Ocular trauma is the leading cause of monocular blindness worldwide [1]. Traumatic endophthalmitis is a vision-threatening complication of open-globe injury, with reported incidences of approximately 4–16% in prior series [2]. Prognosis remains guarded despite advances in ocular trauma care. Core management includes timely microbiologic sampling, broad-spectrum intravitreal antimicrobial therapy, and PPV in selected cases [1, 2]. National guidance emphasizes prompt primary wound closure, adequate Gram-positive and Gram-negative intravitreal coverage, and early referral to vitreoretinal centers [3].

We conducted a multicenter study across seven tertiary hospitals in northwest Spain to characterize presentation and microbiology, quantify visual and anatomic outcomes, and identify predictors of poor vision, situating our findings alongside current evidence and guidelines. In this context, the multicenter design, which covers a considerable number of hospitals within a defined geographical region (northern Spain) with a similar climate and lifestyle, is particularly valuable. It confers great methodological and clinical relevance, allowing for the standardization of diagnostic and therapeutic protocols, thereby reducing the variability associated with local hospital practices.

Methods

We retrospectively reviewed consecutive cases of traumatic endophthalmitis managed between January 2003 and January 2025 across seven tertiary hospitals in northwest Spain. Reporting adheres to the STROBE guidelines.

BCVA was recorded in logMAR units using predefined values for low-vision categories (counting fingers (CF) = 2.00; hand motion (HM) = 2.30; light perception (LP) = 2.80; no light perception (NLP) = 3.00) [1].

The primary outcome was final BCVA (logMAR). Poor visual outcome was defined as final BCVA ≥ 1.0 logMAR.

All patients were evaluated emergently by an ophthalmologist. Penetrating injuries underwent orbital computed tomography to assess for IOFB and prompt surgical repair of the open globe. When an IOFB was present, PPV with IOFB removal was performed. Ocular ultrasonography was performed in suspected endophthalmitis for diagnostic confirmation, and intravitreal antibiotics were administered in high‑risk presentations, consistent with national guidance [3].

The study population was predominantly rural, reflecting the socioeconomic context of Galicia and Asturias, where agricultural, livestock, and industrial activities are common and inherently increase the risk of ocular trauma from metallic and organic debris. The northwest of Spain has a temperate oceanic climate characterized by mild temperatures, high humidity, and frequent rainfall throughout much of the year. These environmental conditions—together with rural occupational exposures—may shape the local microbiological landscape of traumatic endophthalmitis, with a potentially higher burden of soil- and plant-associated organisms compared with drier or urban regions. Male predominance observed in our study, associated with a wide range of ages, is consistent with previous reports and appears to be related to the greater exposure of males to higher-risk work environments (industrial, agricultural, or involving the use of tools) [4, 5].

The nature of the trauma played a critical role in the etiology of the endophthalmitis cases analyzed. In our study, IOFB was identified in 45.9% of cases, with metallic objects being the most frequently implicated (41.9%), followed by organic materials in 24.3%. Zhuang et al. reported comparable rates of IOFB identification in cases of traumatic endophthalmitis (33.9%) [6]. Other authors, consistent with our observations, have also documented a predominance of metallic IOFBs among those identified [7, 8]. Nonetheless, it is important to emphasize that although the presence of an IOFB substantially increases both severity and infection risk, it is not a prerequisite for the development of post-traumatic endophthalmitis.

Functional recovery following traumatic endophthalmitis remained markedly limited. Although most eyes maintained anatomical integrity, only 43.2% achieved a final BCVA ≤ 1.0 logMAR, underscoring the substantial visual morbidity associated with these injuries. These findings are consistent with aggregated evidence from large cohorts assessing prognostic determinants and surgical timing in open-globe trauma [1, 2]. The ocular manifestations documented in our study align with previous reports and are indicative of a marked and often abrupt inflammatory response [4, 5]. Corneal perforation, identified in 60.8% of cases, was most strongly associated with poor baseline VA and the subsequent need for complex, highly invasive reconstructive procedures.

In our series, baseline VA was the only variable significantly associated with poor visual outcome, while culture positivity, IOFB, and RD showed trends aligned with published literature but did not reach statistical significance. The unadjusted association between intravitreal therapy and poorer visual outcomes likely reflects confounding by indication: eyes with greater baseline severity (e.g., dense vitritis, IOFB, or media opacity) were preferentially treated more intensively; therefore, this correlation should not be inferred as causal [2–3, 9–16]. Prospective capture of injection number and timing, alongside standardized severity metrics and multivariable methods to adjust for baseline imbalances, would help clarify this relationship in future studies.

Microbiologically, Gram-positive bacteria predominated, accounting for 93.0% of positive cultures (40/43), with Staphylococcus species being the most frequently isolated organisms (55.8% of positive cultures, 24/43). Among these, coagulase-negative Staphylococcus represented 46.5% of positive samples (20/43). Bacillus cereus constituted 9.3% of positive samples (4/43) and has traditionally been associated with rapidly progressive post-traumatic presentations and infections related to intraocular foreign bodies (IOFBs), particularly in reports from rural or industrial settings where such microorganisms are common contaminants of open-globe injuries [8].

This profile supports empirical intravitreal coverage targeting Gram-positive and Gram-negative organisms, in line with national guidance from the Spanish Vitreo-Retinal Society (SERV), which advocates empirical intravitreal treatment combining vancomycin with a third-generation cephalosporin (or equivalent Gram-negative coverage), and aligns with recently published reviews on traumatic endophthalmitis [3–6].

Fungal infections were less common in our series (16.3% of positive cultures, 7/43) but remained clinically relevant—particularly in injuries involving vegetative material—consistent with published traumatic series [6, 10]. Nevertheless, they occurred at a slightly higher rate than that reported in other series (5%–15%) [7, 8]. This increased incidence may be related to local climatic conditions—specifically a humid and temperate environment—as well as greater agricultural and livestock activity in the region, factors that heighten the risk of trauma involving vegetative material, a well-recognized vector for fungal inoculation.

Most of the microorganisms isolated in our series demonstrated good susceptibility to the antimicrobial regimens commonly used for empirical intravitreal therapy. However, over the course of more than 20 years, some cases exhibited resistance patterns, particularly to macrolides and lincosamides among Staphylococcus, and although Gram-negative organisms were less common, multidrug-resistant isolates were identified during this period, especially Citrobacter freundii and Proteus mirabilis. These findings underscore the need to adjust the therapeutic regimen once antimicrobial susceptibility results became available. Given that initial treatment in most cases of endophthalmitis is administered empirically, the availability of local antimicrobial susceptibility data is essential to guide the selection of first-line therapy and optimize clinical outcomes.

Regarding prophylaxis after open-globe injury, recent meta-analytic data shows no benefit to extending systemic antibiotics beyond 24 hours and no superiority of intravenous or intraocular routes over oral administration; ciprofloxacin monotherapy underperforms versus vancomycin plus a third-generation cephalosporin [16]. These results support concise, risk-adapted prophylaxis—particularly in patients with IOFB, rural injuries, or lens rupture—and reinforce our cohort’s empirical Gram-positive/Gram-negative approach, in harmony with SERV recommendations [1–3, 6, 16].

Because management is initiated empirically, rapid diagnostics can complement culture-based testing. Polymerase chain reaction (PCR) can refine early treatment decisions and support diagnosis when initial cultures are negative [17]. Among host-response biomarkers, vitreous soluble Triggering Receptor Expressed on Myeloid cells-1 (sTREM-1) may help discriminate infectious from non-infectious presentations; elevations have been reported in both culture-positive and culture-negative traumatic cases, supporting escalation when clinical suspicion remains high despite negative cultures [18].

The traditional Endophthalmitis Vitrectomy Study (EVS) paradigm is frequently challenged in the context of traumatic endophthalmitis. The role and timing of PPV should be individualized. In our series, PPV was performed early in 64.9% of cases and was not independently associated with worse visual outcomes. When IOFB, dense vitritis, or media opacity are present, early PPV facilitates source control, direct sampling, and removal of necrotic/toxic material—though definitions of “early” vary across studies [1, 9–11]. Notably, PPV serves both therapeutic and diagnostic purposes by enabling removal of inflamed and infected vitreous, enhancing intravitreal antibiotic penetration, and providing samples for culture. In eyes with total media opacity where a temporary keratoprosthesis is not feasible, ultrasound-guided PPV (USG‑PPV) can enable safer debulking under intraoperative imaging [19]. Recently, the complementary use of intravitreal injections or low-concentration diluted povidone–iodine irrigation (0.025%–1.25%) has been recommended, as it has demonstrated bactericidal, fungicidal, and virucidal activity in infections associated with intraocular foreign bodies, without inducing inflammation or retinal toxicity; however, the overall quality of evidence remains limited [20, 21].

Broader epidemiologic experience similarly emphasizes the ongoing burden of trauma-related endophthalmitis and the importance of coordinated regional care networks, consistent with our multicenter experience in Asturias and Galicia [3, 12–15].

Strengths of this study include its multicenter design, standardized VA conversion, and comprehensive capture of microbiology, treatment, and outcomes across two decades. Limitations inherent to retrospective analyses include potential selection/referral bias, incomplete timing data, and residual confounding by indication (notably for intravitreal therapy and PPV). Inter-hospital variability in practice, small fungal and Gram-negative subgroups, and possible misclassification of IOFB material may limit generalizability. Future prospective studies should incorporate prespecified, penalized multivariable analyses, standardized time-to-intervention metrics, and systematic handling of missing data.

In this multicenter series of traumatic endophthalmitis, infection control and globe preservation were high, yet functional recovery remained limited. The long study period of over 20 years allowed for the demonstration of stability in the spectrum of causative microorganisms, although resistance to certain antibiotics (macrolides and lincosamides) and multidrug-resistant Gram-negative organisms was observed. Epidemiological experience from a specific region of northern Spain, characterized by specific climatic conditions and occupational activities, can inform the development of empirically tailored therapeutic strategies aligned with the local microbial profile and resistance patterns. Early risk stratification, primarily based on initial visual acuity, combined with careful assessment of injury severity, IOFB status, and RD, may assist in patient classification and treatment planning, thereby facilitating prompt source control and management of trauma-associated endophthalmitis.