Systemic Metabolomics in a Framework of Genetics and Lifestyle in Age-Related Macular Degeneration.

Insights into the pathogenesis of age-related macular degeneration (AMD), a leading cause of blindness, point towards a complex interplay of genetic and lifestyle factors triggering various systemic pathways. This study aimed to characterize metabolomic profiles for AMD and to evaluate their position in the trias with genetics and lifestyle. This study included 5923 individuals from five European studies. Blood metabolomics were assessed using a nuclear magnetic resonance platform of 146 metabolites. Associations were studied using regression analyses. A genetic risk score (GRS) was calculated using ß-values of 49 AMD variants, a lifestyle risk score (LRS) using smoking and diet data, and a metabolite risk score (MRS) using metabolite values. We identified 61 metabolites associated with early-intermediate AMD, of which 94% were lipid-related, with higher levels of HDL-subparticles and apolipoprotein-A1, and lower levels of VLDL-subparticles, triglycerides, and fatty acids (false discovery rate (FDR) p-value < 1.4 × 10-2). Late AMD was associated with lower levels of the amino acids histidine, leucine, valine, tyrosine, and phenylalanine, and higher levels of the ketone bodies acetoacetate and 3-hydroxybutyrate (FDR p-value < 1.5 × 10-3). A favorable lifestyle characterized by a healthy diet was associated with higher levels of amino acids and lower levels of ketone bodies, while an unfavorable lifestyle, including smoking, showed opposite effects (FDR p-value < 2.7 × 10-2). The MRS mediated 5% of the effect of the GRS and 20% of that of the LRS on late AMD. Our findings show that metabolomic profiles differ between AMD stages and show that blood metabolites mostly reflect lifestyle. The severity-specific profiles spur further interest into the systemic effects related to disease conversion.

Association of lipid-lowering drugs and antidiabetic drugs with age-related macular degeneration: a meta-analysis in Europeans.

BACKGROUND/AIMS: To investigate the association of commonly used systemic medications with prevalent age-related macular degeneration (AMD) in the general population. METHODS: We included 38 694 adults from 14 population-based and hospital-based studies from the European Eye Epidemiology consortium. We examined associations between the use of systemic medications and any prevalent AMD as well as any late AMD using multivariable logistic regression modelling per study and pooled results using random effects meta-analysis. RESULTS: Between studies, mean age ranged from 61.5±7.1 to 82.6±3.8 years and prevalence ranged from 12.1% to 64.5% and from 0.5% to 35.5% for any and late AMD, respectively. In the meta-analysis of fully adjusted multivariable models, lipid-lowering drugs (LLD) and antidiabetic drugs were associated with lower prevalent any AMD (OR 0.85, 95% CI=0.79 to 0.91 and OR 0.78, 95% CI=0.66 to 0.91). We found no association with late AMD or with any other medication. CONCLUSION: Our study indicates a potential beneficial effect of LLD and antidiabetic drug use on prevalence of AMD across multiple European cohorts. Our findings support the importance of metabolic processes in the multifactorial aetiology of AMD.

The Phenotypic Course of Age-Related Macular Degeneration for ARMS2/HTRA1: The EYE-RISK Consortium.

PURPOSE: Age-related maculopathy susceptibility 2 (ARMS2) is considered the most enigmatic of the genes for age-related macular degeneration (AMD). We investigated the phenotypic course and spectrum of AMD for the risk haplotype at the ARMS2 and high-temperature requirement A serine peptidase 1 (HTRA1) locus in a large European consortium. DESIGN: Pooled analysis of 4 case-control and 6 cohort studies. PARTICIPANTS: Individuals (N = 17 204) aged 55 years or older participating in the European Eye Epidemiology consortium. METHODS: Age-related macular degeneration features and macular thickness were determined on multimodal images; data on genetics and phenotype were harmonized. Risks of AMD features for rs3750486 genotypes at the ARMS2/HTRA1 locus were determined by logistic regression and were compared with a genetic risk score (GRS) of 19 variants at the complement pathway. Lifetime risks were estimated with Kaplan-Meier analyses in population-based cohorts. MAIN OUTCOME MEASURES: Age-related macular degeneration features and stage. RESULTS: Of 2068 individuals with late AMD, 64.7% carried the ARMS2/HTRA1 risk allele. For homozygous carriers, the odds ratio (OR) of geographic atrophy was 8.6 (95% confidence interval [CI], 6.5-11.4), of choroidal neovascularization (CNV) was 11.2 (95% CI, 9.4-13.3), and of mixed late AMD was 12.2 (95% CI, 7.3-20.6). Cumulative lifetime risk of late AMD ranged from 4.4% for carriers of the nonrisk genotype to 9.4% and 26.8% for heterozygous and homozygous carriers. The latter received the diagnosis of late AMD 9.6 years (95% CI, 8.0-11.2) earlier than carriers of the nonrisk genotype. The risk haplotype was not associated with hard or soft drusen < 125 µm (OR, 1.2; 95% CI, 0.9-1.7), but risks increased significantly for soft drusen ≥ 125 µm (OR, 2.1; 95% CI, 1.5-3.0), up to an OR of 7.2 (95% CI, 3.8-13.8) for reticular pseudodrusen. Compared with persons with a high GRS for complement, homozygous carriers of ARMS2/HTRA1 showed a higher risk of CNV (OR, 4.1; 95% CI, 3.2-5.4); risks of other characteristics were not different. CONCLUSIONS: Carriers of the risk haplotype at ARMS2/HTRA1 have a particularly high risk of late AMD at a relatively early age. Data suggest that risk variants at ARMS2/HTRA1 act as a strong catalyst of progression once early signs are present. The phenotypic spectrum resembles that of complement genes, only with higher risks of CNV.

Physical Activity, Incidence, and Progression of Age-Related Macular Degeneration: A Multicohort Study.

PURPOSE: To investigate the impact of physical activity (PA) on the incidence or progression of age-related macular degeneration (AMD) in the general population. DESIGN: Meta-analysis of longitudinal cohort studies. METHODS: We included 14,630 adults with no or early AMD at baseline from 7 population-based studies and examined associations of PA with AMD incidence and progression using multistate models (MSM) per study and subsequent random effects meta-analysis. Age effects were assessed using meta-regression. The main outcome measure was the hazard ratio (HR) for incident early or progression to late AMD. RESULTS: At baseline, mean age was 60.7 ± 6.9 to 76.4 ± 4.3 years, and prevalence of early AMD was 7.7% (range, 3.6%-16.9%) between cohorts. During follow-up, 1461 and 189 events occurred for early and late AMD, respectively. In meta-analyses, no or low to moderate PA (high PA as reference) was associated with an increased risk for incident early AMD (HR, 1.19; 95% CI, 1.01-1.40; P = .04), but not for late AMD. In subsequent meta-regression, we found no association of age with the effect of PA on incident AMD. CONCLUSIONS: Our study suggests high levels of PA to be protective for the development of early AMD across several population-based cohort studies. Our results establish PA as a modifiable risk factor for AMD and inform further AMD prevention strategies to reduce its public health impact.

Evaluating the Occurrence of Rare Variants in the Complement Factor H Gene in Patients With Early-Onset Drusen Maculopathy.

IMPORTANCE: Early-onset drusen maculopathy (EODM) is a severe disease and can lead to advanced macular degeneration early in life; however, genetic and phenotypic characteristics of individuals with EODM are not well studied. OBJECTIVE: To identify genotypic and phenotypic characteristics of individuals with EODM. DESIGN, SETTING, AND PARTICIPANTS: This case-control study collected data from the European Genetic Database from September 2004 to October 2019. A total of 89 patients with EODM diagnosed at 55 years or younger and 91 patients with age-related macular degeneration (AMD) diagnosed at 65 years or older were included. EXPOSURES: Coding regions of CFH, CFI, C3, C9, CFB, ABCA4, PRPH2, TIMP3, and CTNNA1 genes were sequenced, genetic risk scores (GRS) were calculated based on 52 AMD-associated variants, and phenotypic characteristics on color fundus photographs were analyzed comparing patients with EODM and AMD. MAIN OUTCOMES AND MEASURES: GRS, frequency of rare genetic complement variants, and phenotypic characteristics. RESULTS: This case-control study included 89 patients with EODM (mean [SD] age, 51.8 [8.7] years; 58 [65.2%] were female) and 91 patients with AMD (mean [SD] age, 77.6 [6.1] years; 45 [49.5%] female). At a mean (SD) age of 56.4 (7.3) years, 40 of 89 patients with EODM (44.9%) were affected by geographic atrophy or choroidal neovascularization. A lower GRS was observed in patients with EODM compared with patients with AMD (1.03 vs 1.60; P = .002), and 27 of 89 patients with EODM (30.3%) carried rare variants in the CFH gene compared with 7 of 91 patients with AMD (7.7%). Carriership of a rare CFH variant was associated with EODM (odds ratio, 7.2; 95% CI, 2.7-19.6; P < .001). A large macular drusen area (more than 50% covered with drusen) was observed in patients with EODM (24 of 162 eyes [14.8%]) compared with patients with AMD (9 of 164 eyes [5.5%]) (odds ratio, 4.57; 95% CI, 1.5-14.1; P = .008). CONCLUSIONS AND RELEVANCE: A large proportion of patients with EODM in this study carried rare CFH variants, with most of the identified CFH variants clustered in the first 7 complement control protein domains affecting factor H and factor H-like 1. Because EODM frequently leads to advanced macular degeneration at an early age and can result in many years of vision loss, this study supports targeting the complement system and sequencing the CFH gene in patients with EODM to improve genetic counseling and future treatments for AMD.

Smartphone Use Associated with Refractive Error in Teenagers: The Myopia App Study.

PURPOSE: To investigate the association between smartphone use and refractive error in teenagers using the Myopia app. DESIGN: Cross-sectional population-based study. PARTICIPANTS: A total of 525 teenagers 12 to 16 years of age from 6 secondary schools and from the birth cohort study Generation R participated. METHODS: A smartphone application (Myopia app; Innovattic) was designed to measure smartphone use and face-to-screen distance objectively and to pose questions about outdoor exposure. Participants underwent cycloplegic refractive error and ocular biometry measurements. Mean daily smartphone use was calculated in hours per day and continuous use as the number of episodes of 20 minutes on screen without breaks. Linear mixed models were conducted with smartphone use, continuous use, and face-to-screen distance as determinants and spherical equivalent of refraction (SER) and axial length-to-corneal radius (AL:CR) ratio as outcome measures stratified by median outdoor exposure. MAIN OUTCOME MEASURES: Spherical equivalent of refraction in diopters and AL:CR ratio. RESULTS: The teenagers on average were 13.7 ± 0.85 years of age, and myopia prevalence was 18.9%. During school days, total smartphone use on average was 3.71 ± 1.70 hours/day and was associated only borderline significantly with AL:CR ratio (ß = 0.008; 95% confidence interval [CI], -0.001 to 0.017) and not with SER. Continuous use on average was 6.42 ± 4.36 episodes of 20-minute use without breaks per day and was associated significantly with SER and AL:CR ratio (ß = -0.07 [95% CI, -0.13 to -0.01] and ß = 0.004 [95% CI, 0.001-0.008], respectively). When stratifying for outdoor exposure, continuous use remained significant only for teenagers with low exposure (ß = -0.10 [95% CI, -0.20 to -0.01] and ß = 0.007 [95% CI, 0.001-0.013] for SER and AL:CR ratio, respectively). Smartphone use during weekends was not associated significantly with SER and AL:CR ratio, nor was face-to-screen distance. CONCLUSIONS: Dutch teenagers spent almost 4 hours per day on their smartphones. Episodes of 20 minutes of continuous use were associated with more myopic refractive errors, particularly in those with low outdoor exposure. This study suggested that frequent breaks should become a recommendation for smartphone use in teenagers. Future large longitudinal studies will allow more detailed information on safe screen use in youth.

Enlargement of Geographic Atrophy From First Diagnosis to End of Life.

IMPORTANCE: Treatments for geographic atrophy (GA), a late stage of age-related macular degeneration (AMD), are currently under development. Understanding the natural course is needed for optimal trial design. Although enlargement rates of GA and visual acuity (VA) in the short term are known from clinical studies, knowledge of enlargement in the long term, life expectancy, and visual course is lacking. OBJECTIVE: To determine long-term enlargement of GA. DESIGN, SETTING, AND PARTICIPANTS: In this study, participant data were collected from 4 population-based cohort studies, with up to 25 years of follow-up and eye examinations at 5-year intervals: the Rotterdam Study cohorts 1, 2, and 3 and the Blue Mountains Eye Study. Data were collected from 1990 to 2015, and data were analyzed from January 2019 to November 2020. MAIN OUTCOMES AND MEASURES: Area of GA was measured pixel by pixel using all available imaging. Area enlargement and enlargement of the square root-transformed area, time until GA reached the central fovea, and time until death were assessed, and best-corrected VA, smoking status, macular lesions according to the Three Continent AMD Consortium classification, a modified version of the Wisconsin age-related maculopathy grading system, and AMD genetic variants were covariates in Spearman, Pearson, or Mann-Whitney analyses. RESULTS: Of 171 included patients, 106 (62.0%) were female, and the mean (SD) age at inclusion was 82.6 (7.1) years. A total of 147 of 242 eyes with GA (60.7%) were newly diagnosed in our study. The mean area of GA at first presentation was 3.74 mm2 (95% CI, 3.11-4.67). Enlargement rate varied widely between persons (0.02 to 4.05 mm2 per year), with a mean of 1.09 mm2 per year (95% CI, 0.89-1.30). Stage of AMD in the other eye was correlated with GA enlargement (Spearman ρ = 0.34; P = .01). Foveal involvement was already present in incident GA in 55 of 147 eyes (37.4%); 23 of 42 eyes (55%) developed this after a mean (range) period of 5.6 (3-12) years, and foveal involvement did not develop before death in 11 of 42 eyes (26%). After first diagnosis, 121 of 171 patients with GA (70.8%) died after a mean (SD) period of 6.4 (5.4) years. Visual function was visually impaired (less than 20/63) in 47 of 107 patients (43.9%) at last visit before death. CONCLUSIONS AND RELEVANCE: In this study, enlargement of GA appeared to be highly variable in the general population. More than one-third of incident GA was foveal at first presentation; those with extrafoveal GA developed foveal GA after a mean of 5.6 years. Future intervention trials should focus on recruiting those patients who have a high chance of severe visual decline within their life expectancy.

Development of a Genotype Assay for Age-Related Macular Degeneration: The EYE-RISK Consortium.

PURPOSE: To develop a genotype assay to assess associations with common and rare age-related macular degeneration (AMD) risk variants, to calculate an overall genetic risk score (GRS), and to identify potential misdiagnoses with inherited macular dystrophies that mimic AMD. DESIGN: Case-control study. PARTICIPANTS: Individuals (n = 4740) from 5 European cohorts. METHODS: We designed single-molecule molecular inversion probes for target selection and used next generation sequencing to sequence 87 single nucleotide polymorphisms (SNPs), coding and splice-site regions of 10 AMD-(related) genes (ARMS2, C3, C9, CD46, CFB, CFH, CFI, HTRA1, TIMP3, and SLC16A8), and 3 genes that cause inherited macular dystrophies (ABCA4, CTNNA1, and PRPH2). Genetic risk scores for common AMD risk variants were calculated based on effect size and genotype of 52 AMD-associated variants. Frequency of rare variants was compared between late AMD patients and control individuals with logistic regression analysis. MAIN OUTCOME MEASURES: Genetic risk score, association of genetic variants with AMD, and genotype-phenotype correlations. RESULTS: We observed high concordance rates between our platform and other genotyping platforms for the 69 successfully genotyped SNPs (>96%) and for the rare variants (>99%). We observed a higher GRS for patients with late AMD compared with patients with early/intermediate AMD (P < 0.001) and individuals without AMD (P < 0.001). A higher proportion of pathogenic variants in the CFH (odds ratio [OR] = 2.88; P = 0.006), CFI (OR = 4.45; P = 0.005), and C3 (OR = 6.56; P = 0.0003) genes was observed in late AMD patients compared with control individuals. In 9 patients, we identified pathogenic variants in the PRPH2, ABCA4, and CTNNA1 genes, which allowed reclassification of these patients as having inherited macular dystrophy. CONCLUSIONS: This study reports a genotype assay for common and rare AMD genetic variants, which can identify individuals at intermediate to high genetic risk of late AMD and enables differential diagnosis of AMD-mimicking dystrophies. Our study supports sequencing of CFH, CFI, and C3 genes because they harbor rare high-risk variants. Carriers of these variants could be amendable for new treatments for AMD that currently are under development.

Predicting Progression to Advanced Age-Related Macular Degeneration from Clinical, Genetic, and Lifestyle Factors Using Machine Learning.

PURPOSE: Current prediction models for advanced age-related macular degeneration (AMD) are based on a restrictive set of risk factors. The objective of this study was to develop a comprehensive prediction model applying a machine learning algorithm allowing selection of the most predictive risk factors automatically. DESIGN: Two population-based cohort studies. PARTICIPANTS: The Rotterdam Study I (RS-I; training set) included 3838 participants 55 years of age or older, with a median follow-up period of 10.8 years, and 108 incident cases of advanced AMD. The Antioxydants, Lipids Essentiels, Nutrition et Maladies Oculaires (ALIENOR) study (test set) included 362 participants 73 years of age or older, with a median follow-up period of 6.5 years, and 33 incident cases of advanced AMD. METHODS: The prediction model used the bootstrap least absolute shrinkage and selection operator (LASSO) method for survival analysis to select the best predictors of incident advanced AMD in the training set. Predictive performance of the model was assessed using the area under the receiver operating characteristic curve (AUC). MAIN OUTCOME MEASURES: Incident advanced AMD (atrophic, neovascular, or both), based on standardized interpretation of retinal photographs. RESULTS: The prediction model retained (1) age, (2) a combination of phenotypic predictors (based on the presence of intermediate drusen, hyperpigmentation in one or both eyes, and Age-Related Eye Disease Study simplified score), (3) a summary genetic risk score based on 49 single nucleotide polymorphisms, (4) smoking, (5) diet quality, (6) education, and (7) pulse pressure. The cross-validated AUC estimation in RS-I was 0.92 (95% confidence interval [CI], 0.88-0.97) at 5 years, 0.92 (95% CI, 0.90-0.95) at 10 years, and 0.91 (95% CI, 0.88-0.94) at 15 years. In ALIENOR, the AUC reached 0.92 at 5 years (95% CI, 0.87-0.98). In terms of calibration, the model tended to underestimate the cumulative incidence of advanced AMD for the high-risk groups, especially in ALIENOR. CONCLUSIONS: This prediction model reached high discrimination abilities, paving the way toward making precision medicine for AMD patients a reality in the near future.

Genetic Risk, Lifestyle, and Age-Related Macular Degeneration in Europe: The EYE-RISK Consortium.

PURPOSE: Age-related macular degeneration (AMD) is a common multifactorial disease in the elderly with a prominent genetic basis. Many risk variants have been identified, but the interpretation remains challenging. We investigated the genetic distribution of AMD-associated risk variants in a large European consortium, calculated attributable and pathway-specific genetic risks, and assessed the influence of lifestyle on genetic outcomes. DESIGN: Pooled analysis of cross-sectional data from the European Eye Epidemiology Consortium. PARTICIPANTS: Seventeen thousand one hundred seventy-four individuals 45 years of age or older participating in 6 population-based cohort studies, 2 clinic-based studies, and 1 case-control study. METHODS: Age-related macular degeneration was diagnosed and graded based on fundus photographs. Data on genetics, lifestyle, and diet were harmonized. Minor allele frequencies and population-attributable fraction (PAF) were calculated. A total genetic risk score (GRS) and pathway-specific risk scores (complement, lipid, extra-cellular matrix, other) were constructed based on the dosage of SNPs and conditional ß values; a lifestyle score was constructed based on smoking and diet. MAIN OUTCOME MEASURES: Intermediate and late AMD. RESULTS: The risk variants with the largest difference between late AMD patients and control participants and the highest PAFs were located in ARMS2 (rs3750846) and CHF (rs570618 and rs10922109). Combining all genetic variants, the total genetic risk score ranged from -3.50 to 4.63 and increased with AMD severity. Of the late AMD patients, 1581 of 1777 (89%) showed a positive total GRS. The complement pathway and ARMS2 were by far the most prominent genetic pathways contributing to late AMD (positive GRS, 90% of patients with late disease), but risk in 3 pathways was most frequent (35% of patients with late disease). Lifestyle was a strong determinant of the outcome in each genetic risk category; unfavorable lifestyle increased the risk of late AMD at least 2-fold. CONCLUSIONS: Genetic risk variants contribute to late AMD in most patients. However, lifestyle factors have a strong influence on the outcome of genetic risk and should be a strong focus in patient management. Genetic risks in ARMS2 and the complement pathway are present in most late AMD patients but are mostly combined with risks in other pathways.

Integrating Metabolomics, Genomics, and Disease Pathways in Age-Related Macular Degeneration: The EYE-RISK Consortium.

PURPOSE: The current study aimed to identify metabolites associated with age-related macular degeneration (AMD) by performing the largest metabolome association analysis in AMD to date, as well as aiming to determine the effect of AMD-associated genetic variants on metabolite levels and investigate associations between the identified metabolites and activity of the complement system, one of the main AMD-associated disease pathways. DESIGN: Case-control association analysis of metabolomics data. PARTICIPANTS: Five European cohorts consisting of 2267 AMD patients and 4266 control participants. METHODS: Metabolomics was performed using a high-throughput proton nuclear magnetic resonance metabolomics platform, which allows quantification of 146 metabolite measurements and 79 derivative values. Metabolome-AMD associations were studied using univariate logistic regression analyses. The effect of 52 AMD-associated genetic variants on the identified metabolites was investigated using linear regression. In addition, associations between the identified metabolites and activity of the complement pathway (defined by the C3d-to-C3 ratio) were investigated using linear regression. MAIN OUTCOME MEASURES: Metabolites associated with AMD. RESULTS: We identified 60 metabolites that were associated significantly with AMD, including increased levels of large and extra-large high-density lipoprotein (HDL) subclasses and decreased levels of very low-density lipoprotein (VLDL), amino acids, and citrate. Of 52 AMD-associated genetic variants, 7 variants were associated significantly with 34 of the identified metabolites. The strongest associations were identified for genetic variants located in or near genes involved in lipid metabolism (ABCA1, CETP, APOE, and LIPC) with metabolites belonging to the large and extra-large HDL subclasses. Also, 57 of 60 metabolites were associated significantly with complement activation levels, independent of AMD status. Increased large and extra-large HDL levels and decreased VLDL and amino acid levels were associated with increased complement activation. CONCLUSIONS: Lipoprotein levels were associated with AMD-associated genetic variants, whereas decreased essential amino acids may point to nutritional deficiencies in AMD. We observed strong associations between the vast majority of the AMD-associated metabolites and systemic complement activation levels, independent of AMD status. This may indicate biological interactions between the main AMD disease pathways and suggests that multiple pathways may need to be targeted simultaneously for successful treatment of AMD.

Genotype- and Phenotype-Based Subgroups in Geographic Atrophy Secondary to Age-Related Macular Degeneration: The EYE-RISK Consortium.

PURPOSE: Geographic atrophy (GA) secondary to age-related macular degeneration is considered a single entity. This study aimed to determine whether GA subgroups exist that can be defined by their genotype and phenotype. DESIGN: Retrospective analysis of cross-sectional data. PARTICIPANTS: Individuals (196 eyes of 196 patients) 50 years of age or older with GA from the EYE-RISK database. METHODS: Participants were graded for the presence of each of the following fundus features on color fundus photography: large soft drusen, reticular pseudodrusen (RPD), refractile drusen, hyperpigmentation, location of atrophy (foveal vs. extrafoveal), and multifocal lesions. Genotypes of 33 single nucleotide polymorphisms previously assigned to the complement, lipid metabolism, or extracellular matrix (ECM) pathways and ARMS2 also were included, and genetic risk scores (GRSs) for each of those 3 pathways were calculated. Hierarchical cluster analysis was used to determine subgroups of participants defined by these features. The discriminative ability of genotype, phenotype, or both for each subgroup was determined with 10-fold cross-validated areas under the receiver operating characteristic curve (cvAUCs), and the agreement between predicted and actual subgroup membership was assessed with calibration plots. MAIN OUTCOME MEASURES: Identification and characterization of GA subgroups based on their phenotype and genotype. RESULTS: Cluster analyses identified 3 subgroups of GA. Subgroup 1 was characterized by high complement GRS, frequently associated with large soft drusen and foveal atrophy; subgroup 2 generally showed low GRS, foveal atrophy, and few drusen (any type); and subgroup 3 showed a high ARMS2 and ECM GRS, RPD, and extrafoveal atrophy. A high discriminative ability existed between subgroups for the genotype (cvAUC, ≥0.94), and a modest discriminative ability existed for the phenotype (cvAUC, <0.65), with good calibration. CONCLUSIONS: We identified 3 GA subgroups that differed mostly by their genotype. Atrophy location and drusen type were the most relevant phenotypic features.

A Deep Learning Model for Segmentation of Geographic Atrophy to Study Its Long-Term Natural History.

PURPOSE: To develop and validate a deep learning model for the automatic segmentation of geographic atrophy (GA) using color fundus images (CFIs) and its application to study the growth rate of GA. DESIGN: Prospective, multicenter, natural history study with up to 15 years of follow-up. PARTICIPANTS: Four hundred nine CFIs of 238 eyes with GA from the Rotterdam Study (RS) and Blue Mountain Eye Study (BMES) for model development, and 3589 CFIs of 376 eyes from the Age-Related Eye Disease Study (AREDS) for analysis of GA growth rate. METHODS: A deep learning model based on an ensemble of encoder-decoder architectures was implemented and optimized for the segmentation of GA in CFIs. Four experienced graders delineated, in consensus, GA in CFIs from the RS and BMES. These manual delineations were used to evaluate the segmentation model using 5-fold cross-validation. The model was applied further to CFIs from the AREDS to study the growth rate of GA. Linear regression analysis was used to study associations between structural biomarkers at baseline and the GA growth rate. A general estimate of the progression of GA area over time was made by combining growth rates of all eyes with GA from the AREDS set. MAIN OUTCOME MEASURES: Automatically segmented GA and GA growth rate. RESULTS: The model obtained an average Dice coefficient of 0.72±0.26 on the BMES and RS set while comparing the automatically segmented GA area with the graders' manual delineations. An intraclass correlation coefficient of 0.83 was reached between the automatically estimated GA area and the graders' consensus measures. Nine automatically calculated structural biomarkers (area, filled area, convex area, convex solidity, eccentricity, roundness, foveal involvement, perimeter, and circularity) were significantly associated with growth rate. Combining all growth rates indicated that GA area grows quadratically up to an area of approximately 12 mm2, after which growth rate stabilizes or decreases. CONCLUSIONS: The deep learning model allowed for fully automatic and robust segmentation of GA on CFIs. These segmentations can be used to extract structural characteristics of GA that predict its growth rate.

Mediterranean Diet and Incidence of Advanced Age-Related Macular Degeneration: The EYE-RISK Consortium.

PURPOSE: To investigate associations of adherence to the Mediterranean diet (MeDi) with incidence of advanced age-related macular degeneration (AMD; the symptomatic form of AMD) in 2 European population-based prospective cohorts. DESIGN: Prospective cohort study of the Rotterdam Study I (RS-I) and the Antioxydants, Lipides Essentiels, Nutrition et Maladies Oculaires (Alienor) Study populations. PARTICIPANTS: Four thousand four hundred forty-six participants 55 years of age or older from the RS-I (The Netherlands) and 550 French adults 73 years of age or older from the Alienor Study with complete ophthalmologic and dietary data were included in the present study. METHODS: Examinations were performed approximately every 5 years over a 21-year period (1990-2011) in RS-I and every 2 years over a 4-year period (2006-2012) in the Alienor Study. Adherence to the MeDi was evaluated using a 9-component score based on intake of vegetables, fruits, legumes, cereals, fish, meat, dairy products, alcohol, and the monounsaturated-to-saturated fatty acids ratio. Associations of incidence of AMD with MeDi were estimated using multivariate Cox proportional hazard models. MAIN OUTCOMES MEASURES: Incidence of advanced AMD based on retinal fundus photographs. RESULTS: Among the 4996 included participants, 155 demonstrated advanced incident AMD (117 from the RS-I and 38 from the Alienor Study). The mean follow-up time was 9.9 years (range, 0.6-21.7 years) in the RS-I and 4.1 years (range, 2.5-5.0 years) in the Alienor Study. Pooling data for both the RS-I and Alienor Study, participants with a high (range, 6-9) MeDi score showed a significantly reduced risk for incident advanced AMD compared with participants with a low (range, 0-3) MeDi score in the fully adjusted Cox model (hazard ratio, 0.59; 95% confidence interval, 0.37-0.95; P = 0.04 for trend). CONCLUSIONS: Pooling data from the RS-I and Alienor Study, higher adherence to the MeDi was associated with a 41% reduced risk of incident advanced AMD. These findings support the role of a diet rich in healthful nutrient-rich foods such as fruits, vegetables, legumes, and fish in the prevention of AMD.

Increased High-Density Lipoprotein Levels Associated with Age-Related Macular Degeneration: Evidence from the EYE-RISK and European Eye Epidemiology Consortia.

PURPOSE: Genetic and epidemiologic studies have shown that lipid genes and high-density lipoproteins (HDLs) are implicated in age-related macular degeneration (AMD). We studied circulating lipid levels in relationship to AMD in a large European dataset. DESIGN: Pooled analysis of cross-sectional data. PARTICIPANTS: Individuals (N = 30 953) aged 50 years or older participating in the European Eye Epidemiology (E3) consortium and 1530 individuals from the Rotterdam Study with lipid subfraction data. METHODS: AMD features were graded on fundus photographs using the Rotterdam classification. Routine blood lipid measurements, genetics, medication, and potential confounders were extracted from the E3 database. In a subgroup of the Rotterdam Study, lipid subfractions were identified by the Nightingale biomarker platform. Random-intercepts mixed-effects models incorporating confounders and study site as a random effect were used to estimate associations. MAIN OUTCOME MEASURES: AMD features and stage; lipid measurements. RESULTS: HDL was associated with an increased risk of AMD (odds ratio [OR], 1.21 per 1-mmol/l increase; 95% confidence interval [CI], 1.14-1.29), whereas triglycerides were associated with a decreased risk (OR, 0.94 per 1-mmol/l increase; 95% CI, 0.91-0.97). Both were associated with drusen size. Higher HDL raised the odds of larger drusen, whereas higher triglycerides decreases the odds. LDL cholesterol reached statistical significance only in the association with early AMD (P = 0.045). Regarding lipid subfractions, the concentration of extra-large HDL particles showed the most prominent association with AMD (OR, 1.24; 95% CI, 1.10-1.40). The cholesteryl ester transfer protein risk variant (rs17231506) for AMD was in line with increased HDL levels (P = 7.7 × 10-7), but lipase C risk variants (rs2043085, rs2070895) were associated in an opposite way (P = 1.0 × 10-6 and P = 1.6 × 10-4). CONCLUSIONS: Our study suggested that HDL cholesterol is associated with increased risk of AMD and that triglycerides are negatively associated. Both show the strongest association with early AMD and drusen. Extra-large HDL subfractions seem to be drivers in the relationship with AMD, and variants in lipid genes play a more ambiguous role in this association. Whether systemic lipids directly influence AMD or represent lipid metabolism in the retina remains to be answered.

Mechanical Unfolding of an Autotransporter Passenger Protein Reveals the Secretion Starting Point and Processive Transport Intermediates.

The backbone of secreted autotransporter passenger proteins generally attains a stable ß-helical structure. The secretion of passengers across the outer membrane was proposed to be driven by sequential folding of this structure at the cell surface. This mechanism would require a relatively stable intermediate as starting point. Here, we investigated the mechanics of secreted truncated versions of the autotransporter hemoglobin protease (Hbp) of Escherichia coli using atomic force microscopy. The data obtained reveal a ß-helical structure at the C terminus that is very stable. In addition, several other distinct metastable intermediates are found which are connected during unfolding by multiroute pathways. Computational analysis indicates that these intermediates correlate to the ß-helical rungs in the Hbp structure which are clamped by stacked aromatic residues. Our results suggest a secretion mechanism that is initiated by a stable C-terminal structure and driven forward by several folding intermediates that build up the ß-helical backbone.