Ciprofloxacin Versus Tobramycin for the Treatment of Staphylococcal Keratitis

Ciprofloxacin Versus Tobramycin for the Treatment of Staphylococcal Keratitis Michelle C. Callegan* Lee S. Engel,* James M. Hill,*"f and Richard J. O'Callaghan* Purpose. To compare the chemotherapeutic efficacies of ciprofloxacin (0.3%) and fortified (1.36%) tobramycin for the treatment of methicillin-sensitive and methicillin-resistant Staphylococcus aureus keratitis during early and late stages of infection. Methods. Rabbit corneas were intrastromally injected with 10 2 colony-forming units (CFU) of methicillin-sensitive (MSSA) or methicillin-resistant (MRSA). Topical therapy was initiated at either 4 hours postinfection (early stage) or at 10 hours postinfection (late stage). Drops were administered every 15 minutes for 5 hours. Corneal bacterial counts and aqueous humor antibiotic concentrations were determined. Results. Early administration of ciprofloxacin sterilized all MSSA-infected corneas and 83% of MRSA-infected corneas. Late administration of ciprofloxacin reduced the numbers of viable MSSA and MRSA to 3.6 and 3.7 log 10 CFU per cornea, respectively, but did not sterilize any corneas. Early administration of fortified (1.36%) tobramycin sterilized all MSSA-infected corneas but none of the MRSA-infected corneas. Late administration of tobramycin reduced the viable MSSA to very low numbers (0.5 and 0.0 log 10, respectively) and sterilized 33% of MSSA-infected corneas, but had little effect on MRSA-infected corneas. Conclusions. Early in infection, ciprofloxacin was highly effective against MSSA and MRSA, whereas tobramycin was effective only against MSSA. During later stages of infection, tobramycin was more effective than ciprofloxacin against MSSA, and neither antibiotic was effective against MRSA. Thus, ciprofloxacin is limited by the time of application and tobramycin is limited by the resistance of the MRSA strain. Invest Ophthalmol Vis Sci. 1994; 35:1033-1037. istaphylococcus aureus is among the principal pathogens associated with bacterial keratitis, especially as a consequence of extended-wear contact lens use. 1 ' 2 Because of the potential for corneal scarring, antimicrobial treatment (topical application of commercially available cefazolin and an aminoglycoside 3 ' 4 ) must be initiated promptly to prevent loss of vision. In severe cases, the need for frequent administration of fortified tobramycin (13.6 mg/ml) usually requires hospitalization to assure compliance. 3 ' 4 From the * Department of Microbiology, Immunology, and Parasitology, and the \Department of Ophthalmology, Lions Eye Research Laboratories, Louisiana State University Medical Center School of Medicine, New Orleans, Louisiana. Supported in part by U.S. Public Health Service grant EY0887J and Core grant EY02377 from the National Eye Institute, National Institutes of Health, Bethesda, Maryland. Submitted for publication June 22, 1993; revised October 4, 1993; accepted October 5, 1993. Proprietary interest category: N. Reprint requests: James M. Hill, LSU Eye Center, 2020 Gravier Street, Suite B, New Orleans, LA 70112-2234. Edwards and Schlech reported the efficacy of the aminoglycoside, tobramycin, for treatment of experimental keratitis in rabbits. 5 Topical tobramycin drops (0.1% and 0.3%) reduced viable staphylococci and slowed the progression of the infection, as observed by slit-lamp examination. In a clinical study, both 0.3% tobramycin and a fluoroquinolone, 0.3% ofloxacin, were effective against 5. aureus keratitis, resulting in clinical improvement in 85% of the patients, as measured by bacterial culturing techniques. 6 Few quantitative studies of fluoroquinolones in experimental staphylococcal animal models have been published. We 7 ' 8 have shown that, in the therapy of experimental keratitis in the rabbit, topical 0.3% ciprofloxacin is more effective than topical 5.0% cefazolin or 5.0% vancomycin. These studies also demonstrated that with early treatment (4 to 9 hours postinfection), when bacteria are replicating, ciprofloxacin is capable of sterilizing infected corneas. With later treatment (10 to 15 or 10 to 20 hours postinfection), however, when bacterial replication is minimal, nei- Investigalive Ophthalmology & Visual Science, March 1994, Vol. 35, No. 3 Copyright Association for Research in Vision and Ophthalmology 1033

1034 Investigative Ophthalmology & Visual Science, March 1994, Vol. 35, No. 3 ther ciprofloxacin, cefazolin, nor vancomycin effectively sterilized staphylococcus-infected corneas. In the search for a more effective therapeutic regimen against ocular staphylococcal infections, especially for the later stages of infection, we compared the efficacy of early and late treatment with a fluoroquinolone, 0.3% ciprofloxacin, and an aminoglycoside, fortified (1.36%) tobramycin. These studies used both methicillin-sensitive (MSSA) and methicillin-resistant (MRSA) 5. aureus in an intrastromal model of keratitis in the rabbit. MATERIALS AND METHODS Bacterial Strains The MSSA strain used in these studies (ATCC 25923; American Type Culture Collection, Rockville, MD) is a well characterized strain used for antibiotic susceptibility testing, and is susceptible to a wide variety of antibiotics, including tobramycin and ciprofloxacin. Minimal inhibitory concentrations (MIC) for tobramycin and ciprofloxacin, as determined by a broth tube dilution technique, were 0.4 Mg/ m l and 0.2 /ug/ml, respectively. The MRSA strain used in these studies (strain 301) was acquired from a human corneal ulcer 8 ; MICs for tobramycin and ciprofloxacin were 250 Mg/ m l an d 0.5 /iig/ml, respectively. Keratitis New Zealand white rabbits weighing 2.0 to 3.0 kg were used according to guidelines set forth by the ARVO Resolution on the Use of Animals in Ophthalmic and Vision Research. All rabbits were anesthetized by subcutaneous injection of a 1:5 mixture of 100 mg/ml xylazine (Rompun; Miles Laboratories, Shawnee, KS) and 100 mg/ml ketamine hydrochloride (Ketaset; Bristol Laboratories, Syracuse, NY). One drop of proparacaine hydrochloride (0.5%, Alcainer; Alcon Laboratories, Fort Worth, TX) was instilled into each eye before intrastromal injection. Each eye was intrastromally injected with approximately 100 colony-forming units (CFU) of MSSA or MRSA in 10 ^1 tryptic soy broth (Difco, Detroit, MI) via a 30-gauge needle attached to a 100-/il syringe (Hamilton Co., Reno, NV). 78 Preparation of Solutions Commercial ciprofloxacin hydrochloride (0.3%; Ciloxan; Alcon) was used in these studies. Fortified (1.36%) tobramycin was prepared by addition of 1.0 ml of injectable tobramycin (4.0%; Nebcin; Eli Lilly & Co., Indianapolis, IN) to a 2.5-ml bottle of 0.3% tobramycin (Tobrex; Alcon). Sterile saline was used as the antibiotic-negative control. Experimental Design Treatment was initiated at 4 or 10 hours postinfection. Experiment I involved infection with MSSA or MRSA and treatment from 4 to 9 hours postinfection. Experiment II involved infection with MSSA or MRSA and treatment from 10 to 15 hours postinfection. Experiments I and II each consisted often rabbits randomly assigned to four treatment groups as follows: group 1, topical fortified (1.36%) tobramycin (six corneas); group 2, topical 0.3% ciprofloxacin (six corneas); group 3, topical saline (four corneas); and group 4, no topical therapy (untreated, four corneas). Treated rabbits were given a single drop of solution in each eye every 15 minutes for 5 hours, for a total of 21 drops per eye. Corneal manifestations of progressing ocular infection were observed using a slit-lamp biomicroscope (Topcon SL-5D; Kogaku Kikai K.K., Tokyo, Japan). All rabbits were sacrificed 1 hour after the last topical treatment (10 or 16 hours postinfection) by intravenous injection of sodium pentobarbital solution (100 mg/ml; The Butler Co., Columbus, MO). All eyes were proptosed, irrigated with sterile phosphatebuffered saline (PBS, ph 7.4), and corneas and aqueous humor were recovered for determination of viable bacteria (CFU) and antibiotic concentrations, respectively. Aqueous humor (200 /x\) was removed by paracentesis with a 27-gauge needle attached to a 1.0- ml tuberculin syringe. Corneas were removed aseptically by excision at the corneoscleral limbus and dissected into eight to ten small pieces. Recovered corneas ranged from 12 to 16 mm in diameter. Quantification of Viable Per Cornea Quantification of viable MSSA or MRSA per cornea has been previously described. 7 ' 8 Briefly, corneas were aseptically removed, dissected, homogenized, and centrifuged. Aliquots of corneal homogenate were serially diluted in sterile PBS, plated onto tryptic soy agar (Difco) and incubated at 37 C for 48 hours. All colonies were counted, and viable CFU expressed as base 10 logarithms. Statistical analysis of variance (ANOVA) of viable CFU per cornea was performed using a Statistical Analysis Systems (SAS) program. 9 Where a significant ANOVA was found, t tests between the least square means of each group were performed. s in Aqueous Humor Agar-well bioassays for tobramycin and ciprofloxacin developed by Engel et al 10 were used. Briefly, assay plates were prepared by pouring 30 ml No. 2 Oxoid Agar (Oxoid USA Inc., Columbia, MD) into 150 X 15 mm Petri dishes (Curtin Matheson Scientific, Inc., Houston, TX). The agar was allowed to harden

Chemotherapy of Staphylococcal Keratitis 1035 and 10 5 CFU per ml of the indicator bacterial strain was aseptically spread onto the agar surface. Wells were cut into the agar with a sterile trephine (7.5 mm diameter). Aqueous humor samples and antibiotic standards (50 ml) were transferred aseptically into the wells. 5. aureus ATCC strain 25923 and Klebsiellapneumoniae ATCC strain 10031 were used as the indicator strains in tobramycin and ciprofloxacin bioassays, respectively. The Petri dishes were incubated in a humidified chamber at 37 C for 24 hours, and zones of inhibition were measured. All aqueous humor samples and antibiotic standards were assayed in triplicate. concentrations were determined from a standard curve of zone size versus log of concentration, the slope of which was determined from a best-fit curve by least-square means method using an SAS program. 9 Standard curves had a linear regression coefficient of > 0.93. The sensitivities of the tobramycin and ciprofloxacin bioassays were 1.0 mg/ml and 0.1 ml, respectively. 10 RESULTS During the early stages of infection (4 to 9 hours postinfection), corneal changes in staphylococcus-infected eyes included slight stromal edema and minimal accumulation of stromal infiltrate at the site of injection. During the later stages of infection (10 to 15 hours postinfection), marked erosions of the corneal epithelium at the injection site were visible in all infected eyes. With time, the epithelial erosions increased in size (greater than 5 mm) and began to fill with pus. Also during the later stages of infection, punctate stromal infiltrates that were present along the needle tract began to coalesce. Topical application of tobramycin or ciprofloxacin had no effect on the progression of changes in corneal appearance. Early treatment with ciprofloxacin (4 to 9 hours postinfection) sterilized all MSSA-infected corneas and five of six (83.3%) MRSA-infected corneas; the numbers of CFU in these corneas were small (Table 1). Late ciprofloxacin therapy (10 to 15 hours postinfection) resulted in larger populations of residual bacteria for both strains (3.6 and 3.7 log ]0 CFU for MSSA and MRSA, respectively; Table 2). Early treatment with fortified (1.36%) tobramycin sterilized all MSSA-infected corneas; however, the mean number of viable bacteria remaining in MRSAinfected corneas was 4.3 log 10 CFU, and none of these corneas was sterile (Table 1). Late tobramycin treatment of MSSA-infected corneas resulted in 0.5 log 10 CFU and sterilization of two of six (33.3%) corneas, whereas corneas infected with the resistant strain contained 7.0 logjo CFU, and none was sterile (Table 2). There were no significant differences in the numbers of viable bacteria remaining in saline-treated and untreated control corneas infected with either MSSA or MRSA at the end of the early treatment period (10 hours postinfection) or the late treatment period (16 hours postinfection). The concentrations of ciprofloxacin in aqueous humor from MSSA- and MRSA-infected eyes were not significantly different, whether sampled after early or late treatment (Tables 1, 2). The concentration of tobramycin in aqueous humor from MRSA-infected eyes, however, was nearly twice that of the aqueous humor from MSSA-infected eyes after both early and late treatments (Tables 1,2). DISCUSSION In this study, we compared the efficacy of 0.3% ciprofloxacin and fortified 1.36% tobramycin against earlystage and late-stage MSSA and MRSA keratitis in rabbits. Ciprofloxacin was capable of corneal sterilization TABLE l. Viable Staphylococcus aureus in Corneas after Early Treatment* Infection WithMethicillin-Sensitive Infection With Methicillin-Resistant Group Treatment Units (Log, 0 ) (fig/ml) Units (Log lo ) (Hg/ml) 1 2 3 4 Tobramycin (1.36%) Ciprofloxacin (0.3%) Saline None 0.0 ±0.0 0.0 ±0.0 5.7 ±0.1 5.8 ± 0.1 48.0 ± 6.Of 5.5±0.5 4.3 ± 0.2J 0.1 ±0.1J 5.4 ± 0.0 5.8 ±0.1 73.7 ± 5.8f 6.5 ± 1.6 Values are mean ± SEM for six corneas in Groups 1 and 2 and four corneas in Groups 3 and 4. * Early treatment was begun 4 hr after infection and continued for 5 hr. Bacterial counts and antibiotic concentrations were measured 1 hr after treatment was completed (10-hr postinfection). f Significantly different (P = 0.0005). X Significantly different from each other (P = 0.0001) and from all other groups (P < 0.0001). Not significantly different (P = 0.9).

1036 Investigative Ophthalmology & Visual Science, March 1994, Vol. 35, No. 3 TABLE 2. Viable Staphylococcus aureus in Corneas after Late Treatment* Infection With Methicillin-Sensitive Infection With Methicillin-Resistant Group Treatment Units (Logi 0 ) (ng/ml) Units (Log l0 ) (lig/ml) 1 2 3 4 Tobramycin (1.36%) Ciprofloxacin (0.3%) Saline None 0.5 ± 0.2f 3.6 ± O.lf 6.2 ±0.1 6.2 ±0.1 48.1 ± 12.5} 4.4 ± 1.111 7.0 ± 0.1 3.7±0.1 7.4 ±0.1 7.5 ±0.1 88.2 ± 8.9J 15.8 ±2.311 Values are mean ± SEM for six corneas in Groups 1 and 2 and four corneas in Groups 3 and 4. * Late treatment was begun 10 hr after infection and continued for 5 hr. Bacterial counts and antibiotic concentrations were measured 1 hr after treatment was completed (16-hr postinfection). t Significantly different (P = 0.0001). X Significantly different (P = 0.0001). Significantly different (P = 0.0001). 11 Significantly different (P = 0.0001). when treatment for MSSA and MRSA keratitis was begun early (4 to 9 hours postinfection), corroborating our study demonstrating the efficacy of 0.3% ciprofloxacin versus 5.0% cefazolin and 5.0% vancomycin against early-stage staphylococcal keratitis. 7 ' 8 We 78 and others 11 have reported the inefficiency of topical antibiotic drops against staphylococcal keratitis, particularly during later stages of infection. Davis et al 1 ' demonstrated the diminished ability of nafcillin to kill staphylococci in the cornea when treatment was begun at 12 or 16 hours postinfection. As shown in the current study and in previous studies, 7 ' 8 ciprofloxacin was unable sufficiently to reduce staphylococci in the cornea when treatment was initiated during the later stages of infection. Extending therapy from 5 to 10 hours' duration (10 to 20 hours postinfection) had little effect in improving the outcome of the infection. 78 The inability of these antibiotics to kill staphylococci during the late stages of infection could be related to the reduced amount of bacterial replication, the inability of antibiotic to reach staphylococci sequestered in a pus-filled ulcer, or the inhibition of antibiotic activity by pus. Aminoglycosides have been used commonly for treatment of severe staphylococcal keratitis, 4 and have been effective against experimental 5. aureus keratitis during the early stages of infection. 12 Previous studies have shown the ability of gentamicin and neomycin to reduce staphylococci in the cornea to less than 2 logs when treatment was begun during the later stages of infection. 1314 Our study demonstrated that fortified tobramycin is capable of sterilizing 100% of MSSA-infected corneas when therapy is begun early. Corneal sterilization (33%) was also achieved for MSSA keratitis when tobramycin therapy was begun during the late stages of keratitis, a time when other antibiotics have been therapeutically ineffective. 7 ' 8 Fortified 1.36% tobramycin was not effective against MRSA in the cornea, even when treatment was begun at 4 hours postinfection. This result was anticipated, because the MRSA strain used in these studies was determined previously to be resistant to tobramycin; the tobramycin concentrations in aqueous humor did not approach the tobramycin MIC (250 Mg/ml) for this MRSA strain. Tobramycin concentrations achieved in the aqueous humor after 5 hours of treatment of MRSA keratitis were significantly higher than tobramycin concentrations achieved during treatment of MSSA keratitis. Corneal structural alterations, such as stromal edema and infiltration, began during the early stages of keratitis, at approximately 8 hours postinfection. Corneal changes progressed more rapidly, however, in MRSA-infected eyes than in MSSA-infected eyes. Epithelial erosions were present in MRSA-infected eyes at 10 hours postinfection, but were observed in MSSA-infected eyes at 12 hours postinfection. These observations agree with previous studies in which slit-lamp observations revealed more rapidly enlarging epithelial erosions in MRSA-infected eyes, compared to those observed in MSSA-infected eyes. 7 ' 8 Corneal epithelial damage by mechanical debridement has been shown to result in higher concentrations of aminoglycosides in the cornea and aqueous humor. 1415 Overall, antibiotic concentrations in aqueous humor from eyes infected with MRSA were higher than concentrations from MSSA-infected corneas. Because of the early stromal damage and epithelial ulceration seen in progressing MRSA keratitis, more antibiotic probably entered into and diffused through the cornea in these eyes. We have demonstrated that both 0.3% ciprofloxacin and fortified 1.36% tobramycin are highly effective in killing MSSA growing in the cornea during early

Chemotherapy of Staphylococcal Keratitis 1037 stages of infection. Fortified tobramycin also was highly effective against progressing MSSA keratitis during later stages of infection, when other antibiotics have been shown to be therapeutically inadequate. Because MRSA strains also commonly are resistant to aminoglycosides, as seen here with tobramycin, ciprofloxacin should be considered for the treatment of MRSA corneal infections. With a recent increase in quinolone-resistant strains of MSSA and MRSA, 1617 however, evaluation of the efficacies of other antibiotics or combinations for treatment of these infections is needed. Additional studies also are needed to determine the best chemotherapeutic regimen for the treatment of antibiotic-sensitive and antibiotic-resistant staphylococcal keratitis, particularly one that could achieve corneal sterilization during the later stages of infection. Key Words tobramycin, ciprofloxacin, methicillin-resistant, Staphylococcus aureus keratitis, rabbits Acknowledgments The authors thank consultants Drs. Thomas E. Clinch, Michael S. Insler, Herbert E. Kaufman, and Jeffery A. Hobden; and Wayne R. Salsiccia, Bryan P. Hemard, Patrick M. O'Callaghan, and Nataskia O. Lampe for their technical assistance. References 1. Laibson PR, Cohen EJ, Rajpal RK. Corneal ulcers related to contact lenses. CLAOJ. 1993; 19:73-78. 2. Limberg MB. A review of bacterial keratitis and bacterial conjunctivitis. Am J Ophthalmol. 1991; 112 (suppl):2s-9s. 3. Barza M. Antibacterial agents in the treatment of ocular infections. Infect Dis Clin North Am. 1989; 3:533-551. 4. Steinert RF. Current therapy for bacterial keratitis and bacterial conjunctivitis. Am J Ophthalmol. 1991; 112(suppl):10S-14S. 5. Edwards JG, Schlech BA. Efficacy of topical chloramphenicol and tobramycin ophthalmic solutions in preventing severe Staphylococcus aureus keratitis in rabbits. CurrEye Res. 1985;4:821-829. 6. Gwon A. Ofloxacin vs tobramycin for the treatment of external ocular infection. Arch Ophthalmol. 1992;110:1234-1237. 7. Callegan MC, Hobden JA, Hill JM, Insler MS, O'Callaghan RJ. Topical antibiotic therapy for the treatment of experimental Staphylococcus aureus keratitis. Invest Ophthalmol Vis Sci. 1992;33:3017-3023. 8. Callegan MC, Hobden JA, Hill JM, Insler MS, O'Callaghan RJ. Methicillin-resistant Staphylococcus aureus keratitis in the rabbit: therapy with ciprofloxacin, vancomycin, and cefazolin. Curr Eye Res. 1992; 11:1111 1119. 9. SAS User's Guide: Statistics Version, edition 5. Cary, NC: SAS Institute, Inc.; 1985. 10. Engel LS, Callegan MC, Hill JM, O'Callaghan RJ. Bioassays for quantitating ciprofloxacin and tobramycin in aqueous humor. J Ocul Pharmacol. 1993;9:311 320. 11. Davis SD, Sarff LD, Hyndiuk RA. Staphylococcal keratitis: experimental model in guinea pigs. Arch Ophthalmol. 1978;96:2114-2116. 12. Kupferman A, Leibowitz HM. Topical antibiotic therapy of staphylococcal keratitis. Arch Ophthalmol. 1977;95:1634-1637. 13. Leibowitz HM, Ryan WJ, Kupferman A. Route of antibiotic administration in bacterial keratitis. Arch Ophthalmol. 1981;99:1420-1423. 14. Baum J, Barza M. Topical vs subconjunctival treatment of bacterial corneal ulcers. Ophthalmology. 1983;90:162-168. 15. Phinney RB, Schwartz SD, Lee DA, Mondino BJ. Collagen shield delivery of gentamicin and vancomycin. Arch Ophthalmol. 1988; 106:1599-1604. 16. Isaacs RD, Kunke PJ, Cohen RL, Smith JW. Ciprofloxacin resistance in epidemic methicillin-resistant Staphylococcus aureus. Lancet 1988; 2:843. 17. Raviglione MC, BouleJF, Mariuz P, et al. Ciprofloxacin-resistant Staphylococcus aureus in an acute-care hospital. Antimicrob Agents Chemother. 1990; 34:2050-2054.