Summary
Diabetes mellitus has reached an epidemic level worldwide. Cardiovascular diseases are the primary cause of death in diabetic patients, not only because of coronary artery disease and associated hypertension but also because of a direct adverse effect of diabetes on the heart, independent of other potential aetiological factors. However, the existence of this ‘diabetic cardiomyopathy’ remains controversial. We aimed to review current evidence for the existence of diabetic cardiomyopathy, focusing particularly on the clinical setting.
Résumé
Le diabète a atteint un niveau épidémique sur le plan mondial. Les pathologies cardiovasculaires représentent la première cause de décès chez les patients atteints de diabète non seulement du fait des cardiopathies ischémiques et de l’hypertension artérielle associées au diabète mais également du fait d’un effet délétère du diabète lui-même sur le plan cardiaque indépendamment de tout autre facteur étiologique potentiel. Cependant, l’existence de cette « cardiomyopathie diabétique » reste controversée. Le but de cette revue de la littérature est de discuter des preuves de l’existence d’une cardiomyopathie diabétique en s’intéressant principalement au champ clinique.
Background
A tsunami of obesity has been recently described, with 1.46 billion overweight adults (BMI ≥ 25 kg/m 2 ) in 2008, including 500 million who were obese (BMI ≥ 30 kg/m 2 ) . Consequently, diabetes mellitus has reached an epidemic level worldwide, with a prevalence of 4% in 1995 and an anticipated prevalence of 5.4% in 2025, corresponding to 300 million adults with diabetes worldwide . Cardiovascular diseases represent the primary cause of death in this population because of coronary artery disease or associated hypertension but also because of a direct adverse effect of diabetes mellitus on the heart, called DCM. Diabetes mellitus is a complex metabolic disorder that includes insulin resistance (in type 2), often associated with hypertension and obesity . All those conditions are associated with a high risk of coronary artery disease but may also exert a direct negative effect on the myocardium. Among them, the specific negative effect of diabetes mellitus on the heart, leading to DCM, has been extensively investigated in the last three decades. Controversies exist regarding the existence of a specific DCM, as this concept has mainly emerged from experimental models . In this review, we aimed to discuss the current evidence for the existence of DCM, focusing particularly on the clinical setting.
Definition of diabetic cardiomyopathy (DCM)
DCM was first described 40 years ago, based on post-mortem observations in four diabetic patients with heart failure but no coronary artery disease or other aetiological conditions explaining heart failure . The authors observed LV hypertrophy associated with myocardial fibrosis in those patients and introduced for the first time the concept of DCM . Five years later, Regan et al. confirmed those results and described cases of heart failure in familial diabetes without coronary artery disease, hypertension or obesity. A higher myocardial collagen and lipid content was also reported in those cases compared with control subjects .
Since those initial studies, DCM has been defined as the existence of LV dysfunction in diabetic patients without coronary artery disease, hypertension or other potential aetiological condition.
Experimental evidence
Numerous experimental studies, mainly based on rodent models, have demonstrated a direct negative effect of diabetes mellitus on the myocardium . Pathophysiological mechanisms include metabolic alterations with impaired calcium homeostasis, alteration of substrate utilization (increase in lipid use and decrease in glucose oxidation), lipotoxicity, glucotoxicity with intervention of advanced glycation end products, mitochondrial dysfunction, increase in oxidative stress, renin-angiotensin-aldosterone system activation and cardiac dysautonomia. All these mechanisms lead to an increase in myocardial cellular death (necrosis and apoptosis) and to myocardial fibrosis, with the consequent development of myocardial dysfunction and overt heart failure ( Fig. 1 ) .
Epidemiological evidence
In the clinical setting, the most convincing evidence for the existence of DCM comes from large epidemiological studies. Two years after the study of Rubler et al. , the Framingham Heart Study investigators demonstrated that diabetes mellitus was an independent risk factor for heart failure . Risk of heart failure was 2.4-fold and 5-fold higher in diabetic men and women, respectively, than in non-diabetic subjects. The increased incidence of heart failure in diabetic patients persisted even after adjustment for age, hypertension, obesity, coronary artery disease or dyslipidaemia .
Diabetes mellitus as an independent risk factor for heart failure has been confirmed in numerous epidemiological studies . In a prospective study including 2700 elderly subjects (mean age 81 ± 9 years), the incidence of heart failure during a 43-month follow-up was 1.3-fold higher in diabetes mellitus than in euglycaemic subjects after adjustment for age, hypertension, coronary artery disease and sex . The Cardiovascular Health Study, including 5888 subjects aged more than 65 years with a mean delay of 5.5 years follow-up, reported an incidence rate for heart failure of 19.3/1000 person-years, with a substantially higher incidence (approximately 2-fold) associated with diabetes mellitus . In a large case-control study, Bertoni et al. tested the hypothesis that diabetes mellitus was independently associated with idiopathic cardiomyopathy. After adjusting for age, sex, race and hypertension, diabetes mellitus was significantly associated with idiopathic cardiomyopathy (relative odds 1.58, 95% CI 1.55–1.62). Those results were confirmed in a second case-control study . In the French EPICAL Study, prevalence of diabetes mellitus among patients with non-ischaemic advanced heart failure was 19.7% . In the Reykjavik Study, Thrainsdottir et al. explored the associations between heart failure and abnormal glucose regulation (impaired glucose tolerance or impaired fasting glucose) or type 2 diabetes mellitus in a population-based cohort of 19,381 participants. The odds ratio was 2.8 (95% CI 2.2–3.6) for the association between type 2 diabetes mellitus and heart failure and 1.7 (95% CI 1.4–2.1) for the association between abnormal glucose regulation and heart failure.
Finally, the Strong Heart Study recently confirmed diabetes mellitus as an independent risk factor of heart failure . In a population-based cohort of 1204 subjects, the authors showed a 1.5-fold higher risk of heart failure in patients with diabetes mellitus after adjustment for multiple cofactors (age, sex, obesity, central fat distribution, antihypertensive medications, atrial fibrillation, urinary albumin/creatinine ratio, plasma cholesterol, Hb1Ac, smoking habit, alcohol use, educational level and physical activity). Interestingly, for the first time, intercurrent myocardial infarction was censored in this analysis .
Noninvasive evidence for adverse effects of diabetes mellitus on the heart (left ventricular structure, remodelling and function)
Noninvasive evidence for structural myocardial abnormalities associated with diabetes
Myocardial fibrosis
Myocardial fibrosis, as initially described by Rubler et al. and confirmed in both histological studies in humans and experimental studies , is a major consequence of the adverse effects of diabetes mellitus on the heart . Backscatter is an ultrasound tissue characterization technique based on the measurement of myocardial tissue echoreflectivity, which is related to myocardial collagen content . Di Bello et al. showed an increase in myocardial echodensity as assessed by the integrated backscatter index in 26 insulin-dependent diabetic normotensive patients compared with 17 age- and sex-matched control subjects. Fang et al. confirmed these results using calibrated integrated backscatter in a larger study.
Biomarkers of collagen synthesis (PICP; PINP, amino-terminal propeptide of procollagen type I; PIIICP, carboxy-terminal propeptide of procollagen type III; PIIINP, amino-terminal propeptide of procollagen type III) or collagen degradation (CITP) have been showed to be of clinical interest in detecting myocardial fibrosis . In addition, markers of extracellular matrix turnover (such as MMP) and their inhibitors (TIMP, tissue inhibitors of MMP) might also be useful . However, few studies have used these biomarkers in diabetic patients. An increase in CITP was recently described in type 2 diabetic patients compared with control subjects and was correlated with diastolic function (mitral A duration minus pulmonary vein atrial reversal velocity duration) . MMP-7 has also been shown to be associated with diastolic dysfunction with microvascular complications (nephropathy) . Patients with diastolic dysfunction demonstrated an increase in MMP-9 and a decrease in TIMP-1/MMP-9 . Finally, in a small group of diabetic patients, correlation was found between PICP and LV systolic variables (fractional shortening and midwall fractional shortening) .
Myocardial steatosis
An increase in myocardial lipid content (myocardial steatosis) has been demonstrated in diabetic patients using 1H-magnetic resonance spectroscopy. McGavock et al. showed an increase in myocardial triglyceride content in glucose intolerant patients and in patients with type 2 diabetes mellitus compared with controls. Myocardial triglyceride content was associated with diastolic function variables (E/A ratio and E wave deceleration time) . In a group of 42 men with type 2 diabetes mellitus, patients with a high myocardial triglyceride content (superior to the median of the population) presented a decrease in systolic strain and strain rate whereas LV ejection fraction was similar in the two groups .
These structural abnormalities are associated with morphological and functional abnormalities
Concentric remodelling and left ventricular hypertrophy
Concentric remodelling is defined as an increase in relative wall thickness ([2 × posterior wall thickness]/end-diastolic diameter) with normal LV mass index values, whereas LV hypertrophy is defined as abnormal LV mass index values . Concentric remodelling and LV hypertrophy, known to be of adverse prognostic value , are the most frequently described morphological abnormalities associated with diabetes mellitus . In the Framingham Heart Study, Galderisi et al. showed an increase in LV mass and wall thickness independently associated with diabetes mellitus, but in multivariable analysis, statistical significance was reached only in women. In the Cardiovascular Health Study, an increase in LV mass was independently associated with diabetes mellitus even after adjustment for body weight, blood pressure, heart rate and prevalent coronary or cerebrovascular disease . However, in contrast to previous findings, this association was significant in both men and women . The Strong Heart Study confirmed these results in a large cohort of American Indians (1810 participants with diabetes mellitus and 944 glucose-tolerant subjects) . In this population, increases in wall thickness and LV mass were associated with a slight decrease in LV fractional shortening and midwall fractional shortening in diabetic patients compared with glucose-tolerant subjects . All these studies described concentric remodelling and LV hypertrophy associated with diabetes mellitus, independent of other confounding factors such as age, obesity (BMI) and hypertension. However, more recently, the NOMAS Study described diabetes mellitus as an independent determinant of LV mass but in addition to central obesity as assessed by waist circumference .
Impact of diabetes on left ventricular remodelling
The impact of diabetes mellitus on LV remodelling over the course of a lifetime has been demonstrated more recently. Indeed, the Framingham Heart Study investigators showed, in a first report, an increase in LV mass during a 16-year follow-up . Diabetic subjects (without heart failure or previous myocardial infarction) experienced a steeper increase in LV mass with advancing age than non-diabetic subjects . In a second report, Cheng et al. demonstrated an increase in wall thickness with advancing age associated with a decrease in LV diameter and a concomitant increase in fractional shortening. This study pointed out the influence of diabetes mellitus, in addition to sex and hypertension, on LV remodelling over life. Indeed, diabetes mellitus was associated with a greater increase in wall thickness coupled with a smaller decrease in LV diameter .
Finally, Markus et al. also reported the influence of diabetes mellitus on LV remodelling. During a 10-year follow-up in a population-based cohort of 1005 subjects, the authors observed a greater increase in LV mass in patients with prevalent diabetes mellitus associated with an increase in LV end-diastolic diameter, whereas this last variable remained stable over time in euglycaemic subjects ; this was associated with a decrease in LV ejection fraction in prevalent diabetic subjects whereas an increase was observed in euglycaemic subjects .
Functional abnormalities
Longitudinal systolic impairment
Myocardial strain imaging, such as TDI and STI, is useful for detecting ischaemia . These techniques also facilitate demonstration of subclinical impairment of systolic function in asymptomatic diabetic patients without overt heart disease and a normal standard echocardiography compared with controls . Whereas conventional methods such as LV ejection fraction and fractional shortening were insensitive in terms of detecting early preclinical abnormalities, longitudinal dysfunction was demonstrated in many studies using both TDI and STI . For some authors, longitudinal strain decrease was correlated with diabetes imbalance (glycated haemoglobin) or microvascular complications (microalbuminuria) . However, longitudinal alteration is independently associated with diabetes mellitus, regardless of LV hypertrophy or other conventional risk factors .
Radial systolic impairment
Radial systolic function has been investigated less frequently in this population. Indeed, only a few studies have assessed radial function, with conflicting results . The initial studies suggested that radial function was increased or preserved to compensate for longitudinal function alteration. However, most of these studies were based on TDI and its derived velocity and strain rate measurements that depend on Doppler angle . The TDI measurements are limited to the segments in which motion and deformation are aligned with the ultrasound beam and hence do not provide an optimal evaluation of radial function. STI is an angle-independent method and allows a more robust and extensive evaluation of radial function ( Fig. 2 ) . Among reports based on STI, the study by Ng et al. described preserved radial function compared with euglycemic subjects (radial strain 40.6 ± 11.1 vs. 42.7 ± 12.1%, respectively; P not significant). However, this study was based on a highly selected population that included uncomplicated patients and only men with a short duration of diabetes mellitus (4 years) and a strictly controlled haemoglobinA1c concentration (6.4 ± 0.7%). Conversely, our group demonstrated in a population of 114 diabetic subjects, a decrease in both radial strain (50 ± 16 vs. 56 ± 12%; P = 0.003) and longitudinal strain (–19 ± 3 vs.–22 ± 2%; P < 0.001) compared with 88 age-matched controls without cardiovascular risk factors . In multivariable analysis, diabetes mellitus was independently associated with longitudinal strain (in addition to sex) and radial strain. These findings remained stable in a subgroup analysis of 42 strictly age-, sex- and BMI-paired patients and controls .
