Abstract
PURPOSE:
The purpose is to characterize clinical outcomes and management strategies for corneal perforations secondary to microbial keratitis, and to identify potential risk factors for poor visual prognosis and trends in management.
METHODS:
This retrospective case series analyzed patients treated for corneal perforation secondary to documented microbial keratitis at a tertiary care center between July 2023 and July 2025. Patients with traumatic or autoimmune etiologies were excluded. Clinical data, including demographics, visual acuity, microbial culture results, time to perforation, and treatment modalities, were reviewed.
RESULTS:
Of 23 identified corneal perforation cases, 5 were attributed to microbial keratitis. Pathogens included Pseudomonas aeruginosa, herpes simplex virus 1, and suspected fungal organisms; cultures were not obtained in all cases. Time to perforation ranged from immediate presentation to 8 weeks after symptom onset. Most patients (80%) underwent penetrating keratoplasty, while others were managed with corneal gluing, bandage contact lenses, or medical therapy alone. Final visual acuity varied significantly, with outcomes ranging from hand motion to 20/30. Worse visual outcomes were associated with central or large perforations and preexisting ocular comorbidities.
CONCLUSION:
Corneal perforation due to microbial keratitis remains a severe, vision-threatening complication requiring timely and often surgical intervention. Delayed presentation and central ulceration were associated with poorer outcomes. Although limited by a small sample size and a retrospective design, this series highlights the variability in clinical course and visual recovery. Further prospective, multicenter studies are warranted to better define prognostic indicators and to guide standardized treatment algorithms for infectious corneal perforations.
Keywords: Corneal perforation, microbial keratitis, perforation
INTRODUCTION
Corneal perforations are defined as a focal wound through the cornea that extends full-thickness, through the epithelium and the endothelium. These injuries to the cornea can have devastating consequences to vision and can occur secondary to a host of etiologies, including trauma, infections, chemical injury, autoimmune disorders, and tumors. They portend an especially poor prognosis in the setting of microbial keratitis, more commonly referred to as “corneal ulcers.” The insults to the cornea affected by microbial keratitis are multifactorial. Epithelial disruption leads to a deeper infection of the stroma, with subsequent immune responses and tissue destruction. In addition to inflammatory mediators that contribute to this localized destruction, corneal matrix metalloproteinases induce further damage to corneal collagen.[1] Ultimately, continued enzymatic activity can degrade enough stromal tissue that the remaining Descemet membrane and endothelium cannot withstand intraocular pressure or any minor trauma, and perforation becomes imminent.
Perforations in the context of infection usually occur secondary to bacterial keratitis, followed in prevalence by mixed, fungal, and viral keratitis. This pattern mirrors the overall incidence of keratitis, with bacterial keratitis being most common.[2]
Risk factors for corneal perforation often parallel those for microbial keratitis. Well-established risk factors for microbial keratitis include contact lens wear, corneal trauma, ocular surface disease, prior ocular surgery, and systemic immunosuppression, all of which disrupt corneal defenses and predispose to infection.[3] Socioeconomic factors have also been implicated, with lower socioeconomic status associated with delayed presentation and worse clinical outcomes.[3] In the Nottingham Infectious Keratitis Study, ocular surface disease, contact lens wear, and systemic immunosuppression were among the most common predisposing factors, while poor outcomes were associated with delayed presentation, age >50 years, infiltrate size >3 mm, and reduced presenting visual acuity.[4]
In the literature, fungal keratitis has been shown to have a high rate of perforation. In one study out of China, which included 648 patients with fungal keratitis, the rate of perforation was 23.7%.[5] Another study out of India compared differences in clinical outcomes between fungal and bacterial keratitis. While the number of perforations between these groups was not statistically different, the much higher incidence of bacterial keratitis yielded a greater rate of corneal perforation in the fungal keratitis group (16%) compared to the bacterial keratitis group (3%).[6] One study identified several risk factors for corneal perforation from infectious keratitis, including outdoor occupation, illiteracy, alcohol use, vegetable matter trauma, central ulcer location, delayed initiation of fortified antibiotics, and lack of corneal vascularization.[7]
The corneal surgeon has a variety of tools to treat perforations that occur in the setting of infectious keratitis. Amniotic membrane grafts, tissue adhesives, and penetrating keratoplasty (PKP) have all been deployed with variable outcomes. Ultimately, the choice of management depends on numerous key factors, such as the size and location of perforation, pathogen, and the physician’s preference.
While papers have been published on risk factors for perforation after microbial keratitis, this case series will help to evaluate outcomes after infectious corneal perforations. Ultimately, the goal is to better understand this devastating complication, identify patients who may be at higher risk, and recognize clinical characteristics associated with poor outcomes to better inform prognosis and management strategies.
METHODS
This retrospective case series included patients aged 18–95 who were treated for a corneal perforation secondary to documented microbial keratitis between July 1, 2023, and July 1, 2025, at the University Hospitals Eye Institute (Cleveland, OH). This study was approved by the Institutional Review Board of Case Western Reserve University/University Hospitals (IRB protocol number: 20251014).
Cases were identified using ICD-10 codes for corneal perforation: H16.07, H16.071, H16.072, H16.073 and H16.079. Identification of patients with corneal perforation was then followed by manual chart review to confirm that the perforation was preceded by microbial keratitis. Microbial keratitis was defined based on clinical diagnosis by a cornea specialist, supported by slit-lamp findings (epithelial defect with stromal infiltrate, hypopyon, or keratic precipitates), and microbiologic data when available. In cases where cultures were deferred due to concern for imminent perforation, diagnosis was made clinically.
Exclusion criteria included any history of trauma or autoimmune disease contributing to perforation.
For each patient, the following data were collected: age, gender, best-corrected visual acuity (BCVA) at presentation and after treatment, intraocular pressure, size and location of epithelial defect and infiltrate, isolated pathogens (if any), and time from reported symptom onset (or initial presentation when unclear) to perforation. Management strategies and surgical interventions were also recorded where applicable.
RESULTS
Search of the medical record system yielded 23 patients with a diagnosis code of corneal perforation since July 2023. These cases were reviewed, and five cases were determined to have occurred secondary to microbial keratitis. The cases are summarized in Table 1.
Table 1.
Cases of corneal perforation, summarized
| Age and gender | Ocular history | Laterality | Presenting visual acuity | Notable examination findings | Cultures obtained | Isolated pathogen | Time to perforation from initial visit | Treatment | Final known visual acuity | |
|---|---|---|---|---|---|---|---|---|---|---|
| Case 1 | 84 male | None | OS | HM | 4 mm epi defect; 3 mm infiltrate, shallow AC | No | None, suspected bacterial | At presentation | BCL, corneal gluing | 20/400 |
| Case 2 | 75 female | Keratoconjunctivitis sicca, NAION OU | OS | HM | 8 mm × 8 mm epi defect; triangular shaped infiltrate | Yes | Pseudomonas aeruginosa | 1 day | Corneal gluing, PKP | HM |
| Case 3 | 96 female | Ocular trauma with macular scar OS, dry eyes | OD | CF | Mild K edema, D-folds, large white KPs | No K scraping; PCR obtained positive for HSV1 | HSV1 | 8 weeks | BCL, corneal gluing, PKP | 20/400 |
| Case 4 | 45 male | Herpetic keratitis s/p PKP | OD | HM | Edematous PKP graft, central ulceration with perforation | No | None | At presentation | BCL, PKP | 20/100 |
| Case 5 | 79 female | HZO, neurotrophic keratitis | OD | Unable to be recordeda | Descemetocele with perforation | No | None, suspected fungal | At presentation | Corneal gluing, PKP | 20/30 |
aFor case 5, presenting visual acuity was not documented in the medical record due to patient discomfort and inability to complete formal testing. HM: Hand motion, CF: Count fingers, K: Cornea, D-folds: Descemet folds, KP: Keratic precipitates, BCL: Bandage contact lens, PKP: Penetrating keratoplasty, NAION: Nonarteritic ischemic optic neuropathy, HZO: Herpes zoster ophthalmicus, HSV1: Herpes simplex virus 1
Case 1
An 84-year-old male with a history of metastatic cancer of unknown primary, presented to the ED with a 1-week history of severe left eye pain and a gradual decline in vision. Entrance examination was notable in the left eye for hand motion vision. Slit lamp examination was notable in the left eye for a 4 mm corneal epithelial defect with a 3 mm area of infiltrate, and a shallow anterior chamber. Given high concern for a corneal ulcer with perforation, the patient was started on fortified tobramycin and vancomycin. Corneal cultures were deferred. A BCL and shield were placed over the left eye, and the patient was admitted for observation. He was re-examined the following day with corneal gluing performed. One week later, he was tapered down from fortified antibiotics to moxifloxacin drops five times daily. He was seen 2 weeks later with recurrent perforation of his left cornea, and so corneal gluing was repeated. He was seen regularly over the next month and eventually started on Pred-Forte. Given the uncertainty surrounding the etiology of his perforation, consideration was given for herpes simplex virus (HSV) keratitis, and prophylactic Valtrex was started. The patient was last seen 3 months after his initial presentation with reduced vision of the left eye to 20/400 due to corneal scarring.
Case 2
A 75-year-old female with a medical history of Sjogren syndrome and past ocular history of keratoconjunctivitis sicca, bilateral sequential nonarteritic ischemic optic neuropathy, presented to the ED with a 1-week history of left eye pain and discharge. The entrance examination was notable in the left eye for hand motion vision. Slit lamp examination was notable in the left eye for an 8 mm × 8 mm epithelial defect inferiorly with overlying infiltrate, and an anterior chamber with poor view and no apparent hypopyon. There was significant concern for a corneal ulcer; the infiltrate was cultured, and the patient was started on fortified Tobramycin and vancomycin. When she was seen in the clinic the following morning, the area of ulceration was seidel positive with concern for perforation. After discussion of treatment options, she underwent corneal gluing. Cultures returned with pan-sensitive Pseudomonas aeruginosa, and the patient was started on moxifloxacin 400 mg by mouth once daily for a 10-day course. About 1 week later, her fortified antibiotic regimen was de-escalated to moxifloxacin every 1 h. Over the next month, she was tapered off moxifloxacin drops and started on Pred Forte. Her left eye vision remained limited to hand motion due to corneal scarring, although with unclear visual potential. After shared decision making, she eventually underwent PKP ~3 months after initial presentation.
Case 3
A 96-year-old female with a medical history of Alzheimer’s dementia and a past ocular history of remote ocular trauma with macular scar OS, dry eye syndrome, who presented to the clinic with right eye vision changes. The entrance examination was notable in the right eye for counting finger vision. Slit lamp examination was notable in the right eye for mild corneal edema with Descemet folds, and large white keratic precipitates. Dilated fundus examination was unremarkable in the right eye with no evidence of vitritis and a stable macular scar in the left eye. There was significant concern for HSV endothelitis, and so an aqueous tap was performed, which was positive for HSV1. Over the next month, the patient was started on antiviral treatment and closely followed. Six weeks after initial presentation, she had a large epithelial defect with a new central infiltrate and hypopyon. Culture was deferred as it was presumed to have a low yield. She was seen over the following 2 weeks with eventual development of a shallow anterior chamber and concern for impending/prior perforation. She was eventually considered for a PKP to improve vision in her right eye; however, at this time, a dislocated posterior chamber intraocular lens was noted, and so lens exchange was also planned. Six months following the initial presentation, she underwent PKP with IOL exchange and placement of an ACIOL. When last seen 3 weeks following her keratoplasty, she had 20/400 vision in her right eye.
Case 4
A 45-year-old male with a history of herpes keratitis OD status post PKP presented to the clinic with right eye pain and redness. He commented that 2 days prior to the presentation, he felt a “gush of fluid” from his right eye. His entrance examination was notable in his right eye for hand motion vision with no RAPD, and IOP deferred. Slit lamp examination was notable in the right eye for a diffusely swollen PK graft with central ulceration that was Seidel positive. A recurrent herpes keratitis with a perforated corneal ulcer was suspected. After his anterior chamber did not form after BCL placement, he underwent an urgent repeat PKP. Over the next week, he was noted to have a formed anterior chamber and normal IOP. Two months after his keratoplasty, his vision in the right eye was 20/100 with a scleral contact lens.
Case 5
A 79-year-old female with a history of herpes zoster ophthalmicus OD with resultant neurotrophic keratitis presented to the clinic with complaints of right eye pain and blurred vision. Presenting visual acuity was not documented in the medical record due to patient discomfort and inability to complete formal testing. Slit lamp examination was notable for a paracentral descemetocele that was Seidel positive with an unremarkable fundus examination. This perforation was initially assumed to be neurotrophic in etiology, so the patient underwent corneal gluing with aggressive lubrication. However, several days later, she was noted to have a new hypopyon in the right eye, and there was concern for infectious endophthalmitis. She underwent vitreous tap and injection with vancomycin and ceftazidime. Vitreous cultures were negative, and so a fungal keratitis was suspected as the source of infection and etiology of perforation. The patient was started on ketoconazole with a good response to treatment, though eventually complicated by drug-induced acute hepatitis requiring hospitalization. The patient underwent a PKP along with lateral tarsorrhaphy. Over the next year, her examination continued to stabilize with her visual acuity in the right eye improving to 20/30 with correction.
DISCUSSION
This case series highlights the clinical outcomes of corneal perforations secondary to microbial keratitis managed at a tertiary care center over 2 years. Our findings reinforce the complex and multifactorial nature of corneal perforation in the context of infection, and underscore the importance of early intervention, appropriate antimicrobial therapy, and timely surgical management in preserving ocular integrity and visual function.
Across the five cases, delayed presentation appeared to be the most consistent feature in our series. Three of the cases presented with a perforation; one case perforated the day after the initial presentation, and the last patient perforated at 8 weeks from initial presentation. This likely contributed to most patients requiring PKP as part of their management. Poor initial visual acuity was uniformly seen at presentation, and with the exception of Case 1, the other four cases all had varying degrees of known ocular surface disease (keratoconjunctivitis sicca, herpetic keratitis s/p PKP, Herpes Zoster ophthalmicus).
In our cohort, cultures were deferred in most of the cases due to the fact that these patients were already presenting with corneal perforations. In the setting of an open globe/perforation, it is our preference to defer culturing. Suspected microbial etiologies included bacterial, viral, and suspected fungal infections; however, given the lack of microbiologic confirmation in all but one case, definitive classification was limited. The average time from symptom onset to perforation was difficult to ascertain, as several of these patients waited for multiple days to weeks prior to presentation. The fact that 3 of the 5 (60%) patients presented with corneal perforation supports the notion that delayed presentation or inadequate early treatment remains a key contributor to this complication.
Visual outcomes varied significantly depending on the size and location of the perforation, as well as the initial visual acuity at presentation. In addition, concomitant ocular conditions seem to play a significant role in determining final visual acuity. Patients with central perforations and larger epithelial defects seemed to have poorer final BCVA. This aligns with prior studies suggesting that central corneal involvement and more extensive stromal melting are associated with worse visual prognosis. The patient in case 5 had the best final visual outcome, which may reflect that her perforation and descemetocele were more peripheral in nature. This trend may warrant further investigation in larger, prospective studies.
Regarding management, a range of interventions were employed depending on severity, including medical therapy alone, tissue adhesive with bandage contact lens, and surgical approaches such as therapeutic PKP. At our institution, the decision on intervention is primarily dictated by the size of the perforation. If the perforation is <1.5–2 mm, BCL ± cyanoacrylate gluing is preferred as the first-line treatment. Cyanoacrylate glue is also sometimes used as a tamponade to buy time before the patient can be brought to the operating room. PKP can be arranged in 1–2 days rather than same-day, in these situations. With perforations larger than 2 mm, PKP is preferred as the cyanoacrylate glue tends to have trouble keeping these wounds closed. Location is considered, but not a critical decision factor, as even large peripheral perforations are treated with keratoplasty. In addition, visual potential is often not a major factor in the decision. In patients with poor visual potential but with a large perforation, PKP is still pursued to prevent worsening complications such as endophthalmitis or a blind, painful eye. The majority of patients (80%) required surgical intervention with PKP.
These findings are in line with those reported by Lova-Farcia et al., who also noted poor visual outcomes in nontraumatic corneal perforations, and further emphasize the importance of risk stratification and aggressive treatment protocols.[8]
Limitations of this study include its retrospective design, small sample size, and potential referral bias inherent to tertiary care settings. Microbiological data were available in only one of the five cases, which introduces potential misclassification bias, particularly in distinguishing bacterial, viral, and fungal etiologies. The absence of microbiologic confirmation in some cases reflects real-world constraints in managing advanced disease but limits organism-specific conclusions. In addition, most of the patients first presented at a late stage, often with perforation already, limiting our accuracy in deducing the amount of time from the onset of infection to perforation. Late presentation also skews management toward PKP rather than more conservative measures. Despite these limitations, our series offers valuable insight into the real-world outcomes of infectious corneal perforation and adds to the limited existing literature focused specifically on this clinical endpoint.
CONCLUSION
Corneal perforation after microbial keratitis is a sight-threatening complication that often leads to blindness if not identified and treated emergently. Outcomes can be highly variable depending on the underlying pathogen, location, and size of ulceration, time to presentation, and treatment approach. Early recognition of high-risk features and timely escalation to surgical management are critical for preserving vision and preventing globe-threatening sequelae. Further prospective, multicenter studies are needed to better identify predictive factors for poor outcomes and to develop standardized protocols for managing infectious corneal perforations.
Declaration of patient consent
The authors certify that they have obtained all appropriate patient consent forms. In the form, the patient(s) has/have given his/her/their consent for his/her/their images and other clinical information to be reported in the journal. The patients understand that their names and initials will not be published and due efforts will be made to conceal their identity, but anonymity cannot be guaranteed.
Conflicts of interest
There are no conflicts of interest.
Funding Statement
Nil.
REFERENCES
- 1.García-López C, Rodríguez-Calvo-De-Mora M, Borroni D, Sánchez-González JM, Romano V, Rocha-De-Lossada C. The role of matrix metalloproteinases in infectious corneal ulcers. Surv Ophthalmol. 2023;68:929–39. doi: 10.1016/j.survophthal.2023.06.007. [DOI] [PubMed] [Google Scholar]
- 2.Kang L, Lu MC, Niziol LM, Greenwald MF, Mian SI, Thibodeau A, et al. Microbial keratitis isolates at a Midwestern tertiary eye care center. Cornea. 2023;42:1488–96. doi: 10.1097/ICO.0000000000003198. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Ting DS, Ho CS, Deshmukh R, Said DG, Dua HS. Infectious keratitis: An update on epidemiology, causative microorganisms, risk factors, and antimicrobial resistance. Eye (Lond) 2021;35:1084–101. doi: 10.1038/s41433-020-01339-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Ting DS, Cairns J, Gopal BP, Ho CS, Krstic L, Elsahn A, et al. Risk factors, clinical outcomes, and prognostic factors of bacterial keratitis: The Nottingham infectious keratitis study. Front Med (Lausanne) 2021;8:715118. doi: 10.3389/fmed.2021.715118. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Dewang S, Xiaoquan Z, Lu H, Wei S, Wei C, Hua W, et al. Prevalence and prognosis of corneal perforation in patients diagnosed with fungal keratitis. Biomed Res. 2017;28:1251–4. [Google Scholar]
- 6.Prajna NV, Srinivasan M, Lalitha P, Krishnan T, Rajaraman R, Ravindran M, et al. Differences in clinical outcomes in keratitis due to fungus and bacteria. JAMA Ophthalmol. 2013;131:1088–9. doi: 10.1001/jamaophthalmol.2013.1612. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Titiyal JS, Negi S, Anand A, Tandon R, Sharma N, Vajpayee RB. Risk factors for perforation in microbial corneal ulcers in north India. Br J Ophthalmol. 2006;90:686–9. doi: 10.1136/bjo.2005.079533. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Loya-Garcia D, Serna-Ojeda JC, Pedro-Aguilar L, Jimenez-Corona A, Olivo-Payne A, Graue-Hernandez EO. Non-traumatic corneal perforations: Aetiology, treatment and outcomes. Br J Ophthalmol. 2017;101:634–9. doi: 10.1136/bjophthalmol-2016-308618. [DOI] [PubMed] [Google Scholar]
