Risk factors for incident age-related macular degeneration: pooled findings from 3 continents.

Objective: To examine risk factors for incident age-related macular degeneration (AMD) after combining data from 3 population-based cohort studies.

Design: Population-based cohort study.

Population: A population of 9523 adults (age range, 43-95 years at baseline) living in Australia, The Netherlands, and the United States who participated in a baseline examination and a follow-up examination on average 5 or 6 years later.

Methods: Similar procedures were used at all study sites. Examinations included a standardized questionnaire, pupillary dilation, and stereoscopic color fundus photography. Fundus photographs were graded for lesions associated with AMD using the Wisconsin and International Age-Related Maculopathy Grading Systems. Senior investigators from each site adjudicated all photos graded as late AMD.

Main Outcomes Measure: Incidence of late AMD.

Results: Among studies, distributions for most risk factors differed, and overall incidence rates were similar. In the Beaver Dam Eye Study, total serum cholesterol was inversely associated with incident neovascular AMD. In the Blue Mountains Eye Study, current smoking (defined as smoking at the time of the baseline examination) was associated with an increased risk of incident geographic atrophy and late AMD; increased total serum cholesterol, having diabetes, and older age at menopause were positively associated with incident geographic atrophy; and an increase in high-density lipoprotein serum cholesterol was inversely related to incident geographic atrophy. In the Rotterdam Study, current smoking was associated with an increased risk of incident geographic atrophy, neovascular AMD, and late AMD; past smoking was associated with an increased risk of incident neovascular AMD and late AMD; and an increased number of years between menarche and menopause was directly related to incident geographic atrophy. After pooling data, the only statistically significant relationships found were between smoking and total serum cholesterol and incident AMD. Current smoking was associated with an increased incidence of geographic atrophy and late AMD (odds ratios [ORs] relative to nonsmokers: 2.83 and 2.35, respectively; ORs relative to past smokers: 2.80 and 1.82, respectively), and total serum cholesterol was associated directly with incident geographic atrophy (OR: 1.08 per 10 mg/dl) and inversely with incident neovascular AMD (OR: 0.94 per 10 mg/dl).

Conclusions: Pooled data support a growing body of evidence indicating that smoking is related to an increased risk of incident AMD. Current smokers were at higher risk of incident AMD than both past smokers and those who never smoked. The relationships found in this study between total serum cholesterol and incident geographic atrophy and neovascular AMD are not readily explained.