Effects of early hemodynamic resuscitation on left ventricular performance and microcirculatory function during endotoxic shock

The study was conducted at the Experimental Laboratory of the Department of Pathophysiology,
School of Medicine, Universidad de la República. The experimental study was approved
by the Honorary Committee for Animal Research, School of Medicine (CHEA # 071140-000310-07;
www.urbe.fmed.edu.uy/). The Institutional Committee is under the regulations of the National Committee
for Animal Research (CNEA; http://www.cnea.org.uy/).

Animal preparation

Twenty-five female Duroc-Pampa pigs, with mean weight of 24.3?±?4.1 kg, were studied.
All the animals received ketamine (5 mg/kg, i.m.) followed by thiopental (10 mg/kg,
i.v.), and fentanyl (0.01 mg/kg, i.v.). Anesthesia was maintained by continuous thiopental
(5 mg/kg/h) and fentanyl (0.02 mg/kg/h) infusions. After systemic and local anesthesia
(lidocaine 1%), the animals were tracheostomized and mechanically ventilated (Amadeus
Hamilton Medical AG, Rhäzüns, Switzerland). Volume-controlled ventilation was provided
under neuromuscular blockade with atracurium (0.6 mg/kg/h, i.v.) continuous infusion.
The level of anesthesia and neuromuscular blockade was controlled by recording arterial
pressure, heart rate, and respiratory efforts in response to nociceptive stimuli.
Additional anesthesia or analgesia was administered when necessary, according to monitoring.
Tidal volume and positive end expiratory pressure (PEEP) were set at 8 mL/kg and 5 cm
H₂O respectively. The respiratory rate was adjusted to maintain an arterial carbon dioxide
partial pressure (PaCO₂) between 40 and 45 mm Hg. Body temperature was kept stable between 36.5°C and 37.5°C
using a heating pad. After anesthesia, the right femoral artery and vein were catheterized.
Systemic arterial pressure was measured continuously with a solid state catheter (Millar
model SPC-370, 7F, 120 cm, Millar Instruments Inc., Houston, TX, USA). A pulmonary
artery balloon catheter was placed to measure central venous pressure (CVP), pulmonary
artery pressure (PAP), and pulmonary arterial occlusion pressure (PAOP). SvO₂ was also obtained from this catheter. CO was measured by the thermodilution technique
(Oximetrix 3 Computer, Abbott, Chicago, IL, USA), and CI was calculated as CO/body
weight. Arterial lactate, SvO₂, and arterial blood gases were measured with a blood gas analyzer (ABL 700 Series,
Radiometer, Copenhagen, Denmark). All the blood samples were taken simultaneously
and by duplicate to minimize preanalytical errors. The mean value of these samples
was taken for each variable.

The heart was exposed via an anterior sternotomy. An inferior cava vein occlusion
pneumatic device was placed for controlled modifications of preload. A 7-Fr, 12-electrode,
dual-field, combined pressure-volume (P-V) conductance pigtail catheter (Mikro-tip
catheter, model SPC-551, Millar Instruments Inc., Houston, TX, USA) was inserted through
the right carotid artery and advanced to the LV apex for the assessment of LV P-V
relationships. The correct position of the conductance catheter was confirmed by monitoring
individual segmental P-V loops. After completing the experiments, the animals were
euthanized under deep anesthesia with an overdose of potassium chloride.

Experimental protocol

After surgery, a 30-min stabilization period was allowed before baseline measurements.
Then simultaneous recordings of systemic hemodynamics, sublingual microcirculatory
video microscopy, and metabolic parameters were obtained in all the animals (T0).
A representative example of hemodynamic monitoring is presented in the Additional
file 1: Figure 1. After baseline measurements, endotoxic shock was induced by 0.025 mg/kg
i/v Escherichia coli lipopolysaccharide endotoxin (LPS, serotype 0111:B4, Sigma-Aldrich. St. Louis, MO,
USA), given during a 1-h infusion via the femoral vein. Continuous saline solution
administration (8 mL/kg/h) was used to prevent hypovolemia secondary to the previous
fasting period, the surgical preparation, and baseline metabolic loses. All measurements
were repeated after 180 min (T180), once the hemodynamic resuscitation protocol was
completed. The animals were randomly assigned to one of three groups:

LPS group (LPS). These animals received basal saline perfusion but were not treated
with any further hemodynamic support after LPS injection.

Early resuscitation protocol group (ERP). The ERP group was treated from the very
beginning as follows: a) volume expansion with gelatin solution (Haemacell, 250 mL/h;
Sanofi-Aventis, Paris, France) from baseline (T0) and during the first 120 min; b)
noradrenaline 0.1 mcg/kg/min started 60 min after LPS; c) dobutamine 10 mcg/kg/min
started 120 min after LPS. The vasoactive and inotropic drugs doses were defined from
previous pilot experiments. The selected doses were effective to maintain a mean arterial
pressure (MAP) between 50 and 60 mm Hg and to increase CI when compared to baseline
values.

Sham group (SHAM). These animals only received basal saline solution.

Ventricular function data acquisition and analysis

All the ventricular, arterial, and pulmonary signals were simultaneously monitored
in real time and recorded by a multichannel digital acquisition system (200 Hz, SAMAY
MD16, Montevideo, Uruguay). Individual cardiac cycles were identified using the first
minimum that preceded the peak of the first derivative of LV pressure calculated digitally.
Data were recorded at steady-state conditions and during vena cava occlusion. The
P-V catheter was connected to a signal conditioner (Leycom Sigma 5DF, Zoetermeer,
Netherlands) in order to record LV volume. LV pressure was measured simultaneously
(control unit model TC-510, Millar Instruments Inc., Houston, TX, USA). To improve
volume estimation, we used a dual-field excitation method which generates a more homogeneous
intracavity electric field 21]. Measured time-dependent LV volume [(V(t)] was determined using the following equation:

where G(t) is the instantaneous sum of the conductances from each intraventricular electrode
pair, ? is a dimensionless constant, L equals distance between sensing electrodes, ? equals the resistivity of blood which is inversely related to conductivity, and Vp is the correcting volume for the conductance of the surrounding tissues (parallel
conductance). To determine Vp, 10 mL of hypertonic saline (5% NaCl) were injected as a bolus into the right atrium
through the left jugular vein, causing a transient increase in measured volume, G(t), without significantly altering cavity volume. To correct for underestimation of
true volume, the results measured by the conductance catheter were compared to the
results of an independent measurement of CO made by thermodilution. ? was measured at regular intervals using the CD Leycom resistivity meter.

Global LV function was estimated by LV ejection fraction (LVEF). LV preload was assessed
by measuring LV end-diastolic volume (LVEDV). Stroke work (SW) was estimated as the
integral of the P-V loop. Preload recruitable stroke work (PRSW) was used to estimate
LV myocardial contractility 22]. Eight to ten successive beats were selected from the P-V loops measured during vena
cava occlusion. A simple linear regression equation was fit to the relationship of
SW and end-diastolic volume obtained during vena cava occlusion. This slope describes
the PRSW. Among other measurements of LV contractility, this index has been shown
to be the most reproducible and relatively independent of the effects of preload and
afterload 22],23].

Microcirculation analysis

The microcirculatory network was evaluated in the sublingual mucosa of the animals
with the sidestream dark field (SDF) imaging device with a 5× objective (Microscan®,
MicroVision Medical, Amsterdam, Netherlands). Different cautions were followed to
obtain adequate quality images and to ensure good reproducibility. After saliva removal,
steady video images of at least 10 s were obtained while avoiding pressure artifacts.
The video clips were stored as AVI files to allow computerized frame-by-frame image
analysis. SDF images were acquired from at least three different sublingual places.
For each measurement, the average of three videos was taken. Adequate focus and contrast
adjustment were verified. The entire sequence was used to calculate the characteristics
of microvascular blood flow, particularly the presence of stopped or intermittent
flow. Video clips were blindly and randomly processed by the same investigator using
a semi-quantitative analysis 24]-26]. Each image was divided into four equal quadrants, and small vessels (20 ?m diameter)
were identified. Microvascular density (MVD) was determined as the number of vessels
per mm². Flow quantification was performed using a score that distinguishes between no flow
(0), intermittent flow (1), sluggish flow (2), and continuous flow (3), assigning
a value to each individual vessel. The overall score, i.e., the microvascular flow
index (MFI), corresponds to the average of all the individual vessels. The final MFI
was obtained from the average of 12 quadrants (three videos, four quadrants each).
To determine perfusion heterogeneity in each territory, the heterogeneity flow index
(HFI) was calculated as the highest MFI minus the lowest MFI divided by the mean MFI.
The proportion of perfused vessels (PPV) was calculated as ((Total number of vessels
? Number of vessels with flow?=?0 or 1)/Total number of vessels)?×?100. Finally, we
calculated the perfused vessel density (PVD) as MVD × PPV.

Statistical analysis

The normal distribution of all the physiological variables was controlled by the Shapiro-Wilk
test. Accordingly, data are shown as mean?±?SD. Paired t-test was used to analyze changes over time in each group (T0 vs T180). One-way ANOVA
with Bonferroni correction was used to analyze the differences between the three groups
at each time point (SPSS Statistical Package, version 17.0). We used a piecewise linear
function with two segments to study the correlation between MFI and MAP. The coefficients
of the bilinear curves were computed by a least-square nonlinear fitting routine using
the Levenberg-Marquardt procedure (OriginLab, version 9.1.0). Differences were considered
statistically significant when p??.05.