Diuretic resistance is a powerful predictor of adverse outcome in acute heart failure (AHF), irrespectively of underlying glomerular filtration rate. Metrics of diuretic efficacy such as natriuresis, urine output, weight loss, net fluid balance, or fractional sodium excretion, differ in their risk for measurement error, convenience, and biological plausibility, which should be taken into account when interpreting their results. Loop diuretic resistance in AHF has multiple causes including altered drug pharmacokinetics, impaired renal perfusion and effective circulatory volume, neurohumoral activation, post-diuretic sodium retention, the braking phenomenon and functional as well as structural adaptations in the nephron. Ideally, these mechanisms should guide specific treatment decisions with the goal of achieving complete decongestion. Therefore, volume overload needs to be identified correctly to avoid poor diuretic response due to electrolyte depletion or dehydration. Next, renal perfusion should be optimised if possible and loop diuretics should be prescribed above their threshold dose. Addition of thiazide-type diuretics should be considered when a progressive decrease in loop diuretic efficacy is observed with prolonged use (i.e., the braking phenomenon). Furthermore, thiazide-type diuretics are a useful addition in patients with low glomerular filtration rate. However, they limit free water excretion and are relatively contraindicated in cases of hypotonic hyponatremia, where acetazolamide is the better option. Finally, ultrafiltration should be considered in patients with refractory diuretic resistance as persistent volume overload after decongestive treatment is associated with worse outcomes. Whether more upfront use of any of these individually tailored decongestion strategies is superior to monotherapy with loop diuretics remains to be shown by adequately powered randomised clinical trials.
Serum creatinine and cystatin C‐based estimates of glomerular filtration rate are misleading in acute heart failure
