Our results suggest that foveolar ChBFlow is lower than normal in eyes with AMD, confirming the findings in our previous study.
5 In addition, eyes with more AMD fundus features associated with risk for the development of CNV tend to show more pronounced decreases in ChBFlow. All three circulatory parameters—ChBVel, ChBVol, and ChBFlow—decrease progressively with an increase in the severity of AMD features (linear trend test,
P = 0.047, 0.02, and 0.003, respectively). The linear trends remain statistically significant for ChBVel after adjustment for age, and for ChBFlow after adjustment for age and hypertension. Such adjustments are important, because increasing age has been shown to be associated with decreased ChBFlow in normal subjects.
19 20 patients with AMD show ChBFlow decreases that are larger than those produced by aging in normal eyes.
5 In addition, within this group, subjects with a history of hypertension have lower choroidal flow than subjects without a history of hypertension.
Multiple studies have investigated the association between various AMD fundus features and the risk of development of CNV. Large drusen, RPE hyperpigmentary changes in the study eye and the presence of CNV in the fellow eye have been shown to be associated with increased risk. Based on these studies we divided our AMD eyes into three groups, according to increasing risk of CNV.
Several studies have assessed the prognostic information of drusen regarding the development of CNV by looking at drusen size, type, confluence, and drusen area.
21 22 23 Bilateral drusen have been reported to be associated with an increased risk of the development of CNV.
23 24 25 26 Patients with bilateral drusen had cumulative risks of CNV development over the period of 5 years that ranged from 14% to 27%.
21 23 25 26
Smiddy and Fine
25 also reported that the additional presence of focal RPE hyperpigmentation was associated with a 23-fold greater risk of development of an exudative process. Similarly, Leeuwen et al.
27 reported that hyperpigmentation tripled the risk of AMD in subjects with drusen within a 5-year period. In other studies, the 5-year cumulative risk for development of CNV in eyes with both large drusen and RPE changes ranged between 7.1% and 48%.
21 26 28
Several studies have also shown that patients with CNV in one eye have the highest risk for development of CNV in the second eye.
23 27 30 The annual rates vary from 0.6% in a Japanese study
29 to 18% among Americans.
31 This is indeed the type of patient included in group 3 of our study. These patients with the highest risk for CNV showed the lowest ChBFlow of all groups studied, a finding that strongly suggests that decreased flow and ischemia may play a role in the development of CNV in AMD, in the same way that ischemia may trigger neovascularization in other tissues of the body.
3 4
Our results showing the lowest flow in patients with CNV in the fellow eye are in accord with several studies that have suggested lower ChBFlow in AMD eyes with CNV. Mori et al.
9 reported that pulse amplitude in patients with exudative AMD is lower than in patients with nonexudative AMD and in age-matched control subjects. Chen et al.
8 showed that pulse amplitude was significantly decreased in eyes with disciform scarring in comparison to the contralateral eyes with drusen. Although the relationship between pulse amplitude and ChBFlow has not been clearly established, these results suggest that perhaps ChBFlow decreases in AMD eyes with CNV.
Using color Doppler imaging, Uretmen et al.,
10 reported decreased ophthalmic artery and temporal posterior ciliary artery velocities in eyes with CNV in comparison to eyes with nonexudative AMD, and Rigas et al. (
IOVS 2004;45:ARVO E-Abstract 3110) found an increased resistance to blood flow in eyes with neovascular AMD, suggesting decreased ChBFlow. Dimitrova,
32 however, compared both eyes of patients with unilateral neovascular AMD and did not find any significant differences in resistance between neovascular and non-neovascular AMD.
Our findings showing decreased ChBFlow in AMD eyes also fit well with previous histopathologic studies. Sarks
33 and Sarks et al.
34 have shown reduction in the cross-sectional area of the choriocapillaris in AMD. These data are also in agreement with Ramrattan et al.,
35 who demonstrated that the density and diameter of the macular choriocapillaris decreases in AMD in comparison to normal control subjects, the report of Korenzweig
36 showing narrowing of the lumen and loss of cellularity of the choriocapillaris in AMD, and the data of Arnold et al.
37 and Spraul et al.,
38 suggesting the loss of choroidal veins in “reticular pseudodrusen.”
Our results and these histopathologic findings are also in accord with evidence of choroidal perfusion abnormalities in AMD observed by investigators using different techniques. Pauleikhoff et al.
39 and Boker et al.,
40 reported choroidal perfusion abnormalities on fluorescein angiograms of patients with AMD. Chen et al.
41 described areas of delayed choroidal perfusion that were associated with decreased visual function in patients with AMD. Holz et al.
24 showed that slow choroidal filling is a significant risk factor for the development of geographic atrophy.
More recently, Pauleikhoff et al.
42 reported that prolonged filling of the choroidal lobules in the early phases of fluorescein angiography, and reduced choroidal fluorescence on indocyanine angiography are common features in eyes with early AMD. Using color Doppler imaging, Friedman et al.,
43 showed that blood velocity decreases and blood velocity pulsatility increases in the central retinal artery and short posterior ciliary arteries in AMD. Based on some of these findings, they propose that AMD may be associated with an increase in the resistance of the choroidal vasculature caused by a decrease in the compliance of the sclera and the choroidal vessels.
43 Using color Doppler imaging, Ciulla et al.
44 have also shown decreased blood velocities in the retrobulbar vasculature. In a different study using scanning laser ophthalmoscopy and indocyanine green angiography, Ciulla et al.
45 reported increased heterogeneity of choroidal filling time in patients with nonexudative AMD.
45 All these findings, which are in agreement with our results, strongly suggest that vascular impairment and ischemia play an central role in the etiology of AMD.
Disturbances in the choroidal circulation just mentioned could hinder the normal diffusion of substances and gasses across the RPE-Bruch’s membrane complex, which has a crucial role in visual function. This disruption could lead to a situation in which waste materials may not be readily removed and crucial metabolites and gasses may not be adequately supplied to the neural retina. Because the choroidal circulation is the only source of nourishment and waste removal for the outer retina, particularly in the foveola, any alterations in choroidal circulation could be very deleterious.
Because the choroidal circulation decreases with age
19 in the normal eye and because this decrease is further exacerbated by AMD,
5 it is possible that the removal of waste products from the RPE-Bruch’s membrane complex may become impaired in AMD. This process could be part of the mechanism that leads to drusen accumulation in this disease.
Our results show an association between decreased ChBFlow and increased severity of AMD. From our results, however, we cannot conclude whether these decreases have a role in the development of AMD. We also cannot exclude the possibility that the decreases in choroidal flow may be related to the loss of photoreceptors that occurs in AMD. As photoreceptors are lost, the declining demand for oxygen from the choroid could result in decreased blood flow.
The laser Doppler flowmetry technique provides measurements of relative blood velocity, volume, and flow. Because changes in the intensity and coherence of the laser light produced by AMD can theoretically affect the hemodynamic measurements, the comparison of LDF relative blood flow measurements between normal and AMD eyes is open to question.
5 Our results showing that ChBVel decreases with increased risk of CNV support our contention that ChBFlow is decreased in this disease because ChBVel measurement are less affected by changes in the media than ChBVol and ChBFlow measurements. That other histopathologic and circulatory studies have also shown evidence of decreased circulation in AMD further supports our findings.
In summary, our present study suggests that ChBFlow decreases with increase in the severity of AMD features, pointing to a potential role for ischemia in the development of CNV. We cannot conclude from our study, however, whether this decrease in flow triggers the development of CNV. Further studies are necessary to reach a strong conclusion as to whether this association between AMD fundus features and degree decrease in ChBFlow may help identify patients with AMD at risk for visual loss.
The authors thank Judith Alexander, Director, and Readers Revell Whittock and Keith Elsner at the Scheie Image Reading Center for assistance with this project.