Purpose: The gut microbiome is highly influential in diseases with inflammatory components. Multiple studies showed a link between the gut microbiome and age-related macular degeneration (AMD). However, no consistent taxa have been reported.
Methods: We included 1372 participants from the Rotterdam Study (RS). AMD features (e.g. [reticular pseudo-]drusen, retinal pigment epithelium degeneration, and hyperpigmentation) were graded by human graders on color fundus photographs. Next, areas of these features were automatically quantified by a deep learning segmentation model. Stages were determined according to the RS classification (preliminary = 239, early-intermediate = 75, late = 6). Propensity score matching was performed on age, sex and BMI. Multivariable associations with taxonomic and functional profiles were assessed using zero-adjusted models (MaAsLin2; Compound Poisson). p-values were false discovery rate-adjusted (q-values). Taxa associated (q < 0.25) with at least two out of seven AMD(-related) outcomes are reported.
Results: We observed no associations with alpha- or beta-diversity. Nineteen taxa were associated with AMD, of which 7 persisted after additional adjustment for dietary data. Eubacterium xylanophilum group, Lachnoclostridium, Faecalibacterium, Odoribacter splanchnicus and Parabacteroides distasonis were associated with an AMD phenotype. Parabacteroides and Akkermansia were inversely associated with an AMD phenotype. Moreover, 46 MetaCyc pathways were associated with AMD, of which 15 persisted after additional adjustment for dietary data. GLUCARDEG.PWY, PWY.5028, PWY.5347, PWY.5415, PWY.5532, PWY.5971, PWY.6969 and TCA were associated with an AMD phenotype. FOLSYN.PWY, LEU.DEG2.PWY, PANTO.PWY, PANTOSYN.PWY, PWY.6612, PWY0.1586 and PYRIDNUCSYN.PWY were inversely associated with an AMD phenotype.
Conclusions: Several gut microbiota were associated with an AMD phenotype. AMD pathophysiology might be linked to changes in gut-related metabolic pathways.