Empagliflozin raised sodium excretion in diabetic patients with refractory cirrhotic ascites

Design

Randomised, open-label proof-of-concept study, 10 days

Population

21 adults with type 2 diabetes, decompensated cirrhosis and refractory ascites

Primary outcome

Change in fractional excretion of sodium

Effect

+0.44 percentage points vs control (95% CI 0.05 to 0.83); paracentesis −0.70 L/day (−1.17 to −0.23)

Twenty-one adults with type 2 diabetes, decompensated cirrhosis and refractory ascites were randomised to empagliflozin 10 mg daily plus standard care or standard care alone for 10 days, open-label.

Fractional sodium excretion rose 0.44 percentage points more with empagliflozin (95% CI 0.05 to 0.83). Urinary sodium increased by 92 mmol a day and urine volume by 822 mL a day versus control. Mean daily paracentesis volume was 0.6 L with empagliflozin and 1.3 L without, a difference of 0.70 L a day. No serious adverse events were reported.

This matters because refractory ascites has few medical options, and many such patients already have diabetes. But ten days in 21 patients cannot speak to kidney injury, hyponatraemia, euglycaemic ketoacidosis or infection risk in advanced liver disease, and SGLT2 inhibitors are not licensed for this use.

Treat it as a reason to watch for larger trials, not as a reason to start empagliflozin for ascites.

• Do not start an SGLT2 inhibitor for ascites on this evidence; it is a 21-patient pilot
• If a cirrhotic patient is already on an SGLT2 inhibitor, monitor sodium, creatinine and volume status closely
• Check for ketones in any unwell cirrhotic patient taking an SGLT2 inhibitor, even with normal glucose
• Coordinate with hepatology before changing glucose-lowering therapy in decompensated cirrhosis

Why it matters

It raises a testable idea for a group with almost no medical options between paracentesis and TIPS or transplant.

Don't overread it

Ten days in 21 patients says nothing about safety in decompensated cirrhosis.

The statistics, in plain English

The confidence interval for the main outcome (0.05 to 0.83) only just clears zero, so the true effect could be tiny. With 10 and 11 patients per arm, one or two unusual responders can move the averages a lot, and the open-label design leaves room for bias in paracentesis decisions.

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.