ARTICLE IN PRESS

www.elsevier.com/locate/heafai

DTD 5

European Journal of Heart Fa

Independent and incremental prognostic value of endogenous

ouabain in idiopathic dilated cardiomyopathy

Maria Vittoria Pitzalis b, John M. Hamlyn c, Elisabetta Messaggio a, Massimo Iacoviello b,

Cinzia Forleo b, Roberta Romito b, Elisabetta de Tommasi b, Paolo Rizzon b,

Giuseppe Bianchi a, Paolo Manunta a,*

a Division of Nephrology, Dialysis and Hypertension, University ‘‘Vita-Salute’’ San Raffaele Hospital, Via Olgettina 60, 20132 Milan, Italyb Institute of Cardiology, University of Bari, Bari, Italy

c Department of Physiology, School of Medicine, University of Maryland, Baltimore, MD, USA

Received 26 November 2004; received in revised form 3 May 2005; accepted 14 July 2005

Abstract

Increased circulating levels of endogenous ouabain (EO) have been observed in some heart failure patients, but their long term clinical

significance is unknown. This study investigated the prognostic value of EO for worsening heart failure among 140 optimally treated patients

(age 50T14 years; 104 male; NYHA class 1.9T0.7) with idiopathic dilated cardiomyopathy. Plasma EO was determined by RIA and by

liquid chromatography mass spectrometry, values were linearly correlated (r =0.89) in regression analysis. During follow-up (13T5 months),

heart failure progression was defined as worsening clinical condition leading to one or more of the following: sustained increase in

conventional therapies, hospitalization, cardiac transplant, or death. NYHA functional class, age, LVEF, peak VO2 and plasma levels of EO

were predictive for heart failure progression. Heart failure worsened 1.5 fold (HR: 1.005; 95% CI: 1.001–1.007; p <0.01) for each 100 pmol/

L increase in plasma EO. Moreover, those patients with higher plasma EO values had an odds ratio of 5.417 (95% CI: 2.044–14.355;

p <0.001) for heart failure progression. Following multivariate analysis, LVEF, NYHA class and plasma EO remained significantly linked

with clinical events. This study provides the first evidence that circulating EO is a novel, independent and incremental marker that predicts

the progression of heart failure.

D 2005 European Society of Cardiology. Published by Elsevier B.V. All rights reserved.

Keywords: Cardiac failure; Progression; Na pump; Ouabain-like factor

1. Introduction

There are many neurohormonal abnormalities associated

with the progression of left ventricular dysfunction to heart

failure [1], however, their precise roles remain unclear. This

uncertainty [2,3] also applies to endogenous ouabain [EO], a

mammalian steroid hormone [4], that is involved in sodium

homeostasis [5,6] and blood pressure regulation [7]. Several

studies have suggested that EO may have a primary role in

causing cardiac dysfunction and failure. In a study in rats,

1388-9842/$ - see front matter D 2005 European Society of Cardiology. Publishe

doi:10.1016/j.ejheart.2005.07.010

* Corresponding author. Tel.: +39 0226433891; fax: +39 0226432384.

E-mail address: manunta.paolo@hsr.it (P. Manunta).

chronic infusion of very low doses of ouabain to double the

plasma concentration of EO, triggered a signal transduction

pathway that produces cardiac hypertrophy [8]. In a second

study, the young offspring of hypertensive patients had

higher plasma levels of EO than the offspring of normo-

tensive parents which were correlated with diastolic

dysfunction [42]. In another study, in newly diagnosed

patients with mild hypertension, plasma levels of EO were

found to be bimodally distributed [9]. The low (normal)

mode was similar to that of normotensive patients whereas

the high mode had a median value almost twice the normal

value. In the high EO mode patients, left ventricular mass

and stroke volume were increased while heart rate was

lower. In a fourth study of patients with more advanced

ilure xx (2005) xxx – xxx

d by Elsevier B.V. All rights reserved.

HEAFAI-04015; No of Pages 8

ARTICLE IN PRESSM.V. Pitzalis et al. / European Journal of Heart Failure xx (2005) xxx–xxx2

hypertension, circulating levels of EO were directly related

to both blood pressure and total peripheral resistance and

inversely related to cardiac index [10].

These observations prompted us to investigate the

prognostic value of plasma EO in patients with idiopathic

dilated cardiomyopathy, on the assumption that EO may

have a primary role in the progression of heart failure. These

patients were chosen to avoid the confounding effects of

systemic hypertension and cardiac ischemia. EO was

measured using a well-established radio-immunoassay [11]

in addition some samples underwent analysis by mass

spectrometry for additional verification.

2. Methods

2.1. Patients

This was a prospective study of 140 consecutive patients

with idiopathic dilated cardiomyopathy referred to our

Institution as outpatients or for hospitalisation between

January 1998 and March 2003.

In addition, 203 normotensive healthy subjects who

attended the San Raffaele Hospital in Milan gave informed

consent for blood sampling. Subjects with a medical history

of myocardial infarction, heart failure, stroke, diabetes

mellitus, liver disease, use of oral contraceptives, or the

abuse of drugs or alcohol were excluded.

Idiopathic dilated cardiomyopathy was diagnosed on

the basis of the patients’ clinical history, a physical

examination, 12-lead electrocardiography, chest radiogra-

phy, echocardiography, left ventriculography and coronary

angiography according to the WHO criteria [12]. All

patients were in a clinically stable condition and were

taking optimal therapy for at least three months at the time of

study entry.

The study was approved by the local Ethics Committee

and all patients gave written informed consent.

2.2. Echocardiographic examination

Mono and two-dimensional echocardiography record-

ings were obtained using a phased-array echo-Doppler

system (Hewlett Packard Sonos 2500) equipped with a

2.5 MHz transducer. According to the recommendations

of the American Society of Echocardiography [13], left

ventricular end diastolic diameter (LVEDD) was obtained

using a parasternal long axis view; left ventricle ejection

fraction (LVEF) was calculated using Simpson’s rule

[11].

2.3. Cardiopulmonary exercise testing

117 patients underwent symptom-limited bicycle ergo-

meter exercise testing with assessment of oxygen con-

sumption (VO2) by mass spectrometry (Sensormedics

System 2900, Anaheim, CA). The system was calibrated

with a standard gas of known concentration before each test.

The testing protocol consisted of 2 min of free pedalling

followed by 20 W increments every 2 min at a constant

pedal speed of 55–60 rpm. A 12-lead ECG was monitored

continuously and recorded every minute for determination

of heart rate and ST segment changes. Patients were

encouraged to exercise to exhaustion, and all participants

stopped exercise as a result of breathlessness and/or fatigue.

The highest oxygen consumption (peak VO2) at peak

exercise was measured during the last 30 s of symptom-

limited exercise and expressed as millilitres per kilogram

per minute.

2.4. Endogenous ouabain assay and liquid chromatography

mass spectrometry

The blood samples were drawn after the subjects had

rested in a supine position for at least 30 min, collected in

tubes containing EDTA (1.5 mg/ml), and then centrifuged

at 4 -C within 30 min. The plasma was transferred into

plastic tubes and stored at �70 -C prior to analysis. The

assay used an ouabain antiserum with low crossreactivity

for digoxin (¨4%), spironolactone (<0.01%), canrenone

(<0.01%) and canrenoate (0.07%). In addition, all plasma

samples were extracted by C-18 solid phase methods as

previously described [14]. EO was selectively desorbed

using low concentrations of acetonitrile so that digoxin,

spironolactone, and its metabolites canrenone and canrenoate

remained bound. Under these extraction conditions, the

overall assay crossreactivity was minimal (digoxin<0.01%,

spironolactone and related metabolites<0.0001%) so that

EO could be determined with confidence in patients who

were receiving digitalis and/or spironolactone. The dried

sample extracts were reconstituted in water and used for EO

radioimmunoassay as described previously [14]. Briefly, the

detection limit was 25 pmol/L, and other standard curve

parameters were: Kd 3.5T0.2 nmol/L, (inter assay CV 8%,

intra assay CV 5%), Hill coefficient 1T0.01, lower control150 pmol/L (inter assay CV 6%), high control 750 pmol/L

(inter assay CV 3%). In addition, EO levels were assessed

in four randomly selected patients by liquid chromatog-

raphy mass spectrometry (LCMS). The LCMS analysis was

performed using an Agilent 1100 capillary LC system

linked to a Bruker Esquire ion trap mass spectrometer.

Following injection of 1 ml equivalent of the extracted

plasma sample, a gradient of acetonitrile in water was used

to elute bound materials from the capillary LC column. The

effluent was mixed with acetonitrile containing lithium

carbonate and passed to the electrospray interface of the MS

instrument. The column effluent was monitored for molec-

ular ions whose mass to charge ratio matched that for

lithiated EO (i.e., m/z 591). For MS–MS studies, molecular

ions at m/z 591 were isolated and selected for collision

induced dissociation (CID). The intensity of the product

molecular ion corresponding to the lithiated aglycone of EO

ARTICLE IN PRESS

Table 1

Clinical characteristics of patients with and without events during follow-up

All patients Patients with

events

Patients without

events

Number 140 29 111

Age (years) 50T14 57T14 49T14-Men/women 104/36 23 81/20

Body mass index (kg/m2) 26T4 26T4 26T4

Heart rate (bpm) 71T14 70T13 71T14

SBP (mm Hg) 123T13 118T113 124T13*DBP (mm Hg) 78T8 78T8 74T8*

NYHA functional class 1.9T0.7 2.5T0.6 1.7T0.61/2

NYHA functional class (N.)

I 41 1 41

II 74 14 74

III 25 14 25

LVEF (%) 37T10 28T8 39T91/2

LVEDD (cm) 6.07T0.68 6.62T0.84 5.92T0.561/2

EO (pmol/L) 276.3T152.5 382.7T220.5 248.5T115.11/2

Concomitant

medication (%)

ACE inhibitor and/or

AT1 inhibitor

84 100 80-

Beta-blockers 56 48 58

Digitalis 35 59 291/2

Diuretics 54 86 461/2

Spironolactone 13 41 5

Number 117 28 891/2

Peak VO2 (ml/min/kg) 19.1T5.0 16.6T4.3 20.0T4.9-

Mean valuesTSD. NYHA=New York Heart Association. LVEF=left

ventricular ejection fraction; LVEDD=left ventricular end-diastolic diam-

eter; EO=endogeneous ouabain; Peak VO2=O2 consumption at the peak of

exercise during the cardiopulmonary test; *p <0.05 vs. patients with events;

-p <0.01 vs. patients with events; 1/2p <0.001 vs. patients with events.

M.V. Pitzalis et al. / European Journal of Heart Failure xx (2005) xxx–xxx 3

(i.e., m/z 445.4) was determined. Calibration was performed

by injecting known amounts of ouabain over the linear

range (25–500 fmol) of the LCMS combination and

monitoring the retention time and intensity of the lithiated

ouabagenin product ion at m/z 445.4. The EO content of the

samples was obtained by interpolation. The mass spectrom-

etry was performed in Baltimore by an operator who was

blinded to both the clinical status of the patients and the

RIA plasma value.

2.5. Follow-up

The patients were followed up in an outpatient setting,

with scheduled visits every three months and clinical and

instrumental examinations as required. The primary end-

point was the clinical progression of heart failure, which

was prospectively defined as worsening of heart failure

leading to a sustained increase in conventional medication

(beta-blockers, diuretic, ACE inhibitor, AT1 inhibitor,

digitalis), hospitalization, cardiac transplantation or death

[16]. Data on deaths and hospitalizations were collected

regardless of cause. Hospitalisations were classified for

heart failure, and for cardiovascular or noncardiovascular

reasons. Deaths were classified as cardiovascular or non-

cardiovascular; cardiovascular death was defined as death

due to heart failure progression (caused by progressive

hemodynamic deterioration) and as sudden death [17]. For

patients who died outside hospital or in secondary centres,

the relatives were interviewed about the terminal event and

the related charts were collected from the referring physician

or hospital.

2.6. Statistical analyses

Data are presented as meansTSD. Following ANOVA,

normal continuous variables were compared using the t-

test; otherwise, the Mann–Whitney U test was used.

Analysis of covariance was used to compare EO values

between the groups with and without digitalis therapy.

Frequencies were compared using Fisher’s exact test.

Relations between variables were assessed by using the

Pearson correlation coefficient. The Cox proportional-

hazards model was used to assess the association of the

study variables with the events (hazard ratio and 95%

confidence interval, CI, for risk factors are given). The

hazard ratio for a continuous variable refers to the risk

ratio per unit of the analysed variable. To assess the

incremental prognostic value of the variables, an additional

multivariate Cox Regression model was performed in

which the studied variables were added sequentially in

the same order in which they would be considered in

clinical practice. Kaplan-Meyer cumulative survival curves

were also constructed using the median value of plasma

EO to dichotomise the study population into two groups.

The tests were considered statistically significant when the

p value was <0.05.

3. Results

The clinical characteristics of the 140 patients enrolled in

the study are shown in Table 1.

3.1. Clinical correlates

Regression analyses showed that plasma EO was sig-

nificantly correlated with NYHA functional class (r=0.38,

p <0.001), systolic (r=�0.25, p <0.01) and diastolic blood

pressure (r =�0.21, p <0.05), LVEF (r=�0.37, p <0.001),

LVEDD (r =0.27, p =0.001) and peak VO2 (r =�0.33;

p <0.001), but not with age or body mass index. Plasma

EO levels were higher in patients who were taking di-

gitalis therapy (396.43T187.45 vs. 211.59T71.35 pmol/L;

p <0.001).

3.2. Prognostic significance of EO

During follow-up (30T14 months) the following events

were observed: five patients had worsening of heart failure

symptoms, which led to an increase in conventional therapy;

20 patients were hospitalised for heart failure or pulmonary

oedema; three patients underwent urgent cardiac trans-

plantation, one patient died following hospitalisation for

ARTICLE IN PRESS

Table 2

Univariate analysis-predictor (Cox Proportional Hazard Ratio)

HR 95% CI p

Number=140

Age 1.05 1.01–1.08 <0.01

NYHA class 4.78 2.60–8.79 <0.0001

LVEF 0.90 0.86–0.94 <0.0001

LVEDD 3.07 2.01–4.69 <0.0001

EO 1.004 1.003–1.006 <0.0001

Number=117

Peak VO2 0.86 0.79–0.94 0.001

NYHA=New York Heart Association; LVEF=left ventricular ejection

fraction; LVEDD=left ventricular end-diastolic diameter; EO=endogenous

ouabain; peak VO2=O2 consumption at the peak of exercise during the

cardiopulmonary test. Fig. 1. Kaplan-Meier cumulative events reflecting worsening heart failure

in two groups of patients with dilated cardiomyopathy where circulating EO

was either above or below the median value of the study population.

M.V. Pitzalis et al. / European Journal of Heart Failure xx (2005) xxx–xxx4

heart failure. Three patients died suddenly. The clinical

characteristics of patients with and without events are shown

in Table 1.

As shown by the univariate analysis in Table 2, NYHA

functional class, age, LVEF, LVEDD, peak VO2 and plasma

levels of EO were highly predictive for heart failure pro-

gression. EO was predictive for heart failure progression in

patients with (HR: 1.003; 95% CI: 1.001–1.005; p <0.05)

as well as in patients without (HR: 1.006; 95% CI: 1.002–

1.011; p <0.01) digitalis therapy, implying worsening heart

failure at a rate of 1.5 times per 100 pmol/L increase in

plasma EO. Table 3 shows the plasma EO levels in 203

healthy subjects according to age (young/old) and in

patients with idiopathic dilated cardiomyopathy grouped

according to NYHA class. Those patients with the worst

heart failure (NYHA class 3) were found to have higher

circulating EO ( p <0.001). Moreover, when we considered

patients with plasma EO values above or below the median

level (233 pmol/L), those with higher values had an HR of

5.417 (95% CI: 2.044–14.355; p <0.001) for heart failure

progression. The Kaplan-Meier curves for patients with

plasma EO values above or below the median level are

shown in Fig. 1 and illustrate the more rapid decline of

patients with high EO levels.

Circulating EO, when considered as a continuous

variable, remained significantly associated with heart failure

progression after adjustment for age, LVEDD, LVEF,

NYHA functional class and digitalis therapy (Table 4).

The prognostic significance of EO was also evident after

Table 3

Circulating EO levels in healthy subjects according to age and patients with idio

n Sex (f/m) Age (years) Pl

Normal young 151 41/110 38.5 23

Normal old 52 13/39 56.8 22

NYHA 1 41 2/39 40.5 23

NYHA 2 74 25/49 52.7 24

NYHA 3 25 9/16 59.5 42

Pl. EO=Plasma Endogenous Ouabain; SD=Standard Deviation. Normal young a

with IDC were divided according to New York Heart Association class.

peak VO2 was taken into consideration in those patients for

whom this parameter was measured (Table 4). As shown in

Fig. 2, the interactive stepwise procedure revealed the power

of the various relationships to predict major events in

hierarchic order (age; age and NYHA class; age, NYHA

class and LVEF; age, NYHA class, LVEF and EO).

3.3. Endogenous ouabain immunoassay and liquid chro-

matography mass spectrometry

LCMS was used to confirm the presence of EO in the

sample extracts from four patients whose EO was also

determined by radioimmunoassay. Fig. 3 presents the data

obtained for one of the patients and shows the extracted

MS–MS ion current chromatogram for product molecular

ions with m/z 445.4 following CID. The lithiated molecular

ion of the EO aglycone was observed as a large ion current

at 52.6 min. The retention time of the endogenous molecular

ion under the solvent gradient conditions used was similar to

that for the lithiated ouabagenin product ion in this system

(not shown). Fig. 4 shows the product ion scan resulting

from CID of the parent ion of EO at 52.6 min. As expected,

only residual traces of the parent molecular ion (m/z 591)

were observed (large arrow) whereas product molecular ions

at m/z 445.4 and 427.3 corresponding to the lithiated

aglycone of EO and its singly dehydrated counterpart,

respectively, were present. The EO content of the samples

pathic dilated cardiomyopathy grouped by NYHA Class

. EO Mean (pmol/L) SD Median Range

0.36 80.19 210 53–409

2.69 102.78 207 66–726

5.80 88.51 233 105–589

7.96 120.59 225 72–740

6.52 220.51 350 98–956

nd old=healthy subjects matched for age and sex. NYHA 1, 2, 3=patients

ARTICLE IN PRESS

Table 4

Multivariate stepwise Cox proportional hazard analysis for progression of

heart failure among patients with dilated cardiomyopathy

HR 95% CI p

Number=140

Age (years) 1.00 0.97–1.04 NS

NYHA class 2.37 1.13–4.94 0.021

LVEF (%) 0.94 0.89–0.99 0.032

LVEDD (cm) 1.73 1.03–2.89 0.037

EO 1.025 1.003–1.047 0.029

Digoxin therapy 0.78 0.30–2.04 NS

Number=117

EO 1.032 1.013–1.051 0.0008

Peak VO2 0.90 0.82–0.99 0.027

NYHA=New York Heart Association; LVEF=left ventricular ejection

fraction; LVEDD=left ventricular end-diastolic diameter; EO=endogenous

ouabain; peak VO2=O2 consumption at the peak of exercise during the

cardiopulmonary test.

50p<0.001

p<0.02

M.V. Pitzalis et al. / European Journal of Heart Failure xx (2005) xxx–xxx 5

determined by LCMS and the ouabain RIA were correlated

(r =0.89) in linear regression analysis. None of the above

mentioned molecular ions were observed when extracts of

water were used instead of plasma samples.

Neither digoxin nor spironolactone was present in any

significant way in the sample extracts used for the EO

immunoassay. Native plasma samples doped with 10 nM

digoxin (concentrations 5–10 times the normal digitalizing

dose) had EO immunoreactivity after extraction (mean-

TSEM, 265T18 pM) that was indistinguishable from their

(water doped) controls (273T28 pM, n =10). Moreover, no

ouabain immunoreactivity was detected in extracted water

(replacing plasma) irrespective of whether it was doped with

10 nM digoxin or not (i.e., values less than assay thresh-

old¨5 pmol/L). Similar results were obtained when

spironolactone was used up to 500 nmol/L. Thus, it can

be concluded that these drugs do not significantly affect the

EO immunoassay under the experimental conditions of the

present study.

0

40

30

20

10

Age

Age + NYHA class

Age + NYHA class + LVEF

Age + NYHA class + LVEF + EO

p<0.001

Fig. 2. Results of multivariate stepwise analysis. The power (R2) of the

various relationships to predict major events in hierarchic order is shown.

4. Discussion

The primary finding of this study is that increased plasma

levels of EO have an independent and incremental

prognostic value in identifying patients with idiopathic

dilated cardiomyopathy, who are likely to have worsening

heart failure during follow-up. The prognostic value of EO

described in this work is new, and was independent of

demographic, clinical and echocardiographic data and

digitalis administration. Our observations provide more

evidence of a significant role for EO in the complex

scenario of heart failure and may have an impact on therapy.

Several observations, including the natural history of the

relationship between EO and cardiac changes in humans,

together with data from experimental animals, suggest that

EO contributes to the rapid progression of cardiac failure

[2,3,7,8]. In agreement with data from Gottlieb [2], EO

levels were higher in patients with poor hemodynamic status

and these patients were found to have the worst prognosis.

High circulating EO augments the activity of the sympa-

thetic nerves [2,3,18–20], the renin angiotensin system

[21,22], cytokine production [23,24], increases peripheral

vascular resistance [9]. These maladaptive effects of EO are

likely to contribute to the poor prognosis in patients with

high circulating levels of this steroid [25].

EO may impact on various cardiac functional indices

used to define the initial status of patients, as well as the

progression of cardiac failure. For this reason, the inde-

pendent effect of EO may be greater following adjustment

for other indices of cardiac function. Indeed, when EO was

added to the model in Table 4 in which other parameters

were already included, the global R2 value increased

significantly (global R2=44.8). Thus, EO provides additive

prognostic information to the commonly used criteria.

In our patients, left ventricular dysfunction was not due

to ischemia or systemic hypertension and the predictive role

of EO was independent from these and other recognised risk

factors linked with the progression of heart failure. More-

over, as an incremental risk factor, EO seems an attractive

prognostic tool for the identification of patients at the

highest risk, especially when used in combination with the

parameters commonly available in clinical practice. The

informative nature of EO is noteworthy considering that our

evaluations were performed while patients were in a stable

clinical condition and with optimal medical treatment.

Another interesting point is the observation that EO

levels were significantly higher (+88%) in patients who

were receiving digitalis therapy. This is noteworthy for two

reasons. First, the observation does not reflect digoxin

ARTICLE IN PRESS

Fig. 3. LCMS/MS capillary ion chromatogram for a plasma extract from a patient with dilated cardiomyopathy. Shown are the elution characteristics of

positively charged molecular product ions whose mass to charge ratio is equivalent to the lithiated aglycone of ouabagenin (m/z 445.4). The specific current at

52.6 min is the lithiated aglycone of endogenous ouabain.

M.V. Pitzalis et al. / European Journal of Heart Failure xx (2005) xxx–xxx6

interference in the assay for EO. The combination of

differential extraction as well as the limited cross-reactivity

of the EO immunoassay for digoxin effectively excludes

the participation of digitalis glycosides [10,13]. Due to the

design of this study, it was not possible to interrogate the

LC ion chromatograms for evidence of digoxin because

this drug has a mass to charge ratio above the scan range

used to probe for EO (400–650 m/z). Nevertheless, the

absence of digoxin in extracted plasma samples suggests

that we would not have expected to see molecular ions of

digoxin to any significant degree. In addition, in the four

samples that were available for MS analysis, a search for

protonated and lithiated molecular ions of spironolactone

Fig. 4. MS/MS scan of molecular product ions resulting from collision induced dis

scan was performed on the column eluate at 52.6 min. The inset shows the correlat

four patients.

was negative throughout the entire LC run. Therefore,

even when digoxin and/or spironolactone are present in

native plasma, neither drug appears to be present in the

extracted samples used for the EO immunoassay. Secondly,

the prognostic value of EO was independent of digitalis.

This is of great interest because it suggests that cardiac

glycosides such as digoxin and EO may have different

functional consequences in heart failure and because the

Digitalis Investigation Group (DIG) found no overall

impact of digoxin on mortality. However, the DIG study

did not consider EO in its design [26] and whereas heart

failure patients with low and normal circulating levels of

EO may benefit from digoxin, those with higher EO levels

sociation of the lithiated molecular ion of human endogenous ouabain. The

ion (r =0.89) between plasma EO determined by RIA and LC-MS/MS from

ARTICLE IN PRESSM.V. Pitzalis et al. / European Journal of Heart Failure xx (2005) xxx–xxx 7

should, in all likelihood, not be given digitalis [25,26].

Nevertheless, the present study provides the first evidence

that the combination of digoxin and high EO concentrations

in the human circulation is associated with remarkably poor

prognosis in the heart failure setting.

The mechanism of higher circulating EO levels in

digitalised patients likely involves altered secretion and/or

clearance of EO. The kidneys are the primary clearance

route for ouabain and digoxin [27,28] with digoxin being

actively transported into the tubular lumen via p-glycopro-

tein [29]. The basolateral membranes of human and rat

proximal tubular cells actively accumulate digoxin and

ouabain [30] and as the former inhibits ouabain uptake,

digitalis therapy is likely to reduce the renal clearance of

EO. In addition, digoxin may augment EO secretion by

reducing the feedback inhibition of EO on biosynthesis [31].

The relative importance of these mechanisms in digitalized

patients requires further investigation.

Several biomarkers have been described for the pro-

gression of heart failure [32,33]. Among these, brain

natriuretic peptide (BNP) [34,35] appears to be especially

valuable in identifying patients that are likely to have a more

rapid decline in cardiac function. In patients with heart

failure, it seems likely, although not investigated here, that

BNP and EO may be elevated in the same patients.

Moreover, cardiac glycosides augment the secretion of atrial

peptides, including BNP, although this effect appears to be

relatively modest [36]. A recent review of 19 studies

indicated that the risk of death in all cause heart failure

increased 35% for each 100pg/ml increase in BNP [32]. We

did not measure atrial peptides in this study so direct

comparison of the prognostic significance of EO versus

BNP is not feasible. Moreover, the patients in our study

were only followed for 2 years during which time the

increased number of events in the high EO group was

clinically significant but did not include a high number of

fatalities. Therefore, more prolonged studies of the prog-

nostic value of EO are needed to address the question of

mortality.

Nevertheless, the prognostic value of EO in patients

with dilated cardiomyopathy likely reflects a direct func-

tional role for EO. For example, novel pharmacological

agents, along with digoxin antibody fragments that bind

ouabain and EO, block ouabain-induced increases in renal

sympathetic nerve activity, peripheral vascular resistance

and blood pressure in rats [37,38]. Moreover, immuno-

logical neutralization of EO in rats with heart failure

prevented the impairment of baroreflex function, sympa-

thetic hyperactivity, and dilation and dysfunction of the left

ventricle post-myocardial infarction [39,40]. These studies

imply that EO has functional significance in the progres-

sion of heart failure [32].

In this study, we used LCMS to prove that EO was

present in the available sample extracts. The LCMS

paradigm differs from prior work with protonated ions

[4,41] in that the more stable lithiated adducts of ouabain

and EO were monitored. Under dual MS (MS/MS)

conditions, dissociation of the lithiated parent ion led to a

diagnostic product ion representing the lithiated aglycone of

EO (i.e., 445.1 m/z). The appearance of this product ion at

the appropriate LC retention time shows for the first time

that EO can be specifically detected in the human circulation

and quantitated (Fig. 3) from small clinically relevant

volumes of plasma. Moreover, these analytical results

support the overall integrity of several studies that have

used our immunoassay methods.

In conclusion, among patients with otherwise stable

idiopathic dilated cardiomyopathy, high circulating levels of

EO identify those individuals predisposed to progress more

rapidly to heart failure. The prognostic value of EO appears

to be independent of other commonly used parameters. The

mechanism of action that underlies the prognostic signifi-

cance of EO as well as its interactions with other biomarkers

and digoxin requires further study.

Acknowledgements

This work was in part supported by grants from Ministero

Universita e Ricerca Scientifica of Italy (FIRB Grant

RBNE01724C_001 to GB and PRIN Grant 2004069314_01

to GB) and from Ministero della Salute (ICS 110.4/

RF02353), and the USPHS (HL075584) to JMH.

References

[1] Mann DL. Mechanism and models in heart failure. A combinatorial

approach. Circulation 1999;100:999–1008.

[2] Gottlieb SS, Rogowski AC, Weinberg M, Krichten CM, Hamilton BP,

Hamlyn J. Elevated concentrations of endogenous ouabain in patients

with congestive heart failure. Circulation 1992;86:420–5.

[3] Leenen FH, Huang BS, Yu H, Yuan B. Brain Fouabain_ mediates

sympathetic hyperactivity in congestive heart failure. Circ Res

1995;77:993–1000.

[4] Hamlyn JM, Blaustein MP, Bova S, et al. Identification and character-

ization of a ouabain-like compound from human plasma. Proc Natl

Acad Sci U S A 1991;88:6259–66.

[5] Manunta P, Messaggio E, Ballabeni C, Sciarrone MT, Lanzani C,

Ferrandi M, et al. Salt Sensitivity Study Group of the Italian Society of

Hypertension. Plasma ouabain-like factor during acute and chronic

changes in sodium balance in essential hypertension. Hypertension

2001;38:198–203.

[6] Wang JG, Staessen JA, Messaggio E, Nawrot T, Fagard R, Hamlyn

JM, et al. Salt, endogenous ouabain and blood pressure interactions in

the general population. J Hypertens 2003;21:1475–81.

[7] Ferrandi M, Manunta P. Ouabain-like factor: is this the natriuretic

hormone? Curr Opin Nephrol Hypertens 2000;9:165–71.

[8] Ferrandi M, Molinari I, Barassi P, Minotti E, Bianchi G, Ferrari P.

Organ hypertrophic signalling within caveolae membrane subdomains

triggered by ouabain and antagonized by PST 2238. J Biol Chem

2004;279:33306–14.

[9] Manunta P, Stella P, Rivera R, Ciurlino D, Cusi D, Ferrandi M, et al.

Left ventricular mass, stroke volume, and ouabain-like factor in

essential hypertension. Hypertension 1999;34:450–6.

[10] Pierdomenico SD, Bucci A, Manunta P, Rivera R, Ferrandi M,

Hamlyn JM, et al. Endogenous ouabain and hemodynamic and left

ARTICLE IN PRESSM.V. Pitzalis et al. / European Journal of Heart Failure xx (2005) xxx–xxx8

ventricular geometric patterns in essential hypertension. Am J Hyper-

tens 2001;14:44–50.

[11] Richardson P, McKenna W, Bristow M, et al. Report of the 1995

World Health Organization/International Society and Federation of

Cardiology Task Force on the definition and classification of

cardiomyopathies. Circulation 1996;93:841–2.

[12] Shiller NB, Shah PM, Crawford M, et al. Recommendations for

quantitation of the left ventricle by two-dimensional echocardiogra-

phy. J Am Soc Echocardiogr 1989;2:358–67.

[13] Ferrandi M, Manunta P, Balzan S, Hamlyn JM, Bianchi G, Ferrari

P. Ouabain-like factor quantification in mammalian tissues and

plasma: comparison of two independent assays. Hypertension 1997;

30:886–96.

[14] Harris DW, Clark MA, Fisher JF, et al. Development of an

immunoassay for endogenous digitalis-like factor. Hypertension

1991;17(6 Pt. 2):936–43.

[16] Colucci WS, Packer M, Bristow MR, et al. Carvedilol inhibits clinical

progression in patients with mild symptoms of heart failure. US

Carvedilol Heart Failure Study Group. Circulation 1996;94:2800–6.

[17] Narang R, Cleland JG, Erhardt L, et al. Mode of death in chronic heart

failure. A request and proposition for more accurate classification. Eur

Heart J 1996;17:1390–403.

[18] Aileru AA, DeAlbuquerque A, Hamlyn JM, et al. Synaptic plasticity

in sympathetic ganglia from acquired and inherited forms of ouabain-

dependent hypertension. Am J Physiol Regul Integr Comp Physiol

2001;281:R635–44.

[19] Budzikowski AS, Huang BS, Leenen FH. Brain ‘‘ouabain’’, a

neurosteroid, mediates sympathetic hyperactivity in salt-sensitive

hypertension. Clin Exp Hypertens 1998;20:119–40.

[20] Yamazaki T, Akiyama T, Kawada T. Effects of ouabain on in

situ cardiac sympathetic nerve endings. Neurochem Int 1999;35:

439–445.

[21] Zhang J, Leenen FH. AT(1) receptor blockers prevent sympathetic

hyperactivity and hypertension by chronic ouabain and hypertonic

saline. Am J Physiol Heart Circ Physiol 2001;280(3):H1318–23.

[22] Leenen FH, Yuan B, Huang BS. Brain ‘‘ouabain’’ and angiotensin II

contribute to cardiac dysfunction after myocardial infarction. Am J

Physiol 1999;2:H1786–92.

[23] Paulus WJ. Cytokines and heart failure. Heart Fail Monit 2000;1:

50–6.

[24] Matsumori A, Ono K, Nishio R, et al. Modulation of cytokine

production and protection against lethal endotoxemia by the cardiac

glycoside ouabain. Circulation 1997;96:1501–6.

[25] Packer M. The neurohormonal hypothesis: a theory to explain the

mechanism of disease progression in heart failure. J Am Coll Cardiol

1992;20:248–54.

[26] The Digitalis Investigation Group. The effect of digoxin on morbidity

and mortality in patients with heart failure. N Eng J Med 1997;

336:525–33.

[27] Selden R, Margolies MN, Smith TW. Renal and gastrointestinal

excretion of ouabain in dog and man. J Pharmacol Exp Ther 1974;

188:615–23.

[28] St George S, Friedman M, Ishida T. The renal excretion of digoxin in

the normal young subject. J Clin Invest 1958;37:836–7.

[29] Tanigawara Y. Role of p-glycoprotein in drug disposition. Ther Drug

Monit 2000;22:137–40.

[30] Mikkaichi T, Suzuki T, Onogawa T, Tanemoto T, Mizutamari H,

Okada M, et al. Isolation and characterization of a digoxin transporter

and its rat homologue expressed in the kidney. Proc Natl Acad Sci U S

A 2004;101:3569–74.

[31] Laredo J, Hamilton BP, Hamlyn JM. Ouabain is secreted by bovine

adrenocortical cells. Endocrinology 1994;135:794–7.

[32] Jortani SA, Prabbu SD, Valdes Jr R. Strategies for developing

biomarkers of heart failure. Clin Chem 2004;502:265–78.

[33] Doust JA, Piertrzak E, Dobson A, Glasziou P. How well does B-type

natriuretic peptide predict death and cardiac events in patients with

heart failure: systematic review. BMJ 2005;330:625–34.

[34] Richards AM, Doughty R, Nicholls G, et al. Plasma N-terminal Pro-

brain Natriuretic peptide and adrenomedullin: prognostic utility and

prediction of benefit from carvedilol in chronic ischemic left

ventricular dysfunction. Australia–New Zealand Heart Failure Group.

J Am Coll Cardiol 2001;37:1781–7.

[35] Hamada Y, Tanaka N, Murata K, et al. Significance of predischarge

BNP on one-year outcome in decompensated heart failure—compa-

rative study with echo-Doppler indexes. J Card Fail 2005;11:43–9.

[36] Tsutamoto T, Wada A, Maeda K, et al. Digitalis increases brain

natriuretic peptide in patients with severe congestive heart failure. Am

Heart J 1997;1:910–6.

[37] Ferrari P, Torielli L, Ferrandi M, Padoani G, Duzzi L, Florio M, et al.

PST2238: a new antihypertensive compound that antagonizes the

long-term pressor effect of ouabain. J Pharmacol Exp Ther 1998;

285:83–94.

[38] Huang BS, Leenen FH. Brain renin-angiotensin system and ouabain-

induced sympathetic hyperactivity and hypertension in Wistar rats.

Hypertension 1999;34:107–12.

[39] Leenen FH, Yuan B, Huang BS. Brain ‘‘ouabain’’ and angiotensin II

contribute to cardiac dysfunction after myocardial infarction. Am J

Physiol 1999;2:H1786–92.

[40] Huang BS, Yuan B, Leenen FH. Chronic blockade of brain ‘‘ouabain’’

prevents sympathetic hyper-reactivity and impairment of acute

baroreflex resetting in rats with congestive heart failure. Can J Physiol

Pharmacol 2000;78:45–53.

[41] Mathews WR, DuCharme DW, Hamlyn JM, et al. Mass spectral

characterization of an endogenous digitalislike factor from human

plasma. Hypertension 1991;17:930–5.

[42] Manunta P, Lacoviello M, Forleo C, et al. High circulating levels of

endogenous ouabain in the offspring of hypertensive and normoten-

sive individuals. J Hypertens 2005 Sep;23(9):1677–81.