Progress in Understanding Prophylaxis for Endophthalmitis

Recent publications have helped to clarify the role of topical antibiotics in the prophylaxis of endophthalmitis after cataract surgery. Administering topical antibiotic drops, in addition to standard povidone-iodine antisepsis, has become a standard of care in these prophylactic regimens. The need to cleanse the ocular surface and reduce or eliminate surface microflora was emphasized in studies that showed these surface microorganisms were often identical to strains isolated in endophthalmitis cases. Using genetic analysis, investigators found that organisms recovered from the vitreous were genetically identical to isolates from patients’ own eyelids, conjunctiva, or nose.¹ These findings support the need for surgeons to pay close attention to external surfaces of the eye as potential sources of contamination in prophylactic regimens. However, the development of newer antibiotic agents, that offered better corneal penetration and a broader spectrum of action raised interest in the ability of these agents to also provide protection within the eye, in the anterior segment.

STERILIZATION OF THE OCULAR SURFACE
Newer antibiotics, such as the fluoroquinolones, offer advantages over older classes of antibiotics, such as aminoglycosides, by virtue of the former’s broader spectrum of action and relatively low tissue toxicity. Investigators were eager to test the ability of these new agents to eliminate bacteria, not only from the ocular surface prior to surgery, but also perioperatively by examining penetration to inner ocular structures, particularly the aqueous humor. By culturing external ocular surfaces, and measuring aqueous humor penetration, a number of studies have examined the value of adding fourthgeneration fluoroquinolone drops to povidone-iodine in varying topical drop regimens.

Despite increasing doses, He and associates found that no further reduction of conjunctival flora was achieved with a 3-day versus 1-day q.i.d. dosing regimen of moxifloxacin drops.² Another study found similar results using gatifloxacin drops, where q.i.d. dosing for 3 days preoperatively, along with povidone-iodine, showed no advantages over povidone-iodine preparation alone.³ In fact, a 4% to 8% positive conjunctival culture rate remained. Bucci and associates reported that, when both of these fluoroquinolone drops were administered more intensively just prior to surgery (one drop every 10 minutes, for four doses) in addition to a 2-day regimen, the study group receiving more intensive treatment surprisingly showed the higher rate of aqueous humor contamination. Furthermore, complete eradication of bacteria from the conjunctiva or eyelids was not achieved.⁴ For historical perspective, irrigating solutions have also not guaranteed sterility. After use of an irrigating solution with the combination of vancomycin plus gentamicin, a 5% anterior chamber contamination rate remained.⁵

The fact that external flora persist as potential contaminants, despite preoperative antisepsis, is alarming, particularly considering the potential for inflow of extraocular fluid after wound hydrosealing and sutureless corneal incisions, as shown by Herretes and associates.⁶

SAMPLING OF AQUEOUS HUMOR
Studies over recent decades revealed that aqueous humor samples taken at the time of surgery showed persisting rates of bacterial contamination in at least 2% to 40% of cases, despite the use of prophylactic measures that included povidone-iodine.^7-9 Therefore, there is valid interest in our ability to deliver effective antibiotic levels to the anterior chamber, the surgical site during cataract surgery. Topically applied preoperative drops remain a primary focus for delivering these agents, although direct intracameral injection is shown to be clinically effective and superior to drops, as discussed later.

To measure the aqueous humor penetration of antibiotics after the instillation of topical drops, researchers use a familiar study model. Drops are given in various regimens preoperatively, and prior to the surgical incision, aqueous humor is sampled and assayed for antibiotic concentration. A great many preoperative dosing regimens have been tested in this manner, from intermittent dosing for several days preoperatively to frequently pulsed drops just before and after surgery. Results show that some drug penetration does indeed occur and that an increase in measured levels is possible by manipulating the frequency of drop administration. Table 1 displays some reported aqueous humor levels of fluoroquinolones after topical drop instillation. In these studies, inherent differences were present such as dosing regimen, drop conconcentration, drug lipophilicity/ penetration characteristics, or presence of benzalkonium chloride, yet levels did not exceed the range of 1 to 4 &mul;g/mL.

ASSESSING THE DATA
Several factors should be borne in mind when evaluating these reported levels of aqueous humor antibiotic concentrations. First, the levels of aqueous humor reported are means from patient study groups. A closer look at the standard deviations, when they are reported, demonstrates a huge interpatient variability. This naturally follows the very large interpatient variability that occurs to the tears after drop instillation. Table 1 displays some of these ranges, where standard deviations higher than 50% are commonplace. Since these mean concentrations are already low, the likelihood that a percentage of patients will experience even lower levels, due to the large variability, makes this form of dose delivery very imprecise.

In terms of the actual levels reported, what can eye care practitioners expect regarding the potential for bacterial killing in the aqueous humor? Reported aqueous humor antibiotic levels were regarded as sufficient if they exceeded the minimum inhibitory concentrations (MICs) of some bacterial strains. An assumption followed that these levels were adequate and would be clinically effective in eradicating bacteria that gain entry into the eye. However, the MIC is a definition of bacterial susceptibility or resistance generated by laboratory standards that involve incubation times near 24 hours. Therefore, any measured level must be correlated to the amount of time that microbes and antibiotic are in contact to assess the potential for bacterial eradication.

In real-life situations, these reported aqueous humor levels are not only relatively low but are also shortlived because of the approximately 2-hour aqueous turnover rate. Assuming a half-life in aqueous humor of 1 to 2 hours, a peak drug level of 1 mg/mL, for example, would become 0.5 μg/mL in 1 to 2 hours; then 0.25 μg/mL in 2 to 3 hours; then 0.125 μg/mL in 3 to 4 hours; 0.06 μg/mL in 4 to 5 hours, and so forth. Therefore, achieving a peak aqueous humor concentration of 0.5 μg/mL cannot be construed as an effective MIC if that antibiotic level in the aqueous humor is not sustained for a necessary period of time.

BACTERIAL KILLING RATES IN AQUEOUS HUMOR
Two recent studies assessed bacterial killing after the shorter time frames more akin to topical drop dosing in the eye (Figures 1 and 2). Callegan and associates used the full-strength commercially available moxifloxacin and gatifloxacin drop concentrations and evaluated bacterial killing over 5 to 60 minutes.¹⁰ Hyon and associates performed a similar study that also looked at benzalkonium chloride independently.¹¹ Both studies showed that, even at concentrations of 3,000 and 5,000 μg/mL (the equivalent of the 0.3%- and 0.5%-drop concentrations), poor bacterial eradication was accomplished over shorter periods of time, and resistant organisms often persisted beyond the study period. Worth remembering is that the approximately 1- to 4-μg/mL fluoroquinolone concentrations found in aqueous humor after the administration of topical drops are approximately 1,000 times lower than the concentrations used in these studies. Therefore, the likelihood of effecting bacterial kill at these infinitesimally smaller concentrations, shortlived as they are in the aqueous humor, is somewhat remote.

An additional precaution in relating MICs to effective prophylaxis in aqueous humor is being aware of current MICs that accurately reflect the MICs (and MIC₉₀) of isolates taken from clinical cases of endophthalmitis and not from arbitrarily chosen “library” strains. Trends in actual endophthalmitis isolates reflect the true challenges for prophylactic regimens. When study discussions select only bacterial strains with low MICs, or strains with known high susceptibility, then conclusions can be skewed and may be a disservice to the daily needs of the clinician.

A final but important factor is that the aqueous humor antibiotic levels (Table 1) described after preoperative doses are expelled with the surgical incision, so even these low levels are no longer present during the operative and perioperative period. If meaningful antibiotic levels are expected to be present after the surgical procedure (that is, immediately after surgery and during the next week or 2 when patients self-administer drops), then the potential of topical drop regimens to achieve effective levels during these time periods should be viewed separately. Postoperative drop regimens are usually less vigorous than the preoperative regimens tested and are accompanied by the drawbacks of patients’ selfadministration and compliance.

Certainly, the anterior chamber may be capable of clearing a small inoculum of bacteria. However, as bacterial strains grow increasingly resistant, as the spectrum of bacteria identified in endophthalmitis isolates changes, and as the population of cataract patients ages, a closer look is warranted at what should reasonably be expected from topical drop regimens as effective prophylaxis at the aqueous humor site.

CLINICAL EVIDENCEZ
Only one study has offered a prospective analysis of clinical outcomes after the use of topical drops for the prophylaxis of postoperative endophthalmitis. The ESCRS multicenter study of more than 16,000 patients allows comparison of outcomes among four treatment groups, including a control group and patients receiving frequently applied topical fluoroquinolone drops (levofloxacin 0.5%)—as well as patients receiving an intracameral injection of cefuroxime 1 mg.¹² The study group receiving the pulsed-dosing regimen of a topical fluoroquinolone did not achieve a statistically significant reduction in postoperative endophthalmitis rates compared with controls, although somewhat lower absolute rates were seen (Figure 3). While statistical powers were aimed primarily at evaluating the effects of the intracameral injection, the effect of topical drops, as applied in this study, versus controls could be observed in a subgroup of about 4,000 patients each. Here, the lack of statistical difference is of clinical interest and highlights the need to better define where and how topical antibiotics may best exert their effects. In contrast, the intracameral injection unequivocally reduced endophthalmitis rates by approximately fivefold.

An independent study reproduced the identical fluoroquinolone dosing regimen used in the ESCRS study. Investigators measured aqueous humor antibiotic levels over a 90-minute postdose period in pooled patient groups.¹³ This dosing regimen produced the highest reported levels of fluoroquinolone in aqueous humor reported to date (4.4 μg/mL, Table 1); the regimen consisted of two preoperative drops (one each at 60 and 30 minutes preoperatively) plus three pulsed drops (one every 5 minutes at the end of surgery). The pulsed drops also served to replace levels lost at the beginning of surgery. In contrast to topical drops, the intracameral injection (1 mg or 1,000 μg cefuroxime in this case) is capable of delivering a much higher dose instantaneously to the aqueous humor, since drops must first diffuse through several corneal layers.

CONCLUSION
Recent evidence, taken together from reported antibiotic levels in the aqueous humor, studies that examined bacterial killing with high antibiotic concentrations over shorter periods of time, and clinical findings from a large multicenter trial, suggest we should take a closer look at expected antimicrobial effects in the aqueous humor after topical drop administration. Because these reported levels are consistently low, future investigations may benefit from closer scrutiny of antibiotic actions on the ocular surface, as compared with internal sites such as the aqueous humor, for the prophylaxis of postoperative endophthalmitis.

This article is reprinted with permission from Advanced Ocular Care’s January/February 2011 issue.

Susanne Gardner, PharmD, is an independent medical writer, educator, and researcher based in Atlanta. She has served as a writer for the pharmaceutical industry and as a consultant to Bausch + Lomb. Dr. Gardner may be reached at sg.otm@bellsouth.net.

1. Speaker MG,Milch FA,Shah MK,et al.Role of external bacterial flora in the pathogenesis of acute postoperative endophthalmitis.Ophthalmology.1991;98:639-649.
2. He L,Ta CN,Hu N,et al.Prospective randomized comparison of 1-day and 3-day application of topical 0.5% moxifloxacin in eliminating preoperative conjunctival bacteria.J Ocul Pharmacol Ther.2009;25:373-378.
3. Moss JM,Sanisio SR,Ta CN.A prospective randomized evaluation of topical gatifloxacin on conjunctival flora in patients undergoing intravitreal injections.Ophthalmology.2009;116:1498-1501.
4. Bucci FA Jr,Amico LM,Evans RE.Antimicrobial efficacy of prophylactic gatifloxacin 0.3% and moxifloxacin 0.5% in patients undergoing phacoemulsification surgery.Eye Contact Lens.2008;34:39-42.
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.
6. Herretes S,Stark WJ,Pirouzmanesh A,et al.Inflow of ocular surface fluid into the anterior chamber after phacoemulsification through sutureless corneal cataract wounds. Am J Ophthalmol 2005;140:737-740.
7. Sherwood DR,Rich WJ,Jacob JS,et al. Bacterial contamination of intraocular and extraocular fluids during extracapsular cataract extraction.Eye.1989;3:308-312.
8. Dickey JB,Thompson KD,Jay WM. Anterior chamber aspirate cultures after uncomplicated surgery.Am J Ophthalmol. 1991;112:278-282.
9. Bausz M,Fodor E,Resch MD,et al. Bacterial contamination in the anterior chamber after povidone-iodine application and the effect of the lens implantation device. J Cat Refract Surg.2006;32:1691-1695.
10. Callegan MC,Novosad BD,Ramadan RT,et al.Rate of bacterial eradication by ophthalmic solutions of fourthgeneration fluoroquinolones.Adv Ther.2009;26:447-454.
11. Hyon JY,Eser I,O’Brien TP.Kill rates of preserved and preservative-free topical 8-methoxy fluoroquinolones against various strains of Staphylococcus. J Cat Refract Surg.2009;35:1609-16013.
12. ESCRS Endophthalmitis Study Group:Prophylaxis of postoperative endophthalmitis following cataract surgery: results of the ESCRS multicenter study and identification of risk factors.J Cat Refract Surg.2007;33:978-988.
13. Sundelin K,Seal D,Gardner S,et al.Increased anterior chamber penetration of topical levofloxacin 0.5% after pulsed dosing in cataract patients.Acta Ophthalmologica.2009;87:160-165.