The Complexity of Diuretic Resistance

This article refers to ‘Renal tubular resistance is the primary driver for loop diuretic resistance in acute heart failure’ by J.M. ter Maaten et al., published in this issue on pages 1014–1022. How diuretics should be titrated and which targets should be used when adjusting diuretic therapy are the issues that arise most frequently in patients with decompensated heart failure.1 The effects of diuretic agents depend on the delivery of loop diuretics to their intraluminal site of action (pharmacokinetics) and on the dynamics of interaction of the diuretic with its receptor at the site of action (pharmacodynamics).2 The effects of loop diuretics can be assessed using dose–response curves by plotting the logarithm of the fractional urinary diuretic excretion rate vs. sodium excretion rate, or fractional sodium excretion (FENa). Patients with heart failure typically demonstrate a rightward and downward shift of the sigmoid dose–response curve (Figure 1A).3, 4 In this issue, ter Maaten and colleagues report on a study of the mechanism of loop diuretic resistance in 50 haemodynamically stable patients with acute heart failure.5 Following i.v. bumetanide, patients underwent a 6-h urine collection of sodium and bumetanide. Four measures of diuretic response (DR) were calculated: whole kidney DR and whole kidney FENa DR, defined as sodium excreted and increase in FENa per doubling of administered bumetanide, respectively, and tubular sodium DR and tubular sodium FENa, defined as sodium excreted and increase in FENa per doubling of urinary bumetanide, respectively. The first two parameters represent the sum of the pharmacokinetics and pharmacodynamics of bumetanide. The latter two measures represent the pharmacodynamic response and hence the effect of bumetanide at the renal tubular site of action. The loop diuretics dose–response curve guides the clinical use of these agents in oedematous disorders.2, 6 The present analysis and definitions of DR are somewhat different from those in classical diuretic dose–response studies.3, 4, 7, 8 The urinary excretion rate of a loop diuretic is considered a reliable measure of diuretic reaching the site of action and can be used in a typical dose–response analysis (Figure 1A) but requires multiple determinations of bumetanide and sodium in urine in each patient.3, 7 The urinary concentration of diuretic in the final urine (as used in the present study5) is less reliable as it may not accurately represent diuretic concentration at the site of action.6 However, it is a reasonable compromise in a clinical study with a relatively large number of patients. Additionally, as metrics of DR were calculated based on the ratio of natriuresis to log2 bumetanide dose, the slope of the ‘linear’ portion of the dose–response curve (Hill's slope) is ignored. The slope parameter varies among patients (Figure 1B) and is a key determinant of the response to dose increments. A decrease in the slope parameter may represent more resistance than is anticipated from a rightward shift in the curve. In the present study,5 the administered bumetanide dose explained 39% of the variance in bumetanide recovered in urine, with high between-patient variability. Furthermore, bumetanide delivery to its tubular site of action accounted for 23% of whole kidney sodium DR. The tubular sodium DR, representing the effect of bumetanide delivered to the site of action, accounted for most (71%) of the variance in whole kidney DR. The finding that the tubules are responsible for most of the resistance to loop diuretics is consistent with the extensive and complex tubular adaptations that occur in response to diuretic therapy.9-12 Loop diuretics are actively secreted into the lumen via organic anion transporters (OATs) in a proximal tubule epithelial cell (Figure 2A) and exert their action by binding to the apical Na+-K+-2Cl− cotransporter (NKCC2, SLC12A1) in the thick ascending limb (Figure 2B). Multiple studies have shown that diuretic resistance entails compensatory increases in sodium transport in segments of the renal tubule other than that targeted by a particular diuretic compound.9-11 Adaptive responses to diuretic therapy involve the induction of solute carriers, activation of paracrine signalling systems in multiple renal cell types and distal nephron remodelling (Figure 2C–F).12 Acute tolerance refers to a diminished response to the diuretic at its site of action that occurs within hours or within the duration of effect of the first dose.2, 8 It can be prevented by restoring diuretic-induced volume losses,13 and results from several mechanisms including reduced renal blood flow and enhanced proximal solute and fluid reabsorption. In chronic administration of loop diuretics (over several days), the solute rejected from the loop of Henle is reclaimed in the more distal nephron sites.2, 6 This effect is independent of extracellular fluid volume and is characterized by hypertrophy and hyperplasia, an increase in the volume of mitochondria and increases in the basolateral membrane surface area of distal convoluted tubule (DCT) cells, connecting tubule cells and principal cells in the collecting ducts.14-16 Concomitant with these morphological changes, components of the ion transport machinery expressed in the DCT and collecting ducts increase in number and activity, including the thiazide-sensitive NaCl− cotransporter (NCC), the principal determinant of NaCl reabsorption in the DCT, and the amiloride-sensitive epithelial Na+ channel (ENaC) (Figure 2C–F).9 Neurohormonal activation of both angiotensin II and aldosterone can activate sodium transporters in the distal nephron, including NCC and ENaC.17 Recent studies suggest that the Cl−–HCO3− exchanger pendrin (SLC26A4, PDS), located on the apical membrane of intercalated cells in the kidney cortical collecting duct and the connecting tubules, plays an important role in compensatory salt absorption in response to loop diuretics.18 In chronic kidney disease (CKD) endogenous organic acids accumulate and compete with loop diuretics for the OAT, resulting in a reduction in the amount of diuretic entering the nephron. The residual functioning nephrons respond normally when adequate amounts of loop diuretic reach the luminal site of action.7 Consequently, the relationship between natriuresis and the log rate of renal diuretic excretion is shifted to the right (Figure 1A). This pharmacokinetic impairment can be overcome by administering larger doses of diuretics.2 These principles were evident in the present study.5 Patients with a low glomerular filtration rate (GFR) received a higher bumetanide dose (medians of 4 mg and 2 mg in patients with low and high GFRs, respectively). This compensated for the low GFR so that urinary bumetanide excretion was similar in low and high GFR, resulting in equivalent total sodium excretion and urine volume. The tubular sodium DR, representing the total sodium excreted per doubling of renally excreted bumetanide, was also similar across the two groups. The intriguing finding that tubular FENa DR was higher in patients with renal dysfunction may seem counterintuitive. However, this finding is consistent with a previous report of a 60% increase in FENa in response to furosemide with low GFR7 and with the renal adaptations to maintain sodium balance in CKD.19 Given a constant intake of sodium, when GFR is reduced by an intrinsic renal disease, FENa increases because a progressive rise in sodium excretion per nephron is obligatory to compensate for the reduced nephron mass and maintain sodium balance.19 This is reflected in the current study5 by the higher FENa prior to bumetanide administration in patients with low vs. high GFR (medians: 0.8% vs. 0.2%). The addition of diuretics creates a complex situation in which both renal adaptation to the reduced GFR and the diuretics act to increase FENa. Even if the higher baseline is taken into account by assessing the true effect of bumetanide as the increment in tubular FENa DR over baseline (post-diuretic minus baseline FENa DR), the tubular FENa DR remains greater in patients with a reduced GFR. The clinical importance of the increased FENa DR in CKD remains unclear because the major limitation to the cumulative natriuresis that can be achieved in these patients is the diminished filtration of sodium. Can these findings be extrapolated to the care of patients receiving decongestive therapy? The large variability in the delivery of loop diuretics to their renal site of action and relatively small effect of the diuretic reaching the tubules on the net natriuretic response result in a poor correlation between diuretic dose and urine volume. For example, urine output in the Acute Study of Clinical Effectiveness of Nesiritide in Decompensated Heart Failure (ASCEND-HF) was poorly correlated with the loop diuretic dose (Spearman's ρ = 0.154; Figure 1C).20 Although the primary mechanisms that blunt the DR in most patients are tubular adaptations, our initial response to diuretic resistance is often to increase the loop diuretic dose, and this strategy has been used in recent clinical trials (e.g. CARRESS-HF and AVOID-HF). In these studies, the daily i.v. furosemide dose could be progressively advanced to a maximum of 800 mg/day,21 the ceiling dose of furosemide in heart failure with CKD.2 The only known method of overcoming tubular resistance is to add diuretics that block the adaptive nephron sites. Thiazides are often added to prevent reabsorption of NaCl in the DCT, and can result in a ‘synergistic’ response with profound diuresis, albeit at the cost of side effects.22 Few data are available on the efficacy of mineralocorticoid receptor antagonists (Butler J., unpublished data), acetazolamide and aquaretics. Low-dose dopamine, which promotes renal vasodilatation and increases renal blood flow, is not expected to enhance the response to loop diuretics in the presence of tubular adaptations. Diuretic resistance is a complex integral part of diuretic therapy. The data presented by ter Maaten et al.5 highlight the role of tubular alterations in renal NaCl transport capacity in attenuating diuretic-induced sodium losses. These data should provide impetus for future studies that aim to elucidate how these mechanistic insights can be implemented to craft an optimal approach to overcoming diuretic resistance. Conflict of interest: none declared.