Treatment of Experimental Bacterial Keratitis With Topical Trovafloxacin FREE

Microbial keratitis continues to be a common, potentially sight-threateningocular infection.^{1 - 2} Fluoroquinolonesare increasingly selected as initial, broad-spectrum agents for the topicaltherapy of bacterial keratitis.³ The use ofofloxacin and ciprofloxacin, as single agents, was shown in clinical trialsto be comparable to a combination of fortified antibiotics in the treatmentof acute bacterial keratitis.^{4 - 5} However,these 2 quinolones, which currently are often used, have limited in vitroactivity against some gram-positive organisms and various strains of Pseudomonas aeruginosa.^{6 - 7}

Trovafloxacin mesylate, 7-(3-azabicyclo[3,1,0]hexyl)-naphthyridone,is a newer-generation synthetic fluoroquinolone currently available for oraland intravenous administration. Trovafloxacin has broad spectrum antibacterialactivity against gram-negative, gram-positive, and anaerobic bacteria. Invitro and in vivo studies^{8 - 13} insystemic diseases have demonstrated a greater activity against clinicallyimportant gram-positive organisms (most notably streptococci), while maintainingactivity against gram-negative organisms, when compared with ciprofloxacinand ofloxacin.

The purpose of this study is to determine the microbiological efficacyof topical trovafloxacin for potential use in therapy of experimental bacterialkeratitis. After performing susceptibility and pharmacokinetic studies, westudied the microbiological effect of topical trovafloxacin on bacterial keratitiscaused by either gram-positive organisms, such as Staphylococcusaureus and Streptococcus pneumoniae, or agram-negative organism (P aeruginosa) and comparedits efficacy with that of ciprofloxacin and ofloxacin.

Bacterial strains used in this study were clinical ocular isolates collectedfrom corneal specimens at the Wilmer Ophthalmological Institute, Baltimore,Md.

Minimum inhibitory concentrations (MICs) of trovafloxacin, ciprofloxacin,and ofloxacin were determined for various ocular isolates of S aureus, S pneumoniae, Staphylococcus epidermidis, P aeruginosa,and Haemophilus influenzae. The E-test method wasused for determining the MIC for each strain. The bacterial suspension wasprepared by collecting the clinical isolates from a blood agar plate. Theisolate sample was adjusted with 0.9% nonbacteriostatic isotonic sodium chloridesolution to achieve the same density as a 0.5 McFarland standard (1 ×10⁸ colony-forming units/mL). The appropriate E-test strip wasplaced on a Mueller-Hinton II agar plate (BBL, Cockeysville, Md) inoculatedwith the suspension of bacteria. The plates were incubated for 16 to 24 hoursat 37°C, and the MIC was read from the scale on the side of the stripat the point where the ellipse of growth inhibition intercepted the strip.All the tests were performed in duplicate.

The pharmacokinetic studies were performed on rabbit eyes. Twelve rabbitswere used for this study. Epithelial removal (to promote antibiotic entryand to simulate human ulcerative keratitis) was evaluated. In each rabbit,one eye underwent removal of the central (7.5 mm) corneal epithelium, andin the fellow eye the epithelium remained intact. A single drop of trovafloxacinwas applied to both eyes. Rabbits were humanely killed at 15, 30, 60, 120,and 240 minutes (at each time point, 3 rabbits were euthanized), and sampleswere obtained immediately thereafter from aqueous humor, vitreous, and thecentral cornea (7.5 mm). The tissue concentration of trovafloxacin was determinedusing a high-performance liquid chromatography assay.¹⁴

In these studies, 72 (24 in each experiment) New Zealand white rabbits,weighing 2.0 to 2.25 kg, were used in accordance with the guidelines for animalexperimentation established by the Association of Research in Vision and Ophthalmology(Rockville, Md); approval from the appropriate institutional review boardwas obtained for the study, and the institutional guidelines regarding animalexperimentation were followed. All rabbits were sedated and anesthetized withintramuscular injection of ketamine hydrochloride (60 mg/kg) and xylazinehydrochloride (12 mg/kg) and topical 0.5% proparacaine hydrochloride beforehaving only one eye of each animal injected with bacteria.

To produce keratitis, 10³ colony-forming units (0.1 mL) ofclinical bacterial isolates in logarithmic growth phase were injected intothe corneal stroma in one eye. The injection was performed with a 30-gaugeneedle into the stroma in the central part of the cornea under microscopicguidance. The first experiment was performed with 24 rabbits, using a clinicalisolate of a methicillin-sensitive S aureus witha predetermined MIC50 (inhibits 50% of strains) of 0.5 µg/mL for ciprofloxacin.The other 2 experiments were performed in an identical fashion by inoculatingeither S pneumoniae or P aeruginosa.

Twelve hours after the injection, the rabbits were randomized to 4 treatmentgroups, with 6 rabbits in each group: (1) ciprofloxacin, 3 mg/mL (Ciloxan0.3%; Alcon Laboratories, Fort Worth, Tex); (2) ofloxacin, 3 mg/mL (Ocuflox0.3%; Allergan, Irvine, Calif); (3) trovafloxacin mesylate, 5 mg/mL (Trovan-IV;Pfizer, New York, NY); and (4) preservative-free 0.9% nonbacteriostatic isotonicsodium chloride solution for intravenous use (control). Trovafloxacin eyedropswere prepared using the intravenous solution that was transferred into a dropperbottle under a sterile hood.

The eyedrops were administered every hour for 12 hours. One hour afterthe last dose of eyedrops, the rabbits were systemically anesthetized andthen humanely killed by intracardiac injection of pentobarbital sodium (Beuthanasia-DSpecial; Schering-Plough Animal Health Corp, Kenilworth, NJ). The rabbits'corneas were excised using a sterile, disposable, 7.5-mm corneal trephineand irrigated with 3 mL of phosphate-buffered saline to eliminate any residualantibiotic or debris on the surface of the corneas. The corneal buttons wereimmediately homogenized and then serially diluted before plating in duplicateon blood agar for S aureus and S pneumoniae and on Mueller-Hinton II agar for Paeruginosa. The specimens were incubated at 35°C for 24 hours beforequantitative bacteriologic analysis.

The MICs for the various ocular strains are given in Table 1. The MIC for P aeruginosa strainswas significantly lower for ciprofloxacin compared with trovafloxacin (P = .008) and ofloxacin (P = .007)and significantly lower for trovafloxacin compared with ofloxacin (P = .007).

The MIC for S aureus strains was significantlylower for trovafloxacin compared with ciprofloxacin (P<.001)and ofloxacin (P<.001), and no significant differencewas found between the MIC for ciprofloxacin and ofloxacin (P = .10). The MIC for H influenzae strainswas significantly lower for trovafloxacin (P = .005)and ciprofloxacin (P = .005) compared with ofloxacin.No significant difference was found between trovafloxacin and ciprofloxacin(P = .44). The MIC for S epidermidis was lower for trovafloxacin than ciprofloxacin; however, this wasnot statistically conclusive (P = .07). The MIC for S pneumonia strains was significantly lower for trovafloxacin(P = .006) and ciprofloxacin (P = .006) compared with ofloxacin.

After the single trovafloxacin drop application, a substantial penetrationwas found in the corneas at all time points (Table 2). Mean (SD) trovafloxacin levels were higher in eyes withdenuded epithelium (25.9 [22.1] µg/g) compared with intact epithelium(3.7 [2.2] µg/g). The aqueous levels were higher in eyes with denudedepithelium (2.5 [1.4] µg/mL) compared with intact epithelium (0.3 [0.1]µg/mL). The vitreous concentrations after a single topical applicationwere low in most of the eyes. All serum samples for all the time points hadundetectable trovafloxacin concentrations.

The bacterial counts for the 3 studies are given in Table 3. The S aureus keratitis study showeda statistically significant decrease in bacterial counts in the groups treatedwith trovafloxacin compared with the ciprofloxacin and ofloxacin treatmentgroups (P = .002). The S pneumoniae study showed a significant decrease in bacterial counts in the antibiotic-treatedgroups compared with the controls (P = .002) anda significant decrease in bacterial counts in the trovafloxacin groups comparedwith ciprofloxacin but not ofloxacin. The P aeruginosa studyshowed a significant decrease in all antibiotic-treated groups compared withthe control group (P<.001).

Until recently, the fluoroquinolones were primarily agents with excellentactivity against gram-negative bacilli but with questionable activity againstsome important gram-positive pathogens. Newer agents under development havean expanded profile of antibacterial activity against gram-positive pathogens,including staphylococci, streptococci, enterococci, and anaerobic isolates.The fluoroquinolone evolutionary tree has 2 main branches. The first branchis the naphthyridine series, which have a nitrogen atom at the C-8 position(eg, nalidixic acid, enoxacin, tosufloxacin, trovafloxacin); all others arefluoroquinolone agents, with the C-7 position being the most adaptable sitefor chemical substitution.

A major clinical challenge has been the worldwide emergence and spreadof antimicrobial resistance, especially among gram-positive isolates, suchas pneumococci. There is recent information on the mechanism of action offluoroquinolone and how resistance develops. From this, it seems that themost efficient use of these fluoroquinolone compounds is to use a highly activeagent with a favorable pharmacodynamic profile for the shortest treatmentduration consistent with cure.

Trovafloxacin is a newer-generation, expanded-spectrum fluoroquinolone,which was shown in our study to have in vitro potency, favorable pharmacokinetics,and in vivo efficacy, suggesting a potential role in the therapy of bacterialkeratitis. The use of topical fluorinated quinolones has become a frequent,initial, single-agent therapy for bacterial keratitis.^{4 - 5} Themechanism of action of the quinolones is via inhibition of the type II topoisomeraseDNA gyrase, an essential bacterial enzyme that alters the topology of double-strandedDNA within the cell. Inhibition of this activity by fluoroquinolones is associatedwith rapid killing of the bacterial cell.¹⁵ Theimportant advantages in the systemic use of trovafloxacin are the pharmacokineticproperties and its expanded antibacterial spectrum, especially against gram-positiveorganisms, including Streptococcus species. Our resultsshowed that trovafloxacin not only is effective against P aeruginosa, similar to other experimental keratitis studies,^{16 - 17} but also is superior as far as activityagainst gram-positive bacteria.^{18 - 22}

The current clinical use of trovafloxacin so far has been systemic,intravenous, intramuscular, or oral. The intravenous formulation of trovafloxacinis the prodrug alatrofloxacin mesylate, which has little intrinsic antibacterialactivity in vitro but demonstrates good activity in vivo as a result of undergoingrapid hydrolysis and is rapidly converted to trovafloxacin after intravenousadministration. We used the intravenous formulation for preparation of theeyedrops, and this demonstrated a good antibacterial activity in topical usefor the keratitis model.

In vitro activity of trovafloxacin has been shown to be superior toboth ciprofloxacin and ofloxacin against both the penicillin-susceptible andpenicillin-resistant pneumococci.^{23 - 24} Inour study the in vitro antibacterial activity of trovafloxacin against ocularisolates has been demonstrated by using the E-test, a modification of a discdiffusion susceptibility testing.²⁵

Pascual et al²⁶ showed that trovafloxacinpenetrates into phagocytic and nonphagocytic cells, reaching intracellularconcentrations several times greater than the extracellular ones, whereasit remained active intracellularly in human polymorphonuclear leukocytes;this fact may enhance the in vivo antibacterial activity. Lately, the systemicuse of trovafloxacin was restricted to life- or limb-threatening conditionsdue to several reports of hepatic toxicity after systemic use of the drug.^{27 - 28} With topical ocular use, the expectedsystemic levels are extremely low, if detectable at all; moreover, the restrictedsystemic use may offer an advantage with the topical use due to a lesser chanceof emergence of resistance. In conclusion, the high in vitro efficacy againstocular isolates along with the high ocular tissue penetration and nondetectablesystemic levels and in vivo efficacy indicate that trovafloxacin may havean important clinical role in the topical treatment of bacterial keratitis,including streptococcal species.

Corresponding author and reprints: Terrence P. O'Brien, MD, OcularMicrobiology Laboratory, The Wilmer Ophthalmological Institute, Woods 255,600 N Wolfe St, Baltimore, MD 21287-9121 (e-mail: tobrien@jhmi.edu).

Submitted for publication August 1, 2003; final revision received August12, 2003; accepted August 21, 2003.