In praise of povidone-iodine application in ophthalmology

Abstract

Polyvinyl pyrrolidone or povidone-iodine (PVP-I) is a water-soluble complex formed by the combination of iodine and a water-soluble polymer, polyvinyl pyrrolidone. This complex exerts bactericidal, fungicidal, and virucidal action by gradually releasing free iodine at the site of application to react with pathogens. In ophthalmology, PVP-I is used as a disinfectant and antiseptic agent for preoperative preparation of the skin and mucous membranes and for treating contaminated wounds. PVP-I has been shown to reduce effectively the risk of endophthalmitis in various ocular procedures, including cataract surgery and intravitreal injections; however, it has also been used in the treatment of conjunctivitis, keratitis, and endophthalmitis, with promising results especially in low-resource situations. PVP-I has been associated with complications such as postoperative eye pain, persistent corneal epithelial defects, ocular inflammation, and an attendant risk of keratitis. In cases of poor PVP-I tolerance, applying PVP-I at lower concentrations or using alternative antiseptics such as chlorhexidine should be considered. We provide an update on the efficacy of PVP-I in the prophylaxis and treatment of conjunctivitis, keratitis, and endophthalmitis and a comprehensive analysis of the current literature regarding the use of PVP-I in the management of these ocular conditions. Also, PVP-I-related adverse effects and toxicities and its alternatives are discussed. The goal is to present a thorough evaluation of the available evidence and to offer practical recommendations for clinicians regarding the therapeutic usage of PVP-I in ophthalmology.

1. Introduction

Polyvinyl pyrrolidone or povidone-iodine (PVP-I) is a water-soluble complex formed by the combination of iodine and a water-soluble polymer, polyvinyl pyrrolidone. This complex exerts bactericidal, fungicidal, and virucidal actions by gradually releasing free iodine at the site of application.^5,30,107 Free iodine rapidly penetrates microbial cell membranes and interacts with proteins, nucleotides, and fatty acids in the cytoplasm and cytoplasmic membrane, resulting in rapid cell death. PVP-I has a broad antimicrobial spectrum with activity against Gram-positive and Gram-negative bacteria, fungi, protozoa, viruses, bacterial spores, and even emerging pathogens like SARS-CoV-2.^26,107

In ophthalmology PVP-I is used as a disinfectant and antiseptic agent for preoperative preparation of the skin and mucous membranes and for treating contaminated wounds.²⁵ PVP-I is commonly used to sterilize the ocular surface before surgical procedures and has also been used as a treatment for adenoviral conjunctivitis.¹⁹ PVP-I’s bactericidal activity has been demonstrated in numerous clinical studies, with in vivo studies on the ocular surface showing a 96.7% kill rate within 60 seconds.^12,20 The use of PVP-I in ophthalmic surgical procedures has been effective in reducing the risk of postsurgical endophthalmitis.^86,102 Bacterial cultures found in the vitreous of patients with postsurgical endophthalmitis often match species found on the eyelids or conjunctiva, highlighting the importance of reducing the microbial load in these areas before surgery. Numerous studies have evaluated the use of PVP-I in ophthalmic surgeries, particularly during cataract surgery, where it has become a standard of care. Guidelines recommend an application of 5-10% PVP-I solution to the ocular surface, conjunctival sac, and periocular skin for a minimum of three minutes before surgery.⁸ Research has also demonstrated the effectiveness of PVP-I in reducing conjunctival sac colonization before cataract surgery.³⁷ Applying PVP-I 3 times a day for 3 days before surgery has been shown to reduce bacterial species count. For intravitreal injections, using PVP-I as the sole antiseptic agent is recommended, whereas the routine use of topical antibiotics may be harmful.²² PVP-I is also used for conjunctival lavage before vitrectomy and can potentially prevent scleral buckle infection during scleral buckling procedure.⁷⁹ No reports of resistance to PVP-I have been published in the context of topical ophthalmic use. Moreover, PVP-I does not induce new resistance or cross-resistance to antibiotics. PVP-I is widely accepted as an antiseptic agent in ophthalmic surgery due to its broad spectrum of microbicidal activity and strong evidence supporting its efficacy.²⁴

The ideal concentration of PVP-I for clinical practice in ophthalmology has yet to be determined;²⁵ however, a study conducted in 2003 demonstrated that the bactericidal effect of 5% PVP-I was superior to that of 1% PVP-I in cataract surgery patients.²⁰ Interestingly, the time required for killing bacteria is shorter for 0.1 to 1% PVP-I (15 seconds) than for 2.5 to 10% PVP-I (30-120 seconds), which is related to increased number of free diatomic iodine in diluted PVP-Is.⁹ In an in vivo rabbit study, concentrations of 0.5% or less were found to be non-irritating when administered six times per day, although corneal wound healing was delayed by one day in the 0.5% group.¹⁰⁶ Hence, it is essential to balance an effective and non-toxic concentration with appropriate exposure time.

In this review we aim to provide an update on the efficacy of PVP-I in the management of conjunctivitis, keratitis, and endophthalmitis. We will examine the current literature and provide a comprehensive analysis of the use of PVP-I in the management of these ocular conditions. Also, PVP-I-related adverse effects and toxicities and its alternatives will be discussed. Our goal is to present a thorough evaluation of the available evidence and to offer practical recommendations for clinicians.

2. Ophthalmia neonatorum

PVP-I has been compared with a variety of agents for the treatment of ophthalmia neonatorum, including silver nitrate solution,^31,32 tetracycline,¹⁷ erythromycin ointment,^3,31,32 and chloramphenicol ophthalmic drops⁷³ (Table 1). A study conducted by Isenberg and coworkers in California found that 2.5% PVP-I was more effective than silver nitrate or erythromycin and less toxic than silver nitrate in 100 healthy neonates.³¹ These findings were confirmed in a prospective trial of 3,117 infants in Kenya by Isenberg and coworkers.³² A recent study by Ali and coworkers in Iran on 330 infants also demonstrated the superior efficacy of 2.5% PVP-I compared to erythromycin 0.5% on ophthalmia neonatorum.³

Table 1.

Clinical studies on the effect of povidone-iodine on the treatment of ophthalmia neonatorum (PVP-I: povidone-iodine; CFU: colony-forming unit)

Author/Year	Disease	Design	Population	Management	Outcomes	Comments
Isenberg et al. 1994	Ophthalmia Neonatorum Prophylaxis	Interventional study	100 healthy neonates (vaginal delivery, excluded if there was a history of perinatal infection, amnionitis)	A drop of 2.5% PVP-I solution was instilled on one eye Other eye received either one drop of silver nitrate 1% or 0.5% erythromycin ointment. Conjunctival bacterial cultures were taken before treatment and again 2 to 4 hours after birth.	All 3 agents significantly reduced the number of CFUs, but PVP-I caused the most significant decrease.	Silver nitrate showed more ocular toxicity than the other two.
Isenberg et al. 1995	Ophthalmia Neonatorum Prophylaxis	Masked, prospective trial	3117 neonates	2.5 % PVP-I (n=1076) 1 % solution of silver nitrate (n=929) 0.5 % erythromycin ointment (n=1112)	A 2.5 % PVP-I ophthalmic solution for prophylaxis against ophthalmia neonatorum is: more effective than silver nitrate or erythromycin and less toxic and costs less.
Ali et al. 2007	Ophthalmia Neonatorum Prophylaxis	Randomized-clinical trial	310 neonates (100, 103, 107 groups A, B, C respectively)	A: 2.5% PVP-I B: 0.5% erythromycin ointment C: (control group) did not receive any treatment	Conjunctivitis occurred in 52 subjects (17%) 9% of the infected neonates were from group A, 18.4% from group B, and 22.4% from group C. 2.5% PVP-I drops had a significant effect on ophthalmia neonatorum.
Ramirez-Ortiz et al. 2007	Neonatal conjunctivitis Prophylaxis	Randomized-equivalency trial	2004 neonates	PVP-I 2.5% (n = 1024) Ophthalmic chloramphenicol (n = 974)	PVP-I and ophthalmic chloramphenicol showed the same efficacy during the first 48 hrs after birth. From day 3 to 15, PVP-I was 6% less effective than chloramphenicol. After day 16 no significant difference between the groups was seen.	No significant ocular side effects were seen in both groups. There was an increased risk of chlamydial conjunctivitis in the PVI administered group.
David et al. 2011	Ophthalmia Neonatorum Prophylaxis	Prospective, randomized, controlled observational study	394 neonates	PVP-I 2.5% (201) Tetracycline 1% ointment (193)	The incidence of ophthalmia neonatorum was significantly more after PVP-I than tetracycline prophylaxis (15.4% and 5.2% respectively. Noninfective ophthalmia neonatorum was seen in 10 (5%) of the 201 neonates treated with PVP-I and in none (0%) of the neonates treated with tetracycline. Infective ophthalmia neonatorum developed in 21 (10.4%) of the neonates treated with PVP-I and in ten (5.2%) after treatment with tetracycline. Ophthalmia neonatorum developed more commonly in the first 3 days after treatment with PVP-I.	PVP-I was associated with noninfective (sterile) conjunctivitis, probably because of its toxicity to the ocular surface in neonates. Tetracycline was marginally more effective against infective ophthalmia neonatorum.
Khan et al. 2015	Ophthalmia Neonatorum Prophylaxis	Interventional study	100 neonates (200 eyes)	2.5% PVP-I right eye 1.25% PVP-I left eye.	2.5% PVP-I solution caused a statistically significant decrease in the number of CFU. 1.25% PVP-I also reduced the number of CFU to a statistically significant level.	1.25% PVP-I is as effective as the 2.5% PVP-I in reducing the number of CFU in healthy conjunctivae of newborns.

The lower cost of PVP-I compared to other disinfecting agents and its convenient transportability in powder form and easy preparation make it an effective option for use in developing countries.³² A study conducted by Khan and coworkers found that 1.25% PVP-I causes an equal reduction in the number of colony-forming unit (CFU) compared to 2.5% solution in healthy neonates;⁴³ however, a 2.5% PVP-I solution may have toxicity to the ocular surface of infants and lead to non-infectious conjunctivitis.¹⁷ Ramirez-Ortiz and coworkers found that the risk of chlamydial neonatal conjunctivitis was increased by 2.5% PVP-I compared with chloramphenicol ophthalmic ointment in Southern Mexico infants.⁷³ Their study showed that PVP-I was less effective than chloramphenicol for the 3-15 days after birth. Topical 0.5% erythromycin is routinely used in the U.S; however, based on a recent Cochrane review,⁴¹ 2.5% PVP-I could be considered as a strong prophylactic agent in low-resourced situations.

3. Conjunctivitis

A brief review of the clinical studies evaluating the effect of povidone-iodine on the treatment of conjunctivitis is provided in Table 2.

Table 2.

Clinical studies on the effect of povidone-iodine on the treatment of conjunctivitis (VPVP-I: povidone-iodine; QID: 4 times/day; TID: 3 times/day; qPCR: quantitative polymerase chain reaction; CC-IFA: cell culture with confirmatory immunofluorescence; IOP: intraocular pressure; SEI: subepithelial infiltrates; n = number of patients; TEAEs: treatment-emergent adverse events; SD: standard deviation; AT: artificial tears)

Author/Year	Disease	Design	Population	Management	Outcomes	Comments
Isenberg et al. 2002	Acute conjunctivitis	Double-masked, controlled, prospective clinical trial	459 children (mean age 6.6 ± 6.6 years; range, 7 months–21 years)	1.25% PVP-I, QID, 1-2 weeks (230 patients) vs. Neomycin–polymyxin-B-gramicidin (229 patients)	1.25% PVP-I was as effective as antibiotics for treating bacterial conjunctivitis and marginally more effective against chlamydia. Both were ineffective against viral conjunctivitis.	Most cultured organism: Heamophilus sp. (35), Moraxella sp. (21), S. aureus (18), S. pneumonia (15). The mean number of days until cure was similar for PVI-treated eyes (9.4) and antibiotic-treated eyes (9.1). Chlamydia took longer to cure than either viral or bacterial infections regardless of treatment group. There were no statistically significant age or gender differences between the two groups. PVP-I should be strongly considered as a treatment for bacterial and chlamydial conjunctivitis.
Pelletier et al. 2009	Adenoviral conjunctivitis	Prospective, open-label, single-armed, descriptive phase II clinical study	6 patients (9 eyes)	Combined PVP-I 0.4% / dexamethasone 0.1%, QID, minimum of 5 days	In 8/9 eyes, clinical resolution was observed by day 3 or 4. In 6/6 eyes with detectable adenovirus by qPCR, clinical reduction in viral titer was seen by day 3, 4, or 5. In 5/6 eyes with infectious virus confirmed by CC-IFA, elimination of infectivity was seen by day 4 or 5.	No adverse effects related to PVP-I were seen.
Trinavarat et al. 2012	Epidemic keratoconjunctivitis	Prospective, interventional, uncontrolled	61 patients	PVP-I 2%, QID, for 1 week	The recovery rate within a week of treatment was 77%. 28 (45.9%) recovered within a week after the onset.	EKC occurred bilaterally in 40 (66%). The mean time elapsed before treatment was 2.1 ± 1.46 days. Application of PVP-I was sustained until recovery or completing a 7-day trial in 78.7%. No severe adverse effects have been reported.
Pinto et al. 2014	Presumed viral Conjunctivitis	Randomized, masked, and controlled trial	122 patients	Topical dexamethasone 0.1%/PVP-I 0.4%, QID, 1 week (61 patients) vs. Artificial tears, QID, 1 week (61 patients)	Shorter conjunctivitis duration (9.4 ± 4.6 d) in treatment group versus 11.8 ± 4.9 d in the control group.	There was no statistically significant difference between groups in terms of symptoms and incidence of SEI. More stinging was reported in the treatment group.
Tunay et al. 2014	Adenoviral conjunctivitis	Prospective, randomized, controlled	35 infants (35 eyes) Mean age 3.1 months	Conjunctiva irrigated with PVP-I 2.5% then lubrication and netilmicin 0.3% drops QID (15 eyes) vs Lubrications and netilmicin drops without PVP-I irrigation (20 eyes)	The median recovery time was earlier in Group 1 (7 d) than Group 2 (12 d).	No relation was found between gender and clinical scores (lid edema, chemosis, fragility of conjunctival vasculature, and pseudomembrane formation). Significant lower clinical scores were observed from Group 1.
Yazar et al. 2016	Adenoviral conjunctivitis	Retrospective	112 patients	PVP-I solution: 1 mL of Tears Naturale + 1 mL 0.5% PVP-I , 3 drops TID, 2 weeks (56 patients) vs •xsTrifluorothymidine (TFT) 3 drops TID, 2 weeks (56 patients)	54 (96.4%) who received PVP-I drops recovered in two weeks, while 2 (3.6%) had late recovery time. 33 (58.9%) of the control group recovered in two weeks while 23 (41.1%) took more time.	Overall, 92 (82.1%) patients had intra-familial transmission.
Kovalyuk et al. 2017	Adenoviral keratoconjunctivitis	Prospective, randomized, controlled, double-blinded clinical trial	78 eyes (26 eyes in each group)	PVP-I 1.0% and dexamethasone 0.1% vs Dexamethasone 0.1% vs Lubricating eyedrops (hypromellose 0.3%)	The fastest improvement in patients red eyes, discharge, superficial punctate keratitis and pseudomembranes was observed in the study group. Study group reached a near complete recovery in 5–7 days (confirmed by PCR). The slowest improvement was in the control 2 group. The rate of reduction in Adenovirus titers was the slowest in the control 1 group.	Adenovirus type 8 was the most common pathogen (83%). SEI were observed in 44% of the control 1 group, 20% of the control 2 group and in 0% of the study group.
Pepose et al. 2018	Adenoviral conjunctivitis	Multicenter, randomized, vehicle-controlled, double-masked phase 2 trial	176 patients	PVP-I 0.6%/dexamethasone 0.1% (48 patients) vs PVP-I 0.6% (50 patients) vs Vehicle (46 patients)	The proportion of clinical resolution at day 6 was higher with PVP-I/dexamethasone (31.3%) than with vehicle (10.9%) and PVP-I (18.0%). The proportion with adenoviral eradication was higher with PVP-I/dexamethasone than with vehicle at day 3 (35.4% vs 8.7%;) and day 6 (79.2% vs 56.5%) and vs PVP-I (day 3, 32.0%; day 6, 62.0%).	Adverse effects occurred in 69.0% (vehicle), 62.7% (PVP-I), and 53.4% (PVP-I / dexamethasone). Most frequent adverse effects were SEI, punctate keratitis, eyelid edema, and conjunctivitis. Discontinuation due to adverse effects occurred in 37 patients (vehicle, n = 16; PVP-I, n = 12; PVP-I/dexamethasone, n = 9).
Altan-Yaycioglu et al. 2019	SEI in adenoviral keratoconjunctivitis	Retrospective	211 patients	102 patients received 2% PVP-I, BID, 5 days Antibiotics in 93.4%, artificial tears in 88.2%, PVP-I 2% in 48.3%, corticosteroids in 20.3%, ganciclovir in 10.9%, and NSAID in 5.2%.	SEI developed in 33.3% of patients who received PVP-I 2% and in 45.9%, in patients who did not.	Mean age 33.03±14.76 years Clinical signs were conjunctival hyperemia (100%), conjunctival follicles (79.1%), eyelids edema (39.3%), chemosis (16.1%), pseudomembrane (16.6%), and corneal epitheliopathy (29.9%). In total, 13.3% developed conjunctival subepithelial fibrosis, and 39.8% developed SEI.
Pepose et al. 2019	Acute viral conjunctivitis	Randomized, double-masked, parallel-group, vehicle-controlled study	132 patients	PVP-I 0.6%/DEX 0.1% ophthalmic suspension, QID, 5 days (66 patients) vs Vehicle, QID, 5 days (66 patients)	There were no TEAEs In the masked phase, 56.1% of the PVP-I/DEX group experienced at least one TEAE vs 43.9% in the vehicle group 78.9% in the open-label phase experienced at least one TEAE. Most TEAEs were mild.	38 patients continued into the open-label portion of the study. PVP-I/DEX administered for ≤14 days had a favorable safety profile and was well tolerated
Shorter el al. 2019	Adenoviral conjunctivitis	Double-masked randomized trial	56 patients	One-time ophthalmic 5% PVP-I vs Preservative-free AT	There was no change in VA between baseline and day 1 in either group. In the 5% PVP-I group, there was no change in overall discomfort immediately post-administration or on day 1 compared to baseline. In the AT group, overall discomfort was lower immediately post-administration but returned to baseline levels by Day 1.	In the PVP-I group, corneal staining increased immediately post-administration but returned to baseline levels by Day 1.
Ta et al. 2020	Acute bacterial conjunctivitis	Prospective, randomized, double-masked, multicenter, phase 3 clinical trial	753 patients	PVI/DEX (n = 324) PVP-I (n = 108) Placebo (n = 321)	At the day 5 visit, clinical resolution was achieved by 50.5% (111/220) in the PVP-I/DEX group vs 42.8% (95/222) in the placebo group, and bacterial eradication was achieved by 43.3% (94/217) and 46.8% (102/218), respectively. Adverse effects were experienced by 32.8% (106/323), 39.8% (43/108), and 19.0% (61/321) of subjects treated with PVP-I/DEX, PVP-I, and placebo, respectively (most mild in severity).	The modified ITT population for the efficacy analysis comprised 526 subjects. Mean and SD age was 44.3 (22.9) years, and most were female (61.2%) and white (78.1%).
Than et al. 2021	Adenoviral Conjunctivitis	Double-masked pilot randomized trial	56 patients	Group 1: 5% PVP-I once (n=30) Group 2: AT (n=26)	Day 4, viral titers in the 5% PVP-I and AT groups were 2.5% ± 2.7% and 14.4% ± 10.5% of peak respectively. Severity of participant-reported tearing, lid swelling and redness as well as clinician-graded mucoid discharge, bulbar redness and bulbar edema was lower in the 5% PVP-I group than AT group on day 4. After day 4, viral titers, severity of signs and symptoms decreased markedly in both groups and no differences between groups were detected.

3.1. Adenoviral Conjunctivitis

Human adenovirus (HAdV) conjunctivitis is one of the most common types of conjunctivitis, accounting up to 75% of all cases.¹⁵ During the course of the disease, patients may develop corneal subepithelial infiltrates, which can result in significant visual impairment that may persist for an extended period.⁵⁰ Epidemic keratoconjunctivitis (EKC) can be caused by various serotypes of HAdV, with serotypes 8, 19, and 37 being the most prevalent. At present, the treatment of adenoviral keratoconjunctivitis is conservative in nature. The use of cold compresses and artificial tears can alleviate symptoms without reducing viral shedding, thereby limiting the severity and duration of the disease.⁵²

Yates and coworkers investigated the efficacy of PVP-I as a topical treatment for adenoviral eye infections.¹⁰⁴ They observed significant antiviral activity against multiple HAdV serotypes at concentrations of 5%, 2%, and 0.4%. However, the effectiveness varied based on the HAdV serotypes and PVP-I concentration.

Several clinical studies have demonstrated the safety and tolerability of PVP-I in the treatment of adenoviral conjunctivitis. In a study conducted by Yazar and coworkers, the administration of 3 drops of 0.5% PVP-I three times a day, in conjunction with artificial tears at pH 4.2 to increase tolerance, reduced recovery time by 2 weeks compared to the control group.¹⁰⁵ In another study by Trinavarat and coworkers, the outcome of 2% PVP-I four times a day treatment in 61 EKC patients was assessed.⁹⁷ Within the first week of treatment, the rate of recovery was 77%. Recovery was observed within one week after onset in 28 patients (45.9%). The treatment was well-tolerated, and no severe ocular or systemic adverse effects were reported. Within a week, EKC-induced ocular discomfort improved in 75% of patients.

In a study conducted by Özen Tunay and coworkers, the effect of conjunctival irrigation with 2.5% PVP-I (one-time in-office) was assessed in infants with adenoviral conjunctivitis.⁶³ The PVP-I-treated group was found to have lower median clinical scores (lid edema, chemosis of conjunctiva and fragility of its vessels, and formation of pseudomembrane) and significant faster median recovery time compared to the control group. In another study by Altan-Yycioglu and coworkers it was found that although the clinical severity of keratoconjunctivitis was the same in both groups, the incidence of subepithelial infiltrates was significantly reduced in the 2% PVP-I-treated group compared to the control group.⁴

In a double-masked randomized trial, Shorter and coworkers assessed the tolerability and safety of a single instillation of ophthalmic 5% PVP-I solution compared to artificial tears for the treatment of adenoviral conjunctivitis.⁸² The study found that a one-time application of 5% PVP-I had no adverse effects and was safe and tolerable. In a similar study conducted by Than and coworkers to determine the efficacy and safety of a one-time, in office instillation of 5% PVP-I for adenoviral conjunctivitis, it was reported that viral titers in the 5% PVP-I and artificial tear treated groups were 2.5% ± 2.7% and 14.4% ± 10.5% of peak, respectively.⁹⁴ On day 4, the severity score of clinical symptoms and signs was lower in the 5% PVP-I group than in the artificial tear group.

Combination therapy of PVP-I with dexamethasone (DEX) has been shown in several trials to result in faster improvement and a significant reduction in viral titer compared to the control group.^{16,49,66,68,72} In a single-armed non-controlled clinical trial, Pelletier and coworkers assessed the combination of 0.4% PVP-I / 0.1% DEX 4times a day for a minimum of 5days for 9 eyes.⁶⁶ The measured outcomes were clinical resolution of conjunctival injection and discharge, laboratory-confirmed resolution (CC-IFA), and viral titer levels over the time (qPCR). Clinical resolution was observed in eight out of nine eyes on day three or four within the treatment group. In 6 eyes, significant reduction in viral titer from baseline to days 3, 4, or 5 were seen. In 5 out of 6 eyes infected with virus confirmed by CC-IFA method, infection was eliminated by day 4 or 5.

A randomized controlled trial conducted by Pepose and coworkers demonstrated no statistically significant differences in clinical resolution between the treatment group receiving PVP-I 0.6% / DEX 0.1% and the control group (66.1% vs 58.9% on the 6^th day).⁶⁹ In a separate study, Pepose and coworkers used the same treatment of PVP-I 0.6% / DEX 0.1% and observed superior results compared to the control group.⁶⁸ The discrepancy in the findings may be attributed to the variations in the diagnostic criteria employed. Specifically, the diagnosis of adenoviral conjunctivitis was based solely on clinical confirmation in the former study, whereas a positive Rapid Pathogen Screening test was used in the latter.

Kovalyuk and coworkers investigated the efficacy of 1% PVP-I in treating PCR-confirmed adenoviral conjunctivitis.⁴⁹ Patients were assigned to one of three treatment groups: 1% PVP-I / 0.1% DEX, 0.1% DEX alone, and artificial tears administered 4 times daily for 7 days. Evaluations were conducted on days 4, 5, and 7. The inclusion of a DEX-only arm facilitated direct comparison with other treatments. Over the course of 7 days, a reduction in symptoms and clinical signs was observed, with the most significant improvement in clinical signs and symptoms occurring in the PVP-I-DEX treated group, followed by the DEX-only group and then the placebo group. The incidence of subepithelial infiltrates was highest in the DEX-only group (44%) compared to the placebo group (20%) and 0% in the PVP-I-DEX treated group. On day 6, the viral titer decreased by approximately 100% in the PVP-I-DEX treated group compared to a 77% decrease by artificial tears. The group treated only with DEX alone exhibited the least improvement in viral titer, with a reduction of nearly 65%.

In one of our recent studies, the effectiveness of 2% PVP-I eye drops in treating adenoviral keratoconjunctivitis was assessed through patient records and physical examinations.⁸⁴ Patients aged 12 years and older without iodine allergies were included in the study. They received treatment with povidone-iodine 2% eye drops, administered four times a day. Data regarding demographic characteristics, family history of adenoviral keratoconjunctivitis, follicular conjunctivitis, petechial conjunctival hemorrhages, periauricular lymphadenopathy, and the presence of conjunctival pseudomembrane were collected from the patients' medical records. The findings revealed a significant decrease in symptoms and a favorable safety and tolerance profile. Based on our experience, it is necessary to use PVP-I in the fornix to minimize an eye burning sensation.

3.2. Other forms of conjunctivitis

In 2002, Isenberg and coworkers conducted a double-masked, controlled clinical trial to evaluate the treatment of acute conjunctivitis in 459 children (mean age of 6.6 years).³³ They compared the efficacy of 1.25% PVP-I administered four times daily for 1-2 weeks versus neomycin-polymyxin-B-gramicidin. Their results indicated that 1.25% PVP-I was as effective as antibiotics in treating bacterial conjunctivitis, particularly against chlamydia; however, neither agent was effective in treating viral conjunctivitis. The authors concluded that PVP-I could be considered a viable treatment option for bacterial conjunctivitis (especially when caused by chlamydia) in developing countries where the cost of antibiotics is a significant concern.

Pinto and coworkers conducted a randomized, masked, controlled trial to investigate the treatment of presumed viral conjunctivitis.⁷² They compared topical PVP-I 0.4% / DEX 0.1% administered four times daily versus artificial tears administered four times daily for one week. The treatment group exhibited a shorter duration of conjunctivitis (9.4 days) compared to the control group (11.8 days); however, no statistically significant differences were observed between the two groups in terms of patients' symptoms, intraocular pressure and incidence of subepithelial corneal infiltrates.

Ta and coworkers conducted a randomized, double-masked, multicenter clinical trial to assess the treatment of acute bacterial conjunctivitis.⁹¹ They compared PVP-I 0.6% / DEX 0.1%, PVP-I 0.6%, and placebo in three groups. The clinical resolution rate was 50.5% (111/220) in the PVP-I / DEX group versus 42.8% (95/222) in the placebo group. Bacterial eradication occurred more rapidly in the placebo group.

4. Keratitis

4.1. Prophylaxis

Nowadays, refractive surgery is among one of the most common ophthalmic procedures worldwide. Post refractive surgery keratitis is one the most devastating complications, considered as the surgeon’s nightmare.⁸³ One useful strategy to prevent infection after corneal refractive surgery is application of PVP-I. In a French survey,⁹⁵ 77 (92.8%) surgeons used povidone-iodine for skin area disinfection and 54 (65%) for conjunctival fornix disinfection. The contact time of povidone-iodine was less than 3 minutes for 71 (85.0%) surgeons. Notably, we apply PVP-I only on the periocular skin in our routine practice due to the reports of association between PVP-I and complications such as central toxic keratopathy (CTK) (discussed later).

4.2. Treatment

Katz and coworkers conducted a randomized trial to investigate the effects of administering 2.5% PVP-I every 2hours for 2 weeks as an adjunct to antibiotic therapy in the treatment of corneal ulcers. The primary outcome of interest was corrected visual acuity. They concluded that the addition of PVP-I did not improve visual outcomes; however, the efficacy of PVP-I as a standalone therapy remained unknown. One potential limitation of their study was the failure to determine and include the exact microbiological typing. Other limitations included using the same clinician to analyze results and not blinding him to the treatment received by each patient. No significant difference in the side effects was observed.

Gregori and coworkers conducted a clinical trial to evaluate the efficacy of 5% PVP-I in reducing bacterial loads of corneal ulcers prior to administering antibiotics.²³ Their results indicated that a single administration of 5% PVP-I did not significantly reduce bacterial loads compared to rinsing alone. The authors suggested that this may be from inadequate penetration into the corneal stroma and the number of organisms present, among other possible factors.

Isenberg and coworkers conducted a randomized trial to compare the efficacy of 1.25% PVP-I ophthalmic solution with topical antibiotics (neomycin-polymyxin B-gramicidin in the Philippines; ciprofloxacin 0.3% in India) in treating bacterial keratitis.³⁴ The primary outcome measure was the interval from the start of treatment to “presumed cure,” defined as a closed epithelial defect without associated inflammatory signs. The authors also calculated and compared intervals to “recovering,” defined as a residual epithelial defect < 1 mm with only minimal inflammation. Comparisons of intervals to recovering and hazard ratios for both presumed cure and recovering revealed no significant differences between treatment groups. They concluded that there was no significant difference in the treatment of bacterial keratitis between topical 1.25% PVP-I and commonly available topical antibiotics in the developing world. Therefore, 1.25% PVP-I is a viable treatment option for bacterial keratitis when antibiotic treatment is not practical.

Bordin reported a 61-year-old man who experienced pain, redness, watering, and photophobia in his left eye over a 5-month period. The condition initially presented as conjunctivitis and subsequently degenerated into keratitis. Clinical examination revealed an inferior corneal ulcer. This was unresponsive to previous treatments, including antibiotics and antiviral medications; however, the administration of 0.66% PVP-I to the affected eye 3 times daily for 4weeks improved the signs and symptoms and led to the ulcer’s complete resolution.¹⁰ In another case report, Bordin described the successful resolution of peripheral ulcerative keratitis through treatment with 0.66% PVP-I. The patient was an 89-year-old woman with rheumatoid arthritis who presented with a superior corneal ulcer that was unresponsive to treatment with amniotic membrane, topical antibiotics, and antiviral drugs. She was subsequently received 0.66% PVP-I administered 3 times daily for 5 weeks, followed by the addition of serum eyedrops with regression of the ulcer.¹¹

Pedrotti and coworkers conducted a prospective pilot study to evaluate the safety and effectiveness of administering topical PVP-I during the time-to-results period for pathogen identification and susceptibility testing in patients with infectious keratitis.⁶⁵ Their results indicated that administering 0.66% PVP-I during this period is a safe strategy in patients with infectious keratitis and may often obviate the need for broad-spectrum antimicrobial agents. Furthermore, PVP-I was found to be effective in treating eyes with Gram-positive bacterial infections.

An experimental in vitro study was conducted by Redd and coworkers to compare the cytocidal activity of three commonly used antimicrobial agents against Acanthamoeba keratitis: povidone iodine 1%, chlorhexidine 0.04%, and natamycin 5%.⁷⁴ They obtained 9 Acanthamoeba isolates from corneal scrapings and performed a microdilution assay to determine the minimum cytocidal concentration (MCC) for each medication. The MCCs for chlorhexidine ranged from 3.12 mg/ml to 25 mg/ml, with a median of 12.5 mg/ml. Natamycin MCCs ranged from 390.6 to 3125 mg/ml, with a median of 390.6 mg/ml. Povidone iodine MCCs ranged from 0.3 to 78.1 mg/ml, with a median of 2.4 mg/ml. Povidone iodine 1% demonstrated significantly higher potency than chlorhexidine 0.04% and natamycin 5%. These findings suggest that povidone-iodine has the highest potency among the 3 agents and may hold potential for medical therapy in Acanthamoeba keratitis.

Aziztama designed a study to investigate the pharmacokinetics, safety, and efficacy of PVP-I in treating bacterial keratitis.⁷ In Phase I, rabbits received intravitreal injections of 0.1% or 0.3% PVP-I. No retinal damage was observed, and PVP-I had a half-life of approximately 3 hours. In Phase II, rabbits with Staphylococcus epidermidis infection were treated with PVP-I injections. Improvement was seen in all PVP-I treated groups, with bacterial growth only in low-concentration single-injection groups. PVP-I was found to be effective for treating bacterial infections with repeated injections, even at low concentrations.

In another in vitro study, Hadipour Jahromy and coworkers found that PVP-I 1% eye drops and antibiotics effectively reduced bacterial keratitis severity in mice, suggesting povidone-iodine as a potential first-line treatment.³⁵

Table 3 provides a brief review on the clinical studies evaluating the effect of povidone-iodine in the treatment of keratitis. Overall, we employ topical PVP-I as a last-resort option for infectious keratitis that does not respond to conventional treatments, in addition to considering keratoplasty. Nevertheless, in low-resource countries, it can be regarded as a promising treatment approach.

Table 3.

Clinical studies on the effect of povidone-iodine for treatment of keratitis (VA: visual acuity; PVP-I: povidone-iodine; AT: artificial tears; CFU: colony-forming unit)

Author/Year	Disease	Design	Population	Management	Outcomes	Comments
Katz et al. 2004	Corneal ulcers	Randomized trial	358 patients	Group 1: Standard antibiotics therapy (n=173) Group 2: Standard therapy + 2.5% PVP-I every 2 hours for 2 weeks (n=185)	At follow up, 3.9% in group 1 and 6.9% in group 2 had corrected VA < 20/400. 9.4% in group 1 and 13.1% group 2 had corrected VA < 20/60. 17.0% and 18.8% had scars in the visual axis. The addition of PVP-I to standard antibiotic therapy did not improve visual outcomes.
Gregori et al. 2006	Corneal ulcer	Randomized clinical trial	35 patients	Group 1: 5% PVP-I (n=18) Group 2: AT (n=17) Ulcers were cultured before and after a 10-minute application of 5% PVP-I or AT. All patients were then treated with standard antibiotic medications. The number of colony-forming units before and after PVP-I or placebo was compared.	Group 1: 8 (44%) were culture positive, either before and/or after pretreatment Group 2: 9 (53%) were culture positive. There was no statistical difference in CFU change after PVP-I versus AT. Group 1: 3 (17%) showed improvement. Group 2: 7 (41%) showed improvement.
Isenberg et al. 2017	Bacterial keratitis	Randomized, controlled, investigator-masked clinical trial	172 patients	Philippines: Group 1: 1.25% PVP-I (n=40) Group 2: Neomycin-polymyxin B-gramicidin (n=49) India: Group 1: 1.25% PVP-I (n=38) Group 2: Ciprofloxacin 0.3% (n=45)	Median interval to presumed cure in the Philippines was 7 days for both groups. In India was 12 days for PVP-I and 17 days for ciprofloxacin. Hazard ratio for presumed cure among those treated with PVP-I (vs. antibiotics) was 1.46 in the Philippines and 1.70 in India. Comparisons of intervals to recovering and hazard ratio for recovering also revealed no significant differences between treatment groups in either country.
Bordin 2020	Corneal ulcer	Case report	1 patient	0.66% PVP-I TID for 4 weeks	Complete resolution of ulcer	Pathogen was unknown. Previous therapies (antibiotics first and then antiviral) were unable to control the ulcer.
Bordin 2021	Peripheral ulcerative keratitis in rheumatoid arthritis	Case report	1 patient	0.66% PVP-I TID for 5 weeks and then associated with serum eyedrops	Signs and symptoms were alleviated	Previous unsuccessful treatments: amniotic membrane, topical antibiotics, and antiviral drugs.

5. Endophthalmitis

5.1. Prophylaxis

PVP-I is commonly used to prevent endophthalmitis in various ocular procedures, including cataract surgery and intravitreal injections. There is consensus that topical 5% PVP-I reduces the risk of endophthalmitis following cataract surgery, as demonstrated by multiple studies.^61,96,102 PVP-I has also been shown to reduce bacterial load and eliminate more dangerous forms of bacteria from ocular flora.¹⁸

Trinavart and coworkers reported that the application of topical 5% PVP-I reduced the incidence of endophthalmitis after cataract surgery from 0.294% over a 12-month period to 0.097% over a 16-month period.⁹⁶ The most common side effect was ocular irritation, which resolved spontaneously in most cases. No other forms of ocular surface toxicity were reported during follow-up examinations. Another retrospective study compared the incidence of endophthalmitis in extracapsular cataract extractions over an 8-year period and found that skin preparation with 10% PVP-I, in combination with conjunctival 5% PVP-I, significantly reduced the incidence of endophthalmitis compared to skin preparation with 10% PVP-I alone or skin and conjunctival preparation with 5% PVP-I.¹⁰²

In 2005, Netwich and coworkers demonstrated that the risk of endophthalmitis was significantly reduced over a 20-year period in intraocular surgeries through the application of 1% PVP-I drops and even more dramatically through irrigation of the conjunctival sac with 10 mL of 1% PVP-I;⁶¹ however, this reduction may not be solely attributable to the use of PVP-I.

Earlier studies in dogs’ eyes have shown that more dilute forms of PVP-I are as effective as more concentrated forms (0.2% PVP-I was found to be as effective as 1% and 5% PVP-I);⁷⁵ however, a subsequent study in humans found that 5% PVP-I was more effective than 1% PVP-I in reducing conjunctival bacterial flora.²⁰ In contrast, some authors have suggested that 1% PVP-I is at least as effective as 5% PVP-I in preventing postoperative endophthalmitis.¹⁰⁰

Shimada and coworkers compared the efficacy of irrigating the ocular surface with infusion fluid or 0.25% PVP-I during 25G vitrectomy.⁷⁸ The results indicated that irrigating the ocular surface with 0.25% PVP-I significantly reduced the risk of fluid catch bag bacterial contamination compared to infusion fluid (3.8% versus 23.1%). In the context of intravitreal injections, PVP-I has been demonstrated to be an effective and inexpensive method of preventing endophthalmitis.^53,57,77 Another study by Shimada and coworkers achieved a successful outcome, no cases of endophthalmitis in 15,144 intravitreal injections using surgical masks, disinfection of the eyelid with 10% PVP-I, irrigation of the conjunctiva with 10 mL of 0.25% PVP-I (5 mL before and after injection), and topical levofloxacin administered three days before and after injections.⁷⁷ The authors suggest that it is important to wait at least 30 seconds after irrigating the conjunctiva with PVP-I before proceeding with the injection.

A study by Stem and coworkers involving 154,198 cases of intravitreal injections found no difference in the prevalence of endophthalmitis between the use of 5% and 10% PVP-I.⁸⁷ In another retrospective study of 22,046 intravitreal injections, patients who reported PVP-I sensitivity and consequently received reduced or no PVP-I prior to intravitreal injections exhibited significantly higher rates of endophthalmitis compared to those who received the standard 5% PVP-I protocol.⁵⁷ These findings underscore the importance of educating patients about the potential consequences of not using PVP-I.

In terms of the timing of PVP-I application, Stranz and coworkers found that a double application of 5% PVP-I separated by 10-minute intervals was more effective in reducing positive cultures obtained via conjunctival swabs compared to a single application;⁸⁸ however, endophthalmitis did not occur in either group. Another study concluded that reapplying PVP-I after speculum insertion significantly reduced the risk of endophthalmitis in intravitreal injections.⁵³ Additionally, Hosseini and coworkers after an in vitro study suggested that applying 5% PVP-I 15 minutes or 10% for 5 minutes prior to cataract surgery is the most effective approach for endophthalmitis prophylaxis.²⁹

The American Academy of Ophthalmology “Cataract in the Adult Eye” Preferred Practice Patterns Guidelines recommend that topical 5% povidone-iodine drops be instilled into the conjunctival cul-de-sac preoperatively, whereas the Royal College of Ophthalmology Cataract Surgery Guidelines recommend a flush irrigation of 5% povidone-iodine into the conjunctival sac.¹

5.2. Treatment

Nakashizuka and coworkers collected a case series of 4 eyes with postoperative endophthalmitis to investigate the bactericidal effect of 0.025% PVP-I in Balanced Salt Solution PLUS (0.025% PVP-I-BSS PLUS) and its use in vitrectomy followed by postoperative intravitreal and intravenous antibiotics injection.⁵⁹ In all 4 cases, endophthalmitis resolved without adverse events or ocular toxicity. The authors concluded that 0.025% PVP-I-BSS PLUS is effectively bactericidal and non-toxic when used as an irrigation solution in vitrectomy. Otani and coworkers investigated the efficacy and adverse events associated with using BSS PLUS containing 0.025% PVP-I for capsular bag irrigation in a case of postoperative Gram-negative endophthalmitis primarily involving the anterior chamber.⁶² They reported that the endophthalmitis resolved, and no adverse events were observed.

In another case series of 9 eyes, Nakashizuka and coworkers investigated the use of intravitreal injection of 1.25% PVP-I followed by vitrectomy using 0.025% PVP-I irrigation for treating postoperative or endogenous endophthalmitis.⁶⁰ All but one case experienced rapid resolution of endophthalmitis and maintenance of good visual acuity. No adverse events were noted. The authors concluded that intravitreal injection of PVP-I followed by vitrectomy is a safe and effective treatment for endophthalmitis.

Kitagawa and coworkers reported on the use of Balance BSS Plus containing 0.025% PVP-I as an irrigation solution during surgery in 2 cases of bleb-related endophthalmitis that required simultaneous cataract surgery and vitrectomy.⁴⁶ Emergency surgery was performed to remove the infected bleb and reconstruct the filtering bleb. Post surgery, the patients were monitored for 3 and 7 months, respectively, and no recurrences or adverse events were observed. The authors recommended the fresh preparation of 0.025% PVP-I just prior to starting vitrectomy since the solution decolorizes, and its bactericidal effect diminishes after 15 minutes.

Tanaka and coworkers reported successful treatment of a case with endogenous endophthalmitis through intravitreal injection of PVP-I. The patient was an 88-year-old woman who developed endogenous endophthalmitis from bacterial infection following small bowel bypass surgery.⁹² Surgical treatment was not performed due to the patient’s poor general condition and mental status. Instead, an intravitreal injection of 0.1 mL of 1.25% PVP-I was administered. The inflammation rapidly diminished and hypopyon disappeared f4 days after treatment. The fundus became visible 7 days after treatment and no recurrence was observed. Visual acuity in the affected eye recovered to match that of the other eye. The authors concluded that intravitreal injection of PVP-I is a useful and effective alternative to antibiotics for treating endogenous endophthalmitis in patients for whom surgical therapy is difficult.

A study by Lee and coworkers examined the efficacy of intravitreal PVP-I injection in rabbits with Candida albicans endophthalmitis.⁵¹ They showed that intravitreal PVP-I significantly improves Candida albicans endophthalmitis, and the effect are comparable that of the voriconazole; however, no synergistic effect was noted in the combination of PVP-I and voriconazole.

Based on the literature review, it seems that a concentration of 0.025% PVP-I in the irrigation solution during vitrectomy is safe and effective for treatment of endophthalmitis cases. Furthermore, a concentration of 1.25% PVP-I (0.1 mL) can be injected intravitreally in patients whom surgical therapy is not possible; however, further studies are required to establish efficacy and safety profile. Table 4 provides a brief review on the clinical studies evaluating the effect of povidone-iodine in the treatment of endophthalmitis.

Table 4.

Clinical studies on the effect of povidone-iodine for treatment of endophthalmitis (PVP-I: povidone-iodine; BSS: balanced salt solution; AC: anterior chamber; VA: visual acuity)

Author/Year	Disease	Design	Population	Management	Outcomes	Comments
Nakashizuka et al. 2015	Postop endophthalmitis	Case series	4 eyes	Vitrectomy using 0.025% PVP-I–BSS PLUS as irrigation solution Then postop intravitreal and intravenous antibiotics	In all 4 eyes, endophthalmitis was resolved with no adverse events.	Coagulase-negative Staphylococcus sp. was isolated in 1 eye at the beginning but not at completion of surgery. Ocular toxicity was not observed.
Otani et al. 2017	Postop endophthalmitis	Case report	1 eye	Capsular bag irrigation with 0.025% PVP-I–BSS PLUS	Endophthalmitis was resolved and there were no adverse events.	Gram-negative rods were detected from AC fluid. Final visual acuity was 24/20.
Nakashizuka et al. 2019	Postop or endogenous endophthalmitis	Prospective case series	8 patients	Intravitreal injection of 0.1 mL/1.25% PVP-I, followed by AC irrigation and vitrectomy using 0.025% PVP-I	In all but case 7, endophthalmitis resolved rapidly and good VA was maintained.	No adverse events were noted. The periop ERG showed improvements in the oscillatory potential amplitudes and the implicit time of the a-wave after therapy, suggesting functional recovery in the retinal outer and inner layers.
Kitagwa et al. 2020	Bleb-related endophthalmitis	Case report	2 eyes	Combined IOL implantation and vitrectomy using BSS Plus containing 0.025% PVP-I irrigation fluid	No ocular adverse events and no recurrence of endophthalmitis in both cases.
Tanaka et al. 2020	Endogenous endophthalmitis	Case report	1 eye	Intravitreal injection of 0.1 ml of 1.25% PVP-I	The inflammation rapidly diminished and the hypopyon disappeared 4 days after. The fundus became visible after 7 days. No recurrence of endophthalmitis.	The VA in her right eye recovered to that in the left eye (0.2).

6. Adverse effects and toxicity

PVP-I is an effective skin disinfectant and antiseptic often used for preoperative preparations. However, usage for surgical wounds limits fibroblastic activity, wound cellularity, and migration of polymorphonuclear leukocytes. Furthermore, this toxicity depends largely on the concentration. Low concentration solution of 1% PVP-I didn’t affect the skin differently than saline solution.⁹⁹ True allergies to iodine are rare and most adverse reactions are due to direct toxicity, particularly in repeated procedures.³⁹ It has also been reported that after PVP-I application on the skin, patients developed contact dermatitis. The prevalence of contact dermatitis is estimated to be around 1%. Moreover, in patients with a history of allergic reactions to other iodine-containing substances like contrast media, shellfish or seafood, PVP-I can still be safely used because of minimal evidence of cross-reactivity.⁵⁸

PVP-I is a widely used antiseptic agent in ophthalmic surgeries. However, it has been associated with complications such as postoperative eye pain, persistent corneal epithelial defects, an attendant risk of keratitis, and retinal toxicity in cases of intravitreal injection.⁵⁸ Additionally, PVP-I has been associated with chemical burns, cytotoxic effects, and general patient discomfort. Common side effects of PVP-I include mild irritation at the site of application and temporary yellow discoloration of the whites of the eyes.⁵⁸ The use of PVP-I can cause intraocular inflammation. The irrigation of the conjunctival sac before surgery may leak into the anterior chamber during surgery. This can lead to endothelial cell death, corneal decompensation with marked edema and increased intraocular pressure.²⁵ On the other hand, injection of PVP-I can be used to treat endophthalmitis. In this regard, a study showed significant improvement of conjunctival hyperemia, chemosis and vitreous inflammation where a repeated dose of 0.1% was more effective than a single dose of 0.3%.⁴⁴

6.1. Conjunctival toxicity

PVP-I is applied for conjunctival lavage before procedure to prevent endophthalmitis. In one study 10% of subjects experienced mild conjunctival discomfort after treatment with 5% PVP-I.⁵⁸ PVP-I has a pH of about 4 that could be linked to various negative effects on the ocular surface. These effects include a decrease in goblet cell density, an increase in conjunctival epithelial cell size, and squamous metaplasia of the conjunctiva. These result in higher tear film osmolality, more severe corneal staining, and increased ocular surface related symptoms. The severity of these effects is directly related to the duration of ocular exposure to PVP-I. These adverse effects require additional attention especially for dry eye patients undergoing ocular surgery.⁴⁵ Decreasing unnecessary contact time may be the best solution for decreasing toxicity, however, phosphate solutions and liposomal formulations are other applicable measures to decrease toxicity.^93,103 Moreover, although a study found no harm in adding extra PVP-I during donor tissue recovery,⁷⁶ Swift and coworkers showed povidone-iodine's adverse effects on conjunctival cells, indicating potential harm.⁹⁰

6.2. Corneal toxicity

6.2.1. Corneal epithelial toxicity

Central toxic keratopathy (CTK) is characterized as an uncommon, self-limiting, non-inflammatory postsurgical condition. It manifests with central corneal opacity and a notable hyperopic shift. It has been hypothesized that the photoactivation of povidone-iodine by the excimer laser might trigger a toxic response, initiating a sequence of events that ultimately result in the development of CTK.⁵⁶ In the COVID era, we encountered a surge in CTK cases from the use of PVP-I without complete washout during refractive surgery procedures in some centers.

6.2.2. Corneal stromal toxicity

PVP-I can induce stromal toxicity in the form of corneal stromal edema. In donor cornea, penetration of free iodine was found up to the mid-stroma after application of greater than 5% PVP-I; however, corneal fibroblast damage was found after exposure to a concentration of only 0.25% for 2 minutes in vitro.⁶⁷ Recently, we documented that a second time application of 5 minutes of PVP-I 5% during donor preparation resulted in a slight increase of 10% in donor tissues with mild edema, but there was no change in tissues with moderate or severe edema. This change, however, resulted in a significant reduction in the incidence of infectious post-keratoplasty events.⁷⁶

6.2.3. Corneal endothelial toxicity

Shimada used 0.25% povidone-iodine to irrigate the ocular surface every 20-30 seconds during cataract surgery without any endothelial toxicity.⁸¹ In the case of direct injection of 5% PVP-I into the anterior chamber, the effective concentration will be 0.8%, which is toxic for endothelial cells since it has been reported in the previous studies that up to 0.1% intracameral concentration of PVP-I is not toxic for endothelial cells. Hence, up to 1% PVP-I can be injected into the rabbit anterior chamber without any endothelial toxicity. Also in this setting, 5% PVP-I on the ocular surface without entry to the eye can be used without endothelial toxicity.³⁶ In our previous work on the human donor cornea, we found extremely modest decrease in the endothelial cell density after the second time of cornea wash with 5% PVP-I.⁷⁶

6.3. Retinal toxicity

Retinal toxicity primarily impacted the inner retinal layers, particularly in the inferior region (possible maximum contact) in a rabbit model. Highest non-retinotoxic vitreous concentration was associated with the injection of 0.1 mL of 0.05% or more in PVP-I intravitreal injection.⁸⁰ Concentrations of up to 0.5% PVP-I has been reported as non-toxic in rabbit models. A concentration of 0.025% PVP-I in balanced salt saline irrigation solution during vitrectomy has shown to be both bactericidal and non-toxic in human eyes.⁴⁸

7. Alternatives

In cases of poor PVP-I tolerance, it is recommended to apply PVP-I at lower concentrations.³⁹ Although several alternatives are available in this regard,⁵⁸ further research is needed to provide robust evidence regarding the efficacy profiles of these alternatives compared to PVP-I and to demonstrate comparable tolerance in intraocular procedures.

7.1. Chlorhexidine

Chlorhexidine 0.05% or 0.1% is an effective alternative to PVP-I for intravitreal injection prophylaxis, but can cause anaphylaxis, although its incidence is low.⁴⁰ A prospective randomized clinical trial compared PVP-I and aqueous chlorhexidine (AqCHX) as ocular surface antiseptics during intravitreal injections.² The study found that AqCHX lead to less immediate pain and lower corneal epitheliopathy compared to PVP-I. Both agents had similar rates of microbial reduction and adverse events. The results suggest that AqCHX may be a better-tolerated alternative to PVP-I for some patients undergoing intravitreal injections.

7.2. Polyhexamethylene biguanide

Polyhexamethylene biguanide (PHMB) is another alternative to PVP-I. It acts against Gram negative and positive bacteria, fungi, parasites, and certain high therapeutic index viruses.²¹ It has been shown that PHMB will protect human keratinocytes when they are cocultured with Staphylococcus aureus.¹⁰¹ The incidence of adverse effect did not increase with concentration, with the exception of corneal punctate keratopathy at 0.08% PHMB.⁶⁴ In comparison to PVP-I, patients reported significantly decreased pain scores.⁷¹ Additionally, 0.02% PHMB compared to 1.25% PVP-I has shown to have similar antimicrobial efficacy, as well as a longer lasting effect;²⁷ however, the risk of PHMB-induced allergic contact dermatitis explains why a commercial PHMB solution is unavailable in the USA.⁸⁹

7.3. Octenidine

Octenidine is a byspiridine that interacts with the cell membrane like antimicrobial peptides. It has a low risk of resistance due to a non-specific mechanism of action.⁶ In an in vitro setting, octenidine was strongly active against Acanthamoeba and trophozoites.²⁸ Furthermore, its use in blepharoplasty preparation prevented postoperative wound infection without antibiotic prophylaxis.⁵⁴

7.4. Picloxydine

Picloxydine is another alternative that has been shown to be effective in vitro against antibiotic susceptible and antibiotic-resistant strains isolated from conjunctiva strains.^13,42 Both octenidine and picloxydine are lacking in studies to be able to accurately determine efficacy and patient tolerability.

7.5. Ozone

Ozone has been studied for its antiseptic and anti-inflammatory properties.⁷⁰ A formula has been developed and tested in vitro, which showed a growth inhibition of Staphylococcus epidermidis and Escherichia coli after 24 h, and Staphylococcus aureus and Pseudomonas aeruginosa after 7 days. The formula was composed of ozonized oil 0.5% in liposomes and hypromellose.¹⁴ A study investigated the microbial load after treatment with this formula for cataract surgery, which showed a reduction of more than 90%.⁸⁵ Hence this alternative seems to be a promising option for ophthalmic surgery preparation.

7.6. Hypochlorous acid

Another possible alternative is hypochlorous acid (HOCL), which has a natural bactericidal property due to being generated as a portion of cytotoxic myeloperoxidase. It is also nontoxic to the ocular surface, which could make it a promising alternative.⁴⁷ Kanclerz and coworkers showed that HOCL compared to 10% PVP-I has a significantly higher number of positive conjunctival swabs.³⁸

7.7. Biosecur

Lastly, biosecur, an organic non-toxic alcohol-free compound, is also being studied as an alternative. It contains citrus bioflavonoids suspended in glycerin. It has been tested in vitro in a liposomal commercial formula with 0.2% biosecur and found that there were bactericidal and fungicidal effects against microorganisms at 1:2 to 1:16 dilutions.⁵⁵ Additionally, it was tested in vivo on patients undergoing intravitreal injections and after 4 days a significant decrease in bacterial load was detected.⁹⁸

8. Conclusions

PVP-I is a widely used antiseptic agent in ophthalmology owing to its broad spectrum of microbicidal activity and strong evidence supporting its efficacy. It has been shown to reduce effectively the risk of endophthalmitis in various ocular procedures, including cataract surgery and intravitreal injections. PVP-I has also been used in the treatment of conjunctivitis, keratitis, and endophthalmitis with promising results; however, PVP-I has been associated with complications such as postoperative eye pain, persistent corneal epithelial defects, and an attendant risk of keratitis. In cases of suspected toxicity, applying PVP-I at lower concentrations and, in cases of suspected allergy, alternatively usage of chlorhexidine and polyhexamethylene biguanide is advisable since these agents have been better studied and yield more promising results.

9. Methods of literature search

A PubMed and Google scholar search was performed to identify relevant articles (with no time restrictions) using the search terms: “povidone-iodine” or “polyvinyl pyrrolidone” in combination with “eye” or “ophthalmology” or “ocular” or “endophthalmitis” or “keratitis” or “conjunctivitis” or “alternative” or “adverse effect”. Also, reference list of eligible articles was explored for relevant articles. Searches were performed on April 1, 2023. Articles were excluded if they were not referenced in English.