Epidemiology, clinical manifestations, and diagnosis of fungal endophthalmitis

INTRODUCTION — Fungal endophthalmitis due to yeasts (eg, Candida albicans) differs from that due to molds (eg, Aspergillus or Fusarium spp). Yeast-associated endophthalmitis is usually endogenous (secondary to fungemia), and treatment is often successful, whereas mold endophthalmitis is usually exogenous (occurring after trauma, surgery, or keratomycosis) and often results in permanent poor vision in the affected eye.

The epidemiology, pathogenesis, clinical features, and diagnosis of fungal endophthalmitis will be reviewed here. The management of fungal endophthalmitis is discussed separately; bacterial endophthalmitis, Fusarium keratitis, and candidemia and other types of Candida infections are also presented elsewhere. (See "Treatment of endogenous endophthalmitis and chorioretinitis due to Candida species" and "Treatment of exogenous endophthalmitis due to Candida species" and "Treatment of endophthalmitis due to molds" and "Bacterial endophthalmitis" and "Mycology, pathogenesis, and epidemiology of Fusarium infection", section on 'Keratitis' and "Management of candidemia and invasive candidiasis in adults" and "Overview of Candida infections".)

DEFINITIONS — The term "endophthalmitis" means bacterial or fungal infection of the vitreous and/or aqueous (figure 1). It may be endogenous, with pathogens seeding the eye via the bloodstream, or exogenous, with pathogens introduced into the eye via the ocular surface.

In the literature, the term "Candida endophthalmitis" is often used to signify either chorioretinitis or endophthalmitis (ie, with vitritis), but "ocular candidiasis" is the more accurate term for this spectrum of intraocular infection because it encompasses both chorioretinitis and endophthalmitis.

Endogenous fungal endophthalmitis — Endogenous fungal endophthalmitis results from fungemia, which may be persistent or transient. Blood cultures may be negative at the time of endophthalmitis diagnosis. In most cases, the fungi seed the highly vascular choroid first. As a consequence, endogenous fungal endophthalmitis often is preceded by choroiditis or chorioretinitis (infection localized to the choroid or the choroid and retina), with minimal inflammation in the vitreous or aqueous. As the infection progresses, marked vitritis may develop (ie, endophthalmitis) and the aqueous may also become involved.

Exogenous fungal endophthalmitis — Exogenous fungal endophthalmitis usually arises following eye surgery, eye trauma, or extension of keratomycosis (fungal infection of the cornea). In cases following surgery involving the anterior segment (eg, cataract surgery) or following keratomycosis, the aqueous is usually involved first.

EPIDEMIOLOGY AND RISK FACTORS — The incidence and risk factors for endophthalmitis caused by yeast (eg, Candida) are different from those caused by molds. Most fungal endophthalmitis cases in temperate climates are endogenous and caused by Candida species [1], while most cases in tropical climates are exogenous and caused by molds [2].

Candida endophthalmitis — Most cases of Candida endophthalmitis are endogenous, but some cases are exogenous and result from ocular trauma or surgery.

Endogenous Candida endophthalmitis — Nearly all cases of endogenous yeast endophthalmitis are caused by Candida species (other yeasts, such as Cryptococcus, rarely cause endophthalmitis) [3-6]. The most common species of Candida to cause endophthalmitis is C. albicans, which is responsible for over 75 percent of ocular candidiasis cases [6-12]. In a series of 174 cases of candidemia, the albicans species of Candida was twice as common in cases with endophthalmitis than in those without endophthalmitis [13]. (See "Candidemia in adults: Epidemiology, microbiology, and pathogenesis", section on 'Prevalence of Candida species'.)

Ocular candidiasis occurs most commonly in hospitalized patients with risk factors for candidemia. A major risk factor for ocular candidiasis is an indwelling central venous catheter. Other risk factors include prolonged duration of candidemia, total parenteral nutrition, broad-spectrum antibiotics, recent abdominal surgery, neutropenia, and glucocorticoid therapy [7,14-18]. Risk factors for candidemia are discussed in greater detail separately. (See "Candidemia in adults: Epidemiology, microbiology, and pathogenesis".)

In patients who do not have a history of recent hospitalization or an indwelling central venous catheter, intravenous drug use is the major risk factor for endogenous Candida endophthalmitis [19,20]. Over a seven-year period (2001 to 2007) at an eye and ear hospital in Australia, 70 percent of the patients diagnosed with endogenous fungal endophthalmitis had a recent history of injection drug use, and Candida was the primary pathogen in culture-positive cases [19]. Candida species, primarily C. albicans, also caused the majority of injection drug use-associated endophthalmitis in a recent study from Boston [21].

The source of the Candida in cases associated with intravenous drug use may be oral flora. In some cases, patients have acknowledged removing a partially dissolved tablet of buprenorphine from his or her mouth prior to diluting and injecting it [19,22,23]. In other cases, oral flora may be introduced when the person licks the needle prior to intravenous injection. This practice appears to be common; one study from New York City found that one-third of the 40 interviewed people acknowledged licking their needles [24]. An unusual source of Candida occurred in the 1980s, when there was a large outbreak of Candida endophthalmitis among people who injected drugs and who used brown heroin [25,26]. This type of heroin requires an acidic diluent, and the lemon juice used for this purpose was thought to be the major source of the Candida [27].

Exogenous Candida endophthalmitis — Exogenous Candida endophthalmitis is rare. In a series of all patients treated for exogenous fungal endophthalmitis at an eye hospital in Florida over a 16-year period, Candida species accounted for only 6 of the 41 cases (15 percent), whereas molds caused 35 cases (85 percent) [28]. Among the six Candida cases, three occurred after surgery (corneal transplant and intraocular lens [IOL] implant exchange in two, cataract surgery in one), two occurred after trauma, and one occurred as an extension of keratomycosis. The time from surgery or keratitis to the diagnosis of endophthalmitis was long (five weeks to six months) but very short in those cases due to trauma (one day).

Some exogenous Candida endophthalmitis cases have followed corneal transplantation (keratoplasty), in which donor-to-host transmission occurred [29,30]. These post-keratoplasty endophthalmitis cases occur in the acute postoperative period in 0.2 percent of corneal transplant recipients, but this incidence increases to 0.7 percent if late-onset cases are included [31]. Candida causes approximately one-third of post-keratoplasty endophthalmitis cases [31].

In a series of 113 patients with fungal endophthalmitis in India, 90 percent of cases were exogenous, caused either by surgery (53 patients) or trauma (48 patients) [32]. As in the series from Florida, the majority of the exogenous cases were due to molds. Twelve (23 percent) of the 52 culture-positive exogenous cases were due to Candida spp, and 75 percent followed eye surgery, including cataract surgery. C. tropicalis (five cases) and C. albicans (four cases) were the most common species.

All species of Candida have been described as causing exogenous endophthalmitis, but C. parapsilosis has been commonly reported in postsurgical outbreaks [33-35]. The most likely reason for this is that C. parapsilosis has a high propensity to survive in irrigation fluids and prosthetic materials [36]. An outbreak of C. parapsilosis endophthalmitis involving 13 patients was traced to an ophthalmic irrigating solution that was contaminated during the manufacturing process and was used during cataract surgery [34].

Mold endophthalmitis — Mold endophthalmitis is rare in the United States except in hot, humid regions such as Florida, where molds have accounted for 6 percent of culture-positive endophthalmitis cases and 25 percent of endogenous fungal endophthalmitis cases [6,37]. Worldwide, mold endophthalmitis is more common in India and other hot, humid countries. A large series from India reported that fungi caused 19 percent of culture-positive endophthalmitis cases and molds accounted for the majority of the fungal cases [2].

Endogenous mold endophthalmitis — Endogenous mold endophthalmitis occurs rarely and is usually associated with intravenous drug use [21,38,39] or immunosuppression [40]. A review of 86 cases of endogenous Aspergillus endophthalmitis published in the English language literature from 1949 to 2001 noted that 27 percent occurred in people who injected intravenous drugs, 23 percent in solid organ transplant recipients, 17 percent in patients with chronic lung disease, and 43 percent in patients who had been treated with glucocorticoids previously [41]. The predominant species was Aspergillus fumigatus, but A. flavus also occurred commonly and was strongly linked to intravenous drug use.

In a retrospective study at a cancer center in Texas between 1991 and 2007, among 15 patients with mold endophthalmitis, all had underlying hematologic malignancies and over one-half had undergone hematopoietic cell transplantation [40]. Fusarium spp (five cases), Aspergillus spp (four), and Scedosporium spp (four) caused 87 percent of cases; the remaining two cases were caused by Mucor spp and Rhizomucor spp.

Exogenous mold endophthalmitis — Mold endophthalmitis is usually exogenous, occurring as a result of extension of keratomycosis, eye surgery, or penetrating ocular trauma [28]. In a series of exogenous fungal endophthalmitis cases seen at an eye hospital in Florida over 16 years, extension of keratomycosis accounted for nearly one-half of the 35 cases of mold endophthalmitis, whereas eye surgery and eye trauma each accounted for approximately 25 percent of cases [28]. In a series from India, 83 percent of 730 fungal endophthalmitis cases were exogenous (47 percent postoperative and 36 percent post-traumatic), and nearly 85 percent were due to molds [2].

In cases that occur as an extension of fungal keratitis, the aqueous humor is nearly always involved first. The microbiology reflects that of keratomycosis, with Fusarium and Aspergillus spp being the most common etiologies. In the Florida study, 13 of 17 keratitis-related mold endophthalmitis cases were due to Fusarium spp, with four occurring between 2004 and 2006 [28]. Some of these cases may reflect the international outbreak of Fusarium keratitis (over 250 cases) that occurred between 2004 and 2006. This outbreak was linked to the use of a specific lens solution and poor contact lens care; one report of a subset of patients from this outbreak noted that 6 percent of patients developed endophthalmitis [42,43]. (See "Clinical manifestations and diagnosis of Fusarium infection", section on 'Keratitis'.)

In the United States and Europe, mold endophthalmitis is rare after eye surgery. When it does occur, it usually presents subacutely rather than acutely, in contrast with most cases of bacterial postoperative endophthalmitis, which present acutely. Two large series of acute (<6 weeks) postcataract surgery endophthalmitis in Florida and France, totaling 173 cases, reported that all cases were due to bacteria: there were no cases of acute fungal endophthalmitis [44,45]. The onset of symptoms and diagnosis are often delayed; in one review that included 10 cases of postoperative exogenous mold endophthalmitis, 60 percent presented two months or more after eye surgery [28].

In tropical regions of the world, postoperative mold endophthalmitis is more common and may present more acutely. In two retrospective reviews of the microbiology of postoperative endophthalmitis in India, molds caused 22 percent of cases [46,47]. In a series from India of 27 postcataract surgery fungal endophthalmitis cases (22 with mold endophthalmitis), over 80 percent presented within four weeks of surgery. Aspergillus was the major etiology in this series and in others.

Some endophthalmitis cases that follow surgical procedures or intraocular injections have been associated with outbreaks in health care settings as illustrated by the following examples:

●One compounding pharmacy in Florida was associated with two concurrent outbreaks of fungal endophthalmitis in nine states, affecting 45 patients (47 cases) during 2011 to 2012 [48]. The two outbreaks together represent the largest reported outbreak of infectious endophthalmitis in the United States. Twenty-one patients developed Fusarium incarnatum-equiseti endophthalmitis after receiving an intraocular injection of compounded Brilliant Blue dye, used off label to stain epiretinal membranes during retinal surgeries. Twenty-six cases of Bipolaris hawaiiensis endophthalmitis developed after 24 patients received intravitreal injections of triamcinolone; two patients received injections in each eye and developed bilateral Bipolaris endophthalmitis. Follow-up was available for 40 of the patients; 39 suffered vision loss, including two who required enucleation. Both of these outbreaks were traced back to a compounding pharmacy that had supplied the contaminated material used for injection.

●An outbreak of Aspergillus endophthalmitis occurred in five patients who had undergone cataract surgery 4 to 15 days earlier during a period of construction at a hospital in Saudi Arabia [49].

●An outbreak of Fusarium endophthalmitis occurred in nine patients who underwent cataract surgery on a single day in the same operating room in a hospital in Turkey [50]. Cefuroxime taken from the same multidose vial had been injected intracamerally in all patients at the end of surgery.

PATHOGENESIS

Endogenous endophthalmitis — Based on experimental studies of endogenous Candida endophthalmitis in rabbits, it is apparent that Candida spp gain entrance to the eye by first seeding the highly vascular choroid (figure 1) [51]. There, the yeasts proliferate and invoke focal inflammation then extend to the vitreous body, causing further inflammation.

In a study of experimentally induced endogenous Aspergillus endophthalmitis in 14 rabbits, one-half of which were immunosuppressed, the location of infection varied by duration of survival after fungemia [52]. In rabbits that survived less than two weeks after fungemia, ocular fungal infection was localized primarily in the choroid, whether or not the animals were immunosuppressed. Large choroidal abscesses were seen in some of the immunosuppressed rabbits, and serous detachments were common in both groups. In immunocompetent and immunosuppressed rabbits surviving more than two weeks, the fungi were located in granulomas or "fluff balls" in the vitreous but were not seen in the choroid or retina.

Exogenous endophthalmitis — The pathogenesis of exogenous endophthalmitis is inoculation of the aqueous and/or vitreous with fungal species. Usually the aqueous is involved first, since most cases follow anterior eye surgery (eg, cataract surgery, corneal transplantation) or eye trauma (figure 1). The infection may progress posteriorly to involve the vitreous, especially if the posterior lens capsule was ruptured during surgery or trauma.

CLINICAL MANIFESTATIONS — Initial symptoms in fungal endophthalmitis may be floaters and a subtle decrease in vision. Progressive loss of vision in the affected eye usually prompts the patient to seek medical attention. Eye pain may be minimal or absent until late in the course. Unless there is concomitant systemic fungal infection, patients do not have fever, chills, or leukocytosis.

Endophthalmitis is suspected clinically based upon the patient's presenting complaints, risk factors (eg, recent indwelling central catheter or intravenous drug use for endogenous endophthalmitis, eye trauma, or surgery for exogenous endophthalmitis), and the ophthalmic examination. (See 'Ophthalmic examination' below.)

Endogenous fungal endophthalmitis — Although the majority of patients have unilateral eye involvement, both eyes can be affected. Symptoms of vision loss may develop days to weeks after the fungemia. There is often minimal pain until the infection is advanced.

If the eye was seeded by transient fungemia that infected no other organs, the patient may be systemically well at the time of presentation with eye symptoms, have a normal white blood count, and have negative blood cultures. This is often the case in outpatients with endogenous yeast or mold endophthalmitis due to injection drug use. Similarly, candidemia may be transient in patients with indwelling central venous catheters in which the organism seeds the eye but no other organs. Such patients have no evidence of systemic Candida infection at the time of presentation with eye symptoms, and the diagnosis of Candida endophthalmitis may be delayed. In a retrospective review of 15 patients with endogenous Candida endophthalmitis, 11 of whom had had an indwelling venous catheter, the average time from symptom onset to treatment was two months [53]. In patients with risk factors for transient fungemia (eg, injection drug use, recent peripherally inserted central venous catheter even if no longer present), endogenous fungal endophthalmitis should be considered in the differential diagnosis of patients presenting with posterior uveitis or panuveitis.

Hospitalized patients with ocular candidiasis rarely report any eye symptoms. In several large series of inpatients in whom screening eye examinations were performed for candidemia, 81 to 98 percent of patients found to have ocular candidiasis reported no visual or eye symptoms (or were unable to report symptoms) at the time of the examination [8,9,54]. Ocular candidiasis is not uncommon; in most series, ocular candidiasis is detected in 10 to 20 percent (range 2 to 26 percent) of candidemic patients who undergo screening eye examinations [8,13,17,55-58]. Chorioretinitis accounts for most such cases while endophthalmitis (ie, with significant vitritis) only occurs in 1 to 2 percent of candidemic patients.

Endogenous mold endophthalmitis also may be clinically silent, especially initially. In an autopsy study of 85 orthotopic liver transplant recipients, Aspergillus endophthalmitis was noted in six patients (7 percent), only one of whom was diagnosed with endophthalmitis before death [59]. Fourteen patients (17 percent) had invasive pulmonary aspergillosis, and 10 (12 percent) had disseminated disease. The eyes were the second most common site of infection. Two patients had eye involvement as the only nonpulmonary site of infection.

Exogenous fungal endophthalmitis — In patients with exogenous fungal endophthalmitis, the onset of symptoms occurs after a variable period of time. A large series from India found that the average time to presentation was 7 days in 48 posttraumatic fungal endophthalmitis cases and 20 days in 53 postoperative cases, although a significant percentage presented within 48 hours (46 and 21 percent, respectively) [32]. Only 2 percent of posttraumatic cases but 11 percent of postoperative cases presented more than four weeks after trauma or surgery.

In a series from Florida, 10 patients with posttraumatic fungal endophthalmitis developed symptoms an average of 1.8 months after trauma (range one day to six months), and 18 patients with keratomycosis developed endophthalmitis four days to six months after the initial corneal ulcer [28].

OPHTHALMIC EXAMINATION — The ophthalmic examination reflects the endogenous or exogenous origin of infection:

●In endogenous endophthalmitis, the cornea is usually normal and there may be minimal inflammation in the anterior segment of the eye. Initially, the findings may be primarily in the posterior segment of the eye.

•In endogenous Candida endophthalmitis, chorioretinitis may be the only finding initially. The classic findings of chorioretinal involvement are focal, white, infiltrative, often mound-like lesions on the retina (picture 1 and picture 2). When vitreous extension occurs, a vitreal haze is present (picture 3); sometimes, fluffy white balls or "snowballs" or a "string of pearls" in the vitreous are noted (picture 4) [14,15,17]. The degree of vitreous inflammation is graded on a 0 to 4 scale. Inflammation may also extend to the aqueous, where the same scale of inflammation applies. With marked inflammation, the white blood cells in the aqueous may settle out in the bottom as a hypopyon (picture 5) or form an inflammatory membrane within the anterior chamber.

•In endogenous mold endophthalmitis, the initial infection may appear as a choroidal or subretinal mass or as white lesions in the posterior vitreous. There may be significant vitritis and typically there are clumps or "fluff balls" rather than a diffuse, homogenous vitritis. Inflammation may extend to the aqueous.

●In exogenous fungal endophthalmitis due to trauma, the initial surface injury may be apparent. The intraocular inflammation may have a clumped rather than diffuse appearance (picture 6). In postoperative fungal endophthalmitis, there may be white blood cells and "flare" (protein) in the anterior chamber and keratitic precipitates (clumps of white blood cells) on the corneal endothelium and anterior surface of the lens implant, if present [33]. A hypopyon may be present (picture 5). If extension to the vitreous has occurred, there may be vitritis. The view of the retina may be obscured by the intraocular inflammation. In cases of exogenous mold endophthalmitis that occur as an extension of corneal infection (keratomycosis), slit lamp examination will demonstrate the corneal infiltrate and also may reveal filamentous strands extending into the aqueous from the back of the involved cornea. There may be a hypopyon (picture 5) and, in some cases, there may be dense, white, clumped material in the anterior chamber.

DIAGNOSIS

Approach to diagnosis — Fungal endophthalmitis is suspected clinically based on the patient's presenting complaints, risk factors for exogenous (eg, penetrating eye trauma, keratomycosis) or endogenous fungal endophthalmitis (eg, intravenous drug use, recent indwelling central venous catheter, severe immunocompromise), and ophthalmic examination. (See 'Clinical manifestations' above.)

Ocular candidiasis is also detected by screening eye examinations in patients with candidemia. Two national societies in the United States have recently disagreed on the value of routine eye examinations in candidemic patients. The Infectious Diseases Society of America (IDSA) has published guidelines for years (most recently 2016) recommending a dilated retinal examination, preferably by an ophthalmologist, within the first week following the diagnosis of candidemia, regardless of eye symptoms [60]. The support for this approach is that studies which involve routine screening report a significant prevalence of ocular candidiasis (usually at least 10 percent) in candidemic patients, and over 80 percent of these patients have no eye symptoms (or are unable to relay them) at the time of diagnosis (see 'Epidemiology and risk factors' above). However, in 2021 the American Academy of Ophthalmology disagreed with these IDSA guidelines, stating that routine screening eye examinations in candidemia were a "low-value practice" and recommending that an ophthalmologic consultation should only be obtained in patients "with signs or symptoms of ocular infection regardless of Candida septicemia" [61].

Blood cultures should be obtained from all patients with suspected endogenous endophthalmitis. (See 'Blood cultures' below.)

To establish the diagnosis of fungal endophthalmitis, samples of the vitreous (and/or aqueous, if primarily involved) may be obtained by an ophthalmologist and submitted for stains and cultures. Such sampling is especially important for diagnosis in patients who present with acute or subacute intraocular inflammation of unknown etiology and who have eye findings consistent with fungal endophthalmitis but in whom blood cultures are negative. Obtaining intraocular samples may not be necessary for diagnosis in other cases of fungal endophthalmitis, such as in patients with documented candidemia and funduscopic findings typical of Candida chorioretinitis or in patients with keratomycosis of known fungal etiology whose infection has extended into the aqueous.

Any time the vitreous (or aqueous) is sampled, material should be sent for stains and cultures. In patients in whom the diagnosis of fungal endophthalmitis is suspected but not established, the vitreous (and aqueous, if involved) should be sampled. Vitreous samples are obtained either by needle aspiration or vitrectomy (figure 2). Vitrectomy has a higher yield than needle aspiration of the vitreous [1]. Aqueous samples are obtained by needle aspiration or are obtained intraoperatively in cases of keratomycosis requiring corneal transplantation. In keratomycosis cases requiring surgery, samples of cornea should be sent for fungal stain and culture as well as pathology. (See "Treatment of endophthalmitis due to molds", section on 'Vitrectomy and removal of foreign bodies'.)

In many cases of suspected but not yet established fungal endophthalmitis, intraocular injections of antifungal agents are given presumptively, immediately after initial samples of vitreous and/or aqueous are obtained. If the initial sample is a needle aspirate (a "tap"), a subsequent vitrectomy may be required to make a diagnosis, and cultures of the vitrectomy sample may still be positive despite initial intravitreal antibiotic injection [21].

Microbiology studies — Intraocular samples should be sent for fungal stain and culture as well as Gram stain and bacterial culture since it is sometimes not possible to differentiate fungal from bacterial infections based on clinical appearance alone. Fungal cultures should be specifically requested, since molds typically grow best on Sabouraud dextrose agar and at a lower incubation temperature (25°C) than used for bacterial cultures. Stains of intraocular fluids should also be requested. Gram stain identifies yeast but frequently cannot visualize the hyphae of molds. A fungal stain, such as calcofluor white, shows fluorescence of the cell wall of yeasts and molds and should always be requested.

If a vitrectomy has been performed, the vitreous washings should be filtered through a Millipore filter and the filter paper sterilely divided and incubated on various media. The ophthalmologist will also usually obtain a vitreous "biopsy" of undiluted vitreous at the start of the operation, and this material should be stained and cultured as well. Vitrectomy is discussed in greater detail separately. (See "Treatment of endogenous endophthalmitis and chorioretinitis due to Candida species", section on 'Intravitreal antifungal agents and vitrectomy'.)

Blood cultures — Blood cultures should be obtained from all patients with suspected endogenous endophthalmitis, as noted above. However, negative blood cultures do not exclude endogenous fungal endophthalmitis because the eye may have been seeded much earlier during transient fungemia. This is the typical situation in asymptomatic patients who present with Candida endophthalmitis days to weeks after removal of an indwelling central catheter. This is also often the case in intravenous drug use-associated endogenous fungal endophthalmitis; only 1 of 17 such patients in one series had positive blood cultures on presentation [21]. (See "Clinical manifestations and diagnosis of candidemia and invasive candidiasis in adults", section on 'Blood cultures'.)

Polymerase chain reaction — Molecular techniques, such as the polymerase chain reaction (PCR), have been used on intraocular fluids to diagnose mold endophthalmitis. The sensitivity and specificity are not yet known. In one study from India, routine stains and cultures were compared with PCR on intraocular fluid samples (48 vitreous, 2 aqueous) from 50 patients suspected of having postoperative fungal endophthalmitis and 25 uninfected control patients undergoing vitrectomy for another reason [62]. Four of the samples from patients suspected of having postoperative fungal endophthalmitis were positive for fungi by calcofluor white stain (three for hyphae and one for yeast), and, of these, two grew molds; PCR was positive in three cases that were positive for hyphae by calcofluor white stain. PCR was positive in an additional three cases that were negative by stain and culture. None of the control patients had a positive culture or PCR result.

SOCIETY GUIDELINE LINKS — Links to society and government-sponsored guidelines from selected countries and regions around the world are provided separately. (See "Society guideline links: Candidiasis".)

SUMMARY AND RECOMMENDATIONS

●The term "endophthalmitis" means infection within the eye involving the vitreous and/or aqueous (figure 1):

•Endogenous fungal endophthalmitis results from fungal seeding of the eye via the bloodstream. In most cases of endogenous endophthalmitis, the highly vascular choroid is seeded first.

•Exogenous fungal endophthalmitis means that the infection was introduced from the "outside," either by eye surgery, eye trauma, or extension of fungal infection of the cornea (termed "fungal keratitis" or "keratomycosis"). In cases following surgery involving the anterior segment (eg, cataract surgery) or following keratomycosis, the aqueous is typically infected first. (See 'Definitions' above.)

●In temperate climates, endophthalmitis is most commonly caused by Candida spp and is typically endogenous in origin (ie, in the setting of candidemia). In contrast, in tropical climates, mold endophthalmitis is more common than Candida endophthalmitis. Mold endophthalmitis is usually exogenous in origin. (See 'Epidemiology and risk factors' above.)

●The onset of symptoms in endogenous fungal endophthalmitis may occur subacutely, days to several weeks after fungemia. Initially, endogenous fungal endophthalmitis may be associated with minimal eye pain, and vision decrease may be subtle until the infection is advanced. In patients with risk factors for transient fungemia (eg, intravenous drug use, recent indwelling central venous catheter), endogenous fungal endophthalmitis should be considered in the differential diagnosis of patients presenting with posterior uveitis or panuveitis. (See 'Endogenous fungal endophthalmitis' above.)

●Early findings of endogenous Candida endophthalmitis include fluffy white lesions in the choroid and retina (chorioretinitis), often with overlying vitreous inflammation (picture 1 and picture 2). Marked vitritis characterizes advanced Candida endophthalmitis (picture 3) and the vitreous inflammation often has a clumped appearance, described as "fluff balls" or "string of pearls" (picture 4). Endogenous mold endophthalmitis often has a similar appearance, with discrete white infiltrates in the choroid and retina and dense infiltrates ("fluff balls") in the vitreous. The intraocular inflammation in exogenous mold endophthalmitis may also have a clumped appearance (picture 6). (See 'Ophthalmic examination' above.)

●We recommend screening eye examinations for patients with candidemia to allow early detection and treatment of ocular candidiasis. There is some controversy among professional societies about the role of screening [60,61], as noted above. (See "Management of candidemia and invasive candidiasis in adults", section on 'Clinical evaluation' and 'Approach to diagnosis' above.)

●Fungal endophthalmitis is suspected clinically based upon the patient's presenting complaints, risk factors for fungal endophthalmitis (eg, intravenous drug use, indwelling central venous catheter, severe immunocompromise, recent eye trauma or eye surgery) and the ophthalmic examination. Blood cultures should be obtained from all patients with suspected endogenous endophthalmitis. Negative blood cultures do not exclude a diagnosis of fungal endophthalmitis, since seeding of the eye may have occurred during transient fungemia. (See 'Approach to diagnosis' above.)

●To establish the diagnosis of fungal endophthalmitis, samples of the vitreous (and/or aqueous, if primarily involved) may be obtained by an ophthalmologist and submitted for stains and cultures. Such sampling is especially important for diagnosis in patients who present with acute or subacute intraocular inflammation of unknown etiology and who have eye findings consistent with fungal endophthalmitis but in whom blood cultures are negative. Obtaining intraocular samples may not be necessary for diagnosis in other cases of fungal endophthalmitis, such as in patients with documented candidemia and funduscopic findings typical of Candida chorioretinitis or in patients with keratomycosis of known fungal etiology whose infection has extended into the aqueous. (See 'Approach to diagnosis' above and 'Microbiology studies' above.)

ACKNOWLEDGMENT — The UpToDate editorial staff acknowledges Kieren A Marr, MD, who contributed to an earlier version of this topic review.