Smith W, Mitchell P, Leeder SR, Wang JJ. Plasma Fibrinogen Levels, Other Cardiovascular Risk Factors, and Age-Related MaculopathyThe Blue Mountains Eye Study. Arch Ophthalmol. 1998;116(5):583-587. doi:10.1001/archopht.116.5.583
To assess the relationship between stages of age-related maculopathy (ARM) and cardiovascular disease and cardiovascular disease risk factors, including serum lipid and plasma fibrinogen levels, smoking, cardiovascular events, systemic hypertension, diabetes mellitus, and obesity.
A cross-sectional study of 3654 subjects from a defined geographic area identified subjects with late age-related macular degeneration (ARMD) and early ARM from the masked grading of retinal photographs. The history, physical examination findings, and fasting blood samples provided data on possible risk factors. Logistic regression, adjusting for age, sex, and possible confounders, and 2-way analysis of variance were used to assess associations.
The only factors significantly associated with ARM included the 2 established risk factors, smoking and family history of ARMD (odds ratios, 4.1 and 4.2, for late ARMD, respectively), and the 2 variables, body mass index (odds ratio, 1.78 for obese compared with normal body mass index for early ARM) and plasma fibrinogen level (odds ratio, 6.7 for a fibrinogen level of >4.5 g/L [highest quartile] compared with a fibrinogen level of <3.4 g/L [lowest quartile] for late ARMD).
These findings support the concepts that associations exist between plasma fibrinogen levels and late ARMD, a body mass index outside the normal range, and early ARM, and between the family history and smoking and any ARM. We found no other significant associations with any history of cardiovascular disease or other risk factors for cardiovascular disease.
EXCEPT FOR smoking,1- 7 several cardiovascular diseases and risk factors have been found to have statistically significant associations with age-related maculopathy (ARM). These include systemic hypertension,8- 10 a previous diagnosis of vascular events,4,8,10,11 high serum cholesterol levels,2,12 and body mass index (BMI).12,13 Few of the cardiovascular associations found in individual studies have been consistently reproduced, however. These conflicting findings could result from measurement error in the exposure and outcome variables; selection bias, ie, differences in the systematic recruitment of cases and noncases or exposed and nonexposed subjects, leading to a bias in estimating associations; nonadjustment or incomplete adjustment for confounding variables because either the confounder was not measured or it was measured with error; no true associations and the observed associations are due to random noise; or nonuniform and inconsistent relationships across populations with different genetic makeups that may be exposed to different environmental risk factors.
The difficulties in establishing a causal link between a chronic disease outcome and a potential risk factor are magnified when the disease becomes apparent only late in life, as with ARM. Additional problems in these circumstances include a possible recall bias, a long lead time, and survivor cohort effects. Repeated findings of the same risk factors in well-designed studies conducted in different populations are necessary to provide compelling evidence of a real association between ARM and potential risk factors.
The aims of this article are to investigate associations between stages of ARM and the following potential cardiovascular risk factors, after adjusting for the effects of confounders such as age, sex, tobacco smoking, and a family history of ARM:
previous diagnosis of cardiovascular disease (angina, acute myocardial infarction, stroke, hypertension),
potentially modifiable personal characteristics (BMI and systolic and diastolic blood pressures), and
blood factors (fasting serum glucose, creatinine, and lipid levels and coagulation factors, including plasma fibrinogen).
The Blue Mountains Eye Study is a population-based survey of vision and common eye diseases in an urban middle-aged and older population (aged ≥49 years) residing in 2 postal code areas in the Blue Mountains region of Sydney, Australia. This area has a stable and homogeneous population. Details of the survey methods have been described previously.7,14,15 Of the 4433 eligible residents in this area, 3654 (82.4%) gave informed consent and were examined between 1992 and 1994.
A questionnaire covering the medical history was administered, and subjects underwent a detailed eye examination that included stereoscopic macular photographs and a subjective refraction.14,15 Retinal photographs of at least 1 eye were taken of 3582 participants (98.0%).
The Wisconsin Age-Related Maculopathy Grading System, developed by Klein et al16 and incorporated in an international classification,17 was used to grade individual ARM lesions. A grid with a radius of 3000 µm, centered on the fovea on 1 of the pair of stereo photographs, was used. Individual lesion definitions closely followed descriptions of the Wisconsin Age-Related Maculopathy Grading System, as outlined previously.14 All photographs were graded by 1 or both of 2 graders, with queries and discrepancies adjudicated by the chief investigator (P.M.). Intergrader and intragrader reliability was assessed on a random subsample of gradable eyes and was found to be high.14
The 2 late stages of ARM have been termed "age-related macular degeneration" (late ARMD) by the International ARM Epidemiological Study Group.17 The first stage, "neovascular" ("wet") ARMD, includes serous or hemorrhagic detachment of the retinal pigment epithelium or sensory retina, the presence of subretinal or subretinal pigment epithelial hemorrhages, or subretinal fibrous scarring. Cases with minimal subretinal fibrosis and widespread surrounding atrophy were also classified as neovascular because this appearance was considered to indicate previous neovascular ARMD.17 The second stage, "geographic atrophy" ("dry" ARMD), was defined as a discrete area of retinal depigmentation, at least 175 µm in diameter, characterized by a sharp border and visible choroidal vessels.17 We identified 72 cases of late ARMD, including 50 cases of neovascular ARMD and 22 cases of geographic atrophy.
Early ARM was defined as the presence of either indistinct, soft, or reticular drusen within the area of the macula covered by the grid or both distinct, soft drusen within the grid and retinal pigmentary abnormalities in the absence of late ARMD in either eye. This definition of early ARM as an outcome measure was previously reported in the Beaver Dam, Wisconsin,18 and Blue Mountains populations14 and was used to assess relationships between ARM and many potential risk factors in the Beaver Dam Eye Study.1,19- 21 We identified 240 cases of early ARM.
A history of cardiovascular events and cardiovascular risk factors was elicited using the question "Has a doctor ever said that you have had any of the following conditions: high blood pressure, angina, heart attack, stroke, high cholesterol, or diabetes?" A smoking history and a family history of ARMD were ascertained using an interviewer-administered questionnaire described previously.7 Systolic and diastolic blood pressures, height, and weight were measured at the clinic visit by a trained observer, and the BMI was calculated using the equation, BMI equals weight in kilograms divided by height in meters squared. Fasting blood specimens were drawn from subjects, and assays included serum glucose, cholesterol, high-density lipoprotein cholesterol, and triglyceride levels, activated partial thromboplastin time, prothrombin time, and fibrinogen level.
A statistical software package (Statistical Analysis System, SAS Institute, Cary, NC) was used for tabulations and statistical analyses, including a 2-way analysis of variance (ANOVA) (generalized linear models) and logistic regression.22 For logistic regression, each ARM stage was used as a dependent variable, and the potential risk factor, together with age (continuously) and sex, were independent variables. For continuously distributed variables (smoking pack-years), a 2-way ANOVA was used with potential risk factors as dependent variables and ARM and 10-year age groups as class variables. Odds ratios and 95% confidence intervals (CIs) are presented.
All associations between potential risk factors and late ARMD and early ARM are presented in Table 1 and Table 2 in the form of odds ratios, adjusted for age, sex, smoking, and a family history of ARM. Two-way ANOVA, including categories for the disease state (separately for each of the 3 categories described earlier) and 10-year age groups, were also calculated using generalized linear models for BMI and each of the blood test results listed in Tables 1 and 2. As shown in the tables, the only factors statistically significantly associated with late ARMD were the 3 established risk factors: smoking, a family history of ARMD, and the plasma fibrinogen level.
The plasma fibrinogen level was significantly associated with late ARMD, with an increased risk of late ARMD with increasing fibrinogen level (odds ratio, 1.45; 95% CI, 1.18-1.80, for each 1.0-g/L increase in fibrinogen, or an odds ratio of 6.7 for a fibrinogen level of >4.5 g/L [highest quartile] compared with 1 of <3.4 g/L [lowest quartile]). The 2-way ANOVA also showed significant differences in population means for plasma fibrinogen levels between subjects with ARMD (4.68 g/L) and those without ARMD (4.12 g/L), adjusting for age (F1=16.37; P<.001). Because the mean (±SD) plasma fibrinogen level was higher among people with a history of arthritis (4.19 [±1.12] g/L) than among those without arthritis (3.98 [±0.99] g/L; F1=14.25; P=.002), after adjusting for age, the association with late ARMD was reassessed after an additional adjustment for arthritis. The odds ratio was again almost unchanged at 1.49 (95% CI, 1.21-1.85). There was no association between the plasma fibrinogen level and early ARM. No other blood factor was significantly associated with late ARMD or early ARM in either the logistic regression models or the 2-way ANOVA.
Having a BMI either lower or higher than the accepted normal range (20-25) was associated with a significantly increased risk of early ARM. Comparable increases in risk were apparent for late ARMD but did not reach statistical significance (Table 1). An elevated plasma fibrinogen level was associated with an increasing BMI, with the mean fibrinogen level significantly associated with the BMI category (F3=10.46; P<.001). Neither the BMI nor fibrinogen associations with maculopathy changed appreciably, however, when both were included in the multivariate model (results not shown). Although the fibrinogen level was significantly associated with angina (F1=17.03; P<.001) after adjusting for age, it was not significantly associated with a history of either stroke (F1=0.68; P=.41) or acute myocardial infarction (F1=.01; P=.94). Including angina in the logistic model associating fibrinogen level and ARMD did not alter the relationship.
A previously unreported finding from this study is the significant association between increasing plasma fibrinogen levels and late ARMD, although the Eye Disease Case-Control Study Group2 found a nonsignificant increased risk of ARMD associated with increasing plasma fibrinogen levels. We are unable to explain this finding in isolation. Many diseases are associated with elevated fibrinogen levels, including diabetes mellitus,23,24 arthritis,25 and cardiovascular risk factors.26 The plasma fibrinogen level has been found to predict subsequent cardiovascular events,27 but its relation to early carotid atherosclerosis is unclear.28 We and others29 have found no association between the presence of diabetes mellitus and ARMD.
We found no significant associations between cardiovascular diagnoses (hypertension, stroke, angina, and acute myocardial infarction) and ARM. Earlier studies are divided on the association between cardiovascular disease and ARM. Some studies have variously found an increased risk of ARM with a past cardiovascular event,4,8,10,11 systemic hypertension,8- 10 increased serum cholesterol levels,2,12 and carotid atherosclerosis,11 although other studies have found no associations with vascular events,6,12,30 systemic hypertension,6,11,12,30,31 or serum cholesterol levels.6,31,32 Our failure to find any significant association between cardiovascular disease and ARMD may be due to not only statistical power constraints but also perhaps a survivor cohort effect. That is, those who survive to an age when ARMD is likely to develop have survived cardiovascular disease; those who did not survive a cardiovascular event may have been much more likely to have ARMD if they had survived. A final possibility is that there is no true association between cardiovascular disease and ARMD.
Whether elevated plasma fibrinogen levels predispose to ARMD or vice versa cannot be determined from our cross-sectional data. An indicator that ARMD may be a marker for a systemic process that results in elevated fibrinogen levels is the finding that less severe maculopathy was not associated with elevated fibrinogen levels (early ARM odds ratio, 0.99). Thus, no progressive gradient with increasing disease severity was evident, as seen in the associations between ARMD and smoking7 and family history. It is also possible, however, that the association with the plasma fibrinogen level is due to chance, as we have examined many possible associations in our analysis. This possibility is less likely, given the Eye Disease Case-Control Study findings.2
Apart from fibrinogen, none of the other blood factors investigated was associated with any form of ARM. In particular, we found no association between elevated serum cholesterol levels and late ARMD after adjusting for age and sex (F1=1.1; P=.29), in contrast with findings of an increased risk from the Eye Disease Case-Control Study2 and a decreased risk from the Beaver Dam Eye Study,12 but in keeping with reports from other studies.6,31,32 A family history of ARMD showed an increasing association with increasing evidence of maculopathy.
The BMI was significantly associated with early ARM, with both underweight and overweight subjects having a significantly increased risk of early ARM. Although odds ratios were similar for associations with late ARMD, they did not reach statistical significance. Hirvelä et al13 found similar increases in risk with the BMI, but only among men, whereas Klein et al12 found associations between the BMI and retinal pigmentary abnormalities among women but not men. The association found between ARM and the BMI is difficult to explain in terms of cardiovascular risk, as a low BMI conferred an increased risk for maculopathy almost equal to that of obesity. The association with a low BMI could be a statistical artifact resulting from an inadequate adjustment for age or coexisting disease among frail elderly persons, who are more likely to have maculopathy. Although the BMI and fibrinogen levels were associated, neither association with ARMD changed substantially when adjusting for the other. The risk factors identified in this article are based on cross-sectional associations. More convincing evidence for causal associations can be obtained only from follow-up studies, several of which are in progress.
In conclusion, our results support a positive association between plasma fibrinogen levels and late ARMD, a BMI outside the normal range, and early ARM and between family history and smoking and any ARM. We found no other significant associations with any history of cardiovascular disease or other risk factors for cardiovascular disease.
Accepted for publication January 9, 1998.
This study was supported by the Australian Department of Health and Family Services, Canberra, Australian Capital Territory, and the Save Sight Institute, University of Sydney, New South Wales, Australia.
Reprints: Paul Mitchell, MD, FRACO, FRCOphth, Department of Ophthalmology, University of Sydney, Hawkesbury Road, Westmead, New South Wales, 2145, Australia (e-mail: firstname.lastname@example.org).