T H E CLEARANCE OF INTRAVITREAL GENTAMICIN
`
`L. M I C H A E L C O B O , M.D., AND R I C H A R D K. F O R S T E R , M.D.
`Miami, Florida
`
`Intravitreal gentamicin cleared significantly more rapidly from the
`uninflamed aphakic rabbit eye than the phakic control, with a half-life
`of 12 hours in the former and 32 hours in the latter. A similar difference
`was noted in aphakic and phakic eyes with an experimental Staphylo-
`coccus aureus endophthalmitis, but without the same level of signifi­
`cance. Human data from nine aphakic infected eyes undergoing
`reinjection or vitrectomy indicated that therapeutic levels of gentamicin
`are not consistently found 48 hours or more after the initial intravitreal
`injection.
`
`group of patients who underwent repeat
`intravitreal
`injections of antibiotics, a
`procedure that has been used with suc­
`cess in selected cases.1 These data allow a
`critical evaluation of the duration of ther­
`apeutic gentamicin levels and a consider­
`ation of the optimal time for repeating
`intravitreal injections
`in those patients
`who require additional intravitreal antibi­
`otic therapy.
`
`59
`
`Although intravitreal injection of gen­
`tamicin has been used for almost a decade
`in the treatment of postoperative endoph­
`thalmitis in humans, 1 our understanding
`of the dynamics of intravitreal gentami­
`cin, as well as other antibiotics, has been
`based on data derived from studies in
`phakic, uninflamed rabbit eyes. 2 3 Such
`data may be misleading because the es­
`tablishment of a communication between
`the vitreous and the anterior chamber
`might cause more rapid clearance of the
`antibiotic from the vitreous into the aque­
`ous circulation.
`We studied the clearance of gentamicin
`in phakic and aphakic rabbit eyes, both in
`the presence and absence of infection, to
`more closely approximate the most com­
`mon setting for endophthalmitis in the
`postoperative aphakic patient. Addition­
`ally, gentamicin levels were obtained in a
`
`From the Department of Ophthalmology, Univer­
`sity of Miami, School of Medicine, Bascom Palmer
`Eye Institute/Anne Bates Leach Eye Hospital,
`Miami, Florida. This study was supported in part by
`grant G-598 from the Fight for Sight Foundation, and
`grants from the Flournory and Mae Knight Clark
`Research Fund and the Miami Retreat (Dr. Forster).
`Reprint requests to L. M. Cobo, M.D., Duke
`University Medical Center, Box 3802, Durham, NC
`27710.
`
`M A T E R I A L AND M E T H O D S
`Thirty-one albino New Zealand rabbits
`(weighing 2 to 3 kg each) underwent lens
`aspiration in one eye; the unoperated eye
`served as the phakic control. Right and
`left eyes were done alternately, after the
`rabbits were under general anesthesia
`with intravenously administered pento-
`barbitol. A beveled 60-degree corneal
`incision was made, a sector
`iridectomy
`was performed, and
`the anterior
`lens
`capsule incised. A lens fragmenter was
`inserted and all visible lenticular material
`was irrigated from the anterior chamber.
`The posterior capsule was not intention­
`ally incised. The wound was closed with
`resorbable sutures. The rabbits then re­
`ceived a subtenon's injection of 8 mg of
`triamcinalone acetonide and 4 mg of
`gentamicin subconjunctivally. The ani-
`AMERICAN JOURNAL OF OPHTHALMOLOGY 92:59-62, 1981
`
`Exhibit 2065
`Page 01 of 04
`
`

`

`60
`
`AMERICAN JOURNAL OF OPHTHALMOLOGY
`
`JULY, 1981
`
`mais were then observed for two to three
`months with regular slit-lamp examina­
`tions until the eyes were quiet. In the
`first part of the study, 13 of the monocu-
`larly aphakic animals received
`intravit-
`real injections of gentamicin into both
`eyes. A solution of 100 μg/0.1 ml was
`prepared by serial dilutions of stock solu­
`tion of gentamicin. Proparacaine was ap­
`plied topically to both eyes and a 30-
`gauge needle, mounted on a tuberculin
`syringe containing diluted gentamicin so­
`lution, was inserted 2 mm posterior to the
`temporal corneoscleral limbus,
`through
`the pars plana, with the bevel directed
`anteriorly. Before injection of the antibi­
`otic, an anterior chamber paracentesis
`was performed to relieve any increase in
`intraocular pressure. Peyman and associ­
`ates2 demonstrated
`that gentamicin
`is
`cleared exponentially from the rabbit vit­
`reous. By collecting data at fewer points
`(for example, three), more reliable fig­
`ures could be obtained with a given num­
`ber of animals. An exponential decay
`model allowed data to be plotted semilog-
`arithmically to obtain a regression line.
`Rabbits were killed two, 24, or 48 hours
`after
`injection. We isolated the entire
`vitreous body for assay, using a modifica­
`tion of the technique described by Abel
`and Boyle.4 The eyes were promptly enu­
`cleated,
`tagged with an
`identification
`label, and frozen in liquid nitrogen. Each
`globe was dissected, and vitreous was
`isolated and then placed in a graduated
`tube. The concentration of gentamicin
`was subsequently determined by radio-
`immunoassay technique and the concen­
`trations per eye were calculated using the
`known volume of vitreous removed from
`the eye.
`The second group of 18 rabbits was
`subsequently used
`to study antibiotic
`clearance in infected eyes. An endoph-
`thalmitis was consistently created in both
`eyes of each animal, using
`intravitreal
`inoculation of approximately 60 organ­
`
`aureus} Thirty-
`isms of Staphylococcus
`two hours after infection, at which time
`clinical endophthalmitis was apparent,
`the eyes received 100 μg of gentamicin
`intravitreally. The rabbits were killed
`and studied in the same manner as the
`first, noninfected group.
`
`R E S U L T S
`the
`in
`The gentamicin concentration
`rabbit vitreous, measured in micrograms
`per eye, was plotted on a logarithmic
`scale against the hours after injection of
`antibiotic into the vitreous body. Using
`an exponential decay model, a
`least
`squares
`regression
`line was obtained
`from the data. The data for the aphakic
`and the phakic eyes were compared in
`the first group of noninfected
`rabbits
`(Fig. 1), and in the second group of ani­
`mals with experimental endophthalmitis
`(Fig. 2).
`The clearance of gentamicin from the
`rabbit vitreous is equivalent to the slope
`of the
`line derived from plotting
`the
`
`STERILE EYES:
`
`Fig. 1 (Cobo and Forster). The clearance of 100 μg
`of gentamicin injected into the midvitreous cavity of
`uninfected rabbit eyes.
`
`Exhibit 2065
`Page 02 of 04
`
`

`

`VOL. 92, NO. 1
`
`INTRAVITREAL GENTAMICIN CLEARANCE
`
`61
`
`uu
`
`A
`
`INFECTED EYES:
`
`\
`
`*
`
`T
`
`%* Z
`\
`* '
`\
`: \ H
`*
`\
`\
`\
`\
`\
`\
`*
`\
`'
`s
`\ ·
`\
`\
`\
`
`50
`
`25
`
`10
`
`* *
`\
`v
`^
`. *
`*
`\
`\
`s
`*
`\
`%
`*
`
`\
`\
`\
`\
`\
`\
`\
`\
`\
`V \
`\
`\
`•\
`\
`
`5,
`Γ
`
`24
`
`*
`\
`*
`*
`X
`
`72
`
`48
`HOURS
`Fig. 2 (Cobo and Forster). The clearance of 100 μg
`of gentamicin injected into the midvitreous cavity of
`rabbit eyes infected experimentally with Staphylo-
`coccus aureus.
`
`aphakic model compared to 32 hours in
`the control phakic eye (P < .05).
`In the second group of animals infected
`with S. aureus,
`the difference in rate of
`clearance was less distinct. The half-life
`of gentamicin in the infected aphakic eye
`was 14 hours compared to 19 hours in the
`infected phakic eye (.10 > P > .05).
`Gentamicin levels in vitreous fluid ob­
`tained from human eyes at the time of
`reinjection are listed in the Table, includ­
`ing the time that had elapsed from the
`initial intravitreal injection and the isolat­
`ed organism. All eyes had aphakic en-
`dophthalmitis. Vitreous samples were ob­
`tained either by pars plana
`syringe
`aspiration or at the time of vitrectomy.
`The gentamicin level was determined by
`radioimmunoassay technique.
`
`DISCUSSION
`Experimental clearance data provide
`an estimate of the duration of therapeu­
`tic levels for a given, nontoxic dose of
`an intravitreally administered antibiotic.
`Additionally, clearance data are a guide
`in selected cases where
`the physician
`may wish to intervene with either vitrec­
`tomy or reinjection of antibiotic. Such
`data have generally been derived from
`
`concentration of gentamicin per eye on
`the
`logarithmic scale against
`the
`time
`after injection (hours). In the noninfected
`model, gentamicin was cleared more rap­
`idly from the aphakic eye than the phakic
`eye, with a half-life of 12 hours in the
`
`VITREOUS GENTAMICIN CONCENTRATIONS IN INFECTED HUMAN PATIENTS
`
`TABLE
`
`Patient
`No.
`
`1
`2
`
`3*
`4
`5*
`6*
`7*
`8*
`9
`
`Time
`After Injection
`(hrs)
`
`Vitreous
`Gentamicin
`Concentration
`^g/ml)
`
`20
`31
`
`36
`37
`44
`48
`48
`48
`52
`
`0.90
`>20.00
`
`8.00
`5.00
`6.75
`3.50
`0.25
`10.00
`3.50
`
`"These samples were obtained at the time of subsequent vitrectomy.
`
`Organism
`
`Citrobacter diversus
`Aspergillus flavus,
`Staphylococcus epidermidis
`S. aureus
`S. epidermidis
`Enterococcus Group D
`S. epidermidis
`Klehsietta pneumoniae
`Streptococcus viridans
`S. aureus
`
`Exhibit 2065
`Page 03 of 04
`
`

`

`62
`
`AMERICAN JOURNAL OF OPHTHALMOLOGY
`
`JULY, 1981
`
`uninflamed phakic animal models. This
`study indicates that gentamicin is cleared
`more rapidly in the aphakic rabbit eyes.
`Presumably, removal of the lens allows
`intravitreally administered antibiotic to
`diffuse more readily
`into
`the anterior
`chamber where such drugs may be more
`rapidly cleared through the aqueous cir­
`culation. Although studies in the infected
`model again demonstrated more
`rapid
`clearance in the aphakic eyes, the values
`for drug half-life were much closer in the
`phakic and aphakic eyes and the differ­
`ence was not statistically significant as it
`was in the uninfected eyes.
`If all other factors are equal, experi­
`mental clearance data for the intravitreal
`injection of 100 μg of gentamicin may be
`extrapolated from the 1.5-ml rabbit vitre­
`ous body to the 4-ml human vitreous
`body. Using half-life figures from the un­
`inflamed eyes, where a significant differ­
`ence was demonstrated between phakic
`and aphakic clearance rates, we estimat­
`ed that the gentamicin level would reach
`the minimum therapeutic level of 4 μg/ml
`after 32 hours in the aphakic eye com­
`pared
`to 87 hours
`in the phakic eye.
`Comparison with data obtained from hu­
`mans at first glance demonstrated
`the
`variability expected in the clinical setting
`and the problems of strict reliance on
`animal data. However, we then divided
`the group into those with values obtained
`less than 48 hours and those with values
`obtained more than 48 hours after
`the
`initial injection. Only one of four samples
`obtained after 48 hours contained a mini­
`
`mum inhibitory concentration of genta­
`micin, compared to four of five samples
`obtained before 48 hours. We cannot
`account for the unexpectedly low value in
`Patient 1 or the high level in Patient 2
`other than sampling error or a technical
`error in the determination of antibiotic
`levels.
`These data from experimental animals
`and humans indicate that therapeutic lev­
`els of antibiotic are maintained for 24 to
`48 hours after the initial intravitreal in­
`jection of antibiotic in aphakic eyes. An
`optimal time for reinjection of antibiotic
`in a selected aphakic case may be be­
`tween 36 and 48 hours. Although we have
`no human data on the clearance of genta­
`micin in phakic eyes, experimental ani­
`mal data indicate that the drug may be
`cleared more slowly in these eyes. It may
`be reasonable to postpone reinjection in
`the phakic eye to 72 to 96 hours after
`initial injection to avoid potentially toxic
`intravitreal levels.
`REFERENCES
`1. Forster, R. K., Abbott, R. L., and Gelender,
`H.: Management of infectious endophthalmitis. Oph­
`thalmology 87:313, 1980.
`2. Peyman, G. A., May, D. R., Ericson, E. S.,
`and Apple, D.: Intraocular injection of gentamicin.
`Toxic effects and clearance. Arch. Ophthalmol.
`92:42, 1974.
`3. Zachary, I. G., and Forster, R. K.: Experimen­
`tal intravitreal gentamicin. Am. J. Ophthalmol.
`82:604, 1976.
`4. Abel, R., and Boyle, G. L. : Dissecting ocular
`tissue for intraocular drug studies. Invest. Ophthal­
`mol. 15:216, 1976.
`5. Cottingham, A. J., and Forster, R. K.: Vitrec-
`tomy
`in endophthalmitis. Arch. Ophthalmol.
`94:2078, 1976.
`
`Exhibit 2065
`Page 04 of 04
`
`