Prophylaxis of Acute Posttraumatic Bacterial Endophthalmitis

International Ophthalmology 24: 323–330, 2001. © 2003 Kluwer Academic Publishers. Printed in the Netherlands.

323

Prophylaxis of acute posttraumatic bacterial endophthalmitis with or without combined intraocular antibiotics: a prospective, double-masked randomized pilot study Masoud Soheilian1 , Nasrin Rafati1 & Gholam A. Peyman2 1 Ophthalmology

Department, Shahid Beheshti University of Medical Sciences, Labbafinejad Medical Center, Tehran, Iran; 2 Ophthalmology Department, School of Medicine, Tulane University Health Sciences Center, New Orleans, LA, USA

Received 18 May 2001; accepted 25 May 2003

Key words: antibiotics, bacterial endophthalmitis, intraocular injection, trauma

Abstract Purpose: The effectiveness of an intraocular injection of combined gentamicin and clindamycin in the prevention of acute posttraumatic bacterial endophthalmitis following penetrating ocular injuries was evaluated in a prospective, double-masked, randomized pilot study. Methods: Sixty eyes of 60 patients with penetrating ocular injuries were treated at a tertiary care hospital. Following primary repair, the eyes were randomized in two groups. Group 1, the antibiotic injection group (cases), was given an intracameral or intravitreal injection of 0.1 mL antibiotic (40 µg gentamicin and 45 µg clindamycin). Group 2 (balanced saline solution [BSS] injection group [controls]) received intracameral or intravitreal injection of 0.1 mL BSS. All patients received standard prophylactic antibiotic therapy (systemic, subconjunctival, and topical). Result: Although the overall incidence of acute posttraumatic bacterial endophthalmitis was 6.6% (4 eyes), the results of three cultures were negative. All endophthalmitis cases occurred in the BSS injection group; however, there was no statistically significant difference between case and control groups (p = 0.11). The incidence rate for those with retained intraocular foreign bodies was 13.3% and for those without foreign bodies was 4.4%. No retinal toxicity was detected. Conclusion: Intraocular injection of gentamicin and clindamycin in addition to the other methods of prophylaxis may be an effective modality in the prevention of posttraumatic endophthalmitis. Early results suggest that these antibiotics may have a role as adjunct therapy to primary repair of injured globes without significant side effects at the dosage used. Abbreviations: APD – afferent pupillary defect; BSS – balanced salt solution; IOFB – intraocular foreign body.

Introduction Endophthalmitis after penetrating ocular injuries is considered to have a critical prognosis. Its incidence is reported to be from 3.3% to 30% in different series and is highly dependent on the setting and circumstances of trauma [1, 2]. Gram-positive bacteria such as Staphylococcus, Streptococcus, and gram-negative organisms are more common as causative agents. It has been shown that mixed infections may occur in as many as 42% of injured eyes [1]. The most troubling organism in posttraumatic infections is Bacillus, which is ubiquitous and lives particularly in soil. It

has been isolated in approximately 20% to 46% of posttraumatic cases and is most frequently isolated from eyes with a retained intraocular foreign body (IOFB) [3, 4]. The visual prognosis of treated eyes with acute posttraumatic bacterial endophthalmitis is generally poor; only a small percentage of eyes regain ambulatory visual acuity [4–6]. Therefore, preventive strategies have attracted considerable attention in recent years. The value of administering an intraocular antibiotic as prophylaxis following penetrating trauma remains to be determined.

324 In this prospective, double-masked, randomized pilot study, we have evaluated the effectiveness of the intraocular injection of antibiotics (gentamicin and clindamycin) when combined with other prophylactic methods, including systemic, topical, and subconjunctival routes of administration in the prevention of acute posttraumatic bacterial endophthalmitis following penetrating eye injury.

Patients and methods A prospective, double-masked, randomized clinical trial was conducted in our department between 1997– 1999. The patient’s eligibility was ascertained and informed consent was obtained. All patients had a complete physical examination of the eye. The patient was taken to the operating room and after primary repair of the injured globe, was randomly assigned to either the antibiotic injection group, given intracameral or intravitreal injection of 0.1 mL containing 40 µg gentamicin and 45 µg clindamycin (Table 1); or the balanced saline solution (BSS) injection group, given intracameral or intravitreal injection of 0.1 mL BSS (Table 2). The eyes were also classified according to estimated red reflex: poor, less than 20/200; fair, between 20/200 and 20/60; or good, better than 20/50. The degree of cell and flare in the anterior chamber and vitreous was determined according to Hogan and Kimura scores on a scale from 0 to 4 [7]. An orbital X-ray was taken of all patients as well as a high-resolution orbital computed tomography scan for those with retained intraocular foreign bodies (IOFB). All patients were admitted and intravenous gentamicin (3–5 mg/kg) was administered every 8 hours along with cefazolin (50 mg/kg) every 6 hours and was maintained for at least 5 days following surgery. Intracameral injection was indicated when all lacerations were limited to the area anterior to the recti muscle insertion with an intact lens capsule or lacerations that extended beyond the recti muscle. Intravitreal injection was performed for all lacerations that extended beyond the rectus muscle insertion as well as scleral lacerations that were limited posterior to recti muscle insertion.

Surgical technique All cases underwent surgery using general anesthesia. Corneal and scleral lacerations were repaired in the usual manner. At the conclusion of primary repair, the eyes were randomly assigned to either antibiotic or BSS injection. The aqueous and vitreous taps and antibiotic injection were performed. For intravitreal injection, a microvitreoretinal blade was used to enter through the sclera 4 mm from the limbus in phakic eyes, or 3.5 mm from the limbus in aphakic eyes; 0.2 mL of vitreous was aspirated with a 23-gauge needle on a tuberculin syringe with the needle pointing to the midvitreous. The intravitreal antibiotic or BSS was injected slowly into the anterior vitreous at the same site using a 27gauge needle with the bevel facing up. The sclerotomy was then closed. Available vitreous and aqueous specimens were immediately handled for identification of organisms and inoculated on blood and chocolate agar plates, in Sabourad’s medium, thioglycolate medium, and on glass slides for Gram’s and Giemsa stains. At the end of the operation, 20 mg gentamicin and 4 mg betamethasone were injected subconjunctivally. Postoperatively, topical gentamicin was administered four times daily for 1 week; topical betamethasone was used six times daily for the first week, four times daily for the second week, and then gradually stopped. Topical atropine was also given three times daily. Oral prednisolone (1 mg/kg/day) each morning was started 24 hours postoperatively and gradually tapered after 1 week, depending on the amount of inflammation. The patients were followed every day for the first 3 days and on days 5, 7, and 14 by one physician who was masked to the intraocular treatment used. A complete ophthalmic examination was performed at each visit to detect the amount of anterior chamber and vitreous reaction and haziness. Additionally, echography was performed in those cases with poor visibility of the posterior segment 24 hours after repair to detect posterior segment complications. Clinical endophthalmitis was defined as any of the following: lid swelling with pain and +3 or more cells in the anterior chamber accompanied by decreased visual acuity and loss of red reflex in cases without lens opacifications; lid swelling, pain, and +3 or more cells in the anterior chamber accompanied by decreased visual acuity and vitreous involvement detected by echography in cases associated with lens opacity; lid swelling, pain, decreased visual acuity,

11/M

17/M 54/M 27/M 20/M

21/M

20/M 24/M 29/M 10/M 16/M 34/M 14/M 35/M 18/M 10/M 10/M 35/M 31/M 23/M

19/F 13/M 25/M 15/M 6/M 14/M

33/M 7/M

3

4 5 6 7

8

9 10 11 12 13 14 15 16 17 18 19 20 21 22

23 24 25 26 27 28

29 30

OS OS

OD OS OS OD OD OD

OD OS OS OS OS OS OD OS OD OD OS OD OD OD

OD

OD OS OD OS

OS

OD OD

Eye

20/40 LP

HM FC 20/40 HM FC 20/160

LP 20/50 LP 20/80 20/80 20/50 20/40 HM 20/160 HM LP 20/160 LP HM

20/80

LP HM 20/25 HM

20/80

20/200 FC

Initial VA

P G P F† F† G G P F† P P F† P P F† P G P P F† G P

+ − + − − − − − − − + − + + − − − − + − − +

F†

−

F†

−

P P G P

F† P

− −

+ − − −

Red reflex

APD

C+S C+S

C+S S C C S C

C+S C+S S S C C C C C C S C S C

C

S C+S S C

C

C C

Location of wound

+ ?

? ? − ? + −

? − + + + + − ? ? ? ? ? ? ?

−

? ? − ?

−

− ?

Posterior segment involvement

VH VH+Cat.

VH+RD VH+RD

Cat. Cat. Cat. VH+RD Inflammation Cat. VH+ hyphema Cat.+VH VH

RD+VH VH VH VH

VH+RD+Cat.

Cat.+VH

VH+RD+Cat. Cat.

Cat.+ inflammation

Cause of posterior segment involvement

− +

+ − − − − + + + + + − + + + Aphakia + − − + − + 1 1

1 1 1 1 1 1

1 1 1–5 1 1–5 1 1 1 1 1 1 1 1 1

1

1–5 1 1 1

− + − + −

1–5

1–5 1

Wound repair (day)

+

− +

Lens injury

− −

− − + − − −

+ + + − + − − − − − + + − −

−

+ − − −

−

− −

IOFB

− −

+ + − − + +

+ − + + − + − + − − + − + +

−

+ + − +

−

− −

Vitreous prolapse

Anterior Anterior/ posterior

Anterior Posterior Anterior Anterior Posterior Anterior

Anterior Anterior Posterior Posterior Anterior Anterior Anterior Anterior Anterior Anterior Posterior Anterior Posterior Anterior

Anterior

Posterior Anterior Posterior Anterior

Anterior

Anterior Anterior

Relation to rectus muscle

5

>5 >5 5

>5 >5 5 5

5

Wound length (mm)

IC IV

IV IV IC IV IV IV

IV IC IV IV IC IV IV IV IV IV IV IV IV IC

IC

IV IV IV IV

IV

IC IV

Inj. site

− V-R

V-R V-R − Lenx V-R −

Enucleation − Enucleation V-R − − − Lenx Lenx Lenx V-R − V-R −

−

Enucleation V-R − −

Lenx

− Lenx

Additional surgery

FC HA NLP 20/120 20/25 FC Cat. 20/200 HA∗ NLP 20/60 LP 20/60 20/60 20/30 20/30 20/120 20/30 20/40 LP 20/80 HM 20/120 Amblyopia HM 20/60 20/25 20/40 20/80 20/80 Cat. 20/25 FC

20/80 20/60

Final VA

M = male, F = female, VA = visual acuity, FC = finger count, LP = light perception, HM = hand motion, APD = afferent papillary defect, F† = fair, P = poor, G = good, C = cornea, S = sclera, ? = undetectable, VH = vitreous hemorrhage, RD = retinal detachment, HA = high astigmatism, OA = optic atrophy, IC = intracameral, IV = intravitreal, Lenz = anterior lensectomy and vitrectomy + PCIOL, Cat. = cataract, VR = vitreoretinal surgery, Inj. = injection.

57/M 23/M

Age/sex (year)

1 2

Patient No.

Table 1. Clinical characteristics of 30 eyes with intraocular antibiotic injection (cases) – Group I

325

5/F

14/M

15/M 22/M 14/M 16/M 44/M 64/M

26/M 31/M 10/F 30/M 29/M 21/M

10/M 12/M

12/M 17/M 14/F 25/M 33/M 34/M

32/M 7/F 18/M

30/M 41/M

19/M 24/M

18/M

1

2

3 4 5 6 7 8

9 10 11 12 13 14

15 16

17 18 19 20 21 22

23 24 25

26 27

28 29

30

OD

OS OS

OD OD

OS OD OS

OS OD OS OS OD OD

OD OS

OS OS OD OD OD OS

OS OD OS OD OD OD

OS

OD

Eye

20/80

20/60 HM

20/30 HM

20/60 20/30 FC

FC FC FC 20/50 20/160 LP

20/20 LP

20/160 FC FC FC 20/30 20/160

FC FC FC FC 20/25 FC

20/160

FC

Initial VA

Red reflex

F† F† P P P P G P F† P P P G F† G P P P P G F† P F† G P G P F† P F†

APD

− − − − − − − − − − − − − ? − ? − − − − − + − − − − − − + −

S

C S

C C+S

S C C

C C+S C C S C+S

C C+S

C C+S C C C+S C+S

C C C C+S C+S C+S

C+S

C+S

Location of wound

−

− ?

− ?

+ − ?

+ ? − − + +

− ?

− − ? ? − −

? ? ? ? − ?

+

?

Posterior segment involvement

Cat.+VH

Cat.+VH

Inflammation

RD+VH

− VH+ hyphema VH Cat. − − RD+VH RD+VH

Cat. Cat. Cat. Cat. − Inflammation +Cat. − − Cat. Cat. − −

Poor cooperation VH

Cause of posterior segment involvement

+

+ +

− +

− − −

+ + − + − +

+ +

− − + + − −

+ + + − − +

−

−

Lens injury

1

1 1

1 1

1 1 1–5

1 1 1 1 1 1

1 1

1 1 1 1 1 1

1 1 1 1 1 1

1

1–5

Injury wound repair interval (day)

−

− −

− −

− − −

− − − + − −

+ +

+ + − − + −

− − − − + −

−

−

IOFB

+

+ −

+ −

+ − −

+ + − + + +

+ +

− − + − − −

+ − + − − +

+

−

Vitreous prolapse

Posterior

Anterior Posterior

Anterior Anterior

Anterior Anterior Anterior Anterior Posterior Anterior+ posterior Posterior Anterior Anterior

Anterior Anterior Anterior Anterior Anterior Anterior+ posterior Anterior Anterior

Anterior+ posterior Anterior Anterior Anterior Anterior Anterior Anterior

Anterior

Relation to rectus muscle

5 5 5

5

5 5

5 >5 >5 >5 >5

>5

>5

Wound length (mm)

IV

IV IV

IC IV

IV IC IC

IV IV IC IV IV IV

IV IV

IC IC IV IV IC IV

IV IV IV IC IC IV

IV

IC

Inj. site

V-R

Lenx V-R

− V-R

V-R − −

− V-R∗∗ − V-R∗∗ V-R V-R

Lenx V-R

− − Lenx Lenx − V-R∗∗

Lenx Lenx Lenx Lenx V-R.∗∗ V-R

−

−

Additional surgery

20/120 20/20 20/200 HA‡ 20/30 HM VH 20/60 FC HA‡ 20/60

20/25 LP OA† FC NLP FC 20/80 20/200 FC

20/40 20/40 20/30 20/60 20/30 20/200

20/50 20/30 20/120 20/50 20/40 20/80

20/50

20/200

Final VA

M = male, F = female, VA = visual acuity, FC = finger count, LP = light perception, HM = hand motion, APD = afferent papillary defect, F† = fair, P = poor, G = good, C = cornea, S = sclera, ? = undetectable, Cat. = cataract, VH = vitreous hemorrhage, RD = retinal detachment, OA† = optic atrophy, HA‡ = high astigmatism, IC = intracameral, IV = intravitreal, Lenx = anterior lensectomy and victrectomy + PCIOL, IC = intracameral, IV = intravitreal, VR = vitreoretinal surgery, ∗∗ = cases who developed endophthalmitis.

Age/sex (year)

Patient No.

Table 2. Clinical characteristics of 30 eyes with intraocular BSS injection (controls) – Group II

326

327 and possibly a positive culture of harvested intraocular fluids. The clinical diagnosis of endophthalmitis was confirmed by two other independent, masked physicians. Statistical analysis of patient data was performed using SAS statistical software. The data was analyzed using the chi-square, two-tailed Fisher exact tests, and the student T-test. Results The study included 60 eyes of 60 patients. The mean patient age was 22.63 years (± 12.34). There was no statistically significant difference in age between the two groups. Fifty-four patients (90%) were males (the male-to-female ratio was 9 to 1). The right eye was affected in 32 (53%) and the left eye in 28 eyes (47%). Intraocular injection of either antibiotic or BSS was performed intracamerally in 17 eyes (28%) and intravitreally in 43 eyes (72%). Initial visual acuity was light perception only in 13.3% of patients, hand motions in 15%, and counting fingers in 25%. However, visual acuity was better than 20/200 in 46.6%. An afferent pupillary defect (APD) was present in 11 eyes (18.3 %), was not found in 47 eyes (78.3%), and was not obtainable in 2 eyes (3.3%). There was no statistically significant difference between the two groups in the presence of APD. The red reflex was poor in 31 eyes (51.6%), fair in 17 eyes (28.3%), and good in 12 eyes (20%). There was no statistically significant relationship between the initial estimated red reflex and final development of endophthalmitis. The results of cultures of intraocular fluids harvested at the time of primary wound repair demonstrated Staphylococcus epidermidis in 5 eyes, Corynebacterium diphtheria in 2 eyes, and Candida albicans in 1 eye; the rest were negative. Development of endophthalmitis Endophthalmitis developed in 4 eyes (6.6%) in the BSS injection group (controls) and no case of endophthalmitis was detected in the antibiotic injection group (cases). However, this difference is not statistically significant (p = 0.11). The incidence of endophthalmitis in patients with retained IOFB was higher (2 of 15 eyes, 13.3%) compared with those without retained IOFB (2 of 45 eyes, 4.4%). Microbiologic analysis of the vitreous aspirate and of the anterior chamber fluid obtained during a second procedure (pars plana vitrectomy) in all four patients with acute posttraumatic

bacterial endophthalmitis showed Staphylococcus epidermidis in only one patient; this patient had an initial positive culture with this microorganism taken at the time of primary repair also. For the three other patients, microbiologic examinations of the vitreous and anterior chamber fluid were unremarkable. Posterior segment injuries Before primary repair, overall posterior segment involvement, including vitreous hemorrhage and retinal detachment, was identified in 11 eyes (18.3%). In 20 eyes (33.3%) with clear media, the posterior segment was intact. Haziness of the media and/or the presence of cataract precluded a view of the posterior segment in 29 eyes (48.3%). No significant difference in posterior segment involvement could be shown between the groups. Posterior segment involvement, such as vitreous hemorrhage and retinal detachment, was detected in 23 of 60 eyes (38%) by clinical examination and echography (including those 11 eyes with visible posterior segment preoperatively). The postoperative posterior segment complications in the historical control group were also observed in 18 out of 50 eyes (36%) and the difference was not statistically significant. Associated clinical findings The effects of the following factors were studied and no statistically significant relation between the presence of IOFB, lens injury, vitreous prolapse, timing of wound repair, and the length and site of initial laceration could be detected. However, the final visual prognosis for eyes with corneal laceration tended to be better (Table 3). Additional surgeries Overall in 33 eyes (55%), secondary surgeries (lensectomy, vitreoretinal surgeries for IOFB removal, or enucleation) had to be performed before day 15 after the repair. The rate for the BSS injection group was 53% and for the antibiotic injection group was 57%. Three eyes had to be enucleated and all were in the antibiotic injection group. The enucleation was related to the extent of the surgical trauma and was not the result of infection. In these three eyes, initial visual acuity was light perception, APD was present, red reflex was poor, the vitreous was prolapsed in the laceration, and an IOFB was retained. Enucleation took place before day 15 after the primary repair.

328 Table 3. The summary of clinical findings associated with eye injuries in case and control groups Findings

Antibiotic injection group

BSS injection group

Number %

Number %

7 (12%)

8 (26%)