Group B streptococcal sepsis in the piglet: effects of fluid therapy on venous return, organ edema, and organ blood flow.

Abstract

We investigated the physiologic effects of normal saline versus 5% albuminated saline fluid resuscitation on 10-12-day-old piglets infected with group B streptococci for four hours. After intravenously receiving 1 X 10(10) bacteria/kg over 45 minutes, one group was untreated while the two fluid-treated groups received enough intravenous fluid to maintain the baseline cardiac output. An increase in the resistance to venous blood return was the major limitation to cardiac output. The resistance nearly quadrupled in the untreated piglets as shown by a 50% decrease in cardiac output with a nearly doubling of the driving pressure for venous return (mean circulatory pressure was normal and atrial pressures decreased by 70%). In both fluid-treated groups, resistance doubled as shown by an unchanged cardiac output with a doubling of the driving pressure (mean circulatory pressure increased by 50%) and atrial pressures remained at baseline). Blood volume was 9% below control in the untreated group and 13% above control in both fluid-treated groups. Much more crystalloid (155 ml/kg) than colloid (58 ml/kg) was necessary to maintain baseline cardiac output; this resulted in a 36% decrease in the plasma protein oncotic pressure of the former group and a 15% increase in the oncotic pressure of the latter group. Organ edema formation (ileum, pancreas, kidney, adrenal gland, lung) occurred only in the saline-treated animals. We conclude that increased resistance to venous return was the primary cause of shock in our model and that this can be effectively treated by giving enough intravenous fluid to elevate the mean circulatory pressure. However, if the plasma protein oncotic pressure is also lowered (saline group), organ edema results.

Full text

Page 1

MA Bressack, NS Morton and J Hortop
return, organ edema, and organ blood flow
Group B streptococcal sepsis in the piglet: effects of fluid therapy on venous
1524-4571
Copyright © 1987 American Heart Association. All rights reserved. Print ISSN: 0009-7330. Online ISSN:
TX 72514
Circulation Research is published by the American Heart Association. 7272 Greenville Avenue, Dallas,
1987, 61:659-669
Circulation Research
http://circres.ahajournals.org/content/61/5/659
located on the World Wide Web at:
The online version of this article, along with updated information and services, is
http://www.lww.com/reprints
Reprints: Information about reprints can be found online at

journalpermissions@lww.com
410-528-8550. E-mail:
Kluwer Health, 351 West Camden Street, Baltimore, MD 21202-2436. Phone: 410-528-4050. Fax:
Permissions: Permissions & Rights Desk, Lippincott Williams & Wilkins, a division of Wolters

http://circres.ahajournals.org//subscriptions/
Subscriptions: Information about subscribing to Circulation Research is online at
by guest on July 11, 2011http://circres.ahajournals.org/Downloaded from

Page 2

659
Group B Streptococcal Sepsis in the Piglet:
Effects of Fluid Therapy on Venous Return,
Organ Edema, and Organ Blood Flow
Michael A. Bressack, Neil S. Morton, and John Hortop
We investigated the physiologic effects of normal saline versus 5% albuminated saline fluid
resuscitation on 10-12-day-old piglets infected with group B streptococci for four hours. After
intravenously receiving 1 x 10" bacteria/kg over 45 minutes, one group was untreated while the two
fluid-treated groups received enough intravenous fluid to maintain the baseline cardiac output. An
increase in the resistance to venous blood return was the major limitation to cardiac output. The
resistance nearly quadrupled in the untreated piglets as shown by a 50% decrease in cardiac output
with a nearly doubling of the driving pressure for venous return (mean circulatory pressure was normal
and atrial pressures decreased by 70%). In both fluid-treated groups, resistance doubled as shown
by an unchanged cardiac output with a doubling of the driving pressure (mean circulatory pressure
increased by 50%) and atrial pressures remained at baseline). Blood volume was 9% below control
in the untreated group and 13% above control in both fluid-treated groups. Much more crystalloid
(155 ml/kg) than colloid (58 ml/kg) was necessary to maintain baseline cardiac output; this resulted
in a 36% decrease in the plasma protein oncotic pressure of the former group and a 15% increase
in the oncotic pressure of the latter group. Organ edema formation (ileum, pancreas, kidney, adrenal
gland, lung) occurred only in the saline-treated animals. We conclude that increased resistance to
venous return was the primary cause of shock in our model and that this can be effectively treated
by giving enough intravenous fluid to elevate the mean circulatory pressure. However, if the plasma
protein oncotic pressure is also lowered (saline group), organ edema results. (Circulation Research
1987;61:659-669)
G
of 1-4/1,000 live births with a case-fatality ratio of
25-50%. Because the disease can rapidly progress with
eventual death in spite of early treatment with appro-
priate antibiotics,' other forms of therapy have been
investigated. For example, Rojas et al,2 Runkle et al,3
and Meadow et al4 have evaluated indomethacin and
tolazoline in various animal models of GBS infection.
Since fluid resuscitation is basic to the treatment of
septic shock, we wished to see if crystalloid versus
colloid fluid therapy might differentially affect cardio-
vascular function or organ edema. To assess these two
processes, we treated 10-12-day-old piglets, infected
with GBS for 4 hours, with enough normal saline or
5% albuminated saline to maintain baseline cardiac
output. As opposed to most clinicians, who focus on
cardiac factors as the only determinants in shock,5 we
found that the major factor limiting cardiac output in
all infected animals was an increased resistance to
venous blood return. Cardiac output, which fell mark-
edly in the untreated piglets, could be brought back to
normal and organ blood flow improved with either type
of fluid therapy if enough fluid was given to elevate the
blood volume and the upstream driving pressure for
venous return (mean circulatory pressure).
roup B /3-hemolytic streptococcus (GBS) is
the most common infectious cause of neonatal
mortality.' There is an incidence in neonates
From the McGill University, Montreal Children's Hospital
Research Institute and Department of Pediatrics, McGill University,
Montreal, Quebec, Canada.
Address for reprints: Michael A. Bressack, Department of
Pediatrics, Stanford University, Sanford, CA 94305.
There has long been a controversy over the differ-
ential effects of crystalloid versus colloid fluids on
organ edema, especially pulmonary edema, in septic
shock.6 We found organ edema (ileum, pancreas,
kidney, adrenal gland, lung) only in the crystalloid-
treated animals, possibly due to a significant decrease
of plasma protein oncotic pressure in association with
the elevation of mean circulatory pressure.
Materials and Methods
10-12-day-old
(3.24 ± 0.66 kg) were evenly d ivided between 4 groups
(n = 10); age and weight were comparable in all groups:
group I (control): surgically prepared, uninfected, and
untreated; group II: surgically prepared, infected but
untreated; group HI: surgically prepared, infected, and
treated with normal saline (0.9% sodium chloride); and
group IV: surgically prepared, infected, and treated
with 5% human albuminated saline.
Forty male Landrace piglets
Surgical Preparation
After being pretreated with ketamine hydrochloride
(4 mg/kg i.m.), the piglets were weighed, intubated,
then anesthetized with 0.5-0.75% halothane and 50%
nitrous oxide. We ventilated the animals with a
piston-type respirator (model 101, New England Med-
ical Instruments) at an FiO2 of 35%, peak inspiratory
pressure of 14-16 cm H2O, and a rate necessary to keep
the Pco2 35-45 torr and the pH 7.35-7.45. During the
experiments, we did not make further respirator
changes.
by guest on July 11, 2011http://circres.ahajournals.org/ Downloaded from

Page 3

660
Circulation Research Vol 61, No 5, November 1987
Through cutdowns, we placed. 3F polyethylene
catheters (Clay-Adams, Inc.) in the midabdominal
aorta, the right atrium, and the inferior vena cava. Via
a left thoracotomy, we inserted a 3F polyethylene
catheter into the left atrium and a 3F double lumen
thermodilution probe (American Edwards Laborato-
ries) 3 cm into the pulmonary artery. Postoperatively,
the pericardium was left open and the ribs loosely
reapposed. Halothane and nitrous oxide were discon-
tinued, and the animals were allowed to recover for the
next hour. During the recovery and experimental
periods, the piglets were kept supine at a core tem-
perature of 38-39° C (heating pad and lamp) and
maintained on intravenous fentanyl (4 fig/kg) and
pancuronium (0.1 mg/kg) every 45-60 minutes. To
maintain animal comfort, the fentanyl dosage and
interval were selected by its ability to maintain blood
pressure and heart rate during the recovery and baseline
periods at a similar level to that during halothane
anesthesia.
Preparation of Bacteria
Group B /3-hemolytic streptococcus was isolated
from an infected human neonate (Royal Victoria
Hospital, Montreal). So that each experiment would
involve comparable bacteria, we plated the organism on
many blood agar plates, then stored the bacteria in
multiple glycerol vials at — 70° C. For each experi-
ment, a separate vial was thawed, inoculated into 50 ml
of Todd-Hewitt broth (BBL Microbiology Systems),
and were incubated at 37° C for 6 hours. The infected
broth was poured into 500 ml of Todd-Hewitt broth (pH
7.8-8.0) and agitated at 37° C for 16 hours so that the
bacteria could reach their log phase of growth. We
washed the bacteria twice by centrifuging the bacteria
at 2,000 rpm (4° C), pouring off the supernatant, and
adding sterile normal saline. We quantified the final
bacterial suspension by serial dilution plates and
confirmed bacterial purity by blood agar plate inocu-
lation and identification.
Experimental Protocol
After 1 hour of baseline measurements, we followed
all 4 groups for 4 experimental hours. Group I was not
infected. Animals in groups II, ID, and IV received the
bacterial suspension intravenously over 45 minutes (6
ml of approximately 6x10' bacteria/ml). Group II
animals were not treated. During the entire 4-hour
experimental period, groups HI and IV received enough
intravenous fluid therapy (prewarmed to body temper-
ature) to maintain the baseline cardiac output. We
infused the fluid in increments of 10 ml/kg given over
5 minutes at 10-minute intervals for a cardiac output
below baseline. Group III received normal saline
(Abbott Laboratories), and group IV received 5%
human albuminated saline. The albuminated saline was
prepared by diluting 25% human albumin (Cutter
Biological) in normal saline. Small samples of the 5%
albuminated saline (7.5 ml/kg x 2) were injected into
3 nonexperimental piglets to confirm a lack of any toxic
effects.
During the 4 hours of the experiment we continu-
ously monitored aortic (Ao), pulmonary artery (PA),
left atrial (LA), and right atrial (RA) pressures (Bentley
Trantic model #800 transducers, Beckman Dynograph
Recorder R-611). Cardiac output (CO) was measured
by computer (model COM-1, American Edwards
Laboratories) using thermodilution in triplicate every
10-15 minutes (1 ml of iced D5W). We measured
hourly aortic Po2, Pco2, and pH corrected for body
temperature (Radiometer, Copenhagen), hemoglobin
and hematocrit (Coulter Electronics), arterial lactate
(enzymatic UV-method, Boerhringer Mannheim), to-
tal plasma protein (Biuret method, Astra-8), plasma
albumin (Bromcresol green, COB AS B10), and blood
bacterial colony counts (serial dilution plates). Bac-
terial purity was confirmed with blood agar inoculation
and identification.
Derived values included cardiac index (CO/kg
weight), systemic vascular resistance (Ao-RA pres-
sure/Cl), pulmonary vascular resistance (PA—LA pres-
sure/Cl), and plasma protein oncotic pressure (7rmv)
(calculated from total protein and albumin).7
Blood Volume Measurements
Before each experiment began we mixed 10 ml of the
piglet's blood with 50 /xCi of "Cr (Merck Frosst Co.)
for 30 minutes, added vitamin C, washed the cells
twice, and reconstituted the cells in normal saline.
After intravenously injecting a known quantity of the
labelled red cell suspension into the piglet at hour 3 of
the experiment, we sampled blood every 10-15 min-
utes for 1 hour, counted the blood samples and a
reference sample (the original suspension) in a gamma
counter, measured the hematocrit on the blood and
reference samples, and calculated red blood cell
volume by extrapolating the intravascular dilution of
5lCr back to time zero. Total blood volume (V) was
derived using the animal's hematocrit and correcting
for differences between whole body and venous
hematocrit.8
Radionuclide-Labelled Microsphere Injections
Organ blood flow was determined according to the
method of Heymann et al.' We injected approximately
3 x 105 M1Ce-labelled microspheres (3M, diameter
13.5±0.5 fim) at time zero and "Sr-labelled micro-
spheres (3M, diameter 13.4±0.5 pirn) at time 3Vi
hours into the left atrium over 30 seconds. A reference
sample was withdrawn (Harvard pump) from the aortic
catheter, starting 10 seconds before the injection and
continued for 70 seconds at a rate of 1.5 ml/min. In two
animals from each group, ''Nb-labelled microspheres
were also injected into the right atrium at 3!/2 hours; this
confirmed the absence of a significant right-to-left
shunt.
Postmortem studies
After 4 hours of the experiment, we acutely stopped
the heart with a lethal left atrial injection of potas-
sium chloride. Within 7 seconds of cardiac arrest, we
measured the mean circulatory pressure (Pmc) from
by guest on July 11, 2011http://circres.ahajournals.org/Downloaded from

Page 4

Bressack et al Fluid Therapy in Group B Streptococcal Sepsis
661
the right atrial pressure plateau and corrected for any
differences between aortic and atrial pressures.10"12
Histology
The left lung hilum was clamped at an end-expiratory
pressure of 5 cm H2O and a 4-mm wide section of the
left lower lobe was placed into buffered formalin.
Four-millimeter wide sections of liver, left ventricle,
diaphragm, pancreas, adrenal gland, ileum, kidney,
cerebral cortex, and hippocampus were fixed in buff-
ered formalin. All sections were stained with HPS
(hemalum, phloxine, saffron) and examined by light
microscopy.
Organ Blood Flow
Most of the lung, liver, brain, diaphragm, kidney,
heart, pancreas, adrenal gland, and distal ileum were
cut into sections, placed into counting vials, weighed,
and counted along with the reference blood samples in
a gamma counter (Beckman Biogamma II). Counting
window widths were selected to give maximum effi-
ciency as well as good separation for each isotope.
Although a relatively narrow range was chosen for each
isotope, organ counts were corrected for isotopic
interference. Organ blood flow was calculated using the
reference sample method"3 and expressed as milliliters
per minute per gram dry, bloodless tissue to eliminate
any effects due to differences in organ blood or edema
content. Organ blood flow equals
organ radioactivity x reference sample withdrawal rate
reference sample radioactivity
The homogeneity of the microsphere distribution was
verified by comparing the results between kidneys and
cerebral hemispheres. More than 400 microspheres
were verified in each tube counted. Blood flows were
expressed in absolute terms and also in reference to the
cardiac index.
Organ Wet/Dry Weight Ratio
After counting the tissue samples in the gamma
counter, the tissue was dried to a constant weight in a
70° C oven (Fisher, model 106G). We calculated the
ratio of wet/dry organ tissue weight exclusive of blood
by the method described by Pearce et alM using the
"Cr-labelled red blood cells injected premortem and 10
ml of blood withdrawn, weighed and counted at the
time of death. Plasma showed insignificant "Cr activ-
ity. No significant differences in organ blood volume
were found between groups. There was no need to
correct tissue weights for the microspheres because the
weight of the spheres was less than 0.005% of the total
dry weight.
Organ Albumin Content
Keeping the tissues cool at all times, we weighed,
diluted with normal saline (buffered with Tris to pH 8),
then homogenized (Polytron-Brinkman homogenizer
PT 10-35, generator PTA-10 TF) 0.5-2.0-g samples of
the organs studied, excluding the adrenal cortex. We
ultracentrifuged the tissue (7,500 rpmx 15 minutes),
collected the supernatant, resuspended the pellet,
ultracentrifuged again (15,000 rpm X 20 minutes), and
collected the supernatant. A third extraction showed
insignificant albumin content. The two supernatants
were combined and analyzed for porcine albumin
content using Mancini's
immunodiffusion.13 Rabbit anti-porcine albumin anti-
body (Cooper Biomedical) was used for the diffusion
plates and porcine albumin (fraction V, 97% purity,
Sigma Chemical Co., St. Louis, Mo.) for the stan-
dards. For each organ and animal, porcine albumin
content was corrected for tissue blood albumin content
(from the "Cr assay on that organ and the hematocrit
and plasma albumin concentration) and reported as
milligrams albumin per gram bloodless wet tissue
weight.
technique of radial
Statistics
All data are reported as mean ± SD. For comparisons
of repeated measurements within groups, a repeated
measures analysis of variance was used. A one-way
analysis of variance was used for comparisons between
groups at each time period. The means were compared
post hoc by the Tukey test.16 Significance is at a
p<0.05.
Results
Three animals from group II did not survive the
experiment and are not included in the results; the 10
animals reported in group II all survived the 4-hour
experiment. All other groups had 100% survival. There
were no significant differences between groups in
the number of bacteria
(l.l±0.5x 10'°/kg) or the blood bacterial colony
counts (5.1 ±5.Ox 107ml) during the experiment.
injected intravenously
Circulatory Measurements
Mean aortic pressures (Figure 1) remained at base-
line during the entire 4 hours in all groups except for
a brief elevation in the fluid-treated groups during the
VASCULAR
PRESSURE
(torr)
50
TIME (hours)
FIGURE 1. Responses of mean aortic and pulmonary artery
pressure to infusion ofGBS. * Significant difference (p<0.05)
compared with group I at a comparable time; fsignificant
difference (p<0.05) from the baseline value within the same
group.
by guest on July 11, 2011http://circres.ahajournals.org/Downloaded from

Page 5

662
Circulation Research Vol 61, No 5, November 1987
bacterial infusion and a downward but insignificant
trend in the infected, untreated group. In groups II, III,
and IV, the mean pulmonary artery pressure (Figure 1)
greatly increased (3 times baseline) during the bacterial
infusion then plateaued at approximately twice baseline
during the rest of the experiment. Left and right atrial
pressures (Figure 2) decreased in the infected, un-
treated group but did not change significantly in the
other groups.
Cardiac index (Figure 3) was maintained at baseline
values (except briefly during the bacterial infusions) in
all except the infected, untreated animals; the latter
group maintained a low cardiac output during the entire
study with a slow fall to 50% baseline by 4 hours. Heart
rate did not significantly change in any of the groups.
Associated with the above changes, the systemic
vascular resistance (SVR) in group II slowly rose from
1 !/2 to 2 times baseline (Figure 4) as the cardiac output
fell during the study but did not significantly change in
the other groups with stable cardiac outputs except
briefly during the bacterial infusion. On the other hand,
pulmonary vascular resistance (PVR) increased to 5
times baseline (Figure 4) in group II and 2!/2 to 3 times
baseline in the fluid-treated groups; the increase in
group II was significantly larger than in groups HI
orlV.
The normal saline-treated animals required 155 ± 45
ml/kg of fluid during the 4-hour study while the 5%
albuminated saline-treated animals required 58 ±13
ml/kg (Table 1). This resulted in an elevation in blood
volume (V) of approximately 13% and in mean circu-
latory pressure of 50% in both fluid-treated groups com-
pared to the control group. On the other hand, the
infected, untreated animals had a 9% depression in blood
volume with no significant difference in mean circula-
tory pressure compared to the control group. Our values
for mean circulatory pressure in the control group are
comparable to those reported by Guyton et al."
Blood Gas and Lactate Measurements
Aortic Po2 decreased by approximately 15% in all
infected groups (Table 2). Right atria! microsphere
0.30
ATRIAL
PRESSURE
(torr)
• LEFT ATRIAL
' RIGHT ATRIAL
GROUP
•
O 2
A 3
A 4
1
TIME (hourt)
FIGURE 2. Responses of right and left atrial pressure to
infusion of GBS. *Significant difference (p<0.05) compared
with group I at a comparable time; fsignificant difference
(p<0.05) from the baseline value within the same group.
CARDIAC
INDEX
/ Ulen \
ymln*kgy
0.20
Xf'
- /
1 * *
^ * -
i 1 1 1*
GROUP
•
O
A
A
1
2
3
4
TIME (hoin)
inpUbKJN
FIGURE 3. Response of cardiac index to infusion of GBS.
^Significant difference (p<0.05) compared with group I at a
comparable time; fsignificant difference (p<0.05) from the
baseline value within the same group.
injections showed the absence of significant right to left
shunting in all groups (insignificant "Nb counts in
systemic organs). There were no significant changes in
aortic Pco2. All infected groups developed a metabolic
acidosis with a fall in pH and an increase in arterial
lactate levels. The untreated animals had significantly
more metabolic/lactic acidosis than the fluid-treated
animals.
Protein and Hematologic Measurements
Total protein and albumin concentrations decreased
significantly by 12% in group II and by 31 % in group
HI; the difference between these two groups was also
significant (Table 3). In group IV, the total protein
concentration did not change, but the albumin con-
centration increased by 31%. The ;rmv fell by 13% in
group II and 36% in group ID but increased by 15% in
group IV. The hematocrit and hemoglobin concentra-
tions increased by 7% in group II and decreased by 10%
in groups HI and IV.
VASCULAR
RESISTANCE 400
TIMEfhoura)
FIGURE 4. Responses of systemic and pulmonary vascular
resistance to infusion of GBS. *Significant difference (p<0.05)
compared with group I at a comparable time; fsignificant
difference (p<0.05) from the baseline value within the same
group.
by guest on July 11, 2011http://circres.ahajournals.org/Downloaded from