Investigative Ophthalmology & Visual Science, Vol. 29, No. 2, February 1988
Copyright © Association for Research' in Vision and Ophthalmology
A New Method For Oxygen Supply
to Acute Ischemic Retina
Joshua Den-Nun,* Valerie A. Alder, Stephen J. Cringle, and Ian J. Constable
This paper introduces a new method for supplying oxygen directly to ischemic inner retina, using an
oxygen source in the vitreous. Acute retinal vascular occlusion was created in cat eyes by direct
pressure on the optic disk and its margins with a glass probe. The satisfactory occlusion of the retinal
vessels was documented by direct observation, and functionally by recording the ERG. The vascular
occlusion caused a large decrease in the size of the ERG b wave, with no change in the a wave
amplitude. The oxygen source was a catheter made of strands of an oxygen-permeable membrane
which was inserted into the vitreal cavity. After successful vascular occlusion was documented, 100%
gaseous oxygen was perfused through the catheter while recording the ERG. In response to the
perfused oxygen the b wave partially recovered. Ventilating the animal with 100% oxygen when the
retinal vessels were occluded also caused recovery of the b wave amplitude. Termination of the vitreal
oxygen source caused a decrease in b wave amplitude to the level previously observed after the
occlusion of the retinal vessels. When the retinal circulation was restored by removal of the glass
probe the b wave recovered. The results show that it is possible to supply adequate oxygen to the inner
retina via the vitreous to replace the oxygen normally supplied by the retinal circulation. Modification
of this method may be useful for the treatment of recent and incomplete retinal vascular occlusion.
Invest Ophthalmol Vis Sci 29:298-304,1988
The retina in man and cat is nourished by the retinal and the choroidal circulations. The choroidal circulation supplies the avascular outer retina with oxygen by diffusion from the choriocapillaris. The large
blood flow of the choroidal circulation results in a
small arteriovenous oxygen difference which means
that the choriocapillaris has an unusually high PO2
for a capillary bed. The retinal circulation in contrast
supplies only those areas within the inner retina in
which the capillary beds lie. This is demonstrated by
the existence of a balance point between the retinal
and the choroidal circulations for oxygen1'2 which lies
at the boundary of the inner nuclear layer. In clinical
cases of acute occlusion of the retinal circulation several methods have been tried to use the large oxygen-
carrying capacity of the choroidal circulation to supply the ischemic inner retina with sufficient oxygen to
maintain function.3"5 The common basis for all these
methods is that the oxygen tension in the choroidal
circulation is raised by suitable ventilation, often with
hyperbaric oxygen, to allow oxygen to diffuse into the
ischemic retina. With these methods, much of the
retina is exposed to high oxygen levels in order for the
innermost retinal layer to reach the normal range of
oxygen tension. This situation may be toxic to the
outer retina, even within the short time that hyperoxia is used.6'8 Moreover, using the systemic circulation to increase oxygen supply to the choroid places
other body organs at risk from the possible toxic effectsofhyperoxia.12
This paper introduces a new method for supplying
oxygen directly to the inner retina using a transvitreal
approach and a catheter made of an oxygen permeable membrane. A model of acute ischemic inner
retina was produced in cat eyes and the ability of this
method to restore inner retinal function was measured by monitoring the amplitude of the b wave of
the ERG.9-10
From the Lions Eye Institute and Unit of Ophthalmology, Department of Surgery, University of Western Australia, Nedlands,
Western Australia.
* On leave from Department of Ophthalmology, Tel Aviv Medical Centre, Tel Aviv, Israel.
Dr. Ben-Nun holds a Shaw Foundation Fellowship from the Sir
Charles Gairdner Hospital, Western Australia. This research was
supported by the Lion's Eye Institute, the National Health and
Medical Research Council of Australia, and the TVW Telethon
Foundation of Western Australia.
Submitted for publication: April 14, 1987; accepted August 28,
1987.
Reprint requests: Joshua Ben-Nun, Department of Ophthalmology, Tel Aviv Medical Centre, 6 Wiseman Street, Tel Aviv, Israel.
Materials and Methods
Manufacture of the Catheter and the Probe
The catheter was manufactured from hollow fibers
commercially produced for hemodialysis. The fibers
298
Downloaded From: http://iovs.arvojournals.org/ on 07/31/2017
No. 2
OXYGEN SUPPLY TO ISCHEMIC RETINA / Den-Nun er ol.
were made of regenerated cellulose (cuprammonium
Rayon) and the manufacturer's (TERUMO) specifications were: internal diameter 200 /*m; wall thickness 12 /im; pore size 25 A; molecular weight cutoff
about 35,000 at a rejection rate of 95%. Two kinds of
catheters were manufactured that differed only in the
number of filaments used and consequently in the
diameter of the guiding needle required to carry the
catheter. The large catheter had 24 filaments with a
14 gauge needle as a guide. The small catheter had 14
filaments with a 16 gauge needle as a guide. All other
properties and manufacturing procedures for the
catheters were identical. A 150 nm tungsten rod was
inserted into each filament for internal support. In
Figure 1 the main stages in the manufacturing of the
catheter are demonstrated. The fibers (Fig. 1.1) were
inserted into a rigid plastic tube (Fig. 1.2), creating a
central protected area (B) with two ends of exposed
fibers (A and C). The exposed fibers were covered
with epoxy glue (Fig. 1.3). The plastic cover (B) was
then removed, leaving a smooth section on the glued
area (C), and a central 15 mm of uncoated fibers (D)
(Fig. 1.4). The glued section (A) was covered by a
silicon tube and silicon glue was used to fill the space
between thefibersand the tube (Fig. 1.5). This part of
the catheter was used as a connector to the oxygen
source. The catheter was left overnight to dry. The
catheter was inserted into the hypodermic guide with
only the connector section E exposed.
The probe that was used to occlude the retinal vessels was made from a 3 cm length of hollow glass
tube: external diameter 1 mm, internal diameter 0.7
mm. One end of the tube was heat shaped to form a
glass ball, 2 mm in diameter. Arigidmetallic rod, 0.6
mm diameter, 8 cm long, was inserted and glued
along its length to the inner surface of the glass tube.
This metallic rod was used to connect the probe to
the micromanipulator.
General Surgery
Adult domestic cats of weight 3 to 6 kg were used.
Anesthesia was induced by an initial intravenous injection of 2 ml SafFan (alfaxalone 18 mg, alphadalone
acetate 6 mg). After tracheal cannulation, the sympathetic trunk was severed bilaterally. The cats were
ventilated at 33 strokes/min, tidal volume 20 ml,
with 20% oxygen/80% nitrogen (arterial blood gases
PaO2 = 99.5 ± 10.3 mmHg, PaCO2 = 29.2 ± 6.1
mmHg, pH = 7.36 ± 0.20. After an initial dose of 80
mg Flaxedil (gallamine triethiodide) to achieve paralysis, a constant intravenous infusion of Flaxedil (5
mg/kg/hr) and SafFan (2 ml/hr) produced stable anaesthesia and paralysis. Arterial blood gases were
Downloaded From: http://iovs.arvojournals.org/ on 07/31/2017
299
Fig. 1. The stages of catheter manufacture: (1) The oxygen permeable fibers (14 or 24) with tungsten rod inserted into each one.
(2) Rigid plastic tube placed over the centre portion of thefibers,B,
for protection. (3) Both ends A and C of the fibres coated with
epoxy glue. (4) Removal of the plastic tube leaving a 15 mm section
D of oxygen permeablefibres.(5) Silicon tube connector E placed
over A for connection to the oxygen source.
measured at the start of the experiment and after any
change in the oxygen concentration of the ventilation
gas (Corning blood gas analyzer 161). Rectal temperature and ECG were monitored continuously during
the experiment.
Ocular Surgery
A lateral orbitectomy was performed. A metal eye
ring was sutured to the limbus (Fig. 2a) and was rigidly clamped to the head holder to minimize eye
movements. The pupil was dilated with Mydriacyl
(tropicamide 1%) eye drops. The sclera was exposed
in three areas, shown in Figure 2.1: the superior (12
o'clock) (c) the lateral (3 o'clock L.E.; 9 o'clock R.E.)
(c) on the pars plana (b), and the peripheral temporal
retina (5 o'clock L.E.; 7 o'clock R.E.) (d). A 2 mm
diathermy spot was made on the sclera in each area
300
INVESTIGATIVE OPHTHALMOLOGY b VISUAL SCIENCE / February 1988
Vol. 29
Fig. 2. A diagram to represent the main stages of ocular surgery: (1) Fixation of
the eye and entry sites: (a)
metal ring attached near the
limbus, (b) pars plana ciliaris, (c) two rubber diaphragms were glued to the
sclera over pars plana, (d) a
third rubber diaphragm was
glued to the sclera over peripheral retina. (2) The
catheter (g) in its needle
guide (e) is inserted through
the rubber diaphragm. (3)
The needle guide (e) is removed through diaphragm
(d) leaving the catheter in
position in the vitreous. (4)
The occluding probe (f) inserted through the rubber
diaphragm (c).
prior to puncture to minimize subsequent bleeding.
A rubber diaphragm 5 mm diameter and 1 mm thick
was glued over each of the treated areas with cyanoacrylate adhesive to minimize potential vitreous leakage. The catheter guide (e) containing the inserted
catheter (g) (16 gauge or 14 gauge needle) was inserted through the rubber into the eye at the 12
o'clock position (Fig. 2.2). It passed through the vitreous and exited at the 5 o'clock position in the L.E., or
7 o'clock in the R.E. (Fig. 2.3). This resulted in good
fixation of the catheter at both entry and exit points,
with minimal vitreous leakage. After insertion of the
catheter into the globe the occluding probe (f) was
inserted. An 18 gauge needle was used to perform an
initial puncture in the lateral area (3 o'clock L.E.; 9
o'clock R.E.) through the rubber diaphragm and the
diathermized area underneath it. The needle was removed and the occluding probe was inserted into the
eye in an oblique direction (Fig. 2.4) to avoid damage
to the lens. The probe was held in a micromanipulator, and, using indirect ophthalmoscopic control, the
tip of the probe was advanced through the vitreous to
a position in front of, and close to, the optic disk.
ERG Recording
The flash ERG was recorded differentially, using
silver/silver chloride electrodes; the active electrode
was embedded in a corneal contact lens of zero
power, the indifferent electrode was inserted into
temporal tissue and the ground electrode was placed
in neck tissue. The signal was amplified (Neurolog
NL103, NL105) with a gain of 10,000, and was filtered with a bandpass of 0.1 Hz to 1 kHz. The ERG
Downloaded From: http://iovs.arvojournals.org/ on 07/31/2017
was observed and stored on a digital oscilloscope
(Gould OS 1420) and plotted out on an XY recorder
(Advance LR100) when hard copy was required. The
flash stimulus was generated with a Grass PS2B photostimulator using intensity setting 8, and was placed
at a distance of 140 cm from the anterior nodal point
of the eye. The triggering pulse for the photostimulator was provided by the Neurolog NL303, NL402
modules every 20 sec. Background room luminance
was approximately 2 cd/m2. After ophthalmoscopy it
took between 5-10 min for the ERG b wave to return
to its adapted state for this flash intensity and repetition rate.
Experimental Regime
The eye wasflashedcontinuously with the stimulus
and a hard copy of the adapted ERG was made at
each of the relevant stages. After any ophthalmolscopy at least 10 min elapsed before any ERG was
recorded.
After placement of the catheter and probe in the
eye the retinal circulation was occluded. This was
done by moving the probe with the micromanipulator system to press gently on the optic disk. As the
probe was made of glass it was possible to visualize
the disk, and the vessels on its margin through the
probe tip using the indirect ophthalmoscope (Fig. 3,
top). Complete occlusion of the retinal blood vessels
was achieved by this method as demonstrated by the
stasis of the blood flow (Fig. 3, middle). The animal
was then breathed on 100% O2 for 5 min, to determine the capacity of the choroidal circulation to
compensate for the lack of oxygen supply from the
OXYGEN 5UPPLY TO I5CHEMIC RETINA / Den-Nun er ol.
No. 2
301
retinal circulation. Animal ventilation was returned
to 20% O2, 80% N2. After b wave equilibrium was
again established the catheter was perfused with 100%
gaseous O2 at aflowrate of 300 ml/min for a period
of time which varied from 20 to 120 min. If gas bubbles were observed in the vitreous due to a leak in the
catheter it was rejected and replaced with another
catheter. All oxygen was then removed from the
catheter by turning off the flow of oxygen, flushing
out the remaining oxygen with N2O, and clamping
theflexibleend of the catheter to avoid any net gaseous movement within it. The retinal circulation was
restored by withdrawing the occluding probe and the
resumption of blood flow was verified by indirect
ophthalmoscopy and sometimes by fundus photography (Fig. 3, bottom). Finally the animal was ventilated with 100% oxygen for 5 min to determine
whether any increase of the b wave amplitude could
be achieved. Arterial blood gases were sampled at
each alteration in ventilation.
Rectal temperature and ECG were monitored continuously during the experiment. Both eyes were used
and at the completion of the experiment the cat was
killed with a barbiturate overdose.
All experiments performed in this investigation
conformed to the ARVO Resolution on the Use of
Animals in Research.
Results
The experiments were performed on 13 eyes from
seven cats. Several surgical problems were encountered in early experiments which we had to overcome
before obtaining satisfactory results. The major problem was caused by the penetration of the catheter and
probe, where bleeding, retinal detachment and vitreal
leakage leading to ocular hypotony occurred. The
careful use of cautery and rubber seals mentioned in
the ocular surgery methods section solved this problem. Occlusion of the retinal circulation by pressing
on the vessels at the rim of the optic disk also produced retinal detachment and some bleeding if not
performed carefully. It was never possible to completely avoid a narrow ring shaped detachment
around the optic disk. This detachment was usually
accompanied by local subretinal bleeding, especially
in the deocclusion stage. Successful results were obtained from five eyes, most of which were from the
later experiments, after the major technical problems
had been solved.
In Figure 4 the ERG recordings, made at different
stages of one of the experiments, are plotted. The
baseline ERG was measured before vascular occlusion but with the catheter and occluding probe al-
Downloaded From: http://iovs.arvojournals.org/ on 07/31/2017
Fig. 3. Fundus photographs of the posterior pole: Top: Showing
the occluding probe in position close to the optic disk. Middle:
During retinal circulation occlusion with haemostasis. Bottom:
After the retinal circulation was restored.
ready placed in the eye (Fig. 4a). The flash is presented at the beginning of the trace and the response
shows both an a and b wave. The reduction in the
amplitude of the b wave as a consequence of 10 min
occlusion of the retinal vessels is shown in Figure 4b.
The a wave amplitude was relatively unchanged by
this maneuver, implying that the occlusion did not
affect the ability of the choroidal circulation to supply
302
INVESTIGATIVE OPHTHALMOLOGY & VISUAL SCIENCE / February 1988
a.
b.
c.
•*\/
\
d.
Vol. 29
lating the animal with 100% oxygen at the end of the
experiment did not cause an increase in the size of the
b wave.
The results of the experiments to replace inner retinal oxygenation by diffusion of oxygen from the choroidal circulation are presented in Figure 5. The control ERG is plotted in Figure 5a, showing the presence of both a and b wave. Clamping the retinal
circulation again resulted in minimal alteration in the
size of the a wave but a greatly reduced b wave shown
after 15 min occlusion (Fig. 5b). Breathing the cat on
100% O2 for 5 min with the retinal circulation still
occluded resulted in the return of the b wave amplitude to the control value (Fig. 5c). When the ventilation was returned to 21% O2 79% N2 the b wave was
again diminished in amplitude (Fig. 5d).
Discussion
e
*
f.
soav
o*u
50msec
Fig. 4. Sequence of ERG recordings made during one experiment: (a) before vascular occlusion, (b) diminution of b wave after
10 min occlusion of the retinal vessels, (c) restoration of b wave
after 100% oxygen had been perfused through the vitreal catheter
for 25 min, (d) loss of b wave again 4 min after terminating the
oxygen supply from vitreous, (e) recovery of b wave 16 min after
the retinal circulation had been restored by withdrawing the probe,
(0 ERG 10 min after breathing the cat on 100% O2.
the outer retina with sufficient oxygen. Perfusion
with 100% O2 through the catheter for 25 min caused
the b wave amplitude to recover (Fig. 4c). After the
source of vitreal oxygen was stopped the b wave amplitude again fell within 4 min (Fig. 4d), to approximately the same amplitude as it was prior to perfusion (compare Fig. 4b). Deoccluding the retinal circulation resulted in recovery of the b wave amplitude
to nearly the original size, shown in Figure 4e, which
was recorded 16 min after the retinal circulation was
re-established. Breathing the cat on 100% O2 for 10
min produced no further increase in b wave amplitude (Fig. 4f).
In all the successful experiments the final b wave
was smaller than the initial value. The longer the total
occlusion time, the smaller the b wave was as a percentage of its original value after deocclusion. Venti-
Downloaded From: http://iovs.arvojournals.org/ on 07/31/2017
These experiments show that it is possible to partially restore function to an ischemic inner retina by
supplying oxygen via a catheter placed in the vitreous.
All previous attempts to supply the inner ischemic
retina with oxygen have used hyperoxygenation of
the choroidal circulation to increase the diffusion
distance of oxygen into the inner retina. Flower and
Patz," using cats and rhesus monkeys, demonstrated
that during occlusion of the retinal circulation, ventilation with 100% O2 at 1 atmosphere pressure was not
a.
b.
d.
50.V
l_
20msec
Fig. 5. Sequence of ERG recordings made to illustrate the effect
of 100% O2 breathing during retinal circulation occlusion: (a)
preocclusion, (b) after 15 min of retinal circulation occlusion, (c)
after 5 min breathing 100% O2, (d) 4 min after returning the ventilation to 20% O2, 80% N2.
No. 2
OXYGEN SUPPLY TO ISCHEMIC RETINA / Den-Nun er ol.
sufficient to completely restore inner retinal function,
as measured by the ERG b wave amplitude. They
concluded that ventilation with hyperbaric oxygen
was necessary for normal inner retinal function. In
contrast, the results of Landers'14 experiments on cats
and monkeys agree with our data (Fig. 5) that ventilation with 100% O2 at 1 atmosphere can restore the
ERG b wave amplitude when the retinal circulation is
occluded.
Using a distant hyperoxygenated source such as the
choroid to supply ischemic inner retina has the disadvantage that oxygen toxicity may occur. There is
evidence that ventilation with hyperbaric oxygen is
toxic to the retina and there appears to be a relationship between the oxygen pressure, the time for which
the hyperoxygenation occurs and the degree of toxicity. Bridges6 found that breathing 100% oxygen in
hyperbaric conditions caused immediate changes in
the ERG response with complete irreversible extinction of the ERG within a few hours. He found an
inverse relationship between hyperbaric oxygen pressure and the extinction time of the ERG. His estimated extinction time of the ERG for the case of
100% oxygen breathed at 3.3 atmospheres was less
then 3 hr. He also noted that the ERG changes during
hyperoxia are similar to changes caused by poisoning
the respiratory metabolic pathway. Margolis and
Brown8 observed a typical cotton wool spot (axoplasmic cytoid body) in the retina of dogs receiving 100%
hyperbaric oxygen for 4 hr. Yanoff et al5 reported
choroidal and retinal detachments in dogs after 48 hr
of exposure to 100% oxygen in 1 atmosphere pressure. From these data it is reasonable to presume that
any attempt to supply sufficient oxygen to the inner
retinal layers via the choroidal circulation could
cause oxygen toxicity of the outer retina in a short
time. It should also be remembered that hyperoxygenation of the choroidal circulation means hyperoxygenation of the whole body, with all the resultant
toxic effects on other organs of the body.12
Using the transvitreal approach described in this
paper, oxygen is provided directly to the ischemic
tissue, and any systemic influence is avoided. Although the vitreal and inner retinal PO2 were not
measured, it can be predicted that whatever the oxygen diffusion gradient is from the catheter, the most
ischemic region of the retina (the inner retina), because of its propinquity to the oxygen source, would
benefit most from vitreal oxygenation.
The ERG b wave is a good indicator of the functional integrity of the bipolar cells and hence the
inner nuclear layer.9 Fujino and Hamasaki13 found
that the b wave of the ERG in squirrel monkeys was
extinguished within 30 sec after occlusion of the reti-
Downloaded From: http://iovs.arvojournals.org/ on 07/31/2017
303
nal circulation (with an intact choroidal circulation).
Landers14 found that the ERG and VEP were extinguished by occluding the retinal circulation in cats
while the animal was breathing 20% oxygen, and that
these responses were returned to normal by changing
to a 100% inspired oxygen concentration. We have
verified their results in our experiments (Fig. 5). During active central retinal artery occlusion in humans,
both photopic and scotopic b waves are almost extinguished.10 These data support the assumption that
the ERG b wave is a good qualitative measure of the
effect of our transvitreal oxygen source on the ischemic inner layers of the retina. Our interpretation of
the recovery of the b wave during the occlusion of the
retinal vessels is that enough oxygen was supplied to
part of the ischemic inner retina to overcome the
ischemia in that region. The rapid disappearance of
the recovered ERG, after the catheter was clamped,
with the circulation still occluded, supports this assumption.
It should be remembered that the catheter design
and the way that it was used are constrained by its
original purpose (hemodialysis). Modification of the
fiber's structure, in particular pore size, wall thickness
and material could result in a better catheter design
for ocular uses with a more efficient oxygen delivery.
Also, development of a method for accurate placement of the fibers within the vitreous so that the important retinal areas, such as the macula, are closest
to the oxygen source would be advantageous.
We suggest that this method could be useful as a
temporary solution for acute impending central retinal artery occlusion, by giving the retina some extra
hours of survival either for spontaneous recovery to
occur or for complementary treatment to be effective,
without any systemic or retinal damage due to hyperoxygenation. Further investigations and technological improvements are required before clinical use can
be considered. It is important to understand the dynamics of the oxygen transport from the catheter to
the retina and inside the various layers of the retina.
This knowledge undoubtedly will influence the catheter design and the way it will be used.
Key words: retina, oxygen, ischemia, cat, central retinal
artery occlusion
Acknowledgments
We wish to acknowledge the expert technical assistance
of Mr. Michael Brown and Mr. Peter Burrows.
References
1. Dollery CT, Bulpitt CJ, and Kohner EM: Oxygen supply to the
retina from the retinal and choroidal circulations at normal
304
2.
3.
4.
5.
6.
7.
8.
INVESTIGATIVE OPHTHALMOLOGY 6 VISUAL SCIENCE / February 1988
and increased arterial oxygen tensions. Invest Ophthalmol
8:588, 1969.
Alder VA, Cringle SJ, and Constable IJ: The retinal oxygen
profile in cats. Invest Ophthalmol Vis Sci 24:30, 1983.
Patz A: Oxygen inhalation in retinal arterial occlusion. Am J
Ophthalmol 40:789, 1955.
Anderson B, Saltzman HA, and Heyman A: The effects of
hyperbaric oxygenation on retinal arterial occlusion. Arch
Ophthalmol 73:315, 1965.
Anderson B Jr: Ocular effects of changes in oxygen and carbon
dioxide tension. Trans Am Ophthalmol Soc 66:424, 1968.
Bridges WZ: Electroretinographic manifestations of hyperbaric oxygen. Arch Ophthalmol 75:812, 1966.
Yanoff M, Miller WW, and Waldhausen JA: Oxygen poisoning of the eyes. Arch Ophthalmol 84:627, 1970.
Margolis G and Brown IW Jr: Hyperbaric oxygenation: The
eye as a limiting factor. Science 151:466, 1966.
Downloaded From: http://iovs.arvojournals.org/ on 07/31/2017
Vol. 29
9. Brown KT: The electroretinogram: Its components and their
origins. Vision Res 8:633, 1968.
10. Flower RW, Speros P, and Keynon KR: Electroretinographic
changes and choroidal defects in a case of central retinal artery
occlusion. Am J Ophthalmol 83:451, 1977.
11. Flower RW and Patz A: The effect of hyperbaric oxygenation
on retinal ischemia. Invest Ophthalmol 8:605, 1971.
12. Bean WJ: Problems of oxygen toxicity. In Clinical Applications of Hyperbaric Oxygen, Boerema I, Brummelkamp WH,
and Meijne NG, editors. Amsterdam, Elsevier, 1964, pp.
267-276.
13. Fujino T and Hamasaki DI: The effect of occluding the retinal
and choroidal circulations on the electroretinogram of monkeys. J Physiol 180:837, 1965.
14. Landers MB: Retinal oxygenation via the choroidal circulation. Trans Am Ophthalmol Soc 76:528, 1978.