The research, published in BMC Infectious Diseases by a team at Jeonbuk National University, took advantage of one of the world’s most comprehensive health databases, the Korean National Health Insurance Service. The investigators identified 31,175 patients diagnosed with tuberculosis between 2002 and 2013 and compared them with 84,522 matched individuals who had no tuberculosis diagnosis. To ensure the two groups were genuinely comparable, the team used propensity score matching, a statistical technique that balances factors such as age, sex, and underlying illnesses between the cohorts. They then tracked how often either group received a diagnosis of optic neuropathy, the umbrella term for damage to the optic nerve that can blur vision, drain color perception, and in severe cases cause permanent loss of sight.

The headline result was sobering. Even after adjusting for age, sex, comorbidities, and ethambutol exposure, tuberculosis patients were 28 percent more likely than matched controls to be diagnosed with optic neuropathy, with an adjusted hazard ratio of 1.28 and a 95 percent confidence interval of 1.05 to 1.56. That elevation alone would be notable, but the more striking pattern emerged when the researchers stratified patients by how much ethambutol they had received. Compared with tuberculosis patients who had not been exposed to the drug, those given less than 10 milligrams per kilogram of body weight per day had nearly double the risk, at a hazard ratio of 1.94. For doses of at least 10 but less than 15 milligrams per kilogram per day, the risk nearly tripled, reaching 2.98. And at the highest exposure category, 15 milligrams per kilogram per day or more, the hazard ratio soared to 6.98, a nearly sevenfold increase in the likelihood of an optic neuropathy diagnosis.

This dose-dependent gradient is precisely the kind of signal epidemiologists look for when trying to distinguish a genuine drug effect from statistical noise or confounding. If optic neuropathy diagnoses among tuberculosis patients were driven mainly by their underlying illness, their age, or their greater contact with the health care system, one would not necessarily expect risk to scale so cleanly with the milligram-per-kilogram dose of a single medication. The authors argue that the pattern strengthens the case that ethambutol itself is a major driver of the observed nerve damage, a conclusion consistent with decades of smaller clinical reports describing ethambutol-induced optic neuropathy, known in the literature as EON.

The mechanism behind EON has been studied for more than half a century, though it remains incompletely understood. Ethambutol is thought to injure the retinal ganglion cells and the axons of the optic nerve, particularly the small fibers that carry signals from the center of the visual field, which is why patients classically lose color vision and central acuity before they notice broader visual deficits. Zinc chelation and mitochondrial dysfunction have both been proposed as contributing pathways, and the damage often develops insidiously over weeks to months of continuous exposure. What makes the condition clinically treacherous is that it can be irreversible: once the optic nerve fibers degenerate, they do not regenerate, so the only reliable defense is early detection and prompt discontinuation of the drug.

Age emerged as the other dominant risk factor in the Korean data. Patients aged 50 or older faced more than twice the risk of optic neuropathy compared with younger patients, with a hazard ratio of 2.27 and a confidence interval of 1.84 to 2.79. That finding carries real public health weight in a country like South Korea, where the population is aging rapidly and where tuberculosis, though declining, remains concentrated among older adults. Older patients may have reduced renal clearance of the drug, pre-existing subclinical optic nerve vulnerability, or comorbidities such as diabetes and chronic kidney disease that compound the injury. The study adjusted for hypertension, diabetes, and chronic kidney disease among other conditions, yet the age effect persisted, suggesting that aging itself, independent of these measurable diseases, magnifies the eye’s susceptibility to the drug.

Perhaps the most counterintuitive finding in the study is temporal. Even as tuberculosis incidence in Korea steadily declined over the study window, the number of optic neuropathy diagnoses rose markedly after 2009. The authors do not claim that ethambutol alone explains this trend, and claims data cannot fully disentangle why diagnoses increased. Several forces could be at work simultaneously: greater awareness of EON among physicians, expanded access to ophthalmic evaluation, an aging patient population receiving higher cumulative doses, or longer treatment courses for drug-resistant and non-tuberculous mycobacterial infections, for which ethambutol is also widely used. Whatever the mixture of causes, the trend suggests that the burden of drug-associated optic nerve injury is not shrinking in step with the tuberculosis epidemic itself, and may in fact be growing in absolute terms.

The study’s design has strengths and limitations worth weighing. Its scale is a genuine asset: nationwide claims data spanning more than a decade capture a far broader and more representative patient population than any single hospital series, and the propensity-matched control group provides a credible counterfactual for what optic neuropathy rates would look like without tuberculosis or its treatment. On the other hand, administrative claims rely on diagnostic codes from the International Classification of Diseases, tenth revision, rather than on direct ophthalmic examination, so some diagnoses may be misclassified and some cases of subclinical EON may never be coded at all. Exact ethambutol doses and durations recorded in claims may also be imperfect proxies for actual drug exposure. The authors are careful to frame their conclusions in terms of claims-defined optic neuropathy, an honest hedge that does not, however, blunt the central message of a strong dose-response relationship.

What should change in clinical practice? The authors call for strengthened pharmacovigilance and the routine integration of ophthalmic monitoring into tuberculosis management, and the data give that recommendation concrete shape. Patients over 50, those receiving doses at or above 15 milligrams per kilogram per day, and those on prolonged treatment courses would be the natural priority groups for baseline and interval eye examinations, including visual acuity testing, color vision assessment, and where available optical coherence tomography to measure the retinal nerve fiber layer. Clinicians also face a delicate balancing act, because ethambutol remains essential for multidrug regimens and reducing the dose or shortening the course carries its own risks of treatment failure and resistance. The goal is not to abandon the drug but to watch the eyes of the patients most likely to be harmed while the drug does its work against the infection.

The broader lesson extends beyond a single medication. As infectious disease programs succeed in shrinking their target epidemics, the residual patient population tends to be older, sicker, and more susceptible to the toxicities of the very drugs that made the success possible. Tuberculosis is a case in point, but the same dynamic shadows other treatment fields, from antiretroviral therapy to chemotherapy. The Korean cohort is a reminder that the endgame of an epidemic is not only about counting fewer infections; it is about safeguarding the quality of life of the patients who still need treatment. In this case, that means making eye care an explicit, funded, and routine part of tuberculosis therapy, before avoidable and irreversible vision loss becomes the quiet price of a cure.

Subject of Research: Ethambutol-induced optic neuropathy risk among tuberculosis patients in a nationwide Korean cohort

Article Title: Optic neuropathy and ethambutol exposure in a declining tuberculosis population: a nationwide cohort study in Korea

Article References: Yeo, C. D., Kim, J. S., Lee, T.-E., & Lee, H.-J. (2026). Optic neuropathy and ethambutol exposure in a declining tuberculosis population: a nationwide cohort study in Korea. BMC Infectious Diseases. https://doi.org/10.1186/s12879-026-14547-1

Image Credits: AI Generated

DOI: 10.1186/s12879-026-14547-1

Keywords: tuberculosis, ethambutol, optic neuropathy, drug-induced vision loss, pharmacovigilance, ophthalmic monitoring, nationwide cohort study, Korea, dose-response, aging population, National Health Insurance Service, BMC Infectious Diseases