Our challenge is to develop treatments -- including antibiotics, diagnostic techniques and therapeutic modalities -- that can deal with this problem.

The importance of patient history

Any patient must be evaluated for significant lid disease before cataract, corneal or intraocular surgery. However, when you're evaluating an infectious keratitis, it's crucial to determine where the bacteria originated and whether the patient is immunosuppressed, in addition to performing an examination.

The location in which the infection originated is important because organisms such as methicillin-resistant Staphylococcus are common in hospital or nursing home situations. For that reason, any recently hospitalized patient -- or a physician, nurse, or other hospital or nursing home employee -- has to be considered at risk for this type of bacteria.

Similarly, any immunosuppressed patient, such as someone taking corticosteroids or with immunodeficiency syndrome, should be considered at high risk for a resistant infection.

Treating high-risk patients

Most ophthalmologists treat an infectious keratitis with conventional therapy, using broad spectrum antibiotics such as a fluoroquinolone. Some doctors will add additional gram-positive coverage such as Neosporin. Most don't take cultures.

However, if you know that your patient is from a high-risk group and has significant lid disease, it's important to treat the lids prophylactically before considering refractive or intraocular surgery. You may want to bypass standard treatment in favor of antibiotics known to be more effective under these circumstances, such as:

• Bacitracin. Of the medications available commercially, bacitracin ointment is the best choice. Treat high-risk patients with bacitracin before performing surgery.
• Vancomycin. This is the best antibiotic for treating active methicillin-resistant Staphylococcus aureus infections, and I recommend its use for all hospital-aquired infections. (Note that the Center for Disease Control recommends that vancomycin be saved for active infections and not used routinely as an antibiotic prophylaxis.)

If your patient is from a high-risk group and you prefer to play it safe, perform preoperative lid cultures. Cultures will direct you to more specific and effective treatments. In some cases, however, even these treatments won't be sufficient.

Treating nonresponding infectious keratitis

The most important point to remember when you're managing infectious keratitis that isn't responding to conventional treatment is this: Don't keep treating without a proper diagnosis. Don't use "shotgun" therapy. If the drugs aren't working, you need to rethink your original diagnosis.

Sometimes, you can make an educated guess about the identity of the resistant organism. If a patient is only on ofloxacin (Ocuflox) or ciprofloxacin (Ciloxan), for example, the organism is probably Streptococcus or S. viridans.

However, the organism you're dealing with could be an opportunistic bacteria requiring specific therapy. Or, you may not be dealing with bacteria at all. The organism could be a fungus or Acanthamoeba. Under these circumstances, it's even more important to establish the proper diagnosis.

When general treatment fails, I often recommend that the patient be recultured. If it's a slow, indolent infection, try stopping therapy for 12 to 24 hours and then performing cultures. Look for atypical Mycobacteria, Actinomyces, or Nocardia.

When you perform the culture, make sure you culture on the appropriate media. Under normal circumstances, most doctors culture use thioglycolate, sabarauds, chocolate agar and blood agar. Unfortunately, many resistant bacteria won't grow on these cultural mediums. For that reason, it's important to use additional culture materials such as Lowenstein-Jensen or blood agar plates with an e-coli overlay for Acanthamoeba.

Scrapings and corneal biopsy

For those patients who aren't responding to therapy, you should also examine scrapings while waiting for the cultures to grow. If the results are indicative of a particular organism, you may want to take action immediately, perhaps by modifying the therapy while waiting for the culture results.

Keep in mind that any diagnosis made from the scrapings will only be as good as the microbiologist who reads them. Also, it's important that you play a part in these findings. (I'll often look at the scrapings myself to make an unusual diagnosis.) Talk with the pathologist and tell him what you're looking for so he can order the correct stains for the scrapings.

When cultures and scrapings fail to confirm a diagnosis, a corneal biopsy is usually the best followup strategy. A biopsy will often find things that other tests miss.

When resorting to a corneal biopsy, I try to find an active area of infection and employ a 2-mm disposable dermatology punch. I send half of the punch to pathology and spread the other half on different culture media.

Frequent treatment

The other important thing to remember when treating resistant organisms is that therapy should be on an extremely frequent basis. I recommend drops every hour during the day and throughout the night until the patient responds.

During topical treatment, oral therapy is sometimes helpful as well. However, it's not nearly as helpful as topical therapy.

Make sure the patient has a support group at home to ensure that he takes the medication as indicated, and that the patient will be responsible for his own therapy. (This is especially important with elderly patients.) If you can't be sure that your patient will receive the therapy according to your instructions, admit the patient to a hospital for therapy.

The last resort

Patients with severe infectious keratitis who don't respond to therapy and are progressing may require corneal transplantation. Unfortunately, therapeutic keratoplasties are heroic measures with a high incidence of graft failure. They should be reserved for those times when failure to perform surgery will jeopardize the integrity of the eye, such as when the infection has spread to the sclera, or when you find perforations into the anterior chamber.

If a corneal transplant is necessary, wait to perform the transplant until the infection/inflammation is controlled as much as possible (if circumstances allow). Avoid using anti-inflammatory corticosteroids as part of the treatment; they may suppress the immune response and allow the infection to recur in the inner cornea. Instead, try using topical cyclosporine. This helps to suppress the graft rejection without suppressing the body's ability to fight infection.

Fighting the good fight

The problem of bacterial resistance isn't likely to disappear any time soon. But with diligence and aggressive treatment, we can still keep keratitis under control. 

Dr. Donnenfeld is co-chairman of cornea and external disease at Manhattan Eye and Ear Hospital and North Shore University Hospital, Manhattan, N.Y.126