• Summary
  Click here to print a Quick Reference Guide
• Background
• Clinical Diagnosis
• Investigation
• Treatment
• Non-Antimicrobial Treatment
• Empirical Antimicrobial Treatment
• Directed Antimicrobial Treatment (when microbiology results are known)
• Duration of Treatment
• Switch to oral agent(s)

Microbial keratitis (infection of the cornea with ulceration) should be suspected in any patient with risk factors (see box 1.) and a red eye.

History: patients may describe any of the following:

• Pain – foreign body sensation,
• Photophobia,
• Redness,
• Tearing,
• Decreased visual acuity.

Identify any risk factor(s) (check Box 1):

Examination

• Look for conjunctival injection,
• Assess visual acuity,
• Perform slit lamp examination:

Characteristic lesion is a corneal epithelial defect with associated infiltrate into the stroma.

Initial investigations

• An ophthalmologist should undertake Investigation and management of suspected microbial keratitis.
• A corneal scrape should be performed by an appropriately trained person, according to Standard Operating Procedure, (see appendix one) on corneal ulcers in which any of the following features are present:
  • Central ulcer (within 3mm of the visual axis)
  • Ulcer greater than 1mm in diameter
  • Anterior chamber reaction
  • Failure of previous antimicrobial therapy
• For contact lens associated keratitis the contact lens and contact lens case should be sent for culture (Advise patients the case will not be returned).
• If the patient is not responding to treatment at 48 hours and no organism has been cultured, consider stopping topical antibiotics for 12-24 hours and then re-sampling¹, this should be done in conjunction with a consultant ophthalmologist.
• A corneal biopsy should be sent for culture if debridement is required or if there is no response to treatment.

Non-antimicrobial management

• All patients with keratitis should be started on a mydriatic with cycloplegic action.
• For severe cases of microbial keratitis (as judged clinically) atropine 1% eye drops once a day should used.
• For milder cases of microbial keratitis (as judged clinically) cyclopentolate 1% twice daily eye drops may be used as an alternative.  Once the anterior chamber reaction has settled, topical cycloplegia can be discontinued.
• Topical debridement of the cornea (in conjunction with corneal scrape sampling) can be a useful adjunct to therapy as it removes necrotic tissue and enhance antibiotic penetration.
• Good lid hygiene is important in cases of blepharitis.  Ideally patients should receive a patient information leaflet (See below).
• Initially all topical ocular corticosteroids should be stopped or significantly reduced.
• Contact lenses should be temporarily stopped until treatment is complete (see below).
• Contact lens wear may be resumed two weeks after completion of the course of treatment providing the patient is asymptomatic.
• For patient comfort a preservative free lubricant e.g. Carmellose 1% eye drops should be used especially if there are any epithelial defects.

Empirical antimicrobial therapy

• Preservative free products especially benzalkonium chloride free products should be used if there is a history of toxicity or evidence of toxicity develops on therapy.
• Corneal scrapings should be taken before antimicrobial therapy is started (Therapy can be modified once culture or sensitivity results are known – see directed therapy section.)
• All cases of keratitis should be commenced on empirical antibacterial therapy pending cultures results.
• Initial empirical therapy for small ulcers (<1mm) contact lens related keratitis is: Ofloxacin 0.3% eye drops hourly day and night for 24 hours and then hourly by day only. Review at 48hours – see duration section.
• Initial empirical therapy for large ulcers (>1mm) or in patients with ocular surface disease is: Ofloxacin 0.3% eye drops and Cefuroxime 5% eye drops hourly both day and night for 24 hours and then hourly by day only. Review at 48 hours – (see duration section) and onward referral to the corneal service.
• If a preservative free quinolone is required then levofloxacin 0.5% unit dose is an alternative to ofloxacin 0.3%.
• If the patient gives a clear history of anaphylaxis to penicillin or a clear history of allergy to cephalosporin then empirical regimen is Ofloxacin 0.3% eye drops hourly plus Gentamicin 1.5% eye drops (instead of Cefuroxime ).  Gentamicin 1.5% can be preserved or preservative free depending on availability of drops and clinical choice.  The Gentamicin will be hourly day and night for 24 hours then hourly for one day and review at 48 hours.
• Topical Gentamicin 1.5% eye drops should be reserved for second line therapy and used in place of Cefuroxime when clinical response is suboptimal at 48 hours and cultures remain negative (or for cephalosporin allergic patients).
• Intensive topical therapy with drops is recommended.  Hourly at first and then if the patient is responding after 24 hours gradually reduced slowly over several days to four times a day.
• Routine addition of antifungal treatment is NOT required prior to microbiology results unless there is a history of long term topical steroid use OR immunocompromised OR ocular contamination with vegetative matter AND multifocal corneal infiltrates are seen and considered likely to be fungal keratitis by a consultant ophthalmologist.
• Empirical therapy for suspected fungal keratitis: amphotericin 0.15% eye drops applied hourly.

Directed antimicrobial therapy
Modify treatment regimens when results of microbiological investigations are available. (see full guideline)

Duration
Review all treatment regimens at 48 hours.

Treatment failure
If the patient is not responding to treatment at 48 hours and no organism has been cultured, consider stopping topical antibiotics for 12-24 hours and then re-sampling, this should be done in conjunction with a consultant ophthalmologist.

Referral criteria
Suspected microbial keratitis should be managed by an ophthalmologist.

Back to top

Microbial keratitis is inflammation of the cornea due to infection with bacteria, fungi, viruses or protozoa, but is rare in the absence of predisposing factors.  It can be a sight threatening condition and delays in diagnosing and treating this condition have been associated with a 50% reduction in healing with good visual outcomes.  If scarring occurs it can lead to visual loss (especially if central cornea affected) and untreated or severe keratitis can lead to corneal perforation with the potential to develop into an endophthalmitis and the possible loss of an eye.  As this can occur within 24 hours if virulent organisms are present it is vital that there is prompt identification and treatment.^1,2

In the last 10-12 years contact lens wear has become the major cause of keratitis in developed countries, with the risk being 80 times that compared to patients with no ocular risk factors.  Overnight wear of soft contact lenses has a risk five times greater than daily wear soft lenses and twenty times greater than rigid lenses.  Contact lens wear is the greatest risk factor for bacterial keratitis independent of other factors, including hygiene, and it has been reported that 65% of all new cases of keratitis in London were due to contact lens wear.³  Poor hygiene practices double the risk of bacterial keratitis and one study from Moorfields shows that some disposable lenses (Acuvue) may have a higher risk for bacterial keratitis.⁴

Geography and underlying causes determine the most likely causative organisms e.g. fungal keratitis is more common in tropical countries and Acanthameoba more common in contact lens wearers.

Causative organisms
Bacteria
Bacteria account for 90% of cases of keratitis in cool northern climates and 60% in hot southern climates.  The most common organisms implicated in the UK are staphylococci, streptococci and, in contact lens wearers, Pseudomonas aeruginosa.⁵  However, other Gram negatives, particularly Enterobacteriaceae (Proteus, Serratia, Klebsiella, Enterobacter, Citrobacter) can also cause keratitis.⁶

The introduction of refractive surgery, especially laser-assisted in-situ keratomileusis (LASIK), appears to have coincided with the appearance of keratitis caused by more unusual organisms, such as Nocardia and Mycobacterium spp.⁶  Numerous factors may account for the apparent changes in causative organisms but might include: culture techniques; use of single agent therapy (i.e. fluoroquinolones); increased use of topical corticosteroids (in refractive and cataract surgery); increased population of systemically immunodeficient patients; and an expansion in the use of soft contact lenses, especially extended-wear and cosmetic lenses.⁶

Local patterns of antimicrobial resistance can vary and treatment guidelines need to be amended accordingly.⁷  For example, in a US centre, Gentamicin resistance among staphylococci has remained low (<10%) while resistance to quinolones and Cefuroxime increased.

Fungi
Fungal (mycotic) keratitis is more common in tropical countries, particularly in rural areas, and is a relatively uncommon cause of keratitis in developed countries.⁸  The estimated annual incidence in the UK is 0.32-0.5 cases per million.⁹  Risk factors associated with mould causes of fungal keratitis are trauma, contact lens use and corticosteroid use, whereas yeast infections are associated with either local ocular or systemic defects.¹⁰  UK-based data indicates that over half of cases are caused by yeasts (especially C. albicans) with the rest caused by moulds, including Fusarium, Aspergillus and Scedosporium.⁹  Since January 2008, based on isolates recovered from specimens sent to Mycology in Leeds, yeasts are the most common cause of keratitis locally (~70% of isolates) with moulds causing around 30% of infections (unpublished local data).

Amoebae
Acanthamoeba keratitis is rare in Yorkshire, but is more common in London (relative risk 6.4 in 1987-9).  An increasing prevalence of soft contact lens use and lapses in contact lens care (use of tap water to rinse out contact lens case) are proposed explanations for the increasing incidence of Acanthamoeba infections.¹¹

Viruses
The most commonly implicated viruses are the adenoviruses (mostly types 8,19 and 37).  Infections are usually self-limiting.¹²  More serious is herpes simplex virus (HSV) keratitis, which is one of the most common infectious causes of blindness.  It is normally caused by Herpes simplex type 1 virus (HSV-1) although rare cases can be caused by Herpes simplex type 2 (HSV-2).  HSV infections can be primary or recurrent and either epithelial or stromal, with the latter more likely to lead to scaring and loss of vision.  HSV can never be totally eradicated and lays dormant in the trigeminal ganglion.¹³  Latent HSV in the trigeminal ganglion, and sometimes the cornea, can be reactivated by UV exposure, trauma, immunosuppression and ophthalmic procedures e.g. penetrating keratoplasty, Nd:YAG laser peripheral iridotomy.  Varicella-zoster virus (VZV) can cause significant problems if shingles affects the cornea but chicken pox rarely causes ocular problems.  Shingles can cause a stromal keratitis at onset and treatment is usually systemic (see treatment of herpes zoster).

Predisposing factors.

Back to top

Microbial keratitis should be suspected in any contact lens wearer with a red eye.

History: patients may describe any of the following

• Pain – foreign body sensation,
• Photophobia,
• Redness,
• Tearing,
• Decreased visual acuity.

It is important to identify the precipitating factor(s) in each case (see box 1).

Examination

• Conjunctival injection,
• Assess visual acuity,
• Slit lamp examination:

The characteristic lesion of bacterial keratitis is a corneal epithelial defect (ulcer) with associated suppurative infiltrate into the stroma.  (Although some pathogens cause characteristic changes to the cornea, examination findings alone are insufficient to make a microbiological diagnosis).

Clinical features suggestive of fungal keratitis:

• Presence of risk factors (long term topical steroid use, immunocompromised patient).
• Multifocal corneal infiltrates.

The clinical presentation of fungal keratitis is insidious and slowly-progressive ulceration of the cornea with variable degrees of inflammation and pain.  Hypopyon may be seen. Characteristic features include feathery edges to the ulcer, a dry grey elevated infiltrate and satellite lesions.¹⁴

Diagnosis of Acanthamoeba infection is difficult and it is important to have a high index of suspicion especially in contact lens wearers.  As it is often initially diagnosed as herpes simplex, it is worth reviewing contact lens wearers who have been newly diagnosed with herpes simplex if they have significant pain and poor response to therapy. 

Clinical features suggestive of Acanthamoeba keratitis:

• Dendritiform epithelial keratitis
• Perineural infiltrate

(Since early Acanthamoeba keratitis can appear dendritic, never diagnose HSV keratitis in a contact lens wearer without attempting to culture Acanthamoeba keratitis.)

The first symptoms of herpes simplex infection are usually irritation, photophobia and tearing.  As corneal anaesthesia usually occurs early in the course of the infection the symptoms may be minimal.  The most characteristic lesion is the dendritic ulcer, which occurs in the corneal epithelium and has a typical branching, linear pattern with feathery edges and terminal bulbs.  However HSV may form a geographic ulcer in which the edges of the ulcer loose their feathery quality and an ulcerative or vesicular blepharitis may also occur.

Herpes simplex causes epithelial keratitis, an inflammation of the epithelial cells of the cornea and as it usually resolves after 1-2 weeks the rational for treatment is to minimize stromal damage and scaring.  Patients who are immune deficient are more likely to develop stromal keratitis (inflammation of the stromal layer).  The rate of recurrence of ocular herpes for people with one episode is 10% at 1 year, 23% at 2 years and 50% at 10 years.  The risk of recurrence of ocular herpes also increases with the number of previous episodes reported.¹⁵

Back to top

Recommendation: An ophthalmologist should undertake Investigation and management of suspected microbial keratitis.
[Evidence level D]

Recommendation: A corneal scrape should be performed by an appropriately trained person (according to Standard operating procedure – appendix 1), on corneal ulcers in which any of the following features are present:

• Central ulcer (within 3mm of the visual axis)
• Ulcer greater than 1mm in diameter
• Anterior chamber reaction
• Failure of previous antimicrobial therapy

[Evidence level C] ⁶

Recommendation: In cases of contact lens associated keratitis the contact lens and contact lens case should be sent for culture (the lens and contact lens fluid will be cultured).  Advise patients the case will not be returned.
[Evidence level C]

Recommendation: If the patient is not responding to treatment at 48 hours and no organism has been cultured, consider stopping topical antibiotics for 12-24 hours and then re-sampling¹, this should be done in conjunction with a consultant ophthalmologist.
[Evidence level C]

Recommendation: A corneal biopsy should be sent for culture if debridement is required or if there is no response to treatment.
[Evidence level C]

Back to top

Recommendation: All patients with keratitis should be started on a mydriatic with cycloplegic action.
[Evidence level C]

Recommendation: For severe cases of microbial keratitis (as judged clinically) atropine 1% eye drops once a day should be used.
[Evidence level D]

Recommendation: For milder cases of microbial keratitis (as judged clinically) cyclopentolate 1% twice daily eye drops may be used as an alternative.  Once the anterior chamber reaction has settled, topical cycloplegia can be discontinued.
[Evidence level D]

Recommendation: Topical debridement of the cornea (in conjunction with corneal scrape sampling) can be a useful adjunct to therapy as it removes necrotic tissue and enhance antibiotic penetration.
[Evidence level D]

Recommendation:  Good lid hygiene is important especially in cases of blepharitis, all patient should receive patient information leaflet.  A BMJ-recommended blepharitis leaflet can be found at patient.co.uk (www.patient.co.uk/health/Blepharitis.htm)¹⁶
[Evidence level C]

Recommendation: Initially all topical ocular corticosteroids should be stopped or significantly reduced.
[Evidence level C]

Recommendation: Contact lenses should be temporarily stopped until treatment is complete (see below)
[Evidence level C]

Recommendation: Contact lens wear may be resumed two weeks after completion of the course of treatment providing the patient is asymptomatic.
[Evidence level C]

Recommendation: For patient comfort a preservative free lubricant e.g. Carmellose 1% eye drops three to four times a day (adjusted up or down as required to maintain comfort) should be used especially if there are any epithelial defects.
[Evidence level D]

Evidence review.
The use of mydriatics with a cycloplegic action (e.g. atropine and cyclopentolate) is common practice in ophthalmology in anterior eye disease and is strongly recommended where substantial anterior inflammation is present.  Mydriatics relieve pain associated with the inflammatory process by relaxing the ciliary muscle spasm.  It also helps prevent posterior synechiae by dilating the pupil so the area of posterior iris surface in contact with the anterior lens capsule decreases.¹

Debridement is a common practice as it is thought to aid healing by increasing antibiotic penetration.¹⁷  There is little documented evidence of clinical trials to support this.

Contact lenses
Standard ophthalmic textbooks recommend that contact lens wear should be discontinued if there is any sign of infection. 

Corticosteroids
The use of topical corticosteroids is controversial as there is no trial data to guide decisions.  Some clinicians recommend adding in topical steroids after 2-3 days of anti-infective medication providing the patients is responding as they argue that steroids reduce stromal scaring and improve visual outcomes.  If steroids are used they should be at the minimum strength and frequency to control symptoms and the patients should be reviewed at least daily.  The intra ocular pressure should be monitored carefully.¹

Back to top

The aim of treatment is to resolve corneal inflammation and infection and to restore the corneal integrity with minimal scarring and vascularisation thereby restoring visual function.¹

Recommendation: Benzalkonium chloride - free products should be used if there is a history of toxicity or evidence of toxicity develops on therapy.¹²
[Evidence level C]

Recommendation: Corneal scrapings should be taken before antimicrobial therapy is started (Therapy can be modified once culture or sensitivity results are known – see directed therapy section.)
[Evidence level C] 

Recommendation: All cases of keratitis should be commenced on empirical antibacterial therapy pending cultures results.
[Evidence level C]

Recommendation: Initial empirical therapy for small ulcers (<1mm) contact lens related keratitis is: Ofloxacin 0.3% eye drops hourly day and night for 24 hours and then hourly by day only. (Levofloxacin 0.5% unit dose if preservative free required)
Review at 48hours – see duration section.
[Evidence level B] 

Recommendation: Initial empirical therapy for large ulcers (>1mm) or in patients with ocular surface disease is: Ofloxacin 0.3% eye drops hourly (Levofloxacin 0.5% unit dose if preservative free required) and Cefuroxime 5% eye drops hourly day and night for 24 hours and then hourly by day only. Review at 48 hours – (see duration section) and onward referral to the corneal service.
[Evidence level B]

Recommendation: If the patient gives a clear history of anaphylaxis to penicillin or a clear history of allergy to cephalosporin then empirical regimen is Ofloxacin 0.3% eye drops hourly plus Gentamicin 1.5% eye drops (instead of Cefuroxime ) hourly for 24 hours and then hourly day time only and then review.

If a patient is developing an allergy or intolerance to the benzalkonium chloride preservative consider swapping to Levofloxacin 0.5% preservative free and Gentamicin 1.5% eye drops preservative free (if available).  Cefuroxime 5% eye drops is already preservative free.
[Evidence level B]

Recommendation: topical Gentamicin should be reserved for second line therapy and used in place of Cefuroxime when clinical response is suboptimal at 48 hours and cultures remain negative (or for cephalosporin allergic patients).
[Evidence level C] 

Recommendation: Intensive topical therapy with drops is recommended.  Hourly at first and then if the patient is responding after 24 hours gradually reduced slowly over several days to four times a day.
[Evidence level C]

Recommendation: Routine addition of antifungal treatment is NOT required prior to microbiology results unless there is a history of long term topical steroid use OR immunocompromise OR ocular contamination with vegetative matter AND multifocal corneal infiltrates are seen and considered likely to be fungal keratitis by a consultant ophthalmologist.
[Evidence level B]

Recommendation: Empirical therapy for suspected fungal keratitis: amphotericin 0.15% eye drops applied hourly.
[Evidence level D]

Evidence review.
General principles
Microbial keratitis is an ocular emergency and requires early detection and appropriate treatment to preserve as much vision as possible.  In clinical trials it is difficult to establish an end point as the aim of antimicrobial treatment is to sterilize the cornea and this does not always equate with epithelial healing.  Antibiotics may delay healing after sterilization of the cornea due to drug toxicity and preservative toxicity (if preserved antimicrobial agents are used).  Therefore it is important to limit intensive antimicrobial treatment to allow epithelial healing.¹⁸ (See duration)

Intensive topical therapy can lead to high concentrations of drug levels in the cornea maximising the possibility that the minimum inhibitory concentration of infecting bacteria are exceeded.  It is not recommended that after initial intensive therapy the frequency of instillation is reduced to less than four times a day as concerns that sub therapeutic levels may occur and resistance can develop.¹  Ointments do not provide the same concentration of antimicrobial and are therefore not recommended.

The ideal initial antimicrobial regimen should be effective against the likely corneal pathogens, and have low rates of resistance, minimal toxicity to ocular tissues, easy/comfortable to administer and penetrate rapidly into the cornea.⁶

Antibacterial therapy
Evidence to support different regimens for treatment of bacterial keratitis is not current (1998-2003) and sparse.  The general consensus of treatment of keratitis is divided into two options.

• Single therapy with a fluoroquinolone, usually Ofloxacin eye drops.
• Dual therapy with a fortified cephalosporin (Cefuroxime 5%) and an aminoglycoside (Gentamicin 1.5%) eye drop or a fluoroquinolone.

The Ofloxacin study group found that Ofloxacin was as effective as single agent therapy.¹⁹  However, with reports of resistance developing to Ofloxacin it is current practice to use a combination of antibiotics unless the ulcer is small and non sight-threatening.

Aminoglycosides are effective for pseudomonal and coliform infections and have activity against staphylococci; resistance to Gentamicin in these microbial groups is not prevalent.  Unfortunately, with intense treatment with Gentamicin 1.5% eye drops there can be marked epithelial toxicity with pain, redness, punctuate staining and retardation of epithelial healing. Therefore cefuroxime 5% and Ofloxacin ).3% eye drops are used first line and Gentamicin 1.5% eye drops substituted for Ofloxacin if there is insufficient clinical response.^18-20

No robust head to head trials of monotherapy verses dual therapy have been performed. 

Ofloxacin 0.3% eye drops is a preserved eye drop and contains benzalkonium chloride.  Frequent application of benzalkonium chloride is toxic to the cornea and in some patients can lead to ocular surface pathology.  Benzalkonium chloride can disrupt the tear film, decreasing the tear film break up time and lead to corneal erosion.¹²  If a patient is allergic to benzalkonium chloride or is showing signs of toxicity a benzalkonium -free preparation of Levofloxacin 0.5% eye drops is available.  Levofloxacin is the L-enantiomer of Ofloxacin and as it is more soluble than Ofloxacin , it is available as a 0.5% solution (compared to 0.3% Ofloxacin ).   Little clinical evidence is available for the treatment of keratitis however it has been found to be non-inferior to Ofloxacin in the treatment of conjunctivitis.²¹

The positive cultures from corneal scrapes at LTHT have been audited over a four-year cycle from 2006 to 2009. Over that period, three patients with severe pseudomonal keratitis had a poor clinical response to Ofloxacin 0.3% eye drops and Cefuroxime 5% eye drops in combination, but responded well to Ofloxacin 0.3% eye drops and Gentamicin 1.5% eye drops in combination.  As expected, these strains were resistant to Cefuroxime ; they were susceptible to both Ofloxacin and Gentamicin in vitro. Several isolates of staphylococci and streptococci were resistant to Ofloxacin , but sensitive to Cefuroxime .  The Gram positive organisms were all cultured from patients with ocular surface disease.

Systemic antibiotics are not usual in the treatment of bacterial keratitis. However, oral doxycycline is a metalloproteinase inhibitor and its addition to a topical antimicrobial regimen has been advocated by some to prevent corneal perforation in large ulcers with corneal thinning as part of anti-inflammatory therapy for ocular surface disease (Dose 100mg 12-hourly).¹⁹  It is worth noting that this approach has not been shown to be beneficial in infective keratitis but has been described primarily for severe chemical ocular injuries.²²  If it is used, duration should therefore be kept as short as possible in an attempt to reduce adverse effects and colonization with resistant bacteria.  NB. Tetracyclines are becoming an increasingly important therapy for infections caused by multi-resistant bacteria such as meticillin resistant Staphylococcus aureus (MRSA).

Antifungal therapy
Mycotic keratitis, if left untreated, has a substantial risk of progressive tissue damage with corneal and scleral melt, penetration of fungus into the anterior chamber and consequent endophthalmitis.  Severe infections are particularly associated with poor resultant visual acuity or total loss of sight.⁸  Because of this, when there is a high likelihood of fungal keratitis (see recommendations above) empirical antifungal therapy is warranted.

Choice of empirical antifungal
Empirical antifungal therapy needs to cover a range of pathogens until culture results are known.  The chosen agent therefore needs to have activity against moulds and yeasts.  Amphotericin is the most appropriate choice because has an appropriate spectrum (no empirical regimen can be expected to cover all possible agents – hence the need to undertake microbiological sample and modify therapy accordingly) has a shelf life that allows local preparation and storage and is inexpensive.  Topical amphotericin B solution (0.15%) has been widely used for yeast keratitis^8,10 due to its broad spectrum of activity, tolerability and generally favourable outcomes.¹⁰

Anti viral therapy
There is good evidence that topical anti-virals are effective in the treatment of HSV as only 25% of placebo treated patients healed within 1 week and less than 50% in 2 weeks.  Patients treated with topical acyclovir 3% eye ointment and trifluorothymidine 1% eye drops (trifluridine) showed healing of two thirds of patients within 1 week and 90% by 2 weeks.  Trifluorothymidine 1% eye drops(TFT) is unlicensed in the UK but is available in LTHT under specialist request.  The TFT stocked by LTHT contains benzalkonium chloride but it may be possible to obtain a preservative free product if necessary. TFT 1% eye drops and aciclovir 3% eye ointment have also been found to be more effective than topical idoxuridine (no longer available in the UK), with five times as many patients healed within one week with TFT and acyclovir than idoxuridine.  TFT 1% eye drops aciclovir 3% eye ointment and ganciclovir 0.15% eye drops (a licensed product available in the UK).  As most available treatment appear to be as equally effective, then adverse effects and cost should play a major part in choice.  Trial data comparing side effects is sparse.^13,15,23

Aciclovir 3% eye ointment is licensed in the UK, is relatively cheap and even though there are few comparative studies on toxicity it is regarded as having a low toxicity therefore can be recommended as first line treatment.

TFT 1% eye drops is unlicensed in the UK and although widely used in the USA it has poor penetration if the cornea is intact and requires initial administration of every two hours.  It is significantly more expensive than aciclovir and is therefore not recommended first line.

Ganciclovir 0.15% eye drops is also licensed for the treatment of HSV in the UK.  It does not appear to be more effective than aciclovir but is better tolerated and associated with a lower incidence of visual disturbances,²⁴ it is therefore recommended as second line.

There is some evidence that physical debridement plus topical antiviral therapy is more effective than debridement alone but as the quality of the evidence is low firm recommendations cannot be supported.¹⁵

Adding steroids to antivirals in the treatment of stromal keratitis significantly reduced the persistence or progression of stromal inflammation and shortened the duration of stromal keratitis (median duration 26 days with corticosteroid vs 72 days with placebo; difference 46 days 95% CI 14 days to 58 days).  However, it is usual to delay the use of steroids until epithelial healing has occurred as topical steroids applied to a herpetic corneal ulcer can lead to the development of an amoeboid ulcer which can be slow to heal and difficult to treat.  Steroids should never be used without antiviral cover and with very careful monitoring by an ophthalmologist.¹⁵

The addition of oral acyclovir to topical treatment has not been found to be more effective than topical treatment alone.  However, long term oral treatment was found to decrease the incidence of recurrence after one year compared with placebo.¹⁵

Herpes Zoster should be treated aggressively with oral anti-viral drugs.  (See treatment for herpes zoster)

Summary
First line treatment of HSV is topical acyclovir 3% eye ointment applied 5 times a day until at least 3 days after healing is complete.  For stromal infections prednisolone 0.1% or 0.5% eye drops 6-hourly can be added to manage inflammation once epithelial healing has occurred and titrated down.

Second line
Ganciclovir 0.15% eye drops can be used in patients not able to tolerate acyclovir due to side effects.  It is applied 5 times a day until complete epithelial healing and then reduced to three times a day for a further 7 days.

Trifluothymide (unlicensed) 1% eye drops is used in patients who have not responded to aciclovir.  Instilled every two hours up to a maximum of nine times a day for one week then reduced to five times a day. Treatment should be continued for several days after healing has occurred.

Acanthamoeba
(see directed therapy)

Back to top

Recommendation: When culture results are available, antimicrobial therapy should be modified to optimise treatment of the pathogen(s) identified and to minimise unnecessary adverse effects of therapy.
[Evidence level C]

Corneal scrape results are usually available after 48 hours. Discuss the results with microbiology if required and consider altering treatment according to sensitivities.

Coliforms isolated (e.g. Escherichia coli, Serratia, Enterobacter…)
Recommendation: continue single agent Ofloxacin 0.3% eye drops or Cefuroxime 5% eye drops if susceptible, use Gentamicin 1.5% eye drops if isolate is resistant to Ofloxacin and Cefuroxime .
[Evidence level D]

Pseudomonas aeruginosa
Recommendation: For small ulcers (<1mm) and no OSD continue topical Ofloxacin 0.3% eye drops, provided isolate is susceptible.
[Evidence level D]

Recommendation: For large ulcers (>1mm) or OSD, stop the topical Cefuroxime 5% eye drops and commence topical Gentamicin 1.5% eye drops hourly in combination with topical Ofloxacin 0.3% eye drops, provided isolate is susceptible to both agents.
[Evidence level D]

Pseudomonas aeruginosa is inherently resistant to Cefuroxime , but usually susceptible to Gentamicin and Ofloxacin .

Staphylococci:
Recommendation: continue single agent topical Ofloxacin 0.3% eye drops or cefuroxime 5% eye drops if susceptible; use topical Gentamicin 1.5% eye drops if isolate is resistant to Ofloxacin and Cefuroxime .
[Evidence level D]

There is no direct evidence to support directed therapy for staphylococci.  Meticillin-susceptible S. aureus (MSSA) is susceptible to Cefuroxime 5% eye drops but has variable susceptibility to Ofloxacin 0.3% eye drops, while MRSA is often resistant to both.  Coagulase negative staphylococci are also often resistant to both these agents.

Mycobacteria:
Discuss on a case by case basis.

Moxifloxacin 0.5% eye drops have been found to be more potent than other fluoroquinolones against Mycobacterium chelonae and equivalent in activity for M. fortuitum.   Although more clinical trials are required to confirm these results it may be useful in the treatment of mycobacterial keratitis²⁵.

Fungi isolated
Recommendation:  Treatment of keratitis due to yeasts (usually Candida): continue/start amphotericin 0.15% eye drops initially applied hourly for 48 hours, 2 hourly for 72 hours and tapered according to response. (Amphotericin is prepared by the aseptic unit at LTHT).
[Evidence level C].

Recommendation: If a patient with Candida keratitis is responding poorly to topical amphotericin at 48 hours, consider changing to topical voriconazole 1% eye drop instilled hourly initially and titrated down according to response.  This is an unlicensed product obtained from NHS specials units.
[Evidence level D]

Recommendation: Treatment of keratitis due to a mould (usually Fusarium or Aspergillus): Topical natamycin 5% eye drops initially hourly for 48 hours and then reduce to 2 hourly for 72 hours, tapered according to response. (continue empirical amphotericin until natamycin is available, via pharmacy)
[Evidence level A]

Recommendation:  Treatment of severe fungal infections which extend into the sclera or anterior chamber, or lack of response to initial agent after 2-3 days: add chlorhexidine 0.2% eye drops initially hourly and decreased according to response to the initial agent and oral voriconazole 200mg 12-hourly for at least 2 weeks. (Voriconazole may be increased to 300mg 12-hourly if no response occurs) [Evidence level C]

Recommendation: If required due to disease progression,
- consider intralesional voriconazole (50 micrograms in 0.1ml via a 30G needle). Repeat weekly as required.
- intracameral voriconazole (50 micrograms in 0.1ml).
[Evidence level C]

Recommendation: If corneal infection progresses in spite of vigorous antifungal therapy, surgical intervention may be required, i.e. excision biopsy/penetrating keratoplasty¹⁰
[Evidence level C]

Evidence/Justification
Epidemiology differs between geographic areas and although several studies have been conducted on the Indian subcontinent^26-28, the applicability of their results to the UK setting is uncertain. Secondly, availability of appropriate antifungal preparations varies between countries and over time, with little impetus for pharmaceutical companies in developed countries to maintain production or develop new preparations for such an uncommon disease. Older studies have included agents which are no longer used and data on efficacy of the newer agents has often been from case reports. Indeed, a recent Cochrane review found no evidence for the superiority of any single antifungal agent over others.²⁹ The randomised trials included in the review evaluated topical and systemic itraconazole, silver sulphadiazine ointment, miconazole ointment, econazole 2%, natamycin 5% eye drops, chlorhexidine 0.2 and 0.5% eye drops and voriconazole 1% eye drops. None of the trials found statistically significant differences between treatments, but the incidence of side effects was generally very low and usually due to allergic reactions.

Since the Cochrane review was published, for the first time a large randomised, prospective, blinded study has demonstrated that topical 5% natamycin eye drops resulted in significantly better clinical and microbiological outcomes in 162 patients with confirmed fungal keratitis compared to 161 patients receiving topical 1% voriconazole eye drops.³⁰  This difference was largely driven by the success of natamycin in treating Fusarium infections versus voriconazole (p<0.001).  This study demonstrated superiority of natamycin against mould infections, hence its inclusion as first line therapy for a confirmed mould keraritis.  This study did not include any yeasts infections, so amphotericin remains recommended therapy for confirmed yeast keratitis. (Note evidence for natamycin already in)

Voriconazole has activity in vitro against a wide range of fungal pathogens including those likely to cause fungal keratitis in the UK³¹ and in vitro it is generally more active than other antifungals.^31-33  Topical voriconazole penetrates the cornea and achieves therapeutic concentrations in the aqueous, whilst oral voriconazole achieves therapeutic concentrations in the vitreous for most organisms with the possible exception of Fusarium species.³⁴  Outcomes from non-comparative, observational data in patients receiving topical voriconazole +/- oral or injections of voriconazole have recently been reported for 26 patients and only 13 of them responded, with 11 requiring penetrating keratoplasty and two enucleation.³⁵  Although other groups have reported more encouraging results,^34,36 these concerns plus the practical issues of obtaining, storing and funding voriconazole mean that it use should be confined to treatment failures with first line agents or very severe cases.

Although many centres use a combination of topical and oral therapy, particularly in severe infections, there is currently no evidence to support this. One study comparing topical itraconazole with topical and oral itraconazole and found no difference in responses, although it was a small study with no statistical analysis.²⁸  Some authors have suggested that use of combination therapy may reduce the risk of development of resistance,⁸ but again there is no evidence to support this. In spite of this, the addition of oral voriconazole has been recommended for severe infections because the risk of losing the eye is so high.

The use of intracameral and intrastromal voriconazole has been anecdotally reported in the management of fungal keratitis and endophthalmitis, usually in patients resistant to topical and/or systemic therapy^34,37 Various doses have been used, generally with favourable results,^37,38 but reports were all case reports with no comparator.

Acanthamoeba
Recommendation: First line treatment for Acanthaemoeba keratitis is propamidine isothionate 0.1% eye drops with chlorhexidine 0.02% eye drops.  Drops are instilled hourly day and night for two days then by day only for three days then titrate down according to response and side effects/tolerability. [Evidence level C]

Recommendation: Alternative to propamidine isothionate 0.1% is hexamidine 0.1% eye drops only if propamidine is unavailable.Same frequency as propamidine eye drops.

Recommendation: Alternative to chlorhexidine 0.02%, e.g. in allergic patient, is polyhexamethylene biguanide (PHMB) 0.02% and used as the same frequency as chlorhexidine.

Evidence/Justification
As Acanthamoeba exists as both trophozoites and cysts a combination therapy is recommended to cover both forms.  Chlorhexidine is used first line at LTHT but PHMB can be obtained from Moorfields Eye Hospital if necessary.  Initial treatment is Propamidine isothionate 0.1% (Brolene) hourly and chlorhexidine 0.02% hourly day and night for 48 hours, then reduce both to hourly by day only. Treatment is tapered down according to clinical response but usually lasts for 3-6 months.  Chlorhexidine 0.02% eye drops has been reported to be more effective than PHMB.³⁹  If the Acanthamoeba is restricted to the epithelium there is a good chance of recovery however if the stroma is involved there may be necrosis and intense inflammation.     

Herpes simplex virus
Recommendation: First line treatment is topical aciclovir.

If it is a stromal keratitis the addition of a corticosteroid under careful monitoring has been found to reduce inflammation.  For recurrent infections long term prophylactic treatment with oral acyclovir may reduce recurrence.

For pregnant patients
There are very few published data on the potential fetotoxic effects of topical ophthalmic medications in human pregnancy.  Please contact microbiology for specific advice.

Back to top

Empirical therapy
Recommendation: For bacterial keratitis review empirical therapy at 48 hours. If the clinical response is satisfactory and there are no microbiology results that require a change in therapy continue hourly treatment for three more days then reduce to 6-hourly for one week.
[Evidence level D]

Confirmed fungal keratitis
Recommendation: Continue topical treatment for 12 weeks; Continue oral treatment, if started, for 2 weeks.
[Evidence level C]

Back to top

Back to top