For antiseptic prophylaxis in cataract surgery, the European Society of Cataract and Refractive Surgeons (ESCRS) recommends povidone-iodine and, in patients with a povidone-iodine allergy, 0.02% chlorhexidine is listed as an alternative. Conversely, the American Academy of Ophthalmology (AAO) does not recommend chlorhexidine because of ocular-surface toxicity and the risk of irreversible keratitis and does not provide further recommendations for alternatives. [1], [2]
Discussions in relevant review articles suggest aqueous chlorhexidine 0.05% or 0.1% as the best-supported alternative to povidone-iodine for ocular antisepsis when allergy is suspected, while finding no evidence to support routinely replacing povidone-iodine. Most reported intolerance to povidone-iodine reflects irritation or concentration-dependent toxicity rather than true allergy; therefore, a lower povidone-iodine concentration may be considered when toxicity is the concern. However, dilute solutions may require repeated applications to maintain adequate antimicrobial activity. Omitting effective antisepsis because of a self-reported iodine allergy has been associated with endophthalmitis. [3], [4], [5]
Chlorhexidine has been used for cataract surgery, vitrectomy, and intravitreal injections, generally as an aqueous ocular-surface rinse at 0.05%, or less commonly 0.1%, with reported contact times of 30-90 seconds. In a series of 40,535 intravitreal injections using these concentrations, the endophthalmitis rate was 0.0074%, described as comparable to rates reported with povidone-iodine. In Sweden, chlorhexidine combined with intracameral cefuroxime was used in more than 2 million cataract procedures, with postoperative endophthalmitis rates around 0.020%; this outcome cannot be attributed to chlorhexidine alone because cefuroxime and improvements in surgical practice also contributed. Comparative evidence suggests similar reductions in conjunctival bacterial counts with chlorhexidine 0.02%, polyhexanide 0.02%, and povidone-iodine 5%, with more ocular pain and superficial punctate epitheliopathy after povidone-iodine. A separate tolerability comparison found median pain scores of 2 with chlorhexidine 0.5% versus 7 with povidone-iodine 4% (p<0.001), although topical anesthesia was not used and this experimental concentration differs from the 0.05%-0.1% alternatives recommended by the authors. The optimal chlorhexidine concentration and dosing regimen remain undetermined. [3], [4], [5]
Chlorhexidine’s ocular safety depends strongly on its concentration and formulation. Alcohol-containing chlorhexidine preparations and detergent-containing scrubs are unsuitable for direct ocular-surface application because they can cause severe epithelial injury and toxic keratopathy. Chlorhexidine gluconate demonstrated epithelial toxicity at concentrations above 1%, with toxicity occurring at lower concentrations when combined with benzalkonium chloride. Accidental use of chlorhexidine acetate 0.5% with cetrimide 0.1%, instead of aqueous chlorhexidine acetate 0.05%, caused marked epithelial and stromal edema; accidental intraocular exposure to chlorhexidine-containing disinfectants has also caused corneal edema and bullous keratopathy. Chlorhexidine can cause contact dermatitis and, rarely, anaphylaxis, so it does not eliminate allergic risk. The review also notes slower antimicrobial action than povidone-iodine in laboratory testing and reports of reduced microbial susceptibility to chlorhexidine. [3], [4], [5]
Evidence for other alternatives is substantially more limited. Polyhexanide 0.02% produced reductions in preoperative conjunctival bacterial counts similar to povidone-iodine 5% and chlorhexidine 0.02%, with less conjunctival injection, but evidence establishing prevention of postoperative endophthalmitis is lacking. Picloxydine 0.05%, administered from 3 days before through 5 days after intravitreal injection, eliminated baseline positive cultures similarly to tobramycin 0.3%; however, povidone-iodine was still used immediately before injection, so this does not establish picloxydine as a replacement. Hypochlorous acid 0.01% reduced Staphylococcus epidermidis counts on periocular skin by 99.5%, but organic matter rapidly inactivates it, and skin decolonization does not establish efficacy for conjunctival surgical antisepsis. [3], [4], [5]
Additionally, a 2025 systematic review and meta-analysis evaluated chlorhexidine versus povidone-iodine for antisepsis before intravitreal injections, including seven studies identified through July 2024: six observational studies and one randomized controlled trial. Chlorhexidine concentrations were primarily 0.05% or 0.1%, generally in aqueous formulations, while povidone-iodine concentrations were predominantly 5%, with one study using 10%. Four comparative studies encompassing 455,340 injections reported 129 endophthalmitis events. Presumed endophthalmitis occurred after 60/185,849 chlorhexidine injections (0.032%) versus 69/269,491 povidone-iodine injections (0.026%), without a statistically significant pooled risk difference (p= 0.58). Culture-positive endophthalmitis rates were 0.016% versus 0.009% (odds ratio, 2.03; 95% confidence interval, 0.74-5.58; p= 0.17), and culture-negative rates were 0.016% versus 0.017% (odds ratio, 0.97; 95% confidence interval, 0.45-2.06; p= 0.93). Visual acuity after treatment of endophthalmitis also did not differ significantly. Chlorhexidine generally produced less pain and corneal epitheliopathy: in the randomized trial comparing aqueous chlorhexidine 0.1% with povidone-iodine 5%, immediate mean pain scores were 0.44 versus 1.44 out of 10 (p<0.001), ocular surface staining scores were 3.10 versus 4.22 (p<0.001), and confluent corneal staining occurred in 50% versus 78% of eyes (p= 0.001). Both agents produced similar reductions in ocular-surface culture positivity. Chlorhexidine intolerance was uncommon, with only 0.05% of patients requesting a switch to povidone-iodine in one large cohort. The authors concluded that chlorhexidine may offer advantages for patients experiencing povidone-iodine discomfort or with corneal risk factors. However, evidence certainty was low to moderate, most studies were observational, and formulations and injection techniques varied; one study using alcohol-based chlorhexidine reported significantly more endophthalmitis than povidone-iodine. Thus, the absence of statistically significant pooled differences does not establish equivalent infection prevention, and the favorable tolerability findings primarily concern dilute aqueous chlorhexidine. [6]