Abstract
Purpose.:
To describe the association between refractive errors, ocular biometry, and age-related macular degeneration (AMD) in an Asian Malay population in Singapore.
Methods.:
A population-based study of 3280 Malay individuals aged 40 to 80 years was conducted in Singapore. Early- and late-AMD signs were graded from retinal photographs according to the Wisconsin grading system. Autorefraction, followed by subjective refraction, was performed to obtain spherical equivalent refraction (SER) in diopters, with emmetropia defined as SER −0.5 to +0.5 D, hyperopia as > +0.5 D, and myopia as < −0.5 D. Partial coherence laser interferometry was used to measure axial length, anterior chamber depth, and corneal curvature. The association between refractive status, ocular biometry and the prevalence of both early and late AMD were analyzed.
Results.:
Hyperopic refractive error (odds ratio [OR] 1.54; 95% confidence interval [CI] 1.00–2.36; compared with myopia, P = 0.05), shorter axial length (OR, 1.91; CI, 1.05–3.46, comparing 1st vs. 4th quartiles; P = 0.03), and steeper corneal curvature (OR, 1.93; CI, 1.16–3.20, comparing 1st vs. 4th quartiles, P = 0.01) were significantly associated with early AMD, after adjustment for age, sex, smoking, education, height, and systolic blood pressure. Each diopter increase in hyperopic refraction and each millimeter decrease in axial length was associated with an 8% (OR, 1.08; CI, 1.01–1.16; P = 0.03) and 29% (OR, 1.29; CI, 1.06–1.57; P = 0.01) increased risk of early AMD, respectively. No significant association was noted of refractive error and ocular biometry with late AMD.
Conclusions.:
Hyperopic refractive error and shorter axial length are associated with early AMD in Asian eyes.
Age-related macular degeneration (AMD) is the leading cause of blindness in elderly patients, not only in the Western world, but also increasingly so in Asian countries.
1 –9 Although many risk factors for AMD have been examined, only cigarette smoking and genetic markers (e.g., complement factor H) have been consistently found to be associated with AMD.
8 –18
Hyperopia has been reported to be associated with AMD in several case–control studies
19 –26 and a few population-based surveys,
12,27 –31 but the findings have been conflicting. The Blue Mountains Eye study (BMES) reported a weak cross-sectional association between hyperopia and early AMD,
27 but found no association with the 5-year incidence of either early or late AMD.
28 The Rotterdam Study reported an association between increasing hyperopia and 5-year incident AMD,
29 but the Beaver Dam Study found no significant association between baseline refractive status and either the 5- or 10-year incidence of early or late AMD.
30,31 Two other large studies, the Eye Disease Case–Control Study
24 and the Age-Related Eye Disease Study (AREDS),
25 found that subjects with hyperopia were 1.5 to 2.3 times more likely to have exudative AMD than were those with myopia, after adjustment for age and other risk factors. However, few of these studies have been in Asians, where refractive errors, particularly myopia, are more prevalent. There have been no population-based studies that have evaluated the association between ocular biometric factors, such as axial length, and AMD.
The purpose of this study was to evaluate the relationship of refractive error and ocular biometric parameters with the prevalence of AMD in a large population-based study of Malays in Singapore.
The Singapore Malay Eye Study (SiMES) was a population-based, cross-sectional study of adult Malays, conducted to assess prevalence, risk factors, and the public health impact of common age-related eye diseases in an urban Asian population. Detailed population selection and methodology are reported elsewhere.
5,10,32 –35 An age-stratified (by 10-year age group), random sample of the Malay population residing in 15 residential districts aged 40 to 80 years was drawn from a computer-generated random list of 16,069 Malay names provided by the Ministry of Home Affairs. In total 3280 individuals participated in the study (overall response rate, 78.7%). All examinations were conducted, after obtaining informed consent at the Singapore Eye Research Institute, a clinical research facility located centrally in Singapore. The study adhered to the Declaration of Helsinki, and ethics approval was obtained from the Singapore Eye Research Institute Institutional Review Board.
At the study clinic, participants underwent an extensive and standardized examination procedure, which included visual acuity testing, refraction, a detailed clinical slit lamp and fundus examination before and after pupil dilation. Color digital fundus photography of the macular region was also performed.
Presenting distance visual acuity was monocularly measured by using a logarithm of the minimum angle of resolution (logMAR) number chart (Lighthouse International, New York, NY) at a distance of 4 m, with the participant wearing habitual optical correction (spectacles or contact lenses), if any. An automatic lens meter (Auto-Lens Meter, model LM990; Nidek Co., Ltd., Gamagori, Japan) was used to measure the current spectacles. Objective refraction was initially measured with an autorefractor (Canon RK-5 Auto Ref-Keratometer; Canon Inc. Ltd., Tokyo, Japan). Final refraction was determined by subjective refraction by trained and certified study optometrists, with autorefraction readings as the starting point, and was recorded in logMAR scores.
34 The spherical equivalent refraction (SER) of each eye, measured in diopters (D), was calculated with the spherical diopteric power plus half the cylindrical diopteric power. Participants with aphakia or pseudophakia in one or both eyes (
n = 171) were excluded. Emmetropia was defined as SER between ≤ +0.5 and ≥ −0.5 D. Myopia was defined as SER of < −0.5 D and hyperopia as SER of > +0.5 D.
Partial coherence laser interferometry with an ocular biometer (IOLMaster; Carl Zeiss; Meditec AG, Jena, Germany) was performed to measure axial length (AL), anterior chamber depth (ACD), and corneal curvature (CC) in both the horizontal and vertical meridians.
35
Assessment of AMD in the study population has been described in published articles.
5,10 In brief, fundus photographs were reviewed by the same graders who performed AMD grading for the BMES and other studies, including clinical trials at the Centre for Vision Research, University of Sydney, in a standardized manner, according to a modification of the Wisconsin Age-Related Maculopathy grading system.
36
Drusen type was classified on the basis of the size and the sharpness of the edges. Drusen were classified as hard or soft; then, soft drusen were divided into distinct and indistinct soft drusen. Reticular drusen was defined as having ill-defined networks of broad, interlacing ribbons within the grading grid. Retinal pigmentary abnormalities were graded as hypo- and hyperpigmentation. Early AMD was defined as either large (>125 μm in diameter), indistinct soft or reticular drusen or large, distinct soft drusen with retinal pigment epithelium (RPE) abnormalities according to the definition used in the BMES.
3
Neovascular AMD lesions were defined as the presence of RPE detachment; neurosensory detachment; subretinal or sub-RPE hemorrhages; or intraretinal, subretinal, or sub-RPE scar tissue associated with choroidal neovascularization. Subretinal hemorrhages or hard exudates within the macular area were also considered to be signs of neovascular AMD if other retinal vascular diseases were ruled out as alternative causes. Geographic atrophy was defined by the presence of visible choroidal vessels and a discrete atrophic area with a sharp border and a diameter of ≥175 μm. Late AMD was defined as the presence of either neovascular AMD or geographic atrophy.
The relationship between AMD grade and refraction, AL, ACD, and CC was analyzed, after adjustment for age and sex. Data from the right and left eyes were analyzed together in multiple logistic regression with the generalized estimating equation (GEE) model, which adjusts for correlations between the two eyes in a single person.
37 –39 Multivariate analysis was performed after adjustment for age, sex, smoking, education, height, and systolic blood pressure (SPSS, ver. 13; SPSS Inc., Chicago, IL).
P < 0.05 indicated statistical significance.
In total, 3280 participants were examined (overall response rate 78.7%), of whom 3265 (99.5%) had sufficient quality photographs for grading AMD signs. After 171 subjects with aphakia or pseudophakia in one or both eyes and 24 with missing data on refraction or lens status were excluded, our study population included 3070 participants. Characteristics of participants by refractive status are shown in
Table 1. The mean age (SD) of participants was 58.1 ± 10.88 years and 1477 (48.1%) were men. The total number and age-standardized prevalence of early and late AMD cases in our study population was 158 (3.7%; 95% confidence interval [CI] 3.0%–4.4%) and 21 (0.35%; CI, 0.21%–0.63%), respectively.
Table 2 shows the age–sex- and multivariate-adjusted association of refractive errors and ocular biometric parameters and early AMD. After we adjusted for known AMD risk factors (age, sex, education, height, systolic blood pressure, and smoking), a statistically significant association was found between hyperopia and early AMD (multivariate adjusted odds ratio [OR], 1.54;
P = 0.05). Similarly, shorter AL (OR, 1.91;
P = 0.03) and steeper CC (OR, 1.93,
P = 0.01) were significantly associated with early AMD. No significant association was found between ACD and early AMD. The multivariate-adjusted association of refractive errors and ocular biometric parameters with early AMD was also evaluated stratified by age (<65 years and ≥65 years as the second strata). Shorter AL (OR, 2.38; CI, 1.02–5.56;
P = 0.05) and steeper CC (OR, 2.49; CI, 1.10–5.66;
P = 0.03) remained significantly associated with early AMD in subjects <65 years of age but was weaker and not statistically significant in subjects ≥65 years (OR, 1.64; CI, 0.74–3.60;
P = 0.22 for AL and OR 1.60; CI, 0.85–3.01;
P = 0.15 for CC). However, age was not a significant effect modifier.
Table 3 shows the multivariate-adjusted ORs of specific early AMD lesions (any drusen, >63 μm; large drusen, >125 μm; and pigmentary abnormalities) associated with refractive errors. Subjects with hyperopia were 1.8 times (CI, 1.57–2.26;
P < 0.001) more likely to have any drusen, 1.7 times (CI, 1.29–2.23;
P < 0.001) to have any large drusen, and 1.9 times (CI, 1.42–2.60;
P < 0.001) more likely to have pigmentary abnormalities than were subjects with myopia. Similar associations were seen with the ocular biometric parameters.
Finally,
Table 4 shows a significant association between increasingly hyperopic SER and decreasing AL and AMD signs. For each diopter positive increase in the SER, there was an 8% increased risk of AMD (OR, 1.08;
P = 0.03), after adjustment for age, sex, education, height, systolic blood pressure, and smoking. For each millimeter decrease in AL, there was a 29% increased risk of early AMD (OR, 1.29;
P = 0.01). Late AMD was not associated with increasing SE.
No significant association or trend was found between late AMD and hyperopic refractive error (adjusted OR, 1.32; CI, 0.44–3.96, compared with myopia; P = 0.62) or ocular biometric parameters (data not shown).
This population-based study of Asian Malays in Singapore showed a strong association of hyperopic refractive error and shorter AL with early AMD. Subjects with hyperopic refractive errors were 1.5 times more likely to have early AMD than were those with myopia, even after adjustment for AMD risk factors, including age, sex, smoking, education, height, and systolic blood pressure. Each diopter increase in SER was associated with an 8% increased odds of having early AMD. Consistent with this, we demonstrate strong associations between shorter AL and early AMD, with each milliliter decrease in AL associated with a 29% increased odds of early AMD.
Although investigators in several population-based and case–control studies have examined the association between refractive error and AMD, the findings have remained inconsistent and inconclusive. The results of the The Rotterdam Study
29 the Age-Related Eye Disease Study,
25 the French DMLA Study,
19 and the Beijing Eye Study
12 showed an association between hyperopia and AMD, whereas the Beaver Dam study did not show an association between refractive errors and either the 5-
30 or 10-year incidence of AMD.
31 The BMES described a weak cross-sectional association between AMD and hyperopia.
27 However, a subsequent report on the 5-year incidence of AMD did not support the previous observations.
28 Among Asian studies, the Beijing Eye Study found that hyperopic refractive error, apart from age, was the single most important risk factor for early AMD in adult Chinese.
12 Furthermore, in a second report on the characteristics of highly myopic eyes,
40 the highly myopic eyes had a significantly lower prevalence of both early (OR, 3.0; CI, 1.21–7.51;
P = 0.03) and late (OR, 6.33;
P = 0.001) AMD, compared with eyes that did not have high myopia.
In contrast to population-based studies, clinical and case–control studies
19 –25 have reported stronger associations between hyperopia and AMD, particularly exudative AMD. In the Eye Disease Case–Control Study
24 and the AREDS,
25 after adjustment for age and other risk factors, persons with hyperopia were 1.5 to 2.3 times more likely to have exudative AMD than were those who were myopic. Sandberg et al.
20 compared the refractive error of 198 patients with unilateral exudative AMD with the refractive error of 129 patients with bilateral dry AMD. In a comparison of the better eyes of the two groups (better eye was defined as the fellow eye in unilateral cases and the better-seeing eye in bilateral cases), patients with the unilateral exudative AMD had an average spherical equivalent that was 1.0 D more hyperopic than that of patients with the bilateral dry AMD (
P < 0.001). Patients with a refractive error of +0.75 D or greater were more likely to have exudative AMD than were patients with other refractive errors (OR, 2.40;
P < 0.001).
Our findings support an association between hyperopic refractive error and early AMD. More important, our study is the first population-based study to report that subjects with shorter AL and steeper curvature of the cornea are more likely to have early signs of AMD. This association was independent of education level and height. It has been argued that the association of hyperopia and AMD can be confounded by the presence of nuclear cataract. However, the strong correlation of short AL with early AMD validates the link between hyperopic refraction and AMD, independent of the cataract-related myopic shift. The risk estimate of 8% per diopter of progress toward hyperopia and 29% toward each millimeter decrease in AL is considerable, suggesting that refraction and ocular biometric parameters have a plausible role in the pathophysiology of AMD. In view of these findings, older subjects with short AL (≤22.8 mm) and hyperopia may need more careful monitoring for the development of AMD.
Case–control studies have examined AL and AMD without finding any relationship.
41,42 For example, a Norwegian cohort of the EUREYE multicenter community point prevalence study of ARM also found no association between AMD and AL.
43 More recently, Tao and Jonas
26 reported a clinical study examining 379 patients treated with an antivascular endothelial growth factor drug for exudative AMD and compared them with a control group of 191 patients undergoing surgery for age-related cataract, but without ophthalmoscopic signs of exudative AMD. The patients with AMD had a significantly shorter AL (23.31 ± 0.75 vs. 24.20 ± 1.56 mm;
P < 0.001) and were significantly more hyperopic (0.65 ± 2.14 vs. −1.71 ± 4.57 D;
P < 0.001) than were the patients with cataract. Although we found a statistically significant association of hyperopia and biometric parameters only in early AMD in our study and none in late AMD, we believe that this could have resulted from the small number of subjects with late AMD.
A biologically plausible explanation for these observations is not apparent. Boker et al.
21 hypothesized that the association between hyperopia and AMD is related to a reduction in choroidal blood flow in eyes with shorter AL, which could then predispose these eyes to the development of choroidal neovascularization. A hemodynamic model of the pathogenesis of AMD postulated that a degenerative process begins with decreased compliance of the sclera caused by lipoid infiltration.
44 It is therefore hypothesized that shorter, thicker hyperopic eyes, with higher scleral rigidity, cause an increase in resistance of the choroidal venous outflow. An increased choroidal resistance in AMD cases compared with sex- and age-matched controls has been documented by laser Doppler flow measurements
45 in both histologic and in vivo studies. It is possible that the decreased flow prevents easy exchange of nutrients and metabolic products across the RPE and results in drusen formation and thickening of Bruch's membrane. Another speculation is that poorer cooling of the retina by an impaired choroidal blood flow leads to a higher susceptibility to oxidative stress, in a thicker hyperopic eye with a higher demand for oxygen and nutrients.
29 A further possible explanation is a genetic link between the two conditions.
The strengths of this study are the representativeness of the sample of the general population; the high response rate; and the standardized protocol, including the photographic documentation of the macula; and ocular biometric measurements. The limitations include the relatively few cases of late AMD, which decreases the power of the study to identify significant risk factors. Also, being a cross-sectional study, it could not answer temporal questions that a longitudinal study with sufficient power could resolve.
In conclusion, our population-based study in Asian Malays in Singapore demonstrated a significant association between hyperopic refraction and shorter AL and early AMD. These data offer further clues to the pathogenesis of a major blinding condition.
Supported by National Medical Research Council Grants 0796/2003, 0863/2004, and CSI/0002/2005, and Biomedical Research Council Grant 501/1/25-5, with additional support from the Singapore Tissue Network and the Ministry of Health, Singapore.