Clinical background
In acute decompensated heart failure, intravenous loop diuretics are the cornerstone of treatment, but a substantial proportion of patients are discharged with residual congestion. In the ADHERE registry, approximately 20 per cent of patients had no weight loss or even gained weight by discharge [1]. A central problem is that the conventional monitoring methods, fluid balance and weight, are crude surrogates for the actual measure of success: sodium excretion. Sodium is the primary pathophysiological driver of extracellular volume expansion, and a positive sodium balance, even in patients with a documented negative fluid balance, is strongly associated with increased mortality [1]. In addition, fluid balance and weight are difficult to measure accurately in clinical practice and yield data only after several hours' delay, with the result that diuretic doses are often titrated only once daily.
The Natriuretic Response Prediction Equation (NRPE) was developed to solve this problem. From a spot urine sample taken approximately 2 hours after a diuretic dose, the equation can predict the cumulative sodium excretion over the following 6 hours and thereby give an early, objective measure of the diuretic response, long before fluid balance or weight are available.
Calculating the NRPE
The equation is based on established renal physiology. The instantaneous rate of urine formation can be derived from the product of the estimated GFR and the ratio of serum creatinine to urine creatinine, since creatinine undergoes minimal tubular reabsorption and secretion. Multiplication by the urine sodium concentration converts this into an instantaneous sodium excretion (mmol/min). An empirical constant converts the peak instantaneous natriuresis into the cumulative 6-hour excretion.
where the eGFR is calculated with CKD-EPI and the BSA with the Du Bois formula:
and the constant of 3.25 hours converts the peak instantaneous natriuresis into the cumulative 6-hour excretion. The factor 60 converts minutes to hours and the factor 1000 converts mL to litres.
In the original 2016 derivation study, a constant of 2.5 hours was chosen on the basis of the 1 to 1.5 hour half-life of bumetanide [2]. In the subsequent 2021 validation study the constant was optimised data-driven to 3.25 hours on the basis of the original derivation cohort, and this is the constant the calculator uses [1].
The derivation cohort consisted of 50 patients with acute decompensated heart failure at Yale New Haven Hospital who prospectively underwent carefully monitored 6-hour urine collections after intravenous bumetanide [2]. The median age was high, the patients predominantly had heart failure with reduced ejection fraction of non-ischaemic aetiology, and hypertension, diabetes and renal impairment were common. The median dose was 3 mg of intravenous bumetanide. The equation was then validated in a larger prospective cohort (the MDR cohort) with 638 diuretic doses from patients with acute decompensated heart failure [1].
Interpretation in practice
The NRPE gives a predicted sodium excretion in mmol for the 6 hours following the diuretic dose. Three thresholds are used to classify the response:
| Predicted Na excretion (mmol/6 h) | Classification | Clinical meaning |
|---|---|---|
| <50 | Poor response | With twice-daily dosing, total daily excretion is <100 mmol, which on a sodium-restricted diet (3 g/day corresponds to 130 mmol) leads to a positive sodium balance. Uptitration of the diuretic or the addition of a thiazide should be considered. |
| <100 | Suboptimal response | The maximum net sodium excretion is approximately 70 mmol/day with twice-daily dosing, corresponding to less than 0.5 litres of isotonic fluid. Uptitration should be considered in patients with substantial volume overload. |
| >150 | Excellent response | Net sodium excretion exceeds 170 mmol/day with twice-daily dosing, corresponding to more than 1 litre of isotonic fluid. The current dosing is effective. |
In the Yale Diuretic Pathway (YDP), the NRPE was implemented in a nurse-led protocol in which the predicted sodium excretion determined the next diuretic dose. The physician chose a daily sodium target (default 370 mmol, alternatively 230 or 500 mmol) and hold parameters (a rise in creatinine of 0.5 mg/dL, a systolic blood pressure below 90 mmHg). A spot urine sample was taken 1 to 2 hours after each dose, and the EPIC system automatically calculated the predicted sodium excretion and recommended the next dose [1].
Validation and performance
In the MDR cohort (638 diuretic doses) the NRPE showed excellent discrimination, with an AUC ≥0.90 for predicting a poor, suboptimal and excellent natriuretic response. The NRPE outperformed the clinically recorded fluid balance at all thresholds (p <0.05) [1].
An external validation was carried out at the Instituto Mexicano del Seguro Social in Mexico City with 87 diuretic doses from 49 patients with acute decompensated heart failure [3]. The mean age was 57 years, 67 per cent were men, the mean eGFR was 65 mL/min/1.73 m² and the mean ejection fraction 35 per cent. A poor natriuretic response occurred with 39 per cent of doses. The AUC for the NRPE in predicting a poor response was 0.91 (95 per cent CI 0.85 to 0.98), comparable to the original cohort. The NRPE outperformed urine sodium alone (AUC 0.75), urine output over the corresponding shift (AUC 0.74), eGFR, diuretic dose and blood pressure [3].
The NRPE has also been validated for oral loop diuretics in two cohorts [4]. In the MDR cohort (318 oral doses from 237 patients) the AUC for a poor response was 0.87 (95 per cent CI 0.83 to 0.91), not significantly different from the intravenous performance (p = 0.16). In the TRANSFORM-Mechanism cohort (110 oral doses) the AUC was 0.89 (95 per cent CI 0.80 to 1.0). The patients' own assessment of their diuretic response was poorer (AUC 0.57) [4].
In the YDP cohort (161 patients), diuresis improved markedly after implementation of NRPE-guided titration: the mean daily volume rose from 1.8 to 3.0 litres, the net fluid loss from 1.1 to 2.1 litres and the weight loss from 0.3 to 2.5 kg (p <0.001 for all comparisons) [1].
Limitations
The NRPE was derived and validated exclusively in acute decompensated heart failure with volume overload. The results cannot be extrapolated to patients with hypervolaemia of other aetiology, for example liver cirrhosis or renal failure, without further validation [3].
The equation presupposes that the spot urine sample is taken 1 to 2 hours after an intravenous loop diuretic dose. Samples taken at other times have not been validated and may give misleading results, particularly if taken too early, when the urine does not yet reflect the diuretic-induced natriuresis, or too late, when the natriuresis is already waning [1,2].
Patients with bladder dysfunction, urinary incontinence or an inability to cooperate with urine collection were excluded from the validation studies [1,2]. Incomplete bladder emptying at the time of sampling can affect the result, particularly the urine creatinine and hence the calculation. In the external validation only 5 per cent had a urinary catheter and no bladder scans were performed, which is a potential source of error [3].
The YDP protocol excluded patients with chronic kidney disease on dialysis and patients on ongoing thiazide treatment, because of the risk of excessive diuresis with sequential nephron blockade [1]. The calculator should be used with caution in patients with very severely impaired renal function, in whom the eGFR estimate becomes less reliable.
The empirical constant of 3.25 is specific to intravenous bolus dosing of loop diuretics. For oral administration the NRPE has shown similar discrimination, but with a somewhat lower AUC, which may reflect slower and more variable absorption [4].
The NRPE has not been shown to improve survival or reduce readmission rates in a randomised trial. The YDP cohort was a before-and-after comparison without a control group, and the improvements in diuresis may be explained in part by factors other than the equation itself [1].
References
- Rao VS, Ivey-Miranda JB, Cox ZL, et al. Natriuretic Equation to Predict Loop Diuretic Response in Patients With Heart Failure. J Am Coll Cardiol. 2021;77(6):695-708. PMID: 33573739
- Testani JM, Hanberg JS, Cheng S, et al. Rapid and Highly Accurate Prediction of Poor Loop Diuretic Natriuretic Response in Patients With Heart Failure. Circ Heart Fail. 2016;9(1):e002370. PMID: 26721915
- Ramírez-Sánchez P, Falcón-Aguirre A, Tepayotl-Aponte A, et al. External validation of the natriuretic response prediction equation to discriminate diuretic response in heart failure. ESC Heart Fail. 2025;12(1):668-671. PMID: 39135310
- Ivey-Miranda JB, Rao VS, Cox ZL, et al. Natriuretic response prediction equation for use with oral diuretics in heart failure. Eur Heart J. 2025;46(25):2410-2418. PMID: 40272149
