Professional Documents
Culture Documents
com
ABSTRACT
Aim To measure the reproducibility of retinal oxygen
saturation (SaO2) levels among treated glaucomatous
eyes and normal controls in a prospective nonrandomised study.
Methods Patients with perimetric glaucoma (PG) and
normal controls were included. Exclusion criteria for both
groups included visual acuity <20/30, unreliable visual
elds, thyroidopathies, hemoglobinopathies, cardiovascular
and pulmonary diseases. Retinal oximetry was performed
twice consecutively on one randomly selected eye of PG
and normal controls using spectrophotometric retinal
oximeter (SRO; Oxymap ehf., Iceland). Four main retinal
vessel pairs were analysed separately. Coefcients of
variability (CoV), coefcients of repeatability (CoR) and
intraclass correlation coefcients (ICCs) in arteries (a.SaO2)
and veins (v.SaO2) were calculated.
Results 23 PG (mean age 68.310.8 years) and 22
normal subjects (mean age 61.518.2 years; p=0.14)
were included. The intraocular pressure and mean ocular
perfusion pressure in glaucoma (14.44.2 mm Hg; 45.8
5.8 mm Hg) and controls (14.33.3 mm Hg; 45.8
6.1 mm Hg) were similar (p >0.05). In the PG group, the
a.SaO2 had a CoV of 1.6%, a CoR of 4.7 and an ICC of
0.97; the v.SaO2 had a CoV of 5.9%, a CoR of 8.7 and an
ICC of 0.96. In normals, the a.SaO2 had a CoV of 0.98%,
a CoR of 3.3 and an ICC of 0.97; the v.SaO2 had a CoV of
4.8%, a CoR of 7.7 and an ICC of 0.93.
Conclusions Retinal oximetry measurements using SRO
are highly reproducible in both treated glaucomatous and
normal eyes.
INTRODUCTION
To cite: Goharian I,
Iverson SM, Ruiz RC, et al.
Br J Ophthalmol Published
Online First: [ please include
Day Month Year]
doi:10.1136/bjophthalmol2014-305718
GON, researchers are exploring the use of spectrophotometric retinal oximeter (SRO) in the diagnosis
and management of glaucoma.
Recent advancements in SRO technology have
provided new possibilities for the non-invasive
measurement of SaO2 in retinal vessels. Preliminary
studies have shown impaired SaO2 in glaucomatous
eyes, especially in advanced stages, compared with
normal eyes.1214 Impaired retinal oxygen metabolism has also been associated with deeper glaucomatous VF defects,15 decreased neuroretinal rim area
and RNFL thinning.16 These ndings suggest that
altered retinal oxygen metabolism in glaucomatous
eyes may be associated with structural damage.16
Recently, an automatic SRO (Oxymap ehf.,
Iceland) was developed that acquires two images at
two specic wavelengths (570 and 600 nm), to allow
the calculation of SaO2.17 18 The SRO measurements
are subject to intersubject and intrasubject variability
due to confounding variables. These variables
include, but are not limited to, light scattering by
blood cells,17 various light paths through vessels,17
fundus pigmentation,19 vessel width19 20 and ash
intensity21 and require corrections to improve both
sensitivity and reproducibility. Furthermore, it is
notable that SRO is not fully automated, and image
acquisition and the adopted method of image analysis
depend on the trained operator. Few studies have
examined the reproducibility of SaO2 measurements
in healthy individuals,13 17 22 23 but no data have
been published on the reproducibility and the variability of retinal oximetry using SRO in glaucoma
population. The purpose of this study was to
measure the reproducibility of SaO2 levels in both
treated glaucomatous eyes and normal controls.
I, et al. Br J Ophthalmol
doi:10.1136/bjophthalmol-2014-305718
Copyright Article author (orGoharian
their employer)
2014.2014;0:15.
Produced
by BMJ Publishing Group Ltd under licence.
Clinical science
California, USA). All glaucoma patients had prior SAP experience
and at least one conrmatory SAP examination. Systemic blood
pressure and SRO measurements were performed at the same visit.
Good-quality SRO images were dened as an image in focus,
evenly illuminated, with sufcient ash intensity. Mean ocular perfusion pressure was calculated using the formula 2/3 (DBP+1/3
(SBPDBP))IOP.24
Inclusion criteria common to both normal and glaucoma
groups consisted of age 30 and 90 years, resting systolic
blood pressure (SBP) between 90 and 139 mm Hg, resting diastolic blood pressure (DBP) <90 mm Hg, spherical equivalent
refractive error within 7.00 dioptres sphere and3.00 dioptres cylinder, best-corrected visual acuity of 20/30, clear media
dened as lens opacication classication system IIIgrade 1,
reliable SAP dened as <15% rate of xation loss and <20%
rates of false-positive error throughout the study.
Normal subjects had IOP 21 mm Hg in both eyes, healthy
optic disc appearance based on clinical examination and review
of stereodisc photography and two reproducible normal SAP
examinations. Normal SAP was dened as a mean deviation
(MD) and pattern standard deviation (PSD) within 95% CI limits
and a glaucoma hemield test (GHT) result within normal limits.
Absence of GON was dened as vertical cup-disc asymmetry
<0.2, intact neuroretinal rim without peripapillary haemorrhages, notches, localised pallor or NFL defect. Inclusion criteria
for the glaucoma group consisted of GON dened as neuroretinal rim narrowing to the optic disc margin, notching, excavation
or RNFL defect, and corresponding abnormal SAP dened as
abnormal GHT and PSD outside 95% normal limits. Patients
with thyroidopathies, haemoglobinopathies, uncontrolled hypertension, pulmonary diseases and ocular diseases other than glaucoma or cataract were excluded from this study. Taking
medications that would affect systemic blood ow or oxygenation was an exclusion criterion. Poor-quality SRO images and
images that were obtained during eye movements were also
excluded.
I0
I
ODR
ODsensitive
ODisosbestic
RESULTS
Forty-ve subjects were recruited, composed of 23 glaucomatous eyes of 23 glaucoma patients (8 men, 15 women) and 22
normal eyes of 22 healthy subjects (10 men, 12 women).
Table 1 demonstrates the clinical characteristics of the study
cohort. The mean age was similar between patients with glaucoma (68.310.8 years) and normal controls (61.518.2 years;
p=0.14). Glaucomatous and normal eyes did not have a statistically signicant difference in treated IOP (14.44.2 vs 14.3
3.3 mm Hg; p=0.88; table 1). The SAP MD and PSD were
signicantly worse in the glaucoma group compared with
normal ( p<0.00; p=0.001, respectively). The full list of the
ophthalmic medications in the glaucoma group has been provided in table 2.
The ICC, CoV and CoR (table 3) were evaluated for a.SaO2,
and v.SaO2 parameters using SRO. In the glaucoma group, the
a.SaO2 had an ICC of 0.97, a CoV of 1.6% and a CoR of 4.7%.
The v.SaO2 had an ICC of 0.96, a CoV of 5.9% and a CoR of
8.7%. In normal controls, the a.SaO2 had an ICC of 0.97, a
CoV of 0.98% and a CoR of 3.3%; the v.SaO2 had an ICC of
0.93, a CoV of 4.8% and a CoR of 7.7%. To check the reproducibility, the average SDs of repeated measurements of SaO2 in
arteries (a.SaO2) and veins (v.SaO2) were calculated. The mean
SD of a.SaO2 measurements was 0.86% in normal and 1.46%
Goharian I, et al. Br J Ophthalmol 2014;0:15. doi:10.1136/bjophthalmol-2014-305718
Clinical science
Table 1 Clinical features of the study population
Glaucoma
(n=23)
Age (years)
Gender (Male/Female)
IOP (mm Hg)
MOPP (mm Hg)
Diabetes Mellitus (Yes/No)
Systemic antihypertensive
medications (Yes/No)
SAP MD (dB)
SAP PSD (dB)
68.310.8
8/15
14.44.2
45.85.8
5/18
8/15
7.424.6
7.234.7
Table 3
Normal
(n=22)
p-value
61.518.2
10/12
14.33.3
45.86.1
1/21
7/15
0.14
0.55
0.88
0.97
0.19
0.99
1.200.99
1.670.47
<0.001
0.001
dB, decibel; IOP, intraocular pressure; MD, mean deviation; MOPP, mean ocular
perfusion pressure; PSD, pattern standard deviation; SAP, standard automated perimetry.
DISCUSSION
Glaucoma is a multifactorial optic neuropathy, and vascular
factors contribute to the development and progression of GON
and VF loss.3 58 The RGCs and their axons are supplied by the
retinal circulation, through central retinal artery (CRA), and
blood ow has been shown to be reduced in glaucomatous
ONH, and retina compared with normal eyes.25 Several studies
have shown that the arteriovenous SaO2 is different between
advanced glaucoma and normal eyes.12 14 SRO is a novel
imaging technology that uses a modied digital camera to
provide quantitative measurements of oxygenation of retinal
vessels in human eyes.1719 It has been shown that SRO has a
mean SD of 0.8% to 5.3% between repeated SaO2 measurements in normal eyes.17 22 23 However, there are no publications on the reproducibility of SRO in glaucomatous eyes.
There is currently no reference standard for the selection of
vessels in oximetry. Hardarson et al17 selected a single rstdegree vessel pair (retinal location unspecied) to study the
Glaucoma (n=23)
Topical -blockers
Prostaglandin analogues
2-Adrenergic agonists
Non-selective adrenergic agonists
Miotic agents
Topical or systemic carbonic anhydrase inhibitors
17
19
7
0
1
14
a.SaO2
v.SaO2
Normal (n=22)
ICC
CoR*
CoV*
ICC
CoR*
CoV*
0.97
0.96
4.7
8.7
1.6
5.9
0.97
0.93
3.3
7.7
0.98
4.8
reproducibility of SRO. Palsson et al22 obtained their reproducibility data using vessels larger than 8 pixels around the optic
disc. Hammer et al23 examined 10 arterial and 10 venous vessel
sections with a constant length of one-half of the disc diameter
(DD). In the current study, we selected the four main retinal
vessel pairs within >1.5 DD and <3 DD of the optic disc. We
found that this technique provides more robust measurements
because it reduces the chance of potential vessel dropout, and it
increases the reliability by avoiding the peripheral retina.
Moreover, in the peripapillary region there is a possibility of
oxygen diffusion because of short distances between arterioles
and venules or arteriovenous shunts, which may create artefacts
for the SaO2 measurements.22
The BlandAltman analysis showed that the 95% limits of
agreement were narrower in arteries compared with veins in both
glaucoma and control groups. OConnell et al20 examined the
intrasession repeatability of SaO2 in healthy subjects using the
entire image technique and found wider limits of agreement
compared with our data that were worse in veins. Our study
showed that selecting the four main retinal vessel pairs reduced
the variability compared with previous techniques.17 20 23
Because the ocular circulation is a closed circulatory system, and
inow must equal outow like any steady-state system that obeys
the law of conservation of mass,25 using the four main retinal
vessel pairs should theoretically provide the level of oxygen saturation as using higher number of retinal arteries and veins.
We also found that the mean SD of repeated SaO2 measurements was narrower in normal eyes compared with glaucomatous eyes. It is possible that dry eyes and supercial punctate
keratopathy due to the application of topical glaucoma medications contribute to higher variability of SaO2 measurements in
glaucomatous eyes. A study by Hardarson et al17 found the
mean SD of SaO2 measurements was 3.7% and 5.3% in arteries
and veins, respectively, in a group of ve diabetic and ve
healthy subjects. Hammer et al23 reported a mean SD of 2.52%
and 3.25% for arteries and veins, respectively. Palsson et al22
measured oxygen saturation across the entire retina in 26
normal subjects and found the SD of SaO2 measurements was
0.8% in arteries and 1.3% in veins. Differences in sample sizes,
study populations and grading techniques may limit the ability
to thoroughly compare our results with other studies.
The ICC is a measure of intrasubject variance. Lower ICC
values represent greater intrasubject variance. In our study, the
ICC values were quite high, which represent great reproducibility in both normal and glaucomatous eyes. Our values are
higher than those reported by Michelson and Scibor,13 who
used a different imaging spectrometry in a wavelength range of
395710 nm. Other discrepancies such as the variability in
image quality and the measurement area could contribute to
these differences. CoR and CoV are based on the intrasubject
SD and the testretest dispersion. Lower CoV and CoR mean
3
Clinical science
Figure 1 BlandAltman plots illustrating the difference in retinal oxygen saturation (SaO2) measurements vs their mean and 95% limits of
agreement (mean difference1.96 SD). The dotted line represents the zero difference. (A) Arterial SaO2 (a.SaO2) in glaucoma, (B) venous SaO2
(v.SaO2) in glaucoma, (C) a.SaO2 in normal and (D) v.SaO2 in normal.
REFERENCES
1
2
6
7
8
9
10
11
Quigley HA. New paradigms in the mechanisms and management of glaucoma. Eye
2005;19:12418.
Gordon MO, Beiser JA, Brandt JD, et al. The Ocular Hypertension Treatment Study:
baseline factors that predict the onset of primary open-angle glaucoma. Arch
Ophthalmol 2002;120:71420.
Krupin T, Liebmann JM, Greeneld DS, et al; Low-Pressure Glaucoma Study Group.
A randomized trial of brimonidine versus timolol in preserving visual function:
results from the Low-Pressure Glaucoma Treatment Study. Am J Ophthalmol
2011;151:67181.
Shigeeda T, Tomidokoro A, Araie M, et al. Long-term follow-up of visual eld
progression after trabeculectomy in progressive normal-tension glaucoma.
Ophthalmology 2002;109:76670.
De Moraes CG, Liebmann JM, Greeneld DS, et al. Risk factors for visual eld
progression in the low-pressure glaucoma treatment study. Am J Ophthalmol
2012;154:70211.
Chung HS, Harris A, Kagemann L, et al. Peripapillary retinal blood ow in normal
tension glaucoma. Br J Ophthalmol 1999;83:4669.
Hwang JC, Konduru R, Zhang X, et al. Relationship among visual eld, blood ow,
and neural structure measurements in glaucoma. Invest Ophthalmol Vis Sci
2012;53:30206.
Flammer J, Mozaffarieh M. Autoregulation, a balancing act between supply and
demand. Can J Ophthalmol 2008;43:31721.
Khoobehi B, Firn K, Thompson H, et al. Retinal arterial and venous oxygen
saturation is altered in diabetic patients. Invest Ophthalmol Vis Sci
2013;54:71036.
Sehi M, Flanagan JG, Zeng L, et al. Anterior optic nerve capillary blood ow
response to diurnal variation of mean ocular perfusion pressure in early untreated
primary open angle glaucoma. Invest Ophthalmol Vis Sci 2005;46:45817.
Tobe LA, Harris A, Schroeder A, et al. Retinal oxygen saturation and metabolism:
how does it pertain to glaucoma? An update on the application of retinal oximetry
in glaucoma. Eur J Ophthalmol 2013;23:46572.
Clinical science
12
13
14
15
16
17
18
19
20
21
22
23
24
25
OConnell RA, Anderson AJ, Hosking SL, et al. Test-retest reliability of retinal oxygen
saturation measurement. Optom Vis Sci 2014;91:60814.
Heitmar R, Cubbidge RP. The impact of ash intensity on retinal vessel oxygen
saturation measurements using dual wavelength oximetry. Invest Ophthalmol Vis Sci
2013;54:280711.
Palsson O, Geirsdottir A, Hardarson SH, et al. Retinal oximetry images must be
standardized: a methodological analysis. Invest Ophthalmol Vis Sci
2012;53:172933.
Hammer M, Vilser W, Riemer T, et al. Retinal vessel oximetry-calibration,
compensation for vessel diameter and fundus pigmentation, and reproducibility.
J Biomed Opt 2008;13:054015.
Sehi M, Flanagan JG, Zeng L, et al. Relative change in diurnal mean ocular
perfusion pressure: a risk factor for the diagnosis of primary open-angle glaucoma.
Invest Ophthalmol Vis Sci 2005;46:5617.
Sehi M, Goharian I, Konduru R, et al. Retinal blood ow in glaucomatous eyes with
single-hemield damage. Ophthalmology 2014;121:7508.
doi: 10.1136/bjophthalmol-2014-305718
These include:
References
P<P
Email alerting
service
Topic
Collections
Notes
Advance online articles have been peer reviewed, accepted for publication, edited and
typeset, but have not not yet appeared in the paper journal. Advance online articles are
citable and establish publication priority; they are indexed by PubMed from initial
publication. Citations to Advance online articles must include the digital object identifier
(DOIs) and date of initial publication.