Low Cardiac Output Stimulates Vasopressin Release in Patients With Stage D Heart Failure

– Its Relevance to Poor Prognosis and Reversal by Surgical Treatment –

Abstract

Background: Depressed hemodynamics stimulates arginine vasopressin (AVP) release, but the relationship between plasma AVP levels (P-AVP) and cardiac parameters, especially in patients with stage D heart failure (HF) receiving guideline-directed medical therapy, has not examined.

Methods and Results: Data including P-AVP were obtained from 162 in-hospital patients with stage D HF and from 80 patients receiving ventricular assist device (VAD, n=46) or heart transplantation (HTx, n=34) at 3 months after surgery. In the HF group, considerably high P-AVP (5.9±6.1 pg/ml) negatively correlated with serum sodium concentration (S-Na, 135.3±5.8 mEq/L, r=–0.548 [P<0.01]) and cardiac index (CI, 2.2±0.5 L·min^–1·m^–2, r=–0.458 [P<0.01]). After VAD/HTx treatment, improvement in the CI (2.7±0.5 L·min^–1·m^–2 [P<0.01] vs. HF) was accompanied by normalization of serum sodium concentration (S-Na; 138.2±2.0 mEq/L [P<0.01] vs. HF) and suppressed release of AVP (1.7±3.4 pg/ml [P<0.01] vs. HF). P-AVP positively correlated with only S-Na (r=0.454 [P<0.01]), whereas no correlation was observed with CI after VAD/HTx treatment. P-AVP ≥5.3 pg/ml well predicted poor 2-year survival in HF group (60% [P<0.01] vs. 90%).

Conclusions: Low cardiac output stimulates AVP release via a non-osmotic process that results in hyponatremia and poor prognosis in patients with stage D HF. After sufficient recovery of cardiac output by cardiac replacement therapy, AVP release is suppressed and is mainly regulated by serum osmolality.

Arginine vasopressin (AVP) plays a central role in the regulation of water and electrolyte balance, and is an essential hormone for maintaining homeostasis in humans.¹ AVP is secreted by hypothalamic neurons that project to the posterior pituitary gland in response to changes in plasma osmolality, which are detected by osmoreceptors in the hypothalamus (osmotic pathway).^2,3 AVP is also secreted when arterial underfilling caused by hypotension or volume depletion is detected by baroreceptors (non-osmotic pathway). AVP receptors are classified into V_1a (expressed on vascular smooth muscle), V_1b (on pituitary), and V₂ (on collecting duct in the kidney). AVP modulates body fluid regulation through water reabsorption in the collecting duct by stimulation of V₂ receptors, and also regulates vascular tone and cardiovascular contractility via V_1a receptors.⁴

Editorial p ????

In patients with heart failure (HF), the plasma levels of AVP are inappropriately high despite lower plasma osmolality.^5–10 The elevated AVP increases cardiac preload through water retention in collecting duct via stimulation of the V₂ receptor,^2,11 and also stimulates V_1a receptors, which facilitates increasing cardiac afterload by systemic arteriolar vasoconstriction.¹²

Although impaired hemodynamics has been estimated as a key to the inappropriate elevation of AVP levels in patients with HF,^5,7,13 the exact mechanisms and extensive relationship between plasma AVP levels and other cardiac parameters have remained unknown thus far. Furthermore, almost all studies of AVP were executed before the establishment of guideline-directed medical therapy (GDMT) for HF including β-blockers, angiotensin-converting enzyme inhibitors (ACEI) and aldosterone antagonists. If hemodynamics really matter in terms of non-osmotic AVP release, stage D HF should be associated with higher AVP levels in comparison with other stages. Dramatic changes in the plasma levels of AVP after ventricular assist device (VAD) implantation or heart transplantation (HTx) should also be expected in those patients. However, stage classification has recently been developed in the guidelines,¹⁴ and AVP levels have not been measured according to the new classification of HF. Therefore, we here examined the relationship between plasma AVP levels and other cardiac parameters in patients with stage D HF before and after VAD/HTx treatment.

Methods

Patient Selection

Of consecutive patients who were hospitalized for stage D HF and followed at the University of Tokyo Hospital between July 2011 and November 2013, 162 patients who were diagnosed as stage D HF by the Framingham Criteria¹⁵ and the guideline of the American Heart Association¹⁴ were retrospectively enrolled in this study (HF group). All patients were refractory to GDMT consisting of β-blocker, ACEI/ARB, and aldosterone antagonist.

VAD group consisted of consecutive 46 stage D HF patients who implanted either type of pulsatile or continuous flow VAD. After LVAD implantation, we followed GDMT to optimize the dosing of medicine as much as tolerated considering each patient’s hemodynamics. The setting of pulsatile VAD was adjusted as full-fill full-empty mode for maximum support under non-synchronous mode. The rotation speed of continuous flow VAD was also adjusted appropriately considering patients’ hemodynamics and interventricular septum shift observed in transthoracic echocardiography. The HTx group included 34 patients who had once had stage D HF and had been bridged from VAD therapy, but we did not include anyone of HTx group in HF or VAD group because of lack in AVP data during pre-HTx period.

Written informed consent was obtained at admission time from the patients and/or their family members in all cases. The study protocol was approved by the Ethics Committee of Graduate School of Medicine, the University of Tokyo [application no. 779 (1)].

Variables Evaluated

In the HF group, blood samples for AVP measurement were obtained after resting supine for 15 min in the early morning before taking any daily medicine while in a steady state after the treatment of acute exacerbation of HF. Echocardiographic examination was also executed while in a steady state, and cardiac output (CO) was calculated using Doppler-derived aortic flow signals based on the principle that the velocity time integral of blood flow multiplied by the cross-sectional area of the left ventricle outflow estimates CO volume. In patients with VAD support, total CO was calculated as the summation of the CO of the native heart calculated by the Doppler method described above and the estimated VAD flow. There were no patients with moderate or severe aortic regurgitation. In 64 of the HF patients, CO measurement by right heart catheterization was also performed while in the steady state.

In the VAD/HTx groups, echocardiographic examination and hemodynamic study were performed as well as blood sampling for AVP measurement in the same manner as for the HF group at 3 months after the operation. Of them, pairwise data of plasma AVP levels were available for 28 patients between the pre-VAD (ie, stage D HF) and post-VAD periods. No preoperative AVP data were available for patients in the HTx group. Hemodynamic data were obtained for all of the VAD/HTx patients (n=80) before and after operation. All blood samples were centrifuged immediately for 20 min, and the samples were stored at –80°C before the assay. All blood samples for AVP measurements were stored and the AVP levels were determined by a double antibody radioimmunoassay method using AVP RIA Neo from LSI Medicine Corporation available since January 2014.

Statistical Analysis

All statistical analyses were performed using PASW Statistics 18 (SPSS Inc, Chicago, IL, USA). Categorical variables are summarized as frequencies and percentages, and compared using Chi-square test or Fisher’s exact test as appropriate. Continuous variables are represented as mean±standard deviation unless otherwise specified, and compared using unpaired t-test or Mann-Whitney test as appropriate. Each variable of the HF, VAD, or HTx group was compared by ad-hoc Tukey’s test when analysis of variance confirmed significance among the groups. Pearson product-moment correlation coefficient was used to evaluate the relationship between plasma levels of AVP and other clinical parameters. Each clinical parameter was compared by paired t-test or Wilcoxon signed-rank test as appropriate in the VAD group during the study period. Receiver-operating characteristics (ROC) analysis was performed to obtain a cutoff level of plasma AVP for 2-year survival. To examine the effect of the plasma levels of AVP on prognosis, Kaplan-Meier analysis with log-rank test and Cox proportional-hazard model were adopted. Inter-rater reliability between echocardiography and hemodynamic study for CI was analyzed by calculating intraclass correlation coefficients. All hypothesis tests reported are 2-tailed, and used a P-value <0.05 as significant.

Results

Comparison of Patients’ Characteristics Among HF, VAD, and HTx Groups (Table 1)

All patients in the HF group were assigned stage D, and GDMT was introduced and optimized in >80% of patients; inotropes were intravenously infused in >40% of patients. In total, 65 (40.1%) patients were not candidates for VAD/HTx treatment because of their age (>65 years old), 21 patients (13.0%) because of end-organ dysfunction, and 32 patients (19.8%) because of systemic diseases or social problems; 28 patients (17.3%) underwent VAD implantation during the study period. No patients received tolvaptan before enrollment.

Table 1. Characteristics of the Patients in the HF, VAD, and HTx Groups

	HF group (n=162)	VAD group (n=46)	P value vs. HF	HTx group (n=34)	P value vs. HF	P value vs. VAD
Demographic parameters
Age, years	58.0±20.1	37.7±13.0	<0.001*	37.4±14.5	<0.001*	1.000
Male, n (%)	115 (71.0)	36 (78.3)	0.579	23 (67.6)	0.684	0.533
Body surface area, m2	1.68±0.19	1.72±0.18	0.360	1.65±0.23	0.699	0.257
Body mass index	21.9±3.5	20.7±3.5	0.090	20.7±3.4	0.134	0.808
Etiology of ischemia, n (%)	22 (13.6)	7 (15.2)	0.425	8 (23.5)	0.423	0.5313
SBP, mmHg	101.7±13.2	110.0±6.4	<0.001*	128.1±11.2	<0.001*	<0.001*
DBP, mmHg	64.1±6.7	86.4±7.7	<0.001*	72.3±8.2	<0.001*	<0.001*
Heart rate, beats/min	77.3±15.6	81.5±17.3	0.198	84.9±14.7	0.214	0.727
SaO2, %	96.6±3.4	96.8±3.1	0.512	97.2±3.9	0.166	0.211
Concomitant medication
Furosemide, mg/day	43.4±26.7	6.3±11.6	<0.001*	2.1±5.3	<0.001*	0.661
Spironolactone, mg/day	22.3±21.3	25.5±19.4	0.553	7.7±15.5	<0.001*	<0.001*
Trichlormethiazide, mg/day	0.2±0.6	0.0±0.2	0.211	0	0.114	0.973
Administration of furosemide, n (%)	162 (100)	13 (28.3)	<0.001†	4 (11.8)	<0.001†	<0.001†
Administration of spironolactone, n (%)	103 (63.6)	33 (71.7)	<0.001†	8 (23.5)	<0.001†	<0.001†
Administration of trichlormethiazide, n (%)	12 (7.4)	0 (0)	0.041†	0 (0)	0.041†	1.000
Administration of β-blocker, n (%)	150 (92.6)	46 (100)	<0.001†	15 (44.1)	<0.001†	<0.001†
Administration of ACEI/ARB, n (%)	134 (82.7)	34 (73.9)	0.034†	23 (67.6)	0.033†	0.123
Administration of statin, n (%)	104 (64.2)	40 (87.0)	<0.001†	34 (100)	<0.001†	0.099
Catecholamine infusion, n (%)	66 (40.7)	0 (0)	<0.001†	0 (0)	<0.001†	1.000
Laboratory parameters
Plasma AVP, pg/ml	5.9±6.1	2.2±2.8	<0.001*	1.4±1.6	<0.001*	0.694
Hemoglobin, g/dl	11.8±2.3	11.0±1.9	0.062	11.7±2.0	0.952	0.233
Platelets, ×103/μl	19.7±7.7	23.1±7.0	0.010*	22.4±5.6	0.100	0.872
Serum albumin, g/dl	3.4±0.6	3.7±0.6	0.069	3.9±0.7	<0.001*	0.031*
Serum sodium, mEq/L	135.3±5.8	138.2±1.9	0.001*	138.3±2.0	0.002*	0.997
Serum sodium <136 mEq/L, n (%)	72 (44.4)	0 (0)	<0.001†	2 (5.9)	<0.001†	0.822
Serum potassium, mEq/L	4.3±0.5	4.3±0.4	0.987	4.5±0.5	0.021*	0.096
Serum BUN, mg/dl	26.9±14.3	16.1±5.6	<0.001*	19.4±8.5	0.003*	0.435
Serum creatinine, mg/dl	1.2±0.6	0.9±0.3	<0.001*	1.0±0.4	0.137	0.267
Serum total bilirubin, mg/dl	2.1±10.9	0.8±0.4	0.628	0.6±0.2	0.594	0.997
Plasma BNP, log10 pg/ml	2.68±0.44	1.96±0.33	<0.001*	1.85±0.48	<0.001*	0.089
Echocardiographic parameters
LV diastolic diameter, mm	62.8±14.2	56.4±12.9	0.042*	43.0±4.7	<0.001*	<0.001*
Ejection fraction, %	36.5±20.1	24.6±12.9	<0.001*	69.3±7.7	<0.001*	<0.001*
E/e’	19.4±7.9	14.2±4.3	<0.001*	12.4±3.8	<0.001*	<0.001*
CI, L·min–1·m–2	2.2±0.5	2.6±0.4	<0.001*	2.9±0.6	<0.001*	0.115

^*P<0.05 by unpaired t-test or Mann-Whitney test as appropriate. ^†P<0.05 by Chi-square test or Fisher’s exact test as appropriate.

ACEI, angiotensin-converting enzyme inhibitor; ARB angiotensin receptor blocker; AVP, arginine vasopressin; BNP, B-type natriuretic peptide; BUN, blood urea nitrogen; CI, cardiac index; DBP, diastolic blood pressure; E/e’, ratio of the mitral velocity to the early-diastolic velocity of the mitral annulus; HF, heart failure; HTx, heart transplantation; LV, left ventricle; SaO₂, oxygen saturation; SBP, systolic blood pressure; VAD, ventricular assist device.

Patients who had VAD/HTx showed improved hemodynamics (Table 2), together with recovery of end-organ dysfunction and normalization of hyponatremia compared with the HF group (Figure 1). There were no significant differences in hemodynamics before (ie, under VAD treatment) and after HTx (Table 2). There were no significant differences in cardiac index (CI), plasma levels of BNP or the serum sodium concentration (S-Na) level between the VAD and HTx groups (Figure 1). In 64 patients in the HF group and 80 in the VAD/HTx group, CI measured by echocardiography and hemodynamic study had high intraclass correlation coefficients (0.992, P<0.001).

Table 2. Hemodynamic Study Before and After VAD/HTx Treatment

	VAD (n=46)			HTx (n=34)
	Pre-VAD	Post-VAD	P value	Pre-HTx (all VAD support)	Post-HTx	P value
Mean RAP, mmHg	12.3±6.8	7.2±4.5	0.008*	4.6±2.0	4.2±2.2	0.269
Systolic PAP, mmHg	43.8±13.9	21.1±5.2	<0.001*	20.0±4.2	18.7±2.6	0.069
Diastolic PAP, mmHg	24.9±7.6	9.1±3.6	<0.001*	8.3±3.8	7.6±3.0	0.322
Mean PAP, mmHg	32.5±9.5	15.6±6.1	<0.001*	14.0±3.5	12.6±3.4	0.089
PCWP, mmHg	23.9±8.3	7.7±3.7	<0.001*	7.8±3.4	7.3±2.5	0.148
CI, L·min–1·m–2	1.7±0.3	2.7±0.4	<0.001*	2.6±0.3	2.8±0.4	0.121

^*P<0.05 by paired t-test. PAP, pulmonary artery pressure; PCWP, pulmonary capillary wedge pressure; RAP, right atrial pressure. Other abbreviations as in Table 1.

Figure 1.

Comparison of CI (A), plasma levels of BNP (B) and AVP (C), and S-Na (D) in HF (n=162), VAD (n=46), and HTx groups (n=36). ^*P<0.05 by ad-hoc Tukey’s test compared with HF group when analysis of variance proved to be significant. AVP, arginine vasopressin; BNP, B-type natriuretic peptide; CI, cardiac index; HF, heart failure; HTx, heart transplantation; S-Na, serum sodium concentration; VAD, ventricular assist device.

Relationship Between Plasma Levels of AVP and Other Clinical Parameters in HF and VAD/HTx Groups

Plasma AVP levels had a non-normal distribution in the HF and VAD/HTx groups (P<0.05 by Shapiro-Wilk’s test). Unlike the plasma BNP levels, plasma AVP levels were not log-normally distributed. Plasma AVP levels were significantly lower in the VAD/HTx group (n=80) than in the HF group (n=162), whereas plasma AVP levels were not different between the VAD and HTx groups (Table 1, Figure 1). In 28 patients who had VAD implantation, pairwise measurement of plasma AVP levels confirmed that marked decreases in AVP levels were associated with improvement of CI and normalization of plasma BNP levels and S-Na (all P<0.01; Figure 2).

Figure 2.

Changes in CI (A), plasma levels of BNP (B) and AVP (C), and S-Na (D) in patients who had VAD implantation during the study period (n=28). ^*P<0.05 by paired t-test or Wilcoxon signed-rank test compared with preoperative parameters as appropriate. AVP, arginine vasopressin; BNP, B-type natriuretic peptide; CI, cardiac index; S-Na, serum sodium concentration; VAD, ventricular assist device.

In the HF group (n=162), plasma AVP levels significantly correlated with CI, ejection fraction, plasma BNP levels, and S-Na (Table 3). Higher plasma levels of AVP were associated with lower CI (Figure 3A), higher plasma levels of BNP (Figure 3B), and lower S-Na (Figure 3C). There were 5 patients with extremely high plasma AVP levels (>20 pg/ml), whose CI were <2.0 L·min^–1·m^–2, plasma BNP >1,500 pg/ml, and S-Na <130 mEq/L. Although all hemodynamic parameters were improved after VAD implantation (all P<0.05; Table 2), only CI negatively correlated with the plasma level of AVP among the hemodynamic parameters (P=0.008, r=–0.516) (Table 4) in the stage D HF patients in whom a hemodynamic study was performed before VAD implantation (n=64).

Figure 3.

Relationship between plasma levels of AVP and CI, plasma levels of BNP, and S-Na in the HF (n=162, A–C and VAD/HTx groups (n=82, D–F. ^*P<0.05 by Pearson’s product-moment correlation coefficient. AVP, arginine vasopressin; BNP, B-type natriuretic peptide; CI, cardiac index; HF, heart failure; HTx, heart transplantation; S-Na, serum sodium concentration; VAD, ventricular assist device.

Table 3. Correlation Between Plasma AVP Concentration and Other Characteristic Parameters in the HF and VAD/HTx Groups

vs. Plasma AVP	HF (n=162)		VAD/HTx (n=80)
	P value	R value	P value	R value
Demographic parameters
Age, years	0.862	0.014	0.114	0.131
Body surface area, m2	0.304	–0.081	0.161	0.089
SBP, mmHg	0.058	–0.175	0.631	–0.053
DBP, mmHg	0.053	–0.153	0.117	0.131
Heart rate, beats/min	0.068	0.183	0.165	–0.159
SaO2, %	0.438	0.053	0.721	0.079
Concomitant medication
Furosemide, mg/day	0.128	0.120	0.119	0.151
Spironolactone, mg/day	0.910	0.009	0.978	0.003
Trichlormethiazide, mg/day	0.856	–0.014	0.092	0.179
Catecholamine infusion, n (%)	0.755	0.025	–	–
Laboratory parameters
Hemoglobin, g/dl	0.379	–0.07	0.085	0.188
Platelets, ×103/μl	0.290	–0.084	0.895	0.023
Serum albumin, g/dl	0.187	–0.105	0.452	0.088
Serum sodium, mEq/L	<0.001*	–0.548	<0.001*	0.464
Serum potassium, mEq/L	0.676	0.033	0.364	0.105
Serum BUN, mg/dl	0.089	0.132	0.234	0.141
Serum creatinine, mg/dl	0.074	0.164	0.664	0.049
Serum total bilirubin, mg/dl	0.802	0.020	0.359	0.103
Plasma BNP, pg/ml	<0.001*	0.633	0.334	0.109
Echocardiographic parameters
LV diastolic diameter, mm	0.083	0.137	0.215	0.134
Ejection fraction, %	0.018*	–0.230	0.521	–0.071
E/e’	0.118	0.231	0.367	0.096
CI, L·min–1·m–2	<0.001*	–0.458	0.810	–0.026

^*P<0.05 by Pearson’s product-moment correlation coefficient. Abbreviations as in Table 1.

Table 4. Correlation Between Plasma AVP Concentration and Hemodynamic Parameters in the HF and VAD/HTx Groups

vs. Plasma AVP	HF (n=64)		VAD/HTx (n=80)
	P value	R value	P value	R value
Mean RAP, mmHg	0.964	0.028	0.366	0.178
Systolic PAP, mmHg	0.984	0.014	0.387	0.157
Diastolic PAP, mmHg	0.497	0.152	0.570	0.116
Mean PAP, mmHg	0.587	0.143	0.437	0.188
PCWP, mmHg	0.524	0.217	0.564	0.114
CI, L·min–1·m–2	0.008*	–0.516	0.791	0.052

^*P<0.05 by Pearson’s product-moment correlation coefficient. Abbreviations as in Tables 1,2.

In the VAD/HTx group (n=80), plasma AVP levels had no correlation with CI (Figure 3D) or plasma BNP levels (Figure 3E). None of hemodynamic parameters correlated with plasma AVP levels in the VAD/HTx group (Table 4). The only significant correlation was between the plasma AVP level and S-Na (Table 3), but in sharp contrast to the HF patients, higher AVP levels were accompanied by higher S-Na in the VAD/HTx group (Figure 3F).

Clinical Prognosis Stratified by Plasma AVP Levels in Patients With Stage D HF (Figure 4)

In the HF group, 24 patients (14.8%) died during the study period. ROC analyses demonstrated that a cutoff level of plasma AVP for all-cause survival over 2 years was identically 5.3 pg/ml (area under curve, 0.681; sensitivity, 0.708; specificity, 0.674). According to the cutoff level of plasma AVP, patients were stratified into 2 groups, and Kaplan-Meier analyses showed significant differences between them in terms of all-cause survival over 2 years (P<0.001). Cox regression analysis demonstrated that higher plasma levels of AVP were significantly associated with decreased survival (hazard ratio, 4.803; 95% confidence interval, 2.045–11.28; P<0.001). In contrast, plasma levels of AVP had no significant effect on mortality in the VAD/HTx group (P=0.254 by Cox regression analysis).

Figure 4.

Kaplan-Meier curves for 2-year survival according to plasma levels of AVP in the HF group (n=162). ^*P<0.001 by Cox regression analyses. AVP, arginine vasopressin; CI, confidence interval; HF, heart failure; HR, hazard ratio.

Discussion

In this study, we demonstrated that a lower CO had a significant correlation with increased secretion of AVP, and that elevated plasma levels of AVP were significantly associated with hyponatremia and poor prognosis in patients with stage D HF. Plasma AVP levels normalized along with the improvement of CO after VAD/HTx treatment, and positively correlated with S-Na, in contrast to the HF group. VAD treatment was as efficient at improving hemodynamics and reducing plasma levels of AVP as HTx.

Non-Osmotic Regulation of AVP in Stage D HF

AVP secretion is regulated by a non-osmotic pathway that is responsive to such stimuli as decreased circulation, hypoxia, intravenous inotropic infusion, activation of renin-angiotensin-aldosterone, and sympathetic nerve system stimulation.^5,16–18 The non-osmotic pathway has been considered a dominant modulator of plasma AVP levels in hemodynamically sick conditions, including HF, although no quantitative studies have been reported thus far, especially in patients with stage D HF.

This study is a novel quantitative study that demonstrated a significant negative correlation between CO and plasma AVP levels in patients with stage D HF. Other hemodynamic parameters including right atrial pressure and pulmonary capillary wedge pressure did not correlate with plasma AVP levels. As a non-osmotic trigger, arterial underfilling caused by low CO rather than congestion contributes to the secretion of AVP.

Interestingly, improvement of CO by VAD treatment was as sufficient to normalize the plasma levels of AVP as HTx. Hemodynamic parameters in the VAD group were consistently indistinguishable from those of the HTx group. In other words, not only HTx but also VAD treatment can be a cardiac replacement therapy from the viewpoint of sufficient reduction of plasma AVP levels. Our observation was consistent with the fact that the non-osmotic trigger of impaired hemodynamics may be a key to stimulating AVP secretion in patients with advanced HF, as several authors have previously speculated.^5,13 Uretsky et al’s data that demonstrated a significant negative correlation between changes in systemic blood pressure and plasma AVP levels after intravenous infusion of vasodilator accompanied by non-significant increases in CO would support our result,⁷ although the correlation between blood pressure and AVP levels barely reached statistical significance in our analysis.

Neurohumoral activation, which contributes to the malignant cycle of HF,¹⁹ results in higher levels of AVP (6.5–9.5 pg/ml) according to previous studies,^5,10,20,21 but all those studies were executed before the current GDMT was established. Relatively lower plasma levels of AVP in our study (average 5.9±6.1 pg/ml) may be attributable to the effective suppression of neurohumoral activity by the application of GDMT including β-blocker, ACEI, and aldosterone blocker at high rates (92.6%, 82.7%, and 63.6%, respectively). The recently reported subanalysis of the EVEREST study consistently included 78% of patients with plasma AVP levels ≤8 pg/ml receiving high rates of β-blocker and ACEI (83% and 74%, respectively) therapy.⁶

In this study, S-Na negatively correlated with plasma AVP levels in patients with stage D HF. Secreted AVP binds V₂ receptor located in collecting duct, and activates the signaling cascade including aquaporin 2, which facilitates water reabsorption and often causes dilution hyponatremia.^2,11,22 Only a few authors have previously reported such a correlation among small numbers of HF patients or those with acute myocardial infarction.^8,23 We believe that ours is a noteworthy report showing the close relationship of low CO, increased AVP and hyponatremia in advanced HF.

Osmotic Regulation of Plasma AVP After Improvement of Hemodynamics

The osmotic control of AVP release is one of 2 modulators of homeostasis,² and individual variation, genetic, environmental, species differences, circadian rhythm, or the nature of solute providing the osmotic stimuli can significantly affect the release of AVP by altering the threshold and/or the sensitivity of osmoreceptors.^4,24,25 Consistently, plasma levels of AVP had a positive correlation with S-Na but not with CO or plasma BNP levels in our VAD/HTx group whose hemodynamics appeared to be almost normal. In other words, AVP secretion was mainly regulated by S-Na in an osmotic manner after VAD/HTx treatment under hemodynamically stable conditions. Not only HTx but also VAD treatment could reverse the dominant mechanism of AVP secretion from a non-osmotic pathway to an osmotic pathway by improving hemodynamics.

Plasma Levels of AVP and Prognosis

There have been no studies demonstrating increased plasma levels of AVP as a significant predictor for prognosis except for the recent subanalysis of the EVEREST study.⁶ Their cutoff value of plasma AVP was almost consistent with our result (8.0 vs. 5.2 pg/ml). They adopted the upper limit of the healthy population (ie, 8.0 pg/ml) as the cutoff value, whereas we derived our cutoff value of 5.2 pg/ml by ROC analysis. Elevated AVP may have a crucial role in the vicious cycle of worsening HF, because it results from hemodynamic decline, and facilitates worsening of HF including hyponatremia and increased preload and afterload to the heart.²⁶ Consistently, elevated AVP levels were associated with lower CI and S-Na in the present study (Figure 3). Goldsmith et al experimentally demonstrated that intravenous infusion of AVP caused adverse circulatory effects, including decreases in CO and increases in systemic vascular resistance and pulmonary capillary wedge pressure, in patients with HF.⁵ We recently reported that a V₂ receptor antagonist improved congestion and hyponatremia especially in HF patients with preserved function of collecting duct during short-term clinical course.^27,28 It would be a future concern whether a V₂ receptor antagonist improves long-term prognosis in patients with stage D HF.

Study Limitations

All data were analyzed in a retrospective manner, and our result was derived only from the observational analyses. Our result should be tested in a prospective manner. Data for the HF, VAD, and HTx groups were unpaired except for 28 patients who received VAD implantation during the study period. Consecutive observation during HF duration and VAD/HTx treatment in larger populations would strengthen our results. CO was estimated by echocardiography in most of the HF patients, although data were validated by hemodynamic study in 64 of the HF group as well as 80 of the VAD/HTx group. Because we included only patients with stage D HF as the HF group, our results may not be adopted in cases of mild to moderate HF. We evaluated AVP levels in patients with VAD/HTx at 3 months after operation, considering hemodynamic stability. However, future studies are needed to clarify whether osmotic AVP regulation is maintained for longer after surgery.

Conclusions

In patients with stage D HF, low CO stimulates AVP release via a non-osmotic pathway, and elevated AVP results in hyponatremia and poor prognosis. After sufficient recovery of hemodynamics by VAD/HTx therapy, AVP release is suppressed and is mainly regulated by serum osmolality. VAD treatment was as sufficient to improve hemodynamics and reduce plasma AVP levels as HTx.

Acknowledgments

Grant-in-Aid from Fukuda Foundation for Medical Technology to K.K.

Disclosures

The authors have no conflicts of interest.

References

• 1.    Robertson GL. Antidiuretic hormone: Normal and disordered function. Endocrinol Metab Clin North Am 2001; 30: 671–694.
• 2.    Robertson GL, Mahr EA, Athar S, Sinha T. Development and clinical application of a new method for the radioimmunoassay of arginine vasopressin in human plasma. J Clin Invest 1973; 52: 2340–2352.
• 3.    Kinugawa K, Imamura T, Komuro I. Experience of a vasopressin receptor antagonist, tolvaptan, under the unique indication in Japanese heart failure patients. Clin Pharmacol Ther 2013; 94: 449–451.
• 4.    Schrier RW, Berl T, Anderson RJ. Osmotic and nonosmotic control of vasopressin release. Am J Physiol 1979; 236: F321–F332.
• 5.    Goldsmith SR, Francis GS, Cowley AW Jr, Goldenberg IF, Cohn JN. Hemodynamic effects of infused arginine vasopressin in congestive heart failure. J Am Coll Cardiol 1986; 8: 779–783.
• 6.    Lanfear DE, Sabbah HN, Goldsmith SR, Greene SJ, Ambrosy AP, Fought AJ, et al. Association of arginine vasopressin levels with outcomes and the effect of V2 blockade in patients hospitalized for heart failure with reduced ejection fraction: Insights from the EVEREST trial. Circ Heart Fail 2013; 6: 47–52.
• 7.    Uretsky BF, Verbalis JG, Generalovich T, Valdes A, Reddy PS. Plasma vasopressin response to osmotic and hemodynamic stimuli in heart failure. Am J Physiol 1985; 248: H396–H402.
• 8.    Szatalowicz VL, Arnold PE, Chaimovitz C, Bichet D, Berl T, Schrier RW. Radioimmunoassay of plasma arginine vasopressin in hyponatremic patients with congestive heart failure. N Engl J Med 1981; 305: 263–266.
• 9.    Francis GS, Benedict C, Johnstone DE, Kirlin PC, Nicklas J, Liang CS, et al. Comparison of neuroendocrine activation in patients with left ventricular dysfunction with and without congestive heart failure: A substudy of the Studies of Left Ventricular Dysfunction (SOLVD). Circulation 1990; 82: 1724–1729.
• 10.    Goldsmith SR, Francis GS, Cowley AW Jr. Arginine vasopressin and the renal response to water loading in congestive heart failure. Am J Cardiol 1986; 58: 295–299.
• 11.    Nielsen S, Kwon TH, Christensen BM, Promeneur D, Frokiaer J, Marples D. Physiology and pathophysiology of renal aquaporins. J Am Soc Nephrol 1999; 10: 647–663.
• 12.    Goldsmith SR. Vasopressin as vasopressor. Am J Med 1987; 82: 1213–1219.
• 13.    LeJemtel TH, Serrano C. Vasopressin dysregulation: Hyponatremia, fluid retention and congestive heart failure. Int J Cardiol 2007; 120: 1–9.
• 14.    Kinugawa K. How to treat stage D heart failure? When to implant left ventricular assist devices in the era of continuous flow pumps? Circ J 2011; 75: 2038–2045.
• 15.    McKee PA, Castelli WP, McNamara PM, Kannel WB. The natural history of congestive heart failure: The Framingham study. N Engl J Med 1971; 285: 1441–1446.
• 16.    Anderson RJ, Pluss RG, Berns AS, Jackson JT, Arnold PE, Schrier RW, et al. Mechanism of effect of hypoxia on renal water excretion. J Clin Invest 1978; 62: 769–777.
• 17.    Creager MA, Faxon DP, Cutler SS, Kohlmann O, Ryan TJ, Gavras H. Contribution of vasopressin to vasoconstriction in patients with congestive heart failure: Comparison with the renin-angiotensin system and the sympathetic nervous system. J Am Coll Cardiol 1986; 7: 758–765.
• 18.    Schrier RW, Berl T, Harbottle JA, McDonald KM. Catecholamines and renal water excretion. Nephron 1975; 15: 186–196.
• 19.    Levine TB, Francis GS, Goldsmith SR, Simon AB, Cohn JN. Activity of the sympathetic nervous system and renin-angiotensin system assessed by plasma hormone levels and their relation to hemodynamic abnormalities in congestive heart failure. Am J Cardiol 1982; 49: 1659–1666.
• 20.    Goldsmith SR, Francis GS, Cowley AW Jr, Levine TB, Cohn JN. Increased plasma arginine vasopressin levels in patients with congestive heart failure. J Am Coll Cardiol 1983; 1: 1385–1390.
• 21.    Goldsmith SR. Baroreflex loading maneuvers do not suppress increased plasma arginine vasopressin in patients with congestive heart failure. J Am Coll Cardiol 1992; 19: 1180–1184.
• 22.    Rai A, Whaley-Connell A, McFarlane S, Sowers JR. Hyponatremia, arginine vasopressin dysregulation, and vasopressin receptor antagonism. Am J Nephrol 2006; 26: 579–589.
• 23.    Tada Y, Nakamura T, Funayama H, Sugawara Y, Ako J, Ishikawa S, et al. Early development of hyponatremia implicates short- and long-term outcomes in ST-elevation acute myocardial infarction. Circ J 2011; 75: 1927–1933.
• 24.    Rittig S, Knudsen UB, Norgaard JP, Pedersen EB, Djurhuus JC. Abnormal diurnal rhythm of plasma vasopressin and urinary output in patients with enuresis. Am J Physiol 1989; 256: F664–F671.
• 25.    Trudel E, Bourque CW. Central clock excites vasopressin neurons by waking osmosensory afferents during late sleep. Nat Neurosci 2010; 13: 467–474.
• 26.    Schrier RW. Role of diminished renal function in cardiovascular mortality: Marker or pathogenetic factor? J Am Coll Cardiol 2006; 47: 1–8.
• 27.    Imamura T, Kinugawa K, Shiga T, Kato N, Muraoka H, Minatsuki S, et al. Novel criteria of urine osmolality effectively predict response to tolvaptan in decompensated heart failure patients: Association between non-responders and chronic kidney disease. Circ J 2013; 772: 397–404.
• 28.    Imamura T, Kinugawa K, Minatsuki S, Muraoka H, Kato N, Inaba T, et al. Urine osmolality estimated using urine urea nitrogen, sodium and creatinine can effectively predict response to tolvaptan in decompensated heart failure patients. Circ J 2013; 775: 1208–1213.