Postoperative endophthalmitis (POE) is a severe intraocular infection that occurs when microorganisms enter the eye after cataract surgery. Inflammation spreads to both the anterior and posterior segments, potentially causing irreversible vision loss.
Cataract surgery is the most frequently performed ophthalmic procedure, with 3.7 million cases annually in the United States, 7 million in Europe, and 20 million worldwide as of 2021 1). In Japan, the incidence of postoperative endophthalmitis is approximately 0.025–0.052% and continues to decline. In the United States, it decreased from 0.327% in the 1970s to 0.087% in the 1990s, then temporarily increased to 0.265% in 2003 with the adoption of clear corneal incisions, before dropping to 0.04% in 2013–2017 3). 75% of acute postoperative endophthalmitis cases occur within one week after surgery.
Infection prevention consists of the following three-step approach:
In Japan, the incidence of endophthalmitis after cataract surgery is approximately 0.025–0.052%. Although rare, it can lead to blindness, so thorough preventive measures are essential. In recent years, the incidence has been further decreasing due to the widespread adoption of prevention protocols.
When postoperative endophthalmitis develops, the following symptoms may occur.
Bacterial endophthalmitis is broadly classified into three types based on the time of onset.
Acute type
Onset: Within 1 week after surgery
Causative bacteria: Highly virulent bacteria such as Staphylococcus aureus, Enterococcus, and Pseudomonas aeruginosa
Findings: Severe anterior chamber inflammation, hypopyon, vitreous opacity. Symptoms are severe and progress rapidly.
Subacute type
Onset: Approximately 2 weeks after surgery
Causative bacteria: Low-virulence bacteria such as Staphylococcus epidermidis
Findings: Subacute course. Anterior chamber inflammation is milder than the acute type.
Delayed-onset type
Onset: Several weeks to months after surgery
Causative bacteria: Cutibacterium acnes, Corynebacterium species
Findings: A characteristic white plaque is often observed on the lens capsule. Symptoms are mild.
Characteristic ocular findings are as follows:
The most common route is intraoperative introduction of bacteria attached to the intraocular lens (IOL) into the eye. Postoperatively, when wound closure is insufficient, bacteria from the ocular surface can reflux into the anterior chamber. The patient’s own conjunctival and eyelid margin flora are the main source of infection 1).
Gram-positive bacteria account for the majority (94.2%) of postoperative endophthalmitis cases 1).
| Causative Organism | Characteristics | Notes |
|---|---|---|
| CNS (e.g., Staphylococcus epidermidis) | Most common (approx. 70%) | Low to moderate virulence |
| Staphylococcus aureus | Common in acute type | Increasing MRSA |
| Enterococcus | Poor prognosis | Cephems are ineffective |
Other causative organisms include Streptococcus species and Gram-negative rods (e.g., Pseudomonas aeruginosa). In late-onset cases, Propionibacterium acnes is common.
Most cases of acute endophthalmitis develop within one week after surgery. However, subacute and delayed types also exist, so regular follow-up for several weeks to months postoperatively is important. If symptoms such as eye pain or vision loss appear, immediate medical attention should be sought.
The diagnosis of endophthalmitis is based on clinical findings.
Identification of the causative organism constitutes the definitive diagnosis. Aqueous humor or vitreous fluid is collected and submitted for bacteriological examination. Although the identification rate is not necessarily high, it is essential for selecting susceptible antibiotics and differentiating from sterile endophthalmitis, so it must always be performed.
Endophthalmitis prevention in cataract surgery involves a combination of multiple measures.
Preoperative disinfection with povidone-iodine (PVP-I) is the most evidence-based preventive method 1). PVP-I exhibits broad-spectrum bactericidal activity against gram-positive bacteria, gram-negative bacteria, fungi, viruses, and protozoa, and is unlikely to induce drug resistance 1). Since a randomized controlled trial at the New York Eye and Ear Infirmary in 1991 confirmed that preoperative application of PVP-I reduces the incidence of postoperative endophthalmitis, PVP-I has become the standard for surgical field disinfection 2).
The bactericidal activity of PVP-I depends on the amount of free iodine, not the concentration of the drop. There is a “bell-shaped phenomenon” where dilution initially increases the amount of free molecular iodine, resulting in superior bactericidal activity at low concentrations 2).
| Concentration | Free iodine | Characteristics |
|---|---|---|
| 0.1% | 24 ppm (maximum) | Maximum bactericidal activity, effect onset within 15 seconds |
| 1% | 13 ppm | Low tissue toxicity |
| 10% | 5 ppm | For skin and periorbital area; brief contact with cornea |
Implementation in Japan:
Note: Prior use of lidocaine gel before surgery has been shown to reduce the bactericidal effect of PVP-I2). It is preferable to administer topical anesthesia after PVP-I instillation.
“Iodine allergy” is a frequently reported condition in clinical practice, but many cases are based on misunderstanding. Iodine is a component of thyroid hormones and amino acids, and is an essential element for the human body, so a true allergy to elemental iodine is biologically unlikely3).
Contrast media reactions and shellfish allergies have been confused with “iodine allergy,” but the causative agent in shellfish allergy is tropomyosin, and contrast media reactions are primarily due to osmotic pressure and non-immunologic mediator release, with iodine itself not directly involved3). True allergy to PVP-I (IgE-dependent anaphylaxis) is extremely rare, and the burning sensation and hyperemia after PVP-I application are mostly dose-dependent chemical irritant reactions3).
In a group of anti-VEGF injection patients who avoided PVP-I solely due to self-reported “iodine allergy,” the incidence of endophthalmitis was as high as 9.4%, and none of these patients who later received PVP-I experienced an allergic reaction3). Avoiding PVP-I based solely on self-report is dangerous.
If PVP-I allergy is confirmed, chlorhexidine (0.02–0.05%) can be used as an alternative. However, chlorhexidine is toxic to the ocular surface and may cause irreversible keratitis, so careful concentration management is necessary 2).
Indocyanine green (ICG) contains iodine and is contraindicated in patients with a history of iodine hypersensitivity. ICG angiography has been reported to cause shock (0.1%), nausea, vomiting, and urticaria (0.1% to less than 5%) 3).
Preoperative topical antibiotics are an adjunctive measure to reduce resident bacteria in the conjunctival sac 6). However, the evidence for directly reducing the incidence of endophthalmitis is not as strong as that for povidone-iodine disinfection or intracameral antibiotic administration. The use and duration of administration should be determined based on patient risk, wound conditions, resistance risk, and facility protocols.
Intracameral (IC) antibiotic administration at the end of surgery is a prophylactic method that directly controls residual bacteria in the anterior chamber. It is based on povidone-iodine disinfection and wound closure management, and additionally reduces the risk of endophthalmitis.
Cefuroxime
Mechanism of action: Inhibition of cell wall synthesis
Bactericidal property: Time-dependent
Typical dose: 1.0 mg/0.1 mL
Characteristics: Single-use preparations such as Aprokam are available. Less effective against MRSA and enterococci.
Moxifloxacin
Mechanism of action: Topoisomerase II/IV inhibition
Bactericidal property: Concentration-dependent
Typical dose: 150–500 μg/0.1 mL
Features: Broad spectrum against Gram-positive and Gram-negative bacteria, including activity against Pseudomonas species.
Vancomycin
Mechanism of action: Cell wall synthesis inhibition
Target: MRSA, MRSE, etc.
Position: Important as a therapeutic agent.
Caution: Prophylactic intracameral administration raises concerns about association with HORV.
In the ESCRS prospective RCT, the risk of endophthalmitis was higher in the group not receiving intracameral cefuroxime 1 mg/0.1 mL (OR 4.92; 95% CI 1.87–12.9)2). In a pooled analysis of 17 studies involving approximately 900,000 eyes, the use of intracameral antibiotics significantly reduced the risk of endophthalmitis (OR 0.20; 95% CI 0.13–0.32)9).
In the same meta-analysis, the weighted mean incidence of postoperative endophthalmitis was 0.0332% for cefuroxime, 0.0153% for moxifloxacin, and 0.0106% for vancomycin9). However, direct comparisons between drugs are complicated by differences in study backgrounds. The OR for cefuroxime was reported as 0.29–0.30, and for moxifloxacin as 0.26–0.292).
For moxifloxacin, there are also reports supporting its efficacy and safety. A 2026 RCT meta-analysis examined its effectiveness in preventing endophthalmitis and its safety regarding corneal endothelium and central corneal thickness10). Even in high-risk cases such as posterior capsule rupture, considering intracameral administration is meaningful.
On the other hand, intracameral antibiotics alone do not replace all preventive measures. The decision should be made comprehensively based on surgical field disinfection, wound closure, reduction of ocular surface bacterial load, and patient background.
The efficacy of intracameral administration is strongly supported. However, it is not a uniformly mandatory procedure at all facilities. The choice should be made based on drug availability, preparation system, patient risk, and wound conditions. It should be actively considered in high-risk cases such as posterior capsule rupture.
At present, clear superiority of one over the other has not been established. Cefuroxime has the advantage of RCTs and long-term use experience. Moxifloxacin is characterized by broad spectrum and concentration-dependent bactericidal activity. The choice depends on allergies, drug preparation, antibacterial spectrum, and facility protocol.
Intracameral administration is a method of delivering a high concentration of antimicrobial agent directly into the anterior chamber. The turnover time of aqueous humor is 2 to 4 hours, and the injected drug is diluted and eliminated in a relatively short time 3).
Moxifloxacin is concentration-dependent, with bactericidal activity increasing at concentrations well above the MIC. Cefuroxime is time-dependent, and the time above the MIC is important.
Dosage and injection volume are directly related to safety. A model suggests that injection of 0.1 mL of 0.5% moxifloxacin and flushing with 0.5 mL of 0.15% moxifloxacin result in similar intracameral retention times 11). A method of diluting Vigamox 0.5% with BSS and using it as 150 μg/0.1 mL has also been reported 12). Standardize the procedure at each facility and double-check the concentration and injection volume.
Postoperative antibiotic eye drops are used to reduce the bacterial load on the ocular surface until wound closure. When intracameral antibiotics are used, the additional reduction in endophthalmitis incidence with postoperative eye drops is not clear 4). On the other hand, there are reports of reducing conjunctival sac bacterial counts at 1 week postoperatively, and attention should be paid to the selection of highly resistant bacteria with levofloxacin 5).
With the increased use of corneal incisions, some reports indicate a higher risk of endophthalmitis compared to scleral tunnel incisions. The following measures are important.
With intracameral administration, the drug directly contacts the corneal endothelium. In safety evaluation, corneal endothelial cell density and central corneal thickness are important.
Standard-dose cefuroxime 1 mg/0.1 mL is not considered to have significant adverse effects on the corneal endothelium 7). A meta-analysis of RCTs on moxifloxacin also found no significant differences in endothelial cell count or central corneal thickness 10).
On the other hand, toxicity becomes an issue with overdose or preparation errors. Corneal edema and decreased endothelial cell density have been reported with erroneous administration of cefuroxime 12.5 mg/0.1 mL 8). For cefuroxime, TASS, macular edema, and retinal infarction have also been reported 9).
Moxifloxacin is easy to use broadly, but overdose or improper preparation can cause TASS. Toxicity from high-concentration exposure has also been reported in cultured corneal endothelial cells, making concentration management important.
Vancomycin is an important drug for treating MRSA and MRSE. However, prophylactic intracameral administration has been reported to be associated with hemorrhagic occlusive retinal vasculitis (HORV) 3). Routine use for endophthalmitis prophylaxis should be avoided.
Treatment is broadly classified according to the stage of the disease.
This is the stage where inflammation in the anterior chamber is early and has not progressed to hypopyon. In acute or subacute onset, increase the frequency of antibiotic eye drops and observe daily while monitoring for signs of infection.
If hypopyon is present but there is no vitreous opacity, perform anterior chamber irrigation and intravitreal injection of antibiotics.
Conservative treatment (frequent eye drops, subconjunctival injection) is less effective in the hypopyon stage.
When inflammation spreads to the vitreous, vitrectomy is generally indicated. In Japan, early vitrectomy is commonly performed. In addition to vitrectomy, intraocular lens capsule irrigation, posterior capsule removal, and intravitreal antibiotic injection are combined. If a bacterial mass is found in the capsule, the intraocular lens is removed or exchanged.
| Irrigation solution | Concentration | Usage |
|---|---|---|
| Vancomycin | 20 μg/mL | Add to irrigation solution |
| Ceftazidime | 40 μg/mL | Mixed into irrigation fluid |
Postoperatively, frequent instillation (every hour) of vancomycin 1% eye drops and ceftazidime 2% eye drops, along with intravenous imipenem (Tienam®) 0.5–1.0 g twice daily.
Intracameral antibiotics are an effective preventive method, but they are not sufficient alone. Preoperative disinfection with povidone-iodine and appropriate surgical technique form the foundation of prevention, and intracameral antibiotics provide additional benefit when used in conjunction with these measures. A comprehensive approach combining multiple strategies is important.
The development of postoperative endophthalmitis involves the entry of pathogens into the eye during or shortly after surgery. The three-stage model for endophthalmitis prevention proposed by ESCRS is as follows 2).
In recent years, antimicrobial-resistant bacteria (AMR) have become a problem. According to the ARMOR study, 39% of coagulase-negative staphylococci and 59% of MRSA show multidrug resistance1). The four major causative organisms are CNS, enterococci, MRSA, and Propionibacterium acnes.
Effective antimicrobial agents vary depending on the causative organism.
薬剤耐性菌の増加に伴い、消毒薬を中心とした予防戦略が重視されている1)。
Grzybowskiら（2025）のレビューでは、ポビドンヨード（PVP-I）がグラム陽性菌・グラム陰性菌・真菌・ウイルス・原虫に対して広域の殺菌活性を持ち、薬剤耐性を生じにくい点から予防的消毒のゴールドスタンダードとされた1)。PVP-Iはポリビニルピロリドン（PVP）とヨウ素の複合体であり、溶液中の遊離ヨウ素が細菌の細胞膜の飽和脂肪酸に結合して水素イオンと置換し、細胞膜に細孔を形成して細菌細胞死を誘発する2)。PVPは細菌細胞膜への親和性により、ヨウ素を標的に効率よく送達する役割を担う。PVP-Iに対する微生物耐性は現在まで報告されていない2)。
主な消毒薬の特性は以下の通りである。
The “Shimada Technique,” an intraoperative PVP-I irrigation method developed by Shimada and Nakashizuka, involves repeatedly irrigating the ocular surface with 0.25% PVP-I every 20 to 30 seconds during surgery 2). It has been reported to significantly reduce the rate of anterior chamber bacterial contamination without damaging the corneal endothelium (p = 0.0017 compared to the saline group), and is attracting attention as a new standard for continuous intraoperative disinfection 2).
In addition to conventional PVP-I, several new disinfectants are being investigated.
In the OPERA trial by Grassi et al. (2024), combining liposomal ozonated oil (LOO) 0.5% and liposomal foam with PVP-I 5% resulted in a greater reduction in microbial load compared to PVP-I 5% alone (72.31% vs 50.26% on chocolate agar, p<0.0001)13). LOO has antibacterial, anti-inflammatory, and tissue repair-promoting effects.
In the RCT by Romano et al. (2024), 0.02% chlorhexidine showed a greater reduction in bacterial load and less patient discomfort compared to 0.6% PVP-I14).
In the field of intravitreal injections, protocols that omit postoperative antibiotic eye drops and focus on antiseptics are becoming widespread1). Even in cataract surgery, the necessity of postoperative eye drops when using intracameral antibiotics is being reconsidered4). However, preoperative and postoperative eye drops also serve as an adjunct to reduce ocular surface bacterial load, and should not be simply deemed unnecessary5)6).
The role of antibiotic eye drops needs to be evaluated based on the balance of endophthalmitis incidence, ocular surface bacterial load, adherence to eye drops, and drug resistance. The advantages of antiseptics, which are less likely to promote drug resistance, are also attracting attention1)5).
To reduce the burden of eye drops after cataract surgery, a drop-free strategy centered on intracameral antibiotics is being considered. Some reports show that intracameral administration alone has an infection rate equivalent to that of combined eye drops4)9). However, this does not deny the effect of preoperative and postoperative eye drops in reducing the bacterial load on the ocular surface.
Intravitreal combination formulations such as Tri-Moxi (triamcinolone + moxifloxacin) and Tri-Moxi-Vanc are also being investigated. They have the advantage of using the vitreous cavity as a drug reservoir. However, formulations containing vancomycin pose a risk of HORV. Large-scale randomized trials are also lacking.
Some studies show that intracameral administration alone maintains a low infection rate. However, there is insufficient evidence that eye drops can be omitted in all cases. Decisions should be made based on the burden of eye drops, wound condition, infection risk, and risk of resistant bacteria.
