European Heart Journal Advance Access originally published online on May 7, 2008
European Heart Journal 2008 29(11):1342-1343; doi:10.1093/eurheartj/ehn203
Sweet hearts die earlier—lessons from CHARM
Klinik für Innere Medizin III (Kardiologie, Angiologie, Internistische Intensivmedizin), Universitätsklinikum des Saarlandes, Kirrberger Straße, D-66421 Homburg/Saar, Germany
* Corresponding author. Tel: +49 6841 1623372, Fax: +49 6841 1623369, Email: boehm{at}med-in.uni-saarland.de
This editorial refers to ‘Impact of diabetes on outcomes in patients with low and preserved ejection fraction heart failure. An analysis of the Candesartan in Heart failure: Assessment of Reduction in Mortality and morbidity (CHARM) programme’
by M.R. MacDonald et al., on page 1377
Footnotes
The opinions expressed in this article are not necessarily those of the Editors of the European Heart Journal or of the European Society of Cardiology.
doi:10.1093/eurheartj/ehn153 
It has been recognized for nearly a quarter of a century that patients with diabetes who develop cardiovascular disease (CVD) suffer from a worse prognosis than do CVD patients without diabetes mellitus.1 Against this background, the news from the CHARM Investigators2 that diabetes is an independent predictor of cardiovascular morbidity and mortality, not only in heart failure patients with severe systolic dysfunction but also in those with preserved ejection fraction, is not a big surprise. Accelerated coronary atherosclerosis,3 abnormal platelet function, coagulation, fibrinolysis, and endothelial dysfunction which favour coronary thrombosis,4 and autonomic neuropathy associated with an increased rate of ventricular arrhythmia and sudden death5 are possible mechanisms which may independently from left ventricular systolic dysfunction contribute to an increased risk of fatal and non-fatal complications in patients with diabetes and CVD.
Hence, the interesting and not easy to explain issue in the study of MacDonald and co-workers2 is not the new but expected finding that diabetes is harmful in patients with preserved ejection fraction (HF-PEF), but that the detrimental effect of diabetes in these patients is similar to or even more pronounced than in patients with a low left ventricular ejection fraction (HF-LEF).
One might argue that the significant interaction between diabetes and left ventricular ejection fraction seen in the primary combined outcome of cardiovascular death or heart failure hospitalization (P = 0.0009) was mainly driven by the component of heart failure hospitalization (P = 0.0029) while cardiovascular death alone failed to show a significant interaction (P = 0.1559). As overall mortality was considerably higher in the HF-LEF group (29.5%) than in HF-PEF patients (15.9%), the pronounced role of diabetes as a risk factor for heart failure hospitalization particularly in HF-PEF patients could simply be attributed to a prolonged exposure time due to a lower mortality. However, this competing risks hypothesis cannot fully explain the results of the study. If competing risks promoted heart failure hospitalizations in HF-PEF patients, non-diabetic patients with preserved ejection fraction who show a far lower mortality (13.7%) than non-diabetic patients with low ejection fractions (26.1%) should have a higher rate of heart failure hospitalizations, which they clearly do not (12.6% in non-diabetic HF-PEF, 21.7% in non-diabetic HF-LEF). In addition to that, despite a lack of significant interaction P-values for cardiovascular death and all-cause mortality, the relative diabetes-associated increase of risk for HF-PEF patients in comparison with the HF-LEF group is intriguingly similar for the different outcomes: in HF-PEF patients the presence of diabetes consistently augmented the risks by 19–25% more than in HF-LEF patients.
If different life expectancies and competing risks do not explain the particular importance of diabetes as a risk factor in HF-PEF patients, are there any other confounding factors that might have influenced the study results? Dihydropyridine calcium channel blockers and thiazolidinediones (glitazones) both may cause fluid retention and peripheral oedema. From the controversy about cardiovascular safety of glitazone therapy,6 it is well known that in large-scale clinical outcome trials it might be difficult always to differentiate between non-cardiac fluid retention and hydropic decompensation due to progressive heart failure.7 Thus fluid retention caused by glitazones (incidence: 2–5% with monotherapy, 5–15% with concomitant insulin therapy7) or calcium channel blockers (incidence: 2–30% depending on substance and dosage) can easily be misinterpreted as decompensated heart failure and may lead to an increase in the counted number of heart failure hospitalizations.
As expected in the HF-PEF group of the CHARM study population, there was a considerably higher use of calcium channel blockers than in the HF-LEF group (31.2% vs 13.1%) and, within the HF-PEF group, diabetic patients more often received a calcium channel blocker than non-diabetic patients (35.5% vs 29.5%). Although the authors state that ‘glitazones were not widely used during the recruitment period of the CHARM study’, exact data are missing, which is a limitation of this post hoc analysis. Since glitazone therapy is generally considered as contraindicated in New York Heart Association (NYHA) III/IV heart failure stages, it is not implausible that the diabetic HF-PEF patients of the CHARM trial, of whom 50% were in NYHA class II, more often received a glitazone than the diabetic patients of the HF-LEF group, of whom only 30% were in NYHA class II. Hence, calcium channel blockers and glitazones might have introduced a bias favouring increased heart failure hospitalizations in HF patients with preserved ejection fraction.
However, these limitations are certainly not sufficient to compromise the results of MacDonald and co-workers. Interestingly, a recently published report on long-term outcome in diabetic heart failure patients treated with cardiac resynchronization therapy (CRT)8 further put into perspective the role of diabetes mellitus as a predictor of bad outcome in advanced heart failure. In this study, diabetes mellitus did not confer an increased risk neither for all-cause death, nor for cardiovascular death, nor for the time to first hospitalization. As patients eligible for CRT are characterized by NYHA class III or IV status and an ejection fraction 
35%, they are very comparable with the HF-LEF group in the current CHARM substudy.
Which causal mechanisms may provide an explanation for the findings of the CHARM investigators? A closer look at Table 2, which summarizes fatal end-point occurrences and hospital admissions for diabetic and non-diabetic HF-PEF and HF-LEF patients, might provide some insights. While in the HF-LEF group there is no relevant difference in the proportion of cardiovascular deaths between diabetics (83.2%) and non-diabetics (82.8%), in the HF-PEF group diabetes partially changes the mode of death from non-cardiovascular to cardiovascular, with an increase in the proportion of cardiovascular deaths from 67.6% in non-diabetics to 75.7% in diabetics. Although a comparable analysis for the hospitalization end-points in Table 2 is hampered by the fact that these events are in contrast to the fatal end-points not mutually exclusive, one may calculate a ratio of the sum of all cardiovascular events and the sum of all non-cardiovascular events. Doing this, the effect of diabetes on the occurrence of hospital admissions due to cardiovascular causes is very similar to the effect on the cardiovascular death rate: while diabetes has no influence on the basically high ratio of CV to non-CV hospital admissions in the HF-LEF group (1.78:1 in diabetics vs 1.76:1 in non-diabetics), the presence of diabetes increases the ratio of cardiovascular deaths relative to non-cardiovascular deaths in the HF-PEF group from 1.36:1 to 1.57:1. In other words, in populations with a lower general risk such as the HF-PEF group, diabetes specifically promotes cardiovascular complications, while in populations with an already high cardiovascular risk such as the HF-LEF group this effect is attenuated because it is masked by the high basal risk of clinical and haemodynamic status related to heart failure.
The development of diabetic cardiomyopathy may explain the particular role of diabetes mellitus in heart failure patients with preserved left ventricular ejection fraction. In diabetic cardiomyopathy, diabetes might not only be a risk indicator but could represent one mechanism of progressive left ventricular dysfunction. Notably in diabetics in the earlier stages, cardiomyopathy is characterized by a predominance of left ventricular diastolic dysfunction with preserved or only mildy reduced left ventricular ejection fraction.9 Pathological changes include myocyte hypertrophy, perivascular fibrosis, and increased deposition of matrix collagen, cell membrane lipids, and cellular triglycerides. Collagen cross-linking secondary to advanced glycation end-products which happens in the myocardium as well as in the arterial vasculature leads to decreased myocardial and vascular compliance.9,10 With ongoing diabetic cardiomyopathy, diastolic function progressively deteriorates, which predisposes to cardiac decompensation and hospitalization. Diabetes is also recognized as the most common cause of renal failure and an independent risk factor for atrial fibrillation,11 which both are common trigger factors for cardiac decompensation in HF-PEF patients. Besides the development of diabetic cardiomyopathy which represents a disease entity of its own, the above-mentioned structural and functional tissue alterations in diabetes render the heart muscle more susceptible to hypertensive and ischaemia-mediated damage, which explains the potentiating harmful effect of diabetes mellitus in various aetiologies of heart failure.9
In summary, despite some limitations in this post hoc analysis of the CHARM trial, the finding that diabetes mellitus has a particular prognostic impact on heart failure patients with preserved left ventricular ejection fraction is original and deserves further attention in future research. Although some explanations of these findings have been addressed, e.g. the specific characteristics of diabetic cardiomyopathy or a type of ceiling effect in the low ejection fraction group with a pre-existing high basal risk, further studies are needed to clarify the different pathophysiological mechanisms and disease entities caused by diabetes in patients with low vs preserved ejection fraction.
Conflict of interest: none declared.
Footnotes
The opinions expressed in this article are not necessarily those of the Editors of the European Heart Journal or of the European Society of Cardiology.
References
- Smith JW, Marcus FI, Serokman R. Prognosis of patients with diabetes mellitus after acute myocardial infarction. Am J Cardiol (1984) 54:718–721.[CrossRef][Web of Science][Medline]
- MacDonald MR, Petrie MC, Varyani F, Östergren J, Michelson EL, Young JB, Solomon SD, Granger CB, Swedberg K, Yusuf S, Pfeffer MA, McMurray JJV, for the CHARM Investigators. Impact of diabetes on outcome in patients with low and preserved ejection fraction heart failure. An analysis of the Candesartan in Heart Failure: Assessment of Reduction in Mortality and morbidity (CHARM) programme. Eur Heart J (2008) 29:1377–1385. First published on April 14, 2008. doi:10.1093/eurheartj/ehn153.
[Abstract/Free Full Text] - Hambly RI, Sherman L, Mehta J, Aintablian A. Reappraisal of the role of the diabetic state in coronary artery disease. Chest (1976) 70:251–257.[CrossRef][Web of Science][Medline]
- Carr ME. Diabetes mellitus: A hypercoagulable state. J Diabetes Complications (2001) 15:44–54.[CrossRef][Web of Science][Medline]
- Vinik AI, Ziegler D. Diabetic cardiovascular autonomic neuropathy. Circulation (2007) 115:387–397.
[Free Full Text] - Lago RM, Singh PP, Nesto RW. Congestive heart failure and cardiovascular death in patients with prediabetes and type 2 diabetes given thiazolidinediones: a meta-analysis of randomised clinical trials. Lancet (2007) 370:1129–1136.[CrossRef][Web of Science][Medline]
- Wang CH, Weisel RD, Liu PP, Fedak PW, Verma S. Glitazones and heart failure: critical appraisal for the clinician. Circulation (2003) 107:1350–1354.
[Free Full Text] - Fantoni C, Regoli F, Ghanem A, Raffa S, Klersy C, Sorgente A, Faletra F, Baravelli M, Inglese L, Salerno-Uriarte JA, Klein HU, Moccetti T, Auricchio A. Long-term outcome in diabetic heart failure patients treated with cardiac resynchronization therapy. Eur J Heart Fail (2008) 10:298–307.
[Abstract/Free Full Text] - Bell DS. Heart failure: the frequent, forgotten, and often fatal complication of diabetes. Diabetes Care (2003) 26:2433–2441.
[Abstract/Free Full Text] - Poornima IG, Parikh P, Shannon RP. Diabetic cardiomyopathy: the search for a unifying hypothesis. Circ Res (2006) 98:596–605.
[Abstract/Free Full Text] - Kannel WB, Wolf PA, Benjamin EJ, Levy D. Prevalence, incidence, prognosis, and predisposing conditions for atrial fibrillation: population-based estimates. Am J Cardiol (1998) 82.
CiteULike 
Connotea 
Del.icio.us What's this?
Related articles in EHJ:
- Impact of diabetes on outcomes in patients with low and preserved ejection fraction heart failure: An analysis of the Candesartan in Heart failure: Assessment of Reduction in Mortality and morbidity (CHARM) programme
- Michael R. MacDonald, Mark C. Petrie, Fumi Varyani, Jan Östergren, Eric L. Michelson, James B. Young, Scott D. Solomon, Christopher B. Granger, Karl Swedberg, Salim Yusuf, Marc A. Pfeffer, John J.V. McMurray, and and for the CHARM Investigators
EHJ 2008 29: 1377-1385.[Abstract] [Full Text]
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||