Background
Salt and water accumulation leading to fluid overload is associated with increased mortality in intensive care unit (ICU) patients, but diuretics' effects on patient outcomes are uncertain. In this first version of the GODIF trial, we aimed to assess the effects of goal-directed fluid removal with furosemide versus placebo in adult ICU patients with fluid overload.

Methods
We conducted a multicentre, randomised, stratified, parallel-group, blinded, placebo-controlled trial in clinically stable, adult ICU patients with at least 5% fluid overload. Participants were randomised to furosemide versus placebo infusion aiming at achieving neutral cumulative fluid balance as soon as possible. The primary outcome was the number of days alive and out of the hospital at 90 days.

Results
The trial was terminated after the enrolment of 41 of 1000 participants because clinicians had difficulties using cumulative fluid balance as the only estimate of fluid status (32% of participants had their initially registered cumulative fluid balance adjusted and 29% experienced one or more protocol violations). The baseline cumulative fluid balance was 6956 ml in the furosemide group and 6036 ml in the placebo group; on day three, the cumulative fluid balances were 1927 ml and 5139 ml. The median number of days alive and out of hospital at day 90 was 50 days in the furosemide group versus 45 days in the placebo group (mean difference 1 day, 95% CI -19 to 21, p-value .94).

Conclusions
The use of cumulative fluid balance as the only estimate of fluid status appeared too difficult to use in clinical practice. We were unable to provide precise estimates for any outcomes as only 4.1% of the planned sample size was randomised