The “Diffuse-Trickling” Fundus Autofluorescence Phenotype in Geographic Atrophy

Abstract

Purpose.: To further characterize a subgroup of patients exhibiting the fundus autofluorescence (FAF) “diffuse-trickling” phenotype associated with geographic atrophy (GA).

Methods.: In the context of the Fundus Autofluorescence in Age-Related Macular Degeneration (FAM) Study, patients with diffuse-trickling GA were examined and characterized by FAF and spectral-domain optical coherence tomography imaging. Age, sex distribution, and medical history were compared with FAM study patients (n = 288, 60.1% female) with other GA phenotypes (non–diffuse-trickling). In a subset of patients, subfoveal choroidal thickness (SCT) was analyzed.

Results.: Patients with diffuse-trickling (n = 61), compared with patients with non–diffuse-trickling GA, had a significantly younger age at first presentation (68.2 ± 11.6 vs. 75.4 ± 8.1 years, P < 0.001), a shift in the proportion of men from 55% in the age group younger than 65 to 19% in the age group older than or equal to 65, and a significantly higher rate of myocardial infarction (MI) in the age group younger than 65 (24% vs. 0%, P = 0.011); all but one patient with MI were male. Further evaluation revealed that in the age group younger than 65, 54% of patients with diffuse-trickling had been hospitalized due to cardiovascular diseases including hypertensive crisis, angina, and MI. Analysis of choroidal thickness revealed a significantly thinner SCT in diffuse-trickling compared with non–diffuse-trickling GA (135.2 ± 56.4 vs. 191.4 ± 77.8 μm, P < 0.001).

Conclusions.: The results indicate an association of diffuse-trickling GA with systemic cardiovascular disorders in the younger study population. Together with the ocular morphologic characteristics including a lobular appearance and a thin choroid, a vascular insufficiency at the level of the choroid may play a pathogenetic role in this distinct GA phenotype. (ClinicalTrials.gov number, NCT00393692.)

Introduction

Geographic atrophy (GA) represents a morphologic end-stage in various retinal diseases including advanced AMD. While choroidal neovascularization (CNV) is the most common cause of severe visual loss in advanced AMD, approximately 20% of AMD-patients who are legally blind have lost central vision due to GA. ^1–5 Geographic atrophy in AMD usually develops in presence of early disease alterations at the level of the retinal pigment epithelium (RPE), and Bruch's membrane with drusen and/or pigmentary alterations. ^6–11  

Histologically, areas of GA are characterized by loss of the RPE, of outer layers of the neurosensory retina and the choriocapillaris. ^10,12 The initiating event in GA development, however, is yet unknown. It is assumed that alterations in the photoreceptor layer do not develop independently of alterations at the level of the RPE and the reduction in photoreceptor nuclei appears to parallel loss of the RPE. ¹⁰ However, a recent morphologic study in donor eyes with GA by Bird and Hageman (oral communication, 2011) disclosed marked photoreceptor loss outside GA with corresponding normal appearing RPE cells in some eyes with GA. Furthermore, the development of GA in AMD has been ascribed to both, alterations at the level of Bruch's membrane and degenerative changes of the choriocapillaris. ⁹ The facts that the choriocapillaris is the major source of nutrients for the RPE and outer retinal layers, ¹³ and that patches of RPE atrophy may appear as if there is correspondence to the size of choriocapillary lobules, ^14,15 point to the possibility of a role of the choroid and its perfusion in the disease process. Further histologic as well as functional data, including fluorescence angiography and Doppler flowmetry suggest that choroidal insufficiency may have an impact in multifactorial AMD pathogenesis. ¹⁶  

Possibly, different pathways with heterogeneous etiologies may be operative in the development of GA resulting in phenotypically similar appearing end-stage atrophy currently lumped under the terms “atrophic AMD” or “GA.” 

Recent developments in retinal imaging technologies now allow for more refined phenotyping of various retinal diseases. They have been shown useful to discriminate different variants of GA. For example, using fundus autofluorescence (FAF) imaging, distinct patterns of GA secondary to late-onset Stargardt macular dystrophy and central areolar chorioretinal dystrophy have been identified. ^17–19 The presence of such FAF characteristics is related to mutations in the ABCA4 gene and PRPH2 gene, respectively. 

The so called “diffuse-trickling” pattern represents another abnormal FAF pattern associated with GA that has previously been identified in the context of the Fundus Autofluorescence in Age related Macular Degeneration (FAM) Study (ClinicalTrials.gov Identifier: NCT00393692). ^20,21 Compared with other GA subtypes, various characteristic morphologic features in diffuse-trickling GA have been described using different imaging modalities: the most prominent findings are (1) a grayish rather than a markedly decreased FAF signal in areas of atrophy that is typically seen in other GA subtypes ²⁰ ; (2) characteristic coalescent lobular configuration of the atrophic patches ^20,21 ; (3) an obvious separation of the RPE/Bruch's membrane complex in spectral-domain optical coherence tomography (SD-OCT) imaging, which has been interpreted as excessive accumulation of sub-RPE deposits ²⁰ ; and (4) a significantly faster progression of the atrophic areas in the diffuse-trickling phenotype as compared with the other GA subtypes. ^20,21  

Interestingly, a preliminary analysis of patient records pointed toward a high rate of hypertension and myocardial infarction (MI) in younger patients with the diffuse-trickling GA phenotype. 

Given the lobular pattern of the atrophic lesions that might be reminiscent of the lobular configuration of the choriocapillaris in the macula ^14,15 and the phenotypic resemblance with the previously described age-related choroidal atrophy, ²² we hypothesize that choroidal perfusion abnormalities may play a role in evolution of diffuse-trickling GA. To further explore this hypothesis, data in this respect were recorded and analyzed. 

Methods

Subjects

All subjects presented in this study were prospectively recruited in the FAM study, a multicenter, longitudinal, natural history study in patients with AMD. ^21,23–30 The inclusion and exclusion criteria for the GA-arm of this study have been described in detail previously. ²¹ Briefly, patients had to be above 50 years of age (at the time of inclusion) and exhibit GA in at least one eye. Exclusion criteria included any history of retinal surgery, laser photocoagulation, and radiation therapy or other retinal diseases in the study eye, including diabetic retinopathy or hereditary retinal dystrophies. The study followed the tenets of the Declaration of Helsinki and informed consent was obtained from each patient prior to inclusion into the study. 

At each visit, patients underwent a routine ophthalmologic examination, including determination of best corrected visual acuity (BCVA) and each patient was interviewed by a trained study coordinator or investigator using a standardized case report form (CRF) that included ocular history as well as current medications. Furthermore, at the baseline visit, family history for retinal diseases, smoking history, and systemic diseases (including diabetes mellitus, hypercholesterolemia, hypertension, MI, and stroke) were recorded. 

Imaging Acquisition

Serial retinal imaging examinations were performed according to standardized operation protocols. ²¹ Confocal scanning laser ophthalmoscopic (cSLO) examination (HRA classic, HRA2, or Spectralis HRA+OCT; Heidelberg Engineering, Heidelberg, Germany) included FAF (λ = 488 nm, emission 500–700 nm), infrared reflectance (IR), and red-free reflectance. Fluorescein angiography was performed when there was suspicion of CNV. Color fundus photographs with a fundus camera were obtained at least once in every patient. Time-domain OCT (Stratus OCT; Carl Zeiss Meditec, Jena, Germany) became available during the early course of the study and was subsequently included into the imaging protocol. In 2007, high-resolution SD-OCT was incorporated, using a combined instrument (Spectralis HRA+OCT; Heidelberg Engineering) that allows for simultaneous recording of cSLO and SD-OCT. ³¹  

Determination of GA Progression

The total size of GA was measured in processed FAF images by semi-automated imaging analysis software (based on HALCON software; IDS GmbH, Obersulm, Germany) that uses region-growing techniques to segment the areas of GA. ^24,32 The measurement strategy has previously been described in detail. ²⁹ A linear mixed-effects model was used to quantify overall GA growth. A detailed description of the longitudinal modelling process is given elsewhere. ³³  

Measurement of Subfoveal Choroidal Thickness

In order to determine the subfoveal choroidal thickness (SCT), SD-OCT images were analyzed. Using the Heidelberg Eye Explorer software (Heidelberg Engineering), the vertical choroidal extension beneath the foveal depression was measured from the outer border of OCT band 4 (corresponding to the RPE/Bruch's membrane complex) to the inner scleral border. ³⁴  

Phenotyping

Using the previously introduced classification of GA based on abnormal FAF patterns (Fig. 1), ²¹ the specific FAF phenotypic pattern at baseline was independently assessed for each eye by two independent graders (MF, SSV). In case of discrepancy, a third reader (FGH) was asked for arbitration. Briefly, the classification includes five different distinct phenotypic patterns based on FAF features only (Fig. 1). Hereby, the diffuse type (which is split into the diffuse-reticular, diffuse-branching, diffuse-fine granular, diffuse-trickling, and diffuse-GPS types, respectively) is characterized by levels of abnormal increased FAF intensities extending beyond the border zone of GA and may spread outside the vascular arcades and nasally to the optic nerve head, while the other types typically show either no or abnormal FAF only adjacent to the atrophic patches (none, focal, banded, and patchy, respectively). Compared with other diffuse subtypes, the diffuse-trickling type shows coalescent lobular atrophic lesions (Figs. 1, 2). The FAF signal of the atrophic lesions in diffuse-trickling appears rather grayish than dark black opposed to outer retinal atrophy seen in other GA subtypes. Adjacent to atrophic patches, the FAF signal is markedly enhanced with a diffuse “trickling” FAF signal toward the periphery. 

Figure 1

 

Classification of GA based on abnormal FAF patterns. Reprinted from Fritsche LG, Fleckenstein M, Fiebig BS, et al. A subgroup of age-related macular degeneration is associated with mono-allelic sequence variants in the ABCA4 gene. Invest Ophthalmol Vis Sci. 2012;53:2112–2118. The classification includes five different distinct phenotypic patterns. Hereby, the diffuse type (which is split into the diffuse-reticular, diffuse-branching, diffuse-fine granular, diffuse-trickling, and diffuse-GPS types, respectively) is characterized by levels of abnormal increased FAF intensities extending beyond the border zone of GA, while the other types typically show either no or abnormal FAF only adjacent to the atrophic patches (none, focal, banded, and patchy, respectively). Compared with other diffuse subtypes, the diffuse-trickling type shows coalescent lobular atrophic lesions that appear rather grayish than dark black opposed to the atrophy in other GA subtypes.

Figure 2

 

Multimodal imaging (FAF, color fundus, infrared reflectance [IR], and SD-OCT) in patients with diffuse-trickling geographic atrophy. In FAF imaging, the atrophic area appears more grayish (upper and lower panel, FAF image, *) rather than dark black as seen in other GA subtypes. The FAF intensity adjacent to the atrophy is enhanced (upper and power panel, FAF image, <) with a diffuse trickling of the FAF signal toward the periphery. Funduscopically, eyes with the diffuse-trickling phenotype exhibit dense granular hyperpigmentary changes in the central macula (upper panel, color image, +) and the border of the atrophic lesions appears hyperpigmented (lower panel, color image, +); furthermore, adjacent to the atrophic lesions, there are small yellowish dots (upper and lower panel, color image, #). A reticular pattern, best seen on IR images (upper and lower panel, IR image, ˆ) is present corresponding to reticular pseudodrusen (lower panel, color IR image, ˆ). Spectral-domain OCT imaging shows an obvious separation of the RPE/Bruch's membrane complex (arrows).

Funduscopically, eyes with the diffuse-trickling FAF pattern exhibit dense granular hyperpigmentary changes in the central macula and the border of the atrophic lesions appears hyperpigmented (Fig. 2). Furthermore, adjacent to the atrophic lesions, there are small yellowish spots. Reticular pseudodrusen (also known as: reticular drusen, reticular macular disease, or subretinal drusenoid deposits) are present, whereas soft drusen are rather infrequent in this phenotype. ²⁰ Spectral-domain OCT imaging shows an obvious separation of the RPE/Bruch's membrane complex (Fig. 2). ²⁰  

Cohort of Subjects With the Diffuse-Trickling Pattern

A total of 61 subjects were classified to exhibit the diffuse-trickling pattern in the FAM study database. Beyond the established standardized CRF as described above, patients with diffuse-trickling GA were further evaluated in more detail with regard to cardiovascular risk factors, cardiovascular diseases, and age of onset of visual symptoms. This additional interview was either conducted during the study visit or subjects were contacted by phone. Specifically, the following information was obtained in a systematic manner: if hypertension was reported, patients were asked for their age when hypertension was diagnosed initially. Additionally, patients were asked if they had ever suffered from an acute rise in blood pressure of greater than 180/120 mm Hg (hypertensive crisis ³⁵ ), which necessitated immediate pharmacologic blood pressure reduction and hospitalization. Furthermore, if a history of MI was recorded, patients were asked for the age when the MI occurred. 

Moreover, patients were asked for the presence of any other cardiovascular disease (e.g., diabetes mellitus, hypercholesterolemia, cardiac arrhythmia, congestive heart failure (CHF), stroke, peripheral arterial occlusive disease, symptomatic hypotension) and if they had been hospitalized for any cardiovascular event. If so, age and reason for hospitalization were documented. 

To prove the previous observation of a preliminary analysis that there is a high rate of hypertension and MI in younger patients with the diffuse-trickling GA phenotype, patients aged younger than 65 years when diagnosis of diffuse-trickling GA was made were separately analyzed. 

Comparison to Subjects With Other Geographic Atrophy Phenotypes

Demographic data and medical history were compared between subjects with the diffuse-trickling phenotype and other subjects that had been recruited in the FAM study to describe differences between the two groups. For the latter group, a retrospective review of the FAM study database identified 288 GA subjects with phenotypes other than the diffuse-trickling pattern and completed standardized CRFs. 

Differences in age were analyzed using a t-test. Using the Fisher exact test (PASW Statistics 18; IBM SPSS, Armonk, NY, USA) the following parameters were analyzed: sex, smoking history, history of diabetes mellitus, hypercholesterolemia, hypertension, MI, and stroke. Since the analysis was mainly descriptive, P values were not corrected for any kind of multiple testing. 

To compare SCT between diffuse-trickling GA and non–diffuse-trickling GA, 34 patients of the latter group were examined by enhanced depth imaging (EDI) SD-OCT imaging (Spectralis HRA+OCT; Heidelberg Engineering). Subfoveal choroidal thickness was measured as described above. The Mann-Whitney U test was used for comparison of SCT between the two patient groups. Only one eye per patient was included into the analysis of SCT. This was the study eye in unilateral GA patients and the right eye in bilateral GA patients. 

Results

The mean age at first exam of the 61 patients with the diffuse-trickling type was 68.2 ± 11.6 (range, 47.3–89.5; median age: 66.3 [IQR, 58.1–78.3]) years. 

All subjects (n = 43) with bilateral GA showed the diffuse-trickling phenotype bilaterally. In 18 subjects with unilateral diffuse-trickling GA, the fellow eye showed CNV in 15 and hyperpigmentary changes with no signs of either GA or CNV in three subjects. Hence, there were 104 eyes with diffuse-trickling GA at baseline exam. 

A total of eight patients (16 eyes) had typical soft drusen while all eyes with diffuse-trickling GA exhibited reticular pseudodrusen. 

Overall, the mean visual acuity (BCVA) of the better-seeing eye at first exam was logMAR 0.4 ± 0.4 (minimum logMAR 1.6, maximum logMAR 0). 

Data of the spherical equivalent were recorded in 82 eyes (50 patients) with diffuse-trickling GA. Fifty-nine of these 82 eyes (72%) were myopic (spherical equivalent ≤ −0.5 diopters [D]). Thereof 25 eyes (30.5%, 25/82) had a spherical equivalent less than or equal to −3. 

A positive family history for retinal disease was reported by 10 patients. Only one of these relatives could be examined: a patient's mother exhibited bilateral late-stage AMD but other than diffuse-trickling. The exact diagnosis in affected family members of the other nine patients was not available. The degree of relationship did not allow to conclude on a specific inheritance pattern. 

Longitudinal data were available for 41 subjects with the diffuse-trickling phenotype. Over time, five eyes with initial diffuse-trickling GA developed CNV. Overall, there were 20 of 61 patients (33%) with diffuse-trickling GA that had CNV in the fellow eye at baseline or developed CNV during the course of the study, respectively. 

Geographic atrophy progression could be determined in 50 eyes of 29 patients with a median of 2.9 (IQR, [2.11–3.83]) mm²/y. In 14 eyes with GA and serial FAF examination, GA size could not be determined longitudinally, because either the atrophic lesions exceeded the image frames or image quality was not sufficient to perform serial measurements. 

Further Evaluation of Patients With the Diffuse-Trickling Phenotype

A total of 41 patients (mean age 64.6 ± 10.4 [range, 47.3–86.5] years, 28 female and 13 male) with diffuse-trickling GA were additionally interviewed to further assess their cardiovascular risk factors, cardiovascular diseases, and age of onset of visual symptoms in a systematic manner as described above. The other 20 patients with the diffuse-trickling phenotype could not be contacted anymore to conduct this detailed interview. 

Twenty-five of the 41 interviewed patients (61%) had a history of hypertension. The mean age when diagnosis of hypertension was made was 55.1 ± 15.2 (range, 27–83) years. In the age group younger than 65, 9 of 12 female patients (75%) reported suffering from hypertension. 

Seven patients with a history of hypertension reported hospitalization due to hypertensive crisis. Six of them were female and all of them were younger than 65 years when diffuse-trickling GA was diagnosed; hence, 50% of female patients of the age group younger than 65 had suffered from hypertensive crisis. 

There were six patients (one female) with a history of MI. The mean age, when MI occurred was 47.5 ± 14.2 (range, 27–67) years. All five male patients with a history of MI were diagnosed with diffuse-trickling GA at an age younger than 65 years; hence, 42% (5/12) of male patients of the age group younger than 65 reported having suffered from MI. 

A total of 18 patients (44%) reported hospitalization due to cardiovascular diseases. The mean age when these 18 patients were first hospitalized was 49.2 ± 15.0 (range, 27–83) years. The reasons for hospitalization included hypertensive crisis, angina, MI, CHF, peripheral occlusive artery disease, and stroke, respectively. Thirteen of these patients were younger than 65 years when diffuse-trickling GA was diagnosed; hence, 54% (13/24) of patients of the age group younger than 65 had been hospitalized due to cardiovascular disease. 

The 41 interviewed patients with diffuse-trickling GA reported a mean age of onset of visual symptoms of 61.6 ± 10.4 (range, 46–83) years. In the patients reporting a severe cardiovascular event such as hypertensive crisis or MI, there was no obvious chronologic correlation between this event and the onset of visual symptoms. 

Comparison to Subjects With Other Geographic Atrophy Phenotypes

Age and Sex Distribution.

In patients with diffuse-trickling GA, the age at first examination (mean 68.2 ± 11.6 [range, 47.3–89.5], median 66.3 [IQR, 58.1–78.3] years) was significantly younger as compared with patients with other GA subtypes (non–diffuse-trickling; mean 75.4 ± 8.1 [range, 51.0–94.0] years, median 76.0 [IQR, 69–81] years; P < 0.001, t-test; Fig. 3). 

Figure 3

 

Age at first examination of diffuse-trickling GA patients and other GA patients (non–diffuse-trickling). Patients with diffuse-trickling GA are significantly younger at first examination (mean 68.2 ± 11.6 [range, 47.3–89.5], median 66.3 [IQR, 58.1–78.3] years) as compared with patients with non–diffuse-trickling GA (mean 75.4 ± 8.1 [range 51.0–94.0] years, median 76.0 [IQR, 69–81] years; P < 0.001, t-test).

While 48% (n = 29) of diffuse-trickling patients aged younger than 65 years at first exam, only 9% (n = 27) of non–diffuse-trickling GA patients were younger than 65 years at initial visit (Fig. 4). 

Figure 4

 

Distribution of patients with diffuse-trickling GA and non–diffuse-trickling GA by age groups younger than 65 and older than or equal to 65 at first exam.

Analysis of the sex distribution between both groups revealed a marked “sex shift” with age in the diffuse-trickling GA group (Fig. 5): the male proportion dropped from 55% in the age group younger than 65 to 19% in the age group older than or equal to 65; by contrast, in non–diffuse-trickling GA, the male proportion in both age groups was approximately 40%. Hence, the proportion of male patients in the age group older than or equal to 65 was significantly lower in the diffuse-trickling as compared with the non–diffuse-trickling GA group (P = 0.02, χ² test). 

Figure 5

 

Male proportion in diffuse-trickling GA and non–diffuse-trickling GA patients in the age groups younger than 65 and older than or equal to 65 at first exam.

Cardiovascular Risk Factors and Diseases.

In the total diffuse-trickling GA group (n = 61), 39% of patients had a history of smoking, 16% were heavy smokers with greater than 20-pack years; 4.9% had diabetes, 28% hypercholesterolemia, 59% hypertension, 13% had suffered from MI, and 8% from stroke. Comparison of the total study patients revealed no significant difference in the assessed cardiovascular risk factors and diseases between diffuse-trickling (n = 61) and non–diffuse-trickling (n = 288) GA patients (P = 0.081 for diabetes, P = 0.546 for hypercholesterolemia, P = 0.158 for smoking history, P = 0.887 for hypertension, P = 0.343 for MI, and P = 0.386 for stroke, Fisher exact test). Of note, there were large differences in age distribution between both groups and there was a significantly older age of non–diffuse-trickling GA patients (see above). 

Cardiovascular Risk Factors and Diseases in the Age Group Younger Than 65.

Since the detailed interview with diffuse-trickling patients revealed relatively high frequencies of hypertension and MI in the age group younger than 65 at first exam, comparison of cardiovascular risk factors and diseases was also performed between patients of this age group (diffuse-trickling GA n = 29 versus non–diffuse-trickling GA n = 27; Fig. 6): There was an almost equal frequency for smoking history (48% vs. 48%, P = 1.0, Fisher exact test), and an insignificant lower frequency of diabetes in the diffuse-trickling group (0% vs. 7%, P = 0.228). For hypercholesterolemia (35% vs. 30%, P = 0.779), hypertension (52% vs. 37%, P = 0.296), and stroke (10 vs. 0%, P = 0.237), respectively, the frequencies in the diffuse-trickling group were higher, although not significant. 

Figure 6

 

Frequencies of cardiovascular risk factors and diseases in diffuse-trickling GA and non–diffuse-trickling GA patients aged younger than 65 years at first exam. In the non–diffuse-trickling group, there are lower frequencies for hypercholesterolemia, hypertension, MI, and stroke, respectively. For MI, the frequency in the diffuse-trickling GA group is significantly higher (24.1%, seven patients [six male]) as compared with the non–diffuse-trickling GA group. (P = 0.011, Fisher exact test).

For MI, however, there was a significantly higher frequency in the diffuse-trickling group as compared with non–diffuse-trickling GA patients (24% vs. 0%, P = 0.011; Fig. 6). 

Six of the seven patients with a history of MI were male. 

Subfoveal Choroidal Thickness.

There was a significantly thinner SCT in eyes with diffuse-trickling GA (135.2 ± 56.4 [range, 37–285 μm], median 117 [IQR, 99–171] μm, n = 41 eyes of 41 patients with a mean age of 68.2 ± 10.9 years) as compared with non–diffuse-trickling GA (191.4 ± 77.4 μm [range, 55–387 μm], median 173 [IQR, 146.25–241.5] μm, n = 34 eyes of 34 patients with a mean age of 77.2 ± 7.0 years; P < 0.001; Fig. 7). 

Figure 7

 

Subfoveal choroidal thickness in eyes with diffuse-trickling GA (n = 41) compared with eyes with other GA subtypes (non–diffuse-trickling, n = 34). Eyes with diffuse-trickling GA have a significantly thinner subfoveal choroid (135.2 ± 56.4 [range, 37–285 μm], median 117 [IQR, 99–171] μm) as compared with eyes with non–diffuse-trickling GA (191.4 ± 77.4 μm [range, 55–387 μm], median 173 [IQR, 146.25–241.5] μm; P < 0.001, Mann-Whitney U test).

Discussion

The results indicate that patients with diffuse-trickling pattern associated with GA can be distinguished from other GA subtypes beyond ocular characteristics that have been previously described. ^20,21 A young mean age of manifest GA, a shift in the proportion of men in the study population with age, and a high frequency of MI in the young male patients discriminate against other GA subtypes from the FAM study. Furthermore, the hospitalization rate of 54.2% in the age group younger than 65 years due to severe cardiovascular diseases appears to be far above average rates in the general population. Together with the reduced SCT, the findings may point toward an association with systemic vascular abnormalities of this GA variant. 

The lobular appearance of the atrophic lesions in diffuse-trickling GA may also support the assumption of a vascular contributing element. Torczynski and Tso ³⁶ have described the choriocapillaris as a collection of multiple lobules with each lobule functioning as an independent unit in which the precapillary arterial acts as an end artery. The size of these lobules at the posterior pole varied from 420 to 1200 μm in diameter. ³⁶ Due to corresponding size of atrophic lesions in GA, Maguire and Vine ¹⁴ already speculated that focal atrophic areas may be caused by closure of the underlying choriocapillaris lobules. 

Interestingly, the lobular atrophic changes seen in patients with the diffuse-trickling pattern exhibit somewhat similar appearances with reported degenerative RPE lesions in the rhesus monkey model of hypertensive choroidopathy. ^37,38 These degenerative RPE changes were attributed to chronic choroidal ischemia. ^37,38  

There is also resemblance with funduscopic characteristics described in a recently reported entity called age-related choroidal atrophy. ²² An impaired blood supply due to an atrophic choroid in the elderly was assumed as source of the funduscopic changes. These patients exhibited a choroidal thickness less than 125 μm and had a mean age of 80.9 ± 7.3 years. ²² In diffuse-trickling GA, SCT was also severely reduced with a mean of 135 μm. Of note, almost one-third of patients with available data on refraction had a SE of less than or equal to −3 D. We assume that insufficiency at the level of the choroid may also be source for the funduscopic changes in diffuse-trickling. However, compared with patients with age-related choroidal atrophy, patients with diffuse-trickling GA were considerably younger (mean age at SCT measurement was 68.2 ± 10.9 years). Therefore, aging alone would unlikely be the reason for such changes in this phenotype that manifests relatively early in life. Indication of an increased rate of severe cardiovascular events, such as hypertensive crisis and MI in the younger diffuse-trickling patients rather points to an underlying general vascular disorder. The sex-shift with age, that means here the drop of the male study population from 55% in the age group younger than 65 to 19% in the age group older than or equal to 65, could also be in line with the theory of an underlying general disorder in diffuse-trickling that additionally would go along with reduced life expectancy in men. This is supported by the observation that more than one-third of male patients with diffuse-trickling GA of the age group younger than 65 reported a history of MI. It is not a new concept to suggest that a dysfunctional circulation may be a common denominator for cardiovascular diseases and AMD. There are indeed multiple studies that link both conditions. ^39–51 However, a meta-analysis by Chakravarthy and coworkers ⁵² showed that although the pooled estimates for case-control studies were statistically significant for both cardiovascular disease (odds ratio 2.20; 95% confidence interval [CI] 1.49–3.26) and hypertension (odds ratio 1.48, 95% CI 1.22–1.78), the magnitude of the odds ratios were inconsistent across studies. 

These conflicting results may be attributable to heterogeneous patient cohorts. Accurate phenotyping appears to be important to segregate AMD subtypes with a specific risk profile. Klein et al. ⁵³ found that sex (females) and smoking history are associated with the presence of reticular pseudodrusen. Furthermore, they reported that reticular pseudodrusen at baseline were associated with poorer survival. ⁵³ Interestingly, all patients in our study with the diffuse-trickling phenotype had reticular pseudodrusen. Klein et al., ⁵³ however, did not detect a relationship with cardiovascular disease in their population-based study. 

Several lines of evidence now indicate that reticular pseudodrusen are located above the RPE cell monolayer. ^54–56 This is in contrast with the first histopathological report of one eye with reticular pseudodrusen, in which the authors speculated that a loss of choroidal vessels was accompanied by fibrotic remodeling as the correlate for the reticular pattern. ⁵⁷  

However, choroidal changes may indeed play a role in the biogenesis of reticular pseudodrusen irrespective of their anatomical location in relation to the RPE cell monolayer. Sohrab and coworkers ⁵⁸ recently suggested that the arrangement and pattern of reticular pseudodrusen are related closely to the choroidal stroma and the choroidal vasculature. Moreover, Alten et al. ⁵⁹ reported topographic concordance of reticular pseudodrusen with choroidal watershed zones and, therefore, speculated that choroidal hypoxia may contribute to the pathogenesis of this drusen subtype. Querques and coworkers ⁶⁰ described an overall thinned choroid in patients with reticular pseudodrusen. Furthermore, Garg et al. ⁶¹ demonstrated a thinner choroid throughout the macular in eyes with early AMD and reticular pseudodrusen as compared with eyes with early AMD without reticular pseudodrusen. The diffuse-trickling phenotype characterized herein with severely thinned choroid along with the presence of reticular pseudodrusen may further underscore a potential link between choroidal perfusion and reticular pseudodrusen. The high proportion of cardiovascular diseases further points to an association with a generalized vascular disease. 

The significantly younger age of patients with diffuse-trickling as compared with other GA patients and the onset in several patients below the age of 55 years may question the classification into AMD. Interestingly, in 2009, Hamel and coworkers ⁶² described a GA phenotype in 18 patients with a mean age of onset of 47.5 years (range, 41–54 years) and termed it “extensive macular atrophy with pseudodrusen-like appearance” (EMAP). The FAF images show the same GA phenotype, which had been termed “diffuse-trickling.” ²¹ Based on funduscopic features, diffuse-trickling and EMAP also appear to represent the same phenotype. However, no EMAP patient presented with CNV, while approximately one-third of patients with diffuse-trickling GA had either CNV in the partner eye or develop secondary CNV in the study eye. Moreover, in EMAP, an association with cardiovascular diseases was not found, however, it is unclear if respective information was systematically collected. 

Furthermore, there are striking similarities of the diffuse-trickling phenotype with the retinal degeneration described in patients with late-onset retinal macular degeneration, a fully penetrant autosomal-dominant disease due to a mutation in the C1QTNF5 gene on chromosome 11. ^63–65 However, in the majority of patients with diffuse-trickling GA, the family history is not suggestive of autosomal-dominant inheritance. Furthermore, anterior segment abnormalities, such as long anterior inserted zonules and pupillary atrophy, which are described in C1QTNF5 retinopathy were not seen in the patients with diffuse-trickling GA reported herein, whereby the detection of these changes is challenging. 

There are several limitations of the current study. The data on cardiovascular risk factors and diseases was based on the patients' report. The patients with diffuse-trickling GA were not yet systematically examined by cardiologists. However, this is planned to be performed in the ongoing study. Furthermore, the detailed interview that gave important information (e.g., on hospitalization) was only conducted with patients with the diffuse-trickling phenotype. Since these additional information had not been assessed by the original FAM study CRF, a comparison group, especially for the age group younger than 65, is yet lacking. In the ongoing study, however, collection of this further information is now incorporated for all GA patients. 

Moreover, there might be other (e.g., genetic) predisposing factors in the diffuse-trickling phenotype. Genetic analyses including screening for the known AMD-risk alleles and for mutations in the C1QTNF5 gene in patients with diffuse-trickling are currently been performed in the ongoing study. 

Most importantly, there is yet no proof for the hypothesis on the pathogenesis of the diffuse-trickling GA phenotype. The angiographic findings in diffuse-trickling patients indeed indicate a delayed or missing filling of the choriocapillaris (data not shown, angiography has not been systematically performed in all patients). However, in presence of GA, this findings might also be secondary to the atrophy of the RPE. ⁶⁶ Furthermore, massive extracellular material accumulations underneath the RPE, which is seen in all patients with diffuse-trickling GA by SD-OCT imaging ²⁰ may obscure choroidal filling. Further studies are needed including long-term follow-up of young patients with known cardiovascular diseases and a lobular choroidal delayed perfusion to determine a causative association. 

In conclusion, the data indicate an association of the diffuse-trickling GA phenotype with systemic cardiovascular disorders. Together with the ocular morphologic characteristics including a lobular appearance and a thin choroid, a vascular insufficiency at the level of the choroid and subsequent dysfunction of the RPE and outer retina may play a pathogenetic role. Refined phenotyping with various retinal imaging modalities including FAF appears prudent in ongoing and future studies in patients with AMD to identify this retinal phenotype and to further explore discriminating features. 

Acknowledgments

Supported by grants from the Deutsche Forschungsgemeinschaft (German Research Council Ho1926/1-3, FL 658/4-1), and BONFOR 0-137-0012 (MF). 

Disclosure: M. Fleckenstein, Heidelberg Engineering (F, C, R); S. Schmitz-Valckenberg, Heidelberg Engineering (F, C, R); M. Lindner, Heidelberg Engineering (F, C, R); A. Bezatis, Heidelberg Engineering (F, C, R); E. Becker, Heidelberg Engineering (F, C, R); R. Fimmers, None; F.G. Holz, Heidelberg Engineering (F, C, R) 

Footnotes

 MF and SS-V contributed equally to the work presented here and should therefore be regarded as equivalent authors.

Appendix

Fundus Autofluorescence in Age-Related Macular Degeneration Study Group

Department of Ophthalmology, University of Bonn, Bonn, Germany: Frank G. Holz, Steffen Schmitz-Valckenberg, Monika Fleckenstein, Arno P. Göbel, Moritz Lindner, Julia Steinberg, Almut Bindewald Wittich, Joanna Czauderna 

Institute of Biostatistics, University of Bonn, Bonn, Germany: Rolf Fimmers 

Department of Ophthalmology, University of Aachen, Aachen, Germany: Peter Walter, Andreas Weinberger 

Department of Ophthalmology, Inselspital, University of Bern, Bern, Switzerland: Sebastian Wolf, Ute Schnurrbusch Wolf 

Department of Ophthalmology, University of Heidelberg, Heidelberg, Germany: Hans E. Völcker, Friederike Mackensen 

Department of Ophthalmology, University of Leipzig, Leipzig, Germany: Peter Wiedemann, Andreas Mössner 

Department of Ophthalmology, St. Franziskus Hospital Münster, Münster, Germany: Daniel Pauleikhoff, Georg Spital 

Institute of Human Genetics, University of Regensburg, Regensburg, Germany: Bernhard H. F. Weber 

Department of Ophthalmology, University of Würzburg, Würzburg, Germany: Claudia von Strachwitz