Microvascular endothelial function and severity of primary open angle glaucoma

Abstract

Purpose

The role of microvascular endothelial dysfunction on severity of primary open angle glaucoma (POAG) was investigated in this study.

Patients and methods

A prospective cohort study was conducted. One hundred and fourteen ethnically Malay patients (114 eyes) with POAG treated at the eye clinic of Hospital University Sains Malaysia between April 2012 and December 2014 were recruited. Patients aged between 40 and 80 years with two consecutive reliable and reproducible Humphrey visual field 24-2 analyses were selected. Patients who were diagnosed with any other type of glaucoma, previous glaucoma-filtering surgery, or other surgeries except uncomplicated cataract and pterygium surgery were excluded. Humphrey visual field analysis 24-2 was used to stratify the severity of glaucoma using Advanced Glaucoma Intervention Study (AGIS) score at the time of recruitment. Microvascular endothelial function was assessed using Laser Doppler fluximetry and iontophoresis. Iontophoresis process with acetylcholine (ACh) and sodium nitroprusside (SNP) was used to measure microvascular endothelium-dependent and -independent vasodilatation, respectively.

Results

Based on the AGIS score, 55 patients showed mild glaucoma, with 29 moderate and 30 severe. There was statistically significant difference in microvascular endothelial function (ACh% and ACh_max) between mild and moderate POAG cases (P=0.023) and between mild and severe POAG cases (P<0.001). There was negative correlation between microvascular endothelial function and severity of POAG (r=−0.457, P<0.001).

Conclusion

Microvascular endothelial dysfunction may have a role in influencing the severity of POAG in Malay patients.

Introduction

Glaucoma is a complex multi-factorial disease leading to irreversible blindness. The pressure-dependent (mechanical) and pressure-independent (vascular) theories have been postulated to have important role in glaucomatous optic neuropathy.¹ Pressure-independent mechanism holds imbalance of the optic nerve head (ONH) perfusion responsible for ischemia in the retinal nerve fiber layer (RNFL).^{2, 3}

ONH perfusion may be affected by ocular perfusion pressure (OPP). OPP is dynamic but can be mathematically calculated using systemic blood pressure and intraocular pressure (IOP). OPP is affected by the resistance to flow and regulated by the size (caliber) of vessel lumen.⁴ The advancement of technology has made the visualization of ocular blood flow possible using various techniques: color Doppler ultrasound imaging,⁵ scanning laser ophthalmoscopic angiography,⁶ confocal scanning laser Doppler flowmetry,⁷ and retinal vessel analyzer (RVA).⁸ These techniques measure different aspect of ocular blood flow but are not specific to quantify the blood flow to ONH. There was also a lack of gold standard as reference. In addition, most techniques are expensive and require clear media and highly skilled, experienced technician.^{9, 10, 11, 12, 13} For example, RVA provides direct visualization of pulsation of large retinal vessels throughout cardiac cycle without measuring actual blood flow.¹⁴ RVA is used to quantify microvascular changes in systemic and ocular diseases.^{14, 15} There were evidences of reduction in vascular diameter in glaucoma patients.^{15, 16}

Autoregulation is believed to be responsible for maintaining ONH perfusion.^{17, 18} The basic mechanisms underlying autoregulation are still unclear. The proposed mechanisms include metabolic, myogenic, neurogenic, and humoral factors in the ocular circulation.^{19, 20} Endothelium acts as a mediator in regulating the vascular tone. Alterations in the local concentrations of metabolites such as oxygen, carbon dioxide, potassium, and hydrogen may influence ocular vascular tone.¹⁹ The hemodynamic response of the retinal vasculature to hyperoxia is thought to be mediated by endothelin.²¹ Endothelial-derived vasodilator factors are believed to be responsible for hypoxia-induced vasodilatation of the retinal vasculature.²² In general, myogenic autoregulation is considered to be mechanically independent of the endothelium. In the cerebral circulation, however, myogenic-induced vasoconstriction has also been suggested to be at least partly mediated by endothelial factors.²³ Myogenic regulation has been found in the ONH and retina.^{24, 25} Humoral control refers to the potential regulatory influence of numerous vasoactive agents present in the circulating blood. The effect is either through direct interaction with the vascular smooth muscle cells and pericytes or through mediation of endothelial cells causing changes in blood flow.²⁶

Microvascular endothelial function can be assessed non-invasively using laser Doppler fluximetry (LDF) and the process of iontophoresis. LDF is a device that permits real-time continuous measurement of microvascular perfusion.²⁷ Iontophoresis of acetylcholine (ACh) is used to assess endothelial-dependent vasodilatation, while sodium nitroprusside (SNP) iontophoresis assesses endothelial-independent vasodilatation. The LDF signal is a stochastic representation of the number of cells in the sample volume multiplied by their velocities. The output from the LDF is referred to as ‘perfusion flux'.²⁸ Microvascular perfusion flux is the product of the number of blood cells and the mean velocity of blood cells. Mean velocity is proportional to blood flow.

Assuming proportionality between blood cells number and blood volume, the LDF signal should be linearly related to blood flow in the measured volume. In most physiological and clinical situations, LDF flux and volume flow have a good correlation. Many studies have demonstrated a good relationship between LDF measurements and other method of blood flow measurement, such as plethysmography.²⁹ In LDF, the perfusion flux is expressed in terms of arbitrary unit, which do not gives the absolute blood perfusion unit. Measurement of blood perfusion by LDF is therefore less valuable in comparing absolute values of blood perfusion between individuals. However, intraindividual and interindividual comparison of dynamic response to standardized stimuli such as transient ischemic block and local warming can be made.

Microvascular endothelial dysfunction may accelerate the glaucomatous damage especially in the presence of uncontrolled IOP.^{30, 3} More pronounced reduction of ONH blood flow velocities was noted in glaucoma patients with visual field progression when compared with those with non-progression.³¹ Eyes with progressive visual field defects in patients with normal-tension glaucoma had lower blood flow velocities and higher resistive index in the central retinal artery and the short posterior ciliary arteries than did those with stable visual field defects.^{31, 32} There is little knowledge on glaucoma in the Malay population. In addition, there is evidence of the effects of ethnicity on clinical presentation and progression of glaucoma.^{33, 34} The objective of this study was to determine the association between microvascular endothelial dysfunction and severity of primary open angle glaucoma (POAG) in Malay patients. Predictors affecting microvascular endothelial dysfunction in Malay patients with POAG were also evaluated.

Materials and methods

A prospective cohort study was conducted involving 114 Malay patients diagnosed with POAG (114 eyes). These patients were recruited from the glaucoma clinic of Hospital Universiti Sains Malaysia between April 2012 and December 2014. Written consent was obtained after full explanation was given to patients. This study received approval from the research and ethics committee of the School of Medical Sciences, Universiti Sains Malaysia and was conducted in accordance to the Declaration of Helsinki for human research.

POAG is defined according to Asia Pacific Glaucoma Guideline.³⁵ Patients aged between 40 and 80 years with two consecutive reliable (false positive and negative <33%, fixation loss <20%) and reproducible Humphrey visual field 24-2 analyses were selected. Visual field assessment was either conducted within 3 months from the recruitment period or obtained from the medical record of the selected patients. Visual field assessments obtained from the medical record were obtained no more than 6 months before the recruitment period. Patients who were diagnosed with any other type of glaucoma (eg, steroid induced, pseudoexfoliative, and pigmentary glaucoma) were excluded. POAG patients presented with history of previous glaucoma-filtering surgery or other surgeries except uncomplicated cataract and pterygium surgery were excluded. Those with underlying ocular conditions which may interfere with visual field interpretation, were also excluded: media opacities, including cataract (visual acuity <6/60), vitreous hemorrhage, post-pan-retinal photocoagulation, tilted disc, and retinal disease. Those with uncontrolled IOP or failed to achieve target pressure were also excluded. If both eyes were deemed eligible, only the right eye was selected regardless of the severity of glaucoma.

Assessment of the severity of POAG was based on modified Advanced Glaucoma Intervention Study (AGIS) scoring system on Humphrey visual field 24-2 analysis.³⁶ Only one of the two consecutive reliable and reproducible visual fields was selected for AGIS scoring. Primary scoring was conducted by the primary investigator (SMIB) and secondary scoring by the glaucoma specialist (L-SAT). Both investigators were masked. The final score was based on the average of the scoring of two investigators. Glaucoma patients were stratified into mild (0–5), moderate (6–11), and severe (12–20) categories based on the modified scoring system. IOP was also obtained during the recruitment period using Goldmann applanation tonometry (Haag Streit, Koeniz, Switzerland) at sitting position by another investigator (KKY) between 0800 and 1200 hours.

Selected patients were asked to refrain from smoking, eating salty foods, and drinking alcohol and caffeinated beverages for at least 10 h before the vascular measurement.^{37, 38} They were also required to fast for at least 8 h prior to the evaluation and were advised to wear light and loose clothing. Upon arrival, 6 ml of blood were taken to assess the hematocrit level, fasting blood sugar (FBS), and total cholesterol level. Their height and weight were also obtained. Blood pressure was taken using automated sphygmomanometer (Omron, Kyoto, Japan) in sitting position by the primary investigator (SMIB).

Evaluation of microvascular endothelial dysfunction was conducted in a temperature-controlled room (23–24 °C) in the vascular laboratory, Universiti Sains Malaysia. Dual-channel DRT4 LDF (Moor Instruments, Axminster, UK) with DPIT-V2 skin laser probes (Moor Instruments) was used to measure skin perfusion during iontophoresis. LDF generates a low intensity beam of coherent infrared monochromatic 780 nm light. This light was delivered to the site of measurement by a flexible fiber optic probe, and it penetrates to a depth of 1–2 mm into the skin.³⁹ The blood cells in the tissue beneath the probe generate a continuous flux that is linearly correlated to the underlying blood flow.⁴⁰ When this light beam passes through tissues, it undergoes changes in wavelength as it hits moving blood cells. This change is called Doppler shift.⁴¹ LDF uses laser Doppler shift principle to measure perfusion of blood cells, mainly erythrocytes, in the skin microvasculature. Theoretically, blood flow, as estimated by LDF, is determined by the blood flow velocity and the number of moving red cell corpuscle within the surface microvessels of the skin.^{42, 43, 44}

Iontophoresis is a method for non-invasive transdermal drug delivery based on net movements of ions under the influence of small electric current.⁴⁵ This method is based on the principle that charges with the same signal repel and charges with opposite signals attract, facilitating the penetration of ions through the skin.⁴⁶ ACh powder (Fluka Chemie Gmbh, Tokyo, Japan) and SNP powder (Riedel-de Haen, C.O.O., Buchs, Switzerland) were used as endothelium-dependent and -independent vasodilators, respectively. Sodium chloride solution, 0.9% at a physiological strength of 0.154 M (Excel Pharmaceutical, Selangor, Malaysia) was used to dilute both drugs to obtain concentrations of 1% of ACh and 1% of SNP.^{47, 38} During the 10 min acclimatization period, patients were advised to lie down straight in bed to minimize the effect of body position. Body position may affect microvascular circulation. The patient's right forearm was exposed and supported by a hand supporter.

Two iontophoresis chambers attached to the surface of the forearm using double adhesive discs were used simultaneously to deliver the ACh (endothelium-dependent vasodilator) and SNP (endothelium-independent vasodilator) to the skin. The chambers were positioned at least 5 cm apart so that vasoactive response at one site would not influence the other site. The ACh chamber was attached to the anodal lead while the SNP chamber was attached to the cathodal lead. Each chamber was then filled with 0.4 ml of ACh or SNP, respectively, using a micropipette. Laser probes were carefully inserted with minimal pressure into the chambers. An iontophoresis protocol of five current pulses (lasting 2 min each) separated by 1 min current-free intervals was used.⁴³ A current strength of 0.007 mA, density of 0.01 mA/cm², and charge density of 6 mC/cm² were used.⁴³ The primary parameters used to assess the microvascular endothelial function were percentage of change in perfusion owing to ACh iontophoresis (% change) and maximum absolute change in perfusion owing to ACh iontophoresis (ACh_max) as given in these formulas:

To ensure reliability of the test, the intraday and interday coefficient of variation (CV) was measured for both ACh% and ACh_max on 10 healthy subjects at the start of study. The intraday and interday CV was 13.4 and 13.0%, respectively, for ACh%.

The data in this study were analyzed using the SPSS version 22.0 software (IBM SPSS, Inc., Chicago, IL, USA). The Pearson chi-square was used to analyze categorical data. One-way analysis of variance (ANOVA) was used to analyze ocular parameters. Endothelium-independent microvascular function according to the severity of POAG was analyzed using multiple analysis of variance (ANOVA). Factors previously reported to affect microvascular endothelial function include age, sex, body mass index, hematocrit percentage, total cholesterol levels, FBS levels, diabetes, hypertension, hyperlipidemia and systemic medicines namely, beta blockers, calcium channel blockers, angiotensin-converting enzyme inhibitors, statin and oral hypoglycemic agents. These were used as covariates for endothelial dysfunction. Simple linear regression was applied for covariates. Multiple analysis of covariance (MANCOVA) was used to assess the endothelium-dependent microvascular function and severity of glaucoma.

Informed consent

Informed consent was obtained from all individual participants enrolled in this study.

Ethical approval

All procedures performed in this study were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards.

Results

Based on modified AGIS scoring, Malay patients with POAG were divided into mild (55), moderate (29), and severe (30) stage. POAG patients with severe stage of glaucoma were significantly older than those with moderate or mild stage (Table 1). Systemic hypertension remains the most common systemic comorbidity among our Malay patients. There was a significantly higher percentage of mild severity of POAG with systemic hypertension (Table 1). As expected, significant thinner RNFL, larger average cup disc ratio, larger vertical cup disc ratio, and depressed visual field were observed in patients with severe stage of POAG (Table 1). There was no significant difference in IOP during the recruitment period according to severity of POAG (Table 1).

Table 1. Comparison of demographic and clinical data between mild, moderate, and severe POAG in Malay patients.

Demographic characteristics	Mild, N=55	Moderate, N=29	Severe, N=30	P-value
Sex (n (%)
Male	37 (52.9%)	15 (21.4%)	18 (25.7%)	0.373a
Female	18 (40.9%)	14 (31.8%)	12 (27.3%)
Age, years (mean (SD)	64.2 (8.5)	68.6 (8.9)	70.1 (7.4)	0.004b
Systemic diseases (n, %)
Hypertension	39 (51.3%)	23 (30.3%)	14 (18.4%)	0.019a
Diabetes mellitus	23 (54.8%)	12 (28.6%)	7 (16.7%)	0.202a
Hyperlipidemia	27 (57.4%)	13 (27.7%)	7 (14.9%)	0.063a
Body mass index, kg/m2 (mean (SD)	25.2 (4.3)	24.3 (3.5)	24.3 (4.9)	0.484b
Ocular parameters (mean (SD)
Central corneal thickness (μm)	520.8 (23.6)	518.1 (28.6)	514.8(29.8)	0.683b
IOP at recruitment (mm Hg)	16.3 (3.2)	15.7 (3.4)	17.5 (5.2)	0.183b
Retinal nerve fiber layer thickness (μm)	72.0 14.6)	66.3 (14.9)	56.6 (10.3)	<0.001b
Rim area (mm2)	0.87 (0.3)	0.75 (0.1)	0.63 (0.4)	0.011b
Vertical cup disc ratio	0.75 (0.07)	0.78 (0.05)	0.83 (0.08)	<0.001b
Average cup disc ratio	0.77 (0.09)	0.80 0.04)	0.83 (0.06)	0.001b
Humphrey visual field
Mean deviation	−4.0 2.0)	−10.1 (3.1)	−24.3 (4.4)	<0.001b
Pattern SD	4.4 (2.3)	8.1 (3.0)	9.4 (2.6)	<0.001b
Systemic medicines (n (%)
Calcium channel blockers	23 (50%)	14 (30.4%)	9 (19.5%)	0.343a
Beta blockers	8 (57.1%)	4 (28.5%)	2 (14.2%)	0.548a
ACEi	10 (47.6%)	5 (23.8%)	6 (28.5%)	0.971a
Statins	25 (43.8%)	15 (26.3%)	17 (29.8%)	0.058a
OHA	18 (51.4%)	12 (34.2%)	5 (14.2%)	0.109a
Topical medicines (n (%)
Beta blockers	35 (46.0%)	20 (26.3%)	21 (27.6%)	0.800a
Carbonic anhydrase inhibitors	14 (38.8%)	10 (27.7%)	12 (33.3%)	0.358a
Prostaglandin analogs	39 (43.3%)	25 (27.7%)	26 (28.8%)	0.127a
Alpha-2 agonist	01(33.3%)	0 (0%)	02 (66.6%)	0.243a

Abbreviations: ACEi, angiotensin-converting enzyme inhibitors; OHA, oral hypoglycemic agent.

^aP<0.05 is considered statistically significant based on Pearson's chi-square.

^bP<0.05 is considered statistically significant based on one-way ANOVA. Bold values indicate significant values.

Confounding factors such as age, sex, hypertension, diabetes mellitus, hyperlipidemia, cardio-vascular systemic diseases, hematocrit, total cholesterol level, and FBS, which may affect the microvascular endothelial function, were analyzed using simple regression analysis and multiple stepwise linear regression (Table 2). Multiple stepwise linear regressions showed statistically significant negative linear relationship between age and ACh%. For every additional year of age, there was 11.4 (95% CI −22.39, −0.46) unit decrease in ACh%.

Table 2. Simple linear regression between predictors affecting acetylcholine % (endothelial function).

N=114	Coefficient B	95% CI	T-stat	P-value
Age (years)	−11.90	−22.83, −0.96	−2.15	0.033
Sex	−104.61	−302.06, 92.83	−1.05	0.296
Hypertension	−84.22	−288.52, 120.08	−0.81	0.416
Diabetes mellitus	−69.00	−268.83, 130.83	−1.71	0.089
CVSa	55.34	−203.64, 314.33	0.42	0.673
Hyperlipidemia	−30.18	−226.33, 165.97	−0.30	0.761
Systolic blood pressure (mm Hg)	−1.68	−7.28, 3.90	−0.59	0.551
Diastolic blood pressure (mm Hg)	−3.16	−14.51, 8.18	−0.55	0.582
Body mass index (kg/m2)	3.14	−19.38, 25.67	0.27	0.783
Hematocrit %	−14.39	−36.57, 7.78	−1.28	0.201
Total cholesterol (mmol/l)	−5.52	−89.44, 78.39	−0.13	0.896
Fasting blood sugar (mmol/l)	1.49	−31.54, 34.54	0.09	0.929

P<0.05 is considered statistically significant based on Simple Linear Regression.

^aCardio-vascular systemic diseases (myocardial infarction and angina). Bold values indicate significant values.

There was a significant difference in microvascular endothelial function based on severity (P<0.001). There was significant reduction in ACh% and ACh_max values according to severity of POAG (Table 3). In general, POAG patients with severe disease demonstrated lower ACh% and ACh_max values compared with mild and moderate cases. There was a significant negative linear correlation (moderate in strength) between severity and ACh% (r=−0.457, P<0.001) (Figure 1). Based on post-hoc Bonferroni analysis, differences in ACh% and ACh_max were significant between mild and severe patients and mild and moderate patients (Table 4). There was no significant difference between moderate and severe POAG patients. There was significant difference in SNP_max according to severity of POAG; patients with severe disease demonstrated lower SNP_max (Table 3). Based on post-hoc Bonferroni analysis, there was significant difference in SNP_max between mild and severe POAG (Table 4).

Table 3. ACh%, ACh_max, SNP% and SNP_max and severity of POAG in Malay patients.

	Severity			F	P-value
	Mild, N=55	Moderate, N=29	Severe, N=30
Acetylcholine (ACh)-mediated response
ACh% (mean, 95% CI)	715.8 (588.3, 843.3)	372.1 (196.5, 547.7)	239.9 (67.3, 412.6)	11.1	<0.001a
bACh% (mean, 95% CI)	712.9 (573.6, 852.3)	391.2 (200.3, 582.1)	224.5 (27.4, 421.7)	8.2	0.001c
AChmax (AU) (mean, 95% CI)	51.1 (44.0, 58.1)	28.7 (19.0, 38.4)	18.6 (9.1, 27.6)	16.9	<0.001*
bAChmax (mean, 95% CI)	52.2 (44.7, 59.7)	25.8 (15.5, 36.1)	19.1 (8.5, 29.7)	14.4	<0.001c
Sodium nitroprusside (SNP)-mediated response
SNP% (mean, 95% CI)	937.6 (824.1, 1051.1)	732.0 (575.7, 888.2)	765.6 (612.0, 919.3)	2.8	0.062*
SNPmax (AU) (mean, 95% CI)	77.7 (68.8, 86.5)	62.0 (49.8, 74.1)	53.1 (41.1, 65.0)	5.9	0.004*

^aP<0.05 is considered statistically significant based on MANOVA.

^bAdjusted for age, sex, hypertension, diabetes mellitus, hyperlipidemia, cardio-vascular systemic diseases, systolic blood pressure, diastolic blood pressure, body mass index, hematocrit, total cholesterol, fasting blood sugar, calcium channel blockers, beta blockers, angiotensin-converting enzyme inhibitors, statins, and oral hypoglycemic agents.

^cP<0.05 is considered statistical significant based on MANCOVA. Bold values indicate significant values.

Figure 1.

Correlation analysis between severity of POAG and Ach%.

Table 4. Pair wise comparison of ACh%, ACh_max, SNP%, and SNP_max according to severity.

Severity		P-value
aAch%, mean (95% CI)
Mild to moderate	321.7 (23.9, 619.5)	0.030b
Moderate to severe	166.6 (152.3, 508.3)	0.712b
Mild to severe	488.4 (178.2, 798.6)	0.001b
aAChmax, mean (95% CI)
Mild to moderate	26.4 (10.3, 42.4)	<0.001b
Moderate to severe	6.7 (11.7, 25.1)	1.000b
Mild to severe	33.1 (16.3, 49.8)	<0.001b
SNP%, mean (95% CI)
Mild to moderate	205.6 (31.2, 442.5)	0.111c
Moderate to severe	33.6 (235.1, 302.4)	1.000c
Mild to severe	171.9 (62.3, 406.2)	0.231c
SNPmax, mean (95% CI)
Mild to moderate	15.6 (2.7, 34.1)	0.123c
Moderate to severe	8.9 (11.9, 29.8)	0.896c
Mild to severe	24.6 (6.4, 42.8)	0.004c

^aAdjusted for age, sex, hypertension, diabetes mellitus, hyperlipidemia, cardio-vascular systemic diseases, systolic blood pressure, diastolic blood pressure, body mass index, hematocrit, total cholesterol, fasting blood sugar, calcium channel blockers, beta blockers, angiotensin-converting enzyme inhibitors, statins, and oral hypoglycaemic agents.

^bP<0.05 is considered statistically significant based on adjustment of multiple comparison with Bonferroni correction.

^cP<0.05 is considered statistically significant based on post-hoc test with Bonferroni correction. Bold values indicate significant values.

Discussion

In the current study, there was significant association between peripheral microvascular endothelial function (ACh% and ACh_max) and severity of POAG in Malay patients. Long-standing POAG is associated with progression to severe disease.⁴⁸ Thus it is not surprising that Malay patients with severe POAG were older. Age was also identified as a significant confounding factor affecting microvascular endothelial function in this study. Unlike arteriosclerotic changes that increase with age, microvascular endothelial function was reported to reduce with age.⁴⁹ There was 11.4 (95% CI −22.39, −0.46) units reduction of ACh% for every additional year of age.

Perhaps, for similar reason, systemic hypertension was found to be the most common systemic comorbidity in this study. However, there was significantly higher number of patients with hypertension in mild stage of POAG. In general, systemic comorbidities appear to be more common in mild stage of POAG in our study. It was found that the presence of systemic diseases such as hypertension lead to early detection of POAG.^{50, 51} Perhaps, the association between systemic comorbidities and severity of POAG is due to early detection rather than as part of pathogenesis of POAG. Nevertheless, to eliminate the potential effect of systemic hypertension and antihypertensive medications on microvascular endothelial function, our result was further adjusted for systemic hypertension, systemic medications, and other comorbidities.

In addition, systemic comorbidities were not found to be a significant predictor for microvascular endothelial function in this study. On the contrary, reduction of microvascular endothelial function was observed in hypertension, diabetes, and hypercholesterolemia in Caucasians.⁵² Ethnicity has been found to be associated with microvascular endothelial function in heart failure patients even after the adjustment for age, presence of systemic hypertension and diabetes mellitus, blood pressure, and glucose levels.⁵³

To the best of our knowledge, this is the only study on peripheral microvascular endothelial function and severity of POAG using LDF and the process of iontophoresis. ACh stimulates membrane receptor on the surface of endothelial cells leading to an increase in intracellular calcium, which in turn activates the enzyme endothelial nitric oxide (NO) synthase to produce NO. The released NO reaches its target organ (vascular smooth muscle cells) and activates the enzyme guanylate cyclase, which is responsible for the production of cyclic guanosine monophosphate (cGMP) leading to a decrease in calcium thereby causing vasodilatation.^{54, 55} SNP is a direct vasodilator, with equal arterial and venous effects.⁵⁶ SNP stimulates the synthesis of cGMP in smooth muscle that leads to the formation of NO radicals. These radicals initiate a chain of events that lead to dephosphorylation of contractile proteins in smooth muscle, causing relaxation and vasodilatation.⁵⁷

In the current study, data were controlled for potential confounding factors as shown in Table 3 using the MANCOVA statistical analysis. There was reduction in peripheral microvascular endothelial function (as reflected by endothelium-mediated vasodilatation) according to severity of POAG even after controlling for age and other potential confounders, including systemic diseases and systemic medications. This is an indirect quantification of microvascular endothelial function with the assumption that similar endothelial dysfunction occurs in the ocular circulation.^{58, 59} There was also difference in endothelium-independent vasodilatation (SNP_max) between mild and severe POAG. This may indicate impairment of vascular smooth muscle relaxation in microcirculation with severe POAG. Endothelial dysfunction affecting release of endothelial vasoconstrictor and vasodilators supplying the optic nerve head may cause impairment of perfusion leading to RNFL damage.^{60, 61, 62, 63, 64}

Microvascular endothelial dysfunction and impairment of vascular smooth muscle relaxation may accelerate glaucomatous damage responsible for more severe glaucoma in the current study. Trabecular meshwork is also lined by the endothelium.⁶⁵ Any abnormalities or dysfunction of the peripheral endothelium may directly contribute to changes in trabecular meshwork and disturbance in aqueous outflow.^{66, 67, 68, 69} Although this postulation is not thoroughly conclusive, endothelial dysfunction may affect ONH perfusion and impair the aqueous outflow.

Endothelial function in the peripheral vessels and end arteries in the eye may differ. The retinal arteries lose their internal elastic lamina as they bifurcate at optic disc.^{70, 71} That differentiates them from vasculature of other tissues, and as a compensatory mechanism, they have thicker muscularis layer. Thicker muscularis of retinal arteries leads to more vasodilatation in response to chemical stimuli.⁷¹ The endothelial cells of retinal arteries are not fenestrated and linked by tight junctions, but in peripheral vessels there are no tight junctions.⁷² There may be other factors that affect the blood perfusion to ONH contributing to glaucomatous optic neuropathy. Direct evaluation of ocular endothelial function is not possible with current methodologies. The technique adopted in this study allowed indirect evaluation of eye vessel microvascular endothelial function.

Imbalanced distribution in the number of POAG patients according to severity is a possible confounding factor in this study. There were more patients with mild stage compared with severe-stage POAG. There is a possibility that by chance patients with reduction of peripheral microvascular endothelial function may have been recruited among POAG patients with severe disease. An equal distribution of patients according to severity of the disease may eliminate these biases.⁷³

In conclusion, there is a potential role of peripheral microvascular endothelial function as a predictor for the severity of POAG in Malay patients. Impairment in peripheral endothelial-dependent vasodilatation may lead to inadequate perfusion and acceleration in glaucomatous damage.

The authors declare no conflict of interest.