Elsevier

Survey of Ophthalmology

Volume 48, Issue 4, July–August 2003, Pages 403-423
Survey of Ophthalmology

Major review
Endogenous bacterial endophthalmitis: A 17-year prospective series and review of 267 reported cases

https://doi.org/10.1016/S0039-6257(03)00054-7Get rights and content

Abstract

Endogenous bacterial endophthalmitis is a rare but serious condition that occurs when bacteria cross the blood–ocular barrier and multiply within the eye. We provide an overview of endogenous bacterial endophthalmitis by reviewing 267 reported cases and integrating this with our experience of an additional 19 cases. The majority of patients with endogenous bacterial endophthalmitis are initially misdiagnosed and many have an underlying disease known to predispose to infection. This condition is often previously undiagnosed. Blood cultures are the most frequent means of establishing the diagnosis. The most common Gram positive organisms are Staphylococcus aureus, group B streptococci, Streptococcus pneumoniae,and Listeria monocytogenes.The most common Gram negative organisms are Klebsiella spp., Escherichia coli, Pseudomonas aeruginosa, and Neisseria meningitidis. Gram negative organisms are responsible for the majority of cases reported from East Asian hospitals, but Gram positive organisms are more common in North America and Europe. The visual outcome is poor with most cases leading to blindness in the affected eye. Many patients have extraocular foci of infection, with an associated mortality rate of 5%. The outcome of endogenous bacterial endophthalmitis has not improved in 55 years and clinicians need to have a high level of awareness of this commonly misdiagnosed condition.

Section snippets

St Thomas' case series

A 17-year prospective study from 1984 to 2001 was conducted in the departments of ophthalmology and microbiology in a London teaching hospital with a dedicated medical eye unit. All patients with a provisional diagnosis of EBE were admitted to the medical eye unit, and those with positive blood or intraocular culture were then prospectively identified. As Gram negative bacteria are rarely contaminants of blood or intraocular culture, their isolation was considered to be significant. Gram

St Thomas' case series

During the study period 19 patients (21 eyes) were diagnosed with EBE (Table 1). There were 12 men and 7 women with a mean age of 60 years (SD 17.0; range 7–77). Four patients have been reported previously.133 Twelve patients presented to an ophthalmologist with ocular symptoms and 7 to a general physician with manifestations of extraocular infection. All adult patients complained of blurred or altered vision at the onset of infection, and 12 had ocular pain or discomfort. Seventeen patients

Patient Characteristics

Although our literature review suggests that the incidence of EBE peaks at about 50 years of age, there is a wide distribution from neonates to patients in their tenth decade. The last major review by Greenwald et al in 198671 reported a lower mean age of 35 years (range 1 week to 85 years) with 24 of 72 patients aged less than 20 years. They also suggested that the right eye was twice as likely to be affected as the left and postulated that this is because of the more proximal and direct

Clinical Features

The classic features of EBE include ocular pain, blurred vision, swollen eyelids, injected and chemosed conjunctiva, anterior chamber inflammation and hypopyon, elevated IOP with associated corneal edema, a reduced or absent red reflex, and poor fundal view secondary to intraocular inflammation. In our case series all adult patients complained of blurred vision and two-thirds had ocular pain. Hypopyon and elevated IOP were inconstant findings in our patients, but when the IOP was raised it was

Incidence of diagnostic errors

Although some of the above clinical features may help identify patients with EBE, most are nonspecific and it remains a difficult diagnosis, especially given its low incidence. Greenwald's review71 found that 16% (11/67) of patients with EBE were initially misdiagnosed, and our literature review suggests a figure of 22%. Both are probably an underestimate as diagnostic errors may be under-reported. In our case series 63% of cases were initially misdiagnosed.

Adults

The differential diagnosis of EBE

Blood Culture

Blood culture is the most reliable way of establishing the diagnosis. In four large series of EBE,36., 102., 135., 191. blood cultures were more likely to be positive than vitreous. Review of all 267 cases shows that blood culture was the commonest means of confirming the diagnosis, and in our own case series almost three-quarters of blood cultures were positive.

Intraocular Cultures

Although useful, blood cultures alone cannot be relied upon to establish the diagnosis. In the absence of positive cultures from

Causative Organisms

Although our literature review suggests that EBE is more commonly caused by Gram negative bacteria, there was considerable regional variation. In a large literature review, Wong et al191 compared cases of EBE reported from East Asian hospitals with those from the world literature. In East Asia, EBE was overwhelmingly caused by Gram negative organisms, particularly Klebsiella spp. Cases reported from outside that region were more likely to be caused by Gram positive organisms. Unusual organisms

Role in endogenous endophthalmitis

Just over half of the patients in our literature review were treated with intravitreal antibiotics. This reflects debate in the literature regarding their role in EBE. Many topical, subconjunctival, and systemic antibiotics do not reach therapeutic levels within the vitreous12 and this may explain why patients can develop EBE while on appropriate systemic antibiotics, despite therapeutic blood levels.169

Wong et al performed univariable analysis of visual outcome in a large literature review of

Prognosis

The visual outcome of EBE has not improved in 55 years. Review of the literature from 1976–1985 showed that 41% of patients had count fingers vision or better, 26% were blind, and 29% required evisceration or enucleation.71 Similar figures were reported over the preceding 30 years.162 Review of the literature since 1986 also indicates a poor outcome, with equivalent figures of 32%, 44%, and 25% (Fig. 1).

The studies that investigated prognostic factors in EBE71 were retrospective, and although

Pathogenesis

EBE occurs when bacteria cross the blood–ocular barrier, resist host defenses, and multiply within the eye. The blood–ocular barrier may prevent many organisms reaching the internal ocular spaces as relatively few bacteremic patients develop EBE. In one prospective study of 202 patients with bacteremia, none developed EBE although 12 developed minor ocular lesions such as cotton-wool spots and microhemorrhages.17 Another prospective study of 101 patients with septicemia131 reported that 24

Pathology

Animal experiments suggest that tissue damage results from a combination of direct invasion of ocular tissue by bacteria, the release of bacterial toxins, and the host inflammatory response. Callegan et al21 studied the effect of intravitreal injections of rabbit eyes with live S. aureus, Bacillus cereus, and Enterococcus faecalis,as well as bacterial cell walls and exotoxins. B.cereusdirectly invaded the retina and cornea, and both S.aureusand B.cereusreduced the electroretinogram response,

Guidelines for reporting EBE

Large randomized studies of EBE are unlikely to be forthcoming in the near future. The quality of case reports is therefore important, as these form the principal evidence base. Some case reports provide inadequate details of the presentation, management, and outcome of EBE. We therefore advocate that the following minimum information is included when reporting cases of EBE: age; sex; affected side(s); ocular clinical features including classification using Greenwald's system;71 systemic

Conclusions

EBE has a poor visual prognosis that has not improved in over half a century. There are several possible reasons for this. First, EBE is very uncommon yet it may mimic several common ophthalmic conditions. This may explain why many patients are initially misdiagnosed. EBE can be rapidly progressive and delayed treatment may result in a poor outcome. Second, clinicians may fail to appreciate the overlap of ocular and extraocular disease. Third, there have been no large prospective studies of EBE

References (194)

  • M.J Greenwald et al.

    Metastatic bacterial endophthalmitis: a contemporary reappraisal

    Surv Ophthalmol

    (1986)
  • D.P Han et al.

    Endophthalmitis from mycobacterium bovis-Bacille Calmette-Guérin after intraveicular Bacille Calmette-Guérin injections for bladder carcinoma

    Am J Ophthalmol

    (1999)
  • D.P Han et al.

    Spectrum and susceptibilities of microbiologic isolates in the Endophthalmitis Vitrectomy Study

    Am J Ophthalmol

    (1996)
  • S.A Harrison et al.

    Endogenous endophthalmitis caused by Streptococcus mitis

    Am J Ophthalmol

    (1997)
  • I.J Hassan et al.

    Endophthalmitis at the Bristol Eye Hospital: an 11-year review of 47 patients

    J Hosp Infect

    (1992)
  • R.H Havunjian et al.

    Bilateral endogenous endophthalmitis in a patient with diabetes and renal papillary necrosis

    Am J Ophthalmol

    (1991)
  • A Hornblass et al.

    Endogenous orbital cellulitis and endogenous endophthalmitis in subacute bacterial endocarditis

    Am J Ophthalmol

    (1989)
  • M.L Jain et al.

    Metastatic endophthalmitis in a patient with major burns: a rare complication

    Burns

    (1995)
  • L.M Jampol et al.

    Intraocular nocardiosis

    Am J Ophthalmol

    (1973)
  • C.M Knox et al.

    Identification of bacterial pathogens in patients with endophthalmitis by 16S ribosomal DNA typing

    Am J Ophthalmol

    (1999)
  • M.S Kresloff et al.

    Endophthalmitis

    Surv Ophthalmol

    (1998)
  • J.W Kronish et al.

    Orbital infections in patients with human immunodeficiency virus infection

    Ophthalmology

    (1996)
  • F Abousaesha et al.

    Endophthalmitis as a presentation of meningococcal septicemia

    Ir J Med Sci

    (1993)
  • S Al-Amro et al.

    Endogenous bacterial endophthalmitis in sickle cell anemia

    Ann Ophthalmol

    (1997)
  • M.F Al Faran

    Brucella melitensis endogenous endophthalmitis

    Ophthalmologica

    (1990)
  • S.A Al Hazzaa et al.

    Pink hypopyon: a sign of Serratia marcescens endophthalmitis

    Br J Ophthalmol

    (1992)
  • M Algan et al.

    Listeria monocytogenes endophthalmitis in a renal-transplant patient receiving ciclosporin

    Ophthalmologica

    (1990)
  • R Alvarez et al.

    Haematogenous Serratia marcescens endophthalmitis in an HIV-infected intravenous drug addict

    Infection

    (1990)
  • N Anand et al.

    E. coli endophthalmitis

    Eye

    (1996)
  • J.G Arroyo et al.

    Endogenous endophthalmitis initially misdiagnosed as anterior uveitis

    Ann Ophthalmol

    (2000)
  • A.A Assaf et al.

    Clostridium-perfringens endophthalmitis and orbital cellulitis

    Saudi Med J

    (1992)
  • S.B Auerbach et al.

    Meningococcal endophthalmitis without concomitant septicemia or meningitis

    Pediatr Infect Dis J

    (1989)
  • P Berkey et al.

    Group G streptococci as a cause of bacterial endophthalmitis

    Arch Ophthalmol

    (1988)
  • H.L.C Beynon et al.

    Meningococcal endophthalmitis and pericarditis

    J Royal Soc Med

    (1990)
  • H.M Boisjoly et al.

    Metastatic Pseudomonas endophthalmitis following bronchoscopy

    Can J Ophthalmol

    (1987)
  • E Bouza et al.

    A prospective search for ocular lesions in hospitalized patients with significant bacteremia

    Clin Infect Dis

    (2000)
  • G.R Brian et al.

    Listeria monocytogenes endophthalmitis: a case report

    Aust NZ J Ophthalmol

    (1988)
  • R.D Brod et al.

    Endophthalmitis: current approaches to diagnosis and treatment

    Curr Opin Infect Dis

    (1993)
  • T.D Buglass et al.

    Fatal case of group B streptococcal endogenous endophthalmitis

    Can J Ophthalmol

    (1995)
  • M.C Callegan et al.

    Pathogenesis of gram-positive bacterial endophthalmitis

    Infect Immunity

    (1999)
  • P.A Campochiaro et al.

    Aminoglycoside toxicity—a survey of retinal specialists. Implications for ocular use

    Arch Ophthalmol

    (1991)
  • P.A Campochiaro et al.

    Toxicity of intravitreous ceftazidime in primate retina

    Arch Ophthalmol

    (1992)
  • P.A Campochiaro et al.

    Aminoglycoside toxicity in the treatment of endophthalmitis. The Aminoglycoside Toxicity Study Group

    Arch Ophthalmol

    (1994)
  • A.J Cannistra et al.

    Sudden visual loss associated with clostridial bacteraemia

    Br J Ophthalmol

    (1988)
  • J.J Caravalho et al.

    Endogenous endophthalmitis due to Salmonella arizonae and Hafnia alvei

    Southern Med J

    (1990)
  • S.M Carden et al.

    Kingella kingae endophthalmitis in an infant

    Aust NZ J Ophthalmol

    (1991)
  • C Casanova et al.

    Klebsiella pneumoniae liver abscess associated septic endophthalmitis

    Arch Intern Med

    (1989)
  • F.Y Chang et al.

    A clinical study of Klebsiella liver abscess

    J Formosan Med Assoc

    (1988)
  • S.P Chee et al.

    Endogenous Klebsiella endophthalmitis—a case series

    Ann Acad Med Singapore

    (1995)
  • D.L Cheng et al.

    Septic metastatic lesions of pyogenic liver abscess. Their association with Klebsiella pneumoniae bacteremia in diabetic patients

    Arch Intern Med

    (1991)
  • Cited by (460)

    • MRSA panophthalmitis in a brittle diabetic

      2023, Journal of Ophthalmic Inflammation and Infection
    View all citing articles on Scopus

    The authors wish to thank Andrew Hutchings, London School of Hygiene and Tropical Medicine, for statistical support. Dr Jackson was supported by a research grant from the Lady Anne Allerton Fund and a personal training award from the Special Trustees of St Thomas' Hospital. The authors reported no proprietary or financial interest in any product mentioned or concept discussed in this article.

    View full text