J Appl Physiol 97: 1453–1460, 2004.
First published May 28, 2004; 10.1152/japplphysiol.00834.2003.
Candidate mechanical stimuli for hypertrophy during volume overload
Jeffrey W. Holmes
Department of Biomedical Engineering, Columbia University, New York, New York 10027
Submitted 7 August 2003; accepted in final form 21 May 2004
Holmes, Jeffrey W. Candidate mechanical stimuli for hypertrophy
during volume overload. J Appl Physiol 97: 1453–1460, 2004. First
published May 28, 2004; 10.1152/japplphysiol.00834.2003.—A myo-
cyte system that senses and responds to mechanical inputs might be
activated by any number of features of the time-varying length or
force signals experienced by the myocytes. We therefore character-
ized left ventricular volume and wall stress signals during early
volume overload with high spatial and temporal resolution. Left
ventricular pressure and volume were measured in open-chest isoflu-
rane-anesthetized male Sprague-Dawley rats 4 and 7 days after sur-
gical creation of an infrarenal arteriovenous fistula or sham operation.
Mean wall stresses were calculated by using a simple thick-walled
ellipsoidal model. Consistent with previous reports, this surgical
model produced a 66% increase in cardiac output and a 10% increase
in left ventricular mass by day 7. A number of features of the
time-varying volume signal (maximum, mean, amplitude, rates of rise
and fall) were significantly altered during early volume overload,
whereas many other proposed hypertrophic stimuli, including peak
systolic wall stress and diastolic strain, were not. Treating hemody-
namic variables more generally as time-varying signals allowed us to
identify a wider range of candidate mechanical stimuli for hypertro-
phy (including some not previously proposed in the literature) than
focusing on standard time points in the cardiac cycle. We conclude
that features of the time-varying ventricular volume signal and related
local deformations may drive hypertrophy during volume overload
and propose that those features of the volume signal that also change
during pressure overload might be the most interesting candidates for
further exploration.
arteriovenous fistula; sonomicrometry; rat; growth; mechanics; stress
VENTRICULAR MYOCYTES CLEARLY have the ability to sense and
respond to changes in their mechanical environment (21).
These responses are believed to play an important role in
normal cellular growth as well as in remodeling in a number of
disease states. The classic experimental models of hypertrophy
are pressure overload and volume overload. Both models
produce an increase in ventricular mass, but the spatial pattern
of protein assembly differs. In pressure overload, new sarco-
meres are assembled primarily in parallel with existing sarco-
meres, producing an increase in cell cross-sectional area (1, 2);
at the ventricular level, this growth is reflected in an increased
ratio of wall thickness to cavity radius (2, 12). During exper-
imental volume overload, new sarcomere assembly has been
reported to occur either preferentially in series with existing
sarcomeres, producing relative lengthening of the cells and
dilation of the ventricular cavity (2, 11), or proportionately in
series and parallel (15, 16, 18).
Address for reprint requests and other correspondence: J. W. Holmes,
Engineering Terrace 351, Columbia Univ. MC 8904, 1210 Amsterdam Ave.,
New York, NY 10027 (E-mail: jh553@columbia.edu).
Supplemental data for this article may be found at http://jap.physiology.org/
cgi/content/full/00834.2003/DC1. Raw data, processed data, and data analysis
routines are available from the author at http://www.columbia.edu/⬃jh553/
protocols/protocols.html.
http://www. jap.org 8750-7587/04 $5.00 Copyright © 2004 the American Physiological Society
Downloaded from journals.physiology.org/journal/jappl (177.012.197.016) on November 8, 2023.
Because myocytes normally experience cyclic variations in
both length and force, a cellular system that senses these
quantities (or related variables such as strain and stress) might
respond to minima, maxima, mean values, amplitudes of vari-
ation, frequency of variation, or rates of change. At the ven-
tricular level, this suggests that minimum volume and wall
stress, maximum volume and stress, mean volume and stress,
amplitudes of volume and stress variation, heart rate, and rates
of change of volume and stress should all be considered
candidate mechanical stimuli for hypertrophy. Many of these
possibilities or other related quantities have already been pro-
posed in the literature to explain hypertrophy during hemody-
namic overload (Table 1). However, previous studies of exper-
imental volume overload in the rat have not quantified changes
in all of these variables. In particular, most studies have
measured only cardiac output and heart rate, allowing calcu-
lation of stroke volume but not measurement of the time-
varying volume signal or any of its features.
In addition, most studies of volume overload have focused
on time points several weeks after creation of the overload,
when hypertrophy is established or when animals transition to
heart failure. Because we are interested in the signals that
trigger hypertrophy and because hypertrophy itself may alter
those signals, we focused on very early time points, as close as
possible to the onset of hypertrophy. Preliminary experiments
suggested that the first 3– 4 days after creation of the aortocaval
fistula are dominated by postoperative recovery (reduced body
and heart weights) in our experimental model, so we studied
the period immediately after return to preoperative body
weight, days 4 through 7.
The purpose of this study was therefore to completely
characterize changes in the time-varying left ventricular vol-
ume and estimated wall stress signals during early volume
overload with high spatial and temporal resolution. A number
of features of the time-varying volume signal (maximum,
mean, amplitude, rates of rise and fall) were significantly
altered during early volume overload in this model, whereas
many previously proposed candidate hypertrophic stimuli were
not. Treating hemodynamic variables more generally as time-
varying signals allowed us to identify a wider range of candi-
date mechanical stimuli for hypertrophy (including some not
previously proposed in the literature) than focusing on standard
time points in the cardiac cycle.
MATERIALS AND METHODS
All studies were performed in accordance with the Guide for the
Care and Use of Laboratory Animals (14) and approved by Columbia
University’s Institutional Animal Care and Use Committee. A total of
37 adult male Sprague-Dawley rats were used for these studies;
The costs of publication of this article were defrayed in part by the payment
of page charges. The article must therefore be hereby marked “advertisement”
in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
1453
1454 CANDIDATE MECHANICAL STIMULI IN VOLUME OVERLOAD

Table 1. Early changes in proposed hypertrophic
stimuli (rodent models)
Proposed Stimulus Reported PO Reported VO Current
Peak systolic wall stress (12) ⫹ (8, 20) NS
End-diastolic wall stress (12) ⫹ (8) ⫹ (6) ⫹ interest. Animals were anesthetized with isoflurane (3.5% induction,
Systolic fiber strain (27)* ⫺ (20) ⫹ (26) ⫹ 2.5–3.0% maintenance) and ventilated at 60 breaths/min at a flow rate
Diastolic fiber strain (6)* ⫹ (6) NS
Diastolic cross-fiber strain (22)* NS (6) NS
Contractility (3) NS/⫹ (17, 25) NS/⫺ (4, 15, 28) NS
Ventricular work (19) ⫹†‡
Ventricular efficiency (19) NS†
Mean volume§ ⫹ (7, 20) ⫹
Maximum volume§ ⫹ (7) ⫹ (26) ⫹ The chest was entered by midline sternotomy between two straight
Volume amplitude§ ⫺ (7, 20) ⫹ (9) ⫹‡
⫺dV/dt (19) ⫺ (20) ⫹ loss. Bleeding was controlled by electrocautery. A retractor was
⫹dV/dt§ ⫹ placed, and four 1.0-mm sonomicrometers (Sonometrics, London,
Comparison of reported changes in proposed hypertrophic stimuli during the
first week of pressure overload (PO) or volume overload (VO) in rodent
models. ⫹dV/dt and ⫺dV/dt, maximum rates of volume rise (filling) and fall
(ejection), respectively. Reference numbers indicate study that proposed the
stimulus or reported the changes: ⫹, increase; ⫺, decrease; NS, no significant
change. *Midwall circumferential strain was considered equivalent to fiber
strain and midwall longitudinal strain to cross-fiber strain. †Ejection work
(EW) increased, but the ratio of EW to pressure-volume area (PVA) did not
change. ‡Only EW and volume amplitude correlated significantly with left
ventricular weight-to-body weight ratio (LV/BW) in aortocaval fistula (AVF)
animals. §To our knowledge, these have not been proposed as hypertrophic
stimuli previously.
surgical procedures are described in detail below. Experimental
groups consisted of animals studied 4 days (n ⫽ 7) and 7 days (n ⫽
7) after creation of an aortocaval fistula (AVF) and animals studied 4
days (n ⫽ 7) and 7 days (n ⫽ 8) after sham operation.
Surgical procedures. Infrarenal AVFs were created by using the
method of Garcia and Diebold with minor modifications (4, 10).
Animals were anesthetized with ketamine HCl (80 mg/kg) and xyla-
zine HCl (12 mg/kg) and placed on a warming pad (T/Pump, Gaymar,
Orchard Park, NY) to maintain body temperature. The abdomen was
opened in the midline by aseptic technique. The intestines were
externalized onto a sterile 2 ⫻ 2 and covered with a second sterile 2 ⫻
2 to limit drying. The abdominal aorta and vena cava were exposed
between the renal arteries and the femoral arteries by blunt dissection.
The vessels were manually occluded above and below the fistula site,
and a fistula was created by introducing an 18-gauge short-bevel
needle into the abdominal aorta and through the common wall into the
vena cava. After withdrawal of the needle, the vessel and surrounding
field were blotted dry and the hole in the aorta was sealed with
cyanoacrylate glue (Krazy Glue gel, Elmer’s Products, Columbus,
OH). Patency was easily confirmed by the observation of red stream-
ing blood and pulsation in the vena cava rostral to the fistula site. Total
occlusion time was ⬍5 min. Approximately one-third of the surgeries
required a second occlusion to seal leaks, whereas in two cases the
glue failed and the animals died of rapid hemorrhage. The amount of
flow through the fistula and amount of blood loss during the procedure
were qualitatively scored. The intestines were replaced, the abdomen
was irrigated with 5 ml of 0.9% saline containing 50,000 units/ml
penicillin G potassium, and the muscle layer was closed with 3-0
chromic gut suture. The skin layer was closed with glue and staples
(Autoclip 9 mm, Becton-Dickinson, Sparks, MD), and the animals
were allowed to recover on a warming pad. Each animal received a
single dose of 0.3 mg/kg buprenorphine HCl subcutaneously for
postoperative analgesia. Animals were monitored for postoperative
discomfort but in general did not require additional doses of buprenor-
phine. Sham animals underwent an identical operation including blunt
dissection to expose the vessels and occlusion of the vessels for a
similar length of time but no perforation of the aorta or vena cava.
Data collection. Four or 7 days after the initial surgery, animals
underwent a nonsurvival study to measure ventricular dimensions and
pressures, which were then used to calculate other parameters of
of 300 – 400 ml/min and 1:1 ratio of inspiration to expiration time to
achieve a tidal volume of ⬃3.0 ml (SAR-830/P small animal venti-
lator, IITC Life Science, Woodland Hills, CA). Body temperature was
maintained with a warming pad. Great care was taken to minimize
blood loss because of the potential effects on the data to be collected.
hemostats placed from a small hole in the diaphragm to prevent blood
Ontario, Canada) were sewn to the epicardium with 6-0 silk suture in
the following order and locations: 1) on the posterior wall two-thirds
of the distance from apex to base and just lateral to the intraventricular
groove (posterior); 2) on the anterior wall two-thirds of the distance
from apex to base and just lateral to the intraventricular groove
(anterior); 3) on the lateral wall under the left atrial appendage (base);
and 4) at the left ventricular apex (apex). A precalibrated Millar
SP-671 miniature pressure transducer (Millar Instruments, Houston,
TX) was zeroed in warm distilled water and inserted into the left
ventricle through a hole created with a 25-gauge needle. After instru-
mentation, the hemostats were removed, hemostasis was verified, and
a needle ground electrode was placed into the muscle of the upper
right hindlimb and connected to the Sonometrics system. The instru-
mentation period required 40 – 45 min from the start of induction.
A 5-s run of segment length and pressure data was recorded each
5 min for 1 h by using the commercial software SonoLab (Sonomet-
rics) with an acquisition rate of 312 Hz. Anesthesia was maintained at
a depth that just suppressed spontaneous respiration (typically 2.75%),
and depth was monitored frequently by observation and by toe pinch.
After 1 h of data acquisition, a tie was placed around the inferior vena
cava and a single longer data run was acquired during inferior vena
cava occlusion. The animal was then heparinized (100 units injected
directly into each ventricle), and blood gases were measured in
samples taken directly from the right and left ventricles. Blood-gas
analysis was performed by use of an iStat portable blood-gas monitor
(iStat, East Windsor, NJ) to verify the presence of a left-to-right shunt
in fistula animals. Finally, hearts were arrested by retrograde perfu-
sion with cold Krebs-Henseleit buffer containing 30 mM 2,3-butane-
dione monoxime and removed for analysis. The atria were trimmed
away, and overall dimensions as well as the positions of the sonomi-
crometers were measured and recorded. The sonomicrometers were
then removed, and the left and right ventricles were weighed sepa-
rately and fixed in 3.7% formaldehyde for histological analysis.
Data analysis. With attention to signal quality during original
acquisition, very little correction or filtering of the sonomicrometer
signals was required. These signals were viewed by use of SonoView
(Sonometrics), and the best quality base-apex and anterior-posterior
signal pair was selected. Occasionally, single points fell well outside
the length range of a given tracing and were corrected by using
built-in software features; the number of corrections was never more
than 10 in a data run of 1,500 points. Base-apex segment length,
anterior-posterior segment length, and ventricular pressure (Fig. 1)
were then exported to an ASCII file for analysis using custom routines
written in Matlab (v. 5.0, The MathWorks, Natick, MA). In a few
cases, a premature ventricular contraction occurred during data acqui-
sition; in this case, the premature contraction and subsequent beat
were excluded from the exported data.
Custom Matlab software was used to calculate the volume enclosed
by the epicardium, with the assumption that the left ventricle can be

J Appl Physiol • VOL 97 • OCTOBER 2004 • www.jap.org

Downloaded from journals.physiology.org/journal/jappl (177.012.197.016) on November 8, 2023.
 CANDIDATE MECHANICAL STIMULI IN VOLUME OVERLOAD
Fig. 1. Sample data illustrating typical original tracings and calculated pres-
sure-volume loops. A: raw base-apex segment length (BA), anterior-posterior
segment length (AP), and left ventricular pressure (LVP) tracings during the first
0.6 s of a 5-s data run. B: pressure plotted vs. calculated cavity volume for the full
5 s of a data run in an animal with an aortocaval fistula (black) and in a
sham-operated animal (gray). Open circles, automatically selected end diastole;
closed circles, automatically selected end-systole points for each beat.
modeled as a prolate spheroid truncated 50% of the distance from
equator to base as described by Streeter and Hanna (23):
V ⫽ 共␲/3兲关2 ⫹ 3共0.5兲 ⫺ 共0.5兲3兴共AP/2兲2共BA/1.5兲
(1)
⫽ 1.125共␲/6兲AP2BA
where AP is the measured anterior-posterior segment length and BA
is the base-apex segment length. Cavity volume was calculated by
subtracting wall volume on the basis of the measured left ventricular
weight and a density of 1.06 g/cm3 for myocardium. End diastole was
selected automatically as the point immediately preceding the rapid
pressure upstroke and end systole as the point of maximal elastance,
Emax ⫽ left ventricular pressure/left ventricular volume, assuming a zero
volume intercept (Fig. 1). A graphical interface then presented each end
diastole-end systole pair to the user for confirmation or rejection; in the
case of rejection, the beat was excluded from further analysis. Less than
10% of all beats analyzed for this study were rejected.
Mean circumferential (␴c) and longitudinal (␴l) wall stresses were
estimated using the formula suggested by Mirsky and recommended
by Yin (29):
J Appl Physiol • VOL 97 • OCTOBER 2004 •
Downloaded from journals.physiology.org/journal/jappl (177.012.197.016) on November 8, 2023.
1455
␴ c ⫽ 共Pb/h兲共1 ⫺ b2/2a2 ⫺ h/2b ⫹ h2/8a2兲
(2)
␴ l ⫽ 共Pb/2h兲共1 ⫺ h/2b兲2
where a is half the major axis length, b is half the minor axis length
at the midwall, and h is the wall thickness. Wall thickness was
calculated throughout the cardiac cycle from the epicardial segment
lengths and known wall volume with the assumption that the endo-
cardial and epicardial surfaces are confocal nested prolate spheroids
and that wall volume does not change substantially during the cardiac
cycle. Mean circumferential (εc) and longitudinal strain (εl) were
estimated from midwall dimensions as
ε c ⫽ 0.5 ⫻ 关共b/b0兲2 ⫺ 1兴
(3)
ε l ⫽ 0.5 ⫻ 关共a/a0兲2 ⫺ 1兴
For end-diastolic strain, deformed long- (a) and short-axis (b)
midwall diameters were taken at end diastole and reference dimen-
sions a0 and b0 at the point of minimum pressure. For end-systolic
strain, values of a and b at end systole were compared with a0 and b0
at end diastole.
Features of the time-varying volume, pressure, and stress curves
were identified and averaged across all beats in a given data run as
follows. Values of all parameters were calculated at end diastole and
end systole. In addition, the minimum, maximum, mean, greatest rates
of increase and decrease, and frequency of each signal were calculated
for each beat. For each beat, ejection work (EW) was computed as the
area inside the pressure-volume loop and pressure-volume area as EW
plus the area bounded by the pressure-volume loop on the right, a line
from the end-systolic point to the origin on the left, and the volume
axis (24). This analysis was performed for data runs at 15, 30, and 45
min of the 1-h data collection period to confirm stability of the
experimental preparation. Any parameter that varied ⬎10% across
these time points was flagged and reviewed. Typically this occurred
when the absolute value of a parameter (e.g., minimum or diastolic
pressure) was very small. To facilitate display and comparison,
average interpolated volume, pressure, and stress curves were gener-
ated for each group as follows. Given heart rates near 360 beats/min
and a data-acquisition rate of ⬃300 Hz, each heartbeat was repre-
sented in the data set as roughly 50 data points. All parameters were
therefore interpolated for each beat at 50 evenly spaced points by
linear local interpolation. These interpolated curves were averaged
across all beats in a data run, then across three data runs (15, 30, and
45 min) for each animal, and finally across the animals in each group
to create average curves.
Statistics. Statistical analysis was performed using InStat (Graph-
Pad, San Diego, CA). Trends in body weight over the course of the
study were analyzed by one-way repeated-measures ANOVA and
then with a Dunnett’s multiple comparisons test to determine the time
points at which body weight differed significantly from its preopera-
tive value. For all other parameters, unpaired t-tests were used to
determine whether values were significantly different between fistula
and sham groups at each time point.
RESULTS
Twenty-one animals were randomized to fistula groups;
three died during the surgery (two of acute hemorrhage, one
during anesthesia), one died postoperatively of acute cardiac
failure with massive pulmonary edema, and one was killed
because of labored breathing indicative of acute failure, yield-
ing a net survival rate of 76%. Of the 16 animals studied, two
were excluded because right ventricular PO2 fell within the
95% confidence interval of the sham mean, indicating minimal
shunt flow. Of 16 animals that underwent sham surgery, there
was one unexplained postoperative death.
www.jap.org
1456 CANDIDATE MECHANICAL STIMULI IN VOLUME OVERLOAD

Heart and body weights. Preliminary studies had suggested Table 2. Heart weight, body weight, and hemodynamics
that both body weight and left ventricular weight decrease during early volume overload
during the first 2–3 days after fistula creation and sham sur-
Parameter 4-day Sham 4-day Fistula 7-day Sham 7-day Fistula
gery. We therefore focused on the first 3 days after return to
preoperative body weight (postoperative days 4 to 7) as most LV weight, mg 611⫾16 688⫾40 676⫾12 745⫾10*
clearly representing early hypertrophy. As expected on the RV weight, mg 165⫾5 177⫾12 173⫾7 221⫾9*
Body weight, g 246⫾6 241⫾8 277⫾5 261⫾11
basis of preliminary studies, body weight varied significantly LV/BW, mg/g 2.49⫾0.09 2.87⫾0.20 2.44⫾0.04 2.88⫾0.12*
with time in the 7-day fistula group (ANOVA P ⬍ 0.0001), RV/BW, mg/g 0.67⫾0.03 0.74⫾0.06 0.63⫾0.02 0.86⫾0.06*
being reduced relative to preoperative values at days 2 and 3 HR, beats/min 373⫾9 346⫾11 385⫾6 333⫾14*
(P ⬍ 0.05) but not significantly different on days 4 through 7 SV, ␮l 173⫾16 326⫾49* 187⫾17 360⫾27*
CO, ml/min 65⫾6 111⫾14* 72⫾6 119⫾9*
(Fig. 2). Sham body weights were significantly reduced on RV PO2, Torr 66⫾3 81⫾4* 63⫾2 84⫾5*
postoperative day 2 and returned to preoperative values on day EDV, ␮l 445⫾75 570⫾95 319⫾45 535⫾57*
3 (Fig. 2). There were no significant differences in body weight ESV, ␮l 271⫾63 244⫾60 131⫾29 174⫾36
between the sham and fistula groups preoperatively, at 4 days, EDP, cmH2O 6.1⫾0.7 9.1⫾1.5 5.7⫾0.5 7.8⫾1.2
ESP, cmH2O 120⫾11 130⫾12 122⫾6 106⫾8
or at 7 days.
Both absolute and relative left and right ventricular hyper- Changes in heart weight (HR), body weight (BW), and hemodynamic
trophy were absent at day 4 but present by day 7 after creation parameters measured in the present study (means ⫾ SE). LV, left ventricle;
RV, right ventricle; SV, stroke volume; CO, cardiac output; EDV, end-
of the fistula (Table 2). Left ventricular weight was 10% diastolic volume; ESV, end-systolic volume; EDP, end-diastolic pressure;
greater than sham (P ⫽ 0.001), and right ventricular weight ESP, end-systolic pressure. *Statistically significant difference between sham
was 27% greater than sham (P ⫽ 0.001) in the 7-day fistula and fistula groups.
group. Left ventricle-to-body weight ratio was 18% greater in
the 7-day fistula group than in the 7-day shams (P ⫽ 0.003),
whereas the ratio of right ventricular weight to body weight in end-diastolic volume (P ⫽ 0.01) without a change in
was 37% greater (P ⫽ 0.002). end-systolic volume (Fig. 3). At 4 days, neither end-diastolic
Hemodynamics. Consistent with the presence of a large nor end-systolic volume was significantly different between the
arteriovenous fistula, stroke volume was 98 and 92% greater in fistula and sham-operated groups. Diastolic and systolic pres-
4-day (P ⫽ 0.02) and 7-day (P ⫽ 0.0001) fistula groups, sures were not significantly different between fistula and sham-
respectively, than in the corresponding sham groups, whereas operated groups at either 4 days or 7 days.
cardiac output was 81% (P ⫽ 0.01) and 66% (P ⫽ 0.0004) Candidate hypertrophic stimuli. Table 1 lists a number of
greater (Table 2). Heart rate was lower in the fistula groups putative hypertrophic stimuli. Most of these signals showed
than in the sham groups, but this change was only significant at little change 4 or 7 days after creation of a fistula (Table 3).
7 days (P ⫽ 0.08 at 4 days, P ⫽ 0.003 at 7 days). At 7 days, End-diastolic mean wall stresses were higher in the fistula
the increase in stroke volume was due to a significant increase group at 7 days but not at 4 days. End-systolic mean circum-
ferential strain was greater (more negative) in the fistula group
at both 4 and 7 days, and EW was increased in the fistula
groups at both times (Table 3). Peak systolic wall stresses,
end-diastolic strains, and contractility were not significantly
different in the fistula groups. We also examined whether any
of these stimuli were significantly correlated with the ratio of
left ventricular weight to body weight (LV/BW) across the
fistula groups and found that only EW (r2 ⫽ 0.426, P ⫽ 0.011)
and volume amplitude (r2 ⫽ 0.390, P ⫽ 0.017) were signifi-
cantly correlated with LV/BW.
Volume and stress signal features. Average interpolated
volume and stress curves are shown in Fig. 4. Most features of
the time-varying volume signal were significantly different
between fistula and sham groups at 7 days. Maximum volume
(P ⫽ 0.01), the amplitude of volume variation (P ⫽ 0.0002),
mean volume (P ⫽ 0.04), and maximum rates of volume rise
(⫹dV/dt, P ⫽ 0.01) and fall (⫺dV/dt, P ⫽ 0.0002) were all
significantly greater in the fistula animals, whereas minimum
volume was not significantly different (Table 4). Of these
Fig. 2. Trends in body weight (means ⫾ SE) in sham-operated animals (gray signals, only amplitude of volume variation (P ⫽ 0.03) and
circles, n ⫽ 8) and in animals with abdominal aortocaval fistulae (AVF; black
circles, n ⫽ 7). Preoperative and postoperative body weights are plotted at 0 ⫺dV/dt (P ⫽ 0.05) were significantly greater in 4-day fistula
and 0.2 days for clarity. AVF body weight was reduced on days 2 and 3 and animals than in the 4-day shams. None of the features of the
returned to preoperative value on day 4; sham body weight was reduced on day time-varying circumferential or longitudinal wall stress signals
2 and returned to preoperative value on day 3. *Body weight significantly
different from its preoperative value by Dunnett’s multiple-comparison post-
(maximum, minimum, amplitude, mean, or maximum rates of
test (ANOVA P ⬍ 0.0001 for both groups). AVF and sham body weights were rise or fall) were significantly different between fistula and
not different preoperatively, on day 4, or on day 7. sham groups at 4 or 7 days.
J Appl Physiol • VOL 97 • OCTOBER 2004 • www.jap.org
Downloaded from journals.physiology.org/journal/jappl (177.012.197.016) on November 8, 2023.
 CANDIDATE MECHANICAL STIMULI IN VOLUME OVERLOAD
Fig. 3. Average interpolated pressure-volume loops for the 4 groups of
animals studied. F, Fistula groups; E, sham groups. A: 4-day data. B: 7-day
data. Stroke volume was increased in both 4-day and 7-day fistula groups; at
7 days, this was due to increased end-diastolic volume with no significant
change in end-systolic volume, whereas at 4 days neither end-diastolic nor
end-systolic volumes were significantly different. End-diastolic and end-
systolic pressures were not significantly different from sham.
1457
suggested approaches to prioritizing these candidate signals for
subsequent in vitro studies.
Candidate stimuli: narrowing the field. Because so many
mechanical and chemical signals change simultaneously during
experimental volume or pressure overload, most investigators
have focused on signals that both change during overload and
meet other a priori conditions. For example, some investigators
have postulated that a hypertrophic stimulus should be normal-
ized by the hypertrophy it induces, forming a negative feed-
back loop (6, 12). Other possible strategies include focusing on
stimuli that show the largest relative changes, stimuli that
change before the onset of hypertrophy, stimuli that correlate
quantitatively with the degree of hypertrophy in a given model,
or stimuli that explain the responses observed in multiple
different experimental models.
About half of the candidate mechanical signals considered in
this study showed no significant difference between the sham
and the fistula groups (Tables 1, 3, 4). Among the remainder,
features of the time-varying volume signal showed the largest
relative increases, typically 50% at 7 days. Two of these
volume signals, the amplitude of volume variation and ⫺dV/dt,
were also among only four candidate signals significantly
increased by 4 days, closer to the onset of hypertrophy (Table
4); the others were end-systolic circumferential strain and EW,
both related to the amplitude of volume variation (Table 3).
Interestingly, EW and volume amplitude were also the only
two candidate signals from Table 1 that were significantly
correlated with LV/BW across all fistula animals in this study.
Taken together, these results suggest that volume amplitude,
⫺dV/dt, and related signals are the most fruitful candidates for
further investigation.
Given what is already known about myocyte mechanotrans-
duction, it seems unlikely that features of the time-varying
cavity volume signal are transduced directly. Local deforma-
tions associated with the cavity volume signals reported here
are more likely candidates for in vitro study. It is therefore
perhaps surprising at first glance that the estimated strains were
not more prominently altered by volume overload. Because
stroke volume increased in fistula animals entirely by an
increase in end-diastolic volume, measures normalized by
Table 3. Proposed hypertrophic signals during
early volume overload
Parameter 4-day Sham 4-day Fistula 7-day Sham 7-day Fistula

ED␴c, kPa 1.2⫾0.2 2.1⫾0.5 0.8⫾0.2 1.5⫾0.3*

DISCUSSION ED␴l, kPa 0.7⫾0.1 1.2⫾0.3 0.4⫾0.1 0.8⫾0.1*
peak ␴c, kPa 24.2⫾3.7 25.9⫾3.3 16.7⫾1.7 19.8⫾2.4
The overall goal of our work is to identify candidate me- peak ␴l, kPa 13.1⫾2.2 14.2⫾2.3 8.8⫾1.0 10.4⫾1.3
chanical stimuli for hypertrophy and directly test them one at EDεc 0.18⫾0.04 0.28⫾0.03 0.28⫾0.04 0.34⫾0.02
EDεl 0.18⫾0.04 0.20⫾0.03 0.19⫾0.03 0.25⫾0.03
a time in vitro. Because biological systems typically have ESεc ⫺0.12⫾0.02 ⫺0.17⫾0.01* ⫺0.16⫾0.01 ⫺0.20⫾0.01*
nonlinear responses, we consider it critical to test candidate ESεl ⫺0.09⫾0.02 ⫺0.13⫾0.01 ⫺0.13⫾0.02 ⫺0.16⫾0.01
stimuli at physiologically relevant levels and therefore chose to Emax cmH2O/ml 482⫾143 710⫾149 983⫾179 748⫾140
first carefully quantify changes in candidate stimuli in vivo ⫹dP/dt, cmH2O/s 7,790⫾450 7,750⫾270 8,820⫾470 7,490⫾640
during experimental volume overload. We characterized the EW, cmH2O䡠ml 23.8⫾2.2 42.4⫾7.3* 23.8⫾2.5 35.6⫾3.2*
PVA, cmH2O䡠ml 39.8⫾6.6 59.0⫾10.1 32.4⫾3.3 47.1⫾4.9*
time-varying left ventricular volume and estimated wall stress
signals during early volume overload with high spatial and Changes in proposed hypertrophic signals in the present study (means ⫾
temporal resolution. On the basis of these data, we identified SE). ED, end diastole; ES, end systole; ␴c, mean circumferential wall stress; ␴l,
mean longitudinal wall stress; εc, mean circumferential strain; εl, mean longi-
several candidate mechanical stimuli for hypertrophy, includ- tudinal strain; Emax, maximum slope of the end-systolic pressure-volume
ing some that have not been previously recognized or pro- relationship; ⫹dP/dt, maximum rate of pressure rise. *Statistically significant
posed. Below is a discussion of these candidate stimuli and difference between sham and fistula groups.

J Appl Physiol • VOL 97 • OCTOBER 2004 • www.jap.org

Downloaded from journals.physiology.org/journal/jappl (177.012.197.016) on November 8, 2023.
1458 CANDIDATE MECHANICAL STIMULI IN VOLUME OVERLOAD

Fig. 4. Average interpolated volume and

mean circumferential stress curves for the 4
groups of animals studied (means ⫾ SE). F,
Fistula groups; E, sham groups. A and C:
4-day data. B and D: 7-day data. A and B:
amplitude and maximum rate of fall of the
volume signal were different between the
sham and fistula groups at 4 days; by 7 days,
all tested features of the volume signal were
different except its minimum. C and D: by
contrast, no features of the mean stress signal
were significantly different at either time.

end-diastolic quantities, such as ejection fraction, fractional pressure overload (Table 1) and therefore might merit explo-
shortening, or end-systolic strain may minimize apparent ration as stimuli for features of hypertrophy common to these
changes during volume overload. two models, such as net protein accumulation.
One additional strategy suggested above was to search for Comparison to previous reports. Consistent with past re-
signals that explain observed responses in multiple different ports, we found that cardiac output is increased entirely be-
experimental models, for example during both pressure and cause of an increase in stroke volume with unchanged or
volume overload. Among the signals considered here, volume slightly depressed heart rate during the first week after creation
amplitude, ⫺dV/dt, and systolic fiber strain all increase in early of an arteriovenous fistula in the rat (9, 13, 15, 28). We
volume overload and decrease in early pressure overload (Ta- observed smaller elevations in end-diastolic pressure (EDP)
ble 1); these are intriguing candidates for explaining differen- and less depression of end-systolic pressure and contractility
tial aspects of hypertrophy in the two models, such as relative than reported by some other investigators using rat arterio-
cell lengthening in volume overload vs. thickening in pressure venous fistula models; this is largely a function of fistula size.
overload. By contrast, end-diastolic wall stress, mean volume, A review of reports of hemodynamics during the first week
and maximum volume increase in both early volume and early after arteriovenous fistula in rats suggests that shunts larger
than 75% of control cardiac output are associated with acute
Table 4. Volume signal features during depression of systolic pressure generation, whereas smaller
early volume overload shunts are not (4, 9, 13, 15, 28). In most cases, larger fistulae
are also more likely to cause an elevation of EDP; however, at
Parameter 4-day Sham 4-day Fistula 7-day Sham 7-day Fistula least two studies have reported a significant increase in EDP
min V, ␮l 251⫾60 226⫾50 121⫾30 170⫾35 with a shunt size similar to ours (9, 28)
max V, ␮l 479⫾76 593⫾96 352⫾42 556⫾54* The cardiac output and volume values reported here com-
amp V, ␮l 229⫾19 367⫾54* 231⫾18 386⫾24* pare favorably to prior studies using different measurement
mean V, ␮l 359⫾68 411⫾72 230⫾36 366⫾48*
⫺dV/dt, ml/s ⫺6.1⫾0.6 ⫺8.3⫾0.8* ⫺6.7⫾0.3 ⫺9.2⫾0.4*
methods. Dividing our measurements of cardiac output by
⫹dV/dt, ml/s 7.5⫾0.7 10.0⫾1.5 7.2⫾0.6 11.3⫾1.4* body weight produced cardiac index values of 237 ⫾ 31 and
256 ⫾ 22 ml䡠min⫺1 䡠kg⫺1 for the 4-day and 7-day sham
Changes in features of the time-varying volume signal in the present study
(means ⫾ SE). Min V, minimum volume; max V, maximum volume; amp V,
groups, respectively, which fall between Flaim et al.’s sham
amplitude of volume variation; mean V, mean volume. *Statistically signifi- group average of ⬃200 ml䡠min⫺1 䡠kg⫺1 in 500-g rats (9) and
cant difference between sham and fistula groups. Huang et al.’s sham group average of 300 ml䡠min⫺1 䡠kg⫺1 in
J Appl Physiol • VOL 97 • OCTOBER 2004 • www.jap.org
Downloaded from journals.physiology.org/journal/jappl (177.012.197.016) on November 8, 2023.
 CANDIDATE MECHANICAL STIMULI IN VOLUME OVERLOAD
300-g rats (13), both obtained by use of radioactive micro-
spheres. Liu et al. (15) obtained a similar average of 306 ⫾ 28
in 300-g rats using thermodilution. We found end-diastolic
volumes of 445 ⫾ 75 and 319 ⫾ 45 ␮l, respectively, in our
4-day and 7-day sham groups, similar to Brower et al.’s (4)
mean balloon volume of 327 ␮l at a pressure of 12.5 mmHg in
isolated left ventricles from control rats weighing between 400
and 484 g.
Heart and body weight changes during early volume over-
load. Some studies have reported significant increases in ven-
tricular-to-body weight ratios as early as 3 days after creation
of volume overload (5). However, the significant early drop in
body weight we observed on days 2 and 3 suggests that these
increased ratios may reflect the postoperative drop in body
weight rather than a true ventricular hypertrophy. We observed
very similar body weight trends in the AVF and sham groups
in this study. We tested the possibility that the drop we
observed was due to the use of a relatively long-acting anes-
thetic (ketamine/xylazine) for the initial procedure by anesthe-
tizing a subset of five animals and allowing them to recover,
then tracking their weights for the next 7 days. Body weights
in these animals remained constant on day 1, then resumed a
steady increase. It therefore seems likely that the primary cause
of the weight loss observed in both the AVF and sham groups
was the abdominal surgery. It is clearly important to track
actual ventricular and body weights as well as their ratios when
studying hypertrophy in the first days after surgery. Molecular
studies may also need to be interpreted cautiously during the
early postoperative period because acute weight loss and post-
operative stress may produce deceptive results.
Limitations and sources of error. The primary limitation of
this study is that hemodynamics were measured in anesthe-
tized, open-chest animals. It would be preferable to chronically
instrument animals and track the early changes in hemodynam-
ics after creation of volume overload. However, given the
concerns raised by the present study about postoperative stress
confounding the study of early overload, subjecting animals to
chronic implantation of instrumentation seemed as likely to
compromise as to improve data quality. Good agreement with
cardiac output and pressure measurements made by other
investigators in both conscious and anesthetized closed-chest
rats (Comparison to previous reports) suggests the open-chest
anesthetized state did not dramatically alter the conclusions of
this study. The largest source of error in this study was the
ability to accurately place sonomicrometers at selected land-
marks on the beating in situ left ventricle. Inaccurate placement
of the sonomicrometers would produce a random multiplica-
tive error in calculated ventricular volumes. Again, good agree-
ment with volume and cardiac output measurements by other
investigators using different measurement methods (Compari-
son to previous reports) suggests that these errors were rela-
tively small.
The formulas utilized for stress calculation are based on a
simplified model that includes assumptions known to be in-
valid for the myocardium; comparisons to more complex
models have suggested that these equations provide a reason-
able estimate of mean wall stress but not of transmural stress
gradients (29). Without more detailed information on the
deformation, geometry, and material properties of the individ-
ual hearts studied, it would be difficult to construct an accurate
enough finite element model to improve the stress estimates
J Appl Physiol • VOL
Downloaded from journals.physiology.org/journal/jappl (177.012.197.016) on November 8, 2023.
1459
substantially. Another potential source of error was drift or
offset in the Millar miniature pressure transducer. The calibra-
tion slope of the transducer was remarkably stable across all
studies. However, the offset varied considerably from study to
study. This offset was removed by using the assumption that
minimum left ventricular pressure in an open-chest rat should
be approximately zero. The calculated offsets varied consider-
ably but were not significantly different from zero in any
group, confirming that noise rather than signal was removed
with this correction.
Summary. This study characterized standard hemodynamic
measures as well as the time-varying volume and stress signals
present at the onset of hypertrophy in a rat volume overload
model. A number of features of the time-varying volume signal
(maximum, mean, amplitude, rates of rise and fall) were
significantly altered during early volume overload in this
model, whereas many previously proposed hypertrophic stim-
uli were not. Treating hemodynamic variables more generally
as time-varying signals allowed us to identify a wider range of
candidate mechanical stimuli for hypertrophy (including some
not previously proposed in the literature) than focusing on
standard time points in the cardiac cycle. We conclude that
several features of the time-varying ventricular volume signal
and related local deformations may drive hypertrophy during
volume overload and propose that those features of the volume
signal that also change during pressure overload might be the
most interesting candidates for further exploration.
ACKNOWLEDGMENTS
The author thanks Dr. Gregory Brower at Auburn University for help in
learning the surgical volume overload technique and Bozena Malyszko for
help with data analysis for this study.
GRANTS
This work was funded by a Biomedical Engineering Research Grant
(RG-00-0062) from the Whitaker Foundation.
REFERENCES
1. Anversa P, Olivetti G, Melissari M, and Loud AV. Stereological
measurement of cellular and subcellular hypertrophy and hyperplasia in
the papillary muscle of adult rat. J Mol Cell Cardiol 12: 781–795, 1980.
2. Anversa P, Ricci R, and Olivetti G. Quantitative structural analysis of
the myocardium during physiologic growth and induced cardiac hypertro-
phy: a review. J Am Coll Cardiol 7: 1140 –1149, 1986.
3. Arts T, Prinzen FW, Snoeckx LHEH, Rijcken JM, and Reneman RS.
Adaptation of cardiac structure by mechanical feedback in the environ-
ment of the cell: a model study. Biophys J 66: 953–961, 1994.
4. Brower GL, Henegar JR, and Janicki JS. Temporal evaluation of left
ventricular remodeling and function in rats with chronic volume overload.
Am J Physiol Heart Circ Physiol 271: H2071–H2078, 1996.
5. Calderone A, Takahashi N, Izzo NJ Jr, Thaik CM, and Colucci WS.
Pressure- and volume-induced left ventricular hypertrophies are associated
with distinct myocyte phenotypes and differential induction of peptide
growth factor mRNAs. Circulation 92: 2385–2390, 1995.
6. Emery JL and Omens JH. Mechanical regulation of myocardial growth
during volume-overload hypertrophy in the rat. Am J Physiol Heart Circ
Physiol 273: H1198 –H1204, 1997.
7. Esposito G, Rapacciuolo A, Prasad SVN, Takaoka H, Thomas SA,
Koch WJ, and Rockman HA. Genetic alterations that inhibit in vivo
pressure-overload hypertrophy prevent cardiac dysfunction despite in-
creased wall stress. Circulation 105: 85–92, 2002.
8. Fischer TA, McNeil PL, Khakee R, Finn P, Kelly RA, Pfeffer MA, and
Pfeffer JM. Cardiac myocyte membrane wounding in the abruptly pres-
sure-overloaded rat heart under high wall stress. Hypertension 30: 1041–
1046, 1997.
97 • OCTOBER 2004 • www.jap.org
1460 CANDIDATE MECHANICAL STIMULI IN VOLUME OVERLOAD
9. Flaim SF, Minteer WJ, and Zelis R. Acute effects of arterio-venous
shunt on cardiovascular hemodynamics in rat. Pflügers Arch 385: 203–
209, 1980.
10. Garcia R and Diebold S. Simple, rapid, and effective method of produc-
ing aortocaval shunts in the rat. Cardiovasc Res 24: 430 – 432, 1990.
11. Gerdes AM, Campbell SE, and Hilbelink DR. Structural remodeling of
cardiac myocytes in rats with arteriovenous fistulas. Lab Invest 59:
857– 861, 1988.
12. Grossman W, Jones D, and McLaurin LP. Wall stress and patterns of
hypertrophy in the human left ventricle. J Clin Invest 56: 56 – 64, 1975.
13. Huang M, Hester RL, and Guyton AC. Hemodynamic changes in rats
after opening an arteriovenous fistula. Am J Physiol Heart Circ Physiol
262: H846 –H851, 1992.
14. Institute of Laboratory Animal Resources. Guide for the Care and Use
of Laboratory Animals. Washington, DC: National Academy Press, 1996.
15. Liu Z, Hilbelink DR, Crockett WB, and Gerdes AM. Regional changes
in hemodynamics and cardiac myocyte size in rats with aortocaval fistulas.
1. Developing and established hypertrophy. Circ Res 69: 52–58, 1991.
16. Liu Z, Hilbelink DR, and Gerdes AM. Regional changes in hemody-
namics and cardiac myocyte size in rats with aortocaval fistulas. 2.
Long-term effects. Circ Res 69: 59 – 65, 1991.
17. Manning BS, Shotwell K, Mao L, Rockman HA, and Koch WJ.
Physiological induction of a ␤-adrenergic receptor kinase inhibitor trans-
gene preserves ␤-adrenergic responsiveness in pressure-overload cardiac
hypertrophy. Circulation 102: 2751–2757, 2000.
18. Marino TA, Kent RL, Uboh CE, Fernandez E, Thompson EW, and
Cooper GT. Structural analysis of pressure versus volume overload
hypertrophy of cat right ventricle. Am J Physiol Heart Circ Physiol 249:
H371–H379, 1985.
19. Moalic JM, Bercovici J, and Swynghedauw B. Protein synthesis during
systolic and diastolic cardiac overloading in rats: a comparative study.
Cardiovasc Res 15: 515–521, 1981.
J Appl Physiol • VOL 97 • OCTOBER 2004 •
Downloaded from journals.physiology.org/journal/jappl (177.012.197.016) on November 8, 2023.
20. Nakamura A, Rokosh DG, Paccanaro M, Yee RR, Simpson PC,
Grossman W, and Foster E. LV systolic performance improves with
development of hypertrophy after transverse aortic constriction in mice.
Am J Physiol Heart Circ Physiol 281: H1104 –H1112, 2001.
21. Sadoshima J and Izumo S. The cellular and molecular response of
cardiac myocytes to mechanical stress. Annu Rev Physiol 59: 551–571,
1997.
22. Simpson DG, Majeski M, Borg TK, and Terracio L. Regulation of
cardiac myocyte protein turnover and myofibrillar structure in vitro by
specific directions of stretch. Circ Res 85: e59 – e69, 1999.
23. Streeter DD and Hanna WT. Engineering mechanics for successive
states in canine left ventricular myocardium: I. Cavity and wall geometry.
Circ Res 33: 639 – 664, 1973.
24. Suga H. Total mechanical energy of a ventricle model and cardiac oxygen
consumption. Am J Physiol Heart Circ Physiol 236: H498 –H505, 1979.
25. Takaoka H, Esposito G, Mao L, Suga H, and Rockman HA. Heart
size-independent analysis of myocardial function in murine pressure
overload hypertrophy. Am J Physiol Heart Circ Physiol 282: H2190 –
H2197, 2002.
26. Tanaka N, Dalton N, Mao L, Rockman HA, Peterson KL, Gottshall
KR, Hunter JJ, Chien KR, and Ross J Jr. Transthoracic echocardiog-
raphy in models of cardiac disease in the mouse. Circulation 94: 1109 –
1117, 1996.
27. Vendelin M, Bovendeerd PH, Engelbrecht J, and Arts T. Optimizing
ventricular fibers: uniform strain or stress, but not ATP consumption, leads
to high efficiency. Am J Physiol Heart Circ Physiol 283: H1072–H1081,
2002.
28. Wang X, Ren B, Liu S, Sentex E, Tappia PS, and Dhalla NS.
Characterization of cardiac hypertrophy and heart failure due to volume
overload in the rat. J Appl Physiol 94: 752–763, 2003.
29. Yin FCP. Ventricular wall stress. Circ Res 49: 829 – 842, 1981.
www.jap.org