Post Traumatic PolymicrobialEndophthalmitis, Including NeisseriasubflavaShobha Sharma, DO,Norman A. Saffra, MD, FACS,and Edward K. Chapnick, MD

PURPOSE: To report the second known case of post-traumatic endophthalmitis caused by Neisseria subflava.DESIGN: Interventional case report.METHODS: A two-year-old child with post-traumatic cor-neal laceration and uveal prolapse required medical andsurgical therapy for endophthalmitis caused by multipleorganisms including N. subflava.RESULTS: After aggressive therapy, patient had a favor-able outcome without vision compromise.CONCLUSIONS: As there is still not a standard protocol fortherapy for post-traumatic endopthalmitis, we recom-mend the use of broad-spectrum antibiotics via intravit-real, intravenous, and topical routes. Consideration oftypical and unusual bacteria that have been reported tocause endopthalmitis, as well as the source of injury,should guide antibiotic choice. (Am J Ophthalmol2003;136:554–555. © 2003 by Elsevier Inc. All rightsreserved.)

ENDOPHTHALMITIS, A POTENTIALLY BLINDING INTRAOC-

ular infection, can be more devastating in children.Outcome is related to bacterial virulence factors, size of theinoculum, and time to diagnosis and treatment.1

Many microorganisms have been implicated in exog-enous endophthalmitis. An analysis of post-traumaticendophthalmitis in children aged 18 months to 13 yearsidentified streptococci (56.6%) and staphylococci(22.2%) as the most frequent causes.2 The Endoph-thalmitis Vitrectomy Study (EVS) showed that the mostcommon causes of postcataract surgery exogenous en-dophthalmitis were coagulase-negative staphylococci(70%), Staphylococcus aureus (10%), streptococci (9%),other gram-positive organisms (5%) and gram-negativeorganisms (6%).3

For acute postoperative endophthalmitis, vitrectomy orvitreous aspirate with culture is recommended followed byintravitreal vancomycin and amikacin or ceftazidime.3 TheEVS study demonstrated no benefit of systemic antibioticsin acute postoperative endophthalmitis.4 The role of sys-temic treatment in other types of endophthalmitis iscontroversial. Some physicians utilize intravenous vanco-mycin plus ceftazidime, or oral ciprofloxacin.

We report a case of post-traumatic polymicrobial en-dophthalmitis with a gram-positive organism on gram stainbut growth of only Neisseria subflava. A 2-year-old girl wasaccidentally stuck with a fork in her right eye andpresented 24 hours later with pain. Examination revealeddiffuse ocular injection with an irregularly shaped pupil. Atsurgery, a corneal laceration with uveal prolapse into thewound was seen. The corneal injury was sutured. Fibrinousmaterial from the anterior chamber was removed, re-formed, and copiously irrigated. Vancomycin, clindamy-cin, and dexamethasone were instilled subconjunctively,and cefazolin, clindamycin, and methylprednisolone werebegun intravenously along with topical ciprofloxacin,prednisone, and atropine.

Examination of the anterior chamber fibrinous materialrevealed gram-positive cocci in chains. On postoperativeday four, examination failed to reveal significant improve-ment, and post-traumatic endophthalmitis was suspected.The patient was taken to the operating room for apars-plana vitrectomy (PPV) and intravitreal vancomycin;ceftazidime and dexamethazone were administered. Intra-venous antibiotics including ceftazidime, instead of cefa-zolin, were continued. Two days after PPV, cultures fromadmission grew N. subflava. Intravitreal injection of gen-tamicin and vancomycin were repeated. On the ninthhospital day, the patient showed significant improvementof the anterior segment examination. On day 10 thepatient was discharged on oral amoxicillin/clavulanate andtopical medications.

N. subflava, a normal inhabitant of the human naso-phyranx, has been noted to cause rare infections. Welocated one case report of N. subflava endophthalmitis,which occurred in an alcoholic man in whom the patho-genesis of the infection remained uncertain.5 Our casedemonstrates that the object of injury must be consideredwhen assessing empiric antibiotic coverage.

In certain forms of endophthalmitis, for both childrenand adults, the use of a parenteral remains controversial.Antibiotics such as vancomycin, gentamicin, and cepha-losporins do not penetrate the blood–ocular fluid barrierwell.4 Systemic gentamicin, vancomycin, or clindamycinshould be given after penetrating trauma.6 Although thereis no clear consensus regarding the use of broad-spectrumparenteral antibiotics in traumatic injuries of the eye,many clinicians utilize intravenous, intravitreal, and topi-cal routes. Aggressive and prompt treatment with broad-spectrum topical antibiotics should immediately beinitiated after considering common and unusual organisms.

Accepted for publication Feb 24, 2003.From the Division of Infectious Diseases, Department of Medicine

(S.S., E.K.C.), and the Division of Ophthalmology (N.A.S.), Mai-monides Medical Center, Brooklyn, New York.

Inquiries to Shobha Sharma, DO, Division of Infectious Diseases,Department of Medicine, Maimonides Medical Center, 4802 10th Ave-nue, Brooklyn, NY 11219; e-mail: Sshobha@msn.com

AMERICAN JOURNAL OF OPHTHALMOLOGY554 SEPTEMBER 2003

REFERENCES

1. Whitcher JP. The treatment of endophthalmitis—still anexercise in frustration. Br J Ophthalmol 1997;81:713–715.

2. Alfaro DV, Roth DB, Laughlin RM, Goyal M, Liggett PE.Paediatric post-traumatic endopthalmitis. Br J Ophthalmol1995;79:888–891.

3. Han DP, Wisniewski SR, Wilson LA, et al. Spectrum andsusceptibilities of microbiologic isolates in the endophthalmi-tis vitrectomy study. Am J Ophthalmol 1996;122:1–17.

4. Endophthalmitis Vitrectomy Study Group. A randomized trialof immediate vitrectomy and of intravenous antibiotics for thetreatment of postoperative bacterial endophthalmitis. ArchOphthalmol 1995;113:1479–1496.

5. Stricker RB, Pompilio KJ, Axelrod JL, Kochman RS, NewtonJC. Neisseria subflava endophthalmitis. Am J Ophthalmol1982;94:423–424.

6. Hemady R, Zaltas M, Paton B, Foster CS, Baker AS. Bacillus-induced endophthalmitis: new series of 10 cases and review ofthe literature. Br J Ophthalmol 1990;74:26–29.

Potential Pitfalls From VariableOptical Coherence TomographDisplays in Managing DiabeticMacular EdemaDavid J. Browning, MD, PhD

PURPOSE: To describe potential pitfalls in the use ofoptical coherence tomography for management of dia-betic macular edema.DESIGN: Prospective, noninterventional case series.METHODS: Review of optical coherence tomographs in 13eyes with clinically significant diabetic macular edemafrom 11 consecutive patients in a private retina practice.RESULTS: Optical coherence tomography displays arebased on 3.5–mm or 6–mm diameter circular grids thatlook very similar and have identical sector names butschematize different areas of the macula. The numericoutputs for the identically named sectors in the twodisplays do not differ significantly for retinal thicknessbut differ significantly in all sectors except the fovea forretinal volume because of the different areas representedby the sectors.

CONCLUSIONS: Failure to explicitly note the scale ofthe optical coherence tomography display can potentiallymisdirect planned focal and grid laser treatmentfor diabetic macular edema. Failure to explicitly verifyidentical optical coherence tomography display scalesin longitudinal studies of laser treatment for diabeticmacular edema can potentially lead to errors in inter-preting treatment efficacy. (Am J Ophthalmol 2003;136:555–557. © 2003 by Elsevier Inc. All rightsreserved.)

THE POTENTIAL ADVANTAGES OF USING OPTICAL CO-

herence tomography (OCT) in the management ofdiabetic macular edema have been suggested in numerousreports, but the practical aspects of the use of the tool havenot received much attention.1–4 Macular thickening canbe objectively measured, whereas the prevailing stereobiomicroscopic method of edema assessment is inherentlysubjective.5–7 In the past, planning of focal laser treatmenthas depended on fluorescein angiography to identify leak-ing microaneurysms, telangiectatic small vessels, and non-perfused thickened retina. Current clinical practice andcontemporary National Eye Institute–sponsored random-ized trials for treatment of macular edema depart from thisstandard and do not require, but do not preclude, the useof fluorescein angiography for treatment planning (Diabet-ic Retinopathy Clinical Research.net [DRCR.net]). Con-comitantly, the use of OCT in clinical practice andresearch studies is expanding and being defined. The OCTproduces a false-color map based on six radial scans with100 measurements taken along each scan. In the OCTmodels 2 and 3; the display scale can be specified as 3.5 mmor 6 mm. The purpose of this work is to show that theclinician must be aware of differences in data display toavoid both wrong inferences and consequentially wrongplanning for laser treatment and erroneous interpretationof treatment effects in serial scans.

Thirteen eyes with clinically significant diabetic macu-lar edema from 11 consecutively examined patients werescanned with the OCT 3. The data were displayed with a3.5-mm and a 6.0-mm diameter (Figure 1, left). Thefalse-color maps and circular numerical printouts werecompared. One-way analysis of variance was used forstatistical comparison using JMP software (SAS Institute,Cary, North Carolina, USA). Data were accumulated inconformity with all federal and state laws.

The false-color maps and numeric printouts for the twodisplay diameters are very similar, though not identical, asshown in Figure 1, right . The retinal thickness for the ninezones shown in the OCT printout does not vary to asignificant extent (data not shown). The retinal volumesby zone vary significantly for all zones except the fovealzone, as shown in Table 1. Retinal volumes are greater for

Accepted for publication March 20, 2003.From the Charlotte Eye, Ear, Nose, and Throat Associates, Charlotte,

North Carolina.Inquiries to David J. Browning, MD, PhD, 6035 Fairview Rd.,

Charlotte, NC 28210; fax: (704) 295-3186; e-mail: djbrowning@carolina.rr.com.

BRIEF REPORTSVOL. 136, NO. 3 555