Minimum Inhibitory Concentrations of Standard and Novel Antimicrobials for Isolates from Bacterial Keratitis

Abstract

Purpose.: To determine the minimum inhibitory concentrations (MICs) of 12 antimicrobials in current ophthalmic use and 4 potentially new alternatives against isolates from bacterial keratitis.

Methods.: Bacteria were collected from cases of bacterial keratitis in six centers in the United Kingdom between 2003 and 2006. MICs were measured by using susceptibility strips containing a concentration gradient of the antimicrobials penicillin, cefuroxime, ceftazidime, chloramphenicol, gentamicin, amikacin, vancomycin, teicoplanin, ciprofloxacin, ofloxacin, levofloxacin, moxifloxacin, meropenem, linezolid, tigecycline, and daptomycin.

Results.: Isolates (n = 772) were collected including coagulase negative Staphylococcus (CNS) (30%), Pseudomonas aeruginosa (23%), Staphylococcus aureus (14%), Enterobacteriaceae (14%), and streptococci (13%). Meropenem had low MICs for most isolates. All isolates except P. aeruginosa were susceptible to tigecycline. Linezolid was active against the majority of Gram-positive pathogens. Ten percent of S. aureus and 20% of CNS isolates were methicillin resistant. When systemic breakpoints were used, 84% of S. aureus isolates were susceptible to ciprofloxacin and 98% to moxifloxacin. Of the P. aeruginosa isolates, 99% were susceptible to ceftazidime, 96% to gentamicin, 99% to ciprofloxacin and 100% to moxifloxacin. More than 97% of Enterobacteriaceae isolates were susceptible to ceftazidime, gentamicin, ciprofloxacin, and moxifloxacin.

Conclusions.: Based on systemic breakpoint data, resistance to commonly used antimicrobials was apparent. Meropenem is a potentially effective agent for ophthalmic use, with low MICs throughout all the bacterial subgroups. Tigecycline and linezolid showed good activity against particular groups and may be useful for treating bacterial keratitis resistant to current antimicrobials. Of the fluoroquinolones, moxifloxacin showed the lowest MICs and resistance for both Gram-positive and -negative bacteria.

Bacterial keratitis is a serious ocular disease that can lead to profound visual impairment. ^1–3 The likely causative organism and its susceptibility to antimicrobials varies according to the location and the degree of urbanization of the population studied. ^1–6 In the developed world, contact lens wear is a significant risk factor, and the infections are associated with a high proportion of Gram-negative bacterial isolates with a high morbidity. Lam et al. ⁴ reported that the incidence of bacterial keratitis was sixfold higher in contact lens wearers than in the general population. In contrast, the main causative factor in the developing world is ocular trauma. ⁵ Whitcher and Srinivasan ⁶ reported that microbial keratitis accounts for 55% of unilateral blindness in Bangladesh. 

Several antimicrobials are available to treat bacterial keratitis, with new antimicrobials occasionally introduced to provide an improved spectrum of activity against emerging or resistant pathogens. ⁷ Cephalosporins were introduced in the 1960s as resistance of Staphylococcus aureus to penicillin became common. Gentamicin was introduced in the 1970s, mainly because of an upsurge in Pseudomonas aeruginosa infections, associated with increasing contact lens wear. ⁸ The first- and second-generation fluoroquinolones (ciprofloxacin and ofloxacin) have excellent potency against Gram-negative bacteria and reasonably good activity against the Gram-positive bacteria with little corneal toxicity. When the third- and fourth-generation fluoroquinolones (levofloxacin, moxifloxacin, and gatifloxacin) became available in the early 2000s, they offered a broader Gram-positive cover. ^7–9  

Despite the success of the first- and second-generation fluoroquinolones, there has been a trend in increased resistance to both S. aureus ¹⁰ and P. aeruginosa. ¹¹ The fourth-generation fluoroquinolones moxifloxacin and gatifloxacin unfortunately have not been a treatment panacea because of the emergence of resistance. ^12,13 Moshirfar et al. ¹⁴ recently reported two cases of bacterial keratitis after refractive corneal surgery (one P. aeruginosa and the other S. aureus) that were resistant to moxifloxacin and gatifloxacin, and Park et al. ¹⁵ showed a rate of 2% resistance to moxifloxacin and 5% to gatifloxacin in isolates of normal bacterial ocular flora. 

In the face of increasing resistance, there is a need to evaluate new antimicrobials that could be used to treat bacterial keratitis. Meropenem ¹⁶ is a broad-spectrum carbapenem that is currently FDA approved to treat skin infections, intraabdominal infections and bacterial meningitis. It has activity against Gram-positive and -negative pathogens, including extended-spectrum β lactamase (ESBL) and AmpC-producing Enterobacteriaceae. A study looking at intravitreal meropenem in rabbit eyes found it to be as safe as ceftazidime. ¹⁷ Linezolid, ¹⁸ the first of a new class the oxazolidinones, is a synthetic compound with activity against all the major Gram-positive groups, but no activity against Gram-negative bacteria. It has been found to be safer than vancomycin in a rabbit keratitis model (Ekdawi NS, et al. IOVS 2005;46:ARVO E-Abstract 4910). ¹⁹ Tigecycline ²⁰ is a glycylcycline with activity against most aerobic and anaerobic Gram-positive and -negative bacteria but only limited activity against P. aeruginosa. Daptomycin ²¹ is a lipopeptide that is active against Gram-positive bacteria, including methicillin-resistant S. aureus (MRSA). There are no reports of the use of tigecycline or daptomycin in the cornea or indeed any epithelial cell line, although they have both been used successfully to treat resistant systemic infections. 

Topical application of an antimicrobial to the cornea may achieve a very different concentration and bioavailability in the tissue than can be achieved in the serum after systemic administration, although the latter MIC and disc susceptibility criteria are typically used to choose the antimicrobial for treatment of microbial keratitis. ^22–25 There have been reports, however, demonstrating the relationship between the MIC of topically applied antimicrobials and clinical outcome in bacterial keratitis, particularly for accepted pathogenic bacteria. ^26,27 The MIC is therefore, an important measure for evaluating the potential effectiveness of topically applied antimicrobials (new and existing) in the treatment of bacterial keratitis. The Microbiology Ophthalmic Group (MOG; see the 1 for member listing) comprises six centers in the United Kingdom. It was established in April 2003 to investigate the characteristics and antimicrobial susceptibility of bacterial isolates from cases of microbial keratitis. Data in relation to the antimicrobials prescribed for these infections and the clinical outcome are reported elsewhere. ²⁸ In this study, we compared the MICs of existing, recently introduced, and potential novel antimicrobials available for ophthalmic use against bacterial isolates. 

Methods

Six specialist ophthalmology centers (Birmingham, Bristol, Liverpool, London, Manchester, and Newcastle) participated in the study. Bacterial isolates from patients who presented with suppurative keratitis were collected at each center during the period April 2003 to April 2006. Corneal scrape samples were placed onto agar culture plates and also into enrichment culture broth (e.g., brain heart infusion). The plates were incubated overnight at 37°C under both aerobic and enriched carbon dioxide (5%) atmospheric conditions in the local laboratory, as previously described. ^29,30 Bacterial isolates from the six centers were then sent to one microbiology laboratory (Royal Liverpool University Hospital NHS Trust) where they were subcultured and stored on sterile beads (Protect Beads; TSC Ltd., Heywood, Lancashire, UK) at −80°C. The isolates were subsequently subcultured from the beads, and their identity reconfirmed by standard methods. ³¹ MICs were determined for all viable bacteria in each of five major subgroups of microorganisms. The exceptions were the coagulase negative Staphylococcus (CNS), for which a large representative sample was tested based on the spread of MICs to existing antimicrobials, and the two bacterial groups with fewer than 15 isolates, which were excluded (e.g., Listeria spp. and Acinetobacter spp.). 

The MICs were determined with susceptibility strips (E-test; AB Biodisk; Solna, Sweden) which are plastic strips containing a concentration gradient of antimicrobial, according to British Society of Antimicrobial Chemotherapy (BSAC) methods. ²⁴ A bacterial inoculum was prepared with a turbidity equivalent to that of a 0.5 McFarland standard. The following susceptibility media (Iso-sensitest agar; Oxoid, Basingstoke, UK) were inoculated: agar alone for S. aureus, CNS, P. aeruginosa, and Enterobacteriaceae; agar with 5% horse blood for Streptococcus spp. and Moraxella spp.; and agar with 5% horse blood and nicotinamide adenine dinucleotide for Haemophilus spp. E-test strips were placed on the inoculated plates before incubation at 35°C to 37°C for 18 to 24 hours. After incubation, MICs were read from the strips where the edge of bacterial inhibition intersected the side of the strip. Streptococcus and Haemophilus spp. were incubated in a 5% CO₂ rich environment, and all other microorganisms were incubated in air. Quality control strains (S. aureus ATCC 25923, P. aeruginosa ATCC 27853, Escherichia coli ATCC 25922, Enterococcus faecalis ATCC 29212, and Streptococcus pneumoniae ATCC 49619; American Type Culture Collection, Manassas, VA) were included. 

MICs (in milligrams per liter) were determined for the following antimicrobials: penicillin, cefuroxime, ceftazidime, chloramphenicol, gentamicin, amikacin, vancomycin, teicoplanin, ciprofloxacin, ofloxacin, levofloxacin, moxifloxacin, meropenem, linezolid, tigecycline, and daptomycin. MIC₅₀ and MIC₉₀ were calculated by interpolation from graphs of percentage cumulated strains inhibited versus MIC. ³² An isolate was characterized as resistant if the MIC was greater than the systemic breakpoint MIC defined by BSAC. ²⁴ It should be noted that breakpoints for the topical use of antimicrobials to treat bacterial keratitis are not available and may differ significantly from the systemic breakpoints. ²⁸ Methicillin resistance in S. aureus was confirmed by resistance to cefoxitin, determined by disc susceptibility test (10-μg disc). 

Results

A total of 772 bacterial isolates were submitted between 2003 and 2006. The different species isolated and results from three comparable studies are presented in Table 1. The five main groups of bacteria in our study were: CNS (27%), P. aeruginosa (21%), S. aureus (14%), Enterobacteriaceae (13%), and streptococci (13%). 

Table 1.

 

Percentage of Different Bacterial Species from Patients with Bacterial Keratitis from This Study and Three Comparable Studies

	Current Study (UK) n = 772	Tuft and Matheson 36 (UK) n = 1312	Bourcier et al. 1 (France) n = 208	Bharati et al. 5 (India) n = 1109
Gram-positive bacteria
Coagulase negative staphylococci	26.9	NA	48.1	17.4
Methicillin sensitive	21.7	NA	NA	NA
Methicillin resistant	5.2	NA	NA	NA
Staphylococcus aureus	13.6	33.4	7.7	3.9
Methicillin sensitive	12.4	NA	NA	NA
Methicillin resistant	1.2	NA	NA	NA
Streptococci	12.6	19.0	9.2	43.2
S. pneumoniae	3.4	NA	3.4	37.5
Other α-hemolytic	7.8	NA	5.8	4.4
β-hemolytic*	1.4	NA	NA	0.5
Other gram-positive†	4.4	3.5	17.8	5.5
Total gram-positive	57.4	55.9	82.8	70.1
Gram-negative bacteria
P. aeruginosa	20.9	24.8	10.1	18.0
Enterobacteriaceae	13.4	8.5	6.3	5.9
Serratia spp.	4.4	3.0	5.3	1.1
Klebsiella spp.	2.1	0.4	NA	1.1
Citrobacter spp.	1.6	0.4	NA	0.2
Proteus spp.	1.6	0.7	1.0	0.5
E. coli spp.	1.2	0.6	NA	0.4
Enterobacter spp.	1.0	2.1	NA	2.6
Morganella morganii spp.	0.1	NA	NA	NA
Pantoea spp.	0.1	NA	NA	NA
Other	1.2	1.3	NA	NA
Moraxella spp.	2.6	5.9	0.5	0.8
Haemophilus spp.	1.4	2.2	NA	0.5
Other gram-negative‡	4.2	2.8	NA	5.0
Total gram-negative	42.5	44.2	16.9	30.0

*  Lancefield group A (0.3%), group B (0.1%), group C (0.3%), and group G (0.8%).

†   Corynebacterium spp. (2.6%), Bacillus spp. (1.7%), Enterococcus spp. (0.9%), Listeria spp. (0.1%).

‡   Acinetobacter spp. (1.2%), Stenotrophomonas maltophilia (1.2%), Neisseria spp. (0.3%), Pasturella spp. (0.3%), Aeromonas spp. (0.1%), Eikenella spp. (0.1%), Agrobacterium spp. (0.1%), Alcaligines spp. (0.1%), Methylbacterium spp. (0.1%).

The MICs for the five bacterial subgroups are tabulated in Tables 2 through 6. Systemic breakpoint data were not available for cefuroxime, ceftazidime, levofloxacin, and meropenem against S. aureus and CNS; ceftazidime against streptococci; and ofloxacin and tigecycline against P. aeruginosa. All the antimicrobials were tested against all the microorganisms, but the results of isolates that were completely resistant are not shown in the tables. These include penicillin-, vancomycin-, teicoplanin-, linezolid-, and daptomycin-resistant Enterobacteriaceae and cefuroxime-, chloramphenicol-, and tigecycline-resistant P. aeruginosa. Streptococci were also completely resistant to gentamicin and amikacin and are not included. 

Table 2.

 

MICs to S. aureus

Antimicrobial	Min MIC	Max MIC	MIC50	MIC90	% Susceptible
Penicillin	0.016	48.000	0.750	1.500	22.3
Cefuroxime	0.094	512.000	0.750	1.500	NA
Ceftazidime	0.094	512.000	8.000	22.400	NA
Vancomycin	0.500	3.000	1.500	1.850	100.0
Teicoplanin	0.190	3.000	1.000	1.500	100.0
Gentamicin	0.016	512.000	0.380	1.000	95.7
Amikacin	0.500	8.000	3.000	4.000	100.0
Chloramphenicol	0.200	256.000	3.000	8.000	93.6
Ciprofloxacin	0.094	64.000	0.380	3.000	84.0
Ofloxacin	0.190	64.000	0.500	2.700	85.1
Levofloxacin	0.047	64.000	0.125	0.380	NA
Moxifloxacin	0.012	4.000	0.032	0.252	97.9
Meropenem	0.012	3.000	0.032	0.080	NA
Tigecycline	0.016	0.470	0.047	0.064	100.0
Linezolid	0.125	0.750	0.500	0.750	100.0
Daptomycin	0.047	0.640	0.094	0.125	100.0

 Data are in milligrams per liter, n = 94. NA, not available because the systemic break point has not been published and the percentage therefore cannot be calculated.

Methicillin resistance was present in 8.5% of the S. aureus isolates, all of which were resistant to ciprofloxacin. With the use of systemic breakpoints, all MRSA isolates were susceptible to vancomycin, teicoplanin, and amikacin. More than 93% of all isolates were susceptible to chloramphenicol and gentamicin (Table 2). Of the fluoroquinolones, moxifloxacin had higher susceptibility (98%) than did ciprofloxacin (84%) and ofloxacin (85%). One of the two isolates (2%) that was resistant to moxifloxacin was MRSA. Regarding the new antimicrobials, all the isolates were susceptible to linezolid, daptomycin, and tigecycline. For meropenem, the range of MICs of the methicillin-sensitive isolates (0.012–0.32 mg/L) was lower than for the MRSA isolates with MICs between 0.75 and 3.0 mg/L. Although the susceptibility of the isolates to meropenem could not be calculated (systemic breakpoints unavailable), it had the lowest distribution of MICs of the four newer antimicrobials. 

Table 3 shows the MICs for CNS isolates, 19.1% of which were found to be resistant to methicillin. All isolates were susceptible to vancomycin and amikacin. For the fluoroquinolones, susceptibility of the isolates was 95.7% to ciprofloxacin, 93.6% to ofloxacin, and 100% to moxifloxacin. All MIC₉₀ values were below 0.38 mg/L for meropenem and tigecycline, but one isolate had an elevated MIC of 16 mg/L to linezolid and seven isolates were above the systemic breakpoint of 1.0 mg/L for daptomycin, including one isolate with an MIC of 512 mg/L. 

Table 3.

 

MICs to CNS

Antimicrobial	Min MIC	Max MIC	MIC50	MIC90	% Susceptible
Penicillin	0.008	256.000	0.190	1.500	29.8
Cefuroxime	0.125	6.000	0.380	2.000	NA
Ceftazidime	1.500	512.000	6.000	19.200	NA
Vancomycin	0.380	3.000	2.000	2.000	100.0
Teicoplanin	0.250	16.000	2.000	6.000	97.9
Gentamicin	0.023	48.000	0.064	3.500	89.4
Amikacin	0.250	8.000	1.000	2.400	100.0
Chloramphenicol	0.750	512.000	1.500	32.800	87.2
Ciprofloxacin	0.094	8.000	0.190	0.3800	95.7
Ofloxacin	0.250	64.000	0.380	0.7500	93.6
Levofloxacin	0.064	4.000	0.190	0.2500	NA
Moxifloxacin	0.023	0.500	0.047	0.064	100.0
Meropenem	0.023	0.380	0.064	0.302	NA
Tigecycline	0.016	0.380	0.064	0.125	100.0
Linezolid	0.125	16.000	0.380	0.750	97.9
Daptomycin	0.094	512.000	0.380	1.500	85.1

 Data are in milligrams per liter, n = 47. NA, see Table 2.

Table 4 shows the MICs for the streptococci. Penicillin resistance was seen in 22.4% of the streptococci, and all these were α-hemolytic streptococci. One of the penicillin-resistant isolates also had an elevated cefuroxime MIC of 3.0 mg/L. All the S. pneumoniae and β-hemolytic streptococci were penicillin susceptible. All isolates were also susceptible to chloramphenicol, vancomycin, and teicoplanin. Susceptibility to ciprofloxacin was found in 86.6% of isolates, and all were susceptible to ofloxacin, levofloxacin, and moxifloxacin. Of the newer antimicrobials, all were susceptible to meropenem, tigecycline, and linezolid, but for daptomycin, 20.9% had elevated MICs above the systemic breakpoint. 

Table 4.

 

MICs to Streptococci*

Antimicrobial	Min MIC	Max MIC	MIC50	MIC90	% Susceptible
Penicillin	0.008	1.500	0.032	0.214	77.6
Cefuroxime	0.008	512.000	0.047	0.500	98.5
Ceftazidime	0.047	19.000	0.250	4.000	NA
Vancomycin	0.250	1.500	0.750	1.000	100.0
Teicoplanin	0.016	0.125	0.064	0.106	100.0
Chloramphenicol	0.380	8.000	1.500	3.000	100
Ciprofloxacin	0.250	4.000	1.000	3.000	86.6
Ofloxacin	1.000	4.000	2.000	4.000	100.0
Levofloxacin	0.250	1.500	0.750	1.500	100.0
Moxifloxacin	0.047	0.380	0.125	0.250	100.0
Meropenem	0.002	0.380	0.016	0.054	100.0
Tigecycline	0.023	0.380	0.047	0.125	100.0
Linezolid	0.190	1.500	0.500	1.000	100.0
Daptomycin	0.016	12.000	0.500	1.700	79.1

 Data are in milligrams per liter. n = 67. NA, see Table 2.

*  Includes 42 α-hemolytic streptococci, 18 S. pneumoniae, and 7 β-hemolytic streptococci.

The MICs for P. aeruginosa are shown in Table 5. Nearly all (99%) of the P. aeruginosa isolates were susceptible to ceftazidime. Of the fluoroquinolones, 99% of the isolates were susceptible to ciprofloxacin and levofloxacin and 100% to moxifloxacin; ciprofloxacin had the lowest range of MICs. Five isolates (3.3%) were resistant to amikacin and gentamicin and only one (0.7%) to meropenem. 

Table 5.

 

MICs to P. aeruginosa

Antimicrobial	Min MIC	Max MIC	MIC50	MIC90	% Susceptible
Ceftazidime	0.125	256.000	1.000	2.000	99.3
Gentamicin	0.125	256.000	1.000	2.000	96.4
Amikacin	0.125	512.000	2.000	4.000	96.4
Ciprofloxacin	0.016	6.000	0.094	0.500	98.6
Ofloxacin	0.032	12.000	0.625	1.500	NA
Levofloxacin	0.023	4.000	0.250	0.525	99.3
Moxifloxacin	0.012	16.000	0.500	1.000	100.0
Tigecycline	0.047	48.000	12.000	24.000	NA
Meropenem	0.012	16.000	0.094	0.250	99.3

 Data are in milligrams per liter, n = 140. NA, see Table 2.

Table 6 summarizes the MICs for the Enterobacteriaceae. Only 42.9% of the isolates were susceptible to cefuroxime and 68.8% to chloramphenicol. Susceptibility rates to ceftazidime, gentamicin, and amikacin were 98.7%, 98.7%, and 100%, respectively. Fluoroquinolone susceptibility was 98.7% to ciprofloxacin and ofloxacin and 100% to levofloxacin and moxifloxacin. All were fully susceptible to tigecycline and meropenem. 

Table 6.

 

MICs to Enterobacteriaceae

Antimicrobial	Min MIC	Max MIC	MIC50	MIC90	% Susceptible
Cefuroxime	0.190	512.000	12.000	176.000	42.9
Ceftazidime	0.023	512.000	0.190	0.750	98.7
Gentamicin	0.008	48.000	0.500	1.000	98.7
Amikacin	0.094	8.000	1.500	3.000	100.0
Chloramphenicol	0.250	512.000	5.000	2.800	68.8
Ciprofloxacin	0.003	32.000	0.032	0.125	98.7
Ofloxacin	0.023	512.000	0.125	0.380	98.7
Levofloxacin	0.008	1.000	0.047	0.190	100.0
Moxifloxacin	0.008	0.750	0.064	0.250	100.0
Meropenem	0.008	0.380	0.047	0.094	100.0
Tigecycline	0.023	512.000	0.380	1.500	100.0

 Data are in millograms per liter, n = 77. NA, see Table 2.

Discussion

Prompt and effective treatment of bacterial keratitis is critical in reducing the impact of this sight-threatening disease. There are several considerations when choosing the most appropriate antimicrobial therapy for bacterial keratitis. First, because an increase in bacterial resistance to the standard antimicrobials used to treat bacterial keratitis has been reported, it is essential to establish contemporaneous data on the spectrum of causative microorganisms and their expected resistance profile. ^{10,11,13,14,33} Temporal changes in the spectrum of pathogens and resistance have also been noted worldwide, ³⁴ and monitoring the susceptibility of bacterial isolates to current antimicrobials and evaluation of novel therapies is therefore important. Second, the absence of established breakpoint concentrations for antimicrobials when they are used topically makes the interpretation of MIC data difficult. The susceptibility criteria used to select an antimicrobial for treatment are based on the anticipated response of the bacteria against concentrations of the antimicrobial that can be achieved in serum. ^22–24 Topical application of an antimicrobial to the cornea may achieve a very different concentration and bioavailability in the tissue than the serum levels. ²⁵ Although the appropriate disc susceptibility breakpoint for each antimicrobial and bacterial isolate combination has not yet been determined, there is good evidence ^26–28 demonstrating the relationship between the MIC of topically applied antimicrobials and clinical outcome in bacterial keratitis. This relationship is particularly well established for pathogenic bacteria such as P. aeruginosa and S. aureus. ²⁸ The MIC is therefore an important measure for evaluating the potential effectiveness of topically applied antimicrobials in the treatment of bacterial keratitis. In this study, we used systemic breakpoint standards from the BSAC for reference, although these may not be relevant for interpreting the sensitivity of ocular isolates. ²⁴  

The proportion of negative isolation rates varied between centers from 52% to 58% which is similar to some studies, ^30,35 but lower than that reported by Wilhelmus et al. ²⁷ We are not able to comment on the response to treatment in cases of presumed microbial keratitis in which no bacteria were isolated. Sixteen percent of isolates in the study had been treated with combination therapy, which has been discussed in previously published work. ²⁸ Investigation into potential synergy and antagonism of antimicrobial combinations is the subject of further work. 

The range of bacteria isolated in this study from six centers in the United Kingdom demonstrated some differences from the three comparable studies shown in Table 1. ^1,5,36 For example, the proportion of Gram-positive isolates (57%) was similar to that in the study by Tuft and Matheson ³⁶ in London (56%), but much lower than that in Bourcier et al. ¹ in Paris (83%) and Bharati et al. ⁵ in South India (70%). S. aureus varied between 4% in India to 33% in the United Kingdom, and CNS from 17% in India to 48% in France. CNS was not reported in the study by Tuft and Matheson. ³⁶ It is considered part of the normal conjunctival flora ^37,38 and, despite being isolated from the corneas of patients with bacterial keratitis, it is often uncertain whether it is a pathogen. For this reason we included only a representative sample of our total CNS isolates in our study. Although a recent study established that there was a relationship between clinical outcome and the MIC of ciprofloxacin and ofloxacin against S. aureus, such a relationship was not apparent for CNS. ²⁸ Apart from the exclusion of CNS from the data, other differences in the proportion of isolates between studies may reflect differences in patient population, climate of the country, and the prevalence of risk factors such as contact lens use, trauma, or coexistent ocular disease. 

Based on systemic breakpoints, resistance was present to several commonly used antimicrobials in all five bacterial groups. Only 80% of S. aureus, 87% of streptococci, and 96% of CNS isolates were susceptible to ciprofloxacin. Particularly worrying were two S. aureus isolates that were resistant to moxifloxacin, one of which was also methicillin resistant. Although moxifloxacin-resistant S. aureus isolates from keratitis patients are currently uncommon, they are being reported. ^14,15 All the Gram-positive isolates tested were susceptible, however, to vancomycin and teicoplanin, meropenem, and tigecycline. In terms of newer antimicrobials, some CNS and streptococci had MICs above the systemic breakpoint for linezolid and daptomycin. Although all P. aeruginosa and Enterobacteriaceae were susceptible to meropenem, susceptibility levels of P. aeruginosa were high for gentamicin (96.4%) and ciprofloxacin (98.6%), and 99% of Enterobacteriaceae were susceptible to gentamicin, ciprofloxacin, moxifloxacin, and tigecycline. 

Overall, the in vitro data suggest that, of the fluoroquinolones currently in use, moxifloxacin offers the best coverage against both Gram-positive and -negative isolates. Moxifloxacin has also been shown in other studies to have both good patient tolerability ³⁹ and pharmacokinetics. ^40–42 Of the newer antimicrobials potentially suitable for ophthalmic use, meropenem appears to be a good choice for empiric monotherapy in bacterial keratitis, offering broad-spectrum cover against both Gram-positive and -negative microorganisms, with only one of the 772 isolates tested (P. aeruginosa) found to be resistant. Because tigecycline did not have low MICs against P. aeruginosa, it would not be an ideal agent for monotherapy. Linezolid was active against most of the Gram-positive isolates including a small number of MRSA isolates. Tigecycline and linezolid may therefore be useful for dual therapy when used in combination with an antimicrobial with good cover against Gram-negative bacteria. Our finding that up to 20% of streptococci were resistant to daptomycin suggests that it would not provide as good cover against Gram-positive organisms as would the available fluoroquinolones. 

Any conclusions about antimicrobial treatment of bacterial keratitis based on in vitro sensitivity results have inherent limitations. As discussed, the breakpoints for the antimicrobials are determined from expected serum concentrations after systemic administration and breakpoints after topical administration have yet to be standardized. More information about the bioavailability of antimicrobials delivered topically to the cornea is needed, to interpret the MIC levels. This is especially relevant to the novel antimicrobials tested. Topical application of antimicrobials to the cornea and ocular surface can probably achieve much higher concentrations and bioavailability in the tear film and tissue than can be obtained in serum. ²⁵ Proposals and data for establishing ophthalmic breakpoint concentrations are being established, ²⁸ along with studies assessing the pharmacokinetics and biological activity of antimicrobials within the cornea. ^25,40,43,44  

Footnotes

 Presented at the annual meeting of the Association for Research in Vision and Ophthalmology, Fort Lauderdale, Florida, May 2009.

Footnotes

 Supported by a grant from the Foundation for the Prevention of Blindness.

Footnotes

 Disclosure: H. Sueke, None; S. Kaye, None; T. Neal, None; C. Murphy, None; A. Hall, None; D. Whittaker, None; S. Tuft, None; C. Parry, None

Appendix

Members of the Microbiology Ophthalmic Group

Stephen Tuft (Moorfields Eye Hospital); Stephen Kaye and Timothy Neal (Royal Liverpool University Hospital); Derek Tole and John Leeming (Bristol Eye Hospital); Peter McDonnell and the late Timothy Weller (Birmingham and Midlands Eye Hospital); Francisco Figueiredo and Steven Pedler (Royal Victoria Infirmary, Newcastle); and Andrew Tullo and Malcolm Armstrong (Manchester Royal Eye Hospital).