Safety of intravitreal linezolid injection: electroretinographic and histopathologic studies in rabbits

BACTERIAL endophthalmitis is an intraocular infection
that frequently results in loss of vision. Early treatment
and rapid elimination of the infecting organisms are vital
to the preservation of vision. Since systemic antibiotics
do not achieve significant vitreous concentrations, intravitreal administration remains a key strategy in the clinical
management of this condition.
The Endophthalmitis Vitrectomy Study (EVS) has
recommended a combination of vancomycin plus either
amikacin, ceftazidime, or ciprofloxacin, which all have
activity against Gram-negative bacilli, as initial treatment
for bacterial endophthalmitis.
1
Vancomycin is considered
the drug of choice for a range of infections and is nontoxic
at a dose of 1.0 mg in 0.1 mL.
2
Aminoglycosides, such as
amikacin or gentamicin, are often used but intravitreal
injection of gentamicin has been reported to cause
macular toxicity,
3
and the risk also extends to amikacin
even though animal experiments have shown it to be safer
than gentamicin.
4-5
Ceftazidime, in contrast, carries a lower
risk of retinal toxicity and has a broader therapeutic
index,
6-7
but it precipitates in the vitreous humor at body
temperature.
6
The emergence of vancomycin-resistant Enterococci
species and the potential for macular toxicity of the other
intravitreal drugs emphasize the need for alternative drugs
for intravitreal injections. Linezolid is a synthetic
antibacterial agent belonging to a new class of antibiotics,
the oxazolidinones, which is indicated for infections
caused by aerobic Gram-positive bacteria. The in vitro
spectrum of activity also includes certain Gram-negative
and anaerobic bacteria. Linezolid inhibits bacterial
protein synthesis through a mechanism of action different
from that of other antibacterial agents; therefore, crossresistance with other classes of antibiotics is unlikely. It
binds to a site on the bacterial 23S ribosomal RNA of the
50S subunit and prevents the formation of a functional
70S initiation complex, an essential component of the
bacterial-translation process.
8
In vitro studies have demonstrated additivity or
indifference between linezolid and vancomycin,
gentamicin, rifampin, imipenem–cilastatin, aztreonam,
ampicillin, or streptomycin antimicrobial activity.
8
Linezolid has been shown to be active against most isolates
of aerobic and facultative Gram-positive microorganisms
including Enterococcus faecium (including vancomycinresistant strains), Staphylococcus aureus (including
methicillin-resistant strains), Enterococcus faecalis (including
vancomycin-resistant strains), Staphylococcus epidermidis
(including methicillin-resistant strains), Streptococcus
pneumoniae (including multidrug-resistant isolates,
MDRSP); and Gram-negative microorganisms.
9-10
This study evaluated the safety of intravitreal injection
of linezolid (Zyvox, Pfizer, New York, NY, USA) in albino
rabbits based on the following parameters: precipitates
in the vitreous, effect on electroretinography (ERG), and
any inflammatory or necrotic changes in the retina.

METHODOLOGY
Eight albino rabbits were divided randomly into 2 equal
groups. Group 1 received 100 mcg/0.10 mL and Group
2 received 200 mcg/0.1 0 mL. They were handled
according to the tenets of Association for Research in
Vision and Ophthalmology in relation to animal research.
The rabbits were anesthetized with intramuscular
injection of ketamine (Ketaject, PharmAsia-Cuvest, Makati
City, Philippines) at a dose of 65 mg/kg and 2 drops of
0.5% proparacaine hydrochloride (Alcaine, Alcon
Laboratories, Fort Worth, TX, USA). Ciprofloxacin
ophthalmic solution was applied to the eyes before
injection and tobramycin ophthalmic ointment after
injection.
The right eyes of the rabbits received intravitreal
linezolid and the left eyes 0.10 mL balanced salt solution.
Intravitreal injection with a gauge 27 needle on tuberculin
syringe was given 2 millimeters posterior to the temporal
limbus. The needle was angled toward the optic nerve
until the tip was visible in the center of the vitreous,
approximately 3 millimeters in depth.
All eyes were examined 1 day before intravitreal
linezolid injection, 3 hours after injecton, and on day 2
and day 7 postinjection. Pupils were dilated with 0.5%
tropicamide and 2.5% phenylephrine for indirect
ophthalmoscopy to check for vitreous precipitates,
opacities, and hemorrhages.
Prior to ERG (Nihon Kohden Neuropack), the rabbits
were kept in a dark room for 20 minutes. The recording
electrode was placed on the lower lid, the reference
electrode centrally on the shaven forehead, and the
ground electrode on the earlobe. Micropore tapes were
placed on the electrodes for stability. ERG consisting of
photopic, scotopic, bright flash, and 30-Hertz flicker tests
were obtained 3 hours, 2 days, and 7 days postinjection.
The implicit time and amplitude of the corresponding
wave component of each test were measured.
Color and red-free (green) fundus photos of each eye
were also taken after indirect ophthalmoscopy on day 7
postinjection using a Zeiss Visupack FF450 (Zeiss
Corporation, Oberkochen, Germany) fundus camera.
Afterwards, the rabbits were euthanized with an overdose
of intraperineal ketamine. The eyes were enucleated and
fixed immediately in 10% formaldehyde. After 48 hours
in the fixative, gross examinations of the tissues were
performed. The enucleated eyes were trisected parallel
to the cornea-optic-nerve axis. Tissues were then
embedded in paraffin, sectioned at a thickness of 6 µm,
and stained with hematoxylin-eosin. The retinas were
evaluated for inflammatory and necrotic changes by an
ocular pathologist blinded to the study regimen. Light
microscopy was used for examination of the retina.
ERG data were entered into Microsoft Excel (Microsoft
Corp., Redwood, WA, USA) for statistical analysis. The
implicit time and amplitude of the corresponding waves
produced in photopic, scotopic, bright flash, and 30-Hertz
flicker tests of Group 1, Group 2, and control eyes were
compared with one another and at different times
postinjection using paired t-test. A p-value of less than or
equal to 0.05 was considered statistically significant.

RESULTS
Indirect ophthalmoscopy showed no opacity, precipitation, or hemorrhage in the vitreous in all eyes. Vitreous
was clear after linezolid and BSS injection (Figures 1-4).
The characteristic electroretinogram of each test are
shown in Figures 5 and 6. In the scotopic test (rod
response), a statistically significant decrease in the b-wave
amplitude (p < 0.05) was noted in Group 2 between 3
hours and 2 days (p = 0.04) and between 3 hours and 7
days (p = 0.04) after the administration of linezolid,
without a concurrent implicit time change. In the
photopic, bright flash, and 30-Hertz flicker tests, the
implicit time and amplitude of corresponding waves
showed no significant change at any dose or at any time
after drug administration. Table 1 shows the p values of
the dose-dependent ERG implicit time and amplitude
responses between the different time periods and the p
values of time-dependent ERG implicit time and
amplitude responses between the different groups.
Histopathologic findings (Figure 7) included the
presence of minimal inflammatory cells (<3 cells/hpf) in
the nerve-fiber layer in 2 rabbit eyes in Group 2 and in 1
control eye; vacuolizations in the ganglion-cell layer in 3
eyes in Group 1 and in 2 control eyes; minimal decrease
in the cell density in the outer nuclear layer in 2 eyes in
Group 1, in 1 eye in Group 2, and in 3 control eyes; and
partial loss of photoreceptor outer segment in 2 eyes in
Group 1 and in 2 control eyes. The remaining retinal layers
did not show any inflammatory or necrotic changes. The
anterior segments of all eyes were normal.

DISCUSSION
The data in this experiment suggest linezolid has
minimal effects on ERG and few histopathologic changes
in the retina. No vitreous precipitates were seen.
ERG is a diagnostic tool used to detect retinal pathology.
The two parameters of the ERG that are measured
clinically are the amplitude (measured in microvolts, µV)
and the implicit time (measured from the onset of the
stimulus to the peak of the response in milliseconds). The
anatomical and functional similarities between human and
rabbit retinas have made the latter appropriate models
for studying drug toxicity even though the retinal anatomy
of the rabbit is slightly different. For example, the human
retina has three types of cone cells whereas rabbit retina
has only two.
11, 12
Thus, it is preferable to use time- and
dose-dependent ERG changes in a given group to detect
any toxicity.
No ERG implicit time and amplitude changes were
observed in all the ERG tests on time-dependent
comparisons. Except for a statistically significant decrease
in the b-wave amplitude in the scotopic test of Group 2
without a concurrent implicit time change observed
between 3 hours and 2 days and between 3 hours and 7
days after the administration of linezolid, there were no
other implicit time or amplitude changes observed in the
other ERG tests on dose-dependent comparisons. The
decrease in b-wave amplitude was only noted at 3 hours
and not after 2 days, implying a transient effect on the
retina. Morever, the change was noted only in the scotopic
test. Further investigations are needed to determine a true
toxic effect of linezolid on the retina at higher doses.
Histopathologic examination showed minimal
inflammatory changes and absence of necrosis in the
retinal layers in all groups. Mild changes such as
vacuolization, decreased cell density, and partial loss of
photoreceptor outer segment were seen in both the
treated and the control groups.
Histologic studies of the retina are important in
evaluating the safety of a drug because the photoreceptor
and other retinal cells directly adjacent to the vitreous
are highly sensitive not only to the offending pathogen
and the resulting inflammatory response but also to high
doses of antimicrobial agents administered intravitreally
to treat the infection.
1, 13-16

Intravitreal antibiotic therapy has been the principal
mode of treatment for acute bacterial endophthalmitis
since this route results in a far greater intraocular antibiotic concentration than any other method of administration. Because of the importance of prompt treatment
and the inaccuracies of gram-staining results, broadspectrum intravitreal antibiotics effective against both
gram-positive and – negative bacteria are usually
administered even before the culture results are available.
Linezolid has demonstrated a wide spectrum of activity
against different microorganisms, and its ability to kill
resistant strains warrants its evaluation as a possible
alternative to intravitreal vancomycin. Further studies,
however, are needed to determine its effects in human
eyes even though in vitro studies in rabbit eyes showed no
toxic effects on the vitreous and the retina.