Mortality and extent of coronary artery disease in 2776 patients with type 1 diabetes undergoing coronary angiography: A nationwide study

Abstract

Background

In a modern perspective there is limited information on mortality by affected coronary vessels assessed by coronary angiography in patients with type 1 diabetes. The aim of the present study was to characterise distribution of coronary artery disease and impact on long-term mortality in patients with type 1 diabetes undergoing coronary angiography.

Design

The design of this research was a nationwide population-based cohort study.

Methods

Individuals (n = 2776) with type 1 diabetes undergoing coronary angiography 2001–2013 included in the Swedish National Diabetes Registry and Swedish Coronary Angiography and Angioplasty Registry were followed for mortality until 31 December 2013 (mean 7.1 years). In 79% the indication was stable or acute coronary artery disease. Coronary artery disease was categorised into normal (21%), one- (23%), two- (18%), three- (29%) and left main-vessel disease (8%).

Results

Mean age was 57 years and 58% were male. Mean diabetes duration was 35 years, glycated haemoglobin was 67 mmol/mol and 44% had normal or one-vessel disease. In multivariate Cox proportional analyses hazard ratio for mortality compared with normal findings was 1.09 (95% confidence interval 0.80–1.48) for one, 1.43 (1.05–1.94) for two, 1.47 (1.10–1.96) for three and 1.90 (1.35–2.68) for left main-vessel disease. Renal failure 2.29 (1.77–2.96) and previous heart failure 1.76 (1.46–2.13) were highly associated with mortality. Standard mortality ratio the first year was 5.55 (4.65–6.56) and decreased to 2.80 (2.18–3.54) after five years.

Conclusions

In patients with type 1 diabetes referred for coronary angiography mortality is influenced by numbers of affected coronary vessels. The overall mortality rate was higher compared with the general population. These results support early intensive prevention of coronary artery disease in this population.

Introduction

Historically individuals with type 1 diabetes suffered mortality from the effects of insulin deficiency early after diabetes onset and did not survive until a first cardiovascular event. The introduction of insulin in 1923 and the increasing knowledge on the importance of glucose control in the later 20th-century, focusing on preventing microvascular complications, has changed this scenario and life expectancy continues to increase in type 1 diabetes individuals.^1–3 The improved survival in type 1 diabetes will also increase the life-risk for a cardiovascular event, as it does for the general population. The International Diabetes Federation (IDF) estimates that 542,000 children are living with type 1 diabetes and that 86,000 children develop type 1 diabetes yearly although onset can occur at any age.⁴ Cardiovascular disease (CVD) is the leading cause of death in patients with type 1 diabetes and contributes to substantial loss of life as well as to increased morbidity.^5,6 Elevated glucose levels, hypertension, dyslipidaemia and renal complications are well known risk factors contributing to the accelerated atherosclerosis and widespread coronary artery disease (CAD) seen in individuals with type 2 diabetes.^7,8 On the contrary, there is less research related to type 1 diabetes, especially when it comes to CAD where there is limited information on coronary angiography. This is partly explained because older reports may in this respect be hampered by the previous arbitrary definitions of diabetes into insulin and non-insulin dependent diabetes until the new classification of type 1 and type 2 diabetes in 1998.⁹ In these older reports CAD was evaluated by computer beam tomography as calcium score, by intima media thickness in the carotid artery, based on autopsy findings, or coronary angiography in small cohorts.^10,11 Furthermore, these studies were performed in the context of screening for CVD before kidney- or pancreas transplantation.^11–14 To further improve health and life expectancy in type 1 diabetes it is important to extend knowledge on mechanisms, patterns and outcomes of CVD. The aim of the present study was to characterise distribution of CAD and related mortality rates in a national cohort of patients with type 1 diabetes undergoing coronary angiography. Our hypothesis was that individuals with longstanding type 1 diabetes have advanced atherosclerosis and seldom normal coronary angiograms.

Material and methods

Patients

In all, 2776 consecutive patients undergoing coronary angiography during 2001–2013 registered in the Swedish Coronary Angiography and Angioplasty Registry (SCAAR) and in the Swedish National Diabetes Registry (NDR) with type 1 diabetes were included and followed for mortality until 31 December 2013. If repeated coronary angiography occurred during the study period only information from the first was collected. Mean follow-up time was 7.1 (standard deviation (SD) 3.2) years. Due to incomplete information regarding angiographic results 68 individuals (2%) were excluded. The SCAAR is a national quality registry that includes information on every patient who undergoes coronary angiography at the 29 centres in Sweden. The collected data have been validated annually since 2001 by patient hospital records and the entered variables with an overall correspondence rate in data of 95%. The NDR was initiated in 1996 as a tool for quality assurance in diabetes care and has previously been described in detail.¹⁵ About 384,000 patients were reported to the registry in 2015, approximately 10% of those being diagnosed as type 1 diabetes. The NDR is estimated to include more than 95% of all adult individuals, 18 years of age and above, with type 1 diabetes in Sweden. Baseline characteristics are derived from the SCAAR at the first angiography during the study period and from the NDR based on data obtained at the visit closest to the coronary angiography allowing 365 days before or after the coronary angiogram. The long-term follow-up was obtained by merging the SCAAR and the NDR database with the Swedish National Patient Register and the Swedish Cause of Death Registry containing information on underlying causes of death. The Inpatient Registry provides nationwide coverage of all inpatient admissions as of 1987, including information about the International Classification of Disease (ICD) code for comorbidities and cause-specific mortality. It is a reliable, validated alternative to revised certificates of discharge and death.¹⁶

Outcome measures and definitions

Type 1 diabetes was defined both with an epidemiological definition (onset age before 30 years and insulin treatment alone) and using the clinicians’ diagnosis of type 1 diabetes and onset age before 50 years. Glycated haemoglobin (HbA1c) is reported as the International Federation of Clinical Chemistry standard (IFCC; mmol/mol). Estimated glomerular filtration rate (eGFR) was calculated with the MDRD (Modification of Diet in Renal Disease) equation.¹⁷ Coronary vessel disease was visually judged by the local coronary interventionist and categorised into normal (atheromatosis/stenosis < 50%), one-, two-, three- and left main-vessel disease. Causes of mortality were classified according to the ICD-10 and categorised into eight main causes of mortality. Standard mortality ratio (SMR) was the comparison of the mortality in our cohort to the expected mortality in the general Swedish population of the same age.

Statistical analysis

Baseline characteristics are presented as the mean and SD for continuous variables and numbers and percentages for categorical variables. To handle missing data of the following eight variables (indication, body mass index (BMI), smoking, eGFR, microalbuminuria, HbA1c, treated hyperlipidaemia and hypertension) automatic multiple imputation was used where 10 datasets were created with the multiple-choice model (MCM) method. The comparison between patients by numbers of affected coronary vessels was performed by groups and not by pairwise comparison. The chi² or, when appropriate, Fisher’s exact test were used to compare baseline characteristics between the different groups. A Cox’s proportional-hazard regression model was computed to analyse the association between numbers of affected coronary vessels and mortality while adjusting for 19 variables (age, gender, hospital, admission year, indication, treated hypertension, treated hyperlipidaemia, previous myocardial infarction, previous heart failure, previous coronary artery by-pass grafting (CABG), previous stroke, previous renal failure, peripheral artery disease, retinopathy, albuminuria, diabetes duration, HbA1c, BMI and eGFR). Kaplan-Meier curves were computed to illustrate time to mortality by affected coronary vessels on the angiogram. SMR was calculated each year up to five years following the coronary angiography. Mortality estimates for the general Swedish population were obtained from the Swedish Statistical.¹⁸ A two-sided p-value of <0.05 was accepted as statistically significant. All analyses were conducted using the SPSS statistical program (SPSS version 23 software from SPSS Inc., Chicago, Illinois, USA.

Ethical consideration

All patients provided informed consent for their participation in the SCAAR and NDR. The local ethical board at University of Gothenburg approved the study.

Results

Patients

The baseline characteristics of the 2776 patients with type 1 diabetes (58% male) are presented in Table 1. Mean age was 57 years (SD 11), diabetes duration 35 years (SD 14, range 0–76) and HbA1c 8.3% (67 mmol/mol; SD 14). The indications for coronary angiography were stable CAD (31%), non-ST-elevation acute coronary syndrome (NSTE-ACS) (38%) and ST-elevation myocardial infarction (STEMI) (10%), heart failure (4%), atypical chest pain (5%), silent ischaemia (3%) and other rare reasons. Coronary angiography revealed 21% without significant stenosis, 23% with one-vessel, 18% with two-vessel and 29% with three-vessel disease.

Patients with three-vessel- compared to one-vessel disease were older (60 vs 56 years), had longer diabetes duration (39 vs 33 years) and more cardiovascular comorbidity with hypertension, treated hyperlipidaemia, previous myocardial infarction, heart failure, and peripheral artery disease while there was no difference in HbA1c. Patients with three-vessel disease had more frequent signs of microvascular disease such as albuminuria and retinopathy. Complete revascularisation after percutaneous coronary intervention (PCI) was significantly less common in those with three compared to one vessel disease. CABG was the primary decision (44%) after angiography in three-vessel disease.

Figure 1 illustrates angiography findings in patients with stable angina or acute coronary syndromes (NSTE-ACS and STEMI). Among those with stable CAD, 15% had a normal angiography and 21% had one-vessel disease. The corresponding figures for NSTE-ACS were 13% and 23% and for STEMI 4% and 40%, respectively.

Long-term follow-up

In all, 640 patients with type 1 diabetes died (23%; 23% of all men and 23% of all women) during the study follow-up. Kaplan-Meier curves for time to all-cause mortality by affected coronary vessels (Figure 2(a)) demonstrate lower rates in those with normal and one-vessel disease. Patients with two-vessel disease had higher and similar mortality rates compared to those with three-vessel disease during the first five years following angiography. A similar pattern was seen for cardiovascular mortality although at a lower rate (Figure 2(b)). In multivariate Cox proportional hazard regression analysis the risk for mortality was compared to normal findings after adjustments for traditional cardiovascular risk factors as well as indication for coronary angiography (Supplementary Material, Table 1). Except for one-vessel disease, mortality was increased by numbers of affected coronary artery vessels; hazard ratio (HR) 1.07 (95% CI 0.79–1.46) for one-, 1.43 (1.05–1.94) for two-, 1.47 (1.10–1.95) for three- and 1.92 (1.37–2.71) for left main-vessel disease. Variables associated with mortality were renal failure (2.29; 1.77–2.96), STEMI (1.85; 1.35–2.55), previous heart failure (1.76; 1.46–2.13), other indications for coronary angiography (1.68; 1.29–2.18), age (10 years; 1.52, 1.37–1.67), NSTE-ACS (1.41; 1.16–1.73), diabetes duration (10 years; 1.09; 1.01–1.17) and HbA1c (1.01; 1.002–1.02) (Supplementary Material, Table 1). Causes of death are depicted in Figure 3. The more numbers of affected coronary artery vessels, the higher proportion of deaths from ischaemic heart disease. Patients with three- and left main vessel disease had mortality cause from ischaemic heart disease (45–55%) and from endocrine disorders (20–30%). The distribution of age at the time of death from any cause in the different groups was 59 ± 13 years for normal angiography, 64 ± 11 for one-vessel disease, 64 ± 10 for two-vessel disease, 66 ± 10 for three-vessel disease and 69 ± 10 for left main stem disease. The SMR of observed deaths in the study group compared to expected deaths in the general Swedish population for the first year after the index coronary angiography was 5.55 (95% CI 4.65–6.56), second year 4.79 (95% CI 3.96–5.74), third year 3.56 (95% CI 2.86–4.38), fourth year 3.04 (95% CI 2.40–3.81) and fifth year 2.80 (95% CI 2.18–3.54) (Supplementary Material, Figure 1).

Discussion

This nationwide study describes the characteristics of a large cohort of type 1 diabetes patients in a clinical setting undergoing coronary angiography in Sweden 2001–2013. Of the 2776 patients assessed, more than 80% were referred for symptoms or clinical features of CAD. However the coronary angiogram was normal (21%) more often than expected or showed only one-vessel (23%) affected despite the symptoms of CAD and diabetes duration of about 30 years. There are a limited amount of studies to compare our results with, since other studies involving type 1 diabetes patients undergoing coronary angiography are scarce and with few patients included.^13,14 Existing studies are often done in the context of screening patients with CAD before kidney and pancreas transplantation,^11,12 and it is well known from earlier research that diabetes nephropathy comes with the highest risk of CVD and raised mortality in type 1 diabetes patients.^19,20 In 2000, Pajunen et al. published a retrospective study of 64 patients with a clinical indication for coronary angiography where 84% had three-vessel disease or left main disease which is a much higher proportion than in the present study and probably explained by the selected high-risk population.¹³ The higher proportion of coronary artery vessel disease and CVD risk in older studies in type 1 diabetes might also reflect a period of less intensive glycaemic control compared to the situation today.

Even if patient care of type 1 diabetes has improved during the last decades, patients are still at high risk for developing CAD and CVD and remain the main reason for a shorter longevity in type 1 diabetes patients.^21,22 Even recently published studies, have shown that individuals with type 1 diabetes are at risk for a cardiovascular event approximately a decade before the general population have their first events.^23,24 In the present study with a more recent population from clinical practice there was an increase in mortality by numbers of affected coronary vessels. Compared to patients with normal angiography, patients with two-vessel and three-vessel disease had an increased risk for mortality of 43% and 47% respectively. A finding of left main disease almost doubled the mortality during the follow-up time. The majority of death causes among patients with three-vessel and left main disease were from ischaemic heart disease while in patients with normal and one-vessel disease the most common cause of death was from endocrine and metabolic disorders. The strongest association for mortality in our cohort was linked to renal insufficiency consistent with earlier research on CAD in type 1 diabetes individuals.^20,25 Looking at the whole cohort of individuals with type 1 diabetes being referred for coronary angiography, mortality was more than five times higher than for the age-matched general population during the first year after the angiography and still almost three-fold five years thereafter. This was the case despite 70% of those with two- and three-vessel and left main vessel disease having undergone revascularisation, either by PCI or CABG. There is a lack of research regarding the benefits of coronary revascularisation in type 1 diabetes patients but there are at present no reasons to believe that the results would differ much from results in type 2 diabetes where revascularisation has led to improved survival rates.^26,27 Apart from microvascular complications, heart failure and a history of CVD also showed a strong association to mortality during the follow-up period, and this supports all attempts to avoid these complications in patients with type 1 diabetes. Modifiable risk factors associated with mortality were smoking and HbA1c. The importance of glucose control for cardiovascular outcome in patients with type 1 diabetes has been highlighted in the Diabetes Control and Complications Trial (DCCT)/ Epidemiology of Diabetes Interventions and Complications (EDIC) study where tight glycaemic control early in life affected risk of CVD 17 years after and mortality 20 years after start of intensive treatment.²⁸ Modern diabetes care with new types of insulin and technical support, i.e. insulin pumps and continuous glucose monitoring systems, will help individuals with type 1 diabetes achieve even better glycaemic control. A recent study from the NDR have shown that insulin pump therapy is associated with lower cardiovascular mortality compared to treatment with multiple daily injections.²⁹ Equally important is, of course, treating other known risk factors such as dyslipidaemia to even further reduce cardiovascular risk in individuals with type 1 diabetes.³⁰

Strengths and limitations

The strength of the current study is the population-based design where nearly all patients with type 1 diabetes in Sweden were included from everyday clinical practice and with prospective collected detailed information on angiographic findings as well as diabetes-related variables. Also the NDR has good coverage regarding diabetes-related and cardiovascular risk factors. However, there are also limitations to this study, being an observational study there is the possibility of unknown residual confounders that we have not controlled for. Atheromatosis/stenosis <50% was defined as a normal angiogram but may also comprise undiagnosed microvascular disease.

Conclusion

This nationwide study of type 1 diabetes individuals in Sweden referred for coronary angiography shows that the coronary angiogram was normal or with only one-vessel affected in 44% of the population despite lengthy diabetes duration. Mortality was increased by numbers of affected coronary artery vessels and presence of renal failure and cardiovascular co-morbidities. Compared to the general population, the mortality rate is still increased by three to five times in a cohort referred for angiography. These results support a holistic approach with early multifactorial intensive prevention of CAD being important in type 1 diabetes.

Author contribution

All authors have made substantial contributions to this paper. All authors took part in the interpretation of the results. AN, BL, NS, A-MS and KE-O developed the study design. BL managed the database and performed the statistical analyses. VR and AN finalised the manuscript after adjustments by all authors.

Declaration of conflicting interests

The author(s) declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: AN has received honoraria on expert group participation from Astra Zeneca, MSD, Eli Lilly and Boehringer Ingelheim. VR has received an honorarium from Astra Zeneca on expert group participation. BL, NS, A-MS, CH and KE-O have no conflicts of interest to declare in relation to this manuscript.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the Swedish Heart-Lung Foundation, Department of Research and Development Region Kronoberg, the Swedish Diabetes Association and the Kamprad Family Foundation.

References