Risk of age-related macular degeneration: population-attributable fractions in 600 000 individuals without genetic information or baseline ocular phenotypes from two contemporary population-based studies
- 1Department of Clinical Biochemistry, Copenhagen University Hospital Herlev and Gentofte, Herlev, Denmark
- 2Department of Clinical Medicine, University of Copenhagen, Copenhagen, Denmark
Abstract
Objectives To determine absolute risks and population-attributable fractions for risk factors for neovascular and any age-related macular degeneration (AMD) in individuals without genetic information or baseline ocular phenotypes, and to define high-risk profiles for neovascular and any AMD.Methods and analysis In a total of 109 163 individuals (1315 incident neovascular AMD cases) in the Copenhagen General Population Study (CGPS) and 501 255 (10 419 incident any AMD cases) in the UK Biobank, we calculated HRs, 10-year cumulative incidences for absolute risk and population-attributable fractions for risk factors for neovascular and any AMD. High risk profiles were defined using the four risk factors with the highest population-attributable fractions.Results We identified eight concordant modifiable risk factors for neovascular AMD in CGPS and any AMD in UK Biobank: smoking, high alcohol intake, low physical activity, C-reactive protein >2 mg/L, low-density lipoprotein cholesterol ≤3 mmol/L, high high-density lipoprotein cholesterol (≥1.94 mmol/L in CGPS; ≥1.67 mmol/L in UK Biobank (≥4th quartile)), cataract surgery and diabetes. Their population-attributable fractions were 42% in CGPS and 24% in UK Biobank. Sensitivity analyses suggested possible confounding or reverse causation for cataract surgery. The risk profile comprising smoking, high alcohol intake and C-reactive protein >2 mg/L in women was associated with HRs of 2.58 (95% CI 1.70 to 3.92) for neovascular AMD in CGPS and 1.91 (1.63 to 2.24) for any AMD in UK Biobank compared with those with none of the risk factors.Conclusions In individuals without genetic information or baseline ocular phenotypes, eight modifiable risk factors were associated with both neovascular and any AMD in two population-based cohorts and may explain 42% of neovascular AMD cases and 24% of any AMD cases in the population representing possible targets for future prevention trials. Our risk profiles based on commonly available clinical parameters may assist in identifying high-risk individuals in primary care for further clinical follow-up.
Related
- Population burden of diabetes, hypertension, and cardiovascular disease attributable to obesity: a community-based study in Türkiye
- Navigating neonatal thrombosis: contemporary insights into a complex disorder in a unique population
- Global Burden, Temporal Trends, Extrapolation Based on Past Trends, and Population Attributable Fractions of Risk Factors for Type 2 Diabetes Mellitus: An Ecological Study Based on the Global Burden of Disease Study 2021.
- Preventable dementia in the Netherlands; population attributable fraction of modifiable risk factors
- CDKN2A c.146T>C (I49T) in Mexican-ancestry individuals: Findings from melanoma and population cohorts
BibTeX
@article{nordestgaard2026risk,
title = {{Risk of age-related macular degeneration: population-attributable fractions in 600 000 individuals without genetic information or baseline ocular phenotypes from two contemporary population-based studies}},
author = {Børge G Nordestgaard and Shoaib Afzal and Lars Michael Larsen and Sabina O Beheshti},
journal = {BMJ Open Ophthalmology},
year = {2026},
month = nov,
doi = {10.1136/bmjophth-2026-003039},
url = {https://doi.org/10.1136/bmjophth-2026-003039},
note = {CC BY},
}
RIS
TY - JOUR
TI - Risk of age-related macular degeneration: population-attributable fractions in 600 000 individuals without genetic information or baseline ocular phenotypes from two contemporary population-based studies
AU - Børge G Nordestgaard
AU - Shoaib Afzal
AU - Lars Michael Larsen
AU - Sabina O Beheshti
PY - 2026
DA - 2026/11/08
JO - BMJ Open Ophthalmology
DO - 10.1136/bmjophth-2026-003039
UR - https://doi.org/10.1136/bmjophth-2026-003039
AB - Objectives To determine absolute risks and population-attributable fractions for risk factors for neovascular and any age-related macular degeneration (AMD) in individuals without genetic information or baseline ocular phenotypes, and to define high-risk profiles for neovascular and any AMD.Methods and analysis In a total of 109 163 individuals (1315 incident neovascular AMD cases) in the Copenhagen General Population Study (CGPS) and 501 255 (10 419 incident any AMD cases) in the UK Biobank, we calculated HRs, 10-year cumulative incidences for absolute risk and population-attributable fractions for risk factors for neovascular and any AMD. High risk profiles were defined using the four risk factors with the highest population-attributable fractions.Results We identified eight concordant modifiable risk factors for neovascular AMD in CGPS and any AMD in UK Biobank: smoking, high alcohol intake, low physical activity, C-reactive protein >2 mg/L, low-density lipoprotein cholesterol ≤3 mmol/L, high high-density lipoprotein cholesterol (≥1.94 mmol/L in CGPS; ≥1.67 mmol/L in UK Biobank (≥4th quartile)), cataract surgery and diabetes. Their population-attributable fractions were 42% in CGPS and 24% in UK Biobank. Sensitivity analyses suggested possible confounding or reverse causation for cataract surgery. The risk profile comprising smoking, high alcohol intake and C-reactive protein >2 mg/L in women was associated with HRs of 2.58 (95% CI 1.70 to 3.92) for neovascular AMD in CGPS and 1.91 (1.63 to 2.24) for any AMD in UK Biobank compared with those with none of the risk factors.Conclusions In individuals without genetic information or baseline ocular phenotypes, eight modifiable risk factors were associated with both neovascular and any AMD in two population-based cohorts and may explain 42% of neovascular AMD cases and 24% of any AMD cases in the population representing possible targets for future prevention trials. Our risk profiles based on commonly available clinical parameters may assist in identifying high-risk individuals in primary care for further clinical follow-up.
SN - 2397-3269
ER -