Ascites: Under- and Overfill

The Under- versus Overfill Hypothesis

Effective arterial blood volume refers to the part of extracellular fluid present in the arterial part of the circulation that is sensed by baroreceptors and affects renal sodium handling. Although it is a logical physiological concept, there is, unfortunately, no direct, measurable clinical correlate and no specific test that can be easily ordered to clearly define effective arterial blood volume. Instead, it is surmised from a broad constellation of clinical and laboratory parameters, which can make interpretation ambiguous. This is especially so in clinical states with expanded extracellular fluid volume and enhanced sodium retention, such as cardiac failure, liver disease, or nephrotic syndrome. Is the primary problem underfill, that is, a decreased effective arterial blood volume with secondary, physiologic activation of the renin-angiotensin-aldosterone system (RAAS)? Or is there overfill due to pathologic primary renal sodium retention?^1,2 The question is of considerable clinical relevance because aggressive diuretic treatment could be detrimental in an underfill scenario, whereas volume supplementation could be a problem with overfill.

Under- versus Overfill in Liver Disease

In liver disease, the prevailing hypothesis has been that of underfill, albeit modified to the peripheral arterial vasodilation hypothesis.³ In this concept, there is initial arterial underfilling, mainly because of splanchnic vasodilation due to the secretion of vasodilatory compounds, such as nitric oxide from the diseased liver. The underfilling activates the RAAS to maintain euvolemia, but because of portal hypertension, the retained fluid preferentially transudates into the abdominal cavity.³ Consequently, aldosterone antagonists, such as spironolactone, have become a mainstay in the treatment of ascites.¹ Yet, several observations argue against such secondary activation of the RAAS, including that (1) salt retention precedes the fall in systemic vascular resistance in a dog model of cirrhosis, (2) parameters of RAAS activity are not increased in a substantial subgroup of patients with ascites, and (3) provision of a sodium load in patients with cirrhosis results in a positive sodium balance despite suppression of the RAAS.⁴ Instead, these findings suggest primary renal sodium retention, which precedes and contributes to the expansion of extracellular fluid volume and ascites formation. However, what could cause such primary sodium retention? In this issue of JASN, Wang and colleagues provide an intriguing answer: direct activation of the epithelial sodium channel (ENaC) in the collecting duct by bile acids.⁵ Previous research had already shown that ENaC can be activated by bile acids in concentrations below 1 mM.⁶ Those experiments were motivated by the structural relation of ENaC to a bile acid–sensitive ion channel expressed in cholangiocytes and thought to be relevant for regulation of ENaC in the colon, where bile acids are present in millimolar concentrations. Yet, in certain liver conditions, bile acids can also reach such concentrations in blood. Thus, Wang et al. explored whether bile acids in the glomerular filtrate could cause primary salt retention by direct activation of ENaC in the collecting duct.

Evidence for Direct Activation of ENaC with Bile Duct Ligation

The group developed a mouse model on the basis of bile duct ligation and compared this with sham-operated animals. To minimize confounding from secondary RAAS activation, all animals received background mineralocorticoid receptor blockade with spironolactone (20 mg/kg per day). As expected, bile duct ligation resulted in jaundice with bile acid concentrations in blood around 1.6 mmol/L. Consistent with enhanced ENaC activity, bile duct–ligated mice had lower blood potassium and higher bicarbonate concentration, as well as significantly higher urinary sodium excretion after benzamil (an ENaC blocker) administration than sham-operated animals. Moreover, while blood sodium concentrations remained stable, bile duct–ligated mice had a significant increase in total body water compared with sham animals, consistent with salt and water retention. Wang et al. also demonstrated direct activation of ENaC by 1 mM taurocholic acid in split-open tubules from mice and in isolated perfused rabbit collecting ducts. Concentrations of this bile acid in blood had previously been shown to reach levels around 1 mM in mice with bile duct ligation and in human liver conditions. Overall, although there was considerable variability in the individual changes in mice, the data provide solid evidence for a direct activation of ENaC in the collecting duct by bile acids in clinically relevant concentrations.

Analogy to Nephrotic Syndrome

The findings by Wang et al. and their implications are remarkably similar to data on direct ENaC activation in nephrotic syndrome.⁷ Similar to liver disease, there has been a long-running debate about under- versus overfill and primary versus secondary sodium retention in nephrotic syndrome.² There also has been a prevailing theory of underfill, mediated by the decreased oncotic pressure due to albumin loss. This, however, was in conflict with several clinical observations, such as the excretion of edema fluid well before normalization of blood albumin levels in children with nephrotic syndrome going into remission or the complications of fluid overload in some patients receiving albumin supplementation.² These observations were more consistent with overfill and primary sodium retention, and it was, therefore, of great interest when Svenningsen et al. proposed a potential explanation in the form of direct activation of ENaC by proteases that pathologically reached the glomerular filtrate in nephrotic syndrome.⁸

Clinical Relevance

The findings by Wang et al. clearly make a strong case for considering the use of ENaC channel blockers, such as amiloride, rather than mineralocorticoid receptor blockers in the treatment of ascites associated with higher concentrations of bile acid in blood. This contrasts with the current guidance from the American Association for the Study of Liver Diseases, which recommends aldosterone antagonists and loop diuretics as the mainstay of treatment.⁸ So, should we change practice? Unfortunately, the situation is again very analogous to that in nephrotic syndrome. As beautiful and elegant as the experimental results are, their translatability into clinical practice seems to be limited. Just as there is, at best, limited evidence for a beneficial effect of amiloride in the treatment of nephrotic edema, the same applies for the treatment of ascites. Currently, amiloride use is recommended only in cases of aldosterone antagonist–associated severe side effects.⁸ Moreover, in one randomized controlled trial, amiloride was inferior to the aldosterone antagonist canreonate.⁹

Conclusions

Our understanding of molecular physiology has provided fascinating insights into disease mechanisms. The finding of direct activation of ENaC, independent of RAAS activation, provides a beautiful potential explanation for the clinical overfill observations in both nephrotic syndrome and hepatic ascites for which, however, convincing clinical confirmation is still lacking. However, perhaps, the most important insight from all of this is that we cannot treat all the patients the same. Just as not all nephrotic states may be created equal and some patients have evidence of underfill and others of overfill, the same may be true for hepatic ascites. For instance, while in the aforementioned clinical trial, amiloride was overall inferior to canreonate, some individual patients did not respond to canreonate, but well to amiloride (and vice versa).⁹ It gets more complicated because we have even more options for diuretic treatment of ascites: Metolazone, a thiazide-like diuretic thought to also affect the proximal tubule, has been reported to be effective, although these data typically date back some decades.¹⁰ A more modern approach of impairing proximal tubular sodium reabsorption is the use of sodium-glucose transporter 2 inhibitors, for which efficacy has also been reported.¹¹ Part of the science of medicine is for the clinician to use the insights provided by basic science to consider the respective underlying pathophysiology in the individual patient with the aim of identifying the best treatment for that person. However, given the plethora of choices and the lack of clarity with regard to which diuretic or combination thereof is best suited for which patient, it seems that we cannot quite yet dispose of the art of medicine.

Disclosures

Disclosure forms, as provided by each author, are available with the online version of the article at http://links.lww.com/JSN/E818.

Funding

None.

Author Contributions

Writing – original draft: Detlef Bockenhauer.

Writing – review & editing: Detlef Bockenhauer, Giulia Florio.