Special Report

Meeting the Challenge of Bacterial Keratitis: New Wine in Old Battles

Practical advice for managing this potentially sight-threatening condition.

By Kenneth R. Kenyon, MD

Although relatively uncommon in general ophthalmic practice, bacterial keratitis can nonetheless be associated with significant morbidity and sequelae (Figures 1-5). Thus it demands swift recognition, potent treatment and close follow-up. Potential consequences vary depending on the locus of infection and the virulence of the causative organism. For example, a modest Staphylococcus species marginal ulceration may cause focal corneal thinning and scarring, but should not be of visual consequence (Figure 1). In contrast, an ulcer involving the visual axis and caused by a highly aggressive gram-negative organism poses a major risk to vision (Figure 2). Visually significant corneal scarring might require corneal transplantation, or worse, perforation and endophthalmitis might lead to blindness.

Having a life-long professional interest in corneal ulcerative disorders, I am pleased to highlight what I consider to be the important aspects of managing this multifaceted disease.

Identify Risk Factors

As important as initiating effective therapy is determining the etiology of the infectious process. In this respect, the clinician should back away from the acute process evident at the slit lamp and examine the entire eye and indeed the whole patient. Bacterial keratitis simply does not occur in the absence of risk factors. As clinical detectives, we must meticulously and methodically pursue the “CSI” process. Does this student sleep in her contact lenses? What other topical medications, OTC preparations, lubricants, and so on, does this elderly gentleman use in his eyes, and does he administer or store them in a way that predisposes them to contamination? Does this young woman also suffer from rosacea or atopy, which she conceals with make-up?

Indeed, multiple risk factors—including age, trauma, all ocular surface diseases (blepharitis, dry eye, allergy, HSV keratitis), contact lens use (and abuse), multiple in-use topical meds (especially but certainly not solely steroids), corneal sutures or other foreign bodies (Figure 4), eyelid mal-positions, absent corneal sensation (Figure 5), nasolacrimal duct obstruction—must all be considered in order to assess the likelihood of an infectious process and the predicted virulence thereof. In a word, anything that adversely affects the ocular surface can potentially infect the ocular surface. Therefore, risk factor assessment is helpful in confirming a correct diagnosis and ensuring that the treatable risk factors are ultimately addressed.

Figure 1. For this marginal, likely Staphylococcal, ulcer with minimal epithelial defect, topical treatment with an antibiotic and a steroid was the appropriate therapy.

Assess Infectious Aggressiveness

Although the entire spectrum of bacterial species can potentially cause keratitis, most series document the preponderance of gram positive microbes (Figure 3), such as S. epidermidis and S. aureus, with gram negatives less likely, unless their presence is enhanced, for example, by contact lens wear. Importantly, however, the latter, such as Pseudomonas aeruginosa, are far more virulent and hence vision-threatening.

I judge the aggressiveness of each case by its history and current clinical findings. While all cases should be treated promptly, the sense of urgency is diminished if the disease process has been slow-moving, perhaps recurrent, peripheral corneal in location and perhaps simply not entirely eradicated by the initial treatment regimen. A deliberate reassessment of associated factors, readjustment of current medications (e.g., discontinue antiviral, reduce antibiotic frequency, change antibiotic agent, judiciously add steroids), rarely consideration of other microbial classes (e.g., Acanthamoeba or fungi), and commitment to relatively close and possibly ongoing follow-up is adequate.

On the other hand, it is imperatively urgent to “pull out all the stops” for a dense 4-mm white central infiltrate of acute onset in a contact lens wearer, because in this scenario the cause is Pseudomonas until proven otherwise (Figure 2). Here, aggressive corneal culture and topical therapy (as subsequently defined) is imperative, and daily follow-up is mandatory.

Figure 2. Intense inflammatory infiltrate and profound stromalysis characteristic of Pseudomonas keratitis.

Choose Appropriate Therapy

Topical antibiotics are the self-evident, first-line treatment for bacterial keratitis. The decision of which antibacterial agents to use is determined on a case-by-case basis, using what I term “the punishment should fit the crime” model. For example, for a modest peripheral keratits, perhaps associated with chronic blepharitis (Figure 1), an earlier generation fluoroquinolone (such as ofloxacin) or even an ointment (such as polymyxin-bacitracin) having lubricant as well as antibiotic benefits may be efficacious and, as they are generic, their low cost is of increasingly potential benefit and/or necessity. In cases where more extensive corneal involvement and associated risk factors suggest a more aggressive process, I invariably reach for a fourth-generation agent, such as gatifloxacin (Zymar, Allergan), moxifloxacin (Vigamox, Alcon) or levofloxacin 0.5% (Quixin, Vistakon Pharmaceuticals). Their availability to every eyecare provider, as well as acute care physician/practitioner, has been an enormous advance compared with the “Bad Old Days” of having to concoct fortified antibiotics in hospital or compounding pharmacies. It has also reduced the need for referral to specialized centers for even relatively minor infections. This having been said, in the rare situation where a contemporary fluoroquinolone has already been employed with less than prompt arrest of the infectious process, then I may add a second topical agent, usually fortified vancomycin (50 mg/ml prepared by diluting 500 mg of vancomycin with 10 ml BSS), given its synergistic and extended broad-spectrum capabilities.

However, for a rapid onset and potentially visually catastrophic central corneal ulcer, I rely either on the higher concentration preparation of levofloxacin (Iquix 1.5%, Vistakon Pharmaceuticals) that is specifically indicated for the treatment of bacterial corneal ulcers, or on the newest topical antibiotic option, besifloxacin ophthalmic suspension 0.6% (Besivance, Bausch + Lomb). Besifloxacin is unique in being developed and approved only for ophthalmic use. Here again, I may also decide to double my bet with the addition of fortified vancomycin.

Figure 3. Gram-positive bacteria is the likely culprit in this case of peripheral, focal, nonsuppuritive keratitis.

Because besifloxacin, unlike all prior fluoroquinolones, was not used systemically, the likelihood of bacterial resistance development is greatly reduced. As detailed in The Ocular Tracking Resistance in the U.S. Today (TRUST) study¹ and evident in our regional hospital antibiograms, the prevalence of multiple drug-resistant bugs, such as the dreaded methicillin-resistant S. aureus (of which only 15% are currently sensitive to fluoroquinolones) is rising. This is especially ominous for microbial keratitis management, where the window of treatment opportunity may be perilously brief. In vitro studies against 40 species of bacteria demonstrated besifloxacin as the most potent agent tested against gram-positive organisms and anaerobes, as generally equivalent to other fluoroquinolones tested against most gram-negative pathogens, and as retaining potency against gram-positive and gram-negative isolates resistant to other fluoroquinolones and antibacterials.² This activity spectrum has also been verified in the FDA clinical trials of besi-floxacin.³ Apart from its intrinsic potency, the besifloxacin formulation includes a proprietary drug delivery-vehicle (DuraSite), which is designed to prolong its contact time with the ocular surface.

Figure 4. Retained sutures and other foreign material are risk factors for microbial keratitis, as in this case of a broken exposed suture following corneal transplant that led to abscess formation.

The previously mentioned study² also evaluated minimum inhibitory concentrations (MICs) for besifloxacin and found its profile generally favorable to the other antibiotics tested. However, while MICs provide a useful comparative standard of antimicrobial activity, it is also vital to consider MICs in the context of ophthalmic practice. With respect to corneal and conjunctival infections, we have the ability to literally saturate the ocular surface with topically applied medication, thereby driving enormous concentrations of antibiotics directly to the site of the pathogen. By this means, we can attain and sustain high levels of drug, such that even those with somewhat higher MICs will exceed their efficacy thresholds for all but the most virulent bacteria.⁴ The practical issue therefore is to devise and sustain the appropriate dosing strategy.

Thus for all but relatively minor peripheral ulcerations, I commence therapy with a loading dose of the topical antibiotic q10 minutes x 6 and then continue one drop every hour while the patient is awake. In the most severe cases, it may be necessary to require hourly antibiotic administration around the clock. Follow-up within 24 hours is absolutely mandatory, and subsequent close observation, preferably by the same clinician, is key to assess progress, alter dosage regimens, and possibly add corticosteroids, NSAIDs, bandage soft contact lens, and/or other therapies. Given the high potency and efficacy of topical antibiotics administered by an intense “saturation” protocol, subconjunctival antibiotics are almost never justified, and systemic antibiotics are indicated only in the rare instance of actual or impending corneal perforation.

I abide by several other “dos and don'ts” as well. First, and especially in the setting of a contact lens-related corneal ulcer, presume a potentially catastrophic gram-negative infection until proven otherwise. Therapy must include an appropriate fluoroquinlone delivered intensively, not simply a broad but inadequately potent antibiotic ointment. Second, never patch an infected eye. Doing so occludes worsening symptoms and vision from the patient plus it provides the ideal warm and moist conditions for bacterial growth. A therapeutic soft contact lens can always be added subsequently to promote epithelial healing. Third, daily follow-up is mandatory (as stated above) until clinical stabilization and improvement have become unequivocal.

When to Culture?

Twenty years ago, culturing any and every presumed microbial corneal infection was all but mandatory. Antibiotics of the day were of narrower activity range and would too often require special compounding. Today, the ubiquitous availability and broad potency of off-the-shelf “one drop stopping” fluoroquinolones all but obviates the need, much less cost justification, of culturing.

That said, I continue to culture cases in the following settings:

1. rapidly progressive ulceration, threatening the visual axis or corneal perforation
2. inadequate response to initial therapeutic regimen (requiring 24-36 hour discontinuation of current antibiotics prior to culture)
3. suspicion of an additional or unsuspected microbe (such as fungus or Acanthamoeba) requiring specialized culture media.

Given the frequent contamination of in-use topical medications (even antibiotics), lubricants, contact lenses and lens cases,⁵ culture thereof often discloses the infectious organism even when culture of the partially treated cornea itself is not revealing.

Judicious Use of Topical Steroids

Use of topical steroids in the treatment of bacterial keratitis, while eternally controversial, is nonetheless almost always appropriate. Again, steroid use must be tailored to the individual circumstance. For minor peripheral ulcers, which may even have been caused by Staph hypersensitivity, topical steroid in limited doses (e.g. prednisolone acetate 1% qd or bid) may be appropriate on treatment day one or certainly on day two. For most all others, once the stromal infiltrate has been contained, typically within 3 to 5 days, similar steroid dosing is important for improving patient comfort, decreasing inflammation, speeding corneal epithelial healing and infiltrate resolution, and reducing the risk of inflammatory stromal scarring.

Figure 5. Intense microbial stromal infiltrate (left) accompanied by extensive and persistent epithelial defect in a neurotrophic cornea.

The only situations in which I am reluctant to use steroids are when the offending bacteria is likely gram negative (and hence potentially reactivated by steroid) or when there is the suspicion/risk of infectious involvement by an additional occult organism, perhaps an undiagnosed virus, fungus or Acanthamoeba masquerading as a bacterial infection. In these situations, I withhold steroids until culture or confocal microscopy confirms the causative culprit and thus that steroid use might be safe. In such situations, to be sure, the usual caveats of steroid side effects mandate even closer monitoring.

As previously emphasized, the success of bacterial kerati-tis management depends not only on prompt diagnosis but also aggressive management that may require hourly administration of topical medication(s) and daily follow-up for a week or more. Indeed, the patient must become a full partner in the treatment process, understanding the benefits of treatment compliance (no matter how much treatment may resemble water torture) and the sight-threatening risks of inadequate treatment. For patients of certain age, abstention from contact lens wear may also seem like cruel and unusual punishment. They can, however, be reassured that a cautious return to lens wear (utilizing new lenses, cases and solutions of course) will be possible as soon as inflammation reduction and epithelial wound closure have occurred, typically within 1 to 2 weeks.

Calling the patient's attention to patient counseling information included in medication package inserts is also beneficial. For example, they should know to wash their hands before using the drops and to avoid contaminating the bottle tip with material from their eyes, fingers or other source. Finally, as some patients may experience ocular stinging with eyedrop use, it is helpful to inform them that keeping the drops refrigerated or on ice can make them more comfortable.

Apart from the rise of contact lenses as a major risk factor, there is little new under the sun with respect to the clinical causations and presentations of bacterial keratitis. What has changed, however, are the fluoroquinolone antibiotics, which have by way of five generations of evolution continued to enable the drugs to remain one step ahead of the bugs. Thus, as the opening fractured metaphor suggests, the use of much new wine in old battles is certainly a self-evident theme.

References

1. Asbell PA, Colby KA, Deng S, et al. Ocular TRUST: Nationwide antimicrobial susceptibility patterns in ocular isolates. Am J Ophthalmol 2008;145(6):951-958.
2. Haas W, Pillar CM, Zurenko GE, et al. Besifloxacin, a novel fluoroquinolone, has broad-spectrum in vitro activity against aerobic and anaerobic bacteria. Antimicrob Agents Chemother 2009;53(8):3552-3560.
3. Tepedino ME, Heller WH, Usner DW, et al. Phase III efficacy and safety study of besifloxacin ophthalmic suspension 0.6% in the treatment of bacterial conjunctivitis. Curr Med Res Opin 2009;25(5):1159-1169.
4. Ormerod LD, Heseltine PNR, Alfonso E, Becker MI, Kenyon KR, Baerveldt G, Smith RE. Gentamicin-resistant pseudomonal infection: Rationale for a redefinition of ophthalmic antimicrobial sensitivities. Cornea 1989;8:195-199.
5. Schein OD, Hibberd PL, Starck T, Baker AS, Kenyon, KR. Microbial contamination of in-use ocular medications. Arch Ophthalmol 1992;110:82-85.

Kenneth Kenyon, MD, is an Associate Clinical Professor at the Harvard Medical School, Senior Surgeon at the Massachusetts Eye & Ear Infirmary, and Senior Clinical Scientist at the Schepens Eye Research Institute. He practices at Eye Health Vision Centers in Massachusetts and Cornea Consultants internationally. Dr. Kenyon can be reached at kenrkenyon@cs.com or (508) 994-1400.