How Much Furosemide Should be Administered to Prevent Transfusion-Associated Circulatory Overload? Results of a Dose-Finding Study

BACKGROUND: Transfusion-associated circulatory overload (TACO) is a common but under-reported complication of transfusion therapy, and results in significant morbidity and mortality. Pre-transfusion furosemide has been proposed as a mitigating strategy but has not received widespread endorsement, in part due to uncertainty regarding safe and effective dosing. An observational study was therefore conducted with the goal of creating a reliable dose-response curve in patients at high risk of TACO, using the statistical technique of Multiple Comparisons Procedure with Modelling (MCP-MOD).

METHODS: Eligible patients were identified by screening bolus intravenous (IV) furosemide orders at two large academic hospitals. Eligibility criteria were selected to mimic those of a planned randomized controlled trial (RCT) of pre-transfusion furosemide for the prevention of TACO. Inclusion criteria were inpatients 50 years of age or older. Exclusion criteria were active bleeding, hemodynamic instability, glomerular filtration rate (GFR) < 30mL/min or need for dialysis, diuretic therapy (other than furosemide) administered less than 24 hours, furosemide (either IV or PO) administered less than 12 hours, or exogenous albumin administered less than 8 hours prior to index furosemide therapy. The primary outcome measure was the total volume of urine output in the six hours following IV furosemide administration. MCP-MOD analysis was performed after every 50 patients and continued until a weight-adjusted dose-response curve could be established with 100 mL precision, using the following covariates: age, sex, mean arterial pressure (MAP), serum albumin, GFR and history of chronic furosemide use.

RESULTS: A total of 149 participants (53 female and 96 male) were enrolled over 2 years. As too few patients received doses exceeding 0.6 mg/kg to allow for a precise dose-response curve, only 132 patients receiving weight-adjusted doses ranging from 0.1 to 0.6 mg/kg were included in the MCP-MOD analysis. The characteristics of these patients are shown in Table 1. The dose-response curve formula that was derived from these patients was:

Urine Output at 6 hours (mL) = 717.21 + 191.33 ln(dose (mg)/weight (kg)) + 0.73 × Age (years) + 103.37 × Sex (M=1/F=0) + 1.63 × GFR (ml/min) + 5.45 × MAP (mmHg) - 9.56 × albumin (g/L) - 62.31 × chronic diuretic use (Y=1/N-0)

The curve with accompanying 95% confidence intervals is shown in Figure 1

This formula suggests that in an 80 kg patient with a GFR of 75 mL/min, a MAP of 85 mmHg, albumin of 40 g/L and no history of chronic diuretic use, 10 mg of IV furosemide should produce a diuresis of approximately 600 mL, sufficient to prevent circulatory overload from the transfusion of 1 unit of RBCs. A dose of 20 mg would be required to achieve a similar diuresis in patients with either a MAP < 60 mmHg, or a combination of both a GFR < 45 mL/min and a history of chronic diuretic use.

CONCLUSIONS: A statistically precise dose-response curve for IV furosemide is possible with the inclusion of relevant clinical covariates, and can be used to guide dosing decisions in patients at risk of transfusion-associated circulatory overload. This formula derived from this study will inform the dosing recommendations of a planned RCT: Transfusion-Associated Circulatory Overload: Best Eliminated by Lasix® (TACO-BEL).