Open Journal of Pediatrics, 2012, 2, 197-213 OJPed
http://dx.doi.org/10.4236/ojped.2012.23033 Published Online September 2012 (http://www.SciRP.org/journal/ojped/)
The effects of perfluorocarbon dosing strategy on cerebral
blood flow when starting partial liquid ventilation: A
randomised, controlled, experimental study
Mark W. Davies1,2,3, Kimble R. Dunster1,4,5, John F. Fraser2,5, Paul B. Colditz3
1Grantley Stable Neonatal Unit, Royal Brisbane and Women’s Hospital, Brisbane, Australia
2Department of Paediatrics & Child Health, The University of Queensland, Brisbane, Australia
3Perinatal Research Centre, The University of Queensland, Royal Brisbane and Women’s Hospital, Brisbane, Australia
4Medical Engineering Research Facility, Queensland University of Technology, Brisbane, Australia
5Critical Care Research Group, The Prince Charles Hospital, Brisbane, Australia
Received 14 May 2012; revised 16 July 2012; accepted 25 July 2012
ABSTRACT
Introduction: Partial liquid ventilation may benefit
the lung disease in preterm neonates but intratracheal
instillation of perfluorocarbon increases cerebral blood
flow and may cause brain injury. We aimed to deter-
mine if the effects of perfluorocarbon administration
on cerebral blood flow vary by dose-volume, rate of
administration, endotracheal tube portal of entry, or
closely targeting PaCO2. Methods: Forty-two dosing
events (in eleven rabbits) were randomised to different
dosing strategies, including a sham (i.e., placebo/con-
trol) dose of air over 20 min, 20 mL/kg of perfluoro-
carbon slowly over 20 min, 10 mL/kg of perfluoro-
carbon slowly over 20 min, 10 mL/kg of perfluoro-
carbon moderately fast over 10 min, 10 mL/kg of per-
fluorocarbon rapidly over 5 min, 10 mL/kg of per-
fluorocarbon slowly over 20 min via the endotracheal
tube tip lumen (as opposed to the proximal end of the
tube used in all other groups), or 10 mL/kg of per-
fluorocarbon slowly over 20 min whilst targeting a
PaCO2 of 45 - 50 mmHg. Blood gases, haemo-
dynamics, cortical cerebral blood flow and carotid
flow were recorded continuously for 30 minutes from
the start of each dose. Results: Carotid flow increased
with 20 mL/kg perfluorocarbon and cortical cerebral
blood flow was significantly more variable. Carotid
and cortical cerebral blood flow increased using 10
mL/kg or 20 mL/kg with no difference between the
two dose-volumes. There was no difference in cerebral
blood flow by rate of administration, but carotid blood
flow was more variable during slow administration.
There were no differences in the increase in cerebral
blood flow by portal of entry. If PaCO2 was main-
tained between 45 - 50 mmHg there was no increase
in cerebral blood flow and there was less variable
carotid flow. Conclusions: Cerebral blood flow in-
creases with perfluorocarbon dosing. This occurs
regardless of the dose-volume of perfluorocarbon.
These effects were mitigated by closely targeting
PaCO2.
Keywords: Cerebral Blood Flow; Fluorocarbons; Infant;
Newborn Preterm; Partial Liquid Ventilation
1. INTRODUCTION
Preterm infants, especially those born extremely preterm,
require a great deal of support ex utero. They are often
severely ill and their chances of survival are greatly
reduced [1]. Many die because of general immaturity; as
multiple organ systems, including the lungs, cannot adapt
to extrauterine life. Many die of overt lung disease [1].
However, the positive pressure ventilation which they
require to survive is often a cause of significant and
persistent lung injury [2].
Partial liquid ventilation has been touted as an alter-
native form of respiratory support for extremely immature
lungs and severe lung disease: not only to provide a more
effective form of respiratory support but to provide that
support with much less lung injury [3]. A recent exper-
imental study has renewed the promise that partial liquid
ventilation has significant potential to alter the course of
neonatal lung disease and reduce respiratory morbidity
[4].
Extremely preterm infants are also prone to specific
forms of brain injury such as intraventricular haemorrhage,
periventricular leucomalacia and other types of white
matter injury [5]. Rapid changes or disturbance of ven-
tilation can lead to fluctuations in cerebral blood flow
which have been shown to substantially increase the risk
of brain injury [6,7].

M. W. Davies et al. / Open Journal of Pediatrics 2 (2012) 197-213 198
In a previous study we have shown that preterm lambs
receiving a 30 mL/kg dose of tracheal perfluorocarbon
liquid, given over 20 minutes at the start of partial liquid
ventilation, had an increased cortical cerebral blood flow
[8]. This did not seem to be due to any general haemo-
dynamic disturbance, although the lambs that had tracheal
perfluorocarbon did have a small, non-significant increase
in arterial carbon dioxide [8]. The increase in cortical
cerebral blood flow did not differ with varying doses of
tracheal perfluorocarbon (20, 30 or 40 mL/kg given over
20 minutes) [9].
As with any treatment, the benefits of partial liquid
ventilation will need to be weighed against any harm. It
is known that increases in cerebral blood flow put very
preterm infants at greater risk of intraventricular haemor-
rhage and brain injury [6,7]. Any gains in using partial
liquid ventilation to improve gas exchange should not be
at the expense of a higher risk of brain injury. It may be
that smaller doses of perfluorocarbon will allow adequate
treatment of lung disease with minimal effects on cerebral
blood flow. Also, regardless of the dose used, it may be
that perfluorocarbon dosing can be optimised to minimise
any effects on cerebral blood flow.
Before its use in human preterm neonates, the effects
on cerebral blood flow of partial liquid ventilation need
to be determined in a relevant animal model. It has not
been established whether different methods of giving the
initial dose of perfluorocarbon affect the degree and
duration of any changes in cerebral blood flow.
We hypothesised that a 20 mL/kg dose of perfluo-
rocarbon administered when starting partial liquid ven-
tilation would disturb cerebral blood flow. Liquid ven-
tilation starts when the perfluorocarbon liquid is intro-
duced into the lungs. We also hypothesised that the degree
and duration of changes in cerebral blood flow would
vary with:
 the dose of perfluorocarbon administered;
 the duration over which the dose is given (5, 10 or 20
minutes);
 the portal of entry of the perfluorocarbon into the
ETT (side-port on the ETT manifold versus an auxi-
liary lumen further down the ETT); and
 whether, during pressure-controlled ventilation, the
peak inspiratory pressure is adjusted to maintain
PaCO2 or not.
Our aims were to:
1) confirm in a rabbit model (with no lung disease) that
administration of a dose of 20 mL/kg of perfluorocarbon
during commencement of partial liquid ventilation
increased cerebral blood flow and its variability;
2) investigate the effects of a lower initial dose (10
mL/kg versus 20 mL/kg) of perfluorocarbon on cerebral
blood flow and its variability;
3) investigate the effects of different rates of ad-
ministration of the initial dose of perfluorocarbon (10
mL/kg over 5, 10 or 20 minutes) on cerebral blood flow
and its variability;
4) investigate the effects of the ETT portal of entry of
the initial dose of perfluorocarbon on cerebral blood flow
and its variability;
5) investigate the effects of adjusting the peak inspir-
atory pressure to maintain the arterial carbon dioxide
between 45 and 50 mmHg during the initial dose of
perfluorocarbon on cerebral blood flow and its variability.
2. MATERIALS AND METHODS
To study the effects of different dosing strategies of
perfluorocarbon, we used the adult New Zealand white
rabbit which is in the human neonatal weight-range
(similar to the preterm lambs previously studied [8,9]).
The necessary procedures and monitoring are readily
performed in rabbits of this size. The use of an adult
rabbit with no lung disease allows repeated study over
many hours.
2.1. Animal Preparation
Anaesthesia was induced in each rabbit with intra-
muscular atropine (0.1 mg/kg) followed by intra-muscular
Zoletil (tiletamine and zolazepam, 10 mg/kg) and xylazine
(2 mg/kg). An intra-venous line (22 - 24 gauge Insyte
catheter) was inserted into an ear vein and anaesthesia
maintained with an infusion of propofol (0.5 mL/kg/hr).
An intravenous infusion of 3 mL/kg/hr of 10% dextrose
was given throughout. The rabbits were then intubated
via a tracheostomy with a 3.0 HI-LO Jet multi-lumen
endotracheal tube (MuLETT) [Mallinckrodt Inc, Argyle,
New York, USA] with a ligature around the trachea (to
prevent leakage). Mechanical ventilation was commenced
as described below. Respiratory function was monitored
with a VenTrak neonatal pneumotachograph and respir-
atory mechanics monitor (Model 1550, Novametrix
Medical Systems, Wallingford, Connecticut, USA).
A NeoTrend catheter (Diametrics Medical, High
Wycombe, UK), calibrated as per the manufacturer’s
instructions, was then inserted into the descending aorta
via the femoral artery. This allowed continuous mea-
surement of intra-arterial oxygen tension (PaO2), carbon
dioxide tension (PaCO2), pH, temperature, bicarbonate,
base excess and oxygen saturation.
The right common carotid artery was exposed and a 2
mm ultrasonic flow probe (Transonic 2SB) was placed
around the right carotid artery and connected to the T206
Transonic Flowmeter (Transonic Systems Inc., Ithaca,
New York, USA).
Cerebral cortical blood flow was monitored using a
laser Doppler probe (Oxford Array, Oxford Optronix,
Oxford, UK) inserted into the parietal cortex (through a

M. W. Davies et al. / Open Journal of Pediatrics 2 (2012) 197-213 199
burr hole half-way between a line joining the outer
corner of the eye and top of the ear insertion, and the top
of skull) and the probe was held in place as previously
described [10].
Standard physiological monitoring included heart rate
(ECG), oxygen saturation (peripheral pulse oximetry)
and continuous arterial blood pressure.
After completion of all dosing events the rabbit was
killed with 5 mL pentobarbitone sodium (Lethabarb, 325
mg/mL). The mean (SD) duration from anaesthesia in-
duction to pentobarbitone injection was 7.7 (0.74) hours.
2.2. Mechanical Ventilation
The rabbits were ventilated with a neonatal ventilator
(Bear Cub, Bear Medical Systems, Riverside, California)
at 30 breaths per minute, peak inspiratory pressure (PIP)
12 cmH2O, positive end expiratory pressure (PEEP) 5
cmH2O, and inspiratory time of 0.75 seconds. The frac-
tion of inspired oxygen (FiO2) was initially 0.21, in-
creased as necessary to maintain pulse oximetry oxygen
saturations between 94% and 97%. PIP was adjusted to
maintain PaCO2 between 40 - 45 mmHg until the com-
mencement of each dosing strategy phase. No further
adjustment of ventilation, with the exception of FiO2,
were allowed except during the PLV10-slow-proxi-
mal-PC-adjust PIP dosing strategy (see below).
2.3. Perfluorocarbon Dose Volume
The perfluorocarbon dose volumes were chosen to en-
able comparisons with earlier studies in preterm lambs.
We chose 20 mL/kg, as used in our earlier preterm lamb
studies [9], in order to confirm similar effects of per-
fluorocarbon dosing on cerebral blood flow. We then
chose 10 mL/kg as the comparison dose as it was more
easily removed from the lungs between dosing events
and yet was in the range of doses with demonstrated ef-
ficacy in animal studies. Dose-dependent improvement
in lung function has been demonstrated during partial
liquid ventilation but the majority of improvement in
lung function has occurred with relatively small doses
(up to ~10 mL/kg) [11,12]. In rabbits 9 - 12 mL/kg has
been shown to improve lung function [12-14]. In animal
models with no lung disease 10 mL/kg has been shown
to disturb gas exchange less than 30 mL/kg and thus pro-
vide a more stable model to study haemodynamic ef-
fects [15].
2.4. Group Allocation
Each rabbit was allocated to multiple dosing strategies to
reduce the number of animals used. Following pre-
paration, each rabbit was randomly allocated a perfluo-
rocarbon dosing strategy and then monitored for one
minute before and 30 minutes after the start of each
dosing event. After each dosing event (and washout
period) each rabbit was randomly allocated to another
perfluorocarbon dosing strategy.
Eleven rabbits (mean  SD weight of 4.0  0.59 kg)
were randomised to up to six of the seven following
dosing strategies:
1) A CONTROL-PC group which was given a sham
(i.e., placebo/control) dose of air (10 mL/kg) over 20 min
via the wye-piece at the proximal end of the ETT (see
Figure 1) on pressure-controlled ventilation;
2) A PLV20-slow-proximal-PC group which was given
20 mL/kg of FC-77 slowly over 20 min via the wye-
piece at the proximal end of the ETT on pressure-con-
trolled ventilation;
3) A PLV10-slow-proximal-PC group which was given
10 mL/kg of FC-77 slowly over 20 min (0.5 mL/kg/min)
via the wye-piece at the proximal end of the ETT on
pressure-controlled ventilation;
4) A PLV10-faster-proximal-PC group which was
given 10 mL/kg of FC-77 moderately fast over 10 min
(1.0 mL/kg/min) via the wye-piece at the proximal end
of the ETT on pressure-controlled ventilation;
5) A PLV10-rapid-proximal-PC group which was
given 10 mL/kg of FC-77 rapidly over 5 min (2.0 mL/
kg/min) via the wye-piece at the proximal end of the
ETT on pressure-controlled ventilation;
6) A PLV10-slow-tip-PC group which was given 10
mL/kg of FC-77 slowly over 20 min via ETT tip lumen
(see Figure 1) on pressure-controlled ventilation;
7) A PLV10-slow-proximal-PC-adjust PIP group
which was given 10 mL/kg of FC-77 slowly over 20 min
via the wye-piece at the proximal end of the ETT with
the PIP adjusted to maintain the PaCO2 between 45 and
50 mmHg on pressure-controlled ventilation.
In groups 1 to 6 above no further adjustment of venti-
lation (with the exception of FiO2) was allowed after the
commencement of each dosing event studied.
Portal of Entry
The dose of perfluorocarbon is usually given through a
sideport wye-piece on the ETT-connector at the proximal
end of the endotracheal tube. In this study doses of per-
fluorocarbon were given via one of two portals, either
through a wye-piece at the proximal end of the ETT or
through the ETT tip lumen (Figure 1). The use of a
wye-piece was necessary to allow the pneumotachograph
to be fitted between the ETT-circuit manifold and the
ETT connector and prevent reflux of perfluorocarbon
into the pneumotachograph during instillation.
2.5. Washout
Each dosing event was monitored for 30 minutes from
the start of perfluorocarbon dosing. After each dosing

M. W. Davies et al. / Open Journal of Pediatrics 2 (2012) 197-213 200
Figure 1. Configuration of the multi-lumen endotracheal tube (MuLETT), pneumotachograph, ETT con-
nector, ETT circuit manifold and wye-piece.
event as much FC-77 as possible was sucked out of the
ETT and trachea using a 5FG umbilical artery catheter
and syringe. Any remaining FC-77 was then left to
evaporate from the lungs via the expiratory line of the
ventilator circuit. Tracheal gas insufflation at 2 L/min via
the distal ETT tip lumen was also used to increase the
rate of evaporation.
The completeness of perfluorocarbon washout could
be confirmed when the expired tidal volume equalled the
inspired tidal volume. Because there is no expired gas in
the inspiratory limb of the ventilator circuit and FC-77
vapour is known to affect tidal volume measurement
with the VenTrak pneumotachograph [16] it was possible
to determine that all the FC-77 had evaporated from the
lung when the expired tidal volume equalled the inspired
tidal volume (i.e., when there was no more perfluo-
rocarbon vapour in the expired gases). The effect of
FC-77 vapour has been well documented [16]. The mea-
surement of tidal volumes in the range from 5 to 25 mL
using the VenTrak pneumotachograph has coefficients of
variation ranging from 0.008 to 0.04 without perfluo-
rocarbon vapour and from 0.006 to 0.06 with perfluo-
rocarbon vapour. The differences seen in TV measure-
ment are more than 7%.
After this washout period, between each dosing event,
mechanical ventilation continued as before. Once the
rabbit was stable they were then randomly allocated to
another perfluorocarbon dosing strategy as above.
Stabilisation times before each dosing event differed
between those dosing events where there was no prior
perfluorocarbon use (therefore no washout was required)
and those dosing events where a prior washout was

M. W. Davies et al. / Open Journal of Pediatrics 2 (2012) 197-213 201
required. The mean (SD) stabilisation time for the former
was 15.4 (6.96) minutes and for the latter was 64.9
(14.26) minutes.
2.6. Measurements
Two methods of assessing cerebral blood flow were used,
both recorded continuously: cerebral cortical blood flow
(laser Doppler flowmetry), as used in the previous pre-
term lamb studies; and right carotid blood flow (Tran-
sonic Flowmeter). Carotid flow was used to assess more
global brain blood flow to ensure that previous effects
seen were not restricted to the cerebral cortex. The coef-
ficient of variation (CV) for these two measurements
were calculated to assess their variability.
The following were all recorded continuously: heart
rate, blood pressure, peripheral oxygen saturation (pulse
oximetry), PaO2, PaCO2, pH, bicarbonate, base excess,
arterial oxygen saturation, temperature, dynamic com-
pliance, peak inspiratory pressure, and tidal volume.
FiO2 was recorded at the start of each dosing event and
the timing of any changes recorded thereafter.
2.7. Sample Size
Sample size calculation was based on carotid blood flow
data from Takeuchi et al. [17] where mean (SD) carotid
blood flow was 13.8 (3.39) mL/min. Therefore a sample
size of six in each group is required to show an absolute
difference of 6.1 mL/min between groups (
 = 0.05,
power of 80% –
 = 0.2).
2.8. Data Analysis
2.8.1. Data Display
For each dosing event mean values for continuously
measured variables and calculated variables were calcu-
lated at baseline (the average during the minute immedi-
ately prior to perfluorocarbon or air dosing) and every
minute for the 30 minute observation period. For each
1-min period means were calculated from all measure-
ments made over the one minute period. For the six dos-
ing events for each dosing strategy the mean (SD) was
calculated at baseline and every minute to 30 minutes
and these data were then displayed graphically over time.
For laser Doppler flow and carotid blood flow the per-
cent change from baseline was calculated for each datum
and then for each dosing event the data were averaged
for each one minute period. Calculations and graphs
were produced using Microsoft Excel 2002 (Microsoft
Corporation, Redmond, Washington, USA).
To assess the variability of laser Doppler flow and ca-
rotid flow, the coefficient of variation was calculated for
each one minute period for each dosing event. The coef-
ficient of variation is calculated by dividing the standard
deviation by the mean for each one minute period.
2.8.2. Statistical Analysis
To determine if there were any statistically significant
differences between the dosing strategy groups, the data
were compared at two time points: at the end of the dos-
ing period (i.e., at 5, 10 or 20 min) and at the point of
maximal difference between groups. These time points
were chosen as the most clinically relevant points for
between group comparisons. For the purpose of these
statistical tests the data were converted to change from
baseline—this controlled for any differences between
dosing strategy groups at baseline. If the data were nor-
mally distributed and the variances were equal the
groups were compared using Student’s t-test or one-way
ANOVA. If the data were not normally distributed or the
variances were not equal the groups were compared us-
ing the Mann-Whitney test or Kruskal-Wallis test. The
distribution of data were tested using the Kolmo-
gorov-Smirnov normality test with Dallal-Wilkinson-
Lillie for P value. The F test was used to compare vari-
ances. These data analyses were done using GraphPad
Prism version 4.03 (GraphPad Software Inc., San Diego,
CA). A p value of <0.05 was considered to be statistic-
cally significant.
2.9. Ethics Approval
Ethical approval was obtained from the University of
Queensland Animal Experimentation and Ethics Com-
mittee. All of the experimental procedures were per-
formed in accordance with the guidelines established by
the National Health and Medical Research Council of
Australia in its Australian Code of Practice for the Care
and Use of Animals for Scientific Purposes [18].
3. RESULTS
Figures 2 and 3 show the relative changes (% change
from baseline) for both carotid flow and cortical cerebral
blood flow for each perfluorocarbon dosing strategy. In
each figure the vertical scales are identical for ease of
comparison. Tables 1 and 2 summarise the maximum
variations in flow for the two cerebral blood flow vari-
ables by dosing strategy.
3.1. Does 20 mL/kg of Perfluorocarbon Disturb
Cerebral Blood Flow?
We compared the CONTROL-PC group with the
PLV20-slow-proximal-PC group.
Carotid blood flow was increased relative to baseline
during the initial 20 minute dose of 20 mL/kg perfluoro-
carbon when starting partial liquid ventilation compared
with continuing on pressure-controlled conventional
mechanical ventilation—Mann-Whitney test p = 0.04 at
the end of the dose and p = 0.04 at the time of maximum
difference (21 min). The highest mean increase in

M. W. Davies et al. / Open Journal of Pediatrics 2 (2012) 197-213 202
-40
-20
0
20
40
60
80
100
120
pre12 3 45 6 7 89101112131415161718192021222324252627282930
time (minutes)
% change
in
carotid
flow
CONTROL-PC
-40
-20
0
20
40
60
80
100
120
pre12 3 45 6 7 89101112131415161718192021222324252627282930
time (minutes)
% change
in
carotid
flow
CONTROL-PC
CONTROL-PC
(a)
-40
-20
0
20
40
60
80
100
120
pre123456789101112131415161718192021222324252627282930
time (minutes)
% change
in
carotid
flow
PLV20-slow-proximal-PC
FC-77 dose
-40
-20
0
20
40
60
80
100
120
pre123456789101112131415161718192021222324252627282930
time (minutes)
% change
in
carotid
flow
PLV20-slow-proximal-PC
FC-77 dose
PLV20-slow-proximal-PC
FC-77 dose
(b)
-40
-20
0
20
40
60
80
100
120
pre123456789101112131415161718192021222324252627282930
time
(
minutes
)
% change
in
c arotid
flow
PLV10-slow-proximal-PC
FC-77 dose
-40
-20
0
20
40
60
80
100
120
pre123456789101112131415161718192021222324252627282930
time
(
minutes
)
% change
in
c arotid
flow
PLV10-slow-proximal-PC
FC-77 dose
PLV10-slow-proximal-PC
FC-77 dose
(c)

M. W. Davies et al. / Open Journal of Pediatrics 2 (2012) 197-213 203
-40
-20
0
20
40
60
80
100
120
pre123456789101112131415161718192021222324252627282930
time
(
minutes
)
% change
in
carotid
flow
PLV10-faster-proxim al-P C
FC-77 dose
-40
-20
0
20
40
60
80
100
120
pre123456789101112131415161718192021222324252627282930
time
(
minutes
)
% change
in
carotid
flow
PLV10-faster-proxim al-P C
FC-77 dose
PLV10-faster-proximal-PC
FC-77 dose
(d)
-40
-20
0
20
40
60
80
100
120
pre12345678910 1112 13 1415 16 17 1819 202122 23 24 2526 27 2829 30
time (minutes)
% change
in
carotid
flow
PLV10-rapid-proximal-PC
FC-77 dose
-40
-20
0
20
40
60
80
100
120
pre12345678910 1112 13 1415 16 17 1819 202122 23 24 2526 27 2829 30
time (minutes)
% change
in
carotid
flow
PLV10-rapid-proximal-PC
FC-77 dose
PLV10-rapid-proximal-PC
FC-77 dose
(e)
-40
-20
0
20
40
60
80
100
120
pre123456789101112131415161718192021222324252627282930
time (minutes)
% change
in
carotid
flow
PLV10-s low-tip-PC
FC-77 dose
-40
-20
0
20
40
60
80
100
120
pre123456789101112131415161718192021222324252627282930
time (minutes)
% change
in
carotid
flow
PLV10-s low-tip-PC
FC-77 dose
PLV10-slow-tip-PC
FC-77 dose
(f)

M. W. Davies et al. / Open Journal of Pediatrics 2 (2012) 197-213 204
-40
-20
0
20
40
60
80
100
120
pre123456789101112131415161718192021222324252627282930
time
(
minutes
)
% change
in
carotid
flow
PLV10-slow-proximal-PC-adjust PIP
FC-77 dose
-40
-20
0
20
40
60
80
100
120
pre123456789101112131415161718192021222324252627282930
time
(
minutes
)
% change
in
carotid
flow
PLV10-slow-proximal-PC-adjust PIP
FC-77 dose
PLV10-slow-proximal-PC-adjust PIP
FC-77 dose
(g)
Figure 2. Relative changes (% change from baseline) of carotid blood flow for each perfluorocarbon dosing strategy. All vertical
scales are identical for ease of comparison.
-20
0
20
40
60
80
100
pre123456789101112131415161718192021222324252627282930
time (minutes)
% change
in
LDF
CONTROL-PC
-20
0
20
40
60
80
100
pre123456789101112131415161718192021222324252627282930
time (minutes)
% change
in
LDF
CONTROL-PC
CONTROL-PC
(a)
-20
0
20
40
60
80
100
pre12345678910 11 12 1314 15 161718 19 2021 22 23 2425 26 2728 29 30
time (minutes)
% change
in
LDF
PLV20-slow-proximal-PC
FC-77 dose
-20
0
20
40
60
80
100
pre12345678910 11 12 1314 15 161718 19 2021 22 23 2425 26 2728 29 30
time (minutes)
% change
in
LDF
PLV20-slow-proximal-PC
-20
0
20
40
60
80
100
pre12345678910 11 12 1314 15 161718 19 2021 22 23 2425 26 2728 29 30
time (minutes)
% change
in
LDF
PLV20-slow-proximal-PC
FC-77 dose
PLV20-slow-proximal-PC
FC-77 dose
(b)

M. W. Davies et al. / Open Journal of Pediatrics 2 (2012) 197-213 205
-20
0
20
40
60
80
100
pre123456789101112131415161718192021222324252627282930
time (minutes)
% change
in
LD F
PLV10-sl ow-pro ximal-PC
FC-77 dose
-20
0
20
40
60
80
100
pre123456789101112131415161718192021222324252627282930
time (minutes)
% change
in
LD F
PLV10-sl ow-pro ximal-PC
FC-77 dose
PLV10-slow-proximal-PC
FC-77 dose
(c)
-20
0
20
40
60
80
100
pre12 345 67 89101112131415161718192021222324252627282930
time (minutes)
% change
in
LDF
PLV 1 0-fas ter-pro ximal-P C
FC-77 dose
-20
0
20
40
60
80
100
pre12 345 67 89101112131415161718192021222324252627282930
time (minutes)
% change
in
LDF
PLV 1 0-fas ter-pro ximal-P C
FC-77 dose
PLV10-faster-proximal-PC
FC-77 dose
(d)
-20
0
20
40
60
80
100
pre12345678910 1112 13 1415 16 17 181920 2122 23 24 2526 27 2829 30
time (minutes)
% change
in
LD F
PLV10-ra pid-pro ximal-PC
FC-77 dose
-20
0
20
40
60
80
100
pre12345678910 1112 13 1415 16 17 181920 2122 23 24 2526 27 2829 30
time (minutes)
% change
in
LD F
PLV10-ra pid-pro ximal-PC
FC-77 dose
PLV10-rapid-proximal-PC
FC-77 dose
(e)

M. W. Davies et al. / Open Journal of Pediatrics 2 (2012) 197-213 206
-20
0
20
40
60
80
100
pre12 3 456 7 89101112131415161718192021222324252627282930
time (minutes)
% change
in
LD F
PLV 10-slow-tip-P C
FC-77 dose
-20
0
20
40
60
80
100
pre12 3 456 7 89101112131415161718192021222324252627282930
time (minutes)
% change
in
LD F
PLV 10-slow-tip-P C
FC-77 dose
PLV10-slow-tip-PC
FC-77 dose
(f)
-20
0
20
40
60
80
100
pre1 2 3 45 6 7 89101112131415161718192021222324252627282930
time
(
minutes
)
% change
in
LD F
PLV10-slow-proximal-PC-adjust PIP
FC-77 dose
-20
0
20
40
60
80
100
pre1 2 3 45 6 7 89101112131415161718192021222324252627282930
time
(
minutes
)
% change
in
LD F
PLV10-slow-proximal-PC-adjust PIP
FC-77 dose
PLV10-slow-proximal-PC-adjust PIP
FC-77 dose
(g)
Figure 3. Relative changes (% change from baseline) of cortical cerebral blood flow for each perfluorocarbon dosing strategy. All
vertical scales are identical for ease of comparison.
Table 1. The extent of the relative changes of carotid blood flow.
Group PFC dose duration % increase
at end of dose
Maximum %
increase (at time)
Minimum %
decrease (at time)
CONTROL-PC did not get PFC did not get PFC 4 (19 min) 4 (7 min)
PLV20-slow-proximal-PC 20 min 50 52 (27 min) 3 (2 min)
PLV10-slow-proximal-PC 20 min 95 105 (22 min) no decrease
PLV10-faster-proximal-PC 10 min 39 50 (12 min) no decrease
PLV10-rapid-proximal-PC 5 min 1 11 (11 min) 6 (3 min)
PLV10-slow-tip-PC 20 min 57 62 (17 min) no decrease
PLV10-slow-proximal-PC-adjust PIP20 min –10 no increase 14 (14 min)
PFC—perfluorocarbon.

M. W. Davies et al. / Open Journal of Pediatrics 2 (2012) 197-213 207
Table 2. The extent of the relative changes of cortical cerebral blood flow.
Group PFC dose duration % increase
at end of dose
Maximum %
increase (at time)
Minimum %
decrease (at time)
CONTROL-PC did not get PFC did not get PFC 17 (30 min) 1 (1 min)
PLV20-slow-proximal-PC 20 min 46 51 (26 min) 1 (2 min)
PLV10-slow-proximal-PC 20 min 24 29 (24 min) no decrease
PLV10-faster-proximal-PC 10 min 14 18 (11 min) 1 (1 min)
PLV10-rapid-proximal-PC 5 min 6 14 (7 min) 3 (29 min)
PLV10-slow-tip-PC 20 min 47 47 (20 min) no decrease
PLV10-slow-proximal-PC-adjust PIP 20 min –2 2 (2 min) 7 (23 min)
carotid blood flow was around 50% in the PLV20-slow-
proximal-PC (Figures 2(a) and (b)). In the CON-
TROL-PC dosing event group the carotid blood flow
remained with 5% of baseline throughout the 30 min.
A similar increase in cortical cerebral blood flow was
seen in the PLV20-slow-proximal-PC group (Figure
3(b)). In the CONTROL-PC dosing event group the
laser Doppler flow increased slightly up to a maximum
increase of around 18% whereas in the PLV20-slow-
proximal-PC group the laser Doppler flow increased to
a maximum of about 50%. The differences between
groups was not statistically significant—Mann-Whitney
test p = 0.31 at the end of the dose and p = 0.24 at the
time of maximum difference (26 min).
During perfluorocarbon dosing there was more vari-
ability in carotid blood flow but this difference did not
reach statistical significance—Mann-Whitney test p =
0.065 at the end of the dose and Student’s t test
Mann-Whitney test p = 0.18 at the time of maximum
difference (30 min). There was significantly more vari-
ability in cortical cerebral blood flow during perfluoro-
carbon dosing—Mann-Whitney test p = 0.0043 at the
end of the dose and Mann-Whitney test p = 0.041 at the
time of maximum difference (25 min). See additional
data file, pp. 3-5.
See additional data file (pp. 6-17) for the results of
haemodynamics, vital signs and respiratory function.
3.2. Does a Lower Dose of Perfluorocarbon
Disturb Cerebral Blood Flow?
We compared the PLV10-slow-proximal-PC group with
the PLV20-slow-proximal-PC group.
Carotid blood flow increased over the 30 minute ob-
servation period in both groups (Figures 2(b) and (c)) but
there was no difference between groups—Mann-Whitney
test p = 0.82 at the end of the dose and Mann-Whitney
test p = 0.82 at the time of maximum difference (24 min).
The highest mean increase in carotid blood flow was
around 105% in the PLV10-slow-proximal-PC group
and around 50% in the PLV20-slow-proximal-PC group.
Whilst cortical cerebral blood flow increased in both
groups (Figures 3(b) and (c)) there was no statistically
significant difference between groups—Student’s t test p
= 0.48 at the end of the dose and Mann-Whitney test p =
0.59 at the time of maximum difference (27 min).
During perfluorocarbon dosing there was considerable
variability in both carotid blood flow and cortical cere-
bral blood flow (see additional data file, pp. 18-20) but
there were no statistically significant differences between
groups. For the variability of carotid flow the Mann-
Whitney test p = 0.065 at the end of the dose and Stu-
dent’s t test p = 0.058 at the time of maximum difference
(3 min). For variability of cortical cerebral blood flow
the Student’s t test p = 0.48 at the end of the dose and
Mann-Whitney test p = 0.59 at the time of maximum
difference (27 min).
See additional data file (pp. 21-32) for the results of
haemodynamics, vital signs and respiratory function.
3.3. Do Different Rates of Administration Affect
Cerebral Blood Flow?
We compared the PLV10-slow-proximal-PC group, the
PLV10-faster-proximal-PC group and the PLV10-rapid-
proximal-PC group.
Carotid blood flow increased during the initial dose of
10 mL/kg perfluorocarbon when starting partial liquid
ventilation in both the PLV10-slow-proximal-PC and
PLV10-faster-proximal-PC groups (Figures 2(c) and (d)).
In the PLV10-rapid-proximal-PC group the increase in
carotid blood flow was much less pronounced (Figure
2(e)). The extent of the changes of carotid blood flow are
shown in Table 1. The relative increase in carotid blood
flow was greatest when the dose of perfluorocarbon was
given slowly and least when the dose was given rapidly.
However, there were no statistically significant dif-
ferences between groups for carotid blood flow—
Kruskal-Wallis test p = 0.26 at the end of the dose and
Kruskal-Wallis test p = 0.29 at the time of maximum
difference (23 min).
Cortical cerebral blood flow increased in all three
groups (Figures 3(c)-(e)). The pattern of changes were
different between all three groups and the extent of the

M. W. Davies et al. / Open Journal of Pediatrics 2 (2012) 197-213 208
changes of cortical cerebral blood flow are shown in Ta-
ble 2. The relative increase in cortical cerebral blood
flow was greatest when the dose of perfluorocarbon was
given slowly (Figure 3(c)) and least when the dose was
given rapidly (Figure 3(e)). Importantly when the dose
was given rapidly (over 5 minutes) the cortical cerebral
blood flow increased quickly but decreased back to
around baseline by about 12 minutes. When the dose of
perfluorocarbon is given over 10 or 20 minutes cortical
cerebral blood flow remained increased throughout the
30 minute observation period. However, there were no
statistically significant differences between groups—
one-way ANOVA p = 0.40 at the end of the dose and
one-way ANOVA p = 0.12 at the time of maximum dif-
ference (25 min).
During perfluorocarbon dosing there was a statistically
significant difference between groups for variability in
carotid blood flow—Kruskal-Wallis test p = 0.16 at the
end of the dose and one-way ANOVA p = 0.02 at the
time of maximum difference (18 min). There was no
statistically significant difference between groups for
variability in cortical cerebral blood flow—Kruskal-
Wallis test p = 0.22 at the end of the dose and Kruskal-
Wallis test p = 0.87 at the time of maximum difference
(1 min). See additional data file, pp. 33-35.
See additional data file (pp. 36-47) for the results of
haemodynamics, vital signs and respiratory function.
3.4. Does the Portal of Entry of the Dose of
Perfluorocarbon Affect Cerebral Blood
Flow?
We compared the PLV10-slow-proximal-PC group with
the PLV10-slow-tip-PC group. The PLV10-slow-pro-
ximal-PC group had a dose of FC-77 (10 mL/kg) given
over 20 min via the wye-piece at the proximal end of the
ETT (Figure 1) whilst on pressure-controlled ventilation.
The PLV10-slow-tip-PC group had 10 mL/kg of FC-77
given over 20 min via the ETT tip lumen (Figure 1)
whilst on pressure-controlled ventilation.
Carotid blood flow increased over the 30 minute ob-
servation period in both groups (Figures 2(c) and (f)).
The differences were not statistically significant—Mann-
Whitney test p = 0.39 at the end of the dose and Mann-
Whitney test p = 0.48 at the time of maximum difference
(24 min). The highest mean increase in carotid blood
flow was around 110% in the PLV10-slow-proximal-
PC group and around 60% in the PLV10-slow-tip-PC
group.
Cortical cerebral blood flow increased in both groups
(Figures 3(c) and (f)). The differences were not statistic-
cally significant—Mann-Whitney test p = 0.94 at the end
of the dose and Student’s t test Mann-Whitney test p =
0.94 at the time of maximum difference (15 min). The
highest mean increase in cortical cerebral blood flow was
around 47% in the PLV10-slow-tip-PC group and
around 29% in the PLV10-slow-proximal-PC group.
During perfluorocarbon dosing there were no statistic-
cally significant difference between the two groups for
carotid blood flow variability—Mann-Whitney test p =
0.24 at the end of the dose and Mann-Whitney test p =
0.13 at the time of maximum difference (3 min). There
was a little more variability in cortical cerebral blood
flow in the PLV10-slow-proximal-PC group. The dif-
ferences were not statistically significant—Mann-Whitney
test p = 0.80 at the end of the dose and Mann-Whitney
test p = 0.093 at the time of maximum difference (2 min).
See additional data file (pp. 48-50).
See additional data file (pp. 51-62) for the results of
haemodynamics, vital signs and respiratory function.
3.5. Does Tight Control of Arterial Carbon
Dioxide during Perfluorocarbon Mitigate
the Effects on Cerebral Blood Flow?
We compared the PLV10-slow-proximal-PC group with
the PLV10-slow-proximal-PC-adjust PIP group.
The relative change in carotid blood flow (Figure 2(c))
showed a significant increase in the PLV10-slow-
proximal-PC group compared with the PLV10-slow-
proximal-PC-adjust PIP group which had a slight de-
crease (Figure 2(g)). These differences were statistically
significant—Mann-Whitney test p = 0.015 at the end of
the dose and Mann-Whitney test p = 0.015 at the time of
maximum difference (24 min). The highest increase in
mean carotid blood flow was around 105% in the
PLV10-slow-proximal-PC group and the lowest de-
crease was around 14% in the PLV10-slow-proximal-
PC-adjust PIP group.
Cortical cerebral blood flow showed a significant in-
crease in the PLV10-slow-proximal-PC group (Figure
3(c)) compared with the PLV10-slow-proximal-PC-
adjust PIP group (Figure 3(g)) which had a slight de-
crease - Mann-Whitney test p = 0.041 at the end of the
dose and Mann-Whitney test p = 0.065 at the time of
maximum difference (23 min). The highest increase in
mean cortical cerebral blood flow was around 29% in the
PLV10-slow-proximal-PC group and the lowest de-
crease was around 7% in the PLV10-slow-proximal-
PC-adjust PIP group.
During perfluorocarbon dosing there was less variabil-
ity in carotid blood flow in the PLV10-slow-proximal-
PC-adjust PIP group—Student’s t test p = 0.019 at the
end of the dose and Mann-Whitney test p = 0.065 at the
time of maximum difference (5 min). There was no dif-
ference between groups for the variability of cortical
cerebral blood flow—Student’s t test p = 0.94 at the end
of the dose and Mann-Whitney test p = 0.39 at the time
of maximum difference (2 min). See additional data file
(pp. 63-65).

M. W. Davies et al. / Open Journal of Pediatrics 2 (2012) 197-213 209
See additional data file (pp. 66-77) for the results of
haemodynamics, vital signs and respiratory function.
4. DISCUSSION
4.1. Summary of Findings from the Current
Study
Right carotid blood flow was increased during tracheal
perfluorocarbon dosing with 20 mL/kg of FC-77 at the
start of partial liquid ventilation in rabbits. Similar,
non-statistically significant effects were seen in cortical
cerebral blood flow. The variability in cortical cerebral
blood flow was increased. Similar, almost statistically
significant effects were seen in carotid blood flow vari-
ability. There was also a decrease in pH (mirrored by
changes in PaCO2).
Cortical cerebral blood flow and right carotid blood
flow were not different when using 10 mL/kg of per-
fluorocarbon compared with 20 mL/kg. There were no
significant differences in variability of cortical cerebral
blood flow and right carotid blood flow.
Cerebral blood flow did not differ with the rate of ad-
ministration of 10 mL/kg of perfluorocarbon when start-
ing partial liquid ventilation. The variability in carotid
blood flow was increased when the dose is given rapidly.
The effect on cerebral blood flow variability would seem
to be of most clinical importance in the first minutes of
rapid dosing.
Changes in cerebral blood flow are similar regardless
of the portal of entry of the perfluorocarbon dose, i.e., at
the wye-piece at the proximal end of the endotracheal
tube or the tip of the endotracheal tube.
If the peak inspiratory pressure was adjusted during
perfluorocarbon dosing, to maintain the arterial carbon
dioxide between 45 and 50 mmHg, then there was no
increase in cerebral blood flow and there was less vari-
ability in carotid blood flow.
4.2. Interpretation of the Results
Despite the lack of statistically significant between-
group differences in PaCO2 there was a consistent asso-
ciation between increased PaCO2 and increased cerebral
blood flow. In five of the dosing strategies used the
PaCO2 increased during perfluorocarbon dosing, and in
four of those five the tidal volumes decreased. The effect
on tidal volumes, and hence minute ventilation, is not
unexpected in animal models where there is no lung dis-
ease, particularly when no adjustment is made in peak
inspiratory pressures. Partial liquid ventilation is known
to decrease lung compliance in normal lungs, and it also
partially blocks endotracheal tubes during perfluorocar-
bon instillation, increasing the airway resistance [19-21].
Both lead to decreased tidal volumes and/or increased
tidal volumes to maintain PaCO2 [22]. Also PaO2 is
known to be lower during partial liquid ventilation in
animals with normal lungs [15,23]. The most compelling
data that it is simply a decrease in tidal volume that
caused the rise in PaCO2 is that during the dosing strat-
egy where the PaCO2 was closely targeting by adjusting
peak inspiratory pressure the tidal volumes were not de-
creased.
We have no evidence from the current study to suggest
that any other factors were operating during the per-
fluorocarbon dosing episodes in rabbits to increase cere-
bral blood flow. Using adult animals excludes a change
in shunt across a patent ductus arteriosus as a cause for
the increase in cerebral blood flow. Similarly none of the
rabbits had any hypoxic-ischaemic events making any
post-hypoxic-ischaemic hyperperfusion very unlikely
during any of the dosing episodes. None of the rabbits
were hyperthermic at any stage during any of the dosing
episodes; also during those dosing strategies where there
was an increase in cerebral blood flow there was no cor-
responding increase in temperature.
There have been other published reports of studies in-
vestigating the effect of partial liquid ventilation on
cerebral blood flow [8,9,22,24-26]. Two of the reports
are our previous studies in preterm lambs [8,9]. Only one
of the other studies reported measurements during per-
fluorocarbon dosing but in this study there was no con-
trol group for comparison [24]. The studies are discussed
below.
4.3. Comparison with Other Studies
Burkhardt et al. [24] report the only other study that has
investigated the effects of perfluorocarbon dosing on
cerebral concentration of total and oxygenated haemo-
globin using near infrared spectroscopy. They used
newborn piglets of around 730 g and <24 hours old
without lung disease. They were randomised to one of
three different dosing regimens of PF5080:
 10 mL/kg given over about six minutes, target filling
rate of 1.5 mL/min [~2.1 mL/kg/min]; or
 30 mL/kg given over about 16 minutes, target filling
rate of 1.5 mL/min [~2.1 mL/kg/min]; or
 30 mL/kg given over about 41 seconds, target filling
over 30 - 45 s [~40 - 60 mL/kg/min].
The cerebral concentration of oxygenated haemoglo-
bin decreased initially (at 1 minute) when 30 mL/kg was
given rapidly, with return to baseline by 3 min. There
was no decrease with 30 mL/kg given slowly nor when
10 mL/kg was administered. At 5 min and thereafter the
cerebral concentration of oxygenated haemoglobin in-
creased in both 30 mL/kg groups but not in the 10 mL/kg
group. In both 30 mL/kg groups the total cerebral hae-
moglobin concentration increased compared with the 10
mL/kg group. Similarly in both 30 mL/kg groups the
PaCO2 increased and PaO2 decreased but not in the 10

M. W. Davies et al. / Open Journal of Pediatrics 2 (2012) 197-213 210
mL/kg group.
There was no control group, therefore conclusions
cannot be drawn about the effect of perfluorocarbon
dosing compared with gas ventilation alone. The study
design makes direct comparisons between the two vol-
umes of perfluorocarbon given difficult because the
doses were also given over different durations. Interpre-
tation is also difficult as they did not measure cerebral
blood flow directly, did not record and report real-time
haemodynamic variables during the perfluorocarbon liq-
uid dosing, and only reported changes in cerebral oxy-
genation at discrete time points.
Burkhardt et al. [24] concluded that giving perfluoro-
carbon extremely rapidly (20 - 30 times faster than the
rabbits in our “rapid” group) is not recommended because
of the decrease in cerebral blood flow (as determined by
total cerebral haemoglobin concentration) during per-
fluorocarbon dosing.
The three different rates of administration in our rabbit
studies were:
 slow—10 mL/kg over 20 min—0.5 mL/kg/min;
 faster—10 mL/kg over 10 min—1.0 mL/kg/min;
 rapid—10 mL/kg over 5 min—2.0 mL/kg/min.
The volume and rate of administration in our “rapid”
group in rabbits were almost identical to the volume and
rate in the Burkhardt [24] 10 mL/kg group and similar
results were seen: they found no changes in PaCO2 or
total cerebral haemoglobin concentration. We found a
slight increase (maximum 14%) in cerebral blood flow
during dosing with a much lower increase in cerebral
blood flow compared with the slower administration
rates.
Our previous studies [8,9] investigated the effect of
perfluorocarbon dosing in preterm lambs. We showed
that an intratracheal loading dose of 30 mL/kg of per-
fluorocarbon liquid, instilled over 20 min, caused an in-
crease in cortical cerebral blood flow (measured con-
tinuously with laser Doppler) during the administration
of perfluorocarbon and immediately thereafter. The
highest mean increase in the partial liquid ventilation
group was 27%. There was a slightly lower mean arterial
blood pressure in the partial liquid ventilation group; the
heart rate did not differ between groups. The PaCO2 was
marginally higher in the partial liquid ventilation group
at 30 min—the difference between groups was not statis-
tically significant. These lambs who received 30 mL/kg
were compared with lambs that received either 20 or 40
mL/kg. Cortical cerebral blood flow increased over time
in all three groups but there was no difference between
groups. There was no statistically significant difference
in PaCO2 between groups, but this was only measured at
baseline and at 30 minutes.
Therefore, in preterm lambs there was a consistent ef-
fect of perfluorocarbon dosing causing increased cerebral
blood flow. There did not seem to be any haemodynamic
cause for the increased cerebral blood flow but we could
not rule out an increase in PaCO2. Our rabbit studies
have demonstrated similar increases in cerebral blood
flow with perfluorocarbon dosing and associated in-
creases in PaCO2. It is also possible that in a lamb with a
left-to-right shunt across a patent ductus arteriosus (i.e.,
with decreased cerebral blood flow) that an increase in
pulmonary vascular resistance could decrease the left-
to-right shunt and cause a relative increase in cerebral
blood flow. Our rabbit model has ruled out an effect
from a patent ductus arteriosus.
4.4. Choice of Rabbit Model
In order to study the effects of many different dosing
strategies a simple and reliable model of the effects of
perfluorocarbon administration on cerebral blood flow
was needed. The use of adult New Zealand white rabbits
provides an animal model with the following advantages.
They are in the neonatal size-range (similar to the pre-
term lambs previously studied). The procedures involved
are readily performed in rabbits of this size and they are
familiar to the investigators. Using rabbits of this size,
with no lung disease, gives a far more stable animal that
allows repeated study over many hours, as they are ro-
bust enough to tolerate multiple dosing procedures.
There is no requirement to do experiments on particular
days as with date-mated ewes and their lambs. The con-
founding factors of lung disease and/or patent ductus
arteriosus are eliminated. There is no perinatal asphyxia.
Adult New Zealand white rabbits allowed the use of an
animal in which we could measure continuous cerebral
blood flow—laser Doppler as a reliable measure of cere-
bral blood flow has been validated in adult New Zealand
white rabbits. They also allowed the use two different
methods of measuring continuous cerebral blood flow—
carotid flow and cortical cerebral blood flow.
The different dosing strategies were studied in multi-
ple sequences in single animals: others have also used
this approach. Individual rabbits were not allocated dos-
ing strategies: each dosing event was randomly allocated
a different dosing strategy. This allowed an absolute
minimum number of animals to be used. The PLV10-
slow-proximal-PC group served as the reference strat-
egy for perfluorocarbon dosing. All the dosing strategies
experimental runs were done contemporaneously using
identical methods in the same laboratory with the same
personnel. As best we could be assured at the time only
the dosing strategy differed significantly for any given
run. Differences between dosing strategy groups at base-
line were controlled for with the statistical analysis used
with the data converted to “change from baseline”. All
dosing strategies were randomly allocated and allocation
was concealed until just before each episode of per-

M. W. Davies et al. / Open Journal of Pediatrics 2 (2012) 197-213 211
fluorocarbon dosing. A post hoc analysis of the distribu-
tion of individual rabbits within each dosing strategy
group showed no statistically significant difference
compared with that expected by chance for the 11 rabbits
(Chi squared test, p = 0.98). Similarly the distribution of
the order of the dosing events for each rabbit within each
dosing strategy group showed no statistically significant
difference compared with that expected by chance (Chi
squared test, p = 0.32). Furthermore the oxygen indices
were remarkably similar between all comparison dosing
event groups with low average oxygen indices ranging
from 1.5 to 2.0. This is consistent with minimal lung
disease across all dosing strategy groups for all com-
parisons.
4.5. Implications
From the results of the current study no direct conclu-
sions about the effects in preterms can be made. How-
ever, given that the effect of perfluorocarbon dosing is
the same in rabbits as that which we found in preterm
lambs [8,9] it is likely that the effects seen with different
dosing strategies will also been in preterm lambs.
Our studies were born out of a desire to ensure that if
partial liquid ventilation was to be used in extremely
preterm infants that we should first know whether the
instillation of perfluorocarbon disturbs cerebral blood
flow which would put the infants at risk of brain injury.
If cerebral blood flow is disturbed then it is necessary to
determine if the dose of perfluorocarbon can be given
safely without any cerebral blood flow disturbance.
Given the potential for harm associated with perfluoro-
carbon dosing in the most vulnerable of preterm infants it
is reassuring to know that there are ways of giving per-
fluorocarbon that can mitigate any increase in cerebral
blood flow. This can be achieved best by either monitor-
ing PaCO2 continuously during perfluorocarbon dosing
and keeping it under tight control; not giving the dose
over 5 minutes or faster; and not using the secondary
lumen of an endotracheal tube.
If partial liquid ventilation is to be used, then the best
protection for the extremely preterm brain is to start it
safely without disturbing cerebral blood flow. Data from
these studies will enable optimisation of the dosing
method to be used. Given the availability of a suitable
product, randomised controlled trials of partial liquid
ventilation in neonates are both feasible and desirable
[27,28]. Now the design of such trials can incorporate the
findings of this study to make the dosing procedure as
safe as it can be for use in extremely preterm infants.
5. CONCLUSIONS
The following effects were seen during perfluorocarbon
dosing at the start of partial liquid ventilation in rabbits:
 Compared with conventional gas ventilation right
carotid blood flow increased during perfluorocarbon
dosing with 20 mL/kg. Similar, non-statistically
significant effects were seen in cortical cerebral
blood flow. There was also increased variability in
cortical cerebral blood flow and similar, almost
statistically significant, effects were seen in carotid
blood flow variability. The effects on cerebral blood
flow were similar to that seen in preterm lambs with
lung disease.
 Cortical cerebral blood flow and right carotid blood
flow were not different when using 10 mL/kg of
perfluorocarbon compared with 20 mL/kg. There were
no significant differences in variability of cortical
cerebral blood flow and right carotid blood flow.
The lack of any significant differences between dose
volume groups on cortical cerebral blood flow was
similar to that seen in preterm lambs.
 Cerebral blood flow did not differ with the rate of
administration of 10 mL/kg of perfluorocarbon when
starting partial liquid ventilation. The variability in
carotid blood flow was increased, but this did not
seem to be a consistent or clinically important differ-
ence.
 The increase in cortical cerebral blood flow and
right carotid flow (and their variability) were similar
when the dose was given at the endotracheal tube tip
(via a secondary lumen in the wall of the endo-
tracheal tube) compared with doses given more
proximally.
 If the peak inspiratory pressure was adjusted during
perfluorocarbon dosing to maintain the arterial carbon
dioxide between 45 and 50 mmHg then there was no
increase in cerebral blood flow and there was less
variability in carotid blood flow.
6. ACKNOWLEDGEMENTS
We gratefully acknowledge financial support from The Prince Charles
Hospital Foundation, the Royal Women’s Hospital Paul Weedon
Bursary, and the Royal Women’s Hospital Auxiliary.
Many thanks to Kathy Wilson who contributed significantly to the
work in the animal laboratory. Thanks also to Michael Lindeberg and
Bill Sommers who helped out in the animal laboratory. Thanks to Peter
O’Rourke for his statistical advice.
REFERENCES
[1] Costeloe, K., Hennessy, E., Gibson, A.T., Marlow, N.
and Wilkinson, A.R. (2000) The EPICure study: Out-
comes to discharge from hospital for infants born at the
threshold of viability. Pediatrics, 106, 659-671.
doi:10.1542/peds.106.4.659
[2] Hennessy, E.M., Bracewell, M.A., Wood, N., Wolke, D.,
Costeloe, K., Gibson, A. and Marlow, N. for the EPICure

M. W. Davies et al. / Open Journal of Pediatrics 2 (2012) 197-213 212
Study Group (2008) Respiratory health in pre-school and
school age children following extremely preterm birth.
Archives of Diseases in Childhood, 93, 1037-1043.
doi:10.1136/adc.2008.140830
[3] Leach, C.L., Greenspan, J.S., Rubenstein, S.D., Shaffer,
T.H., Wolfson, M.R., Jackson, J.C., DeLemos, R. and
Fuhrman, B.P. for the LiquiVent Study Group (1996)
Partial liquid ventilation with perflubron in premature in-
fants with severe respiratory distress syndrome. The New
England Journal of Medicine, 335, 761-767.
doi:10.1056/NEJM199609123351101
[4] Blassnig, N., Dietl, S., Tschirch, E., Burkhardt, W., Wem-
höner, A. and Rüdiger, M. (2009) Intratracheal applica-
tion of PFC diminishes the hyperoxia-mediated impair-
ment of postnatal lung development. Acta Paediatrica, 98,
7.
[5] Vohr, B. and Ment, L.R. (1996) Intraventricular hemor-
rhage in the preterm infant. Early Human Development,
44, 1-16. doi:10.1016/0378-3782(95)01692-9
[6] Ballabh, P. (2010) Intraventricular hemorrhage in prema-
ture infants: mechanism of disease. Pediatric Research, 67,
1-8. doi:10.1203/PDR.0b013e3181c1b176
[7] Perlman, J.M. (2009) The relationship between systemic
hemodynamic perturbations and periventricular-intraven-
tricular hemorrhage—A historical perspective. Seminars
in Pediatric Neurology, 16, 191-199.
doi:10.1016/j.spen.2009.09.006
[8] Davies, M.W., Dunster, K.R., Wilson, K. and Colditz,
P.B. (2010) Perfluorocarbon dosing when starting partial
liquid ventilation: Haemodynamics and cerebral blood
flow in preterm lambs. Neonatology, 97, 144-153.
doi:10.1159/000239768
[9] Davies, M.W., Dunster, K.R., Wilson, K. and Colditz,
P.B. (2010) The effect of the dose volume of perfluoro-
carbon when starting partial liquid ventilation. Journal of
Paediatrics and Child Health, 46, 714-722.
doi:10 .1111/j.144 0-1754.2010.01828.x
[10] Dunster, K.R. and Davies, M.W. (2004) A novel mount-
ing device to attach intracranial probes to the skull for use
in experimental research models. Physiological Meas-
urement, 25, N11-N14. doi:10.1088/0967-3334/25/2/N01
[11] Tutuncu, A.S., Faithfull, N.S. and Lachmann, B. (1993)
Comparison of ventilatory support with intratracheal per-
fluorocarbon administration and conventional mechanical
ventilation in animals with acute respiratory failure. Ameri-
can Journal of Respiratory and Critical Care Medicine,
148, 785-792. doi:10.1164/ajrccm/148.3.785
[12] Lim, C.M., Domino, K.B., Glenny, R.W. and Hlastala,
M.P. (2001) Effect of increasing perfluorocarbon dose on
VA/Q distribution during partial liquid ventilation in
acute lung injury. Anesthesiology, 94, 637-642.
doi:10.1097/00000542-200104000-00018
[13] Lim, C.-M., Yang, S.-H., Kang, J.-L. and Koh, Y. (2002)
Effect of ventilation mode on gas exchange during partial
liquid ventilation at different perfluorocarbon doses in
surfactant-depleted lung. Lung, 179, 245-255.
doi:10.1007/s004080000065
[14] Suh, G.Y., Chung, M.P., Park, S.J., Koh, Y., Kang, K.W.,
Kim, H., Han, J., Rhee, C.H. and Kwon, O.J. (2000) Par-
tial liquid ventilation shows dose-dependent increase in
oxygenation with PEEP and decreases lung injury associ-
ated with mechanical ventilation. Journal of Critical Care,
15, 103-112. doi:10.1053/jcrc.2000.16463
[15] Fessler, H.E. and Pearse, D. (2000) Accuracy of hemo-
dynamic measurements during partial liquid ventilation
with perflubron. American Journal of Respiratory and
Critical Care Medicine, 162, 1372-1376.
[16] Davies, M.W. and Dunster, K.R. (2000) The effect of
perfluorocarbon vapour on the measurement of respire-
tory tidal volume during partial liquid ventilation. Physio-
logical Measurement, 21, N23-N30.
doi:10.1088/0967-3334/21/3/402
[17] Takeuchi, T., Horiuchi, J. and Terada, N. (1989) Central
vasomotor control of the rabbit portal vein. Pflügers Ar-
chiv European Journal of Physiology, 413, 348-353.
doi:10.1007/BF00584482
[18] National Health and Medical Research Council (Australia)
(2004) Australian code of practice for the care and use of
animals for scientific purposes. 7th Edition, Australian
Government Publishing Service, Canberra.
[19] Weiswasser, J., Lueders, M. and Stolar, C.J.H. (1998)
Pressure- versus volume-cycled ventilation in liquid-ven-
tilated neonatal piglet lungs. Journal of Pediatric Surgery,
33, 1158-1162. doi:10.1016/S0022-3468(98)90551-6
[20] Hummler, H.D., Schulze, A., Pohlandt, F. and Thome, U.
(2000) Dynamics of breathing during partial liquid venti-
lation in spontaneously breathing rabbits supported by
elastic and resistive unloading. Pediatric Research, 47,
392-397. doi:10.1203/00006450-200003000-00018
[21] Shashikant, M.P., Badellino, M.M., Cooper, B., Shaffer,
T.H., Myers, S.I. and Wolfson, M.R. (2002) Physico-
chemical properties of perfluorochemical liquids influ-
ence ventilatory requirements, pulmonary mechanics, and
microvascular permeability during partial liquid ventila-
tion following intestinal ischemia/reperfusion injury.
Critical Care Medicine, 30, 2300-2305.
doi:10.1097/00003246-200210000-00019
[22] Burkhardt, W., Proquitté, H., Krause, S., Wauer, R.R. and
Rüdiger, M. (2004) Changes in FiO2 affect PaO2 with
minor alterations in cerebral concentration of oxygenated
hemoglobin during liquid ventilation in healthy piglets.
Intensive Care Medicine, 30, 315-320.
doi:10.1007/s00134-003-2090-7
[23] Hernan, L.J., Fuhrman, B.P., Papo, M.C., Steinhorn,
D.M., Leach, C.L., Salman, N., Paczan, P.R. and Kahn, B.
Cardiorespiratory effects of perfluorocarbon-associated
gas exchange at reduced oxygen concentrations. Critical
Care Medicine, 23, 553-559.
doi:10.1097/00003246-199503000-00022
[24] Burkhardt, W., Proquitté, H., Krause, S., Wauer, R.R. and
Rüdiger, M. (2002) Cerebral oxygenation is affected by
filling mode and perfluorochemical volume in partial liq-
uid ventilation of healthy piglets. Biology of the Neonate,
82, 250-256. doi:10.1159/000065886
[25] McLaughlin, G.E., Kulatunga, S., Kuluz, J.W., Gelman,
B. and Schleien, C.L. (2001) Cerebral blood flow during
partial liquid ventilation in surfactant deficient lungs un-
der varying ventilation strategies. Pediatric Critical Care

M. W. Davies et al. / Open Journal of Pediatrics 2 (2012) 197-213 213
Medicine, 2, 88-92.
doi:10.1097/00130478-200101000-00017
[26] Dimmitt, R.A., Beckman, S.A., Halamek, L.P., Moss,
R.L., Mickas, N.A., Falco, D.A., Chubb, C. and Skars-
gard, E.D. (2002) Effects of partial liquid ventilation on
cerebral blood flow and cerebral metabolism in neonatal
lambs. Journal of Pediatric Surgery, 37, 840-844.
doi:10.1053/jpsu.2002.32884
[27] Wolfson, M.R. and Shaffer, T.H. (2005) Pulmonary ap-
plications of perfluorochemical liquids: Ventilation and
beyond. Paediatric Respiratory Reviews, 6, 117-127.
doi:10.1016/j.prrv.2005.03.010
[28] Greenough, A. (2006) High frequency oscillation and liq-
uid ventilation. Paediatric Respiratory Reviews, 7, S186-
S188. doi:10.1016/j.prrv.2006.04.207
Abbreviations
ANOVA: Analysis of variance;
Cdyn: Dynamic compliance;
CI: Confidence interval;
CV: Coefficient of variation;
ETT: Endotracheal tube;
FiO2: Fraction of inspired oxygen;
LDF: Laser Doppler flow;
MuLETT: Multi-lumen endotracheal tube;
PaCO2: Arterial carbon dioxide tension;
PaO2: Arterial oxygen tension;
PC: Pressure-controlled;
PEEP: Positive end expiratory pressure;
PFC: Perfluorocarbon;
PIP: Peak inspiratory pressure;
PLV: Partial liquid ventilation.