Background To evaluate the morphological features and prevalence of accompanying late age-related macular degeneration (AMD) according to the fundus distribution of reticular pseudodrusen (RPD).
Methods According to the involved area in the fundus, the distribution of RPD was classified as localised, intermediate, or diffuse type. Morphology of RPD was classified as discrete, branching, or confluent pattern. The presence of late AMD was evaluated.
Results 233 eyes of 121 patients with RPD were included. The distribution of RPD was localised, intermediate and diffuse type in 30.9%, 40.3% and 28.8% of eyes, respectively. The discrete, branching and confluent morphological patterns were found in 45.8%, 44.8% and 9.7% of the localised type, and in 0%, 13.8% and 86.2% of the intermediate type, respectively. In contrast, the diffuse type showed only the confluent morphological pattern. The prevalence of accompanying late AMD was 13.9%, 13.8% and 56.7% in the localised, intermediate and diffuse type, respectively, and it was significantly higher in the diffuse type (p<0.05).
Conclusion RPD with diffuse distribution showed a confluent morphological pattern and a high prevalence of late AMD. RPD can be classified by the fundus distribution for the assessment of visual prognosis.
- Age-related macular degeneration
- fundus distribution
- geographic atrophy
- morphological pattern
- reticular pseudodrusen
Statistics from Altmetric.com
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.
- Age-related macular degeneration
- fundus distribution
- geographic atrophy
- morphological pattern
- reticular pseudodrusen
Reticular pseudodrusen (RPD) were first described in 1990 as ‘pseudodrusen visible in blue light’,1 and several studies using either multimodal imaging techniques or histological examination revealed that RPD are subretinal drusenoid deposits, with characteristic imaging features.2–10 Although most classifications or grading systems of age-related macular degeneration (AMD) did not differentiate the soft drusen from the drusenoid deposits, such as RPD,11 ,12 it became apparent that RPD had clinical features and a prognosis different from soft drusen.7 Several reports revealed a high prevalence of advanced AMD in RPD.2 ,3 ,5 ,6 ,9 ,13–15 Therefore, the early detection of RPD may be important to prevent severe visual loss. A few reports have recently shown that RPD were found not only in the macular area, but also in the extramacular area, and even in the peripheral retina, with either progression or regression during their natural course.4 ,5 ,7 ,9 In addition, RPD appear to have a wide spectrum of morphology, from discrete to confluent lesions.7 ,9
However, little is known about the clinical significance of various types of distribution or morphology, and no standard diagnostic criteria or classification systems of RPD have yet been established. With more knowledge about their distribution or morphology, better diagnostic criteria and a better classification system can be established, based on those data. The purpose of this study was thus to make a classification system of RPD according to the extent of distribution, and to compare morphological features and the accompanying rate of late AMD, which is clinically associated with visual prognosis, among different distribution types of RPD.
This study was conducted according to the Declaration of Helsinki, after approved by the institutional review board of the Samsung Medical Center in Korea. We retrospectively analysed medical records of patients, previously diagnosed as having RPD at Samsung Medical Center, between October 2009 and July 2011. Patients who underwent complete ophthalmological examination, including fundus colour photography, near-infrared, red-free, fundus autofluorescence (FAF) imaging by confocal scanning laser ophthalmoscopy, and spectral domain optical coherence tomography (SD OCT), were included in this study. RPD was defined by multiple lesions with reticular pattern on colour photographs with blue channel examination, hyporeflectant lesions with reticular pattern in near-infrared imaging, light lesions with interlacing network in red-free imaging, multiple reticular hypofluorescent lesions in FAF, and subretinal deposits in SD OCT.1 ,5–7 Eyes with severe epiretinal membrane and laser scars, which disturbed the diagnosis and classification of RPD, were excluded.
Imaging acquisition and analysis
Fundus colour photographs were taken with a Topcon camera IX50 (Topcon, Paramus, New Jersey, USA), and were viewed in the Topcon Image Net program (version 2.56). The blue channel of the colour photograph was examined using high-resolution digital colour fundus photographs, which were viewed in Photoshop (CS3; Adobe System Inc., San Jose, California, USA), as indicated elsewhere.6
Near-infrared, red-free, FAF imaging, SD OCT, fluorescein angiography (FA) and indocyanine green angiography were performed using a Spectralis HRA+OCT (version 22.214.171.124; Heidelberg Engineering, Heidelberg, Germany). The SD OCT images were viewed with Heidelberg software (Spectralis viewing module 126.96.36.199). Images of each eye were assessed independently by two retinal specialists (MYL and DIH), and the diagnosis and classification were determined in agreement. In cases of disagreement, the pictures were reviewed by both observers for discussion and the final adjudication was made by consensus.
Distributional type of RPD
Distribution of RPD was evaluated by five fundus colour photographs, including the central macular field and four adjacent fields (ie, superior, inferior, temporal and medial to the central field). The photographs were taken with a 50° angle camera (IX50, Topcon) (figure 1). According to the extent of involvement of retinal areas, eyes were divided into three types: localised, intermediate and diffuse. The localised type was defined as RPD confined to the central field, or central field and less than half of superior and less than one-third of other adjacent photographic fields. The intermediate type was defined as RPD with features between the localised and the diffuse type. The diffuse type was defined as RPD observed in more than half of superior and more than one-third of three other adjacent fields (figure 2). The presence of RPD in the central field was not considered as an integral part of the diffuse type.
Morphological pattern of RPD
RPD were evaluated by fundus colour photographs, and were classified as discrete, branching, and confluent pattern. Eyes with only discrete, round to oval lesions in reticular pattern were assigned as discrete pattern (figure 3). Eyes with branching or tubular, interconnecting lesions were assigned as branching pattern, which might show discrete lesions, but not a confluent lesion (figure 3). Eyes with merged, confluent lesions, with surrounding branching lesions, were assigned as confluent pattern (figure 4), which might have both discrete and branching lesions.
Definition and grading of AMD
AMD was defined and classified by the international classification and grading system provided by the International ARM Epidemiological Study Group.12 Multimodal imaging tests, including SD OCT and FA, were used to confirm the diagnosis of late AMD.
χ2 tests were used to compare the prevalence of AMD between two different types, and analysis of variance (ANOVA) tests were used for analysing differences between three quantitative means (ie, the patients' age). Statistical analyses were performed with SPSS software V.16.01. For all tests, a p value less than 0.05 was considered to be statistically significant.
In total, 233 eyes of 121 patients diagnosed as having RPD were enrolled in the current study, and the mean age was 71.4±8.7 years (range 43–92 years). One hundred and seven (88.4%) patients were women and 14 (11.6%) patients were men. All patients included in the study were Koreans. One hundred and twelve (92.6%) patients had bilateral RPD and nine (7.4%) patients had unilateral RPD. In patients with unilateral RPD, five patients had a large disciform scar in the opposite eye. All patients underwent colour photography with blue channel examination, near-infrared, red-free, FAF imaging and SD OCT. FA and indocyanine green angiography were performed in 66.5% and 30.0% of eyes, respectively. All eyes showed characteristic imaging features of RPD, such as subretinal deposits in SD OCT.
Fourteen (6.0%) eyes had mild to moderate non-proliferative diabetic retinopathy, 22 (9.4%) eyes had mild epiretinal membrane, and three (1.3%) eyes had retinal vein occlusion, all of which showed bilateral symmetry of distribution of RPD.
Distribution of RPD
All eyes were classified into three types by the fundus distribution of RPD. Therefore, 72 eyes (30.9%) were localised type, 94 eyes (40.3%)were intermediate type, and 67 eyes (28.8%) were diffuse type (table 1). The distribution of RPD was symmetrical in all bilateral RPD, except for two cases (1.8%); they had diffuse type in one eye and intermediate type in the opposite eye. The mean age of localised type, intermediate type and diffuse type was 70.5±8.5 years (range 45–84 years), 71.3±8.5 years (range 43–86 years) and 72.5±9.1 years (range 50–92 years), respectively. There was no statistical difference in age between the three types (ANOVA test, p=0.396).
Morphology of RPD
All eyes were classified into three patterns by the morphological feature of RPD. Therefore, 33 eyes (14.2%) were discrete pattern, 45 eyes (19.3%) were branching pattern, and 155 eyes (66.5%) were confluent pattern. The mean age of discrete, branching and confluent pattern was 72.2±6.8 years (range 54–84 years), 69.7±8.8 years (range 45–84 years) and 71.8±9.0 years (range 43–92 years), respectively. There was no statistical difference in age between the three patterns (ANOVA test, p=0.321).
Localised distribution type showed discrete, branching and confluent morphological pattern in 45.8%, 44.8% and 9.7%, respectively. Intermediate distribution type showed branching morphological pattern in 13.8% and confluent morphological pattern in 86.2%. All eyes in diffuse distribution type, however, showed confluent morphological pattern. To put it differently, all eyes with discrete morphological pattern showed localised distribution. Branching morphological pattern showed localised distribution in 71.1% and intermediate distribution in 28.9%. Confluent morphological pattern showed localised, intermediate and diffuse distribution type in 4.5%, 52.3% and 43.2%, respectively. The proportion of confluent morphological pattern was significantly higher in the diffuse distribution type than in other distribution types (χ2 test, p<0.05) (table 1).
Prevalence of accompanying late AMD
The overall prevalence of late AMD in total RPD eyes was 26.2%. The prevalence of accompanying geographic atrophy was 4.2%, 6.4% and 38.8%, and accompanying neovascular AMD was detected in 9.7%, 7.4% and 17.9% in the localised, intermediate and diffuse type, respectively. Therefore, the prevalence of overall accompanying late AMD was 13.9%, 13.8% and 56.7% in the localised, intermediate and diffuse type, respectively. The prevalence of geographic atrophy and overall late AMD were significantly higher in the diffuse type than in other types (χ2 test, p<0.05). Although the prevalence of neovascular AMD was higher in the diffuse type than in other types, a statistical difference was found only between the diffuse type and the intermediate type (χ2 test, p<0.05) (table 1).
It has recently been reported that a method using multimodal imaging tests is more accurate and sensitive for the detection of RPD than a method using standard colour photography,7 which has been used as the single method in many previous studies.2–4 In this study, multiple imaging tests, including colour photography with blue channel examination, red-free, near-infrared, FAF and SD OCT, were used in all patients for accurate diagnosis of RPD. They were also used for the differential diagnosis of RPD, including early AMD with multiple soft drusen, cuticular drusen and many retinal diseases with multiple flecks.7 ,13 ,16 ,17
The ETDRS grid was not used in this study, because most RPD, even with the localised distribution type, showed distribution outside the ETDRS grid, which made it difficult to differentiate among the three types. Instead of the ETDRS grid, we used five photographic fields, which had been routinely used in our retinal clinic, and were relatively similar to those used in the Central Vein Occlusion Study.18
In the classification according to the fundus distribution in this study, the superior photographic field should be interpreted differently from the temporal, nasal and inferior fields, because the centre of RPD is usually located superior to the fovea. Therefore, RPD were considered to be present in the superior field only when they were distributed over half of the photographic field. In other fields, RPD were considered to be present only when they were distributed over one-third of the photographic field, considering some overlapped area between two adjacent photographic fields. In the definition of diffuse type, the presence of RPD in the central photographic field was not considered as an integral part, because RPD in the central field frequently regressed after the development of neovascular AMD. Red-free or near-infrared photographs were not used in the classification, because every patient did not take them in multiple fields routinely, although it was reported that they showed a higher detection rate of RPD.7
RPD showed various distribution and morphology, as presented in the current study. The distribution of RPD varied within a wide spectrum. One end of the spectrum was an eye with several discrete deposits confined to the superior macular area, and the other end was an eye with numerous deposits in all four quadrants of the fundus. Morphology of RPD also varied, from round to oval shape deposits in the reticular pattern to the confluent deposit with surrounding branching lesions.9 Probably, the duration of disease also varied. RPD may be a progressive disease from the initial, localised type with discrete pattern, to the final, diffuse type with confluent pattern, although it is not clear that all RPD eventually progress from the initial to the final type. Therefore, it appears that each eye with RPD may have a different severity and visual prognosis. However, no grading system of RPD has yet been developed. In the case of soft drusen, the size and the involved area or number of lesions was usually used for grading.19–22 Unfortunately, it is difficult to use the same method in RPD, because RPD have a larger number of lesions of various size by merging, sometimes extending to the periphery. Therefore, three distribution types were used for the severity grading in this study. According to the results using this method, the diffuse distribution type has the higher prevalence, and the localised distribution type has the lower prevalence of accompanying late AMD. In contrast to the classification by distribution, the classification by morphology showed a severely skewed distribution, and had a subjective component to determine the pattern by the existence of borderline features between two adjacent patterns. The classification by distribution thus appears to be more objective. However, discrete or branching morphological pattern may be a sign of a relatively lower risk of accompanying late AMD than confluent morphological pattern, because they were observed only in localised or intermediate distribution type.
Statistically significant differences were found between the diffuse type and the localised or the intermediate type in the prevalence of geographic atrophy and overall late AMD (p<0.001). The prevalence of late AMD increased almost fourfold from the localised type (13.9%) or the intermediate type (13.8%) to the diffuse type (56.7%). Interestingly, the prevalence of geographic atrophy showed almost a six to ninefold increase, much higher than that of neovascular AMD, which showed an approximately twofold increase. RPD appear to have a high risk of geographic atrophy, and probably RPD of the localised or intermediate type can be a good candidate for promising therapeutic drugs against geographic atrophy in the near future.23 Although RPD eyes with the diffuse type showed no statistical difference from the localised type in the prevalence of neovascular AMD, we think that further study is needed to investigate an increased risk of neovascular AMD depending on the distribution type, because there was a significant difference between the diffuse type and the intermediate type. Furthermore, we think that even RPD of the localised distribution type probably need frequent examination, because the accompanying rate of 9.7% of neovascular AMD 4.2% of geographic atrophy is higher than those of the normal healthy population.24–26
This study has several limitations. All patients are Korean, and the study was conducted retrospectively. Other imaging features, such as morphological pattern, were not used in the evaluation of accompanying late AMD.
In conclusion, RPD has distributional and morphological features in a wide spectrum. The prevalence of late AMD in RPD eyes varies according to the distribution type; thus the diffuse distributional type with a confluent morphology showed the highest prevalence of late AMD. The classification of RPD by fundus distribution could be useful, and should be considered in the assessment of the visual prognosis, and in the establishment of standard diagnostic criteria or classification of RPD.
Competing interests None.
Ethics approval Ethics approval was provided by the institutional review board of Samsung Medical Center, Seoul, Korea.
Provenance and peer review Not commissioned; externally peer reviewed.
Data sharing statement The authors declare that some of the study population of this study was previously included in their other article (Clinical characteristics of reticular pseudodrusen in Korean patients, Am J Ophthalmol 2012;153:530-5) because both articles are a retrospective study about reticular pseudodrusen in the same hospital. However, the study design, main outcome measures, inclusion criteria of the study subjects and the study period are different between the two studies.