Human Molecular Genetics

Human Molecular Genetics Advance Access originally published online on September 25, 2006
Human Molecular Genetics 2006 15(21):3206-3218; doi:10.1093/hmg/ddl396

This Article Abstract FREE Full Text (PDF) Supplementary Data All Versions of this Article: 15/21/3206    most recent ddl396v1 Alert me when this article is cited Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Similar articles in ISI Web of Science Similar articles in PubMed Alert me to new issues of the journal Add to My Personal Archive Download to citation manager Search for citing articles in: ISI Web of Science (32) Request Permissions Google Scholar Articles by Conley, Y. P. Articles by Gorin, M. B. Search for Related Content PubMed PubMed Citation Articles by Conley, Y. P. Articles by Gorin, M. B. Social Bookmarking          What's this?

© The Author 2006. Published by Oxford University Press. All rights reserved. For Permissions, please email: journals.permissions@oxfordjournals.org

CFH, ELOVL4, PLEKHA1 and LOC387715 genes and susceptibility to age-related maculopathy: AREDS and CHS cohorts and meta-analyses

Yvette P. Conley^1,2, Johanna Jakobsdottir³, Tammy Mah⁴, Daniel E. Weeks^2,3, Ronald Klein⁶, Lewis Kuller⁵, Robert E. Ferrell² and Michael B. Gorin^2,4,*

¹ Department of Health Promotion and Development, School of Nursing, ² Department of Human Genetics, Graduate School of Public Health, ³ Department of Biostatistics, Graduate School of Public Health, ⁴ Department of Ophthalmology, School of Medicine and ⁵ Department of Epidemiology, Graduate School of Public Health, University of Pittsburgh, The Eye and Ear Institute Building, 203 Lothrop Street, Pittsburgh, PA 15261, USA and ⁶ Department of Ophthalmology and Visual Sciences, School of Medicine and Public Health, University of Wisconsin, Madison, WI 53705, USA

^* To whom correspondence should be addressed. Tel: +1 4126477726; Fax: +1 4126475880; Email: gorinmb{at}upmc.edu

Received August 9, 2006; Accepted September 15, 2006

	ABSTRACT

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS SUPPLEMENTARY MATERIAL REFERENCES

 
Age-related maculopathy (ARM) is an important cause of visual impairment in the elderly population. It is of crucial importance to identify genetic factors and their interactions with environmental exposures for this disorder. This study was aimed at investigating the CFH, ELOVL4, PLEKHA1 and LOC387715 genes in independent cohorts collected using different ascertainment schemes. The study used a case–control design with subjects originally recruited through the Cardiovascular Health Study (CHS) and the Age-Related Eye Disease Study (AREDS). CFH was significantly associated with ARM in both cohorts (P0.00001). A meta-analysis confirmed that the risk allele in the heterozygous or homozygous state (OR, 2.4 and 6.2; 95% CI, 2.2–2.7 and 5.4–7.2, respectively) confers susceptibility. LOC387715 was also significantly associated with ARM in both cohorts (P0.00001) and a meta-analysis confirmed that the risk allele in the heterozygous and homozygous state (OR, 2.5 and 7.3; 95% CI, 2.2–2.9 and 5.7–9.4, respectively) confers susceptibility. Both CFH and LOC387715 showed an allele-dose effect on the ARM risk, individuals homozygous at either locus were at more than two-fold risk compared to those heterozygous. PLEKHA1, which is closely linked to LOC387715, was significantly associated with ARM status in the AREDS cohort, but not the CHS cohort and ELOVL4 was not significantly associated with ARM in either cohort. Joint action of CFH and LOC387715 was best described by independent multiplicative effect without significant interaction in both cohorts. Interaction of both genes with cigarette smoking was insignificant in both cohorts. This study provides additional support for the CFH and LOC387715 genes in ARM susceptibility via the evaluation of cohorts that had different ascertainment schemes regarding ARM status and through the meta-analyses.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS SUPPLEMENTARY MATERIAL REFERENCES

 
Age-related maculopathy (ARM) is a leading cause of central blindness in the elderly of industrialized nations. The prevalence of ARM is expected to increase because of the aging of these populations (1). The etiology of ARM is complex, with environmental as well as genetic susceptibility playing a role. Association-based analyses are generally more sensitive to small genetic effects than linkage-based analyses and are extremely valuable for fine mapping of disease-related genes (2). Case–control association studies with the use of unrelated individuals may have advantages over family-based studies, especially when a multilocus genetic model is anticipated (3,4), however, such studies are potentially sensitive to the ascertainment scheme for the case and control cohorts. For this reason, there is value in assessing candidate genes in populations from projects with different study designs. This current study investigates the complement factor H (CFH) gene, the elongation of very long chain fatty acid-like 4 (ELOVL4) gene, the pleckstrin homology domain-containing protein (PLEKHA1) gene, and the hypothetical LOC387715 gene in two such distinct cohorts.

The association of the CFH gene with ARM susceptibility has been established in samples of European American descent (5–10) as well as in samples from the United Kingdom (11), Germany (12), France (13), Iceland (14) and Japan (15).

Three studies support the PLEKHA1/LOC387715 locus on chromosome 10q26 (12,16,17). The study by Jakobsdottir et al. (16) reported that the PLEKHA1/LOC387715 locus was significantly associated with ARM status, however, strong linkage disequilibrium between PLEKHA1 and LOC387715 in the independent family-based and case–control populations utilized for the study meant that a role for one gene over the other could not be determined (16). Evidence that the hypothetical LOC387715 gene was more likely to be the gene accounting for susceptibility to ARM came from a study by Rivera et al. (12) that utilized two independent case–control samples (12) and a study by Schmidt et al. that utilized both family-based and case–control studies (17). All three studies indicated that the association of this region on chromosome 10q26 with ARM status was independent of the association with CFH that had been previously reported in all three populations (6,9,12). In addition, based on the Schmidt et al. study, the effect of the LOC387715 locus appears to be modified by smoking history (17).

Two studies have evaluated a potential role for ELOVL4 in ARM in humans. Ayyagari et al. (18) evaluated the gene and found no significant association with ARM status in their sporadic case–control analysis. However, Conley et al. found a significant association of ELOVL4 and ARM status in familial and sporadic case–control analyses (9). The difference in findings between these studies may be related to the proportion of cases with exudative ARM in each population, since Conley et al. found that ELOVL4 was especially associated with the exudative sub-phenotype (9). These results indicate that additional studies are needed to establish or refute a relationship between ELOVL4 and ARM.

The two cohorts utilized for this study were the Cardiovascular Health Study (CHS), a population-based cohort of individuals 65 years and older at baseline for which ARM status was not a factor for ascertainment (19), and the Age-Related Eye Disease Study (AREDS), a cohort of individuals aged 55–80 years participating in a randomized controlled clinical trial of anti-oxidant and zinc intervention for which ARM status was a factor for ascertainment (20). These cohorts have been previously described (21,22).

This study was designed to evaluate the CFH, ELOVL4, PLEKHA1 and LOC387715 genes in two independent cohorts with very different ascertainment schemes in relation to ARM status and then to incorporate the findings into meta-analyses. Association of a gene with susceptibility to ARM regardless of ascertainment scheme would further increase the evidence that the association is real and would enhance the likelihood that evaluation of the gene(s) would accurately identify at risk individuals.

	RESULTS

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS SUPPLEMENTARY MATERIAL REFERENCES

 
To further evaluate CFH, ELOVL4, PLEKHA1 and LOC387715 in ARM, we genotyped previously reported SNPs within all four genes in samples from the AREDS and CHS studies. Separate analyses were performed on each data set, using a total of 701 non-Hispanic white ARM patients and 175 controls from the AREDS study, and a total of 126 non-Hispanic white ARM patients and 1051 controls from the CHS study (see Table 1 for sample sizes and other characteristics of the data, and Table 2 for genotype frequencies). The disease status of subjects at their last follow-up visit was the primary endpoint evaluated for AREDS subjects. The AREDS subjects include controls of grade 1 and cases (grades 3–5) with moderate ARM and advanced ARM in one or both eyes. The ARM disease status of CHS subjects was evaluated by Dr Gorin, using monocular, non-mydriatic fundus photographs taken at the 8-year follow-up visit. The majority of CHS cases had moderate ARM including multiple drusen with and without pigment epithelial changes (equivalent to AREDS grade 3) with a small number of cases having geographic atrophy (GA) or choroidal neovascular membranes (CNV) and the CHS controls are of AREDS grade 1 with the exclusion of those cases with significant extramacular drusen.

View this table: [in this window] [in a new window]   Table 1. Characteristics of the study populations

 

View this table: [in this window] [in a new window]   Table 2. Genotype distributions by ARM status

 
Association analyses
For each gene, CFH, ELOVL4, PLEKHA1 and LOC387715, association of one non-synonymous SNP with ARM was assessed by a ² statistic. The magnitude of the effect of each variant was estimated by odds ratios (ORs) and population attributable risks (PARs). To evaluate whether the variants confer risk similarly to mild/moderate and advanced ARM, ORs were calculated for each grade and subtype (GA and CNV) separately using the AREDS data.

CFH.
The association of the Y402H variant in CFH with ARM is extremely significant (P0.00001) in both the AREDS and CHS cohorts (Table 3), confirming earlier findings by ourselves (9,16) and others (5–7,12). The estimated ORs for Y402H in CFH suggest that the variant confers similar risk to all stages of ARM and both forms of advanced ARM, GA and CNV (Fig. 1 and Supplementary Material, Table S2). An allele-dose effect appears to be present, with carriers of two C alleles at higher risk of ARM than carriers of one C allele (Table 4 and Supplementary Material, Fig. S1). Despite the increased risk in carriers of two C alleles, the PAR is similar for the two risk genotypes, owing to relatively high frequency of the CT genotype compared to the CC genotype in the general population. PAR estimates derived from the CHS data set suggest that the CT and CC genotypes explain 27% and 25% of ARM in the non-Hispanic white population, respectively.

View larger version (13K): [in this window] [in a new window] [Download PowerPoint slide]   Figure 1. Estimated crude ORs and 95% CIs for CFH, ELOVL4, PLEKHA1 and LOC387715 genes. Carriers of one or two risk alleles (RR+RN) are compared with those subjects homozygous for the non-risk allele (NN). The solid lines denote the 95% CI corresponding to an OR (open circle). The dotted vertical line marks the null value of an OR of 1. The contrasts that were evaluated in AREDS and CHS cohorts are given on the vertical axis.

 

View this table: [in this window] [in a new window]   Table 3. Results of allele- and genotype-association tests

 

View this table: [in this window] [in a new window]   Table 4. ORs and PAR% for subjects who are hetero- and homozygous for Y402H in CFH and S69A in LOC387715

 
ELOVL4.
The M299V variant in ELOVL4 is significantly associated (P=0.034) with exudative ARM in the AREDS sample (Table 3), in agreement with our previous findings (9). However, no ORs are statistically significant at 95% significance level (Fig. 1 and Supplementary Material, Table S2 and Supplementary Material, Fig. S2). These results do not exclude the potential role of ELOVL4 in ARM, but do not strongly support it. The small number of individuals with exudative ARM did not allow for subphenotype analysis in the CHS cohort.

PLEKHA1 and LOC387715.
The association of the S69A variant in LOC387715 with all presentations of ARM is extremely significant (P0.00001) in both the AREDS and CHS data sets (Table 3), confirming earlier findings by ourselves (16) and others (12,17). The A320T variant in PLEKHA1, which is located on the same haplotype block as LOC387715, is highly significant (P=0.00004) in the AREDS sample but only borderline significant (P=0.08) in the CHS sample. The degree of linkage disequilibrium between A320T and S69A is statistically significant in both AREDS (D'=0.66) and CHS (D'=0.65) controls. In order to identify which gene, PLEKHA1 or LOC387715, more likely harbors the true ARM-predisposing variant, we applied the haplotype method (23). According to the haplotype method, the relative frequency of alleles at neutral variants is expected to be the same in cases and controls for a haplotype containing all the predisposing variants. The results based on applying the method suggest that S69A in LOC387715, and not A320T in PLEKHA1, is an ARM-predisposing variant (Supplementary Material, ‘Distinguishing between PLEKHA1 and LOC387715—Results’). Further, by permutation testing, the null hypothesis: H₀: the S69A variant in LOC387715 fully accounts for the ARM predisposition to the PLEKHA1–LOC387715 haplotype block, is not rejected (P=0.92 in the AREDS data, P=0.45 in the CHS data), while a similar hypothesis for A320T is rejected (P0.0001 in the AREDS data, P=0.0002 in the CHS data).

The S69A variant in LOC387715 shows different risk patterns than Y402H in CFH. The variant appears to increase the risk of severe ARM substantially more than the risk of mild ARM (Fig. 1, Supplementary Material, Table S2 and Supplementary Material, Fig. S4) in the AREDS data where severity of disease is differentiated. For example, the OR for AREDS cases of grade 3, who carry one or two T alleles, is 3.07 (95% CI 1.82–5.17), while the OR for AREDS cases, with CNV in both eyes, who carry one or two T alleles, is 7.21 (95% CI 4.24-12.27). Similar to CFH, S69A shows an allele-dose effect without dramatic differences in the PAR of the GT and TT genotypes (Table 4 and Supplementary Material, Fig. S4). Since only four AREDS controls are TT homozygous at S69A, point estimates and CIs, for recessive and homozygote contrasts, derived from regular logistic regression were compared with estimates from exact regression [models fitted in SAS software release 8.2 (SAS Institute Inc., Cary, NC, USA)]. These quality checks revealed no major differences in point estimates (which is the basis of the PAR estimates) and lower confidence limits (which is the basis of comparison with the ORs), but the upper confidence limits were higher (results not shown).

Interaction analyses
We used logistic regression modeling to build a model of the joint contribution of CFH and LOC387715, CFH and cigarette smoking and LOC387715 and cigarette smoking. A series of models were fitted in order to draw inferences about the most likely and most parsimonious model(s). As described by North et al. (24), models were compared using the Akaike information criterion (AIC). When the most parsimonious model had been identified we estimated joint ORs of the risk factors. Separate estimates were calculated from each cohort. In order to maximize the AREDS sample size, no subphenotype or subgrade analyses were performed; AREDS cases of grade 3–5 were compared with AREDS controls of grade 1.

In a previous article (16), we found no evidence of interacting effects of the CFH and PLEKHA1/LOC387715 loci; the joint action of the two loci was best described by independent multiplicative effects (additive on a log-scale). Rivera et al. (12) reported that S69A in LOC387715 acted independently of Y402H in CFH. Schmidt et al. (17) also arrived at the same most parsimonious model. The AREDS and CHS data also suggest that the two genes contribute independently to disease risk. The best fitting model (the model with the smallest AIC) derived from the AREDS data is an additive model with an interaction term. This model, with AIC of 721.4, does however not provide a significantly better fit (AIC difference <2) than a simpler additive model with AIC of 723.0. The additive model is the most parsimonious model (AIC=635.1) derived from the CHS data and is also the best fitting model (Table 5). Joint ORs for combinations of risk genotypes at Y402H and S69A were computed to further understand the joint action of the two loci (Supplementary Material, Table S4). Using all cases regardless of severity, the AREDS data suggest that individuals heterozygous for the risk allele at one of the loci and homozygous for the non-risk allele at the other are more susceptible to ARM than individuals with no-risk allele at both loci (for the CT–GG joint genotype, OR 2.8, 95% CI 1.6–5.0; for the TT–GT joint genotype, OR 3.2, 95% CI 1.7–6.0). The ARM risk more than doubles if a person is heterozygous at both loci (for the CT–GT joint genotype, OR 7.2, 95% CI 3.8–13.5) and being homozygous for the risk allele for at least one of the loci further increases the risk. The joint ORs estimated from the CHS data show a similar pattern, but having only one risk allele is not sufficient to increase the risk (for the CT–GG joint genotype, OR 1.3, 95% CI 0.6–2.7; for the TT–GT joint genotype, OR 1.2, 95% CI 0.5–2.8).

View this table: [in this window] [in a new window]   Table 5. Results of fitting two-factor models by logistic regression

 
A recent study (17) reported a strong statistical interaction between genotypes at S69A and smoking, both on binary (ever versus never smoked) and continuous scale (pack-years of smoking). We fail to replicate this finding in both the AREDS and CHS data sets (Table 5). Results from the AREDS sample suggests that the joint effects of Y402H and smoking are best described by independent multiplicative effects, without significant dominance or interacting effects. On the other hand, the model that best describes the CHS data includes only additive effects of Y402H. Results from the AREDS data suggest that the joint effects of S69A and smoking are best described by independent multiplicative effects, without significant dominance or interacting effects. The CHS data implicate a model with only S69A. When smoking exposure is a continuous variable (pack-years of smoking) and the S69A genotypes are coded in additive fashion, the interaction term is not significant (P=0.40) in the CHS data. Pack-years of cigarette smoking were not available for participants in the AREDS study. To further understand the combined effect of the genes and cigarette smoking, joint ORs of risk genotypes at each gene and smoking were estimated from the AREDS data (Supplementary Material, Table S7). The results suggest that, while the risk of ARM due to any of the risk genotypes (at Y402H and S69A) is elevated in smokers, both genes have substantially more influence on ARM risk than cigarette smoking. Both the model fitting approach and a simple ² test (P=0.71) show that the main effects of cigarette smoking are insignificant (on binary scale) in the CHS data.

APOE results
Main effects of the APOE gene in ARM were tested using the CHS data. Neither the distribution of APOE-4 carriers (P=0.41) nor APOE-2 (P=0.42) carriers was significantly different between cases and controls, when compared to APOE-3/3.

Meta-analyses
Meta-analysis of CFH.
We used a meta-analysis approach to pool estimated ORs for Y402H from 11 independent data sets [including the CHS and AREDS cohorts reported here (Supplementary Material, Table S10)]. This resulted in the analysis of 5451 cases and 3540 controls all of European or European American descent. The results confirm the increased ARM risk due to the C allele in the non-Hispanic white population (Fig. 2 and Supplementary Material, Table S11). The pooled estimates have narrower CI than any individual study, and non-overlapping CI for hetero- and homozygote ORs: OR_het=2.43 (95% CI 2.17–2.72) and OR_hom=6.22 (95% CI 5.38–7.19), when assuming homogeneity across studies. When the analysis is performed under heterogeneity, the point estimates are essentially the same and the CIs are slightly wider. Leave-one-out sensitivity analyses, under a fixed effect model, show that no study has dramatic influence on the pooled estimates (Supplementary Material, Table S11). The study by Rivera et al. (12) changes the estimates more than any other study; when the study is excluded, the OR_dom and OR_het are approximately 0.2 higher, while the OR_rec and OR_hom are lowered by approximately 0.2. The Rivera et al. study is the only study where the genotype distribution, in the control group, deviates from Hardy–Weinberg equilibrium [HWE (P=0.03)]. The allele and genotype distributions, in cases and controls, are strikingly similar across studies. However, the genotype distribution in CHS cases differs from the other studies and the frequency of the CC risk genotype is lower compared to other cohorts (Supplementary Material, Fig. S5).

View larger version (15K): [in this window] [in a new window] [Download PowerPoint slide]   Figure 2. Estimated ORs and 95% CIs, derived from data sets included in meta-analysis of Y402H in CFH, and pooled estimates from fixed and random effect models. The top figure shows ORhet (OR for CT heterozygotes compared to TT) and the bottom figure shows ORhom (OR for CC homozygotes compared to TT). ‘Hage-C’ and ‘Hage-I’ denote estimates derived from the Columbia and Iowa cohorts of Hageman et al., respectively, and ‘Jakobs’ denotes estimates from the Jakobsdottir et al. paper. ‘Fixed’ denotes pooled estimates derived from all the studies assuming the between-study variability is due to chance. ‘Random’ denotes pooled estimates derived from all the studies allowing for heterogeneity across studies. ‘nARM’ is the total number of ARM cases included in the estimates and ‘ncon’ is the total number of controls without ARM included in the estimates. For the Haines et al. study ‘nARM’ and ‘ncon’ refer to the whole sample (individuals of all genotypes). The dotted vertical line marks the point estimate of the pooled OR under homogeneity (‘Fixed’).

 
Meta-analysis of LOC387715.
Meta-analysis of the risk associated with S69A in ARM included five independent data sets [including the CHS and AREDS cohorts reported here (Supplementary Material, Table S12)]. This resulted in the analysis of 3147 cases and 2381 controls all of European or European American descent. The studies of LOC387715 are more heterogeneous than the studies of CFH; OR_dom and OR_het differ significantly across studies (P<0.01 and 0.02, respectively). The results support earlier findings of the association of the T allele with increased ARM risk (Supplementary Material, Table S13). Carriers of two T alleles are at substantially higher risk than are carriers of one T allele; when accounting for between-study variation, the OR_het and OR_hom are 2.48 (95% CI 1.67–3.70) and 7.33 (95% CI 4.33–12.42), respectively. The genotype distribution is similar across all control populations and across all ARM populations, except the CHS ARM population (Supplementary Material, Fig. S6).

	DISCUSSION

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS SUPPLEMENTARY MATERIAL REFERENCES

 
During the past year, major discoveries of associations of the CFH and PLEKHA1/LOC387715 genes with ARM were published. A number of reports established a strong association of the Y402H coding change in CFH with ARM and three reports found an association, of similar magnitude as the association of Y402H, of the S69A coding change in LOC387715 with ARM. Both of those genes lie within chromosomal regions, CFH on 1q31 and LOC387715 on 10q26, consistently identified by family-based linkage studies (25–31).

Because the majority of the studies of Y402H and all three studies of S69A were specially designed to search for (and find) genes involved in ARM complex etiology, it is possible that they overestimate the effect size of the risk alleles at Y402H and S69A. Therefore, we analyzed two independent case–control cohorts with varying inclusion and exclusion criteria based on ARM status, the AREDS and CHS cohorts. The AREDS cohort did have inclusion and exclusion criteria relevant to severity of ARM and both affected and non-affected individuals were enrolled (20). In contrast, the CHS cohort is a population-based cohort that utilized community-based recruitment of individuals 65 years and older with no inclusion and exclusion criteria relevant to ARM status (19). Retinal assessments in the CHS cohort were not conducted until the 8-year follow-up visit. Given the difference in ascertainment of subjects into the two studies, replication of association of a candidate gene in both cohorts greatly strengthens the support for its causal involvement in ARM pathogenesis.

We evaluated previously reported associations of four genes, CFH (1q31), ELOVL4 (6q14), PLEKAH1 (10q26) and LOC387715 (10q26). Variants in both CFH and LOC387715 are extremely significantly (P0.00001) associated with ARM in both AREDS and CHS cohorts. Both variants show an allele-dose effect on the ARM risk and a model of independent multiplicative contribution of the two genes is most parsimonious in both AREDS and CHS cohorts. The A320T coding change in the PLEKHA1 gene, adjacent to and in linkage disequilibrium with LOC387715 on 10q26, is significantly associated with ARM in the AREDS cohort (P=0.00004), but not in the CHS cohort (P=0.08). Because of extensive linkage disequilibrium between PLEKHA1 and LOC387715 in our initial study population we could not, with reasonable certainty, distinguish between their association signals. Our results based on applying the haplotype method to both the AREDS and CHS cohorts, combined with the findings of Rivera et al. (12), who used conditional haplotype analysis and detected, for the first time, a weak expression of LOC387715 in the retina, and Schmidt et al. (17), who detected only a weak association signal at PLEKHA1, indicate that S69A in LOC387715 is most likely the major ARM-predisposing variant on 10q26. The results of the haplotype method show that PLEKHA1 is not sufficient to account for the ARM-predisposition at 10q26; however, we cannot exclude the possibility that A320T in PLEKHA1 may be on a causative haplotype with S69A and other unknown variants.

The replication of associations of CFH and LOC387715 genes with ARM in AREDS and CHS cohorts, two cohorts with different study designs, continues to provide strong support for their involvement in ARM. Variable findings for PLEKHA1 in AREDS and CHS cohorts do however need to be considered in the light of differences between the two cohorts. In addition to differences in ascertainment of the case and control populations, the evaluation of retinal changes, documentation of retinal findings and prevalence of advanced ARM differed between the two cohorts. In the CHS study, fundus photography was only available for one randomly selected eye and the photography was performed with non-dilated pupils and these limitations could certainly influence the sensitivity to detect disease pathology, although this is more likely to influence the detection of early retinal changes. The proportion of advanced ARM in the entire CHS cohort that was evaluated at the 8-year follow-up evaluation was 1.3% (21) compared to 17% in the AREDS (22) and the variation in the proportion of advanced ARM disease pathology between the two cohorts could lead to variation in findings, especially if a gene is more likely to influence progression of the disease. In addition, one important difference between these two cohorts is the timing of the retinal evaluations. AREDS participants had retinal evaluations conducted at baseline as well as during follow-up evaluations, whereas CHS participants had retinal evaluations done eight or more years after enrollment, when they would have been at least 73 years old. It is possible that survival to the retinal evaluation for the CHS participants could bias the population available for this particular type of study. Potential confounding issues related to the use of the AREDS cohort are that subjects in categories other than the unaffected group were randomized into a clinical trial using vitamin and mineral supplements to evaluate the impact of these on ARM progression and there is some evidence indicating that unaffected subjects in category 1 have different demographic characteristics than affected subjects in the other categories (22). It is not clear whether these could impact the results of our study, but it should be considered when findings are interpreted.

As mentioned previously, most studies that have investigated the genetic etiology of ARM were designed to optimize identification of regions of the genome housing susceptibility genes for ARM and for ARM candidate gene testing. Published attributable risks range from 43 to 68% (5,6,16,17) for the Y402H variant in CFH and from 36 to 57% (16,17) for the S69A variant in LOC387715. Interestingly, the PARs for the CHS population are lower than those previously published: 41% for the Y402H variant in CFH and 27% for the S69A variant in LOC387715 (Supplementary Material, Table S2). Because the majority of the CHS cases have moderate ARM the PAR estimates derived from the CHS data are not completely comparable with estimates from previous studies in which the proportion of patients with advanced ARM was considerably higher. However, they are comparable to estimates derived from using AREDS cases of grade 3. Those estimates are within the previously published range of PARs: 49% for Y402H in CFH and 45% for S69A in LOC387715. These findings may indicate that the ARM attributed to these two susceptibility variants may be lower than previously thought, given that the CHS cohort was not ascertained based on ARM status. A prospective design is needed to more precisely estimate the relative risks, which are approximated by ORs estimated from retrospective case–control designs, and corresponding PARs.

We were not able to replicate the association of ELOVL4 with overall ARM (9). The number of individuals with exudative ARM allowed us to perform subphenotype analysis in the AREDS, but not the CHS cohort. Subphenotype analysis was especially important with regard to ELOVL4, where our previous findings indicated a role for ELOVL4 in exudative ARM; this is trending towards significance in the AREDS cohort. Given the lack of strong association and significant ORs for ELOVL4 in ARM susceptibility in both cohorts and the lack of association reported by Ayyagari et al., it is very unlikely that ELOVL4 plays a substantial role in ARM susceptibility. The power to detect an OR of 0.6 for overall ARM is reasonable, with type I error rate 5%, minor allele frequency 0.15 and population prevalence 6% the power is 81% in AREDS and 69% in CHS. The power to detect the same effect in exudative ARM is only 53% in AREDS data, under the same conditions. Therefore, the possibility that ELOVL4 plays a role in overall ARM is unlikely but mild effect in exudative ARM cannot be refuted. These power estimates were performed using Quanto (32).

We also used the CHS cohort to test whether the 4 or 2 alleles of the APOE gene are associated with ARM. In several studies, the 2 allele is suggested to contribute to disease risk and the 4 allele has been found to protect from ARM. Our results do not reach statistical significance and do not support the hypothesized role of the gene in ARM pathogenesis.

The AREDS and CHS data support the independent contribution of Y402H in CFH and S69A in LOC387715 to ARM susceptibility. A multiplicative risk model for these two variants is the most parsimonious based on evaluation of the AREDS and CHS cohorts; this model was also supported by our previous paper (16) as well as data presented by Rivera et al. and Schmidt et al. (12,17). The ARM risk appears to increase as the total number of risk alleles at Y402H and S69A increases (Supplementary Material, Table S4).

Prior to the discovery of CFH and LOC387715 cigarette smoking was one of the more important known ARM-related risk factors. Cigarette smoking is generally accepted as a modifiable risk factor for ARM; van Leeuwen et al. provide a review of the epidemiology of ARM and discuss the support of smoking as ARM risk factor (33). Schmidt et al. (17) recently reported statistically significant interaction between LOC387715 and cigarette smoking in ARM. Their data suggested that the association of LOC387715 with ARM was primarily driven by the gene effect in heavy smokers. Our own analyses of interaction do not support this finding and the AREDS data suggest that the joint action of S69A and smoking is multiplicative.

A role for CFH and LOC387715 in ARM susceptibility is further supported via the results of our meta-analysis. The meta-analysis, which include the CHS and AREDS cohorts reported in this article, indicates that having one or two copies of the risk allele at CFH or LOC387715 increases the risk of ARM, and those who have two copies are at higher risk. The combined results from all studies as well as the results from each independent study were remarkably tight (Figs 2 and 3). One known limitation of meta-analysis is the susceptibility to publication bias. Generally, such bias is a result of non-publication of negative findings (34). In the case of CFH and LOC387715, all published studies have reported strong association with ARM in the same direction, with the risk allele for CFH being the allele that codes for histidine and the risk allele for LOC387715 being the allele that codes for serine. We expect the preferential publication of statistically significant associations to show random directionality if the significant association is a false-positive result (35). It is therefore unlikely that the consistency of the association of CFH and LOC387715 with ARM is a result of publication bias.

View larger version (15K): [in this window] [in a new window] [Download PowerPoint slide]   Figure 3. Estimated ORs and 95% CIs, derived from data sets included in meta-analysis of S69A in LOC387715, and pooled estimates from fixed and random effect models. The top figure shows ORhet (OR for GT heterozygotes compared to GG) and the bottom figure shows ORhom (OR for TT homozygotes compared to GG). ‘Jakobs’ denote estimates from the Jakobsdottir et al. paper. ‘Fixed’ denotes pooled estimates derived from all the studies assuming the between-study variability is due to chance. ‘Random’ denotes pooled estimates derived from all the studies allowing for heterogeneity across studies. ‘nARM’ is the total number of ARM cases included in the estimates and ‘ncon’ is the total number of controls without ARM included in the estimates. For the Haines et al. study, ‘nARM’ and ‘ncon’ refer to the whole sample (individuals of all genotypes). The dotted vertical line marks the point estimate of the pooled OR under homogeneity (‘Fixed’).

 
While the results of our statistical analyses are in agreement with LOC387715 being the major ARM-related gene on 10q26, they do not prove causality. The possible causal role of CFH in ARM pathogenesis has been further supported by the localization of its protein within drusen deposits of ARM patients and involvement in activation of the complement pathway. Regarding LOC387715, little is currently known about the biology of the gene and nothing about how its protein may affect ARM susceptibility. Until recently the expression of LOC387715 appeared limited to the placenta, but recently weak expression was reported in the retina (12), which opens up the possibility of a tissue-specific role of the gene.

In summary, our results continue to support a role of both CFH and LOC387715 in etiology of ARM, given that both genes harbor variants highly associated with ARM, regardless of how the subjects were ascertained. Evaluation of PLEKHA1 and ELOVL4 in the AREDS and CHS cohorts demonstrates that these genes are much less likely to play role in ARM susceptibility. The CFH and LOC387715 genes appear to act independently in a multiplicative way in ARM pathogenesis and individuals homozygous for the risk alleles at either locus are at highest risk. The continued support for these genes in ARM susceptibility will hopefully bring us closer to being able to utilize the information in these genes to identify at risk individuals and provide a rational basis for future clinical trials to test preventive therapies in high-risk cohorts as well as to provide insights into the basic pathogenesis of this condition.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS SUPPLEMENTARY MATERIAL REFERENCES

 
Cardiovascular health study (CHS) participants—sampling and phenotyping
CHS is a population-based, longitudinal study primarily designed to identify factors related to cardiovascular disease in those aged 65 and older. Retinal assessments were performed at the 8-year follow-up visit. Community-based recruitment took place in Forsyth County, NC; Sacramento County, CA; Washington County, MD; and Pittsburgh, PA. Medicare eligibility lists of the Health Care Financing Administration were utilized to identify individuals who were aged 65 and older. Individuals aged 65 years and older living in the households of list members were also eligible. Inclusion criteria were minimal and included being non-institutionalized, expected to remain in the area for at least 3 years, able to give informed consent, not wheelchair-bound, not receiving hospice care and not receiving radiation or chemotherapy for cancer (19). DNA samples from the CHS from participants who consented for genetic studies were used for this research. Only DNA samples from subjects who had a retinal examination where the findings fit our criteria of a case or control were included in this study.

CHS subjects usually had the retina of one randomly selected eye photographed and the photographs were graded by Dr Gorin using the same classification model that was described in prior publications (29). Only Caucasian individuals are included in the analysis, as the sample size of other groups with ARM is too small for reasonable results: there were 180 black controls but only three cases, and five controls of other races. All CHS cases (n=126) used for analyses are ‘Type A’, which falls into our most stringent model for clinical classification (29). Individuals in this category are clearly affected with ARM based on extensive and/or coalescent drusen, pigmentary changes (including pigment epithelial detachments) and/or the presence of end-stage disease (GA and/or CNV membranes). Very few CHS cases had end-stage ARM, GA or CNV (Table 1); therefore, analyses of specific subtypes of ARM were not conducted. All CHS controls (n=1051) were of AREDS grade 1. A few potential controls (n=22) had unclear signs of GA or CNV and were excluded from analyses.

Age-related eye disease study (AREDS) participants—sampling and phenotyping
AREDS is a prospective, multicenter study of the natural history of ARM and age-related cataract with a clinical trial of high-dose vitamin and mineral supplementation embedded within the study. Individuals recruited into the AREDS study were men and women aged 55–80 years at enrollment; these individuals were required to be free of any condition or illness that would hinder long-term follow-up. Inclusion criteria were minimal and included having ocular media clear enough to allow for fundus photography and either no evidence of ARM in either eye or having ARM in one eye while the other maintained good vision (20/30 or better) (20). DNA samples from subjects who consented for genetic studies from the NEI-AREDS Genetic Repository were used for this research.

ARM status was assigned using the AREDS ARM grading system and based on phenotypes assigned at the most recent follow-up visit. Again, only Caucasian individuals are included in the analysis, as the sample size of other groups is too small for reasonable results: there are only 15 African American, two Hispanic and three individuals of other races. AREDS cases (n=701) consisted of grade 3, 4 and 5. AREDS subjects of grade 3 (n=96) have ARM but do not suffer from end-stage ARM, subjects of grade 4 (n=266) have end-stage ARM in one eye and subjects of grade 5 (n=339) have end-stage ARM in both eyes. AREDS controls (n=175) have AREDS grade 1 (grade 2 individuals were excluded prior to analyses).

Genotyping
The M299V variant in ELOVL4 (rs3812153), the Y402H variant in CFH (rs1061170) and the S69A variant in LOC387715 (rs10490924) were genotyped using RFLP techniques. The primers, annealing temperatures and restriction endonuclease for each assay were: 5'-AGATGCCGATGTTGTTAAAAG-3' (F), 5'-CATCTGGGTATGGTATTAAC-3' (R), 50°C and BspHI for ELOVL4; 5'-TCTTTTTGTGCAAACCTTTGTTAG-3' (F), 5'-CCATTGGTAAAACAAGGTGACA-3' (R), 52°C and NlaIII for CFH; 5'-GCACCTTTGTCACCACATTA-3' (F), 5'-GCCTGATCATCTGCATTTCT-3' (R), 54°C and PvuII for LOC387715.

The A320T variant in PLEKHA1 (rs1045216) was genotyped using 5' exonuclease Assay-on-Demand TaqMan assays (Applied Biosystems Incorporated). Amplification and genotype assignments were conducted using the ABI7000 and SDS 2.0 software (Applied Biosystems Incorporated). For all genotyping conducted for this research, double-masked genotyping assignments were made for each variant, compared and each discrepancy addressed using raw data or by re-genotyping.

Association analyses
SNP-disease association was measured with allele- and genotype ² tests, and P-values were simulated using 100 000 replicates; in cases with one or more expected cell numbers less than five, the Fisher's exact test was used. The strength of the association was estimated by crude OR and PAR. A general formula was used to calculate the PAR: PAR=P_r(OR–1)/(1+P_r(OR–1)), where P_r is the prevalence of the risk factor in the general population. Estimates of P_r were derived from the CHS controls; this is reasonable, because the CHS subjects were not selected on the basis of ARM disease status, and the number of CHS controls is large (n=1,051). Confidence intervals for the PARs were derived using asymptotic normal distribution of log(1–PAR) and transforming to an interval for the PAR. The CIs derived in this way are likely to be too narrow when the risk factor is rare (P_r<0.1) and sample sizes are small (36). For comparison purposes, ORs adjusted (OR_adj) for age and gender were estimated. Logistic regression models were used to calculate both crude and adjusted ORs, using R (37). The less frequent allele in the control group was considered the risk allele, and the OR and OR_adj were calculated by comparing those homozygous for the risk allele (RR) to the baseline group [those homozygous for the normal allele (NN)] and comparing those heterozygous for the risk allele (RN) to the baseline group. The contrasts for dominance (RR and RN versus NN) and recessive (RR versus RN and NN) effects were also evaluated.

Distinguishing between PLEKHA1 and LOC387715
We employed the haplotype method (23) to identify which one of the two loci, A320T in PLEKHA1 or S69A in LOC387715, is more likely the actual disease predisposing variant in the 10q26 region. The basis of the haplotype method is simple and elegant [for a mathematical proof, see Valdes and Thomson (23)]. If all predisposing variants are included on a haplotype, then the neutral variants are expected to be in the same ratio in cases and controls on a particular disease-predisposing haplotype, although the actual frequencies may differ. On the other hand, if not all predisposing variants have been identified, equality in the ratios of haplotype frequencies of non-predisposing variants is not expected. The expected ratios for the A320T–S69A haplotype are formulated in the Supplementary Material. Two null hypotheses were tested: one that A320T fully accounts for the ARM predisposition to the PLEKHA1–LOC387715 haplotype block, and the other that S69A fully accounts for the ARM predisposition to the PLEKHA1–LOC387715 haplotype block (for details on the hypotheses and permutation procedure to generate P-values, see the Supplementary Material). The program SNPHAP (38) was used to estimate haplotype frequencies and individual haplotypes. SNPHAP uses the EM algorithm to calculate a maximum likelihood estimate of haplotype frequencies given the unphased genotype data. The posterior probabilities of individual haplotype assignments exceed 87% for every individual typed at both A320T and S69A. For 80% of the haplotype assignments the underlying genotype at one or both loci is homozygous and hence the posterior probability is 100%.

Interaction analyses
The analyses of interaction were three-fold: first, we tested for interacting genetic effects of Y402H in CFH and S69A in LOC387715 in both CHS and AREDS samples, then we tested for interaction of both Y402H and S69A with smoking history in both CHS and AREDS samples and finally we calculated joint ORs of the three risk factors.

We followed a modeling strategy proposed by North et al. (24). Series of logistic regression models are fitted to the AREDS and CHS data sets in order to find the model that best describes the joint effects of CFH and LOC387715. For each genotype, models allowing for additive effects (ADD1, ADD2 and ADD-BOTH), and models which incorporate dominance effects (DOM1, DOM2 and DOM-BOTH) are fitted. The ADD1 model includes only the term x₁ for additive effects of CFH, coded as –1 for genotype TT at Y402H, as 0 for genotype CT and as 1 for genotype CC. The ADD2 includes only model term x₂ for additive effects of LOC387715, coded as –1 for genotype GG at S69A, as 0 for genotype GT and as 1 for genotype TT. The ADD-BOTH models the joint additive effects of CFH and LOC387715. The DOM1 incorporates dominance effects to ADD1, and includes x₁ and z₁, coded as 0.5 for genotype CT and –0.5 for genotypes TT and CC at Y402H. The DOM2 model similarly incorporates dominance effects to ADD2, and includes x₂ and z₂, coded as 0.5 for genotype GT and –0.5 for genotypes GG and TT at S69A. DOM-BOTH models the joint dominance effects of CFH and LOC387715. Three further models, that model the interaction between CFH and LOC387715 are fitted: ADD-INT includes the product term x₁*x₂, ADD-DOM includes x₁*x₂, x₁*z₂ and z₁*x₂ and DOM-INT includes x₁*x₂, x₁*z₂, z₁*x₂ and z₁*z₂.

The above modeling strategy was modified to investigate the joint effects of CFH and smoking, and the joint effects of LOC387715 and smoking. The modified approach is the same as used by Schmidt et al. (17) to test for interaction between LOC387715 and smoking. The coding scheme is the same, as above, except that smoking is coded as 0 for never smokers and 1 for ever smokers. The models fitted for the effects of CFH and smoking are: ADD1, SMOKE, ADD1-SMOKE, DOM1, ADD1-SMOKE-INT and DOM1-SMOKE-INT, and the models fitted for the effects of LOC387715 and smoking are: ADD2, SMOKE, ADD2-SMOKE, DOM2, ADD2-SMOKE-INT and DOM2-SMOKE-INT.

All models were compared by the AIC. Models for which the AIC differed by <2 are considered indistinguishable (24), and the model with fewer parameters was chosen as the most parsimonious model. Since adjusting for age and gender did not affect the estimates of ORs for Y402H nor S69A (Supplementary Material, Table S3), and to keep number of parameters as small as possible, no adjustment was made for these covariates when modeling interaction. Based on the results of the above interaction analyses, joint ORs were calculated.

APOE analyses
Previous studies have reported possible protective and harmful effects of the apolipoprotein E (APOE) gene in ARM. The 4 allele may have protective effects (39–43), whereas the least frequent allele, 2, may increase the risk of ARM (39,43). The APOE variant was genotyped by CHS and its association with ARM was assessed in this study. Individuals were classified by APOE genotype into individuals with APOE-3/3 genotype, and APOE-2 and APOE-4 carriers (denoted APOE-2/* and APOE-4/*, respectively); individuals with APOE-2/4 genotype were included in both the APOE-2/* and APOE-4/* groups. ² tests were used to test for differences in distributions of APOE-3/3 and APOE-2/*, and APOE-3/3 and APOE-4/*, genotypes in controls and cases.

Meta-analyses
We undertook a meta-analysis approach to pool estimated OR from previously published reports on CFH and LOC387715 and the two reports presented here. Initially data were analyzed, assuming the between-study variation is due to chance, and fixed-effects model was employed. Under the fixed-effect model, the maximum likelihood estimator of the pooled OR is an average of individual estimates, weighted by the inverse of their variances, and the variance of the pooled OR is estimated by the inverse of the sum of individual weights. Meta-analyses under homogeneity were performed in R (37). The assumption of homogeneity was checked using a ² test. However, tests of homogeneity tend to have low power, and therefore, for comparison, we also pooled the OR in a random effects setting. Meta-analyses under heterogeneity were performed using the method of restricted maximum likelihood (REML), as implemented in SAS Proc Mixed [SAS software release 8.2 (SAS Institute Inc.)]. The pooled REML estimator is identical to the DerSimonian-Laird estimator (44,45). The SAS codes by van Houwelingen et al. (45) were modified to perform the analyses under heterogeneity. A literature search was performed in PubMed in May 2006 and was limited to the English language. CFH studies were found by entering the search phrase: (CFH or ‘Complement Factor H’) and (‘Age-related macular degeneration’ or ‘Age-related maculopathy’ or AMD or ARM). Similarly, LOC387715 studies were found using the search phrase: LOC387715 and ‘Age-related macular degeneration’ or ‘Age-related maculopathy’ or AMD or ARM. The only inclusion criterion was that the research participants were Caucasian.

The Y402H variant within CFH has been found strongly associated with ARM in 11 studies (5–14,16); two of these 11 studies are ours, so only the results from our Jakobsdottir et al. (16) paper, that evaluated all contrasts, were used in meta-analysis. The Klein et al. (7) study used a small subset of the AREDS sample, and the Magnusson et al. (14) paper only reported allele-based ORs and no genotype counts. Therefore, these two studies were not included. Results from the Haines et al. (6) study were included in pooled estimates of ORs for hetero- and homozygotes; genotype counts were not available to evaluate contrasts for dominance and recessive effects. Three studies have reported highly associated variant, S69A, within the hypothetical LOC387715 (12,16,17). All three reports on LOC387715 were included in the meta-analysis. Research participants in all studies of CFH and LOC387715 are non-Hispanic whites of European and European American descent. Supplementary Material, Tables S10 and S12 summarize the studies included in the meta-analyses of CFH and LOC387715, respectively.

	SUPPLEMENTARY MATERIAL

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS SUPPLEMENTARY MATERIAL REFERENCES

 
Supplementary Material is available at HMG Online.

	ACKNOWLEDGEMENTS

 
We want to especially acknowledge the AREDS participants and the AREDS Research Group as well as the CHS participants and CHS investigators for their valuable contribution to this research. Regarding the meta-analyses, we are also grateful to Dr Swaroop and colleagues for providing CFH genotype count data, and Dr Schmidt and colleagues for providing LOC387715 genotype count data. This study was supported by NEI grant R01EY009859, The Steinbach Foundation, New York, Research to Prevent Blindness, New York and the Eye and Ear Foundation of Pittsburgh (all to M.B. Gorin). The CHS-related research reported in this article was supported by contracts N01-HC-85079 through N01-HC-85086, N01-HC-35129, N01 HC-15103, N01 HC-55222 and U01 HL080295 from the National Heart, Lung and Blood Institute, with additional contribution from the National Institute of Neurological Disorders and Stroke. A full list of participating CHS investigators and institutions can be found at http://www.chs-nhlbi.org.

Conflict of Interest statement. The authors of this manuscript have no conflicts of interest, financial or otherwise, to disclose, though the University of Pittsburgh has filed a patent application regarding the use of variants in PLEKHA1 and LOC387715 for the determination of genetic risk associated with the development of ARM.

	REFERENCES

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS SUPPLEMENTARY MATERIAL REFERENCES

 

1. The Eye Diseases Prevalence Group Research. (2004) Causes and prevalence of visual impairment among adults in the United States. Arch. Ophthalmol. 122:477–485.[Abstract/Free Full Text]

2. Cordell H.J. and Clayton D.G. (2005) Genetic association studies. Lancet 366:1121–1131.[CrossRef][Web of Science][Medline]

3. Howson J.M., Barratt B.J., Todd J.A., Cordell H.J. (2005) Comparison of population- and family-based methods for genetic association analysis in the presence of interacting loci. Genet. Epidemiol. 29:51–67.[CrossRef][Web of Science][Medline]

4. Risch N. (2001) Implications of multilocus inheritance for gene–disease association studies. Theor. Popul. Biol. 60:215–220.[CrossRef][Web of Science][Medline]

5. Edwards A.O., Ritter R. III, Abel K.J., Manning A., Panhuysen C., Farrer L.A. (2005) Complement factor H polymorphism and age-related macular degeneration. Science 308:421–424.[Abstract/Free Full Text]

6. Haines J.L., Hauser M.A., Schmidt S., Scott W.K., Olson L.M., Gallins P., Spencer K.L., Kwan S.Y., Noureddine M., Gilbert J.R., et al. (2005) Complement factor H variant increases the risk of age-related macular degeneration. Science 308:419–421.[Abstract/Free Full Text]

7. Klein R.J., Zeiss C., Chew E.Y., Tsai J.Y., Sackler R.S., Haynes C., Henning A.K., Sangiovanni J.P., Mane S.M., Mayne S.T., et al. (2005) Complement factor H polymorphism in age-related macular degeneration. Science 308:385–389.[Abstract/Free Full Text]

8. Hageman G.S., Anderson D.H., Johnson L.V., Hancox L.S., Taiber A.J., Hardisty L.I., Hageman J.L., Stockman H.A., Borchardt J.D., Gehrs K.M., et al. (2005) A common haplotype in the complement regulatory gene factor H (HF1/CFH) predisposes individuals to age-related macular degeneration. Proc. Natl Acad. Sci. USA 102:7227–7232.[Abstract/Free Full Text]

9. Conley Y.P., Thalamuthu A., Jakobsdottir J., Weeks D.E., Mah T., Ferrell R.E., Gorin M.B. (2005) Candidate gene analysis suggests a role for fatty acid biosynthesis and regulation of the complement system in the etiology of age-related maculopathy. Hum. Mol. Genet. 14:1991–2002.[Abstract/Free Full Text]

10. Zareparsi S., Branham K.E., Li M., Shah S., Klein R.J., Ott J., Hoh J., Abecasis G.R., Swaroop A. (2005) Strong association of the Y402H variant in complement factor H at 1q32 with susceptibility to age-related macular degeneration. Am. J. Hum. Genet. 77:149–153.[CrossRef][Web of Science][Medline]

11. Sepp T., Khan J.C., Thurlby D.A., Shahid H., Clayton D.G., Moore A.T., Bird A.C., Yates J.R. (2006) Complement factor H variant Y402H is a major risk determinant for geographic atrophy and choroidal neovascularization in smokers and nonsmokers. Invest. Ophthalmol. Vis. Sci. 47:536–540.[Abstract/Free Full Text]

12. Rivera A., Fisher S.A., Fritsche L.G., Keilhauer C.N., Lichtner P., Meitinger T., Weber B.H. (2005) Hypothetical LOC387715 is a second major susceptibility gene for age-related macular degeneration, contributing independently of complement factor H to disease risk. Hum. Mol. Genet. 14:3227–3236.[Abstract/Free Full Text]

13. Souied E.H., Leveziel N., Richard F., Dragon-Durey M.A., Coscas G., Soubrane G., Benlian P., Fremeaux-Bacchi V. (2005) Y402H complement factor H polymorphism associated with exudative age-related macular degeneration in the French population. Mol. Vis. 11:1135–1140.[Web of Science][Medline]

14. Magnusson K.P., Duan S., Sigurdsson H., Petursson H., Yang Z., Zhao Y., Bernstein P.S., Ge J., Jonasson F., Stefansson E., et al. (2006) CFH Y402H confers similar risk of soft drusen and both forms of advanced AMD. PLoS Med. 3:e5.[CrossRef][Medline]

15. Okamoto H., Umeda S., Obazawa M., Minami M., Noda T., Mizota A., Honda M., Tanaka M., Koyama R., Takagi I., et al. (2006) Complement factor H polymorphisms in Japanese population with age-related macular degeneration. Mol. Vis. 12:156–158.[Web of Science][Medline]

16. Jakobsdottir J., Conley Y.P., Weeks D.E., Mah T.S., Ferrell R.E., Gorin M.B. (2005) Susceptibility genes for age-related maculopathy on chromosome 10q26. Am. J. Hum. Genet. 77:389–407.[CrossRef][Web of Science][Medline]

17. Schmidt S., Hauser M.A., Scott W.K., Postel E.A., Agarwal A., Gallins P., Wong F., Chen Y.S., Spencer K., Schnetz-Boutaud N., et al. (2006) Cigarette smoking strongly modifies the association of LOC387715 and age-related macular degeneration. Am. J. Hum. Genet. 78:852–864.[CrossRef][Web of Science][Medline]

18. Ayyagari R., Zhang K., Hutchinson A., Yu Z., Swaroop A., Kakuk L.E., Seddon J.M., Bernstein P.S., Lewis R.A., Tammur J., et al. (2001) Evaluation of the ELOVL4 gene in patients with age-related macular degeneration. Ophthalmic Genet. 22:233–239.[CrossRef][Medline]

19. Fried L.P., Borhani N.O., Enright P., Furberg C.D., Gardin J.M., Kronmal R.A., Kuller L.H., Manolio T.A., Mittelmark M.B., Newman A., et al. (1991) The Cardiovascular Health Study: design and rationale. Ann. Epidemiol. 1:263–276.[Medline]

20. Age-Related Eye Disease Group Study. (1999) The Age-Related Eye Disease Study (AREDS): design implications. AREDS report no. 1. Control Clin. Trials 20:573–600.[CrossRef][Web of Science][Medline]

21. Klein R., Klein B.E., Marino E.K., Kuller L.H., Furberg C., Burke G.L., Hubbard L.D. (2003) Early age-related maculopathy in the cardiovascular health study. Ophthalmology 110:25–33.[CrossRef][Web of Science][Medline]

22. Age-Related Eye Disease Group Study. (2000) Risk factors associated with age-related macular degeneration. A case–control study in the age-related eye disease study: Age-Related Eye Disease Study Report No. 3. Ophthalmology 107:2224–2232.[CrossRef][Web of Science][Medline]

23. Valdes A.M. and Thomson G. (1997) Detecting disease-predisposing variants: the haplotype method. Am. J. Hum. Genet. 60:703–716.[Web of Science][Medline]

24. North B.V., Curtis D., Sham P.C. (2005) Application of logistic regression to case–control association studies involving two causative loci. Hum. Hered. 59:79–87.[CrossRef][Web of Science][Medline]

25. Seddon J.M., Santangelo S.L., Book K., Chong S., Cote J. (2003) A genome-wide scan for age-related macular degeneration provides evidence for linkage to several chromosomal regions. Am. J. Hum. Genet. 73:780–790.[CrossRef][Web of Science][Medline]

26. Majewski J., Schultz D.W., Weleber R.G., Schain M.B., Edwards A.O., Matise T.C., Acott T.S., Ott J., Klein M.L. (2003) Age-related macular degeneration—a genome scan in extended families. Am. J. Hum. Genet. 73:540–550.[CrossRef][Web of Science][Medline]

27. Iyengar S.K., Song D., Klein B.E., Klein R., Schick J.H., Humphrey J., Millard C., Liptak R., Russo K., Jun G., et al. (2004) Dissection of genome-wide scan data in extended families reveals a major locus and oligogenic susceptibility for age-related macular degeneration. Am. J. Hum. Genet. 74:20–39.[CrossRef][Web of Science][Medline]

28. Weeks D.E., Conley Y.P., Tsai H.J., Mah T.S., Rosenfeld P.J., Paul T.O., Eller A.W., Morse L.S., Dailey J.P., Ferrell R.E., et al. (2001) Age-related maculopathy: an expanded genome-wide scan with evidence of susceptibility loci within the 1q31 and 17q25 regions. Am. J. Ophthalmol. 132:682–692.[CrossRef][Web of Science][Medline]

29. Weeks D.E., Conley Y.P., Tsai H.J., Mah T.S., Schmidt S., Postel E.A., Agarwal A., Haines J.L., Pericak-Vance M.A., Rosenfeld P.J., et al. (2004) Age-related maculopathy: a genome-wide scan with continued evidence of susceptibility loci within the 1q31, 10q26, and 17q25 regions. Am. J. Hum. Genet. 75:174–189.[CrossRef][Web of Science][Medline]

30. Klein M.L., Schultz D.W., Edwards A., Matise T.C., Rust K., Berselli C.B., Trzupek K., Weleber R.G., Ott J., Wirtz M.K., et al. (1998) Age-related macular degeneration. Clinical features in a large family and linkage to chromosome 1q. Arch. Ophthalmol. 116:1082–1088.[Abstract/Free Full Text]

31. Kenealy S.J., Schmidt S., Agarwal A., Postel E.A., De La Paz M.A., Pericak-Vance M.A., Haines J.L. (2004) Linkage analysis for age-related macular degeneration supports a gene on chromosome 10q26. Mol. Vis. 10:57–61.[Web of Science][Medline]

32. Gauderman W.J. and Morrison J.M. QUANTO 1.1: a computer program for power and sample size calculations for genetic-epidemiology studies. http://hydra.usc.edu/gxe.

33. van Leeuwen R., Klaver C.C., Vingerling J.R., Hofman A., de Jong P.T. (2003) Epidemiology of age-related maculopathy: a review. Eur. J. Epidemiol. 18:845–854.[CrossRef][Web of Science][Medline]

34. Normand S.L. (1999) Meta-analysis: formulating, evaluating, combining, and reporting. Stat. Med. 18:321–359.[CrossRef][Web of Science][Medline]

35. Lohmueller K.E., Pearce C.L., Pike M., Lander E.S., Hirschhorn J.N. (2003) Meta-analysis of genetic association studies supports a contribution of common variants to susceptibility to common disease. Nat. Genet. 33:177–182.[CrossRef][Web of Science][Medline]

36. Walter S.D. (1975) The distribution of Levin's measure of attributable risk. Biometrika 62:371–37436.[Abstract/Free Full Text]

37. R Development Core Team. (2005) R: A Language and Environment for Statistical Computing(R Foundation for Statistical Computing, Vienna, Austria).

38. Clayton D.G. SNPHAP 1.3.1: a program for estimating frequencies of large haplotypes of SNPs. http://www-gene.cmr.ac.uk/clayton/software.

39. Klaver C.C., Kliffen M., van Duijn C.M., Hofman A., Cruts M., Grobbee D.E., van Broeckhoven C., de Jong P.T. (1998) Genetic association of apolipoprotein E with age-related macular degeneration. Am. J. Hum. Genet. 63:200–206.[CrossRef][Web of Science][Medline]

40. Schmidt S., Saunders A.M., De La Paz M.A., Postel E.A., Heinis R.M., Agarwal A., Scott W.K., Gilbert J.R., McDowell J.G., Bazyk A., et al. (2000) Association of the apolipoprotein E gene with age-related macular degeneration: possible effect modification by family history, age, and gender. Mol. Vis. 6:287–293.[Web of Science][Medline]

41. Schmidt S., Klaver C., Saunders A., Postel E., De La Paz M., Agarwal A., Small K., Udar N., Ong J., Chalukya M., et al. (2002) A pooled case–control study of the apolipoprotein E (APOE) gene in age-related maculopathy. Ophthalmic Genet. 23:209–223.[CrossRef][Medline]

42. Baird P.N., Guida E., Chu D.T., Vu H.T., Guymer R.H. (2004) The epsilon2 and epsilon4 alleles of the apolipoprotein gene are associated with age-related macular degeneration. Invest. Ophthalmol. Vis. Sci. 45:1311–1315.[Abstract/Free Full Text]

43. Zareparsi S., Reddick A.C., Branham K.E., Moore K.B., Jessup L., Thoms S., Smith-Wheelock M., Yashar B.M., Swaroop A. (2004) Association of apolipoprotein E alleles with susceptibility to age-related macular degeneration in a large cohort from a single center. Invest. Ophthalmol. Vis. Sci. 45:1306–1310.[Abstract/Free Full Text]

44. DerSimonian R. and Laird N. (1986) Meta-analysis in clinical trials. Control Clin. Trials 7:177–188.[CrossRef][Web of Science][Medline]

45. van Houwelingen H.C., Arends L.R., Stijnen T. (2002) Advanced methods in meta-analysis: multivariate approach and meta-regression. Stat. Med. 21:589–624.[CrossRef][Web of Science][Medline]

CiteULike    Connotea    Del.icio.us    What's this?

This article has been cited by other articles:

				P J Francis, S C Hamon, J Ott, R G Weleber, and M L Klein Polymorphisms in C2, CFB and C3 are associated with progression to advanced age related macular degeneration associated with visual loss J. Med. Genet., May 1, 2009; 46(5): 300 - 307. [Abstract] [Full Text] [PDF]

				J. J. Wang, E. Rochtchina, W. Smith, R. Klein, B. E. K. Klein, T. Joshi, T. A. Sivakumaran, S. Iyengar, and P. Mitchell Combined Effects of Complement Factor H Genotypes, Fish Consumption, and Inflammatory Markers on Long-Term Risk for Age-related Macular Degeneration in a Cohort Am. J. Epidemiol., March 1, 2009; 169(5): 633 - 641. [Abstract] [Full Text] [PDF]

				P. O. S. Tam, T. K. Ng, D. T. L. Liu, W. M. Chan, S. W. Y. Chiang, L. J. Chen, A. DeWan, J. Hoh, D. S. C. Lam, and C. P. Pang HTRA1 Variants in Exudative Age-Related Macular Degeneration and Interactions with Smoking and CFH Invest. Ophthalmol. Vis. Sci., June 1, 2008; 49(6): 2357 - 2365. [Abstract] [Full Text] [PDF]

				I. Kaur, S. Katta, A. Hussain, N. Hussain, A. Mathai, R. Narayanan, A. Hussain, R. K. Reddy, A. B. Majji, T. Das, et al. Variants in the 10q26 Gene Cluster (LOC387715 and HTRA1) Exhibit Enhanced Risk of Age-Related Macular Degeneration along with CFH in Indian Patients Invest. Ophthalmol. Vis. Sci., May 1, 2008; 49(5): 1771 - 1776. [Abstract] [Full Text] [PDF]

				I Droz, I Mantel, A Ambresin, M Faouzi, D F Schorderet, and F L Munier Genotype-phenotype correlation of age-related macular degeneration: influence of complement factor H polymorphism Br. J. Ophthalmol., April 1, 2008; 92(4): 513 - 517. [Abstract] [Full Text] [PDF]

				J. P. Ioannidis, P. Boffetta, J. Little, T. R O'Brien, A. G Uitterlinden, P. Vineis, D. J Balding, A. Chokkalingam, S. M Dolan, W D. Flanders, et al. Assessment of cumulative evidence on genetic associations: interim guidelines Int. J. Epidemiol., February 1, 2008; 37(1): 120 - 132. [Abstract] [Full Text] [PDF]

				K. Mori, P. L. Gehlbach, S. Kabasawa, I. Kawasaki, M. Oosaki, H. Iizuka, S. Katayama, T. Awata, and S. Yoneya Coding and Noncoding Variants in the CFH Gene and Cigarette Smoking Influence the Risk of Age-Related Macular Degeneration in a Japanese Population Invest. Ophthalmol. Vis. Sci., November 1, 2007; 48(11): 5315 - 5319. [Abstract] [Full Text] [PDF]

				A. Swaroop, K. E. Branham, W. Chen, and G. Abecasis Genetic susceptibility to age-related macular degeneration: a paradigm for dissecting complex disease traits Hum. Mol. Genet., October 15, 2007; 16(R2): R174 - R182. [Abstract] [Full Text] [PDF]

				K. L. Spencer, M. A. Hauser, L. M. Olson, N. Schnetz-Boutaud, W. K. Scott, S. Schmidt, P. Gallins, A. Agarwal, E. A. Postel, M. A. Pericak-Vance, et al. Haplotypes Spanning the Complement Factor H Gene Are Protective against Age-Related Macular Degeneration Invest. Ophthalmol. Vis. Sci., September 1, 2007; 48(9): 4277 - 4283. [Abstract] [Full Text] [PDF]

				C. L. Thompson, B. E.K. Klein, R. Klein, Z. Xu, J. Capriotti, T. Joshi, D. Leontiev, K. E. Lee, R. C. Elston, and S. K. Iyengar Complement factor H and hemicentin-1 in age-related macular degeneration and renal phenotypes Hum. Mol. Genet., September 1, 2007; 16(17): 2135 - 2148. [Abstract] [Full Text] [PDF]

				R. J. Ross, C. M. Bojanowski, J. J. Wang, E. Y. Chew, E. Rochtchina, F. L. Ferris III, P. Mitchell, C.-C. Chan, and J. Tuo The LOC387715 Polymorphism and Age-Related Macular Degeneration: Replication in Three Case-Control Samples Invest. Ophthalmol. Vis. Sci., March 1, 2007; 48(3): 1128 - 1132. [Abstract] [Full Text] [PDF]

				M. B. Gorin A Clinician's View of the Molecular Genetics of Age-Related Maculopathy Arch Ophthalmol, January 1, 2007; 125(1): 21 - 29. [Abstract] [Full Text] [PDF]

This Article Abstract FREE Full Text (PDF) Supplementary Data All Versions of this Article: 15/21/3206    most recent ddl396v1 Alert me when this article is cited Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Similar articles in ISI Web of Science Similar articles in PubMed Alert me to new issues of the journal Add to My Personal Archive Download to citation manager Search for citing articles in: ISI Web of Science (32) Request Permissions Google Scholar Articles by Conley, Y. P. Articles by Gorin, M. B. Search for Related Content PubMed PubMed Citation Articles by Conley, Y. P. Articles by Gorin, M. B. Social Bookmarking          What's this?

Online ISSN 1460-2083 - Print ISSN 0964-6906

Copyright © 2009 Oxford University Press

Oxford Journals Oxford University Press

• Site Map
• Privacy Policy
• Frequently Asked Questions

Other Oxford University Press sites: Oxford University Press Oxford Journals China Oxford Journals Japan American National Biography Booksellers' Information Service Children's Fiction and Poetry Children's Reference Corporate & Special Sales Dictionaries Dictionary of National Biography Digital Reference English Language Teaching Higher Education Textbooks Humanities International Education Unit Law Medicine Music Online Products Oxford English Dictionary Reference Rights and Permissions Science School Books Social Sciences Very Short Introductions World's Classics