Article Text

Download PDFPDF

Microbiological isolates and antibiotic sensitivities in culture-proven endophthalmitis: a 15-year review
  1. Thomas P Moloney,
  2. Joseph Park
  1. Department of Ophthalmology, The Royal Brisbane and Women's Hospital, Herston, Queensland, Australia
  1. Correspondence to Dr Thomas Moloney, Department of Ophthalmology, Royal Brisbane and Women's Hospital, Herston, QLD 4006, Australia; thomas.moloney{at}


Background To describe the microbiological spectrum and antibiotic sensitivities of organisms causing culture-proven endophthalmitis in Queensland, Australia, and to compare results with similar studies from other parts of Australia and other countries.

Methods A retrospective, multicentre, non-comparative, consecutive case series.

Public hospital microbiology records from culture-positive endophthalmitis cases were reviewed over 15 years from June 1998 to June 2013. Outcome measures were type of endophthalmitis, vitreous isolates cultured and antibiotic sensitivities.

Results 205 cases of culture-proven endophthalmitis were identified with a total of 229 isolates cultured. The most common organisms isolated were Staphylococcus epidermidis in 23.1%, Coagulase-negative Staphylococcus in 12.7%, Streptococcus viridans group in 10.0%, Candida species in 6.1%, fungal mold species in 5.7%. For gram-positive organisms, sensitivities were vancomycin 100%, cephazolin 79% and penicillin 47%. For gram-negative organisms, sensitivities were ceftazidime 100%, amikacin 100%, ciprofloxacin 100% and gentamicin 95.5%. For fungal isolates, sensitivities were voriconazole 93%, ketoconazole 89%, caspofungin 70% and amphotericin B 58%.

Conclusions The microbiological spectrum and antibiotic sensitivities of endophthalmitis cases in Queensland, Australia, is similar to the spectrum of organisms causing endophthalmitis in other parts of Australia, North America and Europe. Empirical intravitreal vancomycin, ceftazidime and voriconazole are the most appropriate empirical antibiotics for suspected infective endophthalmitis.

  • Infection
View Full Text

Statistics from


Infectious endophthalmitis is a potentially devastating, sight-threatening intraocular infection. It can be characterised as exogenous if caused by surgery, trauma or spread of infection from another ocular/orbital site and endogenous if caused by haematogenous spread of infection to the eye.1 Treatment usually involves aspiration and subsequent culture of ocular fluids, administration of empirical intravitreal antibiotics, followed by more specific antibiotic therapy once microbiological culture and sensitivities are available.1

The purpose of the current study was to describe the microbiological spectrum and antibiotic sensitivities of organisms causing endophthalmitis in public hospitals across the state of Queensland, Australia, over 15 years and to compare results with the endophthalmitis microbiological spectrum from other parts of Australia and other countries.

Materials and methods

Queensland state-wide public hospital microbiology records were searched for all patients with endophthalmitis and vitreous samples with positive microbiological culture over a 15-year period from June 1998 to June 2013. The medical records for each identified case were reviewed for the clinical setting of the endophthalmitis, microbiology results and the antibiotic sensitivities of the cultured organism or organisms. This review was conducted in accordance with the guidelines set forth by the Declaration of Helsinki and was exempt from institutional review board approval.

Endophthalmitis cases were categorised into the following clinical types—postoperative, postintravitreal injection, post-traumatic, endogenous, bleb-associated and keratitis-associated. Postoperative endophthalmitis cases were then categorised into acute-onset postcataract procedure endophthalmitis (6 weeks or less postprocedure), delayed onset postcataract procedure endophthalmitis (6 weeks or more postprocedure), acute-onset postvitrectomy endophthalmitis and acute-onset postpenetrating keratoplasty endophthalmitis. Postintravitreal injection endophthalmitis cases were also categorised into postbevacizumab/ranibizumab injection endophthalmitis or posttriamcinolone injection endophthalmitis. Post-traumatic endophthalmitis cases were categorised into those cases that involved an intraocular foreign body and those that did not.

Antibiotic sensitivity testing of the isolates was performed using either the VITEK (Automatic Microbial System; Biomerieux Vitek Inc, Hazelwood, Missouri, USA) or the E test (A.B. Biodisk, Remel Inc, Lenexa, Kansas, USA).

Overall antibiotic sensitivities for gram-positive organisms, gram-negative organisms and fungal organisms were calculated by finding the percentage of those organisms that were sensitive to each particular antibiotic.


Over the 15-year period of the study, 205 cases of culture-proven endophthalmitis were identified (table 1). Among these 205 cases, there were 19 cases of polymicrobial endophthalmitis, thus a total of 229 vitreous isolates were cultured. Overall, there were 163 gram-positive organisms isolated (71%), 30 gram-negative organisms isolated (13.1%) and 36 cases cultured fungi (15.7%). The most common organisms isolated were Staphylococcus epidermidis in 23.1% (53/229), Coagulase-negative Staphylococcus in 12.7% (29/229), Streptococcus viridans group in 10.0% (23/229), Candida species in 6.1% (14/229), fungal mold species in 5.7% (13/229), Staphylococcus aureus in 5.24% (12/229), Bacillus species in 5.24% (12/229), Streptococcus pneumonia in 4.8% (11/229) and Pseudomonas aeruginosa in 4.37% (10/229) (table 1). There was a single isolate of methicillin-resistant S. aureus from an acute-onset postcataract procedure endophthalmitis case (0.44%).

Table 1

Endophthalmitis vitreous isolates

The antibiotic sensitivities of the bacterial isolates cultured are summarised in table 2. Sensitivities among the 163 gram-positive organisms isolated were vancomycin 100%, cephazolin 79% and penicillin 47%. Sensitivities among the 30 gram-negative organisms isolated were ceftazidime 100%, amikacin 100%, ciprofloxacin 100% and gentamicin 95.5%. The antibiotic sensitivities of the fungal isolates cultured are summarised in table 3. Among these 36 fungal isolates, overall sensitivities were voriconazole 93%, ketoconazole 89%, caspofungin 70%, natamycin 66%, amphotericin B 58%, 5-flucytosine 50%, posaconazole 46% and itraconazole 42%.

Table 2

Sensitivities of bacterial endophthalmitis vitreous isolates

Table 3

Sensitivities of fungal endophthalmitis vitreous isolates

The clinical settings of the culture-proven endophthalmitis cases are shown in table 4. Over the 15-year period, the type of endophthalmitis was postoperative in 38.4%, endogenous in 21.8%, post-traumatic in 17.0%, postintravitreal injection in 10.0%, keratitis-associated in 9.6% and bleb-associated in 3.1%. In terms of the postoperative endophthalmitis cases, 89.1% were acute-onset postcataract procedure, 7.23% were delayed-onset postcataract procedure, 2.4% were acute-onset postvitrectomy and 1.2% were acute-onset postpenetrating keratoplasty. The most common organisms for each type of endophthalmitis clinical setting are shown in table 5.

Table 4

Clinical setting of culture-positive endophthalmitis cases

Table 5

Endophthalmitis vitreous isolates by clinical setting

Polymicrobial endophthalmitis was most common in post-traumatic endophthalmitis (7/19 cases) followed by endophthalmitis postintravitreal injection (5/19 cases) and postoperative endophthalmitis (5/19 cases) (table 4).


Knowledge of the microbiology of the different organisms causing endophthalmitis is central to the successful treatment of the condition.1 Many previous studies from around the world have described series of endophthalmitis cases where the microbiological spectrum of causative organisms has been variable between the different types of endophthalmitis by clinical setting.2 ,3 By recognising the differences in the frequencies of organisms causing the different types of endophthalmitis, empirical antibiotic and antifungal therapy can then be confidently implemented to cover common organisms.

The current study reviews the microbiological spectrum of organisms causing the different types of endophthalmitis cases across public hospitals in the state of Queensland, Australia, over 15 years and represents the largest series of its kind from Australia. In terms of clinical setting, postoperative endophthalmitis was most common with 38.4% of cases, followed by endogenous endophthalmitis (21.8%), post-traumatic endophthalmitis (17.0%) and postintravitreal injection endophthalmitis (10.0%) (table 4). Overall, the results of the organisms cultured for each type of endophthalmitis in our series show many similarities to results from other reports from Australia, North America and Europe.

Results from the 83 postoperative endophthalmitis cases in our series showed 85.2% were caused by gram-positive organisms, 13.6% were caused by gram-negative organisms and 1.1% were caused by fungal species (table 5). In acute-onset postcataract procedure cases, gram-positive organisms were isolated in 83.1%, gram-negative organisms in 15.6% and fungal organisms in 1.3%. In delayed-onset postcataract procedure endophthalmitis cases, S. epidermidis (42.86%) followed by Propionibacterium acnes (28.57%) were most common. Overall, these results were similar to those from the Endophthalmitis Population Study of Western Australia (EPSWA) where 131 organisms were isolated from 114 culture-positive cases with 86.3% caused by gram-positive organisms, 12.2% gram-negative organisms and 1.5% caused by fungal species.4 These Australian results differ slightly from the Endophthalmitis Vitrectomy Study (EVS) and from the more recent article by Benz et al, both from the USA, which reported higher culture rates of gram-positive isolates (94.2% and 88.6%) and no fungal isolates in each series.2 ,3 Our results also showed a higher rate of polymicrobial postoperative endophthalmitis (4.6%) again compared with the EVS and Benz et al.2 ,3

In terms of postintravitreal injection endophthalmitis, the 23 isolates cultured from the 18 cases in our series were all gram-positive organisms with 5 cases being polymicrobial (28%) (table 5). These results are consistent with other reports from Australia,5 North America6 ,7 and Europe8 ,9 that also reported high gram-positive rates of culture. In our 18 cases, S. viridans group organisms were most common (34.78%) and polymicrobial cases often grew two different organisms from this group. In a major review from North America, McCannel previously reported a threefold increase in the rate of S. viridans group culture in postintravitreal injection endophthalmitis cases compared with postoperative endophthalmitis cases, and our results also demonstrate this.6

As suggested in previous reports,10 our series confirmed a wider spectrum of causative organisms in post-traumatic endophthalmitis cases compared with postoperative endophthalmitis  or postintravitreal injection endophthalmitis. In our series, gram-positive organisms were cultured in 69.2% of cases, gram-negative organisms were cultured in 17.9% of cases and there was a higher rate of fungal cultures (10.25%) (table 5). There was also a high rate of polymicrobial culture (24%) and a high rate of Bacillus species culture (17.2%) consistent with other Australian and North American reports.10–12

Endogenous endophthalmitis was the second most common type of endophthalmitis in our series comprising 21.8% of cases compared with 8% of cases in Benz et al.3 Among our 50 cases, 58% cultured fungal species, 32% cultured a gram-positive organism and 10.0% cultured a gram-negative organism (table 5). Candida species were most common organisms cultured with 28.0% of cases, but there were also a large number of cases caused by molds (18.0%). Overall, the lower rate of bacterial culture in endogenous endophthalmitis compared with fungal culture in endogenous endophthalmitis was consistent with other smaller series from Australian centres 13–15 and North American centres.16–18

Keratitis-associated endophthalmitis with culture-positive vitreous samples made up 8.7% of cases with 77.2% of these culturing gram-positive organisms. Bleb-associated endophthalmitis made up 3.1% of cases with gram-positive organisms causing 57.1% of cases and gram-negative organisms causing 42.8% of cases (table 5).

In terms of overall antibiotic sensitivities, one notable limitation of our study was the highly variable antibiotic sensitivity testing performed on the cultured organisms. As shown in tables 2 and 3, not all organisms were tested against all antibiotics, thus reducing the validity of the calculated overall antibiotic sensitivities. This was especially relevant to ceftazidime testing where only 33% of gram-negative organisms were tested for ceftazidime sensitivity. Despite this, overall bacterial antibiotic sensitivities showed 100% sensitivity to vancomycin among the gram-positive organisms tested and 100% sensitivity to ceftazidime, ciprofloxacin and amikacin among the gram-negative organisms tested (table 2), and these results are similar to reports from North America 2 ,3 ,18 and Europe.8 ,9 ,17 In terms of antifungal sensitivities, 93% of the fungal species tested were sensitive to voriconazole. Importantly, only 58% of fungal species tested were sensitive to the common empirical antifungal agent amphotericin B, with many cultured molds displaying resistance (table 3).

In summary, the results of our series show that the microbiological spectrum of endophthalmitis in Australia is generally similar to the microbiological spectrum of endophthalmitis in North America and Europe. Differences include less gram-positive organisms cultured in postoperative endophthalmitis cases and an increased overall proportion of fungal endophthalmitis when compared with other major series like Benz et al and Schimel et al18 Important similarities include a predominance of S. viridans group organisms in postintravitreal injection endophthalmitis,18 a high rate of Bacillus species culture and polymicrobial culture in post-traumatic endophthalmitis and a high rate of fungal cultures in endogenous endophthalmitis.

In terms of antibiotic sensitivities and empirical antibiotics, our series confirms that there is no single antibiotic that can adequately cover all possible organisms causing endophthalmitis. Current recommendations for empirical treatment with intravitreal vancomycin and ceftazidime for suspected bacterial endophthalmitis would appear to be appropriate for an Australian population, although further studies are needed to confirm the sensitivity of gram-negative organisms to ceftazidime. Although amikacin is not usually used as a first-line agent in Australia due to potential retinotoxicity,19 it does appear to be a suitable alternative to ceftazidime for coverage of gram-negative organisms. If fungal endophthalmitis is suspected, we would recommend the empirical use of intravitreal voriconazole to cover amphotericin B-resistant organisms. However, this would be dependent upon institution-specific restrictions on the use of voriconazole as a first-line agent.


View Abstract


  • Contributors JPF conceived the study and drafted the manuscript. TPM identified the patients, reviewed the medical records and microbiological records and drafted the manuscript.

  • Competing interests None.

  • Ethics approval Human Research Ethics Committee.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.