This retrospective study evaluated the incidence of endophthalmitis following IVT injections of dexamethasone implant and anti-VEGF agents administered by vitreoretinal specialists across multiple outpatient centers within a single large retina practice from January 2019 to January 2024. Injection encounters were identified using Current Procedural Terminology (CPT) and HCPCS Level II codes corresponding to each agent (J0178, J9035, J2778, J7777, Q5128, J7312, and J0179). These structured identifiers allowed accurate categorization of all intravitreal injection types and agents administered during the study period. The exact number of participating centers and physicians could not be retrieved from the electronic dataset; however, all injections were performed by fellowship-trained retina specialists.

Incidence rates of endophthalmitis were calculated for dexamethasone implant injection and anti-VEGF injections. Chi-square test was utilized for comparing the observed and expected frequencies of endophthalmitis cases between dexamethasone implant and anti-VEGF injections. Results indicated a significant difference if p < 0.05.

The study assessed the incidence of endophthalmitis following IVI of dexamethasone implant and anti-VEGF agents. Out of 11,954 dexamethasone implant injections, 15 cases of endophthalmitis were reported, resulting in an incidence rate of 0.125%, or 1.25 occurrences per 1,000 injections. In contrast, 318,618 anti-VEGF injections yielded 106 cases, with an incidence rate of 0.033%, or 0.33 occurrences per 1,000 injections. Chi-square analysis showed a statistically significant difference (χ² = 23.45, p = 2.28534E-07), indicating a higher risk of endophthalmitis following dexamethasone implant injections compared to anti-VEGF injections. The total and yearly distributions of anti-VEGF and dexamethasone implant injections between 2019 and 2024 are summarized in Fig. 1, which illustrates both cumulative injection totals and annual trends for each agent throughout the study period.

A detailed analysis of anti-VEGF agents revealed variability in endophthalmitis incidence rates. Aflibercept evidenced 58 cases of endophthalmitis out of 173,622 injections (0.0334% of total injections); ranibizumab had 7 cases out of 41,936 injections (0.0095% of total injections); bevacizumab had 26 cases out of 79,426 injections (0.0327% of total injections). Faricimab and brolucizumab demonstrated variable incidences, with no cases reported for brolucizumab. The incidence of endophthalmitis per 1,000 injections for each agent analyzed in this study is illustrated in Fig. 3.

While prior studies have established an increased risk of endophthalmitis following intravitreal corticosteroid injections, this present study provides a focused comparison between intravitreal dexamethasone implants and anti-VEGF agents within a large multi-center practice. A U.S. study involving 406,380 injections reported a 0.13% incidence of endophthalmitis after intravitreal steroid use, corresponding to an odds ratio of 6.9 compared with anti-VEGF injections [2]. Similarly, a French national cohort, encompassing more than 3.5 million injections found higher risk with corticosteroids (0.067%) than with anti-VEGF agents (0.020%) [16].

Meta-analyses have confirmed that anti-VEGF injections alone carry low endophthalmitis rates, approximately 0.05% per injection [5]. However, few large studies have directly compared dexamethasone implants with anti-VEGF injections until recently.

Our results align with prior reports demonstrating that endophthalmitis following dexamethasone intravitreal implant is rare but slightly more frequent than after anti-VEGF injections. Stem et al. analyzed 3,593 dexamethasone implant injections and reported four cases of endophthalmitis, two being culture-proven after monoinjections (0.06% risk per injection) [25]. Rajesh et al. analyzed 6,015 injections and observed an incidence of 0.07%, further confirming the low overall risk [26]. Pancholy et al. compared 4,973 dexamethasone implant injections with more than 180,000 ranibizumab injections and found suspected endophthalmitis rates of 0.10% for dexamethasone versus 0.03% for ranibizumab, a statistically significant difference (P = .008), though culture-positive rates were similar across agents [27]. Samuelson et al. identified four cases of endophthalmitis among 3,925 dexamethasone implant injections (0.10%), with three culture-positive for gram-positive bacteria [28]. Kiristioglu et al. reviewed 3,430 dexamethasone implant injections and reported a 1.0% rate of sight-threatening, non-pharmacologic complications including cases of endophthalmitis [29]. Finally, Dhoot et al. analyzed more than 4 million intravitreal injections from the Vestrum Health Database and found suspected endophthalmitis rates of 0.107% for dexamethasone implants and 0.147% for triamcinolone. This is compared to 0.02–0.05% for anti-VEGF agents, confirming a higher relative risk associated with corticosteroid injections [30].

Taken together, these findings reinforce our observation that dexamethasone implants are associated with a slightly higher risk of post-injection endophthalmitis relative to anti-VEGF agents, yet the absolute incidence remains exceedingly low and consistent across studies.

IVI of therapeutic drugs has become one of the most common procedures in the field of ophthalmology during the past decade. At least 6 million IVIs have been carried out yearly since 2016 [17]. An estimated 8 million injections of IVIs, mostly anti-VEGF, corticosteroid, and complement inhibitors, were administered in 2023, and that number is expected to increase to 10 million by 2025 [18]. The anti-VEGF medications implicated in this study are considered first-line treatment options for retinal conditions such as diabetic macular edema, neovascular age-related macular degeneration, and retinal vein occlusion-related macular edema. With an excellent safety profile backed by a low incidence of serious complications, they are effective in improving patients’ visual acuity [19].

Although endophthalmitis occurrences following dexamethasone implants is exceedingly rare [1, 2], research has highlighted that IVT corticosteroid use is linked to a significantly elevated risk—approximately sevenfold—compared to anti-VEGF injections [3]. This increased risk is partly attributable to the procedural and mechanical differences between the two techniques. The standard anti-VEGF injection employs a 30- or 32-gauge needle, whereas corticosteroid preparations use considerably larger needles: 27- or 25-gauge for triamcinolone and approximately 22-gauge for the dexamethasone implant. Moreover, anti-VEGF administration delivers only a small volume of fluid into the vitreous cavity, while corticosteroid injections introduce a solid implant that remains in situ, potentially causing a more sustained alteration of the vitreous environment. The larger needle size may create a more substantial wound tract, facilitating bacterial entry into the vitreous cavity.

In addition, the immunosuppressive effects of steroids are thought to contribute to heightened risk for endophthalmitis by lowering the threshold for bacterial inoculation. One study utilized rabbit eyes to compare the threshold for induction of endophthalmitis at various bacterial loads. It was discovered that eyes that received IVI of steroids required a significantly lower bacterial load when compared to non-steroid groups [4–6]. Another study suggests that corticosteroids act directly on local macrophages, impairing their phagocytic ability and rendering the patient more susceptible to pathogenic infection [7, 8]. Immunosuppressive conditions such as diabetes mellitus are also thought to increase risk of post IVI endophthalmitis [9].

Other studies indicate that corticosteroids may provoke sterile inflammation. This reaction can involve the release of pro-inflammatory cytokines, such as interleukin-6 (IL-6) and interleukin-8 (IL-8), in response to dexamethasone implant exposure [10]. This inflammatory response could imitate or worsen symptoms similar to infectious endophthalmitis, complicating clinical assessment and emphasizing the importance of close monitoring after injections.

Prior literature consistently identifies coagulase-negative Staphylococcus and Streptococcus species, often derived from periocular or oral flora, as the most frequent causative organisms in post-injection endophthalmitis [16]. Effective management involves prompt intravitreal broad-spectrum antibiotics such as vancomycin and ceftazidime, with vitrectomy reserved for severe or non-responsive cases [7]. Preventive measures remain paramount: povidone-iodine antisepsis is the most evidence-based intervention to reduce bacterial load, while adjunctive strategies such as minimizing verbal communication and maintaining strict hand and eyelid hygiene further mitigate risk [20–23]. Routine use of prophylactic topical antibiotics is not recommended, as studies show no reduction in infection rates and potential promotion of resistant organisms [23, 24]. Collectively, adherence to these standard protocols remains the most effective means of minimizing endophthalmitis risk across all intravitreal injection types, including dexamethasone implants.

For this study, we were unable to distinguish between infectious and sterile forms of endophthalmitis. The database used lacked data for laboratory cultures. As a result, the absence of microbiological information, such as bacterial species and antibiotic sensitivities, made it impossible to analyze the specific pathogens involved and their associated risks.

Injection techniques and mitigations strategies also varied among providers in this multicenter study, with differences in needle size and aseptic methods. Minor procedural variations included differences in the type or concentration of povidone-iodine applied, the use of eyelid speculums, and post-procedure antibiotic drop practices. These variations could have affected infection rates and influenced the study’s overall findings.

Additionally, the study did not take into account important patient factors such as existing eye conditions, immune system suppression, or previous treatments, all of which might influence infection risk. The study also did not assess infection rates relative to the frequency of injections (e.g., 4-week versus 8-week intervals). Furthermore, patients receiving both anti-VEGF and corticosteroid injections at the same time were not specifically analyzed as a separate subgroup. In clinical practice, these injections are not typically performed during the same visit, so patients receiving combination therapy were not examined due to the assumption that infection risk will be independent.