Each hour's delay to appropriate antifungal therapy in Candida septic shock was linked to lower survival

Design

Retrospective multicentre cohort

Population

850 adults with culture-positive Candida septic shock at 32 hospitals in Canada, the US and Saudi Arabia

Primary outcome

Hospital survival by delay to pathogen-appropriate antifungal therapy

Effect

Adjusted OR for death 1.045 per hour (95% CI 1.035 to 1.056); 86% survival if treated within 2 hours

This cohort drew on nearly three decades of records from 32 hospitals and timed the delay from the onset of persistent hypotension to the first antifungal active against the isolated Candida species.

The delay was the strongest predictor of death in adjusted analysis, ahead of illness severity. The effect was consistent across infection sites, species and drug classes. Most patients were treated late: the median delay was more than a day.

The data are observational and old, and some patients who were treated late may have had less obvious disease. The message still lines up with the bacterial septic shock literature from the same group: in high-risk patients, antifungal cover belongs in the first hours.

• Survival was 86% when an appropriate antifungal was given within two hours of hypotension.
• Survival fell about 5% for each hour of delay up to 10 to 12 hours.
• The odds of death rose 4.5% per hour of delay after adjustment.
• Only 17% received an appropriate antifungal within six hours.
• Know your unit's Candida species and resistance patterns, including Candida auris, when choosing the first agent.

Why it matters

The median delay was 32 hours, and the survival curve suggests most of the benefit is lost long before cultures turn positive.

Don't overread it

This retrospective cohort spans 1989 to 2018; it shows an association with timing, not proof that earlier treatment caused the difference.

The statistics, in plain English

An odds ratio of 1.045 per hour means the odds of death rise by about 4.5% for every hour; over ten hours that compounds substantially. The interval is narrow, but residual confounding is possible.

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.