Merits of Non-Invasive Rat Models of Left Ventricular Heart Failure

Department of Environmental Sciences and Engineering, Gillings School of Global Public Health, University of North Carolina, Chapel Hill, 27599 USA.
Cardiovascular toxicology (Impact Factor: 1.72). 06/2011; 11(2):91-112. DOI: 10.1007/s12012-011-9103-5
Source: PubMed


Heart failure (HF) is characterized as a limitation to cardiac output that prevents the heart from supplying tissues with adequate oxygen and predisposes individuals to pulmonary edema. Impaired cardiac function is secondary to either decreased contractility reducing ejection (systolic failure), diminished ventricular compliance preventing filling (diastolic failure), or both. To study HF etiology, many different techniques have been developed to elicit this condition in experimental animals, with varying degrees of success. Among rats, surgically induced HF models are the most prevalent, but they bear several shortcomings, including high mortality rates and limited recapitulation of the pathophysiology, etiology, and progression of human HF. Alternatively, a number of non-invasive HF induction methods avoid many of these pitfalls, and their merits in technical simplicity, reliability, survivability, and comparability to the pathophysiologic and pathogenic characteristics of HF are reviewed herein. In particular, this review focuses on the primary pathogenic mechanisms common to genetic strains (spontaneously hypertensive and spontaneously hypertensive heart failure), pharmacological models of toxic cardiomyopathy (doxorubicin and isoproterenol), and dietary salt models, all of which have been shown to induce left ventricular HF in the rat. Additional non-invasive techniques that may potentially enable the development of new HF models are also discussed.


Available from: Alex P Carll
Merits of Non-Invasive Rat Models of Left Ventricular Heart
Alex P. Carll
Monte S. Willis
Robert M. Lust
Daniel L. Costa
Aimen K. Farraj
Ó Springer Science+Business Media, LLC 2011
Abstract Heart failure (HF) is characterized as a limita-
tion to cardiac output that prevents the heart from supplying
tissues with adequate oxygen and predisposes individuals to
pulmonary edema. Impaired cardiac function is secondary
to either decreased contractility reducing ejection (systolic
failure), diminished ventricular compliance preventing
filling (diastolic failure), or both. To study HF etiology,
many different techniques have been developed to elicit this
condition in experimental animals, with varying degrees of
success. Among rats, surgically induced HF models are the
most prevalent, but they bear several shortcomings,
including high mortality rates and limited recapitulation of
the pathophysiology, etiology, and progression of human
HF. Alternatively, a number of non-invasive HF induction
methods avoid many of these pitfalls, and their merits in
technical simplicity, reliability, survivability, and compa-
rability to the pathophysiologic and pathogenic character-
istics of HF are reviewed herein. In particular, this review
focuses on the primary pathogenic mechanisms common to
genetic strains (spontaneously hypertensive and spontane-
ously hypertensive heart failure), pharmacological models
of toxic cardiomyopathy (doxorubicin and isoproterenol),
and dietary salt models, all of which have been shown to
induce left ventricular HF in the rat. Additional non-inva-
sive techniques that may potentially enable the develop-
ment of new HF models are also discussed.
Keywords Heart failure Heart failure model
Cardiomyopathy Echocardiography Rat Isoproterenol
Doxorubicin Spontaneously hypertensive SHHF
Salt diet
ACE Angiotensin-converting enzyme
ANG II Angiotensin II
bAR Beta adrenergic receptor
CO Cardiac output
DM Diabetes mellitus
DOX Doxorubicin
Peak rate of increase in LV pressure
Peak rate of decrease in LV pressure
DOCA Deoxycorticosterone acetate
DS Dahl salt sensitive
E/A Ratio of early-to-late inflow velocities
EF Ejection fraction
ESV End-systolic volume
FS Fractional shortening
A. P. Carll (&)
Department of Environmental Sciences and Engineering,
Gillings School of Global Public Health, University of North
Carolina, 148 Rosenau Hall, CB# 7431, Chapel Hill,
NC 27599, USA
M. S. Willis
McAllister Heart Institute & Department of Pathology
and Laboratory Medicine, School of Medicine,
University of North Carolina, Chapel Hill, NC 27599, USA
R. M. Lust
Department of Physiology, Brody School of Medicine,
East Carolina University, Greenville, NC 27834, USA
D. L. Costa
Office of Research and Development, U. S. Environmental
Protection Agency, Research Triangle Park, NC 27709, USA
A. K. Farraj
Cardiopulmonary and Immunotoxicology Branch,
Environmental Public Health Division, National Health
and Environmental Effects Research Lab, Office of Research
and Development, U. S. Environmental Protection Agency,
Research Triangle Park, NC 27709, USA
Cardiovasc Toxicol
DOI 10.1007/s12012-011-9103-5
Page 1
HF Heart failure
HR Heart rate
i.v. Intravenous
ISO Isoproterenol
LAD Left anterior descending coronary artery
LV Left ventricular
LVP LV pressure
LVEDP LV end-diastolic pressure
LVESP LV end-systolic pressure
LVOT LV outflow tract
MAP Mean arterial pressure
MHC Myosin heavy chain
MI Myocardial infarction
PO Pressure overload
PTU Propylthiouracil
s.c. Subcutaneous
SD Sprauge Dawley
SERCA Sarcoplasmic reticulum Ca
ATPase pump
SERCA2a SERCA type ‘2’’, isoform ‘a’
SH Spontaneously hypertensive
SHHF Spontaneously hypertensive heart failure
SHR Spontaneously hypertensive rat
SV Stroke volume
TAC Transverse aortic constriction
TNF-a Tumor necrosis factor-a
UPS Ubiquitin–proteasome system
VO Volume overload
Heart failure is a global public health problem imposing
considerable disease and economic burdens on individuals
and societies. Cardiac disease in general leads to more
hospitalizations (4.4 million annually) and deaths (a third
of annual fatalities) in the United States than any other
single cause [1, 2]. Heart failure (HF) is the end-phase
syndrome of chronic cardiac disease that occurs when the
left ventricle becomes too weak to provide adequate tissue
perfusion [3, 4]. According to the U.S. Centers for Disease
Control, HF specifically causes more hospitalizations
(1.1 million annually) than any single type of cardiac dis-
ease [1]. Roughly 6 million Americans live with HF, and
hospital discharges for HF, including deaths, nearly tripled
from 1979 to 2005 [3, 5]. Because HF is difficult for
practitioners to define, and its symptomology is excep-
tionally variable [4], currently reported morbidity and
mortality rates may grossly underestimate the impact of HF
due to the difficulty of classifying the disease [6].
The staggering health and economic burdens imposed
by HF (60,000 deaths and $37 billion annually in the
United States [5]) demand an improved understanding of
how potential therapies and toxins may specifically affect
this syndrome. Investigations into the mechanisms of
exacerbation and treatment of HF are often limited by
the inability of animal models to reliably mimic the
pathogenesis of human HF. Several left ventricular HF
models have been developed in rats using surgical, phar-
macological, dietary, and genetic manipulations, all with
notable shortcomings and advantages. In this review, sev-
eral of the surgically induced models are described briefly,
but primarily as counterpoints to the main emphasis on
non-invasively induced models. In order to focus on a few
major non-invasive models and expose lesser known
emerging models, this review omits several models
reviewed previously, including those induced by myocar-
ditis, alcohol, streptozotocin, and hyperhomocysteinemia,
as well as models of right ventricular heart failure
(cor pulmonale) [7].
In particular, the merits of genetic strains (spontane-
ously hypertensive and spontaneously hypertensive heart
failure), pharmacological models of toxic cardiomyopathy
(doxorubicin and isoproterenol), and dietary salt models in
the rat are described in detail. Additionally, less well-
established non-invasive techniques that may enable the
development of new HF models are also described.
The pathogenesis of left ventricular HF has been reviewed
extensively [811]. Briefly, the hallmark of HF is the
heart’s inability to provide adequate tissue perfusion [4].
This state of cardiac insufficiency usually results after an
initial ‘index event’ (e.g., acute myocardial infarct,
gradual onset of systemic hypertension, myocardial
inflammation, valvular insufficiency, a genetic mutation,
coronary heart disease, diabetes mellitus, and/or cardio-
myopathy) alters ventricular wall stress at end-diastole
(preload) and/or at ejection (afterload), eventually impair-
ing cardiac output [10]. The decline in cardiac function
activates the release of neurohormones (e.g., catechola-
mines, angiotensin II, or endothelin) and cytokines,
increasing water retention and cardiac workload in an
effort to preserve blood pressure and kidney perfusion [10].
If persistent, the neurohormonal response may sequentially
elicit the myocardial fetal gene program, cardiomyocyte
hypertrophy, structural remodeling of the heart, cardiac
dilatation, cardiac insufficiency, and circulatory congestion
(Fig. 1)[3, 8, 10
12]. As well, the type of index event
often dictates the course and features of HF pathogenesis.
For instance, aortic regurgitation and ventricular septal
Cardiovasc Toxicol
Page 2
defect stimulate water retention to increase blood volume
(hypervolemia), augmenting stroke volume through
increased preload, ultimately inducing ventricular volume
overload and cardiac dilatation but typically not ventricular
wall thickening. Conversely, unlike physiologic hypertro-
phy, the innocuous and temporary consequence of physical
exertion, prolonged pathologic cardiac hypertrophy is an
adaptation to increased afterload that is characterized by
several features of cardiac remodeling, including increased
cardiomyocyte size, protein synthesis, and sarcomere
number and organization [1012]. With cardiomyocyte
hypertrophy, left ventricular (LV) volume usually decrea-
ses with the thickening of ventricular walls; however,
chamber dilatation can suppress these two gross features
despite severe cardiomyocyte hypertrophy [13]. Ulti-
mately, decompensation ensues when LV dilatation
(indicated by increased chamber volume, and the thinning,
weakening, and fibrosis of the LV wall) and neurohor-
monal compensatory mechanisms fail to maintain suffi-
cient cardiac output for the body’s needs [8, 11, 14]. At or
near decompensation, the myocardium becomes ATP
deficient due to impaired ATP generation. Consequently,
metabolic function often shifts from fatty acid oxidation
to glucose utilization within the myocardium of rodent
HF models; however, fatty acid uptake remains
unchanged within the failing human heart [15]. Notably,
other important pathophysiological differences in HF
progression exist between rodent models and humans,
including a shift from a-myosin heavy chain (MHC) pro-
tein expression to bMHC in the rodent myocardium and the
opposite response in humans. Although congestion (sys-
temic or pulmonary) is not required for HF, it frequently
occurs along with dyspnea (labored breathing) at rest or
upon mild physical exertion. Other HF symptoms can
include lethargy, dizziness, angina pectoris, weight loss
and/or gain, swelling of the limbs, syncope (fainting), and
cyanosis. The modes of fatality in HF include sudden death
(usually from arrhythmia), congestion (especially pul-
monary), pump failure, insufficient perfusion of the heart
(ischemia and infarction), and inadequate cardiac output
Functional, Structural, and Hemodynamic Indicators
Although this review focuses especially on non-invasive
HF models, it must be noted that confirming HF in even
these models often requires invasive techniques such as
implantable telemetry or ventricular catheterization to
obtain arterial and cardiac pressures as well as ECG
measurements. Conversely, echocardiography provides a
non-invasive means for confirming HF. Heart failure is
typically verified by changes in cardiac performance,
dimension, and pressure. When the heart’s compensatory
Fig. 1 Development and progression of decompensated left ventric-
ular hypertrophy. Compensated left ventricular hypertrophy manifests
as a decrease in chamber volume and an increase in LV wall thickness
characterized by cardiomyocyte growth in response to hemodynamic
stress and/or myocardial injury. Neurohormonal and cytokine
activation induces development of LV hypertrophy, often with
enhancements in LV contractile performance maintaining cardiac
output despite decreased chamber volume. Cardiomyocyte death
provokes transition to cardiomyopathic dilation and wall thinning,
corresponding with decreased LV contractile performance. Along
with a host of molecular changes, these events are termed ‘decom-
pensation’ and usually confer a decline in cardiac output
Cardiovasc Toxicol
Page 3
mechanisms fail and decompensation ensues, LV end-
diastolic volume and pressure both increase, while LV end-
systolic pressure (LVESP) and cardiomyocyte contractile
performance usually decline [10, 11, 17]. Elevated LV end-
diastolic pressure (LVEDP) impedes chamber filling and
reduces stroke volume (SV), thereby serving as a major
indicator of HF—particularly, diastolic failure [1821].
Four-fifths of humans with HF symptoms have diastolic
failure, while less than half have systolic failure [5]. In
these instances, elevations in the mitral valve ratio of early-
to-late inflow velocities (E/A, an echocardiographic
parameter) parallels elevated LVEDP to indicate diastolic
dysfunction and failure [22]. Although susceptible to pre-
load and afterload conditions in general, the peak rate of
increase and decrease in LV pressure (LV dP/dt
) represents contractile and diastolic performance,
respectively [9]. Additionally, significant declines in frac-
tional shortening (FS) and ejection fraction (EF)—both
expressed in percentage values—indicate systolic dys-
function; progressive decreases are prognostic of decom-
pensation and death from HF [23]. For instance, in a large
human cohort study, every incremental 5% decrease in EF
below 45% corresponded with a 15% increase in risk of
death within 2 years [24]. EF is estimated from the dif-
ference in LV volume between systole and diastole (stroke
volume) as a fraction of diastolic volume. LV volumes
typically are calculated from ultrasound measurements of
LV dimension, using well-established geometric assump-
tions of LV shape. In healthy experimental animals, EF
typically ranges from 45% to 70%, depending on anes-
thesia, heart rate, and the procedures used [9]. Notably,
anesthetics differentially affect cardiac output and
hemodynamics as well. A comparison on the effects of iso-
flurane, urethane, sodium pentobarbital, and ketamine–
xylazine in mice demonstrated that—in ascending order and
with many significant differences between agents—all
depressed mean arterial pressure and cardiac output relative
to conscious mice [25].
Invasive Models of HF in the Rat
Models Requiring Thoracotomy
A discussion of the limitations and merits of surgical
models of heart failure is required to convey the relative
advantages of non-invasively induced models of heart
failure. Surgical induction of HF often requires opening the
thoracic cavity. One example of such a procedure is coro-
nary artery ligation, which causes myocardial ischemia and
subsequent necrosis that decreases the number of viable
cardiomyocytes, impairs electrical conduction, and results
in HF accompanied by structural remodeling of the lung by
2 weeks post-surgery [2628]. Following a hypervolemic
response to decreased contractility and stroke volume,
elevated preload is primarily responsible for the progression
to HF in this model, which significantly and dramatically
impairs both systolic and diastolic function [19, 2628]. In
one study, diastolic function was significantly impaired at
2 weeks post-surgery, but systolic function (dP/dt
) was
enhanced—likely due to a neurohormonal augmentation of
contractility (Table 1)[28]. In another study, cardiac per-
formance data and gross clinical signs of HF indicated that
systolic and diastolic failure were achieved concomitantly
at 5 weeks post-surgery [19]. The authors used a threshold
LVEDP value to delineate ligated rats as either ‘non-fail-
ing’ (LVEDP \ 15 mmHg) and ‘failing’ (LVEDP C
15 mmHg). They demonstrated that when MI led to
LVEDP C 15 mmHg (62% of ligated rats), multiple indi-
cators of decompensated HF were evident, including pleural
effusion, ascites, tachypnea, increased diameters of the left
atrium and LV at end diastole, and reductions in weight
gain, heart rate, cardiac output (CO), fractional shortening,
and contractility (LV dP/dt
)[19]. Nevertheless, the
average cardiac performance parameters for all ligated rats
(regardless of LVEDP) still differed dramatically from rats
with sham surgery. Survival was not reported in this study,
but other investigators demonstrated similar cardiac dys-
function at 9 weeks post-surgery with only 10% premature
mortality [29]. An additional study lacking cardiac perfor-
mance data achieved even lower mortality rates (\5%) by
specifically ligating the LAD 1–2 mm below the junction of
the pulmonary conus and the left atrial appendage [30].
Despite the low mortality rates reported previously, coro-
nary artery ligation in rats demands surgical expertise and
precision that, if lacking, can rapidly lead to premature
mortality of 30–62% [27, 30]. Furthermore, all coronary
artery ligation procedures require invasive and potentially
traumatic thoracotomy.
A second invasive model involves constrictive banding
or suturing of the thoracic aorta (often called ‘transverse
aortic constriction’ [TAC] or aortic stenosis), which
induces myocardial ‘pressure overload’ and leads to heart
failure after about 15–27 weeks, resulting in about 30%
premature mortality (Table 1)[3136]. Initially, dramatic
increases in afterload cause hypertrophy and thickening of
the posterior wall and interventricular septum, as well as
increased neurohormones [37]. Aortic constriction impairs
cardiac performance in part by causing dysfunction of the
2a isoform of the sarcoplasmic reticulum Ca
pump (SERCA2a)—a pathology common to HF [33, 38
40]. However, thoracic aortic stenosis, like coronary artery
ligation, can carry high mortality rates and typically
requires surgical opening of the thoracic cavity (sternot-
omy or thoracotomy), which may elicit chronic pain, local
and systemic inflammation, epicardial inflammatory cell
Cardiovasc Toxicol
Page 4
Table 1 Key characteristics of surgical models of left ventricular heart failure
Model Surgical
TAC Coronary artery ligation TAC (no thoracotomy
or sternotomy)
Abdominal Aortic
Aortocaval shunt
Etiology of HF PO MI PO PO VO
Significant changes in LV
performance and time
15 weeks: -11% FS,
-16% EF [36]
21 weeks: -7% FS,
EDP = 15 mmHg [34]
26 weeks: -25% FS [31]
27 weeks: -25% FS [35]
2 weeks: ?28% dP/dt
-35% dP/dt
EDP = 23 mmHg [28]
5 weeks: -38% FS, -26%
CO, -40% dP/dt
-15% ESP,
EDP = 18 mmHg [19]
8 weeks: -22% FS, -33%
9 weeks: -32% dP/dt
-24% dP/dt
EDP = 16 mmHg [29]
27 weeks: -6% FS [43]
32 weeks: -12% FS;
-36% dP/dt
; -33%
EDP = 14 mmHg [33]
41 weeks: -17% FS;
-43% dP/dt
-35% dP/dt
EDP = 18 mmHg [33]
7 weeks: -10% EF, ?140%
14 weeks: ?56% ESP,
EDP = 7.4 mmHg [51]
20 weeks: -9% FS in ‘HF’
subgroup (see ‘Caveats’)
4 weeks: EDP = 12 mmHg;
24 weeks: -28% dP/dt
-31% dP/dt
EDP = 15 mmHg [53]
8 weeks: -4% FS, -5% EF;
16 weeks: -3% FS,
-4% EF;
28 weeks: -4% FS; -6% EF
3 weeks:
EDP = 16 mmHg;
5 weeks:
EDP = 11 mmHg;
8 weeks: EDP = 8 mmHg
4 weeks:
EDP = 16 mmHg;
24 weeks: -31% dP/dt
-39% dP/dt
EDP = 28 mmHg [53]
8 weeks: -16% ESP,
-18% dP/dt
-23% dP/dt
EDP = 17 mmHg;
16 weeks: -26% ESP,
-37% dP/dt
-42% dP/dt
EDP = 27 mmHg [55]
8 weeks: -6% FS,
-7% EF;
16 weeks: -7% FS,
-8% EF;
28 weeks: -6% FS,
-7%EF [37]
Gross HF signs Shallow rapid breathing
(SRB): 15 weeks [36]
Tachypnea, ;wt gain,
pleural effusion, ascites:
5 weeks [19]
SRB, pericardial & pleural
effusions, ascites, pulm.
edema: 27 weeks [39]
Pulm. edema: 20 weeks [52];
none: 24 weeks [53]&
28 weeks [37]
SRB, lethargy, edema,
pleural effusion:
12 weeks [53];
pulm. edema, lethargy,
ascites: 0–32 weeks [56],
16 weeks [37, 55]
Cardiovasc Toxicol
Page 5
infiltration, fibrous adhesions, and extensive thoracic
scarring, and may thus interfere with experimentation long
after surgery [39, 41, 42].
Invasive Models Not Requiring Thoracotomy
Importantly, del Monte and colleagues have developed a
less invasive technique of transverse aortic constriction that
circumvents thoracotomy via a suprasternal incision and
the cutting of a clavicle to access the aorta equally as
proximal to the heart as methods requiring a thoracotomy
[39, 43]. Only a limited number of studies have success-
fully applied the supraclavicular approach in rats to elicit
cardiac dilation and dysfunction comparable to the models
requiring thoracotomy or sternotomy (Table 1)[33, 39, 43,
44]. A similar technique requiring a suprasternal notch has
been developed for the mouse and is used much more
commonly [45, 46]. Abdominal aortic constriction provides
an additional alternative approach to stenotic-increased
afterload that is less demanding of surgical precision,
requires a less traumatic and invasive surgery (laparotomy
does not require mechanically assisted ventilation, pneu-
mothorax, or cutting of bone or cartilage), and eventually
leads to HF. At 4–5 weeks after this surgery, researchers
have observed significant evidence of cardiac hypertrophy
with marked increases in LV wall thicknesses but no sig-
nificant change in EF or FS [37, 47]. In contrast, Kobayashi
and colleagues witnessed significantly enhanced LV sys-
tolic function and pressure at this same time-point with no
impairment in diastolic pressure [48]. Importantly, these
values paralleled those of thoracic aortic constricted rats in
both time and magnitude, suggesting strong homology
between thoracic and abdominal aortic stenosis in the early
compensatory stages of HF progression. In another study,
Wistar rats were shown at 7 weeks post-surgery to have
strong indications of systolic failure, including significantly
decreased ejection fraction and sharply increased end-sys-
tolic volume (Table 1)[49]. Less dramatic effects were
observed at 8 weeks post-surgery in another study using
SD rats [37]. Others have noted that 8 weeks of abdominal
aortic constriction caused 15% mortality, increased systolic
pressure, and did not alter LV dP/dt
relative to control
[50, 51]. Declines in FS, marked myocardial fibrosis, and
pulmonary edema have been demonstrated at 20 weeks
post-surgery in 44% of the rats surviving abdominal aortic
constriction; however, survival rate was not reported [52].
At 24 weeks post-surgery, investigators in another study
observed significantly increased LVEDP and depressed
systolic and diastolic performance with no significant
pulmonary edema, but survival data were not reported [53].
Others observed that 28 weeks of abdominal aortic banding
failed to induce significant pulmonary congestion or ascites
Table 1 continued
Model Surgical
TAC Coronary artery ligation TAC (no thoracotomy
or sternotomy)
Abdominal Aortic
Aortocaval shunt
Etiology of HF PO MI PO PO VO
Mortality 21 weeks: 31% [34]
26 weeks: 30% [31]
9 h: 65% [27]
24 h: 50% [165]
2 weeks: 0% [28]
3 weeks: 5% [30]
9 weeks: 10%; 27 weeks:
39% [29]
2 weeks: 57%, 4 weeks:
82% [39]
2 weeks & 27 weeks:
20% [43]
27 weeks: 29% [33]
24 h: 15% [37]
8 weeks: 15% [50, 51]
24 h: 11% [37]
24–72 h: 8%;
16 weeks: 13% [55]
8 weeks: 28%, 12 weeks:
16 weeks: 60%, 21 weeks:
80% [56]
Caveats Weanlings used [34, 36];
Thoracotomy Weanlings used [33, 39,
Weanlings used [49, 51]
‘HF’ in only 44% [52]
HF signs and death vary
greatly in timing and
incidence [56]
Reproducibility Moderate Moderate–low Moderate–low Moderate–low Moderate–low
Morphology LV hypertrophy ?
LVH ? dilatation LVH ? dilatation LVH ? dilatation LVH ? dilatation
Cardiovasc Toxicol
Page 6
Pressure overload, resulting from chronically elevated
afterload, causes concentric hypertrophy, with thickening
of the posterior and septal walls through addition of sar-
comeres in parallel [37]. Because 75% of HF cases in the
United States have antecedent hypertension [3], the pres-
sure overload-induced model of aortic constriction simu-
lates the most common natural pathogenesis to HF;
nevertheless, it remains limited by premature mortality, a
6- to 7-month delay for HF onset, difficult surgical pro-
cedures, and complications from invasive surgeries.
Additional surgical models of heart failure in the rat that
do not require thoracotomy and pericardiectomy include
rapid ventricular pacing, aortocaval shunt (sometimes
termed ‘fistula’), arteriovenous fistula, and induction of
mitral valve regurgitation [54]. In contrast to pressure
overload, rat models of volume overload hypertrophy
simulate the cardiac effects of anemia, heart block, atrial or
ventricular septal defects, mitral or aortic valve regurgita-
tion, or other congenital diseases that lead to dramatic
elevations in preload. Volume overload causes eccentric
hypertrophy (myocyte elongation by series replication of
sarcomeres), with LV dilatation [37]. Aortocaval shunt (or
fistula) has been used in rats to induce volume overload. A
syringe needle (often 18-gauge) is used to create a puncture
that links the aorta and vena cava, leading to the death of
roughly 10% of male rats at surgery, 25% by 8 weeks, a
highly variable fraction (13–60%) by 16 weeks, and up to
80% by 21 weeks post-surgery [37, 5558]. At least two
groups have demonstrated that at 4–8 weeks after shunt
surgery, LVEDP elevates significantly and peaks at
16-24 weeks post-surgery [53, 55]. In contrast, others have
noted peak LVEDP elevations at 3 weeks, with lesser
elevations at 5 and 8 weeks [59]. A handful of studies have
noted pulmonary edema, lethargy, pitted edema, pleural
effusion, and/or ascites in a large portion of rats after
anywhere from 0–32 weeks of shunt surgery [37, 53, 55,
56, 59]. Despite the effective induction of HF in rats via
aortocaval shunt, the variable timing of both HF onset and
subsequent death limits the utility of this model [37, 56].
Arteriovenous fistula is similar to aortacaval shunt, but
instead of a puncture between the vena cava and the aorta,
the fistula is made in the wall between the carotid artery
and jugular vein [54]; however, this application has also
been mostly limited to large animals such as dogs, goats,
and sheep. Rapid ventricular pacing via chronic electrode
stimulation of the LV for 4–6 weeks increases heart rate
to elicit HF by ‘cardiac overdrive’ with minimal pre-
mature mortality [6063]. Although this model facilitates
the study of the sequential progression of HF over several
weeks, it fails to induce myocardial ischemia and hyper-
trophy, and animals tend to recover from treatment with a
reversal of the induced dilated cardiomyopathy [64, 65].
Moreover, the model has been limited to larger animals
such as dogs, sheep, and rabbits and is thus likely not
suited for rats.
Surgical models bear several advantages, such as dra-
matically impaired cardiac function (particularly for TAC
and coronary ligation) through the simulation or induction
of key stimuli of HF progression (i.e., hypertension-induced
increases in afterload, myocardial infarction–induced
increases in preload). Ultimately, however, surgical models
of HF present several challenges to laboratory research,
including long-term complications from highly invasive
surgeries, difficult surgeries requiring expertise, high pre-
mature mortality, ethical concerns, and variable timing in HF
onset within treatment groups.
Non-Invasive Models of HF in the Rat
Genetic Models
Selectively bred and genetically engineered rodent models
of HF often replicate the natural etiologies of and pro-
gression to HF while avoiding potentially confounding
surgical procedures and toxic treatments. Several geneti-
cally engineered mouse strains have been developed that
closely mimic specific aspects of HF, often through altered
protein expression [6668]. Among rats, however, geneti-
cally engineered HF models have scarcely been researched
[69, 70], while selectively bred HF-prone rat strains have
been extensively investigated and incorporated into toxi-
cologic and therapeutic HF research. Chief among these are
the spontaneously hypertensive (SH) and spontaneously
hypertensive heart failure (SHHF) rat strains, which mimic
several key aspects of human HF pathogenesis while pre-
senting fewer challenges to research than invasive models.
Derived from the inbreeding of hypertensive Wistar
Kyoto rats, SH rats (SHRs) are predisposed to systemic
hypertension, concentric hypertrophy, and thus afterload-
driven HF. At 11 and 27 weeks of age, unrestrained
unanesthetized male SHRs have a 24-h mean arterial
pressure (MAP) of 135 and 150 mmHg, respectively, in
contrast to 100 and 105 mmHg for the normal WKY at
these two ages [7173]. At 18–24 months of age, 57% of
male SHRs have been shown to progress from hyperten-
sion-induced compensated hypertrophy to decompensated
HF, with 13% surviving without HF and 30% dying or
euthanized due to non-cardiac reasons (e.g., stroke, tumor,
or debilitation) [74]. Several features common to human
HF emerge during the SHR’s transition to decompensated
HF, including marked increases in myocardial non-cross-
linked collagen, fibronectin mRNA, LV necrosis, LV
fibrosis, and bMHC protein expression, as well as the
complete loss of aMHC [75, 76]. The transition to
decompensation corresponds with tachypnea and shallow
Cardiovasc Toxicol
Page 7
rapid breathing as well as profoundly decreased EF, heart
rate, and LVESP, and increased LVEDP, end-diastolic and
end-systolic volumes relative to age-matched non-failing
SHRs (Table 2)[74]. Within 1 week of onset of labored
breathing and large drops in echocardiographic measures,
most of these rats have pericardial/pleural effusions, atrial
thrombi, and right ventricular hypertrophy [74, 75]. In one
study, systolic failure was less demonstrable in male SHRs
than diastolic failure. At 20 months of age, these rats
lacked declines in FS or EF despite a doubling of E/A ratio,
27% fibrosis, pleural or pericardial effusions, and thick-
ening of the posterior wall and interventricular septum
relative to 12-month-old SHRs [77].
The SHHF/Mcc-fa
strain originates from the seventh
backcross of the normal SHR with ‘Koletsky obese rats
(inbred from the hypertensive offspring of a SD/SH cross
[78, 79]). SHHF rats possess characteristics similar to the
SHR, except 100% eventually acquire dilated cardiomy-
opathy and HF preceded by Type II diabetes mellitus and
consequent diabetic nephropathy accentuated in obese and
male rats [79, 80]. The SHHF’s additional pathology is
attributed to a nonsense mutation, fa, which encodes a
premature stop codon in the leptin receptor [80, 81].
‘Lean’ and ‘obese’ SHHF rats differ in disease severity
and progression primarily by their responsiveness to
leptin—a hormone released upon eating that inhibits
appetite, provokes a sense of satiation, stimulates the
sympathetic nervous system, and increases energy
expenditure in a receptor-dependent manner [82]. The
autosomal recessive corpulence trait (cp) manifests as
obesity in rats homozygous for the fa mutation (fa
while homozygous wild-type (?/?) or heterozygous
) SHHFs are lean [81, 83, 84]. Among lean males,
heterozygotes develop congestive HF and die sooner than
the homozygous wild types [83, 85]. Heterozygosity
confers mild hyperleptinemia and insulin resistance, with
marked effects in homozygous mutant (fa
) rats [83,
86]. Notably, leptin administration has been shown to
induce eccentric dilatation of the left ventricle [87], while
leptin receptor polymorphisms and circulating leptin
associate with human HF [88]. In contrast to 10- to
12-month-old SHRs with concentric hypertrophy, age-
matched lean male SHHF rats develop eccentric hyper-
trophy and lack ventricular wall thickening [89].
Although leptin stimulates sympathetic nerve activity and
may increase arterial pressure, leptin-induced sympathetic
excitation is absent in the obese phenotype of another rat
strain (Zucker) homozygous for the same mutated leptin
receptor gene [82]. Thus, heart failure in the SHHF is not
entirely a result of hypertension-induced increases in
afterload and may result partly from preload-driven vol-
ume overload. Unanesthetized, unrestrained lean male
SHHF rats have hypertension exceeding the SHR (24-h
MAP: 161 mmHg at 10 weeks and 145 mmHg at
15 weeks) [90, 91], while several studies suggest that
unanesthetized, unrestrained obese male SHHFs have less
severe hypertension (MAP: 119 mmHg at 18–26 weeks,
133 mmHg at 40 weeks, and 127 mmHg at 54 weeks)
[92, 93]. Some studies have reported higher systolic
pressure in the obese relative to the lean; however,
pressure measurements in these studies use anesthesia or
restraint (e.g., tail-cuff), which may differentially affect
the two phenotypes [81, 83, 85].
SHHFs express LV hypertrophy at 3 months regardless
of gender or obesity. Decompensated HF with gross
symptoms occurs at 10–13 months in obese males [85, 92,
94], 15 months in obese females [95], 18 months in lean
males [96100], and 24 months in lean females [85, 98,
101, 102]. The overt signs of decompensated HF found in
the SHHF often include subcutaneous edema, tachypnea
and shallow rapid breathing, cold tails, cyanosis, lethargy,
piloerection, pulmonary edema, pleural effusion, ascites,
cardiomegaly, left and right atrial dilatation, and hepato-
megaly [85]. Death typically occurs at 18 months in obese
females [103] and 19 months in lean males [86, 96]asHF
severity increases with decompensation. Although HF
onset is more rapid in obese SHHFs, lean males compare
well to the hypertrophic qualities of obese SHHFs and also
compare closely to human HF pathogenesis [81]. In 18- to
20-month-old lean male SHHFs, a reduction in aMHC and
b-adrenergic receptor (bAR) density as well as increases in
ventilatory rate ([200 breaths/min), bMHC, circulating
, IL-6, natriuretic peptides, and leptin suggests a
profound comparability between HF pathogenesis in SHHF
rats and humans [86, 96, 97, 104]. Furthermore, excep-
tional homology has been demonstrated between lean and
obese male SHHFs in increases in neurohormonal, apop-
totic, fibrotic, inflammatory, metabolic, hypertrophic, and
structural gene expression at 10 months relative to
4 months of age [81]. Thus, the pathogenesis of HF in lean
and obese males is strikingly similar with exception to rate
of progression.
Among SHHFs, the lean male has been the most thor-
oughly studied for cardiac dysfunction. Yet, timing of
systolic and diastolic dysfunction has been inconsistent
between several of these studies (Table 2)[85, 96, 97]. The
variability in observations may stem from different anes-
thetics used during physiologic measurements; however,
similar variability has not been observed in studies using
other models and different anesthetics. Female SHHFs
differ from males in timing to progression and gross signs
of HF. A few studies have noted the absence of several
common HF traits in 24-month-old lean female SHHFs
[98, 101] despite marked declines in LV systolic and dia-
stolic performance and significant cardiomegaly [98, 102]
(Table 2).
Cardiovasc Toxicol
Page 8
Table 2 Key characteristics of genetic models of left ventricular heart failure
Model Genetic
SHHF lean male SHHF obese male SHHF lean female SHHF obese fem. SHR lean male
Etiology of HF HTN, late-onset leptin
dysfunction, DM
HTN, leptin dysfunction,
DM, metabolic syndrome
HTN, late-onset leptin
dysfunction, DM
HTN, leptin dysfunction,
DM, metabolic syndrome
Significant Changes in LV
Performance and Age
14 weeks: EDP = 14 mmHg;
27 weeks: EDP = 15 mmHg;
63 weeks: EDP = 20 mmHg;
90 weeks: EDP = 25 mmHg,
-40% dP/dt
vs. 14 weeks
old [97]
45 weeks: no change FS vs.
18 weeks old [81]
54 weeks: -56% dP/dt
-50% dP/dt
EDP = 8 mmHg vs. Wistar
Furth rats [85]
81 weeks: -8% FS vs.
27 weeks old [98]; -34%
EF, -16% FS vs. 18 weeks
old [96]
45 weeks: no change FS vs.
18 weeks old [81]
54 weeks: -12% EF,
-50% SV, -16% ESP,
-37% dP/dt
41 weeks old [93]
59 weeks: -47% dP/dt
vs. 41 weeks old [94]
54 weeks: -60% dP/dt
-46% dP/dt
EDP = 8 mmHg vs. Wistar
Furth rats [85]
108 weeks: no change FS vs.
27 weeks old [98]
108 weeks: -38% ESP, -48%
, -57% dP/dt
change FS & EDP vs.
54 weeks old [102]
Not available 90 weeks: no change EF &
FS, ?100% E/A [77]
81–108 weeks old: -28%
EF, -27% ESP,
EDP = 10 mmHg [74]
Gross HF signs Pulm. edema, SRB, cold tails,
cyanosis, lethargy, pleural
effusion: 72 weeks [98],
63–81 weeks [85]
SRB: 90 weeks [97]
None: 45 weeks [81]
edema, pulm. edema, SRB,
cyanosis, lethargy, cold
tails, pleural effusion,
ascites: 45–59 weeks [85]
SRB, edema, cyanosis,
lethargy, pleural effusion,
and/or ascites: 108 weeks
[85, 98, 102]
Pulm. edema, cyanosis,
lethargy, SRB, cachexia,
ascites, hydrothorax:
63–90 weeks [103],
63–81 weeks [85]
Tachypnea at rest; lethargy:
81–108 weeks [74]
Mortality 52 weeks: 10% [91] 41 weeks: 0%;
45 weeks: 32%;
50 weeks: 43%;
54 weeks: 56% [93]
52 weeks: 25% [91]
Not available 63 weeks: 0%;
72 weeks: 25%;
81 weeks: 55%;
86 weeks: 85% [103]
104 weeks: 30%
non-cardiac [74]
Caveats Time to gross signs of
decompensation varies
Time to gross signs of
decompensation varies
Time to gross signs of
decompensation varies
Time to gross signs of
decompensation varies
57% Progress to
decompensation in
2 years, time varies [74]
Reproducibility Moderate–high Moderate–high Moderate Moderate Moderate–low
Morphology LV hypertrophy ? dilatation LVH ? dilatation LVH
? dilatation LVH ? dilatation LVH ? dilatation
Cardiovasc Toxicol
Page 9
In addition to the SHR and SHHF, normotensive male
Fischer 344 (F344) rats at 25-29 months acquire the
symptoms, myocardial histopathology, and cardiac dys-
function inherent to HF [105, 106]. Notably, investigators
observed 50% mortality among F344 rats between 24 and
26 months of age [105]. The F344 strain is best suited as a
model of aging-related HF independent of hypertension.
Given that 75% of HF cases have antecedent hypertension
[3], the etiology of HF in the normotensive F344 strain may
render it less pathophysiologically relevant to the human
condition than the SH and SHHF strains. Furthermore, the
F344’s duration to HF onset is highly variable and pro-
longed relative to the SHHF and SH strains. Nevertheless,
with improved health care management of hypertension in
society, age-related HF that is independent of hypertension
may increase in prevalence and thus become more relevant.
Additional strains with potential as HF models exist,
including the Zucker diabetic fatty (ZDF) rat, which
acquires diabetes- and obesity-associated decrements in
cardiac function; nevertheless, some common traits of HF,
including hypertension and decreased SERCA2a, are not
seen in this strain [107].
The utility of the SHR and SHHF strains extends beyond
their abilities to mimic the etiology of cardiac failure in
humans. These strains have minimal premature mortality
and, along with other strain-based models, lack the major
side effects common to surgical, pharmacologic, or dietary
HF models. Moreover, the steady progression to HF in all
three aforementioned strains (1–2.25 years in SHR, SHHF,
and F344 rats) enables long-term therapeutic treatments
or chronic toxic exposures that are more comparable to
human conditions. Yet, in contrast with many other mod-
els, the relatively slow progression to HF and often variable
time to HF onset in these strains demands extensive time,
husbandry, and resources. Finally, the incorporation of
other methods of inducing hypertrophy or HF in the SH,
SHHF, and F344 strains may expedite their otherwise slow
progression to HF at the risk of compromising their path-
ophysiologic relevance.
Pharmacological Models
Models of pharmacologically induced cardiomyopathy and
ensuing HF often involve simpler, easier, and less trau-
matic methods than surgically invasive models of HF.
Doxorubicin (DOX) and isoproterenol (ISO; either acute or
subchronic) have been shown to significantly impair car-
diac function in rats to the point of HF. Other agents with
promise for yielding models of HF through chronic or
subchronic administration include propylthiouracil, angio-
tensin II, and TNF-a.
Doxorubicin. Doxorubicin (DOX; adriamycin) is an
anthracycline antibiotic and antineoplastic agent used in
cancer chemotherapy that has been shown to induce free
radical formation and extensive lipid peroxidation leading
to myocardial inflammation, morphological disorganiza-
tion of myofibrils, necrosis, apoptosis, interstitial fibrosis,
and cardiac dilatation and failure that is both progressive
and irreversible [54, 64, 108115]. Assisted by the enzyme
NADPH cytochrome P450 reductase, the redox cycling of
the quinone structure of DOX and semiquinone generates
electrons that interact with molecular oxygen to produce
reactive oxygen species (ROS) [110, 116]. Free radicals are
also generated non-enzymatically when DOX reacts with
either iron or nitric oxide [110]. Although DOX elicits
cardiotoxicity primarily through oxidative stress, especially
in mitochondria [111], it causes additional cardiac injury
by inducing proteolysis through activation of matrix
metalloproteinases [110] as well as hyperactivation of the
ubiquitin–proteasome system (UPS) [117, 118]. Despite
that UPS hyperactivation and oxidative stress promote
hypertrophy, evidence that DOX elicits hypertrophy
remains scarce [119]. Instead, DNA lesions in cardio-
myocytes, cardiomyocyte apoptosis, intracellular Ca
overload via SERCA2a dysfunction, depletion of endoge-
nous antioxidants, and disruption of mitochondrial struc-
ture and bioenergetic metabolism are all believed to
contribute to DOX-induced HF [110, 115, 120].
The severity of DOX-induced HF in both patients and
experimental animals is highly dependent upon the cumu-
lative dose [110]. In humans, the likelihood of developing
HF with bolus intravenous DOX delivered once every
3 weeks was 3% after a cumulative dose of 430 mg/m
7% after 575 mg/m
, and 21% after 728 mg/m
, equivalent
to 12, 16, and 20.25 mg/kg cumulative doses in the rat,
respectively [121, 122]. In experimental models, serial
daily or weekly doses of DOX may elicit HF in a more
gradual and survivable manner despite similar decrements
in systolic function relative to a single bolus dose [123].
Liu and colleagues administered 3.3 mg/kg DOX per week
intravenously to SD rats for 4 weeks to cause major
declines in LV systolic and diastolic performance and 14%
mortality at completion of the regimen (Table 3)[113].
Among the DOX-treated rats that survived the 4-wk regi-
men, half of them died over the following 4 weeks.
Slightly lower doses may have dramatically improved
survival; no deaths were observed following this same
regimen at a dose of 3 mg/kg/week DOX, i.v.[124].
Intraperitoneal injection of 2.5 mg/kg DOX six times over
2 weeks in Wistar rats resulted in 100% survival 2 weeks
afterward accompanied by significant declines in FS,
, and LV end-systolic pressure as well as increases
in LVEDP and LV end-diastolic and end-systolic diameters
relative to the control group [125]. At 3 weeks after
completion of this same regimen in 8-week-old SD rats,
investigators noted 25% mortality, marked accumulation of
Cardiovasc Toxicol
Page 10
Table 3 Key characteristics of pharmacological models of left ventricular heart failure
Agent/Procedure Pharmacological
Doxorubicin ISO (injected) ISO (infused) PTU (water) TNF-a
Etiology of HF OS, cmyocyte atrophy,
proteolysis, Ca
Relative hypoxia, OS, Ca
overload, coronary spasm,
phosphate exhaustion, necrosis,
Relative hypoxia, OS, Ca
overload, coronary spasm,
phosphate depletion, necrosis
;myocardial blood flow,
cmyocyte atrophy
; mitochondrial function,
cmyocyte htrphy & apoptosis
Significant changes in
LV performance &
time post-regimen
1 mg/kg/day 9 10 days
?32 days: -15% FS,
?67 days: -25% FS;
10 mg/kg ?2 weeks: -15%
FS [123]
3.3 mg/kg/week 9 4 weeks
?0 weeks: -57% dP/dt
-53% dP/dt
, -45% ESP,
EDP = 16 mmHg [113]
2.5 mg/kg 9 6 days over
2 weeks
?2 weeks: -22% FS, -49%
, -21% ESP,
EDP = 9 mmHg [125];
?3 weeks: -17% FS [126]
3 mg/kg/week 9 5 weeks
?2 weeks: -19% FS;
?4 weeks: -22% FS [127]
340 mg/kg 9 2 ?18 weeks:
-30% dP/dt
150 mg/kg ?1 weeks: -12% FS
?2 weeks: EDP = 16–18 mmHg
?18 weeks: EDP = 16 mmHg
[152, 154]
85 mg/kg/day 9 2 days: -30%
SV, -30% dP/dt
EDP = 14 mmHg [204]
20 ?10 ?5 ?3 ?3 mg/kg over
5 days: -50% dP/dt
, -50%
, -24% ESP,
EDP = 18 mmHg [147]
2.5 mg/kg/day 9 3 weeks: -60%
cmyocyte contractile force [205]
0.1 mg/kg/day 9 14 weeks
?24 h: -11% FS[158]
0.04 mg/kg/day 9 32 weeks
?72 h: -9%FS [52]
9.6 mg/kg/day 9 4 days:
-20% dP/dt
1.2 mg/kg/day:
9 3 days: -42% CO, -47%
SV, no change EF & ESP
9 7 days: -29% CO, -41%
SV, no change EF & ESP
9 6–16 weeks: -10% EF &
FS [167]
9 6 weeks: -11.5% FS,
-14% EF, -17% ESP
9 54 weeks: -26% FS,
-31% EF, -25% ESP
9 27 weeks: -14% FS,
-20% EF, -49% dP/
40 lg/kg/day 9 5 days (inj.):
-9% FS, ?27% Tei-index
3.6 mg/kg/day 9 15 days
(infused): -10% FS [187]
Gross HF signs Ascites: 3 weeks [126] Pulm. edema, anorexia: 18 weeks
Cachexia: 52 weeks [178]
Mortality at time
following final
injection or initial
0 weeks: 14%; 4 weeks: 57%
[113] 4 weeks: 0% [124]
10 inj. ?32 days: 40%;
67 days: 70%;
1 inj. ?2 weeks: 20% [123]
2 weeks: 0% [125]; 3 weeks:
25% [126]
24 h: 55% [206]
48 h: 50% [150]
18 weeks: 20% [152]
24 h: 0% [158]
72 h: 0% [52]
2 days infusion: 0%
7 days infusion: 25% [162]
4 weeks infusion: 5% [132]
23 weeks infusion: 0% [76]
6 weeks: 0%
52 weeks: 18% [178]
Not available
Caveats Multiorgan toxicity; unstable High mortality with single
injection or labor intensive with
multiple injections
Possible post-treatment
reversion of HF traits
No hypertrophy nor
diastolic failure
No necrosis or fibrosis
Reproducibility Moderate Moderate Moderate–high High Moderate–low
Morphology LV dilatation LVH ? dilatation? LVH ? dilatation LV dilatation LVH ? dilatation
Cardiovasc Toxicol
Page 11
ascites, and marked declines in FS and heart weight [126].
A 5-week tail vein administration of 3 mg/kg/week DOX
in SD rats achieved similar reductions in FS as well as
significant ventricular wall thinning 4 weeks after dosing,
but survival was not reported [127]. Therefore, at specific
doses, DOX can elicit systolic failure more effectively than
the aforementioned surgical procedures while minimizing
premature mortality.
Despite evidence that serial administration can be less
lethal yet achieve substantially greater decrements in FS
than a bolus injection at an equal cumulative dose, at least
one study has provided contradictory evidence. Serial
administration of DOX (1 mg/kg DOX for 10 days) caused
greater mortality than a 10 mg/kg bolus injection (40 vs.
20%) and took longer (32 vs. 14 days post-treatment) to
reduce FS by 15% [123]. Serial injection impaired FS the
greatest at 67 days post-treatment (-25%), but this effect
resulted in 70% mortality. Ultimately, the utility of bolus
vs. serial DOX administration, as well as other HF models,
depends on the goals of the study (e.g., short vs. long
monitoring period, level of systolic dysfunction, and sur-
vival rate).
The pathophysiologic relevance of the DOX model to
human HF is limited because it results in HF primarily
through eliciting systolic dysfunction and subsequent pre-
load-driven volume overload while only mildly impairing
diastolic function. Furthermore, the frequent occurrence of
ascites and the scarcity of pulmonary edema in the DOX-
HF model indicate that, relative to the left ventricle, the
right ventricle is disproportionately impaired. Finally,
DOX can have major gastrointestinal, renal, hepatic, and
bone marrow toxicities that may be unsuitable for a model
of heart failure [64, 112, 128].
Isoproterenol (also, isoprenaline or isuprel). Isoprote-
renol is a synthetic catecholamine and non-selective bAR
agonist formerly prescribed as a bronchodilator for asth-
matics as well as a treatment for cardiac arrest, heart block,
and bronchospasm. By stimulating the b
AR, ISO increa-
ses chronotropy (heart rate), inotropy (contractile force),
dromotropy (electrical conduction rate of the atrio-ven-
tricular node), and lusitropy (cardiac relaxation), while it
causes vasodilation and bronchodilation through b
stimulation [129]. As such, ISO bears competing short-
term effects on cardiac afterload (increasing it as a positive
inotrope but decreasing it as a vasodilator and positive
lusitrope), while it decreases preload via tachycardia-
impaired diastole and reduced central venous pressure—
the latter a natural adaptation to increased contractility.
As evidenced by the profound success of beta-blockers
in HF treatment, bAR stimulation by catecholamines is
crucial to HF pathogenesis [130]. Regulating cardiac
function, myocyte growth, and apoptosis, bARs are seven
transmembrane-spanning receptors coupled intracellularly
with guanine-nucleotide-binding regulatory proteins (G
proteins). The heart has three bAR subtypes (b
), among
which b
and b
ARs have the most prominent effects on
cardiac function. The binding of the endogenous cate-
cholamines norepinephrine and epinephrine to bARs
enables G proteins to stimulate adenylyl cyclase, which
opens L-type Ca
channels, enabling Ca
to bind to
ryanodine receptors, thereby triggering the release of stored
from the sarcoplasmic reticulum into the contractile
apparatus to provoke cardiomyocyte contraction while
increasing the synthesis of cyclic AMP, PKA, and—with
AR-activation only—MAPK [55, 130]. The consequent
elevation in cyclic AMP increases transport of myocardial
into the cytosol—enhancing myocardial filament
contractility—and activates sarcoplasmic reticulum cal-
cium-dependent ATPases (SERCA), thereby increasing
efflux and myofilament relaxation. In addition to
enhancing contractility, bAR stimulation by catechola-
mines expedites chronotropic depolarization of the sinoa-
trial node, thereby increasing heart rate, cardiac output, and
workload [9, 131]. In this manner, long-term elevations in
catecholamines can deplete high-energy phosphate stores
as well as SERCA [132]. Elevated myocardial Ca
also increase arrhythmias and risk for sudden cardiac death
[9, 133]. Likewise, increased plasma catecholamine levels
occur in response to overload hypertrophy [132] and
myocardial infarction [134] while correlating with the
likelihood of sudden death in HF patients [135] and pro-
moting HF pathogenesis. Acute catecholamine elevations
are the putative cause of Takotsubo stress cardiomyopathy,
characterized by cyclic AMP-mediated Ca
decreased myocyte viability, contraction band necrosis,
angina, and gross myocardial ballooning [136, 137].
Meanwhile, chronic b
AR stimulation (from elevated
norepinephrine) advances cardiac remodeling via cardio-
myocyte hypertrophy, necrosis, apoptosis, and fibrosis, and
further impairs cardiac performance through depletion of
bARs—especially b
ARs—attendant to increased bAR
kinases [130, 138141].
ISO has been used extensively to induce toxic cardio-
myopathy and HF. Although ISO administration increases
contractility through Ca
accumulation in myocytes, Ca
accumulation inevitably reduces the heart’s ability to relax
and stretch, thereby inhibiting ventricular filling and stroke
volume and increasing preload [133]. ISO has been shown
in several studies to cause tachycardia, hypotension, Ca
overload, free radical generation, coronary vasospasm,
high-energy phosphate exhaustion, and impaired myocar-
dial glucose metabolization in the short term, as well as
decreased SERCA expression, ryanodine receptor PKA
hyperphosphorylation, and bAR down-regulation and
uncoupling in the long term that lead to a progression from
myocardial hypoxia to ischemia, inflammation, necrosis,
Cardiovasc Toxicol
Page 12
apoptosis, and fibrosis accompanied by HF-like symptoms
[90, 134, 142149]. Ultimately, these alterations in the
myocardium decrease contractility and heart rate while
increasing water and Na
retention, thereby decreasing
afterload while increasing preload. In contrast to the DOX-
HF model, the predominance of pulmonary edema and
scarcity of ascites suggest that ISO-induced decrements in
cardiac function occur primarily in the left ventricle.
Traditionally, the model of ISO-induced HF has incor-
porated acute administration (i.e. subcutaneous or intra-
peritoneal injections) of high bolus doses to achieve
cardiac necrosis and associated declines in cardiac func-
tion. Zhang and colleagues subcutaneously injected male
SD rats twice with 340 mg/kg ISO. Four months later,
heart failure was strongly indicated by pulmonary con-
gestion, edema, anorexia, and depressed dP/dt
This dose also significantly reduced responsiveness of the
L-type Ca
receptor, for which declines in function and
expression are known to impair contractility in human HF
[151]. Unfortunately, high mortality (*50% after 48 h)
[150] and hepatic and renal necrosis (at ISO
doses C 200 mg/kg) [152] may limit the utility of this
model of HF.
Lower ISO doses can yield more stable models of HF
while more closely approximating catecholamine eleva-
tions common to HF pathogenesis. For example, a single
150 mg/kg sc injection in male SD rats impaired systolic
function at 1 week post-injection [153], while in female SD
rats, it led to 25% mortality and caused significant LV
diastolic dysfunction and hypertrophy after 2 weeks that
persisted after 4 months [152, 154]. Other studies with
similar dose regimens show variations in survival, cardio-
toxicity, symptomatology, and strain dependence (e.g., pre-
existing hypertension appears to enhance the cardiotoxicity
of ISO administration). Importantly, chronic and sub-
chronic administration of very low doses of ISO may mimic
natural HF pathogenesis more accurately than large acute
doses because they can elicit necrosis and fibrotic cardio-
myopathy through free radical-induced extracellular matrix
biosynthesis [150, 155157]. SD rats injected with
0.04–0.1 mg/kg/day ISO for 3–7 months had very limited
sudden cardiac death and several strong similarities to a
subgroup of aortic banded rats showing overt signs of HF:
impaired myocardial collagen cross-linking, LV chamber
dilatation, an equivalent reduction in FS, and impaired LV
developed pressure–volume relations [52] as well as
marked apoptosis and b
and b
AR inotropic downregula-
tion [158]. Moreover, long-term exposure of cardiomyo-
cytes to beta agonists leads to functional ‘uncoupling’ of
bARs and adenylate cyclase [159], characteristic of human
HF. Nevertheless, the stress inflicted on rats by daily han-
dling and injections over several weeks and the protracted
length of such regimens limit this model’s utility.
Alternatively, osmotic pumps provide a convenient
means to circumvent these issues while maintaining levels
more comparable to endogenous neurohormones. A host of
cardiac alterations consistent with the pathogenesis of heart
failure have been revealed following continuous ISO
infusion with subcutaneously implanted osmotic pumps
[76, 90, 132, 139, 142, 144, 146, 157, 160167]. Infusion
of 2.4 mg/kg/day ISO in Wistar rats caused less than 5%
mortality while it altered gene expression in a strikingly
similar magnitude and time-course as aortic constriction
through day 8 of infusion; however, most of these changes
reverted after 26 days of infusion [132]. Although the
myocardial challenge presented by long-term ISO infusion
may mimic the neurohormonal pathogenesis of HF and
enables substantial reductions in premature mortality rel-
ative to other HF models, only a small handful of rat
studies have examined the effects of long-term ISO infu-
sion on in vivo cardiac function [76, 164, 167]. ISO infu-
sion over 3–7 days in Wistar rats caused significant
declines in LV dP/dt
[164], CO [167], and SV [167], but
did not change LVESP [164, 167]orEF[167] (Table 3)—
consistent with a decrease in contractility and a hypertro-
phy-induced decrease in chamber volume. Nevertheless,
even several hours after treatment cessation, temporary
myocardial adaptations to ISO (e.g., hypertrophy) may
transiently mask underlying cardiac pathophysiology. For
instance, LVESP remained unchanged and LV volume was
decreased at 2 h after a 7-day infusion in one of these
models, but it decreased 2 days post-infusion when LV
volume rebounded to normal [167]. Prolonged infusion
may bear more dramatic and persistent effects; 6-16 weeks
of ISO significantly reduced EF and FS and markedly
increased LV systolic diameter, indicative of systolic fail-
ure accompanied by cardiac hypertrophy [167]. Infusion of
0.02 mg/kg/day ISO for 5 months caused no fatalities and
induced a phenotype in 12-month-old SH rats that com-
pared closely to the LV dilatation, fibrosis, and systolic and
diastolic dysfunction seen in untreated 22-month-old SH
rats in decompensated HF [76]. While these extended low-
dose periods may appear prohibitively long to some
researchers, the induction of HF with pathophysiologically
relevant levels of circulating catecholamines may render a
model with exceptional comparability to human HF path-
ogenesis and low premature mortality.
Angiotensin II (ANG II). In both human HF and animal
models of HF, pro-hypertrophic factors typically induce a
progression from hypertrophy to dilated cardiomyopathy
and subsequent HF. By prolonging stimuli of concentric
cardiac hypertrophy, investigators may render new models
of HF if they can induce sufficient fibrosis, apoptosis, and/
or autophagy in the myocardium. ANG II infusion is one
well-established induction method for cardiac hypertrophy
that bears promise for yielding a new HF model. A major
Cardiovasc Toxicol
Page 13
component of the renin-aldosterone-angiotensin system
(RAAS), ANG II originates from angiotensin I via its
converting enzyme (ACE). ANG II causes vasoconstriction
and sodium retention (leading to afterload-dependent
pressure overload and preloaded volume overload,
respectively) as well as myocardial inflammation, oxidative
stress, and fibrosis [168]. As evidenced by the benefits of
ACE inhibitors in humans with HF and of angiotensin
receptor blockade in a hypertensive rat HF model, ANG II
is known to play a key pathogenic role in HF [169]. Several
investigators have administered exogenous ANG II to
rodents by osmotic pump to induce hypertrophic cardio-
myopathy, but very few have examined the effects of long-
term ANG II administration on cardiac function [67, 161,
168, 170172]. Freund and colleagues elicited significant
cardiac hypertrophy in mice via infusion of ANG II
(2.0 mg/kg/day for 14 days) [171]. Despite perivascular
infiltration of inflammatory cells and increased LV wall
thickness, echocardiographic and histological analysis
indicated no cardiac dilatation, myocyte apoptosis, or
declines in cardiac function [171]. Notably, this regimen
risks vascular complications; a lower infusion rate of ANG
II (1.44 mg/kg/day) for 4 weeks has been shown to induce
aortic aneurysm in 63% and aortic rupture in 25% of
apolipoprotein E-deficient mice [173]. Infusion of lower
concentrations (0.29 mg/kg/day) in mice for 8 weeks did
not significantly change cardiac function or dimension
despite causing cardiac fibrosis [170]. While work on
models of Ang II-induced HF has thus far yielded limited
results, a modified regimen that focuses on prolonged Ang
II exposure at higher concentrations may elicit a more
pronounced fibrosis and eventually lead to dilated cardio-
myopathy and/or HF symptomatology.
Propylthiouracil (PTU). Although hypertrophic models
such as that achieved by ANG II may hold promise for
future models of HF, cardiac dysfunction and HF can occur
independent of hypertrophy in less common cases [174].
Hypothyroidism is a common condition affecting 10% of
women and 6% of men [65 that in some cases can lead to
dilated cardiomyopathy and HF without hypertrophy [175].
Hypothyroidism occurs when the thyroid fails to secrete
sufficient levels of the thyroid hormones thyroxine (T
and/or triiodothyronine (T
), and often manifests clinically
as developmental retardation in children (cretinism) and
depressed mental and physical activity in adults (myx-
edema) [176]. Experimental induction of hypothyroidism
by thyroidectomy has been shown to decrease cardiac
output, blood volume, LVESP, heart rate, aortic pressure,
and LV dP/dt
—a measure of cardiac relaxation [177].
Hypothyroidism induces cardiac unloading (decreasing
both afterload and preload), which if sustained can lead to
cardiac atrophy and dilatation [178]. PTU is an antithyroid
medication prescribed for hyperthyroidism that blocks T
synthesis and has been shown to induce hypothyroidism in
rats with similar effects on cardiac output as thyroidectomy
[178180]. Like the HF seen in other models, PTU treat-
ment represses SERCA and the more efficient alpha iso-
form of myosin heavy chain (MHC) while dramatically
increasing expression of bMHC in the rat heart [68, 75,
181]. Tang and associates demonstrated in female SD rats
that 0.025% PTU in drinking water for 6 weeks and 1 year
was associated with dramatic loss of myocardial arterioles,
major declines in FS, EF, LVESP, and body weight, but no
changes in LVEDP (Table 3)[178]. Another study exam-
ined the effects of this same dose of PTU for 6 months in
female SHHFs to determine if hypothyroidism could
accelerate the progression to heart disease in the context of
hypertensive hypertrophy [180]. The treatment elicited
traits similar to those seen in SD rats, including cardiac
dilatation, unchanged LVEDP, and decreased FS, EF, and
LV dP/dt
; however, in contrast to SD rats, myocardial
arteriolar number was not affected by PTU in SHHF rats.
Ultimately, PTU hypothyroidism parallels HF pathogenesis
by impairing myocardial blood flow while inducing car-
diomyocyte atrophy and cardiac dilatation with series
addition of sarcomeres, but it differs from common HF
etiologies by failing to cause hypertrophy and diastolic
dysfunction [178].
Tumor Necrosis Factor-a (TNF-a) is elevated in HF
and believed to play a major role in HF pathogenesis and
mortality [10, 182]. By impairing contractility, decreasing
stroke volume, and causing LV dilatation [183], one
would predict that TNF-a promotes increased preload and
a progression to HF consistent with volume overload.
Although evidence of TNF-a-induced HF is scarce in rats,
studies of TNF-a overexpression in mice provide further
evidence of the promise for a rat model of TNF-a-induced
HF. Specifically, in genetically modified mice over-
expressing cardiac TNF-a, declines in FS have been
observed along with increased cardiomyocyte apoptosis,
mislocalization of desmin and intercalated disc proteins,
and desmin aggregation [184]. TNF-a administration has
been shown to impair rat cardiomyocyte bAR respon-
siveness to catecholamines [185] while causing coronary
vasoconstriction in canine hearts and severe intracellular
and mitochondrial oxidative stress in the rat myocardium,
significantly impairing systolic and diastolic function by
hindering energy utilization for excitation–contraction
coupling [186] (Table 3). In SD rats, subchronic infusion
of 3.6 mg/kg/day of TNF-a (sufficient to maintain levels
comparable to HF patients) for 15 days caused fibrillar
collagen deterioration, cardiomyocyte hypertrophy, car-
diac dilatation, and impaired FS without contraction band
necrosis or fibrosis) [187]. Further characterization of this
HF model may enhance its utility in therapeutic and
toxicological studies.
Cardiovasc Toxicol
Page 14
Dietary Salt Models
Salt Alone. Consumption of dietary salt stimulates water
retention and, potentially, hypertension, increasing both
afterload and preload. A high salt diet increases sympa-
thetic activity and circulating norepinephrine in humans
with salt-sensitive hypertension while having the opposite
effect on normal individuals [188]. Similarly, administra-
tion of 1% NaCl in drinking water for 4 weeks substan-
tially increases mean arterial pressure, noradrenaline (in
urine, plasma, and the myocardium), adrenaline, sympa-
thetic activity, and vasopressin in the SHR but not the
WKY rat [189191]. Studies have indicated that the
administration of a high salt diet (salt loading) is a feasible
and simple means of inducing heart failure in SH [192,
193], SHHF [194], and Dahl salt-sensitive (DS) rats [195].
The combination of increased afterload from salt-sensitive
hypertension and increased preload from hypervolemia is
critical to the development of HF in salt-fed rats. An 8%
NaCl diet fed to 2-month-old SH and WKY rats for
8 weeks elicits marked ventricular fibrosis [192, 193, 196]
and hypervolemia [196]. In one such study, 13% fatality
occurred in SHRs, as well as signs of congestive HF
(labored breathing and lethargy with decreased FS and a
59% increase in right ventricular mass) among a third of
the surviving SHRs [192]. Unexpectedly, cardiac output
and SV declined only in the non-congestive SHRs, while
congestive SHRs had cardiac dilatation and systolic
hypotension. In a separate study, this regimen affected
neither contractile function nor LVEDP, but it impaired the
rate of diastolic relaxation for both ventricles and caused
premature mortality in SHRs with lesser effects in WKY
rats [193].
The effects of dietary salt on the SHHF strain are not as
well characterized. Mediated by leptin resistance and
increased renal endothelin production, salt sensitivity has
been confirmed in obese—but not lean—SHHF males by
the dramatic exacerbation of hypertension and cardiac
hypertrophy from a 7-day 8% NaCl diet [83]. Notably, one
study demonstrated that salt enhances cardiac hypertrophy
in obese SHHFs independent of hypertension and endo-
thelin, potentially by decreasing nitric oxide production
[197]. Despite the obese SHHF’s salt sensitivity, reports of
salt-accelerated HF in the SHHF strain are scarce and
limited to the lean phenotype [194]. Although no cardiac
function data were provided, one laboratory has reported
eliciting HF in 6-week-old lean male SHHF rats fed an 8%
NaCl diet for 5 months. Similar to aortic banding of SHHF
rats, salt significantly increased pro-apoptotic signaling in
SHHF cardiomyocytes [194].
Salt loading in the DS strain has been shown to rapidly
elicit HF with a swift progression from decompensation
to death (with 8% salt diet, death \ 2 weeks after
decompensation) [195, 198]. Collectively, two studies have
shown that 6-week-old DS rats fed an 8% NaCl diet
develop cardiac hypertrophy after 5–6 weeks, transition to
systolic failure on the 9th week, have full-blown dilatation
and decompensated HF with 60-70% mortality by the 11th
week, and have 100% mortality by the 13th week of salt
diet (Table 4)[198, 199]. Unfortunately, the effects of this
regimen on cardiac function are not entirely consistent. In
another study, the same regimen unexpectedly increased
FS, LV thickness, systolic blood pressure, and molecular
markers of cardiac hypertrophy while causing high fatality
(56%) by the 12th week of salt diet [195]. Thus, the DS and
SH strains are both variable in timing and physiologic
manifestation of salt-induced HF. Survivability may be
improved by either a lower salt dose over a longer period of
time or removal of the elevated salt diet prior to anticipated
Salt with Other Agents. While salt loading often
enhances cardiac fibrosis, acute injections of ISO typically
elicit infarct-like cardiac necrosis. Therefore, the combi-
nation of these two treatments may more readily mimic the
HF pathogenesis found among survivors of myocardial
infarction. To date, only one publication appears to have
addressed the effects of concomitant chronic salt load-
ing and ISO treatment. Treatment with salt (1% NaCl
in drinking water for 2 weeks) did not exacerbate
ISO-induced LV hypertrophy in male Wistar rats [166];
nevertheless, data on cardiac function after ISO salt
co-treatment are scarce. Other methods may enhance the
pathologic effects of salt while simulating the human diet.
Co-treatment of SHRs with 1% NaCl and 5% sucrose in
water induced hypertension, tachycardia, renal excretion of
noradrenaline, and responsiveness to pressor substances
exceeding either treatment alone [190]; however, the
effects of this regimen on cardiac pathology were not
examined. Overall, the potential of such combination
treatments to induce HF remains largely unexplored.
Conversely, the co-treatment of rats with salt and deoxy-
corticosterone acetate (DOCA) is an emerging model of
HF. DOCA increases salt retention and thus enables salt-
induced hypertension in non-salt-sensitive strains. The
combination of DOCA with salt after surgical unilateral
nephrectomy has been shown in rats to significantly impair
LV diastole and, in some cases, LV systole. In one study,
1% NaCl in water combined with bi-weekly 15 mg/kg
DOCA injections for 5 weeks impaired diastolic function
(Table 4)[200]. In SD rats, 1% NaCl drinking water and
100 mg/kg/week DOCA (by subcutaneous injection) for
6 weeks significantly impaired both systolic and diastolic
function [201]. Similar to human HF, these changes were
accompanied by increased plasma markers of cardiac
remodeling, including matrix metalloproteinase-2, tissue
inhibitor of metalloproteinase-1, and osteopontin. Despite
Cardiovasc Toxicol
Page 15
Table 4 Key characteristics of dietary models of left ventricular heart failure
Agent/Procedure Dietary salt Acronyms
Salt-SHR Salt-DS Salt-DOCA
Etiology of HF HTN; LV fibrosis; LV
HTN; HTN Cmyocyte—cardiomyocyte;
CO—cardiac output;
—peak LV pressure increase rate
—peak LV pressure decrease rate
DM—diabetes mellitus;
DOCA—deoxycorticosterone acetate
DS—Dahl salt sensitive;
EDP—LV end-diastolic pressure;
EF—ejection fraction;
ESP—LV end-systolic pressure;
ESV—end-systolic volume;
FS—fractional shortening;
HR—heart rate;
LV—left ventricular;
LVH—LV hypertrophy;
LVOT—LV outflow tract;
MI—myocardial infarction;
OS—oxidative stress;
PO—pressure overload;
SRB—shallow rapid breathing;
TAC—transverse aortic constriction
Tei—a correlate of EDP [207]
VO–volume overload.
‘Reproducibility’ = output/speed with
which one can produce HF based on
labor, expertise, and time required
LV Performance & Time
from Beginning of
8% NaCl 9 8 weeks:
-6% FS in ‘failing’ group
8% NaCl in diet:
9 9 weeks: ?165% E/A, no
change EF & FS [208];
9 9 weeks: -7% FS,
9 11 weeks: -35% FS [198];
9 9 weeks: -16% FS,
9 11 weeks: -15% FS [199];
9 12 weeks: ?11% FS [195]
1% NaCl in water ? 30 mg/
kg/week DOCA 9 5 weeks:
-18% dP/dt
EDP = 7 mmHg [200]
1% NaCl in water ? 100 mg/
kg/week DOCA 9 6 weeks:
-25% dP/dt
EDP = 20 mmHg [201]
Gross HF signs SRB, lethargy, pleural
effusion, ascites, pulm.
edema: 8 weeks [192]
Pulm. congestion:
15–18 weeks [198] SRB,
lethargy: 4–10 weeks [208];
rapid breathing, pleural
effusion/ascites, lethargy,
cachexia: 17 weeks [199]
Not available
Mortality 8 weeks: 13–20% [192,
6 weeks: 30%, 8 weeks: 55%,
9 weeks: 72% [208];
9 weeks: 3%, 10 weeks: 30%,
11 weeks: 70% [198];
11 weeks: 60% [199];
8 weeks: 20%, 10 weeks:
30%, 12 weeks: 56% [195]
Not available
Caveats HF in only 27–31% [192] Variable HF onset time; high
fatality; conflicting
physiologic effects between
studies; age of treatment
determines systolic or
diastolic failure [208]
Nephrectomy performed prior
to DOCA ? salt regimen
[200, 201]
Reproducibility Moderate–low Low Unclear
Morphology Dilatation; RV hypertrophy LV hypertrophy ? dilatation
Cardiovasc Toxicol
Page 16
these findings, there remain only a few rat studies dem-
onstrating reduced cardiac function following DOCA and
salt administration. Moreover, this model typically requires
invasive surgery for the removal of a single kidney. Ulti-
mately, the limited literature covering the effects of
DOCA ? salt on cardiac function, survival, and gross HF
signs limit the reliability of this model for HF studies.
Conclusions: The Application of Rat Models of
Non-invasively Induced HF
As noted, most of the previously described models of
non-invasively induced HF have shown exceptional
comparability to more invasive models. Furthermore, they
often involve routine techniques that require less expertise
while reducing the physical stress and premature mortality
commonly inflicted on animals through surgical tech-
niques. Finally, many of these models have been used to
bolster evidence toward the benefits of therapies and/or
the susceptibilities to toxins in humans with cardiac
failure. For instance, aged SHHF rats and salt-loaded DS
rats have been used to demonstrate that exercise enhances
survival and reduces adverse symptoms in HF [86, 194,
195]. Additionally, the aged SHHF has been used to show
the benefits of inhibition of matrix metalloproteinase and
blockade of the A
receptor [84, 94]. Meanwhile, the
model of PTU-induced hypothyroidism has been com-
pared with surgical models of HF to demonstrate mech-
anisms of micro-RNA function in cardiac remodeling
[68]. DOX and ISO have each been used to demonstrate
the cardioprotective effects of a number of antioxidants
[124, 125, 147, 202]. The DOX model has also been used
to show improved survival and cardiac function with
administration of a neuregulin-1/erbB-activating agent
[113]. Also, an endothelin receptor antagonist has been
shown capable of reversing ISO-induced HF [203]. Col-
lectively, animal models have enabled deep mechanistic
insight into the multifactorial condition of HF. A majority
of the non-invasive methods described herein enable
faster turnaround of simpler and more efficient HF models
while remaining pathophysiologically relevant to human
Acknowledgments The authors thank and acknowledge Drs.
Urmila Kodavanti of the U.S. EPA and David Kurtz of Experimental
Pathology Laboratories for their reviews of this manuscript.
Disclaimer This paper has been reviewed and approved for release
by the National Health and Environmental Effects Research Labo-
ratory, U.S. EPA. Approval does not signify that the contents
necessarily reflect the views and policies of the U.S. EPA, nor does
mention of trade names or commercial products constitute endorse-
ment or recommendation for use.
Declaration of interest Alex Carll is supported by UNC/EPA
1. DeFrances, C. J. & Podgornik, M. N. (2006). 2004 National
hospital discharge survey. In Advance data. United States,
pp. 1–19.
2. Minino, A. M., Heron, M. P., & Smith, B. L. (2006). Deaths:
preliminary data for 2004. National vital statistics reports: From
the centers for disease control and prevention, National center
for health statistics. National Vital Statistics System, 54(19),
3. Rosamond, W., et al. (2008). Heart disease and stroke statis-
tics—2008 update: A report from the American heart associa-
tion statistics committee and stroke statistics subcommittee.
Circulation, 117(4), e25–e146.
4. Coronel, R., de Groot, J. R., & van Lieshout, J. J. (2001).
Defining heart failure. Cardiovascular Research, 50(3),
5. Lloyd-Jones, D., et al. (2009). Heart disease and stroke statis-
tics—2009 update: A report from the American heart associa-
tion statistics committee and stroke statistics subcommittee.
Circulation, 119(3), e21–e181.
6. Goff, D. C., Jr., et al. (2000). Congestive heart failure in the
United States: Is there more than meets the I(CD code)? The
corpus christi heart project. In Archives of internal medicine,