Urinary sodium evaluation: the missing target for diuretic treatment optimization in acute heart failure patients? Reply.

Dr Galluzzo et al. commented on our manuscript on urinary sodium content after diuretic initiation in acute heart failure (HF). They expressed their concerns on the absence of a prospective protocol and the fact that there was no estimate of congestion (score) available. The Authors suggest that given these limitations, patients could have been presenting at different positions at the ‘time‐natriuresis’ curve. Certainly, we acknowledge that the observational nature of our study is a limitation. However, our findings are a reflection of ‘real‐world’ data, which also inherently means that there was no structural scoring of congestion at start of treatment. It is also important to realize that our primary findings were based on 6 h urinary sodium content. Although many patients improve clinically during the first hours of treatment, it is not likely assessment of congestion at baseline would have altered our findings at 6 h. We did show that patients who had lower urinary sodium content at 6 h had evidence of more severe HF with higher natriuretic peptide levels at baseline, lower blood pressure, worse renal function, more frequent use of loop diuretics at baseline and more use of inotropes and vasopressors.

The Authors point to their subanalysis of the small DRAIN study in 80 patients, where they evaluated early (2 h) spot urinary sodium content and surrogate outcome measures and found similar results as compared with our analyses. Our results should be interpreted slightly different, considering we evaluated total urinary sodium content over 6 h, which is probably a better representation of total natriuretic response to a given diuretic dose than spot urinary sodium. , , With regard to the inotrope/vasopressor regimen not being standardized in our cohort, we consider these therapies only in very selected patients, according to the European Society of Cardiology HF guidelines, where they are only advocated in a small proportion of acute HF patients with severe hypotension. In our study, the prognostic information of urinary sodium content at 6 h was unchanged by adjusting for inotrope/vasopressor use [hazard ratio (HR) 1.05, 95% confidence interval (CI) 1.01–1.08, P = 0.008, per 10 mmol decrease (n = 146)] or left ventricular ejection fraction [(HR 1.05, 95% CI 1.01–1.08, P = 0.003, per 10 mmol decrease (n = 157)]. Change in natriuretic peptide level was only available in a subset of patients at 24 h after admission, and was not different in tertiles of urinary sodium (P = 0.69 and P = 0.67 for absolute and relative change, respectively).

We do not think worsening renal function (WRF), as used by the Authors, is an appropriate surrogate endpoint in acute HF patients without considering natriuretic/diuretic response. WRF in a patient with good natriuresis is not true WRF and should be termed pseudo‐WRF, and cannot directly be compared with WRF in patients with poor natriuresis/diuretic response, according to most recent position papers on how to evaluate renal function in HF. , For instance, in our cohort and using different definitions, WRF was consistently more frequent in patients with the highest urinary sodium content at 6 h. Yet, the prognostic information of urinary sodium at 6 h was unaffected by the occurrence of WRF, and WRF itself was not independently associated with clinical outcome. This again highlights not to use serum creatinine/WRF as a response marker during decongestion in acute HF.

Finally, the finding that urinary sodium content at 6 h was only associated with all‐cause mortality and appeared not to be associated with HF readmission is probably related to the severity of HF in our tertiary HF clinic, as well as a tendency of some elderly patients with HF at our clinic to decide not to be readmitted after discharge as part of advanced care planning. Therefore, we also evaluated the combined endpoint of mortality and HF rehospitalization, confirming our primary results. We agree with the Authors that we need prospective randomized studies to evaluate whether this cheap and easy marker of natriuretic/diuretic response is a valuable tool to assess treatment effect or can serve as therapeutic target in a cute HF.