Dr Bang V Bui Department of Optometry and Vision Sciences The University of Melbourne Parkville VIC 3010 AUSTRALIA E-mail: firstname.lastname@example.org
Although intraocular pressure (IOP) remains an important risk factor for glaucoma, it is clear that other factors can also influence disease development and progression. More recently, the role that blood pressure (BP) has in the genesis of glaucoma has attracted attention, as it represents a clinically modifiable risk factor and thus provides the potential for new treatment strategies beyond IOP reduction. The interplay between blood pressure and IOP determines the ocular perfusion pressure (OPP), which regulates blood flow to the optic nerve. If OPP is a more important determinant of ganglion cell injury than IOP, then hypotension should exacerbate the detrimental effects of IOP elevation, whereas hypertension should provide protection against IOP elevation. Epidemiological evidence provides some conflicting outcomes of the role of systemic hypertension in the development and progression of glaucoma. The most recent study showed that patients at both extremes of the blood pressure spectrum show an increased prevalence of glaucoma. Those with low blood pressure would have low OPP and thus reduced blood flow; however, that people with hypertension also show increased risk is more difficult to reconcile. This finding may reflect an inherent blood flow dysregulation secondary to chronic hypertension that would render retinal blood flow less able to resist changes in ocular perfusion pressure. Here we review both clinical and experimental studies that have attempted to clarify the relationships among blood pressure, OPP and blood flow autoregulation in the pathogenesis of glaucoma.
Glaucoma describes a group of optic neuropathies that result in progressive loss of ganglion cells. It manifests as characteristic optic disc cupping, nerve fibre layer losses and visual field defects.1 In 1996, the number of patients worldwide with glaucoma was estimated at nearly 66.8 million, of whom 6.7 million were blind.2 This places glaucoma second only to cataract as a global cause of blindness.3
Despite extensive clinical and experimental studies, the mechanisms underlying the development and progression of glaucoma remain unclear. An improved understanding of risk factors and how they interact to produce this disease will help better target potential treatment strategies.
Although raised intraocular pressure (IOP) is a well-known risk factor for glaucoma, the following clinical observations challenge the view that the pathogenesis of glaucoma is dependent purely on intraocular pressure:
Ocular hypertension does not necessarily lead to glaucomatous damage.4
Many patients with glaucoma return IOP readings within the normal range (normal tension glaucoma). Indeed, glaucoma can occur in patients with any IOP. Some investigators5,6 suggest that normal tension glaucoma simply lies at one extreme of the open-angle glaucoma continuum, implying a common mechanism in both conditions.
Although IOP reduction continues to be the only successful treatment to reduce the progression of glaucoma,4,7 many patients with apparently ‘adequate’ IOP reduction still show ongoing vision loss.8
This inconsistent role for IOP in glaucoma may be explained by the involvement of other risk factors, including older age, positive family history, ethnicity and thinner central corneal thickness, as well as the presence of myopia, diabetes and blood pressure (BP) abnormalities. Among these risk factors, those of particular interest are the ones that may be clinically modified (that is, IOP, blood pressure and diabetes). It may be that these risk factors are not independent. Indeed, high IOP, low blood pressure and the presence of long-standing diabetes may all converge to manifest as abnormal blood flow to the optic nerve. Thus a better understanding of the role that blood flow plays in glaucoma might account for these apparent ‘clinical inconsistencies’.
Blood flow to the capillary network of the lamina cribrosa region of the optic nerve can be described by a parameter known as ocular perfusion pressure (OPP). It is known that changes in either IOP or blood pressure will modify ocular perfusion pressure. Specifically, OPP represents a gradient between blood pressure and IOP (in simplistic terms, OPP = BP - IOP). Therefore, low OPP may arise due to elevations in IOP and/or reductions in blood pressure. Consequently, low perfusion pressure can occur even when the IOP is normal (that is, with low blood pressure). Any treatment that improves OPP should be beneficial to ganglion cells. This may be consistent with the fact that lowering IOP provides protection even in those with normal tension glaucoma. This interplay between IOP and blood pressure is important clinically because glaucoma and hypertension often co-exist in ageing populations. This raises the clinical paradox that although there is cardiovascular benefit in the control of systemic hypertension, such interventions would reduce the OPP.
There is a growing body of evidence suggesting that abnormal blood flow and changes in the ability of the eye to buffer against changes in perfusion pressure (a process known as autoregulation) are central to the pathogenesis of glaucoma. A reduced capacity for the eye to maintain blood flow in the face of variations of OPP increases the risk of impaired metabolic supply (reduced oxygen and nutrients),9,10 thus promoting neuronal dysfunction. This article aims to review the clinical and experimental studies that consider whether OPP and blood flow autoregulation have a role in the pathogenesis of glaucoma. The cellular mechanisms underlying autoregulation and how these might change in hypertension will also be discussed.
GLAUCOMA: A MULTI-FACTORIAL DISEASE
Numerous epidemiological studies have contributed to our understanding of the risk factors associated with glaucoma. The best established risk factors are high IOP, advancing age, positive family history and ethnicity (especially African ancestry).11 More recently a number of new risk factors have been identified. These include low blood pressure, thinner central corneal thickness (CCT) and in some studies, the presence of diabetes mellitus.11,12 The potential involvement of these factors in the pathophysiology of glaucoma is schematised in Figure 1. Throughout this review, each of the pathways laid out in Figure 1 will be considered with a description of the experimental evidence for such links. This section briefly covers risk factors such as ageing, myopia, thinner CCT and diabetes mellitus before the roles of blood pressure and blood flow are discussed in greater detail.
Increasing age is a major risk factor for primary open-angle glaucoma, as the disease seldom occurs at ages younger than 40.13 Studies in Australia14,15 and the USA16 show that the prevalence of the disease increases steadily with each decade over the age of 55. The number of retinal ganglion cells is known to decrease gradually with age even in normal human subjects17,18 and animal models.19 This decline in ganglion cell number will reduce neural redundancy and impact on vision in older age.20 One consequence of this reduced neural capacity is that in older individuals fewer ganglion cells need to be lost before there is detectable visual field loss. An alternative hypothesis is that ageing may increase the inherent vulnerability of ganglion cells to IOP insult.19 This is an area of intense interest for glaucoma research. Age-related abnormalities in mitochondria may lead to increased oxidative stress and reduce the capacity for cellular repair.21 In addition, age-related changes in the ability to maintain normal blood flow in the face of fluctuation of OPP (autoregulation) can increase the vulnerability of ganglion cells to hypoxic and metabolic stress. A growing number of studies22–24 has shown that ageing is associated with a progressive reduction in ocular blood flow, which appears to occur in association with an insufficient capacity for autoregulation. This is also supported by data from animal studies25,26 showing that in response to IOP elevation, older rats are less able to maintain normal blood flow than young rats. The exact mechanisms by which older age causes a predisposition to glaucoma are under investigation.
CENTRAL CORNEAL THICKNESS
Both epidemiological studies7,8,11 and case-controlled studies27–29 have shown a strong association between a thinner CCT and the development of glaucoma, even after allowing for the IOP correction associated with applanation tonometry; however, the pathogenic mechanism remains unclear. One hypothesis is that a thin CCT may reflect a thin sclera and thus a thin lamina cribrosa. This connective tissue difference at the level of the lamina cribrosa might lead to greater deformation of the lamina plates and the lamina pores surrounding the axons and blood vessels as they leave the eye. Greater stress and strain in this region can increase susceptibility to IOP-related stress.30 It is of interest that a gradual remodelling of the structures of the lamina cribrosa can occur following IOP elevation and that advancing age31 can potentiate the mechanical disadvantage associated with a thinner outer coat.
An elongated axial length32,33 in highly myopic eyes may lead to thinning of the sclera and lamina cribrosa34 that might exacerbate IOP-related stress at the optic nerve head (ONH). Connective tissue abnormalities have been consistently reported at the ONH of myopic human subjects32,34 and at the posterior pole in animal models of myopia.35–37 In keeping with this theory, subjects with high myopia were found to have a higher prevalence of glaucoma in several population-based studies, including the Barbados Eye Study,38 the Blue Mountains Eye Study39 and the Beijing Eye Study.40
The association between diabetes and glaucoma remains a matter of debate. Some epidemiological studies41–46report a positive association between the two conditions, while others47–50 have failed to find any positive association. Although these findings appear controversial, evidence from experimental studies51,52 suggests that diabetes may enhance the susceptibility of the eye to stress, such as elevated IOP. Widespread vascular damage is implicated in diabetes due to metabolic disturbances, which may exacerbate the ischaemic insult in glaucomatous damage. An elevated nitric oxide (NO) level has been found to contribute to the vascular dysfunction in streptozotocin-induced diabetic rats,53–55 which can lead to compromised autoregulation. The influence of diabetes on glaucoma requires further investigation and this topic has been reviewed elsewhere.56
Despite the fact that the role of IOP in glaucoma is not entirely clear, elevated IOP remains a major risk factor for the development and progression of open-angle glaucoma. Nearly all experimental models of glaucoma involve IOP elevation, therefore much of our understanding of the pathways in glaucoma pathophysiology are those that involve high IOP.
Most theories explaining IOP-induced ganglion cell injury can be grouped into two categories, namely, mechanical (axonal) and ischaemic (vascular) mechanisms. These mechanisms are not likely to be mutually exclusive. In fact, many investigators support a combined mechanical and vascular theory in the pathogenesis of glaucoma,57–60although the relative contribution of the two mechanisms remains unclear.
The mechanical concept suggests that high IOP causes direct or indirect injury to ganglion cell axons or their soma. Specifically, chronic IOP elevation is thought to induce remodelling and deformation of the lamina cribrosa, which gives rise to mechanical compression of ganglion cell axons at the ONH, leading to impaired anterograde and retrograde axoplasmic transport.61 In rats, acute and chronic IOP elevation can lead to accumulation of ‘brain-derived neurotrophic factor’ (BDNF) and its receptor Trkb62 in ganglion cell axons at the level of the lamina cribrosa. These outcomes suggest that, in both acute and chronic IOP elevation, axonal retrograde transport is interrupted and neurotrophins from the lateral geniculate nucleus cannot reach the ganglion cell body. It is believed that this loss of survival signals initiates elements in the cell death (or apoptotic) cascade.62
The vascular mechanism suggests that an important consequence of lamina cribrosa distortion induced by IOP elevation is the compression of blood vessels at the ONH, which in turn reduces the OPP, thus producing regional hypoxia. The presence of hypoxia will activate autoregulation in an attempt to sustain normal blood flow. The failure of such mechanisms produces ischaemia leading to cellular injury. Consistent with these vascular ideas, blood flow at the optic disc rim has been found to be reduced in patients with glaucoma.63 Fluorescein angiography revealed delayed filling of the vessels of the ONH in glaucoma patients with elevated IOP.64 Using corrosion casting, unfilled regions of ONH microvasculature have been detected in donor eyes from patients with glaucoma.65
Ischaemia can lead to glaucomatous neuropathy via a lack of metabolic support, increased oxidative stress66 and excitotoxicity. The lack of metabolic support can lead to stasis of axoplasmic transport. In addition, apoptosis can be initiated by increased oxidative stress arising from elevated pressure and ischaemia. Oxidative stress is mediated by a group of cytotoxic by-products known as ‘reactive oxygen species’ (ROS), which include the free radicals, superoxide and lipid peroxides. When ROS formation exceeds the intrinsic antioxidant capacity of a given cell, damage to the cellular DNA and protein follows.66 Thus, IOP elevation can induce gradual changes at the optic nerve head, resulting in an altered biomechanical and vascular environment, leading to axonal injury.
In addition to these concepts, there is growing evidence that ganglion cells may also be directly and indirectly injured via other pathways. For example, IOP can act directly on pressure sensitive mechanisms on ganglion cells to trigger apoptosis. Evidence for this arises from cell culture work, where a chronic IOP elevation (30 mmHg) can lead to retinal ganglion cell apoptosis. This may occur via the TRAAK receptor, which is found on ganglion cell bodies.67,68 More recent work has shown that hyperbaric pressure (50 mmHg) applied directly to the ganglion cell membrane can activate the N-methyl-D-aspartate (NMDA) channels.69 Furthermore, elevated IOP is known to injure glial cells, impairing their capacity to support neurons. For example, work in experimental models suggests that glutamate uptake in glial cells becomes impaired when IOP is elevated to adequate levels, which would potentiate NMDA activation.70 Finally, recent studies also suggest that an aberrant glial cell mediated immune response may be a secondary contributor to ganglion cell injury.71 Taken together, the above discussion highlights the complexity of glaucomatous neuropathy. In the following sections we will consider how abnormal blood pressure can influence these processes.
BLOOD PRESSURE AS A RISK FACTOR
Vascular insufficiency can be caused not only by IOP elevation but also by low blood pressure, as both conditions act to reduce the OPP. This is an area of intense interest.
BLOOD PRESSURE AND IOP
Population-based studies have consistently found an association between high blood pressure and IOP. In general, each 10 mmHg rise in systolic blood pressure is associated with only a small increase in IOP (approximately 0.27 mmHg).72–76As these studies covered populations with different ethnic backgrounds including Caucasians (Baltimore Eye Studies, Beaver Dam Study, Rotterdam Study, Egna-Neumarkt Study), Africans (Barbados Eye Study) and Asians (Tanjong Pagar Study), it is likely that they are widely applicable. Consistent with these epidemiological associations, animal experiments show that both chronic hypertension in rats77 and short-term high blood pressure in rabbits78 result in elevated IOP. The physiological basis of the relationship between blood pressure and IOP remains unclear. It has been hypothesised that both elevated IOP and blood pressure might be driven by a common extrinsic factor such as an age-related increase in sympathetic tone.78 Alternatively, an increase in blood pressure tends to elevate ciliary artery pressure, thus increasing the ultrafiltration component of aqueous production, resulting in IOP elevation.79,80 Moreover, because increased arterial pressure can produce a small increase in venous pressure, aqueous clearance will be reduced, which can also contribute to a higher IOP.
BLOOD PRESSURE AND GLAUCOMA
Despite the clear association between blood pressure and IOP, the exact relationship between blood pressure and open-angle glaucoma is complex. This is because not only does blood pressure influence IOP and OPP but long-standing hypertension might also reflect a compromised peripheral vascular capacity and autoregulation. Given this complex relationship, it is perhaps not surprising that evidence for the effect of blood pressure on glaucoma remains controversial. A literature review combining epidemiological and case-controlled studies shows that there are 11 studies72,73,81–89 reporting a significant correlation between glaucoma and high blood pressure, whereas 10 other studies90–99 have linked glaucoma with low blood pressure (Table 1). Specifically, the Blue Mountains Eye Study,86 Egna-Neumarkt Glaucoma Study72 and the Rotterdam Eye Study73,82 found that systemic hypertension increases susceptibility to glaucoma. This association between high blood pressure and glaucoma72,73,82,86 is counter intuitive, given that a high blood pressure should produce a high OPP and thus should give a protective effect. Despite the positive correlation between blood pressure and IOP, the actual change in IOP with increasing blood pressure is small. Therefore, it is unlikely that the increased risk of developing glaucoma associated with high blood pressure can be attributed to a blood-pressure-driven IOP elevation. Indeed, other epidemiological studies (Table 1) suggest that systemic hypertension is actually a protective factor in glaucoma.7,90–92 The potential reasons for this will be further discussed in the following sections.
Table 1. A summary of population-based studies examining the role of high blood pressure in glaucoma
High BP and POAG
BP = blood pressure, POAG = primary open angle glaucoma
Both high systolic and low diastolic BP increase risk
Consistent with the possibility that low OPP injures ganglion cells, several studies have found a greater prevalence of glaucoma in people with low blood pressure,100 as well as those that show larger nocturnal dips in blood pressure.94,101 Likewise, the Thessaloniki Eye Study99 shows that overly aggressive treatment with antihypertensive medications (that is, too rapid or too large blood pressure lowering), which produces a large drop in the OPP, increases the risk for glaucoma. This was evidenced by a strong association between aggressive antihypertensive treatment and larger cupping and decreased optic disc rim area.99 The cause of such an association may be related to an increased resistance in arterioles or capillaries in essential hypertension. If high blood pressure is treated without altering capillary resistance, this will produce greater levels of tissue hypoxia. Consistent with this theory, aggressive treatment of hypertension can cause serious harm not only to the eye but also to the heart102,103 and brain.104
The role of high blood pressure in glaucoma remains controversial. The Baltimore Eye Survey76,105 found that the association between blood pressure and glaucoma is age dependent. In particular, systemic hypertension appears to be protective against glaucoma in younger patients, whereas it increases the risk of glaucoma in older patients. The authors speculate that the optic nerve benefits from high perfusion pressure (that is, high blood pressure) when blood vessels are normal early in life, but as the vessels undergo atherosclerosis to become rigid and narrow with age there will be increased resistance to blood flow as well as compromised oxygen and nutrient exchange at the capillary beds, such that high blood pressure is no longer beneficial.105
More recently, the Los Angeles Latino Eye Study106 showed that both low diastolic and high systolic blood pressure are associated with an increased prevalence of open-angle glaucoma. Figure 2, redrawn from that study, shows that the relationship between glaucoma prevalence and diastolic blood pressure is ‘U’ shaped, indicating that patients at both extremes of the blood pressure spectrum are at greater risk of glaucoma.106 This apparent paradox at the extremes can be explained by two factors; one is that patients with hypotension suffer from low OPP at the ONH and the second is that those with chronic hypertension develop atherosclerosis over time leading to increased vascular resistance and compromised vascular autoregulation, as well as impaired nutrient exchange in the capillary beds at the ONH. This hypothesis not only provides a possible explanation for the discrepancy among the epidemiological studies summarised in Table 1 but also indicates that low OPP can occur secondary to any of high IOP, low blood pressure or local atherosclerosis.
Although the epidemiological studies are split over the association between blood pressure and glaucoma, the link between low OPP and glaucoma is consistently borne out in population studies.16,38,72,73,76 This argues for the use of OPP, rather than IOP or blood pressure alone, as a primary risk factor for glaucoma. It suggests that blood pressure monitoring should be integrated into optometric examinations for those at risk of glaucoma. It also indicates the need for sophisticated blood pressure management in patients to sustain peripheral perfusion.
ESTIMATION OF OCULAR PERFUSION PRESSURE
Perfusion pressure usually represents the difference between arterial and venous pressure. In the eye, venous pressure approximates the IOP, so that mean OPP can be taken as the difference between the mean ophthalmic artery pressure (MAPophthalmic) and IOP,107,108 as given in Equation 1.
MAPophthalmic is not readily measured in clinical practice. Instead, the brachial arterial pressure (MAPbrachial) is most commonly measured in clinical settings using an arm cuff sphygmomanometer in an upright position. This is a useful estimate of MAPophthalmic as the blood pressures in both the ophthalmic and brachial arteries are related in the absence of vascular pathology.109 Measured by ophthalmodynamometry, systolic ophthalmic arterial pressure is approximately three-quarters the systolic brachial arterial pressure, whereas diastolic ophthalmic arterial pressure is approximately two-thirds of the diastolic brachial arterial pressure.110,111 Therefore, Equation 1 can be reformulated for clinical application as follows:
MAPbrachial is determined from the systolic blood pressure (SBP) and diastolic blood pressure (DBP) as shown in Equation 3. The scaling factor 2/3 in Equation 2 accounts for the difference in blood pressure between the brachial and ophthalmic arteries in humans when measured in the sitting or standing positions.112 In many non-primate models, such a scaling factor is not required as the animals are habitually in a prone position.
Equation 2 is of clinical significance, as it indicates that the OPP can be reduced either by lowering the mean blood pressure (MAP) or by increasing the IOP.107,108 Thus, low blood pressure can undermine the effect of IOP lowering therapy to improve OPP in glaucoma patients. This may explain why some patients continue to develop visual field loss despite effective therapeutic IOP reduction. Consistent with this idea one study113 has shown that not only IOP but also ocular blood flow correlate with the progression of glaucoma in terms of visual field loss. This association can be schematised as shown in Figure 3.
Several caveats exist when Equation 2 is used to determine the OPP. First, although the use of MAPbrachial as a surrogate for MAPophthalmic is appropriate under normal conditions, it does not account for local vascular pathology, such as atherosclerosis induced by chronic high blood pressure. Once hypertension becomes severe enough to produce atherosclerosis or even total tissue infarction, local perfusion pressure will drop even with high blood pressure. In this setting, systemic blood pressure will not be a good surrogate for local blood pressure. Second, IOP elevation can also induce mechanical compression on axons, which cannot be reversed by lowering blood pressure, as this produces local hypoperfusion.114 The OPP concept cannot account for this mechanical compression.
EXPERIMENTAL EVIDENCE FOR BLOOD PRESSURE INVOLVEMENT
In contrast to the vast body of literature arising from clinical trials and epidemiological studies, few experimental studies have directly addressed the effect of blood pressure on glaucoma susceptibility. To our knowledge, no study has successfully induced ‘normal tension glaucoma’ in an animal model by lowering blood pressure alone. Only a few experiments have considered the effect of IOP and blood pressure modulation simultaneously. The first report by Grehn and Prost115 goes back to 1983. Specifically, ganglion cell function in cats was assessed by axonal impulse conduction, which remained unimpaired when the OPP was above 20 mmHg, regardless of whether IOP was 40 or 135 mmHg. This is of interest as the data suggest that the primary determinant of ganglion cell health is the integrity of the blood flow. There appears to be little influence from a mechanical mechanism of acute IOP, as one would have expected a greater mechanical effect in the presence of an IOP of 135 mmHg than 40 mmHg. This study was then extended to the measurement of full-field electroretinogram (ERG) (the bipolar cell b-wave) and pattern ERG by Siliprandi and associates116 in 1988. The pattern ERG has been shown to depend mainly on ganglion cell integrity.117 Like the Grehn and Prost study,115 the pattern ERG was attenuated either by increasing IOP or reducing blood pressure in cats. These studies indicate that the key determinant to retinal function is OPP, rather than IOP or blood pressure alone.
Studies that measure ocular blood flow are in agreement with the functional data, supporting the hypothesis that OPP plays a dominant role in determining retinal function. Kiel and Van Heuven118 reported that blood pressure and IOP manipulations can produce an equivalent change in choroidal blood flow in rabbits, measured with laser-Doppler flowmetry. Most recently, Liang and associates119 found that impaired ONH blood flow arising from elevated IOP is exacerbated in monkeys with hypotension (mean arterial blood pressure of 56 mmHg), compared with those with normal blood pressure (mean arterial blood pressure of 102 mmHg). These findings are consistent with those reported by Grehn and Prost,115 as they indicate that the primary insult associated with IOP challenge is a vascular compromise.
Taken together, these studies consistently show that OPP is a key determinant of blood flow and visual function, which is in agreement with the concept (Figure 3) that low blood pressure exacerbates IOP-induced ischaemia, whereas high blood pressure, at least in the short term, appears to provide some protection against it.120
It is important to note that these studies altered IOP and blood pressure over only short periods of time (a maximum of 10 minutes). Although they provide crucial insights into the relationship between IOP, blood pressure and ganglion cell integrity, these experiments do not explore the more chronic effects needed to better model open-angle glaucoma and essential hypertension in humans. In particular, acute short-term hypertension does not cause atherosclerosis, cardiac hypertrophy or renal impairment, which frequently occur in long-standing systemic hypertension. Therefore, a model of chronic elevation of IOP and blood pressure would yield greater clinical insight into understanding the relationship between autoregulation, systemic hypertension and glaucoma. In this regard, Hayreh and colleagues121,122 induced experimental atherosclerosis (using a high-cholesterol diet) and systemic hypertension (by renal artery occlusion) in monkeys for several years before chronic IOP elevation was achieved by laser photocoagulation of the trabecular meshwork. Surprisingly, these two reports showed that neither systemic hypertension nor atherosclerosis has much influence on retinal and optic nerve changes induced by IOP elevation. The authors attributed the outcome to inadequate sample size and statistical power. To our knowledge, no other animal experiment has examined the effect of chronic hypertension on glaucoma, possibly due to the technical difficulties associated with producing and monitoring chronic hypertension and IOP elevation simultaneously.
BLOOD FLOW AUTOREGULATION
A small reduction in OPP does not always result in blood flow deficiency, as the retina strives to maintain its circulation even under extreme environmental insults. This has been known since Hill and Flack123 in 1912 found that the choroidal circulation in cats would not arrest until the IOP was raised to levels approximating the carotid artery pressure. Ocular blood flow is determined not only by OPP but also by blood vessel resistance, which in turn is related to blood viscosity (η) and vessel diameter (R), all of which contribute to regulate blood flow (Q) as expressed by the Hagen–Poiseuille law:
ΔP represents the pressure gradient between the two ends of a cylindrical pipe. In the case of blood vessels in the eye, ΔP can be taken to represent ocular perfusion pressure; L is the length of the blood vessel.
Autoregulation of blood flow is defined as the intrinsic ability of an organ to maintain constant blood flow despite changes in perfusion pressure. When there is a change in OPP (ΔP), the retina tends to maintain its blood flow (Q) by adjusting other parameters in Equation 4. Any modification in blood vessel length (L) or blood viscosity (η) is negligible during such an event. Therefore, blood vessel diameter (R) has a major role in determining both the vessel resistance and autoregulatory capacity of blood flow. It is worth noting that Equation 4, in its strictest sense, only applies to fluid flowing through a rigid pipe, where none of the parameters interacts with the others. However, in living tissue, Equation 4 is complicated by other factors. In particular, a change in vessels calibre (R) will interfere with local blood pressure (ΔP). For example, the vasodilatory calcium channel blockers can increase blood vessel diameter (R) and thereby improve blood flow, even though the drug has a systemic blood pressure lowering effect.124 Likewise, despite an increase in systemic blood pressure, L-nitro-arginine methyl ester (L-NAME) causes a reduction in ophthalmic blood flow in rabbits due to its strong local vasoconstriction effect.125,126 This is not surprising given the exponent ‘4’ of parameter R in Equation 4, which means that the vasomotor response has a much greater capacity to regulate local blood flow, compared with a proportional change in any other parameter. Not surprisingly, extensive clinical and experimental evidence has shown that autoregulation acts via changes in vessel diameter. By photographing retinal blood vessels, a compensatory vasodilation of blood vessels has been demonstrated during IOP elevation in glaucoma patients. Furthermore, the degree of change in diameter was associated with the level of IOP elevation.107,108 This vasomotor response to IOP challenge was less effective in patients with glaucoma than in healthy subjects. Mice genetically deficient in endothelial nitric oxide synthase (eNOS; a potent vasoactive peptide), show impaired blood flow autoregulation and increased vulnerability to ischaemic insult such as cerebral infarction, when compared with the wild-type mice.127
More direct evidence of ocular autoregulation comes from measurement of blood flow during OPP variation. As shown in Figure 4A, when autoregulation is removed in cats undergoing euthanasia, ocular blood flow measured with laser-Doppler flowmetry drops linearly with OPP reduction.128 On the other hand, Figure 4B shows that ocular blood flow in normal human subjects is well regulated over a wide range of changes in OPP induced by IOP129 or blood pressure elevation.130 Although measured in different species, the sharp contrast between Figures 4A and 4B illustrates the capacity of the ONH to maintain a relatively constant blood flow in response to blood pressure and IOP challenge.
RANGE OF AUTOREGULATION
While the existence of autoregulation is well established, a question of interest is ‘to what extent can autoregulation sustain function during OPP challenge?’ In normal human subjects, autoregulation at the optic nerve is effective for IOP below 27–30 mmHg. This represents an OPP reduction of some 40–50 per cent from baseline (Figure 4B) for a mean arterial pressure of 100 mmHg (2/3*100–IOP).131 At the other extreme, blood flow remains unchanged until the OPP is elevated by more than 30 ± 8 per cent above baseline (Figure 4B), which in human studies is usually achieved by blood pressure elevation via isometric exercise.130 Other studies have reported that this upper limit of autoregulation can be as high as 34–60 per cent above the baseline ocular perfusion pressure.109,132–134
The autoregulation curve shown in Figure 4B represents the data collected from two different studies, one where OPP was lowered by IOP elevation135 and the other where OPP was increased by raising blood pressure.130 It is worth noting that combining the two sets of data assumes that IOP increase and blood pressure lowering modify the autoregulation curve in the same manner. This assumption has yet to be proven. In fact, at least one study showed that blood flow regulation in the choroidal circulation was more effective with blood pressure reduction rather than IOP elevation.136Figure 5 shows data from Chemtob and colleagues137,138 who characterised the blood flow across a wide range of blood pressure by infusing radioactive microspheres into the retina, choroid and brain of newborn piglets. Although the absolute values of blood flow vary across different tissues, the profiles of the flow–pressure relationship bare a striking resemblance. In particular, a blood flow plateau exists for blood pressures between 50–100 mmHg, consistent with the presence of strong autoregulation in the retina, choroid and brain.
The changes to the autoregulation curve in systemic hypertension, ageing and glaucoma are yet to be fully understood. An age- or disease-related shift in the curve to higher blood pressures, a steeper curve or a narrower autoregulatory range may make neurons more susceptible to reductions in OPP, an issue to be discussed in greater detail in the following sections.
MECHANISMS OF AUTOREGULATION
The mechanisms underlying vascular autoregulation are still under investigation. Our current understanding of ocular blood flow autoregulation is based on investigations in the cerebral and systemic circulations. It is likely that ocular autoregulation involves both myogenic and metabolic mechanisms, through the action of endothelium-derived vasoactive factors that modulate smooth muscle tone and pericytes. In the eye, the role of hormonal components (epinephrine and norepinephrine) in autoregulation is relatively minor as there is no autonomic innervation of retinal and ONH blood vessels; however, it is worth noting that the choroidal circulation has strong autonomic input.139,140 In addition, alpha- and beta-adrenergic as well as cholinergic receptors are present on ocular blood vessels.141 Thus, higher systemic concentrations of catecholamines, as seen in hypertension or higher local concentration due to the use of anti-glaucoma agents (beta-blocker and alpha agonists), may impact local blood flow.142
In myogenic autoregulation, it is thought that changes in perfusion pressure are sensed by mechano-transducers located on the endothelial cells of the blood vessels, which in turn respond to the pressure challenge by releasing vasoactive mediators. By patch-clamping the aortic endothelial cell membrane, Lansman and colleagues143 demonstrated the presence of stretch-activated ion channels, which respond to mechanical pressure, thus forming a set of mechano-transducers. These channels are permeable to Ca2+. One consequence of Ca2+ influx is that this cation acts as a second messenger, which mediates the synthesis and release of endothelium-derived vasoactive factors. In isolated bovine eyes, a graded contraction of retinal arteries occurs when intraluminal pressure is elevated from 10 to 60 mmHg, which acts to maintain vessel diameter. At pressures above 60 mmHg the vessel diameter increases linearly with pressure, signalling the failure of this myogenic mechanism.144 The suppression of this autoregulatory response by the calcium channel blocker nifedipine shows that it is mediated by extracellular calcium.144
Blood flow is also closely coupled with tissue metabolic activity. Evidence for metabolic autoregulation comes from studies showing that blood flow in the eye can be modified in the absence of pressure change. Specifically, retinal vessels constrict in response to hyperoxia (increased PaO2),10,145–147 whereas vasodilation and increased retinal blood flow can be induced by either hypoxia (decreased PaO2) or hypercapnia (accumulation of CO2) in order to meet the needs of neuronal tissues.148,149 Likewise, an increase in retinal metabolism induced by flickering light also leads to compensatory vasodilation.150,151 These findings strongly suggest that a metabolic-dependent mechanism is involved in blood flow autoregulation in the retina and optic nerve.
Both myogenic and metabolic mechanisms of autoregulation are achieved via the release of contracting and relaxing substances from the vascular endothelium, glial cells or neurons, with the latter two cell types mainly involved in the metabolic pathway of autoregulation. Figure 6 is a schematic of the endothelial-dependent mediators, among which nitric oxide (NO; vasodilation) and endothelin-1 (ET-1; vasoconstriction) are probably the most important factors and have opposing actions in ocular blood flow autoregulation.
In vascular endothelial cells, L-arginine is converted to NO via the enzyme NO synthase (NOS). On release from endothelial cells, NO exerts a potent vasodilatory effect via the activation of guanylyl cyclase in the cyclic guanosine monophosphate (cGMP) pathway, thereby reducing blood flow.152 Among the three isoforms of NOS, the endothelial NOS (eNOS or NOS-1) and neuronal NOS (nNOS or NOS-3) are constitutively expressed and are important in sustaining blood flow with normal variations in IOP and blood pressure. The inducible NOS (iNOS or NOS-2) is expressed under pathological conditions such as low OPP. The NO produced with iNOS can be excessive, leading to vascular dysregulation and cellular apoptosis. In human donor eyes, iNOS was found only in glaucomatous laminar cribrosa but not in normal tissue.153 Exposure to excess levels of NO produced by iNOSat the ONH has been implicated in the pathogenesis of glaucoma.153
Studies in both ocular154 and cerebral tissues155 demonstrate the existence of a basal release of nitric oxide, which means that there is active vasodilation under physiological conditions. This basal level of vasodilation allows the vascular tone to be either increased (constricted) or decreased quickly during changes in the perfusion pressure. For example, when NO release was suppressed by systemic infusion of a NOS inhibitor (L-NAME), choroidal blood flow reduced, leading to a downward shift of the autoregulation curve as shown in Figure 7.125 This shows that basal- and pressure-induced NO release are both important in autoregulation.
The vasodilatory effect of NO is counteracted by vasoconstricting factors such as ET-1, which binds to the ETA receptor on the membrane of vascular smooth muscle cells. Altered ET-1 vasoreactivity has been implicated in the pathogenesis of vascular dysregulation and systemic hypertension,156 potentially increasing vulnerability to ischaemic insult. More importantly, clinical and experimental evidence exists to support the role of ET-1 in the development of normal tension glaucoma. Higher plasma levels of ET-1 have been found in patients with normal tension glaucoma than in healthy subjects,157 implying a generalised vasoconstriction and reduction in blood flow. In rabbits, intravitreal injection of ET-1 causes blood flow reduction in the ONH.157 In another study, although plasma ET-1 levels were similar in glaucoma patients and normal controls, ET-1 elevation was greater in those with glaucoma following cold provocation.158 Additionally, Henry and colleagues159 showed that the vasodilation induced by an ETA receptor antagonist (BQ123; intra-arterial infusion) was reduced in patients with normal tension glaucoma when compared with age-matched controls. Taken together, these studies suggest a link between systemic hypertension, vascular dysregulation and glaucoma.
VASCULAR DYSREGULATION AND GLAUCOMA
Abnormalities in autoregulation have been divided into primary and secondary vascular dysregulation (without or with underlying disease, respectively). Flammer and Mozaffareih160 and Moore and associates161 have reviewed how abnormalities of autoregulation might increase the risk of glaucoma. In this section, we focus on the role of systemic hypertension and its implications for secondary vascular dysregulation and glaucoma.
As described above (Table 1 and Figure 2), several studies have found that paradoxically, chronic hypertension may increase the risk of glaucoma, an observation that does not explain the protection that should be afforded by an improved OPP. The presence of secondary vascular dysregulation could increase the vulnerability of the optic nerve to small changes in IOP, blood pressure and metabolic needs. One way that this may occur is if vascular dysregulation reduces the effectiveness of autoregulation. Thus, the same change in OPP can cause a greater reduction in blood flow. Autoregulation confers a wide ‘normal range’ and thus an increased capacity to cope with changes in the OPP. On the other hand, the absence of autoregulation means that smaller changes in the OPP are needed to push blood flow outside of the ‘normal range’.
Systemic hypertension162–164 has been shown to alter autoregulation in systemic circulation, via endothelial cell damage/dysfunction and abnormal release of vasoactive substances; however, the influence of hypertension on ocular blood flow autoregulation is difficult to assess in epidemiological studies, as few studies have measured blood flow over an adequately wide ocular perfusion pressure range. The influence of hypertension has been considered in laboratory studies of cerebral but not of ocular blood flow. In particular, Harper and Bohlen165 showed that the autoregulatory capacity of cerebral blood flow in spontaneously hypertensive rats is rightward shifted to higher pressure (Figure 8), consistent with an effort to compensate for the potential hyperaemia during increased cerebral perfusion pressure. This observation has also been reported in baboons166,167 and humans.168
If the above findings in brain also apply to ocular blood flow, then the lower limit of the ‘normal range’ would be reset to a higher OPP in chronic hypertension. Figure 9B shows the ‘normal range’ compared with Figure 9A, which represents one hypothetical effect of hypertension on the autoregulation curve (rightward shift). This ‘on face’ value means that blood flow will be reduced below the ‘normal range’ (grey area) at a higher OPP in those with hypertension. When the pre-treatment habitual blood pressure (unfilled circles) is taken into account, the actual OPP change needed to reduce blood flow below the ‘normal range’ is actually the same. Thus, the data derived from the cerebral circulation (Figure 8) do not account for an increased risk of optic nerve injury with chronic hypertension. Some other mechanism must be at play. One possibility is that hypertensive treatment dissociates the post-treatment blood pressure (filled circle, Figure 9B) from the ‘set point’ (unfilled circle, Figure 9B) of the autoregulation range. That is, the middle of the autoregulation range is set to the higher pre-treatment blood pressure (unfilled circle), creating a mismatch with the lower post-treatment blood pressure (filled circle). What this would mean is that the medically treated (post-treatment) hypertensive patient is now closer to the lower limit of their ‘normal range’. Thus, a smaller reduction in the OPP would be needed to yield a blood flow that cannot sustain normal function (Figure 9B).
A second possibility is that chronic hypertension affects ocular autoregulation differently from cerebral blood flow, to produce a narrower autoregulatory plateau. Figure 9C shows one extreme where the plateau has been removed, which means that smaller changes in the OPP would be needed to reduce blood flow below the ‘normal range’. A narrower autoregulatory range might arise due to atherosclerosis, such that blood vessels are less able to change their calibre in response to stimulation by vasoactive factors. However, neither epidemiological studies47 nor animal experiments121,122 have identified atherosclerosis as an important risk factor for open-angle glaucoma. An alternative may be aberrant production of vasoactive peptides (Figure 6) involved in autoregulation. Other contributing factors may also play a role in linking glaucoma and impaired autoregulation in hypertension. For example, it has been well established that patients with hypertension also demonstrate increased blood viscosity,169–172 which impacts negatively on blood flow (Equation 4).
Although it is clear that blood pressure plays a role in the pathogenesis of glaucoma, the underlying aetiology and mechanisms are not fully understood. A better understanding of this potential link requires clinical studies of glaucoma that take into account the mode and aggressiveness of treatment for systemic hypertension. The Thessaloniki Eye Study99 has shown that zealous treatment of systemic hypertension in patients without glaucoma is associated with increased cup-to-disk ratio. Further studies should extend this link to patients with concurrent hypertension and glaucoma. Specifically, the aggressiveness of anti-hypertensive treatments may correlate with the progression of glaucoma. If this were the case, it would reinforce the role of OPP as an important risk factor of glaucoma, as well as provide guidance for clinicians in treating these conditions. In terms of laboratory studies, a more comprehensive understanding of the cellular processes and the manner by which they interact with vasoactive peptides to drive the autoregulation of ocular blood flow represents a formidable challenge but one that must be undertaken. Understanding these pathways may eventually lead to the discovery of therapeutic intervention.
During a challenge to the OPP, autoregulation acts as a compensatory mechanism not only to maintain a constant blood flow (vascular reserve), but also in a broader sense to preserve neuronal function even when blood flow is compromised (functional reserve). As shown schematically in Figure 10, the vascular reserve is driven by a compensatory vasodilation. When vasodilation reaches its maximal capacity and the OPP is further reduced, blood flow will decline proportionately with the OPP. At moderate to low OPP, there is an increase in tissue oxygen extraction from the residual arterial blood, which may be adequate to maintain oxygen metabolism for a further period of time. This hypothesis is supported by studies in the brain,173–176 which have frequently found that increased oxygen extraction during ischaemia and/or anaemia allows oxygen metabolism and neural function to be maintained. In the brain, this functional reserve has been shown to depend largely on intact mitochondrial function during the early stage of ischaemia.177–179When ischaemia becomes severe enough to impair mitochondrial function, then irreversible neuronal damage (infarction) ensues. These findings derived from the brain should prompt research to further define the thresholds for both vascular and functional reserves in the eye. The hypothesis laid out in Figure 10181 speculates that optimal visual function can operate with less than 100 per cent blood flow (the redundancy theory). To test this hypothesis, it is essential to measure both ocular blood flow and ganglion cell function simultaneously over a gradient of OPP variation. Furthermore, it will be of interest to see if such a ‘safety buffer’ is reduced with older age or in patients with atherosclerosis or diabetes.
It is clear that IOP alone does not determine the risk of glaucoma development. While epidemiological studies have shown that there is a host of potential risk factors along with IOP, how these risk factors interact with IOP to modify glaucomatous neuropathy remains unclear. Although certainly not the only risk factor, there is adequate evidence to show that abnormalities in blood pressure and blood flow play a central role in glaucoma pathogenesis. In particular, low blood pressure predisposes to low OPP, which increases the likelihood of hypoxic or ischaemic stress. This is pertinent in that nocturnal IOP elevations and blood pressure dips can act synergistically to produce substantial OPP troughs over a diurnal cycle, which have been implicated in the development of normal tension glaucoma.101 An epidemiological study has shown that over-treatment of hypertension increases the risk of glaucoma.99 Although the role of low blood pressure in glaucoma is clearly detrimental, the effect of high blood pressure is more complex. In the short term, high blood pressure can improve the OPP and provide some protection against IOP-induced ischaemia.115,116,118,119 In chronic presentations, the influence of hypertension on glaucoma remains controversial among epidemiological studies (Table 1) and is not well established in animal experiments. The most recent evidence from an epidemiological trial106 showed that hypertension predisposes to the development of glaucoma. This is in line with the widespread vascular damage frequently associated with chronic hypertension, which would act to impair ocular blood flow and autoregulation (secondary vascular dysregulation). The presence of impaired autoregulation means that the eye is less able to cope with episodes of low OPP and over time a cumulative effect could produce ganglion cell loss. In an attempt to adapt to chronic high blood pressure, a rightward shift of the cerebral blood flow autoregulatory curve (Figure 8) has been found in patients with hypertension168 and animal models.166,167 If this can be extrapolated to the eye, it provides an explanation of why patients with hypertension are more vulnerable to low OPP (Figure 9B). From a clinical point of view, it is important to consider not only the IOP but also the blood pressure status in patients with glaucoma. Specifically, it is important to avoid under- or over-treatment of chronic hypertension to achieve an optimal OPP range.