Toxicity of Povidone-iodine to the ocular surface of rabbits

Many factors may affect the ocular surface environment after cataract surgery and many patients have complained of dry eye and symptoms of irritations postoperatively. Khanal et al. reported that corneal sensitivity and tear physiology were significantly deteriorated immediately after phacoemulsification [1]. Li et al. suggested that misuse of eye drops is a pathogenic factor that causes dry eye after cataract surgery [2]. We showed that preservatives in eye drops such as benzalkonium chloride can damage the ocular surface in a dose-dependent manner [3]. And Ram et al. proposed that corneal denervation caused by surgical incision may also be a risk factor of dry eye after cataract surgery [4]. In our previous study, the duration of cataract surgical time was highly correlated with development of ocular surface damage such as goblet cell loss and tear film instability [5]. Also, light from the operating microscope was shown to exert phototoxic effects on the ocular surface and the tear film in vivo. Therefore, excessive light exposure during ophthalmic procedures may be one pathogenic factor causing dry eye syndrome after cataract surgery [6]. Of the many factors causing ocular surface damage during such surgery, we proposed that soaking in 5% (w/v) povidone-iodine (PI) may be one of the major causes a major cause of such damage. Recently, topical polyvinylpyrrolidone-iodine (PI) or chlorhexidine has become widely used as a disinfectant in most fields of ophthalmic surgery, for preoperative prevention of endophthalmitis [7, 8]. Previous studies have shown that application of 5% (w/v) PI to the cornea and conjunctiva effectively decreased the bacterial load of the ocular surface and the adnexae, and thus theoretically reduced the risk of postoperative endophthalmitis [9‐12]. However, the product has an acidic pH (approximately 3.5), and preoperative PI has been shown to exert cytotoxic effects on the ocular surface epithelial cell [13, 14]. Thus, in the present study, we evaluated the effects of 5% (w/v) PI on the corneal and conjunctival surfaces of rabbit eyes, and ocular surface changes were compared based on PI exposure time. Although 5% (w/v) PI has been known to be toxic to the cornea, it has been proven effective to reduce bacteria on the ocular surface and widely used for preoperative preparation. So we decided to do this experiment with 5% (w/v) PI.

All experimental procedures were performed according to the guidelines of the Association for Research in Vision and Ophthalmology (ARVO) as set out in the “Statement for the Use of Animals in Ophthalmic and Visual Research” and the ARRIVE Guidelines for reporting animal research, and approved by the institutional animal care and use committee of Catholic university of Korea (Yeouido-2012-17-01FC). Twenty-three New Zealand white male rabbits (DooYeol Biotech, Seoul, Korea) weighing between 2 and 2.5 kg were used in the study. The rabbits were housed in the individual mesh cages and given free access to standard laboratory food and water. They were housed in an air-conditioned room under a 12 h-light/12 h-dark cycle. No rabbit had either a corneal or conjunctival disorder. A commercially available 10% (w/v) PI solution was diluted with distilled water to 5% (w/v). Rabbits were randomly divided into four groups, with topical administration of phosphate-buffered saline (PBS) for 10 min in the control group (ten eyes), and topical administration of 5% (w/v) PI for 1 min, 3 min, and 10 min in groups 1, 2, and 3, respectively (12 eyes each) (Table 1). The PBS and 5% PI were instilled only single drop. A mixture of tiletamine and zolazepam (Zoletil®, Virbac, Carros, France) at 0.2 mg/kg was injected intramuscularly to immobilize the rabbits prior to performing the examinations, and proxymetacaine hydrochloride 0.5% (w/v) eye drops (Alcaine®, Alcon Laboratories Inc., Fort Worth, TX, USA) were applied topically. And before the collection of cornea and conjunctival tissues, all animals were killed by intraperitoneal injection of a lethal overdose of pentobarbital (100 mg/kg body weight). We conducted the experiments in the same way that we had previously conducted [3, 6]. And all experiments were conducted by the same standard techniques and under the same environmental conditions. The analysis and scoring of the specimens was done by a single masked investigator who did not know the information about each group. Although the number of samples was not large, the experiment was conducted with the smallest sample available for statistical analysis.

Many studies have explored ocular surface damage associated with cataract surgery [1, 2, 4‐6, 18‐22]. Some factors are well-known to be associated with postoperative dry eye. These include corneal damage caused by ultrasound used during cataract surgery, topical anesthesia, eye drops containing preservatives such as benzalkonium chloride, corneal desensitization caused by corneal incision, surgical trauma, inflammatory responses producing chemicals such as free oxygen radicals, and exposure to light from microscopes [1, 2, 4‐6, 18‐22]. In clinical practice, some ophthalmologists have detected corneal damage immediately after preoperative PI instillation. Thus, we hypothesized that PI instillation might be pathogenic factor of dry eye syndrome in terms of inducing ocular surface damage in patients undergoing ophthalmic surgery. The purpose of our study was to evaluate the effects of 5% (w/v) PI on the ocular surfaces and identify possible pathognomonic factor triggering ocular surface damage after cataract surgery.

We evaluated the effects of 5% (w/v) PI applied for various exposure times, after single instillation onto the ocular surface of normal rabbits, by performing the Schirmer test, Rose Bengal staining, corneal fluorescein staining, conjunctival impression cytology and biopsy. We compared the changes of ocular surface among control group and study groups exposed to 5% (w/v) PI for 1, 3, and 10 min. The Rose Bengal staining scores were significantly increased immediately after PI instillation as the PI exposure time increased although the Schirmer test results were not different among the four groups. And on conjunctival impression cytology, the GCD was decreased and the Nelson score was increased with increasing exposure time. The GCD and Nelson score were significantly different between the 1-min and 3-min PI group and the 1-min and 10-min PI group. Conjunctival histology performed on day 7 revealed similar results. The conjunctival epithelial region of MUC5AC staining was reduced markedly in the 3-min and 10-min PI groups compared to the other groups. And SEM-based histological analyses showed loss of microplicae of superficial cornea. This corneal epithelial cell damage is also the reason for the positive staining in the fluorescein. We noted that exposure time-dependent pathophysiological ocular surface changes were similar to those seen in the dry eye syndrome; these included the reduction of GCD, an increase of conjunctival epithelial cell size, squamous metaplasia of conjunctiva and damaged superficial layer of the cornea.

The mucin and the microplicae play a role in tear film adhesion and stabilization to the corneal surface. And the microplicae are also important for the attachment of mucin to the cornea. Decreased production of MUC5AC, as well as changes in membrane-associated mucins also leads to loss of microplicae [23‐26]. The functions of mucous layer play a vital role in the stability of the tear film, converting the hydrophobic corneal epithelium to be hydrophilic and lubricating the ocular and palpebral surfaces. Thus, the damages of microplicae and decreased mucin production induced by PI interfere the formation of the innermost mucous layer of the tear film, and would lead to less retention of fluid even with functional lacrimal glands.

It is important to define the ocular surface toxicity caused by PI because of the recent spotlight cast on ocular surgery-associated dry eye syndrome. PI has been reported to be cytotoxic to the eye. Jiang et al. found that severe corneal epithelial damages were developed after PI instillation into the conjunctival sac, and significant corneal edema was observed after PI injection into the anterior chamber [3]. MacCrae et al. studied rabbit corneas after PI application, and noted moderate transient corneal edema at 5 min, which was resolved 3 h later [14]. Whitacre and Crockett assessed the toxicity of intravitreal PI by injecting 0.1 mL amounts of PI solutions at 0.05, 0.5, and 5% (all w/v) into the vitreous cavity [27]. One of 10 eyes injected with 0.05% PI had iritis, intraretinal hemorrhage and mild retinal necrosis, whereas all four eyes injected with 5% (w/v) PI had dense cataracts and full-thickness retinal necrosis. Apart from the direct toxicity of topical PI for the ocular surface, both contact dermatitis and keratoconjunctivitis sicca have been (rarely) reported [28, 29]. Therefore, intraocular contamination with PI is of concern. To the best of our knowledge, few studies have addressed the cytotoxicity of PI for the ocular surface, especially the effect of PI on conjunctival goblet cell function. And unfortunately, PI exposure times were not compared; all data were compared to only those of the control group. Although ocular toxicity of PI may increase as the exposure time becomes longer, few studies have sought to clearly define an appropriate exposure time.

It is clear that, in vivo, periocular preparation with PI alone (even for a long time) may not completely eradicate all types of microorganism that give rise to postoperative endophthalmitis. And as the results of present study show, the longer the PI exposure time, the more severe the ocular surface damages, so in order to minimize ocular surface toxicity, it would be better to use PI an optimum exposure time along with other prophylactic antibiotics or bactericidal agents that supplement the disinfective effect. In our present study, PI exposure for 1 min was not associated with any microscopically apparent damage to conjunctival epithelial cells, although corneal and conjunctival epithelial cells that were no longer adequately protected by tear films stained with Rose Bengal to a greater extent than control group. Upon conjunctival impression cytology, it was remarkable that no significant difference was apparent between the control and 1-min PI group. Thus, if the antimicrobial effect of 1 min of PI exposure is as effective as that afforded by 3 min of PI exposure, the optimum PI exposure time affording maximum eradication of microorganisms capable of causing endophthalmitis, but without causing ocular surface toxicity, should be 1 min rather than the 3 min of the previous guidelines.

The present study had several limitations. The follow-up period (7 days) was short, and the sample size was small, meaning that possible reversibility of ocular surface toxicity was not observed. The GCD of the 3-min group tended to recover on day 7. Our results indicate that long-term instillation of PI causes acute injury to the ocular surface and may cause chronic injury or induce dry eye. Nevertheless, as the 10-min group, unlike the 3-min group, showed no signs of recovery, further study is required to determine the possibility of chronic conjunctival injury following long-term PI instillation, to explore whether and when damaged ocular surfaces can recover. And in the present study, we applied just one drop of 5% PI for different exposure time. But nowadays, many authors recommend multiple applications of PI for a short exposure time. Therefore, we need more research to determine the effect of multiple short exposures versus one long exposure. Unfortunately, we could not check tear break up time for technical reasons and equipment-related problems. Although we could not demonstrate whether the decrease in GCD and MUC5AC directly caused dry eye, one study suggested a correlation between MUC5AC expression and dry eye clinical test results such as tear break up time [30]. Expression of conjunctival MUC5AC and the squamous metaplasia in CIC are closely correlated with tear break up time as a dry eye severity indicator [30, 31]. Therefore, based on these studies, we consider that the laboratory tests implemented in our study might supersede tear break up time. Use of PI may be a pathogenic factor causing postoperative dry eye resulting from the decrease in GCD and MUC5AC stain. However, we will measure tear break up time for confirmation in our next human clinical trial.

In conclusion, we have shown that longer exposure times with 5% (w/v) PI increasingly damaged the cornea and conjunctiva of normal rabbits (as shown by Rose Bengal staining and fluorescein staining); decreased the GCD, and triggered development of squamous metaplasia, as revealed by impression cytology and corneal and conjunctival histology. We could not confirm whether PI toxicity triggers not only acute injury but also chronic injury. So, further clinical study is required to clarify the relationship between PI and ocular surface toxicity. And also, we evaluated that such an instillation of PI might be pathogenic in terms of inducing ocular surface damage in patients undergoing ophthalmic surgery. Therefore, the toxicities of PI to ocular surface could be a possible pathogenic factor triggering dry eye syndrome after ophthalmic surgery.