The global aim of this project is to expand the knowledge on the multifactorial pathogenesis of AMD. In addition to age, the multifactorial pathogenesis of AMD includes environmental and genetic risk factors. However, how these interact to promote progression remains largely unknown. AMD is a progressive retinal disease characterized by mostly asymptomatic early phases and progression to potentially blinding late forms (choroidal neovascularization or geographic atrophy). Individuals vary in their rate of progression, with some remaining stable for years. The reasons behind this variability, as well as the triggers and mechanisms of AMD progression, are not well understood. Currently, the standard of care for assessment of the risk of progression is solely based on fundus appearance, and is limited in its prediction ability. Our previous work showed that metabolomics enables the identification of specific plasma metabolomic profiles in AMD, which vary with the severity stages. The investigators hypothesize that the plasma and urinary metabolomic profile of subjects who progress over a five-year period (progressors) is distinct from those who remain at the same AMD severity stage (non-progressors). In this proposal, the investigators will follow our existing cohort over five years, comparing the metabolomic profiles of progressors to non-progressors.

Age-related macular degeneration (AMD) is the leading cause of adult blindness in developed countries and the third cause of global blindness. It is an eye-ageing disease with tremendous impact in daily quality of life. Clinically, AMD is a progressive disease, which evolves from early to late stages - geographic atrophy (GA) or choroidal neovascularization (CNV). GA has currently no treatment and CNV requires several costly intravitreal injections. Patients with late stages of AMD represent a huge economic burden to society, with enormous medical and non-medical costs.

Predicting progression is therefore of utmost importance. It is established that AMD progression has a multifactorial nature, combining phenotypic and several environmental and genetic risk factors. Attempts have been made to include the different risk-factors in theoretical predicting models, however they are not practical on the clinical setting and fail to explain biological interactions. Thus, in the current clinical practice, risk prediction is still being assessed on a phenotypic basis, not considering AMD multifactorial nature.

A new approach that may allow the integration of all AMD interacting factors is metabolomics, the simultaneous multiparametric measurement of metabolic changes in living organisms as a response to perturbation (disease, diet, environment, others). The value of metabolomics in medical research has become clear through several studies on cancer, cardiovascular disease, and Alzheimer disease (AD), the latter sharing common pathways with AMD, as well as by increasing investments in this area. The metabolome reflects the occurring biochemical processes, therefore forming a fingerprint of the organism's health status, at a given time, which can be measured through nuclear magnetic resonance (NMR) spectroscopy and/or mass spectrometry (MS). The research team's previous work showed that metabolomics can be a powerful tool to study AMD as well, enabling the identification of specific plasma metabolomic profiles in AMD, which vary with the severity stages, and are primarily lipid metabolites linked to glycerophospholipid and sphingolipid pathways, as well as purines The current study will further elucidate the role of metabolomics in the understanding of AMD, and will also identify potential biologically robust biomarkers that can address the problem of predicting progression. The investigators hypothesize that the plasma and urinary metabolomic profile of subjects who progress over a five-year period (progressors) is distinct from those who remain at the same AMD severity stage (non-progressors). Based on our preliminary data, The investigators hypothesize that a panel of metabolites will distinguish these two groups (progressors vs non-progressors), and that this will mostly consist of lipids and amino acids.

To achieve these purposes, subjects who participated in our group's previous prospective study on AMD metabolomics will be recruited. Eligible subjects will be all that participated in the the IN654 study (between January 2015 and July 2016). These will be recalled and progression will be phenotypically assessed, thus defining the AMD-progressors and non-progressors groups.

Metabolomic signatures of AMD progression and disease will lay the path for the future definition of progression metabolic biomarkers, which can represent a rapid, reliable and potentially affordable methodology for progression prediction. Overall, this approach should offer an opportunity to identify new valuable supplemental tools for routine clinical evaluation. This will allow medical interventions based on preventive strategies to reduce progression to blindness stages, which will, ultimately, also reduce societal costs of AMD.

Preliminary data The researh team pioneered the application of metabolomics to the study of AMD across all stages, and, in 2015, the AIBILI team recruited and collected baseline data on a total of 295 subjects (242 AMD patients and 53 controls). This study cohort will be derived from that study population.

The investigators have published extensively on the hypothesis that metabolomic profiles differ across AMD stages. The latest paper, which includes the entire study population, meta-analyses to combine data from different cohorts was used. This paper provides additional evidence that patients with AMD present an altered plasma metabolomic profile as compared to controls, and that these profiles vary with disease severity. Results revealed that 28 metabolites differed significantly between AMD patients versus controls (false discovery rate (FDR) q-value: 4.1 x 10-2 - 1.8 x 10-5), and 67 across disease stages (FDR q-value: 4.5 x 10-2 - 1.7 x 10-4). Pathway analysis showed significant enrichment of glycerophospholipid, purine, taurine and hypotaurine, and nitrogen metabolism (p-value < 0.04).

To assess the performance of models considering metabolite information, receiving operating curve (ROC) assessments were made. Both a model considering metabolite changes across disease stages (AUC = 0.815; 95% CI:0.771-0.860) and a model comparing patients with AMD versus controls (AUC = 0.789; 95% CI: 0.738-0.840) outperformed (p-value = 3.74 x 10-6 and p-value = 2.07 x 10-4, respectively) a more classical model considering demographic covariates alone (AUC = 0.725; 95% CI: 0.671-0.779) This is evidence that metabolomics can be a useful tool for longitudinal study of AMD progression.

General research design Five-years after the first AMD metabolomic study, all participants will be invited to participate in this study.

All participants will be invited to come to AIBILI to perform the study procedures, which include color fundus photographs (CFP), SD-OCT and swept source OCT (SS-OCT). Additionally, they will also be invited to answer a validated food frequency questionnaire (FFQ)19 and a questionnaire about their regular physical activity. Systemic comorbidities and current medication will be registered. Blood and urine samples will be collected for metabolomic profiling. Functional testing with microperimetry and dark adaptation will be offered as an additional optional study procedure.

Afterwards, the obtained CFP will be graded. According to the procedures for IN654 study classification, all CFP will be standardized using software developed by our group. Two independent graders, masked to any other data, will assess AMD staging of all eyes, according to the AREDS classification.21 Each eye of patients will be assessed separately, and if different, the most advanced eye will be considered as the classification for that subject. Previously performed IN654 study classification will be compared. Progressors will be defined as patients who: (i) had originally been classified as early AMD, and at five years have intermediate or late AMD; (ii) had originally been classified as intermediate AMD and have at five years late AMD. Non-progressors will be defined as those who remain within the same AMD stage at the five-year visit.

Inclusion Criteria:

• The study population will be derived from a cohort enrolled in a previous prospective study by our group on AMD metabolomics

Exclusion Criteria:

• No exclusion criteria will apply.