This prospective observational cohort study was conducted at Taichung Veterans General Hospital (TCVGH) from May 2016 to April 2025. The TCVGH Institutional Review Board and Ethics Committee approved the study (IRB number CE16096A), and all research procedures aligned with the Declaration of Helsinki and adhered to relevant guidelines and regulations. Informed consent was obtained from each patient.

Patients who were aged 20 years or older and were treated with EMB for mycobacterial infection were enrolled in this study. Those who were unable to complete the visual measurement procedures and transferred out during the study period were excluded. Mycobacterial infections included pulmonary and extrapulmonary TB, multidrug-resistant TB, and NTM diseases that were diagnosed and treated on the basis of local and international recommendations [11,12,13].

Figure 1 presents the flowchart of visual assessment procedures for patients. After EMB was prescribed as part of the treatment regimen for a patient with mycobacterial infection, the physician in charge instructed the patient to report any newly perceived subjective ocular abnormalities (e.g., blurred vision and eye pain) to the hospital’s TB case manager and the township public health center’s DOT provider at any time. At each visit, physicians asked whether the patient had experienced any newly perceived subjective ocular abnormalities. Additionally, at each visit following EMB prescription, the hospital’s medical information system sent a compulsory pop-up message to patients to collect objective visual function measurements at the TB health education room. The DOT provider measured the patient’s objective visual function once per week until cessation of EMB treatment. The objective visual function measurements included visual acuity (determined through Landolt’s broken ring test) and color vision (determined through an Ishihara color test). If a patient reported a newly perceived subjective ocular abnormality or if any impairment in objective visual function was detected, EMB was immediately discontinued. Impaired visual acuity was defined as a decrease of ≥ 2 chart lines in either eye per patient; impaired color discrimination was defined as reading < 10 plates correctly in either eye for patients with normal baseline color discrimination or reading ≥ 2 fewer plates correctly in either eye for patients with deficient baseline color discrimination [14].

Flowchart of programmatic assessment of EON. This figure illustrates that a collaborative multidisciplinary team of pulmonologists, TB case managers, DOT providers, and ophthalmologists was established to implement a programmatic monitoring and evaluation of EON. DOT directly observed therapy, EMB ethambutol, EON ethambutol-related optic neuropathy, HIS health information system, TB tuberculosis

Following EMB discontinuation due to subjective and/or objective ocular abnormalities, patients visited the ophthalmologists, who evaluated the etiology of the ocular abnormality through a series of optic examinations, including the Landolt broken ring test, the Ishihara color test, the 30-2 Humphrey visual field test, optical coherence tomography, and fundoscopy.

Our goal was to assess the incidence of, risk factors for, short-term (6-month) outcomes of, and long-term (5-year) outcomes of EON through a prospective programmatic study. EON was diagnosed on the basis of the following criteria: subjective and/or objective visual abnormalities emerged only after the initiation of EMB use; at least one major criterion and two minor criteria for EON diagnosis were met; and other ophthalmologic conditions that may cause visual loss (e.g., cataracts, glaucoma, retinal disease, or diabetic retinopathy) or optic nerve abnormalities (e.g., glaucoma, optic neuritis, or hereditary or acquired optic neuropathy) were excluded. The major criteria are outlined as follows: (1) abnormal Landolt broken ring test results in the absence of other reasonable explanations for abnormal visual acuity, (2) deficient Ishihara color test results in the absence of other reasonable explanations for abnormal color vision, and (3) central or paracentral scotoma on the 30-2 Humphrey visual field test. The minor criteria are outlined as follows: (1) optic disc pallor and (2) visual field defects other than central or paracentral scotomas [15,16,17,18,19,20]. Each patient who received a diagnosis of EON participated in monthly follow-up at the eye clinic for 6 months and a telephone call at 5 years following EMB discontinuation to clarify the short- and long-term outcomes of EON.

Full recovery was defined as the restoration of baseline visual function. Partial recovery was defined as an improvement of ≥ 2 chart lines on the Landolt broken ring test and an increase of ≥ 2 correctly read plates on the Ishihara color test compared with the results at the time of EMB discontinuation in any affected eye. Both full recovery and partial recovery were categorized as favorable outcomes; all other outcomes were considered unfavorable. The unfavorable outcomes were further categorized as unchanged and worse. Worse outcomes were defined as a decrease of ≥ 2 chart lines on the Landolt broken ring test and ≥ 2 fewer correctly read plates on the Ishihara color test in any affected eye after EMB discontinuation; all other unfavorable outcomes were considered unchanged.

In addition to the ocular outcomes of interest, information on the baseline demographic characteristics, diagnosis, therapeutic regimen, comorbidities, relevant laboratory findings, start and end dates of EMB use, and EMB dosage was collected for each patient through patient record forms.

Electronic data were stored on a password-protected personal computer located at TCVGH. SPSS software (version 22.0; SPSS, Chicago, IL, USA) was used for data analysis. Continuous variables are presented as mean and corresponding standard deviation (SD). A paired-samples t test was performed to compare continuous variables between the study groups. Categorical variables are presented as frequencies and percentages and were compared between groups by using a chi-squared test. A two-sided P value of < 0.05 was considered significant. Multiple logistic regression analysis was performed to determine the factors associated with EON, and differences were determined using odds ratios (ORs) and 95% confidence intervals (CIs).