- Email: [email protected]
Eye-banking risk factors for fungal endophthalmitis compared with bacterial endophthalmitis after corneal transplantation
Eye-banking Risk Factors for Fungal Endophthalmitis Compared With Bacterial Endophthalmitis After Corneal Transplantation SOHELA S. HASSAN, DRPH, AND KIRK R. WILHELMUS, MD, PHD, FOR THE MEDICAL REVIEW SUBCOMMITTEE OF THE EYE BANK ASSOCIATION OF AMERICA
● PURPOSE:
To track the relative frequency and explore possible risk factors of fungal compared with bacterial endophthalmitis after corneal transplantation. ● DESIGN: Case-comparison study nested in a surveillance registry. ● METHODS: We examined, among recipients who underwent surgery from January 1991 through December 2003, demographic and eye-banking characteristics of postkeratoplasty endophthalmitis cases that were reported to the Eye Bank Association of America by US eye banks. Potential predictors of fungal compared with bacterial endophthalmitis were examined using multiple logistic regression. ● RESULTS: Of 234 reported cases of postkeratoplasty endophthalmitis reported by US eye bank, 46 cases (19.7%) were fungal, and 130 cases (55.6%) were bacterial. Concordant cultures of the residual donor corneoscleral rim or preservation medium occurred significantly more often with fungal than bacterial endophthalmitis (P ⴝ .009). After the introduction of OptisolGS, the odds of bacterial relative to fungal endophthalmitis decreased by 77% (95% confidence interval, 44% - 91%). After adjustment for the preservation method and other eye-banking variables, the odds of fungal endophthalmitis was 3.4 (95% confidence interval, 1.6 - 7.4) times that of bacterial endophthalmi-
Accepted for publication Dec 3, 2004. From the Department of Ophthalmology, Cullen Eye Institute, Baylor College of Medicine, Houston, Texas; and the Eye Bank Association of America, Washington, DC. A listing of members of the Medical Review Subcommittee of the Medical Advisory Board for the Eye Bank Association of America are listed in the Appendix. This study was supported by a T32 Training Grant (EY07001) and P30 Vision Core Grant (EY02520) from the National Eye Institute and a Senior Scientific Investigator Award from the Research to Prevent Blindness, Inc, New York, New York. Inquiries to Sohela S. Hassan, Dr. P.H., Department of Ophthalmology, Baylor College of Medicine, 6565 Fannin Street, NC-305, Houston, TX 77030; fax: (713) 798-4142; e-mail: [email protected] 0002-9394/05/$30.00 doi:10.1016/j.ajo.2004.12.016
©
2005 BY
tis, when donor corneal preservation was 4 days or longer. ● CONCLUSIONS: The frequency of postkeratoplasty bacterial endophthalmitis compared with fungal endophthalmitis fell after widespread use of a corneal preservation medium supplemented with gentamicin and streptomycin. Further improvements in corneal preservation procedures are needed. (Am J Ophthalmol 2005;139:685-690. © 2005 by Elsevier Inc. All rights reserved.)
M
ICROBIAL ENDOPHTHALMITIS IS A SERIOUS COM-
plication of corneal transplantation.1 Postmortem eyes harbor bacteria and fungi,2–5 and the same microorganisms can be isolated from both donors and recipients who develop endophthalmitis after penetrating keratoplasty.6 –9 Corneal tissue processing involves antiinfective strategies,10 –11 but procurement and storage may allow microbial contamination by and growth of streptococci, enterococci, staphylococci, anaerobes, and yeasts (Shah M.; ARVO meeting; 2001; Abstract).12,13 Case reports have suggested a recent possible increase in donor-to-host transmission of candidal infection after corneal transplantation (Cukrowski C.; EBAA meeting; 2002; Abstract).14 –17 Commercial corneal preservation media for short-term refrigeration do not contain an antifungal agent, and fungal contamination is not reliably detected by colorimetric indicators.18 However, the occurrence and reasons for an apparent emergence of fungal endophthalmitis associated with corneal grafting are unclear. In 1991, the Eye Bank Association of America (EBAA) introduced a mechanism for eye banks to collect and to report serious adverse reactions potentially attributable to donor eye tissue.19 We used this surveillance registry to examine longitudinal trends in postkeratoplasty endophthalmitis and to explore possible risk factors for fungal endophthalmitis in comparison to bacterial endophthalmitis after corneal transplantation.
ELSEVIER INC. ALL
RIGHTS RESERVED.
685
TABLE 1. Eye-banking Characteristics Categorized by Microbial Isolation From Recipients With Postkeratoplasty Endophthalmitis
Variables
Fungi (46)
Bacteria (130)
No Growth or Not Done (58)
Total (234)
Donor age, mean ⫾ SD (yrs) Donor cause of death Cancer Cardiac disease Other medical conditions Trauma Death-to-preservation interval, mean ⫾ SD (hrs) Preservation-to-surgery interval, mean ⫾ SD (days) Preservation method Optisol™ or Dexsol® Optisol™-GS Recipient age, mean ⫾ SD (yrs) Recipient preoperative eye disease Pseudophakic or aphakic corneal edema Corneal graft failure Keratoconus Fuchs’ corneal dystrophy Other corneal disorders
52.0 ⫾ 19.4
48.4 ⫾ 18.3
54.7 ⫾ 14.3
50.9 ⫾ 17.7
13 (28.3%) 11 (23.9%) 13 (28.2%) 9 (19.6%) 7.4 ⫾ 5.1 4.7 ⫾ 1.6
33 (25.4%) 29 (22.3%) 40 (30.8%) 28 (21.5%) 6.8 ⫾ 4.3 4.0 ⫾ 1.9
17 (29.3%) 24 (41.4%) 11 (19.0%) 6 (10.3%) 7.5 ⫾ 3.7 4.2 ⫾ 2.0
63 (26.9%) 64 (27.4%) 64 (27.3%) 43 (18.4%) 7.1 ⫾ 4.3 4.2 ⫾ 1.9
7 (15.6%) 39 (84.4%) 67.8 ⫾ 19.1
62 (48.1%) 68 (51.9%) 67.2 ⫾ 20.5
12 (20.7%) 46 (79.3%) 72.8 ⫾ 14.4
81 (34.9%) 153 (65.1%) 68.6 ⫾ 19.1
12 (26.1%) 5 (10.9%) 6 (13.1%) 9 (19.5%) 14 (30.4%)
44 (33.8%) 26 (20.0%) 12 (9.2%) 10 (7.7%) 38 (29.3%)
13 (22.4%) 3 (5.2%) 5 (8.6%) 6 (10.3%) 31 (53.4%)
69 (29.5%) 34 (14.5%) 23 (9.8%) 25 (10.6%) 83 (35.6%)
SD ⫽ standard deviation.
METHODS
RESULTS
THE STUDY POPULATION COMPRISED CASES OF MICROBIAL
OF 246 CASES OF POSTKERATOPLASTY ENDOPHTHALMITIS
endophthalmitis after penetrating keratoplasty performed between January 1, 1991, and December 31, 2003, that were reported to the EBAA’s adverse reaction registry by US eye banks. Cases of fungal postkeratoplasty endophthalmitis, defined as growth of any yeast or filamentous fungus from recipient intraocular fluid, were compared with eyes with bacterial postkeratoplasty endophthalmitis. Recipient age, donor age, death-to-preservation time, and preservation-to-surgery time were categorized at cut points rounded to the 5-year age, even number of cadaveric hours, or whole number of storage days nearest their respective average values. Causes of donor death were categorized by recorded fatal cancer, cardiac diseases, trauma, and other medical conditions. Recipient preoperative diagnoses were categorized according to whether the preoperative surgical indication was stated to be because of previous anterior segment surgery, defined as pseudophakic or aphakic corneal edema or previous corneal transplant failure, keratoconus, Fuchs’ corneal dystrophy, and other corneal disorders. Data were combined for donor corneas stored in either Dexsol or Optisol (Chiron Ophthalmics, Irvine, California, USA for comparison to Optisol-GS (Bausch & Lomb, Irvine, California, USA). The dataset was transferred into SPSS version 12.0 (Chicago, Illinois, USA) for statistical calculations. Unconditional logistic regression estimated odds ratios with 95% confidence intervals.20,21 All analyses were eye-based.
submitted to the EBAA from 1991 through 2003, 234 were reported by US eye banks. Forty-six (19.7%) were fungal and 130 (55.6%) bacterial. Forty-one (17.3%) yielded no microbial growth from recipient ocular specimens, and 17 (7.3%) had no recipient culture done. Of the 176 culturepositive cases, the mean (⫾ SD) donor age was 49 ⫾ 19 years, and the mean recipient age was 67 ⫾ 20 years (Table 1). Twenty-four (13.6%) culture-positive endophthalmitis cases, 8 fungal and 16 bacterial, had endophthalmitis associated with the corresponding donor mate (i.e., paired corneas from 12 donors were associated with postkeratoplasty endophthalmitis). One hundred forty-eight (84.1%) of the culture-positive endophthalmitis cases had a culture performed of the donor corneoscleral rim or preservation medium, and 55 (37.2%) of these yielded the same microbial species as the recipient culture, including 25 (67.9%) of 39 with fungal endophthalmitis and 30 (27.5%) of 109 with bacterial endophthalmitis (P ⫽ .009). Sixty-nine (39.2%) of all culture-positive postkeratoplasty endophthalmitis cases were preserved in either Optisol (64) or Dexsol (5), and 107 (60.8%) in OptisolGS. Postkeratoplasty endophthalmitis was associated with the former two media between January 1991 and May 1993 and with the latter medium from April 1993 to December 2003. A progressive relative decrease in bacterial compared with fungal postkeratoplasty endophthalmitis (Fig-
686
AMERICAN JOURNAL
OF
OPHTHALMOLOGY
APRIL 2005
FIGURE 1. Relative annual percentage of fungal and bacterial postkeratoplasty endophthalmitis reported by US eye banks.
TABLE 2. Risk Factors for Fungal Endophthalmitis Compared with Bacterial Endophthalmitis After Corneal Transplantation
Variables
No. (%)
Donor age ⬍ 50 years ⱖ 50 years Donor cause of death Other conditions Cancer Death-to-preservation interval ⬍ 8 hours ⱖ 8 hours Preservation-to-surgery interval ⬍ 4 days ⱖ 4 days Preservation method Optisol™ or Dexsol® Optisol™-GS Recipient age ⬍70 years ⱖ 70 years Recipient preoperative eye disease Postsurgical corneal edema or graft failure Other corneal conditions
Adjusted Odds Ratio (95% Confidence Interval)
82 (46.6%) 94 (53.4%)
1.00 1.71 (.86–3.40)
1.00 1.21 (.55–2.66)
130 (73.9%) 46 (26.1%)
1.00 1.16 (.55–2.46)
1.00 1.22 (.52–2.86)
122 (69.3%) 54 (30.7%)
1.00 1.13 (.55–2.32)
1.00 1.21 (.54–2.74)
111 (64.2%) 65 (35.8%)
1.00 3.03 (1.52–6.07)
1.00 3.39 (1.56–7.37)
69 (39.2%) 107 (60.8%)
1.00 5.08 (2.12–12.19)
1.00 4.43 (1.78–11.06)
49 (27.8%) 127 (72.2%)
1.00 1.13 (.53–2.41)
1.00 1.82 (.50–2.56)
87 (49.4%) 89 (50.6%)
1.00 2.00 (1.00–3.97)
1.00 1.73 (.75–4.41)
ure 1) seemed to follow the widespread use of Optisol-GS by US eye banks. Potential risk factors for fungal compared with bacterial endophthalmitis are shown in Table 2. The preservation method and longer preservation-to-surgery interval appeared to be significantly associated with fungal endophthalmitis compared with bacterial endophthalmitis, with unadjusted odds ratios of 5.1 and 3.0, respectively. Following adjustment in a multiple regression model that was VOL. 139, NO. 4
Unadjusted Odds Ratio (95% Confidence Interval)
tested for Hosmer-Lemeshow goodness-of-fit and found to be adequate (P ⫽ .24), the preservation method and preservation duration remained independently and significantly associated with fungal endophthalmitis, with adjusted odds ratios of 4.4 and 3.4, respectively. Thus, the adjusted odds of bacterial endophthalmitis was 77% (95% confidence interval, 44% to 91%) less than that of fungal endophthalmitis with Optisol-GS. The preservation duration was slightly but significantly longer for fungal endoph-
ENDOPHTHALMITIS AFTER CORNEAL TRANSPLANTATION
687
FIGURE 2. Relationship between donor corneal storage duration and postkeratoplasty endophthalmitis.
FIGURE 3. Monthly distribution of fungal and bacterial postkeratoplasty endophthalmitis reported by US eye banks.
thalmitis compared with bacterial endophthalmitis (Figure 2). Seasonal variation in the microbiology of endophthalmitis (P ⫽ .03) suggested that fungal endophthalmitis increased during the summer (Figure 3).
accounted for nearly one fifth of reported culture-positive endophthalmitis, although a gradual increase seemed to occur over the past few years.15–17 This apparent emergence of fungal postkeratoplasty infection may be the result of the relative reduction of bacterial postkeratoplasty endophthalmitis that followed the widespread use of donor corneal preservation medium supplemented with gentamicin and streptomycin. Whether eye-banking factors might be associated with fungal, especially candidal, endophthalmitis after corneal surgery is unclear. This study suggests that a protracted preservation-to-surgery duration increases the relative risk of fungal endophthalmitis in comparison to bacterial
DISCUSSION ENDOPHTHALMITIS AFTER CORNEAL TRANSPLANTATION IS
uncommon.22–24 In this study, reported cases of postkeratoplasty endophthalmitis represented approximately 0.05% of corneal transplants performed in the US during the same period.25 Fungal endophthalmitis, on average, 688
AMERICAN JOURNAL
OF
OPHTHALMOLOGY
APRIL 2005
endophthalmitis. Of reported culture-positive cases, the storage time of donor corneas that resulted in fungal endophthalmitis averaged about a day longer than those leading to bacterial endophthalmitis. The risk of fungal endophthalmitis was approximately triple that of bacterial endophthalmitis among donor corneas preserved for 4 days or more. This finding supports the hypothesis that microorganisms carried by the donor cornea may persist during storage. Antibacterial agents in the preservation medium suppress or eradicate many bacteria, but operational procedures that prolong donor corneal preservation may contribute to an increased opportunity for fungal growth. Furthermore, the apparent seasonality of bacterial endophthalmitis could be explained by fluctuations in monthly surgical volume, but an abrupt increase in yeast endophthalmitis during midsummer may be related to environmental fungal dynamics, as found with some other candidal infections.26 –28 Although our findings cannot be generalized to all corneal graft recipients, aiming toward an optimal preservation-to-surgery interval for refrigerated donor eye tissue seems prudent. This investigation was strengthened by the large number of postkeratoplasty endophthalmitis cases catalogued by the EBAA’s prospective surveillance system. Yet, using the results to revise policy and procedures is limited by study constraints. The endophthalmitis registry relies on clinical recognition, microbiological assessment, and notification, and may be affected by underreporting. Restricting the study to the three fourths of reported postkeratoplasty endophthalmitis cases having a positive recipient culture intended to ensure laboratory validation of a clinical diagnosis, but donor-to-host transmission involves further evidence. When donor cultures were also considered, significantly more cases of fungal endophthalmitis had the same organism isolated from donor cultures, suggesting that selection bias could have an appreciable effect. Furthermore, the analysis was limited to information provided by eye banks. Causes of donor death were not validated by medical records or other sources, and many donor attributes such as diabetes and premortem mechanical ventilation that can be associated with ocular contaminants29,30 were not recorded. Residual confounding by surgical variables such as virectomy is also possible. Whereas corneal grafting does not significantly influence the risk of endophthalmitis associated with cataract surgery,31 further study is needed to examine whether accompanying procedures, such as posterior capsule rupture and vitreous loss, affect the risk of postkeratoplasty endophthalmitis.22 Similarly, the use of broad categories may have obscured preoperative recipient conditions such as ocular surface disease that can predispose to postoperative endophthalmitis.31,32 Eye banks take precautions to reduce microbial contamination of donor eyes, but organisms such as Candida albicans may persist and escape detection.15 The role of antifungal supplementation of the donor media remains VOL. 139, NO. 4
unclear,15 and further assessment of preservation procedures may provide insight into opportunities for preventing postkeratoplasty endophthalmitis. The EBAA’s adverse reaction registry offers a valuable resource for epidemiologic investigations for substantiating and improving quality assurance in eye banking.
REFERENCES 1. Eifrig CW, Flynn HW, Jr., Scott IU, Newton J. Acute-onset postoperative endophthalmitis: review of incidence and visual outcomes (1995–2001). Ophthalmic Surg Lasers 2002; 33:373–378. 2. Polack FM, Locatcher-Khorazo D, Gutierrez E. Bacteriologic study of “donor” eyes. Evaluation of antibacterial treatments prior to corneal grafting. Arch Ophthalmol 1967;78:219 – 225. 3. White JH. Fungal contamination of donor eyes. Br J Ophthalmol 1969;53:30 –33. 4. Escapini H, Jr., Olson RJ, Kaufman HE. Donor cornea contamination with McCarey-Kaufman medium preservation. Am J Ophthalmol 1979;88:59 – 62. 5. Wu T, Mitchell B, Carothers T, et al. Molecular analysis of the pediatric ocular surface for fungi. Curr Eye Res 2003;26: 33–36. 6. Khodadoust AA, Franklin RM. Transfer of bacterial infections by donor cornea in penetrating keratoplasty. Am J Ophthalmol 1979;87:130 –132. 7. Fong LP, Gladstone D, Casey TA. Corneo-scleral rim cultures: donor contamination a case of fungal endophthalmitis transmitted by K-Sol stored cornea. Eye 1988;2:670 – 676. 8. Cameron JA, Antonios SR, Cotter JB, Habash NR. Endophthalmitis from contaminated donor corneas following penetrating keratoplasty. Arch Ophthalmol 1991;109:54 –59. 9. Kloess PM, Stulting RD, Waring GO III, Wilson LA. Bacterial and fungal endophthalmitis after penetrating keratoplasty. Am J Ophthalmol 1993;115:309 –316. 10. Badenoch PR, Alfrich SJ, Wedding TR, Coster DJ. Effectiveness of a decontamination method for donor corneas. Br J Ophthalmol 1988;72:225–227. 11. Mindrup EA, Dubbel PA, Doughman DJ. Betadine decontamination of donor globes. Cornea 1993;12:324 –329. 12. Antonios SR, Cameron JA, Badr IA, Habash NR, Cotter JB. Contamination of donor cornea: postpenetrating kertoplasty endophthalmitis. Cornea 1991;10:217–220. 13. Gomes JA, Dana MR, Dua HS, Goren MB, Liabson PR, Cohen EJ. Positive donor rim sulture in penetrating keratoplasty. Cornea 1995;14:457– 462. 14. Cameron JA, Badr IA, Risco JM, Abboud E, Gonnah ES. Endophthalmitis cluster from contaminated donor corneas following penetrating keratoplasty. Can J Ophthalmol 1998; 33:8 –13. 15. Merchant A, Zacks CM, Wilhelmus K, Durand M, Dohlman CH. Candidal endophthalmitis after keratoplasty. Cornea 2001;20:226 –229. 16. Sutphin JE, Pfaller MA, Hollis RJ, Wagoner MD. Candida albicans after corneal transplantation. Am J Ophthalmol 2002;134:120 –121.
ENDOPHTHALMITIS AFTER CORNEAL TRANSPLANTATION
689
17. Garcia Serrano JL, Dominguez I, Serrano Laborda D. Endoftalmitis recurrente en queratoplastia penetrante, por flora mixta. Arch Soc Esp Oftalmol 2004;79:89 –92. 18. Chu YI, Penland RL, Wilhelmus KR. Colorimetric Indicators of microbial contamination in corneal preservation medium. Cornea 2000;19:517–520. 19. Wilhelmus KR, Stulting RD, Sugar J, Khan MM. Primary corneal graft failure: a national reporting system. Arch Ophthalmol 1995;113:1497–1502. 20. Schlesselmann JJ. Case control studies: design, conduct, analysis. New York: Oxford University Press, 1982:269 –275. 21. Hosmer D, Lemeshow S. Applied logistic regression. New York: John Wiley & Sons, 1989. 22. Aiello LP, Javitt JC, Canner JK. National outcomes of penetrating keratoplasty: risks of endophthalmitis and retinal detachment. Arch Ophthalmol 1993;111:509 –513. 23. Wiffen SJ, Weston BC, Maguire LJ, Bourne WM. The value of routine donor corneal rim cultures in penetrating keratoplasty. Arch Ophthalmol 1997;115:719 –724. 24. Everts RJ, Fowler WC, Chang DH, Reller LB. Corneoscleral rim cultures: lack of utility and implications for clinical decision-making and infection prevention in the care of patients undergoing corneal transplantation. Cornea 2001; 20:586 –589. 25. Eye Bank Association of America. Annual Statistical Reports. Washington, DC: EBAA, 1992–2003. 26. Garcia-Martos P, Delgado D, Marin P, Mira J. Analysis of 40 cases of otomycosis. Enferm Infecc Microbiol Clin 1993;11: 487– 489.
690
AMERICAN JOURNAL
27. Chaudhury A, Nath G, Shukla B, Panda S, Singh TB. Diarrhoea associated with Candida species: incidence and seasonal variation. J Diarrhoeal Dis Res 1996;14:110 –112. 28. Rubio EF. Climate influence on conjunctivital bacteria of patient undergone cataract surgery. Eye 2004;18:778 –784. 29. Dixon DM, Graham CR, Jr., Shaffer RM, Tarantino P. Fungal flora from diabetic and non-diabetic human donor corneas. Cornea 1984;3:281–284. 30. Nelson JD, Mindrup EA, Chung CK, Lindstrom RL, Doughman DJ. Fungal contamination in organ culture. Arch Ophthalmol 1983;101:280 –283. 31. Li J, Morlet N, Ng JQ, Semmens JB, Knuiman MW. Significant nonsurgical risk factors for endophthalmitis after cataract surgery: EPSWA fourth report. Invest Ophthalmol Vis Sci 2004;45:1321–1328. 32. Nouri M, Terada H, Alfonso EC, Foster CS, Durand ML, Dohlman CH. Endophthalmitis after keratoprosthesis: incidence, bacterial causes, and risk factors. Arch Ophthalmol 2001;119:484 – 489.
APPENDIX THE MEDICAL REVIEW SUBCOMMITTEE OF THE MEDICAL
Advisory Board for the Eye Bank Association of America Patricia Dahl, CEBT; David Glasser, MD; Ellen Heck, MT (ASCP), CEBT; Kevin Ross, MS, MPH; Bradley Tennant, CEBT; Woodford Van Meter, MD; and Kirk R. Wilhelmus, MD, PhD
OF
OPHTHALMOLOGY
APRIL 2005
Biosketch Sohela S. Hassan is a post-doctoral fellow in the department of Ophthalmology, Baylor College of Medicine, where she is responsible for overseeing epidemiologic studies focused on risks and outcomes of corneal transplant procedures. Her research commitment is directed to health promotion and improvement. She started her career as a population demographer in Bangladesh and evolved keen interest in community health.
VOL. 139, NO. 4
ENDOPHTHALMITIS AFTER CORNEAL TRANSPLANTATION
690.e1
Biosketch Dr. Kirk R. Wilhelmus specializes in cornea and external disease at the Department of Ophthalmology, Baylor College of Medicine, where he is Medical Director of the Lions Eye Bank of Texas. His research focuses on the clinical epidemiology and microbiology of infectious eye disease, antimicrobial pharmacodynamics, and preventive ophthalmology.
690.e2
AMERICAN JOURNAL
OF
OPHTHALMOLOGY
APRIL 2005
● PURPOSE:
To track the relative frequency and explore possible risk factors of fungal compared with bacterial endophthalmitis after corneal transplantation. ● DESIGN: Case-comparison study nested in a surveillance registry. ● METHODS: We examined, among recipients who underwent surgery from January 1991 through December 2003, demographic and eye-banking characteristics of postkeratoplasty endophthalmitis cases that were reported to the Eye Bank Association of America by US eye banks. Potential predictors of fungal compared with bacterial endophthalmitis were examined using multiple logistic regression. ● RESULTS: Of 234 reported cases of postkeratoplasty endophthalmitis reported by US eye bank, 46 cases (19.7%) were fungal, and 130 cases (55.6%) were bacterial. Concordant cultures of the residual donor corneoscleral rim or preservation medium occurred significantly more often with fungal than bacterial endophthalmitis (P ⴝ .009). After the introduction of OptisolGS, the odds of bacterial relative to fungal endophthalmitis decreased by 77% (95% confidence interval, 44% - 91%). After adjustment for the preservation method and other eye-banking variables, the odds of fungal endophthalmitis was 3.4 (95% confidence interval, 1.6 - 7.4) times that of bacterial endophthalmi-
Accepted for publication Dec 3, 2004. From the Department of Ophthalmology, Cullen Eye Institute, Baylor College of Medicine, Houston, Texas; and the Eye Bank Association of America, Washington, DC. A listing of members of the Medical Review Subcommittee of the Medical Advisory Board for the Eye Bank Association of America are listed in the Appendix. This study was supported by a T32 Training Grant (EY07001) and P30 Vision Core Grant (EY02520) from the National Eye Institute and a Senior Scientific Investigator Award from the Research to Prevent Blindness, Inc, New York, New York. Inquiries to Sohela S. Hassan, Dr. P.H., Department of Ophthalmology, Baylor College of Medicine, 6565 Fannin Street, NC-305, Houston, TX 77030; fax: (713) 798-4142; e-mail: [email protected] 0002-9394/05/$30.00 doi:10.1016/j.ajo.2004.12.016
©
2005 BY
tis, when donor corneal preservation was 4 days or longer. ● CONCLUSIONS: The frequency of postkeratoplasty bacterial endophthalmitis compared with fungal endophthalmitis fell after widespread use of a corneal preservation medium supplemented with gentamicin and streptomycin. Further improvements in corneal preservation procedures are needed. (Am J Ophthalmol 2005;139:685-690. © 2005 by Elsevier Inc. All rights reserved.)
M
ICROBIAL ENDOPHTHALMITIS IS A SERIOUS COM-
plication of corneal transplantation.1 Postmortem eyes harbor bacteria and fungi,2–5 and the same microorganisms can be isolated from both donors and recipients who develop endophthalmitis after penetrating keratoplasty.6 –9 Corneal tissue processing involves antiinfective strategies,10 –11 but procurement and storage may allow microbial contamination by and growth of streptococci, enterococci, staphylococci, anaerobes, and yeasts (Shah M.; ARVO meeting; 2001; Abstract).12,13 Case reports have suggested a recent possible increase in donor-to-host transmission of candidal infection after corneal transplantation (Cukrowski C.; EBAA meeting; 2002; Abstract).14 –17 Commercial corneal preservation media for short-term refrigeration do not contain an antifungal agent, and fungal contamination is not reliably detected by colorimetric indicators.18 However, the occurrence and reasons for an apparent emergence of fungal endophthalmitis associated with corneal grafting are unclear. In 1991, the Eye Bank Association of America (EBAA) introduced a mechanism for eye banks to collect and to report serious adverse reactions potentially attributable to donor eye tissue.19 We used this surveillance registry to examine longitudinal trends in postkeratoplasty endophthalmitis and to explore possible risk factors for fungal endophthalmitis in comparison to bacterial endophthalmitis after corneal transplantation.
ELSEVIER INC. ALL
RIGHTS RESERVED.
685
TABLE 1. Eye-banking Characteristics Categorized by Microbial Isolation From Recipients With Postkeratoplasty Endophthalmitis
Variables
Fungi (46)
Bacteria (130)
No Growth or Not Done (58)
Total (234)
Donor age, mean ⫾ SD (yrs) Donor cause of death Cancer Cardiac disease Other medical conditions Trauma Death-to-preservation interval, mean ⫾ SD (hrs) Preservation-to-surgery interval, mean ⫾ SD (days) Preservation method Optisol™ or Dexsol® Optisol™-GS Recipient age, mean ⫾ SD (yrs) Recipient preoperative eye disease Pseudophakic or aphakic corneal edema Corneal graft failure Keratoconus Fuchs’ corneal dystrophy Other corneal disorders
52.0 ⫾ 19.4
48.4 ⫾ 18.3
54.7 ⫾ 14.3
50.9 ⫾ 17.7
13 (28.3%) 11 (23.9%) 13 (28.2%) 9 (19.6%) 7.4 ⫾ 5.1 4.7 ⫾ 1.6
33 (25.4%) 29 (22.3%) 40 (30.8%) 28 (21.5%) 6.8 ⫾ 4.3 4.0 ⫾ 1.9
17 (29.3%) 24 (41.4%) 11 (19.0%) 6 (10.3%) 7.5 ⫾ 3.7 4.2 ⫾ 2.0
63 (26.9%) 64 (27.4%) 64 (27.3%) 43 (18.4%) 7.1 ⫾ 4.3 4.2 ⫾ 1.9
7 (15.6%) 39 (84.4%) 67.8 ⫾ 19.1
62 (48.1%) 68 (51.9%) 67.2 ⫾ 20.5
12 (20.7%) 46 (79.3%) 72.8 ⫾ 14.4
81 (34.9%) 153 (65.1%) 68.6 ⫾ 19.1
12 (26.1%) 5 (10.9%) 6 (13.1%) 9 (19.5%) 14 (30.4%)
44 (33.8%) 26 (20.0%) 12 (9.2%) 10 (7.7%) 38 (29.3%)
13 (22.4%) 3 (5.2%) 5 (8.6%) 6 (10.3%) 31 (53.4%)
69 (29.5%) 34 (14.5%) 23 (9.8%) 25 (10.6%) 83 (35.6%)
SD ⫽ standard deviation.
METHODS
RESULTS
THE STUDY POPULATION COMPRISED CASES OF MICROBIAL
OF 246 CASES OF POSTKERATOPLASTY ENDOPHTHALMITIS
endophthalmitis after penetrating keratoplasty performed between January 1, 1991, and December 31, 2003, that were reported to the EBAA’s adverse reaction registry by US eye banks. Cases of fungal postkeratoplasty endophthalmitis, defined as growth of any yeast or filamentous fungus from recipient intraocular fluid, were compared with eyes with bacterial postkeratoplasty endophthalmitis. Recipient age, donor age, death-to-preservation time, and preservation-to-surgery time were categorized at cut points rounded to the 5-year age, even number of cadaveric hours, or whole number of storage days nearest their respective average values. Causes of donor death were categorized by recorded fatal cancer, cardiac diseases, trauma, and other medical conditions. Recipient preoperative diagnoses were categorized according to whether the preoperative surgical indication was stated to be because of previous anterior segment surgery, defined as pseudophakic or aphakic corneal edema or previous corneal transplant failure, keratoconus, Fuchs’ corneal dystrophy, and other corneal disorders. Data were combined for donor corneas stored in either Dexsol or Optisol (Chiron Ophthalmics, Irvine, California, USA for comparison to Optisol-GS (Bausch & Lomb, Irvine, California, USA). The dataset was transferred into SPSS version 12.0 (Chicago, Illinois, USA) for statistical calculations. Unconditional logistic regression estimated odds ratios with 95% confidence intervals.20,21 All analyses were eye-based.
submitted to the EBAA from 1991 through 2003, 234 were reported by US eye banks. Forty-six (19.7%) were fungal and 130 (55.6%) bacterial. Forty-one (17.3%) yielded no microbial growth from recipient ocular specimens, and 17 (7.3%) had no recipient culture done. Of the 176 culturepositive cases, the mean (⫾ SD) donor age was 49 ⫾ 19 years, and the mean recipient age was 67 ⫾ 20 years (Table 1). Twenty-four (13.6%) culture-positive endophthalmitis cases, 8 fungal and 16 bacterial, had endophthalmitis associated with the corresponding donor mate (i.e., paired corneas from 12 donors were associated with postkeratoplasty endophthalmitis). One hundred forty-eight (84.1%) of the culture-positive endophthalmitis cases had a culture performed of the donor corneoscleral rim or preservation medium, and 55 (37.2%) of these yielded the same microbial species as the recipient culture, including 25 (67.9%) of 39 with fungal endophthalmitis and 30 (27.5%) of 109 with bacterial endophthalmitis (P ⫽ .009). Sixty-nine (39.2%) of all culture-positive postkeratoplasty endophthalmitis cases were preserved in either Optisol (64) or Dexsol (5), and 107 (60.8%) in OptisolGS. Postkeratoplasty endophthalmitis was associated with the former two media between January 1991 and May 1993 and with the latter medium from April 1993 to December 2003. A progressive relative decrease in bacterial compared with fungal postkeratoplasty endophthalmitis (Fig-
686
AMERICAN JOURNAL
OF
OPHTHALMOLOGY
APRIL 2005
FIGURE 1. Relative annual percentage of fungal and bacterial postkeratoplasty endophthalmitis reported by US eye banks.
TABLE 2. Risk Factors for Fungal Endophthalmitis Compared with Bacterial Endophthalmitis After Corneal Transplantation
Variables
No. (%)
Donor age ⬍ 50 years ⱖ 50 years Donor cause of death Other conditions Cancer Death-to-preservation interval ⬍ 8 hours ⱖ 8 hours Preservation-to-surgery interval ⬍ 4 days ⱖ 4 days Preservation method Optisol™ or Dexsol® Optisol™-GS Recipient age ⬍70 years ⱖ 70 years Recipient preoperative eye disease Postsurgical corneal edema or graft failure Other corneal conditions
Adjusted Odds Ratio (95% Confidence Interval)
82 (46.6%) 94 (53.4%)
1.00 1.71 (.86–3.40)
1.00 1.21 (.55–2.66)
130 (73.9%) 46 (26.1%)
1.00 1.16 (.55–2.46)
1.00 1.22 (.52–2.86)
122 (69.3%) 54 (30.7%)
1.00 1.13 (.55–2.32)
1.00 1.21 (.54–2.74)
111 (64.2%) 65 (35.8%)
1.00 3.03 (1.52–6.07)
1.00 3.39 (1.56–7.37)
69 (39.2%) 107 (60.8%)
1.00 5.08 (2.12–12.19)
1.00 4.43 (1.78–11.06)
49 (27.8%) 127 (72.2%)
1.00 1.13 (.53–2.41)
1.00 1.82 (.50–2.56)
87 (49.4%) 89 (50.6%)
1.00 2.00 (1.00–3.97)
1.00 1.73 (.75–4.41)
ure 1) seemed to follow the widespread use of Optisol-GS by US eye banks. Potential risk factors for fungal compared with bacterial endophthalmitis are shown in Table 2. The preservation method and longer preservation-to-surgery interval appeared to be significantly associated with fungal endophthalmitis compared with bacterial endophthalmitis, with unadjusted odds ratios of 5.1 and 3.0, respectively. Following adjustment in a multiple regression model that was VOL. 139, NO. 4
Unadjusted Odds Ratio (95% Confidence Interval)
tested for Hosmer-Lemeshow goodness-of-fit and found to be adequate (P ⫽ .24), the preservation method and preservation duration remained independently and significantly associated with fungal endophthalmitis, with adjusted odds ratios of 4.4 and 3.4, respectively. Thus, the adjusted odds of bacterial endophthalmitis was 77% (95% confidence interval, 44% to 91%) less than that of fungal endophthalmitis with Optisol-GS. The preservation duration was slightly but significantly longer for fungal endoph-
ENDOPHTHALMITIS AFTER CORNEAL TRANSPLANTATION
687
FIGURE 2. Relationship between donor corneal storage duration and postkeratoplasty endophthalmitis.
FIGURE 3. Monthly distribution of fungal and bacterial postkeratoplasty endophthalmitis reported by US eye banks.
thalmitis compared with bacterial endophthalmitis (Figure 2). Seasonal variation in the microbiology of endophthalmitis (P ⫽ .03) suggested that fungal endophthalmitis increased during the summer (Figure 3).
accounted for nearly one fifth of reported culture-positive endophthalmitis, although a gradual increase seemed to occur over the past few years.15–17 This apparent emergence of fungal postkeratoplasty infection may be the result of the relative reduction of bacterial postkeratoplasty endophthalmitis that followed the widespread use of donor corneal preservation medium supplemented with gentamicin and streptomycin. Whether eye-banking factors might be associated with fungal, especially candidal, endophthalmitis after corneal surgery is unclear. This study suggests that a protracted preservation-to-surgery duration increases the relative risk of fungal endophthalmitis in comparison to bacterial
DISCUSSION ENDOPHTHALMITIS AFTER CORNEAL TRANSPLANTATION IS
uncommon.22–24 In this study, reported cases of postkeratoplasty endophthalmitis represented approximately 0.05% of corneal transplants performed in the US during the same period.25 Fungal endophthalmitis, on average, 688
AMERICAN JOURNAL
OF
OPHTHALMOLOGY
APRIL 2005
endophthalmitis. Of reported culture-positive cases, the storage time of donor corneas that resulted in fungal endophthalmitis averaged about a day longer than those leading to bacterial endophthalmitis. The risk of fungal endophthalmitis was approximately triple that of bacterial endophthalmitis among donor corneas preserved for 4 days or more. This finding supports the hypothesis that microorganisms carried by the donor cornea may persist during storage. Antibacterial agents in the preservation medium suppress or eradicate many bacteria, but operational procedures that prolong donor corneal preservation may contribute to an increased opportunity for fungal growth. Furthermore, the apparent seasonality of bacterial endophthalmitis could be explained by fluctuations in monthly surgical volume, but an abrupt increase in yeast endophthalmitis during midsummer may be related to environmental fungal dynamics, as found with some other candidal infections.26 –28 Although our findings cannot be generalized to all corneal graft recipients, aiming toward an optimal preservation-to-surgery interval for refrigerated donor eye tissue seems prudent. This investigation was strengthened by the large number of postkeratoplasty endophthalmitis cases catalogued by the EBAA’s prospective surveillance system. Yet, using the results to revise policy and procedures is limited by study constraints. The endophthalmitis registry relies on clinical recognition, microbiological assessment, and notification, and may be affected by underreporting. Restricting the study to the three fourths of reported postkeratoplasty endophthalmitis cases having a positive recipient culture intended to ensure laboratory validation of a clinical diagnosis, but donor-to-host transmission involves further evidence. When donor cultures were also considered, significantly more cases of fungal endophthalmitis had the same organism isolated from donor cultures, suggesting that selection bias could have an appreciable effect. Furthermore, the analysis was limited to information provided by eye banks. Causes of donor death were not validated by medical records or other sources, and many donor attributes such as diabetes and premortem mechanical ventilation that can be associated with ocular contaminants29,30 were not recorded. Residual confounding by surgical variables such as virectomy is also possible. Whereas corneal grafting does not significantly influence the risk of endophthalmitis associated with cataract surgery,31 further study is needed to examine whether accompanying procedures, such as posterior capsule rupture and vitreous loss, affect the risk of postkeratoplasty endophthalmitis.22 Similarly, the use of broad categories may have obscured preoperative recipient conditions such as ocular surface disease that can predispose to postoperative endophthalmitis.31,32 Eye banks take precautions to reduce microbial contamination of donor eyes, but organisms such as Candida albicans may persist and escape detection.15 The role of antifungal supplementation of the donor media remains VOL. 139, NO. 4
unclear,15 and further assessment of preservation procedures may provide insight into opportunities for preventing postkeratoplasty endophthalmitis. The EBAA’s adverse reaction registry offers a valuable resource for epidemiologic investigations for substantiating and improving quality assurance in eye banking.
REFERENCES 1. Eifrig CW, Flynn HW, Jr., Scott IU, Newton J. Acute-onset postoperative endophthalmitis: review of incidence and visual outcomes (1995–2001). Ophthalmic Surg Lasers 2002; 33:373–378. 2. Polack FM, Locatcher-Khorazo D, Gutierrez E. Bacteriologic study of “donor” eyes. Evaluation of antibacterial treatments prior to corneal grafting. Arch Ophthalmol 1967;78:219 – 225. 3. White JH. Fungal contamination of donor eyes. Br J Ophthalmol 1969;53:30 –33. 4. Escapini H, Jr., Olson RJ, Kaufman HE. Donor cornea contamination with McCarey-Kaufman medium preservation. Am J Ophthalmol 1979;88:59 – 62. 5. Wu T, Mitchell B, Carothers T, et al. Molecular analysis of the pediatric ocular surface for fungi. Curr Eye Res 2003;26: 33–36. 6. Khodadoust AA, Franklin RM. Transfer of bacterial infections by donor cornea in penetrating keratoplasty. Am J Ophthalmol 1979;87:130 –132. 7. Fong LP, Gladstone D, Casey TA. Corneo-scleral rim cultures: donor contamination a case of fungal endophthalmitis transmitted by K-Sol stored cornea. Eye 1988;2:670 – 676. 8. Cameron JA, Antonios SR, Cotter JB, Habash NR. Endophthalmitis from contaminated donor corneas following penetrating keratoplasty. Arch Ophthalmol 1991;109:54 –59. 9. Kloess PM, Stulting RD, Waring GO III, Wilson LA. Bacterial and fungal endophthalmitis after penetrating keratoplasty. Am J Ophthalmol 1993;115:309 –316. 10. Badenoch PR, Alfrich SJ, Wedding TR, Coster DJ. Effectiveness of a decontamination method for donor corneas. Br J Ophthalmol 1988;72:225–227. 11. Mindrup EA, Dubbel PA, Doughman DJ. Betadine decontamination of donor globes. Cornea 1993;12:324 –329. 12. Antonios SR, Cameron JA, Badr IA, Habash NR, Cotter JB. Contamination of donor cornea: postpenetrating kertoplasty endophthalmitis. Cornea 1991;10:217–220. 13. Gomes JA, Dana MR, Dua HS, Goren MB, Liabson PR, Cohen EJ. Positive donor rim sulture in penetrating keratoplasty. Cornea 1995;14:457– 462. 14. Cameron JA, Badr IA, Risco JM, Abboud E, Gonnah ES. Endophthalmitis cluster from contaminated donor corneas following penetrating keratoplasty. Can J Ophthalmol 1998; 33:8 –13. 15. Merchant A, Zacks CM, Wilhelmus K, Durand M, Dohlman CH. Candidal endophthalmitis after keratoplasty. Cornea 2001;20:226 –229. 16. Sutphin JE, Pfaller MA, Hollis RJ, Wagoner MD. Candida albicans after corneal transplantation. Am J Ophthalmol 2002;134:120 –121.
ENDOPHTHALMITIS AFTER CORNEAL TRANSPLANTATION
689
17. Garcia Serrano JL, Dominguez I, Serrano Laborda D. Endoftalmitis recurrente en queratoplastia penetrante, por flora mixta. Arch Soc Esp Oftalmol 2004;79:89 –92. 18. Chu YI, Penland RL, Wilhelmus KR. Colorimetric Indicators of microbial contamination in corneal preservation medium. Cornea 2000;19:517–520. 19. Wilhelmus KR, Stulting RD, Sugar J, Khan MM. Primary corneal graft failure: a national reporting system. Arch Ophthalmol 1995;113:1497–1502. 20. Schlesselmann JJ. Case control studies: design, conduct, analysis. New York: Oxford University Press, 1982:269 –275. 21. Hosmer D, Lemeshow S. Applied logistic regression. New York: John Wiley & Sons, 1989. 22. Aiello LP, Javitt JC, Canner JK. National outcomes of penetrating keratoplasty: risks of endophthalmitis and retinal detachment. Arch Ophthalmol 1993;111:509 –513. 23. Wiffen SJ, Weston BC, Maguire LJ, Bourne WM. The value of routine donor corneal rim cultures in penetrating keratoplasty. Arch Ophthalmol 1997;115:719 –724. 24. Everts RJ, Fowler WC, Chang DH, Reller LB. Corneoscleral rim cultures: lack of utility and implications for clinical decision-making and infection prevention in the care of patients undergoing corneal transplantation. Cornea 2001; 20:586 –589. 25. Eye Bank Association of America. Annual Statistical Reports. Washington, DC: EBAA, 1992–2003. 26. Garcia-Martos P, Delgado D, Marin P, Mira J. Analysis of 40 cases of otomycosis. Enferm Infecc Microbiol Clin 1993;11: 487– 489.
690
AMERICAN JOURNAL
27. Chaudhury A, Nath G, Shukla B, Panda S, Singh TB. Diarrhoea associated with Candida species: incidence and seasonal variation. J Diarrhoeal Dis Res 1996;14:110 –112. 28. Rubio EF. Climate influence on conjunctivital bacteria of patient undergone cataract surgery. Eye 2004;18:778 –784. 29. Dixon DM, Graham CR, Jr., Shaffer RM, Tarantino P. Fungal flora from diabetic and non-diabetic human donor corneas. Cornea 1984;3:281–284. 30. Nelson JD, Mindrup EA, Chung CK, Lindstrom RL, Doughman DJ. Fungal contamination in organ culture. Arch Ophthalmol 1983;101:280 –283. 31. Li J, Morlet N, Ng JQ, Semmens JB, Knuiman MW. Significant nonsurgical risk factors for endophthalmitis after cataract surgery: EPSWA fourth report. Invest Ophthalmol Vis Sci 2004;45:1321–1328. 32. Nouri M, Terada H, Alfonso EC, Foster CS, Durand ML, Dohlman CH. Endophthalmitis after keratoprosthesis: incidence, bacterial causes, and risk factors. Arch Ophthalmol 2001;119:484 – 489.
APPENDIX THE MEDICAL REVIEW SUBCOMMITTEE OF THE MEDICAL
Advisory Board for the Eye Bank Association of America Patricia Dahl, CEBT; David Glasser, MD; Ellen Heck, MT (ASCP), CEBT; Kevin Ross, MS, MPH; Bradley Tennant, CEBT; Woodford Van Meter, MD; and Kirk R. Wilhelmus, MD, PhD
OF
OPHTHALMOLOGY
APRIL 2005
Biosketch Sohela S. Hassan is a post-doctoral fellow in the department of Ophthalmology, Baylor College of Medicine, where she is responsible for overseeing epidemiologic studies focused on risks and outcomes of corneal transplant procedures. Her research commitment is directed to health promotion and improvement. She started her career as a population demographer in Bangladesh and evolved keen interest in community health.
VOL. 139, NO. 4
ENDOPHTHALMITIS AFTER CORNEAL TRANSPLANTATION
690.e1
Biosketch Dr. Kirk R. Wilhelmus specializes in cornea and external disease at the Department of Ophthalmology, Baylor College of Medicine, where he is Medical Director of the Lions Eye Bank of Texas. His research focuses on the clinical epidemiology and microbiology of infectious eye disease, antimicrobial pharmacodynamics, and preventive ophthalmology.
690.e2
AMERICAN JOURNAL
OF
OPHTHALMOLOGY
APRIL 2005