Excess Cardiovascular Risk in Type 1 Diabetes Mellitus

Article, see p 730

Even at <40 years of age, men and women with type 1 diabetes mellitus have a respective 5- and 10-fold relative risk of experiencing a coronary heart disease (CHD) event.¹ The risk is still higher with a younger age of onset.² By the age of 20 years, life expectancy is reduced by ≈12 years in absolute terms, with approximately one-third of the excess risk being attributable to cardiovascular disease (CVD).³ Although rates of CHD and CVD in type 1 diabetes mellitus may be declining faster than in the general population,⁴ the condition was still more strongly linked to CVD outcomes than any other disease (including type 2 diabetes mellitus) in a recent analysis of >10 million individuals in the UK National Health Service.⁵ It was also linked to a higher incidence of heart failure than type 2 diabetes mellitus across most of the life course in a contemporary population-based analysis in Scotland.⁶

Blood glucose as assessed by hemoglobin A1c is generally considered the most powerful risk factor for CVD in type 1 diabetes mellitus, with other traditional CVD risk factors (blood pressure, low-density lipoprotein cholesterol) having a demonstrable independent contribution only after 15 to 20 years from diagnosis.⁷ Oxidative stress is commonly postulated to be a key mechanism by which hyperglycemia causes vascular damage with the initial mitochondrial generation of excess superoxide anions and subsequent (1) activation of deleterious biochemical pathways (advanced glycation end products, protein kinase C, hexosamine, and polyol pathways) and (2) inhibition of protective pathways (endothelial nitric oxide synthase and prostacyclin synthase).⁸

Although pathways linking hyperglycemia with CVD have been described for some years, it has long been postulated that type 1 diabetes mellitus may carry a unique and specific risk. Direct comparisons with type 2 diabetes mellitus are problematic because of the different ages of onset, but, in this condition, elevated blood glucose is a much less consistent and modifiable CVD risk factor, and only a few agents used to lower it have been proven to reduce the rates of events. In this issue of Circulation, a study by Sousa et al⁹ from the Joslin Clinic may help explain the strong link between hyperglycemia and CVD in type 1 diabetes mellitus by revealing a potential novel pathway that may have been hiding in plain sight.

As the immune system develops, the thymus plays a key role in preventing autoimmunity by removing self-reactive CD4⁺ T cells from the circulation. The author group of the current article has previously described autoimmunity to the α-myosin heavy chain in a mouse model of myocarditis and peripheral blood enrichment of autoreactive cardiac myosin-specific CD4+ T cells in healthy humans.¹⁰ More recently, they have reported a sustained proinflammatory CD4+ T-cell (and autoantibody) response in mice with an experimental form of type 1 diabetes mellitus (nonobese diabetic) following myocardial injury, and the presence of cardiac myosin autoantibodies in serum samples from >80% of people with type 1 diabetes mellitus after a myocardial infarction.¹¹ On the basis of this work, they generated the hypothesis that subclinical myocardial injury resulting from hyperglycemia might be amplified by a dysfunctional immune response specifically in type 1 (but not type 2) diabetes mellitus.

To test their hypothesis, Sousa et al secured access (via the National Institute of Diabetes and Digestive and Kidney Diseases repository) to serial blood samples from young individuals with type 1 diabetes mellitus previously randomly assigned to intensive or conventional glucose control for 6.5 years (in the landmark DCCT study [Diabetes Control and Complications Trial]) and then followed up postrandomization for up to 20 years (EDIC study [Epidemiology of Diabetes Interventions and Complications]).¹² Because peripheral blood mononuclear cells were not available for measuring specific CD4⁺ T-cell populations, they instead measured levels of 5 cardiac muscle–specific autoantibodies in stored serum: 3 related to myosin heavy chain 6, 1 related to myosin heavy chain 7, and 1 related to cardiac troponin. In this way, they were able to track the development of an immune response to cardiac muscle by measuring the rates of autoantibody positivity over time in 2 groups of young adults with type 1 diabetes mellitus well-matched for demographic characteristics (including human leukocyte antigen genotype) and in whom blood glucose was either controlled to target (hemoglobin A1c ≤7.0%) or consistently well above target (hemoglobin A1c ≥9.0%). Because renal microvascular disease is a frequent early herald of future CVD, they excluded not only participants with evidence of complications at baseline, but also those who developed renal or CVD complications during the randomized phase (DCCT).

The first striking result was that the proportion of individuals becoming positive for ≥2 positive cardiac autoantibodies (of a potential 5) during DCCT was markedly higher in the high hemoglobin A1c group (22% versus 1%), a difference that continued to increase over time. Second, almost all those with high antibody levels subsequently developed computed tomography evidence of coronary artery calcification, whereas this was the case for scarcely any of those with low levels. Third, although only 6 CVD events occurred over ≈25 years of follow-up in this young, low-risk cohort, 4 occurred in those with the highest levels of cardiac autoantibodies, and the only fatal event occurred in the only participant to test positive for all 5. Fourth, high-sensitive C-reactive protein levels were elevated ≈3-fold in those with positive cardiac autoantibodies, suggesting, but not proving, subclinical inflammation as a pathway linking specific myocardial inflammation with a more generalized accelerated atherosclerosis. The specificity of this apparently dysfunctional immune response for type 1 diabetes mellitus was demonstrated by low rates of autoantibody positivity in samples from individuals with a 12-year history of type 2 diabetes mellitus and a comparable history of hyperglycemia from the Joslin Heart Study.

Taken together, the results of this compelling series of experiments suggest that impaired central immune tolerance might play a role not only in the development of type 1 diabetes mellitus itself (via a self-reactive response to β-cells), but also in accelerating its cardiovascular complications (via a self-reactive response to α-myosin), particularly in individuals in whom long-term glucose levels are well above target.

Overt autoimmune myocarditis is not as frequent a clinical complication of type 1 diabetes mellitus as might be predicted from these research findings, although a mechanistic link with more frequently observed atherosclerotic CHD (an inflammatory state) might be provided by a secondary cytokine response to myocardial inflammation.¹³ In contrast, myocarditis is a classical manifestation of Chagas disease (American trypanosomiasis), which is characterized by autoimmunity to cardiac myosin triggered by chronic infestation with the protozoal parasite Trypanosoma cruzi.¹⁴ The investigators were able to measure the same panel of antibodies in serum from a Chagas disease cohort; intriguingly, autoantibody levels were similar to those seen in type 1 diabetes mellitus with sustained hyperglycemia. The radioimmunoprecipitation assays used were raised to complementary DNA for isoforms of human α- and β-myosin heavy chain expressed extensively in human ventricles. However, to account for a greater preponderance of atherosclerotic CHD over heart failure events in type 1 diabetes mellitus in comparison with Chagas disease, it is tempting to speculate that these cardiac autoantibodies might at least in some circumstances recognize forms of myosin α-heavy chain expressed in the walls of medium- or large-sized arteries.

The studies by Sousa et al are elegant and represent a credible additional mechanism for accelerated CHD risk in type 1 diabetes mellitus. However, there are a number of important limitations. Longitudinal association is suggestive of causality but is rarely conclusive in its own right; the small numbers of CVD events that occurred in low-risk individuals from the DCCT cohort resulted in very wide CIs in the relative-risk estimates; and samples were not available from the baseline DCCT visit. Nevertheless, the story that emerges is intriguing, urgently requiring validation. Given the postulated link between cardiac antibody status and myocarditis, it will be important for future studies to focus on heart failure (and not just CHD) outcomes. Because there were only 15 heart failure events (4 first events) over 30 years of follow-up in DCCT/EDIC, samples from other cohorts will be required. Unfortunately, few type 1 diabetes mellitus cohorts exist in which (1) appropriate samples have been biobanked (particularly for the measurement of peripheral mononuclear cells) and (2) there has been adequate event ascertainment and duration of follow-up.

With any new biomarker, initial reports can overestimate the strength of association, so an additional element of caution is required. It would be premature to suggest that individuals with cardiac autoantibodies should be targeted for specific antibody therapy. That said, if antimyosin cardiac antibody positivity is ultimately validated as a reliable marker and adds meaningful prediction of CHD and heart failure risk in type 1 diabetes mellitus, the present findings may extend beyond recognition of a dysfunctional immunologic mechanism toward eventual inclusion in risk stratification of individuals with type 1 diabetes mellitus. Identifying truly highest-risk individuals to target for more intensive glucose lowering (eg, insulin pump, continuous glucose monitoring, and closed loop systems), earlier and more aggressive statin therapy, and other cardioprotective adjunct therapy could help reduce the burden of CVD and contribute to improving the outcomes in type 1 diabetes mellitus.

• Download figure
• Download PowerPoint

Footnotes