Association of Antihypertensive Medication

with Retinal Nerve Fiber Layer and Ganglion

Cell-Inner Plexiform Layer Thickness
Rachel S. Chong, MMed(Ophth), PhD,1,2,3 Miao-Li Chee, BSc,1,2,3 Yih-Chung Tham, PhD,1,2,3
Shivani Majithia, OD,1,2,3 Sahil Thakur, MBBS, MS(Ophth),1,2,3 Zhen Ling Teo, MBBS,1,2,3
Zhi Da Soh, BSc, MPH,1,2,3 Jacqueline Chua, PhD,1,2,3 Bingyao Tan, PhD,2,4,5 Damon W.K. Wong, PhD,2,4,5
Leopold Schmetterer, PhD,1,2,3,4,5,6,7,8,9 Charumathi Sabanayagam, MBBS, PhD,1,2,3
Ching-Yu Cheng, MD, PhD1,2,3

Purpose: To evaluate the association between different classes of antihypertensive medication with retinal
nerve fiber layer (RNFL) and ganglion cell-inner plexiform layer (GC-IPL) thickness in a nonglaucomatous multi-
ethnic Asian population.
Design: Population-based, cross-sectional study.
Participants: A total of 9144 eyes for RNFL analysis (2668 Malays, 3554 Indians, and 2922 Chinese) and
8549 eyes for GC-IPL analysis (2460 Malays, 3230 Indians, and 2859 Chinese) aged 44 to 86 years.
Methods: Participants underwent standardized systemic and ocular examinations and interviewer-
administered questionnaires for collection of data on medication and other variables. Intraocular pressure (IOP)
readings were obtained by Goldmann applanation tonometry before pupil dilation for fundoscopy and OCT im-
aging. Blood pressure (BP) was measured with an automatic BP monitor. Mean arterial pressure (MAP) was
defined as diastolic BP plus 1/3 (systolic BP e diastolic BP). Regression models were used to investigate the
association of antihypertensive medication with OCT measurements of RNFL and GC-IPL.
Main Outcome Measures: Average and sectoral RNFL and GC-IPL thickness.
Results: After adjusting for age, gender, ethnicity, MAP, IOP, body mass index (BMI), and presence of
diabetes, we found that participants taking any type of antihypertensive medication (b ¼ 0.83; 95% confidence
interval [CI], 1.46 to 0.02; P ¼ 0.01), specifically angiotensin-converting enzyme inhibitors (ACEIs) (b ¼ 1.66;
95% CI, 2.57 to 0.75; P < 0.001) or diuretics (b ¼ 1.38; 95% CI, 2.59 to 0.17; P < 0.05), had thinner
average RNFL in comparison with participants who were not receiving antihypertensive treatment. Use of a
greater number of antihypertensive medications was significantly associated with thinner average RNFL (P for
trend ¼ 0.001). This association was most evident in the inferior RNFL quadrant in participants using ACEIs
(b ¼ 2.44; 95% CI, 3.99 to 0.89; P ¼ 0.002) or diuretics (b ¼ 2.76; 95% CI, 4.76 to 0.76; P ¼ 0.007). A
similar trend was noted in our analysis of macular GC-IPL thickness.
Conclusions: Use of 2 or more antihypertensive medications, ACEI, and diuretics were associated with a
loss of structural markers of retinal ganglion cell health in a multiethnic Asian population. Ophthalmology 2020;-
:1e8 ª 2020 by the American Academy of Ophthalmology

Supplemental material available at www.aaojournal.org.

Glaucoma and hypertension are chronic, age-related condi- comorbidities,7 although this could have significant
tions that are associated with increased patient morbidity implications for optic nerve perfusion, particularly in
and global healthcare burden.1-4 The relationship between patients who have glaucoma. Various studies have
blood pressure (BP) and glaucoma is complex because of proposed that lower systemic BP is associated with both a
the presence of multiple interacting factors that may modify higher prevalence of glaucoma and a greater chance of
both intraocular pressure (IOP) and ocular perfusion glaucoma progression,8-11 although this is a nonlinear rela-
pressure.5,6 tionship that may be affected by treatment with antihyper-
Diurnal variation of IOP and BP could result in periods tensive medication. However, epidemiologic studies have
of decreased ocular perfusion in certain individuals under presented conflicting evidence on the incidence of glaucoma
normal physiologic or treatment conditions. Research sug- associated with the use of antihypertensive medicines.12-14
gests that more aggressive treatment of systemic hyperten- Different classes of antihypertensive medication pro-
sion significantly lowers the rate of cardiovascular duce varying effects on a patient’s BP profile, in terms of

ª 2020 by the American Academy of Ophthalmology https://doi.org/10.1016/j.ophtha.2020.07.051 1

Published by Elsevier Inc. ISSN 0161-6420/20
 Ophthalmology Volume -, Number -, Month 2020

the duration and magnitude of BP lowering.15 We aimed Blood Pressure Measurement

to further investigate the possible relationship between
Systolic blood pressure (SBP) and diastolic blood pressure (DBP)
BP lowering and glaucoma by examining variations in were measured using a digital automatic BP monitor (Dinamap
structural measures of the peripapillary retinal nerve model Pro Series DP110X-RW, 100V2; GE Medical Systems In-
fiber layer (RNFL) and ganglion cell-inner plexiform formation Technologies Inc, Milwaukee, WI) using a standardized
layer (GC-IPL) in relation to different types of antihy- protocol. Hypertension was defined as SBP  140 mmHg or DBP
pertensive medication. These findings may have impor-  90 mmHg based on the guidelines for treatment or by physician
tant implications in determining which classes of diagnosis. Subjects were seated and rested for at least 5 minutes
systemic antihypertensive medication are most suitable before their BP was taken. Blood pressure was measured twice,
and, conversely, potentially deleterious for patients with with the second reading taken after a lapse of 1 to 2 minutes. A
glaucoma. third measurement was taken if the previous 2 SBP readings
differed by > 10 mmHg or DBP by > 5 mmHg. The SBP and DBP
of the participant were defined as the mean between the 2 closest
readings. Pulse pressure was defined as SBP minus DBP level,
Methods whereas mean arterial pressure (MAP) was defined as DBP plus 1/
3 (pulse pressure).
Written, informed consent was obtained from each participant, and
the studies adhered to the Declaration of Helsinki. Ethical approval
was obtained from the SingHealth Centralized Institutional Review
OCT Imaging
Board.
OCT imaging was performed using commercially available
spectral domain OCT instrument (Cirrus HD-OCT; Carl Zeiss
Study Population Meditec, Dublin, CA) after pupil dilation. An optic disc scan
was acquired using the optic disc cube 200  200 scan protocol,
We conducted a population-based study using data from the
and a macular scan was obtained using macular cube 512  128
Singapore Epidemiology of Eye Diseases Study of adults aged 40
scan protocol, covering a measurement area of 6  6 mm2. The
years or older from 3 major Asian ethnic groups in Singapore:
optic nerve head and RNFL algorithms native to Cirrus HD-
Chinese, Indians, and Malays. The methodologies of these cohort
OCT were used to measure RNFL and GC-IPL parameters
studies have been reported in detail.16-18
(average and quadrants peripapillary RNFL thickness; average,
Data for the current study were derived from the Singapore
minimum, and sectoral GC-IPL thickness) automatically based
Epidemiology of Eye Diseases Study as follows. Spectral-domain
on a reference plane that was determined at 200 mm above the
OCT was incorporated into the examination protocol of our Chi-
level of Bruch’s membrane plane. All images were reviewed
nese cohort as baseline (2009e2011; n ¼ 3353; age range ¼ 44 to
using the Cirrus HD-OCT Review Software, with optic disc and
86 years; response rate ¼ 72.8%), Malay cohort at 6-year follow-
foveal centration and segmentation checked at this time.
up17 (2011e2014; n ¼ 1901; age range ¼ 46e89 years; response
Detailed descriptions of the measurement algorithms have been
rate ¼ 78.7%), and Indian cohort at 6-year follow-up18
described.19,20 To account for the possible effect of
(2013e2015; n ¼ 2200; age range ¼ 48e90 years; response rate
antihypertensive medication or BP on anatomic confounders
75.2%). Participants without OCT scans were excluded for this
including retinal vessels, additional analysis was performed
analysis. After this exclusion, OCT data of RNFL and GC-IPL
using images that were compensated for these structures, as
measures were available for 9936 eyes (2916 Malays, 3892 In-
previously described.21
dians, and 3128 Chinese). In addition, eyes with OCT segmenta-
tion errors and poor scan quality with signal strength less than 6,
eyes with glaucoma or on glaucoma/ocular hypertension treatment,
and eyes with other neurodegenerative diseases or macular pa- Assessment of the Use of Medication and
thology were excluded from OCT analysis. This left a total of 9312 Systemic Medical Conditions
eyes for RNFL analysis (2690 Malays, 3673 Indians, and 2949
Chinese) and 8693 eyes for GC-IPL analysis (2477 Malays, 3332 Relevant sociodemographic and medical information were
Indians, and 2884 Chinese) (Fig 1). collected through a detailed interviewer-administered question-
naire that was conducted in the patient’s language of choice
(English, Chinese, Malay, or Tamil) by bilingual interviewers.
Ophthalmic Assessment Participants were encouraged to bring along a list of medications
for their study visit. Systemic examination was performed, and
Subjects underwent a comprehensive ocular examination at the blood samples were collected for biochemistry analysis. Anti-
Singapore Eye Research Institute. Intraocular pressure measure- hypertensive medications were recorded on the basis of the
ments were obtained using a Goldmann applanation tonometer. number of medications taken (none, 1, 2, or more) and class of
Fundoscopy was performed after pupil dilation with tropicamide medication (angiotensin-converting enzyme inhibitors [ACEIs],
1% and phenylephrine hydrochloride 2.5%, and the optic disc was angiotensin II receptor blockers, calcium channel blockers
evaluated using a þ78 diopter lens at 16 magnification. The [CCBs], diuretics or beta receptor blockers [BRBs]). Diabetes
diagnosis of glaucoma was made according to the International mellitus was defined as random glucose levels of 11.1 mmol/l or
Society for Geographical and Epidemiological Ophthalmology greater, self-reported physician diagnosis of diabetes, self-
criteria, primarily based on optic disc appearance and the results of reported use of hypoglycemic medication, or hemoglobin A1c
static automated perimetry (Swedish Interactive Threshold Algo- value of  6.5. Each participant’s height was measured in
rithm standard 24-2, Humphrey Field Analyzer II). Ocular hyper- centimeters using a wall-mounted measuring tape, and weight
tension was defined as IOP > 21 mmHg. Subjects with glaucoma was measured in kilograms using a digital scale. Body mass
or ocular hypertension were excluded from analysis for the purpose index (BMI) was calculated as weight in kilograms divided by
of this study. height in meters squared.

2
 Chong et al 
Antihypertensive Drug Use with RNFL and GC-IPL Thinning

Figure 1. Study population characteristics. GC-IPL ¼ ganglion cell-inner plexiform layer; IOP ¼ intraocular pressure; ONH ¼ optic nerve head;
RNFL ¼ retinal nerve fiber layer.

Statistical Analysis compensated for anatomic confounders as described in the

“Methods” section.21
All statistical analysis was performed using Stata 14.0 (StataCorp
LP, College Station, TX). The mean and standard deviation were
reported for continuous demographic, systemic, and ocular Results
characteristics of included participants, and frequency and per-
centage were reported for categoric characteristics. Linear A total of 4699 participants were included; of them, the mean age
regression models were performed to evaluate the associations was 58.8  8.5 years and 48.8% were male. Participants taking 2
between antihypertensive medications, which includes the spe- or more types of medication were significantly older (64.43  8.65
cific class of medication and number of medications, with RNFL years) than participants who were not taking antihypertensive
and GC-IPL thickness parameters. Models were initially adjusted medication (56.39  7.41 years) or only 1 type of medication
for age, gender, and ethnicity and subsequently further adjusted (61.83  8.52 years; P < 0.001). However, a subgroup analysis
for age, gender, and selected potential confounders based on past comparing the effect of antihypertensive medication in subjects
literature review, which included MAP, diabetes, BMI, and IOP. aged less than 60 years (b ¼ 1.71; 95% confidence interval
We further generated a locally weighted scatterplot smoothing [CI], 3.15 to 0.27; P ¼ 0.02) and those aged 60 years or more
(LOWESS) plot to evaluate the relationship between MAP and (b ¼ 1.25; 95% CI, 2.49 to 0.02; P ¼ 0.046) showed a
RNFL and did not observe substantial nonlinearity between the consistent association between ACEI use and thinner RNFL
2, suggesting that a linear regression model is appropriate for this measurements in both groups. Among included participants, 64.8%
evaluation. Because eye-specific data were used, generalized were not taking any form of antihypertensive medications, 18.6%
estimating equations with exchangeable correlation structures were taking 1 type of medication, 16.6% were taking 2 or more
and Gaussian link were applied in the linear regression models to types of medications, 11.1% were taking ACEIs, 6.9% were taking
account for inter-eye correlation within individuals. A subgroup angiotensin receptor blockers, 16.5% were taking CCBs, 6.0%
analysis was performed among those images that had been were taking diuretics, and 15.5% were taking BRBs. Subjects who

3
 Ophthalmology Volume -, Number -, Month 2020

required more than 1 type of medication because of inadequate BP- Table 1. Demographic and Characteristics of the 4699 Study
lowering response were most commonly given a combination of Subjects
CCB and BRB. Average peripapillary RNFL thickness was 92.4 
11.0 mm, and average macular GC-IPL thickness was 80.3  7.2 Characteristics Mean ± SD or n (%)
mm (Table 1).
Age, yrs 58.8  8.5
We examined the association of antihypertensive medication
Ethnicity
use and RNFL thickness (Table 2), where we found that the use Chinese 1648 (35.1)
of any antihypertensive medication was significantly associated Malay 1320 (28.1)
with thinner average RNFL (b ¼ 0.83; 95% CI, 1.46 Indian 1731 (36.8)
to 0.02; P ¼ 0.01, adjusted for age, gender, ethnicity, Gender, female 2407 (51.2)
diabetes, IOP, BMI, and MAP). Consistently, a greater number BMI, kg/m2 25.6  4.6
of antihypertensive medication use was associated with thinner Hypertension, n 2754 (58.7)
RNFL (P for trend < 0.001). On the basis of our analysis of Antihypertensive Medication Use, n 1622 (35.2)
specific types of antihypertensive medication, we found that No. of Antihypertensive
the use of ACEIs (b ¼ 1.66; 95% CI, 2.57 to 0.75; P < Medications, n
0.001) and diuretics (b ¼ 1.38; 95% CI, 2.59 to 0.17; P No medication 2987 (64.8)
< 0.05) were significantly associated with thinner RNFL 1 medication 857 (18.6)
(Model 2 in Table 2). These trends of association were  2 medications 765 (16.6)
consistently observed when evaluating RNFL thickness of the Antihypertensive Medication Types
inferior quadrant as the outcome of interest, particularly in ACEIs 511 (11.1)
subjects using ACEIs (b ¼ 2.44; 95% CI, 3.99 to 0.89; Angiotensin II receptor blocker 317 (6.9)
P ¼ 0.002) or diuretic medication (b ¼ 2.76; 95% CI, 4.76 CCBs 762 (16.5)
Diuretics 277 (6.0)
to 0.76; P ¼ 0.007) (Table S1, available at
BRBs 715 (15.5)
www.aaojournal.org).
Systemic Assessment
A similar trend was noted in our analysis of the macular GC- MAP, mmHg 96.9  11.4
IPL thickness (Table 3), where there was a significant association SBP, mmHg 135.3  18.4
between the use of any antihypertensive medication with average DBP, mmHg 77.8  9.9
macular GC-IPL thickness (b ¼ 0.46; 95% CI, 0.89 to 0.03; Diabetes 1271 (27.0)
P ¼ 0.037, adjusted for age, gender, ethnicity, diabetes, IOP, Hemoglobin A1c, mmol/mol 6.2  1.2
BMI, and MAP). Consistently, a greater number of antihyper- LDL-cholesterol, mmol/l 3.4  1.0
tensive medication use was associated with thinner average GC- RNFL Thickness, mm (n [ 9144 eyes)
IPL measurements (P for trend ¼ 0.005). Specifically, the use of Average 92.4  11.0
ACEIs (b ¼ 0.97; 95% CI, 1.62 to 0.31; P ¼ 0.004) or Inferior quadrant 119.0  18.2
diuretics (b ¼ 1.05; 95% CI, 1.81 to 0.29; P ¼ 0.007) had Superior quadrant 115.4  17.5
significant associations with average GC-IPL thickness on OCT GC-IPL Thickness, mm (n [ 8549 eyes)
(Model 2 in Table 3). These trends were also found to be Average 80.3  7.2
consistently observed on comparing the effect of specific Inferior hemisphere 79.7  7.3
antihypertensive medications on inferior and superior Superior hemisphere 80.9  7.4
hemisphere GC-IPL thickness (Table S2, available at
www.aaojournal.org). ACEI ¼ angiotensin-converting enzyme inhibitor; BMI ¼ body mass in-
Additional analysis was performed on a subgroup that had OCT dex; BRB ¼ beta receptor blocker; CCB ¼ calcium channel blocker;
images further compensated for anatomic confounders including DBP ¼ diastolic blood pressure; GC-IPL ¼ ganglion cell-inner plexiform
retinal vessels, as previously described.21 Overall, in the fully layer; LDL ¼ low-density lipoprotein; MAP ¼ mean arterial pressure;
adjusted model, identical trends were observed compared with RNFL ¼ retinal nerve fiber layer; SBP ¼ systolic blood pressure; SD ¼
the main analysis where the use of more antihypertensive standard deviation.
medications (P for trend ¼ 0.045), ACEIs (b ¼ 1.80; 95% Data are mean  SD (%) for continuous variables and number (%) for
categoric variables. MAP ¼ DBP plus 1/3 (SBP e DBP).
CI, 3.064 to 0.56; P ¼ 0.004), and diuretics (b ¼ 1.98;
95% CI, 3.73 to 0.22; P ¼ 0.027) was associated with
thinner average RNFL thickness (Model 2 in Table S3, available
at www.aaojournal.org). Discussion
We further divided the antihypertensive medication user
group into finer subgroups of “treated and well-controlled hy- Previous studies have suggested that there may be an as-
pertension” (defined as SBP  140 and DBP  90) and “treated sociation between BP and glaucoma, either directly or in
but uncontrolled hypertension” (defined as SBP > 140 or DBP conjunction with IOP.9-11 However, several of these reports
> 90) for comparison with nonantihypertensive medication did not take into account the effects of other potentially
users. We observed that antihypertensive medication users, confounding factors, including the use of antihypertensive
whether having well or poorly controlled BP, both had signifi- treatment. More recent research investigating the incidence
cantly thinner RNFL and GC-IPL profiles compared with in-
of glaucoma in relation to systemic drug use has used large
dividuals not taking antihypertensive medication (all P < 0.001,
Table S4, available at www.aaojournal.org). There was also no databases of patient records, in the absence of clinical
significant difference in RNFL and GC-IPL thickness between measurements of BP, IOP, or OCT parameters, with
treated individuals of well-controlled and poorly controlled discordant results.13,14 We report the findings of a major
hypertension after adjusting for relevant confounders (Table S5, population-based study that demonstrates an association
available at www.aaojournal.org). between taking antihypertensive medications, specifically

4
 Chong et al 
Antihypertensive Drug Use with RNFL and GC-IPL Thinning

Table 2. Association between Antihypertensive Medication Use and Average Retinal Nerve Fiber Layer Thickness

Model 1 Model 2
b (95% CI) P b (95% CI) P
Any Antihypertensive Medication 0.98 (1.58 to 0.38) 0.001 0.83 (1.46 to 0.20) 0.010
No. of Antihypertensive Medications
No medications Reference Reference
1 medication 0.48 (1.20 to 0.23) 0.185 0.40 (1.14 to 0.33) 0.279
 2 medications 1.75 (2.55 to 0.94) <0.001 1.58 (2.42 to 0.73) <0.001
P-trend <0.001 <0.001
Antihypertensive Medication Type
ACEIs 1.83 (2.71 to 0.96) <0.001 1.66 (2.57 to 0.75) <0.001
Angiotensin II receptor blocker 0.76 (1.78 to 0.27) 0.150 0.47 (1.51 to 0.58) 0.379
CCBs 0.77 (1.52 to 0.01) 0.046 0.62 (1.38 to 0.13) 0.106
Diuretics 1.65 (2.86 to 0.44) 0.007 1.38 (2.59 to 0.17) 0.025
BRBs 0.61 (1.38 to 0.16) 0.118 0.54 (1.32 to 0.23) 0.167

ACEI ¼ angiotensin-converting enzyme inhibitor; BMI ¼ body mass index; BRB ¼ beta receptor blocker; CCB ¼ calcium channel blocker; CI ¼ con-
fidence interval; IOP ¼ intraocular pressure; MAP ¼ mean arterial pressure; RNFL ¼ retinal nerve fiber layer.
b denotes the change in average RNFL thickness (in mm), per unit change in exposure variables.
Model 1 adjusted for age, gender, ethnicity.
Model 2 adjusted for age, gender, ethnicity, diabetes, IOP, BMI, and MAP.

the use of ACEI and diuretics on RNFL and GC-IPL been effectively random and was unlikely to alter the
thickness measured on OCT after adjusting for BP and IOP. interpretation of our findings. However, our results
Our study has important implications for the use of suggest that the use of certain types of antihypertensive
antihypertensive medications in patients with glaucoma, medication, particularly ACEI as first-line therapy, may
who already possess structurally impaired retinal ganglion need to be carefully considered in patients with vulnerable
cells as a result of their disease. With growing life expec- optic nerves. Treatment efficacy and safety are paramount in
tancy worldwide, the number of individuals with both deciding on the best therapeutic options for patients; how-
glaucoma and hypertension is projected to increase steadily ever, to date, the relationship between antihypertensive
in the near future.2,4 The Ministry of Health Clinical medications and their effect on ocular health has not been
Practice Guidelines for Hypertension in Singapore (2017) studied in depth.
state that in the absence of compelling indications or It has been suggested that lower ocular perfusion pressure
contraindications for a particular drug, any 1 of the 5 or DBP may be associated with greater glaucoma prevalence
major pharmacologic classes of antihypertensive drugs or a higher rate of progression. Studies using ambulatory BP
may be considered as initial treatment. Therefore, the measurements have further concluded that the magnitude or
prescribing practice of ACEI in our cohort should have duration of nocturnal decreases in BP, or “nocturnal dippers,”

Table 3. Association between Antihypertensive Medication Use and Average Macular Ganglion Cell-Inner Plexiform Layer Thickness

Model 1 Model 2
b (95% CI) P b (95% CI) P
Any Antihypertensive Medication 0.57 (0.98 to 0.15) 0.007 0.46 (0.89 to 0.03) 0.037
No. of Antihypertensive Medications
No medications Reference Reference
1 medication 0.39 (0.89 to 0.11) 0.130 0.32 (0.83 to 0.19) 0.226
 2 medications 0.94 (1.49 to 0.40) 0.001 0.82 (1.39 to 0.25) 0.005
P-trend 0.001 0.005
Antihypertensive Medication Type
ACEIs 1.11 (1.74 to 0.48) 0.001 0.97 (1.62 to 0.31) 0.004
Angiotensin II receptor blocker 0.70 (1.42 to 0.02) 0.055 0.56 (1.29 to 0.18) 0.138
CCBs 0.08 (0.59 to 0.44) 0.772 0.02 (0.50 to 0.55) 0.936
Diuretics 1.20 (1.96 to 0.44) 0.002 1.05 (1.81 to 0.29) 0.007
BRBs 0.25 (0.78 to 0.29) 0.370 0.14 (0.69 to 0.41) 0.619

ACEI ¼ angiotensin-converting enzyme inhibitor; BMI ¼ body mass index; BRB ¼ beta receptor blocker; CCB ¼ calcium channel blocker; CI ¼ con-
fidence interval; GC-IPL ¼ ganglion cell-inner plexiform layer; IOP ¼ intraocular pressure; MAP ¼ mean arterial pressure.
b denotes the change in average GC-IPL thickness (in mm), per unit change in exposure variables.
Model 1 adjusted for age, gender, and ethnicity.
Model 2 adjusted for age, gender, ethnicity, diabetes, IOP, BMI, and MAP.

5
 Ophthalmology Volume -, Number -, Month 2020

has the strongest association with progressive glaucomatous also included adjustments for MAP and IOP measurements.
visual loss.11,22 This was in agreement with a meta-analysis This is an important consideration, because systemic anti-
that reported a nocturnal reduction in BP of more than 10%, hypertensives may mask the effect of BP reduction on
but not mean SBP or DBP, was associated with glaucoma structural markers of retinal ganglion cell (RGC) health, in
progression.23 On the basis of our results, it is possible that addition to altering IOP and ocular perfusion pressure.
ACEI and diuretics exert different effects on the ocular Furthermore, a diagnosis of glaucoma was made on the
microvasculature structurally24 or functionally in terms of basis of clinical examination by glaucoma-trained ophthal-
diurnal BP regulation in comparison with other classes of mologists and not electronic records, and we excluded the
antihypertensive drugs. The Hypertension Objective effect of glaucoma medication by selecting participants who
treatment based on Measurement by Electrical Devices of were not receiving IOP-lowering treatment for final
BP (HOMED-BP) study, for example, showed that analysis.
monotherapy with ARB or ACEI antihypertensive agents Our study was limited by a lack of ambulatory BP data or
resulted in greater morning BP variability in comparison information on what time of day antihypertensive medica-
with CCB.25 Certain combinations of antihypertensive tion was administered. Thus, it was not possible to further
drugs also appear superior in reducing BP variability.26 deduce how many of these patients were in fact “nocturnal
How this affects ocular perfusion remains uncertain and dippers” or to what extent this may have been influenced by
should be the subject of future studies in patients with the timing of their antihypertensive medication. Future
glaucoma. studies to determine if these factors may influence RNFL or
Certain types of antihypertensive drugs may also alter GC-IPL measurements in patients would add considerable
neuronal health in the eye, independent of BP or IOP, value to our findings. Our findings were based on cross-
although research in this area has not led to conclusive sectional data at this stage; therefore, further work is
findings thus far. Preclinical models of neurodegenerative required to determine any true causal association. For
disease have implied that application of ACEIs could have a example, it remains to be seen how the duration or dosage of
neuroprotective effect.27,28 However, angiotensin- antihypertensive medications may affect OCT measure-
converting enzyme is also responsible for cleaving amy- ments of RNFL or GC-IPL thickness. It should be noted that
loid beta-protein in vivo,29,30 and longitudinal clinical data chronic history of poorly controlled hypertension is an
have suggested that ACEI use in an elderly Japanese important potential confounder and was not fully taken into
population is associated with increased cognitive account in our study because of the cross-sectional study
impairment.31 Therefore, prolonged use of ACEIs may design. We have attempted to elucidate this aspect by
result in increased deposition of amyloid beta-protein in comparing RNFL and GC-IPL thickness in non-
the retina, which has been observed in patients with glau- antihypertensive medication users, treated hypertensive pa-
coma,32 although there is clearly a need for more extensive tients with well-controlled BP, and treated hypertensive
studies on this topic. patients with poorly controlled BP. The results of these
Our results do not concur with the findings of several analyses suggest that the observed associations between
other published studies that examined the association be- antihypertensive medication use with thinner RNFL and
tween systemic antihypertensive drug use and the incidence GC-IPL were unlikely to be solely due to hypertensive
of glaucoma, where Zhen et al13 notably found that CCB in medications being a surrogate marker for uncontrolled hy-
particular, and not ACEI or diuretics, was associated with pertension. However, a future multivisit longitudinal study
presumed advanced or uncontrolled primary open-angle is warranted to evaluate the sequential relationships among
glaucoma. This discrepancy may arise from the different medication intake status, hypertension control status, and
genetic background of the populations studied. It is well RNFL or GC-IPL profiles.
established that certain ethnic groups are less responsive to In conclusion, we have shown that certain types of
ACEI treatment and may in fact have poorer clinical out- antihypertensive medication, in particular ACEIs and di-
comes on this type of antihypertensive medication,33,34 uretics, may have a strong association with a loss in struc-
although our own subgroup analysis comparing different tural measures of retinal ganglion cell health. Although the
Asian ethnicities did not show any significant difference. observed effects of medication on RNFL and GC-IPL
Furthermore, in contrast to the “nocturnal dipper” pattern thickness may not be clinically substantial at present, our
of BP profile reported in other studies arising from the results could indicate a potential vulnerability of RGCs to
West, Japanese patients with glaucoma appear to have drug-induced BP alterations. Our findings suggest future
increased nighttime BP instead.35 These intriguing avenues of research into how these treatments may poten-
differences suggest a possible genetic basis for variation in tially affect visual outcomes in patients with compromised
the ocular response to antihypertensive medication and optic nerves from glaucoma or other optic neuropathies.
should be investigated further.
References
Study Strengths and Limitations
The main strengths of our study include a large population 1. Kirkland EB, Heincelman M, Bishu KG, et al. Trends in
dataset encompassing multiple Asian ethnic groups and a healthcare expenditures among US adults with hypertension:
comprehensive list of potential confounding factors that national estimates 2003e2014. J Am Heart Assoc. 2018;7:
have previously remained unaccounted for. Our analysis e008731.

6
 Chong et al 
Antihypertensive Drug Use with RNFL and GC-IPL Thinning

2. Bloch MJ. Worldwide prevalence of hypertension exceeds 1.3 population: The Singapore Epidemiology of Eye Diseases
billion. J Am Soc Hypertens. 2016;10:753e754. Study. Ophthalmology. 2020;127(8):1064e1076.
3. Varma R, Lee PP, Goldberg I, Kotak S. An assessment of the 21. Chua J, Schwarzhans F, Nguyen DQ, et al. Compensation of
health and economic burdens of glaucoma. Am J Ophthalmol. retinal nerve fibre layer thickness as assessed using optical
2001;152:515e522. coherence tomography based on anatomical confounders. Br J
4. Tham Y-C, Li X, Wong TY, et al. Global prevalence of Ophthalmol. 2020;104:282e290.
glaucoma and projections of glaucoma burden through 2040: a 22. Costa VP, Jimenez-Roman J, Carrasco FG, et al. Twenty-four-
systematic review and meta-analysis. Ophthalmology. hour ocular perfusion pressure in primary open-angle glau-
2014;121:2081e2090. coma. Br J Ophthalmol. 2017;101:305e308.
5. Langman MJ, Lancashire RJ, Cheng KK, et al. Systemic hy- 23. Bowe A, Grunig M, Schubert J, et al. Circadian variation in
pertension and glaucoma: mechanisms in common and co- arterial blood pressure and glaucomatous optic neuropathyea
occurrence. Br J Ophthalmol. 2005;89:960e963. systematic review and meta-analysis. Am J Hypertens.
6. De Moraes CG, Cioffi GA, Weinreb RN, Liebmann JM. New 2015;28:1077e1082.
recommendations for the treatment of systemic hypertension 24. Chua J, Woon CLC, Hong J, et al. Impact of hypertension on
and their potential implications for glaucoma management. retinal capillary microvasculature using optical coherence to-
J Glaucoma. 2018;27:567e571. mography angiography. J Hypertens. 2019;37:572e580.
7. Wright Jr JT, Williamson JD, Whelton PK, et al. 25. Asayama K, Ohkubo T, Hanazawa T, et al. Does antihyper-
A randomized trial of intensive versus standard blood-pressure tensive drug class affect day-to-day variability of self-
control. N Engl J Med. 2015;373:2103e2116. measured home blood pressure? The HOMED-BP Study.
8. Leske MC, Heijl A, Hyman L, et al. Predictors of long-term J Am Heart Assoc Cardiovasc Cerebrovasc Dis. 2016;5:
progression in the early manifest glaucoma trial. Ophthal- e002995.
mology. 2007;114:1965e1972. 26. Sato N, Saijo Y, Sasagawa Y, et al. Visit-to-visit variability
9. Tielsch JM, Katz J, Sommer A, et al. Hypertension, perfusion and seasonal variation in blood pressure: Combination of
pressure and primary open-angle glaucoma. A population- Antihypertensive Therapy in the Elderly, Multicenter Investi-
based assessment. Arch Ophthalmol. 1995;113:216e221. gation (CAMUI) Trial subanalysis. Clin Exp Hypertens.
10. Topouzis F, Wilson MR, Harris A, et al. Association of open- 2015;37:411e419.
angle glaucoma with perfusion pressure status in the Thessa- 27. Kurosaki R, Muramatsu Y, Imai Y, et al. Neuroprotective
loniki Eye Study. Am J Ophthalmol. 2013;155:843e851. effect of the angiotensin-converting enzyme inhibitor peri-
11. Charlson ME, de Moraes CG, Link A, et al. Nocturnal sys- ndopril in MPTP-treated mice. Neurol Res. 2004;26:644e657.
temic hypotension increases the risk of glaucoma progression. 28. Sonsalla PK, Coleman C, Wong L-Y, et al. The angiotensin
Ophthalmology. 2014;121:2004e2012. converting enzyme inhibitor captopril protects nigrostriatal
12. Muskens RPHM, de Voogd S, Wolfs RCW, et al. Systemic dopamine neurons in animal models of Parkinsonism. Exp
antihypertensive medication and incident open-angle glau- Neurol. 2013;250:376e383.
coma. Ophthalmology. 2007;114:2221e2226. 29. Hemming ML, Selkoe DJ, Farris W. Effects of prolonged
13. Zheng W, Dryja TP, Wei Z, et al. Systemic medication asso- angiotensin-converting enzyme inhibitor treatment on amyloid
ciations with presumed advanced or uncontrolled primary beta-protein metabolism in mouse models of Alzheimer dis-
open-angle glaucoma. Ophthalmology. 2018;125:984e993. ease. Neurobiol Dis. 2007;26:273e281.
14. Horwitz A, Klemp M, Jeppesen J, et al. Antihypertensive 30. Hemming ML, Selkoe DJ. Amyloid beta-protein is degraded
medication postpones the onset of glaucoma. Hypertension. by cellular angiotensin-converting enzyme (ACE) and elevated
2017;69:202e210. by an ACE inhibitor. J Biol Chem. 2005;11(280):
15. Nardin C, Rattazzi M, Pauletto P. Blood pressure variability 37644e37650.
and therapeutic implications in hypertension and cardiovas- 31. Liu S, Ando F, Fujita Y, et al. A clinical dose of angiotensin-
cular diseases. High Blood Press Cardiovasc Prev. 2019;26: converting enzyme inhibitor and heterozygous ACE deletion
353e359. exacerbates Alzheimer’s disease pathology in mice. J Biol
16. Lavanya R, Jeganathan VS, Zheng Y, et al. Methodology of Chem. 2019;294:9760e9770.
the Singapore Indian Chinese Cohort (SICC) Eye Study: 32. Mckinnon SJ. Glaucoma: ocular Alzheimer’s disease? Front
quantifying ethnic variations in the epidemiology of eye dis- Biosci. 2003;8:s1140e1156.
eases in Asians. Ophthalmic Epidemiol. 2009;16:325e336. 33. Materson BJ, Reda DJ, Cushman WC, et al. Single-drug
17. Rosman M, Zheng Y, Wong W, et al. Singapore Malay Eye therapy for hypertension in men. A comparison of six anti-
Study: rationale and methodology of 6-year follow-up study hypertensive agents with placebo. The Department of Veterans
(SiMES-2). Clin Exp Ophthalmol. 2012;40:557e568. Affairs Cooperative Study Group on Antihypertensive Agents.
18. Sabanayagam C, Yip W, Gupta P, et al. Singapore Indian Eye N Engl J Med. 1993;328:914e921.
Study-2: methodology and impact of migration on systemic 34. Wright Jr JT, Dunn JK, Cutler JA, et al. Outcomes in hyper-
and eye outcomes. Clin Exp Ophthalmol. 2017;45:779e789. tensive black and nonblack patients treated with chlorthali-
19. Ho H, Tham YC, Chee ML, et al. Retinal nerve fiber layer done, amlodipine, and lisinopril. JAMA. 2005;6(293):
thickness in a multi-ethnic normal Asian population: The 1595e1608.
Singapore Epidemiology of Eye Diseases Study. Ophthal- 35. Yoshikawa T, Obayashi K, Miyata K, et al. Increased night-
mology. 2019;126:702e711. time blood pressure in patients with glaucoma: cross-sectional
20. Tham YC, Chee ML, Dai W, et al. Profiles of ganglion cell- analysis of the LIGHT Study. Ophthalmology. 2019;126:
inner plexiform layer thickness in a multi-ethnic Asian 1366e1371.

7
 Ophthalmology Volume -, Number -, Month 2020

Footnotes and Financial Disclosures

Originally received: March 24, 2020. HUMAN SUBJECTS: Human subjects were included in this study. Ethical
Final revision: July 23, 2020. approval was obtained from the SingHealth Centralized Institutional Re-
Accepted: July 27, 2020. view Board. All research adhered to the tenets of the Declaration of Hel-
Available online: ---. Manuscript no. D-20-00635. sinki. All participants provided informed consent.
1
Singapore National Eye Centre, Singapore. No animal subjects were used in this study.
2 Author Contributions: Study was performed as part of regular employment
Singapore Eye Research Institute, Singapore.
duties at affiliated institutes. No additional funding was provided.
3
Ophthalmology & Visual Sciences Academic Clinical Program (Eye Conception and design: Chong, Chee, Tham, Cheng
ACP), Duke-NUS Medical School Singapore, Singapore.
Data collection: Chong, Chee, Tham, Majithia, Thakur, Teo, Soh, Chua,
4
SERI-NTU Advanced Ocular Engineering (STANCE), Singapore. Tan, Wong, Schmetterer, Sabanayagam, Cheng
5
NTU Institute of Health Technologies, Singapore. Analysis and interpretation: Chong, Chee, Tham, Majithia, Thakur, Chua,
6 Tan, Wong, Schmetterer, Cheng
School of Chemical and Biomedical Engineering, Nanyang Technological
University, Singapore. Obtained funding: Chong
7
Department of Clinical Pharmacology, Medical University of Vienna, Overall responsibility: Chong, Chee, Tham, Majithia, Thakur, Teo, Soh,
Vienna, Austria. Chua, Tan, Wong, Schmetterer, Sabanayagam, Cheng
8 Abbreviations and Acronyms:
Center for Medical Physics and Biomedical Engineering, Medical Uni-
ACEI ¼ angiotensin-converting enzyme; BMI ¼ body mass index;
versity of Vienna, Vienna, Austria.
BP ¼ blood pressure; BRB ¼ beta receptor blocker; CCB ¼ calcium
9
Institute of Molecular and Clinical Ophthalmology, Basel, Switzerland. channel blocker; CI ¼ confidence interval; DBP ¼ diastolic blood pressure;
Financial Disclosure(s): GC-IPL ¼ ganglion cell-inner plexiform layer; IOP ¼ intraocular pressure;
The author(s) have made the following disclosure(s): R.S.C.: Grants e MAP ¼ mean arterial pressure; RNFL ¼ retinal nerve fiber layer;
Duke-NUS Graduate Medical School and SingHealth; Personal fees e SBP ¼ systolic blood pressure.
Santen Pharmaceutical Co. Ltd., outside the submitted work. Keywords:
Supported by the National Medical Research Council, Republic of hypertension, glaucoma, OCT, anti-hypertensive, medication, RNFL, GC-
Singapore (grant no. NMRC/1249/2010, NMRC/CIRG/1371/2013, IPL, systemic medication.
NMRC/CIRG/1417/2015, NMRC/CG/C010A/2017_SERI, NMRC/CIRG/
Correspondence:
1488/2018; and NMRC/CSA-SI/0012/2017 [C.-Y.C.]); and the Biomedical Ching-Yu Cheng, MD, PhD, Department of Ophthalmology & Visual
Research Council, Singapore, Republic of Singapore (grant no. 08/1/35/19/ Sciences, Singapore Eye Research Institute, Singapore 168751. E-mail:
550).
[email protected].