- Email: [email protected]
Intravenous Drug Use–Associated Endophthalmitis
Intravenous Drug UseeAssociated Endophthalmitis Bobeck S. Modjtahedi, MD,1,* Avni V. Finn, MD,1,* Thanos D. Papakostas, MD,1 Marlene Durand, MD,2,3 Deeba Husain, MD,1 Dean Eliott, MD1 Purpose: To characterize features and outcomes of intravenous drug use (IVDU)-associated endophthalmitis. Design: Retrospective cross-sectional study. Methods: A retrospective chart review of all cases of endophthalmitis seen between September 2006 and November 2014 at a single academic referral center was performed. All cases of IVDU-associated endophthalmitis were identified and characterized. Main Outcome Measures: Ophthalmic examination findings, microbial results, visual outcomes, and secondary ocular sequelae. Results: Thirty patients (32 eyes) with IVDU-associated endophthalmitis were identified, which represented 9% of all endophthalmitis patients (n ¼ 338) and 44% of all endogenous endophthalmitis patients (n ¼ 68). Mean follow-up was 11 months. All eyes had vitritis, 6 had hypopyon, and 2 had subretinal abscesses. Twenty eyes had macular involvement, 8 eyes had no macular involvement, and media opacities precluded macular assessment in 4 eyes. Initial treatment was needle vitreous biopsy with intravitreal antibiotics (“tap and inject”) in 25 eyes (78%) and pars plana vitrectomy (PPV) in 6 eyes (19%); 1 patient refused ocular treatment. An organism was identified from at least 1 source in 75% of eyes (24/32): 59% fungal, 16% bacterial, 22% negative cultures, and 3% refused ocular cultures. Mean visual acuity improved significantly between initial examination and final follow-up (1.64 logMAR to 0.91 logMAR, P < 0.0001). At final follow-up, 90% of eyes had improved vision compared with presentation, 31% of eyes had 20/40 or better vision, and 25% of eyes had 20/200 or worse vision. Twenty-one eyes (66%) required PPV for their infectiond6 initially and 15 secondarily after tap and inject. Sixty-nine percent of eyes (9/13) that had cultures sent from a secondary PPV had positive cultures, despite almost all receiving appropriate intravitreal antibiotic therapy at the time of the tap and inject. Eight patients (27%) had extraocular signs of infection. Twenty eyes (63%) suffered secondary ocular sequelae. Conclusions: This represents the largest series of IVDU-associated endophthalmitis. Bacterial etiologies constitute an important share of cases. A majority of patients eventually required PPV and there was frequent culture positivity even after tap and inject with appropriate antibiotics; therefore, early PPV may have an important role in management. Ophthalmology Retina 2016;-:1e8 ª 2016 by the American Academy of Ophthalmology
Endophthalmitis can be divided into exogenous (postsurgical, posttraumatic, or from a contiguous infection such as keratitis) or endogenous (ocular seeding from a hematogenous infection).1 Endogenous endophthalmitis is encountered either in patients with an ongoing extraocular source of infection or in patients with transient bacteremia or fungemia, such as after an endoscopic procedure or after injection drug use. Although some patients with endogenous endophthalmitis have no apparent risk factors, such an absence is rare.2e5 Most reported cases of endogenous endophthalmitis are in systemically ill or immunocompromised patients. The incidence of endogenous fungal endophthalmitis (EFE) may be declining among hospitalized patients, likely as a result of earlier detection of infection and improved vitreous penetration of systemic antifungals.6 Endogenous endophthalmitis from intravenous drug use (IVDU) would not be expected to be as responsive to the above trend because of the transient nature of septicemia, the lack of 2016 by the American Academy of Ophthalmology Published by Elsevier Inc.
prophylactic treatment, and the fact that patients often do not present until they develop ocular symptoms. IVDU was seen in 14% to 30% of cases of endogenous endophthalmitis in series from Miami, Florida, India, and Australia,5,7e11 but it is seldom described in East Asian series.12e15 The recent surge in heroin use in the United States is a matter of considerable public health concern and necessitates a better understanding of the consequences of IVDU, which include endophthalmitis.16,17 This study examines endophthalmitis associated with IVDU at a large tertiary ophthalmology center in Boston, Massachusetts, and constitutes the largest series to date. We aimed to explore demographics, clinical presentation, microbial etiology, treatments, and outcomes of this condition.
Methods Institutional Review Board approval was obtained by the Human Studies Committee and the study protocol adhered to the http://dx.doi.org/10.1016/j.oret.2016.10.013 ISSN 2468-6530/16
1
Ophthalmology Retina Volume -, Number -, Month 2016 Declaration of Helsinki. International Classification of Diseases 9 (ICD-9) billing codes were used to identify cases of endophthalmitis at the Massachusetts Eye and Ear Infirmary (MEEI) from September 2006 until November 2014. Every medical record was reviewed and patients with a preceding history of IVDU were identified. Inclusion criteria included a history of endophthalmitis, a preceding history of IVDU, no preceding history of ocular surgery/ trauma, and no other noneIVDU-related plausible source of infection. Patient demographics, past ocular history, symptoms (divided into ocular, extraocular, and systemic), intravenous drug used, initial type of physician to evaluate patient, systemic evaluations, visual acuity and complete ocular examination, culture results, treatment, ocular complications, and duration of follow-up were all evaluated. Visual acuity data were analyzed after converting documented Snellen acuity to the equivalent logMAR value based on previously published conversion tables.18,19 Patients with a history of endogenous endophthalmitis not from IVDU were analyzed as well for comparison. Vision data were checked for normality with the D’Agostino-Pearson omnibus normality test and were found to be distributed in a non-Gaussian fashion. The Wilcoxon signed rank test was used for statistical companions within the IVDU and non-IVDU groups. Within each of these groups adjustments for multiple comparisons were not made. For the comparison between IVDU and non-IVDU groups a Mann-Whitney test was used. When looking at all cases of endogenous endophthalmitis (IVDU and non-IVDU together), a Kruskal-Wallis with the Dunn post hoc test for multiple comparisons was used. The level of statistical significance was set at 0.05.
Results Patient Characteristics A total of 338 patients with endophthalmitis were seen during the study period, of which 68 had endogenous endophthalmitis (20%), and 30 patients (32 eyes) were identified as having IVDU-associated endophthalmitis (9% of total, 44% of endogenous cases). There were 36 patients (36 eyes) with non-IVDU endogenous endophthalmitis. One patient did not have visual acuity data available for analysis because presentation preceded electronic medical records. Mean follow-up time was 11 months (range, 1 day to 67 months). The patients’ baseline features are summarized in Table 1. All of the patients admitted to a history of IVDU. Seventeen patients reported heroin use, 5 reported intravenous cocaine use, 1 reported intravenous hydromorphone use, 1 reported polysubstance use, 1 reported intravenous testosterone use, 1 reported intravenous methadone use, and 1 reported crushing and injecting suboxone (buprenorphine and naloxone). For the remaining 3 patients, the drug injected was not recorded or was unknown.
Table 1. Intravenous Drug UseeAssociated Endophthalmitis: Patient Baseline Features, Treatments, and Outcomes Total eyes Total patients Bilateral involvement Baseline features Female Male Hepatitis C Diabetes Mean age (range), years Saw ophthalmologist first Saw alternate provider first Mean (range) duration of symptoms (days) Mean time of follow-up (months) Median (range) time of follow-up (months) Initial macular involvement Macular involvement No macular involvement Indeterminate Treatment Initial tap and inject PPV after tap and inject Initial PPV Total (initial or eventual) PPV No treatment (patient refused) Extraocular involvement Without systemic symptoms or nonocular examination findings Outcomes (n ¼ 31 eyes) Vision Eyes with improvement between presentation and final follow-up Visual acuity of 20/40 (0.3 logMAR) or better at final follow-up Visual acuity of 20/200 or worse at final follow-up Visual acuity of 20/400 or worse at final follow-up Ocular sequelae Total eyes with any problem (in some, more than 1 of the below occurred in 1 eye) More than 1 problem Eyes requiring secondary surgery Epiretinal membrane Visually significant cataract* Retinal tear Retinal detachment/retinal detachment with proliferative vitreoretinopathy Macular scar
32 30 2 Patients 11 19 14 1 36 (17e59) 19 11 12 (1e60) 11 4.3 (0.033e67)
% 37% 63% 45% 3%
Eyes 20 8 4
% 63% 25% 12%
25 15 6 21 1 8 3
78% 60% 19% 66% 3% 27% 10%
N
%
28
90%
10
31%
8
25%
6
19%
20
63%
7 6 12 2 1 5/3
22% 19% 38% 6% 3% 16%/9%
7
22%
63% 37%
Onset of Symptoms and Clinical Characteristics There were no patients with known human immunodeficiency virus (HIV)/AIDS; 17 of 30 patients were tested during their hospital stay and were found to be negative. Nineteen patients were initially evaluated by an ophthalmologist, while in 11 patients another health care provider (such as an internist or emergency medicine physician) was first to see the patient. There was a mean of 12 days of symptoms before presentation to MEEI (range, 1e60 days). The mean duration of symptoms for fungal cases was 16 days (range, 2e60 days) and for bacterial cases was 4.6 days (range, 1e12 days). No patients received intraocular or systemic therapy before presentation at our institution. Five patients were started on topical prednisolone, 1
2
PPV ¼ pars plana vitrectomy. *Three pars plana lensectomies also performed at the time of retinal detachment repair.
patient was started on oral prednisone, and 3 patients were started on antibiotic ointment before presentation to MEEI. All of the patients reported ocular symptoms including decreased vision (27 patients), eye pain (15 patients), redness (14 patients), or floaters (9 patients). Among those patients who were evaluated before presentation to MEEI, 4 patients were diagnosed with anterior uveitis by ophthalmologists, 1 patient was diagnosed with posterior uveitis
Modjtahedi et al
Intravenous Drug UseeAssociated Endophthalmitis Table 2A. Microbial Findings
Microbial Yield from Different Culture Sources
n
Total blood cultures Total aqueous samples sent Total vitreous tap cultures sent
30 patients 22 24
Total vitreous cultures sent from PPV Total eye cultures (aqueous or vitreous from tap or PPV) Microbial results* Any positive result Blood culture positive Intraocular sampley Positive result from tap (aqueous or vitreous)z Aqueous (of the above tap patients)x Vitreous (of the above tap patients) All vitreous samples (tap or PPV) PPV positive Initial PPV PPV positive after initial tap and inject
one dry vitreous tap (no culture sent)/patient had vitreous sample sent with secondary PPV, which was negative 6 from primary PPV, 13 from secondary PPV
19 65 24/32 4/30 21/65 6/24 5/22 2/24 16/43 14/19 5/6 9/13
75% 13% 32% 25% 23% 8% 37% 74% 83% 69%
PPV ¼ pars plana vitrectomy. *All samples also sent for Gram stain and calcofluor-white. y Twenty-one intraocular samples (aqueous or vitreous obtained from a tap and inject procedure or PPV [primary and/or secondary]) were positive in 20 eyes. z One patient with positive aqueous and vitreous samples. x Yeast on calcofluor-white stain, no growth on culture.
by an ophthalmologist, 1 patient was diagnosed with a noninfectious panuveitis by an ophthalmologist, 2 patients were diagnosed with conjunctivitis by nonophthalmologists, 1 patient was diagnosed with a retinal detachment by a nonophthalmologist, and 1 patient was diagnosed with a corneal abrasion by a nonophthalmologist. Four patients were correctly diagnosed as having endophthalmitis when they were seen by outside ophthalmologists and were referred to MEEI for treatment. All eyes had vitritis. Six eyes in 6 patients had a hypopyon on presentation, including 1 patient who was subsequently diagnosed with endogenous bacterial endophthalmitis (EBE), 3 with EFE, and 2 with unknown organisms. Two patients had subretinal abscesses (1 with methicillin-sensitive Staphylococcus aureus and 1 with methicillin-resistant Staphylococcus aureus). There was macular involvement in 20 eyes and no macular involvement in 8 eyes, and the status of the macula could not be determined in 4 eyes because of media opacities.
Ten of the 30 patients reported systemic symptoms, including subjective fevers (9 patients), malaise, and myalgias. Systemic disease evaluation was recommended for all patients, including blood cultures and cardiac evaluation with an echocardiogram. Five patients did not follow up or refused an echocardiogram or cardiac evaluation. There were 8 patients (27%) with extraocular involvement, defined as positive blood cultures, positive systemic disease evaluation, or evidence of an extraocular focus of infection (Table 1). Five of these patients had positive review of systems and/or physical examination findings. One patient had an arm abscess and cellulitis with positive blood cultures for methicillinresistant Staphylococcus aureus (with localized extraocular symptoms but no systemic symptoms). Two patients had evident extraocular infections (1 tooth abscess and 1 orbital cellulitis with sinusitis) with local symptoms but without systemic symptoms or positive blood cultures. Two patients had endocarditis with valvular vegetations on echocardiogram with positive blood
Table 2B. Summary of Culture Results by Source Microbial Results
Source of Positive Sample
Culture results
# Eyes
%
Fungal Candida albicans* Candida dubliniesis Aspergillus niger Yeast on calcofluor-white stain (culture negative)* Bacterial Bacillus cereus Streptococcus mitis MRSA MSSA Negative cultures Refused ocular cultures (blood culture negative)
19 13 1 1 4 5 1 1 1 2 7 1
59% 41% 3% 3% 13% 16% 3% 3% 3% 6% 22% 3%
Blood
Aqueous humor
Vitreous humor
1
1
12 1 1
4 1 1 1 1
1
MRSA ¼ methicillin-resistant Staphylococcus aureus; MSSA ¼ methicillin-sensitive Staphylococcus aureus. *Two patients (1 with yeast and 1 with Candida albicans) had bilateral disease with positive microbial results in both eyes (see Table 5).
3
Ophthalmology Retina Volume -, Number -, Month 2016 cultures and systemic symptoms. Two patients had endocarditis without systemic symptoms or positive blood cultures (1 of these patients refused any cultures and the other patient had Candida albicans endophthalmitis), 1 of which was presumed endocarditis due to septic pulmonary emboli seen on computed tomography scan of the chest, although a transthoracic and transesophageal echocardiogram were negative for vegetations. One patient had a positive blood culture (Candida glabrata) with negative review of systems and no evidence of endocarditis.
Treatment and Microbiology Twenty-five eyes (78%) underwent a needle biopsy and injection of intravitreal antibiotics (“tap and inject”) on initial presentation, 6 eyes underwent initial pars plana vitrectomy (PPV, 19%) with injection of intravitreal antibiotics within 24 hours of presentation, and 1 patient refused ocular cultures and intravitreal treatment (3%, Table 1). Twenty-one eyes (66%) underwent PPV at some point for their infection (6 eyes on presentation, 15 eyes 1e56 days later, with a mean of 15 days). Of 25 eyes initially treated with tap and inject, 15 required secondary vitrectomy because of worsening inflammation noted on examination (28%, 7/25), for diagnostic purposes after a negative culture from the initial tap (8%, 2/25), or to clear the visual axis and promote faster visual recovery (24%, 6/ 25). None of the patients who received initial PPV required secondary vitrectomy for infection-related purposes. Two eyes (of the same patient) underwent repeat tap and inject, with 1 eye eventually also requiring secondary PPV. Among patients with nonIVDU endogenous endophthalmitis, 6 eyes underwent initial PPV (16.7%), 28 eyes underwent tap and inject (77.7%), and 2 patients had intravenous therapy alone (5.6%). Mean and median follow-up time was 9.7 months and 5.1 months, respectively (range, 0.04e45.9). Cultures of intraocular fluid and/or blood cultures were obtained in all patients. Tables 2A and B provide an in-depth summary of culture results. An organism was identified from at least 1 source in 24 of 32 eyes (75%). This represents 73% of patients, because 2 patients had bilateral disease with positive microbial results in both eyes. Only 1 of the fungal cases was due to a mold (Aspergillus), and Candida albicans accounted for nearly all of the cases due to yeast. Blood cultures were positive in 4 patients (13%), which included 60% (3 of 5) of bacterial cases but only 5% (1 of 19) of fungal cases. Seven patients, 2 with EBE, 4 with EFE, and 1 with negative cultures, were febrile at the time of blood cultures. Thirteen eyes had cultures from a secondary PPV sent after having initially undergone tap and inject. Sixty-nine percent (9/13) of these eyes had a positive culture from vitreous obtained from their secondary PPV. In the 9 eyes with positive vitrectomy cultures, the average time from tap and inject until vitrectomy was 11 days (range, 1e56 days), whereas the average for the 4 eyes with negative cultures was 28 days (range, 14e41 days). One eye underwent 2 tap-and-inject procedures before PPV, with negative subsequent cultures (see above). Of the 9 eyes with vitrectomy cultures, 8 received both antibacterial and antifungal intravitreal injections at the time of initial tap and inject, to which the organisms (subsequently identified at PPV) were susceptible. The remaining eye received initial antibacterial intravitreal antibiotics but subsequently grew Candida from the vitrectomy specimen (this eye received intravitreal amphotericin at the time of vitrectomy). In these 9 eyes with positive cultures, PPV was done for worsening disease in 3 eyes, for diagnostic purposes in 2 eyes, and to promote faster visual recovery in 4 eyes. Intravitreal antibiotics were administered to all patients except for the one who refused intravitreal treatment and who was treated with oral fluconazole (Table 3). All of the patients were initially treated with systemic antibiotics, which were frequently broad
4
Table 3. Initial Intravitreal Antibiotic Therapy # Eyes
% Eyes
4 2 2 5 2 3 22 20 2 1
12% 6% 6% 16% 6% 9% 69% 63% 6% 3%
Antibiotics Vancomycin / ceftazidime Vancomycin / amikacin Antifungals Voriconazole, then amphotericin Amphotericin Combination Vancomycin / ceftazidime / amphotericin Vancomycin / amphotericin Refused
Dosages used: vancomycin (1.0 mg in 0.1 mL), ceftazidime (2.0 mg in 0.1 mL), amphotericin (5 mcg or 10 mg in 0.1 mL), and voriconazole (100 mg in 0.1 mL). Note: In the case of penicillin allergy, amikacin was substituted for ceftazidime.
spectrum: 12 patients (40%) received a combination of an antibacterial and antifungal, 11 patients (37%) received an antifungal alone, and 7 patients (24%) received an antibacterial alone (Table 4). Patients on intravenous antifungal therapy were changed to oral azole therapy when fungal identification and sensitivity results were available. Among eyes with non-IVDU endogenous endophthalmitis, 25 (69%) had positive bacterial cultures: methicillin-resistant Staphylococcus aureus (22%, 8 eyes), methicillin-sensitive Staphylococcus aureus (14%, 5 eyes), group B Streptococcus (8.3%, 3 eyes), group G Streptococcus (5.6%, 2 eyes), group C Streptococcus (2.8%, 1 eye), Streptococcus mitis (5.6%, 2 eyes), Streptococcus pneumoniae (5.6%, 2 eyes), Clostridium difficile (2.8%, 2 eyes), and Klebsiella pneumoniae (2.8%, 1 eye). There were 5 eyes (14%) with positive fungal cultures among those with non-IVDU endogenous endopthalmitis: Candida (11%, 4 eyes) and Aspergillus (2.8%, 1 eye). Six patients (17%) had an unknown microbial
Table 4. Initial Systemic Therapy Systemic Antibiotic or Antifungal Oral Fluconazole Moxifloxacin Intravenous Ambisome Vancomycin/ceftazidime/voriconazole Vancomycin/ceftazidime/fluconazole Vancomycin/cefepime/fluconazole Vancomycin/cefepime/voriconazole Vancomycin/cefepime/amphotericin Vancomycin/nafcillin Vancomycin/moxifloxacin Vancomycin/gentamicin/clindamycin Vancomycin/ceftazidime/micafungin/ metronidazole/fluconazole Gentamicin/ceftriaxone Vancomycin/aztreonam and ambisome Gentamicin Voriconazole Vancomycin/ceftazidime Total
# Patients
% Patients
7 1 22 2 6 1 1 1 1 1 1 1 1
23% 3% 73% 7% 20% 3% 3% 3% 3% 3% 3% 3% 3%
1 1 1 1 2 30
3% 3% 3% 3% 7%
Modjtahedi et al
Intravenous Drug UseeAssociated Endophthalmitis Table 5. Individual Patient Summaries
Patient
Age at Presentation
Sex
Presenting VA
1 2 3 4 5 6 7 8 (right eye)
26 27 25 27 34 24 44 32
F M M F M F M F
CF HM 20/600 20/600 HM 20/500 LP 20/125
Aspergillus niger Candida albicans Candida albicans Candida albicans Candida albicans Candida albicans Candida albicans Candida albicans
PPV Tap and Tap and Tap and Tap and Tap and Tap and Tap and
8 (left eye)
32
F
20/40
Candida albicans
Tap and inject
9 10
39 30
F F
CF CF
Candida albicans Candida albicans
11 12 13 14 15 16 (right eye) 16 (left eye) 17 18
57 22 32 23 24 59 59 26 53
M M M M F M M M M
CF LP 20/400 20/200 20/400 CF 20/400 20/100 LP
19 20 21 22 23 24 25 26 27 28 29 30
41 29 54 40 34 53 30 22 17 56 48 29
M M M M M F M F F M F M
** HM 20/50 20/50 20/50 20/300 CF 20/125 20/125 HM 20/80 CF
Organism
Initial Treatment
Subsequent Treatment (if Any)
Final VA 20/100 20/80 20/60 20/80 20/20 20/80 CF 20/80
PPV Tap and inject
n/a PPV* PPV* PPV* PPV* PPV* PPV* Second tap and inject and PPV Second tap and inject n/a PPV*
Candida albicans Candida albicans Candida albicans Candida dubliniesis Yeast Yeast Yeast Yeast Bacillus cereus
PPV PPV PPV Tap and Tap and Tap and Tap and Tap and Tap and
n/a n/a n/a PPV PPV n/a n/a PPV PPV*
20/80 HM 20/40 20/40 20/100 20/300 20/100 20/25 HM
MRSA MSSA MSSA Strep mitis Not identified Not identified Not identified Not identified Not identified Not identified Not identified Refused
Tap and Tap and Tap and Tap and Tap and PPV Tap and Tap and Tap and Tap and Tap and Refused
n/a PPV* n/a n/a n/a n/a PPV PPV n/a n/a n/a Refused
** HM 20/25 20/20 20/30 20/30 20/50 LP 20/60 20/15 20/64 20/160
inject inject inject inject inject inject inject
inject inject inject inject inject inject inject inject inject inject inject inject inject inject inject inject
Ocular Sequelae ERM with foveal traction Macular scar ERM ERM Macula-off RD with PVR Retinal tear, macular scar, ERM
20/20 20/160 NLP
Giant retinal tear, PVR, RD, ERM RD and cataract
ERM ERM Macular scar, ERM Macular scar ERM with traction Total retinal detachment, PVR Macular scar, cataract
ERM and macular scar Tractional RD, ERM ERM Macular scar
CF ¼ count fingers; ERM ¼ epiretinal membrane; F ¼ female; HM ¼ hand motions; LP ¼ light perception; M ¼ male; MRSA ¼ methicillin-resistant Staphylococcus aureus; MSSA ¼ methicillin-sensitive Staphylococcus aureus; n/a ¼ not applicable; NLP ¼ no light perception; PPV ¼ pars plana vitrectomy; PVR ¼ proliferative vitreoretinopathy; RD ¼ retinal detachment; VA ¼ visual acuity. *Positive culture from secondary PPV after initial tap and inject; **vision data not available.
etiology. Fifty-six percent of non-IVDU endogenous endophthalmitis patients had positive blood cultures.
Outcomes Mean visual acuity improved significantly in patients, from 20/800 on presentation to 20/160 on last follow-up (1.64 logMAR to 0.91 logMAR, P < 0.0001), while median visual acuity improved from 20/600 (logMAR 1.5) to 20/80 (logMAR 0.6). Patients with macular involvement had a similar course (mean acuity: 1.56 logMAR to 0.7 logMAR, P ¼ 0.0003; median acuity: 1.4 logMAR to 0.55 logMAR). At final follow-up, 90% of eyes had improved vision compared with presentation, 25% of eyes had 20/200 or worse vision, and 31% of eyes had 20/40 or better vision. Mean and median visual acuity in those who received tap and inject improved from 20/600e20/800 (1.55 logMAR) to 20/160 (0.90 logMAR, P ¼ 0.0019) and from 20/400e20/500 (1.35 logMAR) to 20/60 to 20/80 (0.55 logMAR), respectively. For those who received initial PPV, mean and median acuity improved from
20/1600e20/2000 (1.92 logMAR) to 20/160e20/200 (0.95 logMAR, unable to calculate P value due to small n), and from counting fingers (2 logMAR) to 20/80e20/100 (0.65 logMAR). Among eyes with non-IVDU endogenous endophthalmitis, the average initial vision was 20/1260e20/1600 (1.86 logMAR) and final vision was approximately 20/500 (1.42 logMAR), which was a statistically significant improvement (P < 0.001). Sixty-one percent of eyes had some improvement in vision, 18 eyes (50%) had 20/200 or worse vision, and 12 eyes (33%) had final vision of 20/40 or better. When the IVDU and non-IVDU cases were combined, there was a statistically significant improvement in vision (P ¼ 0.0038). There was not a statistically significant difference in initial vision (P ¼ 0.46) or final vision (0.26) between the IVDU and non-IVDU patients. Twenty eyes (63%) suffered ocular sequelae, with 7 eyes (22%) having more than 1 of the following problems (Table 1). Of the eyes with sequelae, 14 were associated with fungal infections, 2 with bacterial infection, and 4 with unknown organisms. Six eyes (19%) required a secondary surgery: 4 for retinal
5
Ophthalmology Retina Volume -, Number -, Month 2016 detachment repair (with 3 concurrent pars plana lensectomies) and 2 for cataract. Table 5 provides individual patient summaries.
Discussion This is the largest series of IVDU-associated endophthalmitis to date. IVDU-associated endophthalmitis is a significant source of ophthalmic morbidity, accounting for 9% of all cases of endophthalmitis at our academic referral center. Ten patients were initially misdiagnosed before presentation. Patients with IVDU-associated endophthalmitis are typically young and with minimal systemic complaints. These patients may be initially reluctant to volunteer a history of drug use; however, clinicians should have a high suspicion and aggressively pursue the possibility of IVDU in any patient with unexplained intraocular inflammation. This is especially important in younger, otherwise healthy patients, and the examination should include a search for track marks or skin abscesses. The ability to identify a causative organism varied significantly in our series based on source and method of acquisition, with a much higher yield from PPV (including secondary PPV) than from tap and inject. Endogenous endophthalmitis begins as a chorioretinitis that spreads into the vitreous; as a result, most of the microbial load is initially localized in the posterior vitreous and is therefore not accessible via bedside aspirate, especially if there is no posterior vitreous detachment, which is often the case in younger patients. Patients undergoing initial PPV had worse initial vision and may have had more severe infection with greater microbial load, which may have resulted in increased culture positivity. Additionally, the smaller volume of vitreous acquired from bedside aspirate (typically 0.1e0.2 mL) compared with PPV (typically up to 1 mL) results in decreased sensitivity. Seventy-five percent of EFE patients had positive cultures from PPV in William et al’s series.20 Connell et al’s Australian IVDU endophthalmitis cohort9 had positive vitreous cultures in 68.4% of patients (66.7% from vitreous tap and 70% from vitrectomy, all positive samples being for Candida), with the authors finding that repeat aspiration in those with initially negative cultures did not increase the yield. Conversely, vitreous cultures were positive in only 4 of 25 samples in Patel et al’s New Jerseyebased IVDU endophthalmitis cohort (2 of which were thought to be contaminants).21 Polymerase chain reaction had a 100% yield (without a difference in aqueous and vitreous samples), compared with 37.5% with standard techniques,22 in Sowmya and Madhayan’s series and can identify offending organisms in 90 minutes.9,23,24 Most reported cases of IVDU-associated endophthalmitis are fungal, with all cases in several series being attributed to fungal etiology5,7e10; however, bacterial cases constituted a significant disease burden in our series. Five eyes (16%) in our series were caused by bacteria vs. 39% (7/18) of cases in Patel et al’s series (with 11% [2/18] being of uncertain “bacterial vs. fungal etiology”).21 The greater microbial diversity seen in the Boston and New Jersey patients may reflect different microbiomes in colder and more urban environments. In contrast, bacterial etiologies constituted
6
69% of non-IVDU endogenous endophthalmitis cases at our institution. A bacterial etiology may be overlooked by clinicians evaluating patients with endogenous endophthalmitis and a history of IVDU, given the historically strong association with fungal causes.25 It is important to remain mindful of the significant number of bacterial sources of IVDU endophthalmitis. EFE tends to have a more indolent course with a longer duration of symptoms (mean 16 days) compared with EBE (4.6 days). Initial treatment may need to cover a wide spectrum of pathogens, but then should be tailored based on culture results. If a tap and inject is the initial procedure performed and if results from these cultures and blood cultures are negative, a PPV should be promptly considered to make a microbial diagnosis. Although endophthalmitis from IVDU is usually thought to be from transient bacteremia or fungemia, the morbidity of undiagnosed systemic disease is potentially high. As such, systemic disease evaluation and treatment were recommended in all patients, and 27% of cases in our series were found to have extraocular infection. There were 3 patients (3/30, 10%) who had extraocular infections without suggestive symptoms or physical examination findings. The majority of patients (5/8, 63%) with extraocular involvement were symptomatic and/or had positive extraocular examination findings. This highlights the importance of a thorough systemic work-up in all patients with endogenous endophthalmitis, including those with only ocular symptoms. Initial hospitalization is typically necessary to ensure a timely evaluation for systemic foci of infection, to initiate intravenous antibiotics, and to ensure that the involved eye is improving with treatment. Outpatient systemic evaluation for IVDU-associated endophthalmitis is conceivable, especially if the patient does not have extraocular symptoms or examination findings; however, IVDU patients are prone to poor compliance and inconsistent outpatient follow-up, which is mitigated by hospitalization. An interesting subset of patients with IVDU endophthalmitis are those misusing oral heroin substitutes such as buprenorphine.26,27 One patient (3%) in our series was injecting crushed suboxone (buprenorphine and naloxone) and required PPV after initial tap and inject due to worsening disease. Hirsbein et al reported a series of 7 patients who developed endophthalmitis while misusing buprenorphine, all of whom required vitrectomy (5/7 culture positive, all Candida species).27 Connell et al had 5 patients (26%, N ¼ 19) who used buprenorphine, 2 of whom required vitrectomy.9 Ocular morbidity was high in this series, with 63% of patients suffering at least 1 sequela of endophthalmitis and 19% of eyes (6/32) requiring a secondary surgery. Twenty eyes had initial macular involvement, with only a significant scar being noted in 7 eyes at final follow-up, indicating that many lesions may regress without significant scarring. Patients require close follow-up even after the infection is under control because they remain at risk for other ocular sequelae. There was a statistically significant improvement in mean visual acuity in patients, including among those with documented macular involvement. Interestingly, mean final visual acuity was comparable when considering all eyes
Modjtahedi et al
Intravenous Drug UseeAssociated Endophthalmitis
(20/160) and those with documented macular involvement (20/100). This may have been because macular involvement did not affect the fovea (allowing for better than expected vision) or because poor initial view to the posterior pole resulted in failure to diagnose cases of macular involvement (resulting in worse vision in those without documented initial macular involvement). Among all patients with endogenous endophthalmitis in prior series, 56% (10/18) of eyes had final acuity worse than 20/400,5 while the average acuity of EFE patients with yeast infections was 20/200 to 20/250 (logMAR 1.05).10 Initial and final vision was worse in patients with non-IVDU endogenous endophthalmitis than in patients with IVDU endophthalmitis at our institution. This difference did not reach statistical significance, potentially because of small sample size. Similarly, fewer patients (61% vs. 90%) had improvement in vision and more patients had 20/200 or worse vision (50% vs. 25%) among patients with non-IVDU endogenous endophthalmitis compared with IVDU endophthalmitis, respectively. This may have been attributable to the higher incidence of bacterial infections in the non-IVDU endogenous endophthalmitis group, which typically has a worse prognosis than fungal infections. In Connell et al’s series, 83% (15/18) of patients with visual outcome data had acuity of 6/60, 50% achieved 6/18, and mean final acuity was 6/36.9 Patel et al theorized their patients’ poor final acuity (19% of eyes having better than 20/400 vision) was because of the high frequency of initial macular involvement (11/18 eyes, 61%) and subsequent scarring.21 Only 19% of our patients had 20/400 vision or worse at final follow-up, although the percentage of patients with initial macular involvement (20/32 eyes, 63%) was similar to the series by Patel et al.21 Visual acuity outcomes vary significantly and are likely related to differences in duration of disease, microbial etiology, and secondary sequelae. Final visual acuity was similar between eyes initially treated with tap and inject vs. PPV, as was also seen by Connell et al,9 with eyes undergoing initial PPV having a more significant improvement in vision due to worse initial acuity. Although the initial treatment of choice for most patients was tap and inject (25/30, 78%), a majority (15/25, 60%) of these eyes received PPV at a later time. None of the 6 eyes that had initial PPV required repeat PPV for their acute infection. Of Connell et al’s patients, 54% (10/19) required PPV, 7 on initial presentation and 3 due to worsening disease following tap and inject,9 compared with 66% of all eyes in our series (6 initial and 15 delayed).9 There are no clear guidelines for PPV in endogenous endophthalmitis patients. Final acuity was similar in both treatment arms of our study, which is notable for those who underwent PPV, given their worse initial acuity (which presumably indicates worse disease burden). Diagnostic yield was also higher from samples attained from PPV than from tap and inject. Additionally, initial PPV may be less burdensome to patients and may hold cost advantages, because the majority of tap and inject patients eventually require vitrectomy. It is difficult to directly compare outcomes of PPV vs. tap and inject because patients were not randomized to treatment and the 2 groups were different
on presentation. Sixty-nine percent of eyes had a positive vitreous culture from a secondary PPV after having had a tap and inject (an average of 11 days after initial treatment) even though all but 1 eye received appropriate antimicrobial coverage at the time of their tap and inject. Four of the eyes with positive secondary PPV cultures after initial tap and inject underwent PPV to promote faster visual recovery and not because of worsening clinical course or need for diagnostic information; therefore, it is advisable to consider vitreous culture and intravitreal antimicrobials at the time of secondary PPV even if clinical suspicion for active infection is low. The high rate of positive vitreous cultures from these secondary PPV cases raises the possibility that intravitreal injections may not adequately treat the active infection in a sizable number of patients and provides additional support for early PPV. Trends in endophthalmitis are constantly evolving and IVDU-associated endophthalmitis represents a subset that deserves special attention. The limitations of this study include its small sample size and retrospective design. Early vitrectomy may play an important role in management. Future investigations should focus on the role and timing of PPV as well as optimizing management and systemic disease evaluation.
References 1. Durand ML. Endophthalmitis. Clin Microbiol Infec. 2013;19(3):227e234. 2. Kostick DA, Foster RE, Lowder CY, et al. Endogenous endophthalmitis caused by Candida albicans in a healthy woman. Am J Ophthalmol. 1992;113(5):593e595. 3. Mamandhar A, Bajracharya L. Endogenous Aspergillus endophthalmitis in a healthy individual. Nepal J Ophthalmol. 2012;4(1):179e183. 4. Valluri S, Moorthy RS, Liggett PE, Rao NA. Endogenous Aspergillus endophthalmitis in an immunocompetent individual. Int Ophthalmol. 1993;17(3):131e135. 5. Schiedler V, Scott IU, Flynn Jr HW, et al. Culture-proven endogenous endophthalmitis: clinical features and visual acuity outcomes. Am J Ophthalmol. 2004;137(4):725e731. 6. Dozier CC, Tarantola RM, Jiramongkolchai K, Donahue SP. Fungal eye disease at a tertiary care center: the utility of routine inpatient consultation. Ophthalmology. 2011;118(8): 1671e1676. 7. Leibovitch I, Lai T, Raymond G, et al. Endogenous endophthalmitis: a 13-year review at a tertiary hospital in South Australia. Scand J Infect Dis. 2005;37(3):184e189. 8. Chakrabarti A, Shivaprakash MR, Singh R, et al. Fungal endophthalmitis: fourteen years’ experience from a center in India. Retina. 2008;28(10):1400e1407. 9. Connell PP, O’Neill EC, Amirul Islam FM, et al. Endogenous endophthalmitis associated with intravenous drug abuse: seven-year experience at a tertiary referral center. Retina. 2010;30(10):1721e1725. 10. Sridhar J, Flynn Jr HW, Kuriyan AE, et al. Endogenous fungal endophthalmitis: risk factors, clinical features, and treatment outcomes in mold and yeast infections. J Ophthalmic Inflamm Infect. 2013;3(1):60. 11. Connell PP, O’Neill EC, Fabinyi D, et al. Endogenous endophthalmitis: 10-year experience at a tertiary referral centre. Eye (Lond). 2011;25(1):66e72.
7
Ophthalmology Retina Volume -, Number -, Month 2016 12. Lim HW, Shin JW, Cho HY, et al. Endogenous endophthalmitis in the Korean population: a six-year retrospective study. Retina. 2014;34:592e602. 13. Shen X, Xu G. Vitrectomy for endogenous fungal endophthalmitis. Ocul Immunol Inflamm. 2009;17(3):148e152. 14. Wong JS, Chan TK, Lee HM, Chee SP. Endogenous bacterial endophthalmitis: an east Asian experience and a reappraisal of a severe ocular affliction. Ophthalmology. 2000;107(8):1483e1491. 15. Zhang H, Liu Z. Endogenous endophthalmitis: a 10-year review of culture-positive cases in northern China. Ocul Immunol Inflamm. 2010;18(2):133e138. 16. Jones CM, Logan J, Gladden M, Bohm MK. Vital signs: demographic and substance use trends among heroin usersd United States, 2002e2013. Morb Mortal Wkly Rep. 2015;64(26):719e725. 17. Jones CM. Heroin use and heroin use risk behaviors among nonmedical users of prescription opioid pain relieversdUnited States, 2002e2004 and 2008e2010. Drug Alcohol Depend. 2013;132(1-2):95e100. 18. Holladay JT. Proper method for calculating average visual acuity. J Refract Surg. 1997;13(4):388e391. 19. van de Put MA, Hoeksema L, Wanders W, et al. Postoperative vision-related quality of life in macula-off rhegmatogenous retinal detachment patients and its relation to visual function. PloS One. 2014;9(12):e114489. 20. William A, Spitzer MS, Deuter C, et al. Outcomes of primary transconjunctival 23-gauge vitrectomy in the diagnosis and
21. 22.
23. 24.
25.
26. 27.
treatment of presumed endogenous fungal endophthalmitis. Ocul Immunol Inflamm. 2016:1e7. Patel SN, Rescigno RJ, Zarbin MA, Langer P, Bhagat N. Endogenous endophthalmitis associated with intravenous drug abuse. Retina. 2014;34(7):1460e1465. Sowmya P, Madhavan HN. Diagnostic utility of polymerase chain reaction on intraocular specimens to establish the etiology of infectious endophthalmitis. Eur J Ophthalmol. 2009;19(5):812e817. Goldschmidt P, Degorge S, Benallaoua D, et al. New test for the diagnosis of bacterial endophthalmitis. Br J Ophthalmol. 2009;93(8):1089e1095. Hidalgo JA, Alangaden GJ, Eliott D, et al. Fungal endophthalmitis diagnosis by detection of Candida albicans DNA in intraocular fluid by use of a species-specific polymerase chain reaction assay. J Infect Dis. 2000;181(3): 1198e1201. Jackson TL, Eykyn SJ, Graham EM, Stanford MR. Endogenous bacterial endophthalmitis: a 17-year prospective series and review of 267 reported cases. Surv Ophthalmol. 2003;48(4):403e423. Aboltins CA, Allen P, Daffy JR. Fungal endophthalmitis in intravenous drug users injecting buprenorphine contaminated with oral Candida species. Med J Aust. 2005;182(8):427. Hirsbein D, Attal P, Gueudry J, et al. [Endogenous Candida endophthalmitis in drug misusers injecting intravenous buprenorphine]. J Fr Ophtalmol. 2008;31(2):180e183.
Footnotes and Financial Disclosures Originally received: September 12, 2016. Final revision: October 24, 2016. Accepted: October 25, 2016. Available online: ---. Manuscript no. ORET_2016_30. 1
Retina Service, Department of Ophthalmology, Massachusetts Eye and Ear Infirmary, Harvard Medical School, Boston, Massachusetts.
2
Infectious Disease Service, Massachusetts Eye and Ear Infirmary, Boston, Massachusetts.
Author Contributions: Research design: Modjtahedi, Finn, Papakostas, Husain, Eliott Data analysis and/or interpretation: All authors (Modjtahedi, Finn, Papakostas, Durand, Husain, Eliott) Data acquisition and/or research execution: All authors (Modjtahedi, Finn, Papakostas, Durand, Husain, Eliott) Manuscript preparation: All authors (Modjtahedi, Finn, Papakostas, Durand, Husain, Eliott)
3
Department of Medicine, Massachusetts General Hospital, Boston, Massachusetts. *Drs Modjtahedi and Finn contributed equally to this work. Financial Disclosure(s): The author(s) have made the following disclosure(s): There were no funding sources for this project. The authors have no conflicts of interest to report. B.S.M., A.V.F., T.D.P., M.D., D.H.: No financial disclosures. D.E.: Consultant d Alcon, Alimera, Allergan, Arctic, Avalanche, Dutch Ophthalmic, Ophthotech, MacuLogix, Santen, and Thrombogenics; Research grants d Neurotech and Ocata Therapeutics.
8
Abbreviations and Acronyms: EBE ¼ endogenous bacterial endophthalmitis; EFE ¼ endogenous fungal endophthalmitis; HIV ¼ human immunodeficiency virus; IVDU ¼ intravenous drug use; MEEI ¼ Massachusetts Eye and Ear Infirmary; PPV ¼ pars plana vitrectomy. Correspondence: Dean Eliott, MD, Retina Service, Department of Ophthalmology, Massachusetts Eye and Ear Infirmary, Harvard Medical School, Boston, MA 02114. E-mail: [email protected].
Endophthalmitis can be divided into exogenous (postsurgical, posttraumatic, or from a contiguous infection such as keratitis) or endogenous (ocular seeding from a hematogenous infection).1 Endogenous endophthalmitis is encountered either in patients with an ongoing extraocular source of infection or in patients with transient bacteremia or fungemia, such as after an endoscopic procedure or after injection drug use. Although some patients with endogenous endophthalmitis have no apparent risk factors, such an absence is rare.2e5 Most reported cases of endogenous endophthalmitis are in systemically ill or immunocompromised patients. The incidence of endogenous fungal endophthalmitis (EFE) may be declining among hospitalized patients, likely as a result of earlier detection of infection and improved vitreous penetration of systemic antifungals.6 Endogenous endophthalmitis from intravenous drug use (IVDU) would not be expected to be as responsive to the above trend because of the transient nature of septicemia, the lack of 2016 by the American Academy of Ophthalmology Published by Elsevier Inc.
prophylactic treatment, and the fact that patients often do not present until they develop ocular symptoms. IVDU was seen in 14% to 30% of cases of endogenous endophthalmitis in series from Miami, Florida, India, and Australia,5,7e11 but it is seldom described in East Asian series.12e15 The recent surge in heroin use in the United States is a matter of considerable public health concern and necessitates a better understanding of the consequences of IVDU, which include endophthalmitis.16,17 This study examines endophthalmitis associated with IVDU at a large tertiary ophthalmology center in Boston, Massachusetts, and constitutes the largest series to date. We aimed to explore demographics, clinical presentation, microbial etiology, treatments, and outcomes of this condition.
Methods Institutional Review Board approval was obtained by the Human Studies Committee and the study protocol adhered to the http://dx.doi.org/10.1016/j.oret.2016.10.013 ISSN 2468-6530/16
1
Ophthalmology Retina Volume -, Number -, Month 2016 Declaration of Helsinki. International Classification of Diseases 9 (ICD-9) billing codes were used to identify cases of endophthalmitis at the Massachusetts Eye and Ear Infirmary (MEEI) from September 2006 until November 2014. Every medical record was reviewed and patients with a preceding history of IVDU were identified. Inclusion criteria included a history of endophthalmitis, a preceding history of IVDU, no preceding history of ocular surgery/ trauma, and no other noneIVDU-related plausible source of infection. Patient demographics, past ocular history, symptoms (divided into ocular, extraocular, and systemic), intravenous drug used, initial type of physician to evaluate patient, systemic evaluations, visual acuity and complete ocular examination, culture results, treatment, ocular complications, and duration of follow-up were all evaluated. Visual acuity data were analyzed after converting documented Snellen acuity to the equivalent logMAR value based on previously published conversion tables.18,19 Patients with a history of endogenous endophthalmitis not from IVDU were analyzed as well for comparison. Vision data were checked for normality with the D’Agostino-Pearson omnibus normality test and were found to be distributed in a non-Gaussian fashion. The Wilcoxon signed rank test was used for statistical companions within the IVDU and non-IVDU groups. Within each of these groups adjustments for multiple comparisons were not made. For the comparison between IVDU and non-IVDU groups a Mann-Whitney test was used. When looking at all cases of endogenous endophthalmitis (IVDU and non-IVDU together), a Kruskal-Wallis with the Dunn post hoc test for multiple comparisons was used. The level of statistical significance was set at 0.05.
Results Patient Characteristics A total of 338 patients with endophthalmitis were seen during the study period, of which 68 had endogenous endophthalmitis (20%), and 30 patients (32 eyes) were identified as having IVDU-associated endophthalmitis (9% of total, 44% of endogenous cases). There were 36 patients (36 eyes) with non-IVDU endogenous endophthalmitis. One patient did not have visual acuity data available for analysis because presentation preceded electronic medical records. Mean follow-up time was 11 months (range, 1 day to 67 months). The patients’ baseline features are summarized in Table 1. All of the patients admitted to a history of IVDU. Seventeen patients reported heroin use, 5 reported intravenous cocaine use, 1 reported intravenous hydromorphone use, 1 reported polysubstance use, 1 reported intravenous testosterone use, 1 reported intravenous methadone use, and 1 reported crushing and injecting suboxone (buprenorphine and naloxone). For the remaining 3 patients, the drug injected was not recorded or was unknown.
Table 1. Intravenous Drug UseeAssociated Endophthalmitis: Patient Baseline Features, Treatments, and Outcomes Total eyes Total patients Bilateral involvement Baseline features Female Male Hepatitis C Diabetes Mean age (range), years Saw ophthalmologist first Saw alternate provider first Mean (range) duration of symptoms (days) Mean time of follow-up (months) Median (range) time of follow-up (months) Initial macular involvement Macular involvement No macular involvement Indeterminate Treatment Initial tap and inject PPV after tap and inject Initial PPV Total (initial or eventual) PPV No treatment (patient refused) Extraocular involvement Without systemic symptoms or nonocular examination findings Outcomes (n ¼ 31 eyes) Vision Eyes with improvement between presentation and final follow-up Visual acuity of 20/40 (0.3 logMAR) or better at final follow-up Visual acuity of 20/200 or worse at final follow-up Visual acuity of 20/400 or worse at final follow-up Ocular sequelae Total eyes with any problem (in some, more than 1 of the below occurred in 1 eye) More than 1 problem Eyes requiring secondary surgery Epiretinal membrane Visually significant cataract* Retinal tear Retinal detachment/retinal detachment with proliferative vitreoretinopathy Macular scar
32 30 2 Patients 11 19 14 1 36 (17e59) 19 11 12 (1e60) 11 4.3 (0.033e67)
% 37% 63% 45% 3%
Eyes 20 8 4
% 63% 25% 12%
25 15 6 21 1 8 3
78% 60% 19% 66% 3% 27% 10%
N
%
28
90%
10
31%
8
25%
6
19%
20
63%
7 6 12 2 1 5/3
22% 19% 38% 6% 3% 16%/9%
7
22%
63% 37%
Onset of Symptoms and Clinical Characteristics There were no patients with known human immunodeficiency virus (HIV)/AIDS; 17 of 30 patients were tested during their hospital stay and were found to be negative. Nineteen patients were initially evaluated by an ophthalmologist, while in 11 patients another health care provider (such as an internist or emergency medicine physician) was first to see the patient. There was a mean of 12 days of symptoms before presentation to MEEI (range, 1e60 days). The mean duration of symptoms for fungal cases was 16 days (range, 2e60 days) and for bacterial cases was 4.6 days (range, 1e12 days). No patients received intraocular or systemic therapy before presentation at our institution. Five patients were started on topical prednisolone, 1
2
PPV ¼ pars plana vitrectomy. *Three pars plana lensectomies also performed at the time of retinal detachment repair.
patient was started on oral prednisone, and 3 patients were started on antibiotic ointment before presentation to MEEI. All of the patients reported ocular symptoms including decreased vision (27 patients), eye pain (15 patients), redness (14 patients), or floaters (9 patients). Among those patients who were evaluated before presentation to MEEI, 4 patients were diagnosed with anterior uveitis by ophthalmologists, 1 patient was diagnosed with posterior uveitis
Modjtahedi et al
Intravenous Drug UseeAssociated Endophthalmitis Table 2A. Microbial Findings
Microbial Yield from Different Culture Sources
n
Total blood cultures Total aqueous samples sent Total vitreous tap cultures sent
30 patients 22 24
Total vitreous cultures sent from PPV Total eye cultures (aqueous or vitreous from tap or PPV) Microbial results* Any positive result Blood culture positive Intraocular sampley Positive result from tap (aqueous or vitreous)z Aqueous (of the above tap patients)x Vitreous (of the above tap patients) All vitreous samples (tap or PPV) PPV positive Initial PPV PPV positive after initial tap and inject
one dry vitreous tap (no culture sent)/patient had vitreous sample sent with secondary PPV, which was negative 6 from primary PPV, 13 from secondary PPV
19 65 24/32 4/30 21/65 6/24 5/22 2/24 16/43 14/19 5/6 9/13
75% 13% 32% 25% 23% 8% 37% 74% 83% 69%
PPV ¼ pars plana vitrectomy. *All samples also sent for Gram stain and calcofluor-white. y Twenty-one intraocular samples (aqueous or vitreous obtained from a tap and inject procedure or PPV [primary and/or secondary]) were positive in 20 eyes. z One patient with positive aqueous and vitreous samples. x Yeast on calcofluor-white stain, no growth on culture.
by an ophthalmologist, 1 patient was diagnosed with a noninfectious panuveitis by an ophthalmologist, 2 patients were diagnosed with conjunctivitis by nonophthalmologists, 1 patient was diagnosed with a retinal detachment by a nonophthalmologist, and 1 patient was diagnosed with a corneal abrasion by a nonophthalmologist. Four patients were correctly diagnosed as having endophthalmitis when they were seen by outside ophthalmologists and were referred to MEEI for treatment. All eyes had vitritis. Six eyes in 6 patients had a hypopyon on presentation, including 1 patient who was subsequently diagnosed with endogenous bacterial endophthalmitis (EBE), 3 with EFE, and 2 with unknown organisms. Two patients had subretinal abscesses (1 with methicillin-sensitive Staphylococcus aureus and 1 with methicillin-resistant Staphylococcus aureus). There was macular involvement in 20 eyes and no macular involvement in 8 eyes, and the status of the macula could not be determined in 4 eyes because of media opacities.
Ten of the 30 patients reported systemic symptoms, including subjective fevers (9 patients), malaise, and myalgias. Systemic disease evaluation was recommended for all patients, including blood cultures and cardiac evaluation with an echocardiogram. Five patients did not follow up or refused an echocardiogram or cardiac evaluation. There were 8 patients (27%) with extraocular involvement, defined as positive blood cultures, positive systemic disease evaluation, or evidence of an extraocular focus of infection (Table 1). Five of these patients had positive review of systems and/or physical examination findings. One patient had an arm abscess and cellulitis with positive blood cultures for methicillinresistant Staphylococcus aureus (with localized extraocular symptoms but no systemic symptoms). Two patients had evident extraocular infections (1 tooth abscess and 1 orbital cellulitis with sinusitis) with local symptoms but without systemic symptoms or positive blood cultures. Two patients had endocarditis with valvular vegetations on echocardiogram with positive blood
Table 2B. Summary of Culture Results by Source Microbial Results
Source of Positive Sample
Culture results
# Eyes
%
Fungal Candida albicans* Candida dubliniesis Aspergillus niger Yeast on calcofluor-white stain (culture negative)* Bacterial Bacillus cereus Streptococcus mitis MRSA MSSA Negative cultures Refused ocular cultures (blood culture negative)
19 13 1 1 4 5 1 1 1 2 7 1
59% 41% 3% 3% 13% 16% 3% 3% 3% 6% 22% 3%
Blood
Aqueous humor
Vitreous humor
1
1
12 1 1
4 1 1 1 1
1
MRSA ¼ methicillin-resistant Staphylococcus aureus; MSSA ¼ methicillin-sensitive Staphylococcus aureus. *Two patients (1 with yeast and 1 with Candida albicans) had bilateral disease with positive microbial results in both eyes (see Table 5).
3
Ophthalmology Retina Volume -, Number -, Month 2016 cultures and systemic symptoms. Two patients had endocarditis without systemic symptoms or positive blood cultures (1 of these patients refused any cultures and the other patient had Candida albicans endophthalmitis), 1 of which was presumed endocarditis due to septic pulmonary emboli seen on computed tomography scan of the chest, although a transthoracic and transesophageal echocardiogram were negative for vegetations. One patient had a positive blood culture (Candida glabrata) with negative review of systems and no evidence of endocarditis.
Treatment and Microbiology Twenty-five eyes (78%) underwent a needle biopsy and injection of intravitreal antibiotics (“tap and inject”) on initial presentation, 6 eyes underwent initial pars plana vitrectomy (PPV, 19%) with injection of intravitreal antibiotics within 24 hours of presentation, and 1 patient refused ocular cultures and intravitreal treatment (3%, Table 1). Twenty-one eyes (66%) underwent PPV at some point for their infection (6 eyes on presentation, 15 eyes 1e56 days later, with a mean of 15 days). Of 25 eyes initially treated with tap and inject, 15 required secondary vitrectomy because of worsening inflammation noted on examination (28%, 7/25), for diagnostic purposes after a negative culture from the initial tap (8%, 2/25), or to clear the visual axis and promote faster visual recovery (24%, 6/ 25). None of the patients who received initial PPV required secondary vitrectomy for infection-related purposes. Two eyes (of the same patient) underwent repeat tap and inject, with 1 eye eventually also requiring secondary PPV. Among patients with nonIVDU endogenous endophthalmitis, 6 eyes underwent initial PPV (16.7%), 28 eyes underwent tap and inject (77.7%), and 2 patients had intravenous therapy alone (5.6%). Mean and median follow-up time was 9.7 months and 5.1 months, respectively (range, 0.04e45.9). Cultures of intraocular fluid and/or blood cultures were obtained in all patients. Tables 2A and B provide an in-depth summary of culture results. An organism was identified from at least 1 source in 24 of 32 eyes (75%). This represents 73% of patients, because 2 patients had bilateral disease with positive microbial results in both eyes. Only 1 of the fungal cases was due to a mold (Aspergillus), and Candida albicans accounted for nearly all of the cases due to yeast. Blood cultures were positive in 4 patients (13%), which included 60% (3 of 5) of bacterial cases but only 5% (1 of 19) of fungal cases. Seven patients, 2 with EBE, 4 with EFE, and 1 with negative cultures, were febrile at the time of blood cultures. Thirteen eyes had cultures from a secondary PPV sent after having initially undergone tap and inject. Sixty-nine percent (9/13) of these eyes had a positive culture from vitreous obtained from their secondary PPV. In the 9 eyes with positive vitrectomy cultures, the average time from tap and inject until vitrectomy was 11 days (range, 1e56 days), whereas the average for the 4 eyes with negative cultures was 28 days (range, 14e41 days). One eye underwent 2 tap-and-inject procedures before PPV, with negative subsequent cultures (see above). Of the 9 eyes with vitrectomy cultures, 8 received both antibacterial and antifungal intravitreal injections at the time of initial tap and inject, to which the organisms (subsequently identified at PPV) were susceptible. The remaining eye received initial antibacterial intravitreal antibiotics but subsequently grew Candida from the vitrectomy specimen (this eye received intravitreal amphotericin at the time of vitrectomy). In these 9 eyes with positive cultures, PPV was done for worsening disease in 3 eyes, for diagnostic purposes in 2 eyes, and to promote faster visual recovery in 4 eyes. Intravitreal antibiotics were administered to all patients except for the one who refused intravitreal treatment and who was treated with oral fluconazole (Table 3). All of the patients were initially treated with systemic antibiotics, which were frequently broad
4
Table 3. Initial Intravitreal Antibiotic Therapy # Eyes
% Eyes
4 2 2 5 2 3 22 20 2 1
12% 6% 6% 16% 6% 9% 69% 63% 6% 3%
Antibiotics Vancomycin / ceftazidime Vancomycin / amikacin Antifungals Voriconazole, then amphotericin Amphotericin Combination Vancomycin / ceftazidime / amphotericin Vancomycin / amphotericin Refused
Dosages used: vancomycin (1.0 mg in 0.1 mL), ceftazidime (2.0 mg in 0.1 mL), amphotericin (5 mcg or 10 mg in 0.1 mL), and voriconazole (100 mg in 0.1 mL). Note: In the case of penicillin allergy, amikacin was substituted for ceftazidime.
spectrum: 12 patients (40%) received a combination of an antibacterial and antifungal, 11 patients (37%) received an antifungal alone, and 7 patients (24%) received an antibacterial alone (Table 4). Patients on intravenous antifungal therapy were changed to oral azole therapy when fungal identification and sensitivity results were available. Among eyes with non-IVDU endogenous endophthalmitis, 25 (69%) had positive bacterial cultures: methicillin-resistant Staphylococcus aureus (22%, 8 eyes), methicillin-sensitive Staphylococcus aureus (14%, 5 eyes), group B Streptococcus (8.3%, 3 eyes), group G Streptococcus (5.6%, 2 eyes), group C Streptococcus (2.8%, 1 eye), Streptococcus mitis (5.6%, 2 eyes), Streptococcus pneumoniae (5.6%, 2 eyes), Clostridium difficile (2.8%, 2 eyes), and Klebsiella pneumoniae (2.8%, 1 eye). There were 5 eyes (14%) with positive fungal cultures among those with non-IVDU endogenous endopthalmitis: Candida (11%, 4 eyes) and Aspergillus (2.8%, 1 eye). Six patients (17%) had an unknown microbial
Table 4. Initial Systemic Therapy Systemic Antibiotic or Antifungal Oral Fluconazole Moxifloxacin Intravenous Ambisome Vancomycin/ceftazidime/voriconazole Vancomycin/ceftazidime/fluconazole Vancomycin/cefepime/fluconazole Vancomycin/cefepime/voriconazole Vancomycin/cefepime/amphotericin Vancomycin/nafcillin Vancomycin/moxifloxacin Vancomycin/gentamicin/clindamycin Vancomycin/ceftazidime/micafungin/ metronidazole/fluconazole Gentamicin/ceftriaxone Vancomycin/aztreonam and ambisome Gentamicin Voriconazole Vancomycin/ceftazidime Total
# Patients
% Patients
7 1 22 2 6 1 1 1 1 1 1 1 1
23% 3% 73% 7% 20% 3% 3% 3% 3% 3% 3% 3% 3%
1 1 1 1 2 30
3% 3% 3% 3% 7%
Modjtahedi et al
Intravenous Drug UseeAssociated Endophthalmitis Table 5. Individual Patient Summaries
Patient
Age at Presentation
Sex
Presenting VA
1 2 3 4 5 6 7 8 (right eye)
26 27 25 27 34 24 44 32
F M M F M F M F
CF HM 20/600 20/600 HM 20/500 LP 20/125
Aspergillus niger Candida albicans Candida albicans Candida albicans Candida albicans Candida albicans Candida albicans Candida albicans
PPV Tap and Tap and Tap and Tap and Tap and Tap and Tap and
8 (left eye)
32
F
20/40
Candida albicans
Tap and inject
9 10
39 30
F F
CF CF
Candida albicans Candida albicans
11 12 13 14 15 16 (right eye) 16 (left eye) 17 18
57 22 32 23 24 59 59 26 53
M M M M F M M M M
CF LP 20/400 20/200 20/400 CF 20/400 20/100 LP
19 20 21 22 23 24 25 26 27 28 29 30
41 29 54 40 34 53 30 22 17 56 48 29
M M M M M F M F F M F M
** HM 20/50 20/50 20/50 20/300 CF 20/125 20/125 HM 20/80 CF
Organism
Initial Treatment
Subsequent Treatment (if Any)
Final VA 20/100 20/80 20/60 20/80 20/20 20/80 CF 20/80
PPV Tap and inject
n/a PPV* PPV* PPV* PPV* PPV* PPV* Second tap and inject and PPV Second tap and inject n/a PPV*
Candida albicans Candida albicans Candida albicans Candida dubliniesis Yeast Yeast Yeast Yeast Bacillus cereus
PPV PPV PPV Tap and Tap and Tap and Tap and Tap and Tap and
n/a n/a n/a PPV PPV n/a n/a PPV PPV*
20/80 HM 20/40 20/40 20/100 20/300 20/100 20/25 HM
MRSA MSSA MSSA Strep mitis Not identified Not identified Not identified Not identified Not identified Not identified Not identified Refused
Tap and Tap and Tap and Tap and Tap and PPV Tap and Tap and Tap and Tap and Tap and Refused
n/a PPV* n/a n/a n/a n/a PPV PPV n/a n/a n/a Refused
** HM 20/25 20/20 20/30 20/30 20/50 LP 20/60 20/15 20/64 20/160
inject inject inject inject inject inject inject
inject inject inject inject inject inject inject inject inject inject inject inject inject inject inject inject
Ocular Sequelae ERM with foveal traction Macular scar ERM ERM Macula-off RD with PVR Retinal tear, macular scar, ERM
20/20 20/160 NLP
Giant retinal tear, PVR, RD, ERM RD and cataract
ERM ERM Macular scar, ERM Macular scar ERM with traction Total retinal detachment, PVR Macular scar, cataract
ERM and macular scar Tractional RD, ERM ERM Macular scar
CF ¼ count fingers; ERM ¼ epiretinal membrane; F ¼ female; HM ¼ hand motions; LP ¼ light perception; M ¼ male; MRSA ¼ methicillin-resistant Staphylococcus aureus; MSSA ¼ methicillin-sensitive Staphylococcus aureus; n/a ¼ not applicable; NLP ¼ no light perception; PPV ¼ pars plana vitrectomy; PVR ¼ proliferative vitreoretinopathy; RD ¼ retinal detachment; VA ¼ visual acuity. *Positive culture from secondary PPV after initial tap and inject; **vision data not available.
etiology. Fifty-six percent of non-IVDU endogenous endophthalmitis patients had positive blood cultures.
Outcomes Mean visual acuity improved significantly in patients, from 20/800 on presentation to 20/160 on last follow-up (1.64 logMAR to 0.91 logMAR, P < 0.0001), while median visual acuity improved from 20/600 (logMAR 1.5) to 20/80 (logMAR 0.6). Patients with macular involvement had a similar course (mean acuity: 1.56 logMAR to 0.7 logMAR, P ¼ 0.0003; median acuity: 1.4 logMAR to 0.55 logMAR). At final follow-up, 90% of eyes had improved vision compared with presentation, 25% of eyes had 20/200 or worse vision, and 31% of eyes had 20/40 or better vision. Mean and median visual acuity in those who received tap and inject improved from 20/600e20/800 (1.55 logMAR) to 20/160 (0.90 logMAR, P ¼ 0.0019) and from 20/400e20/500 (1.35 logMAR) to 20/60 to 20/80 (0.55 logMAR), respectively. For those who received initial PPV, mean and median acuity improved from
20/1600e20/2000 (1.92 logMAR) to 20/160e20/200 (0.95 logMAR, unable to calculate P value due to small n), and from counting fingers (2 logMAR) to 20/80e20/100 (0.65 logMAR). Among eyes with non-IVDU endogenous endophthalmitis, the average initial vision was 20/1260e20/1600 (1.86 logMAR) and final vision was approximately 20/500 (1.42 logMAR), which was a statistically significant improvement (P < 0.001). Sixty-one percent of eyes had some improvement in vision, 18 eyes (50%) had 20/200 or worse vision, and 12 eyes (33%) had final vision of 20/40 or better. When the IVDU and non-IVDU cases were combined, there was a statistically significant improvement in vision (P ¼ 0.0038). There was not a statistically significant difference in initial vision (P ¼ 0.46) or final vision (0.26) between the IVDU and non-IVDU patients. Twenty eyes (63%) suffered ocular sequelae, with 7 eyes (22%) having more than 1 of the following problems (Table 1). Of the eyes with sequelae, 14 were associated with fungal infections, 2 with bacterial infection, and 4 with unknown organisms. Six eyes (19%) required a secondary surgery: 4 for retinal
5
Ophthalmology Retina Volume -, Number -, Month 2016 detachment repair (with 3 concurrent pars plana lensectomies) and 2 for cataract. Table 5 provides individual patient summaries.
Discussion This is the largest series of IVDU-associated endophthalmitis to date. IVDU-associated endophthalmitis is a significant source of ophthalmic morbidity, accounting for 9% of all cases of endophthalmitis at our academic referral center. Ten patients were initially misdiagnosed before presentation. Patients with IVDU-associated endophthalmitis are typically young and with minimal systemic complaints. These patients may be initially reluctant to volunteer a history of drug use; however, clinicians should have a high suspicion and aggressively pursue the possibility of IVDU in any patient with unexplained intraocular inflammation. This is especially important in younger, otherwise healthy patients, and the examination should include a search for track marks or skin abscesses. The ability to identify a causative organism varied significantly in our series based on source and method of acquisition, with a much higher yield from PPV (including secondary PPV) than from tap and inject. Endogenous endophthalmitis begins as a chorioretinitis that spreads into the vitreous; as a result, most of the microbial load is initially localized in the posterior vitreous and is therefore not accessible via bedside aspirate, especially if there is no posterior vitreous detachment, which is often the case in younger patients. Patients undergoing initial PPV had worse initial vision and may have had more severe infection with greater microbial load, which may have resulted in increased culture positivity. Additionally, the smaller volume of vitreous acquired from bedside aspirate (typically 0.1e0.2 mL) compared with PPV (typically up to 1 mL) results in decreased sensitivity. Seventy-five percent of EFE patients had positive cultures from PPV in William et al’s series.20 Connell et al’s Australian IVDU endophthalmitis cohort9 had positive vitreous cultures in 68.4% of patients (66.7% from vitreous tap and 70% from vitrectomy, all positive samples being for Candida), with the authors finding that repeat aspiration in those with initially negative cultures did not increase the yield. Conversely, vitreous cultures were positive in only 4 of 25 samples in Patel et al’s New Jerseyebased IVDU endophthalmitis cohort (2 of which were thought to be contaminants).21 Polymerase chain reaction had a 100% yield (without a difference in aqueous and vitreous samples), compared with 37.5% with standard techniques,22 in Sowmya and Madhayan’s series and can identify offending organisms in 90 minutes.9,23,24 Most reported cases of IVDU-associated endophthalmitis are fungal, with all cases in several series being attributed to fungal etiology5,7e10; however, bacterial cases constituted a significant disease burden in our series. Five eyes (16%) in our series were caused by bacteria vs. 39% (7/18) of cases in Patel et al’s series (with 11% [2/18] being of uncertain “bacterial vs. fungal etiology”).21 The greater microbial diversity seen in the Boston and New Jersey patients may reflect different microbiomes in colder and more urban environments. In contrast, bacterial etiologies constituted
6
69% of non-IVDU endogenous endophthalmitis cases at our institution. A bacterial etiology may be overlooked by clinicians evaluating patients with endogenous endophthalmitis and a history of IVDU, given the historically strong association with fungal causes.25 It is important to remain mindful of the significant number of bacterial sources of IVDU endophthalmitis. EFE tends to have a more indolent course with a longer duration of symptoms (mean 16 days) compared with EBE (4.6 days). Initial treatment may need to cover a wide spectrum of pathogens, but then should be tailored based on culture results. If a tap and inject is the initial procedure performed and if results from these cultures and blood cultures are negative, a PPV should be promptly considered to make a microbial diagnosis. Although endophthalmitis from IVDU is usually thought to be from transient bacteremia or fungemia, the morbidity of undiagnosed systemic disease is potentially high. As such, systemic disease evaluation and treatment were recommended in all patients, and 27% of cases in our series were found to have extraocular infection. There were 3 patients (3/30, 10%) who had extraocular infections without suggestive symptoms or physical examination findings. The majority of patients (5/8, 63%) with extraocular involvement were symptomatic and/or had positive extraocular examination findings. This highlights the importance of a thorough systemic work-up in all patients with endogenous endophthalmitis, including those with only ocular symptoms. Initial hospitalization is typically necessary to ensure a timely evaluation for systemic foci of infection, to initiate intravenous antibiotics, and to ensure that the involved eye is improving with treatment. Outpatient systemic evaluation for IVDU-associated endophthalmitis is conceivable, especially if the patient does not have extraocular symptoms or examination findings; however, IVDU patients are prone to poor compliance and inconsistent outpatient follow-up, which is mitigated by hospitalization. An interesting subset of patients with IVDU endophthalmitis are those misusing oral heroin substitutes such as buprenorphine.26,27 One patient (3%) in our series was injecting crushed suboxone (buprenorphine and naloxone) and required PPV after initial tap and inject due to worsening disease. Hirsbein et al reported a series of 7 patients who developed endophthalmitis while misusing buprenorphine, all of whom required vitrectomy (5/7 culture positive, all Candida species).27 Connell et al had 5 patients (26%, N ¼ 19) who used buprenorphine, 2 of whom required vitrectomy.9 Ocular morbidity was high in this series, with 63% of patients suffering at least 1 sequela of endophthalmitis and 19% of eyes (6/32) requiring a secondary surgery. Twenty eyes had initial macular involvement, with only a significant scar being noted in 7 eyes at final follow-up, indicating that many lesions may regress without significant scarring. Patients require close follow-up even after the infection is under control because they remain at risk for other ocular sequelae. There was a statistically significant improvement in mean visual acuity in patients, including among those with documented macular involvement. Interestingly, mean final visual acuity was comparable when considering all eyes
Modjtahedi et al
Intravenous Drug UseeAssociated Endophthalmitis
(20/160) and those with documented macular involvement (20/100). This may have been because macular involvement did not affect the fovea (allowing for better than expected vision) or because poor initial view to the posterior pole resulted in failure to diagnose cases of macular involvement (resulting in worse vision in those without documented initial macular involvement). Among all patients with endogenous endophthalmitis in prior series, 56% (10/18) of eyes had final acuity worse than 20/400,5 while the average acuity of EFE patients with yeast infections was 20/200 to 20/250 (logMAR 1.05).10 Initial and final vision was worse in patients with non-IVDU endogenous endophthalmitis than in patients with IVDU endophthalmitis at our institution. This difference did not reach statistical significance, potentially because of small sample size. Similarly, fewer patients (61% vs. 90%) had improvement in vision and more patients had 20/200 or worse vision (50% vs. 25%) among patients with non-IVDU endogenous endophthalmitis compared with IVDU endophthalmitis, respectively. This may have been attributable to the higher incidence of bacterial infections in the non-IVDU endogenous endophthalmitis group, which typically has a worse prognosis than fungal infections. In Connell et al’s series, 83% (15/18) of patients with visual outcome data had acuity of 6/60, 50% achieved 6/18, and mean final acuity was 6/36.9 Patel et al theorized their patients’ poor final acuity (19% of eyes having better than 20/400 vision) was because of the high frequency of initial macular involvement (11/18 eyes, 61%) and subsequent scarring.21 Only 19% of our patients had 20/400 vision or worse at final follow-up, although the percentage of patients with initial macular involvement (20/32 eyes, 63%) was similar to the series by Patel et al.21 Visual acuity outcomes vary significantly and are likely related to differences in duration of disease, microbial etiology, and secondary sequelae. Final visual acuity was similar between eyes initially treated with tap and inject vs. PPV, as was also seen by Connell et al,9 with eyes undergoing initial PPV having a more significant improvement in vision due to worse initial acuity. Although the initial treatment of choice for most patients was tap and inject (25/30, 78%), a majority (15/25, 60%) of these eyes received PPV at a later time. None of the 6 eyes that had initial PPV required repeat PPV for their acute infection. Of Connell et al’s patients, 54% (10/19) required PPV, 7 on initial presentation and 3 due to worsening disease following tap and inject,9 compared with 66% of all eyes in our series (6 initial and 15 delayed).9 There are no clear guidelines for PPV in endogenous endophthalmitis patients. Final acuity was similar in both treatment arms of our study, which is notable for those who underwent PPV, given their worse initial acuity (which presumably indicates worse disease burden). Diagnostic yield was also higher from samples attained from PPV than from tap and inject. Additionally, initial PPV may be less burdensome to patients and may hold cost advantages, because the majority of tap and inject patients eventually require vitrectomy. It is difficult to directly compare outcomes of PPV vs. tap and inject because patients were not randomized to treatment and the 2 groups were different
on presentation. Sixty-nine percent of eyes had a positive vitreous culture from a secondary PPV after having had a tap and inject (an average of 11 days after initial treatment) even though all but 1 eye received appropriate antimicrobial coverage at the time of their tap and inject. Four of the eyes with positive secondary PPV cultures after initial tap and inject underwent PPV to promote faster visual recovery and not because of worsening clinical course or need for diagnostic information; therefore, it is advisable to consider vitreous culture and intravitreal antimicrobials at the time of secondary PPV even if clinical suspicion for active infection is low. The high rate of positive vitreous cultures from these secondary PPV cases raises the possibility that intravitreal injections may not adequately treat the active infection in a sizable number of patients and provides additional support for early PPV. Trends in endophthalmitis are constantly evolving and IVDU-associated endophthalmitis represents a subset that deserves special attention. The limitations of this study include its small sample size and retrospective design. Early vitrectomy may play an important role in management. Future investigations should focus on the role and timing of PPV as well as optimizing management and systemic disease evaluation.
References 1. Durand ML. Endophthalmitis. Clin Microbiol Infec. 2013;19(3):227e234. 2. Kostick DA, Foster RE, Lowder CY, et al. Endogenous endophthalmitis caused by Candida albicans in a healthy woman. Am J Ophthalmol. 1992;113(5):593e595. 3. Mamandhar A, Bajracharya L. Endogenous Aspergillus endophthalmitis in a healthy individual. Nepal J Ophthalmol. 2012;4(1):179e183. 4. Valluri S, Moorthy RS, Liggett PE, Rao NA. Endogenous Aspergillus endophthalmitis in an immunocompetent individual. Int Ophthalmol. 1993;17(3):131e135. 5. Schiedler V, Scott IU, Flynn Jr HW, et al. Culture-proven endogenous endophthalmitis: clinical features and visual acuity outcomes. Am J Ophthalmol. 2004;137(4):725e731. 6. Dozier CC, Tarantola RM, Jiramongkolchai K, Donahue SP. Fungal eye disease at a tertiary care center: the utility of routine inpatient consultation. Ophthalmology. 2011;118(8): 1671e1676. 7. Leibovitch I, Lai T, Raymond G, et al. Endogenous endophthalmitis: a 13-year review at a tertiary hospital in South Australia. Scand J Infect Dis. 2005;37(3):184e189. 8. Chakrabarti A, Shivaprakash MR, Singh R, et al. Fungal endophthalmitis: fourteen years’ experience from a center in India. Retina. 2008;28(10):1400e1407. 9. Connell PP, O’Neill EC, Amirul Islam FM, et al. Endogenous endophthalmitis associated with intravenous drug abuse: seven-year experience at a tertiary referral center. Retina. 2010;30(10):1721e1725. 10. Sridhar J, Flynn Jr HW, Kuriyan AE, et al. Endogenous fungal endophthalmitis: risk factors, clinical features, and treatment outcomes in mold and yeast infections. J Ophthalmic Inflamm Infect. 2013;3(1):60. 11. Connell PP, O’Neill EC, Fabinyi D, et al. Endogenous endophthalmitis: 10-year experience at a tertiary referral centre. Eye (Lond). 2011;25(1):66e72.
7
Ophthalmology Retina Volume -, Number -, Month 2016 12. Lim HW, Shin JW, Cho HY, et al. Endogenous endophthalmitis in the Korean population: a six-year retrospective study. Retina. 2014;34:592e602. 13. Shen X, Xu G. Vitrectomy for endogenous fungal endophthalmitis. Ocul Immunol Inflamm. 2009;17(3):148e152. 14. Wong JS, Chan TK, Lee HM, Chee SP. Endogenous bacterial endophthalmitis: an east Asian experience and a reappraisal of a severe ocular affliction. Ophthalmology. 2000;107(8):1483e1491. 15. Zhang H, Liu Z. Endogenous endophthalmitis: a 10-year review of culture-positive cases in northern China. Ocul Immunol Inflamm. 2010;18(2):133e138. 16. Jones CM, Logan J, Gladden M, Bohm MK. Vital signs: demographic and substance use trends among heroin usersd United States, 2002e2013. Morb Mortal Wkly Rep. 2015;64(26):719e725. 17. Jones CM. Heroin use and heroin use risk behaviors among nonmedical users of prescription opioid pain relieversdUnited States, 2002e2004 and 2008e2010. Drug Alcohol Depend. 2013;132(1-2):95e100. 18. Holladay JT. Proper method for calculating average visual acuity. J Refract Surg. 1997;13(4):388e391. 19. van de Put MA, Hoeksema L, Wanders W, et al. Postoperative vision-related quality of life in macula-off rhegmatogenous retinal detachment patients and its relation to visual function. PloS One. 2014;9(12):e114489. 20. William A, Spitzer MS, Deuter C, et al. Outcomes of primary transconjunctival 23-gauge vitrectomy in the diagnosis and
21. 22.
23. 24.
25.
26. 27.
treatment of presumed endogenous fungal endophthalmitis. Ocul Immunol Inflamm. 2016:1e7. Patel SN, Rescigno RJ, Zarbin MA, Langer P, Bhagat N. Endogenous endophthalmitis associated with intravenous drug abuse. Retina. 2014;34(7):1460e1465. Sowmya P, Madhavan HN. Diagnostic utility of polymerase chain reaction on intraocular specimens to establish the etiology of infectious endophthalmitis. Eur J Ophthalmol. 2009;19(5):812e817. Goldschmidt P, Degorge S, Benallaoua D, et al. New test for the diagnosis of bacterial endophthalmitis. Br J Ophthalmol. 2009;93(8):1089e1095. Hidalgo JA, Alangaden GJ, Eliott D, et al. Fungal endophthalmitis diagnosis by detection of Candida albicans DNA in intraocular fluid by use of a species-specific polymerase chain reaction assay. J Infect Dis. 2000;181(3): 1198e1201. Jackson TL, Eykyn SJ, Graham EM, Stanford MR. Endogenous bacterial endophthalmitis: a 17-year prospective series and review of 267 reported cases. Surv Ophthalmol. 2003;48(4):403e423. Aboltins CA, Allen P, Daffy JR. Fungal endophthalmitis in intravenous drug users injecting buprenorphine contaminated with oral Candida species. Med J Aust. 2005;182(8):427. Hirsbein D, Attal P, Gueudry J, et al. [Endogenous Candida endophthalmitis in drug misusers injecting intravenous buprenorphine]. J Fr Ophtalmol. 2008;31(2):180e183.
Footnotes and Financial Disclosures Originally received: September 12, 2016. Final revision: October 24, 2016. Accepted: October 25, 2016. Available online: ---. Manuscript no. ORET_2016_30. 1
Retina Service, Department of Ophthalmology, Massachusetts Eye and Ear Infirmary, Harvard Medical School, Boston, Massachusetts.
2
Infectious Disease Service, Massachusetts Eye and Ear Infirmary, Boston, Massachusetts.
Author Contributions: Research design: Modjtahedi, Finn, Papakostas, Husain, Eliott Data analysis and/or interpretation: All authors (Modjtahedi, Finn, Papakostas, Durand, Husain, Eliott) Data acquisition and/or research execution: All authors (Modjtahedi, Finn, Papakostas, Durand, Husain, Eliott) Manuscript preparation: All authors (Modjtahedi, Finn, Papakostas, Durand, Husain, Eliott)
3
Department of Medicine, Massachusetts General Hospital, Boston, Massachusetts. *Drs Modjtahedi and Finn contributed equally to this work. Financial Disclosure(s): The author(s) have made the following disclosure(s): There were no funding sources for this project. The authors have no conflicts of interest to report. B.S.M., A.V.F., T.D.P., M.D., D.H.: No financial disclosures. D.E.: Consultant d Alcon, Alimera, Allergan, Arctic, Avalanche, Dutch Ophthalmic, Ophthotech, MacuLogix, Santen, and Thrombogenics; Research grants d Neurotech and Ocata Therapeutics.
8
Abbreviations and Acronyms: EBE ¼ endogenous bacterial endophthalmitis; EFE ¼ endogenous fungal endophthalmitis; HIV ¼ human immunodeficiency virus; IVDU ¼ intravenous drug use; MEEI ¼ Massachusetts Eye and Ear Infirmary; PPV ¼ pars plana vitrectomy. Correspondence: Dean Eliott, MD, Retina Service, Department of Ophthalmology, Massachusetts Eye and Ear Infirmary, Harvard Medical School, Boston, MA 02114. E-mail: [email protected].