Both daily and extended wear soft contact lenses have been associated with an increased frequency of Pseudomonas aeruginosa infectious keratitis.1-4 The corneal epithelial barrier may be damaged by overnight wear of soft contact lenses, thus allowing pathogenic organisms such as Pseudomonas to establish an infection in the corneal stroma.5-7 If effective treatment is not instituted promptly, irreversible corneal scarring and even perforation may result.
The current therapeutic regimen in infectious keratitis includes the use of fortified topical antibiotic drops every 15 to 30 minutes for the first 48 hours to insure that bactericidal concentrations reach the site of infection.8,9 This regimen, however, is difficult for the patient and often requires hospitalization.
Collagen corneal shields (Bio-Cor, Bausch & Lomb, Clearwater, Fla) rehydrated in an antibiotic solution have been demonstrated to be an effective treatment modality for experimental Pseudomonas keratitis.10-13 In patients with corneal infections, the use of collagen shields to deliver antibiotics may reduce the need for frequent application of fortified antibiotics.10-13
In this study, we compared the effectiveness of gentamicin-impregnated collagen shields (800 µg of gentamicin), and frequent drops (14 mg/mL) in treating Pseudomonas keratitis in rabbits.
MATERIALS AND METHODS
New Zealand white rabbits (2 to 3 kg) were anesthetized by an intramuscular injection of a 1:5 mixture of xylazine hydrochloride (100 mg/mL; Miles Laboratories, Shawnee, Kan) and ketamine hydrochloride (100 mg/mL; Bristol Laboratories, Syracuse, NY). Two drops of proparacaine (0.5%) were instilled into each eye and nictitating membranectomies were performed. Two days later, rabbits were again anesthetized with intramuscular xylazine and ketamine, and the corneas anesthetized with topical proparacaine. Two perpendicular intersecting diamond knife cuts (5 mm × 5 mm × 0.2 mm) were made centrally in each cornea. Cuts were 0.20 mm deep and 5 mm in length. Contact lenses (Acuvue; Johnson & Johnson, Jacksonville, FIa) were soaked overnight in filter-sterilized 1% bovine serum albumin (BSA).14 The lenses were then exposed to 10p 8 P. aeruginosa strain 27853 for 1 hour at room temperature, rinsed with sterile PBS and then placed onto the cornea. Partial tarsorrhaphies were performed. Three additional Pseudomonas-soaked lenses were rinsed in PBS, placed into sterile PBS (3 mL), homogenized, and plated in triplicate to determine the number of Pseudomonas adhering to the lenses. Twenty-four hours later, animals were again anesthetized and the contact lenses were removed. The infections were allowed to progress for 12 additional hours. The eyes were rinsed with PBS, photographed, and clinical scores of corneal infection were assigned.
Clinical scoring divided the eye into zones I through III. Zone I consisted of an area 1 mm in all directions from the borders of the intersecting keratotomy incisions. Zone II extended from zone I to the midperiphery, and zone ?? extended from zone II to the limbus. Corneal haze in each zone was graded 1 through 4. Grade 1 was a faint, barely perceptible haze, and grade 4 was a totally opaque cornea. Grades 2 and 3 represent haze of increased intensity beyond grade 1 but less than grade 4.
Microbiologic confirmation of infection uniformity was previously performed by culturing infected corneas at 36 hours and quantifying viable bacteria. No significant difference among corneas was noted.
Three treatment groups were employed. Group 1 (3 rabbits, 6 eyes) received collagen shields hydrated in PBS. Group 2 (3 rabbits, 6 eyes) received gentamicin-impregnated collagen shields (800 µg gentamicin) hydrated in PBS. Group 3 (3 rabbits, 6 eyes) received fortified gentamicin drops (1.4%; 14 mg/ mL) every 30 minutes for 8 hours. All rabbits were kept anesthetized with urethane (1.5 gm/kg) during the treatment period and the eyes were taped shut to avoid shield displacement. Those rabbits receiving drops did not have their eyes taped shut. At the end of the 8-hour treatment period, all rabbits were killed with an overdose of pentobarbital sodium.
The procedure for quantifying viable bacteria per cornea has been described previously.15 Briefly, corneas were rinsed, excised, and homogenized in 3 ml of sterile PBS. An aliquot (0.5 mL) of homogenate was serially diluted and 0.1 mL of each dilution was plated in triplicate on tryptic soy agar plates. Bacterial counts were performed by a masked observer after 24 hrs of incubation at 37°C.
Results were analyzed with a Statistical Analysis Systems program.16 An analysis of variance was performed on the logarithm of the number of colonies per cornea. Where an analysis of variance showed a significant difference, t-tests between least square means from each treatment group were performed. Colony counts per cornea are expressed in base 10 logarithms ± SEM.
The contact lenses (Acuvue®) that were applied to the cornea after 1 hour of exposure to 10p 8 Pseudomonas aeruginosa contained 5.86 ± 0.07 Gog base 10) CFU. Slit-lamp examinations and photographs of each cornea were performed at 0 and 39 hours postinfection (PI). Prior to incision, all corneas were normal. At 39 hours PI, all eyes showed a purulent exudate, severe conjunctivitis, and a fulminant ulcerative keratitis centered on the linear incisions. Ten corneas showed keratitis extending to the midperiphery with penetration into the midstroma (zone II), and seven corneas demonstrated keratitis within 1 mm of the keratotomy incisions (zone I). One cornea was noted to have lost the contact lens and was excluded from the study. The eyes were then randomly assigned to three treatment groups, insuring that each treatment group had approximately equal representation with corneas demonstrating infection in zones I and II. We have previously shown in unpublished data, however, that the corneas at 36 hours irrespective of clinical zones have similar bacterial counts and are not statistically different.
Following 8 hours of treatment (47 hour PI), bacterial counts were determined for all corneas (Figure). Gentamicin-impregnated collagen shields applied to corneas for 8 hours reduced the number of CFUs in corneas by approximately 2 logs compared with corneas receiving shields soaked in only saline (P < .0001). Gentamicin drops (1.4%) applied to corneas every 30 minutes for 8 hours reduced the number of CFUs by 6 logs compared with corneas receiving shields soaked in only saline (P < .0001). In this experiment, antibiotic-soaked collagen shields were less effective than frequent antibiotic drops in reducing the number of CFUs in the cornea (P < .0001).
To achieve a therapeutic effect, a sufficient concentration of antibiotic must reach the microorganisms at the site of infection. Contact time, lipid solubility, and concentration are important factors in determining penetration of a compound from the surface of the eye into the corneal stroma.17 Recently, the corneal collagen shield has been introduced as a promising drug delivery system.18
The collagen shield is a thin membrane made of porcine or bovine scleral collagen that conforms to the shape of the eye and dissolves over time. Shields, hydrated in antibiotic solution or made with antibiotic incorporated into the collagen matrix (currently manufactured and under investigation by Bausch & Lomb, Clearwater, FIa), act to prolong the contact time of the drug with the corneal surface. This improves drug penetration into the cornea and anterior chamber.19 It has been shown in an intrastromal model of Pseudomonas keratitis in the rabbit that a collagen shield hydrated in 4.0% tobramycin was as effective as 4.0% tobramycin drops in reducing the number of colony-forming units of bacteria in the cornea.12 The drops were applied every 30 minutes for 4 hours while only one shield per eye was used.
Several models of Pseudomonas keratitis have been developed previously and employed to evaluate antibiotics and drug delivery systems.20,26 Previous animal models of Pseudomonas keratitis have induced keratitis by introducing epithelial defects,20 trephining into the stroma,21 injecting microorganisms into the cornea,12,13,22,23 or scratching the cornea.24-25 Others have induced keratitis without direct manipulation of the cornea by inducing hypoxic states with tarsorrhaphies prior to the placement of Pseudomonas-contairang contact lenses.26-27
These models are not ideal because they either do not mimic the type of corneal trauma associated with human keratitis or lack precision in the delivery of a known amount of bacteria to a reproducible area of injury. Our model more fully addresses the former, but is at best just as good as the intrastromal model in the exactness of bacterial delivery.
Our keratotomy model employs a reproducible corneal injury introduced with a diamond knife. Then, a known quantity of bacteria is placed over the site of the keratotomy in a precise manner with the Pseudomonas-contaimng contact lenses. The quantity delivered to the cornea, however, may not be as precise as the intrastromal injection of a known quantity of bacteria. Finally, tarsorrhaphies are employed for 24 hours while the Pseudomonas-containing contact lenses are in place. This induces a moderate hypoxic state, thus further injuring corneal epithelium and aiding infection,26'27 as well as holding the contact lens in place. Of note, hypoxia may have been induced during the treatment phase in those eyes which were taped (those with collagen shields) and may have influenced the amount of epithelial damage in those eyes versus eyes which received drops only and were not taped. The hypoxic condition may impair aminoglycoside uptake, as well as reduce bacterial growth and associated susceptibility to antibiotic action.
Figure: Efficacy of gentamicin-impregnated shields (800 µg gentamicin) relative to gentamicin drops (1.4% every 30 minutes) and control (shield hydrated in PBS). Treatment modalities were performed (for 8 hours) on eyes previously keratotomized and infected (for 38 hours) with a known quantity of Pseudomonas.
The ability of a contact lens to bind bacteria in a quantitative fashion is well documented.14,28-30 The bacterial binding of the contact lens is enhanced by presoaking the contact lens in bovine serum albumin (1%). Butrus et al14 have previously shown that presoaking contact lenses in BSA increased the adherence of Pseudomonas aeruginosa to the contact lens by over 300%.
Epithelial damage seems to be a prerequisite for Pseudomonas aeruginosa adherence to the cornea 24,31-35 -phe depth of injury (0.2 mm) in our study exposes corneal stroma to the Pseudomonas bacteria. Pseudomonas aeruginosa has been reported to be more adherent to corneal stroma than to corneal epithelium,20 although Stern et al36 found the opposite to be true. In our model, the adherence was sufficient to produce a fulminating keratitis in all corneas.
This study demonstrates that gentamicinimpregnated collagen shields produced a significant reduction in bacteria, yet this therapy was less effective than 1.4% gentamicin drops administered every 30 minutes. The calculated total quantity of gentamicin contained in a collagen shield was 800 µg. This was less than V20 of the total quantity (21 mg) applied to the cornea using 1.4% drops every 30 minutes for 8 hours. This suggests that gentamicin therapy using impregnated shields should be supplemented, perhaps by changing the shield at least once during the period of treatment or by applying gentamicin drops to the shield.
Differences between drops and shields in efficacy for Pseudomonas keratitis were not observed in studies by Hobden et al.12 They used an intrastromal injection model of Pseudomonas keratitis and showed equal effectiveness of collagen shields hydrated with 4% tobramycin and frequently applied 4% tobramycin drops. The intrastromal model differs from this keratotomy model in that the intrastromal model has only a small or no corneal defect, while the keratectomy model has both large corneal epithelial and stromal defects. These substantial corneal defects may allow increased penetration of drops into the corneal stroma. The combination of increased penetration and 20-fold excess of drug used contribute to the efficacy of the drops relative to the collagen shield.
Sawusch et al37 found collagen shields plus frequent drops (3 mg/mL every 30 minutes) to be more efficacious than frequent drops alone over a 12-hour period. Thus, both groups received the same total amount of drug, and the collagen shield appeared to enhance the effect of the drops. In our model, one group received 21 mg of drug in frequent drops and the other group received 800 µg of drug in a collagen shield with no drops. The shields were effective, but the frequent drops were more effective. However, due to the differences in the amounts of drug applied, the results of the two studies cannot be directly compared.
Future experiments should address the question of what frequency of antibiotic drops applied to collagen shields is as effective in decreasing bacterial counts as fortified antibiotic drops every 30 minutes. Clinically, this finding would have relevance in nursing care and/or home compliance if drops placed over collagen shields were as efficacious even when applied on a far less frequent schedule.
1. Laibson PR. Annual review, cornea and sclera. Arch Ophthalmol. 1972;88:553-574.
2. Krachmer JH, Purcell JJ. Bacterial corneal ulcers in cosmetic soft contact lens wearers. Arch Ophthalmol. 1978;96:57-61.
3. Eichenbaum JW, Feldstein J, Podos SM. Extended-wear aphakic soft contact lenses and corneal ulcers. Br J Ophthalmol. 1982;92:384-390.
4. Adams CP, Cohen EJ, Laibson PR, et al. Corneal ulcers in patients with cosmetic extended-wear contact lenses. Am J Ophthalmol. 1983;96:705-709.
5. Steuhl KP, Döring G, Henni A, et al. Relevance of hostderived and bacterial factors in Pseudomonas aeruginosa corneal infections. Invest Ophthalmol Vis Sci. 1987;28:1559-1568.
6. Watt PJ. Pathogenic mechanisms of organisms virulent to the eye. Transactions of the Ophthalmologic Society of the United Kingdom. 1986;105:26-31.
7. Koidou-Tsiligianni A, Alfonso E, Forster RK. Ulcerative keratitis associated with contact lens wear. Am J Ophthalmol. 1989;108:64-67.
8. Baum JL. Jones DB. Initial therapy of suspected microbial corneal uicers. I. Broad antibiotic therapy based on prevalence of organism. JJ. Specific antibiotic therapy based on corneal smears. Surv Ophthalmol. 1979;24:97-116.
9. Leibowitz HM, Ryan WJ, Kupferman A. Route of antibiotic administration in bacterial keratitis. Arch Ophthalmol. 1981;99:1420-1423.
10. Aquavella JV, Musco PS, Ueda S, et al. Therapeutic applications of a collagen bandage lens: preliminary report. CLAO J. 1988;14:47-50.
11. Unterman SR, Rootman DS, Hill JM, et al. Collagen shield drug delivery: therapeutic concentrations of tobramycin in the rabbit cornea and aqueous humor. J Cataract Refract Surg. 1988;14:500-504.
12. Hobden JA, Reidy JJ, O'Callaghan RJ, et al. Treatment of experimental Pseudomonas keratitis using collagen shields containing tobramycin. Arch Ophthalmol. 1988; 106: 1605-1607.
13. Hobden JA, Reidy JJ, O'Callaghan RJ, et al. Quinolones in collagen shields to treat aminoglycosiderresistant pseudomonal keratitis. Invest Ophthalmol Vis Sci. 199Q;3 1:2241-2243.
14. Butrus SI, Klotz FA, Misra RP. The adherence of Pseudomonas aeruginosa to soft contact lenses. Ophthalmology. 1987;1310-1314.
15. Rootman DS, Hobden JA, Jantzen JA, et al. Iontophoresis of tobramycin for the treatment of experimental Pseudomonas keratitis in the rabbit. Arch Ophthalmol. 1988;106:262-265.
16. SAS Institute, Inc. SAS /STAT Guide for Personal Computers, ed 6. Cary, NC: SAS Institute; 1987.
17. Maurice D, Mîshima S. Ocular pharmacokinetics. In: Sears ML, ed. Handbook of Expérimental Pharmacology: Pharmacology of the Eye. New York, NY; Springer Verlag; 1984: 19.
18. Shofher RS, Kaufman HE, Hill JM. New horizons in ocular drug delivery. Ophthalmol Clin North Am. 1989;2:15-23.
19. Phinney RB, Schwartz SD, Lee DA, et al. Collagen-shield delivery of gentamicin and vancomycin. Arch Ophthalmol. 1988;106:1599-1604.
20. Klotz SA, Au Y-K, Misra RP. A partial-thickness epithelial defect increases the adherence of Pseudomonas aeruginosa to the cornea. Invest Ophthalmol Vis Sci. 1989;30:1069-1074.
21. Wyman M, Swanson C, Kowalski JJ, et al. Experimental Pseudomonas aeruginosa ulcerative keratitis model in the dog. Am J Vet Res. 1983;44:1135-1140.
22. Gray LD, Kreger AS. Rabbit corneal damage produced by Pseudomonas aeruginosa infection. Infect Immun. 1975;12:419-432.
23. Davis SD, Chandler JW. Experimental keratitis due to Pseudomonas aeruginosa: model for evaluation of antimicrobial drugs. Antimicrob Agents Chemother. 1975;8:350-355.
24. Ramphal R, McNiece MT, Polack FM. Adherence of Pseudomonas aeruginosa to the injured cornea: a step in the pathogenesis of corneal infections. Ann Ophthalmol. 1981;13:421-425.
25. Hyndiuk RA, Davis SP, Hatchell DL, et al. Experimental Pseudomonas keratitis: clinical and pathological observations. Cornea. 1983;2:103-114.
26. Koch JM, Refojo MF, Hanninen LA, et al. Experimental Pseudomonas aeruginosa keratitis from extended wear of soft contact lenses. Arch Ophthalmol. 1990;108:1453-1459.
27. Aswad MI, Barza M, Baum J. Effect of lid closure on contact lens-associated Pseudomonas keratitis. Arch Ophthalmol. 1989;107:1667-1670.
28. Duran JA1 Refojo MF, Gipson IK, et al. Pseudomonas attachment to new hydrogel contact lenses. Arch Ophthalmol. 1987;105:106-109.
29. Stern GA, Zam ZS. The pathogenesis of contact lens associated Pseudomonas aeruginosa corneal ulceration. I: The effect of contact lens codings on inherence of Pseudomonas aeruginosa to soft contact lenses. Cornea. 1986;5:41-45.
30. Dart JK, Badenoch PR. Bacterial adherence to contact lenses. CLAOJ. 1986;12:220-224.
31. Van Horn DL, Davis SD, Hyndiuk RA, et al. Experimental Pseudomonas keratitis in the rabbit: bactériologie, clinical, and microscopic observations. Invest Ophthalmol Vis Sci. 1981;20:213-221.
32. DiGaetano M, Stern GA, Zam ZS. The pathogenesis of contact lens associated Pseudomonas aeruginosa corneal ulceration, II: an animal model. Cornea. 1986;5:155-158.
33. Duran JA, Refojo MF, Kenyon KR. Hydrogel contact lensinduced Pseudomonas keratitis in a rabbit model. Cornea. 1987;6:258-260.
34. Hyndiuk RA. Experimental Pseudomonas keratitis. Trans Am Ophthalmol Soc. 1981;79:541-624.
35. Spurr-Michaud SJ, Barza M, Gipson IK. An organ culture system for study of adherence of Pseudomonas aeruginosa to normal and wounded corneas. Invest Ophthalmol Vis Sci. 1988;29:379-386.
36. Stern G, Weitzenkom D, Valenti J. Adherence of Pseudomonas aeruginosa to the mouse cornea: epithelial vs stromal adherence. Arch Ophthalmol. 1982;100:1956.
37. Sawusch MR, O'Brien TP, Dick JD, et al. Usé of collagen corneal shields in the treatment of bacterial keratitis. Am J Ophthalmol. 1988;106:279-281.