- Email: [email protected]
Pars Plana Vitrectomy in the Management of Endogenous Candida Endophthalmitis
PARS PLANA VITRECTOMY IN T H E MANAGEMENT O F E N D O G E N O U S CANDIDA E N D O P H T H A L M I T I S R O B E R T C. S N I P , M.D.,
AND R O N A L D G. M I C H E L S ,
Baltimore,
Endophthalmitis caused by the fungus Candida albicans and diagnosed by histopathologic examination was first report ed in 1943. l The first case of Candida endophthalmitis suspected on clinical grounds was presented in 1958, 2 and suc cessful treatment of this condition by parenteral amphotericin B therapy was reported in I960. 3 Subsequently, the rela tionship of various predisposing factors to Candida sepsis and endophthalmitis has been stressed, 4 and the occurrence of Candida endophthalmitis after sepsis associated with parenteral drug abuse has been documented. 5 - 7 To our knowledge, pars plana (closed) vitrectomy has not been previously re ported in the diagnosis and treatment of seemingly isolated endogenous Candida endophthalmitis, although vitrectomy and intravitreal amphotericin B therapy have been used in the management of exoge nous Candida endophthalmitis after se vere ocular trauma. 8 During pars plana vitrectomy, we obtained a biopsy speci men to confirm the clinical diagnosis of endogenous Candida endophthalmitis and to provide material for culture and antifungal sensitivity studies. Appropri ate antifungal therapy resulted in rapid and seemingly complete resolution of the infection and return to normal vision.
M.D.
Maryland had injected cocaine, diazepam (Valium), sodium secobarbital (Seconal), and opiates recently, al though he was participating in a local methadone program. Five weeks before admission he had been hospitalized elsewhere for treatment of a skin ab scess on the dorsum of the left hand secondary to drug injection. General examination was otherwise unremarkable. Visual acuity was R.E.: 6/6 (20/20) without correc tion, and L.E.: 6/60 (20/200). Vision in the left eye was not improved by refraction. The right eye ap peared normal and the left eye was moderately injected (Fig. 1). The pupillary response to light was normal in both eyes. Ocular motility was normal. Slit-lamp examination revealed that the anterior segment of the right eye was normal. The left cornea was clear and no keratitic precipitates were seen. The anterior chamber of the left eye was of normal depth, and moderate flare and a moderate number of cells were present in the aqueous humor. The iris and lens appeared normal in both eyes, and the intraocular pressure was 18 mm Hg bilaterally. The anterior vitreous body of the right eye appeared normal, and the anterior vitreous body of the left eye contained many cells. Examination of the fundus of the right eye was normal. In the left eye two white "fluff-balls" were present in the cortical vitreous gel located approxi mately 2 disk diameters inferotemporally to the macula (Fig. 2). Vitreous strands connected these two opacities, and a sheet of condensed vitreous strand extended toward the periphery. There was no evidence of retinitis or retinal detachment. A clinical diagnosis of presumed Candida en dophthalmitis was made, and the patient was admit ted for examination. Detailed medical history and
CASE REPORT On Sept. 13, 1975, a 27-year-old white man was seen here with a history of redness and decreased vision in the left eye for four days. Previously he had good uncorrected vision in both eyes. The patient admitted to a history of parenteral drug abuse; he From the Wilmer Ophthalmological Institute, Johns Hopkins University School of Medicine, Bal timore, Maryland. Reprint requests to Ronald Michels, M.D., Wilmer 116, Johns Hopkins Hospital, 601 N. Broadway, Baltimore, MD 21205.
Fig. 1 (Snip and Michels). External appearance of eye with Candida endophthalmitis demonstrating moderate injection with chemosis.
699
700
AMERICAN JOURNAL OF OPHTHALMOLOGY
Fig. 2 (Snip and Michels). Two white fluff-balls (arrows) in the posterior vitreous cavity of the eye with Candida endophthalmitis.
physical examination showed no specific evidence of systemic disease: no cardiac murmurs were noted, and the site of the previous skin abscess was healed. The patient's body temperature was 38°C (100.2°F). Scarring was present on the volar surface of both forearms, the result of numerous injections. Various laboratory tests were performed to search for other foci of infection. Tests included blood chemistry studies, using an automated multiple-analysis sys tem (SMA-6 and SMA-12), which were normal. Serum creatinine values and liver function tests were normal. Hematology profile included a hematocrit value of 44% and a white blood cell count of 18,000/mm. 3 White blood cell differential count showed 5% juvenile forms, 82% polymorphonuclear leukocytes, 12% lymphocytes, and 1% monocytes. Urinalysis, electrocardiogram, and chest x-ray films were normal. A lumbar puncture was performed, and examination of the cerebrospinal fluid was within normal limits. Numerous blood cultures were performed and showed no growth of bacteria or fungi. A pars plana vitrectomy using the Douvas rotoextractor was performed on the left eye on Sept. 17. The instrument, fitted with an overlying fiberoptic sleeve, was introduced into the eye in the superotemporal quadrant through a sclerotomy located 5 mm posterior and parallel to the corneoscleral limbus. An operating microscope, fiberoptic intra ocular illumination, and a contact lens were used to provide visualization throughout the operation. As much of the formed vitreous gel was excised as possible, including the two white fluff-balls in the inferotemporal vitreous body. In addition the pos terior hyaloid of the vitreous body, posterior to the equator, was mechanically separated from the retina
NOVEMBER, 1976
and excised. The operation was uncomplicated and the clear lens was left intact. No intraocular bleed ing occurred. The scierai incision was closed with five interrupted sutures of 8-0 black silk. Material removed from the vitreous cavity was immediately cultured for bacteria and fungi and studied by light microscopy. The white fluff-balls excised from the vitreous cavity showed many cells including polymorphonuclear leukocytes and mononuclear cells, necrotic cells, and dense aggre gates of fungi characterized by pseudohyphae and ovoid, budding yeast forms (Figs. 3 and 4). The fungi stained positively with PAS and Gomori stains. Cultures of material excised from the vitreous cavity grew C. albicans, and sensitivity testing dem onstrated that the fungus was sensitive to flucytosine (5-FC) with a minimum inhibitory concentra tion equal to or less than 2.5 mg/ml. After vitrectomy the patient was treated with amphotericin B in a dosage of 0.6 mg/kg of body weight, administered intravenously every other day. This medication was continued for one month, during which time the patient received a total dose of 678 mg and his serum creatinine level rose from 1.1 to 2.3 mg/100 ml. The left eye healed well and was treated with a cycloplegic medication adminis tered topically. The patient was discharged from the hospital on Oct. 17, at which time intravenous therapy with amphotericin B was discontinued and treatment with flucytosine was begun in a dosage of 2 g taken orally, four times daily. Visual acuity in the left eye was 6/6 (20/20) eight days after pars plana vitrectomy, and there was no evidence of residual fungus growth in the vitreous cavity. Therapy with flucytosine was discontinued six weeks after discharge from the hospital because there was no clinical evidence of residual fungal growth. At that time, liver function tests were nor mal and the blood urea nitrogen level was 15 mg/100 ml. On Jan. 27, 1976, corrected visual acuity in the left eye was 6/5 (20/15), the eye was quiet, and appeared well healed. The vitreous cavity contained a few cells, but there was no evidence of fungal growth or traction band formation in the vitreous cavity. The retina had a mildly increased light reflex from the internal limiting membrane temporal to the macula (Fig. 5). DISCUSSION
Candida endophthalmitis caused by fungi disseminated through the blood stream is being recognized with increas ing frequency. The intraocular infection probably begins with foci of fungal growth in the choroid or retina. The vitre ous body may become involved by local inflammatory reaction in the cortical vit reous gel overlying the retinal lesion, or fungi may break through the internal limiting membrane of the retina and
VOL. 82 NO. 5
PARS PLANA VITRECTOMY
701
Fig. 3 (Snip and Michels). White fluff-balls excised from the vitreous cavity show many necrotic inflammatory cells (asterisk) and many fungi (arrow) (PAS, x500).
Fig. 4 (Snip and Michels). Fungi removed from the vitreous cavi ty demonstrate pseudohyphae and budding blastospores characteristic of Candida species (PAS, x900).
702
AMERICAN JOURNAL OF OPHTHALMOLOGY
Fig. 5 (Snip and Michels). Ophthalmoscopic ap pearance of eye with Candida endophthalmitis four months after pars plana vitrectomy and treatment with antifungal medications administered systemically. The eye appears well healed, and visual acuity is 6/5 (20/15).
grow into the vitreous cavity. Early diag nosis of Candida endophthalmitis and therapy with specific antifungal medi cations seem to be useful in salvaging vision, 4 , 9 - 1 4 although spontaneous resolu tion of Candida retinitis diagnosed clini cally has been reported. 15 However, in some cases of Candida endophthalmitis the results of therapy have been poor despite use of antifungal medications, 7,9,16 a n ( j r e s u l t s are particularly disap pointing when fungal growth in the vitre ous body has become established. 7 * 12,13,16 Currently, Candida endophthalmitis is treated with amphotericin B, which re quires intravenous administration and may be complicated by toxic effects such as renal damage. Occasionally, concur rent medical problems may preclude the use of amphotericin B, or complications caused by amphotericin B may limit the amount that can be given. Flucytosine is a newer antifungal med ication that is less toxic to body tissues and can be administered orally. This
NOVEMBER, 1976
medication is readily absorbed from the gastrointestinal tract and is not metabo lized by mammalian cells. It is converted by enzymes in certain fungi to 5fluorouracil which is toxic to the metabo lism of the fungus. 17 Flucytosine has a narrower spectrum of antifungal activity than amphotericin B, and the emergence of fungus strains resistant to flucytosine may limit its usefulness. Culture and sen sitivity studies are necessary before be ginning treatment with flucytosine because resistant strains of Candida spe cies have been reported. 7 ' 14 · 18 The pars plana vitrectomy technique permits one to obtain biopsy specimens from the vitreous cavity to confirm the clinical diagnosis of fungal endophthal mitis by direct examination of the vitre ous tissue, and to identify the specific fungus and determine its sensitivity to antifungal medications by culture tech niques. In addition, dense colonies of fungi growing in the vitreous body can be mechanically removed, and excision of the vitreous gel may improve diffusion of systemically administered antifungal medications within the vitreous cavity. This may reduce the total dosage of anti fungal medication required to cure the intraocular infection. Also, excision of posterior cortical vitreous gel may pre vent vitreoretinal traction. 19 Dense vitreoretinal membranes and traction bands in the vitreous body are commonly seen after treatment of intravitreal fungal growth with amphotericin B 1 2 and can result in visual loss despite eradication of the fungi (Figs. 6 and 7). Pars plana vitrectomy is not necessary in every eye with fungal endophthalmitis, but it is useful when fungi are present in the vitreous cavity and cannot be isolated from another location. In our patient the left eye was the only site of apparent involvement, and the clinical diagnosis could not be confirmed without direct biopsy. We were not willing to begin a
VOL. 82 NO. 5
PARS PLANA VITRECTOMY
Fig. 6 (Snip and Michels). Dense vitreoretinal membrane partially covers the optic nerve and mac ula in an eye with Candida endophthalmitis treated by intravenous administration of amphotericin B. A fungus-ball persists in the vitreous cavity (arrow) (From Michels 19 ).
potentially long and toxic course of am photericin B therapy without confirma tion of the clinical diagnosis. Also, by culture techniques, we determined that the fungus was sensitive to flucytosine, and this provided an alternative to am photericin B therapy. Excision of the intravitreal aggregates of fungi and removal of the formed vitreous gel enhanced the response of this infection to antifungal medications and prevented development of vitreoretinal traction. Pars plana vitrectomy is a major surgi cal procedure associated with severe com plications, including possible damage to the lens, retinal detachment, or intraocu lar hemorrhage. We currently reserve this operation for cases of mycotic endoph thalmitis with vitreous involvement in which the fungus cannot be isolated from another location or in which there is pro gression of the intraocular infection de spite therapy with antifungal medications administered systemically. Pars plana vit rectomy may also be useful in eyes that
703
develop dense vitreoretinal membranes or traction bands in the vitreous gel as a complication of mycotic endophthalmitis to relieve tractional retinal detachment or distortion of the retina. We have not employed direct intravitreal instillation of amphotericin B in eyes undergoing vitrectomy for mycotic endophthalmitis, although laboratory ex periments have demonstrated that ocular penetration of amphotericin B is poor in the rabbit eye. 20 It has been suggested that penetration of large molecules into the vitreous body is retarded by the molecular-sieve effect of the vitreous gel, 21 and we feel that excision of the formed vitreous gel may improve diffu sion of systemically administered anti fungal medications within the vitreous cavity. Also, there is uncertainty regard ing the concentration of amphotericin B that can be injected into the vitreous cavity without causing retinal damage. One group of investigators found no evi dence of retinal damage in rabbit eyes when 5 or 10 μg of amphotericin B was
Fig. 7 (Snip and Michels). Dense vitreoretinal membrane covers optic nerve and causes traction on adjacent retina in an eye with Candida endophthal mitis treated by intravenous administration of amphotericin B.
704
AMERICAN JOURNAL OF OPHTHALMOLOGY
injected into the center of the vitreous cavity. 22 Other investigators using a simi lar experimental method found histopathologic evidence of retinal necrosis when 1 μg of amphotericin B was injected into the vitreous cavity of rabbit eyes. 23 Even lower amounts of amphotericin B in the vitreous cavity could cause retinal dam age after vitrectomy, because the medica tion could readily diffuse throughout the vitreous cavity, and retinal damage is partly dependent on the proximity of the injected material to the retina. SUMMARY
A 27-year-old white man had endoge nous Candida endophthalmitis with fungi in the vitreous cavity, presumably caused by hematogenous spread related to drug abuse, and was treated by pars plana vitrectomy and antifungal medica tions administered systemically after sur gery. There was no evidence of other systemic involvement, and excision of fungi from the vitreous cavity confirmed the clinical diagnosis and determined the sensitivity of this fungus to antifungal medications. The fungus was sensitive to flucytosine (5-FC), which the patient re ceived orally since it was less toxic to body tissues than amphotericin B. The intraocular infection cleared rapidly after vitrectomy, and visual acuity returned to 6/5 (20/15). REFERENCES 1. Miale, J. B.: Candida albicans infection con fused with tuberculosis. Arch. Pathol. 35:427,1943. 2. Van Buren, J. M.: Septic retinitis due to Candi da albicans. Arch. Pathol. 65:137, 1958. 3. Louria, D. B., and Dineen, P.: Amphoter icin B in the treatment of disseminated moniliasis. J. A. M. A. 174:273, 1960. 4. Michelson, P. E., Stark, W. J., Reeser, F., and Green, W. R.: Endogenous Candida endophthalmi tis. Int. Ophthalmol. Clin. 11:125, 1971. 5. Cherubin, C. E.: The medical sequelae of nar cotic addiction. Ann. Intern. Med. 67:23, 1967. 6. Sugar, H. S., Mandell, G. H., and Shalev, J.: Metastatic endophthalmitis associated with injec
NOVEMBER, 1976
tion of addictive drugs. Am. J. Ophthalmol. 71:1055, 1971. 7. Stone, R. D., Irvine, A. R., and O'Connor, G. R.: Candida endophthalmitis. Report of an unu sual case with isolation of the etiologic agent by vitreous biopsy. Ann. Ophthalmol. 7:757, 1975. 8. Peyman, G. A., Vastine, D. W., and Diamond, J. G.: Vitrectomy in exogenous Candida endoph thalmitis. Albrecht von Graefe's Arch. Ophthalmol. 197:55, 1975. 9. Fishman, L. S., Griffin, J. R., Sapico, F. L., and Hecht, R.: Hematogenous Candida endophthalmi tis. A complication of candidemia. N. Engl. J. Med. 286:675, 1972. 10. Greene, W. H., and Wiernik, P. H.: Candida endophthalmitis. Am. J. Ophthalmol. 74:1100,1972. 1.1. Weinstein, A. J., Johnson, E. H., and Moellering, R. C : Candida endophthalmitis. A complica tion of candidemia. Arch. Intern. Med. 132:749, 1973. 12. Griffin, J. R., Pettit, T. H., Fishman, L. S., and Foos, R. Y.: Blood-borne Candida endophthalmitis. Arch. Ophthalmol. 89:450, 1973. 13. Edwards, J. E., Jr., Foos, R. Y., Montgomerie, J. Z., and Guze, L. B.: Ocular manifestations of Candida septicemia. Review of 76 cases of hematog enous Candida endophthalmitis. Medicine 53:47, 1973. 14. Robertson, D. M., Riley, F. C , and Hermans, P. E.: Endogenous Candida oculomycosis. Arch. Ophthalmol. 91:33, 1974. 15. Dellon, A. L., Stark, W. J., and Chretien, P. B.: Spontaneous resolution of endogenous Candida en dophthalmitis complicating intravenous hyperalimentation. Am. J. Ophthalmol. 79:648, 1975. 16. Sheldon, G. M., and Alexander, R. L.: Candi da endophthalmitis. Can. J. Ophthalmol. 9:146, 1974. 17. Jones, B. R., Richards, A. B., and Morgan, G.: Direct fungal infection of the eye in Britain. Trans. Ophthalmol. Soc. U. K. 89:727, 1969. 18. Fass, R. J., and Perkins, R. L.: 5-fluorocytosine in the treatment of cryptococcal and Can dida mycoses. Ann. Intern. Med. 74:535, 1971. 19. Michels, R. G.: Vitreoretinal and anterior seg ment surgery through the pars plana. Ann. Ophthal mol. In press. 20. Green, W. R., Bennett, J. E., and Goos, R. D.: Ocular penetration of amphotericin B. A report of laboratory studies and a casé report of post-surgical Cephalosporium endophthalmitis. Arch. Ophthal mol. 73:769, 1965. 21. Balazs, E. A.: The molecular biology of the vitreous. In McPherson, A. (ed.): New and Contro versial Aspects of Retinal Detachment. International Symposium. New York, Harper and Row, 1968, pp. 3-15. 22. Axelrod, A. J., Peyman, G. A., and Apple, D. J.: Toxicity of intravitreal injection of amphoteri cin B. Am. J. Ophthalmol. 76:578, 1973. 23. Souri, E. N., and Green, W. R.: Intravitreal amphotericin B toxicity. Am. J. Ophthalmol. 78:77, 1974.
AND R O N A L D G. M I C H E L S ,
Baltimore,
Endophthalmitis caused by the fungus Candida albicans and diagnosed by histopathologic examination was first report ed in 1943. l The first case of Candida endophthalmitis suspected on clinical grounds was presented in 1958, 2 and suc cessful treatment of this condition by parenteral amphotericin B therapy was reported in I960. 3 Subsequently, the rela tionship of various predisposing factors to Candida sepsis and endophthalmitis has been stressed, 4 and the occurrence of Candida endophthalmitis after sepsis associated with parenteral drug abuse has been documented. 5 - 7 To our knowledge, pars plana (closed) vitrectomy has not been previously re ported in the diagnosis and treatment of seemingly isolated endogenous Candida endophthalmitis, although vitrectomy and intravitreal amphotericin B therapy have been used in the management of exoge nous Candida endophthalmitis after se vere ocular trauma. 8 During pars plana vitrectomy, we obtained a biopsy speci men to confirm the clinical diagnosis of endogenous Candida endophthalmitis and to provide material for culture and antifungal sensitivity studies. Appropri ate antifungal therapy resulted in rapid and seemingly complete resolution of the infection and return to normal vision.
M.D.
Maryland had injected cocaine, diazepam (Valium), sodium secobarbital (Seconal), and opiates recently, al though he was participating in a local methadone program. Five weeks before admission he had been hospitalized elsewhere for treatment of a skin ab scess on the dorsum of the left hand secondary to drug injection. General examination was otherwise unremarkable. Visual acuity was R.E.: 6/6 (20/20) without correc tion, and L.E.: 6/60 (20/200). Vision in the left eye was not improved by refraction. The right eye ap peared normal and the left eye was moderately injected (Fig. 1). The pupillary response to light was normal in both eyes. Ocular motility was normal. Slit-lamp examination revealed that the anterior segment of the right eye was normal. The left cornea was clear and no keratitic precipitates were seen. The anterior chamber of the left eye was of normal depth, and moderate flare and a moderate number of cells were present in the aqueous humor. The iris and lens appeared normal in both eyes, and the intraocular pressure was 18 mm Hg bilaterally. The anterior vitreous body of the right eye appeared normal, and the anterior vitreous body of the left eye contained many cells. Examination of the fundus of the right eye was normal. In the left eye two white "fluff-balls" were present in the cortical vitreous gel located approxi mately 2 disk diameters inferotemporally to the macula (Fig. 2). Vitreous strands connected these two opacities, and a sheet of condensed vitreous strand extended toward the periphery. There was no evidence of retinitis or retinal detachment. A clinical diagnosis of presumed Candida en dophthalmitis was made, and the patient was admit ted for examination. Detailed medical history and
CASE REPORT On Sept. 13, 1975, a 27-year-old white man was seen here with a history of redness and decreased vision in the left eye for four days. Previously he had good uncorrected vision in both eyes. The patient admitted to a history of parenteral drug abuse; he From the Wilmer Ophthalmological Institute, Johns Hopkins University School of Medicine, Bal timore, Maryland. Reprint requests to Ronald Michels, M.D., Wilmer 116, Johns Hopkins Hospital, 601 N. Broadway, Baltimore, MD 21205.
Fig. 1 (Snip and Michels). External appearance of eye with Candida endophthalmitis demonstrating moderate injection with chemosis.
699
700
AMERICAN JOURNAL OF OPHTHALMOLOGY
Fig. 2 (Snip and Michels). Two white fluff-balls (arrows) in the posterior vitreous cavity of the eye with Candida endophthalmitis.
physical examination showed no specific evidence of systemic disease: no cardiac murmurs were noted, and the site of the previous skin abscess was healed. The patient's body temperature was 38°C (100.2°F). Scarring was present on the volar surface of both forearms, the result of numerous injections. Various laboratory tests were performed to search for other foci of infection. Tests included blood chemistry studies, using an automated multiple-analysis sys tem (SMA-6 and SMA-12), which were normal. Serum creatinine values and liver function tests were normal. Hematology profile included a hematocrit value of 44% and a white blood cell count of 18,000/mm. 3 White blood cell differential count showed 5% juvenile forms, 82% polymorphonuclear leukocytes, 12% lymphocytes, and 1% monocytes. Urinalysis, electrocardiogram, and chest x-ray films were normal. A lumbar puncture was performed, and examination of the cerebrospinal fluid was within normal limits. Numerous blood cultures were performed and showed no growth of bacteria or fungi. A pars plana vitrectomy using the Douvas rotoextractor was performed on the left eye on Sept. 17. The instrument, fitted with an overlying fiberoptic sleeve, was introduced into the eye in the superotemporal quadrant through a sclerotomy located 5 mm posterior and parallel to the corneoscleral limbus. An operating microscope, fiberoptic intra ocular illumination, and a contact lens were used to provide visualization throughout the operation. As much of the formed vitreous gel was excised as possible, including the two white fluff-balls in the inferotemporal vitreous body. In addition the pos terior hyaloid of the vitreous body, posterior to the equator, was mechanically separated from the retina
NOVEMBER, 1976
and excised. The operation was uncomplicated and the clear lens was left intact. No intraocular bleed ing occurred. The scierai incision was closed with five interrupted sutures of 8-0 black silk. Material removed from the vitreous cavity was immediately cultured for bacteria and fungi and studied by light microscopy. The white fluff-balls excised from the vitreous cavity showed many cells including polymorphonuclear leukocytes and mononuclear cells, necrotic cells, and dense aggre gates of fungi characterized by pseudohyphae and ovoid, budding yeast forms (Figs. 3 and 4). The fungi stained positively with PAS and Gomori stains. Cultures of material excised from the vitreous cavity grew C. albicans, and sensitivity testing dem onstrated that the fungus was sensitive to flucytosine (5-FC) with a minimum inhibitory concentra tion equal to or less than 2.5 mg/ml. After vitrectomy the patient was treated with amphotericin B in a dosage of 0.6 mg/kg of body weight, administered intravenously every other day. This medication was continued for one month, during which time the patient received a total dose of 678 mg and his serum creatinine level rose from 1.1 to 2.3 mg/100 ml. The left eye healed well and was treated with a cycloplegic medication adminis tered topically. The patient was discharged from the hospital on Oct. 17, at which time intravenous therapy with amphotericin B was discontinued and treatment with flucytosine was begun in a dosage of 2 g taken orally, four times daily. Visual acuity in the left eye was 6/6 (20/20) eight days after pars plana vitrectomy, and there was no evidence of residual fungus growth in the vitreous cavity. Therapy with flucytosine was discontinued six weeks after discharge from the hospital because there was no clinical evidence of residual fungal growth. At that time, liver function tests were nor mal and the blood urea nitrogen level was 15 mg/100 ml. On Jan. 27, 1976, corrected visual acuity in the left eye was 6/5 (20/15), the eye was quiet, and appeared well healed. The vitreous cavity contained a few cells, but there was no evidence of fungal growth or traction band formation in the vitreous cavity. The retina had a mildly increased light reflex from the internal limiting membrane temporal to the macula (Fig. 5). DISCUSSION
Candida endophthalmitis caused by fungi disseminated through the blood stream is being recognized with increas ing frequency. The intraocular infection probably begins with foci of fungal growth in the choroid or retina. The vitre ous body may become involved by local inflammatory reaction in the cortical vit reous gel overlying the retinal lesion, or fungi may break through the internal limiting membrane of the retina and
VOL. 82 NO. 5
PARS PLANA VITRECTOMY
701
Fig. 3 (Snip and Michels). White fluff-balls excised from the vitreous cavity show many necrotic inflammatory cells (asterisk) and many fungi (arrow) (PAS, x500).
Fig. 4 (Snip and Michels). Fungi removed from the vitreous cavi ty demonstrate pseudohyphae and budding blastospores characteristic of Candida species (PAS, x900).
702
AMERICAN JOURNAL OF OPHTHALMOLOGY
Fig. 5 (Snip and Michels). Ophthalmoscopic ap pearance of eye with Candida endophthalmitis four months after pars plana vitrectomy and treatment with antifungal medications administered systemically. The eye appears well healed, and visual acuity is 6/5 (20/15).
grow into the vitreous cavity. Early diag nosis of Candida endophthalmitis and therapy with specific antifungal medi cations seem to be useful in salvaging vision, 4 , 9 - 1 4 although spontaneous resolu tion of Candida retinitis diagnosed clini cally has been reported. 15 However, in some cases of Candida endophthalmitis the results of therapy have been poor despite use of antifungal medications, 7,9,16 a n ( j r e s u l t s are particularly disap pointing when fungal growth in the vitre ous body has become established. 7 * 12,13,16 Currently, Candida endophthalmitis is treated with amphotericin B, which re quires intravenous administration and may be complicated by toxic effects such as renal damage. Occasionally, concur rent medical problems may preclude the use of amphotericin B, or complications caused by amphotericin B may limit the amount that can be given. Flucytosine is a newer antifungal med ication that is less toxic to body tissues and can be administered orally. This
NOVEMBER, 1976
medication is readily absorbed from the gastrointestinal tract and is not metabo lized by mammalian cells. It is converted by enzymes in certain fungi to 5fluorouracil which is toxic to the metabo lism of the fungus. 17 Flucytosine has a narrower spectrum of antifungal activity than amphotericin B, and the emergence of fungus strains resistant to flucytosine may limit its usefulness. Culture and sen sitivity studies are necessary before be ginning treatment with flucytosine because resistant strains of Candida spe cies have been reported. 7 ' 14 · 18 The pars plana vitrectomy technique permits one to obtain biopsy specimens from the vitreous cavity to confirm the clinical diagnosis of fungal endophthal mitis by direct examination of the vitre ous tissue, and to identify the specific fungus and determine its sensitivity to antifungal medications by culture tech niques. In addition, dense colonies of fungi growing in the vitreous body can be mechanically removed, and excision of the vitreous gel may improve diffusion of systemically administered antifungal medications within the vitreous cavity. This may reduce the total dosage of anti fungal medication required to cure the intraocular infection. Also, excision of posterior cortical vitreous gel may pre vent vitreoretinal traction. 19 Dense vitreoretinal membranes and traction bands in the vitreous body are commonly seen after treatment of intravitreal fungal growth with amphotericin B 1 2 and can result in visual loss despite eradication of the fungi (Figs. 6 and 7). Pars plana vitrectomy is not necessary in every eye with fungal endophthalmitis, but it is useful when fungi are present in the vitreous cavity and cannot be isolated from another location. In our patient the left eye was the only site of apparent involvement, and the clinical diagnosis could not be confirmed without direct biopsy. We were not willing to begin a
VOL. 82 NO. 5
PARS PLANA VITRECTOMY
Fig. 6 (Snip and Michels). Dense vitreoretinal membrane partially covers the optic nerve and mac ula in an eye with Candida endophthalmitis treated by intravenous administration of amphotericin B. A fungus-ball persists in the vitreous cavity (arrow) (From Michels 19 ).
potentially long and toxic course of am photericin B therapy without confirma tion of the clinical diagnosis. Also, by culture techniques, we determined that the fungus was sensitive to flucytosine, and this provided an alternative to am photericin B therapy. Excision of the intravitreal aggregates of fungi and removal of the formed vitreous gel enhanced the response of this infection to antifungal medications and prevented development of vitreoretinal traction. Pars plana vitrectomy is a major surgi cal procedure associated with severe com plications, including possible damage to the lens, retinal detachment, or intraocu lar hemorrhage. We currently reserve this operation for cases of mycotic endoph thalmitis with vitreous involvement in which the fungus cannot be isolated from another location or in which there is pro gression of the intraocular infection de spite therapy with antifungal medications administered systemically. Pars plana vit rectomy may also be useful in eyes that
703
develop dense vitreoretinal membranes or traction bands in the vitreous gel as a complication of mycotic endophthalmitis to relieve tractional retinal detachment or distortion of the retina. We have not employed direct intravitreal instillation of amphotericin B in eyes undergoing vitrectomy for mycotic endophthalmitis, although laboratory ex periments have demonstrated that ocular penetration of amphotericin B is poor in the rabbit eye. 20 It has been suggested that penetration of large molecules into the vitreous body is retarded by the molecular-sieve effect of the vitreous gel, 21 and we feel that excision of the formed vitreous gel may improve diffu sion of systemically administered anti fungal medications within the vitreous cavity. Also, there is uncertainty regard ing the concentration of amphotericin B that can be injected into the vitreous cavity without causing retinal damage. One group of investigators found no evi dence of retinal damage in rabbit eyes when 5 or 10 μg of amphotericin B was
Fig. 7 (Snip and Michels). Dense vitreoretinal membrane covers optic nerve and causes traction on adjacent retina in an eye with Candida endophthal mitis treated by intravenous administration of amphotericin B.
704
AMERICAN JOURNAL OF OPHTHALMOLOGY
injected into the center of the vitreous cavity. 22 Other investigators using a simi lar experimental method found histopathologic evidence of retinal necrosis when 1 μg of amphotericin B was injected into the vitreous cavity of rabbit eyes. 23 Even lower amounts of amphotericin B in the vitreous cavity could cause retinal dam age after vitrectomy, because the medica tion could readily diffuse throughout the vitreous cavity, and retinal damage is partly dependent on the proximity of the injected material to the retina. SUMMARY
A 27-year-old white man had endoge nous Candida endophthalmitis with fungi in the vitreous cavity, presumably caused by hematogenous spread related to drug abuse, and was treated by pars plana vitrectomy and antifungal medica tions administered systemically after sur gery. There was no evidence of other systemic involvement, and excision of fungi from the vitreous cavity confirmed the clinical diagnosis and determined the sensitivity of this fungus to antifungal medications. The fungus was sensitive to flucytosine (5-FC), which the patient re ceived orally since it was less toxic to body tissues than amphotericin B. The intraocular infection cleared rapidly after vitrectomy, and visual acuity returned to 6/5 (20/15). REFERENCES 1. Miale, J. B.: Candida albicans infection con fused with tuberculosis. Arch. Pathol. 35:427,1943. 2. Van Buren, J. M.: Septic retinitis due to Candi da albicans. Arch. Pathol. 65:137, 1958. 3. Louria, D. B., and Dineen, P.: Amphoter icin B in the treatment of disseminated moniliasis. J. A. M. A. 174:273, 1960. 4. Michelson, P. E., Stark, W. J., Reeser, F., and Green, W. R.: Endogenous Candida endophthalmi tis. Int. Ophthalmol. Clin. 11:125, 1971. 5. Cherubin, C. E.: The medical sequelae of nar cotic addiction. Ann. Intern. Med. 67:23, 1967. 6. Sugar, H. S., Mandell, G. H., and Shalev, J.: Metastatic endophthalmitis associated with injec
NOVEMBER, 1976
tion of addictive drugs. Am. J. Ophthalmol. 71:1055, 1971. 7. Stone, R. D., Irvine, A. R., and O'Connor, G. R.: Candida endophthalmitis. Report of an unu sual case with isolation of the etiologic agent by vitreous biopsy. Ann. Ophthalmol. 7:757, 1975. 8. Peyman, G. A., Vastine, D. W., and Diamond, J. G.: Vitrectomy in exogenous Candida endoph thalmitis. Albrecht von Graefe's Arch. Ophthalmol. 197:55, 1975. 9. Fishman, L. S., Griffin, J. R., Sapico, F. L., and Hecht, R.: Hematogenous Candida endophthalmi tis. A complication of candidemia. N. Engl. J. Med. 286:675, 1972. 10. Greene, W. H., and Wiernik, P. H.: Candida endophthalmitis. Am. J. Ophthalmol. 74:1100,1972. 1.1. Weinstein, A. J., Johnson, E. H., and Moellering, R. C : Candida endophthalmitis. A complica tion of candidemia. Arch. Intern. Med. 132:749, 1973. 12. Griffin, J. R., Pettit, T. H., Fishman, L. S., and Foos, R. Y.: Blood-borne Candida endophthalmitis. Arch. Ophthalmol. 89:450, 1973. 13. Edwards, J. E., Jr., Foos, R. Y., Montgomerie, J. Z., and Guze, L. B.: Ocular manifestations of Candida septicemia. Review of 76 cases of hematog enous Candida endophthalmitis. Medicine 53:47, 1973. 14. Robertson, D. M., Riley, F. C , and Hermans, P. E.: Endogenous Candida oculomycosis. Arch. Ophthalmol. 91:33, 1974. 15. Dellon, A. L., Stark, W. J., and Chretien, P. B.: Spontaneous resolution of endogenous Candida en dophthalmitis complicating intravenous hyperalimentation. Am. J. Ophthalmol. 79:648, 1975. 16. Sheldon, G. M., and Alexander, R. L.: Candi da endophthalmitis. Can. J. Ophthalmol. 9:146, 1974. 17. Jones, B. R., Richards, A. B., and Morgan, G.: Direct fungal infection of the eye in Britain. Trans. Ophthalmol. Soc. U. K. 89:727, 1969. 18. Fass, R. J., and Perkins, R. L.: 5-fluorocytosine in the treatment of cryptococcal and Can dida mycoses. Ann. Intern. Med. 74:535, 1971. 19. Michels, R. G.: Vitreoretinal and anterior seg ment surgery through the pars plana. Ann. Ophthal mol. In press. 20. Green, W. R., Bennett, J. E., and Goos, R. D.: Ocular penetration of amphotericin B. A report of laboratory studies and a casé report of post-surgical Cephalosporium endophthalmitis. Arch. Ophthal mol. 73:769, 1965. 21. Balazs, E. A.: The molecular biology of the vitreous. In McPherson, A. (ed.): New and Contro versial Aspects of Retinal Detachment. International Symposium. New York, Harper and Row, 1968, pp. 3-15. 22. Axelrod, A. J., Peyman, G. A., and Apple, D. J.: Toxicity of intravitreal injection of amphoteri cin B. Am. J. Ophthalmol. 76:578, 1973. 23. Souri, E. N., and Green, W. R.: Intravitreal amphotericin B toxicity. Am. J. Ophthalmol. 78:77, 1974.