In endogenous fungal endophthalmitis, the systemic illness decides the route - not the eye

Design

Retrospective multi-institutional cohort, eight US academic centres, 2010-2025

Population

118 eyes of 87 patients with clinically or culture-proven endogenous fungal endophthalmitis

Primary outcome

Treatment strategy, need for surgery and final visual acuity

Effect

Surgery 21 per cent after intravenous vs 50 per cent after oral therapy, adjusted OR 0.33 (95 per cent CI 0.10 to 1.06); mould infection OR 4.26 for worse final vision

Eight US academic centres pooled 118 eyes in 87 patients with clinically or culture-proven endogenous fungal endophthalmitis treated between 2010 and 2025. Sixty-four per cent were started on intravenous systemic antifungal therapy, most often voriconazole (34 per cent) or fluconazole (30 per cent).

What determined that choice is the interesting part. Patients with systemic symptoms, positive systemic cultures or negative intraocular fluid cultures were the ones started intravenously. Ocular examination findings did not significantly influence the initial route, nor the need for surgery, nor final vision. Eyes started intravenously came to surgery less often on the unadjusted comparison, 21 per cent against 50 per cent (P = 0.001), but once confounders were controlled the effect lost significance: OR 0.33 (95 per cent CI 0.10 to 1.06, P = 0.067). Worse presenting vision (OR 1.78, P < 0.001) and infection with a mould rather than a yeast (OR 4.26, P < 0.001) predicted worse final vision.

There are no standardised guidelines for this disease, and this cohort does not supply one. Its practical value is prognostic and procedural: identify the organism to genus, because mould versus yeast is the single strongest signal you have, and record presenting acuity properly because it carries real predictive weight. The intravenous-versus-oral question remains open, and an odds ratio whose interval brushes 1.06 is a reason to keep asking it, not a reason to switch practice.

• Push for species-level identification; mould infection carried more than four times the odds of a poor visual outcome.
• Record presenting best-corrected acuity carefully - it is a genuine prognostic variable here, not a formality.
• Take the systemic work-up seriously; blood cultures and systemic symptoms drove the treatment route in practice.
• Do not let a quiet-looking eye reassure you into oral therapy; ocular findings predicted nothing in this cohort.
• Agree the antifungal route jointly with infectious diseases rather than deciding it from the slit lamp.

The statistics, in plain English

The surgery difference is the cautionary example: a clear unadjusted gap of 21 against 50 per cent shrank to an odds ratio of 0.33 whose interval reaches 1.06, meaning no effect cannot be excluded once sicker patients are accounted for. That is what confounding by indication looks like. The mould and presenting-vision associations are the more trustworthy findings, but this is still a retrospective cohort, so they identify who does badly rather than what to change.

Read the rest in the app

You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.

Get Daily Doctor — free

Scan to keep reading on your phone. No account needed to start.