Purpose
To report conjunctival bacterial flora antibiotic resistance patterns after serial intravitreal injections performed using a povidone-iodine preparation without the use of preinjection or postinjection topical antibiotics.
Design
Prospective, interventional case series.
Methods
Setting: Single-center clinical practice in Pennsylvania. Study Population: Thirteen eyes of 13 treatment-naïve patients undergoing serial intravitreal anti–vascular endothelial growth factor (VEGF) injections for exudative age-related macular degeneration or macular edema attributable to retinal vein occlusion. Intervention: Conjunctival cultures from the treatment eye were performed prior to each injection preparation. A minimum of 3 monthly conjunctival cultures were obtained per eye over the course of the study. Ocular surface preparation consisted of topical anesthetic and povidone-iodine 5% without the use of preinjection or postinjection topical antibiotics. Main Outcome Measures: Conjunctival flora growth patterns and antibiotic resistance patterns to several common antibiotics tested over the course of the study.
Results
A total of 48 cultures were performed with a 77% culture positivity rate. Over the course of the serial conjunctival cultures in each patient, there was no evidence for emergence of resistant bacteria to any of the tested antibiotics (including fluoroquinolones and azithromycin) or significant alteration from baseline conjunctival flora. Of the 47 bacterial isolates, the most commonly isolated organism was coagulase-negative Staphylococcus both at baseline (73%) and following serial intravitreal injections (78%, P = .73).
Conclusions
Ocular surface preparation for intravitreal injection using povidone-iodine 5% alone in the absence of postinjection topical antibiotics does not appear to promote bacterial resistance or a discernible change in conjunctival flora.
Over the last few years, intravitreal injections have become an increasingly common part of clinical practice in ophthalmology. A variety of medications are available for delivery via intravitreal injection, including anti–vascular endothelial growth factor (VEGF) agents, steroids, and enzymatic protein fragments. Intravitreal injections of anti-VEGF agents performed as often as once a month are the standard of care for patients with exudative age-related macular degeneration (AMD), the most common cause of vision loss in patients older than 65 years of age, as well as for macular edema attributable to diabetic retinopathy and retinal vein occlusion. Given the aging population and increasing incidence of diabetes worldwide, the number of intravitreal injections is likely to continue to increase for the foreseeable future.
Intravitreal injections are generally safe and well tolerated. However, of the complications associated with this procedure, none are more serious than endophthalmitis. The reported incidence of endophthalmitis after intravitreal injection ranges from as low as 0.02% to as high as 0.26%. Despite the low incidence of endophthalmitis following intravitreal injection, the consequences can be devastating as some patients may suffer severe and permanent vision loss. As endophthalmitis is hypothesized to occur through either inoculation of bacteria into the eye at the time of injection or passage of bacteria into the eye through a wound tract, sterilization of the ocular surface is paramount.
In order to achieve this, patients typically undergo a standard ocular surface preparation that includes topical 5% povidone-iodine. Whether or not topical antibiotics should be used during the preparation and/or following intravitreal injection is currently under debate. To date, no prospective study has demonstrated a clear reduction in the incidence of endophthalmitis through the use of postinjection topical antibiotics. Recent studies suggest that repeated short courses of postinjection topical antibiotics not only do not decrease the risk of endophthalmitis but also may actually increase antibiotic resistance among conjunctival flora. One recent study found that topical antibiotics might also impact the nasopharyngeal flora and increase drug resistance in bacteria that contribute to soft tissue infections and pneumonia. Other studies have demonstrated that the causative organism in endophthalmitis is often one of the most important prognostic factors associated with final visual outcome. By increasing the proportion of drug-resistant bacteria on the ocular surface, postinjection antibiotics may place patients at risk of developing infections that are more difficult to treat. Little is known, however, about whether repeat application of povidone-iodine alters conjunctival flora. We sought to determine whether repeated use of topical povidone-iodine 5%, in the absence of pre- or postinjection topical antibiotics, affected antibiotic resistance patterns and/or altered the baseline conjunctival flora.
Methods
After Wills Eye Institute Institutional Review Board approval was prospectively obtained, this prospective, interventional case series was performed at The Retina Service of Wills Eye Institute from August 1, 2011 through July 31, 2012. All participants gave informed consent prior to enrollment and the study was conducted in accordance with Health Insurance Portability and Accountability Act regulations. Our study adhered to the tenets of the Declaration of Helsinki. The study was registered at ClinicalTrials.gov under the identifier NCT01531842 .
Treatment-naïve patients over the age of 18 were enrolled in our study if they presented with a diagnosis of choroidal neovascularization secondary to AMD or macular edema attributable to central or branch retinal vein occlusion and their treating physician determined that the patient would require at least 3 serial injections with an anti-VEGF agent. Exclusion criteria consisted of patients who had received a prior intraocular injection in either eye, current use of contact lenses, chronic use of any ophthalmic medication, ocular surgery within the past 6 months, use of ophthalmic medications for ocular injection in either eye within the past 6 months, use of systemic antibiotics within the past 6 months, and known allergy or contraindication to povidone-iodine or proparacaine.
The following standardized preparation technique was used for all intravitreal injections: patients were anesthetized using topical proparacaine 0.5% (Bausch and Lomb Inc, Tampa, Florida, USA) and the ocular surface and eyelids were sterilized using topical povidone-iodine 5% (Alcon Labs, Fort Worth, Texas, USA). Intravitreal injection was performed 3.5 mm posterior to the limbus in pseudophakic patients and 4 mm posterior to the limbus in phakic patients using a 30 or 32 gauge needle. All patients received at least 3 serial injections at monthly intervals. No pre- or postinjection antibiotic drops were given at the time of the injection and no topical antibiotics were prescribed for patients to use in the days leading up to or in the days following injection.
Conjunctival cultures were performed at baseline and at each subsequent injection. Prior to each injection and before any povidone-iodine drops were given, a conjunctival culture was performed in a standardized fashion by swabbing the inferior fornix using the BBL CultureSwab (Becton Dickinson, and Company, Sparks, Maryland, USA) and stored and transported according to manufacturer instructions. Care was taken to minimize contact with the eyelids and eyelid margin. Swab samples were cultured on blood, chocolate, Columbia CNA, and MacConkey agar plates. Gram staining and testing for the presence or absence of catalase and coagulase/agglutination was performed in order to identify specific bacterial strains. Mean inhibitory concentrations (MIC, mcg/mL) for moxifloxacin, gatifloxacin, and azithromycin were calculated using the Epsilometer test (bioMerieux, Inc, Durham, North Carolina, USA). Mean inhibitory concentrations for oxacillin, rifampin, tetracycline, vancomycin, and gentamicin were calculated using the double disc diffusion test on the Phoenix machine (Becton Dickinson). Antibiotic sensitivity using the categories “susceptible,” “intermediate,” and “resistant,” were determined from the MIC and organism according to the standards of the Clinical and Laboratory Standards Institute (CLSI).
Statistical Analysis
Fisher exact test (GraphPad, La Jolla, California, USA) was used to analyze the following: culture positivity rates at baseline (culture 1) compared to follow-up (cultures 2-4); frequency of coagulase-negative Staphylococcus (CNS) at baseline compared to follow-up; and proportion of resistant cultures to each of the tested antibiotics during baseline compared to follow-up. A P value less than .05 was considered significant. For purposes of this analysis, bacteria that were found to demonstrate “intermediate resistance” according to the CLSI were considered resistant.
Results
Thirteen patients completed the study. The average age was 79 years (range, 28-92 years). Nine of the 13 patients were women. Eleven (85%) were being treated for neovascular age-related macular degeneration and 2 (15%) were being treated for macular edema attributable to central retinal vein occlusion. No patient had a history of prior intraocular injections in the past or topical antibiotic use within the prior 6 months.
Nine patients had serial conjunctival cultures on 4 separate occasions and 4 patients had serial conjunctival cultures on 3 separate occasions just prior to each intravitreal injection, leading to a total of 48 cultures obtained. Thirteen cultures were obtained at baseline, 12 cultures at month 1, 12 cultures at month 2, and 11 cultures at month 3. Of these, 37 grew at least 1 bacterial isolate, resulting in a 77% culture positivity rate. Ten of the 13 baseline cultures (77%) obtained grew at least 1 bacterial isolate. Of the subsequent 35 cultures obtained, 27 (77%) grew at least 1 bacterial isolate ( P > 0.99).
Of the 37 positive cultures, 47 bacterial isolates were identified from the conjunctiva. The most common bacterial isolate was CNS (36 of 47 bacterial isolates, 77%). At baseline, 11 of the 15 bacterial isolates (73%) were CNS, 2 (13%) were Propionibacterium acnes , 1 (7%) was Enterococcus , and 1 (7%) was Corynebacterium . Of the subsequent 32 bacterial isolates, 25 (78%) were CNS, 3 (9%) were P acnes , 2 (6%) were Corynebacterium , 1 (3%) was Actinomyces , and 1 (3%) was Staphylococcus aureus . There was no significant difference in the proportion of CNS at baseline compared to follow-up cultures ( P = .73).
The sensitivities of CNS to the tested antibiotics at baseline (culture 1) were compared to the follow-up period (aggregate of the results of cultures 2-4) and are shown in the Figure . With regard to the fluoroquinolones, 33% (3 of 9) of the CNS were resistant at baseline and 36% (9 of 25) were resistant subsequently ( P > 0.99). Azithromycin resistance in CNS was 56% (5 of 9) at baseline and 44% (11 of 25) subsequently ( P = .70). No differences found in resistance to the other tested antibiotics approached statistical significance.
