Infectious endophthalmitis is a potentially blinding disorder due to contamination of the ocular structures after exogenous trauma, including surgical and intravitreal therapies, or hematogenous spread from other infected tissue sites. Early clearance of the affecting organism is required to minimize tissue destruction, whereas identification of the offending organism and its sensitivity patterns allow a tailored approach to those that fail to respond to empiric antimicrobial treatment. However, identifying the pathogenic organism is not always possible with rates of positive culture ranging between 48% and 75% using traditional gram stain and culture methods.1–3 These findings are primarily based on the results of a vitreous tap or biopsy, which can be difficult to obtain in the acute setting because of the possibility of a dry tap, whereas delays in access to theatre hinder a timely diagnostic and therapeutic vitrectomy. Although the aqueous humor is much easier to sample, the rates of positive culture are significantly lower with only 27% of aqueous specimens culture positive in the Endophthalmitis Vitrectomy Study (EVS).4
We have previously reported on the clinical outcomes of patients that presented to a major tertiary referral centre in—with endophthalmitis.5 The aim of this follow-up study was to identify the factors associated with a positive intraocular culture sample based on clinical presentation and multimodal intraocular sampling.
This study received ethics approval from the review committee and adhered to the tenets of the Declaration of Helsinki. This was a retrospective review of all subjects presenting to a tertiary ophthalmic referral center with presumed endophthalmitis (in one or both eyes), between 1st January 2006 and 31st December 2018. Cases were identified through the admission records, theatre logbooks, and a computerized search of both aqueous and vitreous samples taken during this time period.
The clinical case notes and laboratory findings of all subjects were recorded onto a standard proforma. This included subject demographics, duration of symptoms, etiology (ie, post-injection, perioperative, among others), examination findings, laboratory results and management. Pertinent examination findings included Snellen visual acuity which was converted to LogMAR for the purpose of analysis. Intraocular inflammation was graded using the Standardization of Uveitis Nomenclature (SUN) and the presence/absence of a red reflex was also recorded.6,7
Intraocular Sampling and Treatment
The timing and mode of intraocular sampling methods were recorded. An aqueous sample was obtained via an anterior chamber (AC) paracentesis (25G-30G). A vitreous tap was performed in the procedure room using a 20 to 25G needle located 3.5 to 4 mm behind the corneal limbus, and a vitreous biopsy using a 23 to 25G vitrector at the time of vitrectomy surgery. A vitreous tap (and injection of intravitreal antibiotics) was performed if there are expected delays in getting the patient to theater to perform an immediate (primary) vitrectomy. Patients who received intravitreal antibiotics and then subsequently proceeded to surgery were labeled as having a secondary vitrectomy. All samples were processed by the microbiology laboratory at ADHB where light microscopy examination and microbial culture were performed. Routine culture media included sheep's blood and GC saponin agar, brain heart infusion (BHI), BHI broth with gentamicin, and sabouraud dextrose agar (fungi). Additional media testing was performed if requested by the managing team. A positive growth was reported if the same organism was observed in more than one medium or a confluence of growth on ≥1 solid media.
All patients received intravitreal ceftazidime 2.25% and vancomycin 1% after an intraocular sample was taken (aqueous or vitreous) and/or at the conclusion of vitrectomy surgery.
Data were entered into an Excel spreadsheet (Microsoft Corp, Redmond, WA) and analyzed in STATA version 15 (StataCorp 2017, College Station, TX). Continuous data are presented as median [interquartile range (IQR)] and categorical data as n (%). Logistic regression analysis was used to model predictors of positive culture result, examining role of poor best corrected visual acuity (BCVA) at presentation, AC cellular activity and absence of red reflex. All tests were 2-tailed and a P value of <0.05 was considered significant.
259 subjects were included for analysis. Subject demographics are described in Table 1 and causes of endophthalmitis are reported in Table 2. Mean visual acuity was hand movements (IQR 20/400 – HM) with a hypopyon present in 134 eyes (51.7%)
135 cases were culture-positive (52.1%), with 104 gram-positive (77.0%), 23 gram-negative (17.0%), and 8 fungal (5.9%). Culture results are listed in Table 3. Staphylococcus species were the most common in 68 cases (50.4%) followed by Streptococcus species in 28 cases (20.7%). The culture positive rate was 48.8% for cataract surgery, 41.7% for intravitreal injections, 46.2% for endogenous endophthalmitis, 59.1% for trauma, 42.9% for vitreoretinal surgery, 57.1% for glaucoma surgery, and 81.8% of corneal infections (P = 0.094).
Aqueous tap was performed in 112 eyes and positive in 36 (32.1%). Aqueous tap was positive in 36.4% of cataract surgery, 18.5% of intravitreal injections, 16.7% of endogenous, 40.0% of trauma, 37.5% of vitreoretinal surgery, 35.7% of glaucoma surgery, and 83.3% of corneal infections. Poor BCVA at presentation, AC cellular activity, presence of hypopyon, and absence of red reflex were not associated with likelihood of a positive aqueous tap. There was no association between use of topical antibiotics at presentation and the likelihood of a positive aqueous tap (OR 2.00, P = 0.112). Of the 135 culture-positive cases, 6 eyes (4.4%) had a positive aqueous tap in the setting of a negative vitreous tap, and 4 eyes (3.0%) had a positive aqueous tap in the setting of a negative vitrectomy. If we consider the vitreous tap as the criterion standard, the positive predictive value for aqueous tap was 83%.
Vitreous tap was attempted in 144 eyes but was dry in 22 (15.3%). In those with a successful tap, 56 (45.9%) were culture positive. Vitreous tap was culture positive in 47.2% of cataract surgery, 31.4% of intravitreal injections, 37.5% of endogenous, 50.0% of trauma, 50.0% of vitreoretinal surgery, 70.0% of glaucoma surgery, and 83.3% of corneal infections. On univariate analysis the after were associated with increased likelihood of positive vitreous tap: poor BCVA at presentation [odds ratio (OR) 2.175, P = 0.009]; AC cellular activity (OR 2.405, P = 0.009); and absence of red reflex (OR 3.875, P = 0.001). On multivariate analysis, only absence of red reflex remained significantly associated with positive vitreous tap (OR 2.808, P = 0.023).
Vitrectomy was performed in 169 eyes (primary in 77, secondary in 92). Median time to secondary vitrectomy was 24 hours (IQR 17–72). A specimen was sent for culture in 149 eyes and was positive in 70 (47.0%). 47.7% of primary vitrectomy specimens were culture positive and 45.9% of secondary vitrectomy specimens (P = 0.826). Specimens were sent for 38 subjects for both vitreous tap and secondary vitrectomy. Of these, 13 (34.2%) were culture negative for both, 11 (28.9%) had positive vitreous tap only, 5 (13.2%) had a positive vitrectomy culture despite negative vitreous tap, and 9 (23.7%) were positive for both vitreous tap and vitrectomy culture. For those who had secondary vitrectomy, median time from vitreous tap and inject to vitrectomy was 24.5 hours (IQR 19–72). The rate of positive vitrectomy culture was 55.6% for those receiving vitrectomy 0 to 12 hours after tap and inject, 62.5% for more than 12 to 24 hours, 16.7% for more than 24 to 48 hours, and 37.5% for >48 hours. On logistic regression analysis, only absent red reflex was associated with an increased risk of positive vitrectomy culture (OR 2.935, P = 0.006). There was no association with poor BCVA at presentation or severity of AC activity.
Median time to injection of intravitreal antibiotics was 4 hours (IQR 1–4). Time to injection was shorter if vitreous tap and inject was the primary procedure (median 3 hours IQR 1.6–5.5) compared with if vitrectomy was the primary procedure (median 4.5 hours IQR 2.9–9.4) (P = 0.002).
Factors associated with a positive culture from any of the intraocular specimens sent are reported in Table 4. On multivariate analysis, the following were associated with increased likelihood of positive culture: aqueous tap (OR 2.061, P = 0.020); vitrectomy (OR 2.864, P = 0.001); and absent red reflex (OR 2.732, P = 0.001).
In our study of 259 eyes with a clinical diagnosis of endophthalmitis, the rate of culture positive samples was 50%. The culture yield through vitreous sampling was 47% which is less than an earlier publication from our institution (63%) and that reported in the EVS (69%).1,5 However, the rates of positive aqueous culture were similar to the EVS and comparable to more a more recent publication by de Liano et al where 32% were culture-positive.4,8 The majority of cases were attributed to recent intraocular surgery; however, 23% were post-injection-related which is an increase from our earlier publication (<7%) in keeping with international trends given the rise in number of intraocular injections performed worldwide.9,10 With the results of a vitreous tap set as the criterion standard, we found that a positive aqueous tap (OR 2.1), vitrectomy/vitreous biopsy (OR 2.8), and absent red reflex were associated with a positive vitreous culture on multivariate analysis.
Although the diagnosis of endophthalmitis is made clinically, identification of the responsible organism through culture methods not only confirms the diagnosis, but allows sensitivity analysis to be performed. This is most useful for patients who fail to improve after initial broad-spectrum intravitreal antibiotics due to atypical or resistant organisms. Furthermore, microbial identification also assists with identifying commonality and possible causes during an outbreak of cases, and guides treatment protocols at a time when antimicrobial resistance is on the rise.11 A 2018 publication of bacterial sensitivities in the Auckland region found that the majority of Gram-positive bacteria, which constituted 70% of culture-positive cases, were sensitive to Vancomycin in vitro.12 Although the vast majority of Gram-negative bacteria (including P aeruginosa) were sensitive to Ceftazidime, up to 30% of selected isolates of E coli and K pneumoniae were resistant to the agent.13 Although the concentration of intravitreal antibiotics is typically 50 times higher than the mean inhibitory concentration for most bacteria, endophthalmitis caused by antibiotic resistant organisms generally have poorer visual outcomes and thus identification of the responsible organism is important.14,15
Identifying the causative organism is dependent on a number of factors including the severity of intraocular inflammation, sampling techniques, and microbiological analysis methods. The severity of intraocular inflammation can be graded using clinical signs and ancillary tests.6,7,16 Despite its simplicity, we found that vitritis sufficient to obscure the fundal red reflex was strongly associated with positive microbial culture on both uni- and multivariate analysis. Although other clinical factors such as presenting visual acuity and duration of symptoms have been shown to correlate with positive culture results, we did not observe these associations in our study.17
A vitreous sample has long been considered the criterion standard for identifying the responsible pathogen in endophthalmitis, either via vitreous tap or biopsy.4 Despite being more accessible, the diagnostic yield from an aqueous sample is lower, with sensitivities of 38% reported in the literature.8 This is attributed to a lower obtainable sample volume, rapid regeneration time, and a less favorable medium for microbial proliferation compared with the vitreous. From our study, the aqueous sample has a diagnostic in 8% of all positive cases when the vitreous tap or vitreous biopsy was culture-negative. The positive predictive value for aqueous tap in our study was 83%, with similar values reported in other studies (as high as 95%) of post-surgical endophthalmitis, where the majority involved anterior chamber procedures like cataract surgery or corneal infections, as was also noted in our study.8 In contrast, a vitreous tap can be technically challenging in the acute setting with the possibility of obtaining a dry tap (15% in our study) if the clinician is unable to access a pocket of liquefied vitreous, due to the patient's anatomy (nominal amount of liquefied vitreous present) or poor surgical technique. A vitreous biopsy overcomes this problem and provides a larger diagnostic specimen. Since 2003, we have used the technique described by Russell and Polkinghorne where a sterile tube is used in the aspiration line of the vitrector, which avoids the risks associated with excessive manual aspiration by the assistant surgeon when an undiluted sample is being collected.18 To avoid the risk of hypotony in this setting, Mura et al19 described their technique of performing complete vitrectomy under an air infusion. Other authors have advocated culturing the contents of the vitreous cassette after a complete vitrectomy where the causative organism was cultured in 76% to 43% in those who underwent a core biopsy alone.20 Furthermore, a complete vitrectomy allows the collection of cortical vitreous to provide a much higher cellular count than biopsy of the core alone.21 Although these earlier studies report on the importance of collecting an undiluted sample, Chiquet et al found no difference in the diagnostic yield between undiluted and diluted samples where both culture and polymerase chain reaction (PCR) amplification of microbial DNA segment techniques were used.22
The use of molecular methods such as PCR has been shown to increase the number of samples with a microbiological diagnosis and reduce the wait time for laboratory confirmation. PCR amplification of bacterial (16S rDNA) or fungal (18S rDNA) was historically used for research purposes. However, the diagnostic yield from aqueous humor and vitrectomy samples with traditional culture methods has been demonstrated to increase from 50% to 72% when combined with PCR testing.23,24 Furthermore, PCR techniques allowed a microbiological diagnosis in 25% of cases that were reported as culture-negative.24 Although bacterial and fungal PCR is available at our hospital laboratory, it is not routinely performed and reported on for intraocular samples.
Although the rates of positive culture were higher after a vitreous sample, the yield from a vitreous tap or biopsy in our study was comparable at 46% and 47%, respectively, with similar results reported in the literature.4 Although the EVS found no difference in visual outcomes between tap and inject versus vitrectomy for those with Hand Motions vision or better, a lower re-culture rate was noted in eyes that underwent vitrectomy (13% vs 71%). A vitrectomy not only reduces the bacterial load but removes the inflammatory debris and biofilms, which hinder the passage and action of intravitreal antibiotics otherwise.25 In our study, we found that 37% of eyes had a positive vitrectomy culture, despite initial treatment with intravitreal antibiotics. Despite the advantages of early vitrectomy, the challenges associated with access to theatre resulted in a longer delay in time to treatment for patients undergoing primary vitrectomy (4.5 hours) compared with those that received intravitreal antibiotics alone (3 hours).
This is the largest study of the etiology, clinical presentation, and culture results of endophthalmitis in New Zealand. We will present the surgical and clinical outcomes of patients within this cohort in a subsequent article. The retrospective nature of our study design limited our ability to assess other signs on clinical presentation, such as an afferent pupillary deficit, which have also been associated with positive culture in previous publications.17 Furthermore, we were unable to report on microbial sensitivity as these results were either not available or reported for all patients. Routine use of PCR testing may have allowed a microbiological diagnosis in a greater number of cases; however, this was not reported or performed routinely by our laboratory service.
In conclusion, a multimodal approach to intraocular sampling should be considered in patients presenting with endophthalmitis. The study highlights that taking an aqueous tap taken at the time of a vitreous sample increases the likelihood of obtaining a positive culture. The routine use of molecular diagnostic techniques in conjunction with traditional culture methods should be employed to improve the probability of identifying the pathogenic organism.
The authors acknowledge Salvin Singh, Microbiology Specialist Technician (ADHB), and Luci Kilimasi, Support Services at the Department of Ophthalmology (ADHB).
1. Results of the Endophthalmitis Vitrectomy
Study: a randomized trial of immediate vitrectomy
and of intravenous antibiotics for the treatment of postoperative bacterial endophthalmitis
. Arch Ophthalmol
2. Seal D, Reischl U, Behr A, et al. Laboratory diagnosis of endophthalmitis
: comparison of microbiology
and molecular methods in the European Society of Cataract & Refractive Surgeons multicenter study and susceptibility testing. J Cataract Refract Surg
3. Melo GB, Bispo PJM, Campos Pignatari AC, et al. Real-time polymerase chain reaction test to discriminate between contamination and intraocular infection after cataract surgery. J Cataract Refract Surg
4. Barza M, Pavan PR, Doft BH, et al. Evaluation of microbiological diagnostic techniques in postoperative endophthalmitis
in the Endophthalmitis Vitrectomy
Study. Arch Ophthalmol
5. Fan JC, Niederer RL, Von Lany H, Polkinghorne PJ. Infectious endophthalmitis
: clinical features, management and visual outcomes. Clin Exp Ophthalmol
6. Jabs DA, Nussenblatt RB, Rosenbaum JT, et al. Standardization of uveitis nomenclature for reporting clinical data. Results of the first international workshop. Am J Ophthalmol
7. Trusko BE, Thorne J, Jabs D, et al. The Standardization of Uveitis Nomenclature (SUN) project: development of a clinical evidence base utilizing informatics tools and techniques. Methods Inf Med
8. Sjoholm-Gomez de Liano C, Soberon-Ventura VF, Salcedo-Villanueva G, et al. Sensitivity, specificity and predictive values of anterior chamber tap in cases of bacterial endophthalmitis
. Eye Vis (Lond)
9. Williams GA. IVT Injections: Health Policy Implications. Review of Ophthalmology 2014. Available at: https://www.reviewofophthalmology.com/article/ivt-injections-health-policy-implications
10. Fileta JB, Scott IU, Flynn HW. Meta-analysis of infectious endophthalmitis
after intravitreal injection of anti-vascular endothelial growth factor agents. Ophthalmic Surg Lasers Imaging Retina
11. Laxminarayan R, Duse A, Wattal C, et al. Antibiotic resistance-the need for global solutions. Lancet Infect Dis
12. Department of Clinical Microbiology
ACH. Antimicrobial Susceptibility Test Results for Gram-positive isolates recovered from clinical specimens: 2018. Available at: http://www.labplus.co.nz/assets/Uploads/2018-Gram-Positive-Antimicrobial-Report.pdf
13. Department of Clinical Microbiology
ACH. Antimicrobial Susceptibility Test Results for Gram-negative isolates recovered from clinical specimens: 2018. Available at: http://www.labplus.co.nz/assets/Uploads/2018-Gram-Negative-Antimicrobial-Report.pdf
. Accessed February 2, 2019.
14. Gan IM, Van Dissel JT, Beekhuis WH, et al. Intravitreal vancomycin and gentamicin concentrations in patients with postoperative endophthalmitis
. Br J Ophthalmol
15. Relhan N, Albini TA, Pathengay A, et al. Endophthalmitis
caused by Gram-positive organisms with reduced vancomycin susceptibility: literature review and options for treatment. Br J Ophthalmol
16. Zarranz-Ventura J, Keane PA, Sim DA, et al. Evaluation of objective vitritis grading method using optical coherence tomography: influence of phakic status and previous vitrectomy
. Am J Ophthalmol
17. Johnson MW, Doft BH, Kelsey SF, et al. The endophthalmitis vitrectomy
study: relationship between clinical presentation and microbiologic spectrum. Ophthalmology
18. Russell M, Polkinghorne PJ. Controlled aspiration pressure in methods for collecting pure vitreous samples during vitreous biopsy
19. Mura M, Buschini E, Iannetta D, et al. Vitreous biopsy
under air. Retina
20. Donahue SP, Kowalski RP, Jewart BH, et al. Vitreous cultures in suspected endophthalmitis
. Biopsy or vitrectomy
21. Mudhar HS, Sheard R. Diagnostic cellular yield is superior with full pars plana vitrectomy
compared with core vitreous biopsy
. Eye (Lond)
22. Chiquet C, Maurin M, Thuret G, et al. Analysis of diluted vitreous samples from vitrectomy
is useful in eyes with severe acute postoperative endophthalmitis
23. Cornut PL, Boisset S, Romanet JP, et al. Principles and applications of molecular biology techniques for the microbiological diagnosis of acute post-operative endophthalmitis
. Surv Ophthalmol
24. Chiquet C, Cornut PL, Benito Y, et al. Eubacterial PCR for bacterial detection and identification in 100 acute postcataract surgery endophthalmitis
. Invest Ophthalmol Vis Sci
25. Doft BH, Kelsey SF, Wisniewski SR. Additional procedures after the initial vitrectomy
or tap-biopsy in the endophthalmitis vitrectomy