Diabetes Mellitus, Natriuretic Peptide Axis Dysregulation, and Diabetic Cardiomyopathy

This editorial refers to ‘Pre-clinical diabetic cardiomyopathy: prevalence, screening and outcome’ by S. Kiencke et al., published in this issue on pages 951–957 and ‘Slightly elevated B-type natriuretic peptide levels in a non-heart failure range indicate a worse left ventricular diastolic function in individuals with, as compared to individuals without, type 2 diabetes: the Hoorn Study’ by K. van den Hurk et al., published in this issue on pages 958–965. Early cardiovascular epidemiological studies indicated that diabetes mellitus was an ‘equalizer’ of vascular risk between women and men, i.e. the presence of diabetes in women eliminated the lower rates of cardiovascular disease observed in them relative to those observed in men.1 Subsequent investigations further emphasized the ominous impact of diabetes on vascular risk by labelling the condition as a coronary heart disease ‘risk equivalent’.2 Unfortunately, the global burden of diabetes is projected to increase over the next few decades globally,3 and type 2 diabetes has been observed to occur at much earlier ages, including in adolescents.4 Even in patients with arterial hypertension, type 2 diabetes mellitus is associated with echocardiographic evidence of diastolic dysfunction, a risk marker for future heart failure symptoms and cardiovascular events.5 Furthermore, diastolic dysfunction already occurs in pre-diabetic stages of glucose metabolism abnormalities.6 The latter observations raise serious concerns regarding the future burden of heart disease related to diabetes. In this context, the two reports7,8 in this issue of the journal draw attention to the epidemiology of diabetic cardiomyopathy and to the potential role for natriuretic peptides in screening for the condition. The report by Kiencke et al.7 evaluated the prevalence of diabetic cardiomyopathy in an ambulatory clinic-based sample of middle-aged, overweight-to-obese individuals with prevalent diabetes mellitus for an average duration of over 10 years. The investigators also assessed the utility of plasma B-type natriuretic peptide (BNP) concentrations for screening for the presence of diabetic cardiomyopathy cross-sectionally, and for prognostication with regard to risk of vascular events and death prospectively. The authors report that about a third of their sample met the pre-specified (although somewhat empirical) criteria for diabetic cardiomyopathy and that plasma BNP concentration had inadequate performance characteristics to serve as a reliable screening test for the condition. Yet, higher BNP levels within the ‘normal range’ were markers of an adverse prognosis.7 The second report by van den Hurk et al.8 studied the cross-sectional association of BNP levels with left ventricular (LV) mass, presence of LV diastolic and systolic dysfunction, and left atrial volumes in a population-based cohort of older men and women stratified according to varying degrees of glucose intolerance. The investigators report that LV mass, diastolic dysfunction, and left atrial volume (but not systolic dysfunction) were positively associated with circulating BNP and that the relations were strongest (in terms of effect size) in individuals with diabetes.8 Taken together, these two reports highlight the high prevalence of diabetic cardiomyopathy in multiple samples, and they confirm the association of the condition with plasma BNP levels. The first report7 also illustrates the limited role of plasma BNP for screening for the condition and confirms prior reports that levels within the so-called ‘normal range’ are prognostically important.9 Of note, these reports add to the body of evidence that LV diastolic dysfunction is an early manifestation of diabetic cardiomyopathy,10,11 and the high prevalence of such diastolic dysfunction should be of concern given its association with mortality and with heart failure.12 Several additional points are worth noting as well. These studies demonstrate how risk markers may not always make good screening tests,1,13 given that very high odds ratios are required to yield reasonable increments to screening performance metrics such as the area under the receiver-operating characteristic curves.14 The choice of BNP levels for screening in diabetes also merits comment, given prior evidence indicating that individuals with obesity, metabolic syndrome, and insulin resistance seem to have lower plasma BNP levels (compared with their non-diabetic, non-obese counterparts), likely reflecting a ‘natriuretic handicap’.15 In other words, despite greater myocardial wall stress related to LV diastolic dysfunction and/or the associated burden of risk factors, individuals with diabetes may have lower BNP levels, thereby posing an obvious challenge to the use of plasma BNP levels for screening in this subgroup. The lower levels of BNP in diabetes are of relevance especially given recent evidence that transgenic mice that overexpress BNP are lean and resistant to diet-induced obesity, presumably due to increased lipolysis and fat oxidation due to an up-regulation of mitochondrial biogenesis in muscles.16 The pathogenesis of LV diastolic dysfunction in diabetes and of diabetic cardiomyopathy has been extensively studied.10,11 Intramyocardial accumulation of triglycerides and extracellular deposition of excess collagen and advanced glycation products trigger glucolipotoxicity and activation of several signalling pathways (insulin resistance, oxidative stress, renin–angiotensin system, adipokines, and inflammation) in a milieu of altered substrate metabolism.10,11 It is likely that the presence of a natriuretic handicap may compound these metabolic changes, as noted above. These events translate into myosin isoform switching, mitochondrial dysfunction, and altered calcium homeostasis in cardiac myocytes, which manifest as altered lusitropy.10,11 It is likely that the impact of diabetes is amplified by the presence of high blood pressure and dyslipidaemia, which frequently cluster together. It is also likely that increased vascular stiffness in diabetes17 exacerbates and directly contributes to diastolic dysfunction. The two reports7,8 in this issue draw attention to the notion that LV diastolic dysfunction follows diabetes mellitus as an ominous shadow. However, they also direct us towards several unanswered questions. How early during the course of diabetes mellitus do these early changes in LV diastolic function manifest? What is the natural history of early diabetic cardiomyopathy evident as mild degrees of diastolic function impairment, and is there a time window for reversing these changes? Would use of imaging during exercise stress, or of newer imaging tools such as assessment of strain, strain rate and torsion and dyssynchrony further unmask subtle LV diastolic function alterations earlier in the course of diabetes? Are there better candidate biomarkers (alone or in combination) for screening for diabetic cardiomyopathy, such as urine albumin excretion rate, levels of growth differentiation factor-15, or the latest generation of high sensitivity troponin T or I assays? Will current genomic tools and the notion of genomic convergence (evaluating concordance of genome-wide association signals with gene expression patterns and circulating proteomic and metabolomic markers to identify culprit pathways) yield newer biomarkers for the condition and/or molecular targets? Clearly, additional larger studies that include prospective evaluation of individuals with tools using more sophisticated imaging and other OMIC tools are warranted to address these questions. It is conceivable that rates of cardiovascular disease, including diabetic cardiomyopathy, may escalate as the global burden of diabetes soars over the next two decades. Fortunately, lifestyle measures such as physical activity and diet are associated with lower risk of developing heart failure over the adult life course.18 It is possible that these lifestyle measures, if adopted more widely, could mitigate the future burden of LV diastolic dysfunction as well. Increasing the awareness regarding the potential adverse impact of diabetes (on the heart in general and on cardiac function in particular) among physicians, other health-care providers, and among patients may help stem the tide towards a greater global burden of diastolic dysfunction arising from the epidemic of diabetes. The authors wish to acknowledge the assistance of Ramachandran S. Vasan, MD, the Framingham Heart Study, for his thoughtful comments on the manuscript.