Associations between Elevated Intraocular

Pressure and Glaucoma, Use of Glaucoma

Medications, and 5-Year Incident Cataract
The Blue Mountains Eye Study
Sujatha Chandrasekaran, MBBS, MPH,
1
Robert G. Cumming, MBBS, PhD,
2
Elena Rochtchina, BSc, MApplStat,
1
Paul Mitchell, MD, PhD
1
Purpose: To examine incident relationships between elevated intraocular pressure (IOP), open-angle glau-
coma (OAG), and use of glaucoma medications with 5-year incident cataract.
Design: Population-based cohort study.
Participants: The Australian Blue Mountains Eye Study examined 3654 participants 50 years old at
baseline (82.4% response; 19921994); 2335 eligible participants were reexamined after 5 years (75.1% re-
sponse; 19971999).
Methods: A detailed medical and ocular history, including current medications, was taken, and a compre-
hensive eye examination, including applanation tonometry, automated perimetry, and lens photography, was
performed at each examination. The Wisconsin system was used to grade lens photographs in assessing
incident nuclear, cortical, and posterior subcapsular cataract (PSC). Data from both eyes were assessed using
generalized estimating equation analyses.
Main Outcome Measures: Elevated IOP was dened as 21 mmHg. Open-angle glaucoma was diagnosed
from typical glaucomatous eld loss with matching optic disc cupping, without reference to IOP. Subjects
without OAG or secondary or angle-closure glaucoma with IOP 21 mmHg in either eye were classied as
having ocular hypertension (OH), as were non-OAG subjects with IOP 22 mmHg using glaucoma medications.
Wisconsin levels 4 to 5 were graded as nuclear cataract, at least 5% lens involvement was graded as cortical
cataract, and any PSC dened its presence.
Results: The 5-year incidence of nuclear cataract was 23.4% (592/2532), or 23.1% (574/2486) after
excluding subjects using glaucoma medication. A marginally signicant association was found for elevated IOP
or OH at baseline and incident nuclear cataract (odds ratio [OR], 1.93 [95% condence interval (CI), 0.973.89],
and OR, 1.83 [95% CI, 0.963.48], respectively) in subjects not using glaucoma medications, after multivariate
adjustment. Age- and gender-adjusted analyses showed similar but statistically signicant associations. The
association between elevated IOP or OH and nuclear cataract was signicant in multivariate analyses (OR, 2.07
[95% CI, 1.043.12], and OR, 1.78 [95% CI, 1.053.01], respectively). Use of glaucoma medications was
associated with nonsignicantly increased adjusted odds for incident nuclear cataract (OR, 1.90 [95% CI,
0.923.92]). No associations, however, were found with incident cortical cataract or PSC.
Conclusions: Elevated IOP may increase the risk of nuclear cataract, but not that of other types. The use of
glaucoma medications could magnify this risk. Ophthalmology 2006;113:417424 2006 by the American
Academy of Ophthalmology.
Cataract remains the leading cause of new blindness world-
wide
1
and is the most frequent eye condition in the elderly.
1,2
Glaucoma is another major cause of age-related visual im-
pairment and blindness.
3
Several studies have consistently
identied a causal relationship between elevated intraocular
pressure (IOP) or ocular hypertension (OH) and open-angle
glaucoma (OAG).
46
In the Australian Blue Mountains Eye
Study (BMES), cataract and elevated IOP or OAG were
found to share some common risk factors, such as current
smoking, diabetes, myopia, dark iris color, and use of ste-
roid medications, so that an association between these con-
ditions could be expected.
716
Alternatively, a causal rela-
tionship may exist between the 2 pathologies, in either or
both directions.
It has been suggested that cataract may be a result of
either the glaucomatous process or its treatment. Glaucoma
was previously identied as a risk factor for cataract in a
number of studies,
1723
with reported relative risks ranging
from 1.3 to 6.1 among persons 50 to 79 years old.
1219
The
Originally received: August 26, 2005.
Accepted: October 15, 2005. Manuscript no. 2005-803.
1
Centre for Vision Research, Department of Ophthalmology and West-
mead Millennium Institute, University of Sydney, Westmead, Australia.
2
School of Public Health, University of Sydney, Sydney, Australia.
Correspondence to Paul Mitchell, MD, PhD, Centre for Vision Research,
Department of Ophthalmology, University of Sydney, Westmead Hospital,
Hawkesbury Road, Westmead, NSW 2145, Australia. E-mail: paul_mitchell@
wmi.usyd.edu.au.
417 2006 by the American Academy of Ophthalmology ISSN 0161-6420/06/$see front matter
Published by Elsevier Inc. doi:10.1016/j.ophtha.2005.10.050
Beaver Dam Eye Study (BDES) reported a relative risk of
2.6 for cataract surgery after 5 years in persons with IOP
levels of at least 14 mmHg at baseline. Elevated IOP was
suggested as a possible mechanism by which lens damage is
induced.
24
However, neither the Melbourne Visual Impair-
ment Project nor the Barbados Incident Study of Eye Dis-
eases (BISED) found a signicant relationship between
elevated IOP (21 mmHg) and nuclear cataract.
20,21
Glaucoma medications, both topical and oral, have also
been reported to increase the risk of incident cataract. In the
BISED, participants treated with IOP-lowering medications,
mainly topical -blockers, had a 3-fold relative risk of
developing nuclear cataract,
21
whereas the BDES reported a
borderline increased relative risk of posterior subcapsular
cataract (PSC) (P 0.09) among participants taking oral
or topical -blockers.
25
Other reports have suggested that
both glaucoma and its treatment, primarily topical med-
ications, may contribute to premature cataract.
22,26
Trauma
from glaucoma surgery is also known to accelerate cataract
development.
27
Cataract presence may also be causal in the development
of both elevated IOP and glaucoma.
2832
The BDES re-
ported that higher mean IOP correlated with severity of
nuclear cataract.
28
The mechanical effect of an intumescent
lens that causes angle closure compromising aqueous ow
has been suggested as a causal pathway.
28,29
This phaco-
morphic glaucoma improves after surgical removal of the
lens.
30,32
Alternately, cataractous lenses may leak high
molecular weight protein that can mechanically block the
trabecular meshwork, resulting in elevated IOP and phaco-
lytic glaucoma.
29,32
Rarely, inammatory or phacoanaphy-
lactic responses may complicate this process, elevating IOP
further.
29,31
Although phacolytic glaucoma is typically as-
sociated with hypermature cataract, it has been observed
with immature cataract, possibly due to a leak of lower
molecular weight proteins. Less severe secondary OAG
could also present with immature cataract.
32
This report aims to explore these relationships using
baseline and 5-year data from the BMES in 2 analyses.
First, we aimed to establish whether elevated IOP, OH,
OAG, or use of glaucoma medication could precede and be
independent risk factors for incident cataract. Second, we
wished to determine whether clinically signicant cataract
at baseline (as dened below) was an independent risk
factor for incident elevated IOP.
Materials and Methods
The BMES is a well-known population-based survey of vision and
common eye disease in an urban population 49 years old.
26,33
Participants were permanent residents of 2 postcodes in the Blue
Mountains region, west of Sydney, Australia. Ethical approval was
obtained, and all participants gave written informed consent. De-
tailed questionnaires were administered by trained interviewers
and included general medical, family, and ocular history, medica-
tions, and demographic details. Participants had comprehensive
eye examinations, including subjective refraction,
26
applanation
tonometry, and automated perimetry.
33
Baseline examinations
conducted in 1992 to 1994 included 3654 of 4433 (82.4%) eligible
residents (BMES 1). After a mean period of 5.1 years, 2335
members of this cohort (75.1% of eligible participants) were
reexamined during 1997 to 1999 (BMES 2) (Fig 1). Returning
participants were slightly younger than those who did not return
(mean baseline age, 64.3 vs. 65.6), less likely to be female (57.2%
vs. 65.5%), and less likely to have had previous cataract surgery
(3.8% vs. 6.5%).
At both examinations, cataract was documented from both slit
lamp (Topcon SL-7E camera, Topcon Optical Co., Tokyo, Japan)
and retroillumination (Neitz CT-R cataract camera, Neitz Instru-
ment Co., Tokyo, Japan) lens photographs. Details of photographic
techniques and grading have been described.
34,35
The Wisconsin
cataract grading system
36
was followed closely. Good agreement
was found for intergrader and intragrader reliability.
37
Past cata-
ract surgery was conrmed at each examination and from the
grading. Presence and severity of the 3 main types of age-related
cataractnuclear and cortical cataract and PSCwere assessed
for each eye. Nuclear cataract was dened on a 5-level scale by
comparison with a set of 4 standard slit-lamp photographs. The
percentage area involved by cortical cataract or PSC in each eye
was calculated from the estimated percent area involvement in
each of 9 lens segments, divided by a grid overlay.
36
The 5-year
incidence of each cataract type of interest was dened as devel-
opment of that clinically signicant opacity in at least one eye,
compared with baseline. We used the BDES denition for cataract;
this was considered clinically signicant for Wisconsin grades 4
and 5 of nuclear cataract, when cortical cataract involved at least
5% of the lens, and when any posterior subcapsular opacity was
present.
38
We included persons with different cataract types in
Ungradable or missing nuclear cataract
543 participants (1086 eyes) died
383 participants (766 eyes) moved
393 participants (786 eyes) refused
participation
2335 participants (4670
eyes) eligible for study
at 5 years
Un-gradable or missing nuclear cataract
photographs at follow-up = 252 eyes
Significant nuclear cataract at baseline
= 445 eyes
Baseline and incident aphakic,
pseudophakic, enucleated = 204 eyes
Baseline and incident glaucoma surgery
= 3 eyes
Baseline examination
3654 participants (7308
eyes)
2532/4670 eyes (54.2%) at
risk of nuclear cataract
available for study
Un-gradable or missing nuclear cataract
photographs at baseline = 1234 eyes
Figure 1. Flowchart of participants or eyes available for study at baseline
and follow-up examinations, including losses to follow-up and exclusions
according to study criteria.
Ophthalmology Volume 113, Number 3, March 2006
418
either or both eyes in the primary analyses. Other analyses that
excluded eyes with different types of cataract are also discussed.
Data for cortical cataract and PSC were missing from approx-
imately 3% of subjects at baseline because photographs were not
taken or were unsuitable for grading. An intermittent, random,
nonsystematic camera malfunction resulted in slight underexpo-
sure of photographs in 1045 of the baseline participants (29%). We
felt that these cases could not be reliably graded for nuclear
cataract, and they were excluded. These subjects, however, did not
differ in any important way from subjects with gradable photo-
graphs.
34
Figure 1 demonstrates how ungradable photographs were
divided between baseline and follow-up in determining incident
nuclear cataract. After 5 years, 2532 of 4670 (54.2%) eyes at risk
of nuclear cataract were available for study.
Elevated IOP was dened as 21 mmHg. Open-angle glau-
coma was diagnosed from typical glaucomatous eld loss on a
Humphrey 30-2 visual eld test, in association with matching optic
disc rim thinning and an enlarged cup-to-disc ratio (0.7), or
cupdisc asymmetry (0.3) between eyes.
33
Eyes without OAG
or secondary or angle-closure glaucoma with IOP 21 mmHg in
either eye were classied as having OH. We also included in this
category eyes using glaucoma medications with IOP 22 mmHg,
but without denite signs of OAG. No eyes in this group had
traumatic angle contusion or laser iridotomy that could predispose
to cataract formation. Only 3 of 2497 (0.001%) eyes with OH at
baseline had pseudoexfoliation, recognized as a risk factor for
cataract. We considered that excluding these few patients would
not substantially affect outcomes. In the second set of analyses of
incident elevated IOP, the incidence denition excluded partici-
pants with OH or OAG.
Age, gender, and other factors found associated with age-
related cataract, OAG, elevated IOP, or OH in earlier reports were
considered, including systemic factors such as diabetes and
hypertension. Diabetes was dened from history or fasting
blood glucose 7.0 mmol/l taken within 3 months of the
examination. Hypertension was dened from history of treated
hypertension, and/or systolic blood pressure (BP) 160 mmHg
or diastolic BP 90 mmHg at examination. Other factors,
including current smoking, education, sun exposure, and body
mass index, were assessed. Smokers were dened as participants
who smoked tobacco or had recently stopped within the last year.
Sun exposure was assessed by grading presence of typical signs of
sun-related skin damage as a proxy measure, rated on a 4-level
severity scale. Education status was categorized by attainment of a
trade or qualications after leaving school. Body mass index was
calculated as weight (kilograms)/height (meters squared). Ocular
factors including IOP, iris color, myopia, and presence of pseudo-
exfoliation were assessed on an eye-specic level. Iris color as-
sessment was performed by comparing participants with 4 stan-
dard photographs (courtesy University of Wisconsin) before pupil
dilatation. Myopia was dened as sperical equivalent 1.0
diopter. A baseline history of use of glaucoma medications or oral
and/or inhaled corticosteroids was taken.
The Statistical Analysis System (version 6.12 for Windows,
SAS Institute, Cary, NC) was used. Data from both eyes were
analyzed, using the Liang and Zeger
39
generalized estimating
equation approach. Although mostly bilateral, OAG, elevated IOP,
OH, use of glaucoma medications, and cataract are eye specic.
The generalized estimating equation method permits the use of
data from both eyes while accounting for the correlation between
the two eyes in a single subject. This method confers greater
precision in estimates of association and is less sensitive to missing
data in some eyes relative to other regression models. Cataract was
a dichotomized variable in all analyses. In initial analyses exam-
ining incident cataract, the factors included varied by the type of
cataract, based on their established associations.
In secondary analyses examining incident elevated IOP, we
controlled for elevated IOP and risk factors found for OAG in the
BMES (hypertension, diabetes, smoking, myopia, family history,
history of typical migraine, steroid use, dark iris color, and pseudo-
exfoliation).
711,4042
In all analyses, age was a continuous vari-
able, whereas all other variables were categorical. Odds ratios
(ORs), 95% condence intervals (CIs), and P values are presented;
Ps0.05 indicate statistical signicance.
Results
In analyses examining whether OAG, glaucoma risk factors (par-
ticularly elevated IOP), OH, and use of glaucoma medications
could precede and be independent risk factors for incident cataract,
we excluded 148 eyes (143 with aphakia or pseudophakia, 2
enucleated, and 3 that had undergone glaucoma surgery). Using
data from both eyes, before excluding eyes treated with glaucoma
medications, 592 of 2532 eyes at risk developed nuclear cataract
(23.4%), 350 of 3573 eyes at risk developed cortical cataract
(9.8%), and 99 of 3958 eyes at risk developed PSC (2.5%). After
excluding eyes receiving glaucoma medication, incidence rates for all
3 types of cataract were similar: 574 of 2486 eyes at risk (23.1%)
developed nuclear cataract, 348 of 3512 eyes at risk developed cor-
tical cataract (9.9%), and 97 of 3891 eyes at risk developed PSC
(2.5%).
In our study, 119 participants (3.2%) used glaucoma medica-
tion at baseline, of whom 52 (44.5%) were diagnosed with OAG
and 42 (35.3%) had OH; 60 of 119 participants (2.6%) were using
glaucoma medications at follow-up. Of these 60 participants, 46,
61, and 67 eyes were at risk of developing nuclear cataract, cortical
cataract, and PSC, respectively. Glaucoma medications used at
baseline included timolol (n 88), pilocarpine (n 44), dipiv-
efrin (n 24), betaxolol (n 22), phospholine iodide or other
(n 2), and oral acetazolamide (n 6). More than one glaucoma
medication was used by 52 participants (44%), for a mean duration
of 6.9 years (standard deviation [SD], 6.5). Mean durations of use
by drug type were as follows: timolol, 5.9 years (SD, 4.7); pilo-
carpine, 8.8 years (SD, 8.5); dipivefrin, 4.4 years (SD, 4.2); and
betaxolol, 1.7 years (SD, 0.9).
Incident nuclear cataract was greater in eyes with elevated IOP
at baseline than cortical cataract or PSC (39.1% vs. 11.7% and
3.0%, respectively). This was also true for eyes with OH and OAG
at baseline, before and after excluding eyes on glaucoma medica-
tions. Incident nuclear cataract was also greater in eyes treated
with glaucoma medications at baseline than cortical cataract or
PSC (39.1% vs. 6.6% and 4.5%, respectively) (Tables 13).
Table 1 presents the results of the age-, gender-, and multivariate-
adjusted regression models for associations between 5-year inci-
dent nuclear cataract and elevated IOP, OH, OAG, and the use of
glaucoma medications at baseline. In participants not using glau-
coma medications, elevated IOP and OH at baseline were associ-
ated with 5-year incident nuclear cataract in age- and gender-
adjusted analyses (OR, 2.03 [95% CI, 1.063.91], and OR, 2.07
[95% CI, 1.163.71], respectively). In the multivariate analyses,
elevated IOP and OH at baseline had borderline associations with
5-year incident nuclear cataract (OR, 1.93 [95% CI, 0.973.89],
and OR, 1.83 [95% CI, 0.963.48], respectively). When partici-
pants on glaucoma medications were included, elevated IOP and
OH were signicantly associated with incident nuclear cataract
(ORs, 1.97 [95% CI, 1.053.70] and 1.94 [1.203.12], respec-
tively, and ORs, 2.07 [95% CI, 1.044.12] and 1.78 [1.053.01],
respectively, after multivariate adjustment). Glaucoma medication
use was associated with nonsignicant increased adjusted odds for
incident nuclear opacities (OR, 1.90 [95% CI, 0.923.92]). Open-
Chandrasekaran et al Intraocular Pressure and Incident Cataract
419
angle glaucoma was not associated with signicant increased odds
for incident nuclear cataract in all analyses.
Table 2 presents the results of the age-, gender-, and multivariate-
adjusted regression models for associations between 5-year inci-
dent cortical cataract and elevated IOP, OH, OAG, and the use of
glaucoma medications at baseline. No signicant associations were
noted for cortical cataract or, similarly, for PSC, shown in Table 3.
In secondary analyses examining whether clinically signicant
cataract at baseline could precede and be an independent risk
factor for incident elevated IOP, we excluded eyes with elevated
IOP or on glaucoma medications, those that were aphakic or
pseudophakic at baseline, as well as eyes having incident cataract
surgery. This is known to decrease IOP in both normotensive and
glaucomatous eyes.
4346
At baseline, 284 of 3982 eyes (7.1%) had
nuclear cataract, 547 of 3982 eyes (13.7%) had cortical cataract,
and 91 of 3982 eyes (2.3%) had PSC; 51 of 3982 eyes at risk
(1.3%) developed elevated IOP after 5 years. The relationship
between pure nuclear cataract and incident elevated IOP was
signicant (
2
1
9.98, P 0.002), but it became nonsignicant
after adjusting for age and gender (data not shown).
We then investigated age-, gender-, and multivariate-adjusted
estimates of incident mean IOP according to baseline cataract
status using analysis of covariance. There was no signicant dif-
ference in incident mean IOP according to baseline cataract status,
compared with incident mean IOP in respective reference groups;
P values for incident mean IOP and baseline nuclear cataract,
cortical cataract, and PSC, after agegender adjustment, were
0.87, 0.56, and 0.10, respectively. After multivariate adjustment,
corresponding P values were 0.96, 0.95, and 0.09, respectively.
We also investigated incident mean IOP according to the base-
line grade of nuclear cataract. There was a statistically signicant
association between incident elevated IOP and nuclear cataract
grade (MantelHaenszel
2
1
5.07, P 0.02; grades 4 and 5
analyzed together because of small numbers). This association,
however, became nonsignicant after adjusting for age and gender
(P 0.97) and after multivariate adjustment (P 0.78).
Discussion
Our data indicate a relationship (2-fold increased odds)
between baseline elevated IOP or OH and incident nuclear
cataract in patients using glaucoma medications. No signif-
icant associations were observed for incident cortical cata-
ract or PSC. Suggestive positive associations, of borderline
signicance, were found when each of these baseline factors
was considered independently. Possibly due to insufcient
statistical power, we found no signicant association be-
tween OAG and incident nuclear cataract. Interestingly, the
OR for incident nuclear cataract was lowest for OAG,
compared with OH and elevated IOP after excluding par-
ticipants receiving glaucoma medications at baseline (Table
1). The association (OR, 1.90) between incident nuclear
cataract and use of glaucoma medications at baseline after
multivariate adjustment was nearly signicant, possibly due
Table 1. Adjusted Odds Ratios (ORs) and 95% Condence Intervals (CIs) from General Estimating Equation Models for
Associations between 5-Year Incident Nuclear Cataract and Elevated Intraocular Pressure (IOP) (21 mmHg), Ocular Hypertension
(OH), Open-Angle Glaucoma (OAG), and Use of Glaucoma Medications
Eligible Participants No. of Eyes (%)
Age- and Gender-Adjusted Model Multivariate Model*
OR (CI) P Value OR (CI) P Value
Excluding participants on glaucoma medications
Elevated IOP

No 529/2369 (22.3) 1.0 (referent) 1.0 (referent)

Yes 26/65 (40.0) 2.03 (1.063.91) 0.03 1.93 (0.973.89) 0.06
OH

No 530/2376 (22.3) 1.0 (referent) 1.0 (referent)

Yes 34/85 (40.0) 2.07 (1.163.71) 0.01 1.83 (0.963.48) 0.07
OAG

No 568/2471 (23.0) 1.0 (referent) 1.0 (referent)

Yes 6/15 (40.0) 1.05 (0.333.38) 0.94 1.43 (0.405.12) 0.58
Including participants on glaucoma medications
Elevated IOP

No 529/2369 (22.3) 1.0 (referent) 1.0 (referent)

Yes 27/69 (39.1) 1.97 (1.053.70) 0.04 2.07 (1.044.12) 0.04
OH

No 530/2376 (22.3) 1.0 (referent) 1.0 (referent)

Yes 46/121 (38.0) 1.94 (1.203.12) 0.01 1.78 (1.053.01) 0.03
OAG

No 581/2509 (23.2) 1.0 (referent) 1.0 (referent)

Yes 11/23 (47.8) 1.41 (0.583.40) 0.45 1.80 (0.704.66) 0.22
Using glaucoma medication

No 572/2476 (23.1) 1.0 (referent) 1.0 (referent)

Yes 18/46 (39.1) 1.77 (0.893.55) 0.11 1.90 (0.923.92) 0.08
*Adjusted for age, gender, education, current use of any steroid medication, myopia, dark iris color, smoking, hypertension, diabetes, and sun-related skin
damage.

Referent group: all eyes excluding elevated IOP, OH, and OAG.

Referent group: all eyes excluding OH or OAG.

Referent group: all eyes excluding OAG.

Referent group: all eyes excluding glaucoma medication.

Ophthalmology Volume 113, Number 3, March 2006
420
to insufcient statistical power. This result is quite similar to
the BISED ndings (2-fold vs. 3-fold increased odds).
21
Cataractous human lenses and lens capsules surgically
removed from subjects with glaucoma and systemic hyper-
tension have been studied using Fourier transform infrared
microspectroscopy by quantitatively estimating alterations
in protein structure and composition.
47,48
A signicant in-
crease in protein -sheet composition has been reported in
extracted human lens capsules, suggesting that glaucoma
and systemic hypertension could accelerate formation of the
structure. Enhanced protein aggregation in immature cat-
aractous lens capsules, thought to result from altered inter-
molecular hydrogen bonding of protein components and a
modication of the secondary structure of protein in lens
capsules, was more pronounced in subjects with systemic
hypertension than in glaucomatous eyes. Enhancement in
protein -sheet and random coil structures in human lens
capsules is related to greater protein insolubility and has
been attributed to age-related cataractogenesis.
48,49
Altered
conformation of the lens capsule is also thought to alter
ionic transport through capsule outow channels, consistent
with defective ion transport and accelerated cataractogen-
esis in salt-sensitive hypertensive rats.
50
In addition to the lens capsule, cataractous lenses un-
dergo age-related modications in structural protein and
lipids of the lens membrane and cytoskeleton.
51
Age-related
changes in lens water content seem to inuence lens protein
reconstruction during cataractogenesis.
48,52
Enhanced dif-
ferences in the protein and lipid structures of human lenses
in cataractous eyes with glaucoma are reported when com-
pared with nonglaucomatous eyes.
48
A positive relationship between IOP and hypertension is
well established
53,54
; modication of the structural compo-
nents of the lens capsule enhanced by glaucoma and sys-
temic hypertension could further elevate IOP and, by feed-
back, exacerbate cataract formation.
48,49
This may explain
our reported associations between incident nuclear cataract
and elevated IOP and OH. Glaucoma medications may act
to exacerbate cataractogenic structural changes induced by
elevated IOP or OH.
Strengths of our study are its relatively high participation
rate, the use of data from both eyes on temporal relation-
ships, control for cataract and glaucoma risk factors in the
analysis, careful measurement of study variables using stan-
dardized protocols, masked photographic grading dening
the 3 main cataract types, and a relatively large sample of a
dened population base reducing selection bias.
Limitations of our study include potential underestima-
tion of the true incident cataract rate, as 75.1% of eligible
participants were reexamined after 5 years. Participants lost
to follow-up included those too sick to reattend; cataract is
an established risk factor for mortality.
55,56
Further, our data
do not permit a comparison between eyes developing inci-
dent nuclear cataract and eyes without any nuclear cataract
Table 2. Adjusted Odds Ratios (ORs) and 95% Condence Intervals (CIs) from General Estimating Equation Models for
Associations between 5-Year Incident Cortical Cataract and Elevated Intraocular Pressure (IOP) (21 mmHg), Ocular Hypertension
(OH), Open-Angle Glaucoma (OAG), and Use of Glaucoma Medications
Eligible Participants No. of Eyes (%)
Age- and Gender-Adjusted Model Multivariate Model*
OR (CI) P Value OR (CI) P Value
Excluding participants on glaucoma medications
Elevated IOP

No 326/3354 (9.7) 1.0 (referent) 1.0 (referent)

Yes 11/88 (12.5) 1.37 (0.702.69) 0.36 1.55 (0.793.02) 0.20
OH

No 329/3367 (9.8) 1.0 (referent) 1.0 (referent)

Yes 14/118 (11.9) 1.18 (0.622.25) 0.62 1.21 (0.642.29) 0.55
OAG

No 344/3497 (9.8) 1.0 (referent) 1.0 (referent)

Yes 2/15 (13.3) 0.96 (0.204.56) 0.96 1.36 (0.247.76) 0.73
Including participants on glaucoma medications
Elevated IOP

No 326/3354 (9.7) 1.0 (referent) 1.0 (referent)

Yes 11/94 (11.7) 1.27 (0.652.47) 0.49 1.45 (0.752.82) 0.27
OH

No 329/3367 (9.8) 1.0 (referent) 1.0 (referent)

Yes 17/166 (10.2) 1.02 (0.581.81) 0.95 0.95 (0.521.74) 0.87
OAG

No 347/3549 (9.8) 1.0 (referent) 1.0 (referent)

Yes 3/24 (12.5) 0.91 (0.282.96) 0.87 1.11 (0.323.86) 0.87
Using glaucoma medication

No 343/3504 (9.8) 1.0 (referent) 1.0 (referent)

Yes 4/61 (6.6) 0.63 (0.221.77) 0.38 0.35 (0.081.51) 0.16
*Adjusted for age, gender, education, body mass index, smoking, hypertension, diabetes, and sun-related skin damage.

Referent group: all eyes excluding elevated IOP, OH, and OAG.

Referent group: all eyes excluding OH or OAG.

Referent group: all eyes excluding OAG.

Referent group: all eyes excluding glaucoma medication.

Chandrasekaran et al Intraocular Pressure and Incident Cataract
421
at baseline. Because of a random camera fault causing slight
underexposure of many images, around one third of the
baseline slit-lamp photographs were not able to be reliably
graded for nuclear cataract, potentially reducing statistical
power. This seems unlikely, however, to have introduced
any systematic bias. It is also possible that we have not
adequately controlled for certain confounders, including
other cataract risk factors.
To assess the effect of including eyes with other types of
cataract in addition to specic cataract types assessed, we
performed analyses of pure incident cataract by type, ex-
cluding eyes with other cataract types. Analyses for incident
pure nuclear cataract gave an association of similar magni-
tude, now weaker and nonsignicant, between elevated IOP
at baseline, including participants on glaucoma medications,
and IOP after multivariate adjustment (OR, 1.94 [95% CI,
0.794.78]). In this multivariate model, weaker associa-
tions were found for OH at baseline, including participants
on glaucoma medications (OR, 1.72 [95% CI, 0.744.00]),
and use of glaucoma medication at baseline (OR, 1.52 [95%
CI, 0.633.70]). The lower number of eyes available for
analysis reduced statistical power, which may explain the
loss of statistical signicance for these associations. We had
insufcient data to repeat analyses for incident pure cortical
cataract and PSC.
Both our study and the BISED
21
have assessed incident
nuclear cataract using the same denition for elevated IOP
and also examined incident cataract relationships with OH,
OAG, and use of glaucoma medications. Our ndings for
incident nuclear cataract are comparable to the BISED
ndings. The association between use of glaucoma medica-
tion and incident nuclear cataract only approached statistical
signicance, possibly because of insufcient statistical
power. The methodologies of these 2 studies also differed,
however, in a number of important aspects. For the regres-
sion analyses, we used generalized estimating equation
models including eye-specic data that accounted for the
correlation between fellow eyes. Incidence periods were 4
years in the BISED and 5 years in our study. Incident
nuclear opacities in the BISED were dened as develop-
ment of any nuclear opacity, and their presence was dened
by Lens Opacities Classication System II nuclear scores
2. The BMES used the Wisconsin cataract grading system to
assess presence of nuclear cataract; nuclear cataract was
considered clinically signicant when present for Wisconsin
grades 4 and 5, compared with baseline. We chose the
BDES photographic denition over subjective clinical mea-
sures, such as visual acuity, as the photographic denition is
less subjective and potentially more reliable, particularly in
assessing cataract severity by type.
38
Incident nuclear cata-
ract was 9.2% (241/2609) in the BISED, compared with
23.4% (592/2532) in the BMES, using these denitions. To
reduce possible overlap between participants with factors
related to elevated IOP, the referent group for elevated IOP
Table 3. Adjusted Odds Ratios (ORs) and 95% Condence Intervals (CIs) from General Estimating Equation Models for
Associations between 5-Year Incident Posterior Subcapsular Cataract and Elevated Intraocular Pressure (IOP) (21 mmHg),
Ocular Hypertension (OH), Open-Angle Glaucoma (OAG), and Use of Glaucoma Medications
Eligible Participants No. of Eyes (%)
Age- and Gender-Adjusted Model Multivariate Model*
OR (CI) P Value OR (CI) P Value
Excluding participants on glaucoma medications
Elevated IOP

No 92/3717 (2.5) 1.0 (referent) 1.0 (referent)

Yes 2/95 (2.1) 0.69 (0.095.19) 0.72 0.55 (0.093.25) 0.51
OH

No 94/3732 (2.5) Insufcient data

Yes 0/123 (0)
OAG

No 95/3869 (2.5) 1.0 (referent) 1.0 (referent)

Yes 2/22 (9.1) 2.33 (0.5410.14) 0.26 2.06 (0.656.60) 0.22
Including participants on glaucoma medications
Elevated IOP

No 92/3717 (2.5) 1.0 (referent) 1.0 (referent)

Yes 3/101 (3.0) 1.08 (0.244.79) 0.92 0.53 (0.093.15) 0.49
OH

No 94/3732 (2.5) 1.0 (referent) 1.0 (referent)

Yes 3/177 (1.7) 0.60 (0.142.67) 0.50 0.41 (0.062.96) 0.38
OAG

No 98/3927 (2.5) 1.0 (referent) 1.0 (referent)

Yes 2/31 (6.5) 1.66 (0.416.78) 0.48 1.26 (0.384.25) 0.71
Using glaucoma medication

No 97/3881 (2.5) 1.0 (referent) 1.0 (referent)

Yes 3/67 (4.5) 1.53 (0.346.93) 0.58 1.15 (0.178.08) 0.89
*Adjusted for age, gender, education, current use of any steroid medication, myopia, dark iris color, smoking, hypertension, and diabetes.

Referent group: all eyes excluding elevated IOP, OH, and OAG.

Referent group: all eyes excluding OH or OAG.

Referent group: all eyes excluding OAG.

Referent group: all eyes excluding glaucoma medication.

Ophthalmology Volume 113, Number 3, March 2006
422
in our study excluded persons with OAG and OH, rather
than considering those without elevated IOP alone, as in the
BISED. In relation to use of glaucoma medication, the
proportion of participants receiving glaucoma medications
was higher in our study than that in the BISED, but the
average duration of use was similar. It is possible that
participants on glaucoma medications may be more likely to
be reviewed, conferring a surveillance bias in incident cat-
aract detection. This was not likely in our study, where 121
of 193 eyes on glaucoma medication at baseline (62.7%)
were not reviewed after 5 years, often because of death (86
eyes) or because they were ineligible for the incident anal-
ysis (35 eyes). Further, the BISED studied younger, pre-
dominantly black participants, in contrast to the older, pre-
dominantly white BMES population.
We found no association between cataract at baseline
and incident elevated IOP. Our unadjusted results were
comparable to the BDES report of an association between
increasing nuclear cataract grade and IOP (approximately
2-fold higher risk of incident elevated IOP)
28
; but this was
nonsignicant after adjusting for IOP risk factors. Low
statistical power with wide CIs could explain the difference
with BDES ndings.
In summary, this study found that eyes with elevated IOP
being treated with glaucoma medications had a 2-fold in-
creased risk of 5-year incident nuclear cataract, after adjust-
ing for age and other confounders. This nding was not
found for incident cortical cataract or PSC. Possibly due to
insufcient statistical power, use of glaucoma medications
alone was only a borderline signicant risk factor for inci-
dent nuclear cataract. These ndings are compatible with
previous studies
21,27
suggesting an association. Further re-
search into the cataractogenic potential of different glau-
coma medications may be warranted.
References
1. Thylefors B, Negrel AD, Pararajasegaram R, Dadzie KY.
Global data on blindness. Bull World Health Organ 1995;73:
11521.
2. Hyman L. Epidemiology of eye disease in the elderly. Eye
1987;1:33041.
3. Leske MC. The epidemiology of open-angle glaucoma: a
review. Am J Epidemiol 1983;118:16691.
4. Ekstrom C. Elevated intraocular pressure and pseudoexfolia-
tion of the lens capsule as risk factors for chronic open-angle
glaucoma. A population-based ve-year follow-up study. Acta
Ophthalmol (Copenh) 1993;71:18995.
5. Gordon MO, Beiser JA, Brandt JD, et al, Ocular Hypertension
Treatment Study Group. The Ocular Hypertension Treatment
Study: baseline factors that predict the onset of primary open-
angle glaucoma. Arch Ophthalmol 2002;120:71420.
6. Leske MC, Connell AM, Wu SY, et al. Risk factors for
open-angle glaucoma. The Barbados Eye Study. Arch Oph-
thalmol 1995;113:91824.
7. Lee AJ, Rochtchina E, Wang JJ, et al. Does smoking affect
intraocular pressure? Findings from the Blue Mountains Eye
Study. J Glaucoma 2003;12:20912.
8. Mitchell P, Smith W, Chey T, Healey PR. Open-angle glau-
coma and diabetes: the Blue Mountains Eye Study, Australia.
Ophthalmology 1997;104:7128.
9. Mitchell P, Hourihan F, Sandbach J, Wang JJ. The relation-
ship between glaucoma and myopia: the Blue Mountains Eye
Study. Ophthalmology 1999;106:20105.
10. Mitchell R, Rochtchina E, Lee A, et al. Iris color and intraoc-
ular pressure: the Blue Mountains Eye Study. Am J Ophthal-
mol 2003;135:3846.
11. Mitchell P, Cumming RG, Mackey DA. Inhaled corticoste-
roids, family history, and risk of glaucoma. Ophthalmology
1999;106:23016.
12. Cumming RG, Mitchell P. Alcohol, smoking, and cataracts:
the Blue Mountains Eye Study. Arch Ophthalmol 1997;115:
1296303.
13. Rowe NG, Mitchell PG, Cumming RG, Wang JJ. Diabetes,
fasting blood glucose and age-related cataract: the Blue Moun-
tains Eye Study. Ophthalmic Epidemiol 2000;7:10314.
14. Lim R, Mitchell P, Cumming RG. Refractive associations
with cataract: the Blue Mountains Eye Study. Invest Ophthal-
mol Vis Sci 1999;40:30216.
15. Cumming RG, Mitchell P, Lim R. Iris color and cataract: the
Blue Mountains Eye Study. Am J Ophthalmol 2000;130:
2378.
16. Cumming RG, Mitchell P, Leeder SR. Use of inhaled corti-
costeroids and the risk of cataracts. N Engl J Med 1997;337:
814.
17. Bernth-Petersen P, Bach E. Epidemiologic aspects of cataract
surgery. III: frequencies of diabetes and glaucoma in a cataract
population. Acta Ophthalmol (Copenh) 1983;61:40616.
18. van Heyningen R, Harding JJ. A case-control study of cataract
in Oxfordshire: some risk factors. Br J Ophthalmol 1988;72:
8048.
19. Harding JJ, Harding RS, Egerton M. Risk factors for cataract
in Oxfordshire: diabetes, peripheral neuropathy, myopia, glau-
coma and diarrhoea. Acta Ophthalmol (Copenh) 1989;67:
5107.
20. Weih LM, Mukesh BN, McCarty CA, Taylor HR. Association
of demographic, familial, medical, and ocular factors with
intraocular pressure. Arch Ophthalmol 2001;119:87580.
21. Leske MC, Wu SY, Nemesure B, Hennis A, Barbados Eye
Studies Group. Risk factors for incident nuclear opacities.
Ophthalmology 2002;109:13038.
22. Shaffer RN, Rosenthal G. Comparison of cataract incidence in
normal and glaucomatous population. Am J Ophthalmol 1970;
69:36871.
23. Harding JJ, Egerton M, van Heyningen R, Harding RS. Dia-
betes, glaucoma, sex, and cataract: analysis of combined data
from two case control studies. Br J Ophthalmol 1993;77:26.
24. Klein BE, Klein R, Moss SE. Incident cataract surgery: the
Beaver Dam Eye Study. Ophthalmology 1997;104:57380.
25. Klein BE, Klein R, Lee KE, Danforth LG. Drug use and
ve-year incidence of age-related cataracts: the Beaver Dam
Eye Study. Ophthalmology 2001;108:16704.
26. 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.
27. Lichter PR, Musch DC, Gillespie BW, et al, CIGTS Study
Group. Interim clinical outcomes in the Collaborative Initial
Glaucoma Treatment Study comparing initial treatment ran-
domized to medications or surgery. Ophthalmology 2001;108:
194353.
28. Klein BE, Klein R, Linton KL. Intraocular pressure in an
American community. The Beaver Dam Eye Study. Invest
Ophthalmol Vis Sci 1992;33:22248.
29. Ellant JP, Obstbaum SA. Lens-induced glaucoma. Doc Oph-
thalmol 1992;81:31738.
30. Nakabayashi M. Nuclear cataract as a cause of senile glau-
coma. Ann Ophthalmol 1981;13:597602.
Chandrasekaran et al Intraocular Pressure and Incident Cataract
423
31. Rosenbaum JT, Samples JR, Seymour B, et al. Chemotactic
activity of lens proteins and the pathogenesis of phacolytic
glaucoma. Arch Ophthalmol 1987;105:15824.
32. Epstein DL. Diagnosis and management of lens-induced glau-
coma. Ophthalmology 1982;89:22730.
33. Mitchell P, Smith W, Attebo K, Healey PR. Prevalence of
open-angle glaucoma in Australia. The Blue Mountains Eye
Study. Ophthalmology 1996;103:16619.
34. Mitchell P, Cumming RG, Attebo K, Panchapakesan J. Prev-
alence of cataract in Australia: the Blue Mountains Eye Study.
Ophthalmology 1997;104:5818.
35. Mitchell P, Smith W, Attebo K, Wang JJ. Prevalence of
age-related maculopathy in Australia. The Blue Mountains
Eye Study. Ophthalmology 1995;102:145060.
36. Klein BE, Klein R, Linton KL, et al. Assessment of cataracts
from photographs in the Beaver Dam Eye Study. Ophthalmol-
ogy 1990;97:142833.
37. Panchapakesan J, Cumming RG, Mitchell P. Reproducibility
of the Wisconsin cataract grading system in the Blue Moun-
tains Eye Study. Ophthalmic Epidemiol 1997;4:11926.
38. Klein BE, Klein R, Lee KE. Incidence of age-related cataract: the
Beaver Dam Eye Study. Arch Ophthalmol 1998;116:21925.
39. Zeger SL, Liang KY. Longitudinal data analysis for discrete
and continuous outcomes. Biometrics 1986;42:12130.
40. Mitchell P, Lee AJ, Rochtchina E, Wang JJ. Open-angle
glaucoma and systemic hypertension: the Blue Mountains Eye
Study. J Glaucoma 2004;13:31926.
41. Mitchell P, Wang JJ, Hourihan F. The relationship between
glaucoma and pseudoexfoliation: the Blue Mountains Eye
Study. Arch Ophthalmol 1999;117:131924.
42. Wang JJ, Mitchell P, Smith W. Is there an association between
migraine headache and open-angle glaucoma? Findings from the
Blue Mountains Eye Study. Ophthalmology 1997;104:17149.
43. Hourihan F, Mitchell P. Factors associated with use of glaucoma
medications in a population of older people: the Blue Mountains
Eye Study. Aust N Z J Ophthalmol 1999;27:1769.
44. Hansen TE, Naeser K, Nilsen NE. Intraocular pressure 2 1/2
years after extracapsular cataract extraction and sulcus im-
plantation of posterior chamber intraocular lens. Acta Oph-
thalmol (Copenh) 1991;69:2258.
45. Handa J, Henry JC, Krupin T, Keates E. Extracapsular cataract
extraction with posterior chamber lens implantation in patients
with glaucoma. Arch Ophthalmol 1987;105:7659.
46. Dimitrov PN, Mukesh BN, Taylor HR, McCarty CA. Intraoc-
ular pressure before and after cataract surgery in participants
of the Melbourne Visual Impairment Project. Clin Experiment
Ophthalmol 2001;29:12832.
47. Lee SM, Lin SY, Cheng CL, Liang RC. Progressive changes
in secondary conformation and composition of the senile
cataractous human lens capsules. Acta Ophthalmol Scand
1996;74:5426.
48. Lin SY, Li MJ, Liang RC, Lee SM. Non-destructive analysis
of the conformational changes in human lens lipid and
protein structures of the immature cataracts associated with
glaucoma. Spectrochim Acta A Mol Biomol Spectrosc 1998;
54A:150917.
49. Lee SM, Lin SY, Li MJ, Liang RC. Possible mechanism of
exacerbating cataract formation in cataractous human lens
capsules induced by systemic hypertension or glaucoma. Oph-
thalmic Res 1997;29:8390.
50. Rodriguez-Sargent C, Estape-Wainwright E, Cangiano JL, et
al. Lenticular rubidium uptake in hypertensive cataract-
prone salt-sensitive rats. J Hypertens Suppl 1988;6:S2725.
51. Harding JJ. Cataract: Biochemistry, Epidemiology and Phar-
macology. London: Chapman and Hall; 1991:71124.
52. Tabandeh H, Thompson GM, Heyworth P, et al. Water content,
lens hardness and cataract appearance. Eye 1994;8:1259.
53. Dielemans I, Vingerling JR, Algra D, et al. Primary open-
angle glaucoma, intraocular pressure, and systemic blood
pressure in the general elderly population. The Rotterdam
Study. Ophthalmology 1995;102:5460.
54. Rochtchina E, Mitchell P, Wang JJ. Relationship between age
and intraocular pressure: the Blue Mountains Eye Study. Clin
Experiment Ophthalmol 2002;30:1735.
55. West SK, Munoz B, Istre J, et al. Mixed lens opacities and
subsequent mortality. Arch Ophthalmol 2000;118:3937.
56. Wang JJ, Mitchell P, Simpson JM, et al. Visual impairment,
age-related cataract, and mortality. Arch Ophthalmol 2001;
119:118690.
Ophthalmology Volume 113, Number 3, March 2006
424