Background
Liver cirrhosis, one of the leading health threat worldwide, yields larger years lived with disability (YLDs) globally than either hepatitis or hepatocellular carcinoma (HCC) [
1]. Annually, 5 to 10% of patients with compensated cirrhosis develop ascites [
2]. Hospitalization due to refractory ascites is one of the most frequent reasons for health-care cost in liver cirrhosis [
3]. An unmet need for the treatment of refractory cirrhotic ascites still exists.
Although with possible reported limitations [
4,
5], the non-invasiveness of bioimpedance analysis (BIA) has charmed many researchers to survey its possible application in healthy and diseased populations. BIA has been showed to be an adequate tool for evaluation of total body water (TBW) and extracellular water (ECW) in cirrhotic patients with ascites, [
6,
7] and has been reported to assess body cell mass (BCM) after trans-jugular intrahepatic porto-systemic shunt (TIPS) in liver cirrhosis [
8]. Newly developed ascites is a cardinal symptom of acute decompensation of cirrhosis and may occur in the setting of acute-on-chronic liver failure characterized by dysfunction in multiple organ systems, including renal and cardiovascular failure [
9]. The pathophysiology of ascites due to liver cirrhosis is thought to be multifactorial. Malnutrition, systemic inflammation, and exaggerated activation of the renin-angiotensin-aldosterone system (RAAS) [
10] play roles in the pathogenesis of cirrhotic ascites. These factors are also common pathophysiological features of fluid control in hemodialysis patients due to chronic renal failure. Body composition and fluid status monitoring assessed by non-invasive BIA, especially intracellular water (ICW), have been shown to be of prognostic value in acute decompensated heart failure, [
11] acute kidney injury under continuous hemodiafiltration, [
12] and patients with chronic renal failure under hemodialysis [
13,
14]. Therefore, BIA has also been shown to be an applicable tool for assessment of ICW, even in patients with possible over-hydration or vigorous fluid status changes. However, in decompensated liver cirrhosis, the prognostic value and clinical implications of fluid status monitoring are not fully elucidated.
Tolvaptan, a selective vasopressin 2 (V2) receptor antagonist, has been indicated for treatment of hypervolemic or euvolemic hyponatremia in the US. Its efficacy and safety have been reported for patients with hyponatremia [
15,
16]. Tolvaptan inhibits aquaporin (AQP)-2 channel expression in the collecting ducts and increases the excretion of electrolyte-free water by increasing hypoosmolar urine [
17]. In 2013, tolvaptan was approved for use in cirrhotic ascites, whether hyponatremic or not, in Japan at low doses (3.75 mg to 7.5 mg per day) as an add-on (to conventional diuretics) approach [
18]. Efficacy and safety have been reported in Japanese cirrhotic patients, including cases with Child-Pugh-Turcotte (CPT) grade C cirrhosis with this low-dose and add-on approach [
19,
20]. In a randomized control trial, tolvaptan has been implicated to improve survival of cirrhotic patients with ascites and hyponatremia, especially in cases whose hyponatremia was resolved by tolvaptan [
21]. In recent two observational studies, diuretic responders to tolvaptan showed survival benefit [
22,
23]. However, there were still conflicting previous reports regarding diuretic efficacy and safety of V2 antagonism used for hyponatremia in cirrhotic patients [
24,
25]. Results from pilot studies of heart failure implicate that V2 antagonism not only retracts ECW but also works to decrease ICW, [
26] however, research on how electrolyte-free water excretion that is caused by V2 antagonism affects body composition and fluid status in cirrhotic patients is still sparse. Therefore, the accumulation of evidence for detailed monitoring of fluid status may help its adequate application in cirrhotic individuals.
In this observational study, we hypothesized that the use of add-on low-dose tolvaptan for cirrhotic ascites changes BIA-defined fluid volumes in different compartments, which may be a useful marker for the prediction of its short-term efficacy, and thus long-term survival in these patients.
Discussion
In the current study, we demonstrated a rapid and early decrease in phase angle (reactance), e.g. reduced ICW assessed by BIA after adding low dose tolvaptan, correlated significantly with short-term efficacy of amelioration of body fluid retention and the long-term survival. ΔICWBIA%-6 h predicted short-term efficacy of add-on tolvaptan at an accuracy of 97% and long-term survival in our study cohort. To our knowledge, this is the first real world clinical practice-based study to show the usefulness of BIA-assessed ICW change in response to electrolyte-free water excretion caused by V2 antagonism for both short-term efficacy and long-term survival, in patients of decompensated cirrhosis.
Kogiso et al. reported that ECW/TBW ratio significantly decreased in eight long-term responders to tolvaptan [
19]. We also observed the same tendency in responders in this current study (Fig.
1c). BIA is also a useful tool to demonstrate that the loss of body weight in responders is mainly due to decreased TBW, rather than reduced body cell mass. Selberg et al. showed that higher BIA-derived phase angle (assumed to be determined by tissue cellularity, tissue hydration, and membrane potential) in 305 patients of liver cirrhosis predicted poorer survival [
37]. This is coherent to our finding that a decreased phase angle after add-on tolvaptan (decreased ICW
BIA) predicted better prognosis (Fig.
3).
Masuda et al. reported that tolvaptan reduced both ICW and ECW, evaluated by BIA in six patients with chronic kidney disease [
34]. Masuda et al. also presumed that a rise in serum osmolality and resultant fluid shift from intracellular to interstitial and intravascular spaces would occur immediately after the administration of tolvaptan. With BIA, Nagayama et al. demonstrated different effects on fluid distribution between tolvaptan and furosemide in a patient with liver cirrhosis and chronic kidney disease [
38]. Nomoto et al. showed that in patients with acute decompensated heart failure, diuresis due to tolvaptan caused no significant change of ECW/ICW ratio by using BIA, while that due to conventional diuretics decreased ECW/ICW ratio [
11]. This finding suggested a net reduction of ICW might characterize tolvaptan from conventional diuretics. In the current study, we showed that a decrease in ICW
BIA occurred as soon as 6 h after the first dose of tolvaptan (Fig.
2), and its degree correlated with the increase in BUN and serum sodium (Fig.
5), which are the major determinant factors for both serum osmolality and urine-concentrating mechanisms in the renal medullary loops of Henle [
39]. The increase in serum sodium and BUN (Fig.
4b) might help to induce a shift of free water from the intracellular compartment, which is less prominent in non-responders. Higher baseline BUN has also been reported as a negative predictive factor for the efficacy of tolvaptan for cirrhotic ascites [
40,
41]. In this present study, the baseline BUN tented to be higher (
P = 0.06, Table
1), and the follow-up BUN levels remained high in the non-responders (Fig.
5b).
The reason for why early extractable ICW
BIA after V2 antagonism unexpectedly correlates with survival (Fig.
5) is not easy to explain. Hiramine Y et al. demonstrated a positive effect of add-on tolvaptan on the prognosis of patients with cirrhotic ascites by analyzing 628 patients retrospectively, compared to conventional diuretics alone [
42]. Higher baseline BUN levels, that are supposedly a result of the chronic vascular under-filling state due to cirrhosis, might also be a result of chronic diuretic use for ascites. Sone et al. showed that the chronic administration of furosemide, a loop diuretic agent, greatly reduced renal medullary contents of organic osmolytes in a murine model [
43]. Conversely, a selective V2 antagonist did not produce a sufficient decrease in the content of organic osmolytes; while, an increase in taurine levels, an effective organic osmolyte, was observed following exposure to a selective V2 antagonist in a murine model [
44]. Therefore, a well-preserved urine-concentrating ability that implicates preserved renal medullary osmolyte levels, favors a transient and immediate rise in osmolality after V2 antagonism, and is also assumed to be required for the immediate extraction of ICW
BIA, a positive predictive parameter for tolvaptan efficacy in this study. Moreover, more severe hyponatremia is associated with an increase in ICW
BIA after add-on tolvaptan in our study cohort (Fig.
5b, left). This was coherent with a previous study showing that maintained serum sodium over 140 mmol/L is a significant predictor of response to tolvaptan [
22]. Hyponatremia in cirrhosis is associated with more intractable ascites and greater impairments of renal function [
36]. A higher degree of extractable ICW
BIA with low-dose of tolvaptan treatment might also suggest better-preserved renal function, and therefore, better survival potential.
Base on previous studies until early 2000s, some experts reported that cautions might be needed for the explanation of cross-sectional quantification of ICW or ECW by using BIA, especially in patients with altered body compositions such as kidney, heart, or liver diseases [
4,
45]. In this current study, we used the longitudinal comparison in each individual for assessment, in order to lessen confounding effects such as possible anthropometric factors. In addition, in order to overcome the limitation of the direct use of BIA-defined volumes in diseased state, the use of BIA-derived “phase angle” to define changes in body compositions in cirrhosis has been focused [
37,
46]. We also showed a significant and high correlation between the change of phase angle and ICW
BIA in a cross-sectional analysis (Fig.
3).
Still, there are other major limitations to this study. First, even though add-on low dose tolvaptan for cirrhotic ascites is generally approved and used in daily practice in Japan, it is still not a standard treatment internationally. This point might limit applicability of the results yielded by this study. In addition, although we monitored body composition through serial measurement of BIA within 1 week after administration of low-dose tolvaptan, the small sample size without a proper control group might limit the generalizability of this observation and made it difficult to control for effects of possible confounders with regard to long-term survival, which might cause selection bias. Further investigations are still warranted to determine whether and how low dose tolvaptan improves survival in patients with cirrhosis in a prospective setting. Moreover, although BIA for assessment of body composition is generally considered to be reproducible and accurate with few cost and invasiveness concerns [
5], and the application of phase angle might meet some needs in populations with altered body composition, it is still not fully validated in many edematous states. This is possibly why BIA is not routinely applied in practice until now. Finally, since the concomitant use of conventional diuretics was maintained during the present study, the effect of monotherapy with tolvaptan to change fluid volumes in different compartments was not examined.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.