Rapid and Highly Accurate Prediction of Poor Loop Diuretic Natriuretic Response in Patients With Heart Failure

Abstract

Background: Removal of excess sodium and fluid is a primary therapeutic objective in acute decompensated heart failure and commonly monitored with fluid balance and weight loss. However, these parameters are frequently inaccurate or not collected and require a delay of several hours after diuretic administration before they are available. Accessible tools for rapid and accurate prediction of diuretic response are needed.

Methods and results: Based on well-established renal physiological principles, an equation was derived to predict net sodium output using a spot urine sample obtained 1 or 2 hours after loop diuretic administration. This equation was then prospectively validated in 50 acute decompensated heart failure patients using meticulously obtained timed 6-hour urine collections to quantify loop diuretic-induced cumulative sodium output. Poor natriuretic response was defined as a cumulative sodium output of <50 mmol, a threshold that would result in a positive sodium balance with twice-daily diuretic dosing. Following a median dose of 3 mg (2-4 mg) of intravenous bumetanide, 40% of the population had a poor natriuretic response. The correlation between measured and predicted sodium output was excellent (r=0.91; P<0.0001). Poor natriuretic response could be accurately predicted with the sodium prediction equation (area under the curve =0.95, 95% confidence interval 0.89-1.0; P<0.0001). Clinically recorded net fluid output had a weaker correlation (r=0.66; P<0.001) and lesser ability to predict poor natriuretic response (area under the curve =0.76, 95% confidence interval 0.63-0.89; P=0.002).

Conclusions: In patients being treated for acute decompensated heart failure, poor natriuretic response can be predicted soon after diuretic administration with excellent accuracy using a spot urine sample.

Keywords: diuretic resistance; diuretics; heart failure; poor natriuretic response; sodium.

Figure 1

Correlation plots (panel A & B) and receiver operating characteristics curves (Panel B & C) comparing the relationship between the sodium output prediction equation (A & C) and clinically recorded net fluid output (B &D) with 6-hour cumulative measured sodium output Sodium prediction calculations in panel A and C were made with Equation 1 using the two hour time point. Clinically recorded net fluid output represents the 7AM to 3PM nursing shift where fluid balance is calculated by the clinical staff and the parameter that same day diuretic dosing decisions are frequently made upon.