Purchase this article with an account.
Amirul Islam, Elaine Chong, Allison Hodge, Robyn Guymer, Khin Zaw Aung, Galina Makeyeva, Dallas English, Graham Giles, Paul Baird, Liubov Robman; Data-Driven Pattern Analysis of Dietary Items in Relation to Age-Related Macular Degeneration. Invest. Ophthalmol. Vis. Sci. 2013;54(15):226.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
To identify dietary patterns by use of data-driven principal component analysis and to evaluate their associations with age-related macular degeneration (AMD) in the Melbourne Collaborative Cohort Study (MCCS).
Food frequency data were collected from the ethnically diverse cohort of 41514 older Australians in 1990-1994, and 21,132 of them were assessed in 2003-2007 for early or late AMD prevalence. Principal component analysis (PCA) was used to identify eating patterns from the measured food and beverage items. PCA is a data driven statistical method to reduce a large number of inter-correlated variables (e.g., dietary items) into a few distinct factors of inter-correlated variables which are similar within the factor and dissimilar between the factors. Logistic regression adjusting for age, sex, country of birth, smoking, educational level and total energy intake was used to assess associations with early and late AMD.
Of total, 2508 (12.8%) participants had early AMD and 108 (0.6%) had late AMD. Six factors characterised by predominant intakes of fruits (F1), vegetables (F2), grains, fish, steamed or boiled chicken, vegetables and nuts (F3), red meat products (F4), processed foods comprising of cakes, sweet biscuits and desserts (F5) and salad (F6) explained 33% of the total variation. F3 above median level was strongly associated with lower prevalence of late AMD, odds ratio (OR) (95% confidence interval) 0.59, (0.39-0.89), whereas F4 above median level was associated with higher prevalence of late AMD (OR 1.47, 95% CI 1.0-2.18).
Dietary factor characterised by grains, steamed, grilled and canned fish, steamed or boiled chicken, vegetables and nuts was associated with lower prevalence of late AMD whereas a factor characterised predominantly by red meat intake was associated with higher late AMD prevalence. These results amplify the arguments that the patterns and combinations of dietary intakes could be more informative than individual dietary component.
This PDF is available to Subscribers Only