Endophthalmitis Caused by Corynebacterium Species: Clinical Features, Antibiotic Susceptibility, and Treatment Outcomes

Abstract

Purpose

To report the clinical features, antibiotic susceptibility profiles, treatment, and visual acuity (VA) outcomes of endophthalmitis caused by Corynebacterium species.

Design

Retrospective case series.

Subjects

Patients with endophthalmitis caused by Corynebacterium species.

Methods

Microbiology database records were retrospectively reviewed for all patients with endophthalmitis caused by Corynebacterium species from January 1, 1990 to December 31, 2012 at a large university referral center. The corresponding clinical records were then reviewed to evaluate the endophthalmitis clinical features and treatment outcomes.

Main Outcome Measures

presenting clinical features, visual acuity outcomes, and antibiotic susceptibility patterns.

Results

Of 10 patients identified, clinical settings included post-cataract surgery (n = 6), post-penetrating keratoplasty (n = 2), and post-trabeculectomy (n = 2). The mean time from surgical procedure to presentation with endophthalmitis was 6.8 months (range: 1 day to 28 months). All isolates were vancomycin susceptible. Presenting VA ranged from 7/200 to no light perception. Initial treatment strategies were vitreous tap and intravitreal antibiotic injection (n = 5) and pars plana vitrectomy with intravitreal antibiotic injection (n = 5). VA outcomes were ≥ 20/60 in 5 (50%) of 10 patients and ≤ 20/400 in 5 (50%) of 10 patients.

Conclusions

The most common clinical setting was post-cataract surgery. All isolates were susceptible to vancomycin. Despite prompt treatment with appropriate antibiotics, there were variable visual outcomes.

INTRODUCTION

Corynebacterium is genus of pleomorphic gram positive bacilli or coccobacilli that is ubiquitous in the environment.¹ Fifty-three Corynebacterium species have been identified as causes of human infections.^{2, 3} Corynebacterium species are a relatively rare cause of endophthalmitis and accounted for approximately 1% of culture positive acute-onset, post-cataract surgery endophthalmitis cases in the Endophthalmitis Vitrectomy Study.⁴

Although non-diphtheritic Corynebacterium (NDC) species were thought to be laboratory contaminants in the past, there are reports of NDC infections, including bacteremia, skin infections, urinary tract infections, endocarditis, osteomyelitis, septic arthritis, peritonitis, brain abscess, meningitis, and infections associated with prosthetic device.^{2, 3, 5–20} Corynebacterium species are found in the conjunctiva of five to seven percent of healthy adults.^{21, 22} Endophthalmitis caused by Corynebacterium species was first reported in 1979 and was associated with an intraocular metal foreign body.²³ Since then, there have been several case reports of endophthalmitis caused by Corynebacterium species associated with trauma, cataract surgery (acute- and delayed-onset), and endogenous sources.^24–37 Corynebacterium species have also been identified in cases of scleral buckle-associated infections.^38–40 The purpose of this study is to report the clinical settings, antibiotic susceptibility profiles, and visual acuity (VA) outcomes in series of culture-proven endophthalmitis due to Corynebacterium species in the United States.

MATERIALS AND METHODS

The study protocol for a retrospective review of medical and microbiology records for all patients treated at the Bascom Palmer Eye Institute with vitreous fluid culture-proven endophthalmitis caused by Corynebacterium species between January 1, 1990 and December 31, 2012 was approved by the Institutional Review Board of the University of Miami Miller School of Medicine Medical Sciences Subcommittee for the Protection of Human Subjects. Vitreous cultures were obtained either at the time of vitreous tap and inject or during vitrectomy as previously described.^{41, 42} All vitreous fluid samples were plated on 5% sheep blood and chocolate agars, which were incubated at 35°C f or a period of up to 2 weeks. Additional culture media, including thioglycollate broth, were inoculated at the discretion of the ophthalmologist performing the culture. All cultures were read and classified using standard culturing techniques by Ocular Microbiology Department staff. Patients who only grew Corynebacterium isolates in thioglycollate broth, but not other culture media, did not meet the criteria for a positive culture and were excluded from the study.

The medical records were retrospectively reviewed for all patients with vitreous fluid culture-proven endophthalmitis caused by Corynebacterium species during the study period to ensure clinical course consistent with endophthalmitis. Patient demographics, clinical characteristics, risk factors, treatment strategies, and clinical outcomes were assessed. The treatment strategies were determined by the individual treating physicians and did not follow a standardized protocol.

RESULTS

Patient demographics and clinical settings

Endophthalmitis caused by Corynebacterium species was identified in 10 eyes of 10 patients. The demographics and clinical setting for each case is summarized in Table 1. Of the 10 cases, six (60%) were male and five (50%) were right eyes. Mean age at presentation was 79 years (median: 79 years, range: 66 to 90 years). Clinical settings included six post-cataract surgery (60%), two post-penetrating keratoplasty (PKP, 20%), and two bleb-associated (20%) cases. The mean time from surgical procedure to presentation with endophthalmitis was 6.8 months (median: 19 days, range: 1 day to 28 months). One of the post-cataract surgery cases occurred 14 days after a laser capsulotomy procedure was performed for a dense posterior capsule plaque, eight months after the initial cataract surgery. One of the post-cataract surgery cases was due to a dehisced wound. One of the bleb-associated cases occurred one day after a bleb needling procedure was performed, seven years after the initial trabeculectomy surgery.

Table 1.

Clinical features of patients with endophthalmitis caused by Corynebacterium species.

No.	Age	Clinical Setting	Time After Surgery	Organism	Polymicrobial	Other organism
1	75	Phaco/IOL (uncomplicated)	21 days	Coryne species	No	N/A
2	90	Phaco/IOL (uncomplicated)	9 days	Coryne species	No	N/A
3	73	Phaco/IOL (complicated with RLF)	15months	Coryne xerors	Yes	Staph. epidermidus
4	85	Phaco/IOL (complicated with wound dehiscence)	8 days	Coryne xerosis	Yes	Staph. Aureus*
5	81	ECCE (uncomplicated)	17 days	Coryne species	Yes	Strep. Veridans
6	74	ECCE (uncomplicated)/YAG capsulotomy	8months/14 days	Coryne minutissimum	Yes	Staph. haemolyticus
7	88	Penetrating Keratoplasty	22months	Coryne pseudodiphtheriticum	No	N/A
8	66	Penetrating Keratoplasty	1 month	Coryne macginleyi	No	N/A
9	81	Trabeculectomy with 5-FU	28months	Coryne species	No	N/A
10	77	Trabeculectomy/Needling of bleb	7 years/1 day	Coryne species	No	N/A

5-FU = 5-fluorouricil, Coryne = Corynebacterium, ECCE = extracapsular cataract extraction, IOL = intraocular lens, N/A = non-applicable, Phaco = phacoemulsification, RLF = retained lens fragment, Staph = Staphylococcus, Strep = Streptococcus, YAG = Yttrium aluminium garnet,

^*This isolate was methicillin-resistant.

Microbiology and antibiotic resistance

Corynebacterium species was identified in vitreous samples in all (100%) of the 10 patients, six cultures (60%) were monomicrobial and four cultures (40%) were polymicrobial. The antibiotic susceptibilities of the Corynebacterium isolates are summarized in Table 2. All of the 10 isolates were susceptible to vancomycin. All tested isolates were susceptible to ceftazidime (2 of 2), imipenem (2 of 2), and penicillin (8 of 8). There was a high rate of susceptibility to gentamicin (6 of 7, 86%) and sulfamethoxazole/trimethoprim (6 of 8, 75%). Among the fluoroquinolones tested, isolates were more susceptible to ciprofloxacin (6 of 8, 75%) compared to gatifloxacin (1 of 2, 50%) and moxifloxacin (1 of 2, 50%). There was poor susceptibility to oxacillin (1 of 5, 20%).

Table 2.

Corynebacterium species antibiotic susceptibility.

Antibiotic	Number of Isolates Tested	Susceptible Isolates (%)
Vancomycin	10	10 (100)
Gentamicin	7	6 (86)
Clindamycin	5	3 (60)
Sulfamethoxazole/Trimethoprim	8	6 (75)
Beta-Lactams
Penicillin	8	8 (100)
Oxacillin	5	1 (20)
Ceftazidime	2	2 (100)
Imipenem	3	3 (100)
Fluoroquinolones
Ciprofloxacin	8	6 (75)
Gatifloxacin	2	1 (50)
Moxifloxacin	2	1 (50)

Clinical presentation and management

The initial and subsequent clinical management of patients are summarized in Table 3. The presenting VA was count fingers in three (30%), light perception (LP) in two (20%), hand motion (HM) in two (20%), 1 to 7/200 in two (20%), and no light perception (NLP) in one (10%) of 10 patients. Pain was present in seven (70%) of 10 patients. The mean intraocular pressure (IOP) was 18 mmHg (median: 16, range: 8 to 33). A hypopyon was present in 8 (80%) of 10 patients. A view of the posterior pole was unobtainable in all of the patients due to severe anterior and posterior segment inflammation and media opacities.

Table 3.

Presentation, treatment strategies, and outcomes of patients with endophthalmitis caused by Corynebacterium.

No	Pre- infection VA	Initial VA	IOP	Initial Tx	Initial Intravitreal Injection(s)	Additional Tx (Days After Initial Tx)	Additional Intravitreal Injections	Last VA	Follow-up Time
1	UK	CF	14	T+I	VANC + CTZ + DEX	None	None	20/40	12 mos
2	20/40	1/200	22	T+I	VANC + CTZ + DEX	PPV (6)	VANC + CTZ	20/60	9 mos
3	UK	7/200	16	PPV	VANC + CTZ + DEX	None	None	20/20	4 mos
4	UK	LP	-	T+I	VANC + CTZ + DEX	Enuc (17)	None	Enuc	42 mos
5	UK	HM	14	PPV	VANC + CTZ + DEX	None	None	20/40	30 mos
6	20/80	CF	8	PPV	VANC + DEX	None	None	20/400	1 day
7	NLP	NLP	-	T+I	VANC + CTZ + DEX	Enuc (44)	None	Enuc	2 mos
8	20/50	CF	16	T+I	VANC + CTZ	None	None	CF	10 mos
9	6/200	LP	-	PPV	VANC + CTZ + DEX	None	None	5/200	5 mos
10	20/30	HM	33	PPV	VANC + CTZ + DEX	None	None	20/30	1 mos

CF = count fingers, CTZ = ceftazidime, DEX = dexamethasone, Enuc = enucleation, HM = hand motion, LP = light perception, mos = months, NLP = no light perception, No. = number, PPV = pars plana vitrectomy, T + I = vitreous tap + intravitreal injection, Tx = treatment, VAN = vancomycin, VA = visual acuity, UK = unknown.

Initial treatment consisted of a vitreous tap and intravitreal antibiotics in five (50%) of 10 patients and pars plana vitrectomy (PPV) and intravitreal antibiotics in five (50%). One (10%) of 10 patients underwent a PPV and intravitreal antibiotic injection six days after initial treatment (vitreous tap and injection of antibiotics) due to persistent vitreous opacities, increased height of the hypopyon, and persistent pain.

Vancomycin was used initially for intravitreal antibiotic treatment in all patients. Nine (90%) of 10 patients received intravitreal ceftazidime initially. Additionally, 9 (90%) of 10 patients were treated with intravitreal dexamethasone as part of their initial treatment. All patients were started on topical antibiotic drops: 9 (90%) of 10 on fortified vancomycin (50 mg/ml) and a second antibiotic (fortified tobramycin (14 mg/ml), fortified gentamicin (14 mg/ml), fortified ceftazidime (50 mg/ml), moxifloxacin, or polymyxin B/trimethoprim) and one (10%) of 10 patients on fortified gentamicin alone. A topical steroid drop was started within 48 hours of the initial treatment in all 10 patients.

Clinical outcomes

Clinical outcomes are summarized in Table 3. Enucleation was performed due to blind painful eyes in two (20%) of 10 patients. The VA outcome was 20/60 or better in five (50%) of 10 patients and 20/400 or worse in 5 (50%) of 10 patients (including the two enucleated eyes). The mean follow-up period was 11.5 months (range: 7 months, range: 1 day to 42 months).

DISCUSSION

The current study demonstrates that there are variable VA outcomes in patients with endophthalmitis due to Corynebacterium species, despite prompt and appropriate intravitreal antibiotic treatment. Although there have been reports of vancomycin resistance in Corynebacterium species isolates from non-ocular infections, all of the isolates in the current series were sensitive to vancomycin.^{43, 44} All of the isolates tested were sensitive to imipenem and 75% (6 of 8) were sensitive to ciprofloxacin, the antibiotics used in the reported cases of vancomycin-resistant Corynebacterium species.

The susceptibility pattern of the isolates in our study is similar to those reported by Joseph et al. in a series of 16 cases of endophthalmitis caused by Corynebacterium in India.⁴⁵ In that report, which excluded polymicrobial cases, all of the isolates were susceptible to vancomycin, and as in our study, there was a high rate of susceptibility to cirpofloxacin (Table 4).⁴⁵ A notable difference, however, is the high rate of susceptibility to ceftazidime (2 of 2, 100%) in the current series, compared to the Joseph et al. study (1 or 10, 10%).

Table 4.

Comparison of Endophthalmitis due to Corynebacterium species studies

	Current Study	Joseph et al.LV Prasad Eye Institute45
Antibiotic	No. of Patients (%)	No. of Patients (%)
Clinical setting
Post-cataract surgery	6 (60%)	5 (31%)
Post-PKP surgery	2 (20%)	1 (6%)
Post-Glaucoma procedures	2 (20%)	-
Trauma	-	10 (63%)
Initial treatment
Vitreous tap + antibiotics	5 (50%)	-
PPV + antibiotics	5 (50%)	16a (100%)
Presenting VA
LP or Worse	3 (30%)	7 (44%)
HM or Better	7 (70%)	9 (56%)
Final VA
≥20/400	6 (60%)	11 (69%)
<20/400	4 (40%)	5 (31%)
Antibiotic Susceptibility
Vancomycin	10/10 (100%)	16/16 (100%)
Gentamicin	6/7 (86%)	11/14 (79%)
Ceftazidime	2/2 (100%)	1/10 (10%)
Fluoroquinolones
Ciprofloxacin	6/8 (75%)	12/16 (75%)
Gatifloxacin	1/2 (50%)	15/16 (94%)
Moxifloxacin	1/2 (50%)	7/9 (78%)

HM = hand motion, LP = light perception, PKP = penetrating keratoplasty, No. = number, PPV = pars plana vitrectomy, VA = visual acuity, UK = unknown,

^aIncludes nine patients treated with PPV and intravitreal antibiotics as well as seven patients treated with pars plana lensectomy along with pars plana vitrectomy and intravitreal antibiotics.

Susceptibility of all Corynebacterium isolates to vancomycin is consistent with most studies on endophthalmitis; a study by Gentile et. al. found that among 727 gram-positive isolates, all but two (Enterococcus faecium and Nocardia exalbida, 99.7%) were susceptible to vancomycin.²⁸ A high rate of gram-positive susceptibility to third generation cephalosporins (147 of 156, 94.2%) was also noted in the Gentile et. al. study, similar to susceptibility of Corynebacterium isolates in the current study (2 of 2, 100%).²⁸ The current study demonstrated similar rates of Corynebacterium susceptibility to ciprofloxacin (6 of 8, 75%) as other gram-positive isolates (331 of 466, 71%).²⁸ There was a higher rate of gentamicin sensitivity among Corynebacterium isolates (6 of 7, 86%) in the current study compared to gram positive isolates (448 of 599, 74.8%) in the Gentile et. al. study.²⁸

The most common etiology in the current series is post-cataract surgery, compared to trauma in the Joseph et al. series.⁴⁵ In the current series, only one (10%) of 10 patients presented ≤ 7 days after the associated surgery. One patient developed endophthalmitis 14 days after a laser capsulotomy procedure performed eight months after cataract surgery. Corynebacterium has been previously described to be associated with endophthalmitis after YAG capsulotomy and has been demonstrated to stay sequestered within the capsular bag and even form a capsular hypopyon.^{31, 32, 35} None of the patients in the current series required intraocular lens removal or capsulectomy, which is commonly required when treating endophthalmitis caused by Propionbacterium acnes, the most common cause of delayed-onset endophthalmitis.⁴⁶ Although persistently positive vitreous cultures in endophthalmitis caused by Corynebacterium species have been reported, there were no cases of persistently positive vitreous cultures in the current series.⁴⁷

Six of (60%) of 10 patients with endophthalmitis due to Corynebacterium species in the current series had VA outcomes ≥ 20/400. The Joseph et al. series also had similarly favorable visual outcomes with 11 (69%) of 16 patients achieving VA outcomes ≥ 20/400.⁴⁵ In contrast, only 59 (14%) of 420 in the Endophthalmitis Vitrectomy Study with VA outcomes ≥ 5/200. This suggests that Corynebacterium species are less virulent than other organisms.

There was a higher rate of polymicrobial cases in the current series (4 of 10, 40%), compared to the Endophthalmitis Vitrectomy Study (27 of 291, 9%)⁴ All of the polymicrobial cases occurred in post-cataract surgery patients. Growth of Corynebacterium in multiple plates in these patients demonstrates that bacteria were not merely contaminants. On culture and smear, mycobacteria may be confused for Corynebacterium and may require subsequent DNA analysis.⁴⁸ Of note, DNA sequencing was not required to identify any of the isolates in this series. A study by Uehara and colleagues demonstrated that Corynebacterium species was able to inhibit colonization of nasal cavities by Staphylococcus aureus through bacterial interference.^49,50 Bacterial interference is a term used to describe the antagonism between bacterial species during the process of nutrient acquisition and surface colonization.⁴⁹ Bacterial interference may have contributed to VA outcomes ≥ 20/400 in three (75%) of four polymicrobial cases.

The limitations of the current study include its retrospective design, relatively small number of patients, and use of positive vitreous cultures using standard culture techniques as the inclusion criteria for the study, which could potentially have excluded cases with false negative cultures. The use of polymerase chain reaction-based identification methods for Corynebacterium species have been described, but were not used in this study.²⁰ Despite these limitations, this study provides important prognostic and antibiotic resistance data for endophthalmitis caused by Corynebacterium species.

In conclusion, despite prompt treatment with intravitreal vancomycin, patients in the current study had variable VA outcomes. Corynebacterium species can cause a delayed-onset endophthalmitis. The antibiotic susceptibility data from the current study further supports continued use of vancomycin.