In Vitro Fluoroquinolone Resistance in Staphylococcal Endophthalmitis Isolates

Objective  To evaluate the in vitro susceptibility and cross-resistance of gatifloxacin and moxifloxacin vs older fluoroquinolones among coagulase-negative staphylococci recovered from patients with clinical endophthalmitis.

Methods  A combination of E tests and disk diffusion methods was used to determine in vitro susceptibility and cross-resistance for 111 coagulase-negative staphylococci isolates recovered during a 15-year period (January 1, 1990, to December 31, 2004) against 5 fluoroquinolones.

Results  In vitro susceptibilities (percentage sensitive) in descending order were as follows: gatifloxacin, 74.5%; moxifloxacin, 72.1%; levofloxacin, 69.3%; ciprofloxacin, 65.6%, and ofloxacin, 60.4%. More than 65% of the coagulase-negative staphylococci resistant to ciprofloxacin (n = 38) demonstrated in vitro cross-resistance to gatifloxacin (25 [65.8%] of 38) and moxifloxacin (27 [71.1%] of 38). During the initial 5 years (January 1, 1990, to December 31, 1994), 96.6% of the coagulase-negative staphylococci were sensitive to gatifloxacin and moxifloxacin, with minimal inhibitory concentration required to inhibit or kill 90% of the isolates of 0.19 μg/mL and 0.12 μg/mL, respectively. During the last 5-year period (January 1, 2000, to December 31, 2004), the percentage of sensitive coagulase-negative staphylococci declined to 65.4% for gatifloxacin and moxifloxacin (P=.02). Minimal inhibitory concentration required to inhibit or kill 90% of the isolates was 32 μg/mL or greater for both drugs.

Conclusions  Gatifloxacin and moxifloxacin demonstrated an in vitro efficacy of less than 80% for coagulase-negative staphylococci endophthalmitis in the present study. Ciprofloxacin resistance may serve as a surrogate for concurrent in vitro resistance for gatifloxacin and moxifloxacin. Resistance increased significantly during the last 5 years. Declining in vitro susceptibility to gatifloxacin and moxifloxacin may have important implications for the prevention and treatment of postoperative endophthalmitis.

Endophthalmitis is an uncommon but potentially devastating adverse event following ocular surgery.¹ The incidence of endophthalmitis after phacoemulsification and extracapsular cataract extraction ranges from 0.07% to 0.13%.^1-6 Coagulase-negative staphylococci (CNS) are the most frequent bacterial pathogens recovered from patients with postoperative endophthalmitis.^3,7,8

Techniques to reduce the risk of endophthalmitis following cataract surgery include the use of topical preoperative antibiotics, the application of a combination of povidone and iodine to the ocular surface before surgery, the instillation of antibiotics in the infusion solution, intracameral antibiotic injections, and the administration of postsurgical subconjunctival or topical antibiotics.^9-12 Of this group, only preoperative povidone-iodine antisepsis has been supported by evidence-based medicine to have an effect on clinical outcomes.⁹

The role of preoperative and postoperative topical antibiotics for patients undergoing cataract surgery is controversial.^13,14 Issues include antibiotic selection, pathogen spectrum, emerging resistance, and drug pharmacodynamics.

The fluoroquinolones have been recommended as preferred therapeutic agents to guard against and manage all ocular infections.¹⁵ The newest members of this group, moxifloxacin and gatifloxacin, target DNA gyrase and type IV topoisomerase and are predicted to reduce cross-resistance with older fluoroquinolones as well as provide broader coverage for common gram-positive ocular pathogens.^16,17

The present laboratory study evaluates the in vitro susceptibility and cross-resistance of gatifloxacin and moxifloxacin vs ciprofloxacin, ofloxacin, and levofloxacin against CNS isolates recovered from the vitreous of patients with postoperative endophthalmitis during a 15-year period.

One hundred eleven current and cryopreserved CNS isolates recovered from the vitreous of patients with postoperative endophthalmitis between January 1, 1990, and December 31, 2004, were retrieved from the Bascom Palmer Eye Institute Microbiology Department Bacteria Collection and rehydrated according to standard microbiological protocols. Frozen isolates were thawed, resuspended in trypticase soy broth, and incubated at 35°C for 18 to 24 hours in a non–carbon dioxide incubator. Newly recovered isolates were subcultured onto 5% sheep blood agar and incubated an additional 18 to 24 hours at 35°C.

A combination of E tests (minimal inhibitory concentrations [MICs], measured in micrograms per milliliter; AB Biodisk NA, Inc, Piscataway, NJ) (n = 78) and disk diffusion (antibiotic-impregnated paper disks; Becton Dickinson, Cockeysville, Md) (n = 33) were used to document and report the percentage of isolates with in vitro resistance to gatifloxacin, moxifloxacin, levofloxacin, ofloxacin, and ciprofloxacin during the 15-year study period. Vancomycin susceptibility was also determined for comparison. Available MIC distributions (78 [70.3%]) were used to detect and compare emerging resistance (percentage resistant) and declining susceptibility (changes in modal MIC or MIC₅₀ [the concentration required to inhibit or kill 50% of the isolates]) for January 1, 1990, to December 31, 1994 (n = 29), January 1, 1995, to December 31, 1999 (n = 23), and January 1, 2000, to December 31, 2004 (n = 26).¹⁸ Total counts for the 3 periods were as follows: 1990 to 1994 (n = 40, ciprofloxacin, gatifloxacin, and moxifloxacin; and n = 34, ofloxacin and levofloxacin), 1995 to 1999 (n = 34, ciprofloxacin, gatifloxacin, and moxifloxacin; n = 29, ofloxacin; and n = 33, levofloxacin), and 2000 to 2004 (n = 37, ciprofloxacin, gatifloxacin, and moxifloxacin; and n = 34, ofloxacin and levofloxacin).

Test inoculum for each isolate and quality control strains was prepared using a 0.5 McFarland standard and then swabbed onto a 150-mm Mueller-Hinton agar plate. Either E test strips (AB Biodisk NA, Inc) or antibiotic-impregnated disks (Becton Dickinson) were dropped on a plate for each isolate. Plates were then incubated for 18 to 24 hours in a non–carbon dioxide incubator.

The MIC value (E test range, 0.002-32 μg/mL) was read at 18 to 24 hours as the point where the edge of the growing culture intersected the strip. The zone diameter (measured in millimeters) of growth inhibition around each disk was measured and recorded. Interpretations for sensitive, intermediate, and resistant were in accordance with Clinical Laboratory Standards Institute break points for MICs and zone sizes.¹⁹ Minimal inhibitory concentration values of greater than 2 μg/mL were recorded as resistant for gatifloxacin and moxifloxacin. The resistant cutoff value was greater than 4 μg/mL for ciprofloxacin, levofloxacin, and ofloxacin. The percentage susceptible was less than 1% for the 3 oldest fluoroquinolones and less than 0.05% for gatifloxacin and moxifloxacin. Zone diameters of less than 15 mm (ciprofloxacin), less than 14 mm (gatifloxacin and moxifloxacin), less than 13 mm (levofloxacin), and less than 12 mm (ofloxacin) were recorded as resistant. Zone sizes for percentage susceptible were greater than 21 mm (ciprofloxacin), greater than 19 mm (levofloxacin), greater than 18 mm (ofloxacin), greater than 23 mm (gatifloxacin), and greater than 24 mm (moxifloxacin).¹⁹

Previous and current identification of the isolates were performed using a combination of coagulase tests, analytical profile index manual kits, and a gram-positive identification card (Vitek; BioMerieux, St Louis, Mo).

This was a fixed sample size, at a power of 0.80 and an α of .05; a minimum of 34 isolates per study period could detect a difference of 8 points or more in the number of resistant isolates between the groups.²⁰ An Armitage trends in proportions test and the McNemar paired-sample test were used to compare and determine trends, drug differences, and statistical significance.

Staphylococcus epidermidis (75 [67.6%] of 111) was the most frequently isolated CNS species and was at least twice (39 [52.0%] of 75 vs 6 [16.7%] of 36; P<.001) as likely to be resistant to ciprofloxacin as non–S epidermidis isolates. Comparative in vitro susceptibilities to the 5 fluoroquinolones and vancomycin are displayed in Table 1. Significant differences between in vitro susceptibility were observed for gatifloxacin vs ciprofloxacin (P<.001), levofloxacin (P = .007), and ofloxacin (P = .001), but not moxifloxacin (P = .37). Similarly, significant differences were observed for moxifloxacin vs ciprofloxacin (P = .009) and ofloxacin (P = .02), but not levofloxacin (P = .08) or gatifloxacin (P = .37). No significant difference in in vitro susceptibility was detected for ciprofloxacin vs ofloxacin (P=.89), ciprofloxacin vs levofloxacin (P=.50), or ofloxacin vs levofloxacin (P=.58). None of the 111 isolates were resistant to vancomycin.

Considerable cross-resistance with ciprofloxacin, the oldest of the fluoroquinolones, was confirmed for these isolates (Table 1). At least 65% of the CNS isolates resistant to ciprofloxacin (n = 38) were also resistant to gatifloxacin (n = 25) and moxifloxacin (n = 27). More than 85% of the ciprofloxacin-resistant CNS isolates evaluated were also resistant to levofloxacin (30/34) and ofloxacin (31/32). Cross-resistance with ofloxacin and levofloxacin was greater than 60% for moxifloxacin (69.7% and 76.7%, respectively) and gatifloxacin (62.5% and 70.0%, respectively).

Gatifloxacin provided increased coverage for 34.2% (13/38) of the ciprofloxacin-resistant isolates, and moxifloxacin provided coverage for 28.9% (11/38). The difference in expanded coverage for gatifloxacin and moxifloxacin was not significant (P = .47). Less than 15% of the ciprofloxacin-resistant isolates were sensitive to levofloxacin (11.8%) or ofloxacin (3.1%). Coverage for ciprofloxacin-sensitive CNS isolates was 100.0% for gatifloxacin (73/73) and levofloxacin (67/67), 98.6% (72/73) for moxifloxacin, and 98.4% (63/64) for ofloxacin.

Minimal inhibitory concentration distributions ranged from 0.03 to 32 μg/mL. Comparative 5-year MIC₅₀ and MIC₉₀ (the concentration required to inhibit or kill 90% of the isolates) values are displayed in Table 2. An increase in the MIC₅₀ and a decrease in the percentage of isolates inhibited by the MIC₅₀ were documented for each successive period for all 5 fluoroquinolones. The decline in the number of susceptible isolates from the initial 5 years (1990-1994) through the last 5 years (2000-2004) was significant (by Armitage trends in proportions and 95% confidence interval) for ciprofloxacin (P = .03), levofloxacin (P = .004), gatifloxacin (P = .006), and moxifloxacin (P = .03), but not ofloxacin (P = .06). The greatest decline in in vitro activity for susceptible isolates was documented for levofloxacin (36.1%), followed by gatifloxacin and moxifloxacin (32.2%), ciprofloxacin (30.8%), and ofloxacin (20.6%).

An increase in the prevalence of resistant isolates was documented for each successive 5-year period and was significant for ciprofloxacin (P = .03), levofloxacin (P = .001), gatifloxacin (P = .003), and moxifloxacin (P = .007). Resistance increased for ofloxacin during the 15-year period but did not reach significance (P = .08). The MIC₉₀ in 2000 to 2004 increased by 8 (ciprofloxacin and ofloxacin) to 266 (moxifloxacin) times the initial concentration recorded in 1990 to 1994. The increase in the prevalence of resistant isolates matched the decline in the percentage of susceptible isolates.

During the first 5 years, only low-level resistance (4 μg/mL) was detected for ciprofloxacin and ofloxacin. None of the isolates at baseline demonstrated in vitro resistance to levofloxacin, gatifloxacin, or moxifloxacin. Isolates with intermediate susceptibility ranged from 4% to 7%. In the subsequent 5 years (1995-1999), resistance levels for ciprofloxacin and ofloxacin increased to moderate (8 μg/mL), and low to moderate-level resistance emerged for levofloxacin, gatifloxacin, and moxifloxacin. The appearance of high-level (32-μg/mL) resistance in ciprofloxacin during the last 5-year period correlated with high-level resistance in all other fluoroquinolones.

In this laboratory-based study, gatifloxacin and moxifloxacin demonstrated an in vitro efficiency of less than 80% against CNS recovered from the vitreous of patients with postoperative endophthalmitis. Cross-resistance was documented between gatifloxacin and moxifloxacin and all older fluoroquinolones. Expanded in vitro coverage for fluoroquinolone-resistant CNS vitreous isolates was less than 40% for moxifloxacin and gatifloxacin.

Staphylococcus epidermidis and other CNS remain the most frequent pathogens recovered from patients with postoperative endophthalmitis.^3,8 Declining in vitro susceptibility to the older fluoroquinolones of CNS and other ocular pathogens is problematic and has been documented by our institute and others.^21,22 Gatifloxacin and moxifloxacin, the newest and most potent of the fluoroquinolones, have been offered as replacements to provide expanded coverage for the prevention and management of postoperative endophthalmitis and other ocular infections.^16,17

The development of resistance to the newer fluoroquinolones is thought to be more difficult because of the bulky side chains and the need for mutations in the DNA gyrase and type IV topoisomerase enzymes.^16,17 However, the stage for gatifloxacin and moxifloxacin resistance may have already been set by the ubiquitous use of the older ocular fluoroquinolones, such as ciprofloxacin, and the high level of in vitro resistance reported for these drugs.^22-24

Emerging resistance to moxifloxacin and gatifloxacin has been documented for ocular and nonocular CNS isolates. In a study evaluating ocular isolates from European countries, Morrissey and colleagues²⁵ documented a 2% resistance for gatifloxacin among their CNS isolates. A study by Pong and colleagues²⁶ documented a 30% in vitro resistance for moxifloxacin against nonocular CNS isolates. Mather and colleagues²⁷ documented in vitro resistance of 20% or more among their ocular CNS isolates for moxifloxacin (5 [25%] of 20) and gatifloxacin (4 [20%] of 20). One study²⁸ reported no in vitro resistance among CNS ocular isolates.

In vitro susceptibility results alone may be insufficient to predict in vivo efficacy. A combination of the pharmacodynamics and pharmacokinetics of the drug, infection site, and the MIC₉₀ is needed to predict the in vivo efficacy and clinical applicability. For concentration-dependent antibiotics, such as the fluoroquinolones, the most effective clinical outcomes and microbiological cures have been documented for cases in which the ratio of peak drug concentration–MIC₉₀, or the inhibitory quotient, at the site exceeds 10.^29,30 Others^31-34 recommend reducing the selection of resistant isolates (mutants) by maintaining a concentration at the site of 8 to 10 times the MIC, the mutant prevention concentration. It might be difficult to obtain inhibitory quotients greater than 8, or mutant prevention concentrations 8 to 12 times the MIC, with currently described aqueous³⁵ and/or vitreous levels of gatifloxacin³⁶ and moxifloxacin³⁷ and reported MIC₉₀ values of 0.10 to 32 μg/mL.^24,28,38-40

The present study documents declining in vitro susceptibility and increasing resistance to the fluoroquinolones, including gatifloxacin and moxifloxacin, among CNS isolates recovered from the vitreous of patients with postoperative endophthalmitis. These results may have important implications for the prevention and treatment of endophthalmitis.

Correspondence: Darlene Miller, DHSc, MPH, Bascom Palmer Eye Institute, Miller School of Medicine, University of Miami, 900 NW 17th St, Miami, FL 33136 (dmiller@med.miami.edu).

Submitted for Publication: March 8, 2005; final revision received June 3, 2005; accepted June 12, 2005.

Financial Disclosure: None.

Funding/Support: This study was supported in part by a core grant from Research to Prevent Blindness.

Previous Presentation: This study was presented in part as a poster at the 2004 ARVO Annual Meeting; April 25-29, 2004; Fort Lauderdale, Fla.