Visual Impairment, Age-Related Cataract, and Mortality FREE

VISUAL IMPAIRMENT broadly impacts the ability of people to function and to remain independent as they grow older.^{1 - 2} Impaired vision also severely affects quality of life,³ may be associated with depression,⁴ can affect a person's self-ranking of his or her health,⁵ and increases the need for community support.⁶

Limited data are available to assess whether visual impairment is associated with survival of older persons. Only 3 previous reports have examined this relationship. In a small population in Melton Mobray, Leicestershire (England)(N = 469)⁷ and in the Study of the Well-Being of Older People in Cleveland (Ohio) (N = 1408),⁸ persons with visual impairment had a moderate increased risk of death, a difference that was not statistically significant after adjustment. In the Beaver Dam(Wisconsin) population (N = 4926),⁹ visual impairment was statistically significantly associated with a 57% increase in mortality after adjusting for age and sex. However, the association became nonsignificant after adjusting for multiple systemic characteristics (eg, history of cancer, cardiovascular disease, or diabetes; smoking; diastolic blood pressure) found to be associated with mortality.⁹

Several studies have reported that cataract is associated with increased mortality.^{9 - 13} In the Framingham Eye Study (Mass) (N = 1945), the 3 main cataract subtypes(nuclear, cortical, and posterior subcapsular) were significantly associated with increased mortality after adjusting for age, sex, and presence of diabetes.¹⁰ Later reports have also supported a relationship between the presence of age-related cataract or cataract surgery and mortality in nondiabetic populations,^{11 - 13} although a limited number of confounding variables were adjusted for in these studies. In the Beaver Dam Eye Study, more severe nuclear cataract was significantly associated with reduced survival in people without diabetes after adjusting for multiple systemic characteristics.⁹

Any relationship between visual impairment and mortality could be mediated via cataract as an intermediary or surrogate. Since previous studies have not addressed this question to our knowledge, we aimed to assess whether an association exists between visual impairment and increased mortality and, if so, whether this is independent of the association between age-related cataract and mortality.

The Blue Mountains Eye Study is a population-based cohort study of vision and common eye diseases in residents aged 49 years or older living in a defined area west of Sydney. The population is representative of Australia's ethnic mix and measures of socioeconomic status but is slightly older than the New South Wales (Australia) state average.¹⁴ This research was approved by the Western Sydney Area Human Ethics Committee and written informed consent was obtained from all 3654 participants. Baseline examinations were performed during 1992-1994, with a participation rate of 82.4% overall and 87.9% after excluding people who died or left the area during the examination period.

To identify persons who had died since the baseline examination, demographic information was used to cross-match the 3654 participants with the Australian National Death Index (NDI) data in late 1999. The NDI data had at that time been updated to June 30, 1999, in most Australian states. A probabilistic record linkage package was used, adopting a multiple-pass procedure in which both data sets were grouped based on different characteristics (eg, date of birth, name, sex) each time. Matches were divided into exact and nonexact. All nonexact matched records were examined manually and accepted if there was only 1 nonexact matched characteristic that was noncritical. Information provided by family members during follow-up was also included if the participant was reported to have died on or before June 30, 1999. This increased the number of deaths by 39 persons. Among these additional deaths, most (71%) died in the year prior to June 30, 1999, so that a lag in notification to the NDI was presumed to be the reason this group had not been listed.

During the baseline eye examination, presenting visual acuity was measured using a logMAR chart while the participant wore current distance glasses. Best-corrected visual acuity was then measured after a detailed subjective refraction.¹⁵ For each eye, visual acuity was recorded as number of letters read correctly, from 0 (worse than 20/200) to 70 (20/10) letters. Visual impairment was defined as a visual acuity of 20/40 or worse in the better eye. Presenting visual impairment was defined as impaired vision using current spectacles, if worn, and visual impairment after best correction was defined as impaired vision after best refractive correction. Those whose presenting visual impairment improved after refraction were considered to have "correctable" visual impairment.

Cataract and past cataract surgery were diagnosed from grading of slitlamp and retroillumination lens photographs¹⁶ using the Wisconsin Cataract Grading System.¹⁷ Nuclear opacities were graded by comparison with 4 standard slitlamp photographs and nuclear cataract was defined as an opacity level of 4 or 5. Cortical cataract was defined as opacity involving greater than or equal to 5% of the lens area, and posterior subcapsular cataract was defined as the presence of any such opacity. There were 71 persons without retinal photographs of both eyes, 102 without photographs for grading cortical or posterior subcapsular cataract, and 1045 persons without gradable photographs for assessing nuclear cataract.¹⁶

Diabetic retinopathy or other retinal diseases were diagnosed during a masked grading of the stereo retinal photographs. Grading of age-related maculopathy photographs followed the Wisconsin Age-Related Maculopathy Grading System.¹⁸ Open-angle glaucoma was defined as reported previously.¹⁹ Spherical equivalent refraction, measured in diopters, was calculated using the spherical diopter power plus half the cylindrical power. Myopia was defined when the mean spherical equivalent of the 2 eyes was greater than or equal to −1 diopter.

Baseline characteristics were collected during a face-to-face interview using a standard questionnaire. A history of systemic diseases and a detailed history of smoking (including pack-years), alcohol consumption, and whether participants walked regularly for exercise were recorded. Global self-rated health was assessed by asking: "For someone of your age, how would you rate your overall health; would you say it is excellent, good, fair or poor?" Difficulty in walking or use of a cane, walker, or wheelchair at the clinic visit was recorded. Height, weight, and blood pressure were measured and body mass index calculated.

Age- and sex-standardized mortality rates were directly standardized to the whole study population. Age- and sex-adjusted survival curves were plotted using S-Plus (Version 4.5, MathSoft Inc, Seattle, Wash). Survival analyses, including the log-rank test and the Cox proportional hazards regression analysis were performed using SAS software (Version 6.12, SAS Institute Inc, Cary, NC). Survival time was calculated as days survived since baseline. Systemic factors known to be associated with mortality were screened for associations with mortality using age- and sex-adjusted Cox proportional hazards regression analysis. A basic multivariate model was constructed containing factors significantly associated with mortality after simultaneous multiple adjustment. Each eye condition was assessed separately using age- and sex-adjusted models and then the basic multivariate model.

By June 30, 1999, 604 participants had died, including 565 persons confirmed dead by matching with NDI data (459 exact and 106 nonexact matches on only 1 noncritical variable) and 39 persons confirmed dead by reports from family members.

Age- and sex-adjusted survival curves by presence of visual impairment or nuclear cataract at baseline are shown in Figure 1 and Figure 2. The age- and sex-standardized 7-year cumulative survival rate was 74% among persons with visual impairment after best correction at baseline compared with 84% among persons without visual impairment. Corresponding standardized survival rates were 78% among persons with nuclear cataract at baseline and 86% in those without nuclear cataract.

After adjusting for age and sex, persons with presenting visual impairment(risk ratio [RR], 1.7; 95% confidence interval [CI], 1.4-2.0) or visual impairment after best correction (RR, 1.8; 95% CI, 1.4-2.3) at baseline had a 70% to 80% higher mortality risk than persons without visual impairment. Persons with nuclear cataract (RR, 1.4; 95% CI, 1.1-1.8), posterior subcapsular cataract(RR, 1.4; 95% CI, 1.1-1.8), or cortical cataract (RR, 1.3; 95% CI, 1.0-1.5) at baseline had a 30% to 40% higher mortality risk than those without cataract. The presence of diabetic retinopathy (RR, 2.2; 95% CI, 1.5-3.3) or myopia(RR, 1.4; 95% CI, 1.2-1.8) was also associated with a higher mortality risk. However, presence of age-related maculopathy or glaucoma was not associated with mortality.

Systemic factors that remained statistically significant in the basic multivariate model are presented in Table 1. Fair (RR, 2.2) or poor (RR, 3.2) self-rated health, male sex(RR, 1.7), presence of walking disability (RR, 1.7), current smoking (RR, 2.0), and low body mass index (<20 [RR, 1.9]), were associated with a greater than 50% increased mortality risk.

The association between each eye condition and mortality after adjusting for all significant systemic factors is presented in Table 2. Presence of visual impairment at baseline was independently associated with a 50% to 70% increased mortality risk. Any type of age-related cataract (nuclear, posterior subcapsular, or cortical) was statistically significantly associated with a 30% to 50% increased mortality risk. Myopia at baseline was also associated with a 50% increased mortality risk. After excluding diabetic subjects, the association between visual impairment after best correction and mortality did not change, and the strength of the association between cataract and mortality was slightly reduced (Table 2).

To assess whether the association found between visual impairment and mortality was independent of the association with cataract, we included visual impairment and each type of cataract in the same multivariate model (Table 3). This analysis suggested that the associations found between either visual impairment or cataract and mortality were independent of each other. No significant interaction was found between visual impairment and each type of cataract.

To assess whether the association between myopia and mortality was caused by a myopic shift in refraction associated with developing nuclear cataract, we included nuclear cataract and myopia in the same multivariate model. In this analysis, the hazard RR for persons with myopia weakened from an RR of 1.45 (95% CI, 1.1-1.8) to an RR of 1.35 (95% CI, 1.0-1.8).

We further assessed the association between visual impairment and mortality using 3 categories of visual impairment: none, correctable visual impairment, and visual impairment after best correction, in the same basic multivariate Cox proportional hazards model. Compared with no visual impairment, the presence of correctable visual impairment at baseline was independently associated with a 40% increased mortality risk, and visual impairment after best correction at baseline was associated with an 80% increased mortality risk (results not shown).

The primary causes of bilateral visual impairment among persons with visual impairment after best correction who died (n = 89) and who survived(n = 76) were similar: age-related cataract (54% in those who died and 64% in those who survived), age-related maculopathy (33% and 24%, respectively), and glaucoma or other diseases (13% and 12%).

Cause of death was available for 429 of 604 deaths, owing to a lag in the completion of NDI data. Comparison of causes of death in this group did not reveal any substantial differences between persons with and without visual impairment at baseline. Among those without visual impairment who died, rates were as follows: heart diseases, 36%; cancer, 33%; stroke, 9%; and injury or fractures, 2%. Corresponding rates for those with visual impairment at baseline were 43%, 23%, 12%, and 2%, respectively.

These data indicate that persons with best-corrected visual acuity reduced to 20/40 or worse in their better eye had a 70% increased mortality risk compared with persons without visual impairment. Nuclear, posterior subcapsular, or cortical cataract was also associated with a 30% to 50% increased mortality risk. The visual impairment finding supports and extends the report from the Beaver Dam Eye Study.⁹ In that study (467 deaths), the age- and sex-adjusted association was not statistically significant after adjusting for systemic factors related to mortality. In our study population(604 deaths), the relationship between visual impairment and mortality was maintained after adjusting for a large number of systemic factors found associated with mortality, as well as for the presence of age-related cataract.

There are several possible explanations for an apparent relationship between impaired vision and mortality. First, inadequate control for age and other possible confounding variables could be partly responsible. Prevalence of visual impairment is strongly age-related.^{15 ,20} Although we included age in the Cox proportional hazards model, we may not have completely adjusted for its effects. Although many diseases representing other potential causes of death are included in the multivariate models, they are at best a marker for an underlying etiologic factor, and so we may not have adjusted completely for these other confounding causes of death. Second, decreased vision may be causally associated with other factors related to mortality in elderly persons, such as falling,²¹ hip fracture, and an increased likelihood of motor vehicle accidents,²² although our limited data on causes of death provided no support for this mechanism. Third, visual impairment may contribute to functional disability,^{1 - 3} loss of independence, and need for community support,⁶ as well as reduced social interaction and depression.⁴ Older persons with depression have been reported to have increased mortality.²³

Some previous studies have suggested that age-related cataract may be a predictor of decreased survival in older people.^{10 - 13} Our study confirms the Beaver Dam Eye Study nuclear cataract findings.⁹ We also found that posterior subcapsular and cortical cataract were independently associated with decreased survival, but did not observe a relationship between cataract surgery and mortality, possibly owing to small numbers.

No well-established or biologically plausible theories have been offered to explain the association between age-related cataract and decreased survival. It has been postulated that age-related cataract may reflect the status of systemic processes associated with aging^{10 - 11 ,13} and may be a marker of exposure to systemic factors producing free oxygen radicals, which could be associated with increased physiologic aging and decreased survival.^{9 ,24}

The association found between myopia and mortality could be partly explained by the well-known myopic shift in refraction associated with developing nuclear cataract, since the relationship weakened after including nuclear cataract in the model.

In conclusion, findings from our study indicate that both visual impairment and age-related cataract could be independent risk factors for a moderate increase in mortality risk for older persons. These data support and extend previous findings from the Beaver Dam Eye Study and add to the many described impacts of visual impairment on the older population.

Accepted for publication December 1, 2000.

This study was supported by grant 974159 from the Australian National Health and Medical Research Council (NHMRC), Canberra, Australia, and the Save Sight and Millennium Institutes, University of Sydney. Dr Wang held an NHMRC Public Health Postgraduate Research Scholarship when this study was conducted.

Corresponding author and reprints: Paul Mitchell, MD, PhD, FRACO, FRCOphth, Department of Ophthalmology, the University of Sydney Eye Clinic, Westmead Hospital, Hawkesbury Road, Westmead, New South Wales, Australia 2145(e-mail: paulmi@westgate.wh.usyd.edu.au).