- Design
- Single-centre, prospective, randomised controlled trial
- Population
- 100 analysed ICU patients with sepsis (136 enrolled)
- Primary outcome
- Change in SOFA score after 5 days
- Effect
- No significant difference; 28-day mortality 26% vs 36% (P=0.387)
In a single-centre randomised trial in a Chinese ICU, patients with sepsis received standard care with or without dapagliflozin 10 mg daily for five days; 50 per group were analysed.
Change in SOFA score did not differ between groups. On day five, median ejection fraction was 60% with dapagliflozin against 55% without. Mortality at 28 days was 26% against 36%, which was not statistically significant. Urine ketones were higher with dapagliflozin, although no ketoacidosis occurred.
This is a small pilot trial with a negative primary outcome. It does not support starting an SGLT2 inhibitor for sepsis, and the ketone signal is a reminder of why these drugs are usually paused in acute illness.
- Dapagliflozin did not reduce organ dysfunction in sepsis in this trial.
- Do not start SGLT2 inhibitors to treat sepsis outside a trial.
- Urine ketones rose on dapagliflozin — check ketones in septic patients already taking an SGLT2 inhibitor.
- The mortality difference was not significant and could be chance.
Why it matters
SGLT2 inhibitors have been proposed for almost every organ under stress, and this is an early negative test in sepsis.
Don't overread it
The higher ejection fraction is a secondary measure in a 100-patient trial and says nothing about survival.
The statistics, in plain English
With 50 patients per group, a 10-point mortality difference (P=0.387) is well within what chance alone produces; a trial many times larger would be needed to test mortality.
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.
Scan to keep reading on your phone. No account needed to start.
Tomorrow morning, before your first patient
One edition a day for infectious diseases, written by the desk, every claim tied to its paper. Six minutes.
Get the app — free