606 British Journal of Ophthalmology 1995; 79: 606-609 Topical 0 3% ciprofloxacin, norfloxacin, and ofloxacin in treatment of bacterial keratitis: a new Br J Ophthalmol: first published as 10.1136/bjo.79.6.606 on 1 June 1995. Downloaded from method for comparative evaluation of ocular drug penetration Jeremy P Diamond, Les White, John P Leeming, H Bing Hoh, David L Easty Abstract dry eye, contact lens use, trauma, or the Aims-This study was designed to assess presence of a persistent comeal suture. the relative corneal penetration of topical Subsequent long term visual loss occurs as a drops of three antibiotics and to relate consequence of comeal scarring affecting the those levels to minimum inhibitory con- visual axis. The extent of scarring may be centrations for organisms associated with limited if the infection is identified early and bacterial keratitis. treated adequately. Optimal management Methods-Four drops of each of requires an attempt to isolate the causative ciprofloxacin, norfloxacin, and ofloxacin organism in order to ensure appropriate anti- (0.30/o topical ophthalmic preparations) biotic treatment. However, the aetiological were given to 12 patients undergoing agent is rarely known in time to guide initial corneal transplantation. After the recipi- antibiotic treatment and frequently is never ent tissue was removed, corneal drug identified. Thus, successful treatment of a pre- penetration was measured using high per- sumed bacterial comeal abscess is often depen- formance liquid chromatography. dent on empirical selection of an effective Results-Intracorneal concentrations of broad spectrum antibiotic regimen. For many ofloxacin (geometric mean 0-81 mg kg-1) years the most widely accepted option has been were significantly higher than both the instillation of a combination of a a lactam ciprofloxacin (0-60 mg kg-1; p=0.048) and and an aminoglycoside, both in high concen- norfloxacin (0.54 mg kg-1; p=0.012). tration solutions requiring local formulation. Ciprofloxacin and norfloxacin concentra- The fluoroquinolones are broad spectrum University ofBristol, tions did not differ significantly (p=0*33). bactericidal agents with activity against many of Departnent of Ophthalmology, Conclusions-Review of the minimum the important comeal pathogens including http://bjo.bmj.com/ Bristol Eye Hospital, inhibitory concentrations of the fluoro- staphylococci, Neisseria gonorrhoea, Haemophilus Bristol BS2 2LX quinolones against ocular pathogens influenzae, Enterobacteriaceae, and Pseudo- J P Diamond H Bing Hoh reveals that ciprofloxacin is more potent monas aeruginosa.1 2 Three fluoroquinolone D L Easty than ofloxacin against many bacteria; antibiotics have been formulated as 03%/o topi- ofloxacin is in turn more potent than nor- cal ophthalmic drops: ciprofloxacin (Alcon), Public Health Laboratory, Bristol floxacin. These data favour the selection norfloxacin (Merck, Sharpe and Dohme), and Royal Infirmary, of ciprofloxacin and ofloxacin rather than ofloxacin (Allergan). While these antibiotics on September 24, 2021 by guest. Protected copyright. Upper Maudlin Street, norfloxacin for the empirical treatment of have similar spectra of activity, they are not Bristol BS2 8HW J corneal infection. The greater potency of equally potent against many organisms and vary P Leeming ciprofloxacin offsets the superior penetra- in their pharmacokinetic properties.3 The Antimicrobial tion of ofloxacin. There is a need for success of an antibiotic depends on its ability to Reference Laboratory, improved clinical trial data concerning penetrate the site of infection in concentrations Department of Microbiology, the use of fluoroquinolone eyedrops in effective against the causative bacteria. There- Southmead Hospital, ulcerative keratitis; some encouraging fore, this project was designed to assess the rela- Bristol BS1O 5NB data are available for ciprofloxacin but tive comeal penetration oftopical drops ofthese L White not (in humans) for norfloxacin or antibiotics and to relate those levels to mini- Correspondence to: ofloxacin. mum inhibitory concentrations for organisms John P Leeming, Public Health Laboratory, Bristol (BrJ Ophthalmol 1995; 79: 606-609) commonly associated with bacterial keratitis. Royal Infirmary, Upper Maudlin Street, Bristol BS2 8HW. Accepted for publication Bacterial keratitis may arise secondary to Materials and methods 14 February 1995 comeal epithelial breakdown associated with Twelve patients due to undergo routine pene- trating keratoplasty (with or without cataract extraction) under general anaesthesia at the Table 1 Six quinolone antibiotic administration regimens showing the sequence in which patients received single drops oftopical 0 3% ciprofloxacin, norfioxacin, and ofloxacin, Bristol Eye Hospital were enrolled. Exclusion every 20 minutesfor 1 hour criteria included the presence of a broken corneal epithelium, active comeal inflamma- Sequence Start (0 min) + 20 min + 40 min + 60 min tory disease, topical antibiotic use during 1 Al/Bl Nor/Cip/Oflo Oflo/Nor/Cip Cip/Oflo/Nor Nor/Cip/Oflo week before enrolment, or a history of renal A2/B2 Oflo/Nor/Cip Cip/Oflo/Nor Nor/Cip/Oflo Oflo/Nor/Cip A3/B3 Cip/Oflo/Nor Nor/Cip/Oflo Oflo/Nor/Cip Cip/Oflo/Nor disease. Patients were given an information A4/B4 Nor/Oflo/Cip Cip/Nor/Oflo Oflo/Cip/Nor Nor/Oflo/Cip sheet pertaining to the trial before giving their A5/B5 Cip/Nor/Oflo Oflo/Cip/Nor Nor/Oflo/Cip Cip/Nor/Oflo A6/B6 Oflo/Cip/Nor informed consent. The trial was approved by Nor/Oflo/Cip Cip/Nor/Oflo Oflo/Cip/Nor the local ethics committee. Topical 0 3% ciprofloxacin, norfloxacin, and ofloxacin in treatment ofbacterial keratitis 607 drops and surgery was commenced within 60 minutes after instillation of the last antibiotic 0 4 drop. The host corneal button was snap frozen Br J Ophthalmol: first published as 10.1136/bjo.79.6.606 on 1 June 1995. Downloaded from in liquid nitrogen and stored at -70°C awaiting corneal drug assay. Corneal antibiotic concentrations were assayed by high performance liquid chroma- tography (HPLC) using methods based on those of McDermott4 and Gau.5 Corneal buttons were weighed, cut into small pieces with a scalpel, and homogenised in 500 [±l of water in a ground glass Griffiths tube. The bulk of the first homogenate was removed and a second 500 ,ul ofwater was added. After further homogenisation the solutions were pooled. After brief centrifugation the clear supernatant was analysed by HPLC on Spherisorb 5 ODS II (4X300 mm) (HPLC Technology, Macclesfield) at 50°C with a mobile phase of Water (1 litre) plus phosphoric acid (1 -6 ml) adjusted to pH 3-0 with tetrabutylammonium 0, hydroxide and mixed with acetonitrile (50 ml). d)qt 0 Flow rate was 1 ml min-I and detection was a1) by fluorescence (excitation 310 nm, emission 467 a Perkin-Elmer LC 240 0. nm) using detec- U) a) tor. External calibration was performed with an a) aqueous solution containing ciprofloxacin IL) c (1.67 mg l-l), norfloxacin (173 mg l-1), and a) C) ofloxacin (1 67 mg 1-1). Twenty ,ul ofcalibrator a)Co solution were processed in the same manner as 0 the corneal samples. The drug concentration in each corneal sample was determined by taking C C the dilution factor of the calibrator as 20/1020 0 and the dilution factor of the tests as x/l000+x 00 where x is the weight of the corneal plug in mg. Preliminary experiments (data not shown) indicated >95% recovery of calibrator solution processed in the presence of drug free cornea. http://bjo.bmj.com/ Typical within assay reproducibility was plus or minus 16% or better. A brief literature review was conducted to identify the bacterial species most commonly 0 6 implicated in keratitis and their sensitivity to the o fluoroquinolones studied. Where possible data 0)0 were selected from publications in which the on September 24, 2021 by guest. Protected copyright. C c sensitivities of bacteria of interest to all three ._F-0) 0 antimicrobials were compared. For each fluoro- quinolone the mean intracorneal concentration was expressed as a proportion of the minimum Figure 1 High concentration required to inhibit the growth of performance liquid 90% of isolates of these species (MIC90). The chromatogram of fluoroquinolones. Elution resultant coefficients were used to assess the order ofloxacin, likely in vivo potency of each agent. norfloxacin, ciprofloxacin. Time (seconds) Topical preparations of 0 3% ciprofloxacin, Results norfloxacin, and ofloxacin were obtained from Nine (75 0%) female and three (25.0%) male their manufacturers and rebottled to ensure that subjects were recruited; the mean age was 49 the same drop volume was applied in each case. (range 7-87) years. Each drug administration Patients received four drops of each drug over a permutation listed in Table 1 was used for two 60 minute period as shown in Table 1. To avoid patients. The three fluoroquinolones were all bias resulting from their order of presentation baseline separated in the HPLC analysis. A the three antibiotics were administered in one of typical chromatogram is shown in Figure 1. six sequence permutations. The first patient The antibiotic concentrations found in the enrolled was randomly allocated an administra- corneal buttons are given in Table 2. The tion regimen and thereafter, patients received geometric mean values (and 95% confidence the next regimen in the sequence. Drops were limits) were: ciprofloxacin 0-60 mg kg-' administered at 2 minute intervals to avoid (038-0 95 mg kg-'), norfloxacin 0 54 mg kg-l overspill. Patients received concurrent surgical (0-36-0@95 mg kg-'), and ofloxacin 0-81 mg preparation with unpreserved miotic/mydriatic kg-l (0-51-1-28 mg kg-'). The difference 608 Diamond, White, Leeming, Bing Hoh, Easty Table 2 Corneal antibiotic concentration (mg/kg) for 12 variability; in this study the concentrations of patients receivingfour topical drops of 0 3% ciprofloxacin, norfloxacin, and ofloxacin over 60 minutes before each fluoroquinolone spanned a range of undergoing penetrating keratoplasty approximately tenfold (Table 2). In conse- Br J Ophthalmol: first published as 10.1136/bjo.79.6.606 on 1 June 1995.