Assessment of retinal vascular changes in hypertension: new perspectives

Fundoscopic assessment of retinal microcirculation comprises the classic approach employed to detect structural alterations of retinal arterioles associated with chronic blood pressure elevations, as well as long-lasting abnormalities in insulin sensitivity and the blood glucose profile. The first systematic attempt to use fundoscopic methodology in hypertension dates back approximately 70 years, when Keith and coworkers published a pioneering study aiming to classify the hypertensive state and definine the disease prognosis based on a number of variables, including the retinal vascular network [1,2]. According to Keith's classification, four stages of hypertensive retinopathy can be identified [2]. The initial stage is characterized by an increased retinal arteriolar tone associated with vasospasm and focal narrowing of the small vessels. These alterations are then followed by an increase in intima thickness, hyperplasia of the media wall and more generalized arteriolar narrowing. At this stage, changes in arteriolar and venular function (i.e. arteriovenous nicking or nipping), as well as alterations in arteriolar light reflex (i.e. whickening or potentiation of the central light reflex), are detectable. The overall picture may be complicated by disruption of the blood–retinal barrier, with necrosis of smooth muscle cells, endothelial cells and retinal ischemia. At this more advanced stage, micro-aneurysms, hemorrhages, hard exudates and cotton-wool spots are not unusual fundoscopic findings. Finally, at the fourth stage, papilloedema with swelling of the optic disk takes places, indicating the presence of malignant hypertension. However, over the years, the retinal vascular network has received less attention from clinicians and investigators than other markers of hypertension-related target-organ damage, such as left ventricular hypertophy, carotid intima–media thickening, cerebral lacunae or microalbuminuria [3]. It is likely that this behaviour is due to the poor reproducibility of the approach, which was originally based on the subjective quantification of retinal vascular changes. Indeed, the intra and inter-observer variability of the method is in the range 20–42% and 10–30% (coefficient of variation), respectively, when applied to mild or moderate hypertensives [4]. This implies that the approach cannot reliably assess: (i) the earlier and less pronounced retinal vascular damage induced by hypertension and (ii) the effect of a given blood pressure-lowering intervention on retinal circulation. During recent years, the fundoscopic examination has gained new popularity, thanks to the availability of standardized procedures based on computer-assisted evaluation of digitized retinal photographs [5,6]. In the present issue of the journal, Hughes et al. [7] report the results of a study in which a novel multiscale analysis system, the so-called Retinal Image multiScale Analysis (RISA), was used to detect retinal abnormalities characterizing mild to moderate and malignant hypertensive states. The method automatically picks up and quantifies several topological features of the retinal microcirculation, identifying the skeleton and the order of the vessel trees and, with an operator-assisted procedure, the arteriolar and venular tree [7]. Their study results can be summarized as follows. First, hypertension in its early clinical phases is already characterized by an arteriolar rarefaction and by a narrowing (almost generalized) of the arteriolar vessels. Second, the above-mentioned alterations are magnified in malignant hypertension, which also displays an increased vascular tortuosity and rarefaction. The study results highlight some strengths and limitations of an approach based on the fundoscopic evaluation of hypertension-related retinal vascular changes. Prevalence of retinal vascular abnormalities in hypertension Despite the not inconsiderable amount of information collected over the years, the prevalence of grade 1 and grade 2 retinopathy in the hypertensive population does not appear to be well defined, and the data collected to date are somewhat controversial [3,4]. A recent review of studies aiming to assess the presence of hypertensive retinopathy via fundoscopic examination in individuals aged > 19 years reveals that the prevalence of retinal abnormalities in hypertensive patients is not as high as expected [8]. More importantly, the prevalence is consistently different from one study to the other, with the values ranging from 3 to 21% [8]. As previously mentioned, the heterogeneity of the findings reported in the different studies certainly depends on the different patterns of the retinal abnormalities detected in the various phases of the hypertensive disease. For example, in a recent study, Cuspidi et al. [9] showed that, although retinal arterial narrowings can be found in a large fraction of moderate essential hypertensives (approximately 85%), less than one-half (42%) of more severe hypertensive patients display retinal arteriovenous crossings. Taken together, these findings suggest that standardization of the fundoscopic approach, possibly by making use of a computerized or semi-computerized analysis system, as demonstrated by Hughes et al. [7], is mandatory for obtaining reliable figures with respect to the prevalence of hypertensive retinopathy. New technologies, such as those based on a web telemedicine system, could be of help in screening large population samples, as has been carried out for diabetes [10]. Retinal abnormalities, target-organ damage and programs in hypertension There are very few studies published to date on the relationship between retinal microvascular abnormalities and other types of hypertension-related end-organ damage. However, two recent studies [8,9] have shown that in selected populations: (i) structural changes of the arterial retinal network may be more frequently detected than left ventricular hypertrophy, carotid intima–media hyperplasia and microalbuminuria and (ii) the detection of hypertensive retinopathy doubles the risk of the occurrence of left ventricular hypertrophy and carotid abnormalities. Unfortunately, although the study by Hughes et al. [7] evaluated both moderate and malignant hypertension, it failed to provide new insights on this issue, given the lack of data on renal function in severe hypertensive patients. The evidence available in the literature is also scarce with respect to the relationship between vascular retinal abnormalities and cardiovascular events [11–13]. The Atherosclerotic Risk In Communities (ARIC) study [14,15], performed by employing a computerized assessment of arteriolar and vascular retinal calibres, recently demonstrated that arterovenous crossings, but not arteriolar narrowings, are independently associated with the risk of stroke and cognitive impairment in hypertensive patients. Whether the same conclusion also applies to the risk of congestive heart failure, renal failure or myocardial infarction associated with hypertension remains to be established. Unanswered questions Along with the previously mentioned relationship with target-organ damage and cardiovascular events, several other issues related to fundoscopic examination of the retinal vascular network are worthy of future investigation. For example, this is the case for the ability of retinal vascular abnormalities to reflect, in hypertensive disease, similar abnormalities in other vascular districts, such as the coronary, muscular and subcutaneous circulation. This is also the case for the involvement of metabolic factors, such as hyperinsulinemia and insulin resistance, in hypertension-related retinal vascular changes. Finally, it would be clinically important to determine whether the therapeutic regression of cardiac, renal and vascular target-organ damage seen in hypertension has the same temporal pattern of regression as retinal vascular damage.