In endogenous fungal endophthalmitis, the systemic picture drives treatment and the eye does not

Design

multi-institutional retrospective cohort study across eight US academic centres

Population

118 eyes of 87 patients with clinically or culture-proven endogenous fungal endophthalmitis, 2010 to 2025

Primary outcome

treatment strategies, need for surgery and final visual acuity

Effect

surgery 21% after intravenous vs 50% after oral (adjusted OR 0.33, 95% CI 0.1 to 1.06); mould infection OR 4.26 and worse presenting vision OR 1.78 for worse final vision

Eight US academic institutions pooled 118 eyes of 87 patients with clinically or culture-proven endogenous fungal endophthalmitis between 2010 and 2025 — a condition with no standardised treatment guideline.

Sixty-four per cent were started on intravenous systemic antifungals, most often voriconazole (34 per cent) or fluconazole (30 per cent). What determined that route was systemic: symptoms, positive systemic cultures, and negative intraocular fluid cultures. Ocular examination findings did not significantly influence the initial route, the need for surgery, or final vision. Eyes started on intravenous therapy were less likely to need surgery (21 per cent vs 50 per cent, P = .001), but the effect lost significance after adjustment (OR 0.33, 95% CI 0.1 to 1.06, P = .067). Worse presenting vision (OR 1.78, P < .001) and mould rather than yeast (OR 4.26, P < .001) predicted worse final vision.

The finding that ocular findings do not drive the systemic route is an observation about practice, not a recommendation. It probably reflects that these patients are usually admitted under medicine or infectious diseases with candidaemia, and the eye is found second. The two prognostic variables — presenting vision and mould — are the ones worth carrying into the counselling conversation, and the mould odds ratio of 4.26 argues for pushing hard on species identification early.

• Push for species identification early — mould carried more than four times the odds of a poor visual outcome
• Record presenting visual acuity carefully; it is the other independent prognostic variable
• Do not read the borderline intravenous-versus-oral result as evidence either way
• Expect the systemic team to set the route, and make sure the ocular findings reach them anyway
• In a patient with candidaemia, examine both fundi rather than waiting for symptoms

Why it matters

It gives a condition with no guideline two prognostic variables that can be established on day one.

Don't overread it

Retrospective and non-randomised — the apparent advantage of intravenous therapy did not survive adjustment for confounders.

The statistics, in plain English

The intravenous-versus-oral comparison is the clearest example here of confounding by indication: sicker patients get intravenous therapy, so the crude 21 per cent versus 50 per cent difference in surgery cannot be read as the drug route working. Once adjusted, the confidence interval crosses 1 (0.1 to 1.06) and the finding is inconclusive rather than negative.

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.