The present study confirms that exercise intolerance is related to insulin resistance in heart failure patients. The study further shows that coronary flow reserve, a measure of coronary microvessel function, is correlated to both exercise capacity and insulin resistance and supports a mediating role of CFR in explaining the link between insulin resistance and exercise intolerance. Peripheral endothelial function and arterial stiffness were not associated with exercise capacity and, notably, left ventricular function was not related to either insulin resistance, exercise capacity, coronary- or peripheral vascular function.
Previous studies has shown a reduced CFR in patients with type 2 diabetes compared to patients without diabetes and an association between CFR and the severity of diabetes measured as HbA1c [3
] and fasting glucose [5
]. Only two studies have examined the relationship between the degree of insulin resistance and CFR in patients without diabetes, with one study showing an inverse relationship between CFR and HOMA index in 45 women with suspected coronary artery disease and angiographically normal coronary arteries [20
] and another showing correlation between CFR and insulin sensitivity measured using a hyperinsulinaemic euglycemic clamp in obese patients [21
]. Patients with chronic heart failure are known to have a high degree of insulin resistance and many develop diabetes. The mechanisms of this are unclear, but a common denominator in type 2 diabetes and heart failure is physical inactivity. In the present study patients with type 2 diabetes did not have significantly reduced CFR, although with a larger study population there might have been a difference, but insulin sensitivity was strongly correlated with CFR, indicating pre-diabetes microvascular damage.
Exercise intolerance is the most prevailing symptom in chronic systolic heart failure and VO2
peak is an important prognostic factor in this patient group. However, the link between the degree of cardiac dysfunction and exercise capacity is poorly understood, and previous studies have shown a poor correlation between VO2
peak and LVEF [22
]. In concordance, we found no correlation between these two parameters. We found a positive correlation between CFR and VO2
peak, which has been shown previously in healthy young men but not in patients with heart failure [10
]. It is possible that impaired microvascular function contributes to explain the link between cardiac dysfunction and the impairment inVO2
Insulin resistance is highly prevalent in heart failure [23
]. In accordance with previous studies we found a positive correlation between insulin sensitivity and VO2
peak in heart failure patients [1
], which remained unaffected when leaving out patients with diabetes. The underlying mechanism for the relationship between insulin sensitivity and VO2
peak is not completely understood. We showed that when adjusting for the effect of CFR on VO2
peak, the relationship to insulin sensitivity was weakened and no longer statistically significant. A possible explanation is that increasing insulin resistance influences the cardiac microcirculation resulting in a reduced CFR, which leads to an impaired exercise capacity (Figure ).
Our study did not show any relationship between CFR and peripheral endothelial function measured using flow mediated vasodilation. While a relationship has been shown between CFR and flow mediated vasodilation using brachial ultrasound in other patient populations [25
], this is not well examined in heart failure patients. In one study of patients with dilated cardiomyopathy no correlation between CFR using echocardiography and flow mediated vasodilation was found [26
]. Similarly, a study comparing patients with non-ischemic heart failure to healthy controls found that flow mediated vasodilation and CFR measured using positron emission tomography were correlated in the healthy controls but not in the patients with heart failure [27
]. Interestingly, in a small study of dogs with pacing induced heart failure, CFR was reduced with progression of heart failure, but coronary endothelial function was preserved until a late stage of heart failure [28
]. Thus our results are consistent with the literature in finding no association between CFR and endothelial function.
The present study shows a negative correlation between CFR and augmentation index, which is as measure of stiffness in the large arteries. Only one previous study has related CFR to augmentation index [29
]. In this study of asymptomatic patients with risk factors for ischemic heart disease, the patients with higher augmentation index and pulse wave velocity had a lower CFR. In a normal elastic aorta the pulse wave reflects from the periphery and returns to the heart in diastole, which improves the diastolic filling of the coronary arteries. With aorta stiffening the pulse wave returns during systole resulting in increased afterload and myocardial oxygen demand which could mean, that a reduced CFR in heart failure patients without significant coronary artery stenosis might not only be due to impaired coronary microcirculation, but that a reduced diastolic filling can be a participating factor. Insulin sensitivity and augmentation index were independently associated with CFR, suggesting that they influence CFR through different mechanisms. However, this finding remains to be confirmed in other studies.
Previous studies have shown that about 50% of asymptomatic patients with type 2 diabetes have signs of diastolic dysfunction [30
] and that diastolic dysfunction is associated with insulin resistance in patients with suspected coronary artery disease without diabetes [31
] without systolic heart failure. In patients with systolic heart failure, patients with type 2 diabetes have significantly higher E/e’ than patients without diabetes [32
]. Although we did not include measures of diastolic function the present study, a possible mechanism between insulin sensitivity and diastolic dysfunction might be impaired coronary microvascular function.
A reduced CFR has been shown to be an independent predictor of poor outcome in patients with dilated cardiomyopathy [8
] and in a mixed population of heart failure patients [33
]. In the present study we find a relationship between CFR, VO2
peak and insulin sensitivity, which also are well known prognostic risk factors in heart failure, suggesting that these risk factors are closely linked. Improving VO2
peak and insulin sensitivity through exercise training could also improve CFR as has been shown in a recent study of heart failure patients [14
]. CFR has also been improved through other interventions such as weight loss in obese women [34
] and after treatment with a beta-1 receptor blocker in dilated cardiomyopathy [35
]. However, it has not yet been shown whether improvement in CFR through e.g. exercise training is followed by an improvement in prognosis for heart failure patients. An ongoing study will determine if CFR can improve after different modalities of exercise training in heart failure patients, and if any changes will correspond to changes in peak oxygen uptake, metabolic fitness and in augmentation index.
Limitations: Although the study is limited in size this is somewhat counteracted by the precision of the measures used: insulin sensitivity and glucose clearance were assessed by hyperinsulinemic isoglycemic clamp, the gold standard, and adjusted to fat free mass from DEXA scan. VO2peak was a true maximum test as indicated by the high RER values. The cross-sectional nature of the study impedes causal inference and any conclusions regarding potential effects of intervention must therefore remain speculative. The patients had all had a LAD without significant stenosis at a previous coronary angiography or coronary CT-scan, but the examination was not repeated at the time of the study, meaning that some of the patients potentially could have developed a significant LAD-stenosis. All of the patients did however perform a maximal symptom limited cardiopulmonary exercise test at enrolment in to the study without chest pain or signs of ischemia on the ECG minimizing the chances of a significant LAD-stenosis.