATA-MSU Low Low Low Unclear High Low 56 CANTATA-SU Low Low Low Unclear Low High 57 CANVAS/CANVAS-R Low Low Low Unclear Low Low 58 59 CARMELINA Low Low Low Low Low Low 60 CAROLINA Low Low Low Low Low Low Casner 1988 Unclear Unclear High Unclear High High

https://mc.manuscriptcentral.com/bmj BMJ Page 126 of 244

1 2 Cefalu 2015 Low Low Low Unclear Low Low 3 Chacra 2017 Unclear Low Low Unclear Low Low 4 Chakraborty 2011 Unclear Unclear High Unclear High High 5 6 Charbonnel 2013 Low Low High Unclear High High 7 Charpentier 2009 Unclear Low Low Unclear High Low 8 Chavez 2015 Unclear Unclear High Unclear Unclear High 9 Chen 2015 High Low Low Unclear High High 10 11 Chen 2016 Unclear Unclear High Unclear Low High 12 Chiasson 1994Confidential:Unclear Unclear For ReviewHigh Unclear OnlyHigh High 13 Chiasson 2001 Unclear Low Low Unclear Low Low 14 CHICAGO Unclear Low Low Unclear Low Low 15 16 Chien 2011 Unclear Unclear High Unclear Unclear High 17 Cho 2019 Unclear Unclear High Unclear Low High 18 Choi 2004 Unclear Unclear High Unclear High High 19 Chou 2008 Unclear Low Low Unclear Low Low 20 21 Chou 2012 Unclear Low Low Low High Low 22 CIMT Low Low Low Unclear High High 23 Civera 2008 Unclear Unclear High Unclear Low Low 24 25 COMPASS Low Unclear High Unclear High High 26 CompoSIT-I Low Low Low Unclear High Low 27 CompoSIT-R Low Low Low Unclear Low Low 28 CONFIDENCE Low Low High Unclear High High 29 30 Coniff 1994 Unclear Low Low Unclear High High 31 Coniff 1995 Unclear Low Low Unclear High Low 32 Coniff 1995a Unclear Low Low Unclear Low High 33 Costa 1997 Unclear Low Low Unclear High High 34 35 CREDENCE Low Low Low Low Low Low 36 Cusi 2019 Low Low Low Unclear High High 37 CV181-011 Unclear Low Low Unclear High Low 38 CV181-013 Low Low Low Unclear High Low 39 40 CV181-039 Low Low Low Unclear High Low 41 da Silva 2016 Unclear Unclear High Unclear Low High 42 Dailey 2004 Unclear Low Low Unclear High Low 43 44 Dapagliflozin 006 Low Low Low Low Low Low 45 Dargie 2007 Low Low Low Unclear High Low 46 Davidson 2007 Unclear Low Low Unclear High High 47 Davies 2013 Low Low High Unclear High High 48 49 Davies 2017 Low Low High Unclear Low Low 50 DECLARE-TIMI 58 Low Low Low Low Low Low 51 DeFronzo 1995 Unclear Low Low Unclear High Low 52 DeFronzo 2005 Unclear Low Low Low High High 53 54 DeFronzo 2012 Unclear Low Low Unclear High Low 55 DeFronzo 2015 Low Low Low Unclear High Low 56 Dei Cas 2017 Low Unclear High Unclear High High 57 Dejager 2007 Unclear Low Low Unclear Low High 58 59 Del Prato 2003 Unclear Low Low Unclear High High 60 Del Prato 2011 Unclear Low Low Unclear Low High Del Prato 2014 Unclear Low Low Low High Low

https://mc.manuscriptcentral.com/bmj Page 127 of 244 BMJ

1 2 DELIGHT Low Low Low Low Low High 3 Deng 2017 Unclear Low High Unclear Low High 4 DERIVE Low Low Low Unclear Low Low 5 6 Derosa 2003 Unclear Unclear High Unclear Low High 7 Derosa 2003a Unclear Unclear High Unclear Low High 8 Derosa 2004 Unclear Unclear High Unclear Low High 9 Derosa 2005 Low Low Low Unclear Low High 10 11 Derosa 2007 Low Low Low Unclear Low High 12 Derosa 2009 Confidential:Low Low For ReviewLow Unclear OnlyHigh High 13 Derosa 2009a Low Low Low Unclear Low High 14 Derosa 2010 Unclear Unclear High Unclear Low High 15 16 Derosa 2010a Unclear Unclear High Unclear Low High 17 Derosa 2011 Unclear Unclear High Unclear High High 18 Derosa 2011a Unclear Unclear High Unclear High High 19 Derosa 2012 Unclear Unclear High Unclear High High 20 21 Derosa 2012a Unclear Unclear High Unclear High High 22 Derosa 2013 Unclear Low Low Unclear Low High 23 Derosa 2014 Unclear Low Low Unclear High High 24 25 Derosa 2014a Unclear Low Low Unclear Low High 26 DIA3004 Low Low Low Unclear High Low 27 Distiller 2014 Unclear Unclear High Unclear Low High 28 DIVERSITY-CVR Low Low High Low High High 29 30 Dobs 2013 Low Low Low Unclear High Low 31 Dorkhan 2009 Unclear Unclear High Unclear Low High 32 Douek 2005 Unclear Low Low Unclear High High 33 Drent 2002 Unclear Low Low Unclear High Low 34 35 DUAL-I Low Low High Low Low High 36 DUAL-II Low Low High Unclear Low Low 37 DUAL-III Unclear Unclear High Unclear High High 38 DUAL IX Low Low High Low High Low 39 40 DURATION-2 Low Low Low Unclear High Low 41 DURATION-3 Low Low High Low High Low 42 DURATION-4 Low Low Low Unclear High Low 43 44 DURATION-8 Low Low Low Low Low Low 45 DURATION-NEO-2 Low Low Low Low Low Low 46 EAGLE Unclear Unclear High Unclear High High 47 EASIE Low Low High Unclear High Low 48 49 Ebato 2009 Unclear Unclear High Unclear Low High 50 EDIT Low Unclear High Unclear Low High 51 Efstathiou 2015 Unclear Unclear High Unclear Unclear High 52 ELEGANT Low Unclear High Unclear Low High 53 54 ELIXA Low Low Low Low Low Low 55 ELLENA-IT Low Low High Unclear Low Low 56 EMBLEM Low Low Low Unclear High High 57 EMIT Unclear Low Low Unclear High High 58 59 EMLIFA001 Low Low Low Unclear High High 60 EMPA-REG BASAL Low Low Low Unclear High Low EMPA-REG H2H-SU Low Low Low Unclear High Low

https://mc.manuscriptcentral.com/bmj BMJ Page 128 of 244

1 2 EMPA-HEART CardioLink-6 Low Low Low Low Low Low 3 EMPA-REG MDI Unclear Low Low Unclear Low Low 4 EMPA-REG MET Low Low Low Unclear Low Low 5 6 EMPA-REG METSU Low Low Low Unclear Low Low 7 EMPA-REG MONO Low Low Low Unclear Low Low 8 EMPA-REG OUTCOME Low Low Low Unclear Low Low 9 EMPA-REG PIO Low Low Low Unclear Low Low 10 11 EMPA-REG RENAL - CKD2 Low Low Low Unclear Low High 12 EngelbrechtsenConfidential: 2016 Unclear Unclear For ReviewHigh Unclear OnlyHigh High 13 Erem 2014 Unclear Unclear High Unclear Low Low 14 ESPECIAL-ACS Unclear Unclear High Unclear High Low 15 16 Esposito 2004 Low Unclear High Unclear Low High 17 Esposito 2011 Low Unclear High Unclear Low Low 18 Essen Study Low Unclear High Unclear Low High 19 EUREXA Low Unclear High Unclear High Low 20 21 EUREXA extension Low Unclear High Unclear High Low 22 EXAMINE Unclear Low Low Unclear Low Low 23 Exenatide-113 Unclear Low Low Unclear High High 24 25 EXSCEL Low Low Low Low Low Low 26 Feng 2017 Unclear Unclear High Unclear Low High 27 Ferdinand 2019 Unclear Low Low Unclear High Low 28 Fernandez 2008 Unclear Unclear High Unclear Low High 29 30 Ferrannini 2010 Unclear Low Low Unclear Low Low 31 Filozof 2010 Unclear Low Low Low High High 32 Finn 2009 Low Low Low Unclear Low High 33 Fischer 1998 Unclear Low Low Unclear High High 34 35 Foley 2009 Unclear Low Low Unclear High High 36 Foley 2011 Unclear Low Low Unclear Low High 37 Fonseca 2000 Low Low Low Low Low Low 38 Fonseca 2003 Unclear Low Low Unclear High High 39 40 Fonseca 2007 Unclear Low Low Unclear Low Low 41 Fonseca 2013 Unclear Low Low Unclear Low Low 42 Forst 2003 Unclear Unclear High Unclear Unclear High 43 44 Forst 2005 Unclear Unclear High Unclear High High 45 Forst 2014 Low Low Low Unclear Low Low 46 Forst 2015 Unclear Unclear High Unclear High Low 47 Frederich 2012 Unclear Unclear Unclear Unclear High Low 48 49 FREEDOM-1 Low Low Low Low High High 50 Gaal 2001 Unclear Low Low Unclear High High 51 Gallwitz 2011 Unclear Unclear Unclear Unclear High High 52 Gallwitz 2012 Low Low Low Unclear High Low 53 54 Gantz 2017 Low Low Low Low High Low 55 Gantz 2017a Low Low Low Low High Low 56 Gantz 2017b Low Low Low Unclear High High 57 Garber 2006 Unclear Low Low Unclear High High 58 59 Garber 2007 Unclear Low Low Unclear High High 60 Garber 2008 Unclear Low Low Unclear High Low Gastaldelli 2014 Unclear Low Low Unclear High High

https://mc.manuscriptcentral.com/bmj Page 129 of 244 BMJ

1 2 GENERATION Low Low Low Unclear High Low 3 Gentile 2001 Unclear Low Low Unclear Low High 4 GetGoal-Duo-1 Low Low Low Low Low Low 5 6 GetGoal-Duo-2 Low Low High Unclear High Low 7 GetGoal-F1 Unclear Low Low Unclear Low High 8 GetGoal-L Low Low Low Unclear High Low 9 GetGoal-L-C Low Low Low Unclear High High 10 11 GetGoal-L Asia Unclear Low Low Unclear High Low 12 GetGoal-M Confidential:Unclear Low For ReviewLow Unclear OnlyHigh Low 13 GetGoal-M-Asia Low Low Low Unclear High Low 14 GetGoal-O Unclear Low Low Unclear Low Low 15 16 GetGoal-P Unclear Low Low Unclear Low Low 17 GetGoal-S Unclear Low Low Unclear High Low 18 Giles 2008 Unclear Low Low Unclear High Low 19 Giugliano 1993 Unclear Low Low Unclear Unclear High 20 21 GLAC Unclear Unclear High Unclear High High 22 GLAD Low Low Low Unclear High High 23 GLAL Unclear Unclear High Unclear High High 24 25 Glimepiride Combination Group 1998 Unclear Low Low Unclear High Low 26 Göke 2002 Unclear Low High Unclear High High 27 Göke 2010 Low Low Low Unclear High Low 28 Gómez-Perez 2002 Unclear Low Low Unclear Low High 29 30 Gomis 2011 Low Low Low Unclear High High 31 Goodman 2009 Unclear Low Low Unclear High High 32 Grant 1996 Unclear Unclear High Unclear High High 33 GRAVITAS Unclear Low Low Unclear High Low 34 35 Grey 2014 Low Low Low Unclear High High 36 Gudipaty 2014 Unclear Unclear High Unclear High High 37 Gurkan 2014 Unclear Unclear High Unclear Unclear High 38 Gutniak 1987 Unclear Unclear High Unclear Unclear High 39 40 Güvener 1999 Unclear Low Low Unclear Unclear High 41 Guzman 2017 Low Low Low Low High High 42 Haak 2012 Unclear Low Low Unclear High Low 43 44 Hadjadj 2016 Low Low Low Unclear Low High 45 Halimi 2000 Unclear Low Low Unclear Low High 46 Hällsten 2002 Unclear Low Low Unclear Low High 47 Halvorsen 2019 Unclear Low Low Low High Low 48 49 Han 2018 Low Low Low Unclear High High 50 Handelsman 2017 Low Low Low Low High Low 51 Handelsman 2019 Low Low Low Unclear High Low 52 Haneda 2016 Unclear Low Low Unclear High High 53 54 Hanefeld 1991 Unclear Low Low Unclear Low High 55 Hanefeld 2004 Unclear Low Low Unclear High High 56 Hanefeld 2007 Unclear Low Low Unclear High Low 57 HARMONY 1 Low Low Low Low High Low 58 59 HARMONY 2 Low Low Low Unclear High Low 60 HARMONY 3 Unclear Low Low Unclear High Low HARMONY 4 Low Low High Low High High

https://mc.manuscriptcentral.com/bmj BMJ Page 130 of 244

1 2 HARMONY 5 Low Low Low Low High Low 3 HARMONY 6 Unclear Low Low Unclear Low High 4 HARMONY OUTCOMES Low Low Low Low Low Low 5 6 Hartemann-Heurtier 2009 Unclear Low Low Unclear Low High 7 Hartley 2015 Low Low Low Unclear High Low 8 Hasche 1999 Unclear Low Low Unclear Low High 9 Hattori 2017 Unclear Unclear High Unclear Low High 10 11 Hattori 2018 Unclear Unclear High Unclear Unclear High 12 HEELA Confidential:Unclear Unclear For ReviewHigh Unclear OnlyHigh High 13 Hegele 1995 Unclear Unclear Unclear Unclear High High 14 Heine 2005 Low Low High Unclear High High 15 16 Heliövaara 2007 Unclear Low Low Unclear Unclear High 17 Henry 2012 Low Low Low Unclear High Low 18 Henry 2014 Unclear Unclear Unclear Unclear High Low 19 Hermann 1999 Unclear Low Low Unclear High High 20 21 Hiramatsu 2018 Unclear Unclear High Unclear High High 22 Hirano 2009 Low Low Unclear Unclear Low High 23 Hirano 2012 Low Unclear Unclear Unclear Low High 24 25 Hollander 2007 Unclear Low Low Low High High 26 HOME Low Low Low Unclear Low Low 27 Home 2018 Unclear Low Low Unclear High Low 28 Hong 2013 Unclear Unclear Unclear Unclear High High 29 30 Hong 2015 Unclear Unclear High Unclear High Low 31 Hong 2016 Unclear Low Low Unclear High High 32 Horton 2000 Low Low Low Unclear High Low 33 Hotta 1993 Unclear Low Low Unclear High High 34 35 Hsieh 2011 Unclear Unclear Unclear Unclear High High 36 Hu 2007 Unclear Low Low Unclear Unclear High 37 Hwang 2008 Unclear Unclear High Unclear High High 38 Hygum 2020 Unclear Low Low Low Low High 39 40 Iacobellis 2017 Unclear Unclear High Unclear High High 41 ILLUMINATE Unclear Low Low Unclear Low Low 42 Inagaki 2012 Low Low High Unclear High Low 43 44 Inagaki 2013 Unclear Unclear High Unclear Low Low 45 Inagaki 2014 Unclear Low Low Unclear Low Low 46 Inagaki 2015 Low Low Low Unclear Low High 47 INICOM Unclear Low Low Unclear High High 48 49 Inoue 2019 Low Low High Unclear High High 50 Inoue 2019a Low Low High Unclear Low High 51 Insulin Glargine 4014 Unclear Unclear High Unclear Low High 52 INTERVAL Low Low Low Unclear Low Low 53 54 Ito 2011 Unclear Unclear High Unclear High High 55 Ito 2017 Low Unclear High Unclear Low High 56 Jabbour 2014 Unclear Low Low Unclear High Low 57 Jacob 2007 Unclear Unclear High Unclear High High 58 59 Jain 2006 Unclear Low Low Unclear High Low 60 JEDIS-1 Unclear Unclear High Unclear High High Jeon 2011 Unclear Unclear High Unclear Low High

https://mc.manuscriptcentral.com/bmj Page 131 of 244 BMJ

1 2 Jerums 1987 Unclear Low Low Unclear High High 3 Ji 2014 Low Low Low Unclear High High 4 Ji 2016 Low Low Low Unclear High Low 5 6 Ji 2017 Unclear Low Low Unclear High High 7 Jian 2018 Unclear Unclear High Unclear Unclear High 8 Jibran 2006 Unclear Unclear Unclear Unclear Unclear High 9 Johnston 1998 Unclear Low Low Unclear High High 10 11 Johnston 1998a Unclear Low Low Unclear Low High 12 Johnston 1998bConfidential:Unclear Low For ReviewLow Unclear OnlyLow Low 13 Josse 2003 Unclear Low Low Unclear Low High 14 Jovanovic 2000 Unclear Low Low Unclear High High 15 16 Jovanovic 2004 Unclear Unclear High Unclear Low High 17 Jung 2005 Unclear Unclear Unclear Unclear Low High 18 Juurinen 2009 Unclear Low Low Unclear Low High 19 Kadoglou 2007 Unclear Unclear Unclear Unclear Low High 20 21 Kadoglou 2008 Unclear Unclear High Unclear Low High 22 Kadoglou 2011 Unclear Unclear High Unclear Low High 23 Kadowaki 2011 Unclear Low Low Unclear High Low 24 25 Kadowaki 2017 Unclear Low Low Unclear High Low 26 Kadowaki 2018 Unclear Low Low Unclear High Low 27 Kaku 2009 Unclear Low Low Unclear High High 28 Kaku 2009a Unclear Unclear Unclear Low High Low 29 30 Kaku 2011 Unclear Unclear High Unclear High Low 31 Kaku 2014 Unclear Low Low Unclear High High 32 Kaku 2017 Unclear Low Low Unclear Low Low 33 Kaku 2019 Low Low High Unclear Low Low 34 35 Kaku 2019a Low Low Low Low Low Low 36 Kanazawa 2010 Unclear Unclear High Unclear Low High 37 Kanazawa 2011 Unclear Unclear High Unclear Low Low 38 Kato 2015 Unclear Unclear High Unclear Low High 39 40 Kawamori 2018 Low Low Low Low Low Low 41 Kelly 2007 Unclear Low Low Unclear Low Low 42 Kendall 2005 Unclear Low Low Unclear High High 43 44 Khaloo 2019 Unclear Unclear High Unclear Low High 45 Khanolkar 2008 Unclear Unclear High Unclear Low High 46 Kikuchi 2012 Unclear Low Low Unclear High Low 47 Kim 2014 Unclear Low Low Unclear Low High 48 49 Kim Jeong 2019 Unclear Unclear High Unclear High High 50 KIND-LM Unclear Unclear High Unclear Low High 51 Kiyici 2009 Unclear Unclear High Unclear Low High 52 Ko 2006 Unclear Unclear High Unclear High High 53 54 Koffert 2017 Unclear Unclear High Unclear Unclear High 55 Kohan 2014 Low Low Low Low Low Low 56 Kondo 2016 Unclear Unclear High Unclear High High 57 Kothny 2013 Low Low Low Unclear High Low 58 59 Krawczyk 2005 Unclear Unclear High High High High 60 Kudo-Fujimaki 2014 Unclear Unclear High Unclear Low High Laberge 2016 Unclear Unclear Unclear Unclear Unclear High

https://mc.manuscriptcentral.com/bmj BMJ Page 132 of 244

1 2 Lam 1998 Unclear Low Low Unclear Low High 3 Lambadiari 2018 Unclear Unclear High Unclear Unclear High 4 Langenfeld 2005 Unclear Unclear High Unclear High High 5 6 LANTERN Unclear Low Low Unclear High High 7 Lawrence 2004 Unclear Unclear High Low Unclear High 8 LEAD-1 Unclear Low Low Unclear High Low 9 LEAD-2 Low Low Low Unclear High Low 10 11 LEAD-3 Low Low Low Unclear High Low 12 LEAD-4 Confidential:Low Low For ReviewLow Unclear OnlyHigh Low 13 LEAD-5 Low Low High Unclear Low High 14 LEADER Low Low Low Low Low Low 15 16 Lebovitz 2001 Unclear Low Low Unclear Low High 17 Ledesma 2019 Unclear Low Low Unclear High Low 18 Lee 2013 Unclear Low Low Unclear Unclear Low 19 Lee 2017 Low Low Low Low High Low 20 21 Leiter 2014 Low Low Low Low Low Low 22 Leiter 2014a Low Low Low Unclear Low Low 23 Leonhardt 1991 Unclear Low Low Unclear Low High 24 25 Lewin 2007 Unclear Low Low Unclear High High 26 Lewin 2015 Low Low Low Unclear Low Low 27 Li 2014 Low Unclear High Unclear High High 28 Li 2014b Unclear Unclear High Unclear Unclear High 29 30 Li 2017 Unclear Low Low Unclear Low High 31 Li 2019 Unclear Unclear High Unclear Low High 32 Li 2019a Unclear Unclear High Unclear Unclear High 33 LIBRA Low Low Low Unclear Low High 34 35 Lim 2017 Unclear Low Low Unclear Low High 36 Lin 2003 Unclear Low Low Unclear High Low 37 Lindström 1999 Unclear Low Low Unclear High High 38 Lingvay 2018 Low Low Low Unclear Low Low 39 40 LIPER2 Unclear Low Low Unclear High High 41 Liraglutide-Detemir Study Unclear Low Low Unclear High Low 42 LIRA-ADD2SGLT2i Low Low Low Unclear Low Low 43 44 LIRA-RENAL Low Low Low Unclear High Low 45 LIRA-SWITCH Low Low Low Unclear High High 46 Liu 2013 Unclear Unclear High Unclear Low High 47 Liu 2014 Low Low Low Unclear Low High 48 49 Liu 2020 Low Low High Unclear High Low 50 Liutkus 2010 Unclear Low Low Unclear High Low 51 LixiLan-G Low Low High Low Low Low 52 LixiLan JP-L Low Low High Unclear Low Low 53 54 Lou 2020 Unclear Unclear High Unclear Unclear High 55 Lu 2016 Low Low Low Unclear High High 56 Lukashevich 2011 Unclear Low Low Unclear Low High 57 Lund 2009 Low Low Low Unclear Low High 58 59 LYDIA Low Low High Unclear High Low 60 Ma 2015 Unclear Unclear High Unclear Low High Macauley 2015 Unclear Low Low Unclear Low High

https://mc.manuscriptcentral.com/bmj Page 133 of 244 BMJ

1 2 Madsbad 2001 Unclear Low Low Unclear High High 3 Maffioli 2013 Unclear Low Low Unclear Low High 4 MAGNA VICTORIA Unclear Low Low Unclear Low High 5 6 Marbury 1999 Unclear Low Low Unclear High Low 7 MARCH Low Low Low Unclear High Low 8 Marena 1993 Unclear Low Low Unclear Unclear High 9 Mari 2008 Unclear Low Low Unclear High High 10 11 MARLINA-T2D Unclear Low Low Unclear Unclear High 12 Marre 2002 Confidential:Low Low For ReviewLow Unclear OnlyLow High 13 MASTER Unclear Unclear High Unclear Low High 14 Mathieu 2015 Unclear Low Low Unclear Low Low 15 16 Mathieu 2016 Unclear Low Low Unclear Low Low 17 Matthaei 2015 Unclear Low Low Unclear Low Low 18 Matthaei 2015a Low Low Low Unclear Low High 19 Matthews 2005 Unclear Low Low Unclear High Low 20 21 Matthews 2010 Unclear Low Low Unclear High Low 22 Mattoo 2005 Low Low Low Unclear Low High 23 McCluskey 2004 Unclear Low Low Unclear Low High 24 25 McGill 2013 Unclear Low Low Low High Low 26 McGuire 2010 Unclear Low Low Unclear High High 27 MDI Liraglutide Unclear Low Low Unclear Low High 28 Meneghini 2010 Unclear Unclear High Unclear High High 29 30 Meneilly 2000 Unclear Low Low Unclear Unclear High 31 Mita 2007 Low Unclear High Unclear Low High 32 Mita 2019 Unclear Unclear High Unclear Low High 33 Mitrakou 1998 Unclear Low Low Unclear Low High 34 35 Miyagawa 2015 Low Low Low Low Low Low 36 Miyazaki 2002 Low Low High Unclear Unclear High 37 Mokta 2018 Unclear Unclear High Unclear Low High 38 Moretto 2008 Low Low Low Unclear Low Low 39 40 Moriwaki 2018 Unclear Unclear High Unclear Low High 41 Moses 2014 Unclear Low Low Unclear Low Low 42 Moses 2016 Unclear Low Low Unclear High Low 43 44 Mu 2017 Unclear Low Low Unclear Low Low 45 Müller-Wieland 2018 Low Low Low Unclear High Low 46 Nakamura 2001 Unclear Low Low Unclear Low High 47 Nakamura 2004 Unclear Unclear High Unclear Low High 48 49 Nakamura 2006 Unclear Unclear High Unclear Low High 50 Nar 2009 Unclear Unclear High Unclear Unclear High 51 Nathan 1988 Low Low Low Unclear Unclear High 52 Nauck 2007a Low Low High Unclear High High 53 54 Nauck 2007b Unclear Low Low Unclear High Low 55 Nauck 2009 Low Low Low Unclear High Low 56 Nauck 2011 Low Low Low Unclear High Low 57 Nauck 2016 Low Low Low Unclear High High 58 59 Neff 2016 Unclear Unclear Unclear Unclear High High 60 Negro 2005 Unclear Unclear High Unclear Unclear High Ning 2016 Low Low Low Unclear Low High

https://mc.manuscriptcentral.com/bmj BMJ Page 134 of 244

1 2 Nino 2017 Low Low Low Unclear High Low 3 Nishimura 2016 Unclear Unclear High Unclear Low High 4 Nishio 2006 Unclear Unclear High Unclear Unclear High 5 6 Nogueira 2014 Unclear Unclear High Unclear High High 7 Nowicki 2011 Low Low Low Unclear High Low 8 Ogasawara 2009 Unclear Unclear High Unclear Low High 9 Ohira 2014 Unclear Unclear High Unclear Unclear High 10 11 Ohira 2014a Unclear Low High Unclear Low High 12 Omarigliptin ProtocolConfidential: 2010 Low Low For ReviewLow Unclear OnlyLow Low 13 Onuchin 2010 Unclear Unclear High High Low High 14 Osman 2004 Unclear Low Low Unclear Low Low 15 16 Ovalle 2004 Unclear Unclear High Unclear Low High 17 Owens 2011 Unclear Low Low Unclear Low High 18 Oyama 2008 Unclear Low High Unclear Unclear High 19 Pagano 1995 Unclear Unclear High Unclear Low High 20 21 Pan 2008 Unclear Low Low Unclear High Low 22 Pan 2012 Unclear Low Low Unclear Low Low 23 Pan 2012a Low Low Low Unclear High Low 24 25 Pan 2015 Unclear Low Low Unclear Unclear High 26 Park 2011 Unclear Unclear High Unclear Unclear High 27 Park 2014 Unclear Unclear High Unclear Low High 28 Park 2017 Unclear Low Low Unclear High High 29 30 Parmar Vinendra 2019 Unclear Unclear High Unclear Low Low 31 Parthan 2018 Low Low High Unclear High High 32 Patel 2013 Unclear Low Low Unclear High High 33 Pavithra 2019 Unclear Unclear High Unclear Unclear High 34 35 Pavo 2003 Unclear Low Low Unclear Low High 36 Perez 2009 Unclear Low Low Unclear High Low 37 PERISCOPE Low Low Low Low High Low 38 Periello 2006 Unclear Low Low Unclear Unclear High 39 40 Petrica 2009 Unclear Unclear High High Low High 41 Petrica 2011 Unclear Unclear High Unclear Low High 42 Philis-Tsimikas 2013 Low Low High Low High Low 43 44 Phillips 2001 Unclear Low Low Unclear High High 45 Phillips 2003 Unclear Low Low Unclear High High 46 Phrommintikul 2019 Low Low Low Unclear High Low 47 PIOCOMB Unclear Low Low Unclear Unclear High 48 49 PIOfix Unclear Low Low Unclear High High 50 Pioglitazone 001 Unclear Low Low Unclear High High 51 PIONEER Unclear Unclear High Unclear High High 52 PIONEER 1 Low Low Low Low Low Low 53 54 PIONEER 2 Low Low High Low High Low 55 PIONEER 3 Low Low Low Low Low Low 56 PIONEER 4 Low Low Low Low Low Low 57 PIONEER 5 Low Low Low Low Low Low 58 59 PIONEER 6 Low Low Low Low Low Low 60 PIONEER 7 Low Low Low Low High Low PIONEER 8 Low Low Low Low Low Low

https://mc.manuscriptcentral.com/bmj Page 135 of 244 BMJ

1 2 PIONEER 9 Low Low Low Low Low Low 3 PioRAGE Low Unclear High Unclear Low High 4 PIRAMID 2009 Unclear Low Low Unclear Low High 5 6 Pistrosch 2012 Unclear Low Low Unclear Unclear High 7 Pistrosch 2013 Unclear Unclear High Unclear High High 8 Pi-Sunyer 2007 Unclear Low Low Unclear High High 9 Pop-Busui 2009 Unclear Unclear High Unclear High High 10 11 POPPS Unclear Unclear High Unclear Unclear High 12 Pratley 2014 Confidential:Unclear Low For ReviewLow Low OnlyHigh Low 13 PRESERVE-beta Unclear Low Low Unclear High High 14 PREVENT-J Unclear Unclear High Unclear High High 15 16 PRIDE Low Unclear High Unclear High Low 17 PRIME-V study Unclear Unclear High Unclear Low High 18 PRISMA Unclear Low Low Unclear Low Low 19 PROactive Low Low Low Low Low Low 20 21 PROLOGUE Low Low High Low Low Low 22 QUARTER Low Low Low Unclear Low High 23 Quatraro 1986 Unclear Unclear High Unclear Unclear High 24 25 Rahman 2010 Unclear Low Low Unclear Low Low 26 Rahman 2011 Unclear Unclear High Unclear Unclear High 27 Raskin 2001 Low Low Low Unclear High High 28 Raskin 2004 Unclear Unclear High Unclear High High 29 30 Raskin 2009 Unclear Unclear High Unclear High Low 31 Raz 2008 Low Low Low Unclear Low Low 32 RECORD Unclear Unclear High Low Low Low 33 REGO-F Unclear Unclear High Unclear High High 34 35 RELEASE Unclear Low Low Unclear High High 36 RESULT Unclear Low Low Unclear High Low 37 REWIND Low Low Low Low Low Low 38 Reynolds 2002 Unclear Unclear High Unclear High High 39 40 Reynolds 2007 Unclear Unclear High Unclear Low High 41 Ristic 2006 Low Low Low Unclear High Low 42 Robbins 2007 Unclear Low High Unclear High High 43 44 Roberts 2005 Unclear Low Low Unclear High Low 45 Rodbard 2016 Low Low Low Unclear High High 46 Rosenstock 1998 Unclear Low Low Unclear High High 47 Rosenstock 2006 Unclear Low Low Unclear High Low 48 49 Rosenstock 2007 Unclear Low Low Unclear Low Low 50 Rosenstock 2007a Unclear Low Low Unclear High High 51 Rosenstock 2009 Low Low Low Unclear High Low 52 Rosenstock 2010 Unclear Low Low Unclear High High 53 54 Rosenstock 2012 Unclear Low Low Unclear High Low 55 Rosenstock 2013 Unclear Low Low Unclear High Low 56 Rosenstock 2015 Unclear Low Low Unclear High Low 57 Rosenstock 2016 Low Low Low Unclear Low Low 58 59 Rosenstock 2019 Low Low Low Unclear Low Low 60 Ross 2015 Low Low Low Unclear High Low SAIS1 Unclear Unclear High Unclear High Low

https://mc.manuscriptcentral.com/bmj BMJ Page 136 of 244

1 2 Saleem 2011 Unclear Unclear High Unclear Unclear High 3 Salman 2001 Unclear Unclear High Unclear Low High 4 Saloranta 2002 Unclear Low Low Unclear Unclear Low 5 6 Samson 2011 Low Low High High Unclear High 7 Sato 2019 Unclear Unclear High Unclear Unclear High 8 SAVOR -TIMI 53 Low Low Low Low Low Low 9 Saxagliptin 014 Low Low Low Unclear High Low 10 11 SCALE Diabetes Low Low Low Unclear Low Low 12 Scherbaum 2002Confidential:Unclear Low For ReviewLow Unclear OnlyHigh High 13 Scherbaum 2008 Unclear Low Low Unclear Low Low 14 Schweizer 2007 Unclear Low Low Unclear High Low 15 16 Schweizer 2009 Unclear Low Low Unclear High Low 17 Segal 1997 Unclear Low Low Unclear High High 18 Segal 2005 Unclear Low Low Unclear High Low 19 Seino 2011 Unclear Low Low Unclear High Low 20 21 Seino 2015 Unclear Low Low Unclear Low High 22 Seino 2016 Low Low Low Unclear Low Low 23 Seino 2018 Low Low High Low Low Low 24 25 Shaddinger 2019 Low Low Low Low High Low 26 Shah 2011 Unclear Unclear Unclear Unclear Unclear High 27 Shaienko 2018 Unclear Unclear Unclear Unclear Unclear High 28 Shankar 2017 Unclear Low Low Unclear Low Low 29 30 Shankar 2017a Unclear Low Low Unclear Low High 31 Shestakova 2018 Low Low Low Unclear Low Low 32 Shibuya 2018 Unclear Unclear High Unclear Low High 33 SIMPLE Unclear Unclear High Unclear High Low 34 35 Sit2Mix Unclear Unclear High Unclear High High 36 Sitagliptin 019 Unclear Low Low Unclear High High 37 Sitagliptin 020 Unclear Low Low Unclear High High 38 Sitagliptin 021 Unclear Low Low Unclear Low High 39 40 Sitagliptin 024 Unclear Low Low Unclear High Low 41 Sitagliptin 035 Low Low Low Unclear High High 42 Sitagliptin 036 Unclear Low Low Unclear Low Low 43 44 Sitagliptin 049 Low Low Low Unclear High Low 45 Sitagliptin 051 Low Unclear Unclear Unclear Low Low 46 SMART Low Low High Unclear Low Low 47 Smith 2005 Unclear Low Low Unclear Unclear High 48 49 Søfteland 2017 Low Low Low Unclear Low Low 50 Sohn 2008 Unclear Unclear High Unclear High High 51 Sone 2019 Low Low Low Unclear Low Low 52 South Danish Diabetes Study Low Unclear High High High High 53 54 SPEAD-A Unclear Unclear High Unclear Low Low 55 SPECIFY Low Low High Unclear High High 56 Spengler 1989 Unclear Unclear High risk Unclear Unclear High 57 SPIKE Unclear Unclear High Unclear Low Low 58 59 SPOTLIGHT Unclear Unclear High Unclear High High 60 SPREAD-DIMCAD Low Low Low Low Unclear Low Sridhar 2013 Unclear Low Low Unclear Low High

https://mc.manuscriptcentral.com/bmj Page 137 of 244 BMJ

1 2 St. John Sutton 2002 Unclear Unclear High Unclear Unclear High 3 Standl 1999 Unclear Low Low Unclear High High 4 Standl 2001 Unclear Low Low Unclear High High 5 6 START Low Low Low Unclear High Low 7 START-J Low Unclear High Unclear Low Low 8 Stewart 2006 Unclear Low Low Unclear High High 9 Stocker 2007 Low Unclear High Unclear High High 10 11 Strojek 2009 Low Low High Unclear Low Low 12 Strojek 2011 Confidential:Low Low For ReviewLow Unclear OnlyLow Low 13 Strøm Halden 2019 Low Low Low Unclear High Low 14 Su 2014 Unclear Low Low Unclear High High 15 16 SUCCESS Unclear Unclear High Unclear Low Low 17 SUMER Unclear Unclear High Unclear High Low 18 Sun 2006 Unclear Unclear Unclear Unclear Low High 19 Sun 2016 Unclear Unclear High Unclear Low Low 20 21 SUPER Unclear Low Low Unclear Low Low 22 SUSTAIN 1 Low Low Low Low Low Low 23 SUSTAIN 2 Low Low Low Unclear High High 24 25 SUSTAIN 4 Low Low Low Low Low Low 26 SUSTAIN 5 Low Low Low Low Low Low 27 SUSTAIN 6 Low Low Low Low Low Low 28 SUSTAIN 8 Low Low Low Low Low Low 29 30 SUSTAIN 9 Low Low Low Low Low Low 31 Suzuki 2014 Unclear Unclear High Unclear High High 32 SWIM Low Low High Unclear Low High 33 Takagi 2003 Unclear Unclear High Unclear Unclear High 34 35 Takase 2007 Low Low High Unclear Low High 36 Takashima 2018 Unclear Unclear High Unclear Unclear High 37 Tao 2018 Low Unclear High Unclear High High 38 Taskinen 2011 Unclear Low Low Unclear Low High 39 40 TECOS Low Low Low Low Low Low 41 T-Emerge 1 Unclear Low Low Unclear High Low 42 T-Emerge 3 Unclear Low Low Unclear High Low 43 44 T-Emerge 4 Unclear Low Low Unclear Low High 45 T-Emerge 5 Unclear Unclear High Unclear High High 46 T-Emerge 6 Unclear Low Low Unclear High Low 47 T-Emerge 7 Unclear Unclear Unclear Low High Low 48 49 Teramoto 2007 Unclear Unclear High Unclear Low High 50 Teupe 1991 Unclear Unclear Unclear Unclear High Low 51 Thrasher 2014 Low Low Low Unclear Low Low 52 TIDE Unclear Low Low Low High Low 53 54 Tinahones 2017 Low Low Low Unclear Low Low 55 Tofogliflozin 003 Unclear Low Low Unclear High Low 56 Tolman 2009 Unclear Low Low Unclear High Low 57 TOPSCORE Unclear Unclear High Unclear High High 58 59 TOSCA.IT Low Low High Low High Low 60 Tripathy 2013 Unclear Unclear High Unclear High High TROICA Unclear Low Low Unclear High Low

https://mc.manuscriptcentral.com/bmj BMJ Page 138 of 244

1 2 Truitt 2010 Unclear Low Low Unclear High Low 3 Türkmen Kemal 2007 Unclear Unclear Unclear Unclear Low High 4 UKPDS 44 Unclear Low Low Unclear High High 5 6 Umpierrez 2006 Unclear Unclear High Unclear High High 7 Umpierrez 2018 Low Unclear High Unclear Unclear High 8 Vähätalo 2007 Unclear Unclear Unclear Unclear Unclear High 9 Van Eyk 2019 Unclear Low Low Low Low High 10 11 Van Gaal 2014 Unclear Low Low Unclear High Low 12 Vanderheiden Confidential:2016 Low Low For ReviewLow Unclear OnlyLow High 13 Varghese 2009 Unclear Low Low Unclear High High 14 Veleba 2015 Low Unclear High Unclear High High 15 16 VERTIS Asia Low Low Low Low Low Low 17 VERTIS FACTORIAL Low Low Low Unclear Low Low 18 VERTIS-MET Low Low Low Unclear Low Low 19 VERTIS MONO Low Low Low Unclear Low Low 20 21 VERTIS RENAL Low Low Low Unclear High Low 22 VERTIS SITA2 Low Low Low Unclear Low Low 23 VERTIS SU Low Low Low Unclear High Low 24 25 Vianna 2017 Low Unclear High Low Low High 26 VICTORY Unclear Low Low Unclear Low Low 27 VISION Unclear Unclear High Unclear Low Low 28 VIVIDD Low Low Low Low High Low 29 30 Vongthavaravat 2002 Unclear Unclear High Unclear High High 31 Wajcberg 2007 Unclear Low Low Unclear Unclear High 32 Wang 2005 Unclear Unclear High Unclear Low Low 33 Wang 2013 Low Unclear High Unclear Low High 34 35 Wang 2015 Unclear Unclear High Unclear High High 36 Wang 2016 Low Low Low Unclear Low Low 37 Wang 2016a Unclear Unclear Unclear Unclear Unclear High 38 Wang 2017 Low Low Low Unclear High Low 39 40 Wang 2019 Low Low High Low Low Low 41 Watanabe 2005 Unclear Unclear High Unclear Low High 42 Wolever 2000 Unclear Unclear High Unclear High High 43 44 Wolffenbuttel 1999 Unclear Low Low Unclear High High 45 Wolffenbuttel 2000 Unclear Low Low Unclear High High 46 Wong 2005 Low Unclear High Unclear Low Low 47 Wu 2014 Unclear Unclear High Unclear Unclear High 48 49 Wu 2015 Low Low Low Unclear Low High 50 Xiao 2015 Low Low High Unclear Low High 51 Xiao 2016 Low Low High Unclear Low Low 52 Xu 2017 Low Low High Unclear Low Low 53 54 Yakibu 2017 Low Unclear High Unclear Low High 55 Yale 2013 Low Low Low Unclear High High 56 Yamakage 2019 Low Low High Unclear Low High 57 Yamamoto 2018 Unclear Unclear High Unclear High High 58 59 Yamanouchi 2005 Low Low High Unclear Low High 60 Yamasaki 2005 Unclear Unclear High Unclear Unclear High Yan 2019 Low Unclear High Unclear Low High

https://mc.manuscriptcentral.com/bmj Page 139 of 244 BMJ

1 2 Yang 2002 Unclear Low Low Unclear Unclear High 3 Yang 2011 Low Low Low Unclear High Low 4 Yang 2012 Low Low Low Unclear Low Low 5 6 Yang 2013 Unclear Low Low Unclear Low High 7 Yang 2015 Unclear Low Low Unclear High High 8 Yang 2015a Unclear Low Low Unclear Low Low 9 Yang 2016 Low Low Low Unclear High Low 10 11 Yang 2018 Low Low Low Unclear Low Low 12 Yee 2010 Confidential:Unclear Low For ReviewLow Unclear OnlyLow Low 13 Yki-Järvinen 1999 Low Unclear High Unclear Low High 14 Yki-Järvinen 2013 Low Low Low Unclear Low Low 15 16 Yoon 2011 Unclear Low Low High High Low 17 Yoon 2011a Unclear Low Low Unclear High High 18 Yuan 2012 Low Unclear High Unclear Low High 19 Zang 2016 Low Low Low Unclear Low Low 20 21 ZEUS II Low Low Low Unclear Low Low 22 Zhang 2020 Low Low High Unclear Unclear High 23 Zheng 2019 Unclear Unclear Unclear Unclear Unclear High 24 25 Zhu 2003 Unclear Unclear Unclear Unclear High Low 26 Zib 2007 Unclear Unclear High Unclear Low High 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

https://mc.manuscriptcentral.com/bmj BMJ Page 140 of 244

1 2 3 4 5 6 7 8 9 10 11 12 Confidential: For Review Only 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

https://mc.manuscriptcentral.com/bmj Page 141 of 244 BMJ

1 2 3 Appendix 4 Network plots for each outcome 4 5 The size of the circle in each network is proportional to the number of participants randomly assigned 6 7 to the treatment comparison. The width of each line is proportional to the number of trials comparing 8 9 the two connected treatments. When a line is absent, this indicates that there were no head-to-head 10 trials of the corresponding treatments reporting the outcome of interest. The number provided for each 11 12 treatmentConfidential: class indicates the number of patients For assigned Review to the treatment in Only the network. 13 14 15 Figure 1 Network plot for all-cause mortality 16 17 18 19 20 DPP-4 inhibitor 21 SGLT-2 inhibitor 51,772 22 35,813 23 24 Thiazolidinedione 25 GLP-1 receptor 17,836 26 agonist 27 49,103 28 Sulfonylurea 29 21,719 30 Basal insulin 31 3770 32 33 Metformin 34 5285 35 36 Basal bolus 37 insulin 38 758 39 Glitinide 40 2015 41 42 Bolus insulin Alpha glucosidase 43 551 inhibitor 44 1455 45 46 Standard therapy 47 Placebo 4222 48 89,698 49 50 51 52 53 54 55 56 57 58 59 60

https://mc.manuscriptcentral.com/bmj BMJ Page 142 of 244

1 2 3 Figure 2 Network plot for cardiovascular mortality 4 5 6 7 8 DPP-4 inhibitor 9 10 SGLT-2 inhibitor 40,076 11 27,637 12 Confidential: For Review Only Thiazolidinedione 13 14 GLP1-receptor 14,077 15 agonist 16 38,823 17 Sulfonylurea 18 12,935 19 Basal insulin 20 1996 21 22 23 Metformin 24 1714 25 26 Basal bolus 27 Insulin 28 271 29 Glitinide 30 31 936 32 Bolus insulin 33 551 Alpha glucosidase 34 inhibitor 35 735 36 37 Placebo Standard therapy 38 80,119 3809 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

https://mc.manuscriptcentral.com/bmj Page 143 of 244 BMJ

1 2 3 Figure 3 Network plot for nonfatal myocardial infarction 4 5 6 7 8 DPP-4 inhibitor 9 SGLT-2 inhibitor 49,144 10 33,402 11 12 GLP1-reConfidential:ceptor agonist For Review TOnlyhiazolidinedione 13 45,354 14,884 14 15 16 Sulfonylurea 17 18,661 18 Basal insulin 19 3255 20 21 22 Metformin 23 4250 24 25 Basal bolus insulin 26 464 27 Glitinide 28 1006 29 30 Bolus insulin Alpha glucosidase 31 888 32 inhibitor 33 1225 34 Placebo Standard therapy 35 84,546 5174 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

https://mc.manuscriptcentral.com/bmj BMJ Page 144 of 244

1 2 3 Figure 4 Network plot for nonfatal stroke 4 5 6 7 SGLT-2 inhibitor DPP-4 inhibitor 8 46,637 9 34,741 10 Thiazolidinedione 11 GLP1-receptor agonist 15,032 12 Confidential: For Review Only 45,934 13 14 Sulfonylurea 15 19,245 16 17 Basal insulin 18 3849 19 20 21 Metformin 22 3166 23 24 Basal bolus insulin 25 369 26 Glitinide 27 754 28 29 Bolus insulin 30 398 Alpha glucosidase 31 inhibitor 32 551 Standard therapy 33 Placebo 4102 34 85,381 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

https://mc.manuscriptcentral.com/bmj Page 145 of 244 BMJ

1 2 3 Figure 5 Network plot for kidney failure 4 5 6 7 Thiazolidinedione 8 DPP-4 inhibitor 601 9 24,853 10 11 12 Confidential: For Review Only 13 Sulfonylurea 14 SGLT-2 inhibitor 2783 15 16,357 16 17 18 19 20 21 Metformin 22 237 23 24 25 26 GLP1-receptor 27 28 agonist 29 6768 30 31 Placebo 32 Basal insulin 37,964 33 194 Bolus insulin 34 281 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

https://mc.manuscriptcentral.com/bmj BMJ Page 146 of 244

1 2 3 Figure 6 Network plot for hospitalisation for heart failure 4 5 6 7 DPP-4 inhibitor 8 38,133 9 Thiazolidinedione 10 15,902 11 SGLT-2 inhibitor 12 Confidential:30,876 For Review Only 13 Sulfonylurea 14 17,600 15 16 GLP-1 receptor 17 agonist 18 39,493 19 20 Metformin 21 2422 22 23 24 Basal insulin 25 2015 26 Alpha glucosidase 27 inhibitor 28 36 29 Bolus insulin 30 312 31 Standard therapy 32 3507 33 Placebo 34 79,862 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

https://mc.manuscriptcentral.com/bmj Page 147 of 244 BMJ

1 2 3 Figure 7 Network plot for severe hypoglycaemia 4 5 6 7 8 DPP-4 inhibitor 9 SGLT-2 inhibitor 48,479 10 26,791 11 12 Confidential: For Review Only Thiazolidinedione 13 GLP1-receptor 11,425 14 agonist 15 49,026 16 Sulfonylurea 17 15,986 18 19 Basal insulin 20 21 4774 22 Metformin 23 2921 24 25 Basal bolus 26 insulin 27 157 28 Glitinide 29 819 30 31 Bolus insulin 32 649 33 Alpha glucosidase 34 inhibitor 35 758 36 Placebo Standard therapy 37 82,465 4646 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

https://mc.manuscriptcentral.com/bmj BMJ Page 148 of 244

1 2 3 Figure 8 Network plot for blindness 4 5 6 7 8 SGLT-2 inhibitor 9 4687 10 11 12 Confidential: For Review Only 13 14 15 16 17 18 GLP1-receptor 19 agonist DPP-4 inhibitor 20 19,760 10,826 21 22 23 24 25 26 27 28 29 30 31 32 33 Placebo 34 32,948 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

https://mc.manuscriptcentral.com/bmj Page 149 of 244 BMJ

1 2 3 Figure 9 Network plot for health-related quality of life 4 5 6 7 DPP-4 inhibitor 8 1163 9 10 SGLT-2 inhibitor 11 624 Thiazolidinedione 12 Confidential: For Review33 0Only 13 14 15 GLP1-receptor Sulfonylurea 16 agonist 82 17 5025 18 19 20 21 Basal insulin Metformin 22 293 23 244 24 25 26 27 28 Bolus insulin Glitinide 29 559 374 30 31 32 Alpha glucosidase Placebo 33 Inhibitor 2144 34 Standard 230 35 therapy 36 383 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

https://mc.manuscriptcentral.com/bmj BMJ Page 150 of 244

1 2 3 Figure 10 Network plot for body weight 4 5 6 7 DPP-4 inhibitor 8 SGLT-2 inhibitor 49,369 9 10 20,000 11 Thiazolidinedione 12 GLPConfidential:-1 receptor For Review13 ,Only459 13 agonist 14 42,582 15 Sulfonylurea 16 18,793 17 18 Basal insulin 19 6651 20 21 Metformin 22 8755 23 24 Basal bolus 25 Insulin 26 711 27 Glitinide 28 1644 29 Bolus insulin 30 Alpha glucosidase 488 31 inhibitor 32 2654 33 34 Placebo Standard therapy 35 52,311 3775 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

https://mc.manuscriptcentral.com/bmj Page 151 of 244 BMJ

1 2 3 Figure 11 Network plot for amputation 4 5 6 7 8 SGLT-2 inhibitor 9 11,346 DPP-4 inhibitor 10 7467 11 12 Confidential: For Review Only 13 14 15 16 17 18 GLP1-receptor Thiazolidinedione 19 agonist 4825 20 3031 21 22 23 24 25 26 27 28 29 Standard therapy Placebo 2227 30 24,073 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

https://mc.manuscriptcentral.com/bmj BMJ Page 152 of 244

1 2 3 Figure 12 Network plot for neuropathic pain 4 5 6 7 8 GLP-1 receptor 9 agonist 10 4668 11 12 Confidential: For Review Only 13 14 15 16 17 18 DPP-4 inhibitor 19 7266 20 21 22 23 24 25 26 27 Placebo 28 11,946 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

https://mc.manuscriptcentral.com/bmj Page 153 of 244 BMJ

1 2 3 Figure 13 Network plot for diabetic ketoacidosis 4 5 6 7 8 DPP-4 inhibitor 9 19,917 10 SGLT-2 inhibitor 11 18,171 12 Confidential: For Review Only Sulfonylurea 13 2615 14 15 16 17 18 19 20 GLP-1 receptor 21 agonist Metformin 22 13,113 237 23 24 25 26 27 28 29 30 Glitinide 31 Placebo 551 32 43,485 33 34 35 Alpha glucosidase 36 Inhibitor 37 545 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

https://mc.manuscriptcentral.com/bmj BMJ Page 154 of 244

1 2 3 Figure 14 Network plot for serious hyperglycaemia 4 5 6 7 8 DPP-4 inhibitor 9 32,008 10 Thiazolidinedione 11 2709 12 Confidential:SGLT-2 inhibitor For Review Only 13 11,813 14 15 Sulfonylurea 16 4004 17 GLP-1 receptor 18 agonist 19 19,918 20 21 22 Metformin 23 1098 24 25 Basal insulin 26 1600 27 28 Alpha glucosidase 29 Inhibitor 30 Basal bolus insulin 372 31 248 32 Standard 33 therapy 34 Placebo 2655 35 46,932 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

https://mc.manuscriptcentral.com/bmj Page 155 of 244 BMJ

1 2 3 Figure 15 Network plot for genital infection 4 5 6 7 8 Thiazolidinedione 9 32 10 DPP-4 inhibitor 11 2063 12 Confidential: For Review OnlySulfonylurea 13 2419 14 15 16 17 SGLT-2 inhibitor 18 35,259 19 20 Metformin 21 842 22 23 24 25 26 27 28 GLP-1 receptor 29 agonist Standard therapy 30 10,307 229 31 32 33 Placebo 34 29,886 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

https://mc.manuscriptcentral.com/bmj BMJ Page 156 of 244

1 2 3 Figure 16 Network plot for Fournier gangrene 4 5 6 7 8 SGLT-2 inhibitor 9 16,106 10 11 12 Confidential: For Review Only 13 14 15 16 17 18 19 DPP-4 inhibitor 20 6195 21 22 23 24 25 26 27 28 29 30 31 Placebo 32 19,598 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

https://mc.manuscriptcentral.com/bmj Page 157 of 244 BMJ

1 2 3 Figure 17 Network plot for severe gastrointestinal events 4 5 6 7 8 9 10 Thiazolidinedione 11 DPP-4 inhibitor 1755 12 Confidential:2615 For Review Only 13 14 15 Sulfonylurea 16 1935 17 GLP-1 receptor 18 agonist 19 15,170 20 21 22 Metformin 23 1714 24 25 26 27 Basal insulin 28 597 29 30 Alpha glucosidase 31 inhibitor 32 54 33 Placebo 34 12,760 Standard therapy 35 429 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

https://mc.manuscriptcentral.com/bmj BMJ Page 158 of 244

1 2 3 Figure 18 Network plot for pancreatic cancer 4 5 6 7 8 9 10 DPP-4 inhibitor 27,855 11 SGLT-2 inhibitor 12 Confidential:9222 For Review Only 13 14 15 16 GLP1-receptor 17 agonist 18 33,915 19 20 21 22 Sulfonylurea 23 5654 24 25 26 Basal insulin 27 1853 28 29 30 31 Bolus insulin 32 312 33 34 Placebo 35 54,657 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

https://mc.manuscriptcentral.com/bmj Page 159 of 244 BMJ

1 2 3 Figure 19 Network plot for pancreatitis 4 5 6 7 8 DPP-4 inhibitor 9 33,101 10 11 SGLT-2 inhibitor 12 Confidential:12,512 For Review Only Thiazolidinedione 13 4755 14 GLP-1 receptor 15 agonist 16 40,784 17 Sulfonylurea 18 6472 19 20 21 22 Basal insulin 23 1710 Metformin 24 246 25 26 27 28 29 Basal bolus insulin Glitinide 30 138 551 31 32 33 Alpha glucosidase 34 inhibitor 35 Placebo Standard therapy 36 60,398 2386 545 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

https://mc.manuscriptcentral.com/bmj BMJ Page 160 of 244

1 2 3 Figure 20 Network plot for glycated haemoglobin A1C 4 5 6 DPP-4 inhibitor 7 SGLT-2 inhibitor 49,420 8 21,162 9 Thiazolidinedione 10 18,958 11 GLP-1 receptor 12 aConfidential:gonist For Review Only 13 44,105 Sulfonylurea 14 19,419 15 16 Basal insulin 17 6753 18 19 Metformin 20 8126 21 22 23 Basal bolus insulin 24 606 Glitinide 25 2622 26 27 Bolus insulin 28 1151 Alpha glucosidase 29 inhibitor 30 4305 31 32 Placebo Standard therapy 33 53,679 6024 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

https://mc.manuscriptcentral.com/bmj Page 161 of 244 BMJ

1 2 3 Appendix 5 Evaluations of network inconsistency and heterogeneity 4 5 6 7 Network 8 Global consistency heterogeneity 9 Clinical outcome Chi square P value Tau squared 10 11 All-cause mortality 21.1 0.999 0.08 12 Confidential: For Review Only Cardiovascular mortality 13.1 0.983 0.11 13 14 Nonfatal myocardial infarction 22.0 0.997 0.03 15 Nonfatal stroke 15.7 0.999 <0.001 16 17 Kidney failure 0.15 0.985 <0.001 18 Hospitalisation for heart failure 10.3 0.998 <0.001 19 20 Severe hypoglycaemia 24.6 0.743 0.15 21 Blindness 0.00 0.999 Fixed model 22 23 Eye disease requiring intervention … … … 24 Health-related quality of life 13.5 0.04 0.065 25 26 Body weight 82.4 0.129 <0.001 27 Amputation 0.02 0.897 <0.001 28 Neuropathic pain* … … Fixed model 29 30 Diabetic ketoacidosis 1.66 0.645 <0.001 31 Serious hyperglycaemia 14.6 0.798 0.61 32 33 Genital infection 4.17 0.384 0.09 34 Fournier gangrene* … … <0.001 35 36 Severe gastrointestinal events 3.00 0.558 0.69 37 Pancreatic cancer 3.05 0.549 <0.001 38 39 Pancreatitis 11.1 0.973 0.07 40 HbA1C 86.6 0.522 0.32 41 42 *No source of inconsistency 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

https://mc.manuscriptcentral.com/bmj BMJ Page 162 of 244

1 2 3 Appendix 6 Direct, indirect and network treatment estimates 4 5 Table 1 All-cause mortality 6 7 8 Direct Indirect Incoherence Network 9 Intervention Comparator estimate estimate (p value) estimate 10 0.77 1.13 0.77 SGLT-2 inhibitor v Placebo 0.27 11 (0.71,0.84) (0.57,2.23) (0.71,0.84) 12 GLP-1 receptorConfidential: For0.88 Review0.81 Only 0.88 v Placebo 0.76 13 agonist (0.83,0.94) (0.49,1.34) (0.82,0.94) 14 GLP-1 receptor 0.62 0.88 0.88 SGLT-2 inhibitor v 0.74 15 agonist (0.08,5.04) (0.79,0.98) (0.79,0.98) 16 1.51 0.96 1.03 17 Metformin v Placebo 0.35 (0.64,3.57) (0.67,1.38) (0.74,1.43) 18 1.83 1.10 1.11 19 Sulfonylurea v Placebo 0.56 20 (0.34,9.77) (0.95,1.28) (0.96,1.28) 0.94 1.17 1.01 21 Thiazolidinedione v Placebo 0.24 22 (0.77,1.16) (0.87,1.56) (0.86,1.19) 1.02 0.92 1.01 23 DPP-4 inhibitor v Placebo 0.51 24 (0.95,1.05) (0.70,1.22) (0.94,1.08) 25 Alpha 1.21 0.40 0.89 26 glucosidase v Placebo 0.46 (0.31,4.68) (0.04,4.36) (0.30,2.62) 27 inhibitor 28 1.72 1.61 1.63 Glitinide v Placebo 0.96 29 (0.15,19.01) (0.44,5.84) (0.52,5.08) 30 0.90 0.97 0.93 31 Metformin v Sulfonylurea 0.81 (0.58,1.38) (0.58,1.62) (0.67,1.29) 32 0.94 1.14 1.02 33 Metformin v Thiazolidinedione 0.58 34 (0.60,1.46) (0.68,1.89) (0.73,1.42) 0.95 1.02 1.02 35 Metformin v DPP-4 inhibitor 0.91 36 (0.28,3.25) (0.72,1.45) (0.73,1.42) 1.86 1.31 1.33 37 Metformin v SGLT-2 inhibitor 0.72 38 (0.29,11.83) (0.93,1.86) (0.95,1.87) 39 GLP-1 receptor 3.29 1.15 1.17 Metformin v 0.50 40 agonist (0.16,69.08) (0.82,1.62) (0.84,1.64) 41 3.16 0.50 0.63 Metformin v Glitinide 0.27 42 (0.14,70.80) (0.14,1.73) (0.20,2.03) 43 0.97 1.21 1.10 Sulfonylurea v Thiazolidinedione 0.24 44 (0.74,1.28) (0.95,1.56) (0.91,1.32) 45 1.14 0.95 1.10 46 Sulfonylurea v DPP-4 inhibitor 0.28 (0.98,1.33) (0.71,1.27) (0.96,1.26) 47 1.14 1.44 1.43 48 Sulfonylurea v SGLT-2 inhibitor 0.61 49 (0.46,2.81) (1.22,1.71) (1.21,1.69) GLP-1 receptor 1.07 1.27 1.26 50 Sulfonylurea v 0.71 51 agonist (0.45,2.56) (1.08,1.49) (1.08,1.48) Basal bolus 0.15 2.58 1.39 52 Sulfonylurea v 0.11 53 insulin (0.01,3.31) (0.50,13.17) (0.33,5.87) 54 Alpha 1.56 1.03 1.24 55 Sulfonylurea v glucosidase 0.71 (0.31,7.85) (0.24,4.41) (0.42,3.66) 56 inhibitor 57 0.54 0.90 0.68 Sulfonylurea v Glitinide 0.67 58 (0.12,2.52) (0.16,5.16) (0.22,2.11) 59 60

https://mc.manuscriptcentral.com/bmj Page 163 of 244 BMJ

1 2 3 Direct Indirect Incoherence Network 4 Intervention Comparator estimate estimate (p value) estimate 5 0.46 1.01 1.00 6 Thiazolidinedione v DPP-4 inhibitor 0.33 (0.09,2.23) (0.85,1.20) (0.84,1.19) 7 GLP-1 receptor 2.20 1.14 1.15 8 Thiazolidinedione v 0.50 9 agonist (0.34,14.45) (0.96,1.37) (0.96,1.37) 0.87 0.83 0.87 10 Thiazolidinedione v Standard therapy 0.94 11 (0.69,1.10) (0.24,2.85) (0.69,1.09) 0.66 0.61 0.62 12 ThiazolidinedioneConfidential:v Glitinide For Review Only0.96 13 (0.03,16.38) (0.18,2.08) (0.20,1.94) 14 0.42 1.32 1.31 DPP-4 inhibitor v SGLT-2 inhibitor 0.09 15 (0.12,1.54) (1.18,1.47) (1.17,1.45) 16 GLP-1 receptor 1.21 1.15 1.15 17 DPP-4 inhibitor v 0.88 agonist (0.64,2.30) (1.05,1.26) (1.05,1.26) 18 1.83 0.83 0.89 19 DPP-4 inhibitor v Basal insulin 0.42 20 (0.30,11.29) (0.46,1.48) (0.51,1.55) 21 1.21 0.86 0.87 DPP-4 inhibitor v Standard therapy 0.64 22 (0.29,5.10) (0.64,1.14) (0.65,1.15) 23 GLP-1 receptor 0.70 1.26 0.77 v Basal insulin 0.45 24 agonist (0.39,1.29) (0.32,4.97) (0.45,1.34) 25 GLP-1 receptor Basal bolus 1.33 0.87 1.10 v 0.77 26 agonist insulin (0.19,9.12) (0.10,7.51) (0.26,4.63) 27 GLP-1 receptor 2.98 0.25 0.88 28 v Bolus insulin 0.29 agonist (0.12,73.46) (0.01,6.55) (0.09,8.68) 29 GLP-1 receptor 0.30 0.76 0.75 30 v Standard therapy 0.59 31 agonist (0.01,8.83) (0.57,1.01) (0.57,1.00) 7.03 0.83 1.14 32 Basal insulin v Bolus insulin 0.22 33 (0.36,136.83) (0.15,4.65) (0.12,11.21) 0.32 1.02 0.98 34 Basal insulin v Standard therapy 0.49 35 (0.01,7.99) (0.54,1.91) (0.53,1.81) 36 Alpha 4.09 1.44 0.55 37 glucosidase v Glitinide 0.56 (0.18,91.46) (0.26,7.85) (0.12,2.42) 38 inhibitor 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

https://mc.manuscriptcentral.com/bmj BMJ Page 164 of 244

1 2 3 Table 2 Cardiovascular mortality 4 5 6 Direct Indirect Incoherence Network 7 Intervention Comparator estimate estimate (p value) estimate 8 0.82 1.27 0.83 SGLT-2 inhibitor v Placebo 0.42 9 (0.73,0.93) (0.44,3.62) (0.73,0.93) 10 GLP-1 receptor 0.88 0.72 0.88 v Placebo 0.70 11 agonist (0.80,0.97) (0.27,1.93) (0.80,0.96) 12 Confidential:GLP-1 receptor For0.99 Review0.94 Only 0.94 SGLT-2 inhibitor v 0.97 13 agonist (0.06,15.9) (0.81,1.10) (0.81,1.10) 14 3.05 0.77 0.94 Metformin v Placebo 0.18 15 (0.48,19.3) (0.37,1.64) (0.47,1.88) 16 3.47 0.99 1.00 17 Sulfonylurea v Placebo 0.43 18 (0.15,78.4) (0.78,1.26) (0.79,1.27) 0.94 1.08 0.96 19 Thiazolidinedione v Placebo 0.68 20 (0.72,1.23) (0.60,1.96) (0.76,1.22) 0.98 0.98 0.98 21 DPP-4 inhibitor v Placebo 0.99 22 (0.88,1.10) (0.59,1.62) (0.88,1.09) 23 Alpha glucosidase 1.54 0.17 0.99 v Placebo 0.39 24 inhibitor (0.24,9.80) (0.00,13.4) (0.21,4.69) 25 0.96 0.58 0.65 Glitinide v Placebo 0.82 26 (0.02,48.8) (0.08,4.15) (0.11,3.73) 27 0.77 1.25 0.94 Metformin v Sulfonylurea 0.50 28 (0.31,1.90) (0.43,3.60) (0.48,1.87) 29 0.76 1.03 0.98 30 Metformin v Thiazolidinedione 0.77 (0.12,4.86) (0.47,2.24) (0.48,2.00) 31 0.34 1.01 0.96 32 Metformin v DPP-4 inhibitor 0.51 33 (0.01,8.31) (0.50,2.05) (0.48,1.92) 1.86 34 1.05 1.14 Metformin v SGLT-2 inhibitor (0.29 to 0.58 35 (0.50,2.24) (0.57,2.29) 36 11.85) 37 3.54 0.99 1.46 Metformin v Glitinide 0.49 38 (0.16,80.2) (0.12,8.37) (0.24,8.99) 39 0.85 1.11 1.04 40 Sulfonylurea v Thiazolidinedione 0.47 41 (0.45,1.60) (0.77,1.59) (0.76,1.42) 42 1.03 0.97 1.02 Sulfonylurea v DPP-4 inhibitor 0.86 43 (0.81,1.31) (0.54,1.73) (0.82,1.27) 44 0.84 1.22 1.21 Sulfonylurea v SGLT-2 inhibitor 0.34 45 (0.10,6.87) (0.93,1.59) (0.93,1.57) 46 Basal bolus 0.26 0.76 0.45 Sulfonylurea v 0.65 47 insulin (0.01,6.79) (0.03,19.0) (0.05,4.42) 48 Alpha 49 1.96 0.56 1.01 Sulfonylurea v glucosidase 0.44 50 (0.20,19.2) (0.07,4.79) (0.21,4.81) inhibitor 51 1.99 1.31 1.55 52 Sulfonylurea v Glitinide 0.82 53 (0.12,32.1) (0.14,12.6) (0.27,8.93) 0.18 1.00 0.98 54 Thiazolidinedione v DPP-4 inhibitor 0.14 55 (0.02,1.72) (0.77,1.30) (0.76,1.27) 0.87 0.64 0.86 56 Thiazolidinedione v Standard therapy 0.76 57 (0.61,1.23) (0.10,4.25) (0.61,1.21) 58 0.66 2.11 1.49 Thiazolidinedione v Glitinide 0.56 59 (0.03,16.4) (0.26,17.2) (0.26,8.64) 60 DPP-4 inhibitor v SGLT-2 inhibitor 0.29 1.20 0.12 1.19

https://mc.manuscriptcentral.com/bmj Page 165 of 244 BMJ

1 2 3 Direct Indirect Incoherence Network 4 Intervention Comparator estimate estimate (p value) estimate 5 (0.05,1.72) (1.03,1.41) (1.02,1.39) 6 GLP-1 receptor 1.55 1.11 1.12 7 DPP-4 inhibitor v 0.57 8 agonist (0.50,4.80) (0.97,1.28) (0.97,1.29) 0.33 0.93 0.87 9 DPP-4 inhibitor v Basal insulin 0.54 10 (0.01,8,21) (0.41,2.13) (0.39,1.94) 1.00 0.87 0.88 11 DPP-4 inhibitor v Standard therapy 0.91 12 Confidential: For(0.10,9.68) Review(0.56,1.35) Only (0.57,1.34) 13 GLP-1 receptor 0.71 1.34 0.78 v Basal insulin 0.57 14 agonist (0.30,1.67) (0.17,10.3) (0.35,1.72) 15 GLP-1 receptor Basal bolus 0.66 0.23 0.39 v 0.65 16 agonist insulin (0.03,16.4) (0.01,5.99) (0.04,3.88) 17 GLP-1 receptor 2.98 0.25 1.13 v Bolus insulin 0.29 18 agonist (0.12,73.6) (0.01,6.69) (0.11,11.25) 19 0.34 4.12 0.50 20 Basal insulin v Bolus insulin 0.29 (0.01,8.38) (0.15,112) (0.04,5.68) 21 0.33 1.10 1.00 22 Basal insulin v Standard therapy 0.48 23 (0.01,8.25) (0.44,2.76) (0.41,2.43) 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

https://mc.manuscriptcentral.com/bmj BMJ Page 166 of 244

1 2 3 Table 3 Nonfatal myocardial infarction 4 5 6 Direct Indirect Incoherence Network 7 Intervention Comparator estimate estimate (p value) estimate 8 0.88 0.73 0.87 SGLT-2 inhibitor v Placebo 0.67 9 (0.79,0.97) (0.31,1.70) (0.79,0.97) 10 GLP-1 receptor 0.92 0.97 0.92 v Placebo 0.86 11 agonist (0.85,0.99) (0.50,1.88) (0.85,0.99) 12 Confidential:GLP-1 receptor For1.98 Review0.95 Only 0.95 SGLT-2 inhibitor v 0.55 13 agonist (0.18,22.0) (0.84,1.08) (0.84,1.08) 14 1.16 1.06 1.16 Metformin v Placebo 0.99 15 (0.66,2.04) (0.72,1.86) (0.81,1.66) 16 0.49 1.11 1.09 17 Sulfonylurea v Placebo 0.22 18 (0.13,1.82) (0.89,1.39) (0.88,1.35) 1.19 1.19 1.19 19 Thiazolidinedione v Placebo 0.99 20 (0.45,3.14) (0.85,1.67) (0.87,1.64) 1.01 1.02 1.01 21 DPP-4 inhibitor v Placebo 0.99 22 (0.92.1.12) (0.67,1.55) (0.92,1.11) 23 Alpha 0.15 0.41 0.19 24 glucosidase v Placebo 0.58 (0.03,0.86) (0.02,8.34) (0.04,0.88) 25 inhibitor 26 1.33 0.14 0.29 Glitinide v Placebo 0.22 27 (0.06,27.94) (0.02,1.14) (0.05,1.63) 28 1.16 0.99 1.06 Metformin v Sulfonylurea 0.67 29 (0.68,1.97) (0.59,1.63) (0.74,1.53) 30 0.79 1.24 0.97 31 Metformin v Thiazolidinedione 0.25 32 (0.46,1.34) (0.70,2.20) (0.66,1.44) 2.10 1.10 1.14 33 Metformin v DPP-4 inhibitor 0.42 34 (0.48,8.75) (0.76,1.60) (0.79,1.64) 1.43 1.32 1.32 35 Metformin v SGLT-2 inhibitor 0.96 36 (0.06,35.54) (0.91,1.93) (0.91,1.93) 37 GLP-1 receptor 0.79 1.28 1.26 Metformin v 0.71 38 agonist (0.07,9.43) (0.88,1.85) (0.87,1.83) 39 3.06 1.10 1.12 Metformin v Standard therapy 0.54 40 (0.12,76.98) (0.67,1.79) (0.69,1.82) 41 0.97 0.77 0.91 Sulfonylurea v Thiazolidinedione 0.49 42 (0.70,1.35) (0.43,1.36) (0.69,1.22) 43 1.05 1.18 1.07 Sulfonylurea v DPP-4 inhibitor 0.64 44 (0.83,1.31) (0.76,1.81) (0.88,1.31) 45 1.70 1.23 1.25 46 Sulfonylurea v SGLT-2 inhibitor 0.62 47 (0.49,5.98) (0.97.1.57) (0.98,1.58) GLP-1 receptor 0.94 1.20 1.19 48 Sulfonylurea v 0.66 49 agonist (0.33,2.69) (0.95,1.52) (0.94,1.49) Alpha 50 2.50 7.28 5.63 Sulfonylurea v glucosidase 0.53 51 (0.13,46.99) (1.31,40.49) (1.23,25.71) 52 inhibitor 53 6.06 1.51 3.77 Sulfonylurea v Glitinide 0.42 54 (0.76,48.30) (0.09,25.06) (0.67,21.30) 55 1.16 1.17 1.17 Thiazolidinedione v DPP-4 inhibitor 0.99 56 (0.31,4.42) (0.85,1.62) (0.86,1.61) 57 GLP-1 receptor 3.09 1.26 1.30 Thiazolidinedione v 0.32 58 agonist (0.54,17.86) (0.90,1.75) (0.94,1.80) 59 Thiazolidinedione v Basal insulin 3.05 1.12 0.55 1.18 60

https://mc.manuscriptcentral.com/bmj Page 167 of 244 BMJ

1 2 3 Direct Indirect Incoherence Network 4 Intervention Comparator estimate estimate (p value) estimate 5 (0.12,76.62) (0.53,2.37) (0.57,2.45) 6 1.17 1.00 1.15 7 Thiazolidinedione v Standard therapy 0.76 8 (0.84,1.65) (0.39,2.56) (0.84,1.58) 1.51 1.16 1.16 9 DPP-4 inhibitor v SGLT-2 inhibitor 0.66 10 (0.47,4.87) (1.01,1.33) (1.01,1.33) GLP-1 receptor 0.92 1.11 1.11 11 DPP-4 inhibitor v 0.67 12 Confidential:agonist For(0.39,2.19) Review(0.98,1.26) Only (0.98,1.25) 13 2.70 0.96 1.00 DPP-4 inhibitor v Basal insulin 0.53 14 (0.11,66.7) (0.47,1.94) (0.50,2.00) 15 0.54 1.08 0.98 DPP-4 inhibitor v Standard therapy 0.26 16 (0.18,1.65) (0.69,1.70) (0.65,1.49) 17 GLP-1 receptor 0.83 1.07 0.91 v Basal insulin 0.72 18 agonist (0.35,1.95) (0.34,3.32) (0.46,1.80) 19 GLP-1 receptor Basal bolus 2.84 2.65 2.75 20 v 0.98 agonist insulin (0.12,70.14) (0.10,70.39) (0.28,27.19) 21 GLP-1 receptor 0.88 0.67 0.78 22 v Bolus insulin 0.82 23 agonist (0.18,4.20) (0.13,3.59) (0.25,2.43) GLP-1 receptor 2.88 0.87 0.89 24 v Standard therapy 0.47 25 agonist (0.11,74.24) (0.56,1.34) (0.58,1.36) Basal bolus 2.92 3.14 3.03 26 Basal insulin v 0.98 27 insulin (0.12,72.18) (0.12,83.43) (0.31,29.98) 28 0.76 0.99 0.86 Basal insulin v Bolus insulin 0.82 29 (0.17,3.45) (0.18,5.57) (0.27,2.66) 30 1.35 0.90 0.98 Basal insulin v Standard therapy 0.67 31 (0.25,7.22) (0.38,2.13) (0.46,2.10) 32 Alpha 0.99 0.56 0.67 33 glucosidase v Glitinide 0.81 (0.02,49.96) (0.04,8.53) (0.07,6.31) 34 inhibitor 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

https://mc.manuscriptcentral.com/bmj BMJ Page 168 of 244

1 2 3 Table 4 Nonfatal stroke 4 5 6 Direct Indirect Incoherence Network 7 Intervention Comparator estimate estimate (p value) estimate 8 1.01 1.18 1.01 SGLT-2 inhibitor v Placebo 0.70 9 (0.89,1.14) (0.53,2.65) (0.89,1.14) 10 GLP-1 receptor 0.83 1.16 0.84 v Placebo 0.39 11 agonist (0.75,0.92) (0.55,2.43) (0.76,0.93) 12 Confidential:GLP-1 receptor For0.42 Review1.22 Only 1.20 SGLT-2 inhibitor v 0.23 13 agonist (0.08,2.36) (1.04,1.43) (1.03,1.41) 14 0.95 1.03 1.01 Metformin v Placebo 0.88 15 (0.38,2.39) (0.64,1.66) (0.66,1.55) 16 0.63 1.12 1.10 17 Sulfonylurea v Placebo 0.51 18 (0.12,3.37) (0.87,1.43) (0.87,1.41) 0.90 0.99 0.95 19 Thiazolidinedione v Placebo 0.74 20 (0.59,1.37) (0.66,1.49) (0.71,1.27) 0.94 0.77 0.93 21 DPP-4 inhibitor v Placebo 0.36 22 (0.83,1.08) (0.50,1.18) (0.82,1.05) 23 0.88 0.98 0.92 Metformin v Sulfonylurea 0.81 24 (0.52,1.48) (0.49,1.94) (0.61,1.38) 25 1.29 0.90 1.07 Metformin v Thiazolidinedione 0.43 26 (0.68,2.44) (0.48,1.67) (0.69,1.66) 27 1.21 1.08 1.09 Metformin v DPP-4 inhibitor 0.80 28 (0.27,5.32) (0.69,1.68) (0.71,1.67) 29 0.67 1.01 1.00 30 Metformin v SGLT-2 inhibitor 0.80 (0.03,16.41) (0.65,1.58) (0.64,1.56) 31 1.10 1.26 1.17 32 Sulfonylurea v Thiazolidinedione 0.65 33 (0.74,1.63) (0.81,1.97) (0.87,1.57) 1.26 1.02 1.19 34 Sulfonylurea v DPP-4 inhibitor 0.43 35 (0.97,1.63) (0.65,1.59) (0.95,1.49) 36 0.45 1.14 1.09 Sulfonylurea v SGLT-2 inhibitor 0.16 37 (0.13,1.61) (0.86,1.51) (0.83,1.43) 38 GLP-1 receptor 1.19 1.32 1.32 39 Sulfonylurea v 0.89 40 agonist (0.30,4.80) (1.01,1.72) (1.01,1.71) Alpha 41 0.10 0.16 0.11 42 Sulfonylurea v glucosidase 0.92 (0.00,2.09) (0.00,334) (0.01,1.34) 43 inhibitor 44 0.27 2.75 0.59 Sulfonylurea v Glitinide 0.54 45 (0.03,2.76) (0.10,71.97) (0.09,3.93) 46 3.04 0.89 0.92 Sulfonylurea v Standard therapy 0.45 47 (0.14,68.35) (0.57,1.40) (0.59,1.43) 48 1.43 1.00 1.02 49 Thiazolidinedione v DPP-4 inhibitor 0.59 (0.41,4.99) (0.74,1.35) (0.76,1.37) 50 GLP-1 receptor 1.51 1.12 1.13 51 Thiazolidinedione v 0.75 52 agonist (0.25,9.22) (0.82,1.53) (0.83,1.53) 3.15 1.10 1.18 53 Thiazolidinedione v Basal insulin 0.54 54 (0.12,82.17) (0.46,2.60) (0.51,2.70) Basal-bolus 1.01 2.07 1.55 55 Thiazolidinedione v 0.70 56 insulin (0.06,16.39) (0.21,20.06) (0.27,9.04) 57 0.77 0.89 0.78 Thiazolidinedione v Standard therapy 0.83 58 (0.54,1.11) (0.27,2.96) (0.55,1.11) 59 0.92 0.92 0.92 DPP-4 inhibitor v SGLT-2 inhibitor 0.99 60 (0.28,3.01) (0.77,1.10) (0.77,1.10)

https://mc.manuscriptcentral.com/bmj Page 169 of 244 BMJ

1 2 3 Direct Indirect Incoherence Network 4 Intervention Comparator estimate estimate (p value) estimate 5 GLP-1 receptor 1.03 1.11 1.11 6 DPP-4 inhibitor v 0.89 agonist (0.37,2.88) (0.94,1.30) (0.94,1.30) 7 0.30 1.27 1.15 8 DPP-4 inhibitor v Basal insulin 0.39 9 (0.01,7.35) (0.55,2.92) (0.51,2.59) 0.90 0.76 0.77 10 DPP-4 inhibitor v Standard therapy 0.82 11 (0.21,3.92) (0.48,1.20) (0.50,1.19) GLP-1 receptor 1.23 0.70 1.04 12 Confidential:v Basal insulin For Review Only0.53 13 agonist (0.48,3.18) (0.16,3.06) (0.47,2.32) 14 GLP-1 receptor Basal bolus 0.97 1.74 1.38 15 v 0.75 16 agonist insulin (0.06,15.69) (0.18,16.58) (0.24,7.96) 17 GLP-1 receptor 0.99 Not 0.99 v Bolus insulin 0.99 18 agonist (0.10,9.57) estimable (0.10,9.56) 19 Basal bolus 1.95 0.86 1.32 20 Basal insulin v 0.65 21 insulin (0.18,21.6) (0.07,10.93) (0.23,7.61) 1.00 0.62 0.67 22 Basal insulin v Standard therapy 0.71 23 (0.10,9.63) (0.24,1.61) (0.28,1.61) Alpha glucosidase 4.93 7.74 5.40 24 v Glitinide 0.92 25 inhibitor (0.24,103) (0.00,16589) (0.44,66.36) 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

https://mc.manuscriptcentral.com/bmj BMJ Page 170 of 244

1 2 3 Table 5 Kidney failure 4 5 6 Direct Indirect Incoherence Network 7 Intervention Comparator estimate estimate (p value) estimate 8 SGLT-2 0.70 0.85 0.70 v Placebo 0.86 9 inhibitor (0.56,0.88) (0.11,6.48) (0.56,0.89) 10 GLP-1 receptor 0.67 0.64 0.78 v Placebo 0.74 11 agonist (0.79,0.93) (0.20,2.10) (0.67,0.92) 12 SGLT-2Confidential: GLP-1 receptor For0.99 Review0.90 Only 0.90 v 0.95 13 inhibitor agonist (0.06,15.9) (0.68,1.19) (0.68,1.19) 14 0.95 1.11 0.96 DPP-4 inhibitor v Placebo 0.81 15 (0.72,1.26) (0.34,3.62) (0.73,1.26) 16 0.09 0.09 17 Metformin v SGLT-2 inhibitor NA 0.96 18 (0.01,1.60) (0.01,1.60) 2.97 2.97 19 Sulfonylurea v Thiazolidinedione NA 0.99 20 (0.31,28.60) (0.31,28.60) 0.96 0.07 0.85 21 Sulfonylurea v DPP-4 inhibitor 0.85 22 (0.14,6.57) (0.72,7.13) (0.20,3.72) 23 0.99 1.32 1.16 Sulfonylurea v SGLT-2 inhibitor 0.85 24 (0.10,9.51) (0.19,9.31) (0.26,5.12) 25 GLP-1 receptor 1.52 1.21 1.23 DPP-4 inhibitor v 0.73 26 agonist (0.42,5.49) (0.88,1.67) (0.90,1.68) 27 GLP-1 receptor 1.02 1.02 v Basal insulin NA 0.99 28 agonist (0.18,5.60) (0.18,5.60) 29 GLP-1 receptor 0.33 0.33 30 v Bolus insulin NA 0.99 agonist (0.01,8.07) (0.01,8.07) 31 32 33 NA=not available 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

https://mc.manuscriptcentral.com/bmj Page 171 of 244 BMJ

1 2 3 Table 6 Hospitalisation for heart failure 4 5 6 Direct Indirect Incoherence Network 7 Intervention Comparator estimate estimate (p value) estimate 8 0.68 1.02 0.69 SGLT-2 inhibitor v Placebo 0.45 9 (0.60,0.77) (0.36,2.90) (0.61,0.78) 10 GLP-1 receptor 0.94 0.60 0.93 v Placebo 0.20 11 agonist (0.85,1.04) (0.31,1.17) (0.84,1.03) 12 Confidential:GLP-1 receptor For2.00 Review0.74 Only 0.74 SGLT-2 inhibitor v 0.42 13 agonist (0.18,22.2) (0.63,0.86) (0.63,0.87) 14 0.74 1.26 1.19 Metformin v Placebo 0.66 15 (0.16,3.35) (0.77,2.06) (0.75,1.91) 16 0.68 0.89 0.89 17 Sulfonylurea v Placebo 0.82 18 (0.07,6.98) (0.71,1.21) (0.71,1.12) 1.53 1.66 1.55 19 Thiazolidinedione v Placebo 0.72 20 (1.25,1.87) (1.11,2.48) (1.30,1.86) 1.05 1.03 1.05 21 DPP-4 inhibitor v Placebo 0.91 22 (0.94,1.18) (0.69,1.52) (0.94,1.17) 23 0.33 0.81 0.75 Basal insulin v Placebo 0.53 24 (0.01,8.06) (0.30,2.19) (0.29,1.93) 25 1.43 1.18 1.34 Metformin v Sulfonylurea 0.73 26 (0.78,2.62) (0.50,2.77) (0.84,2.13) 27 0.81 0.72 0.77 Metformin v Thiazolidinedione 0.80 28 (0.44,1.48) (0.36,1.44) (0.49,1.22) 29 0.66 1.15 1.14 30 Metformin v DPP-4 inhibitor 0.79 (0.01,33.61) (0.71,1.84) (0.71,1.82) 31 1.43 1.75 1.74 32 Metformin v SGLT-2 inhibitor 0.90 33 (0.06,35.56) (1.07,2.85) (1.07,2.82) 0.57 0.58 0.57 34 Sulfonylurea v Thiazolidinedione 0.92 35 (0.40,0.79_ (0.41,0.82) (0.45,0.73) 36 0.86 0.82 0.85 Sulfonylurea v DPP-4 inhibitor 0.83 37 (0.66,1.14) (0.56,1.20) (0.68,1.06) 38 39 0.98 1.31 1.30 Sulfonylurea v SGLT-2 inhibitor 0.70 40 (0.23,4.19) (1.01,1.71) (1.01,1.68) 41 GLP-1 receptor 0.59 0.97 0.96 42 Sulfonylurea v 0.51 43 agonist (0.13,2.61) (0.76,1.25) (0.75,1.23) 4.11 1.47 1.48 44 Thiazolidinedione v DPP-4 inhibitor 0.51 45 (0.20,86.6) (1.20,1.80) (1.21,1.81) 46 GLP-1 receptor 7.34 1.65 1.67 Thiazolidinedione v 0.18 47 agonist (0.85,63.13) (1.35,2.02) (1.36,2.05) 48 2.90 2.01 2.08 Thiazolidinedione v Basal insulin 0.83 49 (0.12,72.1) (0.73,5.49) (0.79,5.42) 50 1.93 1.11 1.85 Thiazolidinedione v Standard therapy 0.49 51 (1.26,2.97) (0.25,5.01) (1.22,2.80) 52 0.40 1.53 1.53 DPP-4 inhibitor v SGLT-2 inhibitor 0.41 53 (0.02,9.89) (1.30,1.81) (1.30,1.80) 54 GLP-1 receptor 1.19 1.13 1.13 55 DPP-4 inhibitor v 0.89 56 agonist (0.54,2.65) (0.97,1.31) (0.98,1.31) 1.98 1.23 0.70 1.25 57 DPP-4 inhibitor v Standard therapy 58 (0.18,22.0) (0.77,1.96) (0.79,1.98) 0.33 0.83 0.82 59 SGLT-2 inhibitor v Standard therapy 0.57 60 (0.01,8.09) (0.52,1.33) (0.52,1.30)

https://mc.manuscriptcentral.com/bmj BMJ Page 172 of 244

1 2 3 0.33 1.00 0.89 Basal insulin v Standard therapy 0.52 4 (0.01,8.15) (0.34,2.96) (0.32,2.50) 5 6 7 8 9 10 11 12 Confidential: For Review Only 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

https://mc.manuscriptcentral.com/bmj Page 173 of 244 BMJ

1 2 3 Table 7 Severe hypoglycaemia 4 5 6 Direct Indirect Incoherence Network 7 Intervention Comparator estimate estimate (p value) estimate 8 0.85 1.07 0.89 SGLT-2 inhibitor v Placebo 0.49 9 (0.64,1.12) (0.62,1.84) (0.69,1.14) 10 GLP-1 receptor 0.97 0.84 0.95 v Placebo 0.47 11 agonist (0.83,1.12) (0.58,1.20) (0.83,1.09) 12 Confidential:GLP-1 receptor For0.98 Review0.94 Only 0.94 SGLT-2 inhibitor v 0.93 13 agonist (0.36,2.69) (0.70,1.25) (0.71,1.24) 14 1.68 1.56 1.62 Metformin v Placebo 0.92 15 (0.63,4.45) (0.58,4.17) (0.81,3.23) 16 0.93 6.45 6.03 17 Sulfonylurea v Placebo 0.08 18 (0.11,7.67) (4.33,9.62) (4.07,8.94) 1.53 0.98 1.24 19 Thiazolidinedione v Placebo 0.31 20 (0.85,2.75) (0.52,1.82) (0.81,1.89) 1.08 1.14 1.09 21 DPP-4 inhibitor v Placebo 0.76 22 (0.92,1.26) (0.82,1.60) (0.95,1.26) 23 0.49 1.92 1.90 Basal insulin v Placebo 0.50 24 (0.01,25.1) (1.35,2.72) (1.34,2.69) 25 0.12 0.30 0.27 Metformin v Sulfonylurea 0.38 26 (0.02,0.85) (0.14,0.68) (0.13,0.57) 27 1.00 1.34 1.31 Metformin v Thiazolidinedione 0.85 28 (0.06,16.24) (0.58,3.10) (0.59,2.92) 29 0.98 1.55 1.48 30 Metformin v DPP-4 inhibitor 0.69 (0.11,8.49) (0.74,3.26) (0.73,2.98) 31 3.67 1.46 1.82 32 Metformin v SGLT-2 inhibitor 0.28 33 (0.85,15.74) (0.64,3.31) (0.89,3.71) 7.52 4.22 4.88 34 Sulfonylurea v Thiazolidinedione 0.35 35 (2.63,21.52) (2.28,7.78) (2.87,8.30) 36 5.49 5.58 5.52 Sulfonylurea v DPP-4 inhibitor 0.97 37 (3.45,8.75) (2.87,10.88) (3.77,8.09) 38 6.01 7.01 6.78 39 Sulfonylurea v SGLT-2 inhibitor 0.78 40 (2.33,15.5) (4.26,11.53) (4.36,10.55) GLP-1 receptor 1.25 6.65 6.37 41 Sulfonylurea v 0.19 42 agonist (0.11,14.41) (4.40,10.1) (4.23,9.58) 43 Alpha 1.69 7.93 5.42 44 Sulfonylurea v glucosidase 0.39 (0.07,38.71) (1.25,50.47) (1.06,27.76) 45 inhibitor 46 2.47 1.73 1.94 Sulfonylurea v Glitinide 0.79 47 (0.28,22.1) (0.39,7.74) (0.55,6.77) 48 0.37 1.25 1.13 49 Thiazolidinedione v DPP-4 inhibitor 0.13 (0.08,1.68) (0.80,1.96) (0.73,1.74) 50 GLP-1 receptor 0.83 1.38 1.30 51 Thiazolidinedione v 0.48 52 agonist (0.23,3.06) (0.87,2.18) (0.85,2.01) 0.82 0.62 0.65 53 Thiazolidinedione v Basal insulin 0.67 54 (0.25,2.71) (0.35,1.09) (0.39,1.08) 3.16 1.39 2.47 55 Thiazolidinedione v Standard therapy 0.28 56 (1.40,7.12) (0.40,4.81) (1.25,4.87) 57 0.65 1.31 1.23 DPP-4 inhibitor v SGLT-2 inhibitor 0.15 58 (0.26,1.63) (0.98,1.75) (0.93,1.63) 59 GLP-1 receptor 1.43 1.09 1.15 DPP-4 inhibitor v 0.23 60 agonist (0.96,2.11) (0.88,1.34) (0.96,1.38)

https://mc.manuscriptcentral.com/bmj BMJ Page 174 of 244

1 2 3 0.20 0.60 0.58 DPP-4 inhibitor v Basal insulin 0.24 4 (0.03,1.20) (0.42,0.88) (0.40,0.83) 5 1.36 2.66 2.18 6 DPP-4 inhibitor v Standard therapy 0.41 (0.36,5.20) (1.12,6.32) (1.05,4.52) 7 GLP-1 receptor 0.51 0.44 0.50 8 v Basal insulin 0.77 9 agonist (0.36,0.72) (0.17,1.14) (0.36,0.69) GLP-1 receptor 0.34 0.34 10 v Bolus insulin NA 1.00 11 agonist (0.07,1.78) (0.07,1.78) Basal bolus 0.99 0.99 12 Basal insulinConfidential:v For ReviewNA Only1.00 13 insulin (0.06,16.24) (0.06,16.24) 14 2.13 4.09 3.79 Basal insulin v Standard therapy 0.60 15 (0.22,20.75) (1.79,9.31) (1.75,8.22) 16 Alpha 1.01 0.25 0.36 17 glucosidase v Glitinide 0.55 (0.02,51.5) (0.03,2.46) (0.05,2.56) 18 inhibitor 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

https://mc.manuscriptcentral.com/bmj Page 175 of 244 BMJ

1 2 3 Table 8 Blindness 4 5 6 Direct Indirect Incoherence Network 7 Intervention Comparator estimate estimate (p value) estimate 8 SGLT-2 0.17 0.17 v Placebo NA 1.00 9 inhibitor (0.01,4.07) (0.01,4.07) 10 GLP-1 receptor 1.00 1.00 v Placebo NA 1.00 11 agonist (0.23,4.41) (0.23,4.41) 12 Confidential: For0.99 Review Only 0.99 DPP-4 inhibitor v Placebo NA 1.00 13 (0.57,1.73) (0.57,1.73) 14 15 NA=not available 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

https://mc.manuscriptcentral.com/bmj BMJ Page 176 of 244

1 2 3 Table 9 Health-related quality of life 4 5 6 Direct Indirect Incoherence Network 7 Intervention Comparator estimate estimate (p value) estimate 8 GLP-1 receptor 0.11 0.70 0.12 v Placebo 0.10 9 agonist (0.01,0.22) (0.06,1.34) (0.01,0.24) 10 GLP-1 receptor 0.01 -0.25 0.01 SGLT-2 inhibitor v 1.00 11 agonist (-0.20,0.21) (-88.0,87.5) (-0.20,0.21) 12 Confidential: For0.25 Review-0.05 Only 0.25 Sulfonylurea v Placebo 1.00 13 (-0.17,0.67) (-124,124) (-0.19,0.69) 14 0.26 -0.05 0.02 DPP-4 inhibitor v Placebo 0.21 15 (-0.12,0.63) (-0.24,0.14) (-0.16,0.20) 16 -0.16 -0.29 -0.16 17 Metformin v Thiazolidinedione 0.86 18 (-0.49,0.17) (-1.71,1.12) (-0.48,0.15) -0.03 0.00 -0.02 19 Metformin v DPP-4 inhibitor 0.94 20 (-0.33,0.28) (-0.67,0.68) (-0.28,0.24) GLP-1 receptor -0.11 -0.16 -0.12 21 Metformin v 0.88 22 agonist (-0.43,0.21) (-0.67,0.36) (-0.39,0.14) 23 0.15 0.14 0.15 Thiazolidinedione v DPP-4 inhibitor 0.99 24 (-0.21,0.50) (-0.50,0.78) (-0.15,0.44) 25 GLP-1 receptor 0.05 0.02 0.04 Thiazolidinedione v 0.94 26 agonist (-0.29,0.39) (-0.66,0.70) (-0.25,0.33) 27 0.04 0.32 0.04 Thiazolidinedione v Glitinide 1.00 28 (-0.26,0.34) (-123,124) (-0.26,0.34) 29 GLP-1 receptor -0.14 0.06 -0.11 30 DPP-4 inhibitor v 0.41 agonist (-0.31,0.03) (-0.31,0.42) (-0.25,0.04) 31 0.41 0.25 0.34 32 DPP-4 inhibitor v Standard therapy 0.54 33 (0.10,0.72) (-0.11,0.61) (0.11,0.57) Alpha 34 0.00 0.04 0.00 DPP-4 inhibitor v glucosidase 1.00 35 (-0.30,0.30) (-124,124) (-0.30,0.30) 36 inhibitor 37 GLP-1 receptor 0.20 -0.31 -0.03 v Basal insulin 0.23 38 agonist (-0.12,0.53) (-0.66,0.04) (-0.30,0.24) 39 GLP-1 receptor 0.24 0.75 0.34 40 v Bolus insulin 0.03 41 agonist (0.03,0.45) (0.32,1.17) (0.12,0.56) 42 0.55 0.04 0.37 Basal insulin v Bolus insulin 0.03 43 (0.27,0.82) (-0.35,0.42) (0.11,0.63) 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

https://mc.manuscriptcentral.com/bmj Page 177 of 244 BMJ

1 2 3 Table 10 Body weight 4 5 Direct Indirect Incoherence Network 6 Intervention Comparator 7 estimate estimate (p value) estimate 8 -1.86 -2.11 -1.94 SGLT-2 inhibitor v Placebo 0.47 9 (-2.23, -1.50) (-2.67, -1.55) (-2.24, -1.63) 10 GLP-1 receptor -1.44 -1.24 -1.37 v Placebo 0.48 11 agonist (-1.77, -1.11) (-1.70, -0.77) (-1.64, -1.11) 12 Confidential:GLP-1 receptor For0.11 Review-0.63 Only -0.56 SGLT-2 inhibitor v 0.28 13 agonist (-1.17, 1.39) (-1.03, -0.23) (-0.94, -0.18) 14 -0.88 -0.73 -0.75 Metformin v Placebo 0.79 15 (-1.92, 0.17) (-1.19, -0.26) (-1.17, -0.33) 16 2.84 1.59 1.70 17 Sulfonylurea v Placebo 0.06 18 (1.61, 4.07) (1.20, 1.97) (1.33, 2.07) 2.93 2.36 2.60 19 Thiazolidinedione v Placebo 0.10 20 (2.41, 3.45) (1.91, 2.80) (2.26, 2.94) 0.23 0.50 0.34 21 DPP-4 inhibitor v Placebo 0.25 22 (-0.07, 0.54) (0.11, 0.89) (0.09, 0.58) 23 2.67 2.16 2.20 Basal insulin v Placebo 0.61 24 (0.82, 4.52) (1.62, 2.70) (1.69, 2.72) 25 Alpha glucosidase -0.38 -0.08 -0.24 v Placebo 0.59 26 inhibitor (-1.13, 0.37) (-0.85, -0.69) (-0.77, 0.30) 27 1.12 1.04 1.06 Glitinide v Placebo 0.94 28 (-0.75, 2.98) (0.10, 1.98) (0.22, 1.89) 29 -1.52 -2.69 -2.45 30 Metformin v Sulfonylurea 0.05 (-0.45, -2.58) (-3.23, -2.15) (-2.93, -1.97) 31 -3.80 -3.14 -3.35 32 Metformin v Thiazolidinedione 0.15 33 (-4.63, -2.98) (-3.70, -2.59) (-3.81, -2.89) -1.29 -0.98 -1.09 34 Metformin v DPP-4 inhibitor 0.50 35 (-2.03, -0.56) (-1.50, -0.46) (-1.51, -0.66) 36 1.72 1.07 1.19 Metformin v SGLT-2 inhibitor 0.79 37 (0.57, 2.86) (0.55, 1.60) (0.71, 1.67) 38 GLP-1 receptor -0.25 0.76 0.62 39 Metformin v 0.14 40 agonist (-1.51, 1.00) (0.27, 1.26) (0.17, 1.08) -0.36 -1.42 -1.25 41 Metformin v Standard therapy 0.24 42 (-1.98, 1.27) (-2.12, -0.72) (-1.89, -0.61) 43 Alpha glucosidase -0.55 -0.51 -0.52 Metformin v 0.96 44 inhibitor (-2.05, 0.96) (-1.21, 0.19) (-1.15, 0.12) 45 -2.51 -1.73 -1.81 Metformin v Glitinide 0.62 46 (-5.40, 0.38) (-2.67, -0.80) (-2.69, -0.92) 47 -0.86 -0.93 -0.90 Sulfonylurea v Thiazolidinedione 0.85 48 (-1.44, -0.28) (-1.46, -0.40) (-1.29, -0.51) 49 1.38 1.36 1.36 50 Sulfonylurea v DPP-4 inhibitor 0.95 51 (0.78, 1.98) (0.89, 1.83) (1.00, 1.73) 4.35 3.49 3.64 52 Sulfonylurea v SGLT-2 inhibitor 0.15 53 (3.29, 5.42) (3.01, 3.97) (3.20, 4.08) GLP-1 receptor 2.13 3.29 3.07 54 Sulfonylurea v 0.03 55 agonist (1.19, 3.07) (2.84, 3.73) (2.67, 3.48) 56 -0.99 -0.49 -0.50 Sulfonylurea v Basal insulin 0.12 57 (-4.17, 2.20) (-1.09, 0.12) (-1.09, 0.09) 58 Basal bolus 4.44 -1.73 -0.67 Sulfonylurea v 0.00 59 insulin (1.64, 5.22) (-3.00, -0.46) (-1.85, 0.51) 60

https://mc.manuscriptcentral.com/bmj BMJ Page 178 of 244

1 2 3 Direct Indirect Incoherence Network Intervention Comparator 4 estimate estimate (p value) estimate 5 -2.10 1.14 0.83 6 Sulfonylurea v Bolus insulin 0.21 (-6.95,2.75) (-0.44,2.73) (-0.68, 2.34) 7 Alpha glucosidase 1.00 2.01 1.94 8 Sulfonylurea v 0.39 9 inhibitor (-1.20, 3.20) (1.38, 2.65) (1.32, 2.55) 0.98 0.34 0.64 10 Sulfonylurea v Glitinide 0.45 11 (-0.21, 2.17) (-0.81, 1.48) (-0.18, 1.47) 1.84 2.35 2.26 12 ThiazolidinedioneConfidential:v DPP-4 inhibitor For Review Only0.29 13 (0.99, 2.70) (1.96, 2.75) (1.90, 2.65) 14 3.29 4.57 4.54 Thiazolidinedione v SGLT-2 inhibitor 0.35 15 (0.73, 5.85) (4.14, 5.01) (4.11, 4.96) 16 GLP-1 receptor 3.94 3.98 3.97 Thiazolidinedione v 0.95 17 agonist (2.91, 4.97) (3.56, 4.40) (3.58, 4.36) 18 1.30 0.53 0.39 Thiazolidinedione v Basal insulin 0.48 19 (-0.75, 3.35) (-0.06, 1.12) (-0.18,0.97) 20 Basal bolus -1.01 0.54 0.23 21 Thiazolidinedione v 0.30 insulin (-3.63, 1.62) (-0.78, 1.86) (-0.95, 1.40) 22 0.62 2.38 2.10 23 Thiazolidinedione v Standard therapy 0.04 24 (-0.89, 2.13) (1.72, 2.02) (1.49, 2.70) Alpha glucosidase 2.37 2.91 2.83 25 Thiazolidinedione v 0.53 26 inhibitor (0.82, 3.93) (2.27, 3.55) (2.24, 3.43) 27 1.22 1.60 1.54 Thiazolidinedione v Glitinide 0.82 28 (-0.99, 3.44) (0.68, 2.52) (0.69, 2.39) 29 2.51 2.23 2.27 DPP-4 inhibitor v SGLT-2 inhibitor 0.57 30 (1.62, 3.40) (1.84, 2.62) (1.92, 2.63) 31 GLP-1 receptor 1.65 1.73 1.71 DPP-4 inhibitor v 0.45 32 agonist (1.02, 2.28) (1.37, 2.09) (1.40, 2.02) 33 -2.07 -1.84 -1.87 DPP-4 inhibitor v Basal insulin 0.78 34 (-3.59, -0.55) (-2.41, -1.27) (-2.40, -1.34) 35 Basal bolus -6.20 -2.02 -2.03 36 DPP-4 inhibitor v 0.34 37 insulin (-14.6, 2.26) (-3.17, -0.88) (-3.20, -0.87) -0.47 0.00 -0.17 38 DPP-4 inhibitor v Standard therapy 0.44 39 (-1.43, 0.49) (-0.70, -0.69) (-0.73, 0.40) Alpha glucosidase 0.70 0.52 0.57 40 DPP-4 inhibitor v 0.78 41 inhibitor (-0.35, 1.76) (-0.12, 1.17) (0.02, 1.14) 42 -1.92 -2.51 -2.44 SGLT-2 inhibitor v Standard therapy 0.55 43 (-3.74, -0.09) (-3.16, -1.85) (-3.05, -1.83) 44 GLP-1 receptor -3.56 -3.61 -3.58 v Basal insulin 0.93 45 agonist (-4.13, -3.00) (-4.51, -2.72) (-4.06, -3.10) 46 GLP-1 receptor Basal bolus -4.83 -3.03 -3.74 v 0.47 47 agonist insulin (-6.64, -3.01) (-4.50, -1.55) (-4.89, -2.60) 48 GLP-1 receptor -1.95 -5.19 -2.24 49 v Bolus insulin 0.21 agonist (-3.49, -0.42) (-10.1, 0.32) (-3.71, -0.78) 50 GLP-1 receptor -1.49 -2.00 -1.88 51 v Standard therapy 0.45 52 agonist (-2.64, -0.34) (-2.67, -1.34) (-2.45, -1.30) Basal bolus -0.88 0.36 -0.17 53 Basal insulin v 0.30 54 insulin (-2.66, 0.90) (-1.17, 1.89) (-1.33, 0.99) 2.08 1.63 1.70 55 Basal insulin v Standard therapy 0.66 56 (0.28, 3.88) (0.87, 2.39) (1.00, 2.40) 57 0.84 0.48 0.55 Glitinide v Standard therapy 0.76 58 (-1.22, 2.89) (-0.57, 1.53) (-0.38, 1.49) 59 60

https://mc.manuscriptcentral.com/bmj Page 179 of 244 BMJ

1 2 3 Table 11 Amputation 4 5 Direct Indirect Incoherence Network 6 Intervention Comparator 7 estimate estimate (p value) estimate 8 1.11 1.37 1.11 SGLT-2 inhibitor v Placebo 0.90 9 (0.90, 1.36) (0.05, 34.5) (0.90,1.36) 10 GLP-1 receptor 0.33 0.33 v Placebo NA 1.00 11 agonist (0.01,8.19) (0.01,8.19) 12 Confidential: For1.01 Review Only 1.01 Thiazolidinedione v Placebo NA 1.00 13 (0.59,1.75) (0.59,1.75) 14 0.91 0.91 DPP-4 inhibitor v Placebo NA 1.00 15 (0.64,1.29) (0.64,1.29) 16 Standard 0.33 0.33 17 Thiazolidinedione v NA 1.00 18 therapy (0.12,0.92) (0.12,0.92) SGLT-2 0.82 0.82 19 DPP-4 inhibitor v NA 1.00 20 inhibitor (0.55,1.23) (0.55,1.23) 21 22 NA=not available 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

https://mc.manuscriptcentral.com/bmj BMJ Page 180 of 244

1 2 3 Table 12 Neuropathic pain 4 5 Direct Indirect Incoherence Network 6 Intervention Comparator 7 estimate estimate (p value) estimate 8 GLP-1 receptor 0.33 1.50 0.33 v Placebo 0.002 9 agonist (0.07,1.65) (0.62,3.64) (0.01,8.19) 10 3.00 3.00 DPP-4 inhibitor v Placebo NA 1.00 11 (0.12,73.75) (0.12,73.75) 12 Confidential: For Review Only 13 NA=not available 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

https://mc.manuscriptcentral.com/bmj Page 181 of 244 BMJ

1 2 3 Table 13 Diabetic ketoacidosis 4 5 Direct Indirect Incoherence Network 6 Intervention Comparator 7 estimate estimate (p value) estimate 8 SGLT-2 2.08 0.41 1.92 v Placebo 0.22 9 inhibitor (1.16,3.74) (0.03,5.17) (1.08,3.39) 10 GLP-1 1.10 3.76 1.21 11 receptor v Placebo 0.32 (0.55,2.17) (0.37,38.04) (0.63,2.33) 12 agonist Confidential: For Review Only 13 GLP-1 SGLT-2 0.33 1.79 1.59 14 v receptor 0.32 inhibitor (0.01,8.08) (0.74,4.34) (0.68,3.72) 15 agonist 16 DPP-4 0.71 0.87 0.72 17 v Placebo 0.88 18 inhibitor (0.34,1.49) (0.07,10.88) (0.36,1.47) SGLT-2 0.67 0.67 19 Metformin v NA 1.00 20 inhibitor (0.03,16.41) (0.03,16.40) DPP-4 0.47 1.92 0.73 21 Sulfonylurea v 0.49 22 inhibitor (0.05,4.53) (0.07,53.85) (0.11,4.78) 23 SGLT-2 0.67 0.16 0.28 Sulfonylurea v 0.49 24 inhibitor (0.03,16.56) (0.01,1.90) (0.04,1.93) 25 Alpha 1.49 1.49 26 Sulfonylurea v glucosidase NA 1.00 (0.06,36.63) (0.06,36.63) 27 inhibitor 28 1.51 1.51 Sulfonylurea v Glitinide NA 1.00 29 (0.06,37.03) (0.06,37.03) 30 GLP-1 31 DPP-4 0.33 0.63 0.60 v receptor 0.70 32 inhibitor (0.01,8.15) (0.24,1.70) (0.23,1.54) 33 agonist Alpha 34 1.01 1.01 glucosidase v Glitinide NA 1.00 35 (0.02,51.04) (0.02,51.04) 36 inhibitor 37 38 NA=not available 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

https://mc.manuscriptcentral.com/bmj BMJ Page 182 of 244

1 2 3 Table 14 Serious hyperglycaemia 4 5 Direct Indirect Incoherence Network 6 Intervention Comparator 7 estimate estimate (p value) estimate 8 0.64 0.41 0.59 SGLT-2 inhibitor v Placebo 0.52 9 (0.36,1.16) (0.12,1.42) (0.35,1.01) 10 GLP-1 receptor 0.42 0.34 0.40 v Placebo 0.71 11 agonist (0.24,0.72) (0.13,0.92) (0.25,0.64) 12 Confidential:GLP-1 receptor For1.49 Review1.48 Only 1.48 SGLT-2 inhibitor v 1.00 13 agonist (0.17,12.86) (0.72,3.04) (0.75,2.93) 14 0.31 0.49 0.41 Metformin v Placebo 0.73 15 (0.04,2.64) (0.09,2.78) (0.10,1.60) 16 0.28 0.58 0.45 17 Sulfonylurea v Placebo 0.35 18 (0.09,0.92) (0.24,1.40) (0.22,0.91) 0.62 0.63 0.62 19 DPP-4 inhibitor v Placebo 0.98 20 (0.43,0.91) (0.25,1.56) (0.44,0.88) Alpha glucosidase 0.07 0.07 21 v Placebo NA 1.00 22 inhibitor (0.02,0.28) (0.02,0.28) 23 0.66 0.28 0.43 Metformin v Thiazolidinedione 0.61 24 (0.07,6.57) (0.03,2.64) (0.09,2.11) 25 0.81 0.39 0.66 Metformin v DPP-4 inhibitor 0.63 26 (0.16,4.13) (0.03,4.92) (0.17,2.57) 27 GLP-1 receptor 0.64 1.40 1.02 Metformin v 0.58 28 agonist (0.07,5.51) (0.24,8.29) (0.25,4.11) 29 0.66 0.41 0.47 30 Sulfonylurea v Thiazolidinedione 0.71 (0.08,5.66) (0.10,1.61) (0.15,1.48) 31 0.92 0.53 0.72 32 Sulfonylurea v DPP-4 inhibitor 0.45 33 (0.35,2.40) (0.19,1.51) (0.36,1.45) 1.47 0.67 0.75 34 Sulfonylurea v SGLT-2 inhibitor 0.51 35 (0.17,12.60) (0.27,1.67) (0.32,1.75) 36 GLP-1 receptor 0.44 1.61 0.89 Sulfonylurea v 0.12 37 agonist (0.10,1.82) (0.65,4.00) (0.41,1.96) 38 39 1.00 1.72 1.54 Thiazolidinedione v DPP-4 inhibitor 0.68 40 (0.10,9.99) (0.52,5.70) (0.53,4.43) 41 GLP-1 receptor 1.02 3.13 2.39 Thiazolidinedione v 0.37 42 agonist (0.12,8.77) (0.92,10.71) (0.82,7.00) 43 5.87 1.82 2.81 Thiazolidinedione v Basal insulin 0.35 44 (0.82,41.78) (0.40,8.25) (0.85,9.31) 45 0.65 0.47 0.58 46 Thiazolidinedione v Standard therapy 0.79 (0.16,2.67) (0.07,3.17) (0.19,1.79) 47 1.62 0.97 1.05 48 DPP-4 inhibitor v SGLT-2 inhibitor 0.56 49 (0.34,7.80) (0.50,1.88) (0.57,1.93) GLP-1 receptor 1.21 1.75 1.56 50 DPP-4 inhibitor v 0.54 51 agonist (0.46,3.18) (0.92.3.35) (0.92,2.65) 0.64 0.41 0.38 52 DPP-4 inhibitor v Standard therapy 0.70 53 (0.36,1.16) (0.12,1.42) (0.12,1.18) 54 GLP-1 receptor 1.02 3.30 1.18 v Basal insulin 0.35 55 agonist (0.43,2.41) (0.32,33.63) (0.52,2.64) 56 GLP-1 receptor Basal bolus 1.34 1.34 v NA 1.00 57 agonist insulin (0.16,11.53) (0.16,11.53) 58 59 NA=not available 60

https://mc.manuscriptcentral.com/bmj Page 183 of 244 BMJ

1 2 3 Table 15 Genital infection 4 5 Direct Indirect Incoherence Network 6 Intervention Comparator 7 estimate estimate (p value) estimate 8 3.50 2.18 3.42 SGLT-2 inhibitor v Placebo 0.33 9 (2.83,4.32) (0.86,5.52) (2.79,4.18) 10 GLP-1 receptor 1.14 0.62 0.69 v Placebo 0.56 11 agonist (0.18,7.30) (0.28,1.41) (0.33,1.45) 12 Confidential:GLP-1 receptor For5.44 Review2.97 Only 4.96 SGLT-2 inhibitor v 0.56 13 agonist (2.46,12.02) (0.46,19.22) (2.39,10.29) 14 0.49 0.91 0.74 DPP-4 inhibitor v Placebo 0.25 15 (0.21,1.14) (0.50,1.66) (0.46,1.20) 16 SGLT-2 0.39 0.39 17 Metformin v NA 1.00 18 inhibitor (0.21,0.75) (0.21,0.75) SGLT-2 0.16 0.16 19 Sulfonylurea v NA 1.00 20 inhibitor (0.11,0.22) (0.11,0.22) SGLT-2 3.29 3.29 21 Thiazolidinedione v NA 1.00 22 inhibitor (0.13,84.35) (0.13,84.35) 23 SGLT-2 0.26 0.12 0.22 DPP-4 inhibitor v 0.17 24 inhibitor (0.16,0.42) (0.05,0.31) (0.14,0.34) 25 1.60 1.60 SGLT-2 inhibitor v Standard therapy NA 1.00 26 (0.50,5.09) (0.50,5.09) 27 28 NA=not available 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

https://mc.manuscriptcentral.com/bmj BMJ Page 184 of 244

1 2 3 Table 16 Fournier gangrene 4 5 Direct Indirect Incoherence Network 6 Intervention Comparator 7 estimate estimate (p value) estimate 8 0.56 0.56 SGLT-2 inhibitor v Placebo NA 1.00 9 (0.16,1.92) (0.16,1.92) 10 3.91 3.91 DPP-4 inhibitor v Placebo NA 1.00 11 (0.43,35.37 (0.43,35.37) 12 Confidential: For Review Only 13 NA=not available 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

https://mc.manuscriptcentral.com/bmj Page 185 of 244 BMJ

1 2 3 Table 17 Severe gastrointestinal events 4 5 Direct Indirect Incoherence Network 6 Intervention Comparator 7 estimate estimate (p value) estimate 8 GLP-1 receptor 2.37 2.28 2.33 v Placebo 0.97 9 agonist (1.15,4.91) (0.35,14.96) (1.21,4.48) 10 1.76 1.83 1.77 DPP-4 inhibitor v Placebo 0.97 11 (0.42,7.34) (0.43,7.69) (0.66,4.79) 12 Confidential: For2.32 Review50.36 Only 3.07 Metformin v Sulfonylurea 0.29 13 (034,15.83) (0.21,12162) (0.49,19.36) 14 Thiazolidinedio 1.64 1.16 1.51 Metformin v 0.85 15 ne (0.34,7.84) (0.04,30.71) (0.38,5.98) 16 Standard 1.99 2.81 2.42 17 Metformin v 0.85 18 therapy (0.12,32.52) (0.28,28.56) (0.42,13.78) Alpha 19 0.65 0.65 Metformin v glucosidase NA 1.00 20 (0.07,6.34) (0.07,6.34) 21 inhibitor 1.39 1.39 22 Sulfonylurea v DPP-4 inhibitor NA 1.00 23 (0.18,10.42) (0.18,10.42) 24 GLP-1 receptor 0.63 0.96 0.76 DPP-4 inhibitor v 0.67 25 agonist (0.18,2.26) (0.23,4.11) (0.30,1.96) 26 0.90 0.18 0.28 DPP-4 inhibitor v Basal insulin 0.38 27 (0.04,19.79) (0.03,1.20) (0.05,1.41) 28 GLP-1 receptor 0.26 1.33 0.36 v Basal insulin 0.38 29 agonist (0.05,1.36) (0.05,34.23) (0.08,1.58) 30 31 32 NA=not available 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

https://mc.manuscriptcentral.com/bmj BMJ Page 186 of 244

1 2 3 Table 18 Pancreatic cancer 4 5 Direct Indirect Incoherence Network 6 Intervention Comparator 7 estimate estimate (p value) estimate 8 1.60 3.63 1.77 SGLT-2 inhibitor v Placebo 0.65 9 (0.45,5.63) (0.13,100.43) (0.55,5.75) 10 GLP-1 receptor 1.23 0.52 1.19 v Placebo 0.45 11 agonist (0.81,1.86) (0.06,4.57) (0.78,1.81) 12 Confidential: For0.84 Review1.24 Only 0.86 DPP-4 inhibitor v Placebo 0.69 13 (0.49,1.44) (0.20,7.77) (0.51,1.45) 14 1.35 0.59 1.32 Sulfonylurea v DPP-4 inhibitor 0.65 15 (0.69,2.62) (0.02,19.52) (0.68,2.55) 16 SGLT-2 0.33 0.74 0.64 17 Sulfonylurea v 0.45 18 inhibitor (0.01,8.14) (0.16,3.34) (0.16,2.51) GLP-1 receptor 1.57 0.67 0.73 19 DPP-4 inhibitor v 0.45 20 agonist (0.19,12.89) (0.34,1.32) (0.37,1.42) GLP-1 receptor 1.53 1.53 21 v Basal insulin NA 1.00 22 agonist (0.34,6.77) (0.34,6.77) 23 GLP-1 receptor 2.98 2.98 v Bolus insulin NA 1.00 24 agonist (0.12,74.30) (0.12,74.30) 25 26 NA=not available 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

https://mc.manuscriptcentral.com/bmj Page 187 of 244 BMJ

1 2 3 Table 19 Pancreatitis 4 5 Direct Indirect Incoherence Network 6 Intervention Comparator 7 estimate estimate (p value) estimate 8 0.61 0.63 0.61 SGLT-2 inhibitor v Placebo 0.97 9 (0.24,1.55) (0.20,1.98) (0.30,1.27) 10 GLP-1 receptor 1.12 1.88 1.17 v Placebo 0.34 11 agonist (0.80,1.56) (0.69,5.11) (0.89,1.55) 12 Confidential:GLP-1 receptor For0.62 Review0.51 Only 0.52 SGLT-2 inhibitor v 0.87 13 agonist (0.08,5.05) (0.23,1.15) (0.25,1.12) 14 0.85 1.53 1.37 Sulfonylurea v Placebo 0.67 15 (0.08,9.20) (0.53,4.40) (0.54,3.51) 16 0.45 1.31 1.14 17 Thiazolidinedione v Placebo 0.49 18 (0.03,7.27) (0.45,3.81) (0.42,3.09) 1.54 1.08 1.46 19 DPP-4 inhibitor v Placebo 0.48 20 (1.04,2.28) (0.43,2.70) (1.03,2.08) 0.66 0.21 0.44 21 Metformin v Thiazolidinedione 0.73 22 (0.01,33.6) (0.00,38.69) (0.02,10.06) 23 0.22 2.15 0.34 Metformin v DPP-4 inhibitor 0.39 24 (0.01,5.42) (0.01,394) (0.02,7.15) 25 GLP-1 receptor 1.01 0.11 0.42 Metformin v 0.49 26 agonist (0.02,51.01) (0.00,14.55) (0.02,9.01) 27 2.96 1.02 1.21 Sulfonylurea v Thiazolidinedione 0.55 28 (0.12,72.81) (0.25,4.12) (0.33,4.35) 29 0.89 0.98 0.94 30 Sulfonylurea v DPP-4 inhibitor 0.93 (0.20,4.04) (0.27,3.48) (0.36,2.42) 31 2.83 1.89 2.23 32 Sulfonylurea v SGLT-2 inhibitor 0.70 33 (0.58,13.74) (0.50,7.22) (0.80,6.20) GLP-1 receptor 0.95 1.31 1.17 34 Sulfonylurea v 0.75 35 agonist (0.19,4.68) (0.41,4.19) (0.45,3.01) Alpha 36 0.50 0.50 37 Sulfonylurea v glucosidase NA 1.00 (0.01,25.04) (0.01,25.04) 38 inhibitor 39 0.17 0.17 Sulfonylurea v Glitinide NA 1.00 40 (0.01,4.10) (0.01,4.10) 41 1.20 0.56 0.78 Thiazolidinedione v DPP-4 inhibitor 0.47 42 (0.26,5.64) (0.15,2.15) (0.28,2.14) 43 GLP-1 receptor 0.96 0.98 0.97 Thiazolidinedione v 0.98 44 agonist (0.26,3.55) (0.22,4.38) (0.36,2.62) 45 9.96 46 Basal bolus 1.01 2.20 Thiazolidinedione v (0.04,2408.8 0.51 47 insulin (0.02,51.5) (0.09,53.61) 48 9) 2.51 0.60 2.19 49 Thiazolidinedione v Standard therapy 0.45 50 (0.79,8.03) (0.02,21.44) (0.73,6.62) 1.87 2.50 2.38 51 DPP-4 inhibitor v SGLT-2 inhibitor 0.78 52 (0.29,11.89) (1.07,5.84) (1.10,5.14) 53 GLP-1 receptor 0.81 1.35 1.25 DPP-4 inhibitor v 0.40 54 agonist (0.27,2.41) (0.84,2.16) (0.81,1.91) 55 GLP-1 receptor 2.12 8.92 2.46 v Basal insulin 0.45 56 agonist (0.63,7.13) (0.25,315.04) (0.78,7.76) 57 GLP-1 receptor Basal bolus 2.94 0.30 2.26 v 0.51 58 agonist insulin (0.12,72.71) (0.00,253.94) (0.10,51.09) 59 60

https://mc.manuscriptcentral.com/bmj BMJ Page 188 of 244

1 2 3 0.32 1.36 0.92 Basal insulin v Standard therapy 0.45 4 (0.01,7.99) (0.19,9.76) (0.17,4.92) 5 Alpha 6 0.34 0.34 glucosidase v Glitinide NA 1.00 7 (0.01,8.28) (0.01,8.28) 8 inhibitor 9 10 NA=not available 11 12 Confidential: For Review Only 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

https://mc.manuscriptcentral.com/bmj Page 189 of 244 BMJ

1 2 3 Table 20 Glycated haemoglobin A1C 4 5 Direct Indirect Incoherence Network 6 Intervention Comparator 7 estimate estimate (p value) estimate 8 -0.72 -0.67 -0.61 SGLT-2 inhibitor v Placebo 0.65 9 (-0.84, -0.60) (-0.86, -0.48) (-0.68, -0.54) 10 GLP-1 receptor -0.86 -0.96 -0.90 v Placebo 0.30 11 agonist (-0.98, -0.74) (-1.11, -0.81) (-0.96, -0.83) 12 Confidential:GLP-1 receptor For0.38 Review0.17 Only 0.28 SGLT-2 inhibitor v 0.40 13 agonist (-0.09, 0.85) (0.04, 0.31) (0.19, 0.38) 14 -0.81 -0.81 -0.80 Metformin v Placebo 0.99 15 (-1.08, -0.54) (-0.95, -0.67) (-0.90, -0.71) 16 -1.04 -0.68 -0.66 17 Sulfonylurea v Placebo 0.08 18 (-1.42, -0.65) (-0.80, -0.56) (-0.74, -0.58) -0.73 -0.76 -0.71 19 Thiazolidinedione v Placebo 0.78 20 (-0.89, -0.57) (-0.89, -0.63) (-0.78, -0.63) -0.60 -0.77 -0.61 21 DPP-4 inhibitor v Placebo 0.03 22 (-0.70, -0.51) (-0.90, -0.65) (-0.66, -0.55) 23 -0.75 -0.80 -0.74 Basal insulin v Placebo 0.87 24 (-1.40, -0.09) (-0.98, -0.63) (-0.86, -0.62) 25 Alpha -0.96 -0.34 -0.53 26 glucosidase v Placebo 0.00 (-1.14, -0.77) (-0.55, -0.13) (-0.63, -0.43) 27 inhibitor 28 -0.63 -0.88 -0.74 Glitinide v Placebo 0.42 29 (-1.18, -0.08) (-1.13, -0.63) (-0.91, -0.58) 30 -0.11 -0.10 -0.14 31 Metformin v Sulfonylurea 0.94 32 (-0.38, 0.16) (-0.26, 0.06) (-0.25, -0.04) -0.12 -0.03 -0.10 33 Metformin v Thiazolidinedione 0.48 34 (-0.32, 0.08) (-0.19, 0.14) (-0.19, -0.00) -0.17 -0.14 -0.20 35 Metformin v DPP-4 inhibitor 0.82 36 (-0.44, 0.09) (-0.28, 0.01) (-0.29, -0.10) 37 0.06 -0.13 -0.19 Metformin v SGLT-2 inhibitor 0.41 38 (-0.36, 0.49) (-0.29, 0.03) (-0.30, -0.08) 39 GLP-1 receptor 0.12 0.08 0.09 Metformin v 0.86 40 agonist (-0.29, 0.53) (-0.07, 0.23) (-0.01, 0.20) 41 -0.48 0.02 -0.07 Metformin v Basal insulin 0.19 42 (-1.19, 0.24) (-0.17, 0.22) (-0.21, 0.07) 43 -0.25 -0.72 -0.60 Metformin v Standard therapy 0.02 44 (-0.60, 0.11) (-0.90, -0.53) (-0.72, -0.48) 45 Alpha 46 -0.28 -0.11 -0.27 Metformin v glucosidase 0.54 47 (-0.77, 0.22) (-0.30, 0.07) (-0.40, -0.15) 48 inhibitor -0.26 0.07 -0.06 49 Metformin v Glitinide 0.37 50 -0.94, 0.41) (-0.19, 0.32) (-0.24, 0.12) 0.09 -0.01 0.04 51 Sulfonylurea v Thiazolidinedione 0.36 52 (-0.07, 0.26) (-0.17, 0.14) (-0.04, 0.13) 53 -0.04 -0.05 -0.05 Sulfonylurea v DPP-4 inhibitor 0.96 54 (-0.23, 0.15) (-0.18, 0.09) (-0.14, 0.03) 55 -0.09 0.01 -0.05 Sulfonylurea v SGLT-2 inhibitor 0.63 56 (-0.46, 0.29) (-0.14, 0.16) (-0.15, 0.05) 57 GLP-1 receptor 0.21 0.18 0.23 Sulfonylurea v 0.87 58 agonist (-0.15, 0.58) (0.04, 0.32) (0.14, 0.33) 59 60

https://mc.manuscriptcentral.com/bmj BMJ Page 190 of 244

1 2 3 Direct Indirect Incoherence Network Intervention Comparator 4 estimate estimate (p value) estimate 5 -0.04 0.10 0.07 6 Sulfonylurea v Basal insulin 0.65 (-0.64, 0.56) (-0.09, 0.29) (-0.06, 0.21) 7 Basal bolus 1.08 0.05 0.21 8 Sulfonylurea v 0.11 9 insulin (-0.08, 2.24) (-0.46, 0.56) (-0.14, 0.56) 0.18 0.25 0.26 10 Sulfonylurea v Bolus insulin 0.90 11 (-0.78, 1.14) (-0.16, 0.65) (-0.02, 0.53) -0.56 -0.51 -0.46 12 Sulfonylurea Confidential:v Standard therapy For Review Only0.91 13 (-1.26, 0.15) (-0.68, -0.35) (-0.57, -0.34) 14 Alpha -0.14 0.00 -0.13 15 Sulfonylurea v glucosidase 0.47 (-0.48, 0.20) (-0.18, 0.19) (-0.25, -0.01) 16 inhibitor 17 0.16 0.07 0.08 Sulfonylurea v Glitinide 0.76 18 (-0.11, 0.44) (-0.28, 0.41) (-0.08, 0.24) 19 -0.08 -0.08 -0.10 Thiazolidinedione v DPP-4 inhibitor 0.96 20 (-0.36, 0.20) (-0.20, 0.03) (-0.18, -0.02) 21 -0.17 -0.04 -0.09 22 Thiazolidinedione v SGLT-2 inhibitor 0.72 23 (-0.89, 0.55) (-0.17, 0.10) (-0.19, 0.00) GLP-1 receptor 0.01 0.16 0.19 24 Thiazolidinedione v 0.46 25 agonist (-0.37, 0.40) (0.04, 0.29) (0.09, 0.28) 0.35 0.00 0.03 26 Thiazolidinedione v Basal insulin 0.19 27 (-0.13, 0.82) (-0.19, 0.19) (-0.10, 0.16) 28 Basal bolus 0.00 0.20 0.21 Thiazolidinedione v 0.78 29 insulin (-1.30, 1.30) (-0.29, 0.70) (-0.14, 0.46) 30 -0.34 -0.67 -0.50 Thiazolidinedione v Standard therapy 0.03 31 (-0.58, -0.10) (-0.85, -0.49) (-0.61, -0.39) 32 Alpha -0.60 -0.00 -0.18 33 Thiazolidinedione v glucosidase 0.02 (-1.08, -0.13) (-0.17, 0.17) (-0.29, -0.06) 34 inhibitor 35 0.04 0.10 0.04 36 Thiazolidinedione v Glitinide 0.87 37 (-0.55, 0.63) (-0.15, 0.34) (-0.13, 0.21) 0.06 0.04 0.00 38 DPP-4 inhibitor v SGLT-2 inhibitor 0.88 39 (-0.23, 0.36) (-0.09, 0.17) (-0.08, 0.09) GLP-1 receptor 0.31 0.20 0.29 40 DPP-4 inhibitor v 0.39 41 agonist (0.10, 0.51) (0.08, 0.32) (0.21, 0.36) 42 -0.12 0.17 0.13 DPP-4 inhibitor v Basal insulin 0.25 43 (-0.58, 0.34) (-0.01, 0.36) (0.01, 0.25) 44 Basal bolus -0.70 0.26 0.26 DPP-4 inhibitor v 0.85 45 insulin -10.55, 9.15) (-0.06, 0.58) (-0.08, 0.61) 46 -0.69 -0.35 -0.40 DPP-4 inhibitor v Standard therapy 0.02 47 (-0.93, -0.46) (-0.52, -0.18) (-0.50, -0.30) 48 Alpha -0.41 0.11 -0.08 49 DPP-4 inhibitor v glucosidase 0.01 50 (-0.75, -0.06) (-0.05, 0.25) (-0.18, 0.03) inhibitor 51 -0.39 -0.52 -0.41 52 SGLT-2 inhibitor v Standard therapy 0.66 53 (-0.97, 0.20) (-0.70, -0.35) (-0.53, -0.29) GLP-1 receptor 0.03 0.03 -0.16 54 v Basal insulin 0.99 55 agonist (-0.63, 0.68) (-0.61, 0.68) (-0.27, -0.05) GLP-1 receptor Basal bolus -0.07 0.21 -0.02 56 v 0.45 57 agonist insulin (-0.54, 0.40) (-0.34, 0.77) (-0.37, 0.32) 58 GLP-1 receptor -0.79 -0.68 -0.69 v Standard therapy 0.57 59 agonist (-1.14, -0.45) (-0.85, -0.51) (-0.80, -0.58) 60

https://mc.manuscriptcentral.com/bmj Page 191 of 244 BMJ

1 2 3 Direct Indirect Incoherence Network Intervention Comparator 4 estimate estimate (p value) estimate 5 Basal bolus -0.33 0.28 0.13 6 Basal insulin v 0.26 insulin (-1.26, 0.60) (-0.26, 0.83) (-0.22, 0.49) 7 0.40 0.02 0.18 8 Basal insulin v Bolus insulin 0.35 9 (-0.25, 1.05) (-0.43, 0.48) (-0.09, 0.46) -0.65 -0.60 -0.53 10 Basal insulin v Standard therapy 0.89 11 (-1.30, 0.00) (-0.81, -0.39) (-0.68, -0.39) Alpha 12 Confidential: For-0.37 Review-0.51 Only -0.32 13 glucosidase v Standard therapy 0.60 (-0.86, 0.12) (-0.70, -0.31) (-0.46, -0.19) 14 inhibitor 15 -1.14 0.59 -0.54 Glitinide v Standard therapy 0.20 16 (-1.95, -0.34) (-0.85, -0.33) (-0.72, -0.36) 17 Alpha 0.07 0.17 0.22 18 glucosidase v Glitinide 0.80 (-0.78, 0.64) (-0.10, 0.44) (0.03, 0.40) 19 inhibitor 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

https://mc.manuscriptcentral.com/bmj BMJ Page 192 of 244

1 2 3 Appendix 7 Network meta-analysis treatment estimates 4 5 Table 1 All-cause mortality 6 7 8 SGLT-2 1.14 1.33Confidential: 1.43 1.31 1.31 For1.47 1.03Review 1.28 1.15 Only2.11 1.29 1.50 9 inhibitor (1.02,1.26) (0.95,1.87) (1.21,1.69) (1.09,1.57) (1.17,1.45) (0.84,2.57) (0.24,4.36) (0.13,12.61) (0.39,3.39) (0.68,6.58) (1.19,1.40) (1.12,2.01) GLP-1 10 0.88 1.17 1.26 1.15 1.15 1.29 0.91 1.13 1.02 1.86 1.14 1.33 receptor 11 (0.79,0.98) (0.84,1.64) (1.08,1.48) (0.96,1.37) (1.05,1.26) (0.74,2.24) (0.22,3.83) (0.12,11.08) (0.35,2.99) (0.60,5.79) (1.07,1.21) (1.00,1.77) agonist 12 0.75 0.85 1.08 0.98 0.98 1.10 0.78 0.96 0.87 1.59 0.97 1.13 Metformin 13 (0.53,1.06) (0.61,1.20) (0.78,1.50) (0.70,1.37) (0.70,1.37) (0.58,2.10) (0.18,3.39) (0.10,9.70) (0.28,2.66) (0.49,5.09) (0.70,1.35) (0.76,1.69) 14 0.70 0.79 0.93 0.91 0.91 1.02 0.72 0.90 0.80 1.47 0.90 1.05 Sulfonylurea 15 (0.59,0.82) (0.68,0.93) (0.67,1.29) (0.76,1.10) (0.80,1.04) (0.58,1.81) (0.17,3.05) (0.09,8.83) (0.27,2.37) (0.47,4.57) (0.78,1.04) (0.78,1.41) 16 0.77 0.87 1.02 1.10 Thiazol 1.00 1.12 0.79 0.98 0.88 1.61 0.99 1.15 17 (0.64,0.92) (0.73,1.04) (0.73,1.42) (0.91,1.32) idinedione (0.84,1.19) (0.63,2.00) (0.19,3.36) (0.10,9.69) (0.30,2.61) (0.52,5.05) (0.84,1.17) (0.92,1.45) 18 0.77 0.87 1.02 1.10 1.00 DPP-4 1.12 0.79 0.98 0.88 1.61 0.99 1.15 19 (0.69,0.85) (0.79,0.95) (0.73,1.42) (0.96,1.26) (0.84,1.19) inhibitor (0.64,1.96) (0.19,3.33) (0.10,9.65) (0.30,2.59) (0.52,5.03) (0.92,1.06) (0.87,1.53) 0.68 0.77 0.91 0.98 0.89 0.89 0.70 0.87 0.79 1.44 0.88 1.03 20 Basal insulin 21 (0.39,1.19) (0.45,1.34) (0.48,1.73) (0.55,1.73) (0.50,1.58) (0.51,1.55) (0.16,3.13) (0.09,8.58) (0.23,2.64) (0.41,5.08) (0.51,1.53) (0.55,1.90) 22 0.97 1.10 1.29 1.39 1.27 1.26 1.42 Basal bolus 1.24 1.12 2.04 1.25 1.46 23 (0.23,4.09) (0.26,4.63) (0.29,5.62) (0.33,5.87) (0.30,5.38) (0.30,5.33) (0.32,6.32) insulin (0.08,18.21) (0.19,6.72) (0.33,12.74) (0.30,5.27) (0.34,6.30) 0.78 0.88 1.04 1.12 1.02 1.02 1.14 0.80 0.90 1.64 1.01 1.17 24 Bolus insulin 25 (0.08,7.66) (0.09,8.68) (0.10,10.42) (0.11,11.01) (0.10,10.05) (0.10,10.00) (0.12,11.21) (0.05,11.79) (0.07,11.22) (0.13,21.05) (0.10,9.88) (0.12,11.70) Alpha 26 0.87 0.98 1.15 1.24 1.13 1.13 1.27 0.90 1.11 1.83 1.12 1.31 glucosidase (0.29,2.56) (0.33,2.90) (0.38,3.54) (0.42,3.66) (0.38,3.36) (0.39,3.33) (0.38,4.27) (0.15,5.40) (0.09,13.90) (0.41,8.09) (0.38,3.29) (0.43,3.96) 27 inhibitor 28 0.47 0.54 0.63 0.68 0.62 0.62 0.70 0.49 0.61 0.55 0.61 0.71 Glitinide 29 (0.15,1.48) (0.17,1.68) (0.20,2.03) (0.22,2.11) (0.20,1.94) (0.20,1.93) (0.20,2.46) (0.08,3.05) (0.05,7.80) (0.12,2.42) (0.20,1.91) (0.22,2.28) 30 0.77 0.88 1.03 1.11 1.01 1.01 1.14 0.80 0.99 0.89 1.63 1.16 Placebo 31 (0.71,0.84) (0.82,0.94) (0.74,1.43) (0.96,1.28) (0.86,1.19) (0.94,1.08) (0.65,1.98) (0.19,3.37) (0.10,9.75) (0.30,2.62) (0.52,5.08) (0.88,1.54) 32 0.66 0.75 0.88 0.95 0.87 0.87 0.98 0.69 0.85 0.77 1.40 0.86 Standard 33 (0.50,0.89) (0.57,1.00) (0.59,1.32) (0.71,1.28) (0.69,1.09) (0.65,1.15) (0.53,1.81) (0.16,2.97) (0.09,8.51) (0.25,2.32) (0.44,4.48) (0.65,1.14) therapy 34 35 The columns present the row drug class compared to the column drug class. The rows present the row drug class compared to the column drug class. The 36 effect estimates are expressed as odds ratios and 95% confidence intervals. 37 38 39 40 41 42 43 https://mc.manuscriptcentral.com/bmj 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 193 of 244 BMJ

1 2 3 Table 2 Cardiovascular mortality 4 5 6 SGLT-2 1.06 1.14 1.21 1.16 1.19 1.36 2.69 1.19 1.20 0.78 1.21 1.36 7 inhibitor (0.91,1.23) (0.57,2.29) (0.93,1.57) (0.89,1.52) (1.02,1.39) (0.61,3.04) (0.27,26.61) (0.12,11.99) (0.25,5.70) (0.13,4.53) (1.07,1.36) (0.88,2.09) GLP-1 8 0.94 1.08Confidential: 1.14 1.10 1.12 For1.28 Review2.54 1.13 Only1.13 0.74 1.14 1.28 receptor 9 (0.81,1.10) (0.54,2.16) (0.88,1.47) (0.85,1.42) (0.97,1.29) (0.58,2.83) (0.26,25.01) (0.11,11.25) (0.24,5.37) (0.13,4.26) (1.04,1.25) (0.83,1.96) 10 agonist 11 0.88 0.93 1.06 1.02 1.04 1.19 2.36 1.05 1.05 0.68 1.06 1.19 Metformin 12 (0.44,1.76) (0.46,1.87) (0.54,2.10) (0.50,2.08) (0.52,2.08) (0.42,3.41) (0.22,25.57) (0.09,11.59) (0.19,5.71) (0.11,4.20) (0.53,2.11) (0.54,2.62) 13 0.83 0.88 0.94 0.96 0.98 1.13 2.23 0.99 0.99 0.65 1.00 1.12 Sulfonylurea 14 (0.64,1.08) (0.68,1.13) (0.48,1.87) (0.70,1.31) (0.79,1.22) (0.49,2.57) (0.23,21.95) (0.10,10.01) (0.21,4.71) (0.11,3.72) (0.79,1.27) (0.71,1.78) 15 0.86 0.91 0.98 1.04 Thiazol 1.02 1.17 2.32 1.03 1.03 0.67 1.04 1.17 16 (0.66,1.12) (0.70,1.18) (0.48,2.00) (0.76,1.42) idinedione (0.79,1.32) (0.51,2.67) (0.23,23.06) (0.10,10.40) (0.21,4.96) (0.12,3.89) (0.82,1.32) (0.82,1.65) 17 0.84 0.89 0.96 1.02 0.98 DPP-4 1.15 2.27 1.01 1.01 0.66 1.02 1.14 18 (0.72,0.98) (0.78,1.03) (0.48,1.92) (0.82,1.27) (0.76,1.27) inhibitor (0.51,2.55) (0.23,22.34) (0.10,10.09) (0.21,4.78) (0.11,3.80) (0.92,1.13) (0.74,1.75) 19 0.74 0.78 0.84 0.89 0.85 0.87 Basal 1.98 0.88 0.88 0.57 0.89 1.00 20 (0.33,1.64) (0.35,1.72) (0.29,2.40) (0.39,2.03) (0.37,1.95) (0.39,1.94) insulin (0.18,22.26) (0.09,8.77) (0.15,5.05) (0.08,3.93) (0.40,1.97) (0.41,2.42) 21 0.37 0.39 0.42 0.45 0.43 0.44 0.50 Basal bolus 0.44 0.44 0.29 0.45 0.50 22 (0.04,3.67) (0.04,3.88) (0.04,4.59) (0.05,4.42) (0.04,4.30) (0.04,4.34) (0.04,5.68) insulin (0.02,11.38) (0.03,7.04) (0.02,5.14) (0.05,4.42) (0.05,5.14) 23 0.84 0.89 0.96 1.01 0.97 0.99 1.14 2.26 Bolus 1.00 0.65 1.01 1.14 24 (0.08,8.41) (0.09,8.88) (0.09,10.58) (0.10,10.25) (0.10,9.86) (0.10,9.98) (0.11,11.38) (0.09,57.88) insulin (0.06,16.15) (0.04,11.82) (0.10,10.13) (0.11,11.77) 25 Alpha 26 0.84 0.89 0.95 1.01 0.97 0.99 1.14 2.25 1.00 0.65 1.01 1.13 glucosidase 27 (0.18,3.98) (0.19,4.21) (0.18,5.19) (0.21,4.81) (0.20,4.68) (0.21,4.71) (0.20,6.52) (0.14,35.66) (0.06,16.09) (0.06,6.77) (0.21,4.79) (0.23,5.66) inhibitor 28 1.28 1.36 1.46 1.55 1.49 1.52 1.74 3.45 1.53 1.53 1.55 1.74 29 Glitinide 30 (0.22,7.43) (0.23,7.88) (0.24,8.99) (0.27,8.93) (0.26,8.64) (0.26,8.80) (0.25,11.95) (0.19,61.33) (0.08,27.69) (0.15,15.92) (0.27,8.95) (0.29,10.42) 0.83 0.88 0.94 1.00 0.96 0.98 1.13 2.23 0.99 0.99 0.65 1.12 31 Placebo 32 (0.73,0.93) (0.80,0.96) (0.47,1.88) (0.79,1.27) (0.76,1.22) (0.88,1.09) (0.51,2.49) (0.23,21.95) (0.10,9.89) (0.21,4.69) (0.11,3.73) (0.74,1.70) 33 0.74 0.78 0.84 0.89 0.86 0.88 1.00 1.99 0.88 0.88 0.58 0.89 Standard 34 (0.48,1.14) (0.51,1.20) (0.38,1.86) (0.56,1.42) (0.61,1.21) (0.57,1.34) (0.41,2.43) (0.19,20.29) (0.08,9.14) (0.18,4.41) (0.10,3.45) (0.59,1.35) therapy 35 36 The columns present the row drug class compared to the column drug class. The rows present the row drug class compared to the column drug class. The 37 effect estimates are expressed as odds ratios and 95% confidence intervals. 38 39 40 41 42 43 https://mc.manuscriptcentral.com/bmj 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Page 194 of 244

1 2 3 Table 3 Nonfatal myocardial infarction 4 5 6 SGLT-2 1.05 1.32 1.25 1.36 1.16 1.16 0.38 1.35 0.22 0.33 1.14 1.18 7 inhibitor (0.93,1.19) (0.91,1.93) (0.98,1.58) (0.98,1.90) (1.01,1.33) (0.58,2.32) (0.04,3.80) (0.43,4.27) (0.05,1.01) (0.06,1.88) (1.03,1.27) (0.77,1.82) GLP-1 8 0.95 1.26Confidential: 1.19 1.30 1.11 For1.10 Review0.36 1.29 0.21Only 0.31 1.09 1.13 receptor 9 (0.84,1.08) (0.87,1.83) (0.94,1.49) (0.94,1.80) (0.98,1.25) (0.56,2.19) (0.04,3.61) (0.41,4.04) (0.05,0.96) (0.06,1.78) (1.01,1.18) (0.73,1.73) 10 agonist 11 0.75 0.79 0.94 1.03 0.88 0.87 0.29 1.02 0.17 0.25 0.86 0.89 Metformin 12 (0.52,1.10) (0.55,1.15) (0.65,1.35) (0.70,1.52) (0.61,1.26) (0.41,1.88) (0.03,2.94) (0.31,3.38) (0.04,0.79) (0.04,1.45) (0.60,1.24) (0.55,1.45) 13 0.80 0.84 1.06 1.09 0.93 0.93 0.31 1.09 0.18 0.27 0.92 0.95 Sulfonylurea 14 (0.63,1.02) (0.67,1.06) (0.74,1.53) (0.82,1.46) (0.76,1.14) (0.46,1.89) (0.03,3.07) (0.34,3.47) (0.04,0.81) (0.05,1.50) (0.74,1.14) (0.63,1.43) 15 0.73 0.77 0.97 0.91 Thiazol 0.85 0.85 0.28 0.99 0.16 0.24 0.84 0.87 16 (0.53,1.02) (0.56,1.07) (0.66,1.44) (0.69,1.22) idinedione (0.62,1.16) (0.41,1.77) (0.03,2.83) (0.31,3.23) (0.03,0.76) (0.04,1.40) (0.61,1.15) (0.63,1.19) 17 0.86 0.90 1.14 1.07 1.17 DPP-4 1.00 0.33 1.17 0.19 0.28 0.99 1.02 18 (0.75,0.99) (0.80,1.02) (0.79,1.64) (0.88,1.31) (0.86,1.61) inhibitor (0.50,1.99) (0.03,3.27) (0.37,3.67) (0.04,0.87) (0.05,1.61) (0.90,1.08) (0.67,1.55) 19 0.86 0.91 1.14 1.08 1.18 1.00 Basal 0.33 1.17 0.19 0.29 0.99 1.02 20 (0.43,1.73) (0.46,1.80) (0.53,2.46) (0.53,2.19) (0.57,2.45) (0.50,2.00) insulin (0.03,3.27) (0.38,3.64) (0.04,1.00) (0.04,1.84) (0.50,1.97) (0.48,2.19) 21 2.61 2.75 3.46 3.26 3.56 3.03 3.03 Basal bolus 3.54 0.58 0.86 2.99 3.09 22 (0.26,25.97) (0.28,27.19) (0.34,35.26) (0.33,32.56) (0.35,35.95) (0.31,30.13) (0.31,29.98) insulin (0.29,43.69) (0.04,9.02) (0.05,15.30) (0.30,29.65) (0.30,31.59) 23 0.74 0.78 0.98 0.92 1.01 0.86 0.86 0.28 Bolus 0.16 0.24 0.85 0.87 24 (0.23,2.33) (0.25,2.43) (0.30,3.24) (0.29,2.94) (0.31,3.28) (0.27,2.70) (0.27,2.66) (0.02,3.49) insulin (0.02,1.09) (0.03,1.95) (0.27,2.66) (0.26,2.91) 25 Alpha 26 4.52 4.75 5.99 5.63 6.16 5.25 5.23 1.73 6.12 1.49 5.17 5.34 glucosidase 27 (0.99,20.55) (1.05,21.55) (1.27,28.22) (1.23,25.71) (1.32,28.71) (1.16,23.82) (1.00,27.51) (0.11,26.95) (0.92,40.69) (0.16,14.04) (1.14,23.44) (1.12,25.53) inhibitor 28 3.03 3.18 4.01 3.77 4.13 3.52 3.51 1.16 4.10 0.67 3.47 3.58 29 Glitinide 30 (0.53,17.22) (0.56,18.06) (0.69,23.41) (0.67,21.30) (0.72,23.80) (0.62,19.93) (0.54,22.62) (0.07,20.54) (0.51,32.72) (0.07,6.31) (0.61,19.65) (0.61,21.12) 0.87 0.92 1.16 1.09 1.19 1.01 1.01 0.33 1.18 0.19 0.29 1.03 31 Placebo 32 (0.79,0.97) (0.85,0.99) (0.81,1.66) (0.88,1.35) (0.87,1.64) (0.92,1.11) (0.51,2.01) (0.03,3.31) (0.38,3.71) (0.04,0.88) (0.05,1.63) (0.68,1.58) 33 0.85 0.89 1.12 1.05 1.15 0.98 0.98 0.32 1.15 0.19 0.28 0.97 Standard 34 (0.55,1.31) (0.58,1.36) (0.69,1.82) (0.70,1.58) (0.84,1.58) (0.65,1.49) (0.46,2.10) (0.03,3.31) (0.34,3.82) (0.04,0.89) (0.05,1.65) (0.63,1.48) therapy 35 36 The columns present the row drug class compared to the column drug class. The rows present the row drug class compared to the column drug class. The 37 effect estimates are expressed as odds ratios and 95% confidence intervals. 38 39 40 41 42 43 https://mc.manuscriptcentral.com/bmj 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 195 of 244 BMJ

1 2 3 Table 4 Nonfatal stroke 4 5 6 SGLT-2 0.83 1.00 1.09 0.94 0.92 0.80 0.60 0.84 10.05 1.86 0.99 1.19 7 inhibitor (0.71,0.97) (0.64,1.56) (0.83,1.43) (0.68,1.28) (0.77,1.10) (0.35,1.79) (0.10,3.49) (0.09,8.17) (0.81,124.98) (0.27,12.60) (0.87,1.12) (0.76,1.88) GLP-1 8 1.20 1.21Confidential: 1.32 1.13 1.11 For0.96 Review0.72 1.01 12.10Only 2.24 1.19 1.44 receptor 9 (1.03,1.41) (0.78,1.86) (1.01,1.71) (0.83,1.53) (0.94,1.30) (0.43,2.13) (0.13,4.18) (0.10,9.78) (0.97,150.32) (0.33,15.15) (1.08,1.32) (0.92,2.25) 10 agonist 11 1.00 0.83 1.09 0.94 0.92 0.80 0.60 0.84 10.04 1.86 0.99 1.19 Metformin 12 (0.64,1.55) (0.54,1.28) (0.72,1.65) (0.60,1.45) (0.60,1.40) (0.32,1.95) (0.10,3.63) (0.08,8.46) (0.79,127.24) (0.27,12.91) (0.65,1.51) (0.69,2.08) 13 0.91 0.76 0.92 0.86 0.84 0.73 0.55 0.77 9.19 1.70 0.91 1.09 Sulfonylurea 14 (0.70,1.20) (0.58,0.99) (0.61,1.38) (0.64,1.15) (0.67,1.05) (0.32,1.67) (0.09,3.21) (0.08,7.54) (0.75,112.84) (0.25,11.37) (0.71,1.15) (0.70,1.70) 15 1.07 0.89 1.07 1.17 Thiazol 0.98 0.85 0.64 0.90 10.74 1.99 1.06 1.28 16 (0.78,1.47) (0.65,1.21) (0.69,1.66) (0.87,1.57) idinedione (0.73,1.32) (0.37,1.96) (0.11,3.74) (0.09,8.86) (0.86,133.44) (0.29,13.39) (0.79,1.42) (0.90,1.81) 17 1.09 0.90 1.09 1.19 1.02 DPP-4 0.87 0.66 0.91 10.94 2.02 1.08 1.30 18 (0.91,1.30) (0.77,1.06) (0.71,1.67) (0.95,1.49) (0.76,1.37) inhibitor (0.39,1.95) (0.11,3.79) (0.09,8.89) (0.88,135.48) (0.30,13.65) (0.95,1.22) (0.84,2.02) 19 1.26 1.04 1.26 1.37 1.18 1.15 Basal 0.76 1.05 12.63 2.34 1.24 1.50 20 (0.56,2.83) (0.47,2.32) (0.51,3.09) (0.60,3.15) (0.51,2.70) (0.51,2.59) insulin (0.13,4.35) (0.10,11.70) (0.90,176.56) (0.30,18.40) (0.56,2.77) (0.62,3.61) 21 1.66 1.38 1.66 1.82 1.55 1.53 1.32 Basal bolus 1.40 16.70 3.09 1.64 1.98 22 (0.29,9.63) (0.24,7.96) (0.28,10.02) (0.31,10.60) (0.27,9.04) (0.26,8.83) (0.23,7.61) insulin (0.08,24.54) (0.78,357.11) (0.23,40.97) (0.28,9.49) (0.33,11.85) 23 1.19 0.99 1.19 1.30 1.11 1.09 0.95 0.72 Bolus 11.97 2.21 1.18 1.42 24 (0.12,11.58) (0.10,9.56) (0.12,12.01) (0.13,12.79) (0.11,11.00) (0.11,10.64) (0.09,10.51) (0.04,12.61) insulin (0.40,355.28) (0.11,43.05) (0.12,11.43) (0.14,14.38) 25 Alpha 26 0.10 0.08 0.10 0.11 0.09 0.09 0.08 0.06 0.08 0.19 0.10 0.12 glucosidase 27 (0.01,1.24) (0.01,1.03) (0.01,1.26) (0.01,1.34) (0.01,1.16) (0.01,1.13) (0.01,1.11) (0.00,1.28) (0.00,2.48) (0.02,2.27) (0.01,1.22) (0.01,1.50) inhibitor 28 0.54 0.45 0.54 0.59 0.50 0.49 0.43 0.32 0.45 5.40 0.53 0.64 29 Glitinide 30 (0.08,3.64) (0.07,3.02) (0.08,3.74) (0.09,3.93) (0.07,3.39) (0.07,3.33) (0.05,3.37) (0.02,4.29) (0.02,8.78) (0.44,66.36) (0.08,3.59) (0.09,4.36) 1.01 0.84 1.01 1.10 0.95 0.93 0.80 0.61 0.85 10.15 1.88 1.21 31 Placebo 32 (0.89,1.14) (0.76,0.93) (0.66,1.55) (0.87,1.41) (0.71,1.27) (0.82,1.05) (0.36,1.79) (0.11,3.51) (0.09,8.23) (0.82,125.88) (0.28,12.68) (0.78,1.87) 33 0.84 0.70 0.84 0.92 0.78 0.77 0.67 0.50 0.70 8.41 1.56 0.83 Standard 34 (0.53,1.32) (0.44,1.09) (0.48,1.46) (0.59,1.43) (0.55,1.11) (0.50,1.19) (0.28,1.61) (0.08,3.01) (0.07,7.11) (0.67,106.01) (0.23,10.58) (0.53,1.29) therapy 35 36 The columns present the row drug class compared to the column drug class. The rows present the row drug class compared to the column drug class. The 37 effect estimates are expressed as odds ratios and 95% confidence intervals. 38 39 40 41 42 43 https://mc.manuscriptcentral.com/bmj 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Page 196 of 244

1 2 3 Table 5 Kidney failure 4 5 6 SGLT-2 1.11 0.09 1.16 0.39 1.37 1.10 3.40 1.42 7 inhibitor (0.84,1.47) (0.01,1.60) (0.26,5.12) (0.03,5.88) (0.96,1.95) (0.19,6.18) (0.14,84.84) (1.13,1.78) GLP-1 8 0.90 Confidential:0.08 1.05 0.35 For1.23 Review0.98 3.05 Only1.27 receptor 9 (0.68,1.19) (0.00,1.46) (0.24,4.62) (0.02,5.30) (0.90,1.68) (0.18,5.42) (0.12,75.27) (1.08,1.50) 10 agonist 11 10.69 11.91 12.45 4.20 14.61 11.72 36.36 15.18 Metformin 12 (0.62,183.08) (0.69,206.68) (0.51,306.29) (0.08,212.21) (0.83,255.67) (0.42,325.89) (0.50,2656.71) (0.88,262.28) 13 0.86 0.96 0.08 0.34 1.17 0.94 2.92 1.22 Sulfonylurea 14 (0.20,3.78) (0.22,4.23) (0.00,1.98) (0.03,3.25) (0.27,5.13) (0.10,9.05) (0.09,99.97) (0.28,5.35) 15 2.55 2.84 0.24 2.97 Thiazol 3.48 2.79 8.67 3.62 16 (0.17,38.20) (0.19,42.67) (0.00,12.06) (0.31,28.60) idinedione (0.23,51.98) (0.11,68.72) (0.13,576.29) (0.24,54.14) 17 0.73 0.82 0.07 0.85 0.29 DPP-4 0.80 2.49 1.04 18 (0.51,1.04) (0.60,1.12) (0.00,1.20) (0.20,3.72) (0.02,4.29) inhibitor (0.14,4.55) (0.10,62.30) (0.79,1.36) 19 0.91 1.02 0.09 1.06 0.36 1.25 3.10 1.30 Basal insulin 20 (0.16,5.14) (0.18,5.60) (0.00,2.37) (0.11,10.21) (0.01,8.80) (0.22,7.06) (0.08,117.07) (0.23,7.19) 21 0.29 0.33 0.03 0.34 0.12 0.40 0.32 0.42 Bolus insulin 22 (0.01,7.34) (0.01,8.07) (0.00,2.01) (0.01,11.72) (0.00,7.67) (0.02,10.06) (0.01,12.17) (0.02,10.34) 23 0.70 0.78 0.07 0.82 0.28 0.96 0.77 2.40 24 Placebo (0.56,0.89) (0.67,0.92) (0.00,1.14) (0.19,3.60) (0.02,4.14) (0.73,1.26) (0.14,4.29) (0.10,59.29) 25 26 27 The columns present the row drug class compared to the column drug class. The rows present the row drug class compared to the column drug class. The 28 effect estimates are expressed as odds ratios and 95% confidence intervals. 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 https://mc.manuscriptcentral.com/bmj 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 197 of 244 BMJ

1 2 3 Table 6 Hospitalisation for heart failure 4 5 6 SGLT-2 1.35 1.74 1.30 2.26 1.53 1.09 0.45 4.73 1.46 1.22 7 inhibitor (1.15,1.58) (1.07,2.82) (1.01,1.68) (1.82,2.81) (1.30,1.80) (0.42,2.84) (0.02,11.23) (0.19,120.42) (1.29,1.64) (0.77,1.94) GLP-1 8 0.74 Confidential:1.29 0.96 1.67 1.13For 0.81Review 0.34 3.50 Only 1.08 0.90 receptor 9 (0.63,0.87) (0.80,2.07) (0.75,1.23) (1.36,2.05) (0.98,1.31) (0.31,2.08) (0.01,8.27) (0.14,89.00) (0.97,1.19) (0.57,1.43) 10 agonist 11 0.58 0.78 0.75 1.30 0.88 0.63 0.26 2.72 0.84 0.70 Metformin 12 (0.35,0.93) (0.48,1.25) (0.47,1.19) (0.82,2.06) (0.55,1.41) (0.22,1.80) (0.01,6.67) (0.10,71.49) (0.52,1.34) (0.38,1.30) 13 0.77 1.04 1.34 1.74 1.18 0.84 0.35 3.65 1.12 0.94 Sulfonylurea 14 (0.60,0.99) (0.81,1.33) (0.84,2.13) (1.37,2.21) (0.95,1.47) (0.32,2.22) (0.01,8.70) (0.14,93.29) (0.89,1.41) (0.59,1.51) 15 0.44 0.60 0.77 0.57 Thiazol 0.68 0.48 0.20 2.09 0.64 0.54 16 (0.36,0.55) (0.49,0.73) (0.49,1.22) (0.45,0.73) idinedione (0.55,0.83) (0.18,1.26) (0.01,4.98) (0.08,53.40) (0.54,0.77) (0.36,0.82) 17 0.65 0.88 1.14 0.85 1.48 DPP-4 0.71 0.30 3.10 0.95 0.80 18 (0.56,0.77) (0.76,1.02) (0.71,1.82) (0.68,1.06) (1.21,1.81) inhibitor (0.27,1.85) (0.01,7.34) (0.12,78.71) (0.85,1.06) (0.51,1.26) 19 0.92 1.24 1.60 1.19 2.08 1.40 Basal 0.42 4.35 1.34 1.12 20 (0.35,2.39) (0.48,3.20) (0.56,4.58) (0.45,3.15) (0.79,5.42) (0.54,3.65) insulin (0.01,11.77) (0.15,126.40) (0.52,3.45) (0.40,3.14) 21 2.20 2.98 3.83 2.86 4.98 3.37 2.40 Bolus 10.44 3.21 2.69 22 (0.09,54.59) (0.12,73.53) (0.15,97.93) (0.11,71.26) (0.20,123.76) (0.14,83.45) (0.08,67.93) insulin (0.11,992.13) (0.13,79.26) (0.11,68.59) 23 Alpha 24 0.21 0.29 0.37 0.27 0.48 0.32 0.23 0.10 0.31 0.26 glucosidase 25 (0.01,5.37) (0.01,7.26) (0.01,9.64) (0.01,7.01) (0.02,12.18) (0.01,8.21) (0.01,6.69) (0.00,9.11) (0.01,7.80) (0.01,6.75) inhibitor 26 0.69 0.93 1.19 0.89 1.55 1.05 0.75 0.31 3.25 0.84 27 Placebo (0.61,0.78) (0.84,1.03) (0.75,1.91) (0.71,1.12) (1.30,1.86) (0.94,1.17) (0.29,1.93) (0.01,7.70) (0.13,82.56) (0.54,1.31) 28 29 0.82 1.11 1.42 1.06 1.85 1.25 0.89 0.37 3.88 1.19 Standard 30 (0.52,1.30) (0.70,1.75) (0.77,2.64) (0.66,1.71) (1.22,2.80) (0.79,1.98) (0.32,2.50) (0.01,9.46) (0.15,101.46) (0.76,1.86) therapy 31 32 The columns present the row drug class compared to the column drug class. The rows present the row drug class compared to the column drug class. The 33 effect estimates are expressed as odds ratios and 95% confidence intervals. 34 35 36 37 38 39 40 41 42 43 https://mc.manuscriptcentral.com/bmj 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Page 198 of 244

1 2 3 Table 7 Severe hypoglycaemia 4 5 SGLT-2 6 1.07 1.82 6.78 1.39 1.23 2.13 2.15 3.09 1.25 3.50 1.12 0.56 inhibitor-2 7 (0.80, 1.41) (0.89, 3.71) (4.36, 10.55) (0.85, 2.27) (0.93, 1.63) (1.39, 3.27) (0.13, 36.24) (0.58, 16.33) (0.25, 6.35) (1.00, 12.23) (0.88, 1.44) (0.26, 1.21) 8 GLP-1 Confidential: For Review Only 9 0.94 1.71 6.37 1.30 1.15 2.00 2.01 2.90 1.17 3.29 1.06 0.53 receptor (0.71, 1.24) (0.84, 3.45) (4.23, 9.58) (0.85, 2.01) (0.96, 1.38) (1.45, 2.77) (0.12, 33.54) (0.56, 14.96) (0.23, 5.89) (0.95, 11.39) (0.92, 1.21) (0.25, 1.10) 10 agonist 11 0.55 0.59 3.73 0.76 0.68 1.17 1.18 1.70 0.69 1.93 0.62 0.31 Metformin 12 (0.27, 1.12) (0.29, 1.19) (1.75, 7.97) (0.34, 1.71) (0.34, 1.36) (0.54, 2.54) (0.07, 21.43) (0.28, 10.14) (0.12, 3.89) (0.62, 5.99) (0.31, 1.24) (0.11, 0.84) 13 0.15 0.16 0.27 0.20 0.18 0.31 0.32 0.46 0.18 0.52 0.17 0.08 Sulfonylurea 14 (0.09, 0.23) (0.10, 0.24) (0.13, 0.57) (0.12, 0.35) (0.12, 0.27) (0.19, 0.53) (0.02, 5.42) (0.08, 2.47) (0.04, 0.94) (0.15, 1.81) (0.11, 0.25) (0.04, 0.18) 15 0.72 0.77 1.31 4.88 Thiazol 0.88 1.53 1.54 2.22 0.90 2.52 0.81 0.40 16 (0.44, 1.17) (0.50, 1.18) (0.59, 2.92) (2.87, 8.30) idinedione (0.57, 1.36) (0.92, 2.55) (0.09, 26.41) (0.41, 12.13) (0.17, 4.72) (0.69, 9.23) (0.53, 1.23) (0.21, 0.80) 17 0.81 0.87 1.48 5.52 1.13 DPP- 1.74 1.75 2.51 1.02 2.85 0.92 0.46 18 (0.61, 1.08) (0.72, 1.04) (0.73, 2.98) (3.77, 8.09) (0.73, 1.74) 4_inhibitor (1.21, 2.50) (0.10, 29.23) (0.48, 13.11) (0.20, 5.10) (0.83, 9.84) (0.80, 1.05) (0.22, 0.95) 19 0.47 0.50 0.85 3.18 0.65 0.58 Basal 1.01 1.45 0.59 1.64 0.53 0.26 20 (0.31, 0.72) (0.36, 0.69) (0.39, 1.85) (1.90, 5.32) (0.39, 1.08) (0.40, 0.83) insulin (0.06, 16.44) (0.27, 7.72) (0.11, 3.03) (0.46, 5.92) (0.37, 0.75) (0.12, 0.57) Basal 21 0.47 0.50 0.85 3.16 0.65 0.57 0.99 1.44 0.58 1.63 0.52 0.26 bolus 22 (0.03, 7.86) (0.03, 8.27) (0.05, 15.37) (0.18, 54.12) (0.04, 11.08) (0.03, 9.58) (0.06, 16.24) (0.06, 37.36) (0.02, 14.90) (0.08, 35.30) (0.03, 8.74) (0.01, 4.76) insulin 23 0.32 0.34 0.59 2.20 0.45 0.40 0.69 0.69 Bolus 0.40 1.13 0.36 0.18 24 (0.06, 1.71) (0.07, 1.78) (0.10, 3.51) (0.40, 11.91) (0.08, 2.46) (0.08, 2.07) (0.13, 3.68) (0.03, 18.03) insulin (0.04, 4.04) (0.14, 8.88) (0.07, 1.89) (0.03, 1.10) 25 Alpha 0.80 0.85 1.45 5.42 1.11 0.98 1.70 1.72 2.47 2.80 0.90 0.45 26 glucosidase (0.16, 4.05) (0.17, 4.27) (0.26, 8.22) (1.06, 27.76) (0.21, 5.83) (0.20, 4.92) (0.33, 8.82) (0.07, 43.86) (0.25, 24.65) (0.39, 20.09) (0.18, 4.48) (0.08, 2.62) 27 inhibitor 28 0.29 0.30 0.52 1.94 0.40 0.35 0.61 0.61 0.88 0.36 0.32 0.16 Glitinide 29 (0.08, 1.00) (0.09, 1.05) (0.17, 1.61) (0.55, 6.77) (0.11, 1.45) (0.10, 1.21) (0.17, 2.19) (0.03, 13.25) (0.11, 6.91) (0.05, 2.56) (0.09, 1.10) (0.04, 0.67) 30 0.89 0.95 1.62 6.03 1.24 1.09 1.90 1.91 2.75 1.11 3.12 0.50 Placebo 31 (0.69, 1.14) (0.83, 1.09) (0.81, 3.23) (4.07, 8.94) (0.81, 1.89) (0.95, 1.26) (1.34, 2.69) (0.11, 31.88) (0.53, 14.27) (0.22, 5.55) (0.91, 10.73) (0.24, 1.03) 32 1.78 1.89 3.23 12.05 2.47 2.18 3.79 3.81 5.49 2.22 6.23 2.00 Standard 33 (0.83, 3.82) (0.91, 3.94) (1.19, 8.76) (5.42, 26.82) (1.25, 4.87) (1.05, 4.52) (1.75, 8.22) (0.21, 69.24) (0.91, 33.11) (0.38, 12.93) (1.49, 25.94) (0.97, 4.13) therapy 34 35 The columns present the row drug class compared to the column drug class. The rows present the row drug class compared to the column drug class. The 36 effect estimates are expressed as odds ratios and 95% confidence intervals. 37 38 39 40 41 42 43 https://mc.manuscriptcentral.com/bmj 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 199 of 244 BMJ

1 2 3 Table 8 Blindness 4 5 6 SGLT-2 6.04 5.99 6.03 7 inhibitor (0.18,205.45) (0.23,154.28) (0.25,148.05) 8 0.17 GLP-1Confidential: receptor 0.99 1.00 For Review Only 9 (0.00,5.64) agonist (0.20,4.83) (0.23,4.39) 10 0.17 1.01 DPP-4 1.01 11 (0.01,4.30) (0.21,4.90) inhibitor (0.58,1.75) 12 0.17 1.00 0.99 Placebo 13 (0.01,4.07) (0.23,4.41) (0.57,1.73) 14 15 The columns present the row drug class compared to the column drug class. The rows present the row drug class compared to the column drug class. The 16 17 effect estimates are expressed as odds ratios and 95% confidence intervals. 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 https://mc.manuscriptcentral.com/bmj 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Page 200 of 244

1 2 3 Table 9 Health-related quality of life 4 5 6 SGLT-2 -0.01 -0.13 0.12 0.03 -0.11 0.02 -0.35 -0.11 -0.00 -0.13 -0.45 7 inhibitor (-0.21, 0.20) (-0.46, 0.20) (-0.36, 0.60) (-0.33, 0.39) (-0.37, 0.14) (-0.32, 0.36) (-0.65, -0.05) (-0.51, 0.29) (-0.47, 0.46) (-0.36, 0.11) (-0.77, -0.14) GLP-1 8 0.01 -0.12 0.13 0.04 -0.11 0.03 -0.34 -0.11 0.00 -0.12 -0.45 receptor Confidential: For Review Only 9 (-0.20, 0.21) (-0.39, 0.14) (-0.31, 0.56) (-0.25, 0.33) (-0.25, 0.04) (-0.24, 0.30) (-0.56, -0.12) (-0.44, 0.23) (-0.42, 0.42) (-0.24, -0.01) (-0.68, -0.22) agonist 10 0.13 0.12 0.25 0.16 0.02 0.15 -0.22 0.02 0.13 0.00 -0.32 Metformin 11 (-0.20, 0.46) (-0.14, 0.39) (-0.26, 0.76) (-0.15, 0.48) (-0.24, 0.28) (-0.22, 0.53) (-0.56, 0.12) (-0.38, 0.42) (-0.31, 0.56) (-0.28, 0.29) (-0.66, 0.01) 12 -0.12 -0.13 -0.25 Sulfonyl -0.09 -0.23 -0.10 -0.47 -0.23 -0.13 -0.25 -0.58 13 (-0.60, 0.36) (-0.56, 0.31) (-0.76, 0.26) urea (-0.61, 0.44) (-0.69, 0.22) (-0.61, 0.41) (-0.95, 0.02) (-0.78, 0.32) (-0.73, 0.48) (-0.67, 0.17) (-1.07, -0.08) 14 -0.03 -0.04 -0.16 0.09 Thiazol -0.15 -0.01 -0.38 -0.15 -0.04 -0.16 -0.49 15 (-0.39, 0.33) (-0.33, 0.25) (-0.48, 0.15) (-0.44, 0.61) idinedione (-0.44, 0.15) (-0.41, 0.39) (-0.75, -0.01) (-0.57, 0.28) (-0.34, 0.26) (-0.47, 0.15) (-0.85, -0.13) 16 0.11 0.11 -0.02 0.23 0.15 DPP-4 0.14 -0.23 -0.00 0.11 -0.02 -0.34 17 (-0.14, 0.37) (-0.04, 0.25) (-0.28, 0.24) (-0.22, 0.69) (-0.15, 0.44) inhibitor (-0.17, 0.44) (-0.50, 0.03) (-0.30, 0.30) (-0.31, 0.53) (-0.20, 0.16) (-0.57, -0.11) 18 -0.02 -0.03 -0.15 0.10 0.01 -0.14 Basal -0.37 -0.14 -0.03 -0.15 -0.48 19 (-0.36, 0.32) (-0.30, 0.24) (-0.53, 0.22) (-0.41, 0.61) (-0.39, 0.41) (-0.44, 0.17) insulin (-0.63, -0.11) (-0.57, 0.30) (-0.53, 0.47) (-0.45, 0.14) (-0.83, -0.12) 20 0.35 0.34 0.22 0.47 0.38 0.23 0.37 Bolus 0.23 0.34 0.22 -0.11 21 (0.05, 0.65) (0.12, 0.56) (-0.12, 0.56) (-0.02, 0.95) (0.01, 0.75) (-0.03, 0.50) (0.11, 0.63) insulin (-0.17, 0.64) (-0.13, 0.82) (-0.03, 0.47) (-0.43, 0.21) Alpha 22 0.11 0.11 -0.02 0.23 0.15 0.00 0.14 -0.23 0.11 -0.02 -0.34 glucosidase 23 (-0.29, 0.51) (-0.23, 0.44) (-0.42, 0.38) (-0.32, 0.78) (-0.28, 0.57) (-0.30, 0.30) (-0.30, 0.57) (-0.64, 0.17) (-0.41, 0.63) (-0.37, 0.34) (-0.72, 0.04) inhibitor 24 0.00 -0.00 -0.13 0.13 0.04 -0.11 0.03 -0.34 -0.11 -0.12 -0.45 Glitinide 25 (-0.46, 0.47) (-0.42, 0.42) (-0.56, 0.31) (-0.48, 0.73) (-0.26, 0.34) (-0.53, 0.31) (-0.47, 0.53) (-0.82, 0.13) (-0.63, 0.41) (-0.56, 0.31) (-0.92, 0.02) 26 0.13 0.12 -0.00 0.25 0.16 0.02 0.15 -0.22 0.02 0.12 -0.33 Placebo 27 (-0.11, 0.36) (0.01, 0.24) (-0.29, 0.28) (-0.17, 0.67) (-0.15, 0.47) (-0.16, 0.20) (-0.14, 0.45) (-0.47, 0.03) (-0.34, 0.37) (-0.31, 0.56) (-0.58, -0.07) 28 0.45 0.45 0.32 0.58 0.49 0.34 0.48 0.11 0.34 0.45 0.33 Standard 29 (0.14, 0.77) (0.22, 0.68) (-0.01, 0.66) (0.08, 1.07) (0.13, 0.85) (0.11, 0.57) (0.12, 0.83) (-0.21, 0.43) (-0.04, 0.72) (-0.02, 0.92) (0.07, 0.58) therapy 30 31 The columns present the row drug class compared to the column drug class. The rows present the row drug class compared to the column drug class. The 32 33 effect estimates are expressed as standardised mean difference and 95% confidence interval. 34 35 36 37 38 39 40 41 42 43 https://mc.manuscriptcentral.com/bmj 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 201 of 244 BMJ

1 2 3 Table 10 Body weight 4 5 6 SGLT-2 0.56 1.19 3.64 4.54 2.27 4.14 4.31 2.81 1.70 2.99 1.94 2.44 7 inhibitor (0.18, 0.94) (0.71, 1.67) (3.20, 4.08) (4.11, 4.96) (1.92, 2.63) (3.56, 4.72) (3.12, 5.50) (1.30, 4.32) (1.10, 2.31) (2.12, 3.87) (1.63, 2.24) (1.83, 3.05) 8 GLP-1 Confidential: For Review Only -0.56 0.62 3.07 3.97 1.71 3.58 3.74 2.24 1.14 2.43 1.37 1.88 9 receptor (-0.94, -0.18) (0.17, 1.08) (2.67, 3.48) (3.58, 4.36) (1.40, 2.02) (3.10, 4.06) (2.60, 4.89) (0.78, 3.71) (0.56, 1.72) (1.57, 3.29) (1.11, 1.64) (1.30, 2.45) 10 agonist 11 -1.19 -0.62 2.45 3.35 1.09 2.95 3.12 1.62 0.52 1.81 0.75 1.25 Metformin 12 (-1.67, -0.71) (-1.08, -0.17) (1.97, 2.93) (2.89, 3.81) (0.66, 1.51) (2.32, 3.59) (1.91, 4.33) (0.09, 3.15) (-0.12, 1.15) (0.92, 2.69) (0.33, 1.17) (0.61, 1.89) 13 14 -3.64 -3.07 -2.45 Sulfonyl 0.90 -1.36 0.50 0.67 -0.83 -1.94 -0.64 -1.70 -1.20 15 (-4.08, -3.20) (-3.48, -2.67) (-2.93, -1.97) urea (0.51, 1.29) (-1.73, -1.00) (-0.09, 1.09) (-0.51, 1.85) (-2.34, 0.68) (-2.55, -1.32) (-1.47, 0.18) (-2.07, -1.33) (-1.82, -0.57) 16 -4.54 -3.97 -3.35 -0.90 Thiazol -2.26 -0.39 -0.23 -1.73 -2.83 -1.54 -2.60 -2.10 17 (-4.96, -4.11) (-4.36, -3.58) (-3.81, -2.89) (-1.29, -0.51) idinedione (-2.62, -1.90) (-0.97, 0.18) (-1.40, 0.95) (-3.24, -0.22) (-3.43, -2.24) (-2.39, -0.69) (-2.94, -2.26) (-2.70, -1.49) 18 19 -2.27 -1.71 -1.09 1.36 2.26 DPP-4 1.87 2.03 0.53 -0.57 0.72 -0.34 0.17 20 (-2.63, -1.92) (-2.02, -1.40) (-1.51, -0.66) (1.00, 1.73) (1.90, 2.62) inhibitor (1.34, 2.40) (0.87, 3.20) (-0.96, 2.03) (-1.12, -0.02) (-0.12, 1.56) (-0.58, -0.09) (-0.40, 0.73) 21 -4.14 -3.58 -2.95 -0.50 0.39 -1.87 Basal 0.17 -1.33 -2.44 -1.15 -2.20 -1.70 22 (-4.72, -3.56) (-4.06, -3.10) (-3.59, -2.32) (-1.09, 0.09) (-0.18, 0.97) (-2.40, -1.34) insulin (-0.99, 1.33) (-2.87, 0.21) (-3.17, -1.71) (-2.10, -0.19) (-2.72, -1.69) (-2.40, -1.00) 23 Basal -4.31 -3.74 -3.12 -0.67 0.23 -2.03 -0.17 -1.50 -2.61 -1.31 -2.37 -1.87 24 bolus (-5.50, -3.12) (-4.89, -2.60) (-4.33, -1.91) (-1.85, 0.51) (-0.95, 1.40) (-3.20, -0.87) (-1.33, 0.99) (-3.36, 0.36) (-3.87, -1.34) (-2.72, 0.09) (-3.53, -1.21) (-3.13, -0.61) 25 insulin 26 -2.81 -2.24 -1.62 0.83 1.73 -0.53 1.33 1.50 Bolus -1.11 0.19 -0.87 -0.37 27 (-4.32, -1.30) (-3.71, -0.78) (-3.15, -0.09) (-0.68, 2.34) (0.22, 3.24) (-2.03, 0.96) (-0.21, 2.87) (-0.36, 3.36) insulin (-2.68, 0.47) (-1.50, 1.88) (-2.36, 0.62) (-1.94, 1.20) 28 Alpha 29 -1.70 -1.14 -0.52 1.94 2.83 0.57 2.44 2.61 1.11 1.29 0.24 0.74 glucosidase 30 (-2.31, -1.10) (-1.72, -0.56) (-1.15, 0.12) (1.32, 2.55) (2.24, 3.43) (0.02, 1.12) (1.71, 3.17) (1.34, 3.87) (-0.47, 2.68) (0.32, 2.26) (-0.30, 0.77) (-0.02, 1.49) 31 inhibitor 32 -2.99 -2.43 -1.81 0.64 1.54 -0.72 1.15 1.31 -0.19 -1.29 -1.06 -0.55 Glitinide 33 (-3.87, -2.12) (-3.29, -1.57) (-2.69, -0.92) (-0.18, 1.47) (0.69, 2.39) (-1.56, 0.12) (0.19, 2.10) (-0.09, 2.72) (-1.88, 1.50) (-2.26, -0.32) (-1.89, -0.22) (-1.49, 0.38) 34 -1.94 -1.37 -0.75 1.70 2.60 0.34 2.20 2.37 0.87 -0.24 1.06 0.50 Placebo 35 (-2.24, -1.63) (-1.64, -1.11) (-1.17, -0.33) (1.33, 2.07) (2.26, 2.94) (0.09, 0.58) (1.69, 2.72) (1.21, 3.53) (-0.62, 2.36) (-0.77, 0.30) (0.22, 1.89) (-0.07, 1.07) 36 37 -2.44 -1.88 -1.25 1.20 2.10 -0.17 1.70 1.87 0.37 -0.74 0.55 -0.50 Standard 38 (-3.05, -1.83) (-2.45, -1.30) (-1.89, -0.61) (0.57, 1.82) (1.49, 2.70) (-0.73, 0.40) (1.00, 2.40) (0.61, 3.13) (-1.20, 1.94) (-1.49, 0.02) (-0.38, 1.49) (-1.07, 0.07) therapy 39 40 41 42 43 https://mc.manuscriptcentral.com/bmj 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Page 202 of 244

1 2 3 The columns present the row drug class compared to the column drug class. The rows present the row drug class compared to the column drug class. The 4 effect estimates are expressed as a mean difference (kilograms) and 95% confidence intervals. 5 6 7 8 Confidential: For Review Only 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 https://mc.manuscriptcentral.com/bmj 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 203 of 244 BMJ

1 2 3 Table 11 Amputation 4 5 6 SGLT-2 0.30 0.91 0.82 0.90 2.74 7 inhibitor (0.01,7.44) (0.51,1.64) (0.55,1.23) (0.74,1.11) (0.85,8.83) 8 GLP-1 Confidential: For Review Only 3.32 3.03 2.72 3.00 9.11 9 receptor (0.13,82.15) (0.12,78.00) (0.11,68.09) (0.12,73.67) (0.30,273.49) 10 agonist 11 1.10 0.33 Thiazol 0.90 0.99 3.00 12 (0.61,1.96) (0.01,8.48) idinedione (0.47,1.72) (0.57,1.71) (1.09,8.28) 13 14 1.22 0.37 1.11 DPP-4 1.10 3.35 15 (0.82,1.83) (0.01,9.20) (0.58,2.13) inhibitor (0.78,1.56) (1.01,11.15) 16 1.11 0.33 1.01 0.91 3.04 Placebo 17 (0.90,1.36) (0.01,8.19) (0.59,1.75) (0.64,1.29) (0.96,9.61) 18 19 0.36 0.11 0.33 0.30 0.33 Standard 20 (0.11,1.18) (0.00,3.30) (0.12,0.92) (0.09,0.99) (0.10,1.04) therapy 21 22 The columns present the row drug class compared to the column drug class. The rows present the row drug class compared to the column drug class. The 23 effect estimates are expressed as odds ratios and 95% confidence intervals. 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 https://mc.manuscriptcentral.com/bmj 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Page 204 of 244

1 2 3 Table 12 Neuropathic pain 4 5 GLP-1 6 9.01 3.00 receptor 7 (0.10,832.5) (0.12,73.61) 8 agonist Confidential: For Review Only 9 0.11 DPP-4 0.33 10 (0.00,10.27) inhibitor (0.01,8.17) 11 0.33 3.00 Placebo 12 (0.01,8.19) (0.12,73.75) 13 14 The columns present the row drug class compared to the column drug class. The rows present the row drug class compared to the column drug class. The 15 effect estimates are expressed as odds ratios and 95% confidence intervals. 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 https://mc.manuscriptcentral.com/bmj 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 205 of 244 BMJ

1 2 3 Table 13 Diabetic ketoacidosis 4 5 6 SGLT-2 0.63 0.67 0.28 0.38 0.19 0.18 0.52 7 inhibitor (0.27,1.48) (0.03,16.40) (0.04,1.93) (0.15,0.93) (0.00,7.86) (0.00,7.78) (0.29,0.92) GLP-1 8 1.59 Confidential:1.06 0.44 0.60 For0.29 Review0.29 0.83 Only receptor 9 (0.68,3.72) (0.04,29.09) (0.06,3.30) (0.23,1.54) (0.01,12.97) (0.01,12.83) (0.43,1.59) 10 agonist 11 1.50 0.95 0.42 0.57 0.28 0.28 0.78 Metformin 12 (0.06,36.98) (0.03,26.08) (0.01,17.62) (0.02,15.78) (0.00,38.55) (0.00,38.13) (0.03,20.29) 13 3.62 2.28 2.41 1.36 0.67 0.66 1.89 Sulfonylurea 14 (0.52,25.26) (0.30,17.18) (0.06,102.14) (0.21,8.88) (0.03,16.50) (0.03,16.32) (0.28,12.86) 15 2.65 1.67 1.77 0.73 DPP-4 0.49 0.49 1.38 16 (1.08,6.51) (0.65,4.30) (0.06,49.21) (0.11,4.78) inhibitor (0.01,20.11) (0.01,19.89) (0.68,2.81) 17 Alpha 5.39 3.40 3.59 1.49 2.03 0.99 2.81 18 glucosidase (0.13,228.17) (0.08,149.74) (0.03,496.09) (0.06,36.63) (0.05,83.01) (0.02,49.94) (0.07,117.51) 19 inhibitor 20 5.45 3.43 3.63 1.51 2.05 1.01 2.84 Glitinide 21 (0.13,230.67) (0.08,151.38) (0.03,501.54) (0.06,37.03) (0.05,83.92) (0.02,51.04) (0.07,118.80) 22 1.92 1.21 1.28 0.53 0.72 0.36 0.35 Placebo 23 (1.08,3.39) (0.63,2.33) (0.05,33.08) (0.08,3.62) (0.36,1.47) (0.01,14.89) (0.01,14.73) 24 25 26 The columns present the row drug class compared to the column drug class. The rows present the row drug class compared to the column drug class. The 27 effect estimates are expressed as odds ratios and 95% confidence intervals. 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 https://mc.manuscriptcentral.com/bmj 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Page 206 of 244

1 2 3 Table 14 Serious hyperglycaemia 4 5 6 SGLT-2 0.67 0.69 0.75 1.61 1.05 0.57 0.50 0.11 1.68 2.79 7 inhibitor (0.34,1.33) (0.16,2.95) (0.32,1.75) (0.50,5.22) (0.57,1.93) (0.20,1.62) (0.05,4.81) (0.02,0.53) (0.99,2.86) (0.78,9.95) GLP-1 8 1.48 Confidential:1.02 1.12 2.39 1.56 For 0.85 Review 0.75 0.17 Only2.50 4.14 receptor 9 (0.75,2.93) (0.25,4.11) (0.51,2.45) (0.82,7.00) (0.92,2.65) (0.38,1.91) (0.09,6.43) (0.04,0.77) (1.56,4.00) (1.24,13.84) 10 agonist 11 1.45 0.98 1.09 2.34 1.53 0.83 0.73 0.16 2.45 4.05 Metformin 12 (0.34,6.23) (0.24,3.94) (0.24,4.91) (0.47,11.58) (0.39,5.98) (0.17,4.05) (0.06,9.50) (0.02,1.20) (0.62,9.59) (0.73,22.56) 13 1.33 0.89 0.91 2.14 1.39 0.76 0.67 0.15 2.23 3.70 Sulfonylurea 14 (0.57,3.08) (0.41,1.96) (0.20,4.09) (0.67,6.79) (0.69,2.81) (0.25,2.27) (0.07,6.61) (0.03,0.75) (1.10,4.54) (1.02,13.38) 15 0.62 0.42 0.43 0.47 Thiazol 0.65 0.36 0.31 0.07 1.04 1.73 16 (0.19,2.01) (0.14,1.22) (0.09,2.11) (0.15,1.48) idinedione (0.23,1.88) (0.11,1.18) (0.03,3.46) (0.01,0.42) (0.36,3.04) (0.56,5.35) 17 0.95 0.64 0.66 0.72 1.54 DPP-4 0.55 0.48 0.11 1.60 2.66 18 (0.52,1.75) (0.38,1.09) (0.17,2.57) (0.36,1.45) (0.53,4.43) inhibitor (0.21,1.39) (0.05,4.40) (0.02,0.48) (1.13,2.27) (0.85,8.32) 19 1.74 1.18 1.20 1.31 2.81 1.83 Basal 0.88 0.20 2.94 4.87 20 (0.62,4.92) (0.52,2.64) (0.25,5.84) (0.44,3.93) (0.85,9.31) (0.72,4.68) insulin (0.09,8.75) (0.04,1.09) (1.18,7.32) (1.23,19.27) 21 1.99 1.34 1.37 1.50 3.20 2.09 1.14 Basal bolus 0.22 3.34 5.54 22 (0.21,18.98) (0.16,11.53) (0.11,17.76) (0.15,14.81) (0.29,35.51) (0.23,19.15) (0.11,11.34) insulin (0.02,3.13) (0.37,30.29) (0.47,65.41) 23 Alpha 24 8.88 5.99 6.11 6.69 14.32 9.32 5.09 4.47 14.95 24.78 glucosidase 25 (1.90,41.52) (1.30,27.45) (0.83,44.77) (1.33,33.57) (2.36,86.66) (2.10,41.37) (0.92,28.21) (0.32,62.47) (3.51,63.66) (3.84,159.69) inhibitor 26 0.59 0.40 0.41 0.45 0.96 0.62 0.34 0.30 0.07 1.66 27 Placebo (0.35,1.01) (0.25,0.64) (0.10,1.60) (0.22,0.91) (0.33,2.79) (0.44,0.88) (0.14,0.85) (0.03,2.71) (0.02,0.28) (0.51,5.35) 28 29 0.36 0.24 0.25 0.27 0.58 0.38 0.21 0.18 0.04 0.60 Standard 30 (0.10,1.28) (0.07,0.81) (0.04,1.37) (0.07,0.98) (0.19,1.79) (0.12,1.18) (0.05,0.81) (0.02,2.13) (0.01,0.26) (0.19,1.95) therapy 31 32 The columns present the row drug class compared to the column drug class. The rows present the row drug class compared to the column drug class. The 33 effect estimates are expressed as odds ratios and 95% confidence intervals. 34 35 36 37 38 39 40 41 42 43 https://mc.manuscriptcentral.com/bmj 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 207 of 244 BMJ

1 2 3 Table 15 Genital infection 4 5 6 SGLT-2 0.20 0.39 0.16 3.29 0.22 0.29 0.62 7 inhibitor (0.10,0.42) (0.21,0.75) (0.11,0.22) (0.13,84.35) (0.14,0.34) (0.24,0.36) (0.20,1.99) GLP-1 8 4.96 Confidential:1.95 0.77 16.30 1.08 For 1.45 Review 3.10 Only receptor 9 (2.39,10.29) (0.74,5.15) (0.34,1.75) (0.59,453.86) (0.46,2.54) (0.69,3.05) (0.79,12.18) 10 agonist 11 2.54 0.51 0.40 8.35 0.55 0.74 1.59 Metformin 12 (1.34,4.81) (0.19,1.35) (0.19,0.82) (0.31,228.16) (0.25,1.20) (0.38,1.46) (0.42,5.96) 13 6.42 1.29 2.53 21.10 1.40 1.88 4.01 Sulfonylurea 14 (4.46,9.26) (0.57,2.94) (1.21,5.26) (0.81,553.06) (0.79,2.48) (1.22,2.90) (1.19,13.51) 15 0.30 0.06 0.12 0.05 Thiazol 0.07 0.09 0.19 16 (0.01,7.81) (0.00,1.71) (0.00,3.27) (0.00,1.24) idinedione (0.00,1.75) (0.00,2.30) (0.01,5.97) 17 4.60 0.93 1.81 0.72 15.12 DPP-4 1.35 2.88 18 (2.93,7.22) (0.39,2.19) (0.83,3.95) (0.40,1.27) (0.57,400.46) inhibitor (0.83,2.18) (0.83,9.96) 19 3.42 0.69 1.34 0.53 11.23 0.74 2.14 Placebo 20 (2.79,4.18) (0.33,1.45) (0.69,2.63) (0.35,0.82) (0.43,290.10) (0.46,1.20) (0.66,6.92) 21 1.60 0.32 0.63 0.25 5.26 0.35 0.47 Standard 22 (0.50,5.09) (0.08,1.27) (0.17,2.36) (0.07,0.84) (0.17,164.87) (0.10,1.20) (0.14,1.52) therapy 23 24 25 The columns present the row drug class compared to the column drug class. The rows present the row drug class compared to the column drug class. The 26 effect estimates are expressed as odds ratios and 95% confidence intervals. 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 https://mc.manuscriptcentral.com/bmj 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Page 208 of 244

1 2 3 Table 16 Fournier gangrene 4 5 6 SGLT-2 7.02 1.80 7 inhibitor (0.56,87.98) (0.52,6.21) 0.14 0.26 8 DPP-4Confidential: inhibitor For Review Only 9 (0.01,1.78) (0.03,2.32) 10 0.56 3.91 Placebo 11 (0.16,1.92) (0.43,35.37) 12 13 The columns present the row drug class compared to the column drug class. The rows present the row drug class compared to the column drug class. The 14 effect estimates are expressed as odds ratios and 95% confidence intervals. 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 https://mc.manuscriptcentral.com/bmj 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 209 of 244 BMJ

1 2 3 Table 17 Severe gastrointestinal events 4 5 GLP-1 6 3.24 1.05 2.14 0.76 2.75 4.95 0.43 1.34 receptor 7 (0.18,58.19) (0.11,9.78) (0.12,38.18) (0.30,1.96) (0.63,11.95) (0.13,195.41) (0.22,0.82) (0.06,31.45) 8 agonist Confidential: For Review Only 0.31 0.33 0.66 0.23 0.85 1.53 0.13 0.41 9 Metformin 10 (0.02,5.56) (0.05,2.06) (0.17,2.61) (0.02,3.60) (0.04,20.55) (0.16,14.84) (0.01,2.44) (0.07,2.35) 11 0.95 3.07 2.03 0.72 2.61 4.69 0.41 1.27 Sulfonylurea 12 (0.10,8.79) (0.49,19.36) (0.32,12.64) (0.10,5.42) (0.20,34.87) (0.25,87.48) (0.04,3.85) (0.14,11.90) 13 0.47 1.51 0.49 Thiazol 0.36 1.29 2.32 0.20 0.63 14 (0.03,8.36) (0.38,5.98) (0.08,3.08) idinedione (0.02,5.42) (0.05,30.58) (0.16,32.95) (0.01,3.62) (0.14,2.72) 15 1.32 4.26 1.39 2.81 DPP-4 3.62 6.51 0.56 1.76 16 (0.51,3.38) (0.28,65.35) (0.18,10.42) (0.18,42.81) inhibitor (0.71,18.46) (0.19,227.30) (0.21,1.52) (0.09,35.80) 17 0.36 1.18 0.38 0.78 0.28 Basal 1.80 0.16 0.49 18 (0.08,1.58) (0.05,28.46) (0.03,5.13) (0.03,18.46) (0.05,1.41) insulin (0.04,90.08) (0.03,0.75) (0.02,14.93) 19 Alpha 0.20 0.65 0.21 0.43 0.15 0.56 0.09 0.27 20 glucosidase (0.01,7.98) (0.07,6.34) (0.01,3.97) (0.03,6.14) (0.00,5.36) (0.01,27.80) (0.00,3.48) (0.02,4.73) 21 inhibitor 22 2.33 7.55 2.46 4.98 1.77 6.42 11.55 3.12 Placebo 23 (1.21,4.48) (0.41,138.90) (0.26,23.30) (0.28,89.92) (0.66,4.79) (1.34,30.75) (0.29,464.18) (0.13,74.37) 24 0.75 2.42 0.79 1.60 0.57 2.06 3.70 0.32 Standard 25 (0.03,17.57) (0.42,13.78) (0.08,7.39) (0.37,6.92) (0.03,11.55) (0.07,63.06) (0.21,64.72) (0.01,7.63) therapy 26 27 28 The columns present the row drug class compared to the column drug class. The rows present the row drug class compared to the column drug class. The 29 effect estimates are expressed as odds ratios and 95% confidence intervals. 30 31 32 33 34 35 36 37 38 39 40 41 42 43 https://mc.manuscriptcentral.com/bmj 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Page 210 of 244

1 2 3 Table 18 Pancreatic cancer 4 5 6 SGLT-2 0.67 0.64 0.49 0.44 0.23 0.56 7 inhibitor (0.19,2.33) (0.16,2.51) (0.14,1.72) (0.06,3.07) (0.01,7.08) (0.17,1.83) GLP-1 8 1.49 Confidential:0.96 0.73 0.65 For0.34 Review0.84 Only receptor 9 (0.43,5.19) (0.40,2.30) (0.37,1.42) (0.15,2.90) (0.01,8.36) (0.55,1.28) 10 agonist 11 1.56 1.05 0.76 0.68 0.35 0.88 Sulfonylurea 12 (0.40,6.09) (0.43,2.52) (0.39,1.47) (0.12,3.86) (0.01,9.84) (0.39,1.96) 13 2.05 1.38 1.32 DPP-4 0.90 0.46 1.16 14 (0.58,7.25) (0.70,2.70) (0.68,2.55) inhibitor (0.18,4.63) (0.02,12.35) (0.69,1.95) 15 2.28 1.53 1.46 1.11 0.51 1.28 Basal insulin 16 (0.33,15.90) (0.34,6.77) (0.26,8.23) (0.22,5.69) (0.01,17.74) (0.27,6.05) 17 4.44 2.98 2.85 2.16 1.95 2.51 Bolus insulin 18 (0.14,139.86) (0.12,74.30) (0.10,79.88) (0.08,57.81) (0.06,67.53) (0.10,64.21) 19 1.77 1.19 1.14 0.86 0.78 0.40 Placebo 20 (0.55,5.75) (0.78,1.81) (0.51,2.54) (0.51,1.45) (0.17,3.66) (0.02,10.22) 21 22 23 The columns present the row drug class compared to the column drug class. The rows present the row drug class compared to the column drug class. The 24 effect estimates are expressed as odds ratios and 95% confidence intervals. 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 https://mc.manuscriptcentral.com/bmj 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 211 of 244 BMJ

1 2 3 Table 19 Pancreatitis 4 5 6 SGLT-2 1.91 0.81 2.23 1.85 2.38 0.78 0.84 4.50 13.39 1.63 0.84 7 inhibitor (0.90,4.06) (0.04,18.50) (0.80,6.20) (0.55,6.19) (1.10,5.14) (0.20,3.06) (0.03,20.72) (0.08,258.97) (0.46,385.77) (0.79,3.36) (0.17,4.08) GLP-1 8 0.52 Confidential:0.42 1.17 0.97 1.25 For0.41 Review0.44 2.36 Only7.02 0.85 0.44 receptor 9 (0.25,1.12) (0.02,9.01) (0.45,3.01) (0.36,2.62) (0.81,1.91) (0.13,1.28) (0.02,9.96) (0.04,133.28) (0.25,197.78) (0.65,1.13) (0.11,1.82) 10 agonist 11 1.24 2.37 2.77 2.30 2.95 0.96 1.04 5.59 16.61 2.02 1.05 Metformin 12 (0.05,28.47) (0.11,50.46) (0.12,66.55) (0.10,53.11) (0.14,62.21) (0.04,25.06) (0.01,81.22) (0.04,869.58) (0.18,1512.78) (0.09,42.90) (0.04,28.58) 13 0.45 0.85 0.36 0.83 1.06 0.35 0.38 2.02 5.99 0.73 0.38 Sulfonylurea 14 (0.16,1.24) (0.33,2.20) (0.02,8.67) (0.23,2.99) (0.41,2.75) (0.08,1.52) (0.01,9.70) (0.04,101.71) (0.24,147.32) (0.29,1.86) (0.07,1.95) 15 0.54 1.03 0.44 1.21 Thiazol 1.28 0.42 0.45 2.43 7.23 0.88 0.46 16 (0.16,1.80) (0.38,2.77) (0.02,10.06) (0.33,4.35) idinedione (0.47,3.52) (0.10,1.77) (0.02,11.08) (0.04,150.49) (0.23,227.53) (0.32,2.38) (0.15,1.38) 17 0.42 0.80 0.34 0.94 0.78 DPP-4 0.33 0.35 1.89 5.63 0.68 0.36 18 (0.19,0.91) (0.52,1.23) (0.02,7.15) (0.36,2.42) (0.28,2.14) inhibitor (0.10,1.10) (0.02,8.18) (0.03,107.02) (0.20,158.84) (0.48,0.97) (0.08,1.49) 19 1.29 2.46 1.04 2.88 2.39 3.07 1.09 5.81 17.26 2.10 1.09 Basal insulin 20 (0.33,5.08) (0.78,7.76) (0.04,27.08) (0.66,12.64) (0.56,10.11) (0.91,10.37) (0.04,29.82) (0.09,383.68) (0.51,587.37) (0.65,6.81) (0.20,5.84) 21 1.19 2.26 0.96 2.65 2.20 2.82 0.92 Basal bolus 5.35 15.89 1.93 1.00 22 (0.05,29.20) (0.10,51.09) (0.01,74.40) (0.10,68.23) (0.09,53.61) (0.12,65.16) (0.03,25.28) insulin (0.03,869.37) (0.17,1520.20) (0.08,43.98) (0.03,28.78) 23 Alpha 24 0.22 0.42 0.18 0.50 0.41 0.53 0.17 0.19 2.97 0.36 0.19 glucosidase 25 (0.00,12.77) (0.01,23.92) (0.00,27.87) (0.01,25.04) (0.01,25.47) (0.01,29.83) (0.00,11.38) (0.00,30.42) (0.12,73.13) (0.01,20.35) (0.00,13.15) inhibitor 26 0.07 0.14 0.06 0.17 0.14 0.18 0.06 0.06 0.34 0.12 0.06 27 Glitinide (0.00,2.15) (0.01,4.02) (0.00,5.49) (0.01,4.10) (0.00,4.36) (0.01,5.01) (0.00,1.97) (0.00,6.02) (0.01,8.28) (0.00,3.42) (0.00,2.30) 28 0.61 1.17 0.50 1.37 1.14 1.46 0.48 0.52 2.77 8.23 0.52 29 Placebo 30 (0.30,1.27) (0.89,1.55) (0.02,10.54) (0.54,3.51) (0.42,3.09) (1.03,2.08) (0.15,1.55) (0.02,11.80) (0.05,156.07) (0.29,231.52) (0.13,2.15) 31 1.18 2.26 0.95 2.64 2.19 2.81 0.92 1.00 5.33 15.85 1.93 Standard 32 (0.24,5.72) (0.55,9.29) (0.03,26.03) (0.51,13.62) (0.73,6.62) (0.67,11.78) (0.17,4.92) (0.03,28.61) (0.08,373.77) (0.43,578.51) (0.46,7.98) therapy 33 34 The columns present the row drug class compared to the column drug class. The rows present the row drug class compared to the column drug class. The 35 effect estimates are expressed as odds ratios and 95% confidence intervals. 36 37 38 39 40 41 42 43 https://mc.manuscriptcentral.com/bmj 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Page 212 of 244

1 2 3 Table 20 Glycated haemoglobin A1C 4 5 6 SGLT-2 -0.28 -0.19 -0.05 -0.09 0.00 -0.13 -0.26 -0.31 0.08 -0.13 0.61 0.41 7 inhibitor (-0.38, -0.19) (-0.30, -0.08) (-0.15, 0.05) (-0.19, 0.00) (-0.08, 0.09) (-0.26, 0.01) (-0.61, 0.09) (-0.59, -0.03) (-0.04, 0.20) (-0.31, 0.04) (0.54, 0.68) (0.29, 0.53) GLP-1 8 0.28 0.09 0.23 0.19 0.29 0.16 0.02 -0.02 0.37 0.15 0.90 0.69 receptor Confidential: For Review Only 9 (0.19, 0.38) (-0.01, 0.20) (0.14, 0.33) (0.10, 0.28) (0.21, 0.36) (0.05, 0.27) (-0.32, 0.37) (-0.29, 0.24) (0.25, 0.48) (-0.02, 0.32) (0.83, 0.96) (0.58, 0.80) agonist 10 0.19 -0.09 0.14 0.10 0.20 0.07 -0.07 -0.12 0.27 0.06 0.80 0.60 Metformin 11 (0.08, 0.30) (-0.20, 0.01) (0.04, 0.25) (0.00, 0.19) (0.10, 0.29) (-0.07, 0.21) (-0.42, 0.29) (-0.40, 0.16) (0.15, 0.40) (-0.12, 0.24) (0.71, 0.90) (0.48, 0.72) 12 0.05 -0.23 -0.14 Sulfonyl -0.04 0.05 -0.07 -0.21 -0.26 0.13 -0.08 0.66 0.46 13 (-0.05, 0.15) (-0.33, -0.14) (-0.25, -0.04) urea (-0.13, 0.04) (-0.03, 0.14) (-0.21, 0.06) (-0.56, 0.14) (-0.53, 0.02) (0.01, 0.25) (-0.24, 0.08) (0.58, 0.74) (0.34, 0.57) 14 0.09 -0.19 -0.10 0.04 Thiazol 0.10 -0.03 -0.17 -0.21 0.18 -0.04 0.71 0.50 15 (-0.00, 0.19) (-0.28, -0.10) (-0.19, -0.00) (-0.04, 0.13) idinedione (0.02, 0.18) (-0.16, 0.10) (-0.52, 0.18) (-0.49, 0.06) (0.06, 0.29) (-0.21, 0.13) (0.63, 0.78) (0.39, 0.61) 16 -0.00 -0.29 -0.20 -0.05 -0.10 DPP-4 -0.13 -0.26 -0.31 0.08 -0.14 0.61 0.40 17 (-0.09, 0.08) (-0.36, -0.21) (-0.29, -0.10) (-0.14, 0.03) (-0.18, -0.02) inhibitor (-0.25, -0.01) (-0.61, 0.08) (-0.58, -0.04) (-0.03, 0.18) (-0.30, 0.03) (0.55, 0.66) (0.30, 0.50) 18 0.13 -0.16 -0.07 0.07 0.03 0.13 Basal -0.13 -0.18 0.21 -0.01 0.74 0.53 19 (-0.01, 0.26) (-0.27, -0.05) (-0.21, 0.07) (-0.06, 0.21) (-0.10, 0.16) (0.01, 0.25) insulin (-0.49, 0.22) (-0.46, 0.09) (0.05, 0.36) (-0.21, 0.19) (0.62, 0.86) (0.39, 0.68) Basal 20 0.26 -0.02 0.07 0.21 0.17 0.26 0.13 -0.05 0.34 0.13 0.87 0.67 bolus 21 (-0.09, 0.61) (-0.37, 0.32) (-0.29, 0.42) (-0.14, 0.56) (-0.18, 0.52) (-0.08, 0.61) (-0.22, 0.49) (-0.48, 0.38) (-0.02, 0.70) (-0.25, 0.51) (0.52, 1.22) (0.31, 1.02) 22 insulin 0.31 0.02 0.12 0.26 0.21 0.31 0.18 0.05 Bolus 0.39 0.18 0.92 0.71 23 (0.03, 0.59) (-0.24, 0.29) (-0.16, 0.40) (-0.02, 0.53) (-0.06, 0.49) (0.04, 0.58) (-0.09, 0.46) (-0.38, 0.48) insulin (0.11, 0.68) (-0.14, 0.49) (0.65, 1.19) (0.43, 1.00) 24 Alpha -0.08 -0.37 -0.27 -0.13 -0.18 -0.08 -0.21 -0.34 -0.39 -0.22 0.53 0.32 25 glucosidase (-0.20, 0.04) (-0.48, -0.25) (-0.40, -0.15) (-0.25, -0.01) (-0.29, -0.06) (-0.18, 0.03) (-0.36, -0.05) (-0.70, 0.02) (-0.68, -0.11) (-0.40, -0.03) (0.43, 0.63) (0.19, 0.46) 26 inhibitor 27 0.13 -0.15 -0.06 0.08 0.04 0.14 0.01 -0.13 -0.18 0.22 0.74 0.54 Glitinide 28 (-0.04, 0.31) (-0.32, 0.02) (-0.24, 0.12) (-0.08, 0.24) (-0.13, 0.21) (-0.03, 0.30) (-0.19, 0.21) (-0.51, 0.25) (-0.49, 0.14) (0.03, 0.40) (0.58, 0.91) (0.36, 0.72) 29 -0.61 -0.90 -0.80 -0.66 -0.71 -0.61 -0.74 -0.87 -0.92 -0.53 -0.74 -0.21 Placebo 30 (-0.68, -0.54) (-0.96, -0.83) (-0.90, -0.71) (-0.74, -0.58) (-0.78, -0.63) (-0.66, -0.55) (-0.86, -0.62) (-1.22, -0.52) (-1.19, -0.65) (-0.63, -0.43) (-0.91, -0.58) (-0.31, -0.10) 31 -0.41 -0.69 -0.60 -0.46 -0.50 -0.40 -0.53 -0.67 -0.71 -0.32 -0.54 0.21 Standard 32 (-0.53, -0.29) (-0.80, -0.58) (-0.72, -0.48) (-0.57, -0.34) (-0.61, -0.39) (-0.50, -0.30) (-0.68, -0.39) (-1.02, -0.31) (-1.00, -0.43) (-0.46, -0.19) (-0.72, -0.36) (0.10, 0.31) therapy 33 34 The columns present the row drug class compared to the column drug class. The rows present the row drug class compared to the column drug class. The 35 effect estimates are expressed as a mean difference (%) and 95% confidence interval. 36 37 38 39 40 41 42 43 https://mc.manuscriptcentral.com/bmj 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 213 of 244 BMJ

1 2 3 Appendix 8 GRADE summary of findings for SGLT-2 inhibitors and GLP-1 receptor agonists compared to placebo or each other 4 5 6 Table 1: Cardiovascular mortality 7 8 Confidential:Anticipated absolute effects over 5 years For Review Only 9 Certainty 10 Anticipated in 11 absolute effects treatment 12 Relative Baseline Risk with Risk with (95% CI) over effects 13 Comparison effect risk control intervention 5 years (GRADE) Plain text summary 14 SGLT 2 OR 0.83 Very low Placebo: 13 SGLT-2 2 fewer per Moderate SGLT-2 inhibitor therapy probably reduces 15 16 inhibitor v (0.73 to per 1000 inhibitor: 11 1000 due to cardiovascular mortality in people with diabetes 17 placebo 0.93) per 1000 (from 1 to 4 directness and few or no cardiovascular risk factors. 18 fewer) 19 Low Placebo: 46 SGLT-2 8 fewer per High SGLT-2 inhibitor therapy reduces cardiovascular 20 per 1000 inhibitor: 38 1000 mortality in people with diabetes and 21 per 1000 (from 3 fewer to cardiovascular risk factors. 22 12 fewer) 23 Moderate Placebo: 79 SGLT-2 13 fewer per High SGLT-2 inhibitor therapy reduces cardiovascular 24 per 1000 inhibitor: 66 1000 mortality in people with diabetes and established 25 per 1000 (from 6 fewer to cardiovascular disease. 26 21 fewer) 27 High Placebo: 112 SGLT-2 19 fewer per High SGLT-2 inhibitor therapy reduces cardiovascular 28 per 1000 inhibitor: 93 1000 mortality in people with diabetes and chronic 29 per 1000 (from 8 fewer to kidney disease. 30 30 fewer) 31 Very high Placebo: 175 SGLT-2 30 fewer per High SGLT-2 inhibitor therapy reduces cardiovascular 32 33 per 1000 inhibitor: 145 1000 mortality in people with diabetes and established 34 per 1000 (from 12 fewer cardiovascular disease and chronic kidney 35 to 47 fewer) disease. 36 37 38 39 40 41 42 43 https://mc.manuscriptcentral.com/bmj 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Page 214 of 244

1 2 3 Table 1: Cardiovascular mortality 4 5 Anticipated absolute effects over 5 years 6 7 Certainty 8 Confidential: ForAnticipated Review in Only 9 absolute effects treatment 10 Relative Baseline Risk with Risk with (95% CI) over effects 11 Comparison effect risk control intervention 5 years (GRADE) Plain text summary 12 GLP-1 OR 0.88 Very low Placebo: 13 GLP-1 2 fewer per Moderate GLP-1 receptor agonist therapy probably reduces 13 receptor (0.80 to per 1000 receptor 1000 due to cardiovascular mortality in people with diabetes 14 agonist v 0.96) agonist: 11 (from 1 fewer to directness and few or no cardiovascular risk factors. 15 placebo per 1000 3 fewer) 16 Low Placebo: 46 GLP-1 6 fewer per High GLP-1 receptor agonist therapy reduces 17 per 1000 receptor 1000 cardiovascular mortality in people with diabetes 18 agonist: 40 (from 2 fewer to and cardiovascular risk factors. 19 per 1000 9 fewer) 20 Moderate Placebo: 79 GLP-1 9 fewer per High reduces cardiovascular mortality in people with 21 per 1000 receptor 1000 diabetes and established cardiovascular disease. 22 23 agonist: 70 (from 3 fewer to 24 per 1000 16 fewer) 25 High Placebo: 112 GLP-1 13 fewer per High GLP-1 receptor agonist therapy reduces 26 per 1000 receptor 1000 cardiovascular mortality in people with diabetes 27 agonist: 99 (from 4 fewer to and chronic kidney disease. 28 per 1000 22 fewer) 29 Very high Placebo: 175 GLP-1 21 fewer per High GLP-1 receptor agonist therapy reduces 30 per 1000 receptor 1000 cardiovascular mortality in people with diabetes 31 agonist: 154 (from 7 fewer to and established cardiovascular disease and 32 per 1000 35 fewer) chronic kidney disease. 33 34 35 36 37 38 39 40 41 42 43 https://mc.manuscriptcentral.com/bmj 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 215 of 244 BMJ

1 2 3 Table 1: Cardiovascular mortality 4 5 Anticipated absolute effects over 5 years 6 7 Certainty 8 Confidential: ForAnticipated Review in Only 9 absolute effects treatment 10 Relative Baseline Risk with Risk with (95% CI) over effects 11 Comparison effect risk control intervention 5 years (GRADE) Plain text summary 12 SGLT 2 OR 0.94 Very low GLP-1 SGLT-2 1 fewer per Moderate There is probably no difference between SGLT-2 13 inhibitor v (0.81 to receptor inhibitor: 10 1000 (from 2 due to inhibitor therapy and GLP-1 receptor agonist 14 GLP-1 1.10) agonist: 11 per 1000 fewer to 1 more) directness therapy on cardiovascular mortality in people 15 receptor per 1000 with diabetes and few or no cardiovascular risk 16 agonist factors. 17 Low GLP-1 SGLT-2 2 fewer per High There is no difference between SGLT-2 inhibitor 18 receptor inhibitor: 38 1000 therapy and GLP-1 receptor agonist therapy on 19 agonist: 40 per 1000 (from 8 fewer to cardiovascular mortality in people with diabetes 20 per 1000 4 more) and cardiovascular risk factors. 21 Moderate GLP-1 SGLT-2 4 fewer per High There is no difference between SGLT-2 inhibitor 22 23 receptor inhibitor: 66 1000 therapy and GLP-1 receptor agonist therapy on 24 agonist: 70 per 1000 (from 13 fewer cardiovascular mortality in people with diabetes 25 per 1000 to 7 more) and established cardiovascular disease. 26 High GLP-1 SGLT-2 6 fewer per High There is no difference between SGLT-2 inhibitor 27 receptor inhibitor: 93 1000 therapy and GLP-1 receptor agonist therapy on 28 agonist: 99 per 1000 (from 19 fewer cardiovascular mortality in people with diabetes 29 per 1000 to 10 more) and chronic kidney disease. 30 Very high GLP-1 SGLT-2 9 fewer per High There is no difference between SGLT-2 inhibitor 31 receptor inhibitor: 145 1000 therapy and GLP-1 receptor agonist therapy on 32 agonist: 154 per 1000 (29 fewer to 15 cardiovascular mortality in people with diabetes 33 per 1000 more) and established cardiovascular disease and 34 chronic kidney disease. 35 CI= confidence interval. OR = odds ratio. We used the point estimate of absolute effect for GLP-1 receptor agonist therapy, obtained from GLP-1 receptor 36 agonist therapy versus placebo, to calculate absolute effect for SGLT-2 inhibitors v GLP-1 receptor agonists. 37 38 39 40 41 42 43 https://mc.manuscriptcentral.com/bmj 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Page 216 of 244

1 2 3 4 Table 2 Nonfatal myocardial infarction 5 6 Anticipated absolute effects over 5 years 7 Anticipated Certainty 8 Confidential: Forabsolute Review in Only 9 10 effects (95% treatment 11 Relative Risk with Risk with CI) over 5 effects 12 Comparison effect Baseline risk control intervention years (GRADE) Plain text summary 13 SGLT-2 OR 0.87 Very low Placebo: 30 SGLT-2 4 fewer per Moderate SGLT-2 inhibitor therapy probably reduces 14 inhibitor (0.79 to per 1000 inhibitor: 26 per 1000 due to nonfatal myocardial infarction in people with 15 versus 0.97) 1000 (from 1 fewer indirectness diabetes and few or no cardiovascular risk factors. 16 placebo to 6 fewer) 17 Low Placebo: 58 SGLT-2 8 fewer per High SGLT-2 inhibitor therapy reduces nonfatal 18 per 1000 inhibitor: 50 per 1000 myocardial infarction in people with diabetes and 19 1000 (from 2 fewer cardiovascular risk factors. 20 to 12 fewer) 21 Moderate Placebo: 108 SGLT-2 14 fewer per High SGLT-2 inhibitor therapy reduces nonfatal 22 per 1000 inhibitor: 94 per 1000 myocardial infarction in people with diabetes and 23 1000 (from 3 fewer established cardiovascular disease. 24 to 23 fewer) 25 High Placebo: SGLT-2 16 fewer per High SGLT-2 inhibitor therapy reduces nonfatal 26 27 120 per 1000 inhibitor: 104 1000 myocardial infarction in people with diabetes and 28 per 1000 (from 4 fewer chronic kidney disease. 29 to 25 fewer) 30 Very high Placebo: SGLT-2 25 fewer per High SGLT-2 inhibitor therapy reduces nonfatal 31 190 per 1000 inhibitor: 165 1000 myocardial infarction in people with diabetes and 32 per 1000 (from 6 fewer established cardiovascular disease and chronic 33 to 40 fewer) kidney disease. 34 35 36 37 38 39 40 41 42 43 https://mc.manuscriptcentral.com/bmj 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 217 of 244 BMJ

1 2 3 4 Table 2 Nonfatal myocardial infarction 5 6 Anticipated absolute effects over 5 years 7 Anticipated Certainty 8 Confidential: Forabsolute Review in Only 9 10 effects (95% treatment 11 Relative Risk with Risk with CI) over 5 effects 12 Comparison effect Baseline risk control intervention years (GRADE) Plain text summary 13 GLP-1 OR 0.92 Very low Placebo: 30 GLP-1 receptor 2 fewer per Moderate GLP-1 receptor agonist therapy probably reduces 14 receptor (0.85 to per 1000 agonist: 28 per 1000 (from 0 due to nonfatal myocardial infarction in people with 15 agonist 0.99) 1000 fewer to 5 indirectness diabetes and few or no cardiovascular risk factors. 16 versus fewer) 17 placebo Low Placebo: 58 GLP-1 receptor 5 fewer per High GLP-1 receptor agonist therapy reduces nonfatal 18 per 1000 agonist: 53 per 1000 myocardial infarction in people with diabetes and 19 1000 (from 1 fewer cardiovascular risk factors. 20 to 9 fewer) 21 Moderate Placebo: 108 GLP-1 receptor 9 fewer per High GLP-1 receptor agonist therapy reduces nonfatal 22 per 1000 agonist: 99 per 1000 myocardial infarction in people with diabetes and 23 1000 (from 1 fewer established cardiovascular disease. 24 to 16 fewer) 25 High Placebo: GLP-1 receptor 10 fewer per High GLP-1 receptor agonist therapy reduces nonfatal 26 27 120 per 1000 agonist: 110 per 1000 myocardial infarction in people with diabetes and 28 1000 (from 1 fewer chronic kidney disease. 29 to 18 fewer) 30 Very high Placebo: GLP-1 receptor 15 fewer per High GLP-1 receptor agonist therapy reduces nonfatal 31 190 per 1000 agonist: 175 per 1000 myocardial infarction in people with diabetes and 32 1000 (from 2 fewer established cardiovascular disease and chronic 33 to 29 fewer) kidney disease. 34 35 36 37 38 39 40 41 42 43 https://mc.manuscriptcentral.com/bmj 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Page 218 of 244

1 2 3 4 Table 2 Nonfatal myocardial infarction 5 6 Anticipated absolute effects over 5 years 7 Anticipated Certainty 8 Confidential: Forabsolute Review in Only 9 10 effects (95% treatment 11 Relative Risk with Risk with CI) over 5 effects 12 Comparison effect Baseline risk control intervention years (GRADE) Plain text summary 13 Based on OR 0.95 Very low GLP-1 SGLT-2 1 fewer per Moderate There is probably no difference between SGLT-2 14 SGLT-2 (0.84 to receptor inhibitor: 27 per 1000 (from 4 due to inhibitor therapy and GLP-1 receptor agonist 15 inhibitor 1.08) agonist: 28 1000 fewer to 2 indirectness therapy on nonfatal myocardial infarction in 16 versus GLP- per 1000 more) people with diabetes and few or no cardiovascular 17 1 receptor risk factors. 18 agonist Low GLP-1 SGLT-2 3 fewer per High There is no difference between SGLT-2 inhibitor 19 receptor inhibitor: 50 per 1000 therapy and GLP-1 receptor agonist therapy on 20 agonist: 53 1000 (from 8 fewer nonfatal myocardial infarction in people with 21 per 1000 to 4 more) diabetes and cardiovascular risk factors. 22 Moderate GLP-1 SGLT-2 5 fewer per High There is no difference between SGLT-2 inhibitor 23 receptor inhibitor: 94 per 1000 therapy and GLP-1 receptor agonist therapy on 24 agonist: 99 1000 (from 16 nonfatal myocardial infarction in people with 25 per 1000 fewer to 8 diabetes and established cardiovascular disease. 26 more) 27 28 High GLP-1 SGLT-2 6 fewer per High There is no difference between SGLT-2 inhibitor 29 receptor inhibitor: 105 1000 therapy and GLP-1 receptor agonist therapy on 30 agonist: 110 per 1000 (from 18 nonfatal myocardial infarction in people with 31 per 1000 fewer to 9 diabetes and chronic kidney disease. 32 more) 33 Very high GLP-1 SGLT-2 9 fewer per High There is no difference between SGLT-2 inhibitor 34 receptor inhibitor: 166 1000 therapy and GLP-1 receptor agonist therapy on 35 agonist: 175 per 1000 (28 fewer to nonfatal myocardial infarction in people with 36 per 1000 14 more) diabetes and established cardiovascular disease 37 and chronic kidney disease. 38 39 40 41 42 43 https://mc.manuscriptcentral.com/bmj 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 219 of 244 BMJ

1 2 3 CI= confidence interval. OR = odds ratio. We used the point estimate of absolute effect for GLP-1 receptor agonists, obtained from GLP-1 receptor agonists 4 versus placebo, to calculate absolute effect for SGLT-2 inhibitors v GLP-1 receptor agonists. 5 6 7 8 Confidential: For Review Only 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 https://mc.manuscriptcentral.com/bmj 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Page 220 of 244

1 2 3 4 Table 3 Nonfatal stroke 5 6 Anticipated absolute effects over 5 years 7 Anticipated 8 Confidential: For Review Only 9 absolute Certainty in 10 effects (95% treatment 11 Relative Risk with Risk with CI) over 5 effects 12 Comparison effect Baseline risk control intervention years (GRADE) Plain text summary 13 SGLT-2 OR 1.01 Very low Placebo: 30 SGLT-2 0 fewer per Moderate SGLT-2 inhibitors probably do not reduce 14 inhibitor (0.89 to per 1000 inhibitor: 30 1000 (from 3 due to nonfatal stroke in people with diabetes and few or 15 versus 1.14) per 1000 fewer to 4 indirectness no cardiovascular risk factors. 16 placebo more) 17 Low Placebo: 58 SGLT-2 1 more per High SGLT-2 inhibitors do not reduce nonfatal stroke 18 per 1000 inhibitor: 59 1000 in people with diabetes and cardiovascular risk 19 per 1000 (from 6 fewer factors. 20 to 8 more) 21 Moderate Placebo: SGLT-2 1 more per High SGLT-2 inhibitors do not reduce nonfatal stroke 22 108 per inhibitor: 1000 in people with diabetes and established 23 1000 109 per (from 12 cardiovascular disease. 24 1000 fewer to 15 25 26 more) 27 High Placebo: SGLT-2 1 more per High SGLT-2 inhibitors do not reduce nonfatal stroke 28 120 per inhibitor: 1000 in people with diabetes and chronic kidney 29 1000 121 per (from 13 disease. 30 1000 fewer to 17 31 more) 32 Very high Placebo: SGLT-2 2 more per High SGLT-2 inhibitors do not have an important 33 190 per inhibitor: 1000 effect on nonfatal stroke in people with diabetes 34 1000 192 per (from 21 and established cardiovascular disease and 35 1000 fewer to 17 chronic kidney disease. 36 more) 37 38 39 40 41 42 43 https://mc.manuscriptcentral.com/bmj 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 221 of 244 BMJ

1 2 3 4 Table 3 Nonfatal stroke 5 6 Anticipated absolute effects over 5 years 7 Anticipated 8 Confidential: For Review Only 9 absolute Certainty in 10 effects (95% treatment 11 Relative Risk with Risk with CI) over 5 effects 12 Comparison effect Baseline risk control intervention years (GRADE) Plain text summary 13 GLP-1 OR 0.84 Very low Placebo: 30 GLP-1 5 fewer per Moderate GLP-1 receptor agonist therapy probably reduces 14 receptor (0.76 to per 1000 receptor 1000 (from 2 due to nonfatal stroke in people with diabetes and few or 15 agonist 0.93) agonist: 25 fewer to 7 indirectness no cardiovascular risk factors. 16 versus per 1000 fewer) 17 placebo Low Placebo: 58 GLP-1 9 fewer per High GLP-1 receptor agonist therapy reduces nonfatal 18 per 1000 receptor 1000 stroke in people with diabetes and cardiovascular 19 agonist: 49 (from 4 fewer risk factors. 20 per 1000 to 14 fewer) 21 Moderate Placebo: GLP-1 17 fewer per High GLP-1 receptor agonist therapy reduces nonfatal 22 108 per receptor 1000 stroke in people with diabetes and established 23 1000 agonist: 91 (from 8 fewer cardiovascular disease. 24 per 1000 to 26 fewer) 25 26 High Placebo: GLP-1 19 fewer per High GLP-1 receptor agonist therapy reduces nonfatal 27 120 per receptor 1000 stroke in people with diabetes and chronic kidney 28 1000 agonist: 101 (from 8 fewer disease. 29 per 1000 to 29 fewer) 30 Very high Placebo: GLP-1 30 fewer per High GLP-1 receptor agonist therapy reduces nonfatal 31 190 per receptor 1000 stroke in people with diabetes and established 32 1000 agonist: 160 (from 13 cardiovascular disease and chronic kidney 33 per 1000 fewer to 46 disease. 34 fewer) 35 36 37 38 39 40 41 42 43 https://mc.manuscriptcentral.com/bmj 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Page 222 of 244

1 2 3 4 Table 3 Nonfatal stroke 5 6 Anticipated absolute effects over 5 years 7 Anticipated 8 Confidential: For Review Only 9 absolute Certainty in 10 effects (95% treatment 11 Relative Risk with Risk with CI) over 5 effects 12 Comparison effect Baseline risk control intervention years (GRADE) Plain text summary 13 SGLT-2 OR 1.20 Very low GLP-1 SGLT-2 5 more per Moderate GLP-1 receptor agonist therapy probably reduces 14 inhibitor (1.03 to receptor inhibitor: 1000 due to nonfatal stroke compared to SGLT-2 inhibitor 15 versus GLP- 1.41) agonist: 25 30per 1000 (from 1 more indirectness therapy in people with diabetes and few or no 16 1 receptor per 1000 to 10 more) cardiovascular risk factors. 17 agonist Low GLP-1 SGLT-2 10 more per High GLP-1 receptor agonist therapy reduces nonfatal 18 receptor inhibitor: 59 1000 stroke compared to SGLT-2 inhibitor therapy in 19 agonist: 49 per 1000 (from 1 more people with diabetes and cardiovascular risk 20 per 1000 to 12 more) factors. 21 Moderate GLP-1 SGLT-2 18 more per High GLP-1 receptor agonist therapy reduces nonfatal 22 receptor inhibitor: 1000 stroke compared to SGLT-2 inhibitor therapy in 23 agonist: 91 109 per (from 3 more people with diabetes and established 24 per 1000 1000 to 37 more) cardiovascular disease. 25 26 High GLP-1 SGLT-2 20 more per High GLP-1 receptor agonist therapy reduces nonfatal 27 receptor inhibitor: 1000 stroke compared to SGLT-2 inhibitor therapy in 28 agonist: 121 per (from 3 more people with diabetes and chronic kidney disease. 29 101 per 1000 to 41 more) 30 1000 31 Very high GLP-1 SGLT-2 32 more per High GLP-1 receptor agonist therapy reduces nonfatal 32 receptor inhibitor: 1000 stroke compared to SGLT-2 inhibitor therapy in 33 agonist: 192 per (from 5 more people with diabetes and established 34 160 per 1000 to 66 more) cardiovascular disease and chronic kidney 35 1000 disease. 36 CI= confidence interval. OR = odds ratio. We used the point estimate of absolute effect for GLP-1 receptor agonists, obtained from GLP-1 receptor agonists 37 versus placebo, to calculate absolute effect for SGLT-2 inhibitors v GLP-1 receptor agonists. 38 39 40 41 42 43 https://mc.manuscriptcentral.com/bmj 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 223 of 244 BMJ

1 2 3 4 Table 4 Kidney failure 5 6 Anticipated absolute effects over 5 years 7 Certainty 8 Confidential: ForAnticipated Review in Only 9 10 absolute effects treatment 11 Relative Risk with Risk with (95% CI) over effects 12 Comparison effect Baseline risk control intervention 5 years (GRADE) Plain text summary 13 SGLT-2 OR 0.70 Very low Placebo: 2 SGLT-2 1 fewer per Moderate SGLT-2 inhibitor therapy probably reduces 14 inhibitor (0.56 to per 1000 inhibitor: 1 1000 due to kidney failure in people with diabetes and few or 15 versus 0.89) per 1000 (from 0 fewer to indirectness no risk factors for cardiovascular disease. 16 placebo 1 fewer) 17 Low Placebo: 10 SGLT-2 3 fewer per High SGLT-2 inhibitor therapy reduces kidney failure 18 per 1000 inhibitor: 7 1000 (from 1 to in people with diabetes and cardiovascular risk 19 per 1000 4 fewer) factors. 20 Moderate Placebo: 20 SGLT-2 6 fewer per High SGLT-2 inhibitor therapy reduces kidney failure 21 per 1000 inhibitor: 14 1000 (from 2 to in people with diabetes and established 22 per 1000 9 fewer) cardiovascular disease. 23 High Placebo: 92 SGLT-2 28 fewer High SGLT-2 inhibitor therapy reduces kidney failure 24 per 1000 inhibitor: 64 (from 10 fewer in people with diabetes and chronic kidney 25 to 40 fewer) disease. 26 27 Very high Placebo: SGLT-2 44 fewer per High SGLT-2 inhibitor therapy reduces kidney failure 28 148 per inhibitor: 1000 in people with diabetes and established 29 1000 104 (from 16 fewer cardiovascular disease and chronic kidney 30 to 65 fewer) disease. 31 32 33 34 35 36 37 38 39 40 41 42 43 https://mc.manuscriptcentral.com/bmj 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Page 224 of 244

1 2 3 4 Table 4 Kidney failure 5 6 Anticipated absolute effects over 5 years 7 Certainty 8 Confidential: ForAnticipated Review in Only 9 10 absolute effects treatment 11 Relative Risk with Risk with (95% CI) over effects 12 Comparison effect Baseline risk control intervention 5 years (GRADE) Plain text summary 13 GLP-1 OR 0.78 Very low Placebo: 2 GLP-1 0 fewer per Moderate GLP-1 receptor agonist therapy probably reduces 14 receptor (0.67 to per 1000 receptor 1000 due to kidney failure in people with diabetes and few or 15 agonist 0.92) agonist: 2 (from 0 fewer to indirectness no risk factors for cardiovascular disease. 16 versus per 1000 1 fewer) 17 placebo Low Placebo: 10 GLP-1 2 fewer per High GLP-1 receptor agonist therapy reduces kidney 18 per 1000 receptor 1000 (from 1 failure in people with diabetes and cardiovascular 19 agonist: 8 fewer to 3 risk factors. 20 per 1000 fewer) 21 Moderate Placebo: 20 GLP-1 4 fewer per High GLP-1 receptor agonist therapy reduces kidney 22 per 1000 receptor 1000 (from 2 failure in people with diabetes and established 23 agonist: 16 fewer to 7 cardiovascular disease. 24 per 1000 fewer) 25 High Placebo: 92 GLP-1 20 fewer per High GLP-1 receptor agonist therapy reduces kidney 26 27 per 1000 receptor 1000 failure in people with diabetes and chronic kidney 28 agonist: 72 (from 7 fewer to disease. 29 per 1000 30 fewer) 30 Very high Placebo: GLP-1 33 fewer per High GLP-1 receptor agonist therapy reduces kidney 31 148 per receptor 1000 failure in people with diabetes and established 32 1000 agonist: 115 (from 12 fewer cardiovascular disease and chronic kidney 33 per 1000 to 49 fewer) disease. 34 35 36 37 38 39 40 41 42 43 https://mc.manuscriptcentral.com/bmj 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 225 of 244 BMJ

1 2 3 4 5 Table 4 Kidney failure 6 7 Anticipated absolute effects over 5 years 8 Confidential: For Review Only 9 Certainty 10 Anticipated in 11 absolute effects treatment 12 Relative Risk with Risk with (95% CI) over effects 13 Comparison effect Baseline risk control intervention 5 years (GRADE) Plain text summary 14 SGLT-2 OR 0.90 Very low GLP-1 SGLT-2 0 fewer per Low due to There is may be no difference between SGLT-2 15 inhibitor (0.68 to receptor inhibitor: 2 1000 serious inhibitor therapy and GLP-1 receptor agonist 16 versus GLP- 1.19) agonist: 2 per 1000 (from 0 fewer to imprecision therapy on kidney failure in people with diabetes 17 1 receptor per 1000 1 more) and and few or no cardiovascular risk factors. 18 agonist indirectness 19 Low GLP-1 SGLT-2 1 fewer per Moderate There is probably no difference between SGLT-2 20 receptor inhibitor: 9 1000 (from 3 due to inhibitor therapy and GLP-1 receptor agonist 21 agonist: 10 per 1000 fewer to 2 serious therapy on kidney failure in people with diabetes 22 per 1000 more) imprecision and cardiovascular risk factors. 23 Moderate GLP-1 SGLT-2 2 fewer per Moderate There is probably no difference between SGLT-2 24 receptor inhibitor: 14 1000 (from 5 due to inhibitor therapy and GLP-1 receptor agonist 25 26 agonist: 16 per 1000 fewer to 3 serious therapy on kidney failure in people with diabetes 27 per 1000 more) imprecision and established cardiovascular disease. 28 High GLP-1 SGLT-2 7 fewer per Moderate There is probably no difference between SGLT-2 29 receptor inhibitor: 65 1000 due to inhibitor therapy and GLP-1 receptor agonist 30 agonist: 72 per 1000 (from 23 fewer serious therapy on kidney failure in people with diabetes 31 per 1000 to 14 more) imprecision and chronic kidney disease. 32 Very high GLP-1 SGLT-2 12 fewer per Moderate There is probably no difference between SGLT-2 33 receptor inhibitor: 1000 due to inhibitor therapy and GLP-1 receptor agonist 34 agonist: 115 104 per (from 37 fewer serious therapy on kidney failure in people with diabetes 35 per 1000 1000 to 22 more) imprecision and established cardiovascular disease and 36 chronic kidney disease. 37 CI= confidence interval. OR = odds ratio. We used the point estimate of absolute effect for GLP-1 receptor agonists, obtained from GLP-1 receptor agonists 38 versus placebo, to calculate absolute effect for SGLT-2 inhibitors v GLP-1 receptor agonists. 39 40 41 42 43 https://mc.manuscriptcentral.com/bmj 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Page 226 of 244

1 2 3 4 5 Table 5 Hospitalisation for heart failure 6 7 Anticipated absolute effects over 5 years 8 Confidential: For Review Only 9 Certainty 10 Anticipated in 11 absolute effects treatment 12 Relative Risk with Risk with (95% CI) over effects 13 Comparison effect Baseline risk control intervention 5 years (GRADE) Plain text summary 14 SGLT-2 OR 0.69 Very low Placebo: 5 SGLT-2 2 fewer per Moderate SGLT-2 inhibitor therapy probably reduces 15 inhibitor versus (0.61 to per 1000 inhibitor: 3 1000 due to hospitalisation for heart failure in people with 16 placebo 0.78) per 1000 (from 1 fewer to indirectness diabetes and few or no risk factors for 17 2 fewer) cardiovascular disease. 18 Low Placebo: 30 SGLT-2 9 fewer per High SGLT-2 inhibitor therapy reduces hospitalisation 19 per 1000 inhibitor: 21 1000 (from 7 for heart failure in people with diabetes and 20 per 1000 fewer to 12 cardiovascular risk factors. 21 fewer) 22 Moderate Placebo: 80 SGLT-2 25 fewer per High SGLT-2 inhibitor therapy reduces hospitalisation 23 per 1000 inhibitor: 55 1000 (from 18 for heart failure in people with diabetes and 24 per 1000 fewer to 31 established cardiovascular disease. 25 26 fewer) 27 High Placebo: SGLT-2 33 fewer High SGLT-2 inhibitor therapy reduces hospitalisation 28 105 per inhibitor: 72 (from 23 fewer for heart failure in people with diabetes and 29 1000 per 1000 to 41 fewer) chronic kidney disease. 30 Very high Placebo: SGLT-2 73 fewer High SGLT-2 inhibitor therapy reduces hospitalisation 31 235 per inhibitor: (from 52 fewer for heart failure in people with diabetes and 32 1000 162 per 1000 to 92 fewer) established cardiovascular disease and chronic 33 kidney disease. 34 35 36 37 38 39 40 41 42 43 https://mc.manuscriptcentral.com/bmj 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 227 of 244 BMJ

1 2 3 4 Table 5 Hospitalisation for heart failure 5 6 Anticipated absolute effects over 5 years 7 Certainty 8 Confidential: ForAnticipated Review in Only 9 10 absolute effects treatment 11 Relative Risk with Risk with (95% CI) over effects 12 Comparison effect Baseline risk control intervention 5 years (GRADE) Plain text summary 13 GLP-1 receptor OR 0.93 Very low Placebo: 5 GLP-1 0 fewer per Moderate GLP-1 receptor agonist therapy probably has little 14 agonist versus (0.84 to per 1000 receptor 1000 due to or no effect on hospitalisation for heart failure in 15 placebo 1.03) agonist: 5 per (from 0 fewer to indirectness people with diabetes and few or no risk factors for 16 1000 1 fewer) cardiovascular disease. 17 Low Placebo: 30 GLP-1 2 fewer per High GLP-1 receptor agonist therapy has little or no 18 per 1000 receptor 1000 (from 5 effect on hospitalisation for heart failure in people 19 agonist: 28 fewer to 1 with diabetes and cardiovascular risk factors. 20 per 1000 more) 21 Moderate Placebo: 80 GLP-1 6 fewer per High GLP-1 receptor agonist therapy has little or no 22 per 1000 receptor 1000 (from 13 effect on hospitalisation for heart failure in people 23 agonist: 74 fewer to 2 with diabetes and established cardiovascular 24 per 1000 more) disease. 25 High Placebo: GLP-1 7 fewer per High GLP-1 receptor agonist therapy has little or no 26 27 105 per receptor 1000 effect on hospitalisation for heart failure in people 28 1000 agonist: 98 (from 17 fewer with diabetes and chronic kidney disease. 29 per 1000 to 3 more) 30 Very high Placebo: GLP-1 16 fewer per High GLP-1 receptor agonist therapy probably has little 31 235 per receptor 1000 or no effect on hospitalisation for heart failure in 32 1000 agonist: 219 (from 38 fewer people with diabetes and established 33 per 1000 to 7 more) cardiovascular disease and chronic kidney 34 disease. 35 36 37 38 39 40 41 42 43 https://mc.manuscriptcentral.com/bmj 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Page 228 of 244

1 2 3 4 Table 5 Hospitalisation for heart failure 5 6 Anticipated absolute effects over 5 years 7 Certainty 8 Confidential: ForAnticipated Review in Only 9 10 absolute effects treatment 11 Relative Risk with Risk with (95% CI) over effects 12 Comparison effect Baseline risk control intervention 5 years (GRADE) Plain text summary 13 SGLT-2 OR 0.74 Very low GLP-1 SGLT-2 1 fewer per Moderate SGLT-2 inhibitor therapy probably reduces 14 inhibitor versus (0.63 to receptor inhibitor: 4 1000 (from 1 due to hospitalisation for heart failure compared with 15 GLP-1 receptor 0.87) agonist: 5 per 1000 fewer to 2 indirectness GLP-1 receptor agonist therapy in people with 16 agonist per 1000 fewer) diabetes and few or no risk factors for 17 cardiovascular disease. 18 Low GLP-1 SGLT-2 7 fewer per High SGLT-2 inhibitor therapy reduces hospitalisation 19 receptor inhibitor: 21 1000 for heart failure compared with GLP-1 receptor 20 agonist: 28 per 1000 (from 4 fewer to agonist therapy in people with diabetes and 21 per 1000 10 fewer) cardiovascular risk factors. 22 Moderate GLP-1 SGLT-2 19 fewer per High SGLT-2 inhibitor therapy reduces hospitalisation 23 receptor inhibitor: 55 1000 for heart failure compared with GLP-1 receptor 24 agonist: 74 per 1000 (from 10 fewer agonist therapy in people with diabetes and 25 per 1000 to 27 fewer) established cardiovascular disease. 26 27 High GLP-1 SGLT-2 25 fewer per High SGLT-2 inhibitor therapy reduces hospitalisation 28 receptor inhibitor: 73 1000 for heart failure compared with GLP-1 receptor 29 agonist: 98 per 1000 (from 13 fewer agonist therapy in people with diabetes and 30 per 1000 to 36 fewer) chronic kidney disease. 31 Very high GLP-1 SGLT-2 57 fewer High SGLT-2 inhibitor therapy reduces hospitalisation 32 receptor inhibitor: (from 28 fewer for heart failure compared with GLP-1 receptor 33 agonist: 162 per 1000 to 81 fewer) agonist therapy in people with diabetes and 34 219 per established cardiovascular disease and chronic 35 1000 kidney disease. 36 CI= confidence interval. OR = odds ratio. We used the point estimate of absolute effect for GLP-1 receptor agonists, obtained from GLP-1 receptor agonists 37 versus placebo, to calculate absolute effect for SGLT-2 inhibitors v GLP-1 receptor agonists. 38 39 40 41 42 43 https://mc.manuscriptcentral.com/bmj 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 229 of 244 BMJ

1 2 3 Table 6 Severe hypoglycaemia 4 5 Anticipated absolute effects over 5 years 6 7 Anticipated Certainty in 8 Confidential:absolute For effects Reviewtreatment Only 9 Relative Risk with Risk with (95% CI) over 5 effects 10 Comparison effect control intervention years (GRADE) Plain text summary 11 SGLT-2 OR 0.89 Placebo: 25 SGLT-2 3 fewer per 1000 High SGLT-2 inhibitor therapy does not incur severe 12 inhibitor versus (0.69,1.14) per 1000 inhibitor: 22 (from 8 fewer to hypoglycaemia 13 placebo per 1000 3 more) 14 GLP-1 receptor OR 0.95 Placebo: 25 GLP-1 1 fewer per 1000 High GLP-1 receptor agonist therapy does not incur 15 agonist versus (0.83,1.09) per 1000 receptor (from 4 fewer to severe hypoglycaemia 16 17 placebo agonist: 24 2 more) 18 per 1000 19 SGLT-2 OR 0.94 GLP-1 SGLT-2 1 fewer per 1000 Moderate There probably is no difference between SGLT-2 20 inhibitor versus (0.71,1.24) receptor inhibitor: 23 (from 7 fewer to due to inhibitor therapy and GLP-1 receptor agonist 21 GLP-1 receptor agonist: 24 per 1000 5 more) serious therapy on serious hypoglycaemia 22 agonist per 1000 imprecision 23 24 CI= confidence interval. OR = odds ratio. We used the point estimate of absolute effect for GLP-1 receptor agonists, obtained from GLP-1 receptor agonists 25 versus placebo, to calculate absolute effect for SGLT-2 inhibitors v GLP-1 receptor agonists. 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 https://mc.manuscriptcentral.com/bmj 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Page 230 of 244

1 2 3 Table 7 Blindness 4 5 Anticipated absolute effects over 5 years 6 7 Anticipated Certainty in 8 Confidential:absolute For effects Reviewtreatment Only 9 Relative Risk with Risk with (95% CI) over effects 10 Comparison effect control intervention 5 years (GRADE) Plain text summary 11 SGLT-2 OR 0.17 Placebo: 1 SGLT-2 1 fewer per 1000 Low due to SGLT-2 inhibitor therapy may have no effect on 12 inhibitor versus (0.01,4.07) per 1000 inhibitor: 0 (from 1 fewer to very serious blindness. 13 placebo per 1000 4 more) imprecision 14 GLP-1 receptor OR 1.00 Placebo: 1 GLP-1 0 fewer per 1000 Low due to GLP-1 receptor agonist therapy may have no 15 agonist versus (0.23, 4.41) per 1000 receptor (from 1 fewer to very serious effect on blindness. 16 17 placebo agonist: 1 4 more) imprecision 18 per 1000 19 SGLT-2 OR 0.99 GLP-1 SGLT-2 0 fewer per 1000 Low due to There may be no difference between SGLT-2 20 inhibitor versus (0.57,1.73) receptor inhibitor: 1 (from 0 fewer to very serious inhibitor therapy and GLP-1 receptor agonist 21 GLP-1 receptor agonist: 1 per 1000 1 more) imprecision therapy on blindness. 22 agonist per 1000 23 24 CI= confidence interval. OR = odds ratio. We used the point estimate of absolute effect for GLP-1 receptor agonists, obtained from GLP-1 receptor agonists 25 versus placebo, to calculate absolute effect for SGLT-2 inhibitors v GLP-1 receptor agonists. 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 https://mc.manuscriptcentral.com/bmj 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 231 of 244 BMJ

1 2 3 Table 8 Health-related quality of life 4 5 Standardised mean Certainty in 6 difference treatment 7 Comparison (95% CI) effects Plain text summary 8 SGLT-2 0.13Confidential: (-0.11 to 0.36) Low due to SGLT-2 For inhibitors Review may have no effect on health- Only 9 inhibitor versus very serious related quality of life. 10 placebo imprecision 11 GLP-1 receptor 0.12 (0.01 to 0.24) Low due to GLP-1 receptors may increase health-related quality of 12 agonist versus very serious life. 13 placebo imprecision 14 SGLT-2 0.01 (-0.20 to 0.21) Low due to There may be no difference between SGLT-2 inhibitor 15 16 inhibitor versus very serious therapy and GLP-1 receptor agonist therapy on health- 17 GLP-1 receptor imprecision related quality of life. 18 agonist 19 20 CI= confidence interval. OR = odds ratio. We used the point estimate of absolute effect for GLP-1 receptor agonists, obtained from GLP-1 receptor agonists 21 versus placebo, to calculate absolute effect for SGLT-2 inhibitors v GLP-1 receptor agonists. 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 https://mc.manuscriptcentral.com/bmj 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Page 232 of 244

1 2 3 Table 9 Body weight 4 5 Certainty in 6 Mean difference treatment 7 Comparison (95% CI) effects Plain text summary 8 SGLT-2 -1.94Confidential: kg (-2.24 to -1.63) Low due to SGLT-2 For inhibitors Review may lower body weight. Only 9 inhibitor versus very serious 10 placebo inconsistency 11 GLP-1 receptor -1.37 kg (-1.64 to -1.11) Low due to GLP-1 receptor agonists may lower body weight. 12 agonist versus very serious 13 placebo inconsistency 14 SGLT-2 -0.56 kg (-0.94 to -0.18) Moderate SGLT-2 inhibitors probably reduce body weight more 15 16 inhibitor versus due to than GLP-1 receptor agonists. 17 GLP-1 receptor serious 18 agonist inconsistency 19 20 CI= confidence interval. OR = odds ratio. We used the point estimate of absolute effect for GLP-1 receptor agonists, obtained from GLP-1 receptor agonists 21 versus placebo, to calculate absolute effect for SGLT-2 inhibitors v GLP-1 receptor agonists. 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 https://mc.manuscriptcentral.com/bmj 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 233 of 244 BMJ

1 2 3 4 Table 10 Amputation 5 6 Anticipated absolute effects over 5 years 7 8 Confidential: For ReviewCertainty Only 9 Anticipated in 10 absolute effects treatment 11 Relative Baseline Risk with Risk with (95% CI) over 5 effects 12 Comparison effect risk control intervention years (GRADE) Plain text summary 13 SGLT-2 OR 1.11 Very low Placebo: 10 SGLT-2 1 more per 1000 Low due to SGLT-2 inhibitor therapy may have no effect on 14 inhibitor (0.90, per 1000 inhibitor: 11 (from 1 fewer to 4 serious amputation in people with diabetes and few or no 15 versus 1.36) per 1000 more) imprecision cardiovascular risk factors. 16 17 placebo and 18 indirectness 19 Low Placebo: 12 SGLT-2 1 more per 1000 Moderate SGLT-2 inhibitor therapy probably has no effect 20 per 1000 inhibitor: 13 (from 1 fewer to 4 due to on amputation in people with diabetes and 21 per 1000 more) serious cardiovascular risk factors. 22 imprecision 23 Moderate Placebo: 45 SGLT-2 5 more per 1000 Moderate SGLT-2 inhibitor therapy probably has no effect 24 per 1000 inhibitor: 50 (from 5 fewer to due to on amputation in people with diabetes and 25 per 1000 16 more) serious established cardiovascular disease. 26 imprecision 27 High Placebo: 55 SGLT-2 6 more per 1000 Moderate SGLT-2 inhibitor therapy probably has no effect 28 per 1000 inhibitor: 61 (from 6 fewer to due to on amputation in people with diabetes and 29 per 1000 20 more) serious chronic kidney disease 30 imprecision 31 Very high Placebo: 94 Placebo: 104 10 more per 1000 Moderate SGLT-2 inhibitor therapy probably has no effect 32 per 1000 per 1000 (from 9 fewer to due to on amputation in people with diabetes and 33 34 34 more) serious established cardiovascular disease and chronic 35 imprecision kidney disease. 36 37 38 39 40 41 42 43 https://mc.manuscriptcentral.com/bmj 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Page 234 of 244

1 2 3 Table 10 Amputation 4 5 Anticipated absolute effects over 5 years 6 7 Certainty 8 Confidential: ForAnticipated Review in Only 9 absolute effects treatment 10 Relative Baseline Risk with Risk with (95% CI) over 5 effects 11 Comparison effect risk control intervention years (GRADE) Plain text summary 12 GLP-1 OR 0.33 Very low Placebo: 10 GLP-1 7 fewer per 1000 Very low The effect of GLP-1 receptor agonist therapy on 13 receptor (0.01, per 1000 receptor (from 10 fewer to due to very amputation is uncertain in people with diabetes 14 agonist 8.19) agonist: 3 72 more) serious and few or no cardiovascular risk factors. 15 versus per 1000 imprecision 16 placebo and 17 indirectness 18 19 Low Placebo: 12 GLP-1 8 fewer per 1000 Low due to The effect of GLP-1 receptor agonist therapy on 20 per 1000 receptor (from 12 fewer to very serious amputation is uncertain in people with diabetes 21 agonist: 4 86 more) imprecision and cardiovascular risk factors. 22 per 1000 23 Moderate Placebo: 45 GLP-1 30 fewer per 1000 Low due to The effect of GLP-1 receptor agonist therapy on 24 per 1000 receptor (from 45 fewer to very serious amputation is uncertain in people with diabetes 25 agonist: 15 324 more) imprecision and established cardiovascular disease. 26 per 1000 27 High Placebo: 55 GLP-1 37 fewer per 1000 Low due to The effect of GLP-1 receptor agonist therapy on 28 per 1000 receptor (from 54 fewer to very serious amputation is uncertain in people with diabetes 29 agonist: 18 395 more) imprecision and chronic kidney disease. 30 per 1000 31 Very high Placebo: 94 GLP-1 63 fewer (from 93 Low due to The effect of GLP-1 receptor agonist therapy on 32 per 1000 receptor fewer to 676 very serious amputation is uncertain in people with diabetes 33 agonist: 31 more) imprecision and established cardiovascular disease and 34 35 per 1000 chronic kidney disease. 36 37 38 39 40 41 42 43 https://mc.manuscriptcentral.com/bmj 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 235 of 244 BMJ

1 2 3 Table 10 Amputation 4 5 Anticipated absolute effects over 5 years 6 7 Certainty 8 Confidential: ForAnticipated Review in Only 9 absolute effects treatment 10 Relative Baseline Risk with Risk with (95% CI) over 5 effects 11 Comparison effect risk control intervention years (GRADE) Plain text summary 12 SGLT-2 OR 3.32 Very low GLP-1 SGLT-2 7 more per 1000 Very low Whether there is an important difference between 13 inhibitor (0.13, receptor inhibitor: 10 (from 3 fewer to due to very SGLT-2 inhibitor therapy and GLP-1 receptor 14 versus GLP- 82.2) agonist: 3 per per 1000 244 more) serious agonist therapy on amputation is uncertain in 15 1 receptor 1000 imprecision people with diabetes and few or no cardiovascular 16 agonist and risk factors. 17 indirectness 18 19 Low GLP-1 SGLT-2 9 more per 1000 Low due to Whether there is an important difference between 20 receptor inhibitor: 13 (from 13 fewer to very serious SGLT-2 inhibitor therapy and GLP-1 receptor 21 agonist: 4 per per 1000 987 more) imprecision agonist therapy on amputation is uncertain in 22 1000 people with diabetes and cardiovascular risk 23 factors. 24 Moderate GLP-1 SGLT-2 35 more per 1000 Low due to Whether there is an important difference between 25 receptor inhibitor: 50 (from 13 fewer to very serious SGLT-2 inhibitor therapy and GLP-1 receptor 26 agonist: 15 per per 1000 987 more) imprecision agonist therapy on amputation is uncertain in 27 1000 people with diabetes and established 28 cardiovascular disease. 29 High GLP-1 SGLT-2 42 more per 1000 Low due to Whether there is an important difference between 30 receptor inhibitor: 60 (from 16 fewer to very serious SGLT-2 inhibitor therapy and GLP-1 receptor 31 agonist: 18 per per 1000 982 more) imprecision agonist therapy on amputation is uncertain in 32 1000 people with diabetes and chronic kidney disease. 33 Very high GLP-1 SGLT-2 72 more per 1000 Low due to Whether there is an important difference between 34 35 receptor inhibitor: (from 27 fewer to very serious SGLT-2 inhibitor therapy and GLP-1 receptor 36 agonist: 31 per 103 per 969 more) imprecision agonist therapy on amputation is uncertain in 37 1000 1000 people with diabetes and established 38 cardiovascular disease and chronic kidney 39 disease. 40 41 42 43 https://mc.manuscriptcentral.com/bmj 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Page 236 of 244

1 2 3 4 CI= confidence interval. OR = odds ratio. We used the point estimate of absolute effect for GLP-1 receptor agonists, obtained from GLP-1 receptor agonists 5 versus placebo, to calculate absolute effect for SGLT-2 inhibitors v GLP-1 receptor agonists. 6 7 8 Confidential: For Review Only 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 https://mc.manuscriptcentral.com/bmj 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 237 of 244 BMJ

1 2 3 Table 11 Neuropathic pain 4 5 Anticipated absolute effects over 5 years 6 7 Anticipated 8 Confidential:absolute For Certainty Review in Only 9 effects treatment 10 Relative Risk with Risk with (95% CI) effects 11 Comparison effect control intervention over 5 years (GRADE) Plain text summary 12 SGLT-2 NA Placebo: 2 SGLT-2 NA NA Whether SGLT-2 inhibitor therapy has an effect 13 inhibitor versus per 1000 inhibitor: on neuropathic pain is uncertain. 14 placebo NA 15 GLP-1 receptor OR 0.33 Placebo: 2 GLP-1 1 fewer per Very low due Whether GLP-1 receptor agonist therapy has an 16 17 agonist versus (0.01,8.19) per 1000 receptor 1000 (from to very effect on neuropathic pain is uncertain. 18 placebo agonist: 1 2 fewer to serious 19 per 1000 14 more) imprecision 20 and 21 inconsistency 22 between 23 studies 24 SGLT-2 NA GLP-1 SGLT-2 NA NA Whether there is any difference between SGLT-2 25 inhibitor versus receptor inhibitor: inhibitor therapy and GLP-1 receptor agonist 26 GLP-1 receptor agonist: 1 NA therapy on neuropathic pain or not is uncertain. 27 agonist per 1000 28 29 CI= confidence interval. OR = odds ratio. We used the point estimate of absolute effect for GLP-1 receptor agonists, obtained from GLP-1 receptor agonists 30 versus placebo, to calculate absolute effect for SGLT-2 inhibitors v GLP-1 receptor agonists. 31 32 33 34 35 36 37 38 39 40 41 42 43 https://mc.manuscriptcentral.com/bmj 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Page 238 of 244

1 2 3 Table 12 Diabetic ketoacidosis 4 5 Anticipated absolute effects over 5 6 years 7 8 Confidential:Anticipated For Review Only 9 absolute Certainty 10 effects in 11 (95% CI) treatment 12 Relative Risk with Risk with over 5 effects 13 Comparison effect control intervention years (GRADE) Plain text summary 14 SGLT-2 OR 1.92 Placebo: 2 SGLT-2 2 more per High SGLT-2 inhibitor therapy increases diabetic 15 inhibitor versus (1.08, 3.39) per 1000 inhibitor: 4 1000 (from ketoacidosis. 16 17 placebo per 1000 0 to 12 18 more) 19 GLP-1 receptor OR 1.21 Placebo: 2 GLP-1 0 more per Moderate GLP-1 receptor agonist therapy probably has no 20 agonist versus (0.63, 2.33) per 1000 receptor 1000 (from due to effect on diabetic ketoacidosis. 21 placebo agonist: 2 1 fewer to 3 serious 22 per 1000 more) imprecision 23 SGLT-2 OR 1.59 GLP-1 SGLT-2 1 more per Moderate There probably is no difference between SGLT-2 24 inhibitor versus (0.68, 3.72) receptor inhibitor: 3 1000 (from due to inhibitor therapy and GLP-1 receptor agonist 25 GLP-1 receptor agonist: 2 per 1000 1 fewer to 5 serious therapy. 26 agonist per 1000 more) imprecision 27 28 CI= confidence interval. OR = odds ratio. We used the point estimate of absolute effect for GLP-1 receptor agonists, obtained from GLP-1 receptor agonists 29 versus placebo, to calculate absolute effect for SGLT-2 inhibitors v GLP-1 receptor agonists. 30 31 32 33 34 35 36 37 38 39 40 41 42 43 https://mc.manuscriptcentral.com/bmj 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 239 of 244 BMJ

1 2 3 Table 13 Serious hyperglycaemia 4 5 Anticipated absolute effects over 5 6 years 7 8 Confidential:Anticipated For Review Only 9 absolute Certainty 10 effects in 11 (95% CI) treatment 12 Relative Risk with Risk with over 5 effects 13 Comparison effect control intervention years (GRADE) Plain text summary 14 SGLT-2 OR 0.59 Placebo: 25 SGLT-2 10 fewer per High SGLT-2 inhibitor therapy decreases serious 15 inhibitor versus (0.35, 1.01) per 1000 inhibitor: 15 1000 (from hyperglycaemia. 16 17 placebo per 1000 0 to 16 18 fewer) 19 GLP-1 receptor OR 0.40 Placebo: 25 GLP-1 15 fewer per High GLP-1 receptor agonist therapy decreases serious 20 agonist versus (0.25, 0.64) per 1000 receptor 1000 (from hyperglycaemia. 21 placebo agonist: 10 9 to 19 22 per 1000 fewer) 23 SGLT-2 OR 1.48 GLP-1 SGLT-2 5 more per Moderate There probably is no difference in serious 24 inhibitor versus (0.75, 2.93) receptor inhibitor: 15 1000 (from due to hyperglycaemia between SGLT-2 inhibitor 25 GLP-1 receptor agonist: 10 per 1000 2 fewer to serious therapy and GLP-1 receptor agonist therapy. 26 agonist per 1000 19 more) imprecision 27 28 CI= confidence interval. OR = odds ratio. We used the point estimate of absolute effect for GLP-1 receptor agonists, obtained from GLP-1 receptor agonists 29 versus placebo, to calculate absolute effect for SGLT-2 inhibitors v GLP-1 receptor agonists. 30 31 32 33 34 35 36 37 38 39 40 41 42 43 https://mc.manuscriptcentral.com/bmj 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Page 240 of 244

1 2 3 Table 14 Genital infection 4 5 Anticipated absolute effects over 5 years 6 7 Anticipated 8 Confidential:absolute For Certainty Review Only 9 effects in 10 (95% CI) treatment 11 Relative Risk with Risk with over 5 effects 12 Comparison effect control intervention years (GRADE) Plain text summary 13 SGLT-2 OR 3.42 Placebo: 73 SGLT-2 177 more High SGLT-2 inhibitor therapy increases genital 14 inhibitor versus (2.79, 4.18) per 1000 inhibitor: per 1000 infection. 15 placebo 249 per (from 131 16 17 1000 more to 232 18 more) 19 GLP-1 receptor OR 0.69 Placebo: 73 GLP-1 23 fewer per Moderate GLP-1 receptor agonist therapy probably does not 20 agonist versus (0.33, 1.45) per 1000 receptor 1000 (from due to increase genital infection. 21 placebo agonist: 50 49 fewer to serious 22 per 1000 33 more) imprecision 23 SGLT-2 OR 4.96 GLP-1 SGLT-2 198 more High SGLT-2 inhibitors increase genital infection 24 inhibitor versus (2.39, 10.3) receptor inhibitor: per 1000 compared with GLP-1 receptor agonists. 25 GLP-1 receptor agonist: 50 248 per (from 70 to 26 agonist per 1000 1000 465 more) 27 28 CI= confidence interval. We used the point estimate of absolute effect for GLP-1 receptor agonists, obtained from GLP-1 receptor agonists versus placebo, to 29 calculate absolute effect for SGLT-2 inhibitors v GLP-1 receptor agonists. 30 31 32 33 34 35 36 37 38 39 40 41 42 43 https://mc.manuscriptcentral.com/bmj 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 241 of 244 BMJ

1 2 3 Table 15 Fournier gangrene 4 5 Anticipated absolute effects over 5 years 6 7 Anticipated 8 Confidential:absolute For Certainty Review in Only 9 effects treatment 10 Relative Risk with Risk with (95% CI) effects 11 Comparison effect control intervention over 5 years (GRADE) Plain text summary 12 SGLT-2 OR 0.56 Placebo: 1 SGLT-2 0 fewer per Low due to Whether or not SGLT-2 inhibitor therapy has an 13 inhibitors versus (0.16, 19.2) per 1000 inhibitor: 1 1000 (1 very serious effect on Fournier gangrene is uncertain. 14 placebo per 1000 fewer to 18 imprecision 15 more) 16 17 GLP-1 receptor NA Placebo: 1 GLP-1 NA NA Whether or not GLP-1 receptor agonist therapy 18 agonist versus per 1000 receptor has an effect on Fournier gangrene is uncertain. 19 placebo agonist: NA 20 SGLT-2 NA GLP-1 SGLT-2 NA NA Whether SGLT-2 inhibitor or GLP-1 receptor 21 inhibitor versus receptor inhibitor: agonist therapy have different effects on Fournier 22 GLP-1 receptor agonist: NA NA gangrene is uncertain. 23 agonist 24 25 CI= confidence interval. NA = not available. OR = odds ratio. We used the point estimate of absolute effect for GLP-1 receptor agonists, obtained from GLP- 26 1 receptor agonists versus placebo, to calculate absolute effect for SGLT-2 inhibitors v GLP-1 receptor agonists. 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 https://mc.manuscriptcentral.com/bmj 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Page 242 of 244

1 2 3 Table 16 Severe gastrointestinal events 4 5 Anticipated absolute effects over 5 6 years 7 8 Confidential:Anticipated For Review Only 9 absolute 10 effects Certainty in 11 (95% CI) treatment 12 Relative Risk with Risk with over 5 effects 13 Comparison effect control intervention years (GRADE) Plain text summary 14 SGLT-2 NA Placebo: 44 SGLT-2 NA NA Whether or not SGLT-2 inhibitor therapy has an 15 inhibitor versus per 1000 inhibitor: effect on severe gastrointestinal events is 16 17 placebo NA uncertain. 18 GLP-1 receptor OR 2.33 Placebo: 44 GLP-1 58 more per Low due to GLP-1 receptor agonists may increase severe 19 agonist versus (1.21, 4.48) receptor 1000 (from serious gastrointestinal events. 20 placebo agonist: 102 9 to 153 imprecision 21 per 1000 more) and 22 inconsistency 23 between 24 studies 25 SGLT-2 NA GLP-1 SGLT-2 NA NA Whether there is a difference between SGLT-2 26 inhibitor versus receptor inhibitor: inhibitor therapy and GLP-1 receptor agonist 27 GLP-1 receptor agonist: NA NA therapy on serious gastrointestinal events is 28 agonist uncertain. 29 30 CI= confidence interval. OR = odds ratio. We used the point estimate of absolute effect for GLP-1 receptor agonists, obtained from GLP-1 receptor agonists 31 versus placebo, to calculate absolute effect for SGLT-2 inhibitors v GLP-1 receptor agonists. 32 33 34 35 36 37 38 39 40 41 42 43 https://mc.manuscriptcentral.com/bmj 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 243 of 244 BMJ

1 2 3 Table 17 Pancreatic cancer 4 5 Anticipated absolute effects over 5 6 years 7 8 Confidential:Anticipated For Review Only 9 absolute Certainty 10 effects in 11 (95% CI) treatment 12 Relative Risk with Risk with over 5 effects 13 Comparison effect control intervention years (GRADE) Plain text summary 14 SGLT-2 OR 1.77 Placebo: 4 SGLT-2 3 more per Low due to Whether or not SGLT-2 inhibitor therapy has an 15 inhibitor versus (0.55, 5.75) per 1000 inhibitor: 7 1000 (2 very serious effect on pancreatic cancer is uncertain. 16 17 placebo per 1000 fewer to 19 imprecision 18 more 19 GLP-1 receptor OR 1.19 Placebo: 4 GLP-1 1 more per Moderate GLP-1 receptor therapy probably has little or no 20 agonist versus (0.78, 1.81) per 1000 receptor 1000 (1 due to effect on pancreatic cancer 21 placebo agonist: 5 fewer to 3 serious 22 per 1000 more) imprecision 23 SGLT-2 OR 1.49 GLP-1 SGLT-2 2 more per Low due to Whether there is a difference between SGLT-2 24 inhibitor versus (0.43, 5.19) receptor inhibitor: 7 1000 (3 very serious inhibitor therapy and GLP-1 receptor agonist 25 GLP-1 receptor agonist: 5 per 1000 fewer to 21 imprecision therapy on pancreatic cancer is uncertain. 26 agonist per 1000 more) 27 28 CI= confidence interval. OR = odds ratio. We used the point estimate of absolute effect for GLP-1 receptor agonists, obtained from GLP-1 receptor agonists 29 versus placebo, to calculate absolute effect for SGLT-2 inhibitors v GLP-1 receptor agonists. 30 31 32 33 34 35 36 37 38 39 40 41 42 43 https://mc.manuscriptcentral.com/bmj 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Page 244 of 244

1 2 3 Table 18 Pancreatitis 4 5 Anticipated absolute effects over 5 6 years 7 Anticipated 8 Confidential:absolute For Certainty Review Only 9 effects in 10 (95% CI) treatment 11 Relative Risk with Risk with over 5 effects 12 Comparison effect control intervention years (GRADE) Plain text summary 13 SGLT-2 OR 0.61 Placebo: 5 SGLT-2 2 fewer per Moderate SGLT-2 inhibitor therapy probably has no effect 14 inhibitor versus (0.30, 1.27) per 1000 inhibitor: 3 1000 (4 due to on pancreatitis. 15 16 placebo per 1000 fewer to 1 serious 17 more) imprecision 18 GLP-1 receptor OR 1.17 Placebo: 5 GLP-1 1 more per Moderate GLP-1 receptor agonist therapy probably has no 19 agonist versus (0.89, 1.55) per 1000 receptor 1000 (1 due to effect on pancreatitis. 20 placebo agonist: 6 fewer to 3 serious 21 per 1000 more) imprecision 22 SGLT-2 OR 0.52 GLP-1 SGLT-2 3 fewer per Moderate There is probably no difference in pancreatitis 23 inhibitor versus (0.25, 1.12) receptor inhibitor: 3 1000 (4 due to between SGLT-2 inhibitor therapy and GLP-1 24 GLP-1 receptor agonist: 6 per 1000 fewer to 1 serious receptor agonist therapy. 25 agonist per 1000 more) imprecision 26 27 CI= confidence interval. OR = odds ratio. We used the point estimate of absolute effect for GLP-1 receptor agonists, obtained from GLP-1 receptor agonists 28 versus placebo, to calculate absolute effect for SGLT-2 inhibitors v GLP-1 receptor agonists. 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 https://mc.manuscriptcentral.com/bmj 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 245 of 244 BMJ

1 2 3 Table 19 Glycated haemoglobin A1C 4 5 Certainty in 6 Mean difference treatment 7 Comparison (95% CI) effects Plain text summary 8 SGLT-2 -0.61%Confidential: (-0.68 to -0.54) Low due to SGLT-2 For inhibitor Review therapy may lower glycated Only 9 inhibitor versus serious haemoglobin A1C. 10 placebo imprecision 11 and 12 inconsistency 13 GLP-1 receptor -0.90% (-0.96 to -0.83) Low due to GLP-1 receptor therapy may lower glycated 14 agonist versus serious haemoglobin A1C. 15 16 placebo imprecision 17 and 18 inconsistency 19 SGLT-2 0.28% (0.19 to 0.38) High GLP-1 receptor agonist therapy lowers glycated 20 inhibitor versus haemoglobin A1C more than SLGT-2 inhibitor 21 GLP-1 receptor therapy. 22 agonist 23 24 CI= confidence interval. We used the point estimate of absolute effect for GLP-1 receptor agonists, obtained from GLP-1 receptor agonists versus placebo, to 25 calculate absolute effect for SGLT-2 inhibitors v GLP-1 receptor agonists. 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 https://mc.manuscriptcentral.com/bmj 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60