Global Nature of Incipient Chamber Remodeling and Dysfunction in Diabetic Individuals Living in the Community

See Article by Jensen

Diabetes mellitus is a major risk factor for the development of heart failure, even in asymptomatic populations with no previous history of cardiovascular diseases.^1–3 The principal mechanisms are multiple and include among others (1) epicardial coronary artery atherosclerosis leading to infarction, chronic ischemia, and fibrosis⁴; (2) microvascular dysfunction in the absence of obstructive epicardial atherosclerosis occurring in isolation but more frequently in association with retinal and kidney microvascular disease, interstitial fibrosis, and possibly myocardial scar⁵; (3) myocyte overload associated with exhausted hypertrophy caused by longstanding hypertension, obesity, and dysmetabolism⁶; (4) direct myocyte dysfunction with altered fuel utilization, intracellular lipid accumulation, and increased generation of advanced glycated end products and reactive oxygen species.⁷ In population studies, even when overt coronary artery disease is excluded, most or all the above mechanisms are at work, explaining why diabetes mellitus, obesity, and inflammation are such strong risk factors for heart failure and mortality.⁸

The strong association between myocardial dysfunction and heart failure with diabetes has led to the concept of a diabetic cardiomyopathy,⁹ in which systolic and diastolic dysfunction appear to occur concomitantly, particularly when more sophisticated tools for assessment of myocardial function are employed.¹⁰ The latter observation suggests that while ischemia is long recognized as central to cardiovascular disease in patients with diabetes mellitus,¹¹ particularly type II diabetes mellitus, other mechanisms as listed above may share center stage with ischemia in diabetic patients who develop heart failure with reduced but especially preserved ejection fraction (EF).^9,10 Given the importance of heart failure as a public health problem,^1–3,8 and the fact that heart failure with preserved EF comprises at least half of the population of patients with heart failure, the need to understand the tempo, type, and magnitude of cardiac abnormalities in association with diabetes mellitus cannot be underestimated.

Along this much needed investigative line, Jensen et al¹² demonstrate, in this issue of Circulation: Cardiovascular Imaging, that the early remodeling alterations known to affect the left ventricle (LV) and left atrium in diabetic individuals can also be detected in the right ventricle and right atrium. Such global alterations included a reduction in both left and right end-diastolic volume, reduced maximal and minimal left and right atrial volumes, as well as left and right atrial emptying fraction. Moreover, they confirmed previous observations that myocardial strain is reduced in diabetic UK Biobank participants, while the LV EF differences were of borderline significance, reflecting incipient myocardial systolic dysfunction. These observations suggest that the mechanisms influencing myocardial tissue structure and function in association with diabetes mellitus are global, as opposed to localized to the left heart as discussed by the authors. The work is also important for making such demonstration in a subpopulation of the UK Biobank without history of clinical atherosclerotic coronary or cerebral artery disease. In addition, the investigators used magnetic resonance imaging (MRI) with tissue tagging, probably the most accurate tools that can be applied to population studies for such specific purposes.¹³

Population-scientific studies designed to study risk factor exposures and incipient subclinical disease processes are well known for small effect sizes.^1–3,8,13 The study by Jensen et al is no exception and prompted the authors to perform sensitivity analyses using propensity scores. The latter were useful to confirm the general magnitude of regression coefficients found in corresponding adjusted and unadjusted analyses. The authors acknowledge that not only the population included in the UK Biobank study is healthier than the British general population but the subpopulation selected for inclusion in this specific analysis was even healthier, by the exclusion of participants with history of atherosclerotic disease (n=190) and those with LV EF <50% (n=250). Exclusion of participants with limited imaging studies also tends to select healthier populations that in the case of MRI include individuals who can lie flat for 30 to 60 minutes, hold their breath multiple times, have less extreme obesity, and are not claustrophobic. The absence of contrast-enhanced imaging, while limiting for excluding myocardial scar assessment and postcontrast T1 mapping, is also an advantage for not limiting participation of diabetic individuals with impaired renal function, which has limited the ability to measure replacement fibrosis in diabetic patients before.¹⁴ The cross-sectional study design also represents a limitation of the study by Jensen et al¹² by restricting conclusions to cross-sectional associations, preventing causal inferences, as well as the exploration of longitudinal relationships.

The important findings by Jensen et al,¹² as proposed by the authors, suggest that interstitial fibrosis plays a central role in early myocardial remodeling, particularly at the level of the left and right ventricles. LV end-diastolic volume reduction, measured by MRI, is directly associated with aging, as reported in the journal 10 years ago.¹⁵ Later, longitudinal observations in MESA (Multi-Ethnic Study of Atherosclerosis) in both women and men 10 years after the MRI initial observations confirmed the association of aging with reduction in LV end-diastolic volume, increase in concentric remodeling (increase in LV mass/volume ratio) and reduction in stroke volume with EF preservation.¹⁶ The results from Jensen et al, in analyses adjusted for age and sex, suggest that the diabetic LV remodeling type is similar to that seen in association with aging, possibly reflecting accelerated aging, mediated not only by dysmetabolism but also by inflammation known to accompany both diabetes mellitus and aging.¹² The authors’ hypothesis that this cardiac chamber shrinkage in association with diabetes mellitus is secondary to increased interstitial fibrosis has also been proposed as an explanation for the similar scenario seen during aging, where fibrosis increases particularly in men after the fifth decade of life¹⁷ and is associated with both systolic and diastolic dysfunction.¹⁸ Indeed, in further analyses of the MESA data, Ambale et al¹⁹ demonstrated that the effect of interstitial fibrosis enhancement on LV end-diastolic volume and stroke volume reduction occurs over and above that of aging. Finally, and in direct support to the findings by Jensen et al,¹² Wong et al²⁰ have documented a direct relationship between increased interstitial fibrosis and mortality and incident heart failure hospitalization in patients with diabetes mellitus.

The authors also report an important cross-sectional association between diabetes mellitus and myocardial dysfunction measured by MRI tissue tagging as increased myocardial circumferential strain (reduced circumferential shortening). Other cross-sectional analyses from MESA and other studies have also documented a reduction in myocardial circumferential shortening and sharp increase in LV torque and enhanced or maintained EF with aging.^13,21 However, in contradistinction with aging, LV EF among diabetics in the UK Biobank was preserved despite greater torque, thought to be the compensatory mechanism to reduced circumferential shortening.^13,21 Previous longitudinal studies using speckle tracking and tissue Doppler echocardiography in CARDIA (Coronary Artery Risk Development in Young Adults) showed that exposure to dysmetabolism and dysglycemia is associated with both systolic and diastolic dysfunction.²² At the tissue level, diastolic abnormalities are linked to greater myocardial stiffness due to expanded extracellular space caused by the accumulation of collagen and possibly other alterations of the extracellular matrix.^{6,7,9,18,20,22} Conversely, myocardial systolic dysfunction is commonly associated with contractile apparatus impairment and in the case of chamber dysfunction, possibly also associated with myocardial loss and fibrosis replacement. Only extreme myocardial matrix alterations such as in infiltrative diseases like amyloidosis or hemochromatosis are thought to cause myocyte systolic restrain to the point of reducing shortening. Diabetes mellitus is generally not considered to be one of those conditions. Intracellular dysmetabolism with direct myocyte contractile reduction has been invoked to explain myocardial systolic dysfunction in diabetes mellitus^6,7,9 in combination with or over and above ischemia. As discussed above, this be secondary to multiple mechanisms in patients with diabetes mellitus.^4,5,9

The work by Jensen et al¹² suggests that interstitial space alterations leading to end-diastolic volume reduction occur not only in the LV and left atrium but also in the right ventricle and right atrium. Direct measures of extracellular volume by T1 mapping from atrial myocardium have been fraught with technical challenges largely because of atrial wall thinness, impairing our ability to attribute atrial chamber size reductions to directly measured increased interstitial fibrosis in population studies. Similar challenges have limited the measurement of right ventricular wall interstitial fibrosis in large population studies. However, the similarity of changes and associations with risk factor exposures suggest that the mechanisms are similar throughout myocardium in the entire heart. Moreover, the absence of contrast-enhanced studies limits our ability to assess the contribution of replacement fibrosis to chamber dysfunction in the study by Jensen et cols.¹² Diabetics are known to be more prone to have unrecognized myocardial infarction and may also be more prone to develop myocardial scars of the nonischemic late gadolinium enhancement pattern. Indeed, nonischemic type myocardial scars that have been documented by MRI in patients with aortic stenosis,²³ hypertrophic²⁴ and dilated²⁵ cardiomyopathies, among several other conditions, are also seen in diabetic cohorts.¹⁴ Although it is difficult to completely exclude the role of replacement fibrosis in reducing global circumferential myocardial shortening in the LV, the focal nature of such scars does not appear to completely explain or represent the main mechanism of ventricular dysfunction in diabetic participants of large prospective studies. In the study by Jensen et al,¹² for example, LV EF was basically similar in diabetics and nondiabetics (borderline statistical significance), and right ventricular EF was also similar in both groups. In other cohort studies, like the CARDIA prospective study, LV EF differences were also only slight or absent, whereas myocardial strain differences were clear and highly significant in relation to dysglycemic exposure.²²

Atrial functional changes are more complex, however. While differences in passive conduit function can be explained by stiffer atrial myocardium as in the ventricles, in the atria, such differences tend to be compensated by enhanced atrial active function (atrial kick). Notably, in the work of Jensen et al,¹² total emptying fraction was reduced, reflecting overall reduced atrial performance (data on the different components of atrial function were not provided). Because replacement fibrosis has been documented in the left atrium of diabetics by MRI late gadolinium enhancement,²⁶ the contribution of this mechanism to overall reduction in both left and right (by contiguity) atrial emptying fractions cannot be excluded. In patients who received left atrial ablation, left atrium volumes are reduced and emptying fraction decreased, in association with therapeutically induced replacement fibrosis.²⁷ Therefore, replacement fibrosis in association with diabetes mellitus cannot be ruled out as a main mechanism of atrial volume reduction and dysfunction in the work by Jensen et al.¹² This represents a crucial area for future study as large population studies have demonstrated that both left and right atrial volume and function are independently associated with incident cardiovascular events such as atrial fibrillation, heat failure, and stroke.^28–30 Future studies using MRI late gadolinium enhancement could potentially illuminate this important issue, despite the technical challenges associated with measuring replacement fibrosis in the atria.³¹

Finally, while the motives behind the exclusion of patients with atherosclerotic diseases (prior clinical myocardial infarction, angina pectoris, or stroke) and those with reduced EF were justifiable given the focus of the analysis on early alterations (cardiac chamber remodeling and incipient dysfunction) in association with diabetes mellitus mellitus, future efforts should include a more comprehensive approach that would allow comparative evaluation of the potential role of macrovascular and microvascular disease versus other potential mechanisms of diabetes mellitus–induced cardiac remodeling and dysfunction. Future studies including more detailed phenotyping of vascular disease in association with atherosclerosis and the myocardial consequences of diabetes mellitus could further illuminate the role played by hyperglycemia, inflammation, and other potential mechanisms of cardiovascular disease in association with obesity and type I and type II diabetes mellitus. The work by Jensen et al¹² represents an important step in the quest for a more thorough understanding of the cardiac consequences of this ensemble of dysmetabolic pathologies.

Disclosures

None.