Factors Associated with Changes in Retinal Microcirculation After Antihypertensive Treatment

ORIGINAL ARTICLE Factors associated with changes in retinal microcirculation after antihypertensive treatment

P-R Antonio1, P-S Marta1, DDJ Luı´s2, DPJ Antonio3, ST Manuel4, MS Rafael5, GV Sonia6, GP Manuel6, M-N Isabel7, ENJ Carlos7, C-d-T Gabriel8 and G-UI Francisco9

We report the results of hypertensive treatment over retinal arteriole narrowing in a group of 189 hypertensive patients during a 6-month treatment programme for hypertension. These patients were included in an observational study under clinical practice conditions and analysed using digital photographs of the eye fundus with a previously described semiautomatic snakes method. We analysed the relation between blood pressure control and retinal microcirculation changes during the treatment. We included 189 hypertensive patients; 74.0% of them had BP under control after 6 months of follow-up. The arteriovenous ratio (AVR), measured as the relation between the average retinal arteriole and venous diameter, signiﬁcantly increased after 6 months of follow-up in both eyes (right eye AVR: 0.769±0.065 vs 0.799±0.066 (Po0.0001); left eye AVR: 0.770±0.065 vs 0.796±0.071 (Po0.0001)). AVR changes were caused by increases in arteriole diameter. No linear correlation was found between blood pressure- lowering levels and AVR increase, suggesting that other factors could be involved in the regression of retinal changes. We conclude that in our group of hypertensive patients it was possible to increase the retinal arteriole diameter, expressing an improvement in retinal microcirculation after 6 months of adequate BP control. Although there is an inverse tendency between blood pressure and arteriovenous ratio changes, we could not ﬁnd a linear correlation between these changes.

Journal of Human Hypertension (2014) 28, 310–315; doi:10.1038/jhh.2013.108; published online 21 November 2013 Keywords: arteriovenous ratio; blood pressure; retinal microcirculation

INTRODUCTION locates the region of interest (ROI) – that is, the optic nerve. Second, the Early alterations in retinal microcirculation are observed in most system detects the vascular tree and measures the vessel calibre in several circumferences concentric to the optic disc. Then, the user selects the set hypertensive patients seen in daily practice (that is, retinal of vessels suitable for estimating the arteriovenous ratio and classifies arteriole narrowing or arteriovenous nicking without exudates or these vessels into arteries and veins. Finally, the system estimates the AVR haemorrhage causing damage to the retina itself), and these as the ratio between the average artery and vein vessel calibres. changes are subjective, imprecise and poorly reproducible. As per The snake-based method has high sensitivity and increases the hypertension guidelines, a method is necessary that objectively specificity of the previously described linear method. assesses these changes in hypertensive patients so that a valuable This is a post-authorisation, observational, prospective and multi-centre prognosis can be arrived at. study that was conducted under daily practice conditions at the We described elsewhere a semiautomatic computerised system Department of Internal Medicine of the Hospital de Conxo in Santiago for the evaluation of the calibre of retinal blood vessels that has de Compostela, Hospital Abente Lago and San Rafael de A Corunã, Hospital 1 de Barbanza, Hospital Xeral Calde in Lugo (Spain). Except for the shown very good reproducibility. In a later study, we described a photographs of the retina used as measurement instruments for the main semiautomatic method based on the snakes model that has better 2 objective of the study, the patients were not subjected to any diagnostic or reliability. follow-up procedure other than the usual in daily practice. Photographs Here we report the results of hypertensive treatment on retinal were taken at each hospital by an ophthalmologist and were analysed arteriole narrowing in a group of 189 hypertensive patients during both locally and centrally at Hospital de Conxo in Santiago de Compostela. a 6-month treatment programme with Telmisartan plus Hydro- Digital photographs were taken following pupillary dilatation with chlorothiazide under clinical practice conditions that were pilocarpine and phenylephrine using a TRC-50 IA camera connected to analysed with the previously described semiautomatic method. an Imagenet 1024 system (both from Topcon Instruments, Paramus, NJ, We analysed the relation between blood pressure control and USA). Photographs were taken with a 1280 Â 1024 pixel resolution for an adequate contrast for vessel differentiation. This study was approved by retinal microcirculation changes during the treatment. the authonomic, ethical and clinical investigation committee. Patients in the study were both male and female and were aged 18 or older with grade 1 or grade 2 essential hypertension with novo or previous MATERIALS AND METHODS diagnosis of hypertension, who had uncontrolled BP regardless of whether The semiautomatic snakes method was developed to evaluate retinal they had other vascular risk factors and/or lesions in the target organ and blood vessel calibre, and it is composed of four main steps. First, the user were treated with telmisartan as monotherapy or in combination with a

1Department of Internal Medicine, Hospital de Conxo Medicine Deparment, Complexo Hospitalario Universitario de Santiago de Compostela, Spain; 2Department of Internal Medicine, Hospital Abente y Lago, ACorunã, Spain; 3Department of Internal Medicine, Hospital Barbanza, Ribeira, Spain; 4Department of Internal Medicine, Hospital San Rafael, A Corunã, Spain; 5Department of Internal Medicine, Hospital Lucus Augusti, Lugo, Spain; 6Department of Electronics and Computation (Artificial Vision Group) University of A Corunã, A Corun˜ a, Spain; 7Department of Econometrics, University of Santiago de Compostela, Santiago de Compostela, Spain; 8Assistance Healthcare Institute Angle`s, Girona, Spain and 9Department of Ophthalmology, Hospital de Conxo, Complexo Hospitalario Universitario de Santiago de Compostela, Santiago de Compostela, Spain. Correspondence: Professor P-R Antonio, Department of Internal Medicine, Hospital de Conxo, Complexo Hospitalario Universitario de Santiago de Compostela, Ruá Ramoń Baltar s/n, Santiago de Compostela 15706, Spain. E-mail: [email protected] Received 21 June 2013; revised 31 July 2013; accepted 19 August 2013; published online 21 November 2013 Factors associated with changes in retinal microcirculation P-R Antonio et al 311 diuretic following daily standard practice. All patients gave their written in retinal microcirculation was considered to exist when the arteriovenous informed consent. ratio had increased between the final visit and the baseline visit). Patients whose participation was not recommended in the opinion of Every eye fundus photograph was taken by a local researcher from the the physician investigator or for whom it was not possible to take a digital hospital where the patient was being treated and was analysed by local photograph of the eye fundus were excluded from the study. Also research and by an expert from the main research centre. Concordance excluded were those for whom the treatment was contraindicated in the between the two evaluators regarding the arteriovenous ratio and the data sheet—namely, diabetic patients with advanced diabetic retinopathy. artery and venous diameters was assessed using the Bland–Altman The observation period of the patients was 6 months. Recommendations method, along with a histogram of the difference between measurement in the Telmisartan’s data sheet were observed, and dosage was adjusted to and the Spearman correlation coefficient between measurements. the requirements of each patient and as per medical opinion. If the patient Statistical analyses were carried out using the SAS version 9.2 statistical did not reach target values for BP control, the dose of Telmisartan was package. titrated each month during the study period and if necessary thiazide was added. The decision to prescribe the study drug must have been taken earlier and before suggesting to the patient the possibility of participating in the RESULTS study and must not have been influenced by the patient’s potential A total of 189 patients were included. Table 1 shows the participation in the study. anthropometric characteristics of the sample. Not considered in our study was a washout period in patients who had Of the patients, 53% were male; at baseline, 114 (52.5%) undergone previous antihypertensive treatment because this is a prospective observational study conducted under conditions of routine patients were seen to suffer from dyslipidaemia, 58 (26.7%) were clinical practice. smokers and 32 (14.7%) were diabetic. In 28 patients, a digital Calculation of the sample size was established on the basis of both the photograph of the eye fundus could not be taken, as there was no main objective and the main variable of the study. On the basis of previous adequate fundus camera available. studies3 from a binomial distribution, an estimation was made of the In all, 66.8% of patients received a dose of 80 mg on completion proportion of hypertensive patients treated with telmisartan where the of the study and 27.9% received a dose of 160 mg; 60.9% of retinal arteriovenous ratio increased after 6 months of treatment with a patients were given a fixed combination of Telmisartan with confidence interval of 95%. Data were registered in a Case Report Form 25 mg of Thiazide; 15.7% of patients took associated beta-blockers (CRF) and later introduced in a database. and 2.6% took calcium antagonists; 74.0% of patients had BP The main objective of this study was to determine the factors related to changes in retinal microcirculation after antihypertensive treatment in a under control after a 6-month follow-up and 71.1% of patients group of newly diagnosed hypertensive patients or in patients with were treated with statins throughout the follow-up. uncontrolled hypertension; it also aimed to evaluate the relation between Average baseline systolic BP was 154.6 ±14.6 mm Hg and changes in retinal microcirculation, measured as average arteriole 132.4±10.8 mm Hg at the end of the study. Average baseline diameter, average venular diameter and average arteriovenous ratio, and diastolic blood pressure was 91.7±9 mm Hg and 79±7.2 mm Hg control of hypertension after 6 months of treatment, using software at the end of the study (P 0.0001) (Table 2). 1 o previously validated by our team. Table 3 shows analytical modifications throughout the follow- up stage. Table 4 shows the changes in the arteriovenous ratio (artery/ Statistical methodology vein ratio, AVR). AVR significantly increased in both eyes after 6 Categorical variables were measured by using both relative and absolute months of follow-up. frequencies. The average, s.d., median and minimum and maximum values, Changes in the AVR were caused by increases in arterial including the total number of valid values, were used to describe the diameter (right eye: 5.998±1.539 vs 6.225±1.638, Po0.0001, and continuous variables. left eye: 6.084±1.630 vs 6.287±1.654, P:0.0027), with a significant In order to compare patient subgroups, nonparametric tests (Mann– increase in the venous diameter of the right eye (7.817±1.859 vs Whitney or Kruskal–Wallis) were used for qualitative variables. The 7.837±2.066; P: 0.0025), with no significant changes in the left eye comparison of quantitative variables in matched pairs was made from (7.905±1.912 vs 7.907±1.997). nonparametric tests (Wilcoxon o Friedman), in accordance with the specific characteristics of the variables in the study. In the case of qualitative The AVR increased in 88% of patients in whom BP (systolic and/ variables, the Chi-Square test was performed and the McNemar test was or diastolic) decreased after 6 months of treatment. used for matched pairs. Correlations were observed among all measurements of A logistic regression model was developed to find potential factors both eyes. They were positive and statistically significant from associated with the changes observed in retinal microcirculation (a change zero (Po0.05). Furthermore, between-evaluator correlations were

Table 1. Anthropometric Characteristics

Average s.d. Median Min. Max P-valuea

Age (years) 56.3 11.1 57 27 84 0.5796 Male 55.9 11.2 56 27 80 Female 56.8 11.1 57 31 84 Weight (Kg) 79.9 15.4 79 49.5 146.8 o0.0001 Male 85.8 16.1 85 58 146.8 Female 73.2 11.4 71 49.5 102 Height (cm) 164.4 9.3 164 142 196 o0.0001 Male 170.0 7.2 170 148 196 Female 158.1 7.0 158 142 178 Waist circumference (cm) Á Visit 1 95.032 13.74 96 56 141 0.007 Waist circumference (cm) Á Visit 2 94.307 13.92 94 56 142 BMI (Kg m À 2) Á Visit 1 29.32 4.65 29.2 20.7 44.0 0.121 BMI (Kg m À 2) Á Visit 2 29.218 4.59 28.81 20.43 42.93 aWilcoxon test.

& 2014 Macmillan Publishers Limited Journal of Human Hypertension (2014) 310 – 315 Factors associated with changes in retinal microcirculation P-R Antonio et al 312 Table 2. Blood Pressure modiﬁcations throughout the follow-up stage

Ofﬁce systolic blood pressure Medium DT Median Minimum Maximum P-valuea

Baseline visit 154.6 14.6 150.5 117.5 211 o0.0001 Follow-up visit 132.4 10.8 131 104 175

Ofﬁce diastolic blood pressure Medium DT Median Minimum Maximum P-valuea Baseline visit 91.7 9 91.5 70 122.75 o0.0001 Follow-up visit 79 7.2 78.5 58 99 aWilcoxon test.

Table 3. Analytical modiﬁcations throughout the follow-up stage

Baseline Visit Visit 6 P-valuea

Average (s.d.) Median Range Average (s.d.) Median Range

Cholesterol total (mg dl À 1) 203.2 (43.2) 201.5 112–400 192.9 (42.1) 190,5 116–344 0.0052 Cholesterol—LDL (mg dl À 1) 130.3 (37.1) 129 46–236 119.2 (34.7) 115 53–230 o0.0001 Cholesterol—HDL (mg dl À 1) 46.7 (13.5) 44 19–132 46.7 (13.1) 44 26–131 0.9999 Triglycerides (mg dl À 1) 127.9 (73.3) 111.5 21–557 137.8 (85.1) 118 29–576 0.5117 Fibrinogen (mg dl À 1) 304.2 (82.6) 302 103–515 311.5 (105.2) 309 103–545 0.8341 Glucose (mg dl À 1) 108.5 (29.5) 102 70–292 105.3 (24.7) 101 66–225 0.0351 Urea (mg dl À 1) 41.4 (12.9) 40 14–102 43.6 (14.6) 42 14–119 0.0243 Uric acid (mg dl À 1) 5.8 (1.6) 5.8 2.2–11.5 6.0 (1.7) 5.9 2.1–11.8 0.0071 Creatinine (mg dl À 1) 0.9 (0.2) 0.9 0.4e–1.7 1.0 (0.2) 0.9 0.4–1.9 0.0019 Albumin/creatinine ratio (mg dl À 1) 17.3 (20.4) 14.3 0.9–157.1 20.8 (59.2) 14.3 1.5–500 0.0018 HbA1c (%) 5.9 (1.2) 5.6 4.5–11.7 5.7 (0.9) 5.5 4.6–9.4 o0.0001 aWilcoxon test.

Table 4. AVR changes with the snakes method

Baseline visit Visit 6 P-valuea

n Average (s.d.) Median Range n Average (s.d.) Median Range

AVR right eye 187 0.769 (0.065) 0.767 0.559–0.960 189 0.799 (0.066) 0.803 0.640–1.020 o0.0001 AVR left eye 186 0.770 (0.065) 0.764 0.638–1e.000 187 0.796 (0.071) 0.802 0.600–1.102 o0.0001 aWilcoxon test.

observed in both eyes at every measurement at both the baseline No statistical differences were observed in the decrease in LDL visit and the 6-month follow-up. They were positive and cholesterol and the increase in HDL cholesterol as a function of statistically significant (Po0.05). the change in AVR (Mann–Whitney U test; P40.05). The Bland–Altman analysis provided the agreement limits for With regard to statin taking, for the per-protocol-assessed the difference in measurements of the arteriole diameter recorded sample of patients, statistically significant differences were by the evaluators for the right eye at baseline visit (95% observed only for the AVR of the left eye. At the 6-month CI: À 1.1178 to 1.2114) and for the left eye (95% CI: À 1.1287 to follow-up, those patients who were taking statins at the start 1.1343), and after 6 months of follow-up for the right eye (95% of the study showed greater values in the AVR (Mann–Whitney CI: À 1.5631 to 1.8594) and for the left eye of (95% CI: À 1.5834 U test; Po0.05). to 1.7250). Finally, a study was conducted to gain an understanding of the Although a tendency was observed that clearly associates a potential factors associated with the changes observed in retinal decrease in SBP and DBP in a clinical situation with an increase in microcirculation in each eye separately. A change in retinal AVR (Table 5), no statistical correlation was confirmed between microcirculation was considered to exist when the AVR increased the level of decrease in both systolic and diastolic BP and the level between the final and the baseline visit. of increase in retinal AVR experienced by the patient throughout To this end, a logistic regression model was adjusted with a the follow-up, which suggests that this relation is not linear. We bilateral significance level of 0.05 for all statistical tests. Selection have not found a significant correlation between the decrease in of the associated factors was made using a method of backward BP and the increase in arteriole and venular diameter during steps, using the conditional criterion. follow-up. According to the characteristics analysed for the association of No significant differences were found in the increase in AVR changes observed in retinal microcirculation, having received a among patients with better control of blood pressure—among treatment with diuretics before inclusion in the study was those with BP below 140/90 mm Hg, 135/90 mm Hg and 135/ associated with a change in AVR at the 6-month follow-up with 85 mm Hg. respect to patients who had not been previously treated with

Journal of Human Hypertension (2014) 310 – 315 & 2014 Macmillan Publishers Limited Factors associated with changes in retinal microcirculation P-R Antonio et al 313 Table 5. Contingency table relating AVR to BP in both eyes

We consider relevants increase or decrease over 1% so in blood pressure as AVR

Differences SBP Differences SBP 12 -=+Total - = + Total

+610162 +670269 Dif Dif Left = 500 5 Right =1001 AVR AVR - 301 4 -5016

Total 69 0 2 71 Total 73 0 3 76

Differences DBP Differences DBP 34 -=+Total - = + Total

+552562 +611769 Dif Dif Left = 500 5 Right =1001 AVR AVR - 301 4 -4116

Total 63 2 6 71 Total 66 2 8 76

diuretics OR: 2.805, (95% CI: : 1.26–6.242; P ¼ 0.0115). However, at compromising the parenchyma,15 and only quite recently have the beginning of the study, there were no differences in blood some studies involving very large groups of subjects,16–18 pressure levels between the patients who had received treatment although not others,4 provided evidence that retinal arteriole with diuretics and those who had not. narrowing may predict adverse cardiovascular events inde- pendently of other risk factors. Because of this, and in view of the limitations of qualitative funduscopy for evaluation of patients DISCUSSION with grade I and II hypertension,19 the interpretation of the results The first alteration we observe in the fundus of the eye of of funduscopy recommended by the guidelines is based largely hypertensive patients consists of a narrowing of the artery on studies carried out several decades ago in situations in calibre,4 and indeed as retinal microcirculation can be easily which, for various reasons, interpretation of early hypertensive observed and assessed noninvasively it offers an opportunity for retinopathy was less relevant than it is today.20,21 detailed in vivo study of the structure of small arterioles and Therefore, there is clearly a need for the development of venules. Hence, it has been described in association with objective methods of evaluation of early hypertensive retinopathy, processes such as hypertension and diabetes.5 retinal arteriole narrowing in particular. A recent paper6 argues that the vasculature of the retina is an Hughes et al. reported that the use of quantitative techniques excellent biological model for studying manifestations of extend previous observations by showing that antihypertensive hypertension, and standardized measurement protocols have to treatment with lisinopril or amlodipine in patients with untreated be supplemented with new technology for more objective and essential hypertension is associated with a marked improvement accurate assessment of changes in the retinal vasculature, in arteriolar narrowing and rarefaction, accompanied by a allowing clinical use of images for cardiovascular risk assessment. reduction in BP.22 Systemic associations and clinical significance of these retinal Some years ago, we described for a smaller sample of patients3 abnormalities are not fully understood. Several studies have and a similar time framework an increase in AVR only at the expense shown that these lesions could be associated with increased of the arteriole diameter, expressing the improvement at the retinal cardiovascular mortality in persons with hypertension1,7,8 and be microcirculation level following antihypertensive treatment with an related to an increased number of cerebral white matter lesions9 ARB (Losartan). We were able to see these changes using a and cognitive impairment.10,11 semiautomatic method, which we described and validated for the However, retinal microvascular lesions are also seen in people determination of the retinal vascular calibre and the AVR.1,3 without a history of hypertension and could predict incident This method, ART-VENA, which was a linear method, based on stroke independent of blood pressure.12,13 mathematical models was extremely sensitive for detecting retinal In the last few years, it has been suggested that digital fundus microvessels but lacked similar specificity, and consequently many photographs are useful for observing the retina in order to stratify of the images detected as vessels by the linear method had to be risk and for determining the prognosis of hypertensive and discarded by the observer. We have subsequently introduced a diabetics patients, at least in middle-aged populations, although modification to the method by applying a specialized deformable the connection between initial changes in retinal microcirculation method (snakes) that allows for greater specificity while main- and damage in other target organs has not been definitely taining sensitivity.2 confirmed.14 Hypertensive retinopathy covers a wide range of The current study confirms for a greater group of patients alterations, from narrowing of the retinal arterioles to lesions homogeneously treated with antihypertensive drugs that block

& 2014 Macmillan Publishers Limited Journal of Human Hypertension (2014) 310 – 315 Factors associated with changes in retinal microcirculation P-R Antonio et al 314 the renin–angiotensin system such as Telmisartan (associated or circulation. In this study, daily administration of simvastatin for 7 not to diuretics) the possibility of detecting an increase in the days significantly increased blood velocity and blood flow in retinal arteriole diameter, which suggests a modification of the retinal arteries and veins but did not significantly change the structure of retinal arterioles in a short period of time – namely, vessel diameter.25 6 months. On the other hand, the usefulness of this method is Simvastatin also significantly increased the plasma nitrite/ confirmed for the determination of the calibre of the retinal nitrate levels, and this elevation may be associated with the vascular tree with minimum between-observer variability. increase in blood flow in the retinal vessels. There are currently other methods for studying retinal As most of our patients had taken statins, changes to arterioles microcirculation and its changes with treatment, such as measure- may be partly attributed to statins. In this regard, although no ment of blood flow rate at the level of the retinal central artery.23 connection was found between the changes in LDL-c and HDL-c Other more sophisticated techniques would be the computer- levels and modifications to retinal vascular calibre, a connection based multi-scale retinal image analysis (RISA) system and was indeed found between statin taking and improvement in computed tomographies of the macula and optic nerve, which retinal microcirculation but only in one of the eyes. allow (by using laser) bi-tridimensional tomographic studies of The only factor that could be associated in the multi-variant in vivo structures of the retina and the optic nerve. These analysis to the change in AVR was the treatment with diuretics techniques have an important limitation; that is, they do not have previous to the patients’ inclusion in the study. We believe that applicability for clinical practice because, as more expensive, they there could be two reasons to explain this. Diuretics, along with cannot be applied daily, unless there is the possibility of a specific beta-blockers, are known to be the antihypertensive drugs with the pathology, and they require to be interpreted by a specialist. Our least effect on the general microcirculation; the introduction of a software allows an interpretation of the results in a relatively renin–angiotensin system blocker could have rendered this effect simple way; in addition, it is inexpensive and has a reliability that more apparent. The second hypothesis is that the renin–angiotensin has already been proven in previous studies.2 system in retinal microcirculation could be activated by previous Our technique is based on the application of a software for the treatment with diuretics, placing retinal microcirculation under semiautomatic analysis of a digital photograph of the eye fundus optimal conditions for the effect of its blocking by the angiotensin centred on the papilla and with no mydriasis. This allows for a receptor blocker. This finding would therefore support the remote analysis of any photograph of a patient that is taken with hypothesis of the scarce action of the diuretic on microcirculation. a non-mydriatic retinal camera under these conditions. Therefore, Our study has limitations such as the fact that, as it is a study on its clinical applicability for the morphological study of retinal clinical practice, we do not have a normotensive control group. In microcirculation is perhaps the most interesting aspect of this previous studies conducted by our group, although we have been technique. able to find that in normotensives, and even more so in The algorithm applied by our software to measure AVR has hypertensives, AVR decreases with age, this decrease is more been compared previously with the methods proposed by Parr– apparent in patients aged 40 or older at a rate of two microns per Hubbard. We have analysed the results in a database with 400 decade.1 This makes it highly unlikely that regression to the mean images. The vessel widths were automatically measured by our can account for this arteriolar increase observed within 6 months system, whereas an expert classified the vessels and selected the in this group of hypertensive patients. subset of segments for the calculus. The correlation of our method In reference to the election of antihypertensive treatment, we with the Parr–Hubbard and the revised Parr–Hubbard approaches have thought that all hypertensives should take the same was 0.94 and 0.91, respectively. The correlation between the Parr– medication, because the effect over the microcirculation is not Hubbard and the revised Parr–Hubbard algorithms was 0.95.24 the same for the different antihypertensive groups, even getting Changes in absolute values of the AVR are very small, being a the same antihypertensive control. We selected angiotensin ratio between microvessels of similar calibre. We believe the utility receptor blockers, because they are the most prescribed of the method would be the possibility of assessing the evolution antihypertensives in our area, and they have shown the maximal of the AVR throughout time, which seems to be more important effect on the microcirculation.26 than its absolute value. As we have shown in our study, almost all We chose Telmisartan because it is a long plasma half-life drug, patients (88% of them) in whom BP (systolic and/or diastolic) reaching almost 24 h. It has the highest volume distribution decreased after 6 months of follow-up showed an increase in AVR. among ARB-II. This means that it penetrates the tissues and blocks The application to clinic practice would be evaluating the variation the renin–angiotensin–aldosterone system (RAAS) both locally of the AVR throughout a patient’s follow-up and assessing and systemically. whether there are positive changes within 6 months if BP is It is also surprising that, in a population where hypertension is under control; that is, rather than absolute values, it is a matter of well under control and treated with RAAS blockers, microalbumi- showing an increase in AVR after antihypertensive treatment. nuria increases. If we pay attention to the data, we will see that the Although a tendency was observed that associated the decrease s.d. from the mean is very high and also that, as shown, there were in BP throughout follow-up and changes to the retinal AVR, no no changes to the median with respect to baseline. Consequently, statistically significant correlation could be found between the these results are not conclusive (Table 3). quantitative changes in both variables. In other words, a greater The changes in the arteriovenous ratio could express the decrease in BP throughout the study could not be associated in improvement in the structure of the retinal microcirculation. The any statistically significant manner with a greater increase in AVR at evolution of these changes at the initial phase of hypertension the end of the study, although a decrease in BP seems to be a could translate into an improved prognosis for the hypertensive necessary condition for an increase in retinal AVR. This suggests patient by preventing brain microvascular complications such as that, as well as an adequate control of BP, there are other cognitive impairment or periventricular lacunar infarct, as well as nonestablished variables that may have an influence on changes to an improvement in the prognosis of coronary disease; however, retinal microcirculation. Among the other factors is the retinal the latter should be confirmed in an ad hoc study. renin–angiotensin system and hence we cannot exclude a direct vasodilator effect of Telmisartan in the retinal microcirculation. In this regard, the usefulness of statins for controlling vascular CONCLUSION disease has been reported, as well as their anti-inflammatory In conclusion, we confirm that it is possible to detect changes in properties that can modify the structure of the vessel wall. the retinal microcirculation with antihypertensive treatment Nagaoka et al. investigated the effects of simvastatin on the retinal within a short period using a previously validated simple,

Journal of Human Hypertension (2014) 310 – 315 & 2014 Macmillan Publishers Limited Factors associated with changes in retinal microcirculation P-R Antonio et al 315 objective and economical method. Although there is an inverse 5 Liew G, Sharrett AR, Wang JJ, Klein R, Klein BEK, Mitchell P et al. Relative impor- tendency between blood pressure and arteriovenous ratio tance of systemic determinants of retinal arteriolar and venular caliber. Arch changes, we could not find a linear correlation between these Ophthalmol 2008; 126(10): 1404–1410. changes, suggesting that other factors could be involved in the 6 Cheung CY, Ikram MK, Sabanayagam C, Wong TY. Retinal microvasculature as a model regression of retinal changes. to study the manifestations of hypertension. Hypertension 2012; 60: 1094–1103. 7 Wong TY, Kamineni A, Klein R, Sharret AR, Klein BE, Siscovick DS et al. Quantitative retinal venular caliber and risk of cardiovascular disease in older persons: the cardiovascular health study. Arch Intern Med 2006; 166: 2388–2394. 8 Ikram MK, de Jong FJ, Vingerling JR, Witteman JC, Hofman A, Breteler MM et al. What is known about the topic Are retinal arteriolar or venular diameters associated with markers for cardio- Hypertensive retinopathy is considered by hypertension management vascular disorders? The Rotterdam Study. Invest Ophthalmol Vis Sci 2004; 45: guidelines as target organ damage constituting a cardiovascular risk 2129–2134. factor. 9 Longstreth Jr W, Larsen EK, Klein R, Wong TY, Sharrett AR, Lefkowitz D et al. Retinal microcirculation findings in hypertensive patients are sub- Associations between findings on cranial magnetic resonance imaging and retinal jective, imprecise and poorly reproducible. photography in the elderly: the Cardiovascular Health Study. Am J Epidemiol 2007; We have previously described a semiautomatic method to detect 165: 78–84. early modifications in retinal microcirculation in hypertensive patients 10 Baker ML, Marino Larsen EK, Kuller LH, Klein R, Klein BE, Siscovick DS et al. and their regression with antihypertensive treatment. Retinal microvascular sigs, cognitive function, and dementia in older persons: the Cardiovascular Health Study. Stroke 2007; 38: 2041–2047. What this study adds 11 Hajjar I, Quach L, Yang F, Chaves PHM, Newman A, Mukamal K et al. Although control in hypertension seems necessary, there is no linear Hypertension, White matter hyperintensities and concurrent impairments in relationship between retinal microcirculation improvement (measured mobility, cognition, and mood. The Cardiovascular Health Study. Circulation 2011; by the arteriovenous index) and decreased blood pressure, indicating 123: 858–865. that there could be other factors involved in the modification of the 12 Wong TY, Klein R, Couper DJ, Couper LS, Shahar E, Hubbard LD et al. Retinal retinal structure as well as the level of hypertension. microvascular abnormalities and incident stroke: the Atherosclerosis Risk in Control of the renin–angiotensin system could be an important Communities Study. Lancet 2001; 358: 1134–1140. contributory factor in these changes, in addition to hypertension 13 Wong TY, Klein R, Sharrett AR, Manolio TA, Hubbard LD, Marino EK et al. control, and hence we cannot exclude a direct vasodilator effect of The prevalence and risk factors of retinal microvascular abnormalities in older Telmisartan in the retinal microcirculation. persons. The Cardiovascular Health Study. Ophthalmology 2003; 110: 658–666. 14 Pose-Reino A, Rodriguez Fernandez M, Hayik B, Gomez-Ulla Irazabal F, Carreira Nouche MJ, Gonzalez Barcala J et al. Relation between changes in retinal microcirculation and other organ damage in hypertension and regression with antihypertensive treatment. Med Clin 2007; 128: 605–608. 15 Wong TY, Mitchell P. Hypertensive retinopathy. NEnglJMed2004; 351: 2310–2317. CONFLICT OF INTEREST 16 Wong TY, Hubbard LD, Klein R, Marino EK, Kronmal R, Sharret AR et al. Retinal The authors declare no conflict of interest. microvascular abnormalities and blood pressure in older people: the Cardiovas- cular Health Study. Br J Ophtalmol 2002; 86: 1007–1013. 17 Sharrett AR, Hubbard LD, Cooper LS, Sorlie PD, Brothers RJ, Nieto FJ et al. Retinal ACKNOWLEDGEMENTS arteriolar diameters and elevated blood pressure: the Atherosclerosis Risk in Communities Study. Am J Epidemiol 1999; 150: 263–270. This study was partly financed by Project ETES-FIS 08/90459 of the Fondo de 18 Wong TY, Knudtson MD, Klein R, Klein BE, Hubbard LD. A prospective cohort study Investigacio´ n Sanitaria (FIS); by the Spanish Ministry of Science & Innovation; and by of retinal arteriolar narrowing and mortality. Am J Epidemiol 2004; 159: 819–825. Project PS08/05 of the Department of Health of the Xunta de Galicia. 19 Dimmitt SB, West JN, Eames SM, Gibson JM, Gosling P, Littler WA. Usefulness of ophthalmoscopy in mild to moderate hypertension. Lancet 1989; 1: 1103–1106. 20 Keith NM, Wagener HP, Barker NW. Some different types of essential hyperten- REFERENCES sion: their course and prognosis. Am J Med Sci 1939; 197: 332–343. 1 Pose-Reino A, Gomez-Ulla F, Hayik B, Rodrı´guez Fernańdez M, Carreira Nouche MJ, 21 Breslin DJ, Gifford Jr RW, Fairbairn II JF, Kearns TP. Prognostic importance oph- Mosquera-Gonza´lez A et al. Computerized measurement of retinal blood vessel thalmoscopic findings in essential hypertension. JAMA 1966; 195: 335–338. calibre: description, validation and use to determine the influence of ageing and 22 Hughes AD, Stanton AV, Jabbar AS, Chapman N, Martinez-Perez ME, McGThom hypertension. J Hypertens 2005; 23: 843–850. SA. Effect of antihypertensive treatment on retinal microvascularchanges in 2 Pose-Reino A, Pena-Seijo M, Gonza´lez Penedo M, Ortega Hortas M, hypertension. J Hypertens 2008; 26: 1703–1707. Rodriguez Blanco M, Vega P et al. Estimation of the retinal microvascular calibre in 23 Martinez A, Gonzalez F, Capeans C, Perez R, Sanchez-Salorio M. Dorzolamide hypertensive patients with the snakes semiautomatic model. Med Clıń 2010; effect on ocular blood flow. Invest OphthalmolVis Sci 1999; 40: 1270–1275. 135(4): 145–150. 24 Rodriguez M. Development of an automatic system to measure retinal vascular 3 Pose-Reino A, Rodrı´guez Fernańdez M, Hayik B, Gomez-Ulla F, Carreira Nouche MJ, calibre. Evolution from a manual system to the automation. Doctoral Thesis. Surgery Gonza´lez-Barcala J et al. Regression of alterations in retinal microcirculation Department. University of Santiago de Compostela. A Corunã: Spain, 2013. following treatment for arterial hypertension. J Clin Hypertens 2006; 8: 25 Nagaoka T, Takahashi A, Sato E, Izumi N, Hein TW, Kuo L et al. Effect of Systemic 590–595. Administration of Simvastatin on Retinal Circulation. ArchOphthalmol 2006; 124: 4 Cuspidi C, Meani S, Salerno M, Fusi V, Severgnini B, Valerio C et al. Retinal 665–670. microvascular changes and target organ damage in untreated essential hyper- 26 Schiffrin EL. Remodeling of resistance arteries in essential hypertension and tensives. J Hypertens 2004; 22: 2095–2102. effects of antihypertensive treatment. Am J Hypertens 2004; 12(pt 1): 1192–1200.

& 2014 Macmillan Publishers Limited Journal of Human Hypertension (2014) 310 – 315