Long-term Changes in Visual Acuity in an

Older Population over a 15-Year Period

The Blue Mountains Eye Study
Thomas Hong, MScMed, BAppSc,1 Paul Mitchell, MD, PhD,1 Elena Rochtchina, MAppStat,1
Calvin Sze-un Fong, MBBS,1 Ee-Munn Chia, MBBS, PhD,1 Jie Jin Wang, MBBS, PhD1,2
Purpose: To describe the change in visual acuity (VA) and incidence of visual impairment (VI) in an older
population over a 15-year period.
Design: Population-based cohort.
Participants: Of the 3654 participants of the Blue Mountains Eye Study (BMES) baseline examination from
1992 through 1994, 1149 were re-examined during the 15-year follow-up between 2007 and 2010.
Methods: Best-corrected VA by means of subjective refraction was measured with a logarithm of the
minimum angle of resolution chart using Early Treatment Diabetic Retinopathy Study methods at each
examination.
Main Outcome Measures: Unilateral VI was dened as VA worse than 20/40 and blindness was dened as
VA worse than 20/200 in the worse eye. Incident bilateral VI and blindness were determined according to VA in the
better eye at the 15-year visit. Doubling of the visual angle was dened as a loss of 15 letters or more from
baseline to the 15-year visit. Halving of the visual angle was dened as a VA improvement of 15 letters or more
over the same period. Causes of VI were determined at examination, by photographic grading, and from medical
records.
Results: Cumulative 15-year incidence of unilateral and bilateral VI was 12.3% (95% condence interval [CI],
11.0e13.6) and 5.2% (95% CI, 4.3e6.1), respectively, and for unilateral and bilateral blindness, the cumulative
incidence was 3.7% (95% CI, 3.0e4.4) and 0.9% (95% CI, 0.5e1.3), respectively. These incidence rates
increased signicantly with increasing age (P<0.01 for trend). Doubling and halving of the visual angle occurred in
6.9% (95% CI, 5.9e7.9) and 1.6% (95% CI, 1.0e2.2) of participants, respectively. Cataract accounted for 48.5%
of unilateral and bilateral incident VI, followed by age-related macular degeneration (26.9%). Age-related macular
degeneration accounted for 56.9% of unilateral and bilateral incident blindness cases, followed by cataract
(20.7%).
Conclusions: These data provide population-based estimates of long-term incidence of visual impairment
among older persons. Our estimate for cumulative incidence of blindness, accounting for competing risk of death,
was similar to that of the Beaver Dam Eye Study (BDES) after age standardization. However, our estimate for
cumulative incidence of VI was lower compared with that observed in the BDES population. This difference may
be explained in part by a higher mortality rate among our population.
Financial Disclosure(s): The author(s) have no proprietary or commercial interest in any materials discussed
in this article. Ophthalmology 2013;120:2091-2099 2013 by the American Academy of Ophthalmology.

Previous population-based studies have provided important

data on the prevalence and causes of visual impairment (VI) in
older persons.1e11 The prevalence of VI is known to increase
with age.1,12,13 It has been shown to be associated with a loss of
independence and an increased likelihood of reliance on
community support services among affected older individuals.14 Visual impairment also has been documented
consistently to be associated with an increased risk of
death1,13,15e17 in multiple older population samples. Visual
impairment among the older population thus is a major
condition demanding not only eye health care, but also an
increased need for aged care services for affected individuals.14
A number of studies have examined the long-term
changes in vision and the long-term incidence of VI
 2013 by the American Academy of Ophthalmology
Published by Elsevier Inc.

among older persons in the United States1,11 and

Europe.18,19 The purpose of this study was to describe the
15-year change in visual acuity (VA) and the incidence of
VI in a sample of older Australians participating in the Blue
Mountains Eye Study (BMES).

Methods
Population
The BMES baseline survey was conducted to describe the prevalence of, and risk factors associated with, vision loss and common
eye diseases in a typical older Australian sample consisting of
permanent residents 49 years of age or older residing in the Blue
ISSN 0161-6420/13/$ - see front matter
http://dx.doi.org/10.1016/j.ophtha.2013.03.032

2091

Ophthalmology Volume 120, Number 10, October 2013

Figure 1. Blue Mountains Eye Study (BMES) population ow chart.

Mountains area, west of Sydney. Prospective follow-up of this

population-based sample enabled us to assess the longer-term
incidence of vision loss and common eye diseases and the risk
factors associated with increased risks of these common diseases
among older people.
Detailed methods of the baseline survey were reported previously.20 In summary, at baseline we recruited and examined 3654
participants between 1992 and 1994. Surviving baseline
participants were invited to participate in the 5-, 10-, and 15-year
follow-up examinations. Of these, 2334 returned after 5 years
(75.8% of survivors), 1952 returned after 10 years (76.7% of
survivors), and 1149 returned after 15 years (56.1% of survivors).
Over the period between the 10- and 15-year visits, 364 had
moved, 81 were admitted to a nursing home and were too frail to
participate, 454 declined re-examination, and 496 had died (Fig 1).
All baseline and follow-up examinations of the BMES were
approved by the Human Research Ethics Committees of the
University of Sydney and the Western Sydney Area Health Service
and were conducted adhering to the tenets of the Declaration of
Helsinki. Signed informed consent was obtained from all participants at each examination visit.

Study Procedures
The 15-year follow-up examination methods were similar to those
used at baseline.20 Participants underwent a comprehensive eye
examination after pupil dilation and completed a detailed main
questionnaire during a face-to-face interview. Additional questionnaires were completed at home by study participants.
Visual acuity was measured using a retroilluminated logarithm
of the minimum angle of resolution chart (VectorVision CSV-

2092

100TM; VectorVision, Inc., Daytona, OH). Distance VA was

measured at 8 feet (244 cm) for each eye with current spectacles if
worn, followed by pinhole acuity. Either a previous spectacle
prescription or an autorefraction (Humphrey automatic refractor,
model 597; Humphrey-Zeiss, Oberkochen Germany) provided the
baseline for a subjective refraction. Subjective refraction was performed using the Beaver Dam Eye Study (BDES) modication of
the Early Treatment Diabetic Retinopathy Study protocol.21 Visual
acuity was recorded as the number of letters read correctly in each
eye (from 0e70). If no letters could be read from the chart at 8 feet
(244 cm), VA was assessed further at 2 feet (61 cm) and was
recorded as counting ngers, hand movements, light perception,
or no light perception. Visual acuity used in this report refers to
best-corrected VA after subjective refraction.
Peripheral visual elds were assessed via automated perimetry
in both eyes using Humphrey 24-2 Swedish interactive threshold
algorithm standard tests (Humphrey-Zeiss, model 530). Pupils
were dilated with tropicamide 1% and phenylephrine 10%. Eye
examinations that were performed included a slit lamp (Haag
Streit, Koeniz, Switzerland) and retroillumination (Neitz CT-R
cataract camera; Neitz Instruments, Tokyo, Japan) camera to assess
the lens. Retinal photographs were obtained using a Canon fundus
camera (CF-60DSi, EOS 1Ds Mk III; Canon, Tokyo, Japan) with
a digital back.

Denitions
Visual impairment was dened as VA worse than 20/40 (fewer
than 41 letters read), and blindness was dened as VA worse than
20/200 (0e5 letters read). The denition for VI was in keeping
with the denitions used in previous BMES reports, and it was the

Hong et al

Long-term Changes in VA in Older Population

minimum VA requirement to obtain and maintain an Australian

drivers license. An eye was considered to be at risk of VI developing if VA was 20/40 or better and was considered to be at risk of
blindness developing if VA was 20/200 or better at baseline.
Incident unilateral VI was dened as development of VI in only
1 eye (the worse eye) at follow-up visits where both eyes were at
risk of VI at baseline. Similarly, incident bilateral VI was dened
as development of VI in both eyes when at least 1 eye was at risk of
developing VI at baseline and was dened based on VA in the
better eye at follow-up. In addition, any incident VI was dened as
having at least 1 eye at risk of VI developing at baseline where the
at-risk eye was found to have VI at the follow-up examination.
Incidence of any blindness was dened as having at least 1 eye at
risk of blindness developing at baseline and the at-risk eye being
found to be blind at the follow-up examination.
A change in the number of letters read correctly was dened as
the difference in the numbers of letters read between the 15-year
and baseline examinations. Deterioration or doubling of the
visual angle was dened as a loss of 15 letters or more read
correctly in the better eye from baseline to the 15-year visits.
Improvement in vision (halving of the visual angle) was dened if
there was an improvement of 15 letters or more read correctly over
the same period. Eyes were at risk of deteriorating or improving
VA (doubling or halving the visual angle) if their baseline VA was
light perception or better, or if the baseline VA was 55 letters or
fewer, respectively.

Causes of Visual Impairment

Cataract was diagnosed during the dilated slit-lamp examinations
and also was recorded during lens photographic grading. Age-related
macular degeneration (AMD) was diagnosed at the time of the
dilated fundus examination and was conrmed by retinal photographic grading. Glaucoma was diagnosed when glaucomatous eld
loss was detected, based on repeated visual eld tests in which the
visual eld defect corresponded to optic disc changes, consistent
with typical glaucomatous cupping.22 Previously undiagnosed
ocular conditions detected during the study examinations were
referred to ophthalmologists, who provided regular eye clinic
services to the population of the study area. A survey report of the
ndings for each participants examination was sent to the
participant, treating general practitioner, and ophthalmologist for
those with ongoing care of pre-existing conditions by their doctors
or to assist in treatment of newly diagnosed conditions.
The primary cause of VI was dened as the single condition
that explained at least 50% of the vision loss. It was assessed at all
examinations by the same ophthalmologist (P.M.), who assessed
all VI cases and the severity of each eyes pathologic conditions.
Estimating the causes of measured vision loss was based either on
personal examination by the same ophthalmologist (P.M.) or by his
assessment of photographic images or other data. Based on the
severity of the condition and its likely effect on vision, the
ophthalmologist estimated and allocated the proportion of vision
loss resulting from each individual pathological condition, totaling
100%. Most of the causes were responsible for less than 90% of
reduced vision (e.g., cases with coexisting dense cataract and
glaucoma, or cataract and early AMD).

Statistical Analysis
SAS software version 9 (SAS Inc, Cary, NC) was used for all
analyses, and age was dened as age at baseline. Cumulative
15-year incidence was calculated while considering the competing
risk of death.23 The competing risk regression model is an
adaptation of the Kaplan-Meier method that takes into account
2 competing events; in this case, the events would be either

development of VI or death. Persons eligible for inclusion in

analyses contributed information up to the time when either of
these 2 events occurred. Persons were considered as censored when
reasons other than death prevented them from participation in the
follow-up examinations.
The incidence rates of VI and blindness reported by the BDES
were age standardized to the BMES population for comparison.
For this comparison, we used modied denitions of VI (VA
20/
40) and blindness (VA
20/200) following the denitions used in
the BDES.

Results
The mean age of participants at baseline was 64.5 years, and 57.9%
were female. Table 1 shows the baseline characteristics of the 2501
participants who attended at least 1 follow-up examination after
baseline with complete data for analysis. The main reasons for
not returning for follow-up examination were participant frailty
resulting from multiple comorbidities and relocation.
The overall mean decrease in number of letters read correctly
over the 15 years was similar in right and left eyes (6.9 and 6.8
letters, respectively); there was an inverse relationship between
reduction in the mean number of letters read correctly from baseline to the 15-year examination and increasing age, as shown
in Figure 2. There was no signicant gender difference in the
changes in numbers of letters read correctly in either right or left
eyes (P 0.97 and P 0.25, respectively).
The 15-year cumulative incidence of bilateral VI was 5.2%
(119 persons) and varied from 0.4% in those younger than 55 years
to 10.5% in those 75 years of age or older at baseline. The incidence of bilateral blindness was 0.9% (21 persons) and varied from
0% in those younger than 55 years to 1.7% in those 75 years of age
or older at baseline. Women were more likely than men to have
bilateral VI (6.8% vs. 3.4%; age-adjusted P 0.02) and bilateral
blindness (1.2% vs. 0.5%; age-adjusted P 0.09). The incidences
of VI and blindness were strongly age related in both men and
women (P<0.001; Table 2).
The 15-year cumulative incidence of unilateral VI was 12.3%
(293 persons) and varied from 3.4% in those younger than 55 years
to 15.4% in those 75 years of age or older at baseline. The incidence of unilateral blindness was 3.7% (95 persons) and varied
from 1% in those younger than 55 years to 6.9% in those 75 years
of age or older at baseline. Women were more likely than men to
have unilateral VI (13.3% vs. 10.8%) and blindness (4.7% vs.
2.6%); however, these differences were not statistically signicant
after adjusting for age (P 0.5 and P 0.1; Table 3).
The cumulative incidence of any VI was 17.3% (412 persons),
and incidence of blindness in any eye was 4.6% (116 persons). As
expected, women had a higher incidence of any VI (19.7% vs.
14.2%; age-adjusted P 0.06) and blindness (5.9% vs. 3.1%;
age-adjusted P 0.03) compared with men (Table 3).
Doubling of the visual angle occurred in 183 persons (6.9%)
based on the better eye, which was associated signicantly with
age in both men and women (P<0.001). Halving of the visual
angle occurred in 34 persons (1.6%) based on worse eye VA, and
there was no signicant association of halving the visual angle with
increasing age in either gender. Of these 34 persons, 20 (59%) had
undergone cataract surgery during the follow-up period.
Table 4 shows a comparison of the incidence of VI between our
study and BDES1 ndings using the modied denition of VI and
blindness, after direct standardization to the age distribution of the
BMES population. Incidences of VI were higher in the BDES
population (11.1%) compared with our population (6.4%), after
adjusting for competing risk of death. However, the incidences
of blindness were similar among the BDES population (1.2%)

2093

Ophthalmology Volume 120, Number 10, October 2013

Table 1. Baseline Characteristics of Participants in the Blue Mountains Eye Study: 15-Year Follow-up Examinations
Attended at Least 1 Follow-up
Examination (5, 10, or 15 yrs), n [ 2501

No.

P Value

62.8
55.1

192
157

<0.01
0.10
0.07

1558
1295
2224
1648

63.8
42.8
83.9
62.4

190
125
250
191

0.93
<0.01
<0.01
0.22

5.9
1.8

147
45

6.2
3.3

19
10

0.82
0.08

14.7
4.6
1.8
1.8
63.5

351
105
46
44
1569

8.3
3.9
1.0
1.6
58.7

24
11
3
5
176

<0.01
0.80
0.28
0.87
0.10

3.3
14.3
42.5
6.4
7.4

83
358
1063
160
186

3.0
7.8
41.8
8.5
6.2

9
24
128
26
19

0.74
<0.01
0.82
0.16
0.44

51.24
35.6
13.2
3.8

1239
860
319
91

42.5
34.6
23.0
3.8

124
101
67
11

<0.01
0.74
<0.01
0.99

%
Mean age  SD at baseline (yrs)
Female gender
Education (trade certicate or higher)
Employment
High prestige (Daniels prestige scale <5)
Receiving pension
Home ownership
Currently married
Visual impairment in worse eye
Moderate (VA 6/12e6/60)
Severe (VA 6/60 or worse)
Ocular history of
Cataract
AMD
Diabetic retinopathy
Glaucoma
Recent ophthalmologist/optometrist
examination (<2 yrs)
History of
Stroke
Heart disease (AMI or angina)
Hypertension
Diabetes
Cancer
Smoking status
Nonsmoker
Previous smoker
Current smoker
Heavy drinker (4 drinks/day)

Did Not Attend Any Follow-up

Examination, n [ 306

No.

57.8
60.6

1445
1445

63.5
53.4
90.9
65.9

64.38.6

60.57.8

AMD age-related macular degeneration; AMI acute myocardial infarct; SD standard deviation; VA visual acuity.

and our population (1.1%). Similar gender differences and agerelated trends in incidences also were observed in the BDES
population.
Figure 3 presents the proportions of incident cases by primary
attributable causes for any VI and blindness in any eye (either
rst or second eye). Cataract accounted for most cases (48.5%)
of any VI, followed by AMD (26.9%). However, AMD was
responsible for most blindness cases (56.9%), followed by

cataract (20.7%). Bilateral AMD was responsible for 71.4% of

bilateral blindness.
Of persons attending the 15-year follow-up, 29.9% had undergone cataract surgery and 6.4% had yttriumealuminumegarnet
laser capsulotomy. Among persons who attended the 15-year followup examination, 122 persons had AMD (15.6% underwent laser
treatment, with some more recently undergoing anti-VEGF therapy,
the latter commencing in 2007), 89 had glaucoma (23.6% underwent
laser treatment) and 17 had diabetic retinopathy (41.2% underwent
laser treatment).
Birth cohort effect was examined for persons born in the same
year and is presented graphically in Figure 4. For persons 75 to 79
years of age, the prevalence of VI ranged from 3.0% for those born
between 1930 and 1934 to 6.2% for those born between 1915 and
1919. Similarly, for persons 85 to 89 years of age, the prevalence
of VI ranged from 14.4% for those born between 1920 and 1924 to
32.0% for those born between 1905 and 1909.

Discussion

Figure 2. Mean change in number of letters read correctly in right and left
eyes by age between baseline and the 15-year follow-up.

2094

In our study, the overall cumulative incidence of VI and

blindness over the 15-year period was 5.2% and 0.9%,
respectively. Women had approximately twice the incidence
of any impairment when compared with men. The agerelated increasing trend in incident VI was consistent in

Hong et al

Long-term Changes in VA in Older Population

Table 2. Incidence of Bilateral Visual Changes by Age at Baseline and by Gender in the Blue Mountains Eye Study
Visual Impairmenty

Doubling of the Visual Angle*

Gender
and Age
Range
(yrs)
Female
<55
55e64
65e74
75
Total
Male
<55
55e64
65e74
75
Total
All
<55
55e64
65e74
75
Total

At
Risk

95%
Condence
Interval

217
564
579
375
1735

2.6
3.6
11.8
12.3
8.0

0.1e5.1
0.9e5.3
8.8e14.8
9.2e15.4
6.6e9.4

170
441
464
294
1369

1.4
5.2
6.0
8.1
5.6

0.0e3.3
2.9e7.5
3.7e8.2
5.3e10.9
4.3e6.9

387
1005
1043
669
3104

2.0
4.3
9.0
10.4
6.9

0.4e3.6
2.9e5.7
7.1e10.9
8.3e12.5
5.9e7.9

P
Value
(for
Trend)

At
Risk

Blindnessz

P
95%
Value
Condence
(for
Interval
Trend)

At
Risk

<0.01

212 0.7
541 2.1
534 11.5
238 13.6
1525 6.8

0.0e2.1
0.7e3.5
8.3e14.7
9.5e17.7
5.4e8.2

<0.01

217
559
567
338
1681

0.0
0.3
2.0
2.4
1.2

0.0e0.0
0.0e0.8
0.6e3.4
0.7e4.1
0.6e1.8

<0.01

165
427
413
195
1200

0.0
2.9
3.7
7.0
3.4

0.0e0.0
1.0e4.8
1.8e5.6
3.7e10.3
2.3e4.5

<0.01

167
434
447
264
1312

0.0
0.3
0.5
0.8
0.5

0.0e0.0
0.0e0.9
0.0e1.3
0.0e1.9
0.1e0.9

<0.01

377 0.4
968 2.5
947 7.7
433 10.5
2725 5.2

0.0e1.1
0.4e3.6
5.8e9.6
7.8e13.2
4.3e6.1

384
993
1014
602
<0.01 2993

0.0
0.3
1.3
1.7
0.9

0.0e0.0
0.0e0.7
0.5e2.1
0.6e2.8
0.5e1.3

Halving of the Visual Anglex

P
95%
Value
Condence
(for
Interval
Trend)

P
Value
(for
Trend)

At
Risk

95%
Condence
Interval

<0.01

118
365
475
365
1323

2.7
1.0
1.5
1.4
1.4

0.0e5.7
0.0e2.1
0.4e2.6
0.2e2.6
0.7e2.1

0.43

<0.01

66
231
340
279
916

0.0
2.2
2.1
1.9
1.9

0.0e0.0
0.0e4.4
0.6e3.6
0.3e3.5
1.0e2.8

0.03

<0.01

184
596
815
644
2239

1.7
1.4
1.8
1.6
1.6

0.0e3.6
0.3e2.5
0.9e2.7
0.6e2.6
1.0e2.2

0.04

*Doubling of the visual angle was dened as a loss of 15 letters or more in the better eye at follow-up visits.
y
Incidence of visual impairment was dened a visual acuity of <20/40 at follow-up in the better eye when both eyes were 20/40 at baseline examination.
z
Incidence of severe visual impairment was dened as visual acuity of <20/200 at follow-up in the better eye when both eyes were 20/200 at baseline
examination.
x
Halving of the visual angle was dened as a gain of 15 letters or more in the worse eye at the follow-up visit.

both sexes. Cataract and AMD were predominantly the

primary causes of VI and blindness in this population.
Our report from the BMES cohort complements previous
ndings from the BDES,1 which is the only other study
reporting 15-year longitudinal data on vision changes
among older persons. Both studies included similar samples
drawn from older, predominately white populations. The
overall mean number of letters read correctly was similar
between the 2 study populations. The BDES reported
a mean decrease of 5.4 letters (standard deviation [SD],
12.5 letters) and 7.1 letters (SD, 13.7 letters) in right and
left eyes, respectively, over 15 years, compared with the
corresponding mean decreases of 6.9 letters (SD, 13.6
letters) and 6.8 letters (SD, 13.3 letters) for the 2 eyes,
respectively, in our study. However, the cumulative incidence of VI accounting for the competing risk of death was
substantially higher in the BDES than the BMES across all
age groups (Table 4). The overall mortality rate was
substantially higher in our study sample (43.6%)
compared with the BDES (29.8%). The competing risk of
death may explain partially the discrepancies in VI
incidences between the 2 studies because the BDES had
a younger population at baseline. Without adjusting for
the competing risk of death in the analysis, the cumulative
incidence of VI in our population changed from 6.4% to
9.6%, which is similar to the estimates from the BDES
population (11.1%). Overall doubling and halving of the
visual angles also were similar between the 2 studies:
doubling of the visual angle occurred in 6.9% and 7.2%,

and halving of the visual angle occurred in 1.6% and

1.9% in the BMES and the BDES, respectively. Of eyes
in which blindness developed over the 15-year period, late
AMD was found to be the primary cause of incident
blindness in both the BMES (56.9%) and the BDES (52%).
Cataract was found to be the primary cause of blindness in
20.7% of eyes in the BMES and in 12% of eyes in the
BDES.1
A number of longitudinal studies have examined longterm changes in vision in older persons. The Barbados
Eye Study (BES)24 reported 9-year incidence data from
a predominately African Caribbean population 40 to 84
years of age at baseline, using the same VI denitions as the
BDES.1 The BES found a slightly higher incidence rate of
VI, but double the incidence of blindness (6.4% and
2.1%, respectively) compared with our study estimates,
albeit over a shorter follow-up period than our study
follow-up period. The higher rate of blindness in the BES
may be explained by the higher rates of open-angle glaucoma and cataract (28% for both), which were the primary
causes of blindness. This population originated from the
same region of West Africa as African Americans,6 who
previously were estimated to have approximately a 6-fold
prevalence of glaucoma compared with a white agematched population.25 Using World Health Organization
denitions (VA <20/70 in the better eye for VI, and VA
<20/400 for blindness), 2 population-based longitudinal
studies in Italy reported VI incidence data among adults.
The Ponza Eye Study18 reported a 12-year incidence of

2095

Ophthalmology Volume 120, Number 10, October 2013

Table 3. Incidence of Unilateral and Any Visual Impairment by Age at Baseline and by Gender in the Blue Mountains Eye Study
Unilateral

Any
Blindnessy

Visual Impairment*
Gender
and Age
Range
(yrs)
Female
<55
55e64
65e74
75
Total
Male
<55
55e64
65e74
75
Total
All
<55
55e64
65e74
75
Total

At
Risk

95%
Condence
Interval

212
541
534
238
1525

2.6
10.6
18.7
17.4
13.3

0.1e5.1
7.8e13.4
15.1e22.3
12.9e21.9
11.5e15.1

165
427
413
195
1200

4.4
10.1
13.3
12.7
10.8

1.0e7.8
7.0e13.2
10.0e16.6
8.5e16.9
9.0e12.6

377
968
947
433
2725

3.4
10.4
16.3
15.4
12.3

1.3e5.5
8.3e12.5
13.8e18.8
12.3e18.5
11.0e13.6

P
Value
(for
Trend)

At
Risk

95%
Condence
Interval

<0.01

217
559
567
338
1681

1.2
2.3
6.5
8.4
4.7

0.0e2.9
0.9e3.7
4.1e8.9
5.5e11.3
3.6e5.8

<0.01

167
434
447
264
1312

0.8
1.9
2.6
5.1
2.6

0.0e2.3
0.4e3.4
1.1e4.1
2.6e7.6
1.7e3.5

<0.01

384
993
1014
602
2993

1.0
2.1
4.6
6.9
3.7

0.0e2.1
1.1e3.1
3.2e6.1
4.9e8.9
3.0e4.4

Visual Impairmentz
P
Value
(for
Trend)

At
Risk

95%
Condence
Interval

<0.01

212
541
534
238
1525

3.3
12.5
29.2
30.6
19.7

0.5e6.1
9.5e15.5
25.1e33.3
25.6e35.6
0.0e4.1

<0.01

165
427
413
195
1200

4.4
12.9
16.9
19.4
14.2

1.0e7.8
9.5e16.3
13.3e20.5
14.8e24.0
12.2e16.2

<0.01

377
968
947
433
2725

3.8
12.7
23.5
25.5
17.3

1.6e6.0
10.4e15.0
20.7e26.3
22.1e29.0
15.8e18.8

Blindnessx
P
Value
(for
Trend)

95%
Condence
Interval

P
Value
(for
Trend)

At
Risk

<0.01

217
559
567
338
1681

1.2
2.5
8.5
8.7
5.9

0.0e2.9
1.0e4.0
5.8e11.2
7.6e13.8
4.6e7.2

<0.01

<0.01

167
434
447
264
1312

0.8
2.2
3.1
5.9
3.1

0.0e2.3
0.6e3.8
1.4e4.8
3.2e8.6
2.1e4.1

<0.01

<0.01

384
993
1014
602
2993

1.0
2.4
5.9
8.6
4.6

0.0e2.1
1.3e3.5
4.3e7.5
6.5e10.7
3.8e5.4

<0.01

*Incidence of unilateral visual impairment was dened as visual acuity of <20/40 at follow-up in only 1 eye (worse eye) when both eyes had visual acuity of
20/40 at baseline examination.
y
Incidence of unilateral blindness was dened as visual acuity of <20/200 at follow-up in only 1 eye (worse eye) when both eyes had visual acuity of 20/200
at baseline examination.
z
Incidence of any visual impairment was dened as having at least 1 eye at risk of developing visual impairment (visual acuity <20/40) at baseline and the at
risk eye demonstrating visual impairment at the follow-up examination.
x
Incidence of any blindness was dened as having at least 1 eye at risk of developing blindness (visual acuity <20/200) at baseline and the at-risk eye
demonstrating blindness at the follow-up examination.

bilateral VI of 3.9% and an incidence of bilateral blindness

of 0.7% in persons 40 years of age or older at baseline. The
Priverno Eye Study19 reported a 7-year incidence of bilateral
VI of 1.3% and that of bilateral blindness of 0.2% in
participants between 45 and 69 years of age at baseline.
Both studies had acceptable follow-up rates of 70.7%18 and
81.4%,19 respectively.
Our data conrmed an inverse relationship between
increasing age and worsening VA over the 15-year period.
The higher incidence of VI and blindness in the oldest age
group is consistent with higher prevalence rates of macular
degeneration and cataract26,27 in this age group. Consistent
with previous reports that women were more likely to have
VI, both by prevalence3,8,10 and incidence,1,19 we observed
the same sex difference in the long-term incidence of VI and
blindness. This is likely because women have greater
longevity than men, the latter having mortality as
a competing risk that could have occurred before VI or
blindness developed or was detected at the study follow-up
visits.28,29 As the mortality rate increases with age, these
deaths decrease the demand for health services, at least by
that proportion. Therefore, our estimates of VI burden from
this older population, although subject to survival bias,
likely are closer to the real demand for services by the older
sector of the population. This situation differs from

2096

a randomized controlled trial assessing the efciency of an

intervention, which requires a more conservative estimate of
treatment effect.
We briey assessed birth cohort effect and found a lower
prevalence of VI among those born in more recent years
compared with those born earlier. This suggested a cohort
effect that may be a consequence of improvement of lifestyles, health care service provision, and life expectancy
among Australians.30
The estimated VI and blindness incidences could have
been underestimated in our population. Participants who
did not return to any of the follow-up visits were almost
twice as likely to be current smokers at baseline,
compared with those participants who attended the 15year follow-up examinations. Smoking consistently has
been documented to be associated with an increased risk
of AMD,31,32 the leading cause of blindness in our and
other study populations. Participants who did not attend
any follow-up examinations (including those who died)
after the baseline examination had a higher prevalence of
VI and blindness at baseline. This could have led to an
underestimate of the incidence of blindness, but should
not have affected the estimation of the incidence of VI,
because these prevalent VI cases were no longer at risk
of VI.

Hong et al

Long-term Changes in VA in Older Population

Table 4. Fifteen-Year Incidence of Visual Impairment and Blindness Accounting for the Competing Risk of Death by Age and by Gender
in the Blue Mountains Eye Study Compared with the Beaver Dam Eye Study
Visual Impairment

Blue Mountains Eye Study

Gender and
Age Range at
Baseline (yrs)
Female
<55
55e64
65e74
75
Total
Male
<55
55e64
65e74
75
Total
All
<55
55e64
65e74
75
Total

At
Risk

Crude
%

211
536
523
226
1496

0.7
3.1
12.6
15.7
7.8

165
424
405
182
1176

0.0
4.3
4.8
10.4
4.8

376
960
928
408
2672

0.4
3.6
8.9
13.3
6.4

Beaver Dam
Eye Study

95%
Condence
Interval

At
Risk

Crude
%

6.3e9.3

661
574
569
270
2074

1.5
5.6
18.4
29.7
10.2

3.5e6.1

586
507
436
185
1714

1.4
2.8
12.6
18.5
6.0

5.4e7.4

1247
1081
1005
455
3788

1.4
4.3
15.9
25.1
8.3

Blindness
Standardized to
Blue Mountains
Eye Study

Blue Mountains
Eye Study

% (95%
Condence
Interval)

At
Risk

Crude
%

13.1 (11.6e14.6)

217
559
565
335
1676

0.0
0.5
2.3
3.5
1.3

8.4 (7.1e9.7)

167
433
446
263
1309

0.0
0.3
0.5
1.2
0.5

11.1 (10.1e12.1)

384
992
1011
598
2985

0.0
0.4
1.5
2.5
1.1

Beaver Dam
Eye Study

Standardized to
Blue Mountains
Eye Study

95%
Condence
Interval

At Risk

Crude %

% (95%
Condence
Interval)

0.9e2.3

667
578
598
327
2170

0.2
0.2
0.9
5.0
1.0

1.4 (0.9e3.4)

0.2e1.0

588
510
447
211
1756

0.2
0.0
1.6
1.8
0.6

0.9 (0.2e0.9)

0.7e1.5

1255
1088
1045
538
3926

0.2
0.1
1.2
3.7
0.8

1.2 (0.7e2.4)

The estimated VI incidence also might have been overestimated in our population. Participants who did not return
to any follow-up examination had a lower mean age at
baseline and a lower prevalence of cataract at baseline
compared with those who attended at least 1 follow-up
examination. Our study had a relatively high proportion of
participants lost to follow-up at the 15-year examination.
Assuming losses to follow-up were not deaths, the incidence
estimates potentially could range from 4.6%, if all participants did not have incident VI, to 19%, if all participants had

incident VI. However, it would be unrealistic to assume that

VI developed (or did not develop) in all such participants.
Peripheral vision loss generally was not included in the
study to dene VI, in keeping with previous BMES and
BDES reports. At the BMES baseline examination, only 1
participant had severe peripheral vision loss that could be
classied as VI or blindness according to the World Health
Organization criteria (a visual eld of less than 10 from the
point of xation). Visual acuity is a subjective measure;
inability to cooperate or anything affecting the participants
conscious or mental state (medication, intoxication, or
mental disorders) may have reduced the measurement and
subsequently may have increased our estimates.
In summary, our estimated cumulative incidences of VI
were lower than those reported from the BDES and BES,

Figure 3. Causes of any (unilateral or bilateral) visual impairment and

blindness.

Figure 4. Visual impairment rate by age at examination and year of birth.

2097

Ophthalmology Volume 120, Number 10, October 2013

whereas our study samples had a higher mortality rate,
a competing risk to VI development. Our study showed that
age was associated strongly with the development of VI in
this older population, likely because of the increasing
prevalence of many common ocular diseases associated with
aging. Women were twice as likely as men to have VI and
blindness, even after adjusting for age, which is likely
a result of different age-specic mortality rates between men
and women.
Findings from this report may help to establish effective
allocation of resources toward vision-related treatment,
rehabilitation, and prevention. Based on the population
estimates from the Australian Bureau of Statistics33 (2006),
the United States Census Bureau34 (2000), and the VI
incident rates from our data, we estimate the number of
older persons 50 years of age or older with VI will
increase from 480 000 to 1 million in Australia and from
3 million to 18 million in the United States. This
translates to an estimated demand for approximately
23 000 and 180 000 additional cataract surgical procedures
in the 2 countries, respectively, to treat new unilateral and
bilateral cases of VI and blindness caused by cataract.
However, caution should be applied to these estimates,
given that the BMES population was slightly older and
had a slightly higher socioeconomic status at baseline
compared with the overall Australian population.14

References
1. Klein R, Klein BE, Lee KE, et al. Changes in visual acuity in
a population over a 15-year period: the Beaver Dam Eye
Study. Am J Ophthalmol 2006;142:53949.
2. Klein R, Klein BE, Lee KE, et al. Changes in visual acuity in
a population over a 10-year period: the Beaver Dam Eye
Study. Ophthalmology 2001;108:175766.
3. Zhao J, Ellwein LB, Cui H, et al. Prevalence of vision
impairment in older adults in rural China: the China NineProvince Survey. Ophthalmology 2010;117:40916.
4. Tielsch JM, Javitt JC, Coleman A, et al. The prevalence of
blindness and visual impairment among nursing home residents in Baltimore. N Engl J Med 1995;332:12059.
5. Cedrone C, Ricci F, Nucci C, et al. Age-specic changes in the
prevalence of best-corrected visual impairment in an Italian
population. Ophthalmic Epidemiol 2007;14:3206.
6. Hyman L, Wu SY, Connell AM, et al; Barbados Eye Study
Group. Prevalence and causes of visual impairment in The
Barbados Eye Study. Ophthalmology 2001;108:17516.
7. Huang S, Zheng Y, Foster PJ, et al. Prevalence and causes of
visual impairment in Chinese adults in urban southern China:
the Liwan Eye Study. Arch Ophthalmol 2009;127:13627.
8. Evans JR, Fletcher AE, Wormald RP. Age-related macular
degeneration causing visual impairment in people 75 years or
older in Britain: an add-on study to the Medical Research
Council Trial of Assessment and Management of Older People
in the Community. Ophthalmology 2004;111:5137.
9. Chong EW, Lamoureux EL, Jenkins MA, et al. Sociodemographic, lifestyle, and medical risk factors for visual impairment in an urban Asian population: the Singapore Malay Eye
Study. Arch Ophthalmol 2009;127:16407.
10. Schellini SA, Durkin SR, Hoyama E, et al. Prevalence and
causes of visual impairment in a Brazilian population: the

2098

11.

12.

13.

14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.

Botucatu Eye Study. BMC Ophthalmol [serial online] 2009;

9:8. Available at: http://www.biomedcentral.com/1471-2415/
9/8. Accessed March 19, 2013.
Varma R, Chung J, Foong AW, et al, Los Angeles Latino Eye
Study Group. Four-year incidence and progression of visual
impairment in Latinos: the Los Angeles Latino Eye Study. Am
J Ophthalmol 2010;149:71327.
Wang JJ, Foran S, Mitchell P. Age-specic prevalence and
causes of bilateral and unilateral visual impairment in older
Australians: the Blue Mountains Eye Study. Clin Experiment
Ophthalmol 2000;28:26873.
Karpa MJ, Mitchell P, Beath K, et al. Direct and indirect
effects of visual impairment on mortality risk in older persons:
the Blue Mountains Eye Study. Arch Ophthalmol 2009;127:
134753.
Wang JJ, Mitchell P, Smith W, Leeder SR. Factors associated
with use of community support services in an older Australian
population. Aust N Z J Public Health 1999;23:14753.
Freeman EE, Egleston BL, West SK, et al. Visual acuity
change and mortality in older adults. Invest Ophthalmol Vis
Sci 2005;46:40405.
McCarty CA, Nanjan MB, Taylor HR. Vision impairment
predicts 5 year mortality. Br J Ophthalmol 2001;85:3226.
Foong AW, Fong CW, Wong TY, et al. Visual acuity and
mortality in a Chinese population. The Tanjong Pagar Study.
Ophthalmology 2008;115:8027.
Nucci C, Cedrone C, Culasso F, et al. Incidence of visual loss
in the Ponza Eye Study, Italy. Eye (Lond) 2005;19:17582.
Cedrone C, Culasso F, Cesareo M, et al. Incidence of blindness
and low vision in a sample population: the Priverno Eye Study,
Italy. Ophthalmology 2003;110:5848.
Attebo K, Mitchell P, Smith W. Visual acuity and the causes
of visual loss in Australia. The Blue Mountains Eye Study.
Ophthalmology 1996;103:35764.
Klein R, Klein BEK, Linton KLP, De Mets DL. The Beaver
Dam Eye Study: Visual acuity. Ophthalmology 1991;98:
13105.
Foran S, Wang JJ, Mitchell P. Causes of incident visual
impairment: the Blue Mountains Eye Study. Arch Ophthalmol
2002;120:6139.
Berry SD, Ngo L, Samelson EJ, Kiel DP. Competing risk of
death: an important consideration in studies of older adults.
J Am Geriatr Soc 2010;58:7837.
Hennis AJ, Wu SY, Nemesure B, et al, Barbados Eye Studies
Group. Nine-year incidence of visual impairment in the
Barbados Eye Studies. Ophthalmology 2009;116:14618.
Racette L, Wilson MR, Zangwill LM, et al. Primary openangle glaucoma in blacks: a review. Surv Ophthalmol
2003;48:295313.
Kawasaki R, Yasuda M, Song SJ, et al. The prevalence of agerelated macular degeneration in Asians: a systematic review
and meta-analysis. Ophthalmology 2010;117:9217.
Smith W, Assink J, Klein R, et al. Risk factors for age-related
macular degeneration: pooled ndings from three continents.
Ophthalmology 2001;108:697704.
Klein BE, Klein R, Lee KE. Incidence of age-related cataract: the
Beaver Dam Eye Study. Arch Ophthalmol 1998;116:21925.
Mitchell P, Cumming RG, Attebo K, Panchapakesan J. Prevalence of cataract in Australia: the Blue Mountains Eye Study.
Ophthalmology 1997;104:5818.
Hill DJ, Gray NJ. Patterns of tobacco smoking in Australia.
Med J Aust 1982;1:235.
Smith W, Mitchell P, Leeder SR. Smoking and age-related
maculopathy. The Blue Mountains Eye Study. Arch Ophthalmol 1996;114:151823.

Hong et al

Long-term Changes in VA in Older Population

32. Thornton J, Edwards R, Mitchell P, et al. Smoking and agerelated macular degeneration: a review of association. Eye
(Lond) 2005;19:93544.
33. Australian Bureau of Statistics. Population Projections
Australia, 2006e2101. Updated September 4, 2008. Available
at: http://www.abs.gov.au/AUSSTATS/abs@.nsf/DetailsPage/

3222.02006%20to%202101?OpenDocument. Accessed March

19, 2013.
34. U.S. Census Bureau. California: 2000. Summary Population
and Housing Characteristics. PHC-1-6. Washington, DC: U.S.
Census Bureau; 2002. Available at: http://www.census.gov/
prod/cen2000/phc-1-6.pdf. Accessed March 19, 2013.

Footnotes and Financial Disclosures

Originally received: August 2, 2012.
Final revision: March 1, 2013.
Accepted: March 25, 2013.
Available online: May 30, 2013.

Supported by the National Health and Medical Research Council (grant

nos.: 932085, 974159, 211069, and 457349), Canberra, ACT, Australia
Manuscript no. 2012-1170.

Centre for Vision Research, Department of Ophthalmology and Westmead Millennium Institute, University of Sydney, Sydney, Australia.
2
Centre for Eye Research Australia, University of Melbourne, Melbourne,
Australia.

Correspondence:
Jie Jin Wang, MMed, PhD, Centre for Vision Research, Department of
Ophthalmology, University of Sydney, Westmead Hospital, Hawkesbury
Road, Westmead, New South Wales, Australia 2145. E-mail: jiejin.wang@
sydney.edu.au.

Financial Disclosure(s):
The author(s) have no proprietary or commercial interest in any materials
discussed in this article.

2099