To determine the spectrum of conjunctival flora and the antibiotic susceptibility profiles of patients undergoing cataract surgery at a Midwestern university.
Prospective in vitro laboratory investigation of a patient cohort.
Conjunctival cultures were obtained from patients undergoing cataract surgery at a single ambulatory center on the day of surgery before the instillation of any ophthalmic medications. Isolates and antibiotic susceptibility profiles were identified using standard microbiological techniques.
A total of 183 eyes were cultured, yielding 225 isolates. Twenty-seven eyes (14.8%) showed no growth. Coagulase-negative staphylococci (CNS) were the most commonly isolated organisms (74.8%). Overall susceptibility was highest for gentamicin (94%), which was also true of the CNS isolates (95.0%). A total of 64.5% of CNS isolates were sensitive to ciprofloxacin; 30.1% of CNS isolates were resistant to ≥3 classes of antibiotics; 46.6% of CNS isolates were oxacillin-resistant, and they were more resistant to antibiotics than their oxacillin-sensitive counterparts ( P < .001), including fluoroquinolones ( P < .001). Among eyes with multiple CNS strains, 41.4% had different antibiotic susceptibility profiles even though they were the same species.
Our cohort harbored organisms with similar rates of antibiotic resistance as elsewhere in the country, including oxacillin resistance; however, the rate of fluoroquinolone resistance was less than in other reports. A surprisingly large proportion of different CNS strains from the same eye harbored different antibiotic susceptibility profiles. Our in vitro results, along with those of other investigators, should prompt further dialogue regarding antibiotic of choice for perioperative surgical prophylaxis in ophthalmic surgery.
Cataract extraction is one of the most common surgical procedures performed—certainly in the United States—yet there is still much to be learned about the pathophysiology of postsurgical infections such as endophthalmitis and effective measures of prophylaxis against said infections. While newer evidence suggests the efficacy of intraoperative, intracameral antibiotics, topical, perisurgical antibiotic prophylaxis is still the mainstay of preventive treatment employed in the United States.
The source of microbial pathogens in endophthalmitis is nearly always the patient’s native ocular surface and adnexal flora. Therefore, empiric perioperative antibiotic prophylaxis assumes that the antibiotic used would be efficacious against the patient’s flora. While thankfully rare, endophthalmitis is an extreme, blinding consequence; therefore, resistance among common causal pathogens of endophthalmitis is of significant concern. However, bacterial flora and their antibiotic susceptibility vary from region to region as well as over time. Therefore, for empiric perioperative antibiotic use to be successful, periodic—if not continuous—surveillance is advocated.
There have been several studies examining the antimicrobial susceptibility and resistance of the ocular flora of patients undergoing anterior segment surgeries and endophthalmitis isolates. Most of these studies collected data from outside the United States or along the coasts of the United States. More recently, with the advent of intravitreal injections for macular degeneration, attention has been focused on the ocular flora of patients undergoing repeat injections. Collectively, these papers report a range of antimicrobial susceptibility and resistance rates for the most commonly isolated organisms specific to the individual reporting institutions. Additionally, nationwide surveillance programs are in place and have been published, giving a more “global” picture of ocular flora and antibiotic susceptibility patterns.
Data and knowledge of local, regional, and national antibiotic susceptibility patterns are important for health care providers, hospitals, researchers, pharmaceutical companies, and various government organizations such as public health agencies. For health care providers, including ophthalmologists, the knowledge is important for guiding antibiotic therapy, directing empiric antibiotic use, and monitoring the impact of such usage. Both local and broader geographic surveillances are recommended for clinical care. The ability to predict antibiotic resistance or to gauge the effectiveness of empiric antibiotic usage for individual patients depends on the resistance profile of likely pathogens in the individual’s local community. However, local data are oftentimes limited in terms of mere availability, sample size, and selection bias and may therefore not apply to a specific patient setting. In these situations, broader regional and national data may provide needed information as well as better project susceptibility trends and thereby place local data into perspective.
Therefore, under the aforementioned context, the purpose of the present study is to elucidate the spectrum of ocular flora and their antimicrobial susceptibility profiles in patients undergoing routine cataract surgery at an ophthalmic institution in the Midwest section of the United States. Given a dearth of data from this part of the country, we hope these local data will provide information to educate clinicians in this part of the country to help guide empiric, perioperative antibiotic usage for their patients. Additionally, we hope this report will join the national dialogue for comparison with results from other parts of the country for contemporary and future clinical care and research.
Patients scheduled to have exclusively cataract surgery at the Saint Louis University ambulatory surgery center in Saint Louis, Missouri were approached and asked to participate in this prospective study. Patients over age 18, irrespective of sex, ethnic background, or health status, who could provide their own consents were considered, and all participating subjects consented to the study.
Upon arrival at the surgery center but prior to surgery and before preoperative ophthalmic medications of any kind were administered, a conjunctival swab of the inferior fornix of the eye undergoing surgery was performed using transport media (CultureSwab; Besse Medical, West Chester, Ohio, USA; or TransPorter; Eye Care and Cure, Tucson, Arizona, USA) following the administration of preservative-free tetracaine hydrochloride 0.5% (Alcon Laboratories, Inc, Ft. Worth, Texas, USA) to the subject’s eye. The lag time between the application of preservative-free tetracaine and the conjunctival swab was within 1 minute.
The transport media were sent overnight to the Bascom Palmer Ocular Microbiology Laboratory in Miami, Florida for processing. Upon receipt, each swab was plated onto chocolate and 5% sheep blood agars. Plates were incubated at 35 C in a CO 2 incubator and observed for growth. Isolates were identified using standard microbiological protocols. In vitro susceptibility was determined using a combination of breakpoint minimal inhibitory concentrations (Vitek; BioMerieux, Durham, North Carolina, USA) and disk diffusion in accordance with Clinical Laboratory Standards Institute protocols. Note that these protocols and breakpoint concentrations are based on safe, achievable serum concentrations from systemic administration of antibiotics. There are no standards for ocular tissue concentration of topically applied antibiotics. Systemic serum concentrations and breakpoints have been used routinely in ophthalmic antibiotic and microbiological studies such as the current study. Quality control strains Staphylococcus aureus ATCC 29213 (Vitek) and S. aureus ATCC 25923 (disk diffusion) were included as controls to validate and monitor results. Results were recorded as susceptible, intermediate, or resistant.
With regard to fluoroquinolone susceptibility testing, isolates susceptible to older drugs (ciprofloxacin and levofloxacin) were not all additionally tested against newer drugs (moxifloxacin and gatifloxacin), per previous published reports and standards for in vitro microbiological testing where organisms susceptible to older fluoroquinolones were also susceptible to newer fluoroquinolones.
Descriptive statistics were calculated. Statistical analysis was performed using the Student t test for quantitative variables, Fisher exact test for categorical variables, and the McNemar test for paired-sample categorical variables. Two-tailed P values were calculated, and significance was set at .05. Statistical analyses involving the fluoroquinolones were carried out, factoring in the aforementioned cross-susceptibility between older and newer fluoroquinolones.
Between September 6, 2007 and September 3, 2009, 256 subjects were recruited and 199 completed the study. Sixteen had surgery in both eyes. For these subjects, only their first eye was included in the data analysis. Therefore, for the evaluation of the spectrum of ocular surface flora and its antimicrobial susceptibility, 183 eyes of 183 subjects were included. There were 74 male subjects and 109 female subjects. The average age of these subjects was 67.7 (range 32–93 years). No complications were encountered with the culturing process, and none of the subjects developed endophthalmitis or post–cataract extraction infection.
Two hundred twenty-five organisms were isolated from the 183 eyes. Of the 183 eyes, 27 (14.8%) had no growth. We isolated 201 gram-positive organisms (89.3%), 21 gram-negative organisms (9.3%), and 3 fungi (1.3%). Coagulase-negative staphylococci (CNS) was the most common group of bacterial organisms isolated (166 out of 222; 74.8%), of which 127 were speciated as Staphylococcus epidermidis . We were able to speciate all but 8 of the 166 CNS organisms. From 3 subjects, only 1 colony of CNS was recovered. Since this yield was below the laboratory acceptable threshold of 5 or more colonies, antibiotic susceptibility profiles were not performed for these 3 isolates. In eyes from which CNS was isolated, 27 eyes had 2 different CNS strains and 2 eyes had 3 different CNS strains. From our cohort, we also recovered 11 S. aureus isolates, 4 Pseudomonas species, and 2 Streptococcus species, neither of which was Streptococcus pneumoniae ( Table 1 ).
|Conjunctival Isolates||Percent of Total||Number (N = 225)|
|Coagulase-negative Staphylococcus (other) a||17.3%||39|
|Streptococcus species b||0.9%||2|
|Pseudomonas species c||1.8%||4|
|Other gram-negative species d||7.6%||17|
|Fungus (other) e||0.9%||2|
a Coagulase-negative Staphylococcus (other): S. auricularis (17), unspeciated (8), S. haemolyticus (4), S. simulans (4), S. warneri (2), S. xylosus (1), S. capitis (1), S. cohnii (1), S. lugdunensis (1).
d Gram-negative species: Proteus mirabilis (3), Acinetobacter species (3), Strenotrophomonas maltophilia (2), Moraxella catarrhalis (1), Citrobacter koseri (1), Enterobacter gergoviae (1), Klebsiella oxytoca (1), Enterobacter aerogenes (1), Serratia liquefiaciens (1), Aeromonas hydrophilia (1), Rhizobium radiobacter (1), non-glucose-fermenting bacilli (1).
Tables 2 through 4 present the antibiotic susceptibility of select isolates recovered. As a group, the isolates were least susceptible to penicillin (17%), followed by erythromycin (48%). Susceptibility was highest for the aminoglycosides (94% gentamicin, 90% tobramycin, and 93% amikacin). Focusing on CNS isolates, 55.2% (90 of 163) were resistant to ≥3 antibiotics, and 33.7% (55 of 163) were resistant to ≥5 antibiotics. A total of 30.1% (49 of 163) of CNS were resistant to ≥3 classes of antibiotics out of 8, and 7.4% of CNS were resistant to ≥5 classes of antibiotics. Resistance was defined as being “resistant” or “intermediate” in susceptibility. Aside from vancomycin (100% susceptible), the CNS isolates were most susceptible to gentamicin (95.0%) and tetracycline (91.3%).
|CNS (others) a||39||36||20||36||50||17||59||NT||33||73||34||53||35||100||36||97||NT||NT|
|Gram positive a||201||175||89||173||84||NT|
|CNS (others) a||39||36||92||36||81||NT|
|Gram-positive a||201||176||64||166||65 (66)||86||51 (75)||64||44 (78)|
|S. epidermidis||127||126||63||120||63 (64)||57||47 (75)||40||38 (78)|
|CNS (others) a||39||35||71||33||70 (71)||22||64 (76)||18||56 (77)|
|S. aureus||11||11||45||9||56 (60)||7||43 (60)||6||50 (67)|
|Other gram-negative species||17||16||88||15||93||NT||NT|
Regarding the fluoroquinolones, of the gram-positive isolates, 63.6% (112 out of 176) were susceptible to ciprofloxacin. Looking at just the CNS group of isolates, 64.5% (104 out of 161) of isolates were susceptible to ciprofloxacin, and excepting 1 isolate, all that were additionally tested were susceptible to levofloxacin (99 isolates tested), moxifloxacin (30 isolates tested), and gatifloxacin (20 isolates tested). This 1 exception was a methicillin-resistant Staphylococcus auricularis isolate that was susceptible to ciprofloxacin but was intermediately susceptible to levofloxacin, moxifloxacin, and gatifloxacin. Of those CNS isolates that were resistant to ciprofloxacin, 98.1% of those tested (52 out of 53) were co-resistant or intermediately susceptible to levofloxacin, 75.0% (36 out of 48) were co-resistant or intermediately susceptible to moxifloxacin, and 86.1% (31 out of 36) were co-resistant or intermediately susceptible to gatifloxacin. Looked at another way, of the 57 CNS isolates that were resistant to ciprofloxacin, 13 (22.8%) were subsequently found susceptible to one of the newer fluoroquinolones. Factoring in cross-susceptibility between older and newer fluoroquinolones, the in vitro susceptibility profiles between ciprofloxacin and levofloxacin were not different ( P = .480). They were different between ciprofloxacin and moxifloxacin ( P = .003) and between ciprofloxacin and gatifloxacin ( P = .009). There was no difference between the susceptibility profiles for moxifloxacin and gatifloxacin ( P = .480).
In comparison to gram-positive organisms, gram-negative organisms were more susceptible to ciprofloxacin (80%; 16 out of 20). The 3 isolates that were resistant to ciprofloxacin were also resistant to levofloxacin; the 1 intermediately susceptible isolate was susceptible to levofloxacin (overall 82% levofloxacin-susceptible; 14 out of 17 isolates).
Fifty-two percent of gram-positive organisms were susceptible to oxacillin, and all gram-positive organisms tested were susceptible to vancomycin. More specifically, 46.6% (75 out of 161 isolates) of CNS from our population were oxacillin-resistant CNS (ORCNS), 45.6% (57 out of 125) of S. epidermidis isolates were oxacillin-resistant, and 63.6% (7 out of 11) of S. aureus isolates were oxacillin-resistant (ORSA).
The ORCNS isolates were resistant to an average of 5.8 antibiotics, whereas their oxacillin-susceptible counterparts (OSCNS) were resistant to an average of 1.6 antibiotics ( P < .001). With regard to classes of antibiotics, ORCNS isolates were on average resistant to 2.9 out of 8 classes of antibiotics tested, whereas OSCNS were resistant to an average of 0.7 classes of antibiotics ( P < .001). With regard to fluoroquinolones, 64.9% of ORCNS (48 out of 74) isolates were resistant to ciprofloxacin. Of these, 100% were cross-resistant (26 out of 44) or intermediate (18 out of 44) to levofloxacin, 76.1% were cross-resistant (23 out of 46) or intermediate (12 out of 46) to moxifloxacin, and 76.9% were cross-resistant (23 out of 39) or intermediate (7 out of 39) to gatifloxacin. Compared to OSCNS isolates, ORCNS isolates were more resistant to ciprofloxacin (64.9% vs 9.4%; P < .001) and levofloxacin (64.8% vs 8.1%; P < .001). Again factoring in cross-susceptibility between older and newer fluoroquinolones, ORCNS isolates were more resistant to moxifloxacin (51.4% vs 0%; P < .001) and gatifloxacin (49.3% vs 0%; P < .001) than OSCNS.
The ORSA isolates were resistant to an average of 5.9 antibiotics, whereas their oxacillin-susceptible counterparts (OSSA) were resistant to an average of 2.5 antibiotics ( P = .03). ORSA isolates were resistant to an average of 3 classes of antibiotics, whereas OSSA isolates were susceptible to an average of 1.75 classes of antibiotics. However, this difference was not statistically significant ( P = .18).
In those eyes (29) from which we recovered more than 1 CNS strain, 12 (41.4%) were the same species but with different antibiotic susceptibility profiles. There were 13 eyes (44.8%) from which 1 CNS strain was oxacillin-susceptible but the other was resistant. Only in 4 eyes (13.8%) were the overall antibiotic profiles the same for the isolates; however, more pairs had identical fluoroquinolone susceptibility profiles (15; 51.7%).