Odds ratios under the inverse-variance weighted mendelian randomization method are shown for a 1-SD increase in the logodds of ever having smoked regularly, 1-SD increase in the logodds of smoking cessation (former vs current smoking), 1-SD increase in the lifetime smoking index, 1-SD increase of the age at which an individual started smoking regularly, 1-SD increase of log-transformed alcoholic drinks per week, 1-SD increase in body mass index (calculated as 4.8 kg/m2); 10-mm Hg increase in systolic and diastolic pressure; 1-SD increase in the logodds of having type 2 diabetes; 1% increase in glycated hemoglobin (HbA1c); 18.02-mg/dL (to convert to mmol/L, multiply by 0.0555) increase in fasting glucose; and 1 log-transformed (pmol/L) increase in fasting insulin.
eMethods. Study design, data sources, SNV selection, and statistical analysis
eTable 1. Smoking initiation exposure SNVs and their association with advanced AMD
eTable 2. Smoking cessation exposure SNVs and their association with advanced AMD
eTable 3. Lifetime smoking exposure SNVs and their association with advanced AMD
eTable 4. Age at smoking initiation exposure SNVs and their association with advanced AMD
eTable 5. Alcohol intake per week exposure SNVs and their association with advanced AMD
eTable 6. Body mass index exposure SNVs and their association with advanced AMD
eTable 7. Systolic blood pressure exposure SNVs and their association with advanced AMD
eTable 8. Diastolic blood pressure exposure SNVs and their association with advanced AMD
eTable 9. Type 2 diabetes exposure SNVs and their association with advanced AMD
eTable10. HbA1c exposure SNVs and their association with advanced AMD
eTable 11. Fasting glucose exposure SNVs and their association with advanced AMD
eTable 12. Fasting insulin exposure SNVs and their association with advanced AMD
eTable 13. Mendelian randomization analysis of modifiable risk factors for advanced AMD, geographic atrophy, and neovascular AMD
eFigure 1. Scatter plot of SNV-smoking initiation and SNV-AMD association estimates for the smoking initiation instrument
eFigure 2. Scatter plot of SNV-smoking cessation and SNV-AMD association estimates for the smoking cessation instrument
eFigure 3. Scatter plot of SNV-lifetime smoking and SNV-AMD association estimates for the lifetime smoking instrument
eFigure 4. Scatter plot of SNV-age at smoking initiation and SNV-AMD association estimates for the age at smoking initiation instrument
eFigure 5. Scatter plot of SNV-alcoholic drinks per week and SNV-AMD association estimates for the alcohol intake per week instrument
eFigure 6. Scatter plot of SNV-BMI and SNV-AMD association estimates for the body mass index instrument
eFigure 7. Scatter plot of SNV-systolic blood pressure and SNV-AMD association estimates for the systolic blood pressure instrument
eFigure 8. Scatter plot of SNV-diastolic blood pressure and SNV-AMD association estimates for the diastolic blood pressure instrument
eFigure 9. Scatter plot of SNV–type 2 diabetes and SNV-AMD association estimates for the type 2 diabetes instrument
eFigure 10. Scatter plot of SNV-HbA1c and SNV-AMD association estimates for the HbA1c instrument
eFigure 11. Scatter plot of SNV-fasting glucose and SNV-AMD association estimates for the fasting glucose instrument
eFigure12. Scatter plot of SNV-fasting insulin and SNV-AMD association estimates for the fasting insulin instrument
Customize your JAMA Network experience by selecting one or more topics from the list below.
Kuan V, Warwick A, Hingorani A, et al. Association of Smoking, Alcohol Consumption, Blood Pressure, Body Mass Index, and Glycemic Risk Factors With Age-Related Macular Degeneration: A Mendelian Randomization Study. JAMA Ophthalmol. 2021;139(12):1299–1306. doi:10.1001/jamaophthalmol.2021.4601
Are smoking, alcohol intake, blood pressure, body mass index, and glycemic traits associated with age-related macular degeneration (AMD)?
In this mendelian randomization study, genetically predicted smoking initiation and lifetime smoking were associated with elevated risk of advanced AMD, genetically predicted smoking cessation was associated with decreased risk of advanced AMD, and genetically predicted alcohol intake was associated with increased risk of geographic atrophy.
These findings support a potential causal association of alcohol consumption with an increased risk of geographic atrophy, smoking initiation and lifetime smoking with an increased risk of advanced AMD, and smoking cessation with a decreased risk of advanced AMD.
Advanced age-related macular degeneration (AMD) is a leading cause of blindness in Western countries. Causal, modifiable risk factors need to be identified to develop preventive measures for advanced AMD.
To assess whether smoking, alcohol consumption, blood pressure, body mass index, and glycemic traits are associated with increased risk of advanced AMD.
Design, Setting, Participants
This study used 2-sample mendelian randomization. Genetic instruments composed of variants associated with risk factors at genome-wide significance (P < 5 × 10−8) were obtained from published genome-wide association studies. Summary-level statistics for these instruments were obtained for advanced AMD from the International AMD Genomics Consortium 2016 data set, which consisted of 16 144 individuals with AMD and 17 832 control individuals. Data were analyzed from July 2020 to September 2021.
Smoking initiation, smoking cessation, lifetime smoking, age at smoking initiation, alcoholic drinks per week, body mass index, systolic and diastolic blood pressure, type 2 diabetes, glycated hemoglobin, fasting glucose, and fasting insulin.
Main Outcomes and Measures
Advanced AMD and its subtypes, geographic atrophy (GA), and neovascular AMD.
A 1-SD increase in logodds of genetically predicted smoking initiation was associated with higher risk of advanced AMD (odds ratio [OR], 1.26; 95% CI, 1.13-1.40; P < .001), while a 1-SD increase in logodds of genetically predicted smoking cessation (former vs current smoking) was associated with lower risk of advanced AMD (OR, 0.66; 95% CI, 0.50-0.87; P = .003). Genetically predicted increased lifetime smoking was associated with increased risk of advanced AMD (OR per 1-SD increase in lifetime smoking behavior, 1.32; 95% CI, 1.09-1.59; P = .004). Genetically predicted alcohol consumption was associated with higher risk of GA (OR per 1-SD increase of log-transformed alcoholic drinks per week, 2.70; 95% CI, 1.48-4.94; P = .001). There was insufficient evidence to suggest that genetically predicted blood pressure, body mass index, and glycemic traits were associated with advanced AMD.
Conclusions and Relevance
This study provides genetic evidence that increased alcohol intake may be a causal risk factor for GA. As there are currently no known treatments for GA, this finding has important public health implications. These results also support previous observational studies associating smoking behavior with risk of advanced AMD, thus reinforcing existing public health messages regarding the risk of blindness associated with smoking.
Age-related macular degeneration (AMD) is a leading cause of blindness in Western countries,1-4 accounting for 8.7% of blindness worldwide.5 The prevalence of AMD is projected to increase by 47% in the next 20 years because of population aging, posing a major burden to health care systems across the world.4 Advanced AMD consists of geographic atrophy (GA) and neovascular AMD (nAMD).6 Treatment is currently only available for nAMD and comes in the form of intravitreal injections of anti–vascular endothelial growth factor.3 This treatment is invasive, expensive, of limited effectiveness, and poses a considerable burden on patients.7-9 Therefore, increased public health efforts need to be directed toward prevention of advanced AMD. Identifying causal, modifiable risk factors for advanced AMD is critical to implementing interventions for prevention.
Observational epidemiological studies have investigated the associations of lifestyle and metabolic risk factors (smoking,10-12 alcohol consumption,13-16 obesity,8,17-19 blood pressure,20-22 glycemic traits,21,23-25 and dyslipidemia2,26,27) with AMD, but the findings are inconsistent and cannot be established as causal owing to limitations introduced by confounding and reverse causality. Randomized clinical trials (RCTs) allow reliable causal inferences to be drawn and have shown that dietary intake of specific high-dose combinations of zinc and antioxidants reduces progression from intermediate to advanced AMD.28-30 However, for diseases of aging, such as AMD, where there may be a long lag time between exposure to risk factors and clinical manifestation of disease, RCTs may be prohibitively expensive, time consuming, and infeasible, particularly for exposures such as smoking, alcohol intake, obesity, hypertension, glycemic traits, and blood lipid levels.
Quiz Ref IDMendelian randomization (MR) is a technique that has been used to assess potential causal associations across a wide range of diseases. MR is based on the principle that if a genetic variant causes a change in an exposure (eg, smoking or alcohol intake), and if this exposure is causal for a disease (eg, advanced AMD), then the genetic variant should also be associated with risk of the disease. This method exploits the random allocation of genetic variants at meiosis,31 rendering MR studies similar to RCTs in that they are less prone to confounding than classical observational epidemiological studies. MR studies also mitigate the problem of reverse causation because the genotype is invariant and not modified by disease.32 MR techniques based on multiple genetic variants selected from across the genome have found a causal role for increased levels of high-density lipoprotein cholesterol, but not low-density lipoprotein cholesterol or triglycerides, in advanced AMD risk.33,34
Here we use MR to assess the potential causal role of other exposures that are amenable to intervention (ie, smoking, alcohol intake, body mass index [BMI], blood pressure, and glycemic traits) on the risk of advanced AMD and its subtypes, GA and nAMD, using publicly available data.
Two-sample MR was performed, whereby summary-level data of genetic variants associated with smoking initiation (ever having smoked regularly), smoking cessation (former vs current smoking), lifetime smoking (represented by an index which captures smoking status, duration, heaviness, and cessation), age at smoking initiation, weekly alcohol intake, BMI, systolic and diastolic blood pressure, type 2 diabetes, glycated hemoglobin (HbA1c), fasting glucose level, and fasting insulin level were obtained from study samples that did not overlap with those for advanced AMD and its subtypes (eMethods in Supplement 1). Ethical approval for each data set had been obtained in the original studies.
Summary-level statistics for smoking traits, alcohol intake, BMI, blood pressure, and glycemic traits were obtained from, to our knowledge, the largest genome-wide association studies (GWAS) for these exposures to date (eMethods in Supplement 1). Summary-level genetic association data for advanced AMD, GA, and nAMD were obtained from the International AMD Genomics Consortium (IAMDGC) 2016 data set.35 Data were analyzed from July 2020 to September 2021. The number of participants in each study, the race and ethnicity of the participants, and the number of single-nucleotide variants (SNVs; formerly SNPs) in the genetic instruments for each exposure are provided in the Table. The SNVs used as instrumental variables for the exposures in this study (eTables 1-12 in Supplement 1) were obtained from the studies listed in the Table. SNVs were selected according to criteria outlined in the eMethods in Supplement 1.
We performed univariable inverse-variance–weighted (IVW) 2-sample MR analyses under a multiplicative random-effects model to examine the potential causal associations of smoking, alcohol intake, BMI, blood pressure, and glycemic risk factors with the risk of advanced AMD and its subtypes, GA and nAMD (eMethods in Supplement 1).
Additionally, we conducted sensitivity analyses using the weighted median,36 MR-Egger,37 and MR pleiotropy residual sum and outlier (MR-PRESSO)38 methods, as well as multivariable MR (MVMR) with smoking traits adjusted for alcohol intake and vice versa (eMethods in Supplement 1).
We analyzed 12 potentially modifiable lifestyle and metabolic exposures (Table). Wald test was used to calculate 2-tailed P values. Using a conservative approach, we applied a Bonferroni-corrected significance level of .004 (.05 divided by 12). Analyses were conducted using the MR, MR-PRESSO, and 2-sample MR packages in R version 3.5.0 (the R Foundation).
Quiz Ref IDGenetic predisposition to smoking initiation was associated with overall risk of advanced AMD under the IVW method (OR [odds ratio], 1.26; 95% CI, 1.13-1.40; P < .001) (Figure; eFigure 1 and eTable 13 in Supplement 1). This was supported by the MVMR analysis (OR, 1.19; 95% CI, 1.05-1.36; P = .007) and weighted median analysis (OR, 1.20; 95% CI, 1.02-1.40; P = .03) (Figure; eFigure 1 and eTable 13 in Supplement 1). The results were similar for nAMD (IVW: OR, 1.26; 95% CI, 1.11-1.43; P < .001; MVMR: OR, 1.18; 95% CI, 1.02-1.37; P = .02; weighted median: OR, 1.29; 95% CI, 1.08-1.54; P = .005). Smoking initiation was not associated with GA (IVW: OR, 1.24; 95% CI, 1.03-1.49; P = .02; weighted median: OR, 1.31; 95% CI, 1.00-1.72; P = .05) (Figure; eTable 13 in Supplement 1). There was no indication of heterogeneity using the Cochran Q statistic (Figure) or directional pleiotropy using MR-Egger and MR-PRESSO methods (eTable 13 in Supplement 1).
Genetically predicted smoking cessation was protective for advanced AMD compared with persistent smoking (IVW: OR, 0.66; 95% CI, 0.50-0.87; P = .003; MVMR: OR, 0.71; 95% CI, 0.57-0.88; P = .002; MR-Egger: OR, 0.43; 95% CI, 0.21-0.87; P = .02) (Figure; eFigure 2 and eTable 13 in Supplement 1). There was no association between smoking cessation and nAMD or GA (Figure; eTable 13 in Supplement 1). No evidence was found of heterogeneity using the Cochran Q statistic (Figure) or directional pleiotropy using MR-Egger and MR-PRESSO (eTable 13 in Supplement 1).
Genetically predicted lifetime smoking (represented by a composite index taking into account smoking status, duration, heaviness, and cessation)39 was associated with higher odds of advanced AMD (IVW: OR, 1.32; 95% CI, 1.09-1.59; P = .004; MVMR: OR, 1.48; 95% CI, 1.14-1.94; P = .004; weighted median: OR, 1.42; 95% CI, 1.11-1.81; P = .005) (Figure; eFigure 3 and eTable 13 in Supplement 1). The results were similar for nAMD (IVW: OR, 1.41; 95% CI, 1.14-1.74; P = .001; MVMR: OR, 1.62; 95% CI, 1.19-2.19; P = .002; weighted median: OR, 1.42; 95% CI, 1.08-1.88; P = .01) (Figure; eTable 13 in Supplement 1). The association with GA was not statistically significant. Heterogeneity was detected in the variant-specific estimates for lifetime smoking (Figure). However, the multiplicative random-effects IVW method accounts for heterogeneity and provides valid estimates under the assumption of balanced pleiotropy.40 There was no indication of directional (unbalanced) pleiotropy using the MR-Egger method. The MR-PRESSO test detected no outliers for advanced AMD or nAMD and 1 outlier for GA, with no significant difference in the estimates before and after adjusting for the outlier (eTable 13 in Supplement 1). The age at which smoking was initiated was not found to be associated with risk of advanced AMD and its subtypes (Figure; eFigure 4 and eTable 13 in Supplement 1).
We found no association between increased genetically predicted alcohol consumption and risk of advanced AMD using IVW (OR, 1.57; 95% CI, 1.03-2.40; P = .04), MVMR (OR, 1.65; 95% CI, 1.21-2.26; P = .002), weighted median (OR, 2.04; 95% CI, 1.23-3.39; P = .006), or MR-PRESSO (OR, 1.47; 95% CI, 1.02-2.10; P = .04) (Figure; eFigure 5 and eTable 13 in Supplement 1). Examination of advanced AMD subtypes showed a strong association between higher alcohol intake and GA (IVW: OR, 2.70; 95% CI, 1.48-4.94; P = .001; MVMR: OR, 2.83; 95% CI, 1.64-4.88; P < .001; and MR-PRESSO: OR, 2.50; 95% CI, 1.42-4.42; P = .002), but not nAMD (Figure; eTable 13 in Supplement 1). There was evidence of heterogeneity (Figure), but this was accounted for under the multiplicative random-effects IVW method.40 There was no indication of directional (unbalanced) pleiotropy using MR-Egger. The MR-PRESSO test detected 3 outliers for advanced AMD—1 for GA and 2 for nAMD—but there was no significant difference between the estimates before and after correction for outliers (eTable 13 in Supplement 1).
There was no evidence that BMI, blood pressure, type 2 diabetes, HbA1c, fasting glucose level, or fasting insulin level had a causal association with the risk of AMD using MR (Figure; eFigures 6-12 and eTable 13 in Supplement 1).
Quiz Ref IDThis study used an MR framework to explore potential causal associations between the risk of advanced AMD and the following modifiable risk factors: smoking, alcohol consumption, BMI, blood pressure, and glycemic traits. We found genetic evidence supporting a potential causal association between smoking initiation and advanced AMD risk consistent with previous observational studies. This association was stronger for nAMD than for GA. Similar results were found for lifetime smoking behavior. Additionally, smoking cessation was associated with a decreased risk of advanced AMD, specifically nAMD, compared with persistent smoking. We also found suggestive evidence for a possible causal association between increased alcohol consumption and risk of advanced AMD that was likely driven by a strong association with GA. There was insufficient evidence to suggest a potential causal association with the other exposures, namely BMI, blood pressure, or glycemic risk factors, on advanced AMD risk.
Conventional observational studies have consistently implicated smoking as a risk factor for developing AMD (OR, 1.7-4.6),10,11,41-43 with 1 study reporting the risk of AMD reduced to baseline 20 years after smoking cessation.44 Using MR techniques, we found genetic evidence to support that smoking initiation and lifetime smoking behavior may be causally associated with risk of advanced AMD and that smoking cessation is protective. Several theories have been proposed to explain the association between smoking and AMD. Oxidative stress is thought to play a major role in AMD pathogenesis.45 Smoking is known to decrease levels of antioxidants,46 resulting in the disruption of the retinal pigment epithelium barrier and leading to the formation of drusen and neovascularization.47-50 Smoking can also upregulate endothelial smooth muscle cell proliferation, leading to choroidal neovascularization.51 Atherosclerosis or vasoconstriction secondary to smoking could cause hypoxic conditions in the retina,52 stimulating production of vascular endothelial growth factor and resulting in retinal endothelial cell proliferation and neovascularization.53 Smoking has also been found to be associated with the production of inflammatory mediators and activation of the complement cascade involved in the pathogenesis of AMD.54-56 While these hypotheses may explain how smoking can cause AMD generally, it has been unclear whether and how smoking affects GA and nAMD differently. Our finding that smoking behavior has a greater potential association with nAMD compared with GA may be a reflection of how the pathogenesis of nAMD diverges from what is assumed to be the underlying default pathway of early AMD to GA.57,58 Further investigation is required to provide insights into how these pathways differ.
The evidence in the literature for an association between alcohol consumption and AMD risk is less robust than for smoking. While high alcohol intake has been reported to be associated with an increased risk of AMD, a systematic review concluded that residual confounding effects from smoking could not be ruled out.13,59,60 One study has conversely reported a protective effect with moderate alcohol consumption.61 Our results provide genetic evidence of a potential causal association between increased weekly alcohol consumption and risk of GA. While there are no studies, to our knowledge, looking at the role of alcohol consumption in the pathophysiology of GA specifically, the mechanism of alcohol as a risk factor for AMD is thought to be related to oxidative damage. Alcohol depletes antioxidant levels and results in the production of reactive oxygen species.62,63 Further studies are required to investigate why increased alcohol intake appears to have a causal association with GA but not with nAMD.
The effects of BMI and blood pressure on AMD are uncertain. Some observational studies have found increased BMI and blood pressure to be associated with increased AMD risk, whereas others showed no association.17,26 A meta-analysis of cohort studies found an increased risk of advanced AMD in individuals with obesity.17 One cohort study found antihypertensive medication to be associated with increased AMD risk.64 However, cohort studies are prone to confounding, and meta-analyses are prone to publication bias. Our MR estimates did not indicate potential causal associations for either BMI or blood pressure traits with risk of advanced AMD.
Reports on the association between diabetes and AMD are conflicting. A recent meta-analysis found that diabetes was associated with increased risk of AMD.25 However, the authors stated that it was not possible to rule out underlying confounders, as most of the included studies only adjusted for age and sex. Other studies have reported a protective effect for diabetic retinopathy on AMD risk.65 While both AMD and diabetic retinopathy are inflammatory retinal conditions, AMD is characterized by alterations in the outer blood-retinal barrier (BRB) and diabetic retinopathy by changes in the inner BRB. Damage to the inner BRB in diabetes is associated with upregulation of outer BRB activity,66 which could potentially provide a protective mechanism for AMD. The MR estimates for the 4 glycemic traits analyzed in this study did not reach statistical significance.
This study provides genetic evidence that smoking initiation and lifetime smoking behavior are potential causal risk factors for advanced AMD, and that stopping smoking may be protective against the risk of advanced AMD. The detrimental effects of smoking on multiple conditions, such as cardiovascular disease, cancers, and chronic obstructive pulmonary disease, are well known. Public health campaigns should disseminate information that smoking can also lead to blindness as an additional deterrent against smoking. We also found genetic evidence that increased alcohol consumption has a potential causal association with GA risk. Here again, public health messages and clinical advice regarding the harms of excessive alcohol intake should include the risk of blindness, especially given that there are currently no effective treatments for GA. Given the deterioration in quality of life and the cost to health and social care systems of managing advanced AMD, increased funding should be allocated to smoking cessation and alcohol reduction programs to minimize the health burden of advanced AMD.
This study has several strengths. The use of MR mitigates bias from reverse causation and confounding that can affect findings from observational studies. The 2-sample MR approach increases statistical power as it allows independent large GWAS summary data sets to be used for both exposures and outcomes.67,68 The use of multiple genetic variants as instruments for each of the risk factors enables SNVs across the whole genome to be used to assess the association of each of the modifiable exposures with risk of advanced AMD. This allowed us to perform sensitivity analyses to detect and correct for directional pleiotropy. Most of the participants in the GWAS data sets were of European descent, thus minimizing bias due to population stratification.
Quiz Ref IDThis study also has several limitations. While we used summary statistics from the largest known advanced AMD GWAS to date, the number of advanced AMD cases, especially for GA, was still relatively small compared with other outcomes used in MR studies, such as cardiovascular disease. This might have resulted in insufficient statistical power to detect associations with some of the exposures analyzed. Additionally, effect sizes from MR analyses should be interpreted with caution. This is because the MR estimate is better interpreted as a test statistic for a causal hypothesis rather than the expected impact of a clinical intervention at a specific point in time.69-71 Another potential limitation in MR analyses is directional horizontal pleiotropy, which can occur if genetic variants affect advanced AMD independently of the risk factors investigated. To identify and adjust for pleiotropy, we performed sensitivity analyses using MR-Egger, weighted median, MVMR, and MR-PRESSO methods (eFigures 1-12 in Supplement 1). These tests can provide reliable inferences when some genetic variants are pleiotropic.72-74 There was no evidence for directional pleiotropy using these tests. However, it is not possible to completely rule out the presence of residual pleiotropy. Additionally, we were unable to test the nonlinear association for nonbinary exposures, such as alcohol consumption, blood pressure, BMI, and quantitative glycemic traits, with advanced AMD risk, as MR analysis assumes a linear association between exposure and outcome.
We found genetic evidence that increased alcohol consumption has a potential causal association with risk of GA. We also present genetic evidence that smoking initiation and lifetime smoking behavior may be casually associated with risk of advanced AMD, while smoking cessation results in a lower risk of advanced AMD than persistent smoking. These associations were stronger for nAMD than for GA. To reduce the prevalence of advanced AMD in aging populations, public health campaigns and programs to support smoking abstention, smoking cessation, and reduced alcohol intake should incorporate the evidence that these activities can lead to blindness. The finding that smoking and alcohol have differential effects on nAMD and GA may prompt future studies examining the different pathologies of these 2 forms of advanced AMD.
Accepted for Publication: September 20, 2021.
Published Online: November 4, 2021. doi:10.1001/jamaophthalmol.2021.4601
Open Access: This is an open access article distributed under the terms of the CC-BY License. © 2021 Kuan V et al. JAMA Ophthalmology.
Corresponding Author: Valerie Kuan, PhD, Institute of Health Informatics, University College London, 222 Euston Rd, London NW1 2DA, United Kingdom (email@example.com).
Author Contributions: Dr Kuan had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Concept and design: Kuan, Hingorani, Tufail, Sofat.
Acquisition, analysis, or interpretation of data: Kuan, Warwick, Tufail, Cipriani, Burgess, Sofat.
Drafting of the manuscript: Kuan, Warwick.
Critical revision of the manuscript for important intellectual content: Kuan, Hingorani, Tufail, Cipriani, Burgess, Sofat.
Statistical analysis: Kuan, Warwick, Cipriani, Burgess.
Obtained funding: Cipriani, Sofat.
Administrative, technical, or material support: Kuan, Tufail.
Supervision: Hingorani, Tufail, Sofat.
Conflict of Interest Disclosures: Dr Kuan reports grants from Dunhill Medical Trust. Dr Tufail reports personal fees from Apellis, Bayer, Genentech/Roche, IVERIC Bio, and Heidelberg Engineering, and personal fees and grants from Novartis. No other disclosures were reported.
Funding/Support: Dr Kuan is funded by the Dunhill Medical Trust (RPGF1806\67). Dr Warwick is supported by the Wellcome Trust (220558/Z/20/Z). Dr Hingorani is supported by a BHF Research Accelerator Award (AA/18/6/34223). Dr Tufail is supported by the National Institute for Health Research to Moorfields Eye Hospital and the Biomedical Research Centre for Ophthalmology. Dr Burgess is supported by Sir Henry Dale Fellowship, jointly funded by the Wellcome Trust and the Royal Society (204623/Z/16/Z) Dr Sofat is supported by the National Institute for Health Research, University College London Hospitals Biomedical Research Centre. This research was supported by the NIHR Cambridge Biomedical Research Centre (BRC-1215-20014).
Role of the Funder/Sponsor: The funders had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; or decision to submit the manuscript for publication.
The International AMD Genomics Consortium (IAMDGC): The group members are listed in Supplement 2.
Disclaimer: The views expressed are those of the authors and not necessarily those of the National Institute for Health Research or the Department of Health and Social Care.