British Journal of Anaesthesia 91 (2): 224±32 (2003)

DOI: 10.1093/bja/aeg167

LABORATORY INVESTIGATIONS
Effects of sustained post-traumatic shock and initial ¯uid
resuscitation on extravascular lung water content and pulmonary
vascular pressures in a porcine model of shock
M. Nirmalan1*, T. Willard2, D. J. Edwards2², P. Dark1 and R. A. Little1
1
MRC Trauma Group University of Manchester, University of Manchester, Oxford Road, Manchester
M13 9PT, UK. 2South Manchester University Hospitals, Manchester, UK.
*Corresponding author: University Department of Anaesthesia and Intensive Care, Manchester Royal In®rmary,
Oxford Road, Manchester M13 9WL, UK. E-mail: mahesan.nirmalan@man.ac.uk

Background. The temporal evolution of lung injury following post-traumatic shock is poorly
understood. In the present study we have tested the hypothesis that manifestations of pulmon-
ary vascular dysfunction may be demonstrable within the ®rst hour after the onset of shock.
Methods. Twenty-nine anaesthetized pigs (mean weight 27.4 kg; (SD) 3.2) were randomly
allocated to three groups: control (C, n=9), shock resuscitated with either NaCl 0.9% (S,
n=10), or 4% gelatine (G, n=10). Shock was maintained for 1 h followed by ¯uid resuscitation
with either normal saline or 4% gelatine solution. Cardiac output (CO), mean arterial pressure
(MAP), mixed venous saturation (SvO2), blood lactate concentration, mean pulmonary artery
pressure (MPAP), MPAP/MAP, pulmonary vascular resistance (PVR), extravascular lung water
index (EVLWi), PaO2/FIO2, venous admixture (QÇS/QÇT), and dynamic lung compliance (Cdyn) were
measured at baseline, beginning of shock phase, end of shock phase, and post-resuscitation.
Results. At the end of volume resuscitation CO was restored to control values in both shock
groups. MAP remained signi®cantly below control values (95% CI: C=70±95, S=28±52,
G=45±69 mm Hg) in both shock groups. MPAP/MAP was signi®cantly greater in both shock
groups at the end of the shock phase (95% CI; C=0.15±0.24, S=0.28±0.38, G=0.32±0.42) and at
the post-resuscitation phase (95% CI: C=0.12±0.30, S=0.43±0.61, G=0.32±0.49) indicating the
presence of relative pulmonary hypertension. This was associated with a signi®cant increase in
PVR in Group S (F=3.9; P<0.05). There were no signi®cant changes in PaO2/FIO2, QÇS/QÇT, EVLWi,
or Cdyn. In a small cohort of animals a measurable increase in EVLWi (>30%) and reduction in
Cdyn (>10%) were observed.
Conclusions. Pulmonary vascular injury manifesting as relative pulmonary hypertension and
increased PVR may occur within the ®rst hour after the onset of shock. These changes may
not be accompanied by overt changes in oxygenation, compliance, or EVLWi.
Br J Anaesth 2003; 91: 224±32
Keywords: complications, pulmonary hypertension; complications, pulmonary vascular
resistance; complications, shock; lung, functions; lung, water
Accepted for publication: March 6, 2003

In what is now widely regarded as a landmark publication
Cournand, Riley and co-workers in 1943 ®rst demonstrated
that patients with untreated post-traumatic shock were
characterized by reductions in circulating blood volume,
cardiac output (CO), and oxygen delivery.1 Some of the
subjects in this initial study had reduced arterial oxygen
saturation (<85%) even in the absence of direct pulmonary
²
Declaration of interest: Dr D. J. Edwards acted as a Medical Advisor
to Maelor Pharmaceuticals, Ltd.

Ó The Board of Management and Trustees of the British Journal of Anaesthesia 2003
 Post-traumatic shock and lung functions
trauma. Subsequently, Brewer and colleagues in 1946
described the condition `wet lungs' among war casualties
on initial admission to forward ®eld hospitals.2 Since these
two initial reports, early changes in lung function as a result
of shock have drawn considerable attention. Holcroft and
co-workers,3 Noble,4 and Sturm and colleagues5 have
demonstrated that an increase in extravascular lung water
content (EVLW) may be demonstrated in laboratory models
of shock. Even though the underlying causes for early
increases in EVLW remain unproven intestinal trans-
location of endotoxin brought about by hepato-splanchnic
ischaemia and the consequent activation of in¯ammatory
cascades may provide a potential explanation.6 7 Similar
changes in pulmonary function have been explored in
greater detail in laboratory models of sepsis where seques-
tration of leukocytes and impaired endothelial function has
been demonstrated within the ®rst 4 h after exposure to
endotoxin.8 9
The importance of these pathophysiological processes to
the initial clinical presentation of patients following a period
of sustained post-traumatic shock is not known. Hypo-
xaemia is a common ®nding in such patients even in the
absence of direct chest trauma and consequently the routine
administration of oxygen is recommended in the Advanced
Trauma Life Support protocols.10 While ventilation/perfu-
sion mismatch, aspiration of gastric contents and ¯uid
overload during initial ¯uid resuscitation are considered
possible causes, shock mediated endothelial changes are
rarely considered important at this very early stage.
Characterizing the role of shock per se on pulmonary
function, however, is important in order to initiate appro-
priate therapies for these patients. For example, observed
increases in EVLW (manifesting as hypoxaemia associated
with pulmonary oedema or ®ne basal crepitations over the
lung ®elds) when wrongly attributed to iatrogenic ¯uid
overload frequently leads to ¯uid restriction or even the use
of diuretics with potentially serious consequences in
patients who are already volume depleted. The choice of
resuscitation ¯uids continues to draw considerable interest
and debate amongst critical care physicians in this context.
This is particularly relevant in the light of three recent meta
analyses that showed con¯icting results on the choice of
resuscitation ¯uids on clinical outcome.11±13 The current
clinical consensus, however, is that the choice of resusci-
tation ¯uids does not in¯uence the immediate or long-term
pulmonary outcome in patients with post-traumatic
shock.14 15 In the light of these uncertainties we undertook
the following study in a laboratory model of sustained post-
traumatic shock to test the following hypotheses.
Primary hypotheses
A combination of standardized skeletal injury followed by a
period (1 h) of acute hypovolaemic shock in the anaes-
thetized pig may reproduce the physiological abnormalities
consistent with those described in human victims of post-
225
traumatic shock and thereby be of relevance in the study of
pulmonary changes in shock.
That in such a model an increase in EVLW and other
features of pulmonary vascular dysfunction are demon-
strable before ¯uid resuscitation is commenced and these
changes continue to be an integral component of the overall
pathophysiological pro®le even after initial volume replace-
ment.
Secondary hypothesis
The nature of ¯uids (crystalloids or colloids) used in initial
resuscitation does not in¯uence early changes in EVLW
content or other pulmonary functions.
Methods
With University Ethics Committee approval, 29 immature
female Large-White pigs (mean weight 27.4 (SD 3.2) kg)
were randomly allocated to a control group (C, n=9), a
shock group resuscitated with NaCl 0.9% (S, n=10) and a
shock group resuscitated with succinylated gelatine solution
4% (Maelor Pharmaceuticals Ltd, Wrexham, UK) (G,
n=10). Anaesthesia was induced with halothane, oxygen,
and nitrous oxide 4±5% (FIO2 50%) administered via a
snout mask. After tracheal intubation halothane was reduced
to 1±2% and mechanical ventilation established with
a volume-cycled ventilator (Blease-Brompton-Manley,
Chesham, Bucks, UK) (tidal volume 10±15 ml kg±1; rate
12±15 bpm). At the end of surgery, inspired halothane
concentration was reduced to less than 0.25% and an i.v.
infusion of alphaxalone-alphadolone (Saffan; Pitman-
Moore, Uxbridge, UK; 15 mg kg±1 h±1) commenced.
Inspired oxygen was also reduced to between 0.25 and 0.3.
Using aseptic techniques, the right external jugular vein
was exposed and a pulmonary artery ¯oatation catheter
(Baxter Swan-Ganz CCO/VIP, 7.5F; Edwards Life
Sciences, CA, USA) was introduced. Correct positioning
of the pulmonary artery catheter was con®rmed by inspect-
ing the transduced waveforms. All animals received main-
tenance ¯uids (NaCl 0.9%; 10 ml kg±1 h±1) via the central
vein to replace insensible ¯uid loss (this infusion was
discontinued during the `shock procedure' and restarted
after resuscitation). The femoral artery was then exposed
surgically and the dilution catheter (Pulsiocath PV 2024 4F;
Pulsion Medizintechnik, Munich, Germany) was positioned
in the abdominal aorta. An indwelling suprapubic Foley
catheter was positioned for continuous drainage of urine. At
the end of the instrumentation phase (considered time 0) all
animals were given a rest period of 30 min following which
baseline measurements were made. This was followed by
the `shock procedure' in Groups S and G. A captive bolt
(Cox Universal model V/10165; Temple Cox Development,
Bromley, Kent, UK) was used to achieve bilateral tibial
fractures. The bolt was applied directly to the bone in order
to minimize soft tissue damage. Haemorrhage was then
 Nirmalan et al.
commenced at a rate of 1 ml kg±1 min±1 and continued until
three of the four predetermined end points of shock were
achieved. Shock was achieved in all animals within 30 min
and the end points for shock were as follows.
1. Greater than 30% reduction in CO.
2. Greater than 30% reduction in mean arterial pressure
(MAP).
3. Mixed venous oxygen saturation (SvO2) less than 40%.
4. Blood lactate concentration greater than 3 mmol litre±1.
Shock was maintained for 1 h (shock phase) and further
blood was removed as necessary during this phase in order
to ensure at least three `shock criteria' were maintained
throughout this period. `Start of shock phase' measurements
and `end of shock phase' measurements were made at the
beginning and end of the shock phase, respectively. At the
end of the shock phase, volume resuscitation was achieved
with either NaCl 0.9% (S group) or gelatine (G group). Fluid
administration was stopped when CO was restored and
maintained above 90% of baseline values throughout the
resuscitation phase (60 min) followed by the post-resusci-
tation phase measurements. The animals were killed by
anaesthetic overdose, chest opened, lungs removed, and
EVLW content was determined by a gravimetric method,
which corrects for intravascular volume.16
Measurement of extravascular lung water index
The dual dye dilution method (COLD Z-03; Pulsion
Medizintechnik, Munich, Germany) was used to estimate
extra vascular lung water index (EVLWi) at the following
time points.
1. Baseline: 30 min after completion of the instrumen-
tation phase.
2. Start of shock phase: 60 min after completion of the
instrumentation phase.
3. End of shock phase: 120 min after completion of the
instrumentation phase.
4. Post-resuscitation phase: 180 min after completion of
the instrumentation phase.
Triplicate estimates of EVLWi using a manual injection
of 10 ml of cold indocyanin green (Pulsion; 1 mg ml±1) were
made and the numerical average of the two closest
measurements was taken as the true EVLWi for each
phase. Dynamic lung compliance (Cdyn), PaO2/FIO2 ratio,
venous admixture (Q Ç T), mean pulmonary arterial pres-
Ç S/Q
sure (MPAP), MAP, CO, arterial blood lactate concentra-
tion were also measured at the same time points as EVLWi
in all three groups. A stopwatch was used in all experiments
in order to make the respective measurements at the same
time points in the three groups.
Measurement of dynamic lung compliance (Cdyn)
A tracheal tube with a purpose-built pressure monitoring
port was used to intubate the trachea (Mallinkrodt, Hi-Lo
Jet; size 6.5). An oesophageal balloon (Mallinkrodt, size
226
8.5) was positioned in the mid-oesophagus and the pressure
monitoring ports of both tubes were connected to a
differential pressure transducer (Pneu-01; World Precision
Instruments, Inc., USA) to obtain transpulmonary pressure
signals. A Pneumotachograph (Godart 17212; Gould
Electronics BV, The Netherlands) was used for respiratory
¯ow measurements. Transpulmonary pressure and respira-
tory ¯ow signals were acquired and stored in a personal
computer using standard equipment and software (CED
1902, CED 1401 and Spike 2; Cambridge Electronics
Design, Cambridge, UK). Tidal volume was obtained by
integrating the inspiratory ¯ow signals and Cdyn was
calculated17 for baseline, end of shock, and post-resuscita-
tion phases.
Ç S/
Measurement of pulmonary venous admixture (Q
Ç
QT) and blood lactate concentrations
Arterial and mixed venous blood samples were obtained
simultaneously during the four phases of measurement.
Mixed venous blood was sampled using a low pressure
aspiration technique to prevent contamination with left
atrial blood18 and the samples were analysed immediately in
the blood gas analyser/oximeter (ABL 330/OSM3; Radio-
meter, Copenhagen, Sweden) and lactate analysers (BEG
2300; YSI Ltd, UK).
CO and systemic/pulmonary vascular resistance
Using the Vigilance continuous CO monitor (Baxter
Healthcare Ltd, CA, USA) continuous cardiac output
(CCO) and intermittent thermodilution CO were deter-
mined. CO data used in all subsequent analyses are based on
values obtained using the intermittent technique. The CCO
measurements were used to identify the end points of shock
and adequacy of ¯uid resuscitation only. Pulmonary vascu-
lar resistance (PVR) was derived from MPAP, pulmonary
capillary wedge pressure (PCWP) and CO in the customary
fashion: PVR=[(MPAP ± PCWP)/CO]*80. As right atrial
pressure measurements were not made (because of technical
dif®culties inherent in placing an additional catheter in the
right atrium) a proxy measure of systemic vascular resist-
ance (SVR') was derived in the following way:
SVR'=MAP/CO*80.
Statistical analyses
All variables were analysed using analysis of variance
(ANOVA) for repeated measurements (general linear model;
SPSS 9.0, SPSS, Inc., Chicago, IL, USA). Change in PVR
during the course of the experiment (PVRpost-
resuscitation±PVRbaseline) was compared between the groups
using one-way ANOVA. When using the repeated measures
ANOVA, signi®cant factors were further compared using
95% CI of the estimated means at each of the four phases of
the experiment. Statistical signi®cance was de®ned as
 Post-traumatic shock and lung functions

Table 1 CO, MAP, SvO2, and blood lactate concentration data for the three groups of animals. Mean (SD) and range are used as the summary statistics.
*Indicates test group signi®cantly different from control group (P<0.05)

Baseline Start of shock phase End of shock phase Post-resuscitation phase

CO (litre min±1) C (n=9) 3.6 (1.0) 3.8 (1.1) 3.5 (0.8) 3.5 (0.8)
(2.2±5.5) (2.2±5.6) (2.4±4.9) (2.5±4.9)
S (n=10) 4.2 (0.7) 2.2 (0.6)* 2.0 (0.8)* 3.7 (1.3)
(3.0±5.7) (1.2±2.9) (0.6±2.1) (2.9±6.5)
G (n=10) 4.2 (1.0) 2.4 (0.4)* 2.2 (0.9)* 4.4 (1.5)
(2.5±5.7) (1.6±2.9) (0.7±3.3) (2.3±6.4)
MAP (mm Hg) C (n=9) 92.7 (23.2) 89.1 (17.9) 88.4 (21.5) 82.8 (20.7)
(57±136) (59±107) (54±120) (54±116)
S (n=10) 92.6 (11.6) 41.3 (13.9)* 41.7 (15.8)* 40.2 (15.2)*
(77±118) (22±62.5) (20±62) (11±61)
G (n=10) 89.2 (13.8) 43.9 (10.9)* 43.6 (14.1)* 57.2 (18.7)*
(60.0±110) (30.3±60.0) (21.0±73.0) (34.0±95.0)
SvO2 (%) C (n=9) 60.7 (6.7) 65.2 (7.9) 60.2 (5.1) 58.8 (6.6)
(50±70) (51.6±76.7) (49.9±68.3) (48.9±68.7)
S (n=10) 57.7 (8.1) 23.8 (8.3)* 25.7 (10.9)* 41.9 (10.8)*
(45.5±70.9) (13.9±37.4) (7.0±48.5) (29.4±61.9)
G (n=10) 64.7 (6.2) 30.6 (15.2)* 33.1 (8.9)* 41.3 (13.4)*
(54.0±72.0) (15.7±60.7) (18.7±52.3) (19.1±59.6)
Blood lactate concentration (mmol litre±1) C (n=9) 1.2 (0.2) 1.0 (0.2) 0.97 (0.2) 0.84 (0.1)
(0.8±1.6) (0.7±1.4) (0.7±1.4) (0.6±1.1)
S (n=10) 1.6 (0.8) 2.3 (1.1)* 3.9 (1.7)* 4.7 (2.8)*
(0.8±2.3) (1.3±4.9) (1.7±6.8) (1.4±8.0)
G (n=10) 1.2 (0.3) 2.0 (0.7) 3.0 (1.7)* 4.4 (3.8)*
(0.8±1.9) (0.9±3.2) (1.0±6.4) (0.9±10.9)

P<0.05 (two-tailed). As all data were normally distributed
mean (SD) and corresponding ranges have been used as
summary statistics.
Results
The mean (SD) weights for the three groups were compar-
able (C, 26.8 kg (3.0); S, 27.5 kg (2.5); G, 27.8 kg (3.8)).
Two animals died during the experiments, one each from `S'
and `G' groups. As both animals died during the latter stages
of the resuscitation phase all available measurements from
both animals have been included in the subsequent analyses.
Mean (SD) values for the four parameters used in the
de®nition of shock, i.e. CO, MAP, SvO2, and blood lactate
concentration are summarized in Table 1. All other
measured and derived variables, i.e. PCWP, heart rate
(HR), stroke volume (SV), left ventricular stroke work
(LVSW), MPAP, MPAP/MAP ratio, PVR, SVR', EVLWi,
PaO2/FIO2, QÇ S/Q
Ç T, and Cdyn are shown in Table 2. There
were no signi®cant differences in any of the baseline
parameters (Tables 1 and 2).
As expected from the study design there was a signi®cant
reduction in CO, MAP, and SvO2 associated with the `shock
procedure' (Table 1). As a rise in blood lactate concentra-
tion is a relatively delayed event, all four `shock criteria'
(including blood lactate concentration >3 mmol litre±1)
were achieved during the shock phase in only 12 animals.
The percentage blood volume removed from the two groups
of animals as comparable (S=39.4% (6.7); range 25±50%;
G=39.1% (11.6); range 24±60%). As expected the volume
of saline infused to achieve the resuscitation targets was
greater than the volume of colloids (803 ml (258); range
402±1275 ml compared with 699 ml (274); range 438±1300
ml). This difference, however, did not reach statistical
signi®cance. Volume replacement restored CO to control
values in both shock groups. In spite of this, MAP remained
signi®cantly below control values in both groups (95% CI:
C=70±95, S=28±52, G=45±69 mm Hg). The low MAP in
the presence of a normal CO was associated with a
signi®cant reduction in SVR' in both shock groups at the
end of the resuscitation phase (95%CI: C=1613±2264,
S=664±1315, G=853±1505 dynes s cm±5). A small but non-
signi®cant increase in MPAP was observed for the two
shock groups during the study period. However, the ratio
MPAP/MAP was signi®cantly greater in the two shock
groups at the end of resuscitation (95% CI: C=0.12±0.30;
S=0.43±0.61; G=0.32±0.49). The higher MPAP/MAP ratio
was associated with a signi®cant increase in PVR in Group
S during the study period (F=3.9; P<0.05). Even though the
mean change in PVR for Group G was greater than the
control group this difference was not statistically signi®cant
(C: mean change=±29; 95% CI=±132±77 dynes s cm±5; G:
mean change=56; 95% CI=±91±204 dynes s cm±5).
Mean SV at the post-resuscitation phase was substantially
less in Group S than in Groups C and G even though this
difference did not reach statistical signi®cance (95%
CI: C=23.5±33.9, S=16.6±27, G=23.1±33.5 ml). When
compared with baseline values, however, both shock groups
showed a signi®cant reduction (LVSWpost-resuscitation±
LVSWbaseline) in LVSW (F=19.4, P<0.001) closely follow-
ing the observed trends in MAP. Mean SvO2 recovered to
above 40% after resuscitation but remained signi®cantly

227
 Nirmalan et al.

Table 2 PCWP, HR, SV, LVSW, MPAP, MPAP/MAP, PVR, SVR', PaO2/FIO2, Q Ç S/Q
Ç T, EVLWi, and Cdyn data for the three groups of animals. Mean (SD) and
range are used as the summary statistics. *Indicates test group signi®cantly different from control group (P<0.05). **Indicates signi®cant change (post-
resuscitation value ± baseline value) during the course of the experiments using one-way ANOVA (P<0.05)

Baseline Start of shock phase End of shock phase Post-resuscitation phase

PCWP (mm Hg) C (n=9) 5.0 (2.8) 5.6 (4.1) 5.4 (3.9) 6.6 (2.4)
(0±9) (±3±11) (±3±11) (2±11)
S (n=10) 7.4 (2.8) 3.3 (3.6) 3.1 (2.9) 5.2 (2.6)
(2±12) (0±12) (0±10) (0±10)
G (n=10) 6.6 (1.4) 3.4 (2.3) 3.7 (1.9) 8.2 (4.1)
(5±10) (0±7) (0±7) (0±15)
HR (bpm) C (n=9) 155 (33) 148 (34) 132 (28) 128 (26)
114±216) (98±198) (96±180) (96±168)
S (n=10) 160 (35) 196 (44)* 203 (51)* 173 (57)
(114±240) (132±258) (126±294) (114±282)
G (n=10) 146 (29) 187 (34) 190 (45)* 152 (33)
(114±216) (144±234) (132±264) (102±210)
SV (ml) C (n=9) 23.4 (6.7) 26.3 (7.5) 27.7 (7.9) 28.7 (8.6)
(13.4±31.8) (12.2±35.9) (16.4±39.6) (18.1±45.4)
S (n=10) 26.8 (3.2) 11.3 (3.2)* 9.9 (3.9)* 21.8 (5.1)
(23.8±33.1) (6.9±19.3) (2.9±15.6) (13.6±28.4)
G (n=10) 29.5 (6.6) 13.0 (2.8)* 11.4 (4.4)* 28.2 (8.5)
(16.7±38.4) (9.9±16.9) (4.6±18.1) (12.6±41.1)
±1
LVSW (g m ) C (n=9) 28.0 (10.2) 29.4 (9.5) 30.5 (9.8) 29.1 (10.1)
(10.4±43.1) (17.3±46.4) (20.3±47.9) (19.3±52.4)
S (n=10) 31.1 (5.4) 5.9 (2.8)* 5.4 (2.9)* 11.3 (4.7)*
(25.9±40.1) (2.7±11.9) (0.9±9.9) (5.8±18.9)
G (n=10) 33.1 (9.4) 7.0 (1.9)* 6.8 (4.4)* 20.5 (9.5)
(19.4±44.9) (4.9±11.6) (1.1±15.5) (4.3±34.5)
MPAP (mm Hg) C (n=9) 15.2 (4.7) 16.2 (3.5) 16.0 (3.8) 15.9 (4.3)
(7.0±21.7) (10.3±20.0) (8.0±19.7) (7.7±21.7)
S (n=10) 18.8 (3.9) 11.5 (2.8)* 13.0 (4.9) 21.9 (5.8)
(12.0±25.0) (6.7±15.0) (6.6±20.6) (13.6±30.7)
G (n=10) 18.8 (3.3) 12.7 (3.2) 15.7 (3.9) 21.3 (4.9)
(11.6±22.3) (8.3±18.3) (8.3±20.3) (13.3±28.7)
MPAP/MAP C (n=9) 0.18 (0.06) 0.19 (0.06) 0.19 (0.07) 0.21 (0.08)
(0.10±0.25) (0.12±0.28) (0.08±0.29) (0.08±0.34)
S (n=10) 0.21 (0.05) 0.29 (0.08) 0.32 (0.08)* 0.52 (0.16)*
(0.12±0.29) (0.17±0.42) (0.18±0.46) (0.36±0.88)
G (n=10) 0.22 (0.04) 0.30 (0.1) 0.37 (0.05)* 0.41 (0.1)*
(0.12±0.3) (0.17±0.55) (0.25±0.47) (0.14±0.68)
±5
PVR (dynes s cm ) C (n=9) 249.6 (87) 258.4 (145) 260.6 (111) 221.7 (123)
(58±339) (59±518) (43±437) (43±487)
S (n=10) 222.0 (59) 313.0 (106) 445.8 (178) 424.0 (201)**
(84±308) (104±511) (189±889) (204±889)
G (n=10) 251.1 (108) 324.1 (113) 494.3 (176) 307.5 (221)
(116±490) (157±567) (307±779) (150±778)
±5
SVR' (dynes s cm ) C (n=9) 2231 (952) 1995 (652) 2078 (668) 1939 (611)
(1577±4610) (1325±3438) (1301±3525) (1285±3072)
S (n=10) 1815 (517) 1548 (395) 1816 (572) 990 (265)*
(1392±3146) (928±2173) (1187±2933) (665±1400)
G (n=10) 1774 (544) 1533 (537) 1770 (541) 1179 (479)*
(1294±2912) (887±2550) (1115±2545) (739±2062)
PaO2/FIO2 (kPa) C (n=9) 56.1 (6.1) 54.1 (5.6) 54.2 (6.8) 55.5 (7.1)
(47.5±65.2) (45.2±62.1) (42.4±65.1) (40.7±64.1)
S (n=10) 51.3 (5.5) 52.7 (8.1) 55.2 (7.7) 56.1 (7.1)
(45.2±64.5) (38.5±64.3) (42.9±66.6) (47.3±68.6)
G (n=10) 50.5 (6.0) 54.5 (5.6) 53.9 (6.1) 53.9 (6.5)
(41.1±57.7) (46.0±62.4) (46.0±62.4) (41.7±64.3)
Ç S/Q
Q Ç T (%) C (n=9) 10.1 (4.6) 10.2 (5.8) 9.7 (4.2) 8.5 (4.4)
(4.8±18.2) (5.3±23.3 (4.9±17.9) (3.8±17.7)
S (n=10) 11.3 (3.4) 6.8 (3.6) 5.5 (2.7) 7.5 (3.5)
(3.9±14.7) (0.9±12.5) (1.4±9.7) (0.2±12.5)
G (n=10) 12.8 (4.1) 6.6 (3.2) 6.9 (2.7) 8.0 (5.4)
(7.5±20.6) (2.9±13.8) (3.1±10.5) (0.9±21.4)
±1
EVLWi (ml kg ) C (n=9) 5.2 (1.3) 6.1 (1.1) 6.3 (0.9) 5.1 (0.9)
(3.8±7.3) (5.0±8.1) (5.6±8.1) (3.6±6.3)
S (n=10) 6.2 (3.3) 6.4 (3.2) 7.6 (4.7) 7.0 (5.1)
(2.5±8.3) (4.1±15.2) (4.4±20.5) (2.5±20.3)
G (n=10) 5.8 (1.2) 5.8 (1.6) 5.8 (1.4) 6.3 (1.9)
(4.2±7.4) (3.1±8.5) (4.1±8.4) (2.7±11.4)
Cdyn (ml mmHg±1) C (n=9) 21.6 (4.6) Not measured 23.4 (6.9) 22.9 (8.0)
(15.8±27.9) (13.2±31.1) (14.4±37.3)
S (n=10) 20.6 (4.1) Not measured 20.4 (5.6) 20.2 (6.6)
(11.5±26.7) (11.7±31.7) (10.7±32.8)
G (n=10) 25.0 (5.3) Not measured 26.3 (5.9) 28.8 (6.0)
(14.7±34.6) (18.1±38.0) (17.1±38.8)

228
 Post-traumatic shock and lung functions

below control in both shock groups, re¯ecting haemodilu-
tion and a reduction in oxygen delivery consequent on the
asanguineous ¯uids used.
Fig 1 Comparison of mean blood lactate concentration (SEM) in the three
a priori groups (A) and the post-hoc groups (B). Post-hoc Group A:
animals showing lactate greater than 3 mmol litre±1 during the shock
phase (n=12); Group B: animals not meeting the `lactate criteria' during
the shock phase (n=8); and Group C: control group (n=9). *Indicates test
group signi®cantly different from control group.
Blood lactate concentration continued to rise after
resuscitation in Groups S and G. Post-hoc analysis, how-
ever, showed that this apparent worsening of shock occurred
only in those animals that met all four `shock criteria'
(n=12) before the end of the shock phase (Fig. 1). In order to
explain the underlying reasons for persistent hyper-
lactataemia in these animals further post-hoc comparisons
between oxygen delivery (DO2), oxygen consumption
(VO2) and arterial-mixed venous oxygen content difference
[C(a±v)O2] was made between three post-hoc groups [C:
control group (n=9); B: group of animals that did not meet
the `lactate criteria' during the shock phase (n=8) and A:
group that met the `lactate criteria' during the shock phase
(n=12)]. The results are summarized in Table 3. In both
shock groups the mean DO2 following volume replacement
was less than the control groups with a greater and
statistically signi®cant reduction in Group A (95% CI:
C=313±443, B=215±346, A=176±298 ml min±1). This
reduction in DO2 was, however, not accompanied by a
corresponding increase in C(a±v)O2 in either shock group.
(95% CI for C(a±v)O2 at the post-resuscitation phase:
C=35±50, B=27±41, and A=35±47 ml litre±1). There was no
signi®cant difference in the volume of blood loss between
groups that met all four `shock criteria' and those that met
only three of the four criteria (41% (9); range 25±60 vs 36%
(8); range 24±50% of total blood volume).
Overall, there were no signi®cant changes in PaO2/FIO2,
QÇ S/Q
Ç T, EVLW, or Cdyn during the experiments (Table 2).
However, in ®ve experiments a greater than 30% increase in
EVLWi was observed during the period of study. Four of
these animals showed an increase in blood lactate greater
than 3 mmol litre±1 during the shock phase (S=3, G=1). A
greater than 10% reduction in lung compliance was seen in
four experiments (one in C and three from S and G). Both
animals that died showed blood lactate concentration

Table 3 DO2, VO2, and CO2(a±v) data for the three post-hoc groups based on the lactate response during the shock phase. Group A: animals showing lactate
greater than 3 mmol litre±1 during the shock phase (n=12). Group B: animals not meeting the `lactate criteria' during the shock phase (n=8); and Group C:
control group (n=9). *Indicates test group signi®cantly different from control group

Baseline Start of shock End of shock Post-resuscitation

phase phase phase

DO2 (ml min±1) C (n=9) 433 (161) 450 (154) 387 (105) 378 (95)
(268±731) (259±700) (252±563) (260±549)
B (n=8) 453 (95) 251 (47)* 252 (54)* 281 (97)
(340±593) (170±329) (147±318) (131±479)
A (n=12) 519 (151) 223 (63)* 175 (89)* 237 (75)*
(297±782) (112±296) (70±366) (119±349)
±1
VO2 (ml min ) C (n=9) 163 (42) 154 (37) 147 (46) 152 (62)
(112±231) (107±207) (85±245) (82±284)
B (n=8) 169 (24) 171 (34) 161 (41) 148 (32)
(138±215) (123±232) (98±225) (79±188)
A (n=12) 177 (38) 158 (52) 127 (62) 142 (53)
(129±246) (73±225) (47±257) (88±245)
C(a±v)O2 (ml litre±1) C (n=9) 46 (5) 41 (8) 42 (8) 43 (12)
(37±52) (31±49) (32±51) (24±58)
B (n=8) 44 (10) 70 (13)* 63 (14)* 34 (8)
(31±64) (48±85) (40±79) (25±48)
A (n=12) 42 (8) 76 (18)* 80 (19)* 41 (8)
(32±56) (36±96) (62±126) (32±55)

229
 Nirmalan et al.
greater than 3 mol litre±1 during the shock phase (with
further rise in blood lactate concentration during the
resuscitation phase), greater than 30% increase in EVLWi
and greater than 10% reduction in Cdyn.
A comparison between the COLD system and gravimetric
methods showed that the former method consistently
underestimated EVLWi (Bland-Altman plot: bias=64 ml;
SD=44 ml). At post-mortem examination it was observed
that three animals (S=2; G=1) had increased amounts of
¯uid within the pleural and pericardial spaces. All three
showed a greater than 30% increase in EVLWi and two died
during resuscitation.
Discussion
Acute lung injury is generally recognized as a late
complication and few studies in the literature have explored
its importance in the initial clinical presentation of shock.3±5
The current study reproduces the physiological ®ndings
occurring in patients with a combination of moderate
skeletal trauma and acute haemorrhage and investigates the
pulmonary changes within the ®rst hour after the onset of
shock. Overall, no signi®cant changes in oxygenation,
compliance, or EVLWi were demonstrable either at the end
of the shock period or after initial ¯uid resuscitation. A
signi®cant increase in relative pulmonary pressures and
PVR, however, was evident even within this relatively short
time period. Relative pulmonary hypertension was demon-
strable in both groups at the end of the shock phase (95% CI
for MPAP/MAP ratio at the end of shock phase;
C=0.15±0.24, S=0.28±0.38, G=0.32±0.42) and persisted
even after volume replacement with either crystalloid or
colloid solutions.
Post-traumatic shock has three important facets: hypo-
volaemia, pain, and tissue injury and it is recognized that
haemodynamic responses to haemorrhage are altered
signi®cantly by concomitant nociception and injury.19 20
As both local and systemic factors are important in the
evolution of lung injury bilateral long bone fractures and
haemorrhage were included in the present model. The end
points of shock in most published laboratory work have
been based on removal of a predetermined percentage of
blood volume,19 20 removal of blood until a predetermined
arterial blood pressure was achieved21 or alterations in
resting membrane potential (RMP).3 Clinical evaluation of
shock, however, is based on the extent of global blood ¯ow
reduction and the consequences of tissue ischaemia. To the
best of our knowledge this is the ®rst study that took a
combination of blood ¯ow-based measurements (CO) and
clinical measures of tissue ischaemia (SvO2 and blood lactate
concentration) in the de®nition of shock in laboratory
models. The threshold value for blood lactate concentration
was 3 mmol litre±1 as patients with blood lactate concen-
tration greater than 3 mmol litre±1 on admission to an
emergency department are known to carry a worse progno-
230
sis than others.22 In the UK, victims of major trauma and
shock receive ¯uid resuscitation at the site of injury by
trained paramedical staff as an immediate priority. As a
result, untreated shock lasting for more than an hour is rare
and the shock phase in the present study was restricted to
one hour in keeping with this clinical reality.
The choice of anaesthetic agent is always dif®cult, as all
agents will modify the cardiovascular responses to trauma
and shock to some extent. `Saffan' was chosen for most part
of the study as it is the only agent that has been shown to
preserve the cardiovascular responses to trauma.23
Halothane was used at the beginning to facilitate intubation
without neuromuscular blocking agents and a predomin-
antly `Saffan' based anaesthetic regimen was used as soon
as vascular access was established. Some important limita-
tions of the study, however, require emphasis. Most
importantly, volume replacement was stopped when CO
was restored and maintained to within 90% of baseline as it
was essential to limit the volume of clear ¯uid administered
to prevent unacceptable levels of haemodilution and
premature death. Both shock groups were therefore under-
resuscitated by design. A low MAP associated with
inadequate volume resuscitation and a `normal' MPAP
associated with hypoxic pulmonary vasoconstriction (as a
result of low SvO2)24 may also account for the observed
changes in MPAP/MAP and PVR. Adequate resuscitation
after major haemorrhage requiring blood transfusion and
volume replacement with clear ¯uids always raises the
question of volume overload (absolute or relative to a
constricted circulation) in relatively short experimental
studies such as ours. As the primary hypothesis was that
shock per se results in increased EVLW, under resuscitation
would not have been important had we demonstrated a
signi®cant increase in EVLWi. As an increase in EVLWi
was not demonstrated and the conclusions are based on
changes to pulmonary and systemic arterial pressures,
adequacy of volume resuscitation is an important confound-
ing factor. In this context some of the incidental ®ndings
reported in a more recent study are extremely pertinent and
should be considered along with our ®ndings. In a similar
study by Chiara and colleagues,21 (with a shock period of 1 h
followed by ¯uid/blood resuscitation over 2 h) in spite of
adequate resuscitation (including restoration of haemoglo-
bin and MAP) MPAP, the ratio MPAP/MAP and PVR
remained signi®cantly higher than the corresponding base-
line values. As the study was not designed to evaluate the
pulmonary consequences of shock the above authors did not
explore these incidental ®ndings further or comment upon
their signi®cance. The signi®cant reduction in SVR' in the
present study (at the post-resuscitation phase) suggests that
sympathetically mediated vasoconstriction was overcome
by the effects of reperfusion. Further, in the two animals that
died relative pulmonary hypertension was associated with a
large increase in EVLW (>30%), reduction in Cdyn, and
increased ¯uid within the pleural and pericardial spaces.
These additional features and the data reported by Chiara
 Post-traumatic shock and lung functions
and colleagues21 support our conclusion that pulmonary
vascular changes may begin within the ®rst hour after the
onset of shock. This observation, if con®rmed by further
clinical studies is very relevant to the management of
patients with prolonged shock.
Many of the derived haemodynamic parameters (SV,
LVSW, SVR', and PVR) are invariably linked to changes in
CO, HR, MAP, and MPAP. Nevertheless, such derived
parameters remain popular and useful in the critical care
setting and, therefore, we have applied them to the present
study. Right atrial pressure measurements were not made
and, therefore, a surrogate measure of SVR' has been used
to assess the systemic vascular bed. Inclusion of
right atrial pressure in the derivation of SVR is likely to
amplify the observed differences in SVR' even further
(because of the lower MAP in the two shock groups) and
this makes the reported differences relevant. The derivation
of SVR' during the shock phase (or PVR for that matter) is
based on the assumptions that blood ¯ow through a
constricted circulation is laminar and blood is a
Newtonian ¯uid.25 Both assumptions are invalid in the
presence of profound vasoconstriction and hypovolaemia
and this is a recognized limitation of calculated vascular
resistance. Therefore, all comments on SVR' and PVR are
based on measurements made at baseline and during the
post-resuscitation phase only. Thirdly, the duration of the
instrumentation phase varied during the period of study as a
result of the learning curve inherent in undertaking surgical
instrumentation. This effect, however, was taken into
account in the experimental design by adhering to a strict
randomization protocol.
Early changes in lung function has been reported
previously in pigs by Noble4 and in baboons by Holcroft
and colleagues.3 The lack of a control group in Noble's
study design renders the ®ndings inconclusive and the
clinical signi®cance of shock de®ned by RMP criteria by
Holcroft and colleagues is unclear. We have, therefore, used
more clinically relevant criteria such as CO, MAP, blood
lactate concentration, and SvO2. The study shows that in the
early stages of shock commonly monitored parameters such
as PaO2/FIO2, Q Ç S/Q
Ç T, Cdyn, and EVLWi may remain
unaltered (except in the more severe and irreversible
forms of shock). Changes in relative pulmonary pressures
(MPAP/MAP) may in fact be the earliest evidence of
pulmonary vascular changes in such patients. The present
study clearly highlights the differences between pulmonary
and systemic vascular beds to ischaemic reperfusion.
Translocation of gut-derived endotoxins into the systemic
circulation following mesenteric ischaemia is a well-
recognized cause for intense pulmonary vasoconstriction
associated with systemic vasodilatation.6 26 We believe this
is one of the possible explanations for the differences in the
systemic and pulmonary circulations seen after volume
replacement in the present study.
The reasons for a persistent elevation in blood lactate
concentration in some animals even after ¯uid resuscitation
231
is complex. Post-hoc analysis shows that even though ¯uid
resuscitation resulted in improvements in DO2 post-resus-
citation DO2 values were substantially below that of the
control group and baseline values for each of the post-hoc
groups. In spite of the low DO2 global oxygen extraction
(re¯ected by C(a±v)O2) remained within normal limits. This
implies that a combination of reduced oxygen delivery and
defective extraction contributed to the persistent elevation
in blood lactate concentration. In keeping with current
clinical consensus14 15 no major differences were seen
between the two groups resuscitated with either gelatine or
saline. Both shock groups showed a signi®cant reduction in
mean SVR' associated with an increase in mean PVR
following volume resuscitation. The lack of a statistically
signi®cant rise in PVR for Group G we believe, is
attributable to the relatively small numbers of animals in
each group. The calculation of sample size, however, was
based on the number of animals required within each group
to demonstrate a more than 20% rise in EVLWi during the
study period. Thus, even though the study was not
speci®cally powered to demonstrate differences between
the two ¯uid groups, differences if any are likely to be small
and unimportant. The percentage blood volume removed
from each group was similar (Group S=39.4% (SD=6.7%);
Group G=39.1% (SD=11.6%)) although the amount of blood
removed from individual animals ranged from 24±60% of
total blood volume highlighting the limitations of many
traditionally used shock models based on ®xed percentage
volume loss.19 20 The present ®nding is in keeping with the
clinical observation that the haemodynamic and metabolic
consequences of haemorrhage and shock vary markedly in
healthy subjects.
In conclusion, the present study suggests that pulmonary
and systemic vascular dysfunction may be demonstrable
within the ®rst hour after the onset of shock. As such the
stage for subsequent development of acute lung injury and
or multiple organ failure may be set very early in the clinical
course of shock and should be an important consideration in
the initial resuscitation of patients presenting with shock. As
a result of ethical and ®nancial constraints it was not
possible to include larger numbers of animals in the present
study and therefore further laboratory studies where DO2
and SvO2 values are restored and maintained over a longer
period are necessary to con®rm the above ®ndings and to
elucidate the pathophysiology of early pulmonary changes
in shock.
Acknowledgements
We are grateful to Dr M. O. Columb, South Manchester University
Hospitals for the assistance in data analysis. The role of Mrs H. Marshall
and Mr T. Riney in conducting these experiments is gratefully acknow-
ledged. The studies were funded by the Medical Research Council UK and
Maelor Pharmaceuticals, Wrexham (UK). Data included in this paper were
presented to the Anaesthetic Research Society, UK in March 2001.
 Nirmalan et al.
References
1 Cournand A, Riley RL, Bradley SE, et al. Studies of the circulation
in clinical shock. Surgery 1943; 13: 964±95
2 Brewer LA, Burbank B, Samson PC, et al. The `wet lung' in war
casualties. Ann Surg 1946; 123: 343±62
3 Holcroft JW, Trunkey DD. Extravascular lung water following
haemorrhagic shock in the baboons. Comparison between
resuscitation with Ringer's lactate and Plasmanate. Ann Surg
1974; 180: 408±15
4 Noble WH. Early changes in lung water after haemorrhagic
shock in pigs and dogs. Can Anaesth Soc J 1975; 22: 39±49
5 Sturm JA, Oestern HJ, Trentz O, Neubauer M, Trentz OA, Lewis
FR. Extravascular lung water in traumatic shock of the dog
(author's translation). Chir Forum Exp Klin Forsch 1980; 95±9
6 Sori AJ, Rush BF jr, Lysz TW, Smith S, Machiedo GW. The gut as
a source of sepsis after haemorrhagic shock. Am J Surg 1988;
155: 187±92
7 Kreil EA, Greene E, Fitzgibbon C, Robinson DR, Zapol WM.
Effects of recombinant human tumour necrosis factor alpha,
lymphotoxin and Escherichia coli lipopolysaccharide on
haemodynamics, lung microvascular permeability and
eicosanoid synthesis in anaesthetized sheep. Circ Res 1989; 65:
502±4
8 Haslett, C, Worthen GS, Giclas PC, Morrison DC, Henson JE,
Henson PM. The pulmonary vascular sequestration of
neutrophils in endotoxaemia is initiated by an effect of
endotoxin on the neutrophil in the rabbit. Am Rev Respir Dis
1987; 136: 9±18
9 Worthen GS, Haslett C, Rees AJ, Gumbay RS, Henson JE,
Henson PM. Neutrophil-mediated pulmonary vascular injury.
Synergistic effect of trace amounts of lipopolysaccharide and
neutrophil stimuli on vascular permeability and neutrophil
sequestration in the lung. Am Rev Respir Dis 1987; 136:
19±28
10 Advanced Trauma Life Support Student Manual, 7th Edn. Chicago:
American College of Surgeons, 2002
11 Schierhout G, Roberts I. Fluid resuscitation with colloid or
crystalloid solutions in critically ill patients: a systematic review
of randomised trials. BMJ 1998; 316: 961±4
12 Cochrane Injuries Group Albumin Reviewers. Human albumin
administration in critically ill patients. Systematic review of
randomised controlled trials. BMJ 1998; 317: 235±40
13 Choi PTL, Yip G, Quinonez LG, Cook DJ. Crystalloids vs colloids
in ¯uid resuscitation: a systematic review. Crit Care Med 1999;
27: 200±10
14 Hein LG, Albrecht M, Dworschak M, Frey L, Bruckner UB. Long
232
term observation following traumatic-haemorrhagic shock in the
dog: a comparison of crystalloidal vs colloidal ¯uids. Circ Shock
1988; 26: 353±64
15 Redl H, Krosl P, Schlag G, Hammerschmidt DE. Permeability
studies in a hypovolaemic traumatic shock model: comparison of
Ringer's lactate and albumin as volume replacement ¯uids.
Resuscitation 1989; 17: 77±90
16 Pearce ML, Yamashita J, Beazell J. Measurement of pulmonary
oedema. Circ Res 1965; 16: 482±8
17 Nunn JF. (1993) Applied Respiratory Physiology, 4th Edn. Oxford:
Butterworth-Heinemann, 1993; 55±60
18 Edwards JD, Mayall RM. Importance of the sampling site for
measurement of mixed venous oxygen saturation in shock. Crit
Care Med 1998; 26: 1356±60
19 Rady MY, Kirkman E, Cranley J, Little RA. A comparison of the
effects of skeletal muscle injury and somatic afferent nerve
stimulation on the response to haemorrhage in anaesthetized
pigs. J Trauma 1993; 35: 756±61
20 Mackway-Jones K, Foex BA, Kirkman E, Little RA. Modi®cation
of the cardiovascular response to haemorrhage by somatic
afferent nerve stimulation with special reference to gut and
skeletal muscle blood ¯ow. J Trauma Injury Infect Crit Care 1999;
47: 481±5
21 Chiara O, Pelosi P, Segala M, et al. Mesentric and renal oxygen
transport during haemorrhage and reperfusion: evaluation of
optimal goals for resuscitation. J Trauma Injury Infect Crit Care
2001; 51: 356±62
22 Lecky FE, Little RA, Maycock PF, et al. Effect of alcohol on the
lactate/pyruvate ratio of recently injured adults. Crit Care Med
2002; 30: 981±85
23 Timms RJ. The use of anaesthetic steroids alphaxalone-
alphadolone in studies of the forebrain in the cat. J Physiol
1976; 256: 71±2P
24 Marshall C, Marshall B. Site and sensitivity for stimulation of
hypoxic pulmonary vasoconstriction. J Appl Physiol 1983; 55:
711±6
25 Guyton AC. Overview of the circulation and medical physics of
pressure, ¯ow and resistance. In: Guyton AC ed. Textbook of
Medical Physiology, 8th Edn. Philadelphia: WB Saunders, 1991;
150±8
26 Kreil EA, Greene E, Fitzgibbon C, Robinson DR, Zapol WM.
Effects of recombinant human tumour necrosis factor alpha,
lymphotoxin and Escherichia coli lipopolysaccharide on
haemodynamics, lung microvascular permeability and
eicosanoid synthesis in anaesthetized sheep. Circ Res 1989; 65:
502±4