What you need to know to avoid the most dreaded complication of intraocular surgery.
Mark Speaker, MD New York City
When you consider the effect endophthalmitis can have on the eye, an ounce of prevention is worth many pounds of cure. In this article, I'll explain the best ways, both before and during surgery, to reduce or eliminate the risk of a patient developing this dreaded complication.
Preoperative Preparations
Fortunately, the incidence of endophthalmitis after cataract surgery
is very low, probably less than 1 percent.1 Most cases are caused
by a patient's own ocular flora. The organisms usually responsible
are Staphylococcus epidermidis, Staphylococcus aureus and
streptococcal species. To combat them, bacteriocidal antibiotics,
primarily quinolones, are preferred, since they kill them outright.
Bacteriocidal agents, such as Polytrim, don't eliminate bacteria,
they only inhibit their growth.
One of the most important steps toward avoiding endophthalmitis is catching any signs or symptoms of external disease. If the patient has severe infectious blepharitis or conjunctivitis, you'll need to eliminate the inflammation before surgery to decrease the risk of infection.
Lacrimal obstruction can also be a concern. These obstructions tend to be colonized primarily with pneumococcal bacteria, so lacrimal surgery prior to cataract surgery is usually in order to decrease the endophthalmitis risk.
Another issue arises in patients with prosthetic eyes. Since these patients are monocular, avoiding endophthalmitis in them is crucial. What makes this challenging, however, is that the prosthetic orbit is usually colonized with virulent bacteria, necessitating removal of the prosthesis and culture of the eye. Any virulent organism should be treated for up to a week before surgery, and then the eye should be recultured to ensure no further infection exists.
There is also the potential for glaucoma drops, artificial tears and non-preserved saline drops to be contaminated. If a patient is using any of these, examine him for signs of inflammation. If any signs are present, it's probably best to switch to a fresh bottle of drops and culture the patient's current bottle to identify any potential pathogens.
In addition to treating external disease and the other preop preparation, if you're preparing to perform a filtering procedure for a glaucoma patient, it may be best to plan the procedure so that the filter isn't placed inferiorly. In our practice, we've found that filters placed inferiorly are more likely to become infected than filters placed at other locations. We think this may be due to an increased breakdown of the surface of the bleb in that position that allows organisms access to the tissue. Also, patients who have leaky blebs after surgery may need to be continued on topical antibiotics for a longer period than usual.
As a preop measure, I prescribe quinolone drops to take six times the day before surgery; then I give them a drop that morning. They get the final preop drop when they arrive for the procedure.
Intraoperative Measures
For the procedure itself, consider several points.
All preop sterility procedures are important, especially those that pertain to surgical masks. Make sure your staff is aware of all standard infection control procedures; we've had several instances in which we could directly document the transmission of bacteria from a nurse to a patient. In one case, an asthmatic nurse who coughed a lot without changing her mask touched off a mini-epidemic of endophthalmitis.
For patient preparation, instilling a couple drops of 5% filtered iodine and a quinolone drop can decrease the amount of external flora that could potentially cause endophthalmitis. After the skin prep is complete, we rinse off the iodine with saline.
The four primary intraoperative issues that can impact a patient's risk of endophthalmitis are wound leaks, ruptures in the posterior capsule, phaco burns and the prophylactic use of antibiotics in the BSS.
*Wound leak. A leaky wound is a sign of poor closure, which means there's an opening through which external bacteria can enter the eye. You can help head off this problem by using an incision that's no longer than needed for phacoemulsification. This is applicable for most incision types, including paracentesis-style incisions, a stepped-tunnel or Langerman hinged-style incisions. Also, try not to damage the integrity of the internal lip with the surgical instruments.
If a leak is significant enough to cause iris prolapse or a persistent shallow chamber, you may want to partially suture the incision or close it altogether and move to a new incision site. Finally, at the close of the case, if the leak persists, a suture may be necessary for adequate wound closure.
* Posterior capsule rupture. A capsule rupture can expose the entire eye to infection. To keep capsule ruptures to a minimum, here are some tips.
* Perform a thorough hydrodelineation to separate the cortex from the capsule and lessen your risk of tearing.
* Avoid using high vacuum settings when working near the capsule. This decreases the risk of the capsule "jumping" into the phaco probe.
* Also, since many capsule ruptures occur during cortical removal, perform this step with care, and use a manual I/A technique for difficult-to-remove cortex.
* Phaco burns. A wound burn can affect the healing of the incision, risking infection. To decrease the chance of wound burn:
* Incisions smaller than 3 mm can be prone to friction from the motion of the phaco needle, as well as heat transferred from the compression of the silicone sleeve. To avoid this, start out with a lower phaco power setting, and attempt to minimize the to-and-fro motion of the needle. A smaller diameter needle may help minimize the risk of burn, since it will make less contact with the incision. Remember, however, that the volume of BSS moving through these smaller needles actually decreases and provides less cooling.
* Remember the first sign of an impending phaco burn is the appearance of "lens milk" stagnating by the phaco tip. If you see this you should immediately stop phaco until you find the problem.
* Antibiotics in the BSS. A somewhat controversial prophylactic strategy involves loading the surgical BSS bottle with antibiotics. Though this practice is common, its safety and efficacy haven't been established by anything other than anecdotal reports.
In one report, Tarpon Springs, Fla., surgeon, James Gills, reported no cases of endophthalmitis in 25,000 surgeries since he began using Vancomycin in the bottle.2 In another relatively anecdotal report, researchers found that omitting antibiotics in the infusion bottle increased the odds of having a positive culture in the aqueous by a factor of 2.5. 3 However, a positive culture of the aqueous does not necessarily mean the patient will go on to develop endophthalmitis. For example, a French study that found two cases of endophthalmitis in a group of 15 patients with positive cultures reported no cases in the 47 patients who cultured negative.4
Unfortunately, though some uncontrolled reports seem to support the efficacy of this strategy, certain powerful antibiotics, by their very nature, may not be very effective at all.
For instance, aminoglycocides such as gentamicin are effective against gram-negative organisms. However, 90-95 percent of endophthalmitis cases are caused by gram-positive organisms, which are poorly treated by aminoglycocides. The other popular antibiotic for placement in the infusion is Vancomycin, which is effective against gram-positive organisms. However, Vancomycin needs several hours to do this for some bacteria, but it and other antibiotics only stay in the anterior chamber one or two hours before draining through the trabecular meshwork.5
Another source of problems is the lack of a generally accepted method for formulating the BSS/antibiotic solution. At high enough levels, gentamicin can cause macular infarction and retinal ischemia, and, to make matters more prickly, the therapeutic and toxic levels of the agent are actually fairly close together. Also, since different surgical cases use different amounts of infusion, putting the correct ratio of antibiotic necessary to provide safe, but effective, action can be akin to hitting a moving target.
Also, hovering over all of these concerns is the specter of antibiotic resistance. Though Vancomycin is the best agent for use against gram-positive organisms, it's our only defense against the methicillin-resistant Staph aureus and enterococcus. However, if surgeons regularly use this powerful agent in the cataract infusion bottle, there is a definite increase in the risk of antibiotic resistance. In fact, in 1995, the Centers for Disease Control and Prevention published a warning that Vancomycin shouldn't be used for prophylaxis.
In our practice, the most effective preventive measure is topical ciprofloxacin and povidone iodine.
Postop Prevention
After surgery, I usually recommend the brief use of a quinolone,
due to its spectrum of activity and excellent penetration. Though
some surgeons might reach for the less expensive aminoglycocides,
such as Tobradex, the incidence of allergy and their spectrum
of activity don't compare favorably with the quinolones. Also,
I believe if you use it briefly but intensely, the risk for resistance
is reduced.
My postop quinolone regimen consists of doses four to six times a day for three days, with a longer course if there are any wound abnormalities. However, if the wound is still leaking after three days, it should be closed surgically. I haven't found strong evidence arguing for the injection of a bolus of antibiotic postoperatively. Neither the microbiological efficacy nor the safety of a post-op bolus has been demosntrated and there's always a risk for dilution or dosing errors that could lead to toxicity.
If your patient has diabetes or has had previous vitrectomy surgery, he's at an increased risk for endophthalmitis. Therefore, it's prudent to put these patients on some form of oral antibiotic prophylaxis immediately after surgery. For my patients, this means a quinolone such as ciprofloxacin 750 mg PO bid for five days to a week postoperatively. If the patient had a capsule rupture or any vitreous manipulation during surgery, I recommend ciprofloxacin 500-700 mg PO bid for five to seven days.
Endophthalmitis is a complication every surgeon keeps in the back of his mind, hoping he doesn't encounter it, but preparing for the time when he does. Taking the measures outlined above can go a long way toward keeping the shadow of endophthalmitis from ever darkening your door.
Dr. Speaker is an associate clinical professor of ophthalmology at New York Medical College.
1. Kattan HM, Flynn HW, Pflugfelder SC, et al. Nosocomial endophthalmitis
survey. Current incidence of infection after intraocular surgery.
Ophthalmology 1991;98:227-38.
2. Gills JP. Filter and antibiotics in irrigating solution for
cataract surgery. J Cataract Refract Surg 1991;17:385.
3. Ferro JF, de-Pablos M, Logrono MJ, et al. Postoperative contamination
after using vancomycin and gentamicin during phacoemulsification.
Arch Ophthalmol 1997;115:165-170.
4. Pospisil A, Pospisil F, Dupont MJ, Delbosc B, Montard M. Contamination
bacterienne de la chambre anterieure et chirurgie de la cataracte.
J Fr Ophthalmol 1993;16:10-13.
5. Ferro JF, de-Pablos M, Logrono MJ, et al. Postoperative contamination
after using vancomycin and gentamicin during phacoemulsification.
Arch Ophthalmol 1997;115:165-170.
The Best Place to Kill Bacteria?
A study presented at last year's meeting of the Association for Research in Vision and Ophthalmology (ARVO) found that the best place to prevent endophthalmitis may be on the ocular surface. If surgeons try to kill bacteria within the eye, drug concentrations may not be high enough.
In the study, researchers cultured flora from the inferior bulbar and palpebral conjunctiva of 20 eyes of patients at least 55 years of age. In the treatment group, the cultures were performed before an instillation of ciprofloxacin 0.3% one drop every five minutes x 3. Then, after all the drops had been instilled, researchers repeatedly cultured the eyes of both groups at 15 minutes, 30 minutes, one hour and two hours. They compared the average reduction in colony forming units in the two groups at each time point. The results appear in Table 1. As the table shows, there was a greater reduction of units at each time point in the treatment group.
"When we're talking about preventing endophthalmitis," says Wilmer Eye Institute's Harold Katz, MD, who took part in the study, "killing bacteria on the surface is important before they enter the eye." He cites three studies that suggest the surface is probably the best place for prevention. One study showed that 5% povidone iodine instilled in the cul-de-sac reduced the incidence of endophthalmitis, but didn't penetrate the eye.1 And killing this surface flora will probably be key in preventing infections, according to another study that compared bacteria from an infected eye to the bacteria on the eye's surface. In most cases, they were identical.2
In the third study, surgeons studied the penetration of antibiotic into the anterior chamber, and found that the concentrations were extremely low, well below that necessary to reach the minimum bacteriocidal concentration.3
Says Dr. Katz, "We know that we're achieving only minimal levels in the eye, that 5% povidone iodine on the surface reduces the incidence of endophthalmitis and that the organisms that cause endophthalmitis are genetically identical to the organisms on the patient's own surface flora. Therefore, killing bacteria on the surface is important."
1. Speaker MG, Menikoff JA. Prophylaxis of endophthalmitis
with topical povidone-iodine. Ophthalmology 1991;98:1769-7.
2. Speaker MG, Milch FA, Shah MK, Eisner W, Kreisworth BN. Role
of external bacterial flora in the pathogenesis of acute postoperative
endophthalmitis. Ophthalmology 1991;98:639-49.
3. Donnenfeld ED, Schrier A, Perry HD, Aulicino T, Gombert ME,
Snyder R. Penetration of topically applied ciprofloxacin, norfloxacin,
and ofloxacin into the aqueous humor. Ophthalmology 1994;101:902-5.
Table 1: Reduction in Colony-Forming Units
| TREATMENT | CONTROL | |
| 15 mins. | 97.9 percent | 0.2 percent |
| 30 mins. | 98.0 percent | 15.5 percent |
| 1 hr. | 98.0 percent | 36.6 percent |
| 2 hrs. | 96.3 percent | 7.96 percent |