C ontrary to the widespread belief that there’s a strong association between cardiovascular disease (CVD) and age-related macular degeneration (AMD), the published literature shows a very weak or no association at all. ^, This association between AMD and CVD has been investigated for over 30 years, and while CVD is the leading cause of death in the developed world and AMD is the leading cause of irreversible blindness among the elderly worldwide, the association between them is tenuous. Since both diseases occur in the elderly and both contain lipid-laden lesions with common risk factors, the connection between the 2 diseases was thought to be a forgone conclusion, but the findings have been inconsistent with correlations from studies being positive, neutral, or even negative with the only positive correlations being restricted to specific disease categories such as early AMD or exudative (neovascular) AMD (eAMD). Are we to believe that no strong association exists or have the studies failed to associate the right form of CVD with the appropriate AMD endpoint? ^,,,,,

To identify all the reports that attempted to associate CVD with AMD, we searched PubMed with the MeSH terms (“Macular Degeneration”[Major]) AND (“Heart Diseases”[Major]) OR (“Coronary Artery Disease”[Major]) OR (“Myocardial Infarction”[Major]) OR (“Stroke”[Major])). The search yielded 65 papers, with 28 studies appropriate for review, including 2 large metanalyses. These studies included many of the major population-based studies of AMD and CVD; the Atherosclerosis Risk in Communities Study (ARiC), the Blue Mountain Eye Study (BMES), the Australian Heart Eye Study (AHES), the Multi-Ethnic Study of Atherosclerosis (MESA), and the Rotterdam Study. Of note, each study investigated a different stage of AMD, such as early and/or late, and different categories of CVD, such as all CVD, all CVD and risk factors, and a coronary disease study population of ∼81,000, of which ∼7700 had any AMD. In these studies, no association was found between AMD and CVD. ^,, Furthermore, no associations were found in the largest metanalyses considering AMD with CVD (N = 29,964,334) and AMD with stroke (N = 1,420,978). ^, In summary, the studies themselves all agreed that the results were ambiguous, and more investigations were needed.

Perhaps, the previous association studies didn’t show a strong association between AMD and CVD because the categories of early, intermediate, and late AMD were too broad. One possibility was that if a strong association existed, then it would be between specific phenotypes of AMD and certain high-risk cardiovascular diseases (HRCVDs) within the family of CVDs. Another possibility is that an association exists between CVDs and rates of disease progression in AMD.

Unlike the large population-based studies, several smaller targeted studies have indeed shown an association between a specific form of AMD and certain HRCVDs. ^,, A strong association was reported between the sub-phenotype of AMD known as subretinal drusenoid deposits (SDDs) and a certain class of HRCVDs within the family of CVDs. This may explain why previous studies, no matter how well executed, did not find consistent general associations between all of AMD and all of CVDs. Moreover, these larger studies were confounded by using different stages of AMD that may not have contained SDDs, such as eyes with typical soft drusen in AMD.

We propose that the specific association between SDDs and specific HRCVDs is due to insufficient ophthalmic perfusion over many years that is caused by these HRCVDs. SDDs would then serve as the biomarkers for this decreased perfusion and the presence of HRCVDs. If these HRCVDs were more lethal than other CVDs, then this would explain the observation by Ron Klein that AMD subjects with SDDs have shorter life spans than those with only drusen. If perfusion is the link between CVDs and AMD, then the most likely causes of decreased ophthalmic perfusion would be impaired carotid and ophthalmic artery perfusion that result from atherosclerotic disease, decreased cardiac output, increased choroidal vascular resistance, or increased post-choroidal resistance. We also propose that decreased ocular perfusion isn’t the complete cause of AMD, but rather, this decreased ocular perfusion in individuals, who are genetically at-risk for developing AMD, will result in earlier onset and more rapid disease progression. SDDs, already shown to be associated with faster progression to late AMD than observed in eyes with typical soft drusen, would then serve as the biomarkers for this decreased perfusion. This explanation would be consistent with the association between SDDs and thinner choroid thickness measurements and other indicators of disease progression such as the increased formation and growth rates of GA and the formation of type 3 macular neovascularization (MNV).

SDDs have been identified in several ocular disease such as AMD (intermediate and late-stage disease), inherited retinal dystrophies such as Sorsby fundus dystrophy, pseudoxanthoma elasticum (with angioid streaks), and fundus albipunctatus, vitamin A deficiency retinopathy, preeclampsia with choroidopathy, and central serous chorioretinopathy (CSCR), particularly in older CSCR patients.

While there are different genetic and physiological causes for these aforementioned diseases, we propose that each develops SDDs due to the disruption of the normal metabolic support between the choriocapillaris and photoreceptors (PRs) that results in the accumulation of SDDs between the PRs and the RPE. In the case of AMD in general, this disruption is thought to be due to decreased choroidal perfusion resulting in a thinned choroid and increased in choriocapillaris (CC) flow deficits (CCFDs) combined with a thickened RPE/Bruch’s membrane (BM) complex that results from the deposition of basal laminar deposits, which also causes reduced metabolic support for the RPE and PRs. In the case of diminished CC perfusion causing SDDs in AMD, this localized outer retinal ischemia might also explain the increase in prevalence of type 3 MNV that arises within the retina and the association with the increased growth rate of geographic atrophy (GA). Indeed, this localized ischemia may also explain the greater likelihood of type 3 MNV evolving into atrophy in eyes with SDDs after treatment with intravitreal vascular endothelial growth factor (VEGF) inhibitors.

If SDDs serve as a surrogate marker for regions where choroidal perfusion and the metabolic exchange between the CC and PRs are abnormal, then this may explain why a strong association has been established between HRCVDs and SDDs. This association between HRCVD and SDDs might be found to be even stronger if choroidal hypoperfusion could be shown to correlate with both the onset and growth of SDDs. However, SDDs are notoriously difficult to quantitate due to their appearance and enlargement outside the macula and their tendency to disappear as AMD and other diseases progress.

While the distribution of SDDs in AMD tend to be found outside the central macula in areas dominated by rod photoreceptors, they can also occur in the central macula in AMD, where they are associated with decreased CC perfusion. One possible explanation for this common peripheral distribution of SDDs in AMD may be due to the metabolic fragility of rods compared with cones since cones receive additional metabolic support from Mueller cells in the macula.

While SDDs are known to be associated with disease progression, with HRCVDs, and with increased CCFDs, we need to study other features of AMD disease progression that might be associated with decreased choroidal perfusion and HRCVDs. Such features of AMD progression include the change in macular soft drusen volume and area, the change in the macular burden of OCT hyperreflective foci, ^, also known as hyperpigmentation, the appearance and progression of calcified drusen or drusen with hyporeflective cores, the progression of iAMD to the early changes of atrophic disease such as the appearance and growth of large hypertransmission defects (hyperTDs), changes in the outer retinal thickness and photoreceptor OCT bands, ^, changes in the thickness of the choroid and the perfusion of the CC ^, and the onset of eMNV and the number of anti-VEGF injections needed to control exudation. These association studies may be complicated by the genetics of any given patient with AMD. However, this perfusion model for disease progression may be agnostic to the underlying genotype and just require that an individual be genetically at-risk for AMD.

More detailed and prospective quantitative association studies are needed between HRCVD and AMD. We need methods that can quantify choroidal perfusion such as laser speckle flowgraphy (LSFG), and methods that can reliably measure the rates of AMD progression. Using OCT and OCTA methods, we’ve established a number of previously referenced strategies to measure the progression of AMD; however, LSFG may be the most promising clinical test to measure choroidal perfusion along with the use of advanced magnetic resonance angiography imaging with a 64-channel head coil and algorithms for quantifying blood flow in the ophthalmic artery ^, We predict that a strong association between HRCVD and AMD will be found once we start studying the impact of decreased ocular perfusion on the rate of AMD progression.

R. Theodore Smith: Writing– review & editing, Writing– original draft, Project administration, Methodology, Investigation, Funding acquisition, Formal analysis, Data curation, Conceptualization. Philip J. Rosenfeld: Writing– review & editing, Writing– original draft, Project administration, Methodology, Investigation, Funding acquisition, Formal analysis, Data curation, Conceptualization.