J Appl Physiol 115: 1433–1442, 2013.
First published September 5, 2013; doi:10.1152/japplphysiol.00238.2013.
Cerebrovascular autoregulation after rewarming from hypothermia in a neonatal
swine model of asphyxic brain injury
Abby C. Larson,* Jessica L. Jamrogowicz,* Ewa Kulikowicz, Bing Wang, Zeng-Jin Yang,
Donald H. Shaffner, Raymond C. Koehler, and Jennifer K. Lee
Department of Anesthesiology and Critical Care Medicine, Johns Hopkins University School of Medicine, Baltimore, Maryland
Submitted 22 February 2013; accepted in final form 3 September 2013
Larson AC, Jamrogowicz JL, Kulikowicz E, Wang B, Yang Z-J,
Shaffner DH, Koehler RC, Lee JK. Cerebrovascular autoregulation
after rewarming from hypothermia in a neonatal swine model of asphyxic
brain injury. J Appl Physiol 115: 1433–1442, 2013. First published
September 5, 2013; doi:10.1152/japplphysiol.00238.2013.—After hy-
poxic brain injury, maintaining blood pressure within the limits of
cerebral blood flow autoregulation is critical to preventing secondary
brain injury. Little is known about the effects of prolonged hypother-
mia or rewarming on autoregulation after cardiac arrest. We hypoth-
esized that rewarming would shift the lower limit of autoregulation
(LLA), that this shift would be detected by indices derived from
near-infrared spectroscopy (NIRS), and that rewarming would impair
autoregulation during hypertension. Anesthetized neonatal swine un-
derwent sham surgery or hypoxic-asphyxic cardiac arrest, followed by
2 h of normothermia and 20 h of hypothermia, with or without
rewarming. Piglets were further divided into cohorts for cortical
laser-Doppler flow (LDF) measurements during induced hypotension
or hypertension. We also tested whether indices derived from NIRS
could identify the LDF-derived LLA. The LLA did not differ signif-
icantly among groups with sham surgery and hypothermia (29 ⫾ 8
mmHg), sham surgery and rewarming (34 ⫾ 7 mmHg), arrest and
hypothermia (29 ⫾ 10 mmHg), and arrest and rewarming (38 ⫾ 11
mmHg). The LLA was not affected by arrest (P ⫽ 0.60), temperature
(P ⫽ 0.08), or interaction between arrest and temperature (P ⫽ 0.73).
The NIRS-derived indices detected the LLA accurately, with the area
under the receiver-operator characteristic curves of 0.81– 0.96 among
groups. In groups subjected to arrest and hypothermia, with or without
rewarming, the slope of LDF relative to cerebral perfusion pressure
during hypertension was not significantly different from zero (P ⬎
0.10). In conclusion, rewarming did not shift the LLA during hypo-
tension or affect autoregulation during hypertension after asphyxic
cardiac arrest. The NIRS-derived autoregulation indices identified the
LLA accurately.
cardiac arrest; cerebral blood flow; near-infrared spectroscopy; neo-
natal hypoxic-ischemic encephalopathy; pig
IN THE UNITED STATES, ⬃16,000 children suffer a cardiac arrest an-
nually (30), and neonatal hypoxic-ischemic encephalopathy
(HIE) occurs in approximately three in 1,000 births (11).
Cerebrovascular autoregulation maintains relatively constant
cerebral blood flow (CBF) across fluctuations in cerebral per-
fusion pressure (CPP). The ensurance that blood pressure
remains within the limits of autoregulation after hypoxic brain
injury is critical to preventing secondary injury from pressure-
passive CBF. Several factors can affect the limits of cerebral
autoregulation after pediatric hypoxic-asphyxic (HA) injury.
* A. C. Larson and J. L. Jamrogowicz contributed equally as co-first authors.
Address for reprint requests and other correspondence: J. K. Lee, Dept. of
Anesthesiology and Critical Care Medicine, Johns Hopkins Hospital, Charlotte
R. Bloomberg Children’s Center, 1800 Orleans St., Rm. 6321, Baltimore, MD
21287 (e-mail: jklee@jhmi.edu).
http://www.jappl.org 8750-7587/13 Copyright © 2013 the American Physiological Society
For instance, intracranial hypertension shifted the lower limit
of autoregulation (LLA) to a higher CPP in a neonatal swine
model of hydrocephalus (4). Although therapeutic hypothermia
(27) is increasingly used to treat neonatal HIE (26), and clinical
trials about therapeutic hypothermia after pediatric cardiac
arrest are ongoing [including the Therapeutic Hypothermia
after Pediatric Cardiac Arrest Trials (www.thapca.org)], the
effects of hypothermia on cerebral autoregulation remain un-
clear. After HA cardiac arrest in neonatal swine, hypothermia
decreased the LLA measured 6 h after injury (16). In a rodent
model of traumatic brain injury, 1–2 h of hypothermia pre-
served the cerebral vasodilatory response to hypotension (10).
However, rapid rewarming from hypothermic cardiopulmo-
nary bypass in adults impaired cerebral autoregulation and was
associated with postoperative strokes (14). Moreover, an asso-
ciation between impaired cerebral autoregulation and mortality
has been reported in neonates with asphyxia (25). Whether
rewarming from hypothermia after pediatric HA brain injury
affects autoregulation remains unclear.
The ability to identify the limits of cerebral autoregulation
would assist clinicians in identifying hemodynamic parameters
that maintain pressure-reactive CBF. With the use of near-infrared
spectroscopy (NIRS), we developed the cerebral oximetry (COx)
and hemoglobin volume (HVx) indices (6, 16, 18). COx is
calculated by a correlation coefficient between tissue oxyhemo-
globin saturation and CPP. The theory behind COx assumes that
changes in CBF are proportional to changes in regional tissue
oxyhemoglobin saturation (rSO2). HVx is calculated by a corre-
lation coefficient between the NIRS-derived relative tissue hemo-
globin concentration (rTHb) and mean arterial pressure (MAP).
The theory behind HVx assumes that autoregulatory vasodilation/
vasoconstriction produce changes in cerebral blood volume that
are proportional to changes in rTHb. Positive correlations between
tissue oxyhemoglobin saturation and CPP for COx (5) or between
rTHb and MAP for HVx indicate diminished autoregulatory
responsivity (16 –18).
Whereas the NIRS-derived indices detect the LLA during
hypotension after HA injury and short periods of hypothermia
(16), it is unclear whether the indices will remain accurate after
long durations of hypothermia with rewarming, a scenario rele-
vant to the clinical application of therapeutic hypothermia (24,
28). After asphyxic injury, reoxygenation increases oxidative
stress (22). Whereas hypothermia may attenuate this response,
oxidative stress could resume during rewarming and disturb ce-
rebral autoregulation. Moreover, proinflammatory cytokine shifts
during rewarming (3) could affect vasoreactivity.
Here, we examined the effects of rapid rewarming from
hypothermia on cerebral autoregulation in a neonatal swine
model of HA brain injury. Because injury may continue to
evolve after the insult, we extended the duration of hypother-
1433
1434 Cerebral Autoregulation, Hypothermia, and Asphyxia
mia from 4 h in our previous work (16) to 20 h, to mimic more
closely therapeutic hypothermia after pediatric cardiac arrest or
neonatal HIE. Since rapid rewarming may negate some of the
protection provided by hypothermia, we compared measures of
cerebral autoregulation among sham or HA-injured groups that
received hypothermia alone with sham or HA-injured groups
that received hypothermia with rapid rewarming. Because the
focus of the current study was to compare rewarming with
sustained hypothermia, a normothermic cohort was not used, as
in our prior work (16). We hypothesized that hypothermia
would preserve the vasodilatory response to hypotension but
that rewarming would shift the LLA to a higher CPP. We also
postulated that HVx and COx would identify the LLA during
hypotension. Furthermore, we tested the hypothesis that re-
warming would impair cerebral autoregulation during hyper-
tension compared with sustained hypothermia.
MATERIALS AND METHODS
Animal preparation. All procedures were approved by the Animal
Care and Use Committee at Johns Hopkins University and complied
with the United States Public Health Service Policy on Humane Care
and Use of Laboratory Animals and the Guide for the Care and Use
of Laboratory Animals. Animal care was in accord with National
Institutes of Health Guidelines and ensured animal comfort. Neonatal
male piglets (3–5 days old, 1–2.5 kg) were randomized to undergo HA
injury or sham surgery and to remain hypothermic or undergo re-
warming until the goal sample size was achieved in each treatment
group (HA injury with hypothermia, HA injury with hypothermia and
rewarming, sham surgery with hypothermia, and sham surgery with
hypothermia and rewarming). As illustrated in Fig. 1, one-half of the
piglets in each group underwent cerebral autoregulation testing during
induced hypotension (n ⫽ 6/group), and the others underwent testing
during induced hypertension (n ⫽ 6/group). Anesthesia was induced
with 5% isoflurane with 50%/50% nitrous oxide/oxygen (N2O/O2).
After intubation, the lungs were mechanically ventilated to maintain
normocapnea. Anesthesia was maintained with 2% isoflurane with
70%/30% N2O/O2 during surgery. Femoral arterial and venous cath-
eters were inserted, and fentanyl was given (10 ␮g/kg iv). After
surgery, anesthesia was maintained with 0.4% isoflurane ⫹ 70%/30%
N2O/O2 ⫹ fentanyl (10 –20 ␮g·kg⫺1·h⫺1). Piglets received 5% dex-
trose in 0.45% saline at 10 ml/h iv.
HA cardiac arrest. The inspired O2 was decreased to 10% for 45
min, followed by 5 min of 21% O2 to reoxygenate the heart. This
reoxygenation period is required for cardiac resuscitation. The endo-
tracheal tube was then occluded for 7 min. Piglets were resuscitated
with 100% O2 to mimic common clinical practice, manual chest
Fig. 1. Study design flowchart. The 2nd surgery (cranial
surgery and placement of a balloon catheter in the
inferior vena cava for hypotension cohorts or descend-
ing aorta for hypertension cohorts) was performed 18 h
after resuscitation. 2 BP, induced hypotension blood
pressure; 1 BP, induced hypertension blood pressure.
J Appl Physiol • doi:10.1152/japplphysiol.00238.2013 • www.jappl.org
• Larson AC et al.
compressions, and epinephrine (100 ␮g/kg iv). Piglets without return
of spontaneous circulation (ROSC) within 3 min were excluded. After
resuscitation, inspired O2 was decreased to 30%. Sodium bicarbonate
and calcium chloride were administered to correct metabolic acidosis
and hypocalcemia. Sham-operated piglets received the same anesthe-
sia, surgery, and duration of anesthesia but without arrest.
Temperature. After resuscitation, piglets were maintained at a
rectal temperature of 38.0 –39.0°C (normothermic for swine) for 2 h
with warming blankets and heating lamps. The 2-h period of normo-
thermia was designed to mimic the delay in initiating hypothermia in
clinical practice. At 2 h after ROSC, temperature was reduced to 34°C
over ⬃30 min, as described (1) and as we used previously in our
cerebral autoregulation experiments with short periods of hypother-
mia (16). After 20 h of hypothermia, some piglets were rewarmed
with whole-body warm packs at a goal rate of 4°C/h. Piglets with or
without rewarming remained anesthetized and received muscle relax-
ants. To ensure that CPP remained above the LLA during anesthesia,
low-dose phenylephrine was infused as needed to maintain MAP ⱖ45
mmHg.
Surgical preparation for monitoring cerebral autoregulation. At
18 h after ROSC, a second surgery was performed. A 5-French
balloon catheter was placed in the inferior vena cava via a femoral
vein for later inflation to induce hypotension in piglets in the hypo-
tensive cohort. For piglets in the hypertensive cohort, a balloon
catheter was placed in the descending aorta via a femoral artery for
later inflation to induce cephalic hypertension. Three small cranial
drill holes were made for placement of: 1) a ventricular drain catheter
to measure intracranial pressure (ICP), 2) a 1-mm diameter laser-
Doppler flow (LDF) probe (model DRT4; Moor Instruments, Devon,
UK) in the frontoparietal cortex, 5 mm from the ICP catheter, and
3) a thermistor in the epidural space to monitor temperature. The LDF
probe was secured with rubber washers and cemented to the skull. A
NIRS probe (Covidien, Mansfield, MA; light-emitting diode to shal-
low and deep photodiode distances of 30 mm and 40 mm, respec-
tively) was placed over the frontoparietal cortex contralateral to the
craniotomy sites. The NIRS monitor measured rSO2 and rTHb con-
tinuously. After completion of cranial surgery, data were collected for
⬃3 h before either hypotension or hypertension was induced. To
obtain the baseline LDF, the average LDF during the 1st h of
monitoring across the range of CPP 50 – 60 mmHg was calculated.
Cerebral autoregulation monitoring during hypotension. MAP,
ICP, and LDF were sampled from an analog-to-digital converter at
100 Hz with ICM⫹ software (Cambridge University, Cambridge,
UK) (29). Consecutive averages of nonoverlapping, 10-s recordings
were obtained of MAP and ICP, and CPP was calculated as MAP-ICP
every 10 s. To determine the LLA, we decreased MAP gradually in a
continuous manner over 2–3 h by inflating the balloon catheter in the
inferior vena cava beginning at 22 h after ROSC in the hypothermia
 Cerebral Autoregulation, Hypothermia, and Asphyxia
groups or immediately upon reaching a brain temperature of 38.5°C in
the rewarming groups. LDF was plotted as a function of CPP. The
LLA was identified by dichotomizing the data such that the fit of two
linear regression lines resulted in the lowest combined error squared.
The CPP at the intersection of these two lines was defined as the LLA
(6, 16 –18). To calculate the inflection point from the rSO2 measure-
ments as blood pressure decreased, rSO2 was similarly plotted as a
function of CPP. The data were dichotomized to identify the inflection
point at the CPP where the fit of two linear regression lines had the
lowest combined error squared.
COx and HVx were calculated simultaneously. Cerebral rSO2 and
rTHb were sampled synchronously from the digital output of the
NIRS monitor, as described previously (17). COx was calculated from
a continuously moving 5-min window as the correlation coefficient
between rSO2 and CPP. HVx was similarly calculated from the
correlation of rTHb and MAP (6, 17). COx and HVx have a range of
⫺1 to ⫹1. When the indices are near zero or negative, CPP and rSO2
(for COx) and MAP and rTHb (for HVx) are either not correlated or
are negatively correlated, thereby indicating functional cerebral auto-
regulation. When the indices become positive and approach ⫹1, as
blood pressure and rSO2 or rTHb directly correlate, autoregulation is
less effective.
Monitoring during hypertension. At 22 h after ROSC in the
hypothermia groups without rewarming or immediately upon reaching
a brain temperature of 38.5°C in the rewarming groups, hypertension
was induced over 30 min by inflating the aortic balloon catheter and
infusing phenylephrine and dopamine. The autoregulatory response to
hypertension was evaluated by comparing the change in LDF (as a
percentage of baseline) with the change in CPP (as a percentage of
baseline).
Statistical analysis. Normally distributed data are presented as
means ⫾ SD or in some cases, where noted, as ⫾95% confidence
intervals (CI). Nonparametric data are presented as medians with
interquartile ranges (IQR) when appropriate. Differences were con-
sidered significant at P ⱕ 0.05. Physiologic variables, blood gas
variables, hemoglobin levels, and electrolytes were compared among
groups before and after HA arrest with two-way repeated-measures
ANOVA by treatment group and time with post hoc Newman-Keuls
multiple range test when the F value was significant (SigmaPlot
v11.0, Systat Software, Chicago, IL). Variables during hypoxia and
asphyxia were compared between arrested groups destined to remain
hypothermic or undergo rewarming with t-tests. Differences in weight
and in temperature at each time point were analyzed with one-way
ANOVA or the Kruskal-Wallis ANOVA on ranks when appropriate.
Two-way ANOVA was used to examine the effects of arrest and
temperature on the LLA. The values for the LDF-derived LLA were
compared with the inflection points calculated from the rSO2 vs. CPP
plot during induced hypotension using a Spearman rank correlation.
The effects of arrest and temperature on the rSO2-derived inflection
point were examined using two-way ANOVA.
Average measures of LDF, rSO2, HVx, and COx were binned into
5-mmHg increments of CPP for analysis. The effects of CPP and
treatment group on each measure of cerebral autoregulation (LDF,
HVx, and COx) were analyzed with repeated-measures two-way
ANOVA. Because the LLA differs among individual piglets, we also
analyzed the data using CPP in relation to each animal’s LLA.
Spline regression analyses were performed to evaluate the HVx and
COx responses to hypotension with STATA software v11.1 (Stat-
Corp, College Station, TX). Spline regressions examine slopes above
and below a “cut-point” within a continuous dataset. For this study,
the cut-point was set at each animal’s LLA. We used spline regres-
sions to examine change in ⌬HVx (or ⌬COx)/⌬CPP at levels of CPP
above and below the LLA. A difference between slopes at CPP above
and CPP below the LLA indicates that the index increased at a faster
rate as CPP decreased when CPP was below the LLA compared with
when CPP exceeded the LLA. Receiver operator characteristic (ROC)
curves were calculated to test the sensitivity and specificity of HVx
J Appl Physiol • doi:10.1152/japplphysiol.00238.2013 • www.jappl.org
• Larson AC et al. 1435
and COx in detecting CPP above or below the LLA (31). Both the
spline regression and ROC analyses accounted for multiple intra-
subject measurements. For the ROC curves, CPP was divided into 5
mmHg bins and coded as a binary variable above or below the bin
containing the LLA, and COx and HVx were analyzed as continuous
variables (⫺1 to ⫹1).
To examine cerebral autoregulation during hypertension, regres-
sion analysis was used to compare the percent change in LDF with the
percent change in CPP. The autoregulatory index was defined as the
slope of this relationship. A slope of zero reflects perfect autoregula-
tion, and a slope of one reflects a pressure-passive relationship with
blood flow. Slopes were compared among groups by the Kruskal-
Wallis test.
RESULTS
To examine cerebral autoregulation during hypotension, we
subjected 14 piglets to arrest and 13 piglets to sham surgery.
Two piglets in the arrest group could not be resuscitated,
resulting in a resuscitation rate of 86%. One sham piglet had an
intracranial injury during surgery. Two piglets in the rewarmed
arrest group, three in the hypothermia arrest group, and one in
the hypothermia sham group required phenylephrine during
anesthesia to maintain MAP ⱖ45 mmHg.
To evaluate cerebral autoregulation during hypertension, we
subjected 15 piglets to arrest and 18 piglets to sham surgery.
All piglets were resuscitated after arrest, but three had persis-
tent hypotension during anesthesia and did not survive the
protocol. One sham piglet had persistent hypotension and did
not survive the protocol; one had a cardiac arrest during the
craniotomy; two had intracranial injuries during the surgeries;
and the laser-Doppler probe moved in two piglets. Among
piglets included in the final analysis, one sham and one arrested
piglet in the rewarming groups required phenylephrine during
anesthesia. Data in the hypotensive and hypertensive cohorts
were analyzed on 48 piglets (24 arrests and 24 shams; n ⫽
6/group; Fig. 1).
Hypotensive cohorts. Body weight was similar among groups
[1.6 ⫾ 0.3 kg (SD) in the hypothermic arrest, 1.4 ⫾ 0.3 kg in the
hypothermic sham, 1.6 ⫾ 0.3 kg in the rewarmed arrest, and
1.4 ⫾ 0.2 kg in the sham rewarmed groups; P ⫽ 0.309]. CPP and
ICP were similar among groups during hypothermia (Table 1;
P ⫽ 0.072 for CPP; P ⫽ 0.608 for ICP). Body temperatures were
also similar before rewarming (Table 2; P ⬎ 0.10), with the
exception of a slightly higher temperature at 5 min of asphyxia in
the arrested piglets that were destined to remain hypothermic
(P ⫽ 0.026 for the difference in temperature at 5 min of asphyxia).
At 42 min of hypoxia, hemoglobin O2 saturation levels were 31 ⫾
7% in the hypothermic arrest and 33 ⫾ 8% in the rewarmed arrest
groups (P ⫽ 0.864). At 5 min of asphyxia, O2 saturation levels
were 7 ⫾ 4% in the hypothermic arrest and 8 ⫾ 5% in the
rewarmed arrest groups (P ⫽ 0.760). The severity of hypercapnea
Table 1. Physiologic variables of piglets in hypotensive
groups during hypothermia on day 1 after ROSC
Intracranial Pressure, Cerebral Perfusion Pressure,
Group, n ⫽ 6 mmHg mmHg
Hypothermic arrest 13 ⫾ 7 40 ⫾ 12
Hypothermic sham 10 ⫾ 4 45 ⫾ 5
Rewarm arrest 15 ⫾ 8 46 ⫾ 3
Rewarm sham 11 ⫾ 6 53 ⫾ 10
ROSC, return of spontaneous circulation.
1436 Cerebral Autoregulation, Hypothermia, and Asphyxia • Larson AC et al.

Table 2. Blood Gas, Hb Levels, and Temperatures of Piglets in Hypotensive Groups

Parameter* Group, n ⫽ 6 Baseline Hypoxia 42 min Asphyxia 5 min ROSC 1 h ROSC 3 h ROSC 1 Day†

pH HT arrest 7.35 ⫾ 0.12 7.25 ⫾ 0.03 6.80 ⫾ 0.09 7.20 ⫾ 0.20 7.38 ⫾ 0.04 7.31 ⫾ 0.16
RW arrest 7.45 ⫾ 0.08 7.30 ⫾ 0.14 6.90 ⫾ 0.13 7.36 ⫾ 0.13 7.51 ⫾ 0.10 7.35 ⫾ 0.09
HT sham 7.44 ⫾ 0.05 N/A N/A 7.47 ⫾ 0.06 7.44 ⫾ 0.09 7.35 ⫾ 0.08
RW sham 7.45 ⫾ 0.09 N/A N/A 7.39 ⫾ 0.16 7.41 ⫾ 0.16 7.32 ⫾ 0.10
PaCO2, Torr HT arrest 37 ⫾ 6 40 ⫾ 5 93 ⫾ 11 40 ⫾ 11 39 ⫾ 5 38 ⫾ 5
RW arrest 36 ⫾ 7 45 ⫾ 7 82 ⫾ 14 37 ⫾ 12 32 ⫾ 6 42 ⫾ 3
HT sham 36 ⫾ 6 N/A N/A 35 ⫾ 4 35 ⫾ 9 40 ⫾ 4
RW sham 36 ⫾ 3 N/A N/A 37 ⫾ 3 33 ⫾ 3 43 ⫾ 5
PaO2, Torr HT arrest 170 ⫾ 38 30 ⫾ 4 21 ⫾ 6 126 ⫾ 22 180 ⫾ 57 160 ⫾ 45
RW arrest 133 ⫾ 27 24 ⫾ 5 22 ⫾ 19 134 ⫾ 41 160 ⫾ 53 155 ⫾ 34
HT sham 166 ⫾ 23 N/A N/A 138 ⫾ 24 131 ⫾ 30 144 ⫾ 25
RW sham 151 ⫾ 32 N/A N/A 142 ⫾ 18 152 ⫾ 24 156 ⫾ 18
Hb, g/dl HT arrest 6.8 ⫾ 0.5 8.7 ⫾ 1.6 8.5 ⫾ 1.8 7.6 ⫾ 1.7 9.2 ⫾ 2.1 7.8 ⫾ 1.4
RW arrest 7.0 ⫾ 0.6 8.5 ⫾ 1.6 8.0 ⫾ 2.3 7.7 ⫾ 1.0 8.6 ⫾ 1.6 8.2 ⫾ 0.7
HT sham 6.7 ⫾ 1.2 N/A N/A 8.0 ⫾ 1.3 8.4 ⫾ 1.4 7.9 ⫾ 1.1
RW sham 6.4 ⫾ 2.5 N/A N/A 6.0 ⫾ 1.9 6.2 ⫾ 2.3 6.7 ⫾ 1.8
Body temperature, °C HT arrest 38.8 ⫾ 0.7 38.9 ⫾ 0.4 39.0 ⫾ 0.4‡ 38.6 ⫾ 0.5 34.1 ⫾ 0.3 33.9 ⫾ 0.4
RW arrest 38.3 ⫾ 0.6 38.6 ⫾ 0.1 38.6 ⫾ 0.3 38.6 ⫾ 0.5 33.9 ⫾ 0.3 38.8 ⫾ 0.7
HT sham 37.9 ⫾ 1.0 N/A N/A 38.3 ⫾ 1.4 34.1 ⫾ 0.2 34.1 ⫾ 0.3
RW sham 38.4 ⫾ 0.3 N/A N/A 38.6 ⫾ 0.6 34.0 ⫾ 0.3 38.7 ⫾ 0.4
Hb, hemoglobin; HT, hypothermic; RW, rewarmed; N/A, not applicable; PaCO2, arterial partial pressure of carbon dioxide; PaO2, arterial partial pressure of
oxygen. *Temperature-corrected blood gases are presented during hypothermia; †data are presented at completion of the HT or RW periods and before changing
blood pressure; ‡P ⫽ 0.026 from the RW arrest group by t-test at the time point of asphyxia 5 min.

and acidosis during asphyxia was similar between groups and
largely recovered 1–3 h after ROSC (Table 2; P ⬎ 0.05). Partial
pressure of O2 (PaO2) values recovered to 126 –180 Torr between
1 h and 3 h after ROSC in all groups, and arterial hemoglobin O2
saturation and PaO2 levels did not differ among groups at each
time point (P ⬎ 0.05). All piglets that underwent HA injury
required sodium bicarbonate to correct metabolic acidosis. Base-
line hemoglobin levels ranged from 6.4 ⫾ 2.5 g/dl in the re-
warmed shams to 7.0 ⫾ 0.6 g/dl in the rewarmed arrest group.
Hemoglobin levels were similar among groups at each time point
(P ⬎ 0.05). There was a difference in sodium levels across time
and treatment groups (P ⫽ 0.022). Sodium decreased from 147 ⫾
6 meq/l at baseline to 139 ⫾ 4 meq/l on the 1st day after
resuscitation across the entire cohort. However, there were no
differences in sodium across groups at each time point (P ⬎ 0.05).
Ionized calcium levels were not different among groups (data not
shown).
Cerebral autoregulation during hypotension. Examples of
the relationship of LDF with CPP during induced arterial
hypotension are shown in Fig. 2. Two piglets from each
treatment group were selected as examples to illustrate the
range of slopes in the autoregulatory plateau region above the
calculated LLA. Two-way ANOVA on the LLA indicated no
significant overall effect of arrest (P ⫽ 0.601). Whereas there
was a trend of higher LLA in the rewarmed groups compared
with groups that remained hypothermic (P ⫽ 0.084 for the
effect of temperature), there was no interaction between arrest
and temperature treatment (P ⫽ 0.727; n ⫽ 6). The LLAs were
29 ⫾ 10 mmHg (95% CI: 19, 40) for arrest ⫹ hypothermia;
38 ⫾ 11 mmHg (26, 49) for arrest ⫹ rewarming; 29 ⫾ 8
mmHg (20, 38) for sham ⫹ hypothermia; and 34 ⫾ 7 mmHg
(27, 42) for sham ⫹ rewarming. The LLA was similar among
piglets that received phenylephrine (28 ⫾ 8 mmHg; n ⫽ 6) and
those that did not (34 ⫾ 9 mmHg; n ⫽ 18; P ⫽ 0.188 by t-test).
LDF, HVx, and COx were not different among treatment
groups at each level of CPP (Fig. 3, A–C; P ⬎ 0.05). Because
piglets had different LLAs, and individual differences could
prevent detection of differences among groups, we reanalyzed
data as a function of CPP–LLA. Even after this additional
analysis, no differences in LDF, HVx, or COx were observed
among groups (Fig. 3, D–F; P ⬎ 0.05).
Absolute values of rSO2 decreased with progressive hypo-
tension (Fig. 4, A and B). When piglets from all groups were
combined, there was good correlation between the LLA cal-
culated from the LDF plot and the inflection point in rSO2 vs.
CPP when blood pressure decreased (r ⫽ 0.727, 95% CI:
0.429 – 0.882; P ⫽ 0.0001; n ⫽ 22). An inflection point could
not be identified in two sham piglets because the slope of rSO2
vs. CPP did not change as blood pressure decreased. Two-way
ANOVA on the rSO2-derived inflection point indicated no
effect of arrest (P ⫽ 0.832), temperature (P ⫽ 0.308), or
interaction between arrest and temperature (P ⫽ 0.771).
As CPP decreased, both HVx and COx increased (Fig. 3).
Spline regression analysis was used to evaluate the slope of
the relationships of HVx and COx with CPP above and below
the LLA on a paired basis. Below the LLA, the slope of the
relationship with CPP was significantly less than zero for HVx
in hypothermic arrested piglets (P ⬍ 0.001; Fig. 5A), for HVx
in rewarmed arrested piglets (P ⫽ 0.045; Fig. 5C), and for
COx in rewarmed arrested piglets (P ⫽ 0.016; Fig. 6C). In the
hypothermic arrest group, the relationship for HVx was bipha-
sic, with the slope below the LLA significantly more negative
than the slope above the LLA (P ⬍ 0.001; Table 3 and Fig.
5A). In the rewarmed arrest group, the relationship for COx
was also biphasic, with the slope below the LLA significantly
more negative than the slope above the LLA (P ⫽ 0.047; Table
3 and Fig. 6C). Because baseline levels of CPP (Table 1) were
only 10 –20 mmHg above the LLA, the range of CPP above
the LLA may have been too narrow to determine accurately the
slope of the relationships above the LLA and thereby, detect a
biphasic relationship in every group. Alternatively, ROC
curves can test the sensitivity and specificity of HVx and COx
in predicting the LDF-derived LLA. The areas under the ROC
curves were 0.81– 0.96 for all groups and demonstrated that

J Appl Physiol • doi:10.1152/japplphysiol.00238.2013 • www.jappl.org

 Cerebral Autoregulation, Hypothermia, and Asphyxia
both HVx and COx predicted whether CPP was above or below
the LLA (Table 4).
Cerebral autoregulation during hypertension. Values of
ICP, CPP, and blood chemistry, on the day of assessing
cerebral autoregulation, were in the normal physiological range
and not different among the four hypertensive cohorts (data not
shown). With cephalic hypertension, CPP increased to 99 ⫾ 22
mmHg in the hypothermic arrest, 102 ⫾ 26 mmHg in the
hypothermic sham, 126 ⫾ 26 mmHg in the rewarmed arrest,
and 118 ⫾ 22 mmHg in the rewarmed sham groups (P ⫽ 0.18).
The median slopes of the change in LDF with respect to the
change in CPP were 0.03 (⫺0.09, 0.47; IQR) for hypothermic
arrest, 0.28 (0.15, 0.31) for hypothermic sham, 0.04 (0.03,
0.08) for rewarmed arrest, and ⫺0.06 (⫺0.17, 0.05) for re-
J Appl Physiol • doi:10.1152/japplphysiol.00238.2013 • www.jappl.org
• Larson AC et al. 1437
Fig. 2. Two representative piglets from each
experimental group were selected to illus-
trate variation among individual animals in
the laser-Doppler flow (LDF) responses to
hypotension. From each experimental group,
1 piglet was selected with a relatively hori-
zontal, autoregulatory plateau. As a contrast,
another piglet from the same group was
selected with a LDF curve that was not
horizontal when cerebral perfusion pressure
(CPP) exceeded the lower limit of autoreg-
ulation (LLA). LDF measurements were av-
eraged into 5 mmHg bins of CPP. The cal-
culated LLA is demarcated with a dotted
line. When CPP exceeded the LLA, some
piglets had autoregulatory plateaus that were
relatively horizontal (left column), whereas
others had LDF curves that were not hori-
zontal (right column). A and B: sham, hypo-
thermic piglets; C and D: postarrest, hypo-
thermic piglets; E and F: sham, rewarmed
piglets; G and H: postarrest, rewarmed pig-
lets.
warmed sham groups (P ⫽ 0.162). Only hypothermic shams
had a significant increase in LDF as CPP increased (P ⫽
0.015). The slopes did not differ from zero in the rewarmed
sham, hypothermic arrest, and rewarmed arrest groups (P ⬎
0.10 for all groups). Furthermore, the slopes were significantly
⬍1.0 (P ⬍ 0.05), thereby indicating that blood flow was not
completely passive to CPP (Fig. 7).
DISCUSSION
This study demonstrated several new findings. First, cere-
brovascular autoregulation to hypotension or hypertension was
not impaired consistently after HA arrest and 2 h of normo-
thermic recovery followed by 20 h of hypothermia in neonatal
1438 Cerebral Autoregulation, Hypothermia, and Asphyxia • Larson AC et al.

Fig. 3. A: during hypotension, LDF was similar

among treatment groups at each level of CPP
(treatment effect, P ⫽ 0.96; n ⫽ 6). B: the
hemoglobin volume index (HVx) became more
positive with hypotension, indicating impaired
vasoreactivity. HVx was similar across groups
at each CPP (P ⬎ 0.05; n ⫽ 6). C: the cerebral
oximetry index (COx) became more positive
with hypotension, indicating impaired cerebral
autoregulation. COx was similar among groups
at each CPP (P ⬎ 0.10; n ⫽ 6). To compare
CPP relative to each animal’s LLA, CPP was
centered at 0 on the x-axis. D: with hypoten-
sion, LDF was similar among groups at each
CPP (P ⫽ 0.51; n ⫽ 6). E: HVx became more
positive during hypotension as the LLA was
crossed, indicating impaired vasoreactivity. HVx
was similar among groups at each CPP (P ⬎
0.05; n ⫽ 6). F: COx became more positive
during hypotension as the LLA was crossed, in-
dicating impaired cerebral autoregulation. COx
was similar among groups at each CPP (P ⬎
0.05; n ⫽ 6). Hypothermic sham (open triangles),
hypothermic arrest (solid triangles), rewarmed
sham (open circles), rewarmed arrest (solid cir-
cles). Data are shown as means ⫾ SD.

piglets. Second, rapid rewarming did not shift the LLA signif-
icantly during hypotension. Third, cerebral autoregulation dur-
ing hypertension remained functional after HA injury with
hypothermia and rewarming. Fourth, the NIRS-derived indices
identified levels of CPP that were below the LLA. These
indices may provide noninvasive methods to assess continu-
ously cerebral autoregulation under rapidly changing temper-
ature conditions after pediatric hypoxic brain injury.
This piglet model of whole-body HA produces a pattern of
brain injury (20) similar to those observed in neonatal HIE
(21), including robust cortical injury with a 50% decrease in
viable neurons in the somatosensory cortex (18). We measured

Fig. 4. Regional cerebral oxyhemoglobin saturation (rSO2) measurements were averaged into 5 mmHg bins of CPP. Absolute rSO2 decreased with progressive
hypotension in (A) sham piglets that remained hypothermic (n ⫽ 6) or were rewarmed (n ⫽ 5) and (B) postarrest piglets that remained hypothermic (n ⫽ 6) or
were rewarmed (n ⫽ 6). Hypothermic piglets (open squares); rewarmed piglets (solid circles). Data are shown as means ⫾ SD. C: each piglet’s lower LLA, which
was calculated from the LDF vs. CPP curve, was compared with the inflection point calculated from the rSO2 vs. CPP curve during hypotension. There was good
correlation between the LDF-derived LLA and the rSO2 inflection point. Postarrest rewarmed piglets (circles; n ⫽ 6); sham rewarmed piglets (squares; n ⫽ 5);
postarrest hypothermia piglets (triangles; n ⫽ 6); sham hypothermia piglets (inverted triangles; n ⫽ 5). For the 22 piglets combined: r ⫽ 0.727; 95% confidence
intervals (CI) 0.429 – 0.881; P ⫽ 0.0001. The dotted line demarcates a perfect correlation of r ⫽ 1.0.

J Appl Physiol • doi:10.1152/japplphysiol.00238.2013 • www.jappl.org

 Cerebral Autoregulation, Hypothermia, and Asphyxia • Larson AC et al. 1439

Fig. 5. Spline regression model fit lines

(⫾95% CI) of the HVx index. Each piglet’s
lower LLA is centered at 0 on the x-axis to
compare each animal’s CPP relative to its
LLA. HVx became more positive as CPP
decreased below the LLA in the (A) hypo-
thermic arrest, (B) hypothermic sham, (C)
rewarmed arrest, and (D) rewarmed sham
groups. The slope of the relationship with
CPP was significantly negative in the hypo-
thermic arrest (*P ⬍ 0.001) and rewarmed
arrest (**P ⫽ 0.045) groups. In the hypo-
thermic arrest group, the relationship for
HVx was biphasic, with the slope below the
LLA significantly more negative than the
slope above the LLA (P ⬍ 0.001).

red blood cell flux with the LDF probe positioned over the
vulnerable primary somatosensory cortex where perturbations
in CBF might occur.
We reported that 4 h of hypothermia after HA injury de-
creased the LLA during hypotension, measured 6 h after
resuscitation in neonatal swine (16). During hypothermic car-
diopulmonary bypass, rewarming impaired cerebral autoregu-
lation (14). Proinflammatory cytokine shifts during rewarming
(3) may impact vasoreactivity. Thus we theorized that rapid
rewarming from prolonged hypothermia after HA injury would
affect cerebral autoregulation. However, we did not find a
substantial shift in the LLA during hypotension or differences
in vasomotor function during hypertension between postarrest
hypothermic and rewarmed cohorts, despite a rewarming rate
(4°C/h) that is faster than that typically used in neonates
(ⱕ0.5°C/h). We reasoned that any effect of rewarming on
cerebral autoregulation would be more apparent with fast
rewarming.

A Hypothermic Arrest B Hypothermic Sham

1 1

.5 .5
COx
COx

0 0
Fig. 6. Spline regression model fit lines (⫾95%
CI) of the COx index. Each piglet’s lower LLA
-.5 -.5
is centered at 0 on the x-axis to compare each
animal’s CPP relative to its LLA. COx became
-1 -1 more positive as CPP decreased below the LLA
-30 -20 -10 0 10 20 30 -30 -20 -10 0 10 20 30 in the (A) hypothermic arrest, (B) hypothermic
CPP - LLA (mmHg) CPP - LLA (mmHg) sham, (C) rewarmed arrest, and (D) rewarmed
C Rewarmed Arrest
D Rewarmed Sham
sham groups. The slope of the relationship with
CPP was significantly negative in the rewarmed
1 1
arrest group (*P ⫽ 0.016). In the rewarmed
* arrest group, the relationship for COx was bi-
.5 .5 phasic, with the slope below the LLA signifi-
COx
COx

cantly more negative than the slope above the

LLA (P ⫽ 0.047).
0 0

-.5 -.5

-1 -1
-30 -20 -10 0 10 20 30 -30 -20 -10 0 10 20 30
CPP - LLA (mmHg) CPP - LLA (mmHg)

J Appl Physiol • doi:10.1152/japplphysiol.00238.2013 • www.jappl.org

1440 Cerebral Autoregulation, Hypothermia, and Asphyxia
Table 3. Relationships between Hb volume index (HVx) and
cerebral perfusion pressure (CPP) and between cerebral
oximetry index (COx) and CPP during induced hypotension
(Slope above LLA) ⫺ (Slope below LLA)
Group Difference (95% CI)
HT sham
HVx 0.021 (⫺0.010, 0.052)
COx ⫺0.004 (⫺0.057, 0.049)
RW sham
HVx 0.020 (⫺0.005, 0.046)
COx ⫺0.005 (⫺0.032, 0.022)
HT arrest
HVx 0.035 (0.021, 0.049)
COx 0.015 (⫺0.007, 0.037)
RW arrest
HVx 0.019 (⫺0.014, 0.052)
COx 0.021 (0.000, 0.041)
Relationships were determined by spline regression analysis. Differences in
the slopes of the relationships of HVx and COx with CPP below the lower limit
of autoregulation (LLA) from the slopes above the LLA are shown. CI,
confidence intervals.
The cerebrovascular autoregulation curve is traditionally
illustrated with a horizontal CBF plateau at levels of CPP with
pressure-reactive CBF, a sharp cut-off at the LLA, and a
decline in CBF when CPP is below the LLA. However, when
cerebral autoregulation is functional, the plateau may not have
a slope that is equal to zero (19). Indeed, studies in animals
show a high degree of variability in the slope (13). In individ-
ual animals, the plateau may not be horizontal, because
changes in cerebrovascular resistance in response to changes in
CPP do not maintain perfectly constant CBF. Even when CPP
crosses below the LLA, arteriolar dilation may continue but no
longer be sufficient to maintain constant CBF. Therefore, the
CBF– blood pressure autoregulation curve in individual ani-
mals will have an inflection at the LLA even when a slope in
the autoregulatory range does not equal zero (Fig. 2).
The NIRS-derived indices distinguished between levels of
CPP that were associated with pressure-passive and pressure-
reactive LDF. COx is derived from cerebral oxyhemoglobin
saturation, which is affected by metabolic rate. Hypothermia
decreases O2 consumption (33). Despite changing temperature,
COx accurately detected levels of CPP that were below the
LLA during hypotension, presumably because O2 consumption
remained relatively stable during the 5-min sampling periods
for COx calculations. HVx, which is based on measurements of
deoxygenated and oxygenated tissue hemoglobin and should
be less affected by O2 consumption, also indicated whether
CPP was above or below the LLA. This suggests that both
indices differentiate pressure-reactive from pressure-passive
CBF in superficial cortex.
The limits of CPP that maintain pressure-reactive CBF are
age dependent. Neonatal swine have lower LLAs than juvenile
swine (32). In humans, the LLA is also lower in children (7)
than in adults (15). Intracranial hypertension raises the LLA
(4), which is relevant to clinical situations of evolving cerebral
edema. Therefore, it is critical to identify methods that detect
the limits of cerebral autoregulation and that can be used in
clinical practice.
Measurements of rSO2 from NIRS decreased in all experi-
mental groups during induced hypotension after the completion
J Appl Physiol • doi:10.1152/japplphysiol.00238.2013 • www.jappl.org
• Larson AC et al.
of rewarming or during the continuation of hypothermia. O2
consumption presumably remained relatively constant because
the anesthetic regimen remained unchanged, and the piglets
had stable temperatures during induced hypotension. Arterial
hemoglobin concentration and O2 saturation—two determi-
P nants of O2 supply—also remained constant. Therefore, the
decrease in regional cerebral measurements of oxygenated
0.182 hemoglobin with a concomitant increase in deoxygenated he-
0.879 moglobin concentration was likely secondary to declining CBF
during hypotension when CPP was below the LLA. The
0.119
0.722 correlation between the LDF-derived LLA and the inflection
point when rSO2 was plotted against CPP may have limited
⬍0.001 clinical applicability in defining the CBF LLA. In most clinical
0.172 situations, a patient’s balance of O2 supply and demand may
0.259
change, which would affect the rSO2 measurements and limit
0.047 the ability to use rSO2 to estimate the LLA of CBF.
Piglets resuscitated from HA arrest exhibited pressure-reac-
tive LDF with hypertension during hypothermia and after
rewarming. We reported that cerebral autoregulation is func-
tional during hypertension in piglets after HA arrest and 6 h of
recovery with hypothermia or normothermia (16) or 2 days of
recovery with normothermia (18). With acutely induced hy-
pertension, cardiogenic failure occurred before the cerebral
myogenic vasoconstrictive capacity was exceeded. Thus we
were unable to detect the upper limit of autoregulation. In
unanesthetized neonatal piglets, pressure-reactive CBF was
reported at mean arterial blood pressures between 50 mmHg
and 90 mmHg. Above 90 mmHg (to 120 mmHg), pressure-
passive CBF was observed, which was attributed to increased
prostanoid production (9). In our study of anesthetized piglets,
LDF remained relatively constant at CPP 90 –120 mmHg. It is
possible that anesthesia decreases vasodilator prostanoid pro-
duction or extends the upper limit of autoregulation by increas-
ing vascular myogenic tone. To explore the upper limit of
autoregulation in our model, cardiopulmonary bypass may be
required. Alternatively, juvenile piglets might better tolerate
induced hypertension and serve as an alternative model for
pediatric cardiac arrest that would enable better exploration of
the CBF response to hypertension. Inducing cephalic hyper-
tension with an aortic balloon may cause lower-body ischemia
and metabolic acidosis. Acidosis promotes vasodilation (8),
which could impair the vascular myogenic-constrictive re-
sponse to hypertension. In this study, tissue ischemia during
aortic balloon inflation did not impair cerebral autoregulation.
Table 4. Results of the receiver operator characteristic curves
and area under the curves (AUC) for the HVx and COx indices
in detecting the LLA during induced hypotension
No. Observations, across
Group 6 Piglets/Group AUC
HT sham
HVx 43 0.86
COx 39 0.88
RW sham
HVx 53 0.81
COx 45 0.87
HT arrest
HVx 47 0.88
COx 47 0.90
RW arrest
HVx 41 0.96
COx 41 0.85
 Cerebral Autoregulation, Hypothermia, and Asphyxia
Some piglets required phenylephrine to maintain CPP above
the LLA during anesthesia, and dopamine was administered to
support cardiac contractility during aortic balloon catheter
inflation. ␣-1 Agonism from phenylephrine may directly affect
cerebrovascular resistance and affect CBF autoregulation (23).
Moreoever, phenylephrine may have gender-specific effects on
cerebral autoregulation (2). Only males were used in this study,
and a statistical difference in the LLA was not detected
between those that did or did not receive phenylephrine. Thus
the use of vasopressor agents in our model did not appear to
represent a major confounding influence. Dopamine resulted in
pressure-passive LDF in a swine model of hypovolemia (12)
but was not associated with pressure-passive LDF when com-
bined with aortic balloon inflation in our model of HA injury
(16, 18). All cohorts were ventilated with 30% O2, which
reflects clinical practice and only marginally increases arterial
O2 content.
In our model of HA injury, piglets develop moderate intra-
cranial hypertension and can regain consciousness after resus-
citation (18). Patients with hypoxic brain injuries may be
comatose with high ICP, and intracranial hypertension inde-
pendently affects the LLA (4). Rewarming from hypothermia
in the context of severe intracranial hypertension might affect
cerebral autoregulation, although we were unable to test that
possibility in this study. We did not extend the duration of
hypoxia or asphyxia because cardiac resuscitation becomes
difficult.
This study has limitations. First, we did not include normo-
thermic sham or normothermic HA injury groups, because our
aim was to compare the effects of rewarming with the effects
of hypothermia on cerebral autoregulation. Neuroprotection
from hypothermia (1) may preserve mediators of vasodilation
and vascular tone despite rewarming. Furthermore, the mean
LLAs of 34 mmHg in the rewarmed shams and 38 mmHg in
the rewarmed arrest group are similar to the values of 41
mmHg in shams and 37– 40 mmHg in arrest groups that
recovered for 1–2 days without anesthesia and without hypo-
thermia and rewarming (18). Thus rapid rewarming did not
markedly increase the LLA above that expected without pro-
longed hypothermia and rewarming. Second, our findings of
intact cerebral autoregulation might not be generalizable to
clinical situations where brain injuries can be more severe. The
J Appl Physiol • doi:10.1152/japplphysiol.00238.2013 • www.jappl.org
• Larson AC et al. 1441
Fig. 7. A: during hypertension, LDF remained
relatively constant in postarrest and sham-
operated piglets after hypothermia or hypother-
mia with rewarming (n ⫽ 6). Hypothermic sham
(open triangles); hypothermic arrest (solid trian-
gles); rewarmed sham (open circles); rewarmed
arrest (solid circles). A dashed line demarcates
LDF at 100% of baseline. Data are shown as
means ⫾ SD. B: the autoregulatory index was
calculated as the change in LDF (as a percentage
of baseline) compared with the change in CPP
(as a percentage of baseline; ⌬LDF%/⌬CPP%).
The ⌬LDF%/⌬CPP% was ⬎0 in the hypother-
mic shams (slope 0.24; 95% CI 0.07, 0.42; *P ⫽
0.015). The slopes did not differ from 0 in the
rewarmed sham, hypothermic arrest, and re-
warmed arrest groups. There were no differ-
ences in ⌬LDF%/⌬CPP% among groups (P ⫽
0.162). Data are displayed as means ⫾ SE.
anesthesia provided is not commonly used in intensive care
units. Nonetheless, all animals received the same anesthetic,
permitting comparisons among groups. Histopathologic exam-
ination was not performed, because severe hypotension or
hypertension would distort the pathology. Third, the sample
sizes were small. Based on a previous experiment with acute
hypothermia and HA injury (16) and post hoc power analysis,
a sample size of six for this study was estimated to detect an
11-mmHg change in LLA at 80% power. We selected a shift in
the LLA of ⬎10 mmHg as a threshold that would be clinically
relevant. Whereas there was a trend of increased LLA with
rewarming compared with the remaining hypothermic (P ⫽
0.084), this effect was not specific to conditions of the HA
injury, as there was no interaction between arrest and temper-
ature treatment (P ⫽ 0.727). Moreover, the active autoregula-
tory response to hypertension was consistent with our previous
studies that examined the response, 6 h (16) and 2 days (18)
after HA injury.
In summary, rapid rewarming from hypothermia did not
shift the LLA during hypotension after HA brain injury in a
neonatal swine model with moderate intracranial hypertension.
COx and HVx accurately differentiated CPP below and above
the LLA. These monitors may serve as the basis for noninva-
sive metrics of cerebral autoregulation during therapeutic hy-
pothermia after hypoxic brain injury.
ACKNOWLEDGMENTS
We thank Claire Levine for her editorial assistance.
GRANTS
Support for this work was provided, in part, by a Scientist Development
Grant (to J. K. Lee) and a Postdoctoral Fellowship Award (to B. Wang) from
The American Heart Association (Dallas, TX), The International Anesthesia
Research Society (San Francisco, CA; to J. K. Lee), and a grant from the
National Institute of Neurological Disorders and Stroke NS-060703 (Bethesda,
MD; to R. C. Koehler).
DISCLOSURES
J. K. Lee received a research grant from Covidien for a clinical research
study.
AUTHOR CONTRIBUTIONS
Author contributions: D.H.S., R.C.K., and J.K.L. conception and design of
research; A.C.L., J.L.J., E.K., B.W., Z-J.Y., and J.K.L. performed experiments;
1442 Cerebral Autoregulation, Hypothermia, and Asphyxia
A.C.L., J.L.J., R.C.K., and J.K.L. analyzed data; A.C.L., J.L.J., D.H.S., R.C.K.,
and J.K.L. interpreted results of experiments; J.K.L. prepared figures; J.K.L.
drafted manuscript; A.C.L., J.L.J., R.C.K., and J.K.L. edited and revised
manuscript; A.C.L., J.L.J., E.K., B.W., Z-J.Y., D.H.S., R.C.K., and J.K.L.
approved final version of manuscript.
REFERENCES
1. Agnew DM, Koehler RC, Guerguerian AM, Shaffner DH, Traystman
RJ, Martin LJ, Ichord RN. Hypothermia for 24 hours after asphyxic
cardiac arrest in piglets provides striatal neuroprotection that is sustained
10 days after rewarming. Pediatr Res 54: 253–262, 2003.
2. Armstead WM, Kiessling JW, Kofke WA, Vavilala MS. Impaired
cerebral blood flow autoregulation during posttraumatic arterial hypoten-
sion after fluid percussion brain injury is prevented by phenylephrine in
female but exacerbated in male piglets by extracellular signal-related
kinase mitogen-activated protein kinase upregulation. Crit Care Med 38:
1868 –1874, 2010.
3. Bisschops LL, Hoedemaekers CW, Mollnes TE, van der Hoeven JG.
Rewarming after hypothermia after cardiac arrest shifts the inflammatory
balance. Crit Care Med 40: 1136 –1142, 2012.
4. Brady KM, Lee JK, Kibler KK, Easley RB, Koehler RC, Czosnyka M,
Smielewski P, Shaffner DH. The lower limit of cerebral blood flow
autoregulation is increased with elevated intracranial pressure. Anesth
Analg 108: 1278 –1283, 2009.
5. Brady KM, Lee JK, Kibler KK, Easley RB, Koehler RC, Shaffner DH.
Continuous measurement of autoregulation by spontaneous fluctuations in
cerebral perfusion pressure: comparison of 3 methods. Stroke 39: 2531–
2537, 2008.
6. Brady KM, Lee JK, Kibler KK, Smielewski P, Czosnyka M, Easley
RB, Koehler RC, Shaffner DH. Continuous time-domain analysis of
cerebrovascular autoregulation using near-infrared spectroscopy. Stroke
38: 2818 –2825, 2007.
7. Brady KM, Mytar JO, Lee JK, Cameron DE, Vricella LA, Thompson
WR, Hogue CW, Easley RB. Monitoring cerebral blood flow pressure
autoregulation in pediatric patients during cardiac surgery. Stroke 41:
1957–1962, 2010.
8. Celotto AC, Restini CB, Capellini VK, Bendhack LM, Evora PR.
Acidosis induces relaxation mediated by nitric oxide and potassium
channels in rat thoracic aorta. Eur J Pharmacol 656: 88 –93, 2011.
9. Chemtob S, Beharry K, Rex J, Varma DR, Aranda JV. Changes in
cerebrovascular prostaglandins and thromboxane as a function of systemic
blood pressure. Cerebral blood flow autoregulation of the newborn. Circ
Res 67: 674 –682, 1990.
10. Fujita M, Wei EP, Povlishock JT. Effects of hypothermia on cerebral
autoregulatory vascular responses in two rodent models of traumatic brain
injury. J Neurotrauma 29: 1491–1498, 2012.
11. Graham EM, Ruis KA, Hartman AL, Northington FJ, Fox HE. A
systematic review of the role of intrapartum hypoxia-ischemia in the
causation of neonatal encephalopathy. Am J Obstet Gynecol 199: 587–595,
2008.
12. Hahn GH, Heiring C, Pryds O, Greisen G. Cerebral vascular effects of
hypovolemia and dopamine infusions: a study in newborn piglets. Acta
Paediatr 101: 736 –742, 2012.
13. Jones SC, Radinsky CR, Furlan AJ, Chyatte D, Qu Y, Easley KA,
Perez-Trepichio AD. Variability in the magnitude of the cerebral blood
flow response and the shape of the cerebral blood flow-pressure autoreg-
ulation curve during hypotension in normal rats [corrected]. Anesthesiol-
ogy 97: 488 –496, 2002.
14. Joshi B, Brady K, Lee J, Easley B, Panigrahi R, Smielewski P,
Czosnyka M, Hogue CW Jr. Impaired autoregulation of cerebral blood
flow during rewarming from hypothermic cardiopulmonary bypass and its
potential association with stroke. Anesth Analg 110: 321–328, 2010.
15. Joshi B, Ono M, Brown C, Brady K, Easley RB, Yenokyan G,
Gottesman RF, Hogue CW. Predicting the limits of cerebral autoregu-
lation during cardiopulmonary bypass. Anesth Analg 114: 503–510, 2012.
16. Lee JK, Brady KM, Mytar JO, Kibler KK, Carter EL, Hirsch KG,
Hogue CW, Easley RB, Jordan LC, Smielewski P, Czosnyka M,
J Appl Physiol • doi:10.1152/japplphysiol.00238.2013 • www.jappl.org
• Larson AC et al.
Shaffner DH, Koehler RC. Cerebral blood flow and cerebrovascular
autoregulation in a swine model of pediatric cardiac arrest and hypother-
mia. Crit Care Med 39: 2337–2345, 2011.
17. Lee JK, Kibler KK, Benni PB, Easley RB, Czosnyka M, Smielewski P,
Koehler RC, Shaffner DH, Brady KM. Cerebrovascular reactivity
measured by near-infrared spectroscopy. Stroke 40: 1820 –1826, 2009.
18. Lee JK, Yang ZJ, Wang B, Larson AC, Jamrogowicz JL, Kulikowicz
E, Kibler KK, Mytar JO, Carter EL, Burman HT, Brady KM,
Smielewski P, Czosnyka M, Koehler RC, Shaffner DH. Noninvasive
autoregulation monitoring in a swine model of pediatric cardiac arrest.
Anesth Analg 114: 825–836, 2012.
19. Lucas SJ, Tzeng YC, Galvin SD, Thomas KN, Ogoh S, Ainslie PN.
Influence of changes in blood pressure on cerebral perfusion and oxygen-
ation. Hypertension 55: 698 –705, 2010.
20. Martin LJ, Brambrink A, Koehler RC, Traystman RJ. Primary sen-
sory and forebrain motor systems in the newborn brain are preferentially
damaged by hypoxia-ischemia. J Comp Neurol 377: 262–285, 1997.
21. Miller SP, Ramaswamy V, Michelson D, Barkovich AJ, Holshouser B,
Wycliffe N, Glidden DV, Deming D, Partridge JC, Wu YW, Ashwal S,
Ferriero DM. Patterns of brain injury in term neonatal encephalopathy. J
Pediatr 146: 453–460, 2005.
22. Mueller-Burke D, Koehler RC, Martin LJ. Rapid NMDA receptor
phosphorylation and oxidative stress precede striatal neurodegeneration
after hypoxic ischemia in newborn piglets and are attenuated with hypo-
thermia. Int J Dev Neurosci 26: 67–76, 2008.
23. Ogoh S, Brothers RM, Eubank WL, Raven PB. Autonomic neural
control of the cerebral vasculature: acute hypotension. Stroke 39: 1979 –
1987, 2008.
24. Pietrini D, Piastra M, Luca E, Mancino A, Conti G, Cavaliere F, De
Luca D. Neuroprotection and hypothermia in infants and children. Curr
Drug Targets 13: 925–935, 2012.
25. Pryds O, Greisen G, Lou H, Friis-Hansen B. Vasoparalysis associated
with brain damage in asphyxiated term infants. J Pediatr 117: 119 –125,
1990.
26. Shah PS. Hypothermia: a systematic review and meta-analysis of clinical
trials. Semin Fetal Neonatal Med 15: 238 –246, 2010.
27. Shankaran S, Laptook AR, McDonald SA, Higgins RD, Tyson JE,
Ehrenkranz RA, Das A, Sant’Anna G, Goldberg RN, Bara R, Walsh
MC, Eunice Kennedy Shriver National Institute of Child Health and
Human Development Neonatal Research Network. Temperature profile
and outcomes of neonates undergoing whole body hypothermia for neo-
natal hypoxic-ischemic encephalopathy. Pediatr Crit Care Med 13: 53–
59, 2012.
28. Shankaran S, Pappas A, McDonald SA, Vohr BR, Hintz SR, Yolton K,
Gustafson KE, Leach TM, Green C, Bara R, Petrie Huitema CM,
Ehrenkranz RA, Tyson JE, Das A, Hammond J, Peralta-Carcelen
M, Evans PW, Heyne RJ, Wilson-Costello DE, Vaucher YE, Bauer
CR, Dusick AM, Adams-Chapman I, Goldstein RF, Guillet R, Papile
LA, Higgins RD, Eunice Kennedy Shriver NICHD Neonatal Research
Network. Childhood outcomes after hypothermia for neonatal encepha-
lopathy. N Engl J Med 366: 2085–2092, 2012.
29. Smielewski P, Czosnyka M, Steiner L, Belestri M, Piechnik S, Pickard
JD. ICM⫹: software for on-line analysis of bedside monitoring data after
severe head trauma. Acta Neurochir Suppl 95: 43–49, 2005.
30. Tress EE, Kochanek PM, Saladino RA, Manole MD. Cardiac arrest in
children. J Emerg Trauma Shock 3: 267–272, 2010.
31. van Erkel AR, Pattynama PM. Receiver operating characteristic (ROC)
analysis: basic principles and applications in radiology. Eur J Radiol 27:
88 –94, 1998.
32. Walter B, Aisenpreis K, Fritz HG, Soukup J, Eiselt M, Bauer R.
Age-dependent effects of gradual decreases in cerebral perfusion pressure
on the neurochemical response in swine. Intensive Care Med 36: 1067–
1075, 2010.
33. Zanelli S, Buck M, Fairchild K. Physiologic and pharmacologic consid-
erations for hypothermia therapy in neonates. J Perinatol 31: 377–386,
2011.