Blebitis, Early Endophthalmitis, and Late Endophthalmitis after Glaucoma..-filtering Surgery

Thomas A. Ciulla, MD,I Allen D. Beck, MD/ Trexler M. Topping, MD/ Ann Sullivan Baker, MD 4

Purpose: The differentiating characteristics in blebitis and early and late endophthal­mitis after glaucoma filtration surgery are reviewed.

Methods: All admission records and operative reports, as well as available office notes, on patients with blebitis or bleb-associated endophthalmitis admitted to a large referral eye center from 1985 to 1995 were reviewed retrospectively.

Results: Ten cases of blebitis and 33 cases of bleb-associated endophthalmitis were identified. One patient with blebitis progressed to culture-positive endophthalmitis. Of the 33 cases of bleb-associated endophthalmitis, there were 6 cases of early endoph­thalmitis (before postoperative week 6) and 27 cases of late endophthalmitis. In early endophthalmitis, Staphylococcus epidermidis was isolated on vitreous culture in 4 (67%) of 6 cases, whereas in late endophthalmitis, this organism was isolated in only 1 (4%) of 27 cases.ln the 27 late cases, Streptococcus species and gram-negative organisms comprised 48% of isolates; of 33 cases of endophthalmitis, 15 (45%) demonstrated no growth on vitreous culture. Patients with Eindophthalmitis fared more poorly than those with blebitis in terms of visual outcome.

Conclusions: Because blebitis may be prodromal to endophthalmitis, aggressive antimicrobial therapy, perhaps with oral quinolones, is warranted. In addition, patients with blebitis should be observed closely to identify extension into the vitreous cavity so that intravitreous antibiotics can be administered in a timely fashion. Finally, clinicians should not extrapolate the results of the Endophthalmitis Vitrectomy Study to the postfil­tration surgery endophthalmitis given the differing pathogenesis and unique spectrum of organisms. Ophthalmology 1997; 104:986-995

Originally received: October 29, 1995. Revision accepted: February 5, 1997.

There are two types of infection that occur after glaucoma filtration surgery. Blebitis is a term used to describe a presumed bleb infection without vitreous involvement, which may represent a limited form of endophthalmitis.! Bleb-associated endophthalmitis denotes an ocular infec­tion with involvement of the vitreous, which usually de­velops months or years after glaucoma filtering surgery.2-5 Rarely, endophthalmitis may occur in the early postopera­tive period after glaucoma filtering surgery.6-8

1 Department of Ophthalmology, Indiana University School of Medi­cine, Indianapolis, Indiana.

2 Department of Ophthalmology, Emory University School of Medicine, Atlanta, Georgia.

3 Ophthalmic Consultants of Boston; Department of Ophthalmology, Harvard Medical School, and Department of Ophthalmology, Tufts School of Medicine, Boston, Massachusetts.

4 Infectious Disease Unit, Department of Ophthalmology, Massachusetts Eye and Ear Infirmary, and Medicine Division, Massachusetts General Hospital, Harvard Medical School, Boston, Massachusetts.

Presented in part at the Annual Meeting of the American Academy of Ophthalmology, Atlanta, Georgia, October 1995.

Supported in part by the Heed Ophthalmic Foundation, Cleveland, Ohio.

The authors have no proprietary interest in any of the materials used in this study.

986

Prior reports on these entities have included small sam­ple sizes or have included significant numbers of cases in which unplanned blebs developed after cataract sur­gery. In these unplanned blebs, it may not be possible to

Reprint requests to Thomas A. Ciulla, MD, Retina Section, Department of Ophthalmology, Indiana University School of Medicine, 702 Rotary Circle, Indianapolis, IN 46202-5175.

Ciulla et al . Infection after Glaucoma-filtering Surgery

determine how long the bleb existed before the patient's presentation with infection, and it is not possible to deter­mine whether the bleb resulted from full-thickness or par­tial-thickness wound construction. In addition, the litera­ture on blebitis is limited, and no cases of blebitis pro­gressing to culture-positive endophthalmitis have been documented.

This current study corroborates and amplifies several observations regarding these entities. First, this report di­rectly compares blebitis to bleb-associated endophthal­mitis regarding clinical presentation, recovered organ­isms, and clinical outcomes. It also documents the pro­gression of blebitis to culture-positive endophthalmitis, supporting the position that blebitis may be prodromal to endophthalmitis and may not truly represent a distinct entity. In addition, this study reaffirms the unique spec­trum of organisms in late endophthalmitis and suggests that conjunctival cultures have a limited, but occasionally useful, role in managing these infections. Finally, man­agement and treatment recommendations are made based on these observations. '

Patients and Methods

Retrospective review was undertaken on the admission records and operative reports of patients with blebitis or bleb-associated endophthalmitis admitted from 1985 to 1995. In addition, available records from community oph­thalmologists were reviewed retrospectively to determine preinfection status and postinfection course. Only patients who had undergone actual glaucoma filtration surgery before infection were included. Patients with inadvertent blebs who had gone on to have infection develop were excluded.

Data collected included age, gender, laterality, glau­coma type, glaucoma filter type and date, concomitant eye disease, and other previous or simultaneous eye sur­gery. The operative reports describing the original glau­coma filtration procedure also were reviewed when avail­able. Symptoms and signs on presentation, including best­corrected visual acuity, intraocular pressure, lens status, and Seidel testing, were noted. The results of all available cultures, including the conjunctiva, aqueous, vitreous, and eye medication cultures, also were noted. Treatment regi­men, preoperative and postoperative visual acuity, as well as intraocular pressures were reviewed.

The microbiologic culture techniques have been de­scribed previously.8,9 Briefly, all aqueous and vitreous sam­ples were cultured on aerobic media (e.g., blood agar, choco­late agar, and chopped meat-glucose broth at 37° C), anaero­bic media (e.g., blood agar, chocolate agar, and chopped meat-glucose broth in anaerobic chamber at 37° C), and fungus media (e.g., Sabaroud plate at 37° C). Vitreous and aqueous humor aspirates were inoculated directly. Vitreous washings were filtered under sterile conditions, and seg­ments of the 0.45-p.m filter membranes were placed on appropriate media.1O Organisms were identified by criteria

described previously, and antimicrobial susceptibility was assessed by Kirby-Bauer disc diffusion criteria. ll

-13

Case Report

A 39-year-old man (case 10) presented reporting an irritated, red eye and decreased vision in the right eye. He had a history of severe pigmentary glaucoma unresponsive to medical treatment, and he had undergone filtration surgery 5 years before presenta­tion. Five months before presentation, he underwent extensive redissection and revision of the bleb, which had failed. His postoperative course was unremarkable; he maintained good control of intraocular pressure with a well-formed bleb. He wore a hard contact lens to correct his myopia.

On examination, the visual acuity measured 20/400 and in­traocular pressure measured 16 mmHg. The anterior segment showed intense conjunctival injection around an opalescent bleb, clear cornea, mild cell and flare, and unremarkable iris and lens. Seidel testing was not performed. The vitreous was clear.

He was diagnosed with blebitis and hospitalized, and he was started on intense topical treatment, including topical steroids, fortified cefazolin, and fortified tobramycin. Cultures from the contact lens, contact lens case, and conjunctiva ultimately grew Staphylococcus epidermidis and Serratia marcescens. He re­mained hospitalized for 3 days, during which time the pain, conjunctival injection, and anterior chamber inflammation sub­sided. The vitreous remained clear.

Eight days after being diagnosed with blebitis, he returned with increased pain and redness. The visual acuity measured 201 400 and the examination results showed increased conjunctival injection and anterior chamber inflammation. The vitreous showed moderate cellular infiltrate. He was diagnosed with en­dophthalmitis, and he was readmitted (case 43). He underwent pars plana vitrectomy and intravitreous injection of 1 mg ceftaz­idime with 400 f-lg amikacin. Vitreous cultures grew S. marces­cens. He received intense topical steroids, fortified vancomycin and gentamicin, along with intravenous ceftazidime and ami­kacin initially. The infection ultimately resolved, but he was left with count fingers visual acuity at 3 years.

Results

Epidemiology and Treatment

Ten cases of blebitis and 33 cases of bleb-associated en­dophthalmitis, representing 37 eyes in 35 patients, were identified. Three patients had two distinct episodes of endophthalmitis in the same eye, separated by a minimum of 6 months' time. Endophthalmitis developed in two patients more than 1 year after successful treatment of blebitis. One patient described in the case report was admitted for blebitis, treated medically, discharged, and readmitted 8 days later with signs of endophthalmitis.

Fifteen patients were female and 20 were male. There were 30 cases of primary open-angle glaucoma, 7 with congenital or juvenile onset glaucoma, 2 with combined primary and narrow angle glaucoma, 2 with pigmentary glaucoma, 1 with Fuch heterochromia, and 1 with trau­matic glaucoma. The mean age at presentation with infec-

987

Ophthalmology Volume 104, Number 6, June 1997

Table l. Patient Characteristics

Age (yrs), Glaucoma Other History, Time from Vitreous Vision Vision

Patient Sex, Type, Other Surgery, Filter to Visual Aqueous Washings before after No. Eye Filter type Date Infection Acuity Lens Culture Culture Other Culture Infection Infection

I. Blebitis 1 28, M, JOAG High myopia 5 yrs HM Phakic NO NO Coni CX CF CF

00 Post lip Amblyopia 1 yr sip S. epi 1 yr 6 mas sclerectomy 5 Subsequent bleb yrs prior endophthalmitis repair

Rupture of bleb 1 yr later and repair 1 (case 20) yr prior

2 76, M, POAG 2 yrs 20/100 Phakic NO NO Coni CX OS Mito trab 2 yrs S. epi

prior 3 48, M, POAG 2 yrs 20/100 NO NO Coni CX 20/50

OS T rab 2 yrs prior negative 5 days 4 72, M, Combined mech 2 yrs 20/200 Phakic NO NO Coni CX 20/40 20/40

00 glaucoma S. aureus 2 mas 10 mas Post lip

sclerectomy 2 yrs prior

5 73, M, end-stage 4 yrs LP PCIOL NO NO Coni CX 20/200 LP 00 POAG S. epi and 4 yrs 4 days

ECCE/PCIOL/ S. aureus trab 4 yrs prior

6 66, F, POAG ICCE 15 yrs prior 22 yrs 20/30 ACIOL NO NO Coni CX 20/30 20/30 OS Full thickness Secondary ACIOL S. epi 16 10 mas

filter 22 yr 8 yrs prior mas prior

7 71, M, POAG ECCE/PCIOL 9 mas PCIOL NO NO Coni CX 20/200 OS Filter 10 yrs fro.m below 2 S. aureus

prior yrs prior Bleb revision 7 Subsequent

yrs prior endophthalmitis Mito trab from 17 mas later

below 9 mas (case 37) prior

8 26, M, JOAG 19 mas 20/20 Phakic NO NO Coni CX 20/20 20/20 00 Mito trab 19 S. aureus 3 mas

mas prior 9 48, M, POAG 2 yrs 20/50 Phakic NO NO Coni CX 20/50 20/50

00 Mito trab 2 yrs S. epi 3 mas prior

10 39, M, Pigmentary HCL wearer 5 mas sip 20/400 Phakic NO NO Coni HCL, and CF 00 glaucoma Required second revision without case CX

Filter (unknown admission 8 HCL rare S. type) 5 yrs days later for marcesans, S. prior Serratia epi

Bleb revision 5 endophthalmitis mas prior (case 43)

II. Early endophthalmitis

11 78, F, POAG 6 wks 20/400 Phakic NO S. epi NO NLP OS Trab 6 wks 18 mas

prior 12 66, F, POAG 2 days HM Phakic Negative Negative Coni CX 20/200

00 Trab 2 days S. epi 1 day prior

13 78, F, Narrow angle CRVO 5 mas 2 wks sip HM PCIOL S. epi S. epi Vit BX CX CF 00 glaucoma prior combined S. epi 1 mo

ECCE/PCIOL/ Suture cut 1 wk 1 wk sip Coni CX trab 2 wks prior suture cut negative prior (prior topical

ABX)

{continues

988

Ciulla et al . Infection after Glaucoma-filtering Surgery

Tab Ie 1 (continued).

Age (yrs), Glaucoma Other History, Time from Vitreous Vision Vision

Patient Sex, Type, Other Surgery, Filter to Visual Aqueous Washings before after No. Eye Filter type Date Infection Acuity Lens Culture Culture Other Culture Infection Infection

14 89, M, POAG I mo HM ND S. epi Vit BX CX LP OD Mito trab I mo S. epi (same 7 days

prior bact as cornea)

Corneal ulcer CX P. mirabilis, S. epi, E. cloacae

Eye med CX P. mirabilis, S. epi, E. cloacae

15 47, M, POAG 3 wks 20/60 PCIOL Negative Negative Vit BX CX CF OD Phaco/PCIOL/ negative 3 days

mito trab 3 wks prior

16 82, M POAG I day CF PCIOL S. epi ND Phaco/PCIOL/

trab I day prior

III. Late endophthalmitis A. Full thickness filters 17 81, M, POAG CRAO 4 yrs NLP ND ND Intraocular NLP

OD Post lip Cataract surgery abscess with 4 yrs sclerectomy 4 in remote past purulent yrs prior discharge CX

S. pneumoniae 18 66, F, POAG ICCE/ACIOL 7 12 yrs LP ACIOL Enterococcus Enterococcus Coni CX 20/40 20/400

OD Post lip wks prior 1 wk sip spp. spp. negative 2 days 4 yrs sclerectomy Suture removal 1 suture 12 yrs prior wk prior removal

19 70, M, POAG ECCE/PCIOL 5 yrs 20/50 PCIOL ND Negative Coni CX 20/20 20/80 OD Post lip inferiorly 3 yrs few S. 2 mos 5 mos

sclerectomy 5 prior pneumoniae , yrs prior very rare

S. epi, P. acnes

20 29, M, JOAG High myopia 6 yrs CF Phakic ND Negative Loose coni CF CF OD Post lip Amblyopia 2 yrs sip suture CX 6 mos 1 yr

sclerectomy 6 Blebitis I yr prior bleb H. injluenzae yrs prior treated repair

Spont rupture of medically bleb and (case 1) repair 2 yrs prior

21 30, M, JOAG High myopia 7 yrs HM Phakic ND Negative Coni CX CF CF OD Post lip Amblyopia 3 yrs sip negative 1 yr 6 mos

sclerectomy 7 Past bleb yrs prior endophthalmitis repair

Rupture of bleb 1 yr prior (case and repair 3 20) yrs prior

22 52, M, POAG CE/PCIOL 4 yrs 13 yrs LP PCIOL ND Streptococcus Vit BX HM OS Post lip prior spp. Streptococcus 5 days

sclerectomy spp. LP 13 yrs prior 1.5 yrs

23 87, F, POAG ICCE from below 2 yrs HM Aphakic ND S. epi ND 20/40 20/40 OS Trephine 22 yrs OU 16 yrs prior 16 21 mos

prior mos Leak repair 2

yrs prior

(continues

989

Ophthalmology Volume 104, Number 6, June 1997

Table 1 (continued).

Age (yrs), Glaucoma Other History, Time from Vitreous Vision Vision

Patient Sex, Type, Other Surgery, Filter to Visual Aqueous Washings before after No. Eye Filter type Date Infection Acuity Lens Culture Culture Other Culture Infection Infection

24 50, M, Congen ICCE 42 yrs prior, 44 yrs CF Aphakic Negative Negative Corneal ulcer 20/30 20/40 00 Full thickness PK for and ABK CX S. aureus 18 7 days

filter 44 yrs 13 yrs and 2 yrs Eye med CX mos prior prior rare S. epi,

mod Candida spp. and abundant E. faecalis

B. Partial thickness filters (trabeculectomy)

25 58, F, Fuch 7 yrs LP Phakic Negative Negative NO 20/400 HM OS T rab 7 yrs prior 1 mo 8 days

26 75, M, POAG 3 yrs LP PCIOL NO Streptococcus Coni CX 00 Trab 3 yrs prior spp. Streptococcus

Repeat 3 days spp. later for persist infection: negative

27 63, M, POAG 2.5 yrs LP Phakic Negative Negative after Coni CX LP OS Trab 2.5 yrs prior to IVABX S. epi 10

prior IVABX days 28 77, F, POAG Trab 4 ICCE 3 yrs prior 4 yrs 20/200 Aphakic NO Negative Coni CX CF

00 yrs prior S. epi and 14 S. aureus days

Vit BX CX negative

29 83, F, POAG ICCE 9 yrs prior 10 yrs HM Aphakic NO Negative (vit Vit BX CX CF 00 Trab 10 yrs Past tap only no negative 7 days

prior endophthalmitis PPV) 6 yrs prior (case 28)

30 68, F, POAG 3.5 yrs HM Phakic M. Abundant M. Vit BX CX 20/25 20/100 OS Trab 3.5 yrs catarrhalis catarrhalis . abundant M. 14 8 mos

prior rare P. acnes catarrhalis. mos rare P. acnes

31 48, M, Congen PK, anterior 5 yrs CF 6 ft PCIOL Negative Negative NO CF CF OS Full thickness PPV, ECCE, 7 mos 7 days

filter 42 yrs PCIOL prior iridoplasty 4 yrs

T rab 5 yr prior prior 32 49, M, Congen PK, anterior 6 yrs HM PCIOL Negative Negative NO CF

OS glaucoma PPV, ECCE, 10 Full thickness PCIOL days

filter 43 yrs iridoplasty 5 yrs prior prior

T rab 6 yrs prior Past endophthalmitis 1 yr prior (case 31)

33 67, F, POAG 2 yrs CF PCIOL NO S. rnarcescens Coni CX 20/40 NLP OS ECCE/PCIOL/ S. epi 1 mo 6 days

trab 2 yrs Vit BX CX prior S. rnarcescens

Enucleated for blind painful eye POD 6

34 80, F, POAG 3 yrs LP PCIOL M. M. catarrhalis Coni CX 20/50 CF OS CE/PCIOL/trab catarrhalis S. epi 3 yrs 8 days

6 yrs prior Revision 3 yrs

prior

(continues)

990

Ciulla et al . Infection after Giaucoma~filtering Surgery

Tab Ie 1 (continued).

Age (yrs), Glaucoma Other History, T ime from Vitreous Vision Vision

Patient Sex, T ype, Other Surgery, Filter to Visual Aqueous Washings before after No. E ye Filter type Date Infection Acuity Lens Culture Culture Other Culture Infection Infection

35 59, F, POAG 5 -FU 2 yrs prior 2 y rs LP PCIOL H. injluenzae H. injluenzae NO 20/60 NLP OS Trab/CE/PCIOL 26 6 mos

3 y rs prior mas phthi-Repeat trab 2 sis

yrs prior 36 82, F, POAG Blebitis treated 4 yrs H M PCIOL Negative Negative Conj CX 20/30

00 CE/PCIOL/trab medically 1 day S. epi and 5 mos 4 yrs prior prior S. aureus

Vit BX CX negative

C. Partial thickness filters with mitomycin (trabeculectomy with mitomycin)

37 72, M, POAG ECCEjPCIOL 2 yrs H M OS Filter 11 yrs from below 3

prior yrs prior Bleb revision 8 5-FU 2 y rs prior

yrs prior Blebitis 17 mas Miro trab from prior treated

below 2 yrs topically with prior subsequent

hypotony and 20/200 (case 7)

38 47, M, POAG Scleral buckle 10 1 yr LP OS Phaco/PCIOL/ yrs prior 2 days sip

mito trab 1 A utologus blood blood yr prior injection 2 day injection

prior for persistent hypotony maculopathy

39 66, F, POAG Congen cataracts 2.5 yrs LP 0 0 Ant vit/mito High myopia

trab infe riorly ICCE 20 yrs prior 2.5 yrs prior

40 74, F, POAG Leak 6 wks prior 2.5 yrs LP OS Phaco./PC IOL/ treated with 6 wks sip

mito trab 2.5 bandage leak yrs prior contact lens

and glue 41 32, M, T raumatic ECCE/PCIOL 21 mos C F

OS hyphema and temporally 1 yr cat (hockey) prior 2 yrs prior

Mito trab 21 mos prior

42 75, F, POAG ECCE/PCIOL 3.5 3 yrs C F 0 0 Mito trab yrs prior

inferior-temporally 3 yrs prior

tion was 63 .7 years (range, 26-89 years). Twenty-two cases involved infections of the right eye, and 20 cases involved the left eye. Laterality was not specified in one case.

All ten patients with blebitis underwent conjunctival culture and were treated with intense topical b road-spec­trum antibiotics; seven of these ten patients received topi­cal steroids initially as well. All 33 patients with endoph-

PCIO L Negative Negative NO 20/200 CF 6 mos 10

mas

PCIOL Negative M. catarrhalis NO 20/25 20/60 9 days 6 mos

Aphakic NO Streptococcus NO 20/30 CF spp. 2 y rs 4 mos

PCIOL Streptococcus Streptococcus Conj CX 20/30 LP spp. spp. Streptococcu-s 13 14

spp. mos days

PCIOL Streptococcu-s Streptococcus NO 20/50 HM spp. spp. 5 mos 6 wks

PCIOL N O Negative NO 20/30 20/100 3 mos 3 days

(continues)

thalmitis in this series underwent vitreous culture. Thirty­one of these 33 patients underwent pars plana vitrectomy; one patient (case 29) underwent vitreous tap and another patient underwent primary evisceration (case 17). All 33 patients received broad-spectrum topical and intravenous antibiotics, and all 33 patients except the 1 who underwent primary evisceration received intravitreous antibiotics. The most commonly used antibiotic regimen consisted of

991

Ophthalmology Volume 104, Number 6, June 1997

Tab Ie 1 (continued).

Age (yrs),

Patient Sex, No. Eye

Glaucoma Type,

Filter type

D. Unknown filter type

43 39, M, Pigmentary OD glaucoma

Filter (unknown type) 5 yrs prior

Bleb revision 5 mos prior

Other History, Other Surgery,

Date

HCL wearer Blebitis treated

medically 8 days prior (case 10)

Time from Filter to Visual Infection Acuity

5 mos sip revision

20/400

Lens

Phakic

Aqueous Culture

Negative

Vitreous Washings Culture

S. marcescens

Vision Vision before after

Other Culture Infection Infection

ND CF 3 yrs

M = male; F = female; ND = not done; JOAG = juvenile open-angle glaucoma; POAG = primary open-angle glaucoma; combined mech glaucoma = combined mechanism glaucoma (primary open-angle glaucoma with later superimposed narrow angle glaucoma); congen = congenital glaucoma; Fuch = Fuch heterochromic iridocyclitis; trab = trabeculectomy; mito trab = trabeculectomy with adjunctive mitomycin; CE = cataract extraction; ICCE = intracapsular cataract extraction; ECCE = extracapsular cataract extraction; ACIOL = anterior chamber intraocular lens; PCIOL = posterior chamber intraocular lens; PK = penetrating keratoplasty; ABK = aphakic bullous keratopathy; PPV = pars plana vitrectomy; HCL = hard contact lens; CRAO = central retinal artery occlusion; CRVO = central retinal vein occlusion; 5-FU = 5-fluorouracil; NLP = no light perception; LP = light perception; HM = hand motions; CF = count fingers; CX = culture; BX = biopsy; ABX = antibiotics; IV = intravenous; S. epi = Staphylococcus epidermidis; S. aureus = Staphylococcus aureus; P. mirabilis = Proteus mirabilis; S. marcescens = Serratia marcescens; S. pneumoniae = Streptococcus pneumoniae; E. faecalis = Enterococcus faecalis; M. catarrhalis =

Moraxella catarrhalis; P. acnes = Propionibacterium acnes; H. influenzae = Hemophilis influenzae; E. cloacae = Enterobacter cloacae.

vancomycin plus amikacin (24 patients; 1 mg vancomy­cinl400 f.Lg amikacin in 21 patients, 1 mg vancomycin/ 200 f.Lg amikacin in 3 patients), followed by vancomycin plus gentamicin (4 patients), followed by cefazolin plus gentamicin (3 patients), followed by cefazolin plus ami­kacin (1 patient).

Among these 33 patients with endophthalmitis, there were 6 cases of early endophthalmitis, occurring at or within 6 weeks from filtration surgery (mean, 2.6 weeks; range, 1 day-6 weeks). There were 27' late cases of en­dophthalmitis, occurring after this period (mean, 5.6 years; range, 1-42 years). Of the 33 patients with endoph­thalmitis, there were 8 involving full-thickness filtering blebs, 16 involving trabeculectomy (partial-thickness) filtering blebs, 8 involving mitomycin-treated trabeculec­tomy (partial thickness) filtering blebs, and 1 involving an unknown filter type. Two patients were noted to have undergone 5-fluorouracil injection 2 years before infec­tion. Because records from community ophthalmologists were not available for all patients, however, it is possible that some additional patients had undergone 5-fluoroura­cil injection in the past before infection, in which case, prior injection may not have been noted in the admission record. These patient characteristics, as well as the param­eters listed in the Methods section, are summarized in Table 1.

Characteristics of Blebitis Versus Bleb,associated Endophthalmitis

The characteristic symptoms and signs of blebitis differed from bleb-associated endophthalmitis. All patients with blebitis reported red eyes and photophobia; they presented with conjunctival discharge and severe conjunctival injec-

992

tion centered around the opalescent filtering bleb. Seidel testing results were positive in five of seven patients tested. Anterior chamber reaction was variable. However, none of the ten patients showed vitritis based on examina­tion results. All ten patients received intense topical broad-spectrum antibiotic treatment, and seven of ten pa­tients received frequent prednisolone acetate drops. In nine patients, the infection resolved using this regimen, and in five of these patients, there was prompt return of preinfection visual acuity as noted in Table 1.

In contrast, patients with endophthalmitis had more rapidly progressive presentations, with worsening pain, acuity, and anterior chamber reactions. Seidel testing re­sults were positive in 7 of ] 3 patients tested. As listed in Table 1, postoperative visual acuities generally were poor in the count fingers to light perception range for most cases. The visual outcome is somewhat difficult to assess in this study, given the often poor or unavailable preinfec­tion visual acuity and the varying follow-up periods.

The defining feature in the patients with endophthal­mitis, in contrast with those patients with blebitis, was the presence of vitreous cells in all cases. The degree of cellular infiltration varied from trace to frank intraocular abscess as in the 81-year-old patient (case 17) who pre­sented with a purulent frank intraocular abscess, yielding Streptococcus pneumoniae on culture, and necessitating evisceration.

Conjunctival Cultures in Blebitis

Conjunctival cultures were performed in all ten patients with blebitis. There were four cultures with S. epider­midis, three with S. aureus, one with mixed S. epidermidis and S. aureus, and one negative culture result. One addi-

Ciulla et al . Infection after Glaucoma-filtering Surgery

tional patient with a conjunctival culture yielding S. epid­ermidis and S. marcescens ultimately progressed while being treated medically, growing S. marcescens on vitre­ous culture when readmitted 8 days later for endophthal­mitis (as described in the Case Report). This was the only patient whose conjunctival culture yielded an organism other than Staphylococcus species.

Vitreous Cultures in Early and Late Bleb­associated Endophthalmitis

The organisms isolated on vitreous culture from the 33 patients with bleb-associated endophthalmitis are noted in Table 1. Four (67%) of the six patients with early endophthalmitis yielded S. epidermidis on vitreous cul­ture. Two of the six patients showed no growth on vitre­ous culture. One of these two patients (case 12) showed S. epidermidis on conjunctival culture.

In contrast, in late endophthalmitis, only 1 case of 27 yielded S. epidermidis on vitreous culture (4%). There were six cases with Streptococcus spp., three cases with Moraxella catarrhalis, two cases with S. marcescens, and one case each with Hemophilus influenza and Enterococ­cus spp., as well as 13 cases with negative cultures. The Streptococcus spp. and gram-negative organisms together comprised 48% of isolates. When analyzed by filter type, the vitreous cultures in each group showed similar pat­terns. Fifteen of 33 cases were culture-negative (45%).

Aqueous Cultures in Bleb-associated Endophthalmitis

Of the 33 patients with endophthalmitis, 18 underwent aqueous culture along with vitreous culture. These cul­tures showed good correlation overall, although the aque­ous culture was not as sensitive. There were 16 correlating aqueous and vitreous cultures, 9 of which the results were negative. There were two cases in which the aqueous and vitreous cultures did not correlate. In case 38, the aqueous culture results were negative, whereas the vitreous culture grew M. catarrhalis. In case 43, the aqueous culture re­sults were negative, whereas the vitreous culture grew S. marcescens.

Conjunctival Cultures in Bleb-associated Endophthalmitis

Of the 33 patients with endophthalmitis, 16 underwent conjunctival culture. These cultures correlated in only four cases, with two concomitant Streptococcus spp. In a third case, S. epidermidis and S. marcescens grew on conjunctival culture while yielding S. marcescens on vit­reous culture, as described in the Case Report. In a fourth case, S. epidermidis, Proteus spp., and Enterobacter spp. grew on conjunctival culture, whereas S. epidermidis grew on vitreous culture.

Discussion

The term "blebitis" was introduced to describe an infec­tion centered around the bleb without vitreous involve­ment by Brown et al.! These authors regarded blebitis as a limited form of endophthalmitis that could progress to more fulminant endophthalmitis, although progression to vitreous involvement was not noted in their series.! Bleb infection without vitreous involvement generally re­sponds well to topical and systemic antibiotic treatment, with a good visual outcome. I

,2 Our study confirms the good visual prognosis for blebitis, with return to preinfec­tion visual acuity in all cases for which visual acuity data were available. This study also confirms that blebitis can progress to definite endophthalmitis with a positive vitre­ous culture result, as noted in the Case Report.

The conjunctival cultures in our cases of blebitis showed predominantly Staphylococcus spp., similar to previous reports. I

,2 Intraocular culture data are limited for this condition because of its good response to topical and systemic antibiotic therapy. The relation between con­junctival cultures and intraocular cultures is unclear in bleb-associated endophthalmitis. Mandelbaum et aC noted a positive correlation between conjunctival and in­traocular cultures in only 5 (28%) of 18 cases. Our series was nearly identical in this regard, with positive correla­tion noted in 4 (25%) of 16 cases. However, Wolner et ae noted a positive correlation between conjunctival and intraocular cultures in six (100%) of six cases. Differences in culture technique could explain this discrepancy, but organisms causing endophthalmitis may only be present transiently on the bleb surface. 7

Unlike blebitis, the visual prognosis with bleb-associ­ated endophthalmitis is often poor.5- 7 Streptococcus spp. and gram-negative organisms such as H. injluenzae are isolated most frequently from late-onset endophthalmitis associated with filtering blebs.7,!4 This study confirms these findings, with Streptococcus spp. and gram-negative organisms comprising 48% of vitreous isolates from late­onset cases. The visual outcome of these cases generally was poor, with the majority of cases losing four or more lines of visual acuity. The virulence of these organisms may partially explain why these cases have poor outcomes despite appropriate intravitreous, systemic, and topical antibiotic therapy.15-18 Jett et al!9 have shown the impor­tance of bacterial toxins in ophthalmic infections and their role in treatment failure.

Although most reported cases of endophthalmitis asso­ciated with filtering blebs occur months to years after surgery, cases presenting in the initial perioperative pe­riod have been reported.3,8,9 Coagulase-negative staphylo­cocci are the most common cause of early postoperative endophthalmitis, including filtering surgery, which is markedly different from late-onset endophthalmitis.9 To explain the discrepancy, perioperative introduction of host flora is believed to be responsible for early cases of en­dophthalmitis, whereas late bleb-associated endophthal­mitis is likely caused by transconjunctival migration of

993

Ophthalmology Volume 104, Number 6, June 1997

transiently present bacteria, especially through thin­walled blebs and through bleb leaks.2.7- 9.2o-23 S. epider­midis was isolated in four of six cases of early endophthal­mitis associated with filtering surgery in this report, again confirming previous studies. However, two of these cases were combined cataract extraction and trabeculectomy surgeries, so that it is impossible to determine which por­tion of the procedure was responsible for the endophthal­mitis.

The large number of patients in this series with nega­tive vitreous cultures (45%), despite clinical evidence of endophthalmitis, is intriguing. Although Mandelbaum et al7 found negative culture results in only 17% of eyes, other studies have noted negative culture rates as high as 40% in bleb-associated endophthalrnitis.17 The high frequency of culture-negative endophthalmitis in this study could be explained by the presence of fastidious organisms, smaller numbers of organisms in earlier infec­tions, sterile inflammation, or improper handling of the vitreous sample.

Previous studies of bleb-associated endophthalmitis have included significant numbers of cases in which blebs inadvertently resulted after cataract surgery. In the study by Mandelbaum et at,7 22 of 36 patients had blebs created for control of glaucoma, and information about the type of filtration surgery could only be obtained in 14 of these patients. Inadvertent and deliberate blebs showed no sig­nificant difference in the type of infecting organism. In unplanned blebs, it is unknown whether the bleb resulted from a full-thickness or partial thickness wound, as well as the length of time the bleb existed before patient' s presentation with endophthalmitis. In the current study, there were no obvious differences in the responsible or­ganisms when cultures were compared from patients with full-thickness filters, trabeculectomies, or trabeculecto­mies with adjunctive mitomycin, although the sample size may be too small to detect some differences. This result lends some support to the idea that unplanned blebs, which have varying wound construction, show similar infecting organisms to those in planned blebs.

In summary, this study has several limitations, includ­ing its retrospective design, and different treating physi­cians over a lO-year period. However, the unique spec­trum of organisms in late-onset bleb-associated endoph­thalmitis is reaffirmed, and initial management of these cases should be based on anticipated culture results. The results of the Endophthalmitis Vitrectomy Study (EVS) are not directly applicable to bleb-associated endophthal­mitis because of the low percentage of Streptococcus spp. and gram-negative cases in the study.24 Our current regi­men of intravitreous antibiotics in bleb-associated en­dophthalmitis is identical to the EVS protocol, including vancomycin (1 mg) and amikacin (0.4 mg). Systemic treatment with intravenous vancomycin and ceftazidime merits consideration, despite the EVS finding that sys­temic treatment was of no benefit in postcataract endoph­thalmitis. Vancomycin penetrates the blood-ocular bar­rier and was not tested in the EVS.25 Ceftazidime provides

994

coverage for gram-negative organisms that were encoun­tered infrequently in the EVS.

This study also documents the progression of a case of blebitis to culture-positive endophthalmitis. The best treatment for blebitis is unknown. Because the organisms responsible for this infection may be unknown, treatment with broad-spectrum fortified topical antibiotics, supple­mented in most cases with systemic and subconjunctival antibiotics, has been used with good results. 1.2 The topical treatment regimen in this study also showed generally good outcomes, although one treatment failure was noted. Brown et all included intravenous antibiotics in the treat­ment regimen of 11 of 14 patients with blebitis. It seems reasonable to supplement frequent topical fortified antibi­otic therapy with oral ofloxacin, because this agent is readily available, well tolerated, penetrates the blood­ocular barrier well, and has broad coverage.26 Close ob­servation of blebitis cases for the development of vitritis is strongly recommended. Vitreous culture and injection of intravitreous antibiotics should be performed in those cases that do progress.

The incidence of late bleb-associated endophthalmitis may be increased by the use of 5-fluorouracil and mitomy­cin.2.14.27 Given the widespread use of these agents with filtering surgery, further study of the mechanisms and treatment of bleb-associated infection is warranted.

References

l. Brown RH, Yang LH, Walker SD, et al. Treatment of bleb infection after glaucoma surgery. Arch Ophthalmol 1994; 112:57 -6l.

2. Wolner B, Leibmann JM, Sassan JW, et al. Late bleb­related endophthalmitis after trabeculectomy with ad­junctive 5-fiuorouracil. Ophthalmology 1991;98:1053-60.

3. Katz LJ, Cantor LB, Spaeth GL. Complications of surgery in glaucoma: early and late bacterial endophthalrnitis after filtering surgery. Am J Ophthalmol 1985;92:959-63.

4. Tabbara KF. Late infections following filtering procedures. Ann OphthalmoI1976;8:1228-3l.

5. Freedman J, Gupta M, Bunke A. Endophthalrnitis after tra­beculectomy. Arch Ophthalmol 1978;96:1017-8.

6. Hattenhauer JM, Lipsich MP. Late endophthalmitis after filtering surgery. Am J Ophthalmol 1971;72:1097-1Ol.

7. Mandelbaum S, Forster RK, Gelender H, et al. Late onset endophthalrnitis associated with filtering blebs. Ophthal­mology 1985; 92:964-72.

8. Puliafito CA, Baker AS, Haaf J, et al. Infectious endoph­thalmitis. Review of 36 cases. Ophthalmology 1982; 89:921-9.

9. Ormerod LD, Becker LE, Cruise RI, et al. Endophthalrnitis caused by the coagulase-negative staphylococci. 1. Disease spectrum and outcome. Ophthalmology 1993; 100:715-23.

10. Forster RK. Etiology and diagnosis of bacterial postopera­tive endophthalrnitis. Ophthalmology 1978;85:320-6.

1l. Lennette EH, Belows A, Hausler WJ Jr, et al. Manual of Clinical Microbiology, 3rd ed. Washington, DC: American Society for Microbiology, 1980.

12. Committee on Continuing Education, Board of Education of the American Society of Microbiology. Identification of

Ciulla et al . Infection after Glaucoma-filtering Surgery

aerobic gram-positive and gram-negative cocci. Washing­ton DC: The Society, 1981.

13. National Committee on Clinical Laboratory Standards. Per­formance standards for antimicrobial disc susceptibility tests. Microbiol Rev 1985;49:126-39.

14. Greenfield DS, Suner 11, Miller MP, et al. Endophthalmitis after filtering surgery with mitomycin. Arch Ophthalmol 1996; 114:943-9.

15. Peyman GA, Herbst R. Bacterial endophthalmitis. Treat­ment with intraocular injection of gentamicin and dexa­methasone. Arch Ophthalmol 1974;91:416-8.

16. Peyman GA, Rose M, Sanders D. Intravitreal antibiotic injection and vitrectomy in acute bacterial endophthalmitis. Can J Ophthalmol 1976;11:188-90.

17. Forster R, Abbott R, Gelender H. Management of infectious endophthalmitis. Ophthalmology 1980;87:313-9.

18. Laatikainen L, Tarkkanen A. Early vitrectomy in the treat­ment of post-operative purulent endophthalmitis. Acta Oph­thalmol (Copenh) 1987;65:455-60.

19. Jett B, Jenson H, Atkuri R, et al. Evaluation of therapeutic measures for treating endophthalmitis caused by isogenic toxin producing and toxin nonproducing Enterococcus fae­cali strains. Invest Ophthalmol Vis Sci 1995; 36:9-15.

20. Ashkenazi I, Melamed S, Avni I, et al. Risk factors associ­ated with late infection of filtering blebs and endophthal­mitis. Ophthalmic Surg 1991;22:570-4.

21. Phillips WB, Wong TP, Bergren RL, et al. Late onset en­dophthalmitis associated with filtering blebs. Ophthalmic Surg 1994;25:88-9l.

22. Pillai S. Endophthalmitis following Scheie procedure. Glaucoma 1984;6:96-8.

23. Sugar HS, Zekman T. Late infection of filtering conjuncti­val scars. Am J Ophthalmol 1958;46:155-70.

24. Endophthalmitis vitrectomy study group. 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 1995; 113: 1479-96.

25. MacIlwaine W, Sande M, Mandell G. Penetration of anti­staphylococcal antibiotics into the human eye. Am J Oph­thalmol 1974;77:589-92.

26. Fisch A, Lafaix C, Salvanet A, et al. Ofloxacin in human aqueous humor and lens [letter]. J Antimicrob Chemother 1987; 20:453 -4.

27. Parrish R, Minckler D. "Late" endophthalmitis: the filter­ing surgery time bomb [editorial]. Ophthalmology 1996; 103:1167-8.

995