Age related macular degeneration (AMD) is the leading cause of irreversible visual loss in the industrialised world. Several theories of pathogenesis have been proposed and these include primary retinal pigment epithelium (RPE) and Bruch's membrane senescence, oxidative injury, primary genetic defects, and primary ocular perfusion abnormalities. In this issue of the BJO (p531), Mori and others explore ocular perfusion abnormalities by examining choroidal blood flow in patients with AMD, using pulsatile ocular blood flow (POBF). They used a Langham OBF computerised tonometer in 10 patients with non-exudative AMD, 11 patients with exudative AMD, and 69 age matched control subjects. They found statistically significant differences in the POBF (lower) and pulse amplitude (lower) in patients with exudative AMD compared with those with non-exudative AMD or with the control subjects. The authors conclude that decreased choroidal blood flow may play a part in the development of choroidal neovascular membranes (CNVM) in AMD. Although the technique of POBF carries some limitations as noted by the authors, this work serves to amplify and corroborate previous studies on the role of ocular perfusion perturbations in AMD. Studies of this sort are important with regard to our understanding of the pathogenesis of AMD.

Classically, investigators have postulated that senescence of the RPE, which metabolically supports the photoreceptors, leads to AMD.12 Senescent RPE accumulates metabolic debris as remnants of incomplete degradation from phagocytosed rod and cone membranes leading to drusen formation and further progressive dysfunction of the remaining RPE.12 Bruch's membrane, thickened with drusen, could be predisposed to crack formation.34 Calcification and fragmentation of Bruch's membrane is more prominent in eyes with exudative AMD, and these defects in Bruch's membrane could facilitate development of CNVM.5 This theory is supported by findings in myopic degeneration and angioid streaks in which CNVM develop through breaks in Bruch's membrane. The exact stimulus for CNVM formation is unclear; it is possible that macrophages involved in the initial response to Bruch's membrane injury secrete angiogenic growth factors. In addition, calcification and fragmentation observed in Bruch's membrane, which contains tissue inhibitors of metalloproteinases, may represent a breach in this antiangiogenic barrier, facilitating CNVM development. Whatever the initial stimulus for CNVM formation, it is clear that angiogenic growth factors are ultimately involved, as CNVM and RPE cells have been shown to be immunoreactive for various angiogenic growth factors.

Oxidative insults have also been proposed as a contributing factor and this may involve the macular pigments, lutein and zeaxanthin, which are primarily obtained from dark green, leafy vegetables and account for the yellow pigmentation of the macula lutea. Macular pigment has been hypothesised to have a protective role against the development of AMD through the limitation of oxidative insults by filtering out harmful wavelengths of light or by its antioxidant properties. A recent study showed that primates raised on carotenoid depleted diets had a significantly increased incidence of angiographic transmission defects in the macular regions,6 implying that the RPE is vulnerable to injury in the absence of normal macular pigment. Factors known to decrease macular pigment optical density (MPOD) levels, such as cigarette smoking,7 light iris colour,8and female sex,9 have also been implicated to increase the risk of AMD in epidemiological studies, consistent with a potential protective role of macular pigments in AMD. Previous studies have shown that a higher dietary intake of lutein and zeaxanthin has been associated with a lower risk for AMD,10 although there have been other large studies with conflicting results.

Another theory for AMD pathogenesis includes genetic defects. A variety of genes have been suggested. For example, some investigators recently reported a genetic defect in a gene encoding a retinal rod protein, the ABCR gene, which has also been found to be defective in Stargardt's disease.11 However, there have been other recent publications suggesting that the ABCR mutations might not be linked to AMD.1213 There have also been recent reports of a genetic association between AMD and apolipoprotein E, a protein that has a role in central nervous system lipid homeostasis.1415 Investigators are studying other hereditary dystrophies with some features similar to AMD, such as Doyne's honeycomb retinal dystrophy and Sorsby's dystrophy. Genetic research in AMD is clearly in its infancy and the ophthalmic community can look forward to many new developments in this field.

Another pathogenic theory involves primary vascular changes in the choroid, which then secondarily affect the RPE and lead to AMD. Specifically, it is theorised that lipid deposition in sclera and Bruch's membrane leads to scleral stiffening and impaired choroidal perfusion, which would in turn adversely affect metabolic transport function of the retinal pigment epithelium.1617 The impaired RPE cannot metabolise and transport material shed from the photoreceptors, leading to accumulation of metabolic debris and drusen.1617 This theory is supported by studies demonstrating an association between increased scleral rigidity and AMD.18 Proponents note that the vascular model could account for development of both the non-exudative and exudative forms of AMD. According to this vascular model, there is a generalised stiffening and increase in resistance, not only in the choroidal vasculature, but also in the cerebral vasculature.1617If the choroidal resistance increases more that the cerebral vascular resistance, there is a decrease in choroidal perfusion with an increase in the osmotic gradient against which the RPE must pump, leading to an accumulation of metabolic debris in the form of drusen. If the choroidal resistance increases less than the cerebral vascular resistance, there is higher choroidal perfusion pressure, which facilitates CNVM. This mechanism could partially account for the development of CNVM in the presence of Bruch's membrane cracks that result from senescence, as described above, and this explanation may partially unify these theories.

The vascular theory is also supported by studies demonstrating delayed choroidal filling in AMD using conventional angiographic techniques,19-21 laser Doppler flowmetry,22 colour Doppler imaging,161723 and ICG angiography.24 The study of Mori and others corroborates and amplifies these findings using a different technique. Consequently, there is no doubt that choroidal perfusion abnormalities exist in AMD. However, at the present time, it is not possible to determine if these choroidal perfusion abnormalities have a causative role in non-exudative AMD, if they are simply an association with another primary alteration, such as a primary RPE defect or a genetic defect at the photoreceptor level, or if they are more strongly associated with one particular form of this heterogeneous disease. Clearly, future progress in developing effective treatment strategies for this devastating disorder hinges on a better understanding of disease development.