Culture-proven endogenous endophthalmitis: clinical features and visual acuity outcomes

Culture-proven Endogenous Endophthalmitis: Clinical Features and Visual Acuity Outcomes VIVIAN SCHIEDLER, MD, INGRID U. SCOTT, MD, MPH, HARRY W. FLYNN JR., MD, JANET L. DAVIS, MD, MATTHEW S. BENZ, MD, AND DARLENE MILLER, MA, MPH

● PURPOSE:

To investigate clinical features and visual acuity outcomes associated with endogenous endophthalmitis. ● DESIGN: Retrospective, observational case series. ● METHODS: Twenty-one eyes of 21 patients treated at Bascom Palmer Eye Institute for culture-proven endogenous endophthalmitis between 1996 and 2002 were reviewed. ● RESULTS: Patients were followed a mean of 3 months (range, 1 to 12 months). Fungal isolates occurred in 13 eyes (62%), gram-positive isolates in 7 (33%), and gram-negative isolates in 1 (5%). Twelve patients (57%) were hospitalized at the time of diagnosis and 6 patients (29%) died within 2 months of diagnosis. Initial treatment included tap and injection of intravitreal medication in 10 eyes (48%) and pars plana vitrectomy with injection of intravitreal medication in 11 eyes (52%). Final visual outcomes were obtainable for 18 eyes (two patients died within 10 days of diagnosis, and one patient was lost to follow-up). Eight (44%) of these 18 eyes achieved a visual acuity of 20/400 or better and 10 (56%) of 18 eyes achieved a visual acuity worse than 20/400, including 3 that were either enucleated or eviscerated. Three eyes with Aspergillus endophthalmitis had worse visual outcomes than eyes with either Candida (P ⴝ .036) or bacterial endophthalmitis (P ⴝ .024). ● CONCLUSIONS: Compared with published series of postoperative or post-traumatic endophthalmitis, patients with endogenous endophthalmitis are more likely to have fungal isolates with a predominance of Candida albicans. Endogenous endophthalmitis is generally associated with high mortality and poor visual acuity outcomes, particularly when caused by more virulent species such as

Accepted for publication Nov 3, 2003. From the Bascom Palmer Eye Institute, Department of Ophthalmology, University of Miami School of Medicine, Miami, Florida. Supported in part by Research to Prevent Blindness, Inc., New York, New York. Inquiries to Ingrid U. Scott, MD, MPH, Bascom Palmer Eye Institute, PO Box 016880, Miami, FL 33101; fax: (305) 326 – 6417; e-mail: [email protected] 0002-9394/04/$30.00 doi:10.1016/j.ajo.2003.11.013

©

2004 BY

Aspergillus. (Am J Ophthalmol 2004;137:725–731. © 2004 by Elsevier Inc. All rights reserved.)

E

NDOGENOUS ENDOPHTHALMITIS COMPRISES APPROX-

imately 5% to 7% of cases in large series of endophthalmitis.1,2 Previously published series of patients with endogenous endophthalmitis have reported the most common bacterial organism as Staphylococcus aureus3 and the most common fungal organism as Candida species.4,5 Most patients with endogenous endophthalmitis have one or more predisposing systemic risk factors, although cases among otherwise healthy, immunocompetent persons have been reported.6,7 Endogenous endophthalmitis has been associated with many systemic risk factors, including chronic immune-compromising illnesses (diabetes mellitus, renal failure), indwelling or long-term intravenous catheters, immunosuppressive diseases and therapy (malignancies, human immunodeficiency virus infection or HIV, chemotherapeutic agents), recent invasive surgery, endocarditis, gastrointestinal procedures, hepatobiliary tract infections, and intravenous drug abuse.3–5,8 –14 In this review, clinical features and visual outcomes are reported for patients treated at a single institution for cultureproven endogenous endophthalmitis over a 7-year period.

METHODS THE STUDY WAS APPROVED BY THE INSTITUTIONAL REVIEW

Board of the University of Miami School of Medicine. This retrospective medical record review included all patients treated for culture-proven endogenous endophthalmitis at the Bascom Palmer Eye Institute between January 1, 1996, and December 31, 2002. A positive culture was defined as growth of the same organism on one or more solid media or growth on a single culture medium of an organism detected on initial smear.15 Patients with a history of penetrating ocular trauma, recent ocular surgery (within 1 year before presentation with endogenous endophthalmitis), or evidence of a primary external ocular infection (such as bleb infection or infectious keratitis), were excluded.

ELSEVIER INC. ALL

RIGHTS RESERVED.

725

outside hospital for presumed uveitis before presenting to our institution with bilateral involvement. Two patients (patients 1 and 18) were initially diagnosed with toxoplasmosis uveitis, and after failing treatment with trimethoprim and sulfamethoxasole and pyrimethamine and sulfadiazine, underwent pars plana vitrectomy; Candida albicans was isolated from the vitreous cultures. Candida albicans was not significantly associated with longer time to presentation than Aspergillus or bacterial endophthalmitis. As described above, four of the five patients who initially were thought to have uveitis actually had Candida endophthalmitis. The presence of Candida albicans was significantly associated with incorrect initial diagnosis (P ⫽ .025 ⫻ 2-tailed Fisher test). However, delayed diagnosis was not predictive of visual outcome nor was initial treatment for an incorrect diagnosis of “uveitis.” Fungal isolates (13; 62%) were more common than bacterial isolates, and these included Candida albicans (seven eyes), Aspergillus species (four eyes), Cladophialophora devriesii (one eye), and Pseudallescheria boydii (one eye). There were seven (33%) gram-positive isolates, including Staphylococcus aureus (five eyes), and Streptococcus species (two eyes). There was only one (5%) gramnegative isolate: Klebsiella pneumoniae. One eye had two organisms identified from vitreous cultures: Staphylococcus aureus and Staphylococcus warneri. Fifteen patients had systemic cultures performed at the time of endogenous endophthalmitis diagnosis (Table 1). Ten (67%) of these 15 patients had systemic cultures that confirmed extraocular sources of infection (with the same organism responsible for the ocular infection) but only 5 (33%) of these 15 patients had positive blood cultures. Four of these blood culture-positive cases revealed bacteremia and one revealed a hyperalimentation catheterassociated Candida albicans fungemia. One patient, patient 18, was diagnosed with a urethral stricture related to a remote traumatic urinary catheter placement upon discovery of Candida albicans funguria as the source of endophthalmitis. Patient 12 was treated for a pleural effusion 1 month before diagnosis with endophthalmitis. In this case, pleural fluid cultures were positive for Staphylococcus aureus with the same sensitivity pattern as the organism that was cultured from the eye 1 month later. Two patients (patients 7 and 8) with disseminated Aspergillus had positive sputum cultures for the same organism. Patient 11 had an unspecified biomedical device culture that revealed Aspergillus as the extraocular source of infection. The visual acuity before the onset of endophthalmitis was 20/20 in all five eyes for which this information was available. The median visual acuity at presentation was 2/200 (range, 20/20 to light perception). As shown in Table 4, 10 (48%) of 21 eyes had presenting visual acuities of 20/400 or better, and 11(52%) of 21 eyes had vision worse than 20/400. Presenting visual acuity was not a predictor for visual outcome. Hypopyon was present in 5 (24%) of 21 eyes but was not predictive of final visual

RESULTS THE STUDY INCLUDED 21 EYES WITH CULTURE-PROVEN EN-

dogenous endophthalmitis in 21 patients with a median age of 52 years (range, 26 –79 years). Patients were followed a mean of 3 months (range, 1–12 months). Fourteen patients (67%) were men and seven were women. A summary of the clinical characteristics of these patients is provided in Table 1. Thirteen cases (62%) involved the right eye. Six patients (29%) had clinical signs of bilateral involvement (patients 4, 6, 11, 12, 18, 21), but all six cases had only one eye with culture-positive results (four of the six fellow eyes had culture-negative results, and two did not have cultures performed). These six fellow eyes are described in Table 2. All 21 patients had identifiable potential systemic risk factors for endogenous endophthalmitis (Table 1) and (Table 3). Twelve patients (57%) were hospitalized at the time of endophthalmitis diagnosis and 5 (24%) had been hospitalized within 6 months before developing endophthalmitis. Fourteen patients (67%) had placement of intravenous catheters. The most common medical condition was diabetes mellitus (11 patients [52%], 3 of whom were insulin-dependent, and 2 of whom had corticosteroid-induced hyperglycemia). Nine patients (43%) had recently received or were receiving immunosuppressive therapeutic agents (such as chemotherapy, systemic corticosteroids, or immunomodulating medications) at the time of endophthalmitis diagnosis. Six patients (29%) had received organ transplants within 2 years of developing endogenous endophthalmitis, including three liver transplants, one heart transplant, one kidney/pancreas transplant, and one bone marrow transplant for chronic myelogenous leukemia. Of the 5 patients with a history of malignancy, 1 had numerous myelomas, 1 had recurrent breast cancer, 1 had chronic myelogenous leukemia, 1 had mycosis fungoid, and 1 had lymphoma. Five patients (19%) were on hemodialysis (1 for acute renal failure and 4 for end-stage renal disease), 3 (14%) had a history of intravenous drug use, and 1 (5%) had HIV. Ocular symptoms included decreased vision in 18 (86%) of the 21 eyes , redness (15; 71%), pain (13; 62%), floaters (7; 33%), and photophobia (5; 24%). Thirteen (62%) of the 21 eyes experienced onset of ocular symptoms within 7 days before presentation. The median time between onset of symptoms and presentation with endophthalmitis was 3.5 days (range, 0 to 35 days). Three patients presented initially to outside physicians and were treated with anti-inflammatory agents for “uveitis” until lack of improvement prompted referrals to our institution. Of these three cases, two were caused by Candida albicans, and one was caused by Staphylococcus aureus and Staphylococcus warneri. The patient, patient 21, with the longest time interval from symptom onset to presentation (5 weeks) had endophthalmitis caused by Candida albicans and was treated for 3 weeks with topical corticosteroids at an 726

AMERICAN JOURNAL

OF

OPHTHALMOLOGY

APRIL 2004

VOL. 137, NO. 4

TABLE 1. Clinical Summary

CULTURE-PROVEN ENDOGENOUS ENDOPHTHALMITIS

Patient No.

Eye

Age (yrs)

Gender

Organism

1 2 3 4* 5 6*

OS OS OD OS OD OS

58 69 47 26 75 72

M M M F F M

Candida albicans Cladophialophora devriesii Streptococcus bovis Candida albicans MRSA Candida albicans

DM, IVDU IV, lymphoma, heart TX, IT HIV/AIDS IC, HAL, anorexia nervosa IV, mycosis fungoides, IT foot cellulitis

7

OS

26

M

Aspergillus flavus

IV, DM, leukemia, BMT, GVHD, IT

8 9

OD OS

57 52

M M

Aspergillus terreus Pseudallescheria boydii

IV, DM, liver TX ⫻ 2, HD, IT IV, liver TX ⫻ 2, IT

10 11*

OD OS

28 49

M M

S. aureus, S. warneri Aspergillus fumigatus

12* 13 14 15 16 17 18* 19 20 21*

OD OS OD OD OD OD OD OD OD OD

46 65 30 63 42 60 62 31 26 79

F M M F M F M M F F

S aureus MRSA Aspergillus fumigatus S. aureus Streptococcus oralis Klebsiella pneumoniae Candida albicans Candida albicans Candida albicans Candida albicans

Other Infectious Focus

Systemic Cultures†

(⫺) (⫺) (⫺) (⫹) blood, catheter (⫹) blood, catheter unknown

psoriasis, IT IV, on heart TX list

none endocarditis, vocal cord infection pneumonia Candidemia MRSA line sepsis, pneumonia fungus ball in aorta, cerebral emboli MRSA bacteremia, fungal pneumonia disseminated Candidal skin lesions, fungal brain abscess frequent skin lesions disseminated

IC, DM, HD IV, DM, HD, ESLD IV, DM, liver TX, IT IC, DM, breast cancer, IT IC, DM, HD, kidney/pancreas TX IC, DM, multiple myeloma, IT DM IVDU IVDU IV, DM, HD

S. aureus pleural effusion MRSA socket infection, sepsis none S. aureus line infection kidney abscess Klebsiella, bacteremia Candida UTI none none none

Risk Factors

(⫹) sputum (⫹) sputum unknown unknown (⫹) biomedical tubing (⫹) pleural fluid (⫹) blood unknown (⫹) blood (⫺) (⫹) blood (⫹) urine unknown unknown (⫺)

BMT ⫽ bone marrow transplant; DM ⫽ diabetes mellitus; ESLD ⫽ end-stage liver disease; GVHD ⫽ graft vs host disease; HAL ⫽ hyperalimentation; HD ⫽ hemodialysis; HIV/AIDS ⫽ human immunodeficiency virus/acquired immunodeficiency syndrome; IC ⫽ indwelling catheter; IT ⫽ immunosuppressive therapy; IV ⫽ intravenous catheter; IVDU ⫽ intravenous drug abuse; OD ⫽ right eye; OS ⫽ left eye; TX ⫽ transplant. *Clinical evidence of fellow eye involvement (fellow eyes had negative culture results or no cultures performed). † Positive cultures listed revealed causative organism.

727

TABLE 2. Fellow Eyes Patient No

Eye

Organism*

Initial VA

Systemic Treatment

Ocular Treatment

Final VA

4 6

OD OD

Candida albicans Candida albicans

20/25 20/60

inj AMP 3 PPV/inj AMP inj AMP

IV AMP PO F

20/20 unknown

11

OD

Aspergillus fumigatus

20/400

inj AMP ⫻ 3

IV AMP

20/400

12

OS

Staphylococcus aureus

20/400

inj V,C,AMP,D

none

20/40

18

OS

Candida albicans

20/20

IV AMP, F

20/25

21

OS

Candida albicans

HM

inj AMP 3 PPV/cryo/SBP for RD PPV/inj AMP,D 3 PPL

PO F

HM

Miscellaneous

died in 10 days: aortic fungus ball with cerebral embolization died in 2 months of disseminated Aspergillosis IV cefazolin for S. aureus pleural effusion 1 month prior OD treated for toxoplasmosis uveitis initially; candiduria treated for uveitis by outside physician

AMP ⫽ amphotericin B; C ⫽ ceftazidime; cryo ⫽ cryotherapy; D ⫽ dexamethasone; F ⫽ fluconazole; Inj ⫽ intravitreal injection, IV ⫽ intravenous; OD ⫽ right eye; OS ⫽ left eye; PO ⫽ oral; PPL ⫽ pars plana lensectomy; PPV ⫽ pars plana vitrectomy; RD ⫽ retinal detachment; SBP ⫽ scleral buckle procedure; V ⫽ vancomycin. *Organism cultured from study eye (fellow eye had no cultures performed or had negative culture results.

pergillosis precluded surgical intervention. Five (50%) of the 10 eyes initially treated with intravitreal injection eventually underwent pars plana vitrectomy as a secondary procedure. All seven Candida albicans isolates were not tested for sensitivities to antifungals, because the rate of antimicrobial resistance was extremely low. Only one patient with Aspergillus endophthalmitis (patient 14) was initially treated with antifungal medication (however, sensitivity testing was not done in this case), whereas the other three patients with endophthalmitis caused by Aspergillus were initially treated with antibiotics only. One of these latter isolates, Aspergillus flavus (patient 7), was tested for sensitivities and was found to be resistant to amphotericin. Eight eyes were initially treated with intravitreal antibiotics/antifungals to which the organism was sensitive, and these eyes had better visual outcomes than the 5 eyes that were initially treated with intravitreal antibiotics/antifungals to which the organism was not sensitive (P ⫽ .032 by Mann-Whitney test). Information on systemic treatment with antibiotics or antifungals was not accessible from the medical records of two patients. Of the 19 patients for whom this information was available, 17 (89%) were treated with systemic antibiotics or antifungals or both; the 2 patients who did not receive systemic treatment had a history of intravenous drug abuse and limited follow-up. Seven (33%) of 21 eyes developed ocular complications following treatment for endophthalmitis: 5 had retinal detachments only, 1 developed elevated intraocular pressure, and 1 had a retinal detachment, cataract, and elevated intraocular pressure. The latter, patient 18, first underwent pars plana vitrectomy with scleral buckle procedure for the retinal detachment and subsequently underwent pars plana vitrectomy with removal of silicone oil

TABLE 3. Associated Systemic Medical Conditions Medical Condition

No. of Patients*

Intravenous line or indwelling catheter Diabetes mellitus† Immunosuppressive therapy Organ transplantation Malignancy Hemodialysis Intravenous drug use Human immunodeficiency virus

14 11 9 6 5 5 3 1

*Fourteen patients had more than one associated condition. † Two patients had steroid-induced hyperglycemia.

acuity. The majority of eyes (12 of 21 eyes; 57%) presented with diffuse anterior and posterior involvement. Six eyes (29%) had diffuse posterior involvement only. Two eyes (10%) presented with panophthalmitis (proptosis and restricted motility). One eye (5%) presented with diffuse anterior involvement. Although anterior/posterior and focal/diffuse classification of endophthalmitis at presentation was not a predictor of visual outcome, panophthalmitis was significantly associated with evisceration (P ⫽ .005). Ten (48%) of 21 eyes were treated initially with intravitreal injection of therapeutic agents, and 11 eyes (52%) were treated initially with pars plana vitrectomy and injection of intravitreal therapeutic agents (Table 4). Initial treatment method was not predictive of visual outcome. Only one patient (patient 11) underwent numerous intravitreal injections (a total of six) without pars plana vitrectomy because severe sepsis with systemic as728

AMERICAN JOURNAL

OF

OPHTHALMOLOGY

APRIL 2004

TABLE 4. Treatment and Visual Outcomes Patient No.

Organism

Initial Vision

Classification

Intravitreous Agents

1 2 3 4* 5 6* 7 8 9 10 11* 12* 13 14 15 16

Candida albicans Cladophialophora devriesii Streptococcus bovis Candida albicans MRSA Candida albicans Aspergillus flavus Aspergillus terreus Pseudallescheria boydii S. aureus, S. warneri Aspergillus fumigatus S. aureus MRSA Aspergillus fumigatus S. aureus Streptococcus oralis

20/30 CF LP 20/80 1/200 E 20/60 20/400 20/40 20/25 HM HM 2/200 E 20/400 HM CF LP

Diffuse A/P Diffuse A/P Panophthalmitis Diffuse A Diffuse A/P Diffuse A/P Diffuse P Diffuse P Diffuse A/P Diffuse A/P Diffuse A/P Diffuse P Diffuse P Diffuse A/P Diffuse P Diffuse P

V,C

17 18*

Klebsiella pneumoniae Candida albicans

LP 20/200

Panophthalmitis Diffuse A/P

V,C

19 20 21*

Candida albicans Candida albicans Candida albicans

20/300 20/20 2/200 E

Diffuse A/P Diffuse A/P Diffuse A/P

AMP AMP

AMP

yes yes no yes yes yes yes yes yes yes no yes no yes no yes

AMP

no yes

AMP AMP AMP

yes yes yes

V,C V,C V,C V,C,D V,AK V V,C,D V,C,D V,C V,C,D V V,C,D

D

Vitrectomy

AMP AMP AMP AMP AMP AMP AMP AMP AMP AMP

Treatment of Complication

RD; laser

RD; PPV,SB

RD; none

RD; laser

RD; none high IOP; timolol RD; PPV, SBP cataract; PPL high IOP; SOR

Systemic Treatment

Final Vision

unknown AMP, F V,C,PCN AMP V F V,I,AMP C, AMP V, AMP CIP V,I,CFP,AMP unknown V,G,AMP AMP G CFZ

20/25 20/20 Evisceration 20/80 20/25 unknown†‡ NLP† unknown†§ LP† 20/20 LP† 20/400 CF† Enucleation 20/50 20/800

CFP AMP, F

Evisceration 20/300

none none F

2/200E unknown㛳 CF

*Clinical evidence of fellow eye involvement (fellow eyes had negative culture results or no cultures performed). Died within 2 months of diagnosis. ‡ Transferred to another hospital, medical condition precluded visual examination, died in 10 days. § Medical condition precluded visual examination, died in 4 days. 㛳 Did not return for follow-up after pars plana vitrectomy. Abbreviations: see Table 2 A ⫽ anterior, AK ⫽ amikacin, CFP ⫽ cefepime, CFZ ⫽ cefazolin, CIP ⫽ ciprofloxacin, G ⫽ gentamicin, I ⫽ imipenem, IOP ⫽ intraocular pressure, P ⫽ posterior, PCN ⫽ penicillin, SOR ⫽ silicone oil removal †

(with subsequent normalization of intraocular pressure) and phacoemulsification/intraocular lens implantation for treatment of the cataract. One other eye with retinal detachment was treated with pars plana vitrectomy and scleral buckling. Two eyes with retinal detachments were treated successfully with indirect laser demarcation and cryotherapy. Two patients with retinal detachments did not undergo repair because one died within 10 days, and the other patient’s blind and painful eye was enucleated. The mortality rate among patients in this series was high, particularly among those with fungal infections. Six patients (29%) died within 2 months of diagnosis with endogenous endophthalmitis, five of whom had disseminated fungemia. Three of these patients had disseminated aspergillosis, one had a brain abscess caused by Pseudallescheria boydii, one had candidemia with cerebral embolization of an aortic fungus ball, and one had methicillinresistant Staphylococcus aureus sepsis. Another two patients VOL. 137, NO. 4

died 10 months (fungemia with Cladophialophora devriesii) and 27 months (presumptive disseminated tuberculosis and stroke in a patient with the AIDS syndrome) following treatment for endophthalmitis. Only 18 eyes had information available on final visual acuity (defined as the last visual acuity recorded in the medical record after initial treatment). Two patients succumbed to systemic fungal infections 4 and 10 days after diagnosis, respectively, and became unresponsive soon after initial examination and treatment. The third patient for whom a final visual acuity was unavailable had an intravenous drug abuse-related endogenous endophthalmitis and failed to return for follow-up after pars plana vitrectomy. Overall, 8 (44%) of 18 eyes achieved a visual acuity of 20/400 or better. Ten (56%) of 18 eyes achieved a visual acuity of worse than 20/400, including 2 eyes that were eviscerated and 1 that was enucleated. Presentation with panophthalmitis was associated with evisceration

CULTURE-PROVEN ENDOGENOUS ENDOPHTHALMITIS

729

TABLE 5. Comparison of Fungal and Bacterial Endogenous Endophthalmitis FUNGAL* Clinical Features

1. 2. 3. 4. 5.

Onset of symptoms Bilateral disease† (⫹) Blood cultures Final VA ⱖ 20/400 Final VA - LP/NLP

BACTERIAL

(Essman, 1997)

Current Series

(Okada, 1994)

Current Series

Slow 6/18 (33%) 0% 65% 30%

Slow 5/11 (45%) 1/11 (9%) 4/11 (27%) 6/11 (45%)

Rapid 4/28 (14%) 72% 22% 38%

Rapid 1/8 (13%) 4/8 (50%) 4/8 (50%) 4/8 (50%)

*Includes Candida species and Aspergillus species only; molds excluded. Fellow eyes with clinical evidence of involvement included (fellow eyes either had negative culture results or no cultures performed). †

(P ⫽ .005 ⫻ 2-tailed Fisher test), as both eyes that presented with panophthalmitis were eventually eviscerated (1 had Streptococcus bovis and 1 had Klebsiella pneumoniae). Of the eyes with bacterial endophthalmitis, four (50%) of eight had final visual acuities worse than 20/400, including two eviscerations. Eyes with endophthalmitis caused by Aspergillus fared the worst; all three with recorded final visual acuities were light perception or worse (no light perception and enucleation). Aspergillus endophthalmitis had worse visual outcomes than either Candida (P ⫽ .036) or bacterial endophthalmitis (P ⫽ .024 by Mann-Whitney test). Eyes with endophthalmitis caused by Candida had a trend toward better visual outcome than the bacterial cases, but this was not statistically significant (P ⫽ .58); three (60%) of the five eyes with endophthalmitis caused by Candida achieved final visual acuities of 20/400 or better.

Table 5 compares fungal and bacterial endogenous endophthalmitis. All three series appear to demonstrate that bilateral disease is more common among patients with fungal than bacterial endogenous endophthalmitis. Better visual outcomes in fungal endophthalmitis were reported in the Essman series4; however, this may reflect their smaller proportion of Aspergillus cases (three or 15% of 20 eyes), a more virulent organism, compared with the current series (four or 27% of 11 eyes). Patients with endogenous endophthalmitis caused by Aspergillus generally had worse visual outcomes compared with those caused by Candida; this finding agrees with the Essman series,4 in which 76% of the Candida cases but none of the Aspergillus cases achieved a final visual acuity of 20/400 or better. In the current series, 50% of bacterial cases had final visual acuities worse than 20/400, which is somewhat less than the Okada series3 which reported 78% of patients with final vision worse than 20/400. Of note, two cases of bacterial infection presented with panophthalmitis and were eventually eviscerated. This is consistent with the poor visual prognosis associated with panophthalmitis according to the classification system proposed by Greenwald and associates7 and their observation that every case of bacterial panophthalmitis in their series resulted in either blindness and phthisis or enucleation. The high mortality rate among patients with endogenous endophthalmitis in the current series is not surprising given that most patients who develop endogenous endophthalmitis are severely debilitated. This is particularly true for the patients with disseminated fungemia. A recent study reported a 77% mortality rate among patients with Candida endophthalmitis and known systemic candidemia, suggesting that ocular involvement is a good predictor of mortality for systemically ill patients.16 In our study, however, only one patient who died soon after diagnosis had Candida endophthalmitis. Mortality is probably related to the extent and severity of (transient vs disseminated) candidal infection. Endogenous endophthalmitis is associated with a high frequency of fungal isolates, poor visual outcomes, and

DISCUSSION THIS REVIEW OF 21 EYES OF 21 PATIENTS WITH ENDOGE-

nous endophthalmitis at our institution revealed a predominance of fungal isolates. Candida albicans was the most common causative organism, accounting for 33% of all cases, which is similar to a recent review of 34 eyes with endogenous endophthalmitis,5 in which 38% of cases were caused by this organism. It is known that Candida endophthalmitis can masquerade as uveitis and have a gradual onset with a relatively indolent course. We found that Candida endophthalmitis was associated with an incorrect initial diagnosis of uveitis. Previously published series have reported rates of incorrect initial diagnosis for Candida endophthalmitis approaching 50%.3,5 This underscores the need for the ophthalmologist to maintain a high suspicion of endogenous endophthalmitis for patients with intraocular inflammation and a recent history of hospitalization, significant medical comorbidities, or a history of candidal infection. 730

AMERICAN JOURNAL

OF

OPHTHALMOLOGY

APRIL 2004

high mortality. Most patients are seriously ill and hospitalized; however, any patient with intraocular inflammation and a history of recent hospitalization or systemic risk factors should raise suspicion of endogenous endophthalmitis. Unfortunately, visual outcomes remain largely influenced by the causative organism, with Aspergillus and gram-negative infections having the worst prognosis. ACKNOWLEDGMENTS

The authors thank Giovanni Gregori, PhD, and Bill Feuer, MS, for their invaluable help with the statistical analysis of this case series.

REFERENCES 1. Bohigian GM, Olk RJ. Factors associated with a poor visual result in endophthalmitis. Am J Ophthalmol 1986;101:332– 334. 2. Puliafito CA, Baker AS, Haaf J, et al. Infectious endophthalmitis. Review of 36 cases. Ophthalmology 1982;89:921– 929. 3. Okada AA, Johnson RP, Liles WC, et al. Endogenous bacterial endophthalmitis. Report of a 10-year retrospective study. Ophthalmology 1994;101:832–838. 4. Essman TF, Flynn HW Jr., Smiddy WE, et al. Treatment outcomes in a 10-year study of endogenous fungal endophthalmitis. Ophthalmic Surg Lasers 1997;28:185–194. 5. Binder MI, Chua J, Kaiser PK, et al. Endogenous endophthalmitis. An 18-year review of culture-positive cases at a tertiary care center. Medicine 2003;82:97–105.

VOL. 137, NO. 4

6. Kostick DA, Foster RE, Lowder CY, et al. Endogenous endophthalmitis caused by Candida albicans in a healthy woman. Am J Ophthalmol 1992;113:593–595. 7. Valluri S, Moorthy RS, Liggett PE, et al. Endogenous Aspergillus endophthalmitis in an immunocompetent individual. Int Ophthalmol 1993;17:131–135. 8. Flynn HW Jr., Brod RD, Han DP, Miller D. Endophthalmitis: diagnosis, treatment, prevention: In: Spaeth GL, editor. Ophthalmic surgery, principles and practice, 3rd ed. Philadelphia: WB Saunders, 2003:663–677. 9. Chee SP, Jap A. Endogenous endophthalmitis. Curr Opinion Ophthalmol 2001;12:464 –470. 10. Weishaar PD, Flynn HW Jr., Murray TG, et al. Endogenous Aspergillus endophthalmitis: clinical features and treatment outcomes. Ophthalmology 1998;105:57–65. 11. Greenwald MJ, Wohl LG, Sell CH. Metastatic bacterial endophthalmitis: a contemporary reappraisal. Surv Ophthalmol 1986;31:81–101. 12. Wong JS, Chan TK, Lee HM, et al. Endogenous bacterial endophthalmitis: an East Asian experience and a reappraisal of a severe ocular affliction. Ophthalmology 2000;107:1483– 1491. 13. Brod RD, Flynn HW Jr., Miller D. Endogenous fungal endophthalmitis. In: Tasman W, Jaeger EA, editors. Duane’s clinical ophthalmology. Volume 3. Chapter 11. Philadelphia: J. B. Lippincott, 2000. 14. Michelson JB, Friedlander MH. Endophthalmitis of drug abuse. Int Ophthalmol Clin 1987;27:120 –126. 15. Olson JC, Flynn HW Jr., Forster RK, et al. Results in the treatment of postoperative endophthalmitis. Ophthalmology 1983;90:692–699. 16. Menezes AV, Sigesmund DA, Demajo WA, et al. Mortality of hospitalized patients with Candida endophthalmitis. Arch Intern Med 1994;154:2093–2097.

CULTURE-PROVEN ENDOGENOUS ENDOPHTHALMITIS

731