The Relationship between Ischemic Heart Disease and Diabetes - A population-based approach

Transcription

1 Faculty of Health Sciences, University of Copenhagen Ph.D. thesis The Relationship between Ischemic Heart Disease and Diabetes - A population-based approach Gentofte University Hospital, Department of Cardiology

2 Ph.D. thesis The Relationship between Ischemic Heart Disease and Diabetes - A population-based approach Faculty of Health Sciences, University of Copenhagen Gentofte University Hospital, Department of Cardiology

3 Academic advisors Professor Professor Christian Torp-Pedersen, M.D, DMSc. Department of Cardiology, Gentofte University Hospital, Denmark Associate Professor Gunnar H. Gislason, M.D, Ph.D. Department of Cardiology, Gentofte University Hospital, Denmark Members of the assessment committee Professor and Chief Physician Jan Skov-Jensen Department of Cardiology, Gentofte University Hospital (Chairman) Chief Physician Peter Rossing, Steno Diabetes Center, Gentofte Professor Nick Finer, University College London Hospitals, Centre for Weight Loss, Metabolic and Endocrine Surgery The Faculty of Health Sciences, Copenhagen University, Denmark has approved this Ph.D. dissertation for public defense. The public lecture and defense will take place February 10th, 2012 at in the Auditorium Gentofte University Hospital.

4 Preface My interest in cardiology evolved during my medical studies at Copenhagen University. Initially, I actually wanted to become a specialist in Ophthalmology. However, I found myself increasingly fascinated by Cardiology and the eye did not catch my interest as much as the heart. During my studies, I started working at The Heart Centre in Rigshospitalet as an assistant in Cardiology, a job that only made my interest in Cardiology even greater. After completion of my internship in clinical medicine, I commenced my first year of clinical specialty in Cardiology at Bispebjerg Hospital. Previously, At Bispebjerg Hospital, I had the great fortune to meet Professor Christian Torp-Pedersen, who introduced me to research by giving me an assignment which I by the way never managed to complete. Nevertheless, Christian offered me a position as a research assistant. At the time, Christian s research-unit had moved from Bispebjerg to Gentofte Hospital. In September 2008 I started working on my PhD-thesis. Shortly thereafter, I had the great privilege to meet Dr. Gunnar H. Gislason Dr. G, who became my mentor on the project. I also had the great fortune being introduced to Professors Lars Køber, Rigshospitalet, Allan Vaag, Steno Diabetes Center and Dr. Peter Riis Hansen, Gentofte hospital who soon became invaluable co-advisers on the Ph.D. thesis. Part of being associated with Christian Torp-Pedersen s research group includes collaboration with the DANTRIP (Danish Trends in Pharmacological and Invasive Treatment of Cardiovascular Diseases) investigators; A collective of health researchers consisting of epidemiologists, clinicians and statisticians from Gentofte Hospital, The National Institute of Public Health (NIPH), Køge Hospital, and Rigshospitalet. Through monthly meetings where the younger researchers were given the opportunity to present their projects in differing stages of development with senior researchers imparting, valuable feedback as well as giving lectures on relevant topics provided an inspirational educational environment that has greatly influenced the research work. The DANTRIP collaboration has undoubtedly played a large role furthering the publication of many papers in leading peer reviewed medical journals by the participating DANTRIP researchers. I feel privileged to be a part of this group. The first half a year of researching was quite difficult and frustrating, since both my computer skills and ability to do SAS programming were apparently lacking. I would like to thank my friend and colleague Dr. Søren Skøtt Andersen, who spent unreasonable amounts of time programming with me on the data for my first article and who listened patiently as ever on my daily outbursts of frustration. You are truly the best friend ever. I am forever grateful. I also received invaluable assistance and guidance from my talented fellow research assistant colleague Dr. Fredrik Folke with whom I had the great privilege of sharing office. Thank you for putting up with my talkative personality! Unfortunately, the evident worldwide financial crisis also had an impact on my research project, making the financing of the project problematic. That the project was able reach its conclusion in a satisfactory manner within the available time frame is in large part due to the diligent encouragement and support provided by my academic advisors Christian Torp-Pedersen and Gunnar Gislason to whom I owe a great thanks. I am also very grateful for the economic support provided by the FUKAP foundation. Finally, I would like to thank Head of Department Jan Kyst Madsen, who is doing an excellent job in leading the research department along with my academic advisors. During my tenure as aspiring researcher, I have also been graced with great colleagues, the fellow research assistants who never failed to provide key insights and keeping up the spirit when the data did not compute. Last but certainly not least, I would like to thank the Harvard Study Group at the Fjernarkiv for their fine companionship, support and friendship. Furthermore, I am forever indebted to my excellent colleague and dear friend Dr. Charlotte Andersson, without whom this project would not have been possible within the time frame available. Finally, I would like to thank my loving daughter Filippa as well as Johnny, my family, and my best friend Dr. Mette Rauhe Mouridsen, who has believed in me and supported me through the good times as well as the rough times. July 2011 The Relationship between Ischemic Heart Disease and Diabetes A population-based approach 3

5 Table of Contents Preface 3 Table of Contents 4 List of Papers 4 Introduction 5 The role of diabetes in cardiovascular disease. 5 Diabetes and Heart Failure 6 Hypotheses 7 Objectives 7 Methods 8 Databases 8 Study Population 8 Pharmacotherapy 9 Co-morbidity 9 Outcomes 10 Statistical analyses 11 Ethics 11 Results overview of papers 12 Paper I: 12 Paper II: 14 Paper III: 17 Methodological considerations 19 Study design 19 Poisson regression vs. Cox regression 19 Selection bias, protopathic bias, and confounding by indication 19 Discussion 20 Diabetes and cardiovascular risk 20 Temporal trends in type of first initiated GLM 21 MI and Cardiovascular risk 21 The development of diabetes in MI patients 21 Post-MI development of HF and risk of diabetes. 22 Development of diabetes in heart failure patients and the risk of death. 23 Study strengths 24 Study weaknesses 24 Novelty of the results 25 Conclusions/Implications 25 Future research 26 Summary 27 Dansk Resumé 28 References 29 Appendices 33 Appendix 1 33 Appendix 2 34 Appendix 3 35 Appendix 4 36 List of Papers The present thesis is based on the following publications: I. Norgaard ML; Andersen SS ; Schramm TK; Folke F; Jørgensen CH; Hansen ML; Andersson C; Bretler DM; Vaag A Køber L; Torp- Pedersen C; Gislason GH; Changes in short- and long-term cardiovascular risk of incident diabetes and incident myocardial infarction - a nationwide study. Diabetologia Aug; 53(8): Epub 2010 May 9. II. Norgaard ML; Andersson C; Andersen SS; Vaag A; Schramm TK; Folke F; Køber L; Riis Hansen P; Torp-Pedersen C; Gislason GH; Temporal trends in the initiation of glucose-lowering medications after first-time myocardial infarction a nationwide study between 1997 and CardiovascDiabetol Jan 19;10:5. III. Andersson C; Norgaard ML; Riis Hansen P; Fosbøl EL; Schmiegelow M; Weeke P; Bjerring Olesen J; Raunsø J; Jørgensen CH; Vaag A; Køber L; Torp-Pedersen C; Gislason GH; Heart failure severity, as determined by loop diuretic dosages, predicts the risk of developing diabetes after myocardial infarction: a nationwide cohort study. Eur J Heart Fail Dec; 12(12): Epub 2010 Sep 23. 4

6 Introduction A possible causal relationship between diabetes mellitus and the risk of developing cardiovascular disease was already suggested in 1922 (a published report on glucosuria in patients with myocardial infarction (MI)). 1 The wealth of knowledge regarding the association of diabetes and the incidence of cardiovascular disease have increased massively during the last couple of decades firmly establishing diabetes as an important riskfactor as well as prognostic factor for cardiovascular disease. 2-5 Furthermore, it is now wellestablished that diabetes is associated with impaired prognosis 2, 3, 6-9 after MI. The role of diabetes in cardiovascular disease Diabetes is a metabolic disorder of multiple etiologies characterized by chronic hyperglycaemia with disturbances of carbohydrate, fat, and protein metabolism resulting from defects of insulin secretion, insulin action, or a combination of both. 10 It is a chronic disease that causes serious health complications including renal failure, heart disease, stroke and blindness. Type 2 diabetes is caused by a combination of decreased insulin secretion and decreased insulin sensitivity. Typically, the early stage of type 2 diabetes is characterized by insulin resistance and a decreased ability for insulin secretion that causes excessive post-prandial hyperglycaemia. In the early stage of type 2 diabetes the fasting glucose levels are usually normal or in near-normal range (impaired glucose tolerance (IGT)). It is not within the scope of the present thesis to account for the pathophysiology in depth, but in brief the abnormal metabolic state, including chronic hyperglycaemia, dyslipidemia, and insulin resistance that accompanies diabetes, causes arterial dysfunction and makes the arteries susceptible to atherosclerosis. Diabetes alters function of multiple cell types, including platelets, smooth muscle cells, and endothelium, indicating the extent of vascular disorder in this disease, see Figure 1 for an overview. Figure 1 Diabetes Mellitus Hyperglycemia Excess Free Fatty Acids Insulin Resistance Oxidative Stress Protein Kinase C Activation Receptor for Advanced Glycation End Product (RAGE) Activation ENDOTHELIUM Nitric Oxide Nitric Oxide Activation of NF-кB Nitric Oxide Endothelin-1 Angiotensin II Tissue Factor Angiotensin II Activation of Plasminogen Activator Activator Protein-1 Inhibitor-1 Prostacyclin Vasoconstriction Hypertension Vascular Smooth Muscle Cell Growth Inflammation Release of Chemokines Release of Cytokines Expression of Cellular Adhesion Molecules Thrombosis Hypercoagulation Platelet Activation Decreased Fibrinolysis Atherogenesis Thus the perception of type 2 diabetes encompasses a global metabolic disorder characterized by hyperlipidaemia, hypertension, and hypercoagulability in addition to hyperglycaemia. Each of these abnormalities plays an important role in cardiovascular disease development and progression and provides targets for therapy. Because the history of diagnostic criteria in diabetes is of interest for the present thesis, the pathophysiology and diagnostic criteria will be briefly covered in appendix 1. Of particular interest, The Relationship between Ischemic Heart Disease and Diabetes A population-based approach 5

7 the criterion on diabetes / pre-diabetes was not changed in Denmark during the study period of the present thesis ( ), 11 opposite to what happened in the US in 2003, where the cut-off point of impaired fasting glucose was further lowered to 5.6 mmol/l by the American Diabetes Association (ADA). 12 Diabetes is recognized as an important risk factor for the development of cardiovascular disease and a strong predictor of an adverse outcome after myocardial infarction. 3, 13 Patients requiring Glucose lowering medications (GLM) have the same long-term risk of major cardiovascular outcomes as patients with a previous MI. 4, 5 This increase in risk in diabetes patients is independent of age, gender and type of diabetes. 4, 5 A decline in all-cause- and cardiovascular mortality has occurred during the last decades among both diabetes patients and individuals without diabetes, although this decline appears markedly lower among diabetes patients than in non-diabetic patients. 14, 15 Explanations for this decline involve e.g. the effects of cardiovascular disease prevention by tight blood pressure control, aspirin, statins and blockers of the reninangiotensin system, which are well-established in patients at high cardiovascular risk. 5, Similar, intensive glycaemic control on microvascular complications in patients with diabetes is today wellestablished Whereas previous studies have indicated a high risk of cardiovascular adverse outcomes early in diabetes course, 26, 27 at least one previous cohort study of MI patients demonstrated that diabetes had no independent influence on mortality immediately following an acute MI, but increased in relative risk over time, conferring one of the strongest risk factors for mortality in MI patients after a couple of years. 28 This could mean that diabetes is an accelerating disease and that primary prophylactic intervention with cardiovascular pharmacotherapy may first be beneficial some time after established diabetes. An increase in relative risk over time associated with diabetes may however not only reflect a progression of diabetes disease, but may partly or mainly reflect a decline in mortality risk in MI risk over time. Thus, for answering such a question, the relative risk of diabetes and MI needs to be compared with the background population at well-defined times. To the best of my knowledge, no previous study has done this. Furthermore, to date, studies comparing the risk in patients with MI and diabetes, respectively, have primarily focused on prevalent diagnoses and not on the risk following incident MI or incident diabetes, which, given the progressive nature of diabetes, may differ from a prevalent disease. Intensive glycaemic control on microvascular complications in patients with diabetes is today well-established During the last couple of decades several novel treatment strategies have been implemented and new GLM have been introduced. Although the initial results from the UK Prospective Diabetes Study (UKPDS) did not conclusively document any beneficial effects of intensified glycaemic control on prevention of macrovascular complications, the epidemiological follow-up data from the UKPDS indicated that long term benefits on cardiovascular risk do occur. 29 Also, the Steno-2-Study showed that intensive multi-factorial treatment including blood pressure control, glucose and lipid lowering therapies along with lifestyle intervention prevents (or postpones) the development of micro- and macrovascular complications in type 2 diabetes patients. 30, 31 This increasing amount of knowledge is assumed to have influenced diabetes management and prognosis over years, but studies on this subject are sparse. Therefore, in the present thesis, temporal changes in prognosis associated with diabetes and MI between 1997 and 2006 was investigated (paper I) as was the trends in use of glucose-lowering agents following MI between 1997 and 2006 (paper II). The prevalence of diabetes in MI patients has been reported with a prevalence as high as 20 % 13 and several studies have shown that smaller degrees of dysglycaemia is common in patients with MI, who have previously not been diagnosed with diabetes. In fact, previous studies have documented that abnormal glucose metabolism is more common than normal glucose metabolism among patients with MI. 32, 33 Among patients with MI without known diabetes up to 65% exhibit abnormal glucose regulation when studied using an oral glucose tolerance test (OGTT) 34 of whom 25% have values diagnostic for diabetes and 40 % have IGT. 32, 33, 35, 36 However, information on the actual number of patients receiving GLM (as a proxy for the development of diabetes) after first-time MI on a population basis is lacking, both prior to and after implementation of the novel guidelines (using OGTT as gold standard in revealing diabetes in first-time MI patients). Diabetes and Heart Failure MI is characterized by myocyte necrosis, loss of contractile function and impaired myocardial relaxation. Thus, functional myocytes may be replaced by fibrotic tissue affecting both the left ventricular systolic and diastolic function leading to various degrees of heart failure (HF). HF is common after MI and the incidence of HF 30 days post-mi was recently reported to be greater than 23% in a community-based sample. 37 Diabetes has at least for three decades been established as an independent risk factor for development of HF. 38 More recently, a linear relationship between fasting plasma glucose and the risk of hospitalization for HF has also been demonstrated among patients without previous MI, with or without overt diabetes. 39 Opposite the well-established relationship of diabetes increasing the risk of HF, the prevalence of diabetes appears to increase with worsening of New York Heart Association (NYHA) functional class, 40 indicating that HF may increase the risk of diabetes. Furthermore, it was recently reported that HF patients without diabetes have increased insulin resistance with worsening of NYHA class. 41 Present thesis (paper III) aims to enlighten whether patients developing HF after MI are at particular high risk of developing diabetes. This is at present time not well-known and remains to be established on a large-scale population. In summary, although much is already known about the influence of diabetes on cardiovascular disease, information on changes in the cardiovascular risk over time in MI and diabetes is insufficiently investigated. Further, it is mainly unknown whether increased focus on diabetes in cardiovascular disease has increased diabetes-treatment in MI populations and whether the prognosis associated with diabetes has changed in recent years. Finally, recent evidence has suggested that heart failure increases the risk of developing diabetes, revealing a potential new population that could be target for early diabetes prevention and management but this is also still insufficiently investigated. 6

8 Hypotheses Objectives The present thesis aimed to investigate the following hypotheses: 1. Patients with a first-time MI have a high short-term risk of cardiovascular adverse outcomes compared with the general population. The risk may decline rapidly after some time and then stabilize. 2. Patients initiating GLM (as a proxy for incident diabetes) exhibit an approximately constant and elevated short- and long term risk of cardiovascular adverse outcomes compared with the general population. 3. For both patient groups (first-time MI or incident diabetes) a decrease in risk may be expected when comparing the early with the late study period. The present thesis had the following objectives on the subject: The relationship between ischemic heart disease and diabetes. 1. To assess secular trends of cardiovascular outcomes following first-time MI or following incident diabetes compared with the general population. 2. To investigate the temporal trends in initiation of GLM (as a proxy for diabetes) among patients discharged after first-time MI and among the general population. 3. To assess whether HF severity predicts the risk of developing diabetes after MI. 4. The incidence rates of GLM initiation in patients with first-time MI as well as in the general population is increasing during the 10 year study period ( ). 5. Development of HF after first-time MI increases the risk of development of diabetes requiring GLM. The Relationship between Ischemic Heart Disease and Diabetes A population-based approach 7

9 Methods Databases In Denmark there are unique opportunities for epidemiological register-based research because there is a long tradition for thorough registration in the national healthcare system. Nationwide administrative registers can be cross-linked on the individual level through access on Statistics Denmark where all information is held. This is possible in Denmark since all residents have a unique and permanent individual personal registration number. In this thesis following Danish registers were employed: The Danish National Patient Register keeps records of all patients admitted to Danish hospitals since At discharge, one primary, and if applicable one or more secondary diagnoses are registered according to the International classification of Diseases (ICD) (ICD-8) the 8 th revision until 1994 and the 10 th revision (ICD-10) from The Danish Register of Medicinal Product Statistics (national prescription register) contains all prescriptions dispensed from pharmacies since Each medication is coded according to an international classification of pharmaceuticals, the Anatomical Therapeutical Chemical classification (ATC) System. The prescription register also includes information about the date of dispensing, the strength, formulation and quantity dispensed, and the affiliation of the doctor issuing the prescription (hospital-based doctor or general practitioner). The database has previously been described in detail and validated. 42 The Danish Cause of Death Register contains data concerning immediate, contributory and underlying causes of death classified using the ICD-10. The information held in this register is based on information obtained from the death certificate filled out by a doctor. The Central Population Register contains information on vital status (dead or alive) and the date of death. All deaths are recorded in this register within two weeks of their occurrence. Study Population Overall this thesis studies individuals 30 years of age and alive on January 1, 1997 who are hospitalized for a first-time MI during the study period (January 1, 1997 until December 31, 2006). In paper I, we furthermore study individuals initiating GLM for the first time during the study period. Paper II includes individuals who develop diabetes requiring GLM after first-time MI and finally, in paper III MI patients who develop heart failure are studied. Since the populations in the papers in this thesis are somewhat different, the populations will be described in detail separately. Paper I The general population: All inhabitants in Denmark who were at least 30 years of age and alive on January 1, 1997 were identified. Patients with prior history of myocardial infarction (MI) or patients claiming glucose-lowering medication (GLM) prior to January 1, 1997 were not included in the study. The incident MI-population: Incident MI was identified as firsttime hospitalization with MI (ICD-10 code I21 or I22). First-time diagnosis of MI was considered, if the diagnosis of MI was not registered during the last 19 years prior to the study period. The incident diabetes population: Patients with incident diabetes requiring anti diabetic treatment were identified as individuals initiating GLM (oral or insulin) (ATC code A10) for the first time during the study period (January 1, 1997-December 31, 2006). Patients claiming prescriptions on GLM during the period were not included. Depending on which diagnosis came first, patients were registered as either having incident MI or incident diabetes. Paper II The general population: Comprised all individuals, aged 30 years and alive on January 1, The population was updated on a yearly basis allowing entrance of new subjects aged 30 years of age per January 1, the following year. Patients with previous MI, incident MI as well as patients claiming GLM prescriptions (ATC code A10) prior to January 1, 1997 were excluded from the general population. The MI-population: Comprised all patients, aged 30 years alive at discharged after a first-time MI (ICD-10 code I21 or I22) between 1997 and 2006 (January 1, 1997-December 31, 2006). First-time admission for MI implied that the National Patient Register had not registered any prior admission for MI in the previous 19 years. Patients initiating GLM (ATC code A10) during the study period were assessed. Patients previously claiming any prescriptions on GLM (during the period ) were not included. 8

10 Paper III The MI-population: All patients hospitalized for MI (ICD-10 code I21 or I22) between 1997 and 2006 and who were alive 90 days after discharge were included. Patients with a prior MI in were excluded (ICD-10 code I21 or I22, or ICD-8 code 410) to ensure a homogenous population. Patients claiming prescriptions of any GLM (ATC code A10) prior to, or during the first 90 days after discharge were also excluded, as were patients who had claimed a prescription of loop diuretics (ATC code C03C) prior to hospitalization. Development of diabetes: Was defined as the first claimed prescription for GLM (ATC code A10). The diagnosis of MI in the National Patient Register has proved to be valid; with a sensitivity of 91 % and a positive predictive value of 93 %. 43 Pharmacotherapy From the Danish Register of Medicinal Product Statistics medical treatment was defined as followed: Glucose-lowering medications Paper I & II: Oral or insulin treatment commencing during the study period (ATC-code A10). Applying for Paper II only: For determination of first initiated GLM, the following agents (ATC-codes) were identified: insulin (A10A), metformin (A10BA02), sulfonylureas (A10BB), glitazones (A10B); acarbose (A10F01), repaglinide (A10BX0) and metformin+repaglinide (A10BD03).No concomitant treatment was assessed. Paper III: Treatment initiation of oral or insulin agents (ATC-code A10). Loop diuretics Paper III: As a proxy for the diagnosis heart failure in patients with MI, we used dispensed prescription on loop diuretics (ATC C03C) used at day 90 after discharge. The approximated dosages were achieved by consideration of the amount of dispensed tablets, their strengths, and dispensing time intervals (for up to 7 consecutively claimed prescriptions), as done before 44, 45 HF severity was defined according to arbitrary loop diuretic dosages: no HF (no loop diuretics), mild HF (loop diuretic dosage 40 mg/day), moderate HF (> mg/day), and severe HF (>120 mg/day). This way of defining HF severity was supported by previous work, which demonstrated loop diuretic dosages to be positively correlated with worsened NYHA functional class and increased risk of mortality, and not correlated with glomerular filtration rate in HF patients. 46 Moreover, as patients with use of loop diuretics prior to myocardial infarction hospitalization were excluded from the present study cohort, the indication for loop diuretic treatment was in most cases likely to be HF. The sensitivity of using loop diuretics as a proxy for HF was considered acceptable, as the majority of HF 47, 48 patients previously have been reported to use loop diuretics. Concomitant treatment in paper III: Patients were considered to be in treatment with a specific agent if they claimed a prescription during the first 90 days following discharge, or in the year preceding MI hospitalization. The following ATC codes were used: beta-blockers (C07); statins (C10A); renin angiotensin system inhibitors (RASi) (C09); thiazides (C03A); aldosterone antagonists (spironolactone/eplerenone) (C03D); calcium channel blockers (C08); digoxin (C01AA05); vitamin K antagonists (B01AA0); aspirin (B01AC06); and clopidogrel (B01AC04). Co-morbidity Paper I: Incident MI-population and incident diabetes population: Comorbidities were assessed by registration of pre-specified hospital discharge diagnoses up to one year before the inclusion date (date of the incident). The general population: Co-morbidity was assessed one year prior to June 30, 2001 for individuals surviving beyond June 30, 2001, and a year prior to January 1, 1997 for individuals who died before June 30, 2001, respectively. Applying for both the incident MI, incident diabetes and the general population following diagnoses were used: Chronic obstructive pulmonary disease (ICD-10 code J42 and J44 ), renal disease (ICD-10 code N03, N04, N17 N18, N19, R34 I12, I13 ) cancer C00 - C97 and congestive heart failure I50. Paper II and III: Comorbidity was assessed using the Ontario Myocardial Infarction Mortality Prediction Rules for ICD-10 codes based on 9 predefined diagnoses from the index hospitalization with MI and up to one year prior to discharge from index hospitalization (diabetes diagnoses excluded in paper II and III and heart failure excluded in paper III).The Ontario Myocardial Infarction Mortality Prediction Rules are validated for predicting prognosis for post MI populations. 49 The Relationship between Ischemic Heart Disease and Diabetes A population-based approach 9

11 Outcomes The following endpoints were assessed: Paper I All-cause mortality, cardiovascular death (ICD-10 codes I00-I99), and new onset MI (ICD-10 codes I21, I22) Paper II Development of diabetes in the MI population: Defined as initiation of GLM (defined as claimed prescription of GLM (ATC-code A10) within the first year following discharge for first-time MI according to the year of the MI. Development of diabetes in the general population: Since the general population had no index event, the GLM initiation was registered according to the year of GLM commencement. Initiation of GLM was defined as claimed prescription of GLM (ATC-code A10). Determination of temporal trends in type of first initiated GLM: The following agents (ATC-codes) were identified: insulin (A10A), metformin (A10BA02), sulfonylureas (A10BB), glitazones (A10B); acarbose (A10F01), repaglinide (A10BX0) and metformin+repaglinide (A10BD03). Paper III Development of diabetes after study start and the subsequent risk of death associated with diabetes: Diabetes being defined as first claimed prescription of GLM (ATC-code A10). 10

12 Statistical analyses Ethics Paper I Multivariable analysis assessing risk of all-cause mortality, cardiovascular death and MI/recurrent MI was performed using Poisson regression analysis, adjusted for age, gender and calendar year. Kaplan-Meier estimates were used to assess temporal risk of cardiovascular death stratified by sex. The multivariable analysis revealed statistically significant interaction with sex and calendar time. Consequently, the study period was divided into two time periods January 1, 1997 to June 30, 2001 and July 1, 2001 to December 31, 2006 and analyzed separately as were the two genders. Paper II Incidence rates of GLM initiation in the MI population as well as in the general population were calculated. In order to assess temporal trends in GLM initiation, the study period was divided into five two-year periods (not applied in the incidence rates analyses). Multivariable Cox proportional-hazard models, adjusted for age, gender and calendar year, were used to investigate the likelihood of initiating GLM within a year post-mi according to the different year-groups using the year as the reference group. Test for temporal trends in pharmacological types of first GLM initiated were doneby Cochran-Armitage trend test. All the used Danish registers are placed in Statistics Denmark. Researchers can apply for access and when granted access is provided, the researcher has access to selected data. The access is provided with encrypted person identification and tight rules are preventing detailed output of individual people from being downloaded by researchers. Retrospective register studies require permission from the Danish Data Protection Agency and this is granted for Statistics Denmark to allow research use with the above mentioned limitations. The Danish Data Protection Agency approved the study (No ). According to the Danish regulations no ethical approval is required for retrospective register studies in Denmark. Paper III Kaplan-Meier s method was used to generate cumulative incidence curves of diabetes. Equality over strata was tested using the logrank test. Cox proportional-hazard models were used for analyses of the risk of developing diabetes and the subsequent risk of death associated with diabetes. The mortality analysis was additionally adjusted for HF severity group. A time-dependent Cox analysis with age as the underlying time variable was performed to investigate the prognostic importance of having developed diabetes. The model allowed for a continuous update on diabetes status (every day) and all patients were initially classified as not having diabetes. At the day of first claimed prescription of GLM, patients were classified as having diabetes. For all analyses, the statistical level of significance was set at Applying for all three papers analyses were performed using SAS version 9.1 (SAS institute Inc., Cary, NC, USA). For the incidence rates analyses in Paper II Stata version 11.0 (StataCorp, College Station, Tx, USA) was used. Continuous variables are presented as mean with standard deviation (SD). Categorical data are presented as percentages. The Relationship between Ischemic Heart Disease and Diabetes A population-based approach 11

13 Results overview of papers This section provides an abstracted overview of the three papers which are included in the thesis. For each paper a brief presentation of the aims, material and methods, results, and main conclusions is provided. For details of the studies please refer to the appendix section wherein all 3 papers are apparent in full. Paper I Changes in short- and long-term cardiovascular risk of incident diabetes and incident myocardial infarction - a nationwide study. Aim The aim of this study was to assess secular trends of cardiovascular outcomes following first diagnosis of MI or diabetes in an unselected population. Materials and methods All Danish residents aged 30 years without prior diabetes or MI were identified by individual-level-linkage of nationwide registers. Individuals hospitalised with MI or claiming a first-time prescription for a glucose-lowering medication (GLM) during the period from 1997 to 2006 were included. Analyses were done by Poisson regression models. Primary endpoints were death by all causes, cardiovascular death and MI. Results The study included 3,092,580 individuals of whom 77,147 had incident MI and 118,247 new-onset diabetes, Table 1. MI patients had an increased short-term risk of all endpoints compared with the general population. The rate ratio (RR) for cardiovascular death within the first year after MI was 11.1(95% CI ) in men and 14.8( ) in women, respectively. The risk rapidly declined and 1 year after the index MI, RR was 2.1 ( ) and 2.80 ( ) in men and women, respectively, Table 2. Patients with diabetes carried a constantly elevated risk of all endpoints compared with the general population. The cardiovascular death RR was 1.90( ) and 1.92( ) in men and women, respectively during the 1st year after GLM initiation. Main conclusions Incident MI is associated with high short-term risk, which decreases rapidly over time. Incident diabetes is associated with a persistent excessive cardiovascular risk after initiation of GLM therapy. This further strengthens the necessity of early multifactorial intervention in diabetes patients for long-term benefit. Table 1.Demographic characteristics according to exposure-group and time period Variables Incident MI Incident diabetes Reference Early period a Late period b Early period a Late period b Participants All (n) Men (n) Women (n) Male sex (%) Age (years) c All participants 69.7 (13.0) 70.5 (13.3) 62.7 (13.6) 63.0 (12.5) 52.7 (15.7) Men 66.9 (12.8) 67.5 (12.9) 61.0 (12.9) 61.7 (11.8) 51.1 (14.7) Women 74.2 (12.1) 75.1 (12.5) 64.8 (14.1) 64.7 (13.1) 54.1 (16.4) Prior hospital admissions d COPD (%) Renal disease (%) Cancer (%) Congestive heart failure (%) Values are mean + (SD) unless otherwise indicated a Time period 1 January 1997 to 30 June 2001, b Time period 1 July 2001 to 31 December 2006, c Defined on 1 January 1997, d Hospitalisation up to 1 year before incident event; for the reference population the date was set to 1 year prior to 30 June, 2001, unless death occurred before that date, in which case date was set to 1 year prior to 1 January, 1997 COPD, chronic obstructive pulmonary disease. 12

14 Table 2. Age-adjusted Poisson regression analyses using the background population as a reference Variables Incident MI Incident diabetes All-cause mortality Early period a Late period b Early period a Late period b Events (n) RR 0 1 years RR 1 3 years RR 3 5 years Cardiovascular death MI/recurrent MI Men 6.46 ( ) 6.17 ( ) 2.23 ( ) 2.29 ( ) Women 8.53 ( ) 8.67 ( ) 2.16 ( ) 2.51 ( ) Men 1.42 ( ) 1.47 ( ) 1.63 ( ) 1.42 ( ) Women 1.90 ( ) 2.02 ( ) 1.50 ( ) 1.49 ( ) Men 1.38 ( ) 1.46 ( ) 1.59 ( ) 1.42 ( ) Women 1.84 ( ) 1.80 ( ) 1.51 ( ) 1.42 ( ) Events (n) RR 0 1 years RR 1 3 years RR 3 5 years Men 11.1 ( ) 10.4 ( ) 1.90 ( ) 1.97 ( ) Women 14.8 ( ) 14.8 ( ) 1.92 ( ) 2.10 ( ) Men 2.11 ( ) 2.14 ( ) 1.79 ( ) 1.51 ( ) Women 2.80 ( ) 2.92 ( ) 1.70 ( ) 1.67 ( ) Men 1.99 ( ) 2.10 ( ) 1.89 ( ) 1.61 ( ) Women 2.63 ( ) 2.77 ( ) 1.78 ( ) 1.65 ( ) Events (n) RR 0 1 years RR 1 3 years RR 3 5 years Men 21.7 ( ) 20.2 ( ) 1.78 ( ) 1.72 ( ) Women 43.3 ( ) 42.6 ( ) 1.93 ( ) 2.05 ( ) Men 2.99 ( ) 2.92 ( ) 1.62 ( ) 1.29 ( ) Women 5.67 ( ) 5.64 ( ) 1.93 ( ) 1.72 ( ) Men 2.67 ( ) 2.70 ( ) 1.58 ( ) 1.33 ( ) Women 4.33 ( ) 5.15 ( ) 1.97 ( ) 1.89 ( ) Values are RR (95% CI) unless otherwise indicated a Time period 1 January 1997 to 30 June 2001; b time period 1 July 2001 to 31 December 2006 The Relationship between Ischemic Heart Disease and Diabetes A population-based approach 13

15 Paper II Temporal changes in the initiation of glucoselowering medications after a first-time myocardial infarction- a nationwide study between 1997 and 2006 Aim The aim of the study was to investigate temporal trends in the initiation of glucose-lowering medications (GLM) among patients discharged after first-time myocardial infarction and among the general population. Furthermore, temporal trends in type of first GLM were assessed. Material and methods All Danish residents aged 30 years without prior diabetes hospitalised with first-time MI between 1997 and 2006 were identified by individual-level-linkage of nationwide registers. Initiation of GLM during follow-up was assessed by claimed prescriptions from pharmacies. The study population was divided into five two-year periods (with exception of the calculated incidence rates). Temporal trends in initiation of GLM were assessed by incidence rate calculations in the MI population as in the general population. Temporal trends in initiation of different pharmacological types of first GLM was done in percentages also dividing the study period into five two year periods. Results In total 77, 147 patients 30 years of age were hospitalized with first-time MI, of whom 66,788 patients were alive at discharge and had previously never used GLM. Within the first year post discharge from first-time MI 1567 patients initiated GLM. Of these 1452 patients were included in the incidence rate analysis since 115 patients initiated GLM in 2006 were excluded due to incomplete follow-up time. Demographic characteristics are shown in Table 3.The incidence rate of GLM initiation within the first year after discharge for first-time MI in 1997 was approximately 19.6 per 1000 person-years. This incidence rate increased until year After year 2001 the incidence rate stabilized, Figure 1. A similar trend was observed in the general population. The incidence rate increased from 2.8 to 4.0 per 1000 person-years in Following year 2004 the incidence rate stabilized, Figure 2.In the entire population initiating GLM after first-time MI, the two predominantly prescribed GLMs were sulfonylureas and metformin, Table 4. During the study period the use of metformin as first GLM increased significantly from 30.5% in to 40.0% in (p for trend < ). Conversely, the use of sulfonylureas decreased from 59.0% in to 45.5% in (p for trend < ). The use of insulin as first-glm also increased from 9.1% in to 14.3% in (p for trend < 0.002). Table 3.Demographic characteristics at the time of first-time myocardial infarction according to year of occurrence Patients with first-time MI a Patients initiating GLM b within a year post-mi Men (%) Mean age (SD) 67.8 (12.9) 68.5 (13.1) 69.3 (13.1) 69.4 (13.2) 69.4 (13.2) Comorbidity c Congestive heart failure (%) Cardiac arrhythmias (%) PVD d (%) CVD e (%) Renal disease (%) Cancer (%) Shock (%) COLD f (%) Peptic ulcer a MI: myocardial infarction; b GLM :glucose-lowering medication; c comorbidity up till one year prior to discharge from index hospitalization; d PVD: peripheral vascular disease; e CVD: cerebrovascular disease; f COLD: chronic obstructive lung disease. 14

16 Figure 1. Incidence rate per 1000 person-years of initiation of glucose-lowering medication (GLM) within a year following for first-time myocardial infarction. Figure 2. Incidence rate per 1000 person-years of initiation of glucose-lowering medication (GLM) in the general population. Table 4. Distribution of first prescription claimed for glucose-lowering medications according to year of first-time myocardial infarction p for trend. Metformin (%) < Insulin (%) Sulfonylurea (%) < Other GLM(%) a a GLM: glucose-lowering medications: Other GLM: glitazones, acarbose, repaglinide, and repaglinide+metformin. The Relationship between Ischemic Heart Disease and Diabetes A population-based approach 15

17 Main conclusions The study demonstrated increased incidence rates of GLM initiation within the first year post-mi. A similar trend was shown in the general population. The results of the study suggest that the increase in the initiation of GLM in the MI population was primarily the effect of a general increased awareness of diabetes in the population as a whole. From a public health perspective, this study underscores a continuous need for diagnostic and therapeutic improvement in the care of MI patients that develop diabetes, and such strategy is likely to provide long-term benefits. 16

18 Paper III Heart failure severity, as determined by loop diuretic dosages, predicts the risk of developing diabetes after myocardial infarction: a nationwide cohort study. Aim The aim of the study was therefore to investigate the relationship between HF severity and risk of diabetes in patients discharged after MI and to investigate the prognostic importance of incident diabetes after MI. Material and methods All Danish residents, without prior diabetes (claiming GLM prior to 1997) or previous use of loop-diuretics, hospitalized with first-time MI between 1997 and 2006 and alive 90 after discharge were identified by individual-level-linkage of nationwide registers. Patients were followed until first claimed prescription of GLM, death, or until the end of The risks of diabetes and the subsequent risk of death associated with diabetes were analyzed by multivariable Cox proportional-hazard models. Additionally, the mortality analysis was adjusted for HF severity group. HF severity was determined by loop diuretic dosage after discharge. Results The study included 50,874 patients. Of these, 9,516 (19%) had HF at study baseline. In this HF population 3,006 (6%) had mild ([loopdiuretic dosage] 40 mg/day), 5,383 (11%) moderate (> mg/day), and 1,127 (2%) severe ( 120 mg/day) HF. In total, 2,531 (5%) patients developed diabetes during follow-up. Table 5 presents the baseline characteristics stratified by HF severity. Overall, patients with HF were older than patients without HF (73±11 years, vs. 64±13 years, p<0.0001) and a greater proportion of patients with HF were women, compared to patients without HF (43% vs. 31%, p<0.0001). Increasing HF severity was associated with increased risk of diabetes, Table 6 (in selection). Patients who developed diabetes had an adjusted HR for death of 2.85 (95% confidence interval ). Increasing HF severity was associated with increasing HR for death, i.e., 1.35 ( ), 1.57 ( ), and 2.13 ( ), for mild, moderate, and severe HF, respectively, compared to patients without HF. Main conclusions This nationwide study indicates that HF increases the risk of developing diabetes in a severity-dependent manner after MI. The results of the present study suggest that increased focus should be put on prevention of diabetes in post-mi HF patients. This strategy will enable effectuation of early and aggressive evidence based treatment, hereby reducing the mortality and morbidity in these high risk patients. The Relationship between Ischemic Heart Disease and Diabetes A population-based approach 17

19 Table 5. Baseline characteristics: No HF Mild HF ( 40 mg/day) Moderate HF (> mg/day) Severe HF (>120 mg/day) Number: 41,358 3,006 5,383 1,127 % of study population: 81% 6% 11% 2% Gender, female (%) 31% 41% 43% 46% Age at hospitalization (years) 64 (±13) 72 (±11) 74 (±11) 75 (±11) Year of hospitalization: % 8% 13% 3% % 7% 12% 3% % 7% 12% 2% % 6% 11% 2% % 2% 5% 1% Concomitant diseases*: Pulmonary edema 0% 2% 2% 3% Atrial flutter/fibrillation 4% 8% 10% 11% Cerebro-vascular disease 3% 3% 5% 6% Peripheral occlusive artery disease 1% 1% 1% 1% Renal disease 1% 1% 1% 3% COPD 3% 7% 7% 9% Concomitant medications: Statins 58% 41% 40% 32% RASi 39% 60% 62% 67% Beta blockers 81% 76% 72% 69% Calcium blockers 24% 28% 30% 35% Thiazides 20% 20% 26% 34% Spironolactone 3% 14% 17% 25% Digoxin 5% 16% 22% 32% Aspirin 59% 55% 57% 57% Clopidogrel 44% 25% 23% 19% Vitamin K antagonists 6% 13% 15% 17% (*) refers to a hospitalization diagnosis in one year prior to acute myocardial infarction hospitalization. HF: heart failure. RASi: renin angiotensin system inhibitors. COPD: chronic obstructive pulmonary disease. Table 6. Risk of diabetes according to HF severity group. Overall HR (95% CI): No HF 1.00 (ref.) Mild HF ( 40 mg/day) 1.39 ( ) Moderate HF (> mg/day) 1.71 ( ) Severe HF (>120 mg/day) 1.99 ( ) 18

20 Methodological considerations Study design We decided only to include new onset diabetes and first-time MI and this approach was particularly important since we made a direct comparison between the cardiovascular risks in patients with first-time MI, incident diabetes and incident heart failure. This model enabled the establishment of incident diabetes and firsttime MI study populations with comparable disease durations, although it should be recognized that the precise time of onset of chronic diseases, e.g. diabetes, is not possible to determine in these administrative databases. This composes a limitation of our study. However, it is reasonable to assume duration of diabetes of 50, 51 7 to 10 years prior to initiation of GLM treatment. Poisson regression vs. Cox regression Cox regression is the most frequently used analysis for epidemiological cohort studies of censored data but is very inefficient and time consuming when large datasets are analyzed, as it considers one day at time. It furthermore prioritizes one time scale (the underlying time of the analysis) beyond other time-scales. In contrast, Poisson regression is an efficient alternative to Cox regression analyses of cohort studies with multiple timescales, time varying covariates, and is easier to handle when the aim is to investigate risks in certain time-bands (although time-stratification it is doable in Cox regression analyses). The interpretation of incidence rate ratios provided by Poisson regression and hazard ratios provided by Cox regression is comparable. Unlike a timedependent Cox regression, Poisson regression assumes a constant incidence rate in each of the time-bands investigated. If these circumstances are fulfilled, both analyses provide very similar results. 52 conclusions. The term protophatic bias refers to the situation in which the initial symptoms of the outcome of interest is the very reason for the initiation of the pharmacological treatment used as the exposure measure, i.e if GLM was initiated because of early manifestation of cardiovascular disease. GLM however, with exception of metformin as monotherapy in young women presenting with PCOS (polycystic ovarian syndrome), are only used in patients with diabetes. Therefore, there is no reason to suspect any impact of such bias in the presented studies. Along this line, confounding by indication, a source of bias in non-randomized studies with potentially important differences in determinants of the outcome of interest between exposed and unexposed, is unlikely to be an important issue in our studies. Indeed in our study design, the treatment used to define exposure (GLM) was initiated because of diabetes (which is associated with increased cardiovascular risk) but merely served as a specific indicator of diabetes. Selection bias, protopathic bias, and confounding by indication Selection bias is a systematic error in epidemiological studies that arises from the population selection where the association between exposure and outcome differs between those included and those not included. In this thesis, the data cover the entire population of Denmark independent of socioeconomic status, age or participation in specific health-insurance programs. Therefore, the risk of selection bias is reduced, and the study notably includes citizens both in and out of the labor market. The Danish health care system partially reimburses drug expenses, and all Danish pharmacies are required to register all dispensed drug prescriptions, ensuring complete registration. The use of pharmacological treatment to identify the main exposure (in this thesis - GLM treatment as a proxy for diabetes) in epidemiological studies may lead to important bias and misleading The Relationship between Ischemic Heart Disease and Diabetes A population-based approach 19

21 Discussion Overall, the present thesis aimed to enlighten several uncertainties in the relationship between ischemic heart disease and diabetes: Explore the short and long-term prognosis in patients with incident diabetes and incident MI. To clarify the extent to which incident MI patients developed diabetes requiring GLM both as dependent on HF severity and within a year after discharge for first-time MI in a nationwide setting. In the thesis following was demonstrated: 1. Incident diabetes is associated with a persistent twofold higher risk of cardiovascular adverse outcomes after initiation of GLM. 2. A trend towards a risk reduction in men but not in women with diabetes was demonstrated when comparing the 1997-mid 2001 cohort (early period) with the mid cohort (late period). 3. Incident MI is associated with a high short-term risk, which decreases rapidly over time. 4. The incidence rates of diabetes (GLM-initiation) have increased in the MI population from year 1997 to Following year 2001 the incidence rates stabilized. 5. The incidence rates of diabetes (GLM-initiation) have increased accordingly in the general population from year 1997 to Corresponding to the MI population a stabilisation in incidence rate was observed in the general population after The use of metformin as first initiated GLM has increased from 30.5% to 40.0 % during the study period. Conversely, the use of sulfonylurea has decreased from 59.0% to 45.5%. The use of insulin as first initiated GLM has increased from 9.1% to 14.3%. 7. HF predicts development of diabetes in a severity-dependent manner among patients with MI. 8. Incident diabetes was found to be associated with nearly a threefold increase in risk of death in patients with MI. Diabetes and cardiovascular risk In this thesis incident diabetes was found to be associated with a persistent two-fold higher risk of cardiovascular adverse outcomes after initiation of GLM. Our results are consistent with previous studies by Haffner et al. and Schramm et al, respectively. Haffner et al. have demonstrated that patients with diabetes without prior MI exhibit the same risk of MI as patients without diabetes, but with prior MI. 5 Schramm et al. confirmed those results in a nationwide cohort of diabetes patients and provided evidence that diabetes patients requiring GLM therapy have a cardiovascular risk comparable to non-diabetic patients with a prior MI in all ages and among both sexes. 4 These findings applied for the prevalent diagnosis and only in patients requiring GLM therapy. As for the diabetes population, the results of our study extend the findings of Schramm et al. in that we document that patients with diabetes already carry a twofold higher risk of cardiovascular death, from the point in time when GLM therapy is initiated. Furthermore, this thesis documents an increase in the incidence rates of diabetes (GLM-initiation) in the MI population from year 1997 to 2001, stabilizing following year Accordingly, the incidence rates of diabetes (GLM-initiation) have increased in the general population from year 1997 to 2004, stabilizing after year We also document only a trend towards a risk reduction in men but not in women with diabetes when comparing the mid 2001 cohort (early period) with the mid cohort (late period). During the ten year study period several novel treatment modalities have been implemented and new glucose lowering agents have been introduced. Moreover, the use of all GLM, including insulin alone or in supplement to oral GLM has increased substantially following the UKPDS. Initially, the results from the UKPDS did not conclusively document any beneficial effect of intensified glycaemic control on prevention of macrovascular complications. However, epidemiological data from the UKPDS and observational follow-up data from Epidemiology of Diabetes Intervention and Complications (DCCT-EDIC) found that intensive glycaemic control was associated with a 16 % and 57% reduction in major CVD outcomes, respectively. The results from the abovementioned study support the possibility that such benefits occur. Furthermore, recent data have shown a beneficial legacy effect of early intensive glycaemic control on prevention of cardiovascular disease in patients with type 2 diabetes. 29 The Steno-2 study showed that early multi-factorial treatment including blood-pressure-, glucose- and lipid-lowering therapies, along with lifestyle interventions prevent (or postpone) both micro- and macrovascular complications in patients with type 2 diabetes. 30, 31 Over the last decade, these findings have been implemented in both national and international guidelines. Generally, the awareness on diabetes and its cardiovascular complications has increased significantly during the study period launching a broad range of large randomized clinical trials in the search of documenting a beneficial effect of intensified glycaemic control on prevention of macrovascular complications. Recently, three randomized controlled trials compared the effect of intensive vs. standard glycaemic control on cardiovascular disease (CVD) outcomes in relatively high risk participants with type 2 diabetes Surprisingly, all three studies failed to show a reduction in CVD outcomes. Speculations on why these three randomized trials failed to demonstrate a reduction in CVD outcomes have been many. First, all three trials were carried out on patients with established diabetes (mean duration 8-11 years) with either a large burden of risk factors or established CVD. Secondly, the additive beneficial effect of intensive glycaemic control may be difficult to demonstrate on top of the effects of CVD prevention by tight blood pressure control, aspirin, statins and blockers of the reninangiotensin system. Finally, it can be hypothesized that if these 20

22 randomized trials had included participants with shorter duration of diabetes and/or absence of CVD, a benefit from intensive glycaemic control could have been evident. Despite the negative outcome of the abovementioned studies, The American Diabetes Association (ADA) currently stated that glycaemic control early in the course of type 2 diabetes may provide CVD benefits and that glycaemic control may play a greater role for CVD prevention in patients with a shorter duration of diabetes before macrovascular disease is established. (class 2b recommendation, level of evidence A, for the benefit of glycaemic control on cardiovascular disease). 56 A metaanalysis in the Lancet in 2009 found that intensive glucose-lowering treatment significantly reduced events of non-fatal MI by 17% and a reduction of coronary heart disease events by 15%. The effect of co-morbidity on the relationship between glycaemic control and cardiovascular outcomes in diabetes has also recently been studied. Not surprisingly, this study revealed that the burden of co-morbidity does indeed affect this relationship. 57 This study concludes that patients with high levels of co-morbidity may have diminished cardiovascular benefits from intensive blood glucose control. Taking the body of the abovementioned evidence into account, it is somewhat surprising and disappointing that this thesis document only a trend towards a risk reduction in men but not in women with diabetes when comparing the 1997-mid 2001 cohort (early period) with the mid cohort (late period). Furthermore, since current evidence points toward a beneficial effect on cardiovascular events in patients with a shorter duration of diabetes we could have hoped for a steeper increase in incidence rates in GLM initiation in the post MI patient group than in the general population, since these results indicate that the increase is primarily an effect of a general increased awareness of diabetes in the population as a whole. On the basis of extensive trial evidence, following conclusion can be drawn: Glucose-lowering therapy may require many years to yield apparent benefits. Further evidence is needed from studies with extended follow-up before the true benefits of achieving a near normoglycaemia in patients with type 2 diabetes can be assessed. For the time being, until specific guidelines are established, a pragmatic approach differentiating between new-onset diabetes patients with a lesser burden of co-morbidity and patients with a longer duration of diabetes/higher degree of co-morbidity should be made. Since the beneficial effect on cardiovascular adverse events are seemingly more apparent in relative low risk patients the intensive glucose-lowering regimes should be preserved for this patients group and lesser stringent targets might be appropriate for patients with more advanced disease of longer duration and higher baseline HbA1c concentration. Importantly, the glucose-lowering treatment should aim for the lowest glycaemia that can safely be sustained. Temporal trends in type of first initiated GLM This thesis also briefly examined the temporal trends in initiation of different pharmacological type of first initiated GLM during our 10-year study period and found that the use of metformin has increased significantly from 30.5% in to 40.0% in Conversely, the use of sulfonylureas decreased from 59.0% in to 45.5% in The use of insulin as first- GLM also increased from 9.1% in to 14.3% in The change in the pattern of type of GLM used in MI patients during the period is likely to reflect the results of the UKPDS as well as the DIGAMI1 study signaling increased survival among users of metformin in general as well as the value of early insulin treatment after a recent MI. 24, 58 This has shown to be valid also for patients undergoing percutaneous intervention (PCI), in whom the use of metformin or thiazolidinediones seem to attenuate the re-stenosis formation. 59 Opposite, the effect of sulfonylureas or insulin on outcomes in MI patients treated with and without PCI seems to be less beneficial. 60, 61 It is therefore somewhat surprising that although the proportion of patients treated with metformin has increased significantly during the study period the relative risk of all-cause mortality, cardiovascular death and recurrent MI in the incident diabetes population largely remain constant. The extents to which these changes of clinical practice may change the longtime survival in Danish MI patients with diabetes remain to be determined. MI and Cardiovascular risk The main focus of the present thesis was not to explore the relationship between MI and cardiovascular risk, but nevertheless, no relative risk reduction was documented when comparing the incident MI patients from the two time periods ((1997 to mid 2001) and (mid 2001 to 2006). Significant improvements in the treatment of patients with acute MI have been implemented during the last decade. In 2003 percutaneous coronary intervention was applied a gold-standard in Denmark for acute ST-elevation MI 62 and the use of beta-blockers, platelet inhibitors, and statins has increased significantly Insufficient treatment with beta-blockers and statins following hospitalization after MI could be a plausible explanation for the lack of a risk reduction. It has been documented previously that if treatment with these agents are not initiated in close proximity to discharge, 44, 67 the probability of ever receiving treatment is low. Possibly, the lack of risk reduction could partially be explained by the need for a longer observation time after the improved use of secondary prophylactic treatment before the prognosis changes. Nevertheless, our study emphasizes the need to maintain focus on intensifying the secondary prophylactic treatment in extension to hospitalization and post-discharge for first-time MI. Interestingly, our study indicates that women with first-time MI have a higher risk of recurrent MI than men event after adjusting for age. The poorer outcomes in women than in men after MI have widely been attributed to the fact that women tend to have infarctions at an older age and more co-morbidities than men. Sex has generally not been interpreted as constituting an independent association. 68 In part, the explanation of the higher risk in women could be explained by the fact that women with acute MI are approached in a much less aggressively invasive way and receive less interventional treatment. 69 However, full reason remains to be explored taking into account discrepancies in both invasive treatment as well as discrepancies in concomitant medical treatment. 69 Future studies on the subject are greatly warranted. The development of diabetes in MI patients In paper II we aimed to investigate the temporal trends in the initiation of GLM in the period 1997 to 2006 prior to the guidelines recommending an OGTT in both MI patients as well as in the gen- The Relationship between Ischemic Heart Disease and Diabetes A population-based approach 21

23 eral population. As outlines previously, our study demonstrated an increase in the incidence rate of GLM initiation within the first year after first-time MI from the year 1997 until Following year 2001 the incidence rates stabilized. Our finding was confirmed in adjusted Cox-analyses. Moreover, our study demonstrated a similar trend in the initiation of GLM in the general population during the 10 year study period. Corresponding to the MI population a stabilisation in incidence rate was observed in the general population after We expected to find an increase in the incidence rates of GLM initiation in both the MI population as well as in the general population. Furthermore, we anticipated that the increase in GLM initiation would be more pronounced in the MI-population population viewed in the light of an increased focus on diabetes as an independent risk factor for further cardiovascular adverse events in patients with established cardiovascular disease. The increased incidence of diabetes in the general population was in consistency with a previous study employing the Danish National Diabetes Register, where Carstensen et al. demonstrated an increase in the incidence of diabetes. 70 It is questionable whether the increase in incidence rates of diabetes in Denmark represents a true epidemiologic increment. Other possible explanations include 1) increased diagnostic intensity due to an augmentation of the awareness of diabetes in the general population, and 2) a large undiagnosed population of patients with diabetes early in the study period being diagnosed during the rest of the study period, and 3) a decrease in mortality rates in the general population leading to increased numbers of subjects at risk of diabetes. The extent to which these individual factors each contribute to the increase in incidence of diabetes is speculative. Emphasis on differences in the inclusion criteria between our study and that of Carstensen et al. must be acknowledged in that we only documented an increase in incidence of patients with diabetes requiring GLM. Several previous studies have indicated that a large proportion of post-mi patients have undiagnosed diabetes or develop diabetes during admission or within a short time-span after discharge. 32, 33, 35, 36, 71 We therefore anticipated that a much larger proportion of patients would have initiated GLM shortly after discharge for MI, than observed in our current study. There are several plausible explanations for the finding. First, diabetes in many patients was probably left undiagnosed, because the majority of patients presented with normal fasting glucose values and therefore in accord with contemporary guidelines, did not undergo an OGTT, although many would here have impaired glucose tolerance or overt diabetes. 35 Importantly, however, because our study was conducted prior to publication of prevailing international guidelines, these numbers should not be interpreted as representing a predominant negligence from clinicians. Second, guidelines at the time of the study generally recommended lifestyle intervention and dietary treatment as first-line treatment for patients with newly detected diabetes. Conceivably, a substantial number of patients not initiating GLM were in fact being managed by non pharmacological intervention. However, since diabetes is well-established as an important independent long-term prognostic factor after MI that predicts increased mortality even 17 years after MI, 3 our study underscores the importance of an aggressive diagnostic and therapeutic approach in diabetes patients with MI. Lifestyle interventions were highly unlikely to be sufficient treatment in reaching target blood glucose values for all diabetes patients not initiating GLM. Our results indicate that a considerable number of post-mi patients do not receive GLM as appropriate. Indeed the incidence rates are, not surprisingly, much higher in the MI population than in the general population. MI patients are at much higher risk of developing diabetes. If the higher incidence rates in the MI population compared with the general population were solely to be explained by a true increased awareness of diabetes in the MI population and not in the general population as a whole, a steeper, linear or even exponential increment of incident diabetes in the MI population would be expected. Since we did not find such steep increase, it is most likely that the increased initiation of GLM in the MI population was, in fact, primarily the effect of a general increased awareness of diabetes in the population as a whole. Focus on early initiation of GLM is of utmost importance and the current results shows that when it comes to early GLM initiation in those with newly detected diabetes, there remains a considerable room for improvement in modern clinical practice. Post-MI development of HF and risk of diabetes Several factors may, in part, explain the current findings of increasing diabetes incidence with increasing HF severity. First, heart failure, the metabolic syndrome, and diabetes are low-grade inflammatory states, and it is notable that tumor necrosis factor-a, a central pro-inflammatory mediator, can promote both insulin resistance and HF in experimental models. 72, 73 Moreover, in patients with heart failure, circulating levels of tumor necrosis factora correlate with disease severity and prognosis, and renin- angiotensin-system inhibitors (RASi) are associated with reduced levels of tumor necrosis factor-a in these patients. 73 Another contributing factor may relate directly to reduced cardiac output in heart failure causing diminished blood flow and thereby diminished delivery of insulin and glucose to major insulin-sensitive tissues, such as the skeletal muscle 74. Heart failure is also associated with physical inactivity and peak oxygen capacity has previously been shown to correlate with insulin resistance in patients with heart failure 75. This latter association may also to some extent be mediated by reduced peripheral blood flow. A high adrenergic tone, as is commonly seen in heart failure patients, may also contribute to our findings by increasing peripheral insulin resistance in the skeletal muscle and increasing hepatic gluconeogenesis. 76, 77 It is also possible that diminished skeletal muscle blood flow by a- adrenergic-mediated vasoconstriction plays a role in this relationship. While it is well known that treatment with thiazides increases the risk of development of diabetes, no such direct relationship has, to our knowledge, been documented in relation to treatment with loop-diuretics The mechanisms accounting for the increased risk of diabetes when treated with thiazides is not completely understood. However, it has been suggested that hypokalaemia and/or excess aldosterone levels play part in this interrelationship. 79, 83, 84 Indeed, increased plasma aldosterone levels have been reported in patients with untreated HF subjected to loop diuretics, 85 and increasing aldosterone levels were recently confirmed to be positively correlated with insulin resistance in patients with HF. 86 Therefore, a direct effect of loop-diuretics on the risk of developing diabetes must be considered. One mechanism underlying the association between high aldosterone levels and risk of diabetes could potentially relate to diminished insulin secretion secondary to hypokalaemia, as several experimental studies have 22

24 found insulin secretion to decrease with low-potassium state. 83 In line with these observations, we found that the increase in the risk of diabetes was attenuated by the concomitant use of RASi, which potentially could be a consequence of either decreased aldosterone levels or increased potassium levels. The latter observation should be interpreted with caution since the indication for treatment with RASi were unknown to us (e.g. some patients receiving RASi, but no diuretics, may have suffered from hypertension as part of the metabolic syndrome). The findings are in concordance with the findings in the Candesartan in Heart failure: Assessment of Reduction in Mortality and morbidity (CHARM) trial. In this study heart failure patients (mainly NYHA class II III) were randomly assigned to candesartan or placebo. They found candesartan to decrease the risk of diabetes by 22% (relative risk 0.78, 95% CI ). 87 However, other studies exploring the effect of RASi on glucose metabolism (in different populations) have demonstrated varying results Studies including hypertensivepatients receiving thiazides reported that RASi attenuated the risk of diabetes, 89, 90 but the Diabetes Reduction Assessment with Ramipril and Rosiglitazone Medication (DREAM) trial, (aiming to compare the effects of ramipril with placebo on the risk of developing diabetes in patients without cardiovascular disease, but with glucose intolerance at baseline), found no differential risk of diabetes in the two groups. 88 This could suggest that the potential diabetes risk-lowering effect of RASi is limited to patients with elevated aldosterone concentrations, such as those with HF and/or on diuretic treatment. Development of diabetes in heart failure patients and the risk of death In paper III in this thesis, we document that patients who developed diabetes have a nearly three-fold increased risk of death. This result emphasizes the importance of identification of individuals at high risk of diabetes, whom potentially would benefit from primary diabetes prevention programmes and a more aggressive therapeutic approach, where metformin and acarbose may be attractive pharmacological options. Acarbose inhibits intestinal glucose uptake, leading to diminished post-prandial hyperglycaemia and improved insulin sensitivity, and has been documented to decrease the risk of diabetes in patients at high risk. 91 Treatment with metformin has also been reported to decrease the risk diabetes by improves insulin sensitivity 92 indeed, because insulin resistance may compromise myocardial metabolism, 93 metformin has been suggested as a new option for improving outcomes in patients with HF, irrespective of diabetes status. 94 In the presence of stressors such as pressure/volume overload and tachycardia, a compromised myocardial metabolism is incapable of meeting energy demand, leading to worsened HF 93. In this context, we and others have previously found metformin to be associated with better outcomes than sulfonylureas or insulin agents in patients with heart failure and established diabetes. 95, 96 Although it is well known that a poor glucose regulation increases the risk of HF, 39 it remains unknown whether a strict glycaemic control can improve cardiac function in these patients. The mechanisms underlying the blood glucoselowering effects of available pharmacological agents may therefore be of greater importance for their clinical effects than specific target glucose levels. The Relationship between Ischemic Heart Disease and Diabetes A population-based approach 23

25 Study strengths Study weaknesses The studies presented in this thesis were strengthened by the large number of participants, the nationwide coverage of prospectively recorded registries, the use of validated endpoints, the real life contemporary clinical setting, and the completeness of follow up. The data cover the entire population of Denmark independent of socioeconomic status, age or participation in specific healthinsurance programs. Notably, the studies in the thesis included citizens in and out of the labour market. Therefore, the risk of selection bias is avoided. Moreover, the Danish Health Care System partially reimburses drug expenses, and all Danish pharmacies are required to register all drug prescriptions, ensuring complete registration. The cardiovascular endpoints of the studies were strengthened by the study design which controlled for important measured confounders (i.e comorbidity, concomitant pharmacotherapy, and age). Observational studies may reveal disease associations but cannot establish causal relationships, and the results of this thesis should be interpreted in this context. The potentially important sources of bias have been mentioned in section selection bias, protophaticbias, and confounding by indication. As mentioned previously these types of bias are unlikely to have important influence on the results in this thesis. Still, it is impossible to completely refute an impact of one or more of these biases on the results presented in this thesis. The Danish population is predominantly of Caucasian descent; therefore, the results presented in this thesis should only be extrapolated to other ethnicities with caution. In addition, the possibility of residual confounding is present in all observational studies, and the registers, hold no information on no well known risk factors for cardiovascular disease such as hypertension, body mass index, physical activity, lipid disorders, dietary factors, microalbuminuria, blood glucose levels, HbA 1c levels, smoking or results from echocardiography (latter in paper III). Therefore, these potentially important confounders are not controlled for in this thesis. An important limitation of this thesis is that diabetes was defined by a claimed prescription for GLM. This fact is likely to have underestimated the true incidence/prevalence of diabetes (diet-treated and undiagnosed patients could not be identified). Moreover, it is important to mention the fact that the diabetes population was not diagnosed in connection to an admission to a hospital. Therefore, the initial risk in the diabetes population (in paper I) is not completely comparable to that of the MI population of the study. The diagnostic criteria for MI changed in 1999, with more sensitive markers being introduced (i.e. troponins) during the study period. Nevertheless, this has not resulted in dramatic changes in the prognosis for MI patients, and it is unlikely that this has affected our study. 97 Loop-diuretic treatment and dosages were used as a proxy for heart failure, since the diagnosis of heart failure in The Danish National Patient Register has a low sensitivity (29%) and we also had no information on clinical symptoms or left ventricular ejections fraction. 98 We furthermore, acknowledge the heterogeneity in the baseline characteristics across the study groups and cannot exclude that the relationship between heart failure severity and risk of diabetes to be influenced by unmeasured factors. 24

26 Novelty of the results Conclusions/Implications The present thesis extents the knowledge from the previous work by Haffner et al. and Schramm et al., that patients with established diabetes carry a twofold higher risk of cardiovascular adverse outcomes compared with patients without diabetes 4, 5.We document that this twofold higher risk also applies for the incident diabetes patient. The general awareness of diabetes has increased and the thesis shows that there is an increased incidence rate of GLM initiation in the general population as well as in the MI population suggesting that the increased incident rate of GLM initiation primarily reflects the generally increased awareness of diabetes more than an increased awareness among high risk patients with established cardiovascular disease. Furthermore, this thesis describes the proportion of patients hospitalized with first-time MI who within a year post MI develops diabetes requiring GLM. The poor prognosis in patients with MI and heart failure is wellknown as is the increased risk of HF in diabetes patients. 38 However, to our knowledge it has not previously been documented that HF increases the risk of diabetes according to the severity of HF after MI. This thesis suggests that heart failure increases the risk of diabetes in a severity-dependent manner after MI. This thesis focuses on the relationship between ischemic heart disease and diabetes. This thesis confirm previous knowledge that diabetes patients exhibit an elevated risk of death and cardiovascular morbidity and mortality and extend this knowledge to applying also to the incident diabetes patients. This knowledge is of utmost importance in caring for the diabetes patients and also for the ischemic patient since special attention is needed when treating these patient groups in order to reduce their morbidity and mortality. The extent to which glycaemic control adds to the reduction of future cardiovascular events and reduction of macrovascular complications still remains to be determined. However, it is of great importance to emphasize that control of hyperglycaemia should not take precedence over control of hypertension and cholesterol levels among adults with diabetes as part of an evidence-based global cardiovascular risk reduction effort. Nevertheless, screening for diabetes in patients hospitalized with first-time MI cannot be stressed enough. It has only recently been confirmed that fasting glucose measurements are insufficient in finding patients with diabetes in MI-population settings 102. In 2007 and 2008 the OGTT was implemented in the international and Danish national guidelines, respectively as the gold-standard test for revealing diabetes and dysglycaemic states in MI patients. 103, 104 The secondary prophylactic treatment in diabetes patients with established cardiovascular disease differs from that of the non-diabetic patient with cardiovascular disease in both target blood pressure values as in target cholesterol levels. These target values cannot be reached if diabetes remains undiagnosed. Importantly, if only fasting blood glucose levels are measured; twothirds of the MI patients with abnormal glucose metabolism remain undiagnosed. 32 Hopefully, the present thesis underscores the importance of usage of OGTT testing in patients with MI. Furthermore, this thesis emphasizes the importance of diabetes prevention through lifestyle intervention a continuous monitoring for the development of diabetes in patients with heart failure. The Relationship between Ischemic Heart Disease and Diabetes A population-based approach 25

27 Future research The primary objective of the future research on the role of glycaemic control in the prevention of CVD must be to conduct randomized clinical trials including patients with a short duration of diabetes, preferably at the point in time, where the patient is diagnosed with diabetes. The first study of this thesis demonstrated that the diabetes patients exhibit an elevated risk of death and cardiovascular morbidity and mortality from the point of time when treatment with GLM is initiated. Previously, the elevated risk has been documented in patients with prevalent diabetes. The randomized clinical trials ACCORD, VADT, and ADVANCE failed to document a reduction in CVD outcomes with intensive glycaemic control (HbA1c less than 6,0 % and 6,5 %, respectively). However, all three trials comprised patients with long diabetes duration from 8 to 11 years. At this time, it is plausible that the additive beneficial effect of intensive glycaemic control may be difficult to demonstrate on top of the effects of CVD prevention by tight blood pressure control, aspirin, statins and blockers of the renin-angiotensin system. It can be hypothesized that if these randomized trials had included participants with shorter duration of diabetes and/or absence of CVD, a benefit from intensive glycaemic control could have been evident. Therefore, large randomized clinical trials including patients with low prevalence of comorbidity and short duration of diabetes are urgently warranted. However, on the basis of current information, and the urgent need to address the residual risk of CVD in a rapidly expanding population with type 2 diabetes, it is premature to conclude that glucose control has no part to play especially in patients with short duration of diabetes. These prospective studies will more precisely define the diagnostic, therapeutic and prognostic implications of glucose controlin the context of cardiovascular disease and such data will form the background for guidelines aimed at cardiovascular risk management in patients with diabetes. A second objective of future research is to reinvestigate the true impact of the change of both diabetes treatment guidelines and changes in the guidelines on diagnosing diabetes in this real life population-scale setting. Paper II was conducted prior to publication of prevailing guidelines on diagnosing diabetes in MI patients with an OGTT, and all three papers in this thesis prior to prevailing guidelines on management of diabetes (where lifestyle intervention was first line treatment of newly detected diabetes). Hopefully, as time goes by, a larger number of patients are tested with an OGTT after hospitalization after first-time MI, thereby, enabling both early diagnosing of diabetes in patients with CVD but certainly also enabling early and aggressive multi-factorial intervention in these high-risk patients ultimately resulting in reduction of future cardiovascular events. 26

28 Summary Diabetes is a well-established risk factor for cardiovascular disease and is common among patients with acute myocardial infarction (MI), where the prevalence is as high as 20%. Patients with diabetes requiring glucose-lowering medication (GLM) have been reported as having the same long-term risk of major cardiovascular outcomes as patients with a previous MI. Focus on the adverse prognosis carried by diabetes in the post-mi patients has caused new guidelines to be developed recommending similar preventive strategies to reduce the risk in these patients. Furthermore, heart failure is common after myocardial infarction and smaller studies have recently indicated that heart failure (HF) may be associated with insulin resistance. Nevertheless, larger scale studies investigating this topic are currently missing. Because of the adverse prognosis carried by patients with both diabetes and MI it is of great importance to explore the short and long-term risk of having these diseases and furthermore to investigate the extent to which patients with MI develop diabetes in order to commence early and aggressive evidence based therapy providing long term benefit in these high risk patients. The present thesis had the following objectives on the subject: The relationship between ischemic heart disease and diabetes 1. To examine the short- and long-term risk of death and cardiovascular outcomes in patients with incident diabetes and in patients with first-time MI during a 10 year period in Denmark, using the general population as a reference. 2. To investigate the temporal trends in initiation of GLM following discharge for first-time MI between 1997 and To assess whether HF severity, as determined by loop diuretic dosages, predicts the risk of developing diabetes after myocardial infarction. The thesis revealed the following main results 1. 1) Patients developing GLM requiring diabetes carried an approximately two-fold risk of death of all causes, cardiovascular death and MI immediately following initiation of GLM treatment, compared to the general population. There was a high short-term risk of death, cardiovascular death and recurrent MI in the incident MI population with a rapid decline during the 1 st year. No decline of relative risk was demonstrated when comparing the MI population in the two the time-periods 1997 to mid-2001 and mid-2001 to ) An increased incidence rate of GLM initiation within the first year following discharge for first-time MI from the year 1997 until From 2001 and onwards the incidence rate stabilized. A similar trend was documented in the general population with an increased incidence rate from the year 1997 until Following 2004 the incidence rate stabilized. A temporal shift in type of first initiated GLM to first-time MI patients was demonstrated. The use of metformin and insulin increased significantly during the study period. Correspondingly, the use of sulfonylurea declined. 3) Our study demonstrated a linear relationship between HF severity post-mi and risk of developing diabetes, with patients using loop diuretics > 120 mg/day carrying an approximately two-fold increase in risk of diabetes, compared to patients without HF. Incident diabetes was found to carry nearly a threefold increase in relative risk of death. Conclusion 1. Patients developing GLM requiring diabetes carry an approximately two-fold increased risk of cardiovascular death compared with the general population. Patients with first-time MI have a high short-term risk of cardiovascular death rapidly declining and nearly resembling the risk carried by the diabetes population. The thesis was based on national administrative registers. The registers primarily employed for this thesis were: The Danish Register of Medicinal Product Statistics (the National Prescription Register), the Danish National Patient Register, The Central Population Register and the Danish Causes of Death Register. Patients admitted to the hospital with first-time MI or patients developing diabetes requiring GLM during the study period of 10 years (1. January December 2006) were included in the study. 2. Development of diabetes is frequent after first-time MI and probably under-diagnosed. 3. The risk of developing diabetes increases with increasing severity of heart failure. Focus on the development of diabetes in patients with ischemic heart disease with or without the presence of heart failure still compose a public health matter, because early and aggressive evidence-based therapy is thought to reduce the morbidity of patients with diabetes thereby improving the prognosis for these patients. The Relationship between Ischemic Heart Disease and Diabetes A population-based approach 27

29 Dansk Resumé Diabetes er hyppigt forekommende hos patienter, der har haft akut myokardieinfarkt(ami) med en rapporteret prævalens op mod 20 %. Patienter med behandlingskrævende diabetes har samme kardiovaskulære risiko som patienter, der tidligere har haft et AMI og samtidig forekomst af begge sygdomme er associeret med en dårlig prognose. Øget fokus omkring den dårlige prognose hos patienter med både iskæmisk hjertesygdom og diabetes har afstedkommet udarbejdelsen af guidelines, der anbefaler ensartede forbyggende tiltag hos de to patientgrupper. Endvidere er hjertesvigt hyppigt efter et AMI og mindre studier har indikeret at hjertesvigt er associeret med insulin resistens. Der mangler viden om området fra større studier. Set i lyset af den markant dårligere prognose hos patienter med både AMI og diabetes er det af stor samfundsmæssig værdi at få klarlagt korttids- samt langtidsrisikoen forbundet med AMI og nydiagnosticeret diabetes samt udviklingen af diabetes efter AMI, således at man i fremtiden kan målrette den forbyggende indsats og dermed nedsætte risikoen for kardiovaskulære hændelser hos disse højrisikopatientgrupper. Afhandlingen havde til formål at belyse følgende områder inden for forholdet mellem iskæmisk hjertesygdom og diabetes Korttids- og langtidsrisiko for død og kardiovaskulære hændelser i tiden efter indlæggelse for førstegangs AMI og i tiden efter opstart af antidiabetisk medicin (oral/insulin). Ændringer over tid i opstart af antidiabetika efter udskrivelse for førstegangs AMI. Betydningen af hjertesvigt for udvikling af behandlingskrævende diabetes hos patienter indlagt med førstegangs AMI. Afhandlingen bygger på data fra nationale registre. De nationale registre er samlet hos Danmarks Statistik og krypterede CPRnumre muliggør sammenføring af de forskellige registre på individniveau. Afhandlingen bygger på data fra primært 4 registre: lægemiddeldatabasen, landspatientregistret, CPR-registret samt dødsårsagsregisteret. Vi identificerede patienter indlagt med et førstegangs AMI samt personer, der opstartede antidiabetisk behandling indenfor den 10-årige studieperiode (1. januar 1997 til 31. december 2006). risiko reduktion for kardiovaskulær død, AMI eller død af alle årsager i AMI populationen, når perioden 1997 til medio 2001 og medio 2001 til 2006 sammenlignes. 2) at incidensraten af patienter, der opstarter i antidiabetika inden for et år efter udskrivelse efter et AMI er stigende fra 1997 til 2001, hvorefter raten stabiliseres. I baggrundsbefolkningen ses samme tendens med en stigning i incidensraten fra 1997 til 2004, hvorefter denne stabiliseres. Der er sket et skift i valg af første initierede antidiabetiske medicin til behandling af diabetes hos patienter med iskæmisk hjertesygdom. Brugen af metformin og insulin er signifikant steget, hvorimod brugen af sulfonylurinstoffer er faldet. 3) at der er en lineær sammenhæng mellem sværhedsgraden af HF (vurderet ud fra loop diuretika dosis) og risikoen for udvikling af behandlingskrævende diabetes. Risikoen for død øges med stigende grad af HF, ligesom at udvikle diabetes markant øger risikoen for død. Konklusion 1. Patienter, der udvikler behandlingskrævende diabetes, udviser dobbelt så stor risiko for kardiovaskulær død i forhold til den raske danske baggrundsbefolkning. Patienter med et førstegangs AMI har en høj korttidsrisiko, der er rapidt faldende nærmende sig en sammenlignelig risko med diabetes patienterne. 2. Udvikling af diabetes efter et førstegangs AMI er hyppigt men formentlig underdiagnosticeret. 3. Der er stigende risiko for udvikling af diabetes ved tiltagende sværhedsgrad af hjertesvigt. Fokus på udviklingen af diabetes hos patienter med iskæmisk hjertesygdom både med eller uden tilstedeværelsen af hjertesvigt har fortsat stor samfundsmæssig vigtighed, idet tidlig opsporing og behandling af diabetes hos denne patient gruppe potentielt vil reducere antallet af kardiovaskulære hændelser og dermed bedre prognosen hos disse højrisiko patienter. Afhandlingen har vist følgende 1) at patienter, der udvikler behandlingskrævende diabetes, har en konstant fordobling i risiko for kardiovaskulær død, AMI samt død af alle årsager sammenlignet med den danske baggrundsbefolkning. Den øgede risiko er allerede til stede ved opstart af antidiabetika. Patienter med førstegangs AMI derimod har en høj korttidsrisiko men denne risiko falder hurtigt inden for det første år efter indlæggelsen for et AMI. Til trods for store forbedringer i behandlingsregimerne for AMI patienter samt i behandlingen af diabetikere i løbet af studieperioden er der intet, der tyder på en relativ 28

30 References 1. SA L. Some clinical considerations. JAMA 1922: Zuanetti G, Latini R, Maggioni AP, Santoro L, Franzosi MG. Influence of diabetes on mortality in acute myocardial infarction: Data from the gissi-2 study. J Am Coll Cardiol. 1993;22: Kumler T, Gislason GH, Kober L, Torp-Pedersen C. Diabetes is an independent predictor of survival 17 years after myocardial infarction: Follow-up of the trace registry. Cardiovasc Diabetol. 2010;9:22 4. Schramm TK, Gislason GH, Kober L, Rasmussen S, Rasmussen JN, Abildstrom SZ, Hansen ML, Folke F, Buch P, Madsen M, Vaag A, Torp- Pedersen C. Diabetes patients requiring glucose-lowering therapy and nondiabetics with a prior myocardial infarction carry the same cardiovascular risk: A population study of 3.3 million people. Circulation. 2008;117: Haffner SM, Lehto S, Ronnemaa T, Pyorala K, Laakso M. Mortality from coronary heart disease in subjects with type 2 diabetes and in nondiabetic subjects with and without prior myocardial infarction. N Engl J Med. 1998;339: Malmberg K, Ryden L. Myocardial infarction in patients with diabetes mellitus. Eur Heart J. 1988;9: Malmberg K, Yusuf S, Gerstein HC, Brown J, Zhao F, Hunt D, Piegas L, Calvin J, Keltai M, Budaj A. Impact of diabetes on long-term prognosis in patients with unstable angina and non-q-wave myocardial infarction: Results of the oasis (organization to assess strategies for ischemic syndromes) registry. Circulation. 2000;102: Mukamal KJ, Nesto RW, Cohen MC, Muller JE, Maclure M, Sherwood JB, Mittleman MA. Impact of diabetes on long-term survival after acute myocardial infarction: Comparability of risk with prior myocardial infarction. Diabetes Care. 2001;24: Abbott RD, Donahue RP, Kannel WB, Wilson PW. The impact of diabetes on survival following myocardial infarction in men vs women. The framingham study. JAMA. 1988;260: DeFronzo R. International textbook of diabetes mellitus. 3rd. Edition Definition and diagnosis of diabetes mellitus and intermediate hyperglycaemia 12. Diagnosis and classification of diabetes mellitus. Diabetes Care. 2010;33 Suppl 1:S Norhammar A, Malmberg K, Ryden L, Tornvall P, Stenestrand U, Wallentin L. Under utilisation of evidence-based treatment partially explains for the unfavourable prognosis in diabetic patients with acute myocardial infarction. Eur Heart J. 2003;24: Preis SR, Hwang SJ, Coady S, Pencina MJ, D'Agostino RB, Sr., Savage PJ, Levy D, Fox CS. Trends in all-cause and cardiovascular disease mortality among women and men with and without diabetes mellitus in the framingham heart study, 1950 to Circulation. 2009;119: Thomas RJ, Palumbo PJ, Melton LJ, 3rd, Roger VL, Ransom J, O'Brien PC, Leibson CL. Trends in the mortality burden associated with diabetes mellitus: A population-based study in rochester, minn, Arch Intern Med. 2003;163: Mrc/bhf heart protection study of cholesterol lowering with simvastatin in 20,536 high-risk individuals: A randomised placebocontrolled trial. Lancet. 2002;360: Collaborative meta-analysis of randomised trials of antiplatelet therapy for prevention of death, myocardial infarction, and stroke in high risk patients. BMJ. 2002;324: Hayden M, Pignone M, Phillips C, Mulrow C. Aspirin for the primary prevention of cardiovascular events: A summary of the evidence for the u.s. Preventive services task force. Annals of Internal Medicine. 2002;136: Køber L, Torp-Pedersen C, Carlsen JE, Bagger H, Eliasen P, Lyngborg K, Videbæk J, Cole DS, Auclert L, Pauly NC, Aliot E, Persson S, Camm AJ. A clinical trial of the angiotensin-converting enzyme inhibitor trandolapril in patients with left ventricular dysfunction after myocardial infarction. New England Journal of Medicine. 1995;333: LaRosa JC, He J, Vupputuri S. Effect of statins on risk of coronary disease. JAMA: The Journal of the American Medical Association. 1999;282: Reichard P, Nilsson BY, Rosenqvist U. The effect of long-term intensified insulin treatment on the development of microvascular complications of diabetes mellitus. N Engl J Med. 1993;329: The effect of intensive treatment of diabetes on the development and progression of long-term complications in insulin-dependent diabetes mellitus. The diabetes control and complications trial research group. N Engl J Med. 1993;329: Ohkubo Y, Kishikawa H, Araki E, Miyata T, Isami S, Motoyoshi S, Kojima Y, Furuyoshi N, Shichiri M. Intensive insulin therapy prevents the progression of diabetic microvascular complications in japanese patients with non-insulin-dependent diabetes mellitus: A randomized prospective 6-year study. Diabetes Res Clin Pract. 1995;28: Effect of intensive blood-glucose control with metformin on complications in overweight patients with type 2 diabetes (ukpds 34). Uk prospective diabetes study (ukpds) group. Lancet. 1998;352: Intensive blood-glucose control with sulphonylureas or insulin compared with conventional treatment and risk of complications in patients with type 2 diabetes (ukpds 33). Uk prospective diabetes study (ukpds) group. Lancet. 1998;352: Muhlestein JB, Anderson JL, Horne BD, Lavasani F, Allen Maycock CA, Bair TL, Pearson RR, Carlquist JF. Effect of fasting glucose levels on mortality rate in patients with and without diabetes mellitus and coronary artery disease undergoing percutaneous coronary intervention. Am Heart J. 2003;146: Kannel WB, McGee DL. Diabetes and glucose tolerance as risk factors for cardiovascular disease: The framingham study. Diabetes Care. 1979;2: Melchior T, Kober L, Madsen CR, Seibaek M, Jensen GV, Hildebrandt P, Torp-Pedersen C. Accelerating impact of diabetes mellitus on mortality in the years following an acute myocardial infarction. Trace study group. Trandolapril cardiac evaluation. Eur Heart J. 1999;20: Holman RR, Paul SK, Bethel MA, Matthews DR, Neil HA. 10-year follow-up of intensive glucose control in type 2 diabetes. N Engl J Med. 2008;359: Gaede P, Vedel P, Larsen N, Jensen GV, Parving HH, Pedersen O. Multifactorial intervention and cardiovascular disease in patients with type 2 diabetes. N Engl J Med. 2003;348: Gaede P, Lund-Andersen H, Parving HH, Pedersen O. Effect of a multifactorial intervention on mortality in type 2 diabetes. N Engl J Med. 2008;358: Bartnik M, Ryden L, Ferrari R, Malmberg K, Pyorala K, Simoons M, Standl E, Soler-Soler J, Ohrvik J. The prevalence of abnormal glucose regulation in patients with coronary artery disease across europe. The euro heart survey on diabetes and the heart. Eur Heart J. 2004;25: Norhammar A, Tenerz A, Nilsson G, Hamsten A, Efendic S, Ryden L, Malmberg K. Glucose metabolism in patients with acute myocardial infarction and no previous diagnosis of diabetes mellitus: A prospective study. Lancet. 2002;359: The Relationship between Ischemic Heart Disease and Diabetes A population-based approach 29

31 34. Who consultation. Definition, diagnosis and classification of diabetes mellitus and its complications. Geneva: World health organisation. 35. Egstrup M, Henriksen FL, Hofsten DE, Jensen BK, Hangaard J, Egstrup K. [prevalence of glucometabolic abnormalities in connection with myocardial infarction in a danish coronary care unit]. Ugeskr Laeger. 2008;170: Hofsten DE, Logstrup BB, Moller JE, Pellikka PA, Egstrup K. Abnormal glucose metabolism in acute myocardial infarction: Influence on left ventricular function and prognosis. JACC Cardiovasc Imaging. 2009;2: Velagaleti RS, Pencina MJ, Murabito JM, Wang TJ, Parikh NI, D'Agostino RB, Levy D, Kannel WB, Vasan RS. Long-term trends in the incidence of heart failure after myocardial infarction. Circulation. 2008;118: Kannel WB, Hjortland M, Castelli WP. Role of diabetes in congestive heart failure: The framingham study. Am J Cardiol. 1974;34: Held C, Gerstein HC, Yusuf S, Zhao F, Hilbrich L, Anderson C, Sleight P, Teo K. Glucose levels predict hospitalization for congestive heart failure in patients at high cardiovascular risk. Circulation. 2007;115: Amato L, Paolisso G, Cacciatore F, Ferrara N, Ferrara P, Canonico S, Varricchio M, Rengo F. Congestive heart failure predicts the development of non-insulin-dependent diabetes mellitus in the elderly. The osservatorio geriatrico regione campania group. Diabetes Metab. 1997;23: AlZadjali MA, Godfrey V, Khan F, Choy A, Doney AS, Wong AK, Petrie JR, Struthers AD, Lang CC. Insulin resistance is highly prevalent and is associated with reduced exercise tolerance in nondiabetic patients with heart failure. J Am Coll Cardiol. 2009;53: Gaist D, Sorensen HT, Hallas J. The danish prescription registries. Dan Med Bull. 1997;44: Madsen M, Davidsen M, Rasmussen S, Abildstrom SZ, Osler M. The validity of the diagnosis of acute myocardial infarction in routine statistics: A comparison of mortality and hospital discharge data with the danish monica registry. J Clin Epidemiol. 2003;56: Gislason GH, Rasmussen JN, Abildstrom SZ, Gadsboll N, Buch P, Friberg J, Rasmussen S, Kober L, Stender S, Madsen M, Torp-Pedersen C. Long-term compliance with beta-blockers, angiotensin-converting enzyme inhibitors, and statins after acute myocardial infarction. Eur Heart J. 2006;27: Fosbol EL, Gislason GH, Jacobsen S, Folke F, Hansen ML, Schramm TK, Sorensen R, Rasmussen JN, Andersen SS, Abildstrom SZ, Traerup J, Poulsen HE, Rasmussen S, Kober L, Torp-Pedersen C. Risk of myocardial infarction and death associated with the use of nonsteroidal anti-inflammatory drugs (nsaids) among healthy individuals: A nationwide cohort study. Clin Pharmacol Ther. 2009;85: Eshaghian S, Horwich TB, Fonarow GC. Relation of loop diuretic dose to mortality in advanced heart failure. The American Journal of Cardiology. 2006;97: Flu W-J, van Kuijk J-P, Galal W, Kuiper R, van de Ven LL, Verhagen HJM, Bax JJ, Poldermans D. Prevalence and pharmacological treatment of left-ventricular dysfunction in patients undergoing vascular surgery. European Journal of Heart Failure. 2010;12: McMurray JJV, Carson PE, Komajda M, McKelvie R, Zile MR, Ptaszynska A, Staiger C, Donovan JM, Massie BM. Heart failure with preserved ejection fraction: Clinical characteristics of 4133 patients enrolled in the i-preserve trial. European Journal of Heart Failure. 2008;10: Tu JV, Austin PC, Walld R, Roos L, Agras J, McDonald KM. Development and validation of the ontario acute myocardial infarction mortality prediction rules. J Am Coll Cardiol. 2001;37: Harris MI, Klein R, Welborn TA, Knuiman MW. Onset of niddm occurs at least 4-7 yr before clinical diagnosis. Diabetes Care. 1992;15: Turner RC, Cull CA, Frighi V, Holman RR. Glycemic control with diet, sulfonylurea, metformin, or insulin in patients with type 2 diabetes mellitus: Progressive requirement for multiple therapies (ukpds 49). Uk prospective diabetes study (ukpds) group. JAMA. 1999;281: Levesque LE, Hanley JA, Kezouh A, Suissa S. Problem of immortal time bias in cohort studies: Example using statins for preventing progression of diabetes. BMJ. 2010;340:b Duckworth W, Abraira C, Moritz T, Reda D, Emanuele N, Reaven PD, Zieve FJ, Marks J, Davis SN, Hayward R, Warren SR, Goldman S, McCarren M, Vitek ME, Henderson WG, Huang GD. Glucose control and vascular complications in veterans with type 2 diabetes. N Engl J Med. 2009;360: Patel A, MacMahon S, Chalmers J, Neal B, Billot L, Woodward M, Marre M, Cooper M, Glasziou P, Grobbee D, Hamet P, Harrap S, Heller S, Liu L, Mancia G, Mogensen CE, Pan C, Poulter N, Rodgers A, Williams B, Bompoint S, de Galan BE, Joshi R, Travert F. Intensive blood glucose control and vascular outcomes in patients with type 2 diabetes. N Engl J Med. 2008;358: Gerstein HC, Miller ME, Byington RP, Goff DC, Jr., Bigger JT, Buse JB, Cushman WC, Genuth S, Ismail-Beigi F, Grimm RH, Jr., Probstfield JL, Simons-Morton DG, Friedewald WT. Effects of intensive glucose lowering in type 2 diabetes. N Engl J Med. 2008;358: Skyler JS, Bergenstal R, Bonow RO, Buse J, Deedwania P, Gale EA, Howard BV, Kirkman MS, Kosiborod M, Reaven P, Sherwin RS. Intensive glycemic control and the prevention of cardiovascular events: Implications of the accord, advance, and va diabetes trials: A position statement of the american diabetes association and a scientific statement of the american college of cardiology foundation and the american heart association. Diabetes Care. 2009;32: Greenfield S, Billimek J, Pellegrini F, Franciosi M, De Berardis G, Nicolucci A, Kaplan SH. Comorbidity affects the relationship between glycemic control and cardiovascular outcomes in diabetes: A cohort study. Ann Intern Med. 2009;151: Malmberg K. Prospective randomised study of intensive insulin treatment on long term survival after acute myocardial infarction in patients with diabetes mellitus. Digami (diabetes mellitus, insulin glucose infusion in acute myocardial infarction) study group. BMJ. 1997;314: Lexis CP, Rahel BM, Meeder JG, Zijlstra F, van der Horst IC. The role of glucose lowering agents on restenosis after percutaneous coronary intervention in patients with diabetes mellitus. Cardiovasc Diabetol. 2009;8: Jorgensen CH, Gislason GH, Andersson C, Ahlehoff O, Charlot M, Schramm TK, Vaag A, Abildstrom SZ, Torp-Pedersen C, Hansen PR. Effects of oral glucose-lowering drugs on long term outcomes in patients with diabetes mellitus following myocardial infarction not treated with emergent percutaneous coronary intervention--a retrospective nationwide cohort study. Cardiovasc Diabetol. 2010;9: Jorgensen CH, Gislason GH, Bretler D, Sorensen R, Norgaard ML, Hansen ML, Schramm TK, Abildstrom SZ, Torp-Pedersen C, Hansen PR. Glyburide increases risk in patients with diabetes mellitus after emergent percutaneous intervention for myocardial infarction - a nationwide study. Int J Cardiol Andersen HR, Nielsen TT, Rasmussen K, Thuesen L, Kelbaek H, Thayssen P, Abildgaard U, Pedersen F, Madsen JK, Grande P, Villadsen AB, Krusell LR, Haghfelt T, Lomholt P, Husted SE, Vigholt E, Kjaergard HK, Mortensen LS. A comparison of coronary angioplasty with fibrinolytic therapy in acute myocardial infarction. N Engl J Med. 2003;349: Gislason GH, Abildstrom SZ, Rasmussen JN, Rasmussen S, Buch P, Gustafsson I, Friberg J, Gadsboll N, Kober L, Stender S, Madsen M, Torp-Pedersen C. Nationwide trends in the prescription of betablockers and angiotensin-converting enzyme inhibitors after myocardial infarction in denmark, Scand Cardiovasc J. 2005;39: Kontny F. Improving outcomes in acute coronary syndromes--the frisc ii trial. Clin Cardiol. 2001;24:I Rasmussen JN, Gislason GH, Abildstrom SZ, Rasmussen S, Gustafsson I, Buch P, Friberg J, Kober L, Torp-Pedersen C, Madsen M, Stender S. Statin use after acute myocardial infarction: A nationwide study in denmark. Br J Clin Pharmacol. 2005;60:

32 66. Fox KA, Mehta SR, Peters R, Zhao F, Lakkis N, Gersh BJ, Yusuf S. Benefits and risks of the combination of clopidogrel and aspirin in patients undergoing surgical revascularization for non-st-elevation acute coronary syndrome: The clopidogrel in unstable angina to prevent recurrent ischemic events (cure) trial. Circulation. 2004;110: Lee HY, Cooke CE, Robertson TA. Use of secondary prevention drug therapy in patients with acute coronary syndrome after hospital discharge. J Manag Care Pharm. 2008;14: Malacrida R, Genoni M, Maggioni AP, Spataro V, Parish S, Palmer A, Collins R, Moccetti T. A comparison of the early outcome of acute myocardial infarction in women and men. The third international study of infarct survival collaborative group. N Engl J Med. 1998;338: Hvelplund A, Galatius S, Madsen M, Rasmussen JN, Rasmussen S, Madsen JK, Sand NP, Tilsted HH, Thayssen P, Sindby E, Hojbjerg S, Abildstrom SZ. Women with acute coronary syndrome are less invasively examined and subsequently less treated than men. Eur Heart J. 2010;31: Carstensen B, Kristensen JK, Ottosen P, Borch-Johnsen K. The danish national diabetes register: Trends in incidence, prevalence and mortality. Diabetologia. 2008;51: Seibaek M, Sloth C, Vallebo L, Hansen T, Urhammer SA, Burchardt H, Torp-Pedersen C, Pedersen O, Hildebrandt P. Glucose tolerance status and severity of coronary artery disease in men referred to coronary arteriography. Am Heart J. 1997;133: Dandona P, Aljada A, Chaudhuri A, Mohanty P, Garg R. Metabolic syndrome: A comprehensive perspective based on interactions between obesity, diabetes, and inflammation. Circulation. 2005;111: Feldman AM, Combes A, Wagner D, Kadakomi T, Kubota T, Li YY, McTiernan C. The role of tumor necrosis factor in the pathophysiology of heart failure. J Am Coll Cardiol. 2000;35: Barrett EJ, Eggleston EM, Inyard AC, Wang H, Li G, Chai W, Liu Z. The vascular actions of insulin control its delivery to muscle and regulate the rate-limiting step in skeletal muscle insulin action. Diabetologia. 2009;52: Swan JW, Anker SD, Walton C, Godsland IF, Clark AL, Leyva F, Stevenson JC, Coats AJ. Insulin resistance in chronic heart failure: Relation to severity and etiology of heart failure. J Am Coll Cardiol. 1997;30: Triposkiadis F, Karayannis G, Giamouzis G, Skoularigis J, Louridas G, Butler J. The sympathetic nervous system in heart failure physiology, pathophysiology, and clinical implications. J Am Coll Cardiol. 2009;54: Kostis JB, Sanders M. The association of heart failure with insulin resistance and the development of type 2 diabetes. Am J Hypertens. 2005;18: Andersson OK, Gudbrandsson T, Jamerson K. Metabolic adverse effects of thiazide diuretics: The importance of normokalaemia. J Intern Med Suppl. 1991;735: Bakris G, Stockert J, Molitch M, Zhou Q, Champion A, Bacher P, Sowers J. Risk factor assessment for new onset diabetes: Literature review. Diabetes Obes Metab. 2009;11: Goldner MG, Zarowitz H, Akgun S. Hyperglycemia and glycosuria due to thiazide derivatives administered in diabetes mellitus. N Engl J Med. 1960;262: Plavinik FL, Rodrigues CI, Zanella MT, Ribeiro AB. Hypokalemia, glucose intolerance, and hyperinsulinemia during diuretic therapy. Hypertension. 1992;19:II Sagild U, Andersen V, Andreasen PB. Glucose tolerance and insulin responsiveness in experimental potassium depletion. Acta Med Scand. 1961;169: Zillich AJ, Haines ST. Ashp therapeutic position statement on the treatment of hypertension. Am J Health Syst Pharm. 2006;63: Gorden P. Glucose intolerance with hypokalemia. Failure of shortterm potassium depletion in normal subjects to reproduce the glucose and insulin abnormalities of clinical hypokalemia. Diabetes. 1973;22: Bayliss J, Norell M, Canepa-Anson R, Sutton G, Poole-Wilson P. Untreated heart failure: Clinical and neuroendocrine effects of introducing diuretics. Br Heart J. 1987;57: Freel EM, Tsorlalis IK, Lewsey JD, Latini R, Maggioni AP, Solomon S, Pitt B, Connell JM, McMurray JJ. Aldosterone status associated with insulin resistance in patients with heart failure--data from the aloft study. Heart. 2009;95: Pfeffer MA, Swedberg K, Granger CB, Held P, McMurray JJ, Michelson EL, Olofsson B, Ostergren J, Yusuf S, Pocock S. Effects of candesartan on mortality and morbidity in patients with chronic heart failure: The charm-overall programme. Lancet. 2003;362: Bosch J, Yusuf S, Gerstein HC, Pogue J, Sheridan P, Dagenais G, Diaz R, Avezum A, Lanas F, Probstfield J, Fodor G, Holman RR. Effect of ramipril on the incidence of diabetes. N Engl J Med. 2006;355: Mancia G, Parati G, Borghi C, Ghironzi G, Andriani E, Marinelli L, Valentini M, Tessari F, Ambrosioni E. Hypertension prevalence, awareness, control and association with metabolic abnormalities in the san marino population: The smooth study. J Hypertens. 2006;24: Sarafidis PA, Bakris GL. Antihypertensive therapy and the risk of newonset diabetes. Diabetes Care. 2006;29: Chiasson JL, Josse RG, Gomis R, Hanefeld M, Karasik A, Laakso M. Acarbose for prevention of type 2 diabetes mellitus: The stop-niddm randomised trial. Lancet. 2002;359: Knowler WC, Barrett-Connor E, Fowler SE, Hamman RF, Lachin JM, Walker EA, Nathan DM. Reduction in the incidence of type 2 diabetes with lifestyle intervention or metformin. N Engl J Med. 2002;346: Witteles RM, Fowler MB. Insulin-resistant cardiomyopathy clinical evidence, mechanisms, and treatment options. J Am Coll Cardiol. 2008;51: Wong AK, AlZadjali MA, Choy AM, Lang CC. Insulin resistance: A potential new target for therapy in patients with heart failure. Cardiovasc Ther. 2008;26: Andersson C, Olesen JB, Hansen PR, Weeke P, Norgaard ML, Jorgensen CH, Lange T, Abildstrom SZ, Schramm TK, Vaag A, Kober L, Torp- Pedersen C, Gislason GH. Metformin treatment is associated with a low risk of mortality in diabetic patients with heart failure: A retrospective nationwide cohort study. Diabetologia. 2010;53: Eurich DT, McAlister FA, Blackburn DF, Majumdar SR, Tsuyuki RT, Varney J, Johnson JA. Benefits and harms of antidiabetic agents in patients with diabetes and heart failure: Systematic review. BMJ. 2007;335: Abildstrom SZ, Rasmussen S, Madsen M. Changes in hospitalization rate and mortality after acute myocardial infarction in denmark after diagnostic criteria and methods changed. Eur Heart J. 2005;26: Kumler T, Gislason GH, Kirk V, Bay M, Nielsen OW, Kober L, Torp- Pedersen C. Accuracy of a heart failure diagnosis in administrative registers. Eur J Heart Fail. 2008;10: Andersson C, Weeke P, Pecini R, Kjaergaard J, Hassager C, Kober L, Torp-Pedersen C. Long-term impact of diabetes in patients hospitalized with ischemic and non-ischemic heart failure. Scand Cardiovasc J De Groote P, Lamblin N, Mouquet F, Plichon D, McFadden E, Van Belle E, Bauters C. Impact of diabetes mellitus on long-term survival in patients with congestive heart failure. Eur Heart J. 2004;25: Gustafsson I, Brendorp B, Seibaek M, Burchardt H, Hildebrandt P, Kober L, Torp-Pedersen C. Influence of diabetes and diabetes-gender interaction on the risk of death in patients hospitalized with congestive heart failure. J Am Coll Cardiol. 2004;43: Glucose tolerance and cardiovascular mortality: Comparison of fasting and 2-hour diagnostic criteria. Arch Intern Med. 2001;161: The Relationship between Ischemic Heart Disease and Diabetes A population-based approach 31

33 103. Ryden L, Standl E, Bartnik M, Van den Berghe G, Betteridge J, de Boer MJ, Cosentino F, Jonsson B, Laakso M, Malmberg K, Priori S, Ostergren J, Tuomilehto J, Thrainsdottir I, Vanhorebeek I, Stramba-Badiale M, Lindgren P, Qiao Q, Priori SG, Blanc JJ, Budaj A, Camm J, Dean V, Deckers J, Dickstein K, Lekakis J, McGregor K, Metra M, Morais J, Osterspey A, Tamargo J, Zamorano JL, Deckers JW, Bertrand M, Charbonnel B, Erdmann E, Ferrannini E, Flyvbjerg A, Gohlke H, Juanatey JR, Graham I, Monteiro PF, Parhofer K, Pyorala K, Raz I, Schernthaner G, Volpe M, Wood D. Guidelines on diabetes, prediabetes, and cardiovascular diseases: Executive summary. The task force on diabetes and cardiovascular diseases of the european society of cardiology (esc) and of the european association for the study of diabetes (easd). Eur Heart J. 2007;28: Diabetes og hjertesygdom Deedwania PC, Fonseca VA. Diabetes, prediabetes, and cardiovascular risk: Shifting the paradigm. Am J Med. 2005;118: Glucose tolerance and mortality: Comparison of who and american diabetes association diagnostic criteria. The decode study group. European diabetes epidemiology group. Diabetes epidemiology: Collaborative analysis of diagnostic criteria in europe. Lancet. 1999;354: Coutinho M, Gerstein HC, Wang Y, Yusuf S. The relationship between glucose and incident cardiovascular events. A metaregression analysis of published data from 20 studies of 95,783 individuals followed for 12.4 years. Diabetes Care. 1999;22: Danaei G, Lawes CM, Vander Hoorn S, Murray CJ, Ezzati M. Global and regional mortality from ischaemic heart disease and stroke attributable to higher-than-optimum blood glucose concentration: Comparative risk assessment. Lancet. 2006;368: Gillett MJ. International expert committee report on the role of the a1c assay in the diagnosis of diabetes: Diabetes care 2009; 32(7): Clin Biochem Rev. 2009;30: WHO. Hba1c report

34 Appendices Appendix 1 Pathophysiology of diabetic vascular disease Diabetes alters function in multiple cell types, including endothelium, smooth muscle cells and platelets, indicating the extent of vascular disarray in this disease. The endothelium consists of a single layer of endothelium cells, providing a metabolically active interface between blood and tissue and modulates blood flow, nutrient delivery, thrombosis and leukocyte diapedesis. This cell layer synthesizes important bioactive substances including nitric oxide (NO), other reactive oxygen species, prostaglandins, endothelin, and angiotensin II that regulate vessel structure and function. NO potently dilates vessels and mediates much of the endothelium s control of vascular relaxation. Furthermore, it inhibits platelet activation and limits inflammation by reducing leukocyte adhesion to endothelium and migration into the vessel wall, and diminishes vascular smooth muscle cell proliferation and migration. Taken together, these properties inhibit atherogenesis and protect the blood vessel. Diabetes impairs endothelium-dependent NO-mediated vasodilation. Insulin resistance leads to excess liberation of free fatty acids (FFA) from adipose tissue that indirectly through other complex mechanisms results in impaired NO production and diminished bioavalability, thereby resulting in vasoconstriction, inflammation and thrombosis leading to accelerated atherogenesis. In smooth muscles diabetes alters the function in ways that promote atherosclerotic lesion formation, plaque instability and clinical events. Diabetes also impairs the platelet function and this circumstance participates significantly in thrombus formation. Elevated blood glucose levels lead to activation of protein kinase C, decreased production of platelet-derived NO. Through complex mechanisms diabetic abnormalities increase intrinsic platelet activation and decrease endogenous inhibitors of platelet activity. Moreover, in diabetes a tendency toward coagulation, coupled with impaired fibrinolysis, favors formation and persistence of thrombi. Thus the perception of type 2 diabetes encompasses a global metabolic disorder characterized by hyperlipidaemia, hypertension, and hypercoagulability in addition to hyperglycaemia. Each of these abnormalities plays an important role in cardiovascular disease development and progression and provides targets for therapy. arterial disease and ischaemic heart disease. It has been demonstrated that the current diagnostic threshold poorly reflect the risk of developing the above mentioned macrovascular complications, as the progressive relationship between glucose levels and cardiovascular risk extends far below the current diabetic threshold. 103, The diagnostic criteria on diabetes have previously differed throughout the world and the differences in diagnostic criteria play an important role in the diagnostic process in patients presenting with ischaemic heart disease. Therefore, the criteria will be clarified for. Table 1: Criteria for glucometabolic classification according to the WHO and ADA. Glucometabolic Source Classification criteria category Normal glucose regulation mmol/l WHO FPG < h PG < 7.8 ADA (1997) FPG < 6.1 ADA (2003) FPG < 5.6 Impaired fasting glucose (IFG) WHO FPG < 6.1 and < h PG < 7.8 ADA (1997) FPG < 6.1 and < 7.0 ADA (2003) FPG < 5.6 and < 7.0 Impaired glucose tolerance (IGT) WHO FPG < h PG 7.8 and 11.1 Diabetes Mellitus WHO FPG h PG 11.1 ADA (1997) FPG 7.0 ADA (2003) FPG 7.0 ADA (2010) HbA1c > 6,5 % WHO 2011 HbA1c > 6,5 % FPG= fasting plasma glucose. 2-h PG= two hour post load plasma glucose. ADA = American Diabetes Association. WHO = World Health Organization. Diagnostic criteria on diabetes The diagnosis of diabetes has traditionally been based on a fasting glucose level exceeding an arbitrary threshold, which has mainly been based on data from epidemiological studies and reflecting the risk of developing microvascular complications in the form of retinopathy and nephropathy.103 However, the majority of health related complications in patients with diabetes, is caused by macrovascular disease 105 e.g. cerebrovascular disease, peripheral The Relationship between Ischemic Heart Disease and Diabetes A population-based approach 33

35 34

36 The Relationship between Ischemic Heart Disease and Diabetes A population-based approach 35

37 36