Corneal ulcer

What it is

A corneal ulcer is a defect of the corneal epithelium that exposes the underlying stroma. Ulcers are classified by depth and behaviour:

• Superficial: only the epithelium is lost; uncomplicated ulcers heal within about a week.
• Stromal: part of the stroma is lost; deep ulcers threaten the integrity of the eye.
• Descemetocele: the whole stroma is lost and only Descemet's membrane remains; perforation is imminent.
• Perforated: the full thickness of the cornea is breached and the iris may prolapse.
• Melting ulcer (keratomalacia): proteases from bacteria (such as Pseudomonas or streptococci) and from neutrophils digest the stroma, which becomes soft and gelatinous; it can progress within hours.
• Spontaneous chronic corneal epithelial defect ("indolent" or "boxer" ulcer): a superficial ulcer that fails to heal because the epithelium does not adhere to an abnormal anterior stroma, with a loose lip of epithelium at its edge.

Causes include trauma and foreign bodies, eyelid and eyelash disorders (entropion, distichiasis, ectopic cilia), keratoconjunctivitis sicca, exposure from incomplete eyelid closure, facial nerve paralysis and infection.

Who it affects

Dogs of any breed. Brachycephalic breeds (pug, shih tzu, French and English bulldog, Pekingese) are at high risk because of their prominent eyes, incomplete blinking, nasal folds and reduced corneal sensitivity. Indolent ulcers are typical of middle-aged and older dogs, with the boxer as the classic breed.

Anatomy involved

The canine cornea is the transparent front part of the outer fibrous tunic of the eye. From front to back it consists of the anterior epithelium, the collagenous stroma (most of its thickness), Descemet's membrane and the endothelium. It has no blood vessels and is richly innervated by branches of the ophthalmic nerve (trigeminal), so ulcers are very painful. Pain also triggers a reflex uveitis through an axonal reflex, which causes miosis and ciliary spasm.

Clinical signs

Blepharospasm, epiphora, photophobia, rubbing of the eye, conjunctival hyperaemia, corneal oedema around the lesion and miosis. Deep or infected ulcers may show a crater, a soft grey-white or gelatinous stroma, corneal vascularisation, and hypopyon.

Diagnosis

• Fluorescein stain: the hydrophilic stroma retains the dye, while the intact epithelium and Descemet's membrane do not. A descemetocele shows a clear centre with stained walls.
• Assessment of depth with magnification or a slit lamp.
• Cytology and culture of deep, melting or infected ulcers.
• A search for the cause: eyelids, eyelashes, foreign bodies under the third eyelid, Schirmer tear test.

Treatment

Superficial ulcers: remove the cause, topical broad-spectrum antibiotic, topical atropine for the pain of reflex uveitis, an Elizabethan collar and systemic analgesics; reassess within a few days. Indolent ulcers: debridement of the loose epithelium, often followed by grid or punctate keratotomy or diamond-burr debridement. Melting ulcers: intensive antibiotic therapy based on cytology and culture plus anti-collagenase agents such as autologous serum. Deep ulcers, descemetoceles and perforations need surgery: conjunctival pedicle grafts, corneoconjunctival transposition or other grafts. Topical corticosteroids are contraindicated in any ulcer.

In the cat

In the cat the most important cause of corneal ulceration is feline herpesvirus type 1, which produces very characteristic dendritic (linear and branching) ulcers or broader geographic ulcers, often recurrent; trauma (scratches in fights), entropion and exposure in brachycephalic breeds are other causes. The feline cornea has peculiarities that change management:

• Corneal sequestrum: a chronic ulcer can progress to a plaque of necrotic stroma, amber to black, typical of the cat, especially the Persian and other brachycephalic breeds, and associated with herpesvirus.
• Grid and punctate keratotomy must not be performed in cats, because they promote sequestrum formation; non-healing superficial ulcers are treated with cotton-swab debridement, a therapeutic contact lens or superficial keratectomy.
• Antivirals (oral famciclovir, topical cidofovir or ganciclovir) are added for ulcers suspected to be herpetic, and corticosteroids are avoided.
• Melting ulcers are less common than in the dog, but are treated just as intensively.

Key studies in the cat

• La Croix et al. (2001), retrospective study of 29 cats with 36 non-healing ulcers. Adding grid keratotomy to debridement did not shorten healing (mean 42 versus 30 days), and a sequestrum formed in 4 of 13 eyes with grid keratotomy compared with 2 of 21 after debridement alone. A small study, but the main source of the advice against keratotomy in cats.
• Goldreich et al. (2020), 81 cats with culture-confirmed bacterial keratitis. Staphylococcus species made up 55% of the 102 isolates. Susceptibility was highest to ofloxacin (100%), ciprofloxacin (94%), chloramphenicol (93%) and doxycycline (92%), and very low to polymyxin B (2%) and bacitracin (15%).
• Multari et al. (2021), 172 cats (175 eyes) operated on for corneal sequestrum, 71.5% of them Persians and 12.2% Exotic Shorthairs. After keratectomy combined with various grafts or flaps, all eyes kept their vision, 86% had no or mild corneal opacity, and the sequestrum recurred in 11.4% within 2 years.
• Mironovich et al. (2023), placebo-controlled study in 132 shelter cats with FHV-1 treated for 7 days. None of the corneal ulcers in the placebo group healed, whereas 77% healed in cats given an antiviral.
• Frejlich and Payen (2026), 392 referral cats with ulcerative keratitis or corneal sequestrum. Ulcerative keratitis was diagnosed in 36% of brachycephalic cats seen and 15% of non-brachycephalic cats. Brachycephalic cats had more central ulcers and more sequestra, but not more deep or perforated ulcers.
• Reynolds et al. (2026), retrospective study of 71 spontaneous superficial ulcers in 68 cats. A sequestrum developed in 28%, and was more likely after cotton-tipped swab debridement (odds ratio 4.1). Oral lysine, topical NSAIDs and topical EDTA were associated with delayed healing. These are retrospective associations, not proof of cause.