Endophthalmitis complicates approximately one in every 1,000 cataract operations. Risk factors cited in the peer-reviewed literature include extracapsular surgery, intracapsular surgery, clear corneal incisions, diabetes mellitus, prolonged surgical time, previous or concurrent trabeculectomy, repeated instrument entry and exit, chronic blepharitis, chronic conjunctivitis, keratitis sicca, ocular surface disease, capsular rupture, vitreous prolapse, and vitrectomy surgery. The potential for this risk may rise to one in every 100 cases with vitreous loss. Although rapid diagnosis and expeditious surgical intervention can preserve excellent visual function in many patients with endophthalmitis, preventive measures are the cornerstone of any surgical management strategy.
REVOLUTIONIZING ENDOPHTHALMITIS TREATMENT
The Endophthalmitis Vitrectomy Study (EVS), completed in 1995, found that 70% of endophthalmitis cases were caused by coagulase-negative, Gram-positive micrococci, overwhelmingly Staphylococcus epidermidis.1 This study has revolutionized our treatment algorithm for postcataract surgery endophthalmitis, recognizing the essential aspects of vitreous-tap diagnosis and expeditious injection of intravitreal antibiotics, while surprisingly raising the threshold for pars plana vitrectomy for patients with light perception or worse-quality vision.
Since the completion of the EVS, new data have extended our understanding of the pathogenesis and prevention of postoperative endophthalmitis. Postcataract infections originate by one of three routes: (1) introduction through instrumentation at the time of surgery; (2) inoculation through the wound after cataract surgery; and (3) (although extremely rarely) by endogenous spread from concurrently infected extraocular tissues, such as a tooth abscess or infected diverticulum. Material presented at the 2002 ARVO meeting in Fort Lauderdale, Florida, in particular offered insight into bacteriologic factors relevant to cataract surgery.
CURRENT DATA
Many surgeons believe that postcataract infections are introduced into the eye from the ocular surface. This belief brings into question the traditional use of topical perioperative aminoglycosides for cataract patients, especially when most endophthalmitides are Gram-positive and aminoglycosides are so insoluble. In our analysis, Gram-positive isolates from 163 patients with bacterial conjunctivitis were only 85% sensitive to tobramycin, while 97% were sensitive to levofloxacin, a third-generation fluoroquinolone, 83% to sulfasoxazole, 77% to ciprofloxacin, and only 75% to trimethoprim, commonly used in combination with the Gram-negative agent, polymyxin B.2
Recchia and colleagues indicated that an increasingly higher percentage of postcataract infections are due to Gram-positive organisms.3 In a study of 493 consecutive patients with postcataract endophthalmitis, researchers cultured an organism from the vitreous in 318 cases (65%). During the last decade of the 20th century, Gram-positive isolates increased from 92% to 97%. Furthermore, resistance rates to commonly used prophylactic antibiotics increased; resistance among all isolates to ciprofloxacin rose significantly (23% to 38%), while resistance to ciprofloxacin and cefazolin rose among coagulase-negative staphylococci (18% to 38%).
Another new study from Stanford University compared the ability of 21 different antibiotics to cover coagulase-negative Staphylococcus organisms.4 Researchers took preoperative conjunctival swabs from 66 patients prior to applying antibiotics or antiseptic. Their analysis concluded that, among the four fluoroquinolones tested, levofloxacin had the highest antistaphylococcal susceptibilty (91%) compared to norfloxacin (79%), ofloxacin (75%), and ciprofloxacin (73%). Conversely, resistance patterns also favored levofloxacin at only 5%, whereas norfloxacin was 18%, ciprofloxacin 20%, and ofloxacin 23%.
CONTEMPORARY FINDINGS
New in vivo data from Frank Bucci, MD, in Wilkes-Barre, Pennsylvania, demonstrate that levofloxacin reaches therapeutic aqueous concentrations, therefore exceeding the mic90 for both Staphylococcus and Streptococcus.5 Dr. Bucci found that 0.5% levofloxacin reached four- to sevenfold higher aqueous concentrations than 0.3% ciprofloxacin when administered according to identical preoperative regimens. The ciprofloxacin levels were below the established NCCLS MIC90 for both Staphylococcus and Streptococcus. Dr. Bucci also noted that, higher intracameral levofloxacin concentrations could be achieved with a regimen of administering five drops every 10 minutes immediately prior to surgery, when compared to administering the drug four times per day for 2 days preoperatively. He achieved an additional 50% increase in aqueous levels by combining the two regimens.
Jensen and Fiscella6 showed that, of 24 endophthalmitis cases in 9,079 patients, eyes receiving topical ofloxacin postoperatively developed endophthalmitis significantly less often than those receiving topical ciprofloxacin (P<.0009). According to these investigators, this difference in endophthalmitis rates may reflect differences in pharmacological and bioavailability properties that exist among fluoroquinolone antibiotics. Ciprofloxacin, the least soluble of available topical fluoroquinolones, achieves the lowest intraocular levels. Levofloxacin, with 3.3 times more active drug per drop than ofloxacin, might be the preferred choice at this time because of superior Gram-positive coverage and solubility.
Although some surgeons have popularized the use of antibiotic infusion through balanced saline-irrigating solutions during cataract surgery, a group of researchers in Arizona, led by Robert Snyder, MD, do not see the efficacy of this approach.7 Dr. Snyder and his colleagues noted that antibiotics chosen for infusion should be fast-acting, due to the limited time exposure to purported intracameral bacterial contaminants. The fluoroquinolones showed dose-dependent killing. On the other hand, vancomycin killing did not correlate with drug concentration relative to the MIC of Staphylococcus species tested. Fluoroquinolones may be more suitable for killing bacteria seeded into the anterior chamber than vancomycin. Because vancomycin concentration decreases rapidly in the anterior chamber following surgery completion, residual surviving organisms with exposure to this antibiotic of last resort could have a high likelihood of vancomycin resistance. Those who advocate aminoglycoside antibiotic infusion during routine surgery ignore both the severe potential retinal toxicity of this class, and waning Gram-positive sensitivity.
RESISTANCE
Sound clinical analysis dedicated to each prospective cataract patient by a knowledgeable, caring surgeon provides the best solution to endophthalmitis risk. There is no single agent capable of killing every microbe known to cause postoperative infections.8 Even in this brief review of recent ARVO abstracts, epidemiologic patterns differ between hospitals, cities, and regions, a fact that renders each surgeon uniquely capable of understanding the peculiarities of their own bacteriologic environs. Although newer fourth-generation fluoroquinolones, such as moxifloxacin and gatifloxacin, may demonstrate increased potency for Gram-positive bacteria over second- and third-generation drugs, the fourth-generations demonstrated no advantage for Gram-negative coverage in a keratitis study conducted by Kowalski et al.9 Gram-negative resistance appears to cross all fluoroquinolone generations. Thus, miniscule but significant holes have appeared in the once-invincible fluoroquinolone family's Gram-negative coverage spectrum. The best protection of all may be a thorough povidone-iodine preparation,10 including the periorbital skin, lids, lashes, and conjunctival cul-de-sac.
Meticulous iodine preparation and reliable surgical technique, coupled with highly effective and penetrating topical antibiotics given frequently prior to surgery, provide our patients with the best defense against infection.
John D. Sheppard, MD, MMSc, serves as Professor of Ophthalmology, Microbiology & Immunology, as well as Program Director for Ophthalmology Residency Training at the Eastern Virginia Medical School in Norfolk, Virginia. He is also Clinical Director of the Thomas R. Lee Center for Ocular Pharmacology. Dr. Sheppard may be reached at (757) 622-2200; docshep@hotmail.com
1. Han DP, Wisniewski SR, Wilson LA, et al: Spectrum and susceptibilities of microbiologic isolates in the Endophthalmitis Vitrectomy Study. Am J Ophthalmol 122:1-17, 1998
2. Sheppard JD, Oefinger PE, Wegerhoff PE: Susceptibility patterns of conjunctival isolates to newer and established anti-infective agents. IOVS 2002 (abstr 1588) (suppl)
3. Recchia FM, Busbee BG, Pearlman RB, et al: Changing trends in the epidemiology and microbiology of post-cataract endophthalmitis. IOVS (2002 (abstr 4447) (suppl)
4. Ta CN, Mino de Kaspar H, Chang RT, et al: Antibiotic susceptibility pattern of coagulase-negative staphylococci in patients undergoing intraocular surgery. IOVS 2002 (abstr 4444) (suppl)
5. Bucci FA: An in vivo comparison of the ocular absorption of levofloxacin versus ciprofloxacin prior to phacoemulsification. IOVS 2002 (abstr 1579) (suppl)
6. Jensen MK, Fiscella RG: Comparison of endophthalmitis rates over four years associated with topical ofloxacin vs. ciprofloxacin. IOVS 2002 (abstr 4429) (suppl)
7. Snyder RW, Krueger T, Nix DE: Kill curves for vancomycin versus 3rd generation quinolones. IOVS 2002 (abstr 4452) (suppl)
8. Benz MS, Scott IU, Flynn HW, et al: In vitro susceptibilities to antimicrobials of pathogens isolated from the vitreous cavity of patients with endophthalmitis. IOVS 2002 (abstr 4428) (suppl)
9. Kowalski RP, Karenchak LM, Romanowski EG, et al: An in vitro comparison of 2nd, 3rd, and 4th generation fluoroquinolones against bacterial keratitis isolates. IOVS 2002 (abstr 1585) (suppl)
10. Ciulla TA, Starr MB, Masket S: Bacterial endophthalmitis prophylaxis for cataract surgery: an evidence-based update. Ophthalmology 109(1):13-24, 2002 (Jan)
