Intraocular Penetration of Systemic Antibiotics in Eyes with Penetrating Ocular Injury

Abstract

Purpose:

To determine whether penetrating scleral or corneal injury can enhance intraocular penetration of systemic moxifloxacin, vancomycin, and ceftazidime.

Methods:

Thirty rabbits were divided into 3 groups for each antibiotic and then further subdivided to receive either scleral or corneal injury to the right eye. The left eye served as a control. Intravenous antibiotics were given following injury, and eyes were subsequently enucleated. Vitreous antibiotic concentration was determined by high-performance liquid chromatography analysis. Plasma concentration was measured for comparison.

Results:

Intravitreal moxifloxacin concentration was unchanged by injury. Minimum inhibitory concentration (MIC₉₀) was achieved in the vitreous against the most common gram-positive endophthalmitis-causing organisms. Intravitreal vancomycin levels were not enhanced by injury and did not reach the MIC₉₀ for gram-positive organisms commonly causing intraocular infection. Intravitreal ceftazidime was increased in the injured eyes, 67% and 73% higher in scleral and corneal injury eyes. It reached MIC₉₀ of many gram-negative bacteria.

Conclusions:

Intravitreal antibiotic penetration of systemic antibiotics with or without penetrating ocular injury varies depending on the antibiotic. For prevention or treatment of gram-positive-bacteria-causing endophthalmitis, intravitreal vancomycin is necessary and provides the most reliable coverage. Systemic ceftazidime can be used for many gram-negative bacteria, but intravitreal injection is recommended for better coverage, especially for more-potent organisms. Systemic moxifloxacin can be considered for most gram-positive and -negative infections due to its excellent intraocular penetration and broad coverage, but the patient's previous history of its topical use and increasing resistance patterns must be considered.

Introduction

Penetrating ocular injury can lead to devastating visual impairment. Post-traumatic endophthalmitis (TE) is a major complication that can often be prevented with timely diagnosis and early intervention.¹ TE rates range from 0% to 16.5% in the current literature, especially higher when with retained intraocular foreign bodies (IOFBs).²

The majority of organisms causing TE are gram-positive organisms, including Staphylococcal and Bacillus species.^3,4 Gram-negative organisms cause a minority of these infections, but typically herald a poor visual prognosis despite aggressive treatment with intraocular antibiotics and vitrectomy.^4,5

Review of recent literature suggests indications for intravitreal antibiotics in cases of traumatic globe rupture with IOFBs.⁶ These indications include a gram stain showing organisms from an intraocular specimen, suspected endophthalmitis, and high-risk cases. In cases of TE, there are various opinions regarding the role and benefit of intravenous antibiotics in prophylaxis or treatment. In fact, the endophthalmitis vitrectomy study has shown that there is no role for systemic antibiotics in the treatment of postsurgical endophthalmitis.⁷ Nevertheless, a recent study that looked at penetrating injury with retained IOFBs showed that systemic antibiotics, specifically fourth-generation fluoroquinolones, resulted in no cases of endophthalmitis even with delayed removal of the IOFBs.⁸

Newer-generation fluoroquinolones have demonstrated broad spectrum coverage with excellent intraocular penetration in animal studies when given systemically.^9,10 Two antibiotics, vancomycin and ceftazidime, commonly used intravitreally to treat endophthalmitis for their excellent gram-positive and gram-negative coverage, have limited intraocular penetration when systemically administered.^4,11–15 Provided good intraocular penetration, these antibiotics would be ideal choices for prophylaxis in penetrating ocular injury. Our study investigates whether penetrating ocular injury can improve intraocular penetration of 3 systemic antibiotics—moxifloxacin, vancomycin, and ceftazidime—in a rabbit animal model.

Methods

Thirty New Zealand white rabbits weighing 1.72–2.70 kg were obtained from Myrtle's Rabbitry, Inc. (Thompson Station, TN). The animals were housed in room lighting with a 12-h light-dark cycle. All experiments were performed in accordance with the ARVO statement for the use of animals in ophthalmic and vision research, and the protocol was approved by the Institutional Animal Care and Use Committee of Henry Ford Health System.

Anesthesia was induced with a ketamine/xylazine (50/10 mg/kg) mixture given intramuscularly. Surgical anesthesia was maintained with isoflurane using a pediatric facemask. Proparacaine hydrochloride 0.5% (Alcon Laboratories, Inc., Fort Worth, TX) was used for additional topical anesthesia prior to penetrating injury.

Ocular injury was performed using nonsterile techniques to simulate real-world traumatic conditions. Thirty rabbits were divided into 3 experimental groups, each with 10 rabbits. Each group was designated to receive 1 of the 3 antibiotics studied, vancomycin, moxifloxacin, or ceftazidime.

Within each group, half (5 rabbits) received a scleral incision as a traumatic scleral injury model, and the other half (5 rabbits) received a central clear corneal incision as a traumatic corneal injury model. All injuries were performed in right eyes, while the left eyes were uninjured to serve as controls. All penetrating ocular injuries were performed using a clean, but nonsterile, No. 11 scalpel. Scleral injuries, 3–4 mm in length, were made 5-mm posterior to the limbus in the infero-temporal quadrant using a stab and twist technique to create an irregular wound. Linear corneal injuries, 4 mm in length, were made in the central cornea, being careful to avoid violating the limbus, using a stab incision.

Ten minutes following ocular injury, each animal was administered intravenous antibiotics based on their experimental group: vancomycin (15 mg/kg), moxifloxacin (20 mg/kg), or ceftazidime (50 mg/kg). Thirty minutes following ocular injury, rabbit eyes were enucleated bilaterally and immediately frozen in liquid nitrogen and then placed at −80°C for storage. Blood samples were obtained concomitant with enucleation to determine serum antibiotic concentrations. This was done using direct, transcutaneous, intracardiac aspiration using a long 18-gauge needle. The animals were then euthanized with a lethal dose of Euthasol (pentobarbital sodium and phenytoin sodium; Virbac AH, Inc., Fort Worth, TX).

Ocular dissection and isolation of the entire frozen vitreous was then performed according to the technique described by Abel and Boyle.¹⁶ Analysis for antibiotic concentration was done by the Anti-infective Research Laboratories at the University of Houston using high-performance liquid chromatography (HPLC). One rabbit without ocular injury or systemic antibiotic was euthanized and its vitreous was harvested to serve as an HPLC control.

HPLC analysis of the samples was carried out in a masked fashion by the HPLC operator. Pure antibiotic standard samples were used to create standard curves for comparison to vitreous and plasma specimens.

Statistical analysis, using the Student's t-test, was performed to determine whether a significant difference existed between injured and control eyes in the 3 experimental groups and between scleral and corneal injury groups. Plasma antibiotic concentrations served as reference controls to determine the percentage of penetration of each antibiotic.

Results

Moxifloxacin

In the moxifloxacin group, the average serum concentration was 6.34 and 7.32 μg/mL in the corneal and scleral injury groups, respectively. The eyes receiving corneal injury had an average intravitreal concentration of 3.28±0.54 μg/mL. The scleral injury eyes had an average intravitreal concentration of 3.34±0.44 μg/mL. In comparison, the intravitreal concentration in control eyes without injury was 3.26±0.50 and 3.48±0.83 μg/mL in the corneal and scleral injury eyes, respectively (Table 1). There is no statistically significant difference between the injured and control eyes in intravitreal concentration of moxifloxacin (P=0.37 for corneal injury; P=0.62 for scleral injury). In the corneal injury group, the intravitreal concentration reached 51% of plasma concentration. In the scleral injury group, the intravitreal concentration reached 47% of plasma concentration.

Table 1.

Results of Antibiotic Concentration

	Scleral injury		Corneal injury
	Injury eye	Control eye	Injury eye	Control eye
Moxifloxacin
Vitreous concentration	3.34 μg/mL	3.48 μg/mL	3.28 μg/mL	3.26 μg/mL
Serum concentration	7.32 μg/mL		6.34 μg/mL
Vitreous/serum percentage^a	47%		51%
Injury vs. control eye^b	P=0.62		P=0.37
Sclera vs. cornea injury eye^c	No statistical difference
Vancomycin
Vitreous concentration	0.44 μg/mL	0.28 μg/mL	0.22 μg/mL	nd
Serum concentration	20.4 μg/mL		18.8 μg/mL
Vitreous/serum percentage^a	1%–2%		1%–2%
Injury vs. control eye^b	Concentration too low^d		Concentration too low^d
Sclera vs. cornea injury eye^c	Concentration too low^d
Ceftazidime
Vitreous concentration	1.78 μg/mL	1.07 μg/mL	1.50 μg/mL	0.74 μg/mL
Serum concentration	127.0 μg/mL		135.0 μg/mL
Vitreous/serum percentage^a	1.40%		1.10%
Injury vs. control eye^b	P=0.04		P=0.03
Sclera vs. cornea injury eye^c	19% increase in scleral than in corneal injury eyes

Concentrations as detected by HPLC.

Serum concentration percentage of total vitreous concentration.

Statistical analysis of the difference between injury and control eyes.

Analysis of the difference between scleral and corneal injury eyes.

Concentrations being analyzed are in the range of HPLC background noise.

HPLC, high-performance liquid chromatography; nd, levels not detectable by HPLC.

Vancomycin

In the vancomycin group, the average serum concentration was 18.80 and 20.40 μg/mL in the corneal and scleral injury groups, respectively. The eyes with corneal injury had an average intravitreal concentration of 0.22±0.01 μg/mL. The scleral injury eyes had an average intravitreal concentration of 0.44±0.10 μg/mL. In comparison, the intravitreal concentration in control eyes without injury was 0.28±0.03 μg/mL to nondetectable. These levels are very low and fall within the range of standard error due to HPLC baseline noise, so the statistical and clinical differences may be irrelevant. In the corneal injury group, there is no calculated statistically significant difference between the intravitreal concentration of vancomycin in the control and injured eyes (P=0.178). In the scleral injury group, the calculated P value is lower at 0.019, and this may hint at a statistically significant difference between the intravitreal concentration of vancomycin in the control and injured eyes; however, no reliable analysis can be done since the levels fall within the range of HPLC baseline noise. These concentrations represent 1% (corneal injury group) to 2% (scleral injury group) of plasma concentration. Thus, intravitreal penetration of systemic vancomycin is very limited, even with scleral or corneal penetrating injury.

Ceftazidime

In the ceftazidime group, the average serum concentration was 135.00 and 127.00 μg/mL in the corneal and scleral injury groups, respectively. The eyes with corneal injury had an average intravitreal concentration of 1.50±0.56 μg/mL. The scleral injury eyes had an average intravitreal concentration of 1.78±0.54 μg/mL. In comparison, the intravitreal concentration in control eyes without injury was 0.74±0.33 and 1.07±0.35 μg/mL in the corneal and scleral injury groups, respectively. There is a statistically significant difference between the intravitreal concentration of ceftazadime in the control and injured eyes in both the corneal and scleral injury groups (P=0.03 and 0.04, respectively). There is an increase in intravitreal concentration caused by corneal injury of 73% and 67% for scleral injury. These concentrations represent 1.1% (corneal injury group) to 1.4% (scleral injury group) of plasma concentration.

Discussion

TE rates range from 2.4% to 16.5% in the current literature and can be affected by numerous factors.^2,17–20 These risk factors include increased age, rural areas, vegetative matter, virulence of the organism, disruption of the crystalline lens, and delay in primary wound closure or delay in removal of IOFBs.^2,3,21,22 In cases with retained IOFBs, rates of endophthalmitis occupy the higher end of the spectrum.^1,2,22–25 The difficulty in treating TE necessitates obtaining a better understanding of the pharmacokinetics of antibiotics administered to patients with penetrating ocular trauma, and whether systemic antibiotics can serve a prophylactic or therapeutic role to improve outcomes.

The degree of ocular penetration of a systemic antibiotic is an important consideration when selecting an antimicrobial for empiric treatment in these cases of penetrating ocular injury, especially when many physicians give systemic antibiotics routinely. Our study showed superior penetration of intravenous moxifloxacin into the vitreous cavity compared with ceftazadime, and poor penetration of vancomycin in a rabbit, traumatic ocular injury model. The rabbit model has previously been shown to be suitable for evaluation of the activity, toxicity, and pharmacokinetics of antimicrobials in experimental endophthalmitis, as the globe size, aqueous humor turnover rate, and blood-ocular barriers are comparable to those of human eyes.^26,27

Moxifloxacin

Fluoroquinolones have grown in popularity for treatment of a variety of infections. They have a broad spectrum of bacterial coverage but are most efficacious against aerobic gram-negative bacilli, such as Pseudomonas aeruginosa. Moxifloxacin is a newer, fourth-generation fluoroquinolone that has enhanced coverage against gram-positive cocci and anaerobic bacteria.^28,29 However, this is at the expense of decreased efficacy against Ps. aeruginosa (MIC₉₀=8 μg/mL) compared with other fluoroquinolones, such as ciprofloxacin (MIC₉₀=4 μg/mL).³⁰

Various studies have shown robust intravitreal concentrations of moxifloxacin after systemic administration. Lott et al. found intravitreal concentrations in patients undergoing surgery to be as high as 0.811 μg/mL after 1 dose of 400 mg oral moxifloxacin 3 h after administration, and 1.845 μg/mL after 5 doses of 400 mg of oral moxifloxacin over 5 days.³¹ This exceeds the MIC₉₀ levels of most endophthalmitis-causing bacteria, such as Staphylococcus and Streptococcus species, Bacillus cereus, and Enterococcus faecalis, with the exception of Pseudomonas.^31,32 Tzepi et al. also found good intraocular penetration of moxifloxacin after intravenous administration in a rabbit model where intravitreal concentrations reached 1.68 μg/mL ∼30 min after injection.¹⁰ Bronner et al. showed increased intravitreal penetration of moxifloxacin in a rabbit model with ocular inflammation from endophthalmitis, 1.75 μg/mL 30 min following intravenous injection, compared with 1.56 μg/mL in the control eyes.³³ Vitreous levels peaked at 2 h (2.50 μg/mL in the infected group versus 1.66 μg/mL in the uninfected group), showing that moxifloxacin will continue to penetrate into the inflamed eye.

In our study, the intravitreal concentration of moxifloxacin reached 3.34±0.44 μg/mL in the scleral injury group and 3.28±0.54 μg/mL in the corneal injury group. In the control groups, the intravitreal concentration averaged 3.37 μg/mL. These concentrations represent 47%–51% of plasma concentration, which was the highest intravitreal:plasma concentration achieved by any of the 3 antibiotics used in our study, reflecting its excellent intraocular penetration. However, the intravitreal concentrations in the injury groups were not significantly different from the control groups (P=0.37 and 0.62 for the corneal and scleral injury groups). These levels reached the MIC₉₀ for all common endophthalmitis-causing organisms, including Staphylococcal species (MIC₉₀=0.06–0.2 μg/mL), Streptococcus species (MIC₉₀=0.12–0.25 μg/mL), Proteus (MIC₉₀=0.25–0.8 μg/mL), and Moraxella (MIC₉₀=0.06–0.12 μg/mL). However, it did not reach effective levels against Pseudomonas species (MIC₉₀=8 μg/mL), and Acinetobacter (MIC₉₀=16–32 μg/mL).³⁰

Bronner et al. showed that moxifloxacin penetrated the vitreous cavity better in eyes with inflammation from endophthalmitis in his rabbit model as described earlier.³³ In their model, the iatrogenic ocular infections were allowed to fester for 24 h, allowing for a significant immune response to the inoculated bacteria, including recruitment of inflammatory mediators that likely allowed increased permeability of the intraocular vasculature as suggested by the authors. Since moxifloxacin has excellent penetration even into the intact eye, this increased permeability would explain the improved penetration of the antibiotic in the infected eye. Penetrating ocular injury will also result in inflammation over time; however, moxifloxacin levels were not increased in the injured eyes in our study. This may be explained by the short window (30 min) from the time of ocular wound creation to enucleation for sample retrieval. In theory, if the vitreous specimen was to be isolated much later after injury, allowing time for more inflammation to develop, then moxifloxacin levels should be higher, consistent with Bronner et al.'s late findings in his study.³³ Since real-world patients with ruptured globe injuries are taken for repair emergently, the short timeframe used in our model may be more appropriate for evaluating the pharmacodynamics of intraocular penetration of systemic antibiotics solely from penetrating ocular injury.

More recently, there has been concern for growing resistance to the newer-generation flouroquinolones. Prior studies have demonstrated that in vitro susceptibilities of isolates from bacterial keratitis and endophthalmitis have increasing resistance to fluoroquinolones.^34,35 However, these studies largely tested older-generation fluoroquinolones. A review by Hwang found that moxifloxacin is less likely to be affected by single-step topoisomerase mutations that cause resistance to older fluoroquinolones.³⁶ Park et al. showed that normal bacterial flora from the ocular surface are highly susceptible to fourth-generation fluoroquinolones, including gatifloxacin and moxifloxacin.³⁷ Moxifloxacin has a 4.5% overall resistance rate to all organisms isolated, with no resistance seen in gram-negative and Pseudomonas species, whereas nearly a quarter of all isolates were resistant to ciprofloxacin. However, prior use of topical moxifloxacin has been shown to significantly alter the resistance of ocular flora. In a recent study, Yin et al. showed that patients receiving prophylactic topical moxifloxacin after intravitreal injections for macular degeneration have a 4-fold increase in the average MIC₉₀ levels of the normal ocular flora.³⁸ A subanalysis of the Steroids for Corneal Ulcers Trial showed a 3.5-fold increase in the MIC₉₀ for various isolates when patients received a short pretreatment course of moxifloxacin.³⁹ The authors concluded that any prior use of flouroquinolones is associated with increased resistance patterns. In addition, a recently published 10-year review of patients with postoperative endophthalmitis by Schimel et al. has shown an increasing resistance trend to the flouroquinolones at large, ranging from 0% to 10% in the early 1990s, 30%–40% by 2000 to 2005, and about 60% from 2005 to 2011.⁴⁰ Based on this data, the use of prophylactic or therapeutic systemic moxifloxacin in patients with ruptured globe injuries is condonable since most of these patients have not had prior topical moxifloxacin exposure based on demographic trends. Alternative antibiotics may need to be considered in older patients with prior ocular surgery or procedures who may have been treated with topical moxifloxacin in the past.

A retrospective review by Colyer et al. highlights the potential role of newer-generation flouroquinolones, such as moxifloxacin, in preventing post-TE.⁸ The study looked at the incidence of endophthalmitis in 79 eyes of soldiers during the Iraq war with penetrating ocular injury and delayed removal of IOFBs. Mean time to IOFB removal was 38 days (range of 2–661 days), and no patients developed TE. These patients were largely treated with systemic newer-generation fluoroquinolones. The absence of endophthalmitis in these cases is likely due to a combination of factors, including excellent intravitreal penetration of fluoroquinolones, fluoroquinolone-naive ocular flora among soldiers, and low resistance to newer fluoroquinolones in Iraq due to poor availability of these antimicrobials. Arguments have been made that war-related injuries with IOFBs due to high-velocity projectiles may self-sterilize prior to entering the eye.⁴¹ However, 43% of the IOFBs in Colyer study were nonmunitions particles, including glass, stone, and organic material, unlikely to be sterilized. Although these results are not directly translatable to ocular injury we routinely see since IOFBs are quickly removed, it does provide some indication for the effectiveness of the systemic treatment of fourth generation of fluoroquinolones.

We found that concentrations of moxifloxacin in the vitreous exceeded the MIC₉₀ levels for the most prevalent endophthalmitis-causing microorganisms except Pseudomonas species. Our data suggest that moxifloxacin is a viable option for broad-spectrum empirical coverage in eyes with penetrating injury. Care must be taken to assure that patients do not have an extensive prior history of topical moxifloxacin use, in which case ocular flora that may have contaminated the wound would show greater resistance. Nevertheless, moxifloxacin provides a convenient and widely available option to the armamentarium against intraocular infection.

Vancomycin

Post-TE is most commonly caused by gram-positive organisms, such as Staphylococcus and Streptococcus species, which are routinely found on the normal ocular surface.^41–45 B. cereus has been recovered in 25%–50% of culture-positive cases and is known to lead to a very fulminant infection.⁴¹ For this reason, many prior studies have focused on evaluating the efficacy of antibiotics that are effective against gram-positive bacteria in treating or preventing TE.^27,46,47 Due to resistant strains, such as Methicillin-resistant Staphylococcus aureus, and also their broad gram-positive coverage, intravitreal vancomycin has been the antibiotic of choice for treating presumed or fulminant endophthalmitis from trauma or otherwise. Intraocular drug penetration is dependent upon the physical and chemical characteristics of the drug, the manner of drug administration, and the drug's ability to pass through the blood-ocular barrier. Most systemically administered drugs, including vancomycin, do not achieve therapeutic intraocular levels.⁴⁶ Studies have shown that vancomycin has unreliable penetration into the vitreous cavity in the uninflammed eye.^12,46 If penetrating ocular trauma increased systemic vancomycin penetration into the eye, then it would offer a prophylactic and therapeutic option for TE.

Early animal studies looked at intraocular penetration of vancomycin when given intravenously, avoiding doses above 2 g, which can lead to ototoxicity and hearing loss in humans.⁴⁶ Pryor et al. showed that vancomycin levels are detectable in the aqueous in a rabbit model, more so in the inflamed eye, but undetectable in the vitreous up to 4 h after intravenous administration.⁴⁶ Meredith et al. showed that surgical lensectomy in a rabbit model will increase penetration of systemic vancomycin into the vitreous cavity to levels well above the MIC₉₀ of common gram-positive endophthalmitis-causing organisms, more so in the vitrectomized eye.¹³ Ferencz et al. studied ocular penetration in patients with postoperative endophthalmitis and found erratic penetration by a single dose of intravenous vancomycin.¹² Six of 14 eyes showed adequate vitreous concentrations against Staphylococcus epidermidis, and the authors concluded that intravenous vancomycin therapy should not be solely relied upon in cases of endophthalmitis due to gram-positive cocci.

In our study, the intravitreal concentration of vancomycin reached 1%–2% of plasma concentration, which was the lowest intravitreal:plasma concentration achieved by any of the 3 studied antibiotics. There was no difference (P=0.178) between the vitreous concentrations in control, uninjured eyes (0.28±0.03 μg/mL) and the corneal injury group (0.22±0.01 μg/mL). However, we did find a statistical difference (P=0.019) between the vitreous concentrations in the control eyes (0.28±0.03 μg/mL) and scleral injury group (0.44±0.10 μg/mL). Again, this difference may be irrelevant due to such low concentrations being within the standard error of HPLC baseline noise. Additionally, this result is clinically insignificant since it did not reach the MIC₉₀ for the most prevalent gram-positive, endophthalmitis-causing organisms, including Sta. aureus (1–2 μg/mL), S. pneumonia (0.5 μg/mL), S. epidermis (4 μg/mL), E. faecalis (2 μg/mL), and B. cereus (2 μg/mL).⁴⁸

Pryor et al. showed that surgical inflammation can increase systemic vancomycin penetration into the eye.⁴⁶ Our results suggest that the acute inflammation from a penetrating corneal wound does not result in sufficient breakdown of the blood-ocular barrier to allow improved vancomycin penetration into the vitreous cavity. Scleral injury may result in minimally increased vitreous penetration of vancomycin. Higher systemic dosing may lead to higher vitreous concentrations in the setting of inflammation; however, this is limited by significant side effects when treating with levels over 2 g per day. Ferencz et al. demonstrated that intravitreal vancomycin levels were higher 4–5 h after intravenous injection in eyes with endophthalmitis, suggesting that there is continued penetration into the vitreous cavity in the inflamed eye.¹² Similar to moxifloxacin, in theory, if the vitreous specimen was isolated later after injury, allowing time for more inflammation to develop, then vancomycin levels should be higher. However, the increases in levels with time were inconsistent in Ferencz et al. study, and the majority of specimens did not lead to therapeutic concentrations. On the contrary, a single intravitreal dose of vancomycin provides a very reliable response, leading to effective levels for at least 3 days.¹² For this reason, we cannot advocate systemic vancomycin therapy for prophylaxis or treatment of TE due to gram-positive organisms, and intravitreal injection of vancomycin should remain the mainstay in treatment with concurrent vitrectomy when indicated.

Ceftazidime

Post-TE caused by gram-negative organisms is much less common than those caused by gram-positive organisms, with the incidence ranging anywhere from 7.7% to 33% of all cases.^44,49–51 However, these infections tend to have a poorer prognosis due to the fastidious nature of the organisms, especially Ps. aeruginosa that accounts for up to 28% of all gram-negative trauma-related infections either with or without an associated IOFBs.^{5,17,44,52–54} Reports have shown that eyes with gram-negative rod infections lose vision quickly.^41,55 Thus, efficacy against Pseudomonas should be a priority when selecting an antibiotic with gram-negative coverage. Other pertinent gram-negative endophthalmitis-causing organisms include Haemophilus influenza, although this is isolated less commonly in trauma-related endophthalmitis. Proteus mirabilis has been isolated in up to 7.7% of trauma-related cases.^17,19,55,56 Other infrequent organisms include Moraxella, other Pseudomonas species, Acinetobacter, and Neisseria species.^41,57–60

The high virulence of Ps. aeruginosa is felt to be multifactorial, including its high motility, its glycocalyx covering that protects against antibiotic penetration and phagocytosis, as well as toxin, protease, and β-lactamase production that results in host cell destruction and antibiotic resistance.^5,61–63 Ceftazidime is a third-generation cephalosporin that is favored for empiric treatment of endophthalmitis by intravitreal route due to its broad gram-negative coverage and anti-Pseudomonal properties. It is frequently given in conjunction with vancomycin for broad-spectrum, blanket coverage.

Previous animal studies have shown poor intravitreal penetration of ceftazidime when given intravenously even with high doses to achieve comparable plasma concentrations to peak plasma levels in humans after a single 1–2 g infusion (69–170 μg/mL).⁶⁴ Aguilar et al. found that ceftazidime levels were undetectable in the vitreous in rabbits after intravenous infusion even after redosing every 8 h for up to 70 h.¹¹ Schech et al. found vitreous concentrations of ceftazidime to be undetectable to low (1.2 μg/mL) when administered high doses in a swine model.⁴ Both studies showed increased penetration into the vitreous with ocular inflammation or postvitrectomy in the rabbit (35 μg/mL 2 h after a single dose), or with penetrating injury in the swine (7.5–8.4 μg/mL 1 h after 2 doses 12 h apart).^4,11 These concentrations are well above the MIC₉₀ for H. influenza (0.25–0.39 μg/mL), Pr. mirabilis (0.1 μg/mL), and Neisseria gonorrheae (0.78 μg/mL).^4,65 Ps. aeruginosa was once highly susceptible but increasing resistance to both ceftazidime and the flouroquinolones is developing.^66,67 The MIC₉₀ of ceftazidime for Ps. aeruginosa ranges from 4 to 25 μg/mL for susceptible strains, and up to 50 μg/mL in resistant strains.^15,65,68 Acinetobacter also has a high MIC₉₀ (25 μg/mL). The vitreous concentration of ceftazidime achieved in the previous animal studies is insufficient to treat an infection caused by these latter organisms.

In our study, the intravitreal concentration of ceftazidime reached 0.98% to 1.36% of plasma concentration, which is consistent with the previous studies that have shown poor intraocular penetration of this antibiotic.^4,11 Although low levels reach the vitreous cavity in control eyes (0.74±0.33–1.07±0.35 μg/mL), this reaches the MIC₉₀ for some gram-negative organisms, including Proteus and H. influenza. With penetrating injury, ceftazidime levels increase by 67% and 73% in the scleral and corneal groups, respectively, which would further increase its efficacy against the more-susceptible organisms, as well as provide adequate coverage for Neisseria species. Although this is a statistically significant increase, the levels are not sufficient to treat Pseudomonas or Acinetobacter infections.

Aguilar et al. found that ceftazidime penetrated the vitreous cavity in rabbits up to 35 μg/mL in their model of severe ocular inflammation in a vitrectomized eye.¹¹ Indeed, ocular penetration may be greater in cases of massive inflammation from fulminant endophthalmitis. However, such eyes usually have not previously been vitrectomized, so effective diffusion and high vitreous cavity concentrations may be limited by dense pyogenic and fibrin material. Schech et al. also found higher levels in the vitreous cavity after penetrating injury in their swine model (7.5–8.4 μg/mL), although this was after 2 intravenous doses using very high concentrations ranging from 130 to 175 mg/kg,⁴ much higher than the standard clinical dosage, 50 mg/kg, as used in our study.

Although vitreous concentrations of intravenously delivered ceftazidime are increased with penetrating ocular injury, if there is a concern for an ocular wound contaminated by gram-negative organisms, we cannot advocate treatment or prophylaxis with intravenous ceftazidime alone based on our study. The fastidious nature of gram-negative organisms and high percentage of cases involving Pseudomonas species necessitate ceftazidime delivered by intravitreal injection for higher concentrations to be achieved. This would provide better coverage for any gram-negative organism, especially for more-potent organisms, such as Ps. aeruginosa.

Summary

Intravitreal antibiotic penetration of systemic moxifloxacin, vancomycin, and ceftazidime with or without penetrating ocular injury varies. This is likely due to the innate properties of the drug itself, including its permeability across the blood-ocular barrier, and how this changes with penetrating ocular injury and resultant acute inflammation. For prevention or treatment of gram-positive TE, intravitreal vancomycin therapy is required due to its poor penetration when given systemically. Systemic ceftazidime can be used for many gram-negative bacteria, but intravitreal injection is recommended for better coverage, especially for more-potent organisms, such as Ps. aeruginosa, which is a dangerous and common gram-negative endophthalmitis-causing organism. Systemic moxifloxacin has excellent intraocular penetration and is not significantly altered with penetrating ocular injury. This broad-spectrum antibiotic can be considered in all cases of traumatic bacterial endophthalmitis, as long as its shortcomings are kept in mind. If the ocular flora has contaminated an ocular wound, there is increased resistance in patients with a history of topical moxifloxacin use, and it has poor efficacy against Ps. aeruginosa.

Footnotes

Author Disclosure Statement

No competing financial interests exist.

References

Chow

D.R.

, Garretson

B.R.

, Kuczynski

, et al. External versus internal approach to the removal of metallic intraocular foreign bodies. Retina., 20:364–369, 2000.

Ahmed

, Schimel

A.M.

, Pathengay

, Colyer

M.H.

, and Flynn

H.W.

Jr . Endophthalmitis following open-globe injuries. Eye., 26:212–217, 2012.

Thompson

J.T.

, Parver

L.M.

, Enger

C.L.

, Mieler

W.F.

, and Liggett

P.E.

Infectious endophthalmitis after penetrating injuries with retained intraocular foreign bodies. National Eye Trauma System. Ophthalmology., 100:1468–1474, 1993.

Schech

J.M.

, Alfaro

D.V.

3rd , Laughlin

R.M.

, Sanford

E.G.

, Briggs

, and Dalgetty

Intravenous gentamicin and ceftazidime in penetrating ocular trauma: a swine model. Retina., 17:28–32, 1997.

Irvine

W.D.

, Flynn

H.W.

Jr. , Miller

, and Pflugfelder

S.C.

Endophthalmitis caused by gram-negative organisms. Arch. Ophthalmol., 110:1450–1454, 1992.

Yeh

, Colyer

M.H.

, and Weichel

E.D.

Current trends in the management of intraocular foreign bodies. Curr. Opin. Ophthalmol., 19:225–233, 2008.

Results of the Endophthalmitis Vitrectomy Study. A randomized trial of immediate vitrectomy and of intravenous antibiotics for the treatment of postoperative bacterial endophthalmitis. Endophthalmitis Vitrectomy Study Group. Arch. Ophthalmol., 113:1479–1496, 1995.

Colyer

M.H.

, Weber

E.D.

, Weichel

E.D.

, et al. Delayed intraocular foreign body removal without endophthalmitis during Operations Iraqi Freedom and Enduring Freedom. Ophthalmology., 114:1439–1447, 2007.

Gao

, Pennesi

M.E.

, Qiao

, et al. Intravitreal moxifloxacin: retinal safety study with electroretinography and histopathology in animal models. Invest. Ophthalmol. Vis. Sci., 47:1606–1611, 2006.

10.

Tzepi

, Vergados

, Kanellakopoulou

, et al. Pharmacokinetics of intravenously administered moxifloxacin in eye compartments: an experimental study. Int. J. Antimicrob. Agents., 33:160–162, 2009.

11.

Aguilar

H.E.

, Meredith

T.A.

, Shaarawy

, Kincaid

, and Dick

Vitreous cavity penetration of ceftazidime after intravenous administration. Retina., 15:154–159, 1995.

12.

Ferencz

J.R.

, Assia

E.I.

, Diamantstein

, and Rubinstein

Vancomycin concentration in the vitreous after intravenous and intravitreal administration for postoperative endophthalmitis. Arch. Ophthalmol., 117:1023–1027, 1999.

13.

Meredith

T.A.

, Aguilar

H.E.

, Shaarawy

, Kincaid

, Dick

, and Niesman

M.R.

Vancomycin levels in the vitreous cavity after intravenous administration. Am. J. Ophthalmol., 119:774–778, 1995.

14.

Mochizuki

, Yamashita

, Torisaki

, Komatsu

, Tanahashi

, and Kawasaki

Intraocular kinetics of ceftazidime (Modacin). Ophthalmic Res., 24:150–154, 1992.

15.

Walstad

R.A.

, and Blika

Penetration of ceftazidime into the normal rabbit and human eye. Scand. J. Infect. Dis. Suppl., 44:63–67, 1985.

16.

Abel

Jr. , and Boyle

G.L.

Dissecting ocular tissue for intraocular drug studies. Invest. Ophthalmol., 15:216–219, 1976.

17.

Brinton

G.S.

, Topping

T.M.

, Hyndiuk

R.A.

, Aaberg

T.M.

, Reeser

F.H.

, and Abrams

G.W.

Posttraumatic endophthalmitis. Arch. Ophthalmol., 102:547–550, 1984.

18.

Duane

T.D.

Clinical Ophthalmology. Rev. ed. Hagerstown, MD: Medical Dept., Harper & Row; 1978.

19.

Essex

R.W.

, Yi

, Charles

P.G.

, and Allen

P.J.

Post-traumatic endophthalmitis. Ophthalmology., 111:2015–2022, 2004.

20.

Zhang

, Zhang

M.N.

, Jiang

C.H.

, Yao

, and Zhang

Endophthalmitis following open globe injury. Br. J. Ophthalmol., 94:111–114, 2010.

21.

Mieler

W.F.

, Ellis

M.K.

, Williams

D.F.

, and Han

D.P.

Retained intraocular foreign bodies and endophthalmitis. Ophthalmology., 97:1532–1538, 1990.

22.

Wani

V.B.

, Al-Ajmi

, Thalib

, et al. Vitrectomy for posterior segment intraocular foreign bodies: visual results and prognostic factors. Retina., 23:654–660, 2003.

23.

Jiang

C.H.

, and Zhang

M.N.

Traumatic endophthalmitis following penetrating ocular injuries with retained intraocular foreign bodies. Chin. J. Traumatol., 6:167–170, 2003.

24.

Jonas

J.B.

, and Budde

W.M.

Early versus late removal of retained intraocular foreign bodies. Retina., 19:193–197, 1999.

25.

Narang

, Gupta

, Dogra

M.R.

, Pandav

S.S.

, and Das

Role of prophylactic intravitreal antibiotics in open globe injuries. Indian J. Ophthalmol., 51:39–44, 2003.

26.

Barza

Antibacterial agents in the treatment of ocular infections. Infect. Dis. Clin. North Am., 3:533–551, 1989.

27.

Baum

, Peyman

G.A.

, and Barza

Intravitreal administration of antibiotic in the treatment of bacterial endophthalmitis. III. Consensus. Surv. Ophthalmol., 26:204–206, 1982.

28.

Wolfson

J.S.

, and Hooper

D.C.

Fluoroquinolone antimicrobial agents. Clin. Microbiol. Rev., 2:378–424, 1989.

29.

Hooper

D.C.

, and Wolfson

J.S.

Quinolone Antimicrobial Agents. 2nd ed. Washington, DC: American Society for Microbiology; 1993.

30.

Lalitha

, Srinivasan

, Manikandan

, et al. Relationship of in vitro susceptibility to moxifloxacin and in vivo clinical outcome in bacterial keratitis. Clin. Infect. Dis., 54:1381–1387, 2012.

31.

Lott

M.N.

, Fuller

J.J.

, Hancock

H.A.

, et al. Vitreal penetration of oral moxifloxacin in humans. Retina., 28:473–476, 2008.

32.

Hariprasad

S.M.

, Shah

G.K.

, Mieler

W.F.

, et al. Vitreous and aqueous penetration of orally administered moxifloxacin in humans. Arch. Ophthalmol., 124:178–182, 2006.

33.

Bronner

, Jehl

, Peter

J.D.

, et al. Moxifloxacin efficacy and vitreous penetration in a rabbit model of Staphylococcus aureus endophthalmitis and effect on gene expression of leucotoxins and virulence regulator factors. Antimicrob. Agents Chemother., 47:1621–1629, 2003.

34.

Goldstein

M.H.

, Kowalski

R.P.

, and Gordon

Y.J.

Emerging fluoroquinolone resistance in bacterial keratitis: a 5-year review. Ophthalmology., 106:1313–1318, 1999.

35.

Graves

, Henry

, O'Brien

T.P.

, Hwang

D.G.

, Van Buskirk

, and Trousdale

M.D.

In vitro susceptibilities of bacterial ocular isolates to fluoroquinolones. Cornea., 20:301–305, 2001.

36.

Hwang

D.G.

Fluoroquinolone resistance in ophthalmology and the potential role for newer ophthalmic fluoroquinolones. Surv. Ophthalmol., 49 Suppl 2:S79–S83, 2004.

37.

Park

S.H.

, Lim

J.A.

, Choi

J.S.

, Kim

K.A.

, and Joo

C.K.

The resistance patterns of normal ocular bacterial flora to 4 fluoroquinolone antibiotics. Cornea., 28:68–72, 2009.

38.

Yin

V.T.

, Weisbrod

D.J.

, Eng

K.T.

, et al. Antibiotic resistance of ocular surface flora with repeated use of a topical antibiotic after intravitreal injection. JAMA Ophthalmol., 131:456–461, 2013.

39.

Ray

K.J.

, Prajna

, Srinivasan

, et al. Fluoroquinolone treatment and susceptibility of isolates from bacterial keratitis. JAMA Ophthalmol., 131:310–313, 2013.

40.

Schimel

A.M.

, Miller

, and Flynn

H.W.

Jr . Endophthalmitis isolates and antibiotic susceptibilities: a 10-year review of culture-proven cases. Am. J. Ophthalmol., 156:50–52.e1, 2013.

41.

Bhagat

, Nagori

, and Zarbin

Post-traumatic infectious endophthalmitis. Surv. Ophthalmol., 56:214–251, 2011.

42.

Fisch

, Salvanet

, Prazuck

, et al. Epidemiology of infective endophthalmitis in France. The French Collaborative Study Group on Endophthalmitis. Lancet., 338:1373–1376, 1991.

43.

Han

D.P.

, Wisniewski

S.R.

, Wilson

L.A.

, et al. Spectrum and susceptibilities of microbiologic isolates in the Endophthalmitis Vitrectomy Study. Am. J. Ophthalmol., 122:1–17, 1996.

44.

Ness

, and Schneider

Endogenous endophthalmitis caused by methicillin-resistant Staphylococcus aureus (MRSA). Retina., 29:831–834, 2009.

45.

Somani

, Grinbaum

, and Slomovic

A.R.

Postoperative endophthalmitis: incidence, predisposing surgery, clinical course and outcome. Can. J. Ophthalmol., 32:303–310, 1997.

46.

Pryor

J.G.

, Apt

, and Leopold

I.H.

Intraocular penetration of vancomycin. Arch. Ophthalmol., 67:608–611, 1962.

47.

Smith

M.A.

, Sorenson

J.A.

, Lowy

F.D.

, Shakin

J.L.

, Harrison

, and Jakobiec

F.A.

Treatment of experimental methicillin-resistant Staphylococcus epidermidis endophthalmitis with intravitreal vancomycin. Ophthalmology., 93:1328–1335, 1986.

48.

Rolston

K.V.

, Nguyen

, and Messer

In vitro activity of LY264826, a new glycopeptide antibiotic, against gram-positive bacteria isolated from patients with cancer. Antimicrob. Agents Chemother., 34:2137–2141, 1990.

49.

Benz

M.S.

, Scott

I.U.

, Flynn

H.W.

Jr. , Unonius

, and Miller

Endophthalmitis isolates and antibiotic sensitivities: a 6-year review of culture-proven cases. Am. J. Ophthalmol., 137:38–42, 2004.

50.

Kunimoto

D.Y.

, Das

, Sharma

, et al. Microbiologic spectrum and susceptibility of isolates: part II. Posttraumatic endophthalmitis. Endophthalmitis Research Group. Am. J. Ophthalmol., 128:242–244, 1999.

51.

Tran

T.P.

, Le

T.M.

, Bui

H.T.

, Nguyen

T.M.

, Kuchle

, and Nguyen

N.X.

[Post-traumatic endophthalmitis after penetrating injury in Vietnam: risk factors, microbiological aspect and visual outcome]. Klin. Monatsbl. Augenheilkd., 220:481–485, 2003.

52.

Affeldt

J.C.

, Flynn

H.W.

Jr. , Forster

R.K.

, Mandelbaum

, Clarkson

J.G.

, and Jarus

G.D.

Microbial endophthalmitis resulting from ocular trauma. Ophthalmology., 94:407–413, 1987.

53.

Alfaro

D.V.

, Roth

, and Liggett

P.E.

Posttraumatic endophthalmitis. Causative organisms, treatment, and prevention. Retina., 14:206–211, 1994.

54.

Colyer

M.H.

, Chun

D.W.

, Bower

K.S.

, Dick

J.S.

, and Weichel

E.D.

Perforating globe injuries during operation Iraqi Freedom. Ophthalmology., 115:2087–2093, 2008.

55.

Rowsey

J.J.

, Newsom

D.L.

, Sexton

D.J.

, and Harms

W.K.

Endophthalmitis: current approaches. Ophthalmology., 89:1055–1066, 1982.

56.

Vedantham

, Nirmalan

P.K.

, Ramasamy

, Prakash

, and Namperumalsamy

Clinico-microbiological profile and visual outcomes of post-traumatic endophthalmitis at a tertiary eye care center in South India. Indian J. Ophthalmol., 54:5–10, 2006.

57.

Berrocal

A.M.

, Scott

I.U.

, Miller

, and Flynn

H.W.

Jr . Endophthalmitis caused by Moraxella species. Am. J. Ophthalmol., 132:788–790, 2001.

58.

Essex

R.W.

, Charles

P.G.

, and Allen

P.J.

Three cases of post-traumatic endophthalmitis caused by unusual bacteria. Clin. Exp. Ophthalmol., 32:445–447, 2004.

59.

Sabaci

, Bayer

, Mutlu

F.M.

, Karagul

, and Yildirim

Endophthalmitis after deadly-weapon-related open-globe injuries: risk factors, value of prophylactic antibiotics, and visual outcomes. Am. J. Ophthalmol., 133:62–69, 2002.

60.

Sharma

, Saffra

N.A.

, and Chapnick

E.K.

Post traumatic polymicrobial endophthalmitis, including Neisseria subflava. Am. J. Ophthalmol., 136:554–555, 2003.

61.

Schaechter

, and Lederberg

The Desk Encyclopedia of Microbiology. 1st ed. Amsterdam; Boston: Elsevier/Academic Press; 2004.

62.

Stratton

C.W.

, and Tausk

Intrinsic resistance of Pseudomonas aeruginosa. Antibiot. Chemother., 42:275–286, 1989.

63.

Zinsser

, and Joklik

W.K.

Zinsser Microbiology. 19th ed. Norwalk, CT: Appleton & Lange; 1988.

64.

Kucers

, and Bennett

N.M.

The Use of Antibiotics: A Comprehensive Review with Clinical Emphasis. 3d ed. Philadelphia: Lippincott; 1979.

65.

Watanabe

, Katsu

, Moriyama

, and Kitoh

In vitro evaluation of E1040, a new cephalosporin with potent antipseudomonal activity. Antimicrob. Agents Chemother., 32:693–701, 1988.

66.

Breidenstein

E.B.

, de la Fuente-Nunez

, and Hancock

R.E.

Pseudomonas aeruginosa: all roads lead to resistance. Trends Microbiol., 19:419–426, 2011.

67.

Cunha

B.A.

Pseudomonas aeruginosa: resistance and therapy. Semin. Respir. Infect., 17:231–239, 2002.

68.

Sader

H.S.

, Gales

A.C.

, Pfaller

M.A.

, et al. Pathogen frequency and resistance patterns in Brazilian hospitals: summary of results from three years of the SENTRY Antimicrobial Surveillance Program. Braz. J. Infecti. Dis., 5:200–214, 2001.