Renal function and natriuresis-guided diuretic therapy – a pre-specified analysis from the PUSH-AHF trial

Aim: In a randomized controlled trial, we recently showed that a natriuresis-guided diuretic approach improved natriuresis and diuresis in patients with acute heart failure (HF). In this pre-specified analysis, we investigated the association between (worsening) renal function, outcomes and the effect of intensive natriuresis-guided loop diuretic therapy as compared with standard of care. 

Methods and results: The Pragmatic Urinary Sodium-based algoritHm in Acute Heart Failure (PUSH-AHF) trial randomized patients to natriuresis-guided diuretic therapy or standard of care. Serum creatinine and estimated glomerular filtration rate (eGFR) were assessed at fixed timepoints, and worsening renal function (WRF) was assessed at 72 h. The primary outcome was the interaction between randomized treatment allocation, baseline eGFR and the dual primary outcome of PUSH-AHF: total natriuresis at 24 h and time to all-cause mortality or HF rehospitalization at 180 days. In 309 patients, median baseline eGFR was 53 (35–73) ml/min/1.73 m², and 58% had eGFR <60 ml/min/1.73 m². Baseline eGFR did not significantly modify the treatment effect of natriuresis-guided diuretic therapy on natriuresis at 24 h (p for interaction = 0.730). However, baseline eGFR significantly modified the effect on all-cause mortality and HF rehospitalization (p for interaction = 0.017): the risk of this second primary outcome was lower in patients with lower eGFR who were randomized to the natriuresis-guided group. In the natriuresis-guided arm, eGFR decreased more (−11.0 vs. −6.91 ml/min/1.73 m²; p = 0.002) during the first 3 days, but this effect was attenuated at discharge (−10.3 vs. −8.69 ml/min/1.73 m²; p = 0.38). WRF was more frequently observed in patients randomized to natriuresis-guided treatment, but was not associated with worse clinical outcomes. 

Conclusions: Natriuresis-guided diuretic treatment improved diuresis and natriuresis irrespective of baseline eGFR and occurrence of WRF, was effective even in patients with low eGFR, and the observed effect on eGFR was transient and not associated with worse clinical outcomes.