Ординатура / Офтальмология / Английские материалы / Diabetes and Ocular Disease Past, Present, and Future Therapies 2nd edition_Scott, Flynn, Smiddy_2009

54 Diabetes and Ocular Disease

acids, and lipids), inflammation, and excitotoxicity. Müller cells and astrocytes control glutamate metabolism, so glutamate accumulation in the extracellular fluid between neurons and glia implies that glial cells are defective, and the clearance of retinal glutatmate is impaired in experimental diabetes [34,35]. Glutamate is a well-recognized cause of neuronal cell death in cerebral ischemia (glutamate excitotoxicity) [36].

Vascular changes that begin shortly after the onset of insulin-deficient diabetes include delayed leukocyte migration in the perifoveal capillaries [37], increased blood–retina barrier permeability [38], and increased retinal blood flow compared to nondiabetic control subjects [39]. Studies in diabetic animals have shown increased blood–retina barrier permeability and alterations in retinal blood flow within 1 to 3 months [40,41]. These findings suggest that vascular autoregulation is impaired before clinically evident vascular lesions appear [42]. Thus, humans and rodents exhibit similar cellular alterations in the preclinical phases of diabetic retinopathy.

Further evidence for early pathophysiologic abnormalities in the preclinical phase arises from studies in experimentally diabetic dogs. Engerman and Kern [43] showed that the intensive control of diabetes in dogs for the first 2.5 years determined the subsequent development of vascular lesions, whether or not the animals were subsequently treated with high or low doses of insulin to achieve tight or poor metabolic control, respectively. Thus, while this early phase of diabetic retinopathy appears to be innocuous from a clinical standpoint, numerous cellular and metabolic processes are active that lead to the development of clinically evident nonproliferative diabetic retinopathy (NPDR). Indeed, a recent demonstration that retinal flavoprotein fluorescence increases in diabetic patients before the onset of visible retinopathy is strong evidence for early onset of metabolic dysregulation [44].

While it is reassuring that patients with diabetes may have no visible retinopathy, the absence of microaneurysms or hemorrhages should not lead to complacency on the part of patients or physicians. In fact, aggressive control of the metabolic and systemic cardiovascular risk factors known to exacerbate retinopathy onset and progression provides an ideal opportunity to prevent vision-threatening changes. Patients who have not developed retinopathy should have a treatment strategy designed to optimize the chance to maintain vision. These patients with healthy appearing retinas and good vision represent the greatest therapeutic opportunity, particularly in light of the emerging diabetes epidemic.

NONPROLIFERATIVE DIABETIC RETINOPATHY

NPDR is defined and staged by ophthalmoscopic features such as vascular lesions, including microaneurysms, intraretinal hemorrhages, and vasodilation. Table 4.2 summarizes the manifestations of NPDR.

Implicit in these classification terms is the concept of a primary vascular disorder. The definitions (nonproliferative and proliferative) are useful clinically because they permit evaluation of ophthalmoscopically visible ocular risk factors for moderate and severe visual loss. The specific sequence of cellular events that lead to the features of NPDR remain uncertain because they occur below the resolution

Glial cell occlusion of capillaries

of any currently available clinical imaging tools. The sum of experimental and clinical studies strongly suggests that all retinal cells are affected in the preclinical stage of diabetic retinopathy (DR), and certainly by the time of development of NPDR, but there is no empirical evidence that retinopathy results from a specific or isolated vascular cell defect or biochemical pathway [45].

Capillary closure in the peripheral retina may lead to shunting of retinal blood flow into the posterior pole, where it increases the propensity for developing diabetic macular edema (DME) [46]. Capillary closure is a characteristic element of progressive NPDR, but it is unclear whether formed vascular elements—erythrocytes, leukocytes, or platelets—initiate vascular occlusion. Experimental studies demonstrate that transient leukocyte adherence to endothelial cells increases in diabetes [47], and this change may be part of a retina-wide chronic inflammatory process. Histopathologic studies have shown that glial cells migrate through the vessel wall and occlude vascular lumens in patients with diabetic retinopathy [48]. Whether this is a primary event related to glial cell proliferation or secondary to intraluminal capillary plugging is not known. Basement membrane thickening is a characteristic histopathologic feature of diabetic retinopathy and may contribute to capillary closure, but its cause is also unknown.

It has long been held that pericytes are among the first retinal cells to die in diabetes. Although pericytes and endothelial cells clearly undergo programmed cell death (apoptosis) [49], it is unproven whether pericytes are uniquely susceptible to diabetes. The original light microscopic study [50] of trypsin digest preparations

from small retinal depressions that reflect light away from the observer so that the macula appears slightly darker than the surrounding retina. The feature is best observed by slit-lamp biomicroscopy and is also noted on fundus photographs, particularly with red-free filters. It is more easily recognized in young patients who have a bright foveal reflex than in older persons. The thinning may result from macular ischemia and/or nonischemic neuroretinal degeneration (apoptosis). This finding may contribute to paracentral scotomas and may be confused with epiretinal membranes or macular edema.

The biochemical and cellular events that initiate vascular lesions in diabetic retinopathy are complex and uncertain in humans. Most of the available information is derived from studies in animals with experimental diabetes induced by streptozotocin or alloxan, or from vascular cell culture experiments. While it is clear that intensive treatment of diabetes in humans or animals significantly delays the onset and progression of retinopathy [56], it is not known whether the development of retinopathy represents a direct effect of insulin deficiency or insulin resistance, a consequence of hyperglycemia, or another metabolic derangement associated with diabetes, such as hyperlipidemia. The metabolic pathways that have been associated with diabetic retinopathy include activation of the polyol pathway, nonenzymatic glycosylation, and activation of the ß isoform of protein kinase C (PKC-ß) [57,58].

Increased glucose metabolism via the polyol pathway [59], first suggested as a cause of cataracts in diabetes, has also been considered to account for diabetic retinopathy and peripheral neuropathy. The hypothesis suggests that increased glucose metabolism via this pathway results in the accumulation of sorbitol, reduction of myo-inositol, and/or reduction in activity of sodium-potassium-ATPase, which may account for vascular dysfunction. Aldose reductase is a key enzyme in the polyol pathway. However, specific vascular functional or neuronal abnormalities, such as barrier breakdown or capillary closure, have not been fully explained by this hypothesis. Studies of aldose reductase inhibitors in diabetic dogs [60] and rats [61] have shown conflicting results. Several clinical trials of aldose reductase inhibitors (sorbinil, tolrestat) have failed to show a benefit on slowing human retinopathy progression [62]. After three decades of aldose reductase clinical trials, aldose reductase inhibitors have not yet proven to be a useful treatment for diabetic retinopathy.

Another theory for the development of diabetic retinopathy involves vascular damage by advanced glycosylation end products (AGEs). According to the concept of nonenzymatic glycosylation [63], sugar molecules bond covalently to reactive molecules and cause alterations in the functions of proteins, nucleic acids, and cells, such as macrophages. This reaction gives rise to the glycohemoglobin (hemoglobin A1c) test, which measures integrated glucose levels over 3 months. Nonenzymatic glycosylation has been proposed to account for cross-linking of long-lived proteins such as collagens, which are found in vascular basement membranes and vitreous. Collagen cross-linking may reduce the turnover of collagen and allow for basement membrane thickening or may contribute to vitreous collagen contraction. Advanced glycation end products increase in Müller cells in experimental diabetes, and a soluble AGE receptor that blocks its activation, decreases neuronal cell death [64]. However, to

become insulin deficient when their pancreatic ß-cells fail. Recent studies now show that impaired insulin action also occurs in the retinas of experimentally diabetic animals [73,74], indicating that diabetes itself directly impacts the retina. It is not certain how this change impacts the retina but it is likely to impair normal anabolic processes required for vision [45].

DIABETIC MACULAR EDEMA

The physiologic factors that govern the development of DME are similar to those involved in tissue edema elsewhere in the body, and understanding the pathophysiology of DME allows construction of a set of risk factors and treatment principles for DME.

Starling’s law of the capillary states that edema formation in tissues from fluid flux across the capillary wall is related to the hydrostatic pressure gradient (blood pressure minus tissue pressure) less the oncotic pressure that draws water into the vessels. This relationship has recently been shown to also operate in the retina for DME [75]. That is, increased intravascular hydrostatic pressure from hypertension or intravascular fluid overload drives fluid across the vascular wall (Fig. 4.5) and

Capillary breakdown

Figure 4.5. Relationship of altered vascular physiology to development of macular edema. Capillary occlusion with resulting nonperfusion has been confirmed as capillary dropout. Resulting retinal hypoxia produces autoregulatory arteriolar vasodilation with reduced pressure in arterioles and increase in capillary and venular hydrostatic pressure. Vessels dilate and increased capillary hydrostatic pressure leads to edema development, according to Starling’s law. (Source: Redrawn from Kristinsson JK, Gottfredsdottir MS, Stefansson E: Retinal vessel dilatation and elongation precedes diabetic macular edema. Br J Ophthalmol. 1997;81:274–278, with permission from the BMJ Publishing Group.)

60 Diabetes and Ocular Disease

leads to increased fluid accumulation in the macula. The oncotic force that pulls water from tissue into capillaries is determined by the plasma albumin concentration, so when albumin levels decrease below 3.0 mg/dL, the oncotic pull is sufficiently diminished to contribute to tissue edema.

Patients with diabetes frequently have impaired Starling’s equilibria. As shown in Table 4.3, the clinical risk factors for DME include increased intravascular volume due to hypertension, fluid overload (congestive heart failure and renal failure) and hypoalbuminemia from diabetic nephropathy.

Venous tortuosity and dilation are frequently noted in patients with progressive retinopathy. The physiologic basis of this feature owes to autoregulatory vasodilation of arterioles that causes intravascular pressure in the arterioles to decrease and that in the venules to increase, according to Poiseuille’s Law. The increased hydrostatic pressure also leads to greater blood vessel length and tortuosity, per LaPlace’s Law. Serial observations in patients with diabetes have shown that retinal vascular diameter and length increase prior to the onset of DME and improve following macular photocoagulation for DME [76] and after panretinal photocoagulation for proliferative diabetic retinopathy (PDR) [46].

In addition to altered autoregulation of vascular flow, the intrinsic integrity of the blood–retina barrier is also impaired. Studies with vitreous fluorometry in humans show that breakdown of the inner blood–retina barrier (formed by tight junctions between endothelial cells) predominates over changes in the outer barrier (tight junctions between retinal pigment epithelial cells) in early DME [77]. The outer barrier breaks down in patients with chronic DME. The proteins that comprise the tight junctions between vascular endothelial cells are reduced in early experimental diabetes, and this may account for increased vascular permeability [78]. As such, the hemodynamic abnormalities in the retina are analogous to those that occur in the kidney in early diabetes; that is, increased renal blood flow and increased glomerular permeability, with resultant albuminuria [79].

Other factors may also aggravate the overall severity of retinopathy. For example, hyperlipidemia has been associated with an increased risk of hard exudates and macular edema [80–82], and anemia is associated with worsening of retinopathy in general [83]. Anemia may impair oxygen delivery to the retina. In addition, erythropoietin may serve as a trophic factor for retinal cells [84] and its deficiency

Table 4.3. Mechanisms of Diabetic Macular Edema

Poor metabolic control

Increased hydrostatic pressure

Hypertension

Intravascular fluid overload (congestive heart failure, renal failure)

Decreased colloid oncotic pressure

Hypoalbuminemia

Hyperlipidemia

Anemia

might aggravate retinal cell death. Conversely, excessive intraocular erythropoietin levels may contribute to the development of DME and PDR [85,86].

Together, these risk factors give rise to principles of DME treatment, including improving metabolic control, blood pressure, fluid overload, anemia, and hyperlipidemia, as shown in Table 4.4.

Microaneurysms are the most characteristic ophthalmoscopic features of diabetic retinopathy. They occur throughout the posterior pole and are often first noted temporal to the macula. Their importance lies in their association with the retinopathy severity and as sources for leakage of fluid and lipid transudates. Histologically, they are outpouchings of the capillaries, with focal endothelial cell proliferation and pericyte loss, often adjacent to areas of nonperfusion. The factors that contribute to microaneurysm formation likely include structural features (loss of supporting pericytes and astrocytes), hemodynamic alterations (increased capillary intramural pressure), and local production of vasoproliferative factors, such as VEGF. Like cotton wool spots, retinal thickening, and hemorrhages, microaneurysms can wax and wane through the course of retinopathy [87].

Understanding the pathophysiology of DME allows construction of a set of systemic risk factors, such as poor diabetes control, systemic arterial hypertension, hyperlipidemia, and hypoalbuminemia.

PROLIFERATIVE DIABETIC RETINOPATHY

PDR, characterized by neovascularization of the optic disc, retina, and/or iris, may be an aberrant attempt to alleviate hypoxia in eyes with severe capillary closure or other retinal ischemia. However, despite the appearance of nonperfused retinal vessels in patients with PDR, retinal hypoxia has not been documented directly in patients [88]. Table 4.5 outlines the features of PDR. The new vessels grow perpendicular to the plane of the retina into the scaffolding provided by the vitreous cortex, typically from venules at the junction of perfused and nonperfused retina (Fig. 4.6). In contrast to normal retinal vessels, which are ensheathed by intact astrocytes, neovascularization is associated with reactive glial cells [89], which do

Table 4.4. Clinical Risk Factors for Diabetic

Macular Edema and Retinopathy

Poor metabolic control

Hypertension (>130/80 mm Hg)

Intravascular fluid overload

Congestive heart failure

Renal failure

Hypoalbuminemia

Anemia—Erythropoietin effects on retina

Hyperlipidemia