Purpose
We determined whether Besivance (Bausch & Lomb), AzaSite (Inspire Pharmaceuticals, Inc; both with DuraSite bioadhesive [InSite Vision, Inc]) and ciprofloxacin are toxic inside the anterior chamber.
Design
Randomized, masked, placebo-controlled animal study.
Methods
Twenty New Zealand white rabbits (40 eyes) were randomized to 1 of 4 study groups: Besivance, AzaSite, ciprofloxacin, and balanced salt solution. Each eye was injected with 0.1 mL of the study medication. Clinical slit-lamp examinations were conducted at 24 and 48 hours after injection. All rabbits then were killed and all eyes were enucleated. We randomized eyes to either corneal vital staining or histopathologic examination. The main outcome measures were clinical and pathologic signs of toxicity.
Results
The 2 DuraSite-based study groups (Besivance and AzaSite) showed clinically and pathologically significant differences when compared with the ciprofloxacin and balanced salt solution groups. Besivance and AzaSite eyes exhibited significantly similar and severe clinical damage, including severe corneal edema. Ciprofloxacin and balanced salt solution eyes appeared very similar and had only mild conjunctival injection and limbal vascularity. Vital staining and histopathologic evaluation revealed glaucomatous and toxic damage in eyes given DuraSite-based medications, whereas non-DuraSite groups showed minimal changes.
Conclusions
DuraSite blocks the trabecular meshwork and may be additionally toxic when introduced as a large bolus. Until the safety of these medications is established with further studies using smaller injected volumes, we recommend placement of a suture over a clear corneal wound if DuraSite-based medications are used.
Besifloxacin (Besivance; Bausch & Lomb, Rochester, New York, USA), a fourth-generation fluoroquinolone, recently was approved by the United States Food and Drug Administration for topical treatment of bacterial conjunctivitis. The formulation includes the DuraSite system (InSite Vision, Inc, Alameda, California, USA), which may prolong effective therapeutic drug delivery to the surface of the eye. AzaSite (Inspire Pharmaceuticals, Inc, Durham, North Carolina, USA), a formulation of azithromycin, was the first Food and Drug Administration-approved drug with a DuraSite bioadhesive delivery system; Besivance is the second such drug. This novel system embeds the active antimicrobial agent in a liquid polymer. After the eye drop is administered, the DuraSite polymer covers the cornea and the medicine gradually is released from the matrix. As physical trauma (i.e., blinking) and time progress, the polymer breaks down and sloughs off into the normal lacrimal drainage, thereby giving the corneal surface continued and direct contact with a fresh layer of drug.
Sutureless clear corneal surgery can result in the tear film moving in and out of the eye with every blink if the wound leaks. Currently, it is unclear what effect, if any, DuraSite embedded with these antibiotics might have on anterior segment structures if it were to gain entry to the anterior chamber after sutureless clear corneal surgery. This study was undertaken to assess potential toxicity concerns.
Methods
This study was comprised of 4 randomly assigned study groups: Besivance 0.6% (10 eyes), AzaSite 1% (10 eyes), ciprofloxacin ophthalmic solution 0.3% (Falcon Pharmaceuticals, Fort Worth, Texas, USA; 10 eyes), and balanced salt solution (BSS; Alcon Laboratories, Fort Worth, Texas, USA; 10 eyes).
We used 20 New Zealand rabbits of the same age and sex, each weighing 2.8 to 3.2 kg. The animals were quarantined for 7 days before their use in the study. Each rabbit was examined before testing for ophthalmologic disease or abnormal variance in eye shape and size.
Throughout the study, each rabbit was anesthetized with an intramuscular injection of ketamine hydrochloride (50 mg/kg) and xylazine (7 mg/kg) in a mixture of 7:1, respectively, for the intracameral injections of the antibiotic or placebo and before the animals were killed humanely. Before the intracameral injection, topical proparacaine hydrochloride anesthetic agent also was applied to each cornea. Eye movement and animal respiration were monitored during surgery to ensure that adequate levels of anesthesia were maintained. The area around the eye was draped in an aseptic manner and a lid speculum was placed to retract the lids. Using an aseptic technique and a Zeiss surgical microscope, we inserted a 30-gauge needle through the anterior sclera and conjunctiva into the anterior chamber of the study eye. We then drew out 0.1 mL of aqueous from the anterior chamber into the syringe. The syringe containing aqueous was removed from the needle, leaving the needle in the anterior chamber, and the syringe containing 0.1 mL of the predetermined study treatment (the solution content was masked) was attached and injected.
Slit-lamp examination (SLE) was performed 24 and 48 hours after intracameral injections for assessment of clinical signs of toxicity and inflammation. The following parameters were analyzed and scored from grades 0 to 4 in a masked fashion by the same observer (L.W.): conjunctival and limbal injection, corneal edema and opacity, hypopyon, anterior chamber cells, anterior chamber flare, fibrin formation, iris vascularity, posterior synechia, and crystalline lens opacity. After the second SLE, each rabbit was anesthetized and then was killed humanely with a 1.0-mL intravenous injection of pentobarbital sodium–phenytoin sodium.
All study eyes were enucleated. In 8 rabbit eyes (2 from each treatment group), the cornea was removed and evaluated for any evidence of morphologic damage to the corneal endothelial cells by using vital staining. They were prepared according to a technique used in previous studies. Corneas and a 2.0-mm rim of sclera were removed from the globe. Scissors were used to complete the full-thickness scleral section. The anterior uvea (iris diaphragm) was peeled from the sclera and the corneal button immediately was placed endothelial side up. The endothelium then was stained immediately using a modification of a method by Spence and Peyman as follows: immersion in trypan blue 0.25% in normal saline for 90 seconds; rinsing in 2 baths of saline 0.9%, immersion in Alizarin red 1% for 60 seconds, rinsing with saline 0.9%. A central 8.25-mm corneal button then was punched out using a trephine. The specimens were placed on a microscopic slide, endothelium face up, for microscopic evaluation. The overall size and shape of the corneal endothelial cells in the eyes from the different treatment groups qualitatively were compared. Quantification of the number of corneal endothelial cells was performed by taking photomicrographs of the endothelium under light microscopy (Olympus Optical Company, Ltd, Center Valley, Pennsylvania, USA) at a standard magnification of ×200. The photomicrographs were obtained at the center and the periphery of each specimen in a standardized fashion. All photographs used the same pixel size to allow for comparison of corneas from the different groups. They were analyzed by a digital imaging system (Evaluation of Posterior Capsule Opacification), which originally was designed for the evaluation of posterior capsule opacification but contains standard processing functions. A reference circle of the same radius was used to quantify the number of corneal endothelial cells in different areas of each cornea. After projecting the circle on the appropriate photographs, the number of cells inside the circle was counted for each cornea, both in the center and in the periphery.
The other 32 globes were formalin fixated and processed for histopathologic examination of the corneal endothelium and anterior segment structures. We followed the previously established protocol for estimation of corneal endothelial cell count by counting the number of endothelial cells within a ×40 objective focused on the central cornea. The mean endothelial cell counts from 5 central corneal sections were calculated. This method has been shown to have a consistent correlation with specular microscopic cell density.
Statistical Analysis
We completed statistical analyses for all SLE data, gross measurements of globes after enucleation, and endothelial cell counts. The SLE data are nonparametric because the numbers represent a descriptive score of severity in each category. Thus, we used the median test and interquartile range to summarize each study group and the Mann–Whitney U test for comparisons between groups. All measurements and cell counts are parametric data; so, we calculated the mean and standard deviation for each study group, and then we compared the groups with the Student t test. With a Bonferroni correction for multiple comparisons, statistical significance was set at P < .003.
Results
Clinical
SLE 24 and 48 hours after surgery showed statistically significant differences between the DuraSite-based study eyes (Besivance and AzaSite) and the non-DuraSite eyes (ciprofloxacin and BSS). The Besivance and AzaSite study eyes showed profound conjunctival injection, moderate limbal vascularity, and severe, diffuse corneal edema. Corneal ectasia and bullous keratopathy were noted to varying degrees in all (20/20) DuraSite-injected eyes ( Figure 1 ). Because of the corneal opacification, no anterior segment structures could be assessed. There was no statistically significant or clinically apparent difference between the Besivance and AzaSite groups.
No statistically significant or clinically apparent differences were observed between the non-DuraSite eyes. Examination of the ciprofloxacin and BSS eyes showed very mild conjunctival injection in 60% (12/20) of eyes after 24 hours. Forty-eight hours after injection, mild limbal vascularity had developed in 80% (16/20) of eyes. All other areas were normal, aside from 1 eye (1/20) injected with ciprofloxacin that exhibited mild conjunctival injection and discharge, increased limbal vascularity, and moderately diffuse corneal opacification along with minimal aqueous fibrin. The slit-lamp results statistically were worse in every parameter that we could monitor for both products with DuraSite when compared with ciprofloxacin and BSS ( Table ).
| Besivance (10 Eyes) | AzaSite (10 Eyes) | Ciprofloxacin (10 Eyes) | BSS (10 Eyes) | |
|---|---|---|---|---|
| Conjunctival injection, limbal vasculature b | 3 (0), 2 (0) | 3 (0), 2 (0) | 0.5 (0.5), 1 (0) | 0.5 (0.5), 1 (0) |
| Corneal opacity, extent of opacity b | 4 (0), 4 (0) | 4 (0), 4 (0) | 0 (0), 0 (0) | 0 (0), 0 (0) |
| Aqueous cell, flare, fibrin, hypopion, iris vasculature, crystalline lens opacity, posterior synechia b , c | C/A | C/A | 0 (0) | 0 (0) |
| Globe volume (cm 3 ) d , e | 3.05 (0.17) | 3.16 (0.52) | 2.46 (0.13) | 2.56 (0.17) |
| Corneal horizontal, vertical dimensions (mm) d | 14.10 (0.57), 13.68 (0.96) | 13.85 (0.73), 13.58 (0.46) | 13.06 (0.35), 11.99 (0.96) | 13.09 (0.59), 12.25 (0.85) |
| Corneal endothelial cell density by histopathologic analysis (mm) d , f | 11.49 (5.39) | 8.26 (4.17) | 23.27 (7.46) | 24.40 (4.51) |
a P values for all categories were not statistically significant between Besivance and AzaSite or ciprofloxacin and BSS, whereas they were statistically significant between DuraSite-based groups (Besivance and AzaSite) and non–DuraSite-based groups with P ≤ 0.001.
b The median score is reported with the (interquartile range), calculated in each category for each study group based on the scored data (0 = no signs; 4 = very severe signs).
c All scores for these categories were exactly the same, and therefore have been compressed.
d The mean and (standard deviation) have been calculated.
e Globe volume = 4/3*π*( APR * HR * VR ), where APR is anteroposterior radius, HR is horizontal radius, and VR is vertical radius.
f Density = number of endothelial cells within a ×40 objective.
During SLE, it also was noted that some eyes appeared much larger than others, although this observation was not noted at the time of injection. The Table contains the gross measurements of the globes on enucleation after 48 hours. A statistically significant increase in size was found between the DuraSite-based study groups when compared with the non-DuraSite study groups.
Vital Staining
All aspects of the corneas in the ciprofloxacin group were very similar to those of the BSS group. Analyses of the 2 corneas showed corneal endothelial cells exhibiting a normal hexagonal pattern ( Figure 2 ). A mild edema of the intercellular borders was observed in some areas of the endothelium. Mild alterations of the size and shape of the endothelial cells also were observed in some areas, especially at the periphery of the specimens. One of the corneas in this group exhibited a few areas of linear endothelium damage, which was probably related to the manipulation of the specimen for the staining.
