‘Keep an eye’ on the heart: retinal microcirculation disarray in congestive heart failure

This editorial refers to ‘Retinal microvascular dysfunction in heart failure’†, by M.P. Nägele et al., on page 47. Assessment of the retinal microvascular network has been recognized for many years as a tool that can provide information of both pathophysiological and clinical relevance in a number of cardiovascular diseases. This has been shown to be the case in essential hypertension, in which evaluation of retinal vessels started about a century ago via approaches that were much less sophisticated and technologically refined than those currently adopted in the clinical and research setting. Indeed, it was through retinal vessel examination that it was found that in hypertensive patients retinal arteriolar diameters undergo a narrowing process whose magnitude is related to the severity of the hypertensive state.1–3 Furthermore, it was via retinal vessel studies that the arteriolar narrowing accompanying a chronic blood pressure elevation was found to be not only an adaptive process to the higher intravascular pressure but also a marker capable of independently predicting and contributing over the years to the development and progression of hypertension.1–2 , 4 Alterations in the microvascular retinal network have also been documented in diabetes mellitus, obesity, renal insufficiency, and renal failure, their detection being associated with target organ damage at the cardiac and vascular level as well as with an increased risk of future fatal and non-fatal cardiovascular and renal events.2 , 5–7 The present issue of the journal reports the intriguing results of the study by Nagele et al. 8 which addresses for the first time the structural and functional characteristics of retinal arterioles and venules in patients with congestive heart failure and an impaired left ventricular function due to either myocardial ischaemia or idiopathic dilated cardiomyopathy. To this aim, the authors recruited 222 individuals, distributed in three groups, i.e. patients with congestive heart failure, patients with cardiovascular risk factors but without any cardiovascular disease, and healthy controls. Heart failure patients in New York Heart Association class II were under optimized pharmacological treatment and exhibited several associated cardiovascular risk factors, such as obesity or overweight, hypertension, dyslipidaemia, and diabetes mellitus. The prevalence of these conditions was similar, however, to that seen in the group of patients with cardiovascular risk factors, and significantly and markedly greater than that detected in healthy controls. In each subject, different parameters of retinal morphology and function were collected. By using a mydriatic approach, a dynamic retinal analyser, and a fundus camera, the authors measured the diameters of the retinal arterioles and venules anatomically located in the area of the ocular optic disc and calculated the central retinal artery and vein equivalents (CRAE and CRVE, respectively) as well as the arteriovenous ratio (AVR) value. In addition, by employing a sophisticated approach based on the assessment of the arteriolar and venular dilation in response to repeated local flicker stimulations (localized visual stimuli delivered by projecting flicker patterns inside or outside the downstream region of the target vessel), the authors were able to collect functional data on the retinal network. The results showed that in patients with heart failure (i) there was a marked reduction (–75% compared with controls and –60% compared with patients with a high cardiovascular risk profile) in the dilatory responses of both arteriolar and venular retinal microvessels to the flicker stimuli and (ii) the alterations (which were statistically significant after data adjustments for confounders) were common to both ischaemic heart failure and idiopathic dilated cardiomyopathy, although somewhat more pronounced in the former case. This provides evidence that a major functional disarray of the retinal microcirculation characterizes the failing heart, regardless of its aetiology. Interestingly, in the case of the retinal venular alterations, a significant inverse relationship was found to echocardiographic markers of left ventricular filling pressure. It was also found that the functional alterations were not accompanied by significant differences in CRAE, CRVE, and AVR values between groups of patients, indicating that the main structural features of the retinal microcirculation were not deranged in the failing heart syndrome. As shown in Take home figure, a number of factors may participate in the development of the functional microvascular disarray of the retinal microvascular network that occurs in heart failure patients. A factor of primary importance is probably the endothelial dysfunction because (i) endothelial release of nitric oxide has a powerful dilatory effect on vessels of all dimensions9 and (ii) a reduced ability of endothelial cells to release nitric oxide and affect vasomotor tone has been consistently reported in heart failure.10 Indeed, the observations made in the present study provide direct support for the involvement of the endothelial factor in the microvascular alterations of heart failure because the flicker-induced arteriolar and venular vasodilatation is specifically mediated by nitric oxide release.8 Schematic drawing illustrating the potential mechanisms responsible for the retinal microcirculatory disarray in congestive heart failure. Heart failure is characterized, along with haemodynamic alterations (left arrows), by insulin resistance, sympathetic (SNS) overactivity, renin–angiotensin–aldosterone system (RAAS) activation, and nitric oxide (NO) impairment. All these metabolic and humoral alterations may trigger functional alterations of the retinal microvessels, by directly acting on the retinal network or, more probably, by favouring and potentiating the NO impairment at the endothelial level. Several other factors may also participate, however. This is the case for the sympathetic nervous system either because in heart failure patients there is a generalized increase of sympathetic activity11 or because sympathetic neural influences modulate the structure and the dynamic function of the microcirculation.12 It is also the case for insulin resistance, both because insulin resistance is also common in heart failure13 and because it leads to an increase of circulating levels of insulin, a hormone that in obesity and the metabolic syndrome has shown an association with microcirculatory alterations.14 It is finally the case for the renin–angiotensin–aldosterone system, whose activation in heart failure11 leads to elevated circulating levels of both angiotensin II and aldosterone, i.e. two substances with established adverse effects on microvessels.12 It is important to add that these additional factors may contribute to the development of the functional retinal disarray separately, but that they may also reinforce each other because of their positive feed-back type of relationship, as exemplified by the evidence that while sympathetic activity stimulates renin release and aggravates insulin resistance, angiotensin II and insulin both have a marked sympathoexcitatory effect.15 In addition, because of their modulation of a variety of endothelial functions, all of them can modulate microvascular vasomotion and structure directly, but also operate indirectly through the common pathway represented by the endothelial factor. In recent years important technical refinements in the in vivo assessment of retinal vessels have renewed and stimulated interest in the evaluation of microvasculature alterations in cardiovascular diseases. The scanning laser flow meter method, for example, represents a clear methodological improvement because, via a non-midriatic approach, retinal vessel diameters can be studied in a dynamic fashion, i.e. during the whole cardiac cycle through an automatic full-field perfusion imaging analysis.16 As mentioned above, the method used by Nagele and co-workers8 also carries an important advantage because it allows direct examination of nitric oxide-dependent retinal vasodilatation, i.e. addressing a factor of primary pathophysiological importance for cardiovascular diseases. Some general elements of weakness need also to be mentioned, however. First, the method used by Nagele et al. 8 needs pharmacological dilatation of the pupil via anticholinergic drugs, which may affect, to an unknown and uncontrolled extent, microvascular reactivity. Secondly, ‘normal’ values for retinal vessel structure and function have been provided for the software analysis of CRAE, CRVE, and AVR values17 but are not available for all methodological approaches or functions, including that studied by Nagele and co-workers.8 Thirdly, whether and to what extent the microcirculatory pattern described at the level of the retinal network reflects other microvascular areas (e.g. those in the subcutaneous and visceral tissues) remain to be determined. This is critical to the detection of which area shows more sensitivity and specifically the earliest the microvascular alterations associated with the disease. The study by Nagele and co-workers8 represents an important initial step for considering retinal examination as a procedure of potential interest in the diagnostic approach to the failing heart syndrome. The examination will not replace the classic evaluation of the heart failure state but it might represent a tool useful for a more complete assessment of this clinical condition. Future studies will be needed to determine the prognostic value of the observed endothelial alterations as well as their potential utility as markers of the disease severity and progression. Furthermore, assessment of the impact of the pharmacological treatment aimed at improving the heart failure condition will be also worth carrying out as one of the major goals of future investigations. Conflict of interest: none declared.