
Endophthalmitis is a detrimental surgical complication that often leads to irreversible visual impairment. The recommended treatment of choice for this condition is intravitreal antibiotics [1]. However, due to multidrug-resistant pathogens, the treatment may not be effective, leading to poor visual outcomes. Povidone–iodine (PI), on the other hand, shows action against bacteria, fungi, and viruses [2,3] without development of resistance, and it does not induce either new or cross-resistance to antibiotics [4].
PI is widely used externally as a disinfectant and antiseptic agent for preoperative preparation of the skin and mucous membranes before ophthalmic surgeries. Recognizing the need for an effective agent that does not encourage resistance with continued use, PI has been intravitreally used in endophthalmitis patients, leading to promising results. It was first used intravitreally in humans by Nakashizuka et al. [5] who, in 2015, mixed 0.025% PI in a balanced salt solution for vitreous irrigation in four eyes and achieved complete resolution of post-operative endophthalmitis with no complications. In 2019, the same group reported treating eight eyes with endophthalmitis with intravitreal injection of 1.25% PI followed by pars plana vitrectomy (PPV) using PI irrigation [6]. More reports of endophthalmitis successfully treated with a single intravitreal PI (IPI) injection have been published [7,8]. In addition, two post-traumatic endophthalmitis cases due to intraocular foreign bodies were successfully treated with 0.025% PI irrigation in China [9]. These reports show resolution of endophthalmitis and demonstrate the treatment to be safe in terms of visual acuity, electroretinogram outcomes, visual field defects, and endothelial cell density (ECD).
There are limited documented reports of IPI injection for endophthalmitis. We report the results of 0.625%/0.1 mL of IPI injection (0.013% PI vitreous concentration) as treatment for acute endophthalmitis (AE). Based on in vivo studies, the effective and safe concentration range of PI in the vitreous is estimated from 0.013%–0.025% [5,6,10,11]. We adopted the lowest effective concentration within this range, which was 0.013%, corresponding to 0.625% per 0.1 mL of PI.
We referred to the IPI injection preparation method presented by the Nakashizuka group to prepare 1.25%/0.1 mL of PI and diluted it by half [6]. We drew 0.05 mL of 10% PI (Povidine ®; Firson) into a 1-mL syringe, along with 0.75 mL of normal saline. The solution was mixed thoroughly to achieve a uniform concentration, and then 0.1 mL of the solution was used for IPI injection. The final concentration was 0.625%.
This study was approved by the appropriate institutional review board and adheres to the Declaration of Helsinki. Due to the retrospective nature of the study, informed consent was not needed.
Patient 1 was a 77-year-old woman who visited for a postoperative day 1 checkup after a non-complicated simple cataract surgery. Descemet’s membrane folds were identified in her left eye, with an anterior chamber (AC) cell grade of 3+ with cyclitic membrane. She presented with pain and headache. Only slight conjunctival injection was noted. Her best-corrected visual acuity (BCVA) was 0.6, and her intraocular pressure was 11 mmHg. After two days, her BCVA had decreased to 0.01, and her AC cell grade had increased to 4+, alongside a grade 1 hypopyon (Fig. 1A). Immediate IPI injection was performed and, 4 h later, emergency PPV was performed, along with AC irrigation and intravitreal injection of antibiotics. Her pain and infection improved; on postoperative day 7, the BCVA of her left eye was 0.4, her AC cell grade had decreased to 1+, with no hypopyon, although some cyclitic membranes remained. On postoperative day 28, the BCVA of her left eye was 0.8, with near-complete resolution of inflammation (Fig. 1B). Complete resolution was observed on postoperative day 60 and persisted until her last follow-up at 10 months. The patient’s central macular thickness (CMT) before cataract surgery was 257 μm, which decreased to 251 μm post-PPV. Her ECD before cataract surgery was 2,433 cells/mm2, increasing to was 2,513 cells/mm2 after PPV (Fig. 2A, B).
Patient 2 was an 85-year-old woman referred from a local clinic one month after a non-complicated simple cataract operation on her right eye. She experienced persistent inflammation after surgery despite continued use of moxifloxacin and prednisolone acetate eye drops. Her BCVA was 0.02, and her intraocular pressure was 9 mmHg. Slit-lamp examination revealed an AC cell grade of 3+, with an edematous cornea, Descemet’s membrane folds, and keratoprecipitates on the endothelium (Fig. 1C). There was no hypopyon observed. Fundus examination revealed a vitreous cell grade of 3+, vitreous opacity, and a small retinal hemorrhage temporal to the macula. Optical coherence tomography revealed some soft drusen on the macula. Although the patient did not report pain, the chronic nature and inflammation in the AC and the vitreous enticed a prompt action of IPI injection and PPV with intravitreal injection of antibiotics. On postoperative day 7, her BCVA was 0.2 and her AC cell and vitreous cell grades had decreased to 1+. On postoperative day 28, her BCVA was 0.5, and no sign of inflammation was observed in the AC or vitreous (Fig. 1D). The retinal hemorrhage initially observed was resolved eight months after surgery. The patient’s right eye remained inflammation-free, with a BCVA of 0.6, through the last follow-up at eight months. Her CMT before cataract surgery was 265 μm and her CMT post-PPV was 245 μm, with respective ECD of 2,398 cells/mm2 and 2,101 cells/mm2 (Fig. 2C, D).
Patient 3 was a 76-year-old man being treating for diabetic macular edema and proliferative diabetic maculopathy in both eyes. Following treatment with two intravitreal bevacizumab injections for macular edema, he underwent cataract surgery on his left eye. His preoperative cataract grade was C4N3P1. At the time of surgery, the patient’s macula was free of edema, and a simple, non-complicated cataract extraction with intraocular lens implantation within the capsular bag was performed. However, the next day, slit-lamp examination revealed an AC cell grade of 4+, many fibrotic membranes, and grade 1 hypopyon with edematous cornea (Fig. 1E). The conjunctiva was also severely infected, his BCVA was only light perception–positive, and the fundus was hazy due to AC reaction. An immediate IPI injection was performed, followed by vitrectomy, AC irrigation, and intravitreal injection of antibiotics. The interval between IPI injection and vitrectomy was 24 h due to the patient’s schedule. On postoperative day 12, the patient’s AC cell grade had decreased to 1+, there was no infection in the conjunctiva, and his BCVA had improved to 0.15 (Fig. 1F). The patient frequently missed subsequent appointments and developed neovascular glaucoma, leading to aggravation of visual acuity. Finally, at the last follow-up 16 months after surgery, he reported a sudden decrease in vision to light perception– negative, although there was no recurrence of endophthalmitis. This patient’s CMT before cataract surgery was 145 μm, and his CMT post-PPV was 142 μm, while his ECD before cataract surgery was 3,125 cells/mm2 and his ECD after PPV was 2,625 cells/mm2 (Fig. 2E, F).
Patient 4 was an 81-year-old woman who was referred from a local clinic two days after a non-complicated simple cataract operation on the right eye. Her BCVA was 0.01, and her intraocular pressure was 16 mmHg. Slit-lamp examination revealed an edematous cornea with Descemet’s membrane folds, an AC cell grade of 4+, and a grade 1 hypopyon (Fig. 1G). Her fundus was not visible due to an AC reaction (Fig. 1G). Immediate IPI injection was performed; after 3 h, vitrectomy, AC irrigation, and intravitreal injection of antibiotics were completed. Unlike the aforementioned patients, who had negative cultures, intravitreal sampling in this patient yielded growth of Achromobacter xylosoxidans. On postoperative day 7, her BCVA was 0.03, her AC cell grade was 2+, and the hypopyon had resolved. Her corneal edema had improved as well, and the AC cell grade continued to gradually decrease (Fig. 1H). The patient’s eye condition appeared to be stable until postoperative day 30, when her AC cell grade was 3+. B-scan ultrasonography revealed vitreous opacity, and her BCVA was measured as finger count. Immediate PPV, AC irrigation, and intravitreal injection of antibiotics were completed to resolve the vitreous opacity. Initially, the inflammation appeared to be controlled; however, one month postoperatively, the patient experienced a recurrence of abrupt AC inflammation with a grade 1 hypopyon. The patient’s BCVA was 0.01. Considering that there was a microorganism growth from culture and recurrent endophthalmitis, the surgeon decided to remove the intraocular lens. A third vitrectomy was conducted with intraocular lens removal and silicon oil injection, followed by intravitreal antibiotics injection. The inflammation resolved completely after seven days and remained stable until 14 months after the first operation, with a BCVA of 0.2. Finally, silicon oil removal and intraocular lens scleral fixation were performed, and her final BCVA was 1.0. The patient’s CMT before cataract surgery was 229 μm and her CMT post-PPV was 243 μm, while her ECD before cataract surgery was 2,725 cells/mm2 and her ECD after PPV was 2,551 cells/mm2 (Fig. 2G, H).
The four patients in this series were diagnosed with AE early and were promptly treated with IPI injection and vitrectomy. All cases achieved complete resolution of inflammation, and no definite damage caused by IPI injection was apparent. Measured CMT and ECD values did not show significant changes. These are the only indices of toxicity and relatively simple examinations that reflect retinal and corneal damage; thus, these examinations have been used in previous PI evaluations [10]. Other evaluation indices used in other studies include electroretinography and Goldman perimetry, but the data necessary for these evaluations were not collected in all cases at the time of treatment [6]. The visual prognosis was poor in case 3 due to the development of neovascularization; however, this patient had underlying proliferative diabetic retinopathy with poor compliance, and it appears that his vision deteriorated rapidly due to the development of neovascularization at the iris due to changes in the vascular endothelial growth factor concentration after vitrectomy [12]. Case 4 experienced two recurrences. Achromobacter xylosoxidans, a rare, slow-growing but serious bacterium, was identified in the aqueous culture. Studies have recommended complete capsulectomy and intraocular lens removal for patients with recurrent and recalcitrant endophthalmitis associated with Achromobacter xylosoxidans [13,14]. One report detailed two recurrences of endophthalmitis, as in the current case, but final visual acuity was measured as 0.8 after intraocular lens removal and injection of antibiotics [14]. In our case study, the culture results became available three weeks after the initial surgery, but intraocular lens removal was performed only two months after the first surgery. A second operation was performed after confirming the growth of bacterium and the recurrence of endophthalmitis. Referring to previous studies [13,14], we proceeded with the reoperation to remove the intraocular lens and capsule.
IPI injection has emerged as a potential treatment for endophthalmitis. While good effectiveness of IPI is crucial, ensuring the proper concentration is equally important to avoid ocular toxicity. In humans, Nakashizuka et al. reported successful treatment of endophthalmitis with minimal retinal damage using 0.125% intravitreal PVI injection (0.025% intravitreal concentration), a meticulously calculated safe and effective concentration [5,6]. The patients experienced no complications as reflected by visual field examination, electroretinogram results, and ECD measurement. Based on in vivo studies, an effective yet non-toxic PI concentration in the vitreous is estimated to range from 0.013%–0.025% [5]. Given the limited data available on safety and efficacy, this study focused on minimizing side effects and employed the lowest effective concentration, which was 0.013% PI delivered intravitreally, corresponding to 0.625%/0.1 mL of PI. The method involving a 1.25% IPI injection demonstrated effective treatment of endophthalmitis [6]. However, concerns regarding ocular toxicity associated with IPI use remain. Thus, efficacy and the potential risk of side effects should both be considered. A lower concentration of IPI (0.625%) may be a viable alternative in such scenarios. This potentially reduces the risk of retinal toxicity while maintaining efficacy.
This case series cannot fully evaluate the effect of IPI injection because the surgery was performed immediately after IPI injection. To properly evaluate this effect, PI treatment must be conducted independently. After proving the safety and effectiveness of this method, it should then be performed on its own.
Here, a 0.625%/0.1-mL IPI injection followed by vitrectomy was performed in four AE patients after successful cataract surgery. Further large-scale controlled studies are needed to establish the effect of this treatment option.
A part of this paper was presented at the 130th annual meeting of Korean Ophthalmology Society.
The authors declare no conflicts of interest relevant to this article.
Conception (H.S.P.); Design (Y.J.K.); Data acquisition (M.Y.C.); Interpretation (M.Y.C., Y.J.K.); Writing (M.Y.C.); Review (All authors); Final approval of the article (All authors)
![]() |
![]() |