Abstract
This medical record review study compares the clinical characteristics of endogenous endophthalmitis in patients with and without injection drug use.
The United States experienced a tripling in the number of opioid overdose deaths from 2000 to 2014. In particular, the Northeast region has the highest age-adjusted rate of deaths associated with drug use, at 16.1 per 100 000 persons.1 In New Hampshire, the rate is even higher, at 26.2 per 100 000 persons.1
A devastating sequela of injection drug use (IDU) is endogenous endophthalmitis (EE).2 Injection drug use can lead to transient microbial bloodstream infection from use of a nonsterile injection apparatus, which can seed ocular infection.3 The prognosis of EE is poor, with nearly 50% of affected eyes having no light perception despite treatment.2 We investigated clinical characteristics of IDU vs non-IDU EE at Dartmouth-Hitchcock Medical Center (DHMC), Lebanon, New Hampshire, during the opioid epidemic.
Methods
A retrospective, manual medical record review identified EE cases from January 1, 2012, through December 5, 2016, at DHMC with codes 360.0 and 360.1 from the International Classification of Diseases, 9th Revision, and codes H44.0 and H44.1 from the International Statistical Classification of Diseases and Related Health Problems, 10th Revision. Patient demographics, IDU history, microbial data, and clinical courses were recorded. Patients were classified as having EE if they had intraocular inflammation without recent eye surgery, had trauma with a positive intraocular microbial culture or concurrent positive blood culture, or experienced clinical resolution after treatment with intraocular antibiotics or antifungals. The study was approved by the committee for protection of human subjects at DHMC with adherence to Declaration of Helsinki tenets.4 The committee waived the need for informed consent for this medical record review. Statistical analysis was performed using χ2 and paired and unpaired t tests, with no adjustment to the P values to take into account multiple comparisons.
Results
From 2012 through 2016, 15 patients with EE were identified (5 male and 10 female; age range, 24-83 years), of whom 9 (56.3%) had a history of IDU. During the study period, the number of non-IDU EE cases remained at 0 to 2 cases per year, whereas the number of IDU EE cases per year ranged from 0 to 4 per year, with the most IDU cases occurring in 2016.
Reduced vision was the most common presenting symptom, including all 9 patients with IDU EE and 5 of 6 patients with non-IDU EE. Less common symptoms included pain, floaters, photophobia, and conjunctival injection. The patients with IDU were younger, had fewer comorbidities, and tended to delay seeking medical care (Table). Patients without IDU were more likely to be encountered during hospitalization or shortly after discharge and less likely to warrant surgical intervention because of more frequent resolution of vitritis. More negative intraocular and blood culture findings for EE occurred among patients with IDU, and patients with IDU experienced significantly more improvement in visual acuity after intervention than did patients without IDU (Figure).
Table. Comparison of Clinical Characteristics of Patients With IDU and Non-IDU EEa .
| Characteristic | IDU EE Group (9 Patients/11 Eyes) |
Non-IDU EE Group (6 Patients/7 Eyes) |
Difference Between Groups (95% CI) | P Value |
|---|---|---|---|---|
| Age, y | ||||
| Mean (SE) | 31 (2.3) | 63 (4.0) | 32.0 (22.7 to 41.3) | <.001 |
| Range | 24-42 | 55-83 | NA | NA |
| Sex, No. (%) of patients | ||||
| Male | 2/9 (22) | 3/6 (50) | 27.8 (−26.0 to 70.1) | .28 |
| Female | 7/9 (78) | 3/6 (50) | 27.8 (−26.0 to 70.1) | .28 |
| Chronic comorbidities, No. (%) of patients | 4/9 (44) | 6/6 (100) | 55.6 (−1.76 to 86.3) | .03 |
| Time from symptom onset to presentation, d | ||||
| Mean (SE) | 24.7 (9.2) | 2.0 (1.1) | 22.7 (−2.00 to 47.4) | .07 |
| Range | 5-90 | 1-11 | NA | NA |
| BCVA, mean (SE), logMAR [Snellen equivalent]b | ||||
| Presenting | 1.39 (0.27) [20/491] | 0.95 (0.20) [20/178] | 0.44 [22] (−0.36 to 1.24) [−64 to 20] | .26 |
| Final | 0.82 (0.21) (20/132) | 1.05 (0.19) (20/222) | 0.23 [11] (−0.42 to 0.87) [−22 to 45] | .47 |
| Eyes treated with surgical intervention, No. (%) | 7/11 (64) | 2/7 (29) | 45.5 (−9.04 to 76.2) | .07 |
| Positive culture result, No. (%) of eyes | ||||
| Intraocular | 3/11 (27) | 2/7 (29) | 9.1 (−39.3 to 46.9) | .67 |
| Anterior chamber tap | 0/1 | 1/2 (50) | 50 (−59.0 to 98.7) | .48 |
| Vitreous tap | 3/7 (43) | 1/4 (25) | 17.9 (−46.7 to 63.7) | .57 |
| Vitrectomy with vitreous | 2/4 (50) | 1/2 (50) | 0 (−65.2 to 65.2) | >.99 |
| Blood | 2/11 (18) | 4/7 (57) | 39.0 (−12.3 to 75.6) | .10 |
| Combined intraocular and blood | 5/11 (45) | 6/7 (86) | 40.2 (−13.4 to 72.1) | .10 |
| Cases with bacterial involvement, No. (%) of eyes | 5/11 (45) | 6/7 (86) | 40.2 (−13.4 to 72.1) | .10 |
| Cases with fungal involvement, No. (%) of eyes | 5/11 (45) | 1/7 (14) | 31.2 (−21.0 to 65.3) | .18 |
| Cases with mixed bacterial and fungal involvement, No. (%) of eyes | 4/11 (36) | 1/7 (14) | 22.0 (−28.5 to 57.7) | .32 |
| Consultation inpatient or <1 wk of inpatient discharge, No. (%) of patients | 2/9 (22) | 5/6 (83) | 61.1 (0.4 to 86.4) | .02 |
| Left against medical advice during evaluation, No. (%) of patients | 2/9 (22) | 0/6 | 22.2 (−27.7 to 60.0) | .23 |
Abbreviations: BCVA, best-corrected visual acuity; EE, endogenous endophthalmitis; IDU, injection drug use; NA, not applicable.
Patients presented with EE to the Dartmouth-Hitchcock Medical Center from January 1, 2012, through December 31, 2016. P values were calculated using the χ2 test or unpaired t test.
For reference, logMAR = 0 is the equivalent of Snellen visual acuity of 20/20; logMAR 1.0, 20/200. Differences are expressed as logMAR (approximate Early Treatment Diabetic Retinopathy Study letter score, 85 – 50 × logMAR).5
Figure. Mean Visual Acuity Before and After Treatment in Patients With and Without Injection Drug Use (IDU).
Includes 18 eyes, 11 of patients with IDU (61%) and 7 of patients without IDU (39%). Error bars represent SE. BCVA indicates best-corrected visual acuity.
aP = .02, paired t test, compared with after follow-up.
bP = .59, paired t test, compared with after follow-up.
Discussion
The peak in IDU EE cases observed in 2016 mirrors the increased in IDU-related deaths in New Hampshire. Patients with IDU EE were young and ambulatory and presented later but were more likely to experience improved vision with treatment compared with patients with non-IDU EE, who fared worse likely because of more chronic comorbidities and advanced age. In addition, microbes were less likely to be recovered in patients with IDU. These observations support the notion that patients with IDU subvert typical patterns of self-care by normalizing injection-related harms and delaying medical treatment until emergencies occur.6 This delay also makes microbial diagnosis challenging, because the injection-induced transient bacteremia may have resolved by the time of presentation. Furthermore, organisms may be sequestered in tissues difficult to access even with repeated sampling, which may contribute to the frequent nonclearing vitritis seen in patients with IDU and account for the observed higher need for surgical intervention.7 Although further interpretation is limited by the small sample size at a single hospital, the contrast in initial clinical impression between patients with and without IDU was substantial. The patients with IDU represent a younger and healthier subset of the population with EE and may regain vision with prompt recognition and treatment.
References
- 1.Rudd RA, Aleshire N, Zibbell JE, Gladden RM. Increases in drug and opioid overdose deaths: United States, 2000-2014. MMWR Morb Mortal Wkly Rep. 2016;64(50-51):1378-1382. [DOI] [PubMed] [Google Scholar]
- 2.Wu ZH, Chan RP, Luk FO, et al. Review of clinical features, microbiological spectrum, and treatment outcomes of endogenous endophthalmitis over an 8-year period. J Ophthalmol. 2012;2012:265078. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Shamsuddin D, Tuazon CU, Levy C, Curtin J. Bacillus cereus panophthalmitis: source of the organism. Rev Infect Dis. 1982;4(1):97-103. [DOI] [PubMed] [Google Scholar]
- 4.World Medical Association World Medical Association Declaration of Helsinki: ethical principles for medical research involving human subjects. JAMA. 2013;310(20):2191-2194. doi: 10.1001/jama.2013.281053 [DOI] [PubMed] [Google Scholar]
- 5.Gregori NZ, Feuer W, Rosenfeld PJ. Novel method for analyzing Snellen visual acuity measurements. Retina. 2010;30(7):1046-1050. [DOI] [PubMed] [Google Scholar]
- 6.Morrison A, Elliott L, Gruer L. Injecting-related harm and treatment-seeking behaviour among injecting drug users. Addiction. 1997;92(10):1349-1352. [PubMed] [Google Scholar]
- 7.Tirpack AR, Duker JS, Baumal CR. An outbreak of endogenous fungal endophthalmitis among intravenous drug abusers in New England. JAMA Ophthalmol. 2017;135(6):534-540. [DOI] [PMC free article] [PubMed] [Google Scholar]
