Discussion

Do most practitioners treat suspected endophthalmitis with intravitreal injection of more than 1 antibiotic?

David R. Hardten, MD: Yes, and most are administered as separate injections. The pharmacy may supply these drugs in a syringe containing 0.5 mL, so 0.4 mL is discarded to leave 0.1 mL containing the proper individual dose of 1 mg.

Richard L. Lindstrom, MD: Often, in areas where specialty practices are abundant, ophthalmologists may refer patients for injection if a retina-vitreous office is conveniently located. However, many ophthalmologists are in a setting where they initiate the therapy themselves.

Why are the outcomes of European studies different from those in the United States? For example, the Moran Eye Center study showed topical antibiotics have value, which was not supported by the Swedish and ESCRS studies?

Edward E. Manche, MD: One of the concerns about the ESCRS study was the high baseline rate of endophthalmitis, which was approximately 1 in 300. This rate is higher than anything that has ever been reported in the United States. The baseline rate in the study in Sweden was about 1 in 2,000; that is, they had a low rate of endophthalmitis. If a practitioner is operating under poor conditions, intracameral cefuroxime should probably be used.

Lindstrom: However, it is confusing because even with the high endophthalmitis incidence, levofloxacin treatment did not show any value. Perhaps one explanation is the different agents used. The US studies that showed value used moxifloxacin and gatifloxacin, while the ESCRS study used levofloxacin.

Hardten: In addition, in the ESCRS study most patients were patched overnight, so they had no antibiotics from the time of surgery until the next day when the patch was removed, at which point the drops were initiated. Most practitioners in the United States do not patch their patients, and start the antibiotics immediately postoperatively. I believe this practice is a major contributor to our lower endophthalmitis rates in the United States.

Lindstrom: Typically, I do not see the patient after day 7. Because 20% of the patients who get endophthalmitis are diagnosed between day 7 and day 14 postoperatively, like many other practitioners, I also increased my duration of therapy to 14 days.

What counseling do you give patients regarding symptoms to watch out for or that require a return office visit?

Hardten: Part of typical counseling is, "If there's a problem, come back." That is the generic philosophy of it. More importantly, what do your staff on the phone do when the patient calls into the office and says, "I think I have a problem; I had cataract surgery." Just because they are 10 days or 2 weeks postoperative, the redness, irritation, and light sensitivity should not automatically be blamed on dry eye syndrome or the prescription eye drops. These patients should visit the office, as endophthalmitis can develop that long after surgery.

The patient is probably aware of something going on, but they may be prevented from making a critical follow-up visit if they are told, "Well, don't worry about it; if it doesn't get better, come on back in." This response may be acceptable if the problem is intermittent, but if the patient is having a progressive issue, they must be examined in the office. Therefore, for most of us, it is essential to properly train the phone staff, the call center, or the front desk — whoever takes those calls — to be diligent in identifying cases that warrant a follow-up office visit.

Lindstrom: I use the mnemonic "RSVP," which, appropriately, means "respond back." When? "R" for when the eye gets red, "S" for light sensitivity or photophobia, "V" for loss of vision, and "P" for pain. This is something I encourage my keratoplasty patients to remember and follow. They can be counted on to come in if there is a rejection, but they need to be alert for other indications that there could be a problem that is worthy of making an office visit.

If a patient's preoperative culture is MRSA-positive, how does that change your prophylactic strategy?

Jay S. Pepose, MD, PhD: Testing a confirmatory nasal swab on some of these patients may be warranted. If the swab is positive, bacitracin or mupirocin are sometimes used. The patient should be counseled that the risks are higher than for other patients. Some patients should be sent to an infectious disease specialist before surgery to reduce the chances of autoinfection.

Lindstrom: I am an advocate of trimethoprim/polymyxin B ophthalmic solution. This treatment is inexpensive and very effective, which may be because it has not been used extensively. I agree that bacitracin ointment also is effective.

How do you handle povidone-iodine allergy?

Manche: If the patient is allergic to povidone-iodine, chlorhexidine could be used, as was used in the Sweden study. That is what we do at our center for patients with povidone-iodine allergies. When we use chlorhexidine, we just clean the lids with it and keep it out of the eye.

Hardten: These compounds are formulated as both solutions and soaps. The chlorhexidine and povidone-iodine soaps are extremely toxic, so it is very important to not use the soaps around the eyes.

If a person is allergic to lobster and shrimp will they also be allergic to povidone-iodine?

Hardten: Shrimp and iodine are not the same allergies, which is also true for IV contrast dye. A person can be allergic to any of the shellfish or to a radioactive dye, and they will not necessarily be allergic to povidone-iodine. Therefore, although there are some people who have contact allergy to povidone-iodine, it is not related to having an allergy to crustaceans or to IV contrast dye.

How do you handle the allergies to fluoroquinolones?

Manche: I use trimethoprim/polymyxin B in those cases.

Hardten: First, I ask what their allergy is, because often the allergy symptoms they report are nausea and stomach upset. If the patient was on 5 different medicines, it may be difficult to know what caused the allergic reaction. Next, I record the allergy in the chart where it will be noticeable, because the pharmacist will call for all events (eg, if the patient vomited because they took ciprofloxacin on an empty stomach). As one of your staff will most likely take the pharmacist's call, if it is written on the chart the proper actions can be taken.

Pepose: I have tested the product on some patients in the office if I am not sure that there truly is an allergy. If they report historically having a mild reaction, for example hives or an upset stomach, I may put a drop in their eye in the office. That is the best place for them to have a reaction. After a few minutes of observation, many of these patients display no sensitivity reactions.

Do you presume endophthalmitis following intravitreal injections is infectious? Would you recommend empirically treating it? Or do you culture?

Hardten: The recommendations are basically to culture and inject. Even though there is a high incidence of gram-positive organisms, culture for early identification is critical. Also, in that setting, if the practitioner does not prepare for an injection as he does for a surgical procedure, he might not be as diligent cleaning the lids and lashes and performing other antiseptic preparations. This is a situation that requires careful attention, particularly when the injections will be given repeatedly. In addition, the lids and lashes should be carefully inspected to ensure that blepharitis is controlled.

How do you deal with situations where a patient cannot afford an expensive newer fluoroquinolone, or it is not available on their insurance plan?

Manche: If it is a hardship case I give the patients samples. If the newer fluoroquinolone is not in their insurance provider's formulary and the patient does not want to pay for it themselves, I write an alternative prescription, for example trimethoprim/polymyxin B.

Hardten: I have a substitution list, and discussion with the patient will include telling them that the substitution probably is not as effective, but that we do not have solid proof. I let the patient go down the list to whatever drug they want, and work diligently, as always, to prevent endophthalmitis intraoperatively.

Lindstrom: We allow our nursing personnel to substitute, so I have a list of 3 antibiotics, a list of 3 NSAIDs, and a list of 3 steroids. Patients that need a substitute can go to number 2 or number 3 on the list. Some practitioners worry about doing that because of informed consent, and will tell the patient that the substitute drug may not be as effective. This is a difficult situation that we all deal with.

What do you do differently for a patient who is at higher risk for infection and macular edema (eg, one who has vitreous loss, which is commonly seen)?

Hardten: Whenever I patch the patient I give a subconjunctival injection of steroid and antibiotic. If I had inadvertent vitreous loss intraoperatively, I give a subconjunctival injection and consider giving an intracameral injection. I then have them patched for 3 or 4 hours, take off the patch, and start the drops the same day.

Pepose: In those situations I maintain the patient on drops for a longer period.

Manche: Practitioners who do not routinely use intracameral antibiotics should consider them in high-risk cases, as there is a lot of support in the literature for their use, especially if you break the capsule, because the relative risks increase fairly dramatically with that.

Pepose: When using triamcinolone suspension to visualize the vitreous, some triamcinolone may remain, which may provide some anti-inflammatory benefit also.

What doses and volumes are used for intracameral injections?

Hardten: The desired volume of moxifloxacin can be taken out of the bottle. For vancomycin, I use 0.05 cc of the 10 mg/mL formulation. For cefuroxime, which is 125 mg/mL, I use 0.1 cc.

Do you put in a suture if a complication such as vitreous loss occurs?

Hardten: If the wound is self-sealing, I do not put in a suture.

Manche: Nor do I.

Pepose: If indicated by the results of a Seidel test or if I had to enlarge the wound, I may put in a suture.

Lindstrom: I definitely suture if I drop a piece of nucleus or something occurs that will require an immediate referral to a vitreal retinal surgeon, but I do not suture for a capsular tear or vitreous loss.